Protein Kinase C-δ Negatively Regulates T Cell Receptor-induced NF-κB Activation by Inhibiting the Assembly of CARMA1 Signalosome*

Background: TCR-induced NF-κB activation is crucial for T cell activation. Results: PKCδ inhibited TCR-induced NF-κB activation by a mechanism independent of its kinase activity. Conclusion: PKCδ negatively regulates T cell activation by inhibiting the assembly of CARMA1 signalosome. Significance: It provides a novel mechanism on negative regulation of T cell activation. T-cell receptor (TCR)-induced T-cell activation is a critical event in adaptive immune responses. The engagement of TCR complex by antigen along with the activation of the costimulatory receptors trigger a cascade of intracellular signaling, in which caspase recruitment domain-containing membrane-associated guanylate kinase 1 (CARMA1) is a crucial scaffold protein. Upon stimulation, CARMA1 recruits downstream molecules including B-cell CLL/lymphoma 10 (Bcl10), mucosa-associated lymphoid tissue lymphoma translocation gene 1 (MALT1), and TRAF6 to assemble a specific TCR-induced signalosome that triggers NF-κB and JNK activation. In this report, we identified protein kinase Cδ (PKCδ) as a CARMA1-associated protein by a biochemical affinity purification approach. PKCδ interacted with CARMA1 in TCR stimulation-dependent manner in Jurkat T cells. Overexpression of PKCδ inhibited CARMA1-mediated NF-κB activation, whereas knockdown of PKCδ potentiated TCR-triggered NF-κB activation and IL-2 secretion in Jurkat T cells. Reconstitution experiments with PKCδ kinase-dead mutant indicated that the kinase activity of PKCδ was dispensable for its ability to inhibit TCR-triggered NF-κB activation. Furthermore, we found that PKCδ inhibited the interaction between MALT1 and TRAF6, but not the association of CARMA1 with PKCθ, Bcl10, or MALT1. These observations suggest that PKCδ is a negative regulator in T cell activation through inhibiting the assembly of CARMA1 signalosome.

TCR-induced activation of T cells plays a central role in adaptive immune responses. Activated T cells undergo clonal expansion, differentiation, and begin to execute their immunological functions either by producing cytokines for regulating immune responses or by killing target cells infected with pathogens directly. The tightly controlled activation of transcription factors of the nuclear factor-B (NF-B) 4 family plays a crucial role in these processes (1).
TCR-induced activation of NF-B signaling pathway is initiated from the interaction of TCR with specific antigen in the context of MHC and costimulation of CD28 by its ligands on antigen-presenting cells. Upon stimulation, a cascade of tyrosine phosphorylation of signaling components is successively triggered, which results in the recruitment and activation of a number of signaling proteins, including phosphoinositide-dependent kinase 1 (PDK1) (2). PDK1 phosphorylates PKC and recruits it to lipid rafts in the plasma membrane (3). Concurrently PDK1 recruits CARMA1 to lipid rafts. Phosphorylated PKC phosphorylates the close-by CARMA1 at specific serine residues, and leads to the conformational changes of CARMA1 from an inactive form to an active one (4 -6). Active CARMA1 assembles a CBM complex by associating with Bcl10, which in turn recruits MALT1 (7,8). The CBM complex subsequently recruits downstream signal proteins, such as TRAF6, caspase 8, and TAK1, to collaboratively activate the inhibitor of B kinase (IKK) complex. IB is then phosphorylated and undergoes degradation, which leads to the activation of NF-B.
Among the CBM signalosome, the scaffold protein CARMA1 is crucial for the recruitment of downstream signal proteins, as well as their plasma membrane localization (9 -11). Its activity requires fine tuning to elicit proper activation of NF-B. Posttranscriptional modifications of CARMA1, such as phosphorylation and ubiquitination, determine the outcome of TCR stimulation. For examples, kinases such as PKC, IKK␤, Ca 2ϩ / calmodulin-dependent protein kinase II, and hematopoietic progenitor kinase 1 can phosphorylate CARMA1 at different serine residues in the PKC-regulated domain (PRD), and subsequently lead to NF-B activation (4,5,(12)(13)(14). However, casein kinase 1␣ phosphorylates CARMA1 at S608, which impairs its ability to activate NF-B (15). In addition, monoubiquitination of CARMA1 catalyzed by Cbl-b disrupts the for-mation of CBM complex (16). Although much has been reported, the subtle regulation of CARMA1 signalosome remains not so clear thus far.
To identify potential regulators of CARMA1-mediated NF-B activation, we performed biochemical affinity purification with CARMA1 as a bait protein, and identified PKC␦ as a protein specifically associated with CARMA1. Knockdown of PKC␦ accelerated TCR-induced NF-B activation and IL-2 secretion, suggesting an inhibitory role of PKC␦ in TCRmediated NF-B activation. Furthermore, we demonstrated that the kinase activity of PKC␦ was not required for its inhibitory role. Instead, PKC␦ inhibited TCR-induced NF-B activation by steric hindrance of recruiting TRAF6 to CARMA1 signalosome.
Constructs-NF-B luciferase reporter plasmid was provided by Dr. Gary Johnson. Mammalian expression plasmids for human CARMA1 and its mutants, PKC␦ and its mutants, PKC, Bcl10, or MALT1, fused with Flag-, HA-, or RFPepitope, were constructed by standard molecular biology techniques. Double-stranded oligonucleotides for RNA interference corresponding to the target sequences were inserted into pSUPER. Retro vector (Oligoengine) according to protocols recommended by the manufacturer. The target sequence for human PKC␦-siRNA #1 is 5Ј-AAACACTGGTGCAGAAGAA-3Ј; for PKC␦-siRNA #3 is 5Ј-GCAGCAAGTGCAACAT-CAA-3Ј.
Protein Purification and Mass Spectrometry Analysis-HEK293 cells (5 ϫ 10 8 ) stably transfected with pCTAP-CARMA1 (Stratagene) were collected and the cell lysates were subjected to tandem affinity purification procedures. The purified CARMA1-associated proteins were digested by trypsin in solution. The tryptic peptides were analyzed by HPLC-ESI/ MS/MS with a Thermo Finnigan LTQ adapted for nanospray ionization. The tandem spectra were searched against Homo sapiens National Center for Biotechnology Information reference data base using SEQUEST. Results was filtered by Xcorr ϩ1 Ͼ 1.9, ϩ2 Ͼ 2.2, ϩ3 Ͼ 3.5, sp Ͼ500, Deltcn Ͼ0.1, Rsp Ͻ ϭ 5.
Transfection and Reporter Assays-HEK293 cells were transfected with mammalian expression plasmids by standard calcium phosphate precipitation. Jurkat T cells were transfected with siRNAs or expression plasmids by retroviral transduction. For luciferase reporter assays, HEK293 cells (1 ϫ 10 5 ) were seeded on 24-well dishes and transfected the next day. In the same experiment, we added empty control plasmid to ensure that each transfection received the same amount of total DNA. To normalize the transfection efficiency, 0.01 g pRL-TK (Renilla luciferase) reporter plasmid was added to each transfection. Luciferase assays were performed with a dual-specific luciferase assay kit. Firefly luciferase activities were normalized on the basis of Renilla luciferase activities. All reporter assays were repeated for at least three times. Data shown were average values Ϯ S.D. from one representative experiment.
Coimmunoprecipitation and Immunoblot Analysis-For transient transfection and coimmunoprecipitation experiments, HEK293 cells (2 ϫ 10 6 ) were transfected for 20 h. The transfected cells were lysed in Nonidet P-40 lysis buffer (20 mM Tris, 150 mM NaCl, 1% Nonidet P-40, 1 mM EDTA, pH 7.5) supplemented with proteases inhibitors (10 g/ml aprotinin, 10 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride). For each immunoprecipitation, 0.4 ml aliquot of lysates was incubated with 0.5 g of antibody or control IgG and 25 l of GammaBind G Plus-Sepharose at 4°C for 2 h. Sepharose beads were washed three times with 1 ml lysis buffer containing 0.5 M NaCl. The precipitates were analyzed by standard immunoblotting procedures.
For endogenous coimmunoprecipitation experiments, Jurkat T cells (5 ϫ 10 7 ) were costimulated by mouse anti-human CD3 (10 g/ml), mouse anti-human CD28 (10 g/ml), and 10 g/ml goat anti-mouse IgG for crosslinking at 37°C for the indicated times. Cells were then lysed and subjected to coimmunoprecipitation and immunoblot analysis as described above.
ELISA-IL-2 production in the culture medium of Jurkat cells was measured after CD3/CD28 crosslinking for 48 h using a human IL-2 ELISA kit.

RESULTS
Identification of PKC␦ as a CARMA1-associated Protein-CARMA1 plays a central role in signaling to NF-B activation by TCR or BCR engagement. For a better understanding of signaling events in this pathway, we purified CARMA1-associated proteins by a tandem affinity purification system, and identified the eluted proteins by a shotgun mass spectrometry analysis method. By comparing with other unrelated purifications using the same method, we identified several candidates, including PKC␦, as specific CARMA1-associated proteins. Coimmunoprecipitation results confirmed the interaction between ectopic PKC␦ and CARMA1 in HEK293 cells (Fig. 1A). Besides, PKC␦ also interacted with MALT1 and TRAF6, but not Bcl10 (Fig. 1B).
To determine whether CARMA1 and PKC␦ interact in untransfected cells, we stimulated Jurkat T cells by crosslinking CD3 and CD28 with antibodies, and performed endogenous coimmunoprecipitation. The results showed that endogenous PKC␦ was recruited to CARMA1 after CD3/CD28 costimulation ( Fig. 1, C and D). Immunofluorescent confocal microscopy also indicated that a population of PKC␦ was translocated to the plasma membrane upon TCR stimulation (supplemental Fig.  S1).
PKC␦ Negatively Regulates TCR-induced Activation of NF-B-To address whether PKC␦ is involved in CARMA1dependent NF-B activation, we performed NF-B reporter assays. Overexpression of PKC␦ alone had no marked effect on the basal level of NF-B activation. However, when cotransfected with CARMA1, PKC␦ inhibited CARMA1-induced NF-B activation in a dose-dependent manner ( Fig. 2A).
To evaluate the role of PKC␦ in TCR-induced NF-B signaling, we used RNAi to knockdown PKC␦ in human Jurkat T cells. We made four PKC␦ siRNA constructs, and determined their effects on knockdown of PKC␦. As shown in Fig. 2B, the #3 PKC␦-siRNA construct could inhibit the expression of transiently transfected PKC␦ to 20% of the control sample (p Ͻ 0.01). The #1 siRNA constructs could reduce PKC␦ level to 60% of the control sample, whereas #2 and #4 siRNA plasmid had minimal effects. We thus chose #3 and #1 siRNA constructs for further experiments.
We generated stable Jurkat T cells expressing PKC␦-siRNAs by retroviral transduction. Comparing with the controls, PKC␦-knockdown Jurkat T cells exhibited enhanced phosphorylation and degradation of IB upon the stimulation of CD3/ CD28 crosslinking (Fig. 2C). Consistently, knockdown of PKC␦ potentiated the production of IL-2, which is a pro-inflammatory cytokine induced by NF-B activation (Fig. 2D). The extent of IL-2 enhancement was correlated with the efficiency of PKC␦ knockdown (Fig. 2D). These evidences suggest that PKC␦ itself plays a negative role in TCR-induced activation of NF-B.
The Kinase Activity of PKC␦ Is Not Required for Its Inhibitory Role in TCR-induced NF-B Activation-It has been reported that upon TCR ligation, PKC␦ is phosphorylated and regulates granule exocytosis in a kinase-dependent manner (17). Therefore, we tested whether the kinase activity of PKC␦ was required for TCR-induced NF-B activation. We reconstituted PKC␦-knockdown Jurkat T cells with PKC␦-wild type (WT) or PKC␦-kinase dead (KD) mutant (18). The reconstitutions were confirmed by immunoblotting (Fig. 3A). The reconstituted cells, PKC␦-knockdown cells and controls were stimulated by CD3/CD28 crosslinking, and then the culture medium of each sample was collected to measure IL-2 production, while the cell lysates were subjected to immunoblots to detect NF-B activation. The results showed that the ectopic expression of both WT and KD PKC␦ reversed the promotion of IL-2 production and IB phosphorylation and degradation in PKC␦-knockdown cells (Fig. 3, B and C), indicating that the kinase activity was dispensable for the inhibitory role of PKC␦ in TCR-induced NF-B activation.

PKC␦ Is Associated with the MAGUK Region of CARMA1, and Has No Marked Effect on the Interaction between CARMA1
and PKC-The N terminus of CARMA1 contains a CARD domain and a coiled-coil domain. The C-terminal region contains a MAGUK-homology structure, which is composed of postsynaptic density 95/dislarge/zona occludens 1 (PDZ), Src homology 3 (SH3), and guanylate kinase (GUK) domains (10). PKC, a novel PKC isoform, associates with the PRD region of CARMA1, which links the coiled-coil domain and PDZ domain. PKC phosphorylates CARMA1 at Ser-552 and Ser-645 in PRD, which is essential for TCR-induced NF-B activation (4,5,7). PKC␦, as another novel PKC isoform, shares 85% in catalytic domain and 62.5% in overall protein sequence homology with PKC, thus it may interfere with the interaction between CARMA1 and PKC through competitive binding.
However, domain-mapping experiments suggested that PKC␦ interacted with all the domains consisted in the MAGUK region of CARMA1 but not PRD region (Fig. 4A). CARMA1 associated with the catalytic domain (aa 353-668) of PKC␦, but not the regulatory domain (aa 1-352) (Fig. 4B). As an unexpected result, PKC␦ had no marked effect on the CARMA1-PKC interaction, determined by competitive coimmunoprecipitation assays (Fig. 4C).

. The kinase activity of PKC␦ is dispensable for its inhibitory role in TCR-induced activation of NF-B.
A, PKC␦-knockdown Jurkat T cells were transfected with MIGR-GFP-PKC␦-WT (wild type) or MIGR-GFP-PKC␦-KD (kinase-dead) mutant by retroviral transduction. Seventy-two hours later, reconstituted cells were sorted by FACS. The efficiency of reconstitution was examined by immunoblotting with an antibody to PKC␦ or ␤-actin. B and C, PKC␦knockdown Jurkat T cells, reconstituted Jurkat T cells with PKC␦ WT or KD mutants, and control cell lines were stimulated by CD3/CD28 crosslinking. The culture medium was collected and subjected to ELISA for measuring the production of IL-2 (B). The cell lysates were applied to immunoblots with indicated antibodies (C).

PKC␦ Inhibits the Interaction between MALT1 and TRAF6-
We next investigated whether PKC␦ acted through inhibiting the assembly of CARMA1 signalosome. The competitive coimmunoprecipitation experiments showed that PKC␦ had no marked effect on the interactions between CARMA1-Bcl10, CARMA1-MALT1, or Bcl10-MALT1 (Fig. 5A). In contrast, PKC␦ inhibited the association between MALT1 and TRAF6 in a dose-dependent manner (Fig. 5A).
We further performed endogenous coimmunoprecipitation experiments on PKC␦-knockdown Jurkat T cells or control cells to confirm the effect of PKC␦ on MALT1-TRAF6 interaction. The results showed that TRAF6 in controls was associated with MALT1 moderately after CD3/CD28 crosslinking, peaked at 30 min, and became weaker after 60 min. While in PKC␦knockdown cells, obviously more TRAF6 was associated with MALT1 after CD3/CD28 crosslinking. The association peaked at 10 min, and became weaker after 30 min (Fig. 5B). These results suggested that PKC␦ inhibited the assembly of CARMA1 signalosome by interfering the association between MALT1 and TRAF6.

DISCUSSION
In this study, we investigated the role of PKC␦ in NF-B activation in T cells and reported several observations. First, PKC␦ participates in the negative control of TCR-induced NF-B activation. Second, the kinase activity of PKC␦ is dispensable for the negative regulation. Third, PKC␦ is recruited to CARMA1 upon TCR stimulation, and inhibits the subsequent recruitment of TRAF6 by MALT1.
T lymphocytes express multiple PKC isoforms, among which PKC is considered as the most important isoform participating in T cell activation (19). It is essential for CARMA1 phosphorylation and further NF-B activation. It also phosphorylates RapGEF2 and enhances T cell adhesion (20). Although PKC was reported to be the only PKC isoform translocated to immunological synapse after TCR stimulation (21)(22)(23), more evidences suggest that other PKCs, such as PKC␣, also translocate to the plasma membrane upon TCR stimulation and possibly participate in T cell activation (24). Besides, PKC and PKC are PKC-associated PKCs in T cells, and possibly assist in PKC-mediated signaling (25). In contrast to these PKCs, PKC␦ is reported to serve a negative role in TCR-induced IL-2 cytokine production and T cell proliferation (26). However, the mechanism of this negative regulation is not clear so far.
By identifying the CARMA1-associated proteins, we found that PKC␦ dynamically interacted with CARMA1 and was translocated to the plasma membrane in a TCR stimulation-dependent manner. It has been shown that following TCR ligation, PKC␦ is rapidly phosphorylated at Thr-505 on the activa- tion loop (17,27). This phosphorylation leads to the activation of PKC␦, which regulates lytic granule exocytosis in CTLs in a kinase-dependent manner (17). In this study, however, we found that the kinase activity of PKC␦ was not required for its regulation of TCR-induced NF-B activation, because the promotion of either IB phosphorylation or IL-2 production in PKC␦-knockdown Jurkat T cells could be reversed by reconstitution with kinase-dead mutant of PKC␦, similar to wild-type PKC␦. These surprising findings suggest that negative regulation of TCR-signaling by PKC␦ may be mediated through steric hindrance of the assembly of CARMA1 signalosome.
Although PKC␦ and PKC share high protein sequence homology, they are associated with different region of CARMA1: PKC␦ interacts with the MAGUK region, and PKC interacts with the PRD region. These findings indicated that PKC␦ did not function through competitive binding to CARMA1 with PKC, which was confirmed by our competitive binding assays. Upon TCR stimulation, phosphorylated CARMA1 forms a CBM complex with Bcl10 and MALT1. Then MALT1 directly recruits TRAF6 and TAK1 into the CARMA1 signalosome (28,29). TRAF6 functions as an E3 ligase to catalyze the K63-linked polyubiquitination of MALT1, Bcl10, and IKK␥, and thus promotes signaling to NF-B activation (28,30). Our results showed that PKC␦ had no apparent effect on the association of CARMA1 with PKC, Bcl10, or MALT1. Instead, PKC␦ inhibited the interaction between MALT1 and TRAF6 in a dose-dependent manner, whereas knockdown of PKC␦ enhanced the TCR-induced recruitment of endogenous TRAF6 to MALT1. These results provide an explanation for the inhibitory role of PKC␦ in TCR-induced NF-B activation: PKC␦ is recruited to CARMA1 signalosome upon TCR stimulation and inhibits the recruitment of essential downstream signal components, such as TRAF6.
PKC␦-deficient mice were reported to be susceptible to autoimmune disease, indicating that PKC␦ plays a regulatory role in immune systems. Although the development of immune organs in PKC␦-deficient mice appears to be normal, cultured PKC␦-deficient lymphocytes, both B and T cells, display enhanced proliferation in response to various mitogenic stimuli (26,31). Since signal transduction from antigen receptors in both B and T cells to NF-B shares the same downstream signaling events through CARMA1-Bcl10-MALT1 complex, our results might also explain the enhanced proliferation of PKC␦deficient B cells. Taken together, our findings suggest that PKC␦ negatively regulates TCR-induced NF-B activation and IL-2 production by inhibiting the assembly of CARMA1 signalosome.