The Ca2+-dependent Phosphatase Calcineurin Controls the Formation of the Carma1-Bcl10-Malt1 Complex during T Cell Receptor-induced NF-κB Activation*

T cell receptor (TCR) ligation induces increased diacylglycerol and Ca2+ levels in T cells, and both secondary messengers are crucial for TCR-induced nuclear factor of activated T cells (NF-AT) and NF-κB signaling pathways. One prominent calcium-dependent enzyme involved in the regulation of NF-AT and NF-κB signaling pathways is the protein phosphatase calcineurin. However, in contrast to NF-AT, which is directly dephosphorylated by calcineurin, the molecular basis of the calcium-calcineurin dependence of the TCR-induced NF-κB activity remains largely unknown. Here, we demonstrate that calcineurin regulates TCR-induced NF-κB activity by controlling the formation of a protein complex composed of Carma1, Bcl10, and Malt1 (CBM complex). For instance, increased calcium levels induced by ionomycin or thapsigargin augmented the phorbol 12-myristate 13-acetate-induced formation of the CBM complex and activation of NF-κB, whereas removal of calcium by the calcium chelator EGTA-acetoxymethyl ester (AM) attenuated both processes. Furthermore, inhibition of the calcium-dependent phosphatase calcineurin with the immunosuppressive agent cyclosporin A (CsA) or FK506 as well as siRNA-mediated knockdown of calcineurin A strongly affected the PMA + ionomycin- or anti-CD3 + CD28-induced CBM complex assembly. Mechanistically, the positive effect of calcineurin on the CBM complex formation seems to be linked to a dephosphorylation of Bcl10. For instance, Bcl10 was found to be hyperphosphorylated in Jurkat T cells upon treatment with CsA or EGTA-AM, and calcineurin dephosphorylated Bcl10 in vivo and in vitro. Furthermore, we show here that calcineurin A interacts with the CBM complex. In summary, the evidence provided here argues for a previously unanticipated role of calcineurin in CBM complex formation as a molecular basis of the inhibitory function of CsA or FK506 on TCR-induced NF-κB activity.

Upon engagement of the T cell receptor (TCR), 2 phospholipase C␥1 (PLC␥1) is induced, which in turn hydrolyzes phos-phoinositol 4,5-bisphosphate, resulting in elevated levels of diacylglycerol and inositol trisphosphate. Diacylglycerol is crucial for the activation of members of the PKC family, with PKC⍜ being the key PKC family member required for the activation of AP1, NF-AT, and NF-B signaling pathways in T cells. The other product generated by phosphoinositol 4,5-bisphosphate hydrolysis, inositol trisphosphate, is an inducer of calcium influx into the cytoplasm of T cells by a process known as storeoperated calcium entry. After an initial inositol trisphosphateinduced depletion of the calcium stored in the endoplasmic reticulum, CRAC channels at the plasma membrane are opened in a STIM-dependent fashion, leading to a more sustained calcium influx (1). Calcium acts as a second messenger in the T cell necessary for the activation of several signaling molecules, including the protein kinases CaMKII and PKC␣ as well as the protein phosphatase calcineurin, which in turn are crucial for the activation of the transcription factors NF-AT, AP1, CREB1, and NF-B. Especially calcineurin has been demonstrated to be crucial for the activation of the AP-1, NF-AT, and NF-B signaling pathways (1)(2)(3)(4). In contrast to the NF-AT pathway, however, which is activated due to a direct dephosphorylation of the NF-AT transcription factors by the calcium-dependent protein phosphatase calcineurin, the molecular mechanisms by which calcineurin affects the NF-B pathway is far less understood.
The transcription factor NF-B is crucial for development, survival, homoeostasis, and function of T lymphocytes. One essential step in the antigen receptor-induced NF-B signaling pathway is the formation of the CBM complex, which is composed of Carma1, Bcl10, and Malt1 (5). Whereas Bcl10 and Malt1 interact constitutively, Carma1 is recruited to this Bcl10-Malt1 complex in an inducible manner, depending on site-specific phosphorylation of Carma1. Besides PKC⍜, which seems to be the main Carma1 kinase following TCR engagement, also HPK1 and IKK2 have been reported to phosphorylate Carma1, albeit at different serine residues (6 -8). Following CBM complex formation, Malt1 is ubiquitinated by TRAF6, and this Lys 63 -conjugated ubiquitin chain functions as an anchor for the recruitment of the IB-kinase complex, which is subsequently activated by a still unknown mechanism (9,10). The activated IB-kinase complex phosphorylates IB proteins and thereby targets these proteins for proteasomal degradation. NF-B is subsequently liberated and translocates to the nucleus, where it supports the expression of various NF-B target genes. Finally, TCR-induced NF-B signaling is terminated by the phosphorylation-dependent proteasomal degradation of Carma1 and Bcl10 (5,11,12). The calcium-dependent phosphatase calcineurin is composed of a catalytically active A subunit (CnA) of ϳ61 kDa and a regulatory B subunit (CnB) of 19 kDa. In its inactive state, the active site of calcineurin is blocked by a carboxyl-terminal localized autoinhibitory domain (1,13). For its activation, calcineurin requires the binding of the CnB subunit as well as the calcium-calmodulin complex. The binding of the regulatory proteins causes a structural alteration of CnA and a release of the autoinhibitory domain. Immunosuppressive drugs like cyclosporin A (CsA) and tacrolimus (FK506) target calcineurin and impair the activation of key signaling pathways in lymphoid cells. For instance, pretreatment of T cells with pharmacological inhibitors specific for calcineurin, like CsA or FK506, led to a marked reduction of the TCR-induced NF-AT, JNK, and NF-B activity (2)(3)(4). However, although calcineurin dependence of the TCR-induced NF-B signaling pathway has been known for over a decade, the molecular mechanism by which this calcium-dependent protein phosphatase regulates NF-B has not been clarified.
Here, we demonstrate that calcium influx is essential for the formation of the CBM complex in activated T cells. Furthermore, we provide evidence for a crucial role of the calcium-dependent protein phosphatase calcineurin for the assembly of the CBM complex and NF-B activation. Blocking calcineurin activity with CsA or FK506 or suppression of the catalytic calcineurin A subunit by siRNA led to a drastic decrease in CBM complex formation. Moreover, calcineurin constitutively interacts with components of the CBM complex like Malt1, and inhibition of calcineurin augmented the basal Bcl10 phosphorylation in Jurkat T cells. Furthermore, basal Bcl10 phosphorylation was diminished by calcineurin coexpression in HEK293 cells. In conclusion, our data imply that calcineurin regulates the TCR-induced NF-B signaling pathway by controlling the CBM complex assembly, most likely by the removal of inhibitory phosphate groups from the CBM complex component Bcl10.

EXPERIMENTAL PROCEDURES
Cell Culture, Transfection, and siRNA-mediated Knockdown-Jurkat T cells and Jur4 cells were cultivated using RPMI supplemented with 10% FCS, glutamine, penicillin, and streptomycin. HEK293 cells were kept in DMEM supplemented with 10% fetal bovine serum, penicillin (50 units/ml), and streptomycin (50 g/ml). Primary mouse T lymphocytes were isolated from lymph nodes and spleens of 6-week-old BL6 mice using the CD4 ϩ /CD8 ϩ T cell isolation kit from Dynal according to the manufacturer's protocol. For the stimulation of T cells, 50 ng/ml PMA and 500 ng/ml ionomycin were used. HEK293 cells were transfected using the CaPO 4 transfection method. In brief, cells were plated the day before transfection to obtain a cell density of about 50% at the day of transfection. For the transfection of one well of a 6-well plate, 1-2.5 g of DNA was mixed with 90 l of sterile water and 10 l of a 2 M CaCl 2 solution. After incubation for 5 min, 100 l of 2ϫ HeBS buffer was added, mixed, and incubated at room temperature for an additional 5 min. Subsequently, the transfection sample was added to the cells. For the suppression of calcineurin A isoforms by siRNA, Jurkat T cells were transfected with either 100 nM control siRNA or the SMARTpool siRNAs for CnA␣ or CnA␤ as indicated, using the Nucleofection Kit V (Amaxa/ Lonza). If both CnA isoforms were suppressed, a 75 nM concentration of each SMARTpool siRNA was used. The cells were subsequently incubated for 72 h prior to analysis. Jurkat T cells were transfected using the Lipofectamine LTX ϩPLUS reagent (Invitrogen) according to the manufacturer's protocol.
Real-time PCR Analysis-RNA from 2 ϫ 10 6 cells/sample was isolated using the RNeasy Protect kit (Qiagen), and cDNA was synthesized using an avian myeloblastosis virus reverse transcriptase kit (Invitrogen). Real-time PCR analysis to determine interleukin-2 (IL2) expression levels was performed using the SYBR Green kit from Qiagen in a Lightcycler 480 system (Roche Applied Science). All measurements were performed in triplicate, and the expression of IL2 was normalized to ␤-actin expression. The PMA plus ionomycin (PϩI)-induced increase in IL2 expression was determined using the ⌬⌬Cp method.
Immunoprecipitation and Immunoblotting-Immunoprecipitation and immunoblotting procedures were performed as described previously (17). In brief, 250 -500 g of protein extracts were mixed with 1 g/sample of the appropriate antibody, and samples were incubated overnight at 4°C with agitation. After incubation, 10 l of a 50% protein G slurry was added, and the samples were further incubated for 1 h. Subsequently, the precipitates were washed extensively in TNT buffer (20 mM Tris, pH 8.0, 200 mM NaCl, 1% Triton X-100, 1 mM DTT, 50 mM NaF, 50 mM ␤-glycerophosphate, 50 M leupeptin, 1 mM PMSF). The resulting immunopurified proteins were used for immunoblotting experiments. For the immunoblotting analysis, either the immunopurified protein complexes or, as indicated, 50 -100 g of a protein extract were loaded onto a standard SDS-polyacrylamide gel. SDS-PAGE and the transfer to nitrocellulose (Schleicher & Schuell) or nylon membranes (Immobilon PVDF membrane, Millipore) were performed using standard protocols. The membrane was blocked with 5% milk powder in TBS ϩ Tween 20 prior to the incubation with the primary antibody (1:1000 in TBS ϩ Tween 20), subsequently washed three times for 5 min each, and incubated in a TBS-Tween 20 solution containing either horseradish peroxidase-conjugated or IRDye700/800-conjugated secondary antibody (1:5000). The detection was performed using either ECL substrates from Amersham Biosciences or the Odyssey infrared scanning system (LICOR).
In Vitro Kinase Assay and in Vivo Phosphorylation Studies-For the in vitro kinase assays, the IKK complex was purified from untreated or PϩI-stimulated Jurkat T cells with 1 g of anti-NEMO antibody. Resulting immunocomplexes were washed extensively with TNT and finally with kinase assay buffer to equilibrate the samples. The kinase reaction was performed at 30°C for 30 min after adding 10 Ci of [␥-32 P]ATP and 0.5 g of a bacterial expressed GST-IB␣ (amino acids 1-53) fusion protein in kinase reaction buffer. Samples were subsequently washed extensively with TNT buffer and PBS prior to a separation by SDS-PAGE. The separated proteins were transferred to nitrocellulose membrane, and the phosphorylation was monitored by autoradiography. For the in vivo phosphorylation studies, 2 ϫ 10 7 Jurkat T cells were incubated for 18 h in phosphate-free DMEM with 5% dialyzed calf serum prior to incubation with 2 mCi/ml [ 32 P]orthophosphate for a further 6 h. For in vivo phosphorylation studies using HEK293 cells, the cells were kept in phosphate-free medium, including dialyzed FCS, for 1 h prior to the addition of [ 32 P]orthophosphate and a further incubation for 2 h. The cells were treated as indicated and lysed in TNT, and the resulting extracts were subjected to an immunoprecipitation analysis. Precipitated proteins were separated by SDS-PAGE and transferred to nitrocellulose membrane, and the resulting membrane was used for autoradiography to monitor the phosphorylation and subsequently subjected to immunoblot analysis.
In Vitro Phosphatase Assay-For the in vitro phosphatase assay, 3 g of purified GST-Bcl10 were subjected to a kinase assay reaction either with or without FLAG-tagged IKK2 to generate phospho-Bcl10. For this, HEK 293 cells were tran-siently transfected with expression vectors encoding either FLAG-IKK2 or only the FLAG epitope. The ectopically expressed proteins were purified by immunoprecipitation using anti-FLAG beads (M2, Sigma), washed extensively with TNT buffer and kinase buffer to remove contaminating proteins, and finally included in an in vitro kinase assay with soluble GST-Bcl10 following the standard kinase assay protocol as described above. After the kinase reaction, soluble GST-Bcl10 was transferred to a new Eppendorf tube, and the GST-Bcl10 proteins were pulled down using GSH-Sepharose (Sigma) and washed twice with TNT buffer and once with phosphatase assay buffer (100 mM HEPES, pH 7.4, 100 mM NaCl, 20 mM potassium acetate, 2 mM magnesium acetate, 2 mM CaCl 2 , 2 mM MnCl 2 ). Finally, GSH-coupled GST-Bcl10 was subjected to a phosphatase reaction, including either only calmodulin (catalog no. 208690, Merck) or additionally recombinant calcineurin A-calcineurin B heterodimers (catalog no. BML-SE163-5000, Enzo Life Sciences). The reaction was performed for 1 h at 30°C, and the protein-coupled beads were washed twice with PBS prior to a standard SDS-PAGE. Phospho-GST-Bcl10 was monitored by autoradiography after transfer of the proteins to a nitrocellulose membrane. To ensure equal loading of GST-Bcl10, the membrane was subjected to a standard Bcl10 immunoblot assay.
Luciferase Reporter Assay-For the reporter gene assays, a Jurkat T cell clone was used, which was stably transfected with a luciferase reporter gene under the control of a multimerized B binding site (Jur4 cells). Cells were treated with the individual reagents as indicated, and luciferase activity was generally estimated after 6 h of treatment. Luciferase values were normalized for protein concentration (relative luciferase units/g of protein). The experiments were done in duplicates and were repeated at least three times with similar results.
Gel Shift Analysis-For gel shift analysis (EMSA), 5 g of nuclear proteins or whole cell extracts (DignamC extracts) from untreated or stimulated cells were incubated on ice for 20 min in a reaction containing 0.3 ng of 32 P-labeled B-specific or Oct-specific oligonucleotide, 1 g of poly(dI:dC), and 3 l of a binding buffer. The samples were separated on a native 5% polyacrylamide gel, and the gel was dried and subjected to autoradiography.
Determination of Relative Carma1-Bcl10 Complex Formation-To estimate the relative amount of Carma1 bound to Bcl10, autoradiographs of the appropriate immunoblots were scanned, and resulting TIFF files were analyzed by densitometry using the ImageJ software. The intensities of the Carma1 and Bcl10 signals were determined, and background intensity was subtracted. Subsequently, the ratio of Carma1 signal to Bcl10 signal was calculated, and Carma1-Bcl10 complex formation relative to the control transfection was determined. The value calculated with the control transfection was set arbitrary to 100%. A similar procedure was applied for the estimation of NF-B activity in siRNA-transfected Jurkat T cells measured by EMSA.

RESULTS
Calcium Influx Is Crucial for the Assembly of the CBM Complex-A critical role of increased cellular calcium levels for the TCR-or phorbol ester-induced NF-B signaling pathway has been reported previously. This augmentative effect of increased intracellular calcium levels is reflected in an augmented IB␣ degradation in cells stimulated with a combination of PMA and ionomycin when compared with cells stimulated with PMA alone (Fig. 1A). Although an IB␣-degradation is detectable in both cases, the onset of IB␣ degradation is much faster in cells additionally treated with ionomycin (compare lanes 2-4 with lanes 5-7). To unravel the molecular mechanism underlying the enhancing effect of calcium on PMA-induced NF-B activity, we set out to analyze the impact of elevated cellular calcium levels on upstream signaling events. One crucial step in the TCR-induced NF-B signaling pathway is the formation of a multisubunit protein complex consisting of Carma1, Bcl10, and Malt1 (the CBM complex). Because TCR engagement already causes calcium entry, we used the calcium ionophore ionomycin in combination with the phorbol ester PMA to dissect a potential effect of calcium on CBM complex formation. Jurkat T cells were stimulated with either ionomycin or PMA separately or in combination prior to the analysis of the CBM complex. As expected, the Bcl10-Malt1 interaction remained largely unaltered upon stimulation with either PMA or ionomycin or with a combination of both agents (Fig. 1B,  top). In contrast, Carma1 recruitment to Bcl10 was moderately enhanced in Jurkat T cells stimulated with PMA alone (Fig. 1B, compare lane 1 with lanes 4 and 5). A combined treatment with PMA and the calcium ionophore ionomycin strongly augmented the Carma1 interaction with Bcl10 (Fig. 1B, lanes 2 and  3), whereas treatment of the cells with ionomycin alone had no significant effect on the Carma1-Bcl10 interaction (Fig. 1B, lanes 6 and 7). Similar effects were observed in experiments with the A3.01 T cell line (data not shown). The observed effect of PMA and/or ionomycin on the Carma1-Bcl10 interaction strongly correlated with the NF-B activity, as monitored by the degree of IB␣ degradation (Fig. 1, A and B, bottom) or by a reporter analysis (Fig. 1C) using Jurkat T cells stably transfected with a luciferase gene under the control of a multimerized NF-B binding site (Jur4 cells). Consistently, ionomycin cotreatment also augmented Bcl10 degradation, which is a feature of activated T cells, when compared with PMA treatment alone (Fig. 1B, top, lanes 2-5). In contrast, no significant Bcl10 degradation was observed in T cells treated with ionomycin alone (Fig. 1B, top, lanes 6 and 7).
Because these results suggest an important role of elevated cellular calcium levels for the formation of the CBM complex, we next thought to use an alternative agent to increase cellular calcium levels in order to verify the observed effects. Thapsigargin, a sesquiterpene lactone from the plant Thapsia garganica, is known to block the Ca 2ϩ -ATPase at the endoplasmic reticulum, thus leading to increased calcium levels in the cytoplasm (18). Indeed, pretreatment of Jurkat T cells with thapsigargin (50 -500 nM) augmented the PMA-induced Carma1-Bcl10 interaction drastically (Fig. 1D, compare lanes 2 and 3 with  lanes 4 and 5), similar to the levels seen with PMA ϩ ionomycin stimulation (Fig. 1D, lanes 10 and 11). Again, we observed an enhanced Bcl10 degradation (Fig. 1D) upon thapsigargin pretreatment in comparison with PMA stimulation alone. By contrast, neither a CBM complex formation nor a Bcl10 degrada- tion was observed when cells were treated with thapsigargin alone (supplemental Fig. S1). Of note, thapsigargin pretreatment led also to an electromobility shift of the Bcl10-bound Malt1, which most likely represents the previously reported ubiquitinated Malt1 (Fig. 1D, top). Furthermore, the cumulative effect of the thapsigargin-induced calcium influx on NF-B activity was also seen in luciferase reporter assays using Jur4 cells (Fig. 1C). Again, thapsigargin augmented the PMA-induced NF-B activity similar to ionomycin. Although treatment with thapsigargin alone had no effect on the Carma1-Bcl10 interaction, similar to the situation upon ionomycin treatment (Fig. 1B), a slight NF-B activation was observed in NF-B-dependent reporter gene assays using either thapsigargin or ionomycin to increase cellular calcium levels (Fig. 1C).
Calcium Withdrawal Attenuates PMA ϩ Ionomycin-induced CBM Complex Assembly and NF-B Activation-To further support our findings regarding the crucial role of calcium for the CBM complex formation and NF-B activation, we next conducted a set of reciprocal experiments by using the cellpermeable calcium chelator EGTA-AM. Pretreatment of cells with EGTA-AM (50 and 100 M) had a profound effect on the PMA ϩ ionomycin-induced NF-B activation as measured by EMSA, NF-B-dependent reporter gene assay, and IB␣ degradation (Fig. 2, A and C). However, this negative effect is stim-ulus-specific because the EGTA-AM pretreatment only impaired the NF-B activity upon PMA or PϩI stimulation and not upon TNF-␣ stimulation (Fig. 2, B and C). In accordance with the results obtained with ionomycin or thapsigargin cotreatment, we observed a drastic reduction in Carma1-Bcl10 interaction by pretreating the cells with EGTA-AM prior to stimulation with PMA ϩ ionomycin (Fig. 2D, top). Taken together, these results clearly suggest that the molecular basis of the calcium impact on the NF-B signaling pathway is the modulation of the CBM complex formation.
The Calcium-dependent Protein Phosphatase Calcineurin Is Required for CBM Complex Formation-To explore the potential role of calcineurin as mediator of the calcium effect on the NF-B activity and especially the CBM complex formation, we first analyzed the PϩI-induced NF-B activity in Jurkat T cells alone or upon pretreatment with CsA, a well known pharmacological inhibitor of calcineurin. As described previously (3), the PϩI-induced NF-B activity was dramatically attenuated by CsA pretreatment as monitored by EMSA and IB␣ degradation (Fig. 3B) or NF-B-dependent luciferase reporter assays (Fig. 3A). Interestingly, CsA pretreatment impaired not only the NF-B activity induced by PϩI but also that induced by PMA or ionomycin alone (Fig. 3A). Furthermore, pretreatment of Jurkat T cells with CsA or FK506, another pharmacological calcineurin inhibitor, impaired the activity of the IB-kinase complex as monitored by an in vitro kinase assay (Fig. 3C). To determine whether the observed negative effect of the calcineurin inhibitors is based on reduced CBM complex formation, we analyzed the Carma1-Bcl10 interaction in Jurkat T cells that were either stimulated with PϩI alone or additionally pretreated with CsA (Fig. 3D). In accordance with the results shown in Figs. 1 and 2, we observed a marked reduction in the Carma1-Bcl10 interaction by pretreatment with CsA (Fig. 3D, compare lanes 2-4 with lanes 5-7) or with FK506 (data not shown), which was accompanied by an attenuated Bcl10 degradation in CsA-pretreated Jurkat cells (Fig. 3D). In addition, the appearance of a slower migrating Bcl10 signal was only observed with PϩI-stimulated cells and not with CsA-pretreated cells (Fig. 3D, top, compare lanes 3 and 4 with lanes 6  and 7). Because this Bcl10 band shift results most likely from an IKK2-mediated Bcl10 hyperphosphorylation, which depends on a prior CBM complex assembly (14), we concluded that calcineurin is indeed crucial for the CBM complex formation. A similar effect of CsA on Carma1-Bcl10 interaction as well as Bcl10 degradation was observed in primary murine T lymphocytes (Fig. 3E). To further support our findings regarding the inhibitory effect of calcium withdrawal or calcineurin inhibition on CBM complex assembly, Jurkat T cells were pretreated with EGTA-AM or CsA prior to stimulation with agonistic anti-CD3 and anti-CD28 antibodies. Again, both inhibitors led to a distinct reduction in signal-induced Carma1-Bcl10 interaction (Fig. 3F). To determine whether calcineurin inhibition by CsA would impair other signaling pathways, including calcium-dependent signaling pathways, in PMA ϩ ionomycinstimulated Jurkat T cells, the activation status of different PKCs as well as CaMKII was determined by using phospho-specific antibodies for the activated isoforms of these kinases (Fig. 4). As expected, CsA pretreatment caused a significant decrease in IB␣ phosphorylation (Fig. 4A, compare lanes 3 and 4 with  lanes 7 and 8). Moreover, CsA pretreatment also caused a moderate reduction in p65 phosphorylation at Ser 536 (Fig. 4C). However, the phosphorylation pattern of CaMKII showed only a slight reduction after CsA pretreatment (Fig. 4B) (data not shown). In contrast, we detected a moderate increase in PKC⍜ phosphorylation, whereas the phosphorylation levels measured by pPKC␣ and pan-phospho-PKC antibodies showed, if anything, only a weak increase and no reduction. Collectively, the differences seen in the phosphorylation patterns of the different PKC isoforms or of CaMKII cannot explain the reduced CBM complex formation observed in CsA-pretreated cells.
Reduced NF-B Activation and CBM Complex Formation by siRNA-mediated Suppression of Calcineurin A-To ensure that the negative effects of CsA and FK506 on Carma1-Bcl10 interaction as well as NF-B activation depend on their inhibitory effect on calcineurin, siRNAs specific for the two isoforms of the catalytic subunit CnA, CnA␣ and CnA␤, were used to sup-press their expression in Jurkat T cells. Both isoforms are expressed in Jurkat T cells with nearly identical levels (Fig. 5A,  lanes 1 and 2), and only the suppression of both catalytically active calcineurin isoforms led to a marked attenuation of the PMA ϩ ionomycin-induced Carma1-Bcl10 interaction (Fig.  5A, compare lanes 1 and 2 with lanes 7 and 8; for a quantification, see Fig. 5A, right). Consistent with the results seen in the Carma1-Bcl10 interaction analysis, the NF-B activity as monitored by EMSA was only reduced when both CnA isoforms were suppressed in combination (Fig. 5B). A similar effect was observed after siRNA-mediated suppression of both CnA isoforms in Jurkat T cells stimulated with agonistic anti-CD3 ϩ CD28 antibodies (supplemental Fig. S2). IL2 expression is regulated by NF-AT, NF-B, and AP1 transcription factors and is highly sensitive to CsA treatment (19). To estimate the degree of calcineurin inhibition by siRNA-mediated suppression of the CnA isoforms, quantitative PCR analysis (Fig. 5C,  top) of untreated and PϩI-stimulated Jurkat T cells transiently . The same extracts were used for immunoblot analyses (IB) with antibodies recognizing IB␣, phospho-p65 (Ser 536 ), and ERK2 (bottom). C, the IKK complex was immunopurified from Jurkat T cells pretreated with CsA or FK506 prior to PϩI stimulation using an anti-NEMO antibody, and the kinase activity (KA) was estimated. Phosphorylation was determined by autoradiography. IB␣ degradation was determined by an additional immunoblot. D, anti-Bcl10 immunoprecipitation analysis (IP) (top) of whole cell extracts from Jurkat T cells pretreated with CsA (200 ng/ml) for 1 h. Selected samples were additionally stimulated with PϩI as indicated. A fraction of the whole cell extracts (50 g/sample) was subjected to control immunoblot analyses (bottom). E, a similar anti-Bcl10 immunoprecipitation analysis as described in D was performed using primary murine T cells. F, Jurkat T cells were pretreated with CsA (200 ng/ml) or EGTA-AM (100 M) for 1 h prior to the stimulation with agonistic anti-CD3 (1 g/ml, plate-bound) and anti-CD28 (5 g/ml, soluble). Subsequently, an anti-Bcl10 immunoprecipitation analysis was performed (top). To monitor the cellular expression levels of the analyzed proteins, additional immunoblot analyses were performed (bottom).
transfected with the different CnA-specific siRNAs were performed. Again, the combined suppression of both CnA isoforms had the strongest effect on PϩI-induced IL2 expression, whereas the suppression of the individual CnA isoforms led to a gradual decrease in IL2 expression. Notably, parallel immunoblot analysis revealed that PϩI stimulation induced calcineurin A expression (Fig. 5C, bottom), which suggests that a remaining calcineurin activity even in siRNA-transfected cells might be possible. However, in general, the effect of siRNA-mediated CnA knockdown on the IL2 expression reflected the degree of NF-B and CBM complex inhibition.
Inhibition of Calcineurin by CsA or EGTA-AM Causes Increased Basal Phosphorylation of Bcl10-Having established a role of calcineurin for the PMA ϩ ionomycin-or anti-CD3 ϩ CD28-induced assembly of the CBM complex, we next aimed to identify the molecular mechanism(s) by which calcineurin affects the CBM complex formation. Because the CBM complex formation is highly regulated by the phosphorylation of Carma1 and Bcl10, an in vivo phosphorylation study was performed by labeling Jurkat T cells with [ 32 P[orthophosphate prior to a treatment with either CsA or EGTA-AM alone or in conjunction with PMA ϩ ionomycin stimulation (Fig. 6) to gain further insight into the effect of calcium-calcineurin on the phosphorylation status of Carma1, Malt1, and Bcl10. Subsequently, endogenous Bcl10, Malt1, or Carma1 were isolated by immunoprecipitation with the appropriate antibodies. If calcineurin affects the CBM complex formation by rendering the phosphorylation status of one of these CBM complex components, treatment with the calcineurin inhibitor CsA should increase the phosphorylation of the targeted protein. Indeed, pretreatment of Jurkat T cells with either CsA or EGTA-AM distinctively increased the basal Bcl10 phosphorylation (Fig. 6,  top, lanes 1-3). By contrast, the PMA ϩ ionomycin-induced increase of Bcl10 phosphorylation was not dramatically affected by CsA or EGTA-AM pretreatment (Fig. 6, top, lanes  4 -6). Furthermore, neither CsA nor EGTA-AM pretreatment led to an increase in basal Carma1 phosphorylation; instead, a slight decrease was observed (Fig. 6, bottom, lanes 1-3). In cells stimulated with PMA ϩ ionomycin, CsA and EGTA-AM either had no effect or led to a mild reduction in Carma1 phosphorylation (Fig. 6, bottom, lanes 4 -6). In addition, the phosphorylation status of Malt1 remained unchanged upon treatment with CsA or EGTA-AM irrespective of PMA ϩ ionomycin stimulation (data not shown).
The increased basal Bcl10 phosphorylation in Jurkat T cells upon CsA or EGTA-AM treatment points to the possibility that Bcl10 might be a substrate for the phosphatase activity of calcineurin. Bcl10 has been shown to be a target of different kinases, like CaMKII or IKK2. Important, however, is the notion that all described Bcl10 phosphorylations exert a negative effect on CBM complex formation and NF-B activity (14,20). To further analyze the potential role of calcineurin for Bcl10-dephosphorylation, Bcl10 was ectopically expressed in HEK293 cells either alone or in combination with IKK2 or calcineurin A␣. As expected, the coexpression of IKK2 led to a distinct increase in one slowly migrating Bcl10-specific signal (Fig. 7A, compare lanes 2 and 4, signal III), which resembles a IKK2-mediated Bcl10 phosphorylation (14,20). This slower migrating signal, however, was reduced upon coexpression of calcineurin A␣ (Fig. 7A, lanes 3 and 5), again arguing for a role of calcineurin in Bcl10 dephosphorylation. To prove that this slower migrating Bcl10 signal indeed represents a Bcl10 phosphoisoform, two different experimental strategies were employed. First, another in vivo phosphorylation analysis was performed using HEK293 cells cotransfected with FLAG-Bcl10 together with either a constitutive active ⌬Cam mutant or an inactive ⌬CamH151Q mutant alone or in conjunction with Xpress-IKK2 prior to a labeling of the cells with [ 32 P]orthophosphate. As shown in Fig. 7B, the coexpression of ⌬CamWT, but not of ⌬CamH151Q, led to a distinct reduction of the basal Bcl10 phosphorylation in HEK293 cells (Fig. 7B,  lanes 2-4, reduced signals II and III). Coexpression of IKK2, on the other hand, augmented the Bcl10 phosphorylation, especially signal III. However, the addition of ⌬CamWT caused again a decrease in the IKK2-mediated Bcl10 phosphorylation (Fig. 7B, lanes 5 and 6). As a second experimental setup to determine whether signal III represents a phosphoisoform of Bcl10, immunopurified exogenously expressed Bcl10 was treated with shrimp alkaline phosphatase, and the Bcl10 pattern was subsequently analyzed by immunoblot. As expected, coexpression of IKK2 further enhanced the appearance of the slowest migrating Bcl10 signal, whereas the coexpression of the constitutive active ⌬Cam mutant reduced this signal (Fig. 7C,  lanes 6 -8 and 10 -12). Moreover, treatment of the immunopurified Bcl10 with shrimp alkaline phosphatase led to a strong reduction of the slower Bcl10 signal (signal III), further sup- A, Jurkat T cells were either left untreated or pretreated with CsA (200 ng/ml) prior to stimulation with PMA ϩ ionomycin. Resulting whole cell extracts were subjected to immunoblot analysis (IB) with the indicated phospho-specific antibodies to monitor the activation of PKC proteins, especially PKC⍜ and PKC␣, and phospho-IB␣. B, to monitor the activation status of CaMKII, cytoplasmic extracts also used in Fig. 3B were analyzed with anti-phospho-CaMKII and anti-CaMKII antibodies. C, cytoplasmic extracts from Jurkat T cells either left untreated or pretreated with CsA for 30 min prior to PϩI stimulation were analyzed with the phospho-p65-specific antibody.
The opposite effect of calcineurin and IKK2 on the phosphorylation pattern of Bcl10 implied that known IKK2 target sites in Bcl10 might be dephosphorylated by calcineurin. To prove this hypothesis and to test whether calcineurin affects the basal Bcl10 phosphorylation also in T cells, Jurkat T cells were transiently transfected with expression vectors for FLAG-Bcl10 WT or a mutant of Bcl10 with serine to alanine substitutions at positions Ser 134 , Ser 136 , Ser 138 , Ser 141 , and Ser 144 , which are known IKK2 target sites (FLAG-Bcl10 S5A ). Subsequently, the cells were either left untreated or stimulated with PϩI for 3 h to induce Bcl10 degradation. Inactivation of the five serine residues in Bcl10 caused a loss of two specific Bcl10 signals, which correspond to phosphoisoforms of Bcl10 (signals II and III, Fig.  7D, lanes 1 and 5). As expected, coexpression of ⌬Cam reduced the intensity of signal III and partially of signal II (compare lanes 1 and 2 with lanes 3 and 4). Interestingly, only signal I was decreased upon PϩI stimulation irrespective of whether it was caused by Bcl10 WT or Bcl10 S5A . Together, these results suggest that calcineurin causes a dephosphorylation of Bcl10 also in Jurkat T cells, and the central serine residues at the positions Ser 134 , Ser 136 , Ser 138 , Ser 141 , and Ser 144 are likely to be targets of the calcineurin-mediated Bcl10 dephosphorylation.
Phosphorylation of Bcl10, especially at Ser 138 , has been linked to a signal-induced degradation of Bcl10. Because pre-

FIGURE 5. siRNA-mediated suppression of CnA␣ and CnA␤ attenuates P؉I-induced NF-B activity.
A, anti-Bcl10 immunoprecipitation analyses (IP) were performed using whole cell extracts from untreated or PϩI-stimulated Jurkat T cells that were transfected with different siRNAs. Control immunoblot analyses (IB) were performed to determine the cellular levels of the indicated proteins. Two independent experiments were used for the quantification of the relative PϩI-induced Carma1-Bcl10 interaction (bottom). Calculation was performed on the basis of the signal intensity measured by densitometry. B, Jurkat T cells transfected with the indicated siRNA were either left untreated or stimulated with PMA ϩ ionomycin. Resulting whole cell extracts were subjected to EMSA experiments using NF-B-or Oct-specific probes (top). To monitor the siRNA-mediated knockdown of CnA isoforms, additional immunoblot analyses were performed using the indicated antibodies (bottom). n.s., nonspecific. Relative NF-B binding activity was determined by normalizing the background subtracted signals of the NF-B-specific bands to the intensity of the Oct-specific signals. C, to estimate the functionality of the siRNA-mediated suppression of CnA isoforms, the IL2 expression in untreated or PϩI-stimulated (8 h) Jurkat cells, which were transfected with the stated siRNAs, was determined by quantitative PCR analysis (top). All samples were measured in triplicate and normalized for ␤-actin expression. Protein extractions of transfected cells were done in parallel, and control immunoblots were performed with the indicated antibodies (bottom). Error bars, S.E.  4 -6). IB, immunoblot. treatment with CsA led to a partial stabilization of Bcl10 during PMA ϩ ionomycin stimulation (Fig. 3D), we next analyzed whether IKK2-mediated phosphorylation or calcineurin-mediated dephosphorylation would affect the stability of Bcl10, especially the Bcl10 phosphoisoform (signal III). HEK293 cells were transfected with expression vectors either for FLAG-Bcl10 alone (Fig. 7E, lanes 1-4), together with ⌬Cam (lanes 5-8), with Xpress-IKK2 (lanes 9 -12), or with a combination of ⌬Cam and Xpress-IKK2 (lanes [13][14][15][16]. To monitor the stability of the ectopically expressed Bcl10 protein, the cells were either left untreated or were treated with cycloheximide for different times. As expected, coexpression of IKK2 or ⌬Cam altered the phosphorylation pattern of Bcl10. Whereas signal III was increased upon IKK2 coexpression, it was diminished in all ⌬Cam-expressing cells. Bcl10 appears to be a relatively stable protein in HEK293 cells because its levels remained stable upon 8 h of cycloheximide treatment, whereas endogenous c-Myc protein or HA-⌬Cam showed a shorter half-life (Fig. 7E, lanes  1-8). Importantly, neither coexpression of IKK2 nor coexpression of ⌬Cam had a marked effect on the stability of exogenously expressed Bcl10, especially signal III. Thus, these results suggest that the stability of Bcl10 is not directly regulated by IKK2-mediated Bcl10 hyperphosphorylation or calcineurinmediated Bcl10 dephosphorylation. To further determine whether the reduced phosphorylation of Bcl10 upon calcineurin coexpression in intact cells is indeed caused by a direct FIGURE 7. Calcineurin affects the phosphorylation status but not the stability of Bcl10 in HEK293 cells. A, immunoblot analyses using whole cell extracts from transiently transfected HEK293 cells with antibodies recognizing IKK2 (top), calcineurin A (middle), or Bcl10 (bottom). B, FLAG-Bcl10 was ectopically expressed in HEK293 cells either alone or in conjunction with HA-⌬CamWT, a catalytically inactive ⌬Cam mutant (⌬CamH151Q) or Xpress-IKK2, as indicated. Transfected cells were radiolabeled (100 Ci/ml [ 32 P]orthophosphate, 2 h), and resulting whole cell extracts were subjected to anti-FLAG immunoprecipitation experiment. Phosphorylation was determined by autoradiography following an anti-Bcl10 immunoblot to ensure the successful precipitation. C, FLAG-Bcl10 was immunoprecipitated (IP) from transiently transfected HEK293 cells, and immunocomplexes were incubated at 37°C for 1 h with shrimp alkaline phosphatase (SAP; 1 unit, lanes 1-4). Subsequently, the immunocomplexes were subjected to anti-Bcl10 immunoblot analysis (IB). Expression levels of the FLAG-Bcl10 (top, lanes [13][14][15][16], HA-⌬Cam (middle), or Xpress-IKK2 (bottom) were determined by immunoblot analyses. D, Jurkat T cells were transiently transfected with expression vectors for FLAG-Bcl10 WT , FLAG-Bcl10 S5A , and FLAG-⌬Cam, as indicated. The cells were either left untreated or were stimulated with PϩI for 3 h prior to immunoblot analysis with the indicated antibodies. E, HEK293 cells transiently transfected with expression vectors for FLAG-Bcl10, HA-⌬Cam, or Xpress-IKK2 were either left untreated or treated with cycloheximide (50 ng/ml) for different times prior to immunoblot analyses with antibodies for Bcl01, IKK2, or HA epitope. As a positive control for cycloheximide efficacy, c-Myc levels were analyzed. dephosphorylation of Bcl10 by calcineurin, we performed an in vitro phosphatase experiment using recombinant proteins (Fig.  8). GST-Bcl10 fusion protein was phosphorylated by IKK2 in vitro, and phospho-GST-Bcl10 was purified and subsequently subjected to an in vitro assay including either calmodulin alone or a combination of calmodulin with increasing amounts of the calcineurin A-calcineurin B heterodimer. Consistent with our previous results, we observed also in this in vitro phosphatase assay a distinct reduction of Bcl10 phosphorylation by the addition of active calcineurin (Fig. 8, compare lanes 1, 3, and 5). Taken together, these data imply that calcineurin can act as a Bcl10 phosphatase and open up the possibility that one of the molecular mechanisms by which calcineurin positively affects the CBM complex assembly is the removal of negative acting phosphate groups from Bcl10.
Calcineurin A Interacts with the CBM Complex-Having identified calcineurin as potential Bcl10 phosphatase, we next aimed to determine whether calcineurin might interact with components of the CBM complex. To this end, interaction studies were performed upon ectopical expression of CnA␣, CnA␤, or the constitutive active deletion mutant of calcineurin A␣ (⌬Cam) in conjunction either with Carma, Malt1, or Bcl10 in HEK293 cells (Fig. 9, A-C). Indeed, a CnA-Malt1 proteinprotein interaction was observed with both calcineurin A isoforms, CnA␣ and CnA␤, whereas Carma1 was only capable of a CnA␣ interaction and did not form a complex with CnA␤ ( Fig.   FIGURE 8. Calcineurin is a Bcl10 phosphatase in vitro. For the in vitro phosphatase assay, bacterial expressed GST-Bcl10 fusion protein was coupled to GSH-agarose prior to a kinase assay with FLAG-IKK2. The phosphorylated GST-Bcl10 proteins were subsequently washed twice with TNT buffer and once with phosphatase reaction buffer. The purified phosphorylated GST-Bcl10 protein was then subjected to an in vitro phosphatase assay by adding either recombinant calmodulin alone (150 units/sample, lanes 1 and 2) or in conjunction with increasing amounts of recombinant calcineurin A-calcineurin B (lanes 3-6). The samples were separated by SDS-PAGE and transferred to nitrocellulose membrane, and the resulting membrane was subsequently exposed to x-ray films to determine the phosphorylation status of Bcl10 ( 32 P). To control for equal GST-Bcl10 loading, the membrane was additionally subjected to an anti-Bcl10 immunoblot analysis (IB) (bottom).  A and B). In contrast to Malt1 and Carma1, an interaction of calcineurin A with Bcl10 was not detectable upon ectopical coexpression in HEK293 cells (Fig. 9C), possibly due to a low stability of this enzyme-substrate complex. However, similar to the results shown in Fig. 7, coexpression of calcineurin A␣ and more pronounced ⌬Cam led again to a marked decrease in the slower migrating phospho-Bcl10 signal (signal III). To further analyze the interaction of calcineurin with the CBM complex, an additional interaction study was performed by immunoprecipitating endogenous Malt1 from untreated Jurkat T cells or from Jurkat T cells stimulated with PMA ϩ ionomycin. Here, we observed a constitutive interaction of calcineurin A with Malt1 (Fig. 9D, lanes 1-3). These results consistently suggest that CnA interacts physically with the CBM complex.

9,
(which suggests that calcineurin dephosphorylates Bcl10 in vivo). In addition, the decrease in Bcl10 phosphorylation in HEK293 and Jurkat T cells upon coexpression of calcineurin A or more pronounced upon coexpression of the constitutive active mutant ⌬Cam further supports our hypothesis. Moreover, we observed a protein-protein interaction of the calcineurin A subunit or a constitutive active deletion mutant of calcineurin A␣ (⌬Cam) with Malt1 or Carma1. In addition, calcineurin A forms a protein complex with the CBM complex in Jurkat T cells as demonstrated by the Malt1-calcineurin A interaction (Fig. 9D). The reason for the lack of Bcl10-CnA interaction with ectopically expressed proteins remains unclear. However, one explanation could be that Malt1 is required for a stable Bcl10-CnA interaction. Alternatively, the recently published Bcl10-calmodulin interaction might also contribute to the recruitment of calcineurin to Bcl10 (25). Experiments are under way in our laboratory to clarify this question.
Taken together, the evidence provided in this study clearly demonstrates that calcium influx is crucial for the formation of the CBM complex during TCR-induced NF-B activation. Furthermore, our data strongly suggest that calcineurin is one of the calcium-dependent signaling proteins involved in the assembly of the CBM complex (Fig. 10). However, calcineurin might not be the only calcium-dependent enzyme required for this process. Blocking CaMKII with the specific inhibitor KN93, for instance, had a profound effect on the PMA-and TCR-induced NF-B activity (26). Furthermore, PKC␣ has also been demonstrated to affect the antigen receptor-induced NF-B activation in T cells. However, PKC␣ seems to modulate the TCR-induced NF-B activity indirectly by affecting PKC⍜, the central regulator of CBM complex assembly (27). CaMKII, on the other hand, has been identified as both a Bcl10 and a Carma1 kinase and could participate in the control of CBM complex assembly by phosphorylating Carma1 at Ser 109 (23,28). The reduced Carma1 phosphorylation upon EGTA-AM pretreatment would support the hypothesis of an additional calcium-dependent but CsA-insensitive effect on Carma1 phosphorylation. However, although PKC␣ or CaMKII might contribute to the overall calcium dependence of the TCR-induced NF-B activation, they seem not to be negatively affected by CsA (Fig. 4). From these results, especially the calineurinmediated dephosphorylation of Bcl10 in vitro (Fig. 8), we concluded that the positive effect of calcineurin is rather direct and does not involve the activation of other calcium-dependent signaling pathways. Although we provide evidence that CBM complex formation might be affected by the calcineurin-mediated dephosphorylation of Bcl10, further work is needed to FIGURE 10. Hypothetical model of the calcineurin effect on CBM complex assembly. In resting T cells, Carma1 is hypophosphorylated and Bcl10, included in a heterodimer with Malt1, displays a low basal phosphorylation (gray boxes). Furthermore, calcineurin and PKC⍜ activity remains low in unstimulated T cells. Engagement of T cell receptor, however, induces PLC␥1-dependent Ca 2ϩ influx and diacylglycerol (DAG)-mediated PKC⍜ activation (white box, center). Activated PKCQ phosphorylates Carma1, and activated calcineurin dephosphorylates Bcl10. Hypophosphorylated Bcl10 and phospho-Carma1 are a prerequisite for the efficient formation of the CBM complex. As a negative feedback mechanism, Bcl10 gets rephosphorylated or even hyperphosphorylated, for example by IKK2 or by CaMKII. The CBM complex becomes inactive, and Bcl10 is subsequently degraded.
clarify the exact role of calcineurin for the TCR-induced NF-B signaling pathway. For instance, it will be necessary to identify the serine residues in Bcl10 that are dephosphorylated by calcineurin. To this end, it seems likely that one or more of the five already identified carboxyl-terminal IKK2 phosphorylation sites are targeted by calcineurin because calcineurin coexpression reverted the IKK2-mediated Bcl10 hyperphosphorylation (Fig. 7, A-C). Furthermore, the Bcl10 S5A mutant is not affected by calcineurin (Fig. 7). Another question pertains to the kinases responsible for basal Bcl10 phosphorylation. IKK2 and CaMKII have been identified as Bcl10 kinases, which are responsible for Bcl10 phosphorylation upon cell activation. However, whether the basal Bcl10 phosphorylation is also conducted by these kinases or conducted by still unknown kinases needs to be clarified in future experiments. The answers to these questions will not only provide a better understanding of the regulation of the TCR-induced NF-B signaling pathway but might also lead to the development of more specific inhibitors for the different CnA isoforms in order to avoid the toxicity of the currently used immunosuppressors CsA and FK506.