Tyrosine-Phosphorylation-Dependent Translocation of the SLAT Protein to the Immunological Synapse Is Required for NFAT Transcription Factor Activation

SUMMARY

SWAP-70-like adaptor of T cells (SLAT) is a guanine nucleotide exchange factor for Rho GTPases that regulates the development of T helper 1 (Th1) and Th2 cell inflammatory responses by controlling the Ca²⁺-NFAT signaling pathway. However, the mechanism used by SLAT to regulate these events is unknown. Here, we report that the T cell receptor (TCR)-induced translocation of SLAT to the immunological synapse required Lck-mediated phosphorylation of two tyrosine residues located in an immunoreceptor tyrosine-based activation motif-like sequence but was independent of the SLAT PH domain. This subcellular relocalization was coupled to, and necessary for, activation of the NFAT pathway. Furthermore, membrane targeting of the SLAT Dbl-homology (catalytic) domain was sufficient to trigger TCR-mediated NFAT activation and Th1 and Th2 differentiation in a Cdc42-dependent manner. Therefore, tyrosine-phosphorylation-mediated relocalization of SLAT to the site of antigen recognition is required for SLAT to exert its pivotal role in NFAT-dependent CD4⁺ T cell differentiation.

INTRODUCTION

Engagement of the T cell receptor (TCR) by specific antigen-major histocompatibility complex (MHC) complexes presented by antigen-presenting cells (APCs) is central to the effective induction of an antigen-specific T cell response. This cognate antigen presentation, taking place at the T cell-APC interface called the immunological synapse (IS), triggers biochemical signaling cascades involving multiple cellular proteins, such as protein tyrosine kinases, adapters, or cytoskeletal proteins, and activates in turn a number of transcription factors, notably NFAT, NF-κB, and AP-1. On a longer time scale, these pathways result in changes of gene expression that ultimately lead to T cell activation, proliferation, and differentiation.

Recently, we isolated a TCR-regulated protein called SWAP-70-like adaptor of T cells (SLAT) (Tanaka et al., 2003) on the basis of its abundant expression in T helper 2 (Th2) cells and its homology with SWAP-70, a B cell-enriched guanine nucleotide exchange factor (GEF) involved in B cell activation, immunoglobulin class switching, and migration to lymphoid organs (Borggrefe et al., 1998; Pearce et al., 2006; Shinohara et al., 2002). SLAT (also called Def-6 or IBP) is abundant in central and peripheral lymphoid tissues, with high amounts displayed in thymocytes and in peripheral T cells (Becart et al., 2007; Gupta et al., 2003b; Tanaka et al., 2003), and it translocates to the IS upon antigen stimulation (Gupta et al., 2003a; Tanaka et al., 2003). The human paralog of SLAT, termed IRF-4-binding protein (IBP), was independently isolated by another group (Gupta et al., 2003b) and later found to function as a TCR-regulated GEF for the Rho GTPases Rac1 and Cdc42 (Gupta et al., 2003a). In addition, SLAT cooperates with activated Rac1 to induce a change in cell shape, most probably independently of its GEF activity (Oka et al., 2007). Structurally, SLAT harbors, beginning at its N terminus, a potential Ca²⁺-binding EF-hand domain and an immunoreceptor tyrosine-based activation motif (ITAM)-like sequence of unknown function, a PI(3,4,5)P₃-binding pleckstrin-homology (PH) domain (Gupta et al., 2003a; Oka et al., 2007), and a Dbl-homology (DH) domain exhibiting GEF activity (Gupta et al., 2003a).

Examination of SLAT-deficient mice on a mixed genetic background revealed spontaneous development of systemic lupus in aged female mice (Fanzo et al., 2006). Our recent analysis of SLAT-deficient mice on a homogenous C57BL/6 background revealed a role of SLAT in thymic DN1 cell expansion, T cell activation, and Th1 and Th2 cell inflammatory responses (Becart et al., 2007). The defect in Th1 and Th2 cell responses was traced to defective Ca²⁺-NFAT signaling (Becart et al., 2007). However, the molecular basis by which SLAT contributes to NFAT activation is unknown. Here, we reported that SLAT specifically activated NFAT, but not NF-κB or AP-1, upon TCR triggering and that this NFAT activation correlated with, and depended upon, its membrane and IS translocation. This localization of SLAT required Lck-dependent phosphorylation of two tyrosine residues in its ITAM-like sequence. Furthermore, enforced targeting of the SLAT DH domain to the membrane promoted TCR-induced NFAT activation in a Cdc42- and, to a lesser extent, Rac1-dependent manner, and it restored NFAT activation and Th1-Th2 cell differentiation in SLAT-deficient CD4⁺ T cells.

RESULTS

SLAT Enhances TCR-Induced NFAT Activity and Is Recruited to the Membrane and IS

SLAT regulates Th1-Th2 cell differentiation by controlling NFAT activation in CD4⁺ T cells (Becart et al., 2007). To further understand the function of SLAT, we examined the effect of ectopic SLAT expression on the TCR-mediated activation of NFAT, NF-κB, and AP-1. SLAT-transfected Jurkat T cells showed a dose-dependent increase in NFAT-reporter activity relative to control transfectants (Figure 1A), which was abrogated by cyclosporin A, an inhibitor of calcineurin (Figure 1B), confirming the involvement of SLAT in modulating TCR-induced NFAT transcriptional activity. In contrast, neither NF-κB nor AP-1 activity was affected by SLAT (Figures 1C and 1D), and we confirmed these results by demonstrating that the anti-CD3- and anti-CD28-stimulated activation of these two transcription factors was intact in primary SLAT-deficient CD4⁺ T cells (Figure S1 available online).

Figure 1. SLAT Enhances NFAT Activity and Translocates to the Membrane and IS.

(A–D) Jurkat-TAg cells were cotransfected with empty pEF vector (10 μg) or with pEF vector encoding Xpress-tagged WT SLAT (1–10 μg), together with NFAT- (A), NF-κB- (B), or AP-1- (C) luciferase reporters (10 μg) plus a β-Gal (5 μg) reporter gene. In (D), cells were also transfected with a CD28-encoding pSRα vector. Cells were left unstimulated or stimulated with OKT3 (A–C) or OKT3 plus anti-CD28 (D) for the final 6 hr of a 24 hr culture. In (B), the cells were also treated with Cyclosporin A (CsA; 10 μM) during the 6 hr stimulation period. Normalized luciferase activity was determined in duplicates, and graphs show mean ± SD. Expression of transfected proteins was detected by anti-Xpress immunoblotting.

(E and F) Cytosolic (c) and membrane (m) fractions were prepared from Jurkat (JE6.1) T cells (E), peripheral T cells (F), or thymocytes (F) stimulated with anti-CD3 and anti-CD28 for the indicated times. Extracts were immunoblotted with the indicated antibodies (Abs). Numbers under the SLAT blot indicate the relative expression of SLAT in the membrane fraction as determined by densitometry. SLAT expression in unstimulated membrane fraction was assigned a value of 1. Expression of p38 MAPK and Lck in the cytosolic or membrane fractions, respectively (E), confirmed proper separation of the different fractions. Vav1 served as a normalization control.

(G) Jurkat-TAg cells were transfected with Xpress-SLAT. CMTMR-labeled Raji B cells were pulsed or not with SEE and mixed at a 1:1 ratio with the transfected Jurkat cells. Conjugates were bound to poly-L-lysine-coated coverslips, fixed, and stained with anti-Xpress plus a secondary Alexa 488-coupled anti-mouse Ig. Overlay of the green (Xpress-tagged SLAT) and red (CMTMR) images along with differential interference contrast (DIC) images are shown. Data shown are representative of eight (A–D), five (G), and three (E and F) independent experiments.

Next, we explored the subcellular localization of SLAT and the mechanisms that regulate this distribution. Anti-CD3 and anti-CD28 costimulation induced rapid SLAT translocation to the membrane fraction; this translocation was still evident after 5 min (Figure 1E). Indicative of proper activation, stimulation induced membrane translocation of PKC-θ. CD3+CD28 costimulation induced similar membrane translocation of SLAT in murine peripheral T cells or thymocytes (Figure 1F). Furthermore, we confirmed our earlier finding that TCR stimulation induces translocation of SLAT to the IS (Figure 1G) (Gupta et al., 2003a; Tanaka et al., 2003).

SLAT Localization and NFAT Activation Are Independent of PI3-K and PIP₃ Binding to the PH Domain

To gain more insight into the mechanism of SLAT-mediated NFAT activation, we assessed the effect of various SLAT mutants on NFAT activation (Figure 2A). A SLAT mutant lacking the first 122 aminoterminal residues (ΔEF) stimulated NFAT activity upon TCR stimulation to the same extent as wild-type (WT) SLAT. Further deletion of residues 123–215 (PH+DH) abolished NFAT activation (Figure 2B). Likewise, the isolated DH domain also did not support NFAT activity (Figure 2B). These results suggest a requirement for the region between the EF and PH domains in linking TCR signals to NFAT activation. A point mutation of Arg-236 (R236C), shown to be critical for binding of the phosphoinositide 3-kinase (PI3-K) product PI(3,4,5)P₃ (PIP₃) to the SLAT PH domain (Gupta et al., 2003a), as well as pretreatment with the PI3-K inhibitor LY294002, had no effect on NFAT activation (Figures 2B and 2C). Thus, PIP₃ binding to the PH domain of SLAT and, hence, PI3-K activity are not required for SLAT-mediated NFAT activation.

Figure 2. Mapping of SLAT Domains that Mediate NFAT Activation and Membrane Localization.

(A) Schematic representation of Xpress-tagged SLAT mutants. The EF hand, ITAM-like, PH, and DH domains are shown.

(B) Jurkat-TAg cells were cotransfected with an empty vector or the indicated SLAT plasmids (10 μg) plus NFAT-luciferase and β-Gal reporter genes. Stimulation and determination of relative luciferase units (RLU) were performed as in Figure 1A, and graphs show mean ± SD. SLAT expression was analyzed by anti-Xpress immunoblotting (inset).

(C) NFAT-luciferase activity was determined in the presence or absence of LY294002 (10 μM). Expression of SLAT was analyzed after 6 hr stimulation (right, top). Cells were stimulated or not for 5 min with a CD3 monoclonal antibody (mAb) in presence or absence of LY294002. Specificity of the PI3-K inhibitor was confirmed by its inhibition of Akt phosphorylation and lack of effect on ERK1 and ERK2 phosphorylation (right, bottom). Graph shows mean ± SD.

(D and E) Jurkat cells (D) or CD4⁺ primary WT T cells (E) were pretreated or not for 45 min with wortmannin (D) or LY294002 (D and E) and stimulated or not with anti-CD3 and anti-CD28 for 2 min. Cytosol and membrane fractions were prepared, resolved by SDS-PAGE, and immunoblotted with the indicated Abs. The relative expression of SLAT in the membrane fraction (D) was determined as in Figure 1E.

(F) Jurkat-TAg cells transfected with the indicated SLAT plasmids were stimulated with anti-CD3 (2 min), and subcellular fractions were analyzed by immunoblotting. The relative expression of SLAT in the membrane fraction was determined as in Figure 1E. Data shown are representative of six (B and C) and three (D, E, and F) independent experiments.

(G) SLAT-deficient CD4⁺ T cells preactivated with CD3+CD28 mAbs plus IL-2 were infected with WT SLAT or R236C-expressing retrovirus and cultured for 6 days. The GFP⁺ population was sorted and restimulated with CD3+CD28 mAbs for 2 min; then cytosol and membrane fractions were prepared, resolved by SDS-PAGE, and immunoblotted with the indicated Abs.

We then examined the subcellular localization of endogenous SLAT or recombinant SLAT mutants. LY294002 or wortmannin pretreatment had no effect on the TCR-induced translocation of endogenous SLAT to the membrane in Jurkat or primary CD4⁺ T cells (Figures 2D and 2E, respectively). The NFAT-activating ΔEF mutant was also recruited to the membrane after anti-CD3 stimulation, whereas the nonactivating PH+DH and DH proteins were not (Figure 2F). In contrast, the non-PIP₃-binding R236C SLAT mutant underwent intact stimulus-induced membrane translocation in transfected Jurkat T cells (Figure 2F) as well as in stimulated SLAT-deficient CD4⁺ T cells, which were infected with retrovirus expressing the R236C mutant (Figure 2G). Further analysis of the localization of the SLAT mutants vis-à-vis the IS in staphylococcal enterotoxin E (SEE)-stimulated Jurkat T cells revealed that the R236C mutant was properly localized to the IS (Figures 3A and 3B), and, moreover, LY294002 pretreatment did not prevent the translocation of WT SLAT in Jurkat T cells to the IS; conversely, the transfected PH+DH and DH protein failed to undergo IS translocation (Figures 3A and 3B). To further evaluate the potential role of PIP₃ binding to the PH domain in the context of an antigen-specific primary T cell response, we crossed SLAT-deficient mice with OT-II TCR-transgenic mice expressing an ovalbumin (Ova)-specific TCR and evaluated the localization of retrovirally transduced R236C SLAT mutant. Ova-peptide stimulation induced translocation of the R236C SLAT mutant, but not the DH mutant, to the IS (Figure 3C). The formation of an intact IS was confirmed by our showing that CD4 was also localized in the IS after antigen stimulation (Figure 3C). Together, these results reveal a strict correlation between TCR-mediated SLAT localization to the membrane and IS and SLAT’s ability to activate NFAT; furthermore, they indicate that a mechanism distinct from PH-domain binding to membrane-associated PIP₃ is critical for the membrane and IS recruitment of SLAT.

Figure 3. SLAT Recruitment to the IS.

(A) Jurkat-TAg cells were transfected with Xpress-tagged SLAT plasmids as indicated. After 24 hr, the cells were pretreated or not with LY294002 for 1 hr, then stimulated or not with SEE-pulsed CMTMR-labeled Raji B cells and analyzed as in Figure 1G. Overlay of the green (SLAT) and red (CMTMR) images along with DIC images are shown.

(B) Quantitative analysis of the results shown in (A). SLAT localization in the IS was analyzed in 200–250 T-APC conjugates. SLAT recruitment to the IS was classified in three categories: complete recruitment (represented by black bars), partial recruitment (represented by gray bars), or no recruitment (represented by white bars). The graph represents the mean percentage of imaged cells scored in each group ± SD from three experiments.

(C) Preactivated SLAT-deficient OT-II CD4⁺ T cells were infected with WT SLAT-, DH-, or R236C-expressing retrovirus. Sorted GFP⁺ CD4⁺ T cells were incubated with OVA_323–339 peptide-pulsed (+OVA) or nonpulsed (−OVA) DCs for 30 min. T-DC conjugates were then stained for CD4 (blue) and SLAT (red) expression. Corresponding DIC images are also shown. CD4 staining serves as a control of proper peptide-specific IS formation. Asterisks indicate T cells.

Phosphorylation of the ITAM-like Sequence of SLAT

The region between the EF and PH domains of SLAT contains an ITAM-like sequence (residues 133–144). This sequence contains negatively charged amino acids at conserved positions, the first YxxL (or YxxI) motif, a spacer sequence of seven amino acids, and a second Tyr-containing sequence, YxxK, in which a lysine residue replaces the leucine or isoleucine typically found in ITAM motifs (Figure 4A). A search of potential tyrosine-phosphorylation sites with the NetPhos2.0 algorithm revealed that Tyr-133 and Tyr-144 display a high score. Therefore, we generated tyrosine-to-phenylalanine SLAT mutants (Figure 4B) and examined their inducible phosphorylation. Pervanadate (PV) stimulation of COS-7 cells resulted in SLAT phosphorylation only when Lck was coexpressed (Figure 4C). In contrast, the phosphorylation of the Y144F mutant and, to a lesser extent, the Y133F mutant was drastically decreased (Figure 4D). Lck-dependent phosphorylation of the Y133-144F double mutant was completely abolished both in PV-stimulated COS-7 and in TCR-stimulated Jurkat T cells (Figures 4C–4E). The importance of Lck for SLAT phosphorylation was evidenced by the lack of this TCR-stimulated phosphorylation in Lck-deficient JCam1.6 Jurkat cells (Figure 4F).

Figure 4. Lck-Dependent Phosphorylation of Tyr-133 and Tyr-144 in SLAT.

(A) Comparison of the ITAM-like sequence of SLAT with the consensus ITAM. Gray boxes indicate conserved amino acids.

(B) Schematic representation of Tyr-mutated SLAT plasmids.

(C–E) COS-7 (C and D) or Jurkat-TAg (E) cells were transfected with the indicated plasmids and stimulated for 2 min with PV (C and D) or anti-CD3 (E). SLAT immunoprecipitates (top two panels) or cell lysates (bottom two panels) were analyzed by immunoblotting. Numbers under the p-Tyr blot indicate the relative phosphorylation of SLAT as determined by densitometry.

(F) WT (JE6.1) and Lck-deficient (JCaM1.6) Jurkat cells were stimulated with anti-CD3 for the indicated periods of time. SLAT immunoprecipitates (top two panels) or cell lysates (bottom two panels) were analyzed as in (C)–(E).

(G) Schematic diagram of retroviral pMIG-SLAT constructs.

(H) Primary SLAT-deficient T cells preactivated with anti-CD3 and anti-CD28 plus IL-2 were infected with WT SLAT or Y133-144F-expressing retrovirus and cultured for 6 days. GFP⁺ and GFP⁻ populations were sorted and restimulated with anti-CD3 and anti-CD28 for 5 min, and SLAT immunoprecipitates were analyzed by immunoblotting with the indicated Abs (top two panels). Cell lysates were blotted with anti-SLAT (bottom panel). The GFP⁻ population served as a negative control. Numbers under the p-Tyr blot indicate the relative phosphorylation of SLAT proteins. The data shown are representative of three independent experiments.

We confirmed the physiological relevance of these results by infecting SLAT-deficient primary T cells with retroviruses expressing WT or Y133-144F-mutated SLAT (Figure 4G). Phosphorylation of the Y133-144F mutant was decreased by 80% relative to that of WT SLAT (Figure 4H). These results indicate that SLAT undergoes Lck-dependent phosphorylation on Tyr-144 and Tyr-133 upon TCR and CD28 stimulation.

Membrane and IS Recruitment of SLAT and NFAT Activation Depend on Tyr-133+144 Phosphorylation

We next investigated the consequences of SLAT phosphorylation on SLAT’s membrane and IS translocation and NFAT activation. A Src-family-kinase inhibitor, PP2, impaired the anti-CD3-stimulated, SLAT-promoted NFAT activation in a dose-dependent manner (Figure 5A). The Y133F mutant and, to a much larger extent, the Y144F and Y133-144F mutants displayed reduced ability to activate NFAT (Figure 5B). We also generated a “pseudoactive” mutant by replacing Tyr-133 and Tyr-144 with a negatively charged aspartic-acid residue (Y133-144D) to mimic the phosphorylated state. This mutant, in contrast to Y133-144F, restored TCR-induced NFAT activation (Figure 5B).

Figure 5. Role of Tyr-133 and Tyr-144 Phosphorylation in SLAT Membrane and IS Recruitment and NFAT Activation.

(A) Luciferase-reporter activity in lysates of Jurkat-TAg cells cotransfected with empty or WT SLAT-expressing vector, along with an NFAT-luciferase reporter gene and a β-Gal reporter. Eighteen hr after transfection, the cells were incubated with PP2 and left unstimulated or were stimulated for 6 hr with anti-CD3. The graph shows mean of RLU ± SD. Inset, upper panels: Immunoblot analysis of lysates from transfected, PP2-treated and -stimulated cells. Inset, lower panels: Cells were pretreated with 1 or 10 μM PP2 for 45 min and stimulated with anti-CD3 for 5 min. Lysates were immunoblotted with the indicated Abs for confirmation of the effectiveness and specificity of PP2 pretreatment.

(B) Luciferase-reporter activity in lysates of Jurkat-TAg cells transfected with the indicated SLAT mutants plus NFAT-luciferase and β-Gal reporter genes. The graph shows mean of RLU ± SD. Expression of SLAT proteins was analyzed by anti-Xpress immunoblotting (inset).

(C) Cytosolic (c) and membrane (m) fractions from WT (JE6.1) and Lck-deficient (JCaM1.6) Jurkat cells stimulated with anti-CD3 and anti-CD28 for 2 min were immunoblotted as indicated. Relative SLAT expression in the membrane fraction was determined as in Figure 1E.

(D) Jurkat-TAg cells transfected with the indicated SLAT plasmids were stimulated with anti-CD3. Subcellular fractions were analyzed by immunoblotting. The relative expression level of SLAT proteins in the membrane fraction was determined as in Figure 1E.

(E) Jurkat-TAg cells transfected with the indicated SLAT plasmids were pretreated or not with PP2 (10 μM) for 45 min, then stimulated with CMTMR-labeled SEE-pulsed (or control) Raji B cells (red). SLAT localization (green) was analyzed as in Figure 1G.

(F) Quantitative analysis of SLAT localization in the IS was performed as in Figure 3B.

(G) Preactivated SLAT-deficient OT-II CD4⁺ T cells were infected with WT SLAT- or Y133-144F-expressing retrovirus. Sorted GFP⁺ CD4⁺ T cells were stimulated and analyzed as in Figure 3C. Asterisks indicate T cells. The data shown are representative of four (A and B), three (C–F), and two (G) independent experiments.

In parallel, we analyzed the membrane and IS translocation of tyrosine-mutated SLAT mutants. The Y133F mutation did not have a substantial effect on the membrane translocation of SLAT (Figure 5D) and only mildly inhibited IS relocalization (Figure 5E), in agreement with its partial effect on the tyrosine phosphorylation of SLAT (Figure 4D) or NFAT activation (Figure 5B). However, the Y144F and Y133-144F mutants were grossly impaired in their membrane or IS translocation upon TCR engagement (Figures 5D–5F). Interestingly, and in contrast to the Y133-144F mutant, the phosphorylation-mimicking Y133-144D mutant also translocated to the cell membrane in a stimulus-dependent manner (Figure 5D). The importance of Lck-dependent SLAT phosphorylation for its proper localization is evident from the findings that CD3+CD28 costimulation failed to induce membrane recruitment of endogenous SLAT in Lck-deficient Jurkat cells (Figure 5C) and that PP2 blocked relocalization of SLAT to the IS in SEE-stimulated WT Jurkat cells (Figure 5E). We extended these findings to primary CD4⁺ T cells by demonstrating that, unlike WT SLAT, the Y133-144F SLAT mutant did not localize to the IS in retrovirus-transduced, Ova-peptide-stimulated SLAT-deficient OT-II CD4⁺ T cells (Figure 5G). Thus, TCR+CD28-induced tyrosine phosphorylation of Tyr-144 and, to a lesser extent, Tyr-133 regulates SLAT translocation to the membrane and IS and, consequently, is critical for NFAT activation.

NFAT Activation Requires Membrane Recruitment of SLAT DH Domain and Cdc42-Rac1 Activation

The DH domain of SLAT possesses TCR-stimulated GEF activity for Rac1 and Cdc42 (Gupta et al., 2003a) (Figure S2), raising a question as to whether the inability of SLAT-DH to activate NFAT (Figure 2B) results from its failure to undergo membrane and IS localization (Figures 2F and 3) or, rather, reflects GEF independence of SLAT-mediated NFAT activation. We generated a construct (MyrDH) in which a membrane-localizing acylation sequence, consisting of the seven NH₂-terminal amino acids of Lck (MGCVCSS), was fused to the DH domain of SLAT (Figure 6A). As a negative control, we also mutated the myristoylation site of Lck (Gly-2) to alanine (mMyr), a mutation that abolishes Lck membrane localization (Yasuda et al., 2000). We confirmed that MyrDH was constitutively localized at the membrane and showed that it was recruited to IS in SEE-stimulated Jurkat cells, whereas the unmodified DH domain or the mMyrDH mutant were confined to the cytosol irrespective of TCR stimulation (Figures 6B and 6C). Furthermore, and as expected, the retrovirus-transduced MyrDH mutant, but not the soluble DH domain, localized to the IS in Ova-stimulated SLAT-deficient OT-II CD4⁺ T cells (Figure 6D). This localization pattern was paralleled when the ability of these mutants to activate NFAT was analyzed. Thus, in contrast to the cytosol-confined DH or mMyrDH mutants, MyrDH activated NFAT in stimulated T cells to the same extent as did WT SLAT (Figure 6E).

Figure 6. Membrane Targeting of the SLAT DH Domain Is Sufficient to Induce Cdc42- and Rac1-Dependent NFAT Activation.

(A) Schematic representation of retroviral Myc-tagged SLAT constructs. The Lck residues that are modified by myristoylation (Myr) or palmitoylation (Palm) are indicated.

(B) Jurkat-TAg cells were transfected with the indicated Myc-tagged vectors, and cytosolic and membrane fractions were analyzed by immunoblotting.

(C) Jurkat-TAg cells were transfected with the indicated Myc-tagged SLAT plasmids and stimulated with CMTMR-labeled SEE-pulsed (or control) Raji B cells (red). Overlays of the green (SLAT) and red (CMTMR, B cell) images are shown. The data shown are representative of four independent experiments.

(D) Preactivated SLAT-deficient OT-II CD4⁺ T cells were infected with DH- or MyrDH-expressing retrovirus. Sorted GFP⁺ CD4⁺ T cells were stimulated and analyzed as in Figure 3C. Asterisks indicate T cells.

(E and F) Jurkat-TAg cells were transfected with Myc-tagged SLAT plasmids plus NFAT-luciferase and β-Gal reporter genes. In (F), transfected cells were pre-treated or not with latrunculin B (0.25 μM) for 1 hr prior to 6 hr of anti-CD3 stimulation. Expression of SLAT proteins was analyzed by anti-Myc immunoblotting (top right panel in [F]). Efficiency of latrunculin B treatment was assessed by phalloidin staining of Jurkat-TAg cells (bottom right panel in [F]). Latrunculin B inhibited actin polymerization but not the expression of the SLAT or MyrDH, nor did it alter the constitutive membrane targeting of the MyrDH protein (right panel in [F]). Graphs show mean of luciferase activity ± SD and are representative of five independent experiments.

(G) Jurkat-TAg cells were transfected with SLAT plasmids along with Rac1^N17 or Cdc42^N17 plus NFAT-luciferase and β-Gal reporter genes. Normalized luciferase activity was determined as in Figure 1A, and the graph shows mean ± SD representative of four independent experiments.

(H) WT and SLAT-deficient (knockout [KO]) CD4⁺ T cells were activated with CD3+CD28 mAbs plus IL-2 for 48 hr, rested overnight, transduced or not with 1 μM recombinant Tat-Cdc42^CA protein for 30 min, and restimulated with CD3+CD28 mAbs for 2 hr. Cytoplasmic (C) and nuclear (N) fractions were immunoblotted with NFATc1 or NFATc2 Abs. Fractions were also immunoblotted with α-tubulin and lamin B Abs to confirm purity of the fractions, respectively. Whole-cell lysates were also immunoblotted with Cdc42, -SLAT, and -Vav1 Abs for assessing the efficiency of Tat-mediated transduction of Cdc42^CA ([H], bottom).

(I) [Ca²⁺]_i was measured by flow cytometry in indo1-loaded CD4⁺ T cells from WT and SLAT-deficient (KO) mice. Tat-Cdc42^CA transduction was performed as in (H), and TCR+CD28 costimulation was performed with a crosslinking goat anti-hamster (gαh) Ig Ab, followed by the addition of thapsigargin (Thapsi). Data are expressed as a histogram displaying FL5/FL4 ratio versus time (s).

(J) Preactivated WT and SLAT-deficient (KO) primary CD4⁺ T cells were transduced with retroviral vectors expressing either GFP alone (Empty) or GFP plus pMX Cdc42Q61L (Cdc42^CA) and differentiated under Th1 cell conditions. Sorted GFP⁺ populations were restimulated with anti-CD3 and anti-CD28 for 24 hr, and IFN-γ in culture supernatants was quantified by an ELISA. Graphs show mean of IFN-γ production ± SD. Retroviral gene transduction was used in this experiment (instead of Tat-mediated protein transduction) because this approach allows for stable, long-term gene expression (in contrast to Tat-Cdc42^CA, which is stable for only ~4 hr).

GEF-mediated activation of Rho GTPases induces actin-cytoskeleton reorganization, which is required for TCR-induced T cell activation. Therefore, we determined whether SLAT-induced NFAT activation depends on actin-cytoskeleton reorganization by analyzing the effect of Latrunculin B, an inhibitor of actin polymerization. This drug inhibited the TCR-stimulated NFAT activation in empty-vector-transfected cells and, moreover, also inhibited TCR-dependent NFAT activation mediated by WT SLAT or by MyrDH by ~75% and ~95%, respectively (Figure 6F, left), suggesting that actin-cytoskeleton reorganization is required for SLAT-mediated NFAT activation.

To further investigate the role of SLAT GEF activity in NFAT activation, we examined the effect of dominant-negative RhoA (RhoA^N19), Rac1 (Rac1^N17), or Cdc42 (Cdc42^N17) mutants on SLAT-mediated NFAT activation. RhoA^N19 had no effect on NFAT activation (data not shown), whereas Cdc42^N17 and, to a lesser extent, Rac1^N17 inhibited the SLAT- and MyrDH-induced NFAT activation (Rac1^N17: 25%–40% inhibition; Cdc42^N17: 75%–95% inhibition) (Figure 6G). Given this result, we also used the reverse approach by determining whether constitutively active Cdc42 (V12Cdc42; Cdc42^CA) can restore the impaired NFAT-Ca²⁺ signaling in SLAT-deficient primary CD4⁺ T cells. For this purpose, we used a protein-transduction approach, in which the cells were transduced with recombinant Cdc42^CA fused to a Tat-derived transduction peptide (Tskvitaria-Fuller et al., 2007). This recombinant fusion protein is transduced into primary T cells with high efficiency, and it was previously found to be functional (Tskvitaria-Fuller et al., 2007). Indeed, SLAT-deficient CD4⁺ T cells transduced with the Tat-Cdc42^CA protein regained their ability to undergo NFAT nuclear translocation (Figure 6H) and Ca²⁺ mobilization (Figure 6I) in response to anti-CD3 and anti-CD28 costimulation. Furthermore, Th1 cell differentiation measured by IFN-γ expression, which is normally impaired in SLAT-deficient CD4⁺ T cells (reflecting the critical role of NFAT in T cell differentiation) (Becart et al., 2007), was restored when SLAT-deficient primary CD4⁺ T cells were transduced with retrovirus expressing a Cdc42^CA mutant (Figure 6J). These data establish a causal link between the GEF activity of SLAT and the resulting activation of Cdc42 on one hand, and its critical function in promoting activation of the Ca²⁺-NFAT signaling pathway and, consequently, Th cell differentiation, on the other.

Restoration of Th Differentiation and NFAT Activation in SLAT-Deficient Primary T Cells

To determine the physiological relevance of our findings, we infected SLAT-deficient CD4⁺ T cells with bicistronic GFP retroviruses expressing WT (full-length) SLAT or various SLAT mutants and analyzed NFAT activation and Th1-Th2 cell differentiation. As controls, both NFATc1 and NFATc2 effectively translocated to the nucleus in WT T cells and in WT SLAT-reconstituted SLAT-deficient T cells (but not in empty-vector-transduced SLAT-deficient T cells) upon TCR+CD28 ligation (Figure 7A). Strikingly, however, reconstitution with a retroviral vector expressing the membrane-targeted (but not the cytosolic) DH domain fully rescued nuclear NFAT translocation (Figure 7A). Similarly, the R236C SLAT mutant also restored NFAT nuclear translocation; however, the Y133-144F mutant did not rescue NFAT activation (Figure 7A), consistent with our previous results (Figures 2 and 5).

Figure 7. Restoration of NFAT Activation and Th1-Th2 Cell Differentiation by Membrane-Targeted SLAT DH Domain.

(A) WT and SLAT-deficient CD4⁺ T cells were activated with anti-CD3 plus anti-CD28 plus IL-2 and transduced with retroviral pMIG vectors expressing either GFP alone (Empty) or GFP plus WT SLAT, DH, MyrDH, R236C, or Y133-144F mutants. Sorted GFP⁺ populations were restimulated with anti-CD3 and anti-CD28 for 2 hr. Cytoplasmic (C) and nuclear (N) fractions were immunoblotted with NFATc1 or -NFATc2 Abs. Fractions were also immunoblotted with anti-α-tubulin and anti-lamin B to confirm purity of the fractions, respectively.

(B) Preactivated WT and SLAT-deficient (KO) primary CD4⁺ T cells were infected with the indicated retroviruses and differentiated under Th1 conditions. Cells were harvested after three rounds of infection and rested for 24 hr, and GFP⁻ or GFP⁺ populations were sorted. IFN-γ in culture supernatants was quantified by ELISA 24 hr after anti-CD3+CD28 restimulation, and graphs show mean of IFN-γ production ± SD. Statistical differences were determined using two-tailed Student’s t test. ^#p < 0.001, WT versus SLAT-deficient mice.

(C) Preactivated SLAT-deficient OT-II CD4⁺ T cells were infected with empty pMIG retrovirus or with the indicated WT SLAT or SLAT mutant-expressing retrovirus. As a control, WT OT-II T cells were transduced with empty pMIG retrovirus. Sorted GFP⁺ cells prepared after 7 days were adoptively transferred into naive C57BL/6 mice, which were immunized 1 day later with Ova+CFA. Spleen cells harvested 6 days later were restimulated with Ova_323–339 peptide, and IFN-γ⁺ cells were enumerated. The background level of IFN-γ-expressing GFP⁺ WT OT-II cells, which were not stimulated in vitro, was ~2.5% (not shown).

Because impaired activation of both NFATc1 and NFATc2 accounts for the defective Th1 and Th2 cell responses in SLAT-deficient mice (Becart et al., 2007), we further investigated the ability of the same SLAT mutants to rescue Th1 (Figure 7B) or Th2 cell differentiation (Figure S3) of SLAT-deficient T cells. As shown before (Becart et al., 2007), SLAT-deficient CD4⁺ T cells transduced with the empty vector displayed a ~50% reduction in both IFN-γ and IL-4 production in comparison to that of their WT counterparts, and both GFP⁻ and GFP⁺ populations displayed this defect (Figure 7B and Figure S3). In contrast, infection of these cells with WT SLAT-expressing retrovirus restored Th1 and Th2 cell differentiation to normal, WT T cell levels (Figure 7B and Figure S3). As an internal control, the GFP⁻ (nontransduced) populations of SLAT-deficient CD4⁺ T cells still displayed a defect in IL-4 or IFN-γ expression. In full correlation with the effect of the SLAT mutants on NFAT nuclear translocation (Figure 7A), the MyrDH and R236C SLAT mutants, but not DH or Y133-144F, fully rescued the Th1 (Figure 7B) and Th2 (Figure S3) cell-differentiation defect.

Lastly, we confirmed the restoration of Th1 cell differentiation of SLAT-deficient T cells in the context of an in vivo antigen-specific response by using an adoptive-transfer system (Figure 7C). As a positive control, ~24% of adoptively transferred CD4⁺ WT OT-II T cells were IFN-γ⁺, whereas, in contrast, only ~5% of the transferred SLAT-deficient T cells reconstituted with empty pMIG expressed IFN-γ, confirming the defective Th1 cell differentiation of SLAT-deficient CD4⁺ T cells (Figure 7C). However, retroviral reconstitution with full-length SLAT, membrane-targeted MyrDH, or R236C, but not Y133-144F, resulted in similar proportions (~15%–20%) of IFN-γ⁺ T cells, approaching the level of IFN-γ⁺ cells in the positive control group (Figure 7C). These results indicate that membrane targeting of the isolated SLAT DH domain is necessary and sufficient for proper TCR-induced NFAT activation and CD4⁺ T cell differentiation.

DISCUSSION

NFAT is a major mediator of productive T cell activation, and its activation is required for production of various cytokines, including IL-2 (Masuda et al., 1998), IL-4 (Chuvpilo et al., 1993), and IFN-γ (Campbell et al., 1996). Therefore, analysis of the regulation of SLAT function was of major interest because we found SLAT to be a key component of the Ca²⁺-NFAT pathway in T lymphocytes (Becart et al., 2007). In addition, SLAT plays a role in actin polymerization through its ability to activate the small Rho GTPases Rac1 and Cdc42 (Fanzo et al., 2006; Gupta et al., 2003a).

The current study establishes SLAT as a key protein linking actin-cytoskeleton reorganization and Cdc42-Rac1 activation to Ca²⁺-NFAT signaling and, consequently, to Th1 and Th2 cell differentiation. SLAT translocated to the membrane and IS and potently activated NFAT in a TCR-dependent manner, but it was dispensable for AP-1 and NF-κB activation, in agreement with the phenotype of SLAT-deficient primary T cells (Becart et al., 2007). Consistent with the findings that IS assembly and translocation of signaling molecules to the membrane and IS greatly influence T cell activation (Grakoui et al., 1999), we observed a strict correlation between the structural features required for SLAT membrane and IS recruitment and SLAT’s ability to activate NFAT. Both events required the Lck-dependent phosphorylation of Tyr-133 and Tyr-144. However, the combined PH-DH domains were not sufficient, and PI3-K-dependent binding of the PH domain to PIP₃ was not even required. Furthermore, enforced membrane targeting of the isolated DH domain was sufficient to enhance TCR-stimulated NFAT activation in a Cdc42-Rac1-dependent manner, controlling in turn Th1 and Th2 cell differentiation. The corroboration of these findings in primary T cells establishes the physiological relevance of these findings.

The lack of any apparent regulatory role for SLAT PH-domain binding to membrane-associated PIP₃, as well as the importance of Tyr-144 in these events, are at odds with another study (Gupta et al., 2003a). This study demonstrated that the IS translocation of SLAT depended on PI3-K and, furthermore, that SLAT GEF activity was stimulated by Lck-mediated tyrosine phosphorylation of another tyrosine residue, Tyr-210 (Gupta et al., 2003a). We found, however, that mutation of Tyr-210 did not alter the Lck-dependent phosphorylation of SLAT (Figure S4), consistent with the low score obtained for Tyr-210 with the NetPhos2.0 algorithm. The reasons for these discrepancies are unclear. However, our results were obtained by several independent and complementary approaches, including retroviral transduction of SLAT-deficient primary T cells, thereby establishing the robustness of our findings. Furthermore, the previous finding of Lck-mediated phosphorylation of SLAT on Tyr-210 was based on analysis of Lck-overexpressing non-T (293T) cells (Gupta et al., 2003a).

Interestingly, phosphorylation of Tyr-133 and Tyr-144 was required, but not sufficient, for inducing NFAT activation because a SLAT mutant mimicking phosphorylation of these residues (Y133-144D) still needed TCR stimulation in order to activate NFAT and translocate to the cell membrane, suggesting that SLAT membrane and IS translocation is required but not sufficient. Similarly, enforced membrane targeting of the isolated DH domain with the MyrDH mutant also was not sufficient to activate NFAT and required TCR stimulation. Furthermore, the MyrDH- (or WT SLAT-) induced NFAT activation was dependent on functional Cdc42 (and Rac1?), as indicated by the ability of a DN Cdc42 mutant to block SLAT-dependent NFAT activation and, conversely, the restoration of NFAT nuclear translocation, Ca²⁺ mobilization, and Th1 cell differentiation in SLAT-deficient CD4⁺ T cells by transduced constitutively active Cdc42. These findings make a strong case for a causal role of the GEF activity of SLAT in inducing its Ca²⁺- and Th-differentiation-promoting activities and, thus, provide a mechanistic basis for our previous findings that SLAT-deficient T cells display impaired Ca²⁺-NFAT signaling and Th cell differentiation (Becart et al., 2007). So far, we (and others) have not identified DH domain residue(s) essential for the GEF activity of SLAT (or SWAP-70), mutation of which would be required to directly establish the importance of the GEF activity of SLAT for Ca²⁺ signaling and Th cell differentiation. However, it is important to note that the DH domain of SLAT lacks sufficient primary amino acid sequence homology with “conventional” DH domains or even with the DH domain of its close relative, SWAP-70, thereby making this task the goal of future analysis.

Together, these findings indicate that additional TCR-induced signals, beyond the mere membrane and IS translocation of SLAT, are required for its proper function. One likely regulatory model is that the Lck-mediated phosphorylation of SLAT creates a binding site or sites for an undefined critical TCR-regulated adaptor protein and/or induces a conformational change in SLAT (that is not fully mimicked by the Y133-144D mutant). As a result of such a modification, SLAT is recruited to the IS, where it can access and activate Rho GTPases. In any case, the ability of MyrDH SLAT to fully activate NFAT and to reconstitute Th1 and Th2 cell differentiation in SLAT-deficient primary T cells clearly demonstrates that, first, mislocalization of the SLAT DH domain in the cytosol upon TCR engagement underlies its inability to activate NFAT. Second, it substantiates the conclusion that membrane and IS association of SLAT is a prerequisite for NFAT activation.

SLAT-deficient T cells display impaired IS formation and defective actin polymerization (Fanzo et al., 2006), most likely due to impaired Cdc42-Rac1 activation, along with a defect in Ca²⁺-NFAT signaling (Becart et al., 2007). As mentioned above, the dependence of Ca²⁺-NFAT activation on Cdc42 (and, to a lesser extent, Rac1) activity implicates SLAT as a key actin-cytoskeleton regulator linking activation of Rho GTPases to NFAT activity, and it is consistent with Cdc42 and Rac being substrates for the GEF activity of SLAT (Gupta et al., 2003a). In this context, antigen-receptor-stimulated NFAT activation was previously found to require Rac1 activity (Turner et al., 1998). The importance of Rho GTPase-mediated actin-cytoskeleton reorganization in assembly of the IS and in controlling early events preceding gene transcription (such as migration, integrin activation, or conjugate formation) is well established (Cannon and Burkhardt, 2002; Tybulewicz, 2005; Wang et al., 2003). However, the pathways linking actin rearrangements to Ca²⁺-NFAT signaling are far less clear. The finding that cytochalasin D treatment impairs TCR-stimulated NFAT, but not NF-κB or AP-1, activation (Nolz et al., 2007) suggests that the actin cytoskeleton does not globally regulate all TCR-stimulated signaling pathways. It is intriguing that the Slat mutation results in a similar phenotype, i.e., impaired NFAT activation in the face of intact NF-κB or AP-1 activation.

Several Cdc42-Rac1 effectors can selectively regulate Ca²⁺-NFAT signaling. For example, the Wiskott-Aldrich syndrome protein (WASP), a Cdc42 effector, plays a role in actin-cytoskeleton rearrangements as well as in NFAT activation via its association with WIP (Dong et al., 2007; Huang et al., 2005; Silvin et al., 2001). Similarly, the WAVE2 complex regulates actin-cytoskeleton reorganization but is also required for NFAT (but not NF-κB or AP-1) activation, reflecting its essential role in calcium-release-activated calcium (CRAC)-channel-mediated Ca²⁺ entry (Nolz et al., 2006). Finally, Pak1 kinase, a common Rac1 and Cdc42 effector, is involved in NFAT activation via its Lck-mediated phosphorylation and resulting association with the adaptor Nck (Yablonski et al., 1998). Therefore, further studies will be required in order to identify downstream effectors involved in Cdc42-Rac1-dependent, SLAT-mediated NFAT activation and determine whether they are linked to any of the regulators of the actin cytoskeleton and Ca²⁺ signaling discussed above. In this regard, we found that MyrDH-mediated NFAT activation was totally abrogated by an ERK, but not a p38, kinase inhibitor (Figure S5). This data is consistent with accumulating evidence showing that Rho GTPases regulate gene transcription through diverse distal effector elements of the TCR signaling pathway, including extracellular signal-regulated kinase (ERK) (Frost et al., 1997; Jiang et al., 2000; Yablonski et al., 1998), whose activation is also decreased in SLAT-deficient T cells (Becart et al., 2007). Notably, Yablonski et al. (Yablonski et al., 1998) demonstrated that Rac and Cdc42 could influence NFAT by activating ERK through its target Pak1 in T cells. Cdc42 has also been recently shown to specifically modulate ERK activation (Marques et al., 2008), and other studies provide evidence of a functional coupling of NFAT transcriptional activity and Ras-MEK-ERK activation (Sanna et al., 2005; Tsukamoto et al., 2004; Villalba et al., 2000; Yang et al., 2005).

Our finding that enforced membrane targeting of the SLAT DH domain is sufficient to restore NFAT activation and Th1 or Th2 cell differentiation in SLAT-deficient CD4⁺ T cells suggests that the failure to stimulate sufficient Cdc42 and/or Rac1 activity accounts for the NFAT-dependent Th1-Th2 cell defects of SLAT-deficient T cells. This notion is further reinforced by our findings that the downstream defects observed in SLAT-deficient T cells, i.e., the impaired NFAT activation and Th cell differentiation, can all be reversed by expression of a constitutively active Cdc42 mutant. This causality between activation of Rho-family GTPases and Th cell differentiation is consistent with reports showing that (1) Rac2 activates Th1 cell-specific signaling (Li et al., 2000), and (2) CD4⁺ T cells from Wiskott-Aldrich syndrome patients (mutated in WAS gene encoding WASP) display defective Th1 cell response due to impaired NFATc2 activation (Trifari et al., 2006). Importantly, however, our findings go one step further by implicating SLAT as a critical regulator of not only Th1 cell responses, but also Th2 cell responses and, furthermore, link this regulatory function in a causal manner to the small GTPase-stimulating GEF activity of SLAT and, hence, to actin-cytoskeleton reorganization.

In summary, our data provide evidence that TCR+CD28 stimulation induce Lck-mediated membrane localization of SLAT. Thereafter, TCR-specific signals facilitate the triggering of SLAT GEF activity toward Cdc42 and Rac1, leading to NFAT activation and Th1 and Th2 cell differentiation. Further work aiming at elucidating SLAT-interacting partners may thus shed light on the poorly understood events that coordinate and link actin-cytoskeletonreorganization to Ca²⁺ signaling and gene transcription in T cells. Moreover, strategies aimed at inhibiting the translocation of SLAT to the IS or, more directly, its GEF activity may provide an efficient means of blocking the machinery of Ca²⁺-NFAT signaling in T cells for the treatment of NFAT-dependent autoimmune and inflammatory diseases. In that regard, SLAT may turn out to be an attractive drug target with potentially similar importance to that of the recently discovered T cell CRAC channel, Orai1 (Feske et al., 2006; Vig et al., 2006; Yeromin et al., 2006), and the Ca²⁺ sensor, Stim1 (Roos et al., 2005; Zhang et al., 2005).

EXPERIMENTAL PROCEDURES

Antibodies and Reagents

SLAT-specific antisera were prepared as described (Tanaka et al., 2003). Monoclonal antibodies (mAbs) specific for human CD3 (OKT3) and pTyr (4G10) were affinity purified from hybridoma-culture supernatants as described (Liu et al., 1997). The human CD28 mAb and anti-mouse CD4 (RM4-5) were from PharMingen. The Xpress mAb was from Invitrogen. Lck, ERK2 (D-2) -c-Myc (9E10; for immunoblotting), NFATc1 (7A6), NFATc2 (4G6-G5), Cdc42 (P1), and lamin B Abs were from Santa Cruz Biotechnology. Rat α-tubulin (YL1/2) Ab was purchased from Serotec, and PKC-θ Ab was obtained from Transduction Laboratories. Phospho-Tyr319 ZAP-70, phospho-Thr202 and Tyr204 ERK1+2, Akt, and p38 polyclonal Abs and phospho-Ser473 Akt and ZAP-70 (99F2) rabbit mAbs were obtained from Cell Signaling Technology. Mouse CD3ε (145-2C11) and CD28 (37.51) mAbs were purchased from Biolegend. Vav1 Ab was from Upstate Biotechnologies, and c-Myc mAb used for immunofluorescence was obtained from Roche Applied Science. The IκB-α Ab was from Santa Cruz Biotechnology. SEE was from Toxin Technology. The cell tracker orange (CMTMR), Alexa-488-conjugated anti-mouse Ig Ab, Alexa-555-conjugated anti-rabbit Ig Ab, and Alexa-647-conjugated anti-rat Ig Ab were obtained from Molecular Probes. The Src inhibitor (PP2), latrunculin B, LY294002, and wortmannin were obtained from Calbiochem. The recombinant Tat-Cdc42^CA fusion protein was a kind gift from C. Wülfing (Tskvitaria-Fuller et al., 2007).

Plasmids

Xpress-tagged WT SLAT (pEF-SLAT-Xpress) plasmid was generated by subcloning of the murine SLAT cDNA into the EcoRI sites of the pEF4-His C expression vector (Clontech) encoding an in-frame Xpress tag epitope upstream of the insert. The pEF-SLAT-Myc expression plasmid was constructed by cloning of the SLAT cDNA, fused in frame with a c-Myc epitope-coding sequence at its 3^′ terminus, into the EcoRI and NotI sites of the pEF-Myc-His A vector. Various Xpress-tagged deletion mutants of SLAT were constructed by PCR amplification from full-length SLAT vectors with the following primers: SLAT-ΔEF (Δ1–90) forward primer (FP1): 5^′-GGTACCGGATGAGTTGTGCTGGACCCT-3^′; reverse primer (RP1): 5^′-GCGGCCGCCTAATTCCCTGGTGCTGGATCC-3^′. SLAT-PH+DH (Δ1–216) FP2: 5^′GGTACCGGTCCTGAAGCAGGGCTATCT-3^′; reverse primer RP1. SLAT-DH (Δ1–312) FP3: 5^′-GAATTCCTGCAGGCGGAGGGGAAGAC-3^′; reverse primer RP1. PCR products were inserted into KpnI-EcoRI (for SLAT-ΔEF and SLAT-PH+DH) or EcoRI-NotI (for SLAT-DH) sites of pEF4-His C vector and sequenced to confirm correct nucleotide sequence. Underlining indicates unique restriction sites used for subcloning. The pEF-MyrDH vector was generated by amplification of the SLAT DH domain with a 5^′ primer encoding the membrane-localizing acylation sequence of Lck. The Y133F and/or Y144F or D, R236C, and mutated MyrDH (mMyrDH) point mutations were introduced by use of a site-directed mutagenesis kit (Stratagene). Presence of single mutations was confirmed by DNA sequencing. Dominant-negative forms of Rac1 (Rac1^N17) and Cdc42 (Cdc42^N17) were expressed in pEF or pCAN vectors, respectively.

Mice, Cell Culture, Purification, and Transfection

Mice were maintained under specific pathogen-free conditions in accordance with guidelines of the Association for Assessment and Accreditation of Laboratory Animal Care International. SLAT-deficient mice on a C57BL/6 background have been previously described (Becart et al., 2007). SLAT-deficient OT-II TCR-transgenic mice were generated by intercrossing OT-II TCR-transgenic mice and SLAT-deficient mice (both on a C57BL/6 background), and their T cells were used as a source of Vβ5-Vα2 CD4⁺ T cells specific for amino acid residues 323–339 of Ova (OVA_323–339). Six- to 10 week-old mice were used in all experiments. After erythrocyte lysis, lymph node and spleen T cells were enriched to >90% purity on mouse T cell-enrichment columns (R&D Systems). CD4⁺ T cells were isolated by positive selection with CD4 (L3T4) mAb-coated microbeads (MACS). T cells were cultured in RPMI-1640 medium (Life Technologies) supplemented with 10% heat-inactivated fetal bovine serum (FBS), 2 mM glutamine, 1 mM sodium pyruvate, 10 mM HEPES, 1 mM MEM nonessential amino acid solution, and 100 U/ml each of penicillin G and streptomycin (Life Technologies). Short stimulation of murine T cells was performed as described (Becart et al., 2007).

Simian virus 40 large T antigen-transfected human leukemic Jurkat T cells (Jurkat-TAg), WT Jurkat cells (clone JE6.1), Lck-deficient (JCaM1.6) Jurkat T cells, and Raji B cells were obtained from ATCC and grown in the same RPMI-1640 medium. Jurkat-TAg cells in logarithmic growth phase were transfected with plasmid DNAs by electroporation (Villalba et al., 2001). The cells were washed with serum-free RPMI-1640 medium, serum starved, incubated with the human CD3 and CD28 mAbs on ice, and stimulated at 37°C for the indicated times by crosslinking with a secondary goat anti-mouse IgG (Pierce), as described (Charvet et al., 2005). Short stimulation of Jurkat-TAg cells (subcellular fractionation and immunoprecipitation experiments) was performed with crosslinked OKT3 mAb for the indicated times at 37°C. COS-7 cells were grown in DMEM medium plus 10% FBS and transfected with the TransIT®-LT1 transfection reagent (Mirus). After 24 hr, transfected cells were harvested for stimulation with 1% (w/v) PV and cell-extract preparation.

Immunofluorescence Microscopy

Conjugation of transfected Jurkat-TAg cells and Raji B cells, staining, and preparation of the samples for microscopy were performed as described(Charvet et al., 2005). For F-actin visualization, Jurkat-TAg cells were fixed, permeabilized, and stained with TRITC-conjugated phalloidin (Sigma-Aldrich). For antigen-specific stimulation, retrovirally transduced CD4⁺ T cells from SLAT-deficient or WT OT-II TCR-transgenic mice were incubated with 10 μg/ml OVA_323–339 peptide-pulsed or nonpulsed dendritic cells (DCs). Cells were briefly centrifuged to promote conjugate formation and incubated at 37°C for 30 min in serum-free medium on poly-L-lysine-coated slides, then fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100 in PBS, and stained with mouse CD4 (RM4-5; PharMingen) and SLAT Ab. Goat anti-rat and anti-rabbit IgG secondary antibodies conjugated to Alexa Fluor 647 or Alexa Fluor 555 (Molecular Probes), respectively, were used to visualize primary antibodies. Immunofluorescence images were recorded by a Marianas digital fluorescence-microscopy system (Intelligent Imaging Innovations) as described (Charvet et al., 2005). A minimum of 100 randomly selected conjugates per slide were examined in three to five independent experiments.

Retroviral Transduction

The full-length SLAT cDNA in the pMIG retroviral vector was described (Tanaka et al., 2003). The pMIG-DH vector was generated as follows: First, the SLAT DH domain (residues 313–630) was PCR-amplified with the forward and reverse primers 5′-AGATCTACCATGACTGATTTTTATCTGAAGCTGCAGGCGGAGGGGAAGAC-3′ and 5′-AGATCTCTAATTCCCTGGTGCTGGATCC-3′, respectively, then subcloned into the pGEM vector before being finally subcloned into the BglII site of pMIG. Single mutants (pMIG-Y133-144F or D) were generated by site-directed mutagenesis with the pMIG-SLAT vector as a template. pMIG-MyrDH vector was obtained by first adding a HpaI site to pEF-MyrDH vector via site-directed mutagenesis with the forward and reverse primers 5′-CAGTAGCTTGGTACCGTTAACATGGGCTGTGTCTGC-3′ and 5′-GCAGACACAGCCCATGTTAACGGTACCAAGCTACTG-3′, respectively, followed by subcloning into the HpaI and EcoRI sites of pMIG. The retroviral vector of Cdc42^CA (pMX GFP Cdc42Q61L) was purchased from Addgene.

Platinum-E packaging cells (Morita et al., 2000) (0.4 × 10⁶ cells) were plated in a six-well plate in 2 ml DMEM plus 10% FBS. After overnight incubation, the cells were transfected with 3 μg retroviral plasmid DNA with TransIT®-LT1 transfection reagent. After 24 hr, the medium was replaced with RPMI plus 10% FBS. Cultures were maintained for 24 hr, and the retroviral supernatant was harvested, filtered (0.45 μm), supplemented with 5 μg/ml polybrene and 100 U/ml IL-2, and then used to infect CD4⁺ T cells that had been preactivated for 18 hr with CD3+CD28 mAbs plus recombinant IL-2 (100 U/ml) under Th-1 or Th2-polarizing conditions. Plates were centrifuged for 1 hr at 2000 rpm and incubated for 8 hr at 32°C and for 16 hr at 37°C, followed by two additional retroviral infections at daily intervals.

Adoptive Transfer and Ova Immunization

CD4⁺ T cells from OT-II WT or SLAT-deficient mice were isolated and cultured with plate-bound anti-CD3 (5 μg/ml), soluble anti-CD28 (2.5 μg/ml), and IL-2 (100 U/ml). The cells were retrovirus-infected on days 2–4 and cultured for another 3 days. GFP⁺ CD4⁺ T cells were sorted, and 2 × 10⁵ cells were injected intravenously into naive C57BL/6 mice. One day later, the mice were immunized subcutaneously with Ova (50 μg; Grade V; Sigma-Aldrich) emulsified in complete Freund’s adjuvant (CFA) (Difco; BD Diagnostics). Flow-cytometric enumeration of transferred IFN-γ⁺ T cells was performed 6 days later, after in vitro restimulation with OVA_323–339 peptide (10 μg/ml) for 8 hr at 37°C in presence of Golgi Plug (BD Biosciences). Samples were analyzed on a FACSCalibur flow cytometer with CellQuest (BD Biosciences) and FlowJo software (TreeStar) after gating on GFP⁺ and CD4⁺ T cells.