Protein Kinase C and Calcineurin Synergize to Activate IκB Kinase and NF-κB in T Lymphocytes*

The nuclear factor of κB (NF-κB) is a ubiquitous transcription factor that is key in the regulation of the immune response and inflammation. T cell receptor (TCR) cross-linking is in part required for activation of NF-κB, which is dependent on the phosphorylation and degradation of IκBα. By using Jurkat and primary human T lymphocytes, we demonstrate that the simultaneous activation of two second messengers of the TCR-initiated signal transduction, protein kinase C (PKC) and calcineurin, results in the synergistic activation of the IκBα kinase (IKK) complex but not of another putative IκBα kinase, p90 rsk . We also demonstrate that the IKK complex, but not p90 rsk , is responsible for thein vivo phosphorylation of IκBα mediated by the co-activation of PKC and calcineurin. Each second messenger is necessary, as inhibition of either one reverses the activation of the IKK complex and IκBα phosphorylation in vivo. Overexpression of dominant negative forms of IKKα and -β demonstrates that only IKKβ is the target for PKC and calcineurin. These results indicate that within the TCR/CD3 signal transduction pathway both PKC and calcineurin are required for the effective activation of the IKK complex and NF-κB in T lymphocytes.

The nuclear factor of B (NF-B) is a ubiquitous transcription factor that is key in the regulation of the immune response and inflammation. T cell receptor (TCR) cross-linking is in part required for activation of NF-B, which is dependent on the phosphorylation and degradation of IB␣. By using Jurkat and primary human T lymphocytes, we demonstrate that the simultaneous activation of two second messengers of the TCR-initiated signal transduction, protein kinase C (PKC) and calcineurin, results in the synergistic activation of the IB␣ kinase (IKK) complex but not of another putative IB␣ kinase, p90 rsk . We also demonstrate that the IKK complex, but not p90 rsk , is responsible for the in vivo phosphorylation of IB␣ mediated by the co-activation of PKC and calcineurin. Each second messenger is necessary, as inhibition of either one reverses the activation of the IKK complex and IB␣ phosphorylation in vivo.

Overexpression of dominant negative forms of IKK␣ and -␤ demonstrates that only IKK␤ is the target for PKC and calcineurin. These results indicate that within the TCR/CD3 signal transduction pathway both PKC and calcineurin are required for the effective activation of the IKK complex and NF-B in T lymphocytes.
Identification of the molecular events that ensue in T lymphocytes following antigen presentation is paramount to understanding the regulation of the immune response. The signal transduction pathways triggered by antigen presentation lead to the immediate activation of transcription factors that further amplify the process of lymphocyte activation, ultimately leading to cell proliferation and division. Dysregulation of this process results in T lymphocyte anergy, autoimmunity, and disruption of T lymphocyte homeostasis (1,2). Two additional settings that would benefit from a better understanding of the molecular events triggered by T cell activation are viral pathogenesis and drug discovery. For example, HIV 1 integrates in the chromosome of T lymphocytes where it remains latent. Its reactivation by transcription factors that are activated following T cell receptor cross-linking is relevant to the pathogenesis of AIDS (3,4). Finally, development of improved immunosuppressive agents that target T cell function will be accelerated by identifying the exact molecular events that result from the molecular events regulating T cell activation.
Effective antigen presentation to T lymphocytes involves not only the engagement of the T cell receptor-CD3 complex but also other receptors that mediate co-activation signals such as CD28. How these two separate signal transduction pathways (TCR/CD3 and CD28) converge to result in the maximal activation of the T lymphocyte is under active study. Engagement of the T cell receptor (TCR) by its cognate peptide-major histocompatibility complex induces phospholipase C activation which hydrolyzes phosphatidylinositol 4,5-bisphosphate to inositol 1,4,5-triphosphate and diacylglycerol (5). Diacylglycerol activates protein kinases C (PKC), whereas inositol 1,4,5triphosphate leads to the release of Ca 2ϩ from intracellular stores (6,7). The same cellular events initiated by phospholipase C activation can be mimicked by treatment with a combination of a Ca 2ϩ ionophore that raises intracellular Ca 2ϩ levels and phorbol esters that activate PKC isoforms (8). Free intracellular Ca 2ϩ targets the Ca 2ϩ /calmodulin-activated phosphatase calcineurin which mediates a critical positive signal necessary for IL-2 induction through its synergy with PKC activation, synergy that can be reversed by targeting calcineurin with immunosuppressive drugs such as cyclosporin A or FK506 (9,10). The initial studies that investigated the targets of the synergy between PKC and calcineurin identified the transcription factors, nuclear factor of activated T cells and nuclear factor of B (NF-B) (9,10,11), as being activated by the combination of these separate signal transduction pathways.
NF-B is a heterodimer of transcription factors that belong to the Rel family of proteins. The canonical NF-B is a heterodimer of p65 (RelA) with p50 or p52 (12,13,14). This heterodimer is anchored by a group of proteins named IB, which function to retain NF-B in the cytosol by masking its nuclear localization signal (15)(16)(17)(18). IB␣ is a prototype IB molecule known to control the subcellular localization of NF-B (p50/p65). Following activation of certain signal transduction pathways, a site-specific hyperphosphorylation of IB␣ at Ser-32 and Ser-36 renders the inhibitor molecule susceptible to site-specific ubiquitination and subsequent degradation by the proteasome complex (19 -23). This releases NF-B to undergo nuclear translocation. Two novel IB␣ kinases, IKK␣ (24 -26) and IKK␤ (27,28), contained within a high molecular weight complex termed the signalsome target the phosphorylation of Ser-32 and Ser-36 of IB␣ and mediate the TNF-␣-induced IB␣ hyperphosphorylation and NF-B activation. An additional N-terminal IB␣ kinase, the mitogen-activated ribosomal S6 protein kinase RSK-1 or p90 rsk (29,30), has been shown to mediate the activation of NF-B by phorbol esters and to phosphorylate IB␣ preferentially at Ser-32 both in vivo and in vitro. However, its functional relevance in vivo is yet unclear (30).
Previous studies addressing how the PKC-and calcineurindependent signal transduction pathways triggered by TCR/ CD3 cross-linking lead to the synergistic activation of NF-B identified IB␣ as a target molecule (11). Whereas PKC activation by phorbol myristate acetate (PMA) resulted in a moderate degree of IB␣ hyperphosphorylation-degradation and activated calcineurin alone had no effect on IB␣, the combined activation of these two second messengers leads to a synergistic and highly effective hyperphosphorylation and degradation of IB␣ (31). Moreover, whereas TCR and TNFR triggered separate signal transduction pathways, both ultimately target IB␣, and the use of specific inhibitors of calcineurin and PKC enabled the separation of the signaling pathways of TCR and TNFR (31). Despite this novel observation, the mechanisms whereby two separate second messengers, PKC and calcineurin, lead to the lymphocyte-specific hyperphosphorylation of IB␣ remains unknown.
The recent identification of the N-terminal IB␣ provides the opportunity to advance in our understanding of the molecular mechanisms, whereas signal transduction pathways triggered by TCR/CD3 cross-linking, separate from those downstream of CD28, lead to IB␣ degradation and NF-B activation in T lymphocytes. In this study we have investigated the synergy between two second messengers of the TCR/CD3 pathways, PKC and calcineurin, as regulators of the N-terminal IB␣ kinases. By using primary human T lymphocytes and T lymphocyte cell lines, we demonstrate that PKC-and Ca 2ϩ -dependent pathways synergistically activate both the IKK complex and p90 rsk . In contrast to the activation of the IKK complex, p90 rsk activation is calcineurin-independent, and only the IKK complex (IKK␤) but not p90 rsk mediates IB␣ phosphorylation and NF-B activation in vivo. Moreover, the synergistic activation of the IKK signalsome by PKC and calcineurin is inhibited by either calcineurin-or PKC-specific inhibitors suggesting that either signal transduction pathway is required and essential for effectively activating the IKK complex and, hence, NF-B activation in T lymphocytes.

EXPERIMENTAL PROCEDURES
Plasmids-The B-luc reporter plasmid consists of three B concatamers from the HIV-long terminal repeat cloned upstream of a concanavalin-A minimal promoter driving the expression of luciferase (32). The pREP4/CAT plasmid, which consists of the Rous sarcoma virus promoter-enhancer driving the transcription of the CAT gene, was used to control for transfection efficiency (Invitrogen, Carlsbad, CA). IKK␤ KD (K44A) was obtained from M. Roth (Tularik, South San Francisco, CA). IKK␣ kinase dead (KD) (D144N) was a kind gift from Alain Israel, Institute Pasteur, Paris, France. Wild-type PKC␣ cDNA was kindly provided by Dr. Altman, La Jolla, CA. The expression vector, pSR␣4⌬CaM-AI, encoding a constitutively active calcineurin catalytic subunit is similar to the previously described pSR␣-⌬CaM-AI (9) except that it contains an additional 75 base pairs of 5Ј-untranslated sequence from CN4a (34).
To isolate CD3 ϩ T cells, peripheral blood mononuclear cells from healthy donors were obtained from buffy coats by density gradient centrifugation (Ficoll-Paque, Amersham Pharmacia Biotech). Peripheral blood mononuclear cells were then depleted of monocytes by two cycles of plastic adherence, and CD3 ϩ T cells were purified by neuraminidase-treated sheep red blood cell rosetting. The remaining cell population was repeatedly found to be 98% CD3 ϩ T cells as determined by flow cytometry analysis. CD3 ϩ T cells used in the various experiments were maintained in RPMI 1640 supplemented with 10% fetal bovine serum (Invitrogen, Carlsbad, CA), 2 mM L-glutamine, and antibiotics (penicillin 100 units/ml, streptomycin 100 g/ml) at 0.5 ϫ 10 6 cells/ml. CD3 ϩ T cells were stimulated and harvested on the 2nd day after isolation.
Where indicated, cells were pretreated with 2 M GF109203X and Gö 6976 for 15 min, 30 M PD 098059 for 20 min. FK506 was used at 20 ng/ml. For Jurkat T cells, PMA was used at 20 ng/ml, ionomycin at 3.5 g/ml, and TNF-␣ at 10 ng/ml. For CD3 ϩ T cell activation, PMA was used at 2.5 ng/ml, ionomycin at 0.7 g/ml, and TNF-␣ at 10 ng/ml. TCR/CD3 cross-linking was performed with 3 g/ml anti-CD3 antibody as previously demonstrated by our group (34).
For Western immunoblots, equal amounts of whole cell extract (WCE) protein were loaded and separated by 10% SDS-polyacrylamide gel electrophoresis (PAGE) and transferred to Immobilon-P membranes (Millipore, Bedford, MA). Immunoblotting was performed with specific antibodies and visualized by using the ECL Western blotting detection kit (Amersham Pharmacia Biotech, Buckinghamshire, UK).
For the immunocomplex kinase assay, 100 g of Jurkat WCE and 50 g of CD3 ϩ T cell WCE were rotated with specific antibodies for 1 h and then for an additional 1 h more with protein A-agarose beads (Life Technologies, Inc.) at 4°C. The immunoprecipitations were performed in WCE buffer with high NaCl concentrations (0.5 M). The beads were washed 3 times with 0.5 M NaCl WCE buffer followed by 1 wash of kinase assay washing buffer (50 mM Tris-HCl, pH 7.4, 40 mM NaCl). The beads were mixed with 15 l of kinase buffer (33) (20 mM Hepes, pH 7.4, 2 mM magnesium chloride, 2 mM manganese chloride, 10 M ATP, 10 mM NaF, 10 mM p-nitrophenyl phosphate, 10 mM ␤-glycerophosphate, 300 M sodium orthovanadate, 2 M phenylmethylsulfonyl fluoride, aprotinin at 10 g/ml, leupeptin at 1 g/ml, pepstatin at 1 g/ml, 1 mM dithiothreitol) containing 2 g of GST-IB␣-(1-53) substrate and 1 Ci of [␥-32 P]ATP. The 30-min kinase reaction at 30°C was stopped by adding 4ϫ SDS-PAGE sample buffer. The proteins were separated by SDS-PAGE and transferred to Immobilon-P membrane. The top part of the membrane was used for immunoblots of IKK␣ or Rsk-1; on the bottom part of the membrane, the amount of GST-IB␣-(1-53) and the levels of its phosphorylation were visualized by staining with Coomassie Blue and autoradiography, respectively. Preparation of Recombinant IB␣-The IB␣ MAD3 cDNA (35) plasmid was obtained from Chiron and used as a template for subsequent polymerase chain reaction amplification. The N-terminal IB␣ MAD3 (1-53) sequence was amplified using wild-type primer A (CGGGATC-CATGTTCCAGGCGGCCGAG), as the 5Ј sense primer, creating a BamHI site upstream of the coding sequence, and wild-type primer B (GGAATTCCTCAGCGGATCTCCTGCAGCT) as antisense primer, creating an EcoRI site downstream of the coding sequence. A double S32A/S36A mutant was amplified from the full-length cDNA using polymerase chain reaction primers which created alanines at amino acids 32 and 36. Following digestion with BamHI and EcoRI, these sequences were ligated into pGEX-KG (derived from pGEX-2T from Amersham Pharmacia Biotech). These constructs were transformed into Escherichia coli DH5␣ cells, which were grown exponentially, and after 60 min of stimulation with isopropylthiogalactopyranoside (Sigma) cells were lysed. Proteins were isolated by affinity chromatography on glutathione-bonded 4% cross-linked agarose (Sigma). The purity of GST-IB␣-(1-53) and GST-IB␣-(1-53) 32A/36A was analyzed with 10% SDS-PAGE and subsequent Coomassie Blue staining. The purity of both proteins was greater than 90%.
Gene Transfection and Reporter Assays-FuGENE6 was used to express plasmids transiently in Jurkat T cells. In brief, 8 l of FuGENE6 (Roche Molecular Biochemicals) were mixed with 92 l of plain RPMI 1640 media and incubated for 5 min. FuGENE6/RPMI 1640 solution was added to sterile tube containing 0.4 g of B-luc reporter plasmid, 0.6 g of pREP4/CAT, and 0 -1 g of a plasmid of interest (total is 2 g) and incubated for 15 min. The DNA/sFuGENE6 solution was added to 1 ϫ 10 6 log phase Jurkat T cells.
Jurkat cells were transfected with the indicated plasmids and grown for 40 h. Cells were stimulated for 4 h with PMA (20 ng/ml), ionomycin (3.5 g/ml), TNF-␣ (10 ng/ml), or PMA and ionomycin together. After stimulation, cells were washed twice in cold phosphate-buffered saline and lysed with 210 l lysis solution (100 mM K 2 PO 4 , pH 7.8; 0.2% Triton X-100; 5 mM dithiothreitol, 2 g/ml aprotinin). Equal amounts (100 l) of extract were assayed for luciferase and CAT expression. CAT expression was determined by the Roche Molecular Biochemicals CAT en-zyme-linked immunosorbent assay kit using the manufacturer's protocol. Luciferase activity was assayed using the Promega Luciferin reagent and a Berthold Lumat. Luciferase activity is normalized to CAT expression. All transfection experiments were performed in duplicate.

PKC-and Ca 2ϩ -dependent Pathways Synergize to Activate the IKK Complex in T Cells-Cross-linking of the TCR/CD3
results in the activation of PKC-and Ca 2ϩ -dependent pathways that synergize in T cells to activate NF-B by targeting the phosphorylation and degradation of its inhibitor, IB␣ (11,31). To determine whether TCR cross-linking can lead to IKK activation, we measured the IB␣ kinase activity of the IKK complex immunoprecipitated from Jurkat T cells that were activated or not following TCR cross-linking. Cross-linking of the TCR with anti-CD3 but not IgG antibodies results in a FIG. 1. PKC-and Ca 2؉ -dependent pathways are required for the IKK complex activation. A, the activation of the IKK complex following TCR/CD3 cross-linking is calcineurin-dependent. Jurkat T cells (2 ϫ 10 6 cells per sample) were incubated with 3 g/ml anti-CD3 (lanes 2 and 3) or isotype control IgG (lane 1) antibody for 45 min at 4°C in 1 ml of media. Cells were then cross-linked on goat anti-mouse-coated plates for 20 min at 37°C. Jurkat T cells were pretreated with 20 ng/ml FK506 before incubation with anti-CD3 antibodies (lane 3). IKK activity was measured in an in vitro kinase assay (IVK) as described under "Experimental Procedures," using GST-IB␣ (1-53 amino acids) as substrate (IB␣ 32 P). Coomassie staining of the polyvinylidene difluoride membrane containing the IVK (IB␣) and immunoblotting (IB) for IKK␣ (IKK␣ IB) were performed. B, PMA and ionomycin synergize to activate IKK in Jurkat and CD3 ϩ T cells. Primary CD3 ϩ (lanes 1-5) and Jurkat T cells (lanes 6 -15) were stimulated (ϩ) or not (Ϫ) for 8 min with ionomycin (IONO), PMA, or TNF. For CD3 ϩ T cell activation, PMA was used at 2.5 ng/ml, ionomycin at 0.7 g/ml for 8 min, and TNF␣ at 10 ng/ml for 4 min; for Jurkat T cell stimulation PMA was used at 20 ng/ml, ionomycin at 3.5 g/ml for 8 min, and TNF-␣ at 10 ng/ml for 8 min. IKK activity was measured in an IVK using GST-IB␣-(1-53) as substrate (IB␣ 32 P). The specificity of IKK kinase activity toward to Ser-32/Ser-36 was demonstrated by using GST-IB␣ (1-53 amino acids) with substituted serines for alanines (IB␣ 32A/36A , lanes 11-15). In parallel, the level of endogenous IB␣ from the same cell extracts were detected by immunoblotting with anti-IB␣ antibodies (IB␣ IB in vivo). C, Jurkat T cells were either untreated (lanes 1-5) or pretreated with 4 g/ml neutralizing anti-TNF antibody (lanes 6 -8) or with 4 g/ml isotype control IgG antibody (lanes 9 -11) 1 h prior to stimulation. Efficiency of the neutralizing anti-TNF antibody was demonstrated by using the mixture of recombinant TNF (10 ng/ml) with neutralizing anti-TNF antibody (4 g/ml) incubated for 1 h at 4°C before stimulation (lane 5). IKK activity was measured in IVK as described above. Equal amounts of the substrate and the immunoprecipitated kinase complex were present in the assay confirmed by Coomassie staining of the polyvinylidene difluoride membrane containing the IVK (IB␣), and immunoblotting (IB) for IKK␣ (IKK␣ IB), respectively. moderate activation of the IKK complex (Fig. 1A, 2nd lane), which is inhibited by pretreatment of Jurkat T cells with the calcineurin-specific inhibitor FK506 (Fig. 1A, lane 3). From these results, and based on previous studies from our group (11,31), we conclude that calcineurin participates in the TCR/ CD3-initiated signal transduction pathway that leads to IKK activation.
To investigate the role of PKC and calcineurin in the activation of the IKK complex, an in vitro kinase assay using IKK complex immunoprecipitates from resting freshly isolated peripheral human T lymphocytes and Jurkat T cells that were or were not stimulated with ionomycin, PMA, or their combination. Ionomycin stimulation of CD3 ϩ and Jurkat T cells does not affect the IKK complex activity (Fig. 1B, lanes 2 and 7), whereas PMA induces a moderate IB␣ kinase activity (Fig.  1B, lanes 3 and 8). However, stimulation of CD3 ϩ and Jurkat T cells with the combination of PMA and ionomycin increased the IKK complex kinase activity beyond that observed in PMAtreated cells (Fig. 1B, lanes 4 and 9). When TNF treatment was used as a control of IKK activation (24), it was observed that the degree of TNF activation was similar to that of PMA but significantly lower than that achieved by the combination of PMA and ionomycin (Fig. 1B, lanes 5 and 10). The IKK complex kinase activity was specific for Ser-32/Ser-36, as it was not observed when an IB␣ substrate in which both serines were substituted with alanines was used in the in vitro kinase assay (Fig. 1B, lanes 11-15). Equal amount of IB␣ substrate present in the in vitro kinase assay (Fig. 1B, IB␣) or the amount of IKK complex immunoprecipitated (Fig. 1B, lanes 1-5) confirmed that the increased in vitro phosphorylation of IB␣ is a function of enhanced IKK complex kinase activity. Similar results were obtained when the IKK complex was immunoprecipitated with additional anti-IKK␣ or IKK␤ antibodies (data not shown). In addition, Raf-1 immunoprecipitates of cell lysates from PMA, PMA and ionomycin, or TNF-stimulated cells did not result in the phosphorylation of IB␣ at Ser-32/Ser-36, despite evidence of Raf-1 activation following PMA treatment (data not shown).
To determine whether the qualitative differences in the IKK activation triggered by PMA or PMA and ionomycin correlated with the in vivo IB␣ hyperphosphorylation, the same cell lysates used for immunoprecipitation of cell kinases were separated by SDS-PAGE and immunoblotted for endogenous IB␣ using anti-IB␣-specific antibodies. Hyperphosphorylation of IB␣, determined by the slower migrating form of IB␣, was mainly observed in the PMA-and ionomycin-treated cells (Fig.  1B, IB␣ IB in vivo, lanes 4 and 9). This observation establishes a direct correlation between the qualitative activation of the IKK complex by PKC-and Ca 2ϩ -dependent pathways and IB␣ hyperphosphorylation in vivo.
To demonstrate that the observed synergistic activation of the IKK complex is the direct effect of PMA and ionomycin co-stimulation, and not due to their secondary induction of TNF, we measured the IB␣ kinase activity of the IKK complex from PMA-and ionomycin-treated Jurkat T cells in the presence or not of neutralizing anti-TNF antibodies (Fig. 1C). Pretreatment of Jurkat T cells with such antibodies abrogates the activation of the IKK complex by recombinant TNF (Fig. 1C,  lanes 4 and 8) but has no effect on the IKK complex activation by the combination of PMA and ionomycin (Fig. 1C, lanes 2, 3  and 6, 7). The specificity of the neutralizing anti-TNF antibodies was confirmed by their preincubation with recombinant TNF (Fig. 1C, lane 5) and by using IgG isotype antibodies as control (Fig. 1C, lanes 9 -11).
p90 rsk Is Synergistically Activated by PKC-and Ca 2ϩ -dependent Pathways-The mitogen-activated p90 rsk phosphoryl-ates IB␣ at Ser-32 (29,30), and it is questioned whether it directly results in IB␣ phosphorylation and degradation in vivo (20 -23, 36). Because p90 rsk is a second messenger that is activated by PKC (37), we investigated whether Ca 2ϩ -dependent pathways synergistically activated the presumed IB␣ kinase activity of p90 rsk that would be induced following PKC stimulation.
Jurkat T cells and freshly isolated peripheral blood CD3 ϩ T lymphocytes were treated with PMA and/or ionomycin, and p90 rsk and IKK complexes were subsequently immunoprecipitated and analyzed for their ability to phosphorylate IB␣ at Ser-32/Ser-36. In Jurkat T cells, PMA alone significantly induces the p90 rsk IB␣ activity ( Fig. 2A, p90 rsk IP, lane 2). However, the combination of PMA and ionomycin only weakly increased the kinase activity of p90 rsk beyond that induced by the stimulation with PMA alone (Fig. 2A, p90 rsk IP, lane 3) suggesting that at this concentration of PMA, p90 rsk is already fully activated. In primary CD3 ϩ T cells, and in contrast to that observed in Jurkat T cells, strong synergistic activation of p90 rsk toward the IB␣ substrate by the combination of PMA and ionomycin was observed (Fig. 2B, p90 rsk IP, lane 4). Again, stimulation by ionomycin alone does not induce p90 rsk activation (Fig. 2B, p90 rsk IP, lane 2).
Calcineurin Is Required for Synergistic Activation of the IKK Complex but Not p90 rsk in T Lymphocytes-The role of Ca 2ϩdependent signaling and of calcineurin on the activation of the IKK complex and p90 rsk kinase activity was investigated in the presence or absence of the calcineurin-specific inhibitor FK506 (38). Jurkat T cells and freshly isolated T lymphocytes were pretreated with FK506 for 1 h followed by their stimulation with PMA, ionomycin, or their combination. As shown in Fig. 2, FK506 reverses the synergistic activation of the IKK complex by PMA and ionomycin in both Jurkat T cells ( Fig. 2A, IKK IP,  lanes 3 and 5) and in primary CD3 ϩ T cells (Fig. 2B, IKK IP,  lanes 3, 4, and 8, 9) and has no effect on p90 rsk activation ( Fig.  2A, p90 rsk IP, lanes 2-5; Fig. 2B, p90 rsk IP, lanes 3, 4 and 8, 9). The reduction of IKK kinase activity in vitro by FK506 correlates with the decrease of the in vivo IB␣ hyperphosphorylation ( Fig. 2A, in vivo IB␣ IB, lanes 3 and 5; and Fig. 2B, in vivo

FIG. 2. Calcineurin mediates synergistic IKK activation, but not p90 rsk , in T cells. A, Jurkat T cells were pretreated (ϩ) or not (Ϫ)
with 20 ng/ml specific calcineurin inhibitor FK506 1 h before stimulation with PMA and/or ionomycin. Immunoprecipitated p90 rsk (p90 rsk IP) and the IKK complex (IKK␣ IP) were analyzed in IVK. IVK and IB are as described in Fig. 1. B, same as A except that primary CD3 ϩ T cells were used. Lymphocytes were pretreated with 100 ng/ml FK506 1 h before stimulation. 4 and 9). The specificity of FK506 as an inhibitor of the T cell receptor-initiated signaling leading to IKK activation was tested in the TNF-mediated IKK activation and in vivo IB␣ phosphorylation. TNF-induced IKK activation and IB␣-induced hyperphosphorylation in vivo was not affected by the pretreatment of T lymphocytes with FK506 (Fig. 2B, in vivo  IB␣, lanes 5 and 10). Altogether, these results demonstrate a direct correlation between the degree of IKK activation and the in vivo IB␣ hyperphosphorylation. Moreover, they point to the necessary role of calcineurin in mediating the hyperphosphorylation of IB␣ by IKK following the activation of PKC-and Ca 2ϩ -dependent signaling that is triggered by TCR cross-linking.

IB␣ IB, lanes
The IKK Complex and Not p90 rsk Controls Inducible IB␣ Phosphorylation and NF-B Activation in T Cells-To evaluate the in vivo role of p90 rsk as an IB␣ kinase involved in the NF-B activation following T cell activation, we used the specific MEK-1 inhibitor PD 098059 (PD) (39) to block p90 rsk activation (40). In Jurkat T cells a strong activation of p90 rsk IB␣ activity by PMA was reversed by PD (Fig. 3A, IB␣ 32 P, lanes 2 and 5). The combination of PMA and ionomycin only moderately increased the kinase activity of p90 rsk (Fig. 3A, IB␣ 32 P, lane 3). Interestingly, PD did not completely eliminate the PMA-and ionomycin-induced activation of the p90 rsk IB␣ kinase activity, allowing the detection of a degree of synergy when compared with PMA and PD-treated cells (Fig.  3A, IB␣ 32 P, lane 6). This residual effect was completely blocked by PD treatment for a longer time (1 h) (data not shown), suggesting that MEK-1 mediates both PKC-and Ca 2ϩdependent signaling. The specificity of PD as a p90 rsk inhibitor in this model was further evaluated by measuring its inhibitory activity on the IKK complex kinase activity from the same lysates. The synergy between PKC-and Ca 2ϩ -dependent pathways that results in the activation of the IKK complex as an IB␣ kinase (Fig. 3A, lane 11) was minimally modified by PD (Fig. 3A, lane 14), confirming that PD inhibits p90 rsk activation but not the IKK complex.
In primary CD3 ϩ T cells, the IB␣ kinase activity of p90 rsk toward IB␣ substrate was synergistically increased by the combination of PMA and ionomycin (Fig. 3B, IB␣ 32 P, lane 4), an effect that was reversed when cells were treated with PD (Fig. 3B, lane 9). The increased p90 rsk kinase activity was independent of the amount of IB␣ substrate present in the in vitro kinase assay or the amount of p90 rsk immunoprecipitated (Fig. 3B, p90 rsk IP, IB␣, and p90 rsk IB, respectively). As expected, TNF did not increase the IB␣ kinase activity of p90 rsk (Fig. 3B, IB␣ 32 P, lane 5). The specificity of PD was once again verified in primary CD3 ϩ T cells by determining whether the synergistic activation of the IKK complex activity by PMA and ionomycin was PD-insensitive. As shown in Fig. 3B (IKK IP,  lanes 4 and 9), PMA and ionomycin resulted in the synergistic activity of IKK that was not modified by pretreatment of cells   FIG. 3. PD 098059 inhibits PMA/ ionomycin-induced p90 rsk activation but does not affect IB␣ phosphorylation and degradation in vivo. A, Jurkat T cells were pretreated (ϩ) or not (Ϫ) for 20 min with 30 M PD 098059 (PD) and subsequently stimulated (ϩ) or not (Ϫ) with PMA and/or ionomycin. Immunoprecipitates of p90 rsk (p90 rsk IP) or of IKK␣ (IKK IP) were analyzed in the in vitro kinase assay (IVK). IVK and IB are as described in Fig. 1B. B, same as in A except that primary CD3 ϩ T cells were used. C, Jurkat T cells (1 ϫ 10 6 ) were transfected with B-luc-(0.4 g) and REP4/CAT (0.6 g) reporter plasmids. Forty hours later cells were pretreated with 30 M inhibitor PD 098059 for 20 min and then stimulated (ϩ) or not (Ϫ) with PMA and/or ionomycin for 4 h. Equal amounts (100 l) of extracts were assayed for luciferase and CAT expression. Luciferase activity is normalized to CAT expression (relative luciferase activity). All transfection experiments were performed in duplicate.
The specificity of PD as an inhibitor of the in vitro IB␣ kinase activity of p90 rsk but not of the IKK complex allowed us to address the relative role of each IB␣ kinase in mediating the PMA or PMA-and ionomycin-dependent IB␣ hyperphosphorylation and subsequent degradation in primary T cells in vivo. The same cytosolic extracts from purified CD3 ϩ T cells that were used for the p90 rsk and IKK immunoprecipitations experiments were subjected to SDS-PAGE analysis and immunoblotting with anti-IB␣-specific antibodies. PMA and ionomycin combined, but not PMA, or ionomycin alone, induced a slower migration form of IB␣ that was not reversed in the presence of PD (Fig. 3B, in vivo IB␣ IB, lanes 4 and 9).
To test the in vivo relevance of the regulation of IB␣ phosphorylation in vitro and in vivo by these two kinases, Jurkat T cells were transfected with a B-dependent reporter gene, and cells were treated or not with PMA alone and combination of PMA and ionomycin in the presence or absence of PD. As shown in Fig. 3C, the synergistic activation of NF-B transcriptional activity by combination of PMA and ionomycin was minimally sensitive to PD. Altogether, while these results indicate that PMA and ionomycin synergize to activate in vitro the IB␣ kinase activity of both the IKK complex and of p90 rsk , the latter is not regulated by calcineurin and does not appear to play a role in the activation of NF-B in vivo.
Conventional PKC Isoforms Are Required for the Activation of the IKK Complex in T Lymphocytes-Recent studies demonstrated that both the conventional isoform, PKC␣, and the novel isoform, PKC, may be potentially involved in NF-B activation (41,42). PKC␣ associates with and phosphorylates IKK␤ (41) suggesting that the activation of PKC␣ following TCR/CD3 cross-linking could result in the activation of the IKK complex. PKC, a novel PKC isoform was recently identified as the PKC isoform that synergizes with calcineurin leading to JNK activation (42) and that is recruited to the TCR at the antigen-presenting cells docking region during antigen presentation (43).
To characterize the type of PKC isoform that synergizes with calcineurin to activate the IKK complex, a series of pharmacological inhibitors of PKC were utilized (44). GF109203X (GF) inhibits the conventional and novel isoforms, whereas Gö 6976 (Gö) inhibits only the conventional isoforms (45). The specificity of the PKC inhibitors was verified by using TNF as a PKC-independent stimulus that leads to IB␣ hyperphosphorylation through the activation of the IKK complex (46). Jurkat T cells and freshly isolated CD3 ϩ T cells were pretreated with GF followed by cell stimulation with PMA, ionomycin, or the combination of PMA and ionomycin, followed by the analysis of the IB␣ kinase of immunoprecipitated IKK. GF abrogated the IKK complex kinase activation triggered by PMA or the combination of PMA with ionomycin in both Jurkat T cells (Fig. 4A,  IB␣ IVK, lanes 2-4, 6, and 7) and in primary CD3 ϩ T lymphocytes (Fig. 4B, IB␣ IVK, lanes 3, 4, and 8, 9). TNF-induced IKK activation was not reversed by GF pretreatment of Jurkat T cells (Fig. 4A, IB␣ IVK, lanes 4 and 8) or of primary CD3 ϩ T cells (Fig. 4B, IB␣ IVK, lanes 5 and 10). As expected from previous results, the loss of IKK complex kinase activity in  Fig. 1. B, same as A except that CD3 ϩ T cells were used. C, Jurkat T cells (1 ϫ 10 6 ) were transfected with B-luc-(0.4 g) and REP4/CAT (0.6 g) reporter plasmids using FuGENE6 method. Forty hours later, cells were pretreated with 2 M GF or Gö for 15 min and then stimulated (ϩ) or not (Ϫ) with PMA and/or ionomycin for 4 h. Equal amounts (100 l) of extracts were assayed for luciferase and CAT expression. Luciferase activity is normalized to CAT expression.
PMA and ionomycin, but not in TNF-treated cells preincubated with GF directly, correlates with the lack of IB␣ hyperphosphorylation in vivo observed in IB␣ immunoblots of the same cytosolic samples (Fig. 4, A and B, in vivo IB␣ IB). These observations confirm the requirement for a conventional or novel isoform of PKC for the synergistic IKK activity.
Because GF can inhibit both conventional (␣, ␤, ␥) and novel (⑀, , ␦, , ) PKC isoforms (44), we sought to identify which subgroup of the PMA-responsive PKC isoforms was mediating the synergy with calcineurin. We predicted that if PKC would be the PKC isoform that synergized with calcineurin in T cells, pretreatment with Gö would not inhibit the PMA and ionomycin-induced activation of the IKK complex kinase activity and, hence, the in vivo IB␣ hyperphosphorylation. To test this, Jurkat T lymphocytes and freshly isolated primary CD3 ϩ T cells were preincubated with Gö, followed by their stimulation with PMA, ionomycin, or the combination of PMA and ionomycin and analysis of the IKK complex activity. Gö selectively inhibited the synergistic activation of the IKK complex kinase activity mediated by the combination of PMA and ionomycin in both Jurkat T cells (Fig. 4A, IB␣ IVK, lanes 2, 3 and 10 , 11) and in primary CD3 ϩ T cells (Fig. 4B, IB␣ IVK lanes 3, 4 and  13, 14). Gö pretreatment did not affect the TNF-induced IKK complex kinase activation in Jurkat T cells (Fig. 4A, IB␣ IVK, lanes 4 and 12) or in primary CD3 ϩ T cells (Fig. 4B, IB␣ IVK, lanes 5 and 15). The Gö-dependent inhibition of IKK complex activity triggered by PMA and ionomycin directly correlated with the abrogation of the PMA-and ionomycin-induced IB␣ hyperphosphorylation in vivo in both types of T cells (Fig. 4, A and B, in vivo IB␣ IB). As expected, TNF-mediated IB␣ hyperphosphorylation in vivo was not reversed by Gö treatment (Fig. 4, A and B, in vivo IB␣ IB).
The functional relevance of these observations was evaluated in Jurkat T cells that were transfected with a B reporter gene and stimulated or not with PMA or PMA and ionomycin in the presence or not of GF and Gö. As shown in Fig. 4C, both GF and Gö specifically inhibited the PMA and PMA-and ionomycininduced but not the basal NF-B-dependent transcriptional activity, suggesting that the conventional PKC isoforms mediate the PMA effects on the activation of IKK.
To confirm the contribution of this subfamily of PKC isoforms to the synergistic activation of NF-B by calcineurin, wild-type of PKC␣ was overexpressed alone or in combination with the constitutively active form of calcineurin ⌬Cam-AI (Fig. 5). Overexpression of PKC␣-wt or ⌬Cam-AI alone has no effect on NF-B. However, stimulation of PKC␣-wt-transfected Jurkat T cell with PMA resulted in NF-B activation to a degree similar to that induced by PMA and ionomycin costimulation in mock-transfected cells (Fig. 5). Furthermore, co-expression of PKC␣-wt and ⌬Cam-AI resulted in a synergistic NF-B activation (Fig. 5).
These results indicate that, different from the activation of another cellular kinase, JNK, conventional PKC isoforms such as PKC␣ can mediate the synergistic interaction with calcineurin following TCR/CD3 cross-linking that results in the activation of the IKK complex kinase activity in T cells.
A Dominant Negative Form of IKK␤ Blocks NF-B Activation Triggered by PKC-and Calcineurin-dependent Pathways-By having demonstrated that calcineurin synergizes with PKC to activate the IKK complex, we next investigated the effect of dominant negative forms of IKK␣ and IKK␤ on the NF-B activation that follows PMA-ionomycin cell stimulation. Jurkat T cells were transiently co-transfected with expression vectors of dominant negative IKK genes and luciferase reporter genes driven by NF-B concatamers. Rous sarcoma virus-CAT was used as a control of transfection efficiency and cell toxicity.
Overexpression of IKK␤ kinase dead (KD), but not the wildtype isoforms of IKK␣ or -␤ (data not shown), selectively impaired both PMA and PMA/ionomycin-induced up-regulation of NF-B-driven transcription (Fig. 6). Interestingly, overexpression of the IKK␣-DN alone did not inhibit the NF-B activation by PMA and ionomycin, and its combination with the IKK␤-DN did not enhance the inhibition achieved by IKK␤-DN alone (Fig. 6). These observations extend and confirm the involvement of the IKK complex in the convergence of the PKC and calcineurin signal transduction pathways to mediate IB␣ hyperphosphorylation-degradation and NF-B activation. DISCUSSION The results presented in this study identify the molecular targets and mechanisms whereby two TCR/CD3-dependent Jurkat T cells (1 ϫ 10 6 ) were transfected with B-luc-(0.4 g) and REP4/CAT (0.6 g) reporter plasmids, pME18S-PKC␣-wild-type (PKC␣-wt) (0.5 g), and pSR␣4⌬CaM-AI (⌬CaM-AI) encoding a constitutively active calcineurin catalytic subunit (0.5 g). Forty hours later cells were stimulated or not with PMA and/or ionomycin (IONO) for 4 h as described above. Equal amounts (100 l) of cytosolic extracts were assayed for luciferase and CAT expression. Luciferase activity is normalized to CAT expression (relative luciferase activity). All transfection experiments were performed in duplicate. second messengers, PKC and calcineurin, lead to the activation of NF-B in T lymphocytes. The identification that cyclosporin A or FK506 are effective inhibitors of IKK activation advances our knowledge as to the function of these commonly used immunosuppressive agents. Moreover, our results highlight the essential role that conventional PKC isoforms play in the TCRmediated NF-B activation, molecules that should be considered as targets for future drug development with the aim of interfering with T cell activation.
NF-B is a ubiquitous transcription factor that is involved in multiple immune and inflammatory responses (47). T cell cross-linking results in NF-B activation (48 -51) and together with NF-AT, AP-1, and octomer leads to IL-2 expression in in vitro experimental settings (52)(53)(54). However, NF-B may play a more significant role in regulating T cell function in vivo than that inferred from studies analyzing the transcriptional regulation of the IL-2 promoter. The fact that IL-2 production is greatly impaired in c-Rel-deficient lymphocytes (55) and in T cells with constitutive repression of NF-B activity (56) suggests that NF-B, rather than only the nuclear factor of activated T cells, is required for an adequate T cell function. Hence, the identification of calcineurin as a necessary component in the activation of IKK by TCR engagement and the complete inhibition of NF-B activation by cyclosporin A and FK506 may explain the effectiveness of such drugs as NF-B-specific T cell activation inhibitors.
The relevance of NF-B as a target of TCR engagement is not restricted to understanding the immune response but to other relevant areas such as HIV pathogenesis. Recent studies indicate that T lymphocytes serve as a reservoir of latent HIV provirus in patients effectively responding to highly active antiretroviral therapy. The fact that NF-B is a key transcription factor in reactivating HIV from latency in T lymphocytes explains the HIV reactivation and viral production that ensues following T cell receptor activation of latent HIV-infected T cells (57). Identification of calcineurin or of conventional PKC isoform as potential targets of this process could be of future value in the study of HIV reactivation.
In the present study, we find that the IKK complex, and not p90 rsk , mediates IB␣ hyperphosphorylation at Ser-32 and Ser-36 and thus NF-B activation in vivo following PMA and ionomycin stimulation. This observation highlights that while both kinases are activated by PKC-dependent pathways and further amplified in a synergistic manner by Ca 2ϩ -dependent pathways, only the IKK complex appears to be responsible for NF-B activation via IB␣. Although IB␣ Ser-32 and Ser-36 may prove to be a good in vitro substrate to measure p90 rsk activity, its in vivo extrapolation to NF-B activation may be less certain. We conclude this from the fact that p90 rsk and IKK do not co-immunoprecipitate (data not shown) and, more importantly, that a MEK inhibitor (PD) does not affect IKK activation or IB␣ phosphorylation in vivo followed by PMA and ionomycin treatment, whereas it completely inhibits p90 rsk activation. The observation that the MEK inhibitor spares the signal transduction pathway leading to NF-B activation from the mitogen-activated protein kinase pathway may be of future value in selectively inhibiting and differentiating specific target functions of T cell activation, such as NF-B versus AP1 mitogen-activated protein kinase-dependent activation.
IKK␣ and -␤ are contained within a high molecular weight complex with multiple components (24,27). The inhibitory effect of dominant negative forms of IKK␤ on NF-B activation suggests that this kinase is relevant in mediating the synergistic activation of NF-B by the combination of PKC and calcineurin. The role of IKK␣ in this process is less clear. Although overexpression of dominant negative forms of IKK␣ had little effect on the PMA and ionomycin-induced NF-B activity, both endogenous IKK␣ and IKK␤ become in vivo hyperphosphorylated following T cell activation (data not shown), thus suggesting that activation of both kinases may be needed for full signalsome activity. Prior studies (27,28) demonstrated that IKK␤ rather then IKK␣ played a major role in IB␣ phosphorylation by TNF, potentially explaining the stronger effect of overexpressed IKK␤-DN on induced NF-B activation observed in these studies (25).
The mechanism whereby calcineurin converges with PKCdependent pathways to activate the IKK complex is unknown. Previous studies from our group indicated that calcineurin alone had no effect on the activation of NF-B in T lymphocytes (11). However, its presence was required in order for PKC-dependent pathways to induce a maximal level of NF-B activation (11,31). Results presented here extend and confirm those observations by documenting that the level of IKK activation triggered by PKC-dependent pathways is only moderate and that increased (Ca 2ϩ ) levels alone are not sufficient to activate IKK. Rather, increased (Ca 2ϩ ) levels need to be present at the time of PKC activation to result in maximum IKK activity and hence in vivo IB␣ phosphorylation. The fact that calcineurin alone does not activate IKK does not exclude that it does not target the IKK complex. Potentially, calcineurin may modify the composition or interaction of proteins with the signalsome. This could allow for a more effective activation of the signalsome by PKC-dependent pathways. Alternatively, calcineurin may function upstream of the signalsome by modifying transducers of the PKC-dependent pathway resulting in a more effective downstream activation of IKK by PKC. Future studies need to address these and other possibilities, which should ultimately lead to the identification of potential targets of new immunosuppressive agents.
The identification of conventional PKC isoforms in the activation of the IKK complex activation is of potential relevance. By using specific pharmacological inhibitors in primary CD3 ϩ T cells, we conclude that T cell-specific classical PKC isoforms such as ␣ or ␤I must be involved in this process (58 -62). Identification of which of these PKC isoforms that mediate the activation of IKK needs to be pursued. The recent observation that PKC␣ can directly interact and activate IKK␤ but not IKK␣ (41) suggests that PKC␣ can directly mediate TCR/CD3generated signals to the IKK complex. On the contrary, whereas PKC is activated during antigen presentation (43) and involved in the induction of AP-1 transcriptional activity (42,63), its lack of cytoplasmic membrane translocation following TCR/CD3 activation, together with results presented here, suggests that this T lymphocyte-specific PKC isoform may participate in signal transduction pathways activated following antigen presentation, separate from those initiated from TCR/ CD3. Antigen presentation requires not only the activation of TCR/CD3, but also of other co-stimulatory receptors such as CD28. The recent report (64) demonstrating that the combination of CD3-and CD28-generated signals converge on the mitogen-activated protein 3-type kinase, Cot, suggests that the process of antigen presentation that leads to NF-B activation may require the separate but coordinated activation of at least TCR/CD3 and CD28 signaling pathways, each one with distinct but necessary second messages. Future studies should address how two necessary components of the TCR/CD3-initiated pathways, conventional PKC isoforms and calcineurin, interact with the CD28-dependent second messengers to effectively activate NF-B. discussions. We thank Teresa Hoff for excellent manuscript preparation.