The Physical Association of Protein Kinase Cθ with a Lipid Raft-Associated Inhibitor of κB Factor Kinase (IKK) Complex Plays a Role in the Activation of the NF-κB Cascade by TCR and CD28¹ Free

CD28 is a key accessory receptor that enhances cellular activation via the TCR. Not only is CD28 an integral component of the supramolecular activation clusters (SMAC)⁴ that assemble at the TCR contact site with the APC, but it also plays an active role in the recruitment of signaling molecules to the TCR (1, 2, 3, 4, 5). The latter effect is mediated by the coclustering of lipid rafts to the TCR synapse. Lipid rafts are detergent-resistant, cholesterol and glycosphingolipid-enriched membrane domains which contain proteins that contribute to signal transduction (6). These rafts also include signaling molecules belonging to the src-family, LAT (linker for activation of T cells) and Ras (1, 4, 6). Through this action, CD28 enhances signal transduction by the TCR, including engagement of signaling pathways that cannot be activated by TCR alone. One example is activation of the c-Jun N-terminal kinase (JNK) cascade and another engagement of the NF-κB pathway (7, 8). These cascades appear to be functionally linked through the action of a mitogen-activated protein kinase kinase kinase (MAPKKK), MEKK1, which induces JNK as well as inhibitor of κB factor (IκB) kinase (IKK)β activity during CD28 coligation (7, 8, 9, 10). Both cascades are critical for the activation of the IL-2 gene, including activation of the CD28 response element (RE) in the promoter of that gene (7, 11, 12). This composite RE, which includes c-Rel and AP-1 binding sites, is activated in a synergistic fashion by the NF-κB and JNK cascades during TCR/CD28 coligation (7, 13). Although the mechanism of NF-κB activation by the TCR/CD28 is unclear, we have recently shown that CD28 synergizes with TCR in the activation of homo- and heterodimeric IKK complexes in primary human T cells (8). Activated IKKs phosphorylate the inhibitory proteins, IκBα and IκBβ, leading to their degradation and release of NF-κB transcription factors into the nucleus (8, 14, 15).

IKKα, IKKβ, and IKKγ are the core components of a 700- to 900-kDa multimolecular kinase complex, which is responsible for stimulus-induced phosphorylation of IκBs (16, 17, 18, 19, 20, 21, 22). Although proteins like MEKK1 and the NF-κB-inducing kinase have been shown to associate with the IKKs, there is considerable debate as to whether these are functionally involved in the regulation of IKK activity (18, 23, 24). In addition, two other serine-threonine kinases, namely Cot kinase and protein kinase C θ (PKCθ), have now been implicated in IKK activation by CD3/CD28 (25, 26, 27, 28). Whether these kinases are involved in individual signaling pathways or act in hierarchical fashion is still uncertain. In this communication, we will focus on PKCθ because this is the only PKC isoform that is recruited to the TCR synapse and the SMAC (1, 5, 29, 30). Moreover, PKCθ has been shown to be involved in NF-κB activation in T lymphocytes, in addition to its ability to induce the JNK/AP-1 pathway (6, 26, 27, 28, 31, 32, 33). This suggests that PKCθ plays a key role in the regulation of signaling cascades influenced by CD28 receptor.

Because activated PKCθ is recruited to the T cell membrane, it is possible that IKK activation may commence in a signalsome that assembles at the TCR synapse. We demonstrate that CD3/CD28 coligation in primary human T lymphocytes induces relocation of endogenous PKCθ as well as IKKs to a GM1 ganglioside-enriched, detergent-insoluble membrane compartment that overlaps with lipid rafts. IKKs recruited to lipid rafts were actively capable of phosphorylating the recombinant IκBα in vitro. Although exogenously expressed PKCθ induced IKKβ activity in Jurkat cells, it was also possible to show that endogenous PKCθ associated with activated IKK complexes in primary human CD4⁺ T cells. Expression of dominant-active (DA) and dominant-negative (DN) PKCθ had stimulatory and inhibitory effects, respectively, on the CD28RE. Taken together, these data suggest that activation of PKCθ by TCR and CD28 plays an important role in the assembly and activation of IKK complexes in the T cell membrane.

OKT3 (anti-CD3) was obtained from Ortho Pharmaceuticals (Raritan, NJ), and the 9.3 (anti-CD28) and 187.1 mAb were generously provided by Bristol-Meyer Squibb (Princeton, NJ). Polyclonal anti-PKCθ and anti-Lck were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), and anti-IKKα, anti-IKKβ, and anti-IKKγ were purchased from PharMingen (San Diego, CA). Anti-Flag (M2) and FITC-conjugated anti-Flag Abs were obtained from Sigma (St. Louis, MO). Tosyl-activated magnetic beads and M-450 anti-CD4 beads were purchased from Dynal (Great Neck, NY). The PKC inhibitors Gö6976 and rottlerin were purchased from Calbiochem (San Diego, CA). Recombinant IL-2 was from Chiron (Emeryville, CA).

Human CD4⁺ T cells were purified from PBLs and were stimulated with anti-CD3 and anti-CD28 mAbs coupled to tosyl-activated magnetic beads as previously described (8, 34). After stimulation for ∼14 days, beads were removed and cells returned to their resting state over a 48-h time period (8). These rested cells were restimulated as described below.

The Jurkat T cell clone BMS2 and resting primary human CD4⁺ T lymphocytes were stimulated with 2 μg/ml each anti-CD3(OKT3) or OKT3 plus anti-CD28 (9.3) mAb, secondarily cross-linked with 10 μg/ml mAb 187.1 for 30 min. Stimulation with P (100 nM) plus I (1 μg/ml) was used as positive control. Alternatively, Jurkat T cells were stimulated by OKT3/9.3 Abs coupled to magnetic beads for 30 min before conducting confocal microscopy.

Flag-IKKα and IKKβ (wild-type and kinase-inactive mutants) were provided by Tularik (San Francisco, CA) (19). PKCθ, wild-type (wt), DA, and DN (kinase-inactive) were provided by Dr. Baier (University of Innsbruck, Innsbruck, Austria) (31). The CD28RE/AP-1 luciferase reporter was previously described (8, 13). For cellular transfection, we used the indicated amounts of cDNA for electroporation (240 V, 950 μF) as previously described (8, 35)

A total of 2 × 10⁷ primary CD4⁺ or Jurkat T cells were stimulated with anti-CD3 (aCD3) plus anti-CD28 (aCD28) mAb (2 μg/ml each), secondarily cross-linked with 187.1 (10 μg/ml), or P + I (100 nM and 1 μg/ml, respectively). Cells were lysed in buffer A (50 mM HEPES (pH 7.6), 250 mM NaCl, 1% Triton X-100, 2 mM MgCl₂, 2 mM DTT, 0.1 mM Na₃VO₄, 20 μM β-glycerophosphate, and 20 μM p-nitrophenylphosphate) and cleared by centrifugation. Lysate (200 μg) was treated with 2 μg of the indicated Abs, bound to protein A-Sepharose, and rocked for 2 h at 4°C. Immunoprecipitated complexes were washed, separated by SDS-PAGE, and transferred to nitrocellulose membrane (8). Membranes were sequentially overlaid with the indicated concentrations of the primary Abs, followed by a 1:2000 dilution of the secondary HRP-conjugated Ab.

Cellular fractionation was performed as previously described (36). Briefly, 1 × 10⁸ primary CD4⁺ cells were suspended for 15 min in hypotonic buffer E (10 mM Tris (pH 7.4), 10 mM KCl, 1.5 mM MgCl₂, 0.2 mM PMSF, 1 mM sodium vanadate, 10 μg/ml leupeptin, and 0.01 TIU/ml aprotonin), then lysed with 50 mechanical strokes in a Dounce homogenizer. Lysates were spun for 30 min at 100,000 × g, and the supernatant, designated cytosol, was collected. The pellet was rinsed twice in buffer E and resuspended in buffer E, containing 1% Nonidet P-40 for 30 min. Samples were recentrifuged at 100,000 × g, and the supernatant, designated detergent-soluble material (DSM), was collected (33). The remaining pellet was rinsed once in buffer E plus detergent, followed by sonication in RIPA buffer (20 mM Tris (pH 7.5), 250 mM NaCl, 10 mM DTT, 10 mM MgCl₂, 1% Nonidet P-40, 0.1% SDS, 0.5% sodium deoxycholate), and cleared by centrifugation as described above. This supernatant was designated detergent insoluble material (DIM) (36).

Aliquots of 1 × 10⁸ Jurkat or primary human CD4⁺ lymphocytes stimulated or unstimulated were lysed in 1 ml of ice-cold MBS (25 mM MES pH6.5, 150 mM NaCl, 0.5% Triton X-100, 1 mM Na₃VO₄, 1 mM PMSF, and 10 μg/ml aprotinin). Lysates were homogenized using 15 strokes of a Dounce homogenizer, mixed with an equal volume of 85% sucrose (w/v) in MBS, and transferred to a SW41 centrifuge tube. The mixture was first overlaid with 6 ml of 35% sucrose, then 3 ml of 5% sucrose in MBS with 1 mM Na₃VO₄, and centrifuged at 200,000 × g for 16 h at 4°C. Following centrifugation, eleven 1-ml fractions were sequentially collected from the top of the centrifuge tube. The low-density fractions (i.e., fractions 2–4) contained lipid raft material, whereas cytosolic material and soluble membranes were predominantly recovered in fractions 9–11. Primary human CD4⁺ T cells were similarly processed, except cells were stimulated with anti-CD3 + anti-CD28 mAb coupled to magnetic beads.

To determine the efficiency of the isolation of lipid rafts via sucrose gradient centrifugation, 2 μl of each fraction was dot blotted onto Immobilon-P Nitrocellulose Transfer membrane (Millipore, Bedford, MA). After drying, the membrane was blocked in 6% BSA for 1 h at room temperature and washed once in PBS with 0.1% Tween 20 (PBS-T). HRP-linked cholera toxin β subunit (Sigma) was diluted in PBS-T to 4.2 μg/μl and was used to overlay the membrane for 1 h. The membrane was washed three times, incubated with SuperSignal Chemiluminescent Substrate (Pierce, Rockford, IL), and subjected to autoradiography.

A total of 10⁷ Jurkat or resting CD4⁺ T cells in 1 ml of RPMI were either unstimulated or stimulated with anti-CD3 + CD28 mAb as described above (8). For the effects of PKC inhibitors on IKK activity, CD4⁺ T cells were incubated with the indicated concentrations of these drugs for 1 h before stimulation. Two hundred microgram lysate was incubated with 2 μg anti-IKKγ Ab, adsorbed onto protein A-Sepharose for 2 h. In vitro kinase assays using GST-IκB1–55(1–55) as substrate were performed as previously described (8, 17).

A total of 10⁷ Jurkat cells were transfected with 10 μg of the consensus NF-κB or CD28RE luciferase (Luc) reporters in the absence or presence of one of the following constructs: wt-PKCθ, DA-PKCθ, or DN-PKCθ. In separate experiments, CD28RE Luc reporter was coexpressed with DA-PKCθ and either DN-IKKβ or DN-IKKα. Cells were rested for 24 h and then stimulated for 6 h with anti-CD3 + CD28 mAb or P + I. Luciferase assays were performed as previously described (8). Transfection efficiency was monitored by cotransfection of a β-galactosidase plasmid (CMV-β-gal).

Jurkat T cells were transfected with 10 μg of Flag-IKKβ plus 10 μg of wt PKCθ by electroporation as described. (35). Cells were rested for 24 h and stimulated for 1 h with anti-CD3 + CD28 mAb coupled to magnetic beads, at a ratio of 4 beads/cell (34). Cells were vigorously resuspended and the magnetic beads were removed. In a separate experiment, Jurkat T cells transfected with Flag-IKKβ and DA-PKCθ were used without any stimulations. Cells spun onto a glass slide were fixed in 4% glutaraldehyde for 30 min and permeabilized in 70% methanol for 1 h. Slides were overlaid with 5 μg/ml goat anti-PKCθ for 2 h. After washing in PBS (three times), slides were treated with 5 μg/ml of PE-conjugated anti-goat plus 5 μg/ml FITC-conjugated anti-Flag for 2 h. Appropriate controls consisted of omitting the primary Abs to rule out nonspecific binding of the secondary reagent. Samples were washed extensively and examined under a LEICA inverted TCS-SP confocal microscope, using a ×100 objective lens.

We have previously shown that CD28 costimulation is essential for the activation of heterogeneous IKK complexes, as well as activation of the IL-2 promoter (8). We focused on the role of PKCθ, because this is the only PKC isoform that is recruited to the TCR (29, 30). Moreover, it was recently demonstrated by the cellular transfection approach that PKCθ plays a role in the activation of the IKKs in Jurkat cells (26, 27). To demonstrate the relevance of these findings in NF-κB activation in primary human CD4⁺ T cells, we examined the effect of PKC inhibitors on IKK activation during CD28 costimulation (Fig. 1). As previously shown, CD3/CD28 coligation but not anti-CD3 alone, induced IKK activation (Fig. 1, lanes 1–3). Rottlerin, a Ca²⁺-independent PKC isoform inhibitor, interfered in IKK activation (lanes 4–6), whereas Gö6976, a Ca²⁺-dependent PKC isoform inhibitor, had no effect (lanes 7–9). Similar findings were made in Jurkat cells (data not shown). These data suggest that Ca²⁺-independent PKCs participate in the activation of IKK complexes by the TCR. These data do not rule out the role of other PKC isoforms in NF-κB pathway activation, e.g., stimulation by cytokines.

Because PKCθ has been implicated in CD28 costimulatory events (29, 32), we examined whether this PKC isoform physically associates with the IKK signalsome. PKCθ immune precipitates were analyzed for the presence of IKK subunits by Western blotting. Compared with resting CD4⁺ T cells, which did not show any association between these proteins, a small amount of IKKβ coprecipitated with PKCθ in anti-CD3-treated cells (Fig. 2,A). The amount of IKKβ coprecipitaiton was dramatically increased by CD28 costimulation (lane 4), which is in agreement with the key role of this receptor in IKK activation (Fig. 1). To show the specificity of PKCθ in this costimulatory event, we also performed an analysis in which we compared PKCθ with PKCα immune complexes. As shown in Fig. 2,B, both IKKα and IKKβ coprecipitated with PKCθ in anti-CD3 + CD28 or P + I-stimulated CD4⁺ T cells. In contrast, PKCα immune complexes did not contain any IKKs under resting or stimulated conditions (Fig. 2,A), showing a specific role for PKCθ in CD3/CD28-induced NF-κB activation (26, 27). In a reverse protocol, anti-IKKα immune complexes could be seen to coprecipitate with PKCθ in CD3 + CD28 or P + I-treated cells (Fig. 2,C). Parallel in vitro kinase assays showed that PKCθ was associated with active IKK complexes capable of phosphorylating the substrate, IκBα (Fig. 2,D). In contrast, anti-PKCα precipitates were not associated with active IKK complexes (Fig. 2 D). Taken together, these data show that PKCθ associates with a complex that also includes activated IKKs in CD3/CD28-stimulated T cells. Although we do not know whether this is a direct interaction or an indirect association via intermediary components in a macromolecular complex, other workers have failed to show a direct interaction of PKCθ with IKKs (27).

In an extension of these studies in Jurkat cells, we asked whether coexpression of wt, DA-, or DN-PKCθ with epitope-tagged IKKs could induce IKK activity. Although DA-PKCθ induced IKKβ activity, no activation was obtained with wt or DN-PKCθ (Fig. 3, top). HA-IKKβ was equally expressed in these samples (Fig. 3, middle). Blotting for PKCθ confirmed overexpression of wt, DA-, and DN-PKCθ (Fig. 3, bottom). In contrast, a parallel experiment using Flag-IKKα cotransfection failed to show an effect of PKCθ on IKKα activity (data not shown). This is in agreement with the findings of Lin et al. (27) who showed that DA-PKCθ induced the phosphorylation of coexpressed IKKβ but not IKKα. Taken together, the above data show that PKCθ plays a key role in IKK activation in primary human CD4⁺ lymphocytes as well as in overexpressing Jurkat cells.

A biologically relevant target for the CD28 accessory receptor is the IL-2 gene, in particular the CD28RE in the promoter of that gene (7, 13). We have recently shown that IKKβ, but not IKKα, contributes to the activation of that RE during CD3/CD28 costimulation (7, 8). The importance of PKCθ in the activation of that RE was confirmed by the use of wt, DA-, or DN-PKCθ, which were equally expressed in Jurkat cells together with a CD28RE-Luc reporter (Fig. 4,A). Although the response of this reporter was minimally affected by wt-PKCθ, DA-PKCθ itself was sufficient in CD28RE activation (Fig. 4,B). This activity was further enhanced by CD3 + CD28 or P + I costimulation (Fig. 4,B). In contrast, DN-PKCθ interfered in CD28RE activation (Fig. 4,B). Similar inhibitory effects on CD28RE activation were obtained with rottlerin, but not Gö6976 (Fig. 4,C). To demonstrate that the IKKs function downstream of PKCθ in the activation of that RE, we used DN-IKKα and DN IKKβ cotransfection with DA-PKCθ. Although DN-IKKβ interfered in the DA-PKCθ-induced response, DN-IKKα had minimal effects on CD28RE-Luc activity (Fig. 4,D). These findings are in agreement with the ability of DA-PKCθ to induce IKKβ, but not IKKα, activation under cotransfection conditions (Fig. 3).

Because PKCθ is recruited to the signal assembly complex at the TCR (29, 30), it raised the question whether IKKs and their associated components are also recruited to that membrane site. We have previously shown that TCR ligation leads to the translocation of several signaling components, including the ζ-chain of the TCR, cytoskeletal components, p56^lck, p36(LAT), PLC-γ₁, and ZAP-70 to DSM and DIM (36). Interestingly, the same signaling molecules are recruited to the SMAC, where PKCθ localizes in the center cluster, whereas talin appears in the peripheral zone (29, 30). Talin is a member of the cytoskeletal complex that assembles at the TCR synapse (37). To determine whether PKCθ and talin are corecruited to the DIM, we fractionated primary human CD4⁺ T cells into cytosolic DSM and DIM components. Western blot analysis showed talin relocation from the cytosol to the DIM during CD3 + CD28 or P + I, and, to a lesser extent, anti-CD3 stimulation (Fig. 5,A). Although endogenous PKCθ underwent similar redistribution to the DIM during CD3 + CD28 or P + I stimulation (Fig. 3), this kinase also relocated to the DSM under these stimulatory conditions (Fig. 5,A). In a related set of experiments, we also asked whether IKK components relocate to these cellular fractions. IKKα, IKKβ, and IKKγ were recruited to the DIM, and to a variable degree the DSM, during cellular stimulation with anti-CD3 + CD28 or P + I (Fig. 5 B).

Recently, it has been shown that lipid rafts are important in controlling the assembly of signaling molecules at the TCR synapse. Moreover, co-cross-linking of the CD28 receptor enhances the clustering of lipid rafts at the TCR synapse, thereby enhancing TCR signaling (1, 5, 38). Lipid rafts are defined by their insolubility in cold, nonionic detergents, unlike the bulk of the cell membrane. Because this suggests that DIM may overlap with lipid rafts, we used sucrose density fractionation to prepare raft fractions from unstimulated and anti-CD3 + CD28-treated Jurkat cells. Using the cholera toxin β-subunit to identify the fractions containing GM1 ganglioside, we showed that lipid rafts were mostly present in the hypodense layers (fractions 3–5) from resting as well as stimulated T cells (Fig. 6,A). Immunoblotting showed PKCθ and talin recruitment to the hypodense fractions during cellular stimulation (Fig. 6,B). Although PKCθ and talin did not appear in the low density fractions from unstimulated cells, these fractions did show the constitutive presence of Lck, which was further increased by CD3 + CD28 costimulation (Fig. 6,B). Moreover, Western blotting for IKKβ and IKKα showed their relocation to lipid rafts in CD3 + CD28-treated cells (Fig. 6 B). Taken together, these data indicate that IKK components are recruited to lipid rafts in parallel with PKCθ and talin.

Although for logistic reasons the entire experiment in Fig. 6,B could not be repeated in primary CD4⁺ T cells, we did perform sucrose gradient fractionation on CD3 + CD28-stimulated primary T cells. As for Jurkat cells, the lipid rafts were identified by cholera toxin staining and shown to reside in fractions 3–5. Immunoblotting showed the presence of IKKs in these low density fractions, in addition to their presence in the DSM (fractions 10 and 11; Fig. 6,C). Moreover, immune precipitation with anti-IKKγ and performance of in vitro kinase assays showed the presence of activated IKK complexes in fractions 2–4 as well as 10 and 11 (Fig. 6 C, lower panel). These findings indicate that lipid rafts from CD3/CD28 costimulated T cells contain activated IKK complexes. Similar results were obtained in Jurkat T cells.

To confirm the recruitment of PKCθ and IKKs to the membranes, we used confocal microscopy. Exogenously transfected wt-PKCθ and Flag-IKKβ were expressed in a diffuse manner in unstimulated cells (Fig. 7). However, upon stimulation with anti-CD3 + CD28-conjugated beads, both PKCθ and IKKβ were redistributed to membrane cap sites (Fig. 7). Overlay of the green (IKKβ) and red (PKCθ) images yielded a composite yellow fluorescent pattern, which indicates colocalization of these proteins in the membrane cap sites of stimulated cells (Fig. 7). When cotransfected with DA-PKCθ, Flag-IKKβ colocalized in the membrane in a diffuse distribution without the need for cellular stimulation (Fig. 7, right panel). These data suggest that PKCθ activity is involved in the membrane recruitment of the IKK signalsome, which is also consistent with coimmunoprecipitation studies shown in Fig. 2. In accordance with this notion, DN-PKCθ resides in the cytosol of unstimulated cells, similar to wt-PKCθ (data not shown). Although endogenous PKCθ undergoes a similar redistribution as exogenously expressed PKCθ during CD28 costimulation (data not shown), it was not possible with the available anti-IKK Abs to obtain a strong enough signal to study endogenous IKK redistribution. This was, however, achieved by the biochemical and coimmunoprecipitation studies shown in Figs. 2 and 6.

In this paper, we have demonstrated that PKCθ participates in CD3/CD28-induced activation of IKKs in primary human CD4⁺ lymphocytes. DN-PKCθ and a Ca²⁺-independent PKC inhibitor interfered in the activation of the IKK complexes as well as the CD28RE of the IL-2 promoter. Moreover, IKKs and PKCθ were recruited from the cytosol to a talin-enriched, detergent-resistant cellular compartment during CD28 costimulation. Similarly, PKCθ, talin, and IKKs were recruited to GM1-enriched lipid rafts, which represent a detergent resistant membrane compartment. IKKs recruited to the lipid rafts were active. It was also possible to show that the activated IKK signalsome coimmunoprecipitate with PKCθ from CD3/CD28 costimulated primary CD4⁺ lymphocytes. Moreover, active PKCθ and IKKβ were recruited to and colocalized in the membrane of stimulated Jurkat T cells as determined by confocal microscopy. It was also demonstrated that exogenously expressed DA-PKCθ activated IKKβ, but not IKKα, in Jurkat cells. Similarly, DN-IKKβ, but not DN-IKKα, interfered in the activation of the CD28RE by DA-PKCθ. Taken together, these findings demonstrate stimulus-dependent interaction between PKCθ and IKK components, and present a novel mechanism to explain synergistic activation of IKKs by TCR and CD28.

Although some components of the IKK signalsome have been identified, its overall composition and regulation have not been fully characterized (17, 18, 19, 20, 21, 22, 39). Indeed, the evolving scenario points to heterogeneous IKK complexes responding to different stimuli (24, 39, 40), including TCR/CD28 costimulation in primary T cells (8). Stimulus-dependent assembly and activation of IKK complexes is an attractive model, considering the multitude of stimuli involved in NF-κB activation (24, 40). Four different serine-threonine kinases have now been implicated in the activation of the IKKs by CD3/CD28, namely MEKK1, NF-κB-inducing kinase, Cot, and PKCθ (7, 25, 26, 27, 28). Although our overexpression studies in Jurkat cells confirm the recently published data about the role of PKCθ in IKK activation (Fig. 3) (26, 27), the current communication extends those findings by showing the recruitment and colocalization of PKCθ and IKKs in the membrane of T cells ( Figs. 5–7). Moreover, we demonstrate that PKCθ and IKKs are physically associated with a macromolecular complex that can be reversibly precipitated with Abs to PKCθ or IKKs from lysates prepared from CD3 + CD28-stimulated cells (Fig. 2). Coudronniere et al. (26) recently reported that rottlerin, an inhibitor of novel PKC isoforms, interfered in the activation of the NF-κB pathway and the CD28RE by the TCR. We show here that treatment of primary human CD4⁺ with rottlerin inhibits the CD3/CD28-mediated activation of endogenous IKK complexes (Fig. 1). Gö6976, an inhibitor of classical PKC isoforms, had no effects on IKK activity and transcriptional activation of the CD28RE (Figs. 1 and 4). Although the role of other PKC isoforms in the activation of the NF-κB pathway has been demonstrated, our findings implicate PKCθ in IKK activation and recruitment during CD3/CD28 costimulation T cells. We found that endogenous PKCθ physically associated with the IKK signalsome in the cell membrane (Figs. 2 and 7). Although this interaction was minimal affected by of CD3 ligation, the association was markedly augmented by CD28 costimulation (Fig. 2 A). Although the importance of PKCθ in IKK activation was recently reported (26, 27), this is the first evidence for CD3/CD28-induced association of PKCθ with the IKK complex in a physiologically relevant system. Although a previous study could not demonstrate that PKCθ directly interacts with IKK components (27), it is possible that this interaction involve additional components of a macromolecular complex, e.g., cytoskeletal or scaffolding proteins. This aspect requires further study.

We propose that PKCθ/IKK association takes place in the immunological synapse. PKCθ along with IKKα, IKKβ, and IKKγ were recruited to the membrane fractions and lipid rafts in primary human CD4⁺ lymphocytes (Figs. 5 and 6). Lipid rafts, which are cholesterol- and sphingolipid-enriched membrane domains, are docking sites for the attachment of costimulatory receptors and signaling molecules in the cell membrane (1, 29, 30, 38). Formation of the TCR synapse and the assembly of productive signaling complexes in the lipid rafts require reorganization of the cortical cytoskeleton (30, 35), and may explain our finding that the IKKs and PKCθ are recruited to a detergent-insoluble cellular compartment that includes talin and spectrin (Figs. 5 and 6). CD28 participates in the organization of lipid rafts and recruitment of the associated signaling molecules to the TCR synapse (1, 5, 38). Recruitment of the IKK components to the lipid rafts represents a novel aspect of CD28 function, which may also be responsible for activation of the IKK complex (Fig. 6,C) and the NF-κB pathway. The mechanism of CD28-induced recruitment of signaling molecules to the lipid rafts is poorly understood. However, recent studies have demonstrated that CD28 contributes to the rearrangement of the cytoskeletal complex at the TCR contact site via the activation of Vav (41). Vav has been shown to participate in the recruitment of PKCθ to the T cell membrane and the cytoskeleton during TCR engagement (41) We hypothesize that this process is essential for the recruitment of IKKs to the lipid rafts. Although ectopically expressed wt PKCθ and IKKβ colocalized in the membrane after CD3/CD28 costimulation, DA-PKCθ and IKKβ spontaneously localized in the membrane in the absence of any stimuli (Fig. 7). This suggests that, in addition to its role in IKK activation (26, 27), PKCθ may also play a role in the recruitment of IKKs to the T cell membrane. The idea that the IKKs are activated in the vicinity of the TCR is further strengthened by the demonstration that TNF-α induces the IKK holoenzyme to associate with components of the p55 TNF receptor site in the cytoplasmic membrane (42).

In conclusion, we have shown that, during CD28 costimulation, PKCθ-induced IKK activation in primary human CD4⁺ T cells involves their physical interaction with an activation complex that localizes in the T cell membrane. This may represent an early event in the activation of the NF-κB cascade during CD3/CD28 costimulation. The signaling molecules that attract IKKs to the lipid rafts remain to be identified. However, because DA-PKCθ facilitated the relocation of IKKβ to the membranes (Fig. 7), it is tempting to speculate that PKCθ plays an active role in this process. PKCθ activity has also been linked to JNK activation, and this pathway synergizes with the NF-κB cascade in the activation of the IL-2 promoter (7, 31, 32, 33). In this regard, PKCθ may act as a master switch responsible for the regulation of key signaling pathways involved in the delivery of signal two.

We thank Ivonne Castaneda and Peggy Shih for secretarial assistance and Martin Kasubowski for technical support.