Emt/Itk Associates with Activated TCR Complexes: Role of the Pleckstrin Homology Domain¹

Expressed in mast and T-cells/inducible T cell tyrosine kinase (Emt/Itk)³ belongs to the Tec family of protein tyrosine kinases that are structurally organized in a N-terminal pleckstrin homology (PH) domain, followed by Tec homology (TH), Src homology 3 (SH3), SH2, and catalytic (SH1) domains (1). Emt/Itk is involved in critical roles in TCR-induced calcium flux (2), IL-2 production (2), proliferation (3), CTL responses (4), and thymic development (3).

After TCR stimulation, inositol phospholipids generated by phosphatidylinositol 3-kinase recruit Emt/Itk via PH domain interactions to the plasma membrane (5) where it comes in proximity with Lck, which phosphorylates and activates Emt/Itk (6, 7). Emt/Itk can also be activated by ligation of CD2 or CD28 on T cells (8, 9, 10, 11, 12) and by engagement of FcεRI on mast cells (13).

Although the downstream targets of Emt/Itk are unknown, Emt/Itk-deficient mice display compromised phospholipase C-γ1 phosphorylation and reduced extracellular calcium flux in response to TCR stimulation (2). These events, along with others such as activation of Src tyrosine kinases, are important early signals that follow the engagement of the TCR (14). The interaction of the TCR with antigenic peptide-MHC on APC induces the formation of molecular clusters that, in addition to TCR/CD3, contain additional cell surface and intracellular proteins that are critical for the propagation of T cell activation signals (15, 16, 17). A number of proteins, including the CD4 and CD8 coreceptors and the Src family kinases Lck and Fyn, have been shown to colocalize with the TCR/CD3 complex and facilitate transduction of the antigenic signal (17, 18, 19).

Stimulation of T cells through CD28 induces the physical association of this coreceptor with Emt/Itk, as demonstrated by coimmunoprecipitation analysis (11, 12). Similar experiments, however, where stimulation through TCR/CD3 has been attempted, have failed to reproducibly reveal an inducible association of Emt/Itk with the TCR complex (Refs. (6, 9) and our own unpublished observations). These negative data do not necessarily rule out an inducible physical association between Emt/Itk and TCR/CD3, but instead they may point out the possibility that such an interaction may not be amenable to biochemical analysis utilizing detergents and coimmunoprecipitation techniques. To assess this, we have utilized laser confocal microscopy where the interaction of Emt/Itk and TCR/CD3 is studied in its natural milieu at the single cell level. The data indicate that Emt/Itk colocalizes with TCR/CD3 molecular clusters upon CD3ε engagement in a manner similar to that for CD28-Emt/Itk colocalization. This phenomenon depends on the membrane association of Emt/Itk through its PH domain and correlates with TCR-induced activation of Emt/Itk as assessed by tyrosine phosphorylation. Finally, the PH domain is indispensable for the activation of Emt/Itk through TCR/CD3.

SV40 T Ag-transfected human leukemic Jurkat T cells (JTAg) were cultured at 37°C with RPMI 1640 (Irvine Scientific, Irvine, CA) containing 10 mM HEPES, 2 mM l-glutamine (Sigma, St. Louis, MO), 8% FBS (HyClone, Logan, UT) in a humidified 5% CO₂ atmosphere. Jurkat cells, 20 × 10⁶ in 400 μl, were transfected with 20 μg of the desired plasmid DNA by electroporation (Bio-Rad gene pulser, Hercules, CA) in a 0.4-cm-gap electrocuvette (DocFrugal’s, San Diego, CA) at 960 μF and 240 mV. After 48 h of culturing, ∼10% of the cells expressed visible levels of the transfected protein, as assessed by green fluorescent protein (GFP) fluorescence.

Wild type (wt) murine Emt/Itk (20) was cloned into the EcoRI-SpeI sites of the pME18 expression vector (21). Truncation of the PH domain (ΔPH) was performed by deletion of the PvuII fragment, containing the PH and part of the TH domain; religation; and subcloning into the EcoRI-EcoRV sites of pME18s-wt-Emt/Itk.

The H902 epitope tag, consisting of aa 319–333 of HIV gp120, was added by PCR amplification (30 cycles, 94°C, 1 min; 51°C, 2 min; 72°C, 2 min) of pME18s-Wt-Emt/Itk using the following primers: sense primer 5′-GGAAGCCATGGCCCGTATCCAGAGAGGACCAGGGAGAGCATTTGTTACAATAGGAAAAATGAACAACTTCATCCTCCT-3′ (codingfor aa 319–333 of gp120 and aa 1–7 of Emt/Itk) and antisense primer 5′-TTCCCATATCTTAGCCCTGC-3′ (coding for aa 348–354 of Emt/Itk). Addition of H902 tag to ΔPH-Emt/Itk was performed similarly with the same sense primer and the antisense primer 5′-CTCATAGGTTTCAAGGT-3′ (coding for aa 233–238 of Emt/Itk). The PCR products were resolved on a 1% agarose gel, isolated and purified with the Geneclean II kit (Bio 101, Vista, CA), and cloned into the pCR2.1 vector with the use of the TA cloning kit (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. Finally, the EcoRI-ScaI fragments of the pCR2.1 constructs were subcloned into the EcoRI-ScaI sites of pME18s-Wt-Emt/Itk.

GFP was added to the C terminus of Emt/Itk by removing the stop codon and creating an in-frame KpnI site by PCR amplification of H902-tagged pME18s-Wt-Emt/Itk or H902-tagged pME18s-ΔPH-Emt/Itk. The following primers were used: sense primer 5′-TTGAATTCCGTTGGAAGCCATGGCCCGTAT-3′ (containing an EcoRI site and aa 1–4 of the H902 epitope tag), antisense primer 5′-TTTCTAGAGGTACCCAAGCCCAGCTTCTGCGATTT-3′ (containing aa 613–619 of Emt/Itk, KpnI, and XbaI sites). The PCR products were purified as described above and subcloned into the EcoRI-KpnI sites of pEGFP-N2 expression vector (Clontech, Palo Alto, CA).

Addition of the first 12 amino acids of Lck to the pEGFP-ΔPH-Emt/Itk-GFP was performed by PCR amplification with the following primers: sense primer 5′-AAGAATTCGAAGCCATGGGCTGTGTCTGCAGCTCAAACCCTGAAGATGACGCCCGTATCCAGAGAGGACC-3′ (containing aa 1–12 of Lck and aa 1–7 of H902 tag) and the antisense primer described in the previous paragraph for the addition of GFP. The PCR product was cloned into pCR2.1, and the EcoRI-ScaI fragment was subcloned to pEGFP-Wt-Emt/Itk-GFP.

The above constructs were confirmed by sequencing.

JTAg cells, 30 × 10⁶, were lysed with 1 ml of lysis buffer (1% Nonidet P-40, 150 mM NaCl, 20 mM Tris (pH 7.3), 0.4 mM EDTA, 5 μg/ml leupeptin, 5 μg/ml pepstatin A, 1 mM sodium o-vanadate, and 1 mM PMSF) for 1 h at 4°C. Cell lysates were incubated for 2 h at 4°C with 5 μg monoclonal anti-H902 (IgG1) Ab (National Institutes of Health AIDS Reference Reagent Program, Rockville, MD), followed by overnight incubation with 30 μl protein G-Sepharose at 4°C. Immune complexes were washed three times with lysis buffer, resolved by 8% SDS-PAGE, and transferred to polyvinylidene difluoride membranes (Gelman Sciences, Ann Arbor, MI). Western blotting analysis was done as previously described (22). Monoclonal anti-phosphotyrosine Ab PY-20 was provided by Dr. Bartholomew Sefton (Salk Institute, La Jolla, CA), and anti-Emt/Itk antiserum was produced in rabbits immunized with a GST fusion protein as previously described (23).

Emt/Itk immune complexes, prepared as described above, were washed once with kinase reaction buffer (50 mM HEPES (pH 7.4), 2 mM MnCl_2, 200 μM ATP, 10 mM MgCl₂, and 10 mM DTT) and then resuspended in 100 μl of the same buffer. The immune complexes were incubated for 1 h at 37°C, resolved by SDS-PAGE (8%), and analyzed by Western blotting with anti-phosphotyrosine Abs.

For stimulation, JTAg cells, 30 × 10⁶, were incubated (30 min on ice) in RPMI 1640 culture medium containing 20 μg/ml concentrations of the anti-CD3ε mAb OKT3 (hybridoma obtained from American Type Culture Collection, Manassas, VA) or an isotype (IgG2a) control Ab UPC-10 (Bionetics, Charleston, SC). Cells were washed, resuspended in culture medium containing 20 μg/ml of rabbit anti-mouse IgG (cross-linking Ab, Jackson Immunoresearch, West Grove, PA), and incubated at 37°C for 3 min. After incubation, the cells were placed on ice until further analysis.

For fluorescence microscopy, 2 × 10⁶ cells were incubated (30 min on ice) in 250 μl of culture medium containing 40 μg/ml Ab OKT3. After a washing, the cells were incubated similarly in culture medium containing 20 μg/ml Texas Red-conjugated goat anti-mouse IgG (Molecular Probes, Eugene, OR). The cells were then exposed to 37°C for 10 min to induce capping and were immediately cooled with 1 ml ice-cold culture medium containing 0.1% sodium azide to block TCR internalization. After two washes in azide-containing medium, the cells were placed on poly-l-lysine-coated glass slides and fixed with 2% paraformaldehyde in PBS. The slides were mounted with Prolong Anti-fade (Molecular Probes) according to the manufacturer’s instructions.

Confocal imaging was performed using the Bio-Rad MRC1024 LSCM system. GFP and Texas Red were imaged separately to avoid channel bleed-through. GFP was excited at 488 nm and read through the 522–535 nm channel PMT, whereas Texas Red was excited at 568 nm and detected by the 598–640 nm channel PMT. Images were acquired with the Bio-Rad LaserSharp software, converted with GraphicConverter (Lemke Software, Peine, Germany), and pseudocolored and overlaid with Adobe Photoshop 4.0 (Adobe, San Jose, CA). The relative amount of membrane-associated Emt/Itk was calculated by measuring pixel intensity with ImageQuant software version 1.1 (Molecular Dynamics, Sunnyvale, CA). To this end, the total amount of GFP fluorescence was determined by defining a gate that included the whole cell. The intracytoplasmic GFP fluorescence was determined by setting a gate as above but excluding the cell membrane. The fraction of membrane-associated GFP was calculated as the total minus intracytoplasmic over the total GFP.

Emt/Itk is important in the propagation of signals initiated both through the TCR/CD3 molecular complex and through the CD28 coreceptor molecule (6, 11, 24). Cross-linking of CD28 with specific Abs induces the coimmunoprecipitation of Emt/Itk and CD28 (11, 12). However, under the same experimental conditions, an association between Emt/Itk and TCR/CD3 has not been demonstrated (6, 9). In our own studies (unpublished observations), we have observed association of CD3ε and Emt/Itk after anti-CD3ε Ab immunoprecipitation of Brij 97 cell lysates. However, this observation is not consistently reproducible. A possible explanation of such findings is that a putative association among these proteins may not be amenable to analysis by detergent solubilization and immunoprecipitation techniques.

Here, we have reexamined this issue by utilizing GFP-tagged Emt/Itk and fluorescence confocal microscopy. We have prepared fusion Emt/Itk constructs, where GFP has been cloned at the C terminus of the kinase, and analyzed expression in transiently transfected cells by laser scanning confocal microscopy. The presence of GFP increases the relative molecular size of the protein to 102 kDa and does not interfere with either the TCR/CD3-induced phosphorylation (Fig. 3,E) or the kinase activity (Fig. 3,F) of Emt/Itk. In transiently transfected JTAg cells, wt-Emt/Itk-GFP is primarily extranuclear, and it is distributed diffusely throughout the cytoplasm with a distinct perimembranous localization (Fig. 1 A).

We examined the effect of TCR/CD3-induced stimulation on the intracellular localization of Emt/Itk. JTAg cells, transiently transfected with wt-Emt/Itk-GFP, were treated with anti-CD3ε Ab OKT3 and cross-linked with Texas Red-conjugated rabbit anti-mouse IgG. Cells were examined by confocal microscopy before (nonstimulated cultures) or after (stimulated cultures) incubation at 37°C for 10 min. Before exposure to 37°C, CD3ε (red) appears as a punctate circumference around the plasma membrane (Fig. 1,C), whereas the membrane-associated Emt/Itk (green) forms a contiguous perimembranous ring (Fig. 1,A). Under these conditions, the two signals do not colocalize (Fig. 1,B). In cells incubated at 37°C, both Emt/Itk and CD3ε form similarly shaped clusters (“caps”) of identical cellular localization (Fig. 1, D and F, respectively). Superimposition of the two images confirms the OKT3-induced colocalization of Emt/Itk and CD3ε (Fig. 1 E). This colocalization event was verified in four independent experiments where every examined cell displaying CD3ε “capping” also displayed Emt/Itk coclustering with the features described above.

To obtain an estimate of the fraction of total Emt/Itk-GFP that is membrane associated, we quantified GFP fluorescence in both stimulated and nonstimulated cells as described in Materials and Methods. Analysis of nine individual nonstimulated cells revealed that 28.9 ± 4.3% (average ± SD) of total Emt/Itk-GFP was membrane associated. Similar analysis of 10 individual stimulated cells revealed a significant (p < 0.001) increase in the fraction of membrane-associated kinase to 40.4 ± 6.2%.

Other investigators have previously established the inducible association of Emt/Itk and the CD28 coreceptor utilizing a biochemical approach (11, 12). In experiments similar to those described above, we were able to confirm this association at the single-cell level using confocal microscopy. Thus, CD28 cross-linking induces membrane redistribution of Emt/Itk and its colocalization with CD28 (Fig. 1, G–L).

To further explore the association of Emt/Itk with CD3ε, we examined the role of the Emt/Itk PH domain in this event. PH do- mains can interact with membrane phospholipids and thus allow recruitment of the PH domain-containing protein to the cell membrane for potential interaction with other membrane-associated structures (25). We constructed a truncation mutant of Emt/Itk-GFP where the PH domain was deleted and transiently transfected it into JTAg cells. In contrast to wt-Emt/Itk-GFP, the ΔPH-Emt/Itk-GFP does not display enhanced membrane localization, and it is rather diffusely cytoplasmic (Fig. 2,A). In stimulated cells (incubation with Ab OKT3 and cross-linking with Texas red-conjugated anti-mouse IgG at 37°C), ΔPH-Emt/Itk fails to localize in distinct patches as those seen in cells transfected with wt-Emt/Itk (compare Fig. 1,D with Fig. 2,D). This occurs despite the “capping” of CD3ε (Fig. 2,F). Thus, no colocalization of Emt/Itk with CD3ε occurs (Fig. 2,E). The above experiments were also performed with CD28-induced activation using Ab 9.3 with identical results. Thus, ΔPH-Emt/Itk fails to colocalize with CD28 also (Fig. 2, G–L). In the above experiments, we cannot totally exclude the possibility that small, undetectable amounts of ΔPH-Emt/Itk associate with the cell membrane through a non-PH domain-mediated mechanism. The ΔPH-Emt/Itk-GFP transfectants look uniformly brighter than their wt-Emt/Itk-GFP counterparts not because of higher levels of expression but because the sensitivity of the photomultiplier tube was calibrated to the strongest signal emitted by the cell, caused by the relatively uniform and diffuse cytoplasmic fluorescence of the ΔPH-Emt/Itk-GFP.

Lck has been shown to be critical for the TCR/CD3-induced activation of Emt/Itk, in that it is needed for trans-phosphorylating and subsequently activating Emt/Itk (6, 7). Therefore, association with the cell membrane is an important requirement for Emt/Itk to be targeted by Lck. Thus, as it might be predicted, ΔPH-Emt/Itk transiently transfected into JTAg cells fails to become tyrosine phosphorylated on incubation with Ab OKT3 (Fig. 3,A, top). In contrast, similarly transfected wt-Emt/Itk becomes phosphorylated (Fig. 3 A, top). This latter finding is in agreement with previously published observations (6, 7). Similar experiments performed with anti-CD28 stimulation indicate that ΔPH-Emt/Itk is not phosphorylated in response to Ab 9.3 stimulation either (data not shown). The lack of ΔPH-Emt/Itk phosphorylation is specific, given that other target proteins do become tyrosine phosphorylated upon stimulation (indicated by anti-phosphotyrosine Western blotting of total cell lysates; data not shown).

Tyrosine phosphorylation of Emt/Itk correlates with its kinase activity (6, 7, 11). However, when we tested kinase activity by autophosphorylation assay, we were surprised to see that in spite of its lack of anti-CD3-induced trans-phosphorylation (Fig. 3,A, top), ΔPH-Emt/Itk has a significantly higher spontaneous kinase activity than did the wt control (Fig. 3,C, top). Densitometric analysis of the signals indicates that the baseline in vitro kinase activity of ΔPH-Emt/Itk is 3.5-fold higher than that of wt-Emt/Itk. It is interesting that increases in basal kinase activity upon removal of the PH domain have been also observed with ΔPH mutants of protein kinase B (26) and protein kinase D (27). The addition of GFP does not alter the inducible trans-phosphorylation of Emt/Itk (Fig. 3,E). Furthermore, GFP linkage does not interfere with autophosphorylation of the ΔPH mutant (Fig. 3,F). In Fig. 3,A, we do not detect autophosphorylation of the ΔPH mutant because these experiments are performed in vivo and they only reflect trans-phosphorylation presumably by Src family kinases. The results of the above experiments cannot be attributed to unequal protein concentrations in each lane, because probing with anti-Emt/Itk antiserum reveals equal loading (Fig. 3, A and C, bottom).

The ΔPH-Emt/Itk gene product migrates as two bands (Fig. 3, A and C, bottom). We do not know the reason, but possibly it might be due to alternative splicing. The same is true for ΔPH-Emt/Itk-GFP (data not shown). In addition, the higher m.w. form appears to be more intensely phosphorylated than the lower form (data not shown).

The lack of ΔPH-Emt/Itk membrane localization raised the question whether the PH domain serves merely as a membrane targeting signal or whether it is also necessary for other events related to Emt/Itk activation. To address this, we further modified the ΔPH-Emt/Itk-GFP by adding to its N terminus a membrane localization sequence consisting of the first 12 amino acid residues of Lck, which provides a myristoylation and acylation target sequence (28). JTAg cells were transiently transfected with this construct (mΔPH-Emt/Itk) and stimulated as described above. The addition of the Lck-membrane localization sequence allows the PH domain truncation mutant to localize to the cell membrane (Fig. 4,A), and to colocalize with the TCR/CD3 complex on stimulation with Ab OKT3 (Fig. 4 D–F). Similar results were obtained when transfectants were stimulated with Ab 9.3 (Fig. 4, G–L). To ensure that colocalization of the mΔPH-Emt/Itk with TCR/CD3 and CD28 is specific, we added the Lck membrane localization signal to GFP alone (mGFP) and tested its localization in transiently transfected JTAg cells. Even though mGFP localized to the membrane (Fig. 4,M), there was no colocalization with TCR/CD3 on OKT3 stimulation (Fig. 4, M–R). Similar results were obtained with 9.3 stimulation (data not shown). Surprisingly, however, transiently transfected JTAg cells do not display tyrosine phosphorylation of mΔPH-Emt/Itk after incubation with Ab OKT3 (Fig. 3,B, top). Furthermore, mΔPH-Emt/Itk did not display the high basal autokinase activity seen with the ΔPH-Emt/Itk mutant (Fig. 3 D, top). Stimulation of mΔPH-Emt/Itk transfected cells with Ab 9.3 also failed to reveal trans-phosphorylation and kinase activity of the transfected mutant protein (data not shown). The possibility that lack of phosphorylation of mΔPH-Emt/Itk is caused by inactivation of endogenous Lck is not likely. Lowin-Kropf et al. (29) have demonstrated that TCR/CD3 clustering (“cap” formation) and cytoskeletal reorganization during T cell and APC interaction depend on the presence of active Lck. Lck-deficient Jurkat T cells do not display these events, but they do so when the Lck gene is reconstituted by transfection (29). Because expression of the ΔPH-Emt/It mutant allows “cap” formation, we conclude that Lck must be active.

Our data constitute the first demonstration of inducible colocalization between Emt/Itk and the TCR/CD3 molecular complex at the single-cell level. Furthermore, the data underscore the importance of the PH domain not only as a membrane targeting signal but also as an indispensable component for the inducible activation of Emt/Itk.

PH domains are important targeting signals that allow proteins containing them to be recruited to the cell membrane through interaction with membrane phospholipids (25). The biological importance of PH domains is underscored by observations with the tyrosine kinase Btk. Mutations in this Tec family kinase have been associated with human X-linked agammaglobulinemia (30, 31) and murine X-linked immunodeficiency (32, 33). Many of the identified mutations have been localized in the PH domain of Btk (34), and they interfere with the ability of the mutant kinase to interact with specific membrane phospholipids and become activated (35, 36, 37).

After TCR-mediated stimulation, inositol phospholipids recruit Emt/Itk via PH domain interactions to the plasma membrane (5) and presumably bring it to close proximity with Lck which in turn phosphorylates and activates Emt/Itk (6, 7). Thus, deletion of the PH domain would not allow Emt/Itk to localize to the appropriate molecular clusters on the cell membrane upon TCR engagement, come to proximity with Lck, and become phosphorylated. In fact, the data presented here support this contention, in that ΔPH-Emt/Itk fails to localize to the cell membrane and it does not become phosphorylated on tyrosines upon TCR stimulation. This result raised the question of whether the PH domain simply plays a membrane-targeting role or has additional functions related to the activation of Emt/Itk. To address this issue, we modified our ΔPH-Emt/Itk construct by adding to its N terminus the first 12 amino acids of Lck (mΔPH-Emt/Itk). These residues are targets for myristoylation/acylation and membrane localization of Lck (28). Indeed, the mΔPH-Emt/Itk was targeted to the cell membrane, and it coclustered with TCR/CD3 and CD28 upon ligation of these receptors. However, to our surprise, it did not become tyrosine phosphorylated upon stimulation.

August et al. (5) targeted a ΔPH-Emt/Itk mutant to the cell membrane of COS-7 cells by the addition of the murine c-Kit extracellular and transmembrane domains and found that upon coexpression of Src, the targeted ΔPH-Emt/Itk becomes tyrosine phosphorylated and enzymatically active. This seems to be in disagreement with our data where membrane-targeted mΔPH-Emt/Itk does not display significant trans-phosphorylation or enzymatic activity upon TCR/CD3 ligation (Fig. 3). However, there might be several reasons to explain the discrepancy between the results of August et al. and those presented here. These include the use of different cellular systems (nonlymphoid COS vs lymphoid Jurkat cells), different modes of Emt/Itk activation (Src overexpression vs TCR/CD3 stimulation), and different membrane localization signals (c-kit vs Lck signal).

One explanation of why mΔPH-Emt/Itk, although colocalized with TCR/CD3, cannot become trans-phosphorylated on stimulation could be that the PH domain may be critical for endowing the appropriate conformation for Emt/Itk to become phosphorylated by Lck. This contention is supported by the data of Andreotti et al. who analyzed a fragment of Emt/Itk by multidimensional nuclear magnetic resonance and found an intramolecular association between the SH3 domain and a proline-rich sequence located between the PH and SH3 domains (38). Thus, intramolecular interactions may be important in regulating the interaction of Emt/Itk with its targets/substrates and its subsequent activation. If the absence of the PH domain causes conformational changes that affect receptor-induced phosphorylation of Emt/Itk, such changes are unlikely to be global so to affect the conformation of the whole molecule, because mΔPH-Emt/Itk can colocalize inducibly with both TCR/CD3 and CD28 in a manner similar to that of the nonmutant protein (Fig. 4).

Another explanation for the lack of phosphorylation of mΔPH-Emt/Itk could be the inability of the mutant kinase to associate with other critical signaling molecules through an interaction that is dependent on the PH domain. The interaction of PH domains with membrane phospholipids has been well documented (25). For example, the PH domain of Emt/Itk can bind to inositol D3 phosphates, most notably inositol 1,2,3,4,5-hexaphosphate (5). The PH domain of Btk can interact with IP₄, IP₅, IP₆ (35), and PI(3, 4, 5)P₃ in a manner dependent on the activation of phosphatidylinositol 3-kinase (39, 40, 41). Mutations found in the PH domain of Btk from X-linked agammaglobulinemia patients impair lipid-PH interactions and hamper enzymatic activity (35, 41).

Other signaling proteins with which Emt/Itk and Btk have been shown to interact are heterotrimeric GTP-binding proteins. Thus, the Gα_q, Gα₁₂ and Gβγ subunits of GTP-binding proteins can interact with and activate the enzymatic activity of Emt/Itk and Btk both in vivo and in vitro (42, 43, 44, 45). The interaction of Btk with Gα_q and Gα₁₂ is mediated through distinct domains. Gα_q binds to a region composed of the TH and SH3 domains (42), whereas Gα₁₂ interacts with the PH domain and the adjacent Btk motif (43). In our own studies, a Jurkat cell line expressing a “dominant-negative” mutant of Gα₁₆ (46) reveals deficient TCR/CD3-induced tyrosine phosphorylation of Emt/Itk (unpublished observations) consistent with the interpretation that heterotrimeric GTP-binding proteins are involved in the activation of this Tec kinase. In addition to the above, there are still other signaling proteins that have been shown to interact with Emt/Itk and Btk including protein kinase C (13, 47) and BAP-135 (48).

The inducible association between Emt/Itk and CD28, previously demonstrated by coimmunoprecipitation (11, 12), has been confirmed here at the single-cell level by laser scanning confocal microscopy. The fact that a similar association between Emt/Itk and TCR/CD3 can be demonstrated only by confocal microscopy, but not by biochemical means (Refs. 6, 9 and our own unpublished results), suggests that this association may be sensitive to the manipulations required (e.g., detergent solubilization) for coimmunoprecipitation analysis. Another association that has been refractory to coimmunoprecipitation analysis but revealed by confocal microscopy is the one between Btk and the B cell Ag receptor (49).

In our experiments, we noticed that about one-third of total cellular Emt/Itk-GFP was associated with the cell membrane in nonstimulated Jurkat T cells; this amount increased significantly on stimulation through TCR/CD3. The association of Emt/Itk-GFP with the cell membrane, under nonstimulating conditions, cannot be attributed to the fact that the protein has been modified by the addition of GFP because ΔPH-Emt/Itk-GFP does not accumulate in the plasma membrane (Fig. 2). Furthermore, other investigators using different techniques have seen association of both Emt/Itk and Btk with membrane components. Thus, use of a biochemical cell fractionation approach reveals that ∼20% of the total Emt/Itk in resting Jurkat T cells is associated with light vesicles (L. Berg, unpublished observations). Nisitani et al. (49) studied the colocalization of Btk with the B cell receptor complex in Ramos B cells using specific Abs and confocal microscopy. It is clear from their data that under “no stimulation” conditions there exists significant association of Btk with the cell membrane that increases on Ag receptor cross-linking. Thus, it appears that in the cell lines studied, although the majority of Btk and Emt/Itk remains cytoplasmic under nonstimulating conditions, a significant proportion associates with the cell membrane.

Monks et al. (17) have recently demonstrated that on T cell activation the Ag receptor complex along with other signal transduction proteins, including protein kinase C-θ, Fyn, Lck, and LFA-1, are clustered into segregated three-dimensional domains within the contacts between T cells and APC. Monks et al. have termed these clusters supramolecular activation clusters (SMAC), and they have demonstrated that upon activation proteins in the SMACs are compartmentalized into central and peripheral clusters in reference to the TCR (17). Thus, Lck and Fyn kinases seem to colocalize in the central SMACs along with TCR/CD3, whereas LFA-1 remains peripherally confined in the peripheral SMACs. The data presented here support a two-step model where, upon stimulation, Emt/Itk first becomes recruited to the cSMACs and then, in the close proximity of Src kinases, becomes trans-phosphorylated and activated. The PH domain seems to serve a dual role in this process; it mediates targeting of Emt/Itk to the cell membrane and it is also involved in its activation.

We thank M. Wood for assistance with confocal microscopy.