Protein Kinase Cθ Modulates Nuclear Receptor-Corepressor Interaction during T Cell Activation*

Transcriptional repression by nuclear receptor corepressors plays a critical role in T cell development. However, the role of these corepressors in T cell activation is poorly understood. We report that T cell activation silenced transcription driven by nuclear receptors retinoic acid receptor, retinoid X receptor, and thyroid hormone receptor and induced silencing mediator of retinoic acid and thyroid hormone receptors (SMRT)-receptor interaction. Whereas the expression of a dominant active mutant of protein kinase Cθ (PKCθ) induced strong SMRT-receptor interaction in the absence of T cell activation, a dominant negative mutant of PKCθ decreased the interaction. Loss of PKCθ expression by induction of “RNA interference” resulted in the attenuation of basal and activation-induced SMRT-receptor interaction. We suggest that T cell activation silences nuclear receptor-dependent transactivation in part through PKCθ-dependent enhancement of SMRT-receptor interaction.

Transcriptional repression by nuclear receptor corepressors plays a critical role in T cell development. However, the role of these corepressors in T cell activation is poorly understood. We report that T cell activation silenced transcription driven by nuclear receptors retinoic acid receptor, retinoid X receptor, and thyroid hormone receptor and induced silencing mediator of retinoic

acid and thyroid hormone receptors (SMRT)receptor interaction. Whereas the expression of a dominant active mutant of protein kinase C (PKC) induced strong SMRT-receptor interaction in the absence of T cell activation, a dominant negative mutant of PKC decreased the interaction. Loss of PKC expression by induction of "RNA interference" resulted in the attenuation of basal and activation-induced SMRT-receptor interaction. We suggest that T cell activation silences nuclear receptor-dependent transactivation in part through PKC-dependent enhancement of SMRT-receptor interaction.
Engagement of T lymphocytes through T cell receptor-CD3 complex and the CD28 costimulatory molecule is the initial step in a complex process that determines the outcome of such diverse responses as thymocyte development, lineage commitment, antigen-specific activation, and activation-induced apoptosis (1)(2)(3)(4)(5)(6). The physiological outcome is determined based on the signaling context and involves integration of multiple signaling pathways. Recent gene expression studies have identified protein kinase C (PKC) 1 , a novel member of the PKC family, to assume a center stage in the supramolecular activa-tion complex associated with membrane rafts at the contact site between an antigen-presenting cell and a T lymphocyte (7)(8)(9)(10). Loss of PKC in a transgenic mouse model has shown that PKC is essential for AP-1, NF-B activation, CD25 upregulation, and IL-2 production (11). In vitro experiments have shown that PKC synergizes with calcineurin to activate a number of regulatory elements in the IL-2 promoter (12)(13)(14)(15)(16). This cooperation has also been identified to induce Fas ligand expression during activation-induced apoptosis (17). T cell activation was recently shown to induce PKC-dependent phosphorylation of WIP ((Wiskott-Aldrich syndrome protein (WASP)-interacting protein)), releasing WASP from WIP inhibition, an essential feature of normal T cell signaling (18).
Retinoic acid receptor (RAR), retinoid X receptor (RXR), thyroid hormone receptor (TR), and vitamin D receptor (VDR) are members of the nuclear receptor superfamily that control diverse cellular functions (19 -24). These receptors function as ligand-dependent transcription factors by binding as homodimers or heterodimers to the DNA-response elements to either activate or suppress target genes. In the presence of ligand, when bound to the target DNA, the receptors recruit coactivators to enhance gene expression. When the ligand is lacking, the receptors recruit corepressors and induce transcriptional repression. Two primary corepressors, known to be involved in the transcriptional silencing by RAR, RXR, and TR, are silencing mediator of retinoic acid and thyroid hormone receptor (SMRT) and nuclear hormone receptor corepressor (N-CoR) (25)(26)(27)(28)(29). Transcriptional repression by these corepressors plays a crucial role in T cell development as recent studies have shown that T cells in N-CoR Ϫ/Ϫ mouse embryos exhibit arrest at the double-negative stage (30).
We have shown previously that the activation of T cells undergoing active cycling led to the accumulation of RXR␣ protein. Transcriptional activation analysis using reporterbased assays showed that the accumulated RXR␣ protein was transcriptionally inactive in these cells (31). We now show that the activation of human Jurkat T cells by simultaneous engagement with anti-CD3 and CD28 antibodies inhibited not only RXR element (RXRE)-dependent transcription but also RAR element-(RARE) and TR element (TRE)-driven transactivation. We present evidence that antibodies to CD3 and CD28 induce interaction between SMRT and ligand binding domains of all three (RXR, RAR, and TR) receptors via PKC-dependent pathways. This enhanced interaction between SMRT and the receptors may in part explain the loss of receptor-driven transactivation during T cell activation.
Western Blot-Protein extracts were electrophoresed in a 10% Nu-PAGE BisTris gel using NuPAGE MES-SDS running buffer (Invitrogen) and transferred to a polyvinylidene difluoride membrane using XCell Blot Module (Invitrogen). The membrane was blocked with 5% milk overnight at 4°C and incubated with the appropriate antibody diluted in 3% milk. The protein was detected using the ECL Western blotting detection system from Amersham Biosciences.
Plasmids and Transfections-RARE-, RXRE-, TRE-, and VDR element (VDRE)-driven transcriptional activations were studied by transfection using luciferase-based reporter plasmids TKRARE-Luc, TKCRBPII-Luc, TKDR4-Luc, and TKDR3-Luc, respectively. Dr. K. Ozato, National Institutes of Health, Bethesda, provided TKCRBPII-Luc and TKRARE-Luc plasmids. TKDR3-Luc and TKDR4-Luc were constructed by cloning AAGGTTCAcgaGGTTCACG (sequence from mouse osteopontin promoter) and AGGACActcaAGGACActcaAGGA-CActcaAGGACA sequence, respectively, in KpnI and BglII sites of TK-Luc plasmid. TKCRBPII-M-Luc plasmid containing RXRE mutant sequence was constructed similarly to serve as negative control. Plasmids expressing human RAR␣, RXR␣, TR␣, and VDR were provided by Dr. P. Chambon  (10 7 ) were transfected by electroporation using a Gene Pulser II (Bio-Rad) at 0.250 kV and 975 microfarads as described (31,32). CV-1 cells were transfected using CalPhos Mammalian Transfection Kit from Clontech (Palo Alto. CA). After transfection cells were treated with the indicated reagents and time periods prior to harvest and determination of luciferase activity using the Luciferase Assay System (Promega, WI). Transfection efficiency was normalized to protein concentrations in the extracts. The values in the figures represent the mean of three independent experiments with standard error calculated for each value.
RNA Interference (RNAi)-Three short interfering RNA (siRNA) duplexes, containing 21 nucleotides with (2Ј-deoxy)-thymidine 3Ј overhangs, corresponding to nucleotides 477-497, 802-822, and 1954 -1974 of human PKC mRNA (GenBank TM accession number L07032), and a randomly scrambled sequence were obtained from Xeragon Inc. (Germantown, MD). Jurkat cells (10 7 ) were electroporated in Opti-MEM I reduced serum medium (Invitrogen) in 4-mm cuvettes at 0.25 kV and 975 microfarads with the indicated concentrations of the siRNA either in the presence or absence of various plasmid constructs. Cells were incubated in RPMI medium containing 10% fetal bovine serum and 12 h later treated with anti-CD3 and anti-CD28 antibodies for 24 h prior to harvest and determination of luciferase activity.

T Cell Activation by CD3 and CD28 Ligation Induces Loss of RARE-, RXRE-, and TRE-dependent Transactivation-We
have reported previously that CD3 cross-linking of Jurkat cells induced loss of RXRE-dependent transcription (31). In this report we extended this work to study the effect of activation by CD3 and costimulatory molecule CD28 on RARE-, RXRE-, TRE-, and VDRE-mediated transactivation. Jurkat cells were transfected with TKRARE-Luc, TKCRBPII-Luc, TKDR4-Luc, or TKDR3-Luc either in the presence or absence of receptor expressing plasmids and the cognate ligands. Twelve hours later the cells were treated with antibodies to CD3 and CD28 for 24 h. Fig. 1A shows that CD3/CD28 ligation resulted in a significant inhibition in transcriptional activation by RAR, RXR, and TR receptors. The inhibition was observed with endogenous as well as exogenously expressed receptors and was independent of the presence or absence of the ligand. Under similar conditions, VDR-mediated activation was only weakly inhibited. Little or no luciferase activity was observed when cells were transfected with TKCRBPII-M-Luc, a plasmid containing mutations in RXRE site, indicating the specificity of receptor-dependent activation assay. To demonstrate that this transcriptional loss was not a generalized phenomenon, a nonnuclear hormone receptor-based reporter assay was tested under similar conditions. Jurkat cells were transfected with NFAT-Luc plasmid for 12 h and then treated with antibodies to CD3 and CD28 for 24 h. Fig. 1B shows that CD3/CD28 engagement induced a strong induction of NFAT-dependent transcription.
Enhancement of SMRT-Receptor Interaction Upon CD3/ CD28 Engagement-In order to understand the mechanism of transcriptional loss induced by CD3/CD28 ligation, we explored the possibility that T cell activation may enhance the level of interaction between the receptors and the corepressor SMRT, thereby explaining the loss of transcriptional activation. Jurkat cells express readily detectable levels of SMRT protein that do not change significantly after ligation by CD3/CD28 crosslinking ( Fig. 2A). To study the change in the functional interaction between the SMRT and the receptors during T cell ligation, we used a mammalian two-hybrid system to quantitate the levels of interaction in Jurkat cells. A luciferase reporter construct containing yeast GAL4 (17-mer) DNA binding sequence, a mammalian expression vector pSG5 containing ligand binding domains of human RAR␣, RXR␣, TR, or VDR fused to the GAL4 activation domain (GAL4-AD), and a plasmid bearing specific regions of human SMRT (751-1495 or 1291-1495 amino acids corresponding to receptor interaction domain (RID) 1 ϩ 2 or RID 2, respectively) fused to the GAL4 DNA binding domain (GAL4-DBD-SMRT) in pSG5 (33) were transfected in Jurkat cells. After 12 h of incubation, the cells were treated with antibodies to CD3 and CD28 for 24 h. RAR␣ and TR␣ showed significant basal level interaction with GAL4-DBD-SMRT construct containing SMRT amino acids from 751 to 1495 (Fig. 2B), whereas a lower interaction was observed with RXR␣ (Fig. 2C). ATRA and T 3 , ligands specific for RAR and TR␣, respectively, inhibited the interaction. Interaction between RXR␣ and SMRT was not affected by RXR-specific ligand, 9-CRA, a finding also reported earlier in a non-T cell line (33). A very weak or no interaction was observed with VDR and SMRT construct containing 751-1495 amino acids and served as a negative control (Fig. 2C). SMRT construct containing amino acids from 1291 to 1495 (RID 2) showed very little interaction with RAR␣, and TR␣ no interaction (VDR), but a slightly higher interaction with RXR␣ (data not shown).
Interestingly, CD3/CD28 ligation strongly increased the basal level interaction between the SMRT and the ligand binding domains of RAR␣ and T 3 R␣ (Fig. 2B). There was a relatively lower increase in the interaction between SMRT and ligand binding domain of RXR␣ in cells treated with anti-CD3 and anti-CD28 antibodies (Fig. 2C). As with the basal interaction, the CD3/CD28-induced interaction of SMRT with RAR and TR␣ was attenuated by their cognate ligands, whereas the CD3/CD28-induced interaction between RXR␣ and SMRT was not affected by the RXR-specific ligand 9-CRA. There was no effect of CD3/CD28 ligation on the weak interaction seen between VDR and SMRT (Fig. 2C). Together, these data indicate that T cell engagement via CD3 and CD28 receptors induces a significant interaction between SMRT and the ligand binding domains of RAR␣, RXR␣, and TR␣.

. CD3 and CD28 ligation increases SMRT-receptor interaction.
A, Jurkat cells were treated with antibodies to CD3 and CD28 for 24 h, harvested, and cell extracts subjected to Western blot analysis using antibodies to SMRT. B and C, Jurkat cells were transfected with 2.5 g of GAL4 -17-mer-Luc, 2 g of GAL4-DBD-SMRT-(751-1495), and 2 g of PSG5-GAL4AD fusion constructs containing ligand binding domains of RAR␣ or TR␣ (B), RXR␣, or VDR (C). Twelve hours later the cells were treated as indicated for 24 h, and luciferase activity was measured as described under "Experimental Procedures." transfected cell extracts (Fig. 3B). Expression of a DA-PKC␣, a conventional PKC isoform, had no significant effect on the SMRT-receptor interaction (Fig. 3C). These data indicate that activation of PKC, a hallmark of T cell activation process, may play an important role in the positive modulation of SMRTreceptor interaction.
PKC Induces SMRT-Receptor Interaction in a T Cell-dependent Manner-PKC is not expressed ubiquitously and is found predominantly in hematopoietic cells and muscle (37). We next investigated whether PKC-dependent induction of SMRT-receptor interaction was a feature specific to T cells or could also be observed in PKC non-expressing cells. CV-1, a kidney fibroblast cell line which is not known to express PKC, was transfected with GAL4 -17-mer-Luc, GAL4-DBD-SMRT-(751-1495), and pSG5GAL4-AD-RAR␣, pSG5GAL4-AD-RXR␣, or pSG5GAL4-AD-T 3 R␣ in the presence or absence of WT-PKC, DA-PKC, or empty vector. A strong interaction between SMRT and RAR␣ was detected in CV-1 cells that did not show any increase in the presence of WT-or DA-PKC (Fig. 5). CV-1 cells also exhibited significant interaction between SMRT and ligand binding domains of RXR␣ and TR␣ that did not increase in the presence of WT-or DA-PKC. As expected from an earlier report (33), the interaction of SMRT with RAR␣ and TR␣ was abolished by ATRA and T 3 , ligands specific for RAR and TR, respectively, whereas SMRT interaction with RXR␣ did not decrease in the presence of 9-CRA (data not shown). Together, these data provide evidence that PKC, which is uniquely expressed in T cells, induces SMRT-receptor interaction involving factors that are T cell-specific.

Partial Loss of PKC Expression by PKC-specific RNAi Attenuates Basal and T Cell Activation-mediated SMRT-Receptor
Interaction-siRNA duplexes have been shown to silence gene expression in a sequence-specific manner by a process called RNA interference (RNAi) (38 -44). We tested three doublestranded siRNAs, directed at nucleotides 477-497, 802-822, and 1954 -1974 of human PKC mRNA, for silencing PKC expression. Transfection of Jurkat cells with these siRNAs showed varying degrees of inhibition of PKC expression within 24 -36 h of transfection. After 36 h, the levels of PKC slowly regained and reached normal levels within 72 h after transfection (data not shown). Of the three, siRNA directed against nucleotides 802-822 was the most effective in inhibiting PKC expression (50 -70% inhibition, Fig. 6 A) and was used in further studies. Transfection of a nonspecific scrambled RNA duplex did not significantly alter the levels of PKC.
We next studied the effect of siRNA-mediated silencing of PKC expression on basal and CD3/CD28-induced SMRT-receptor interaction. Jurkat cells were transfected with GAL4 -17-mer-Luc, GAL4-DBD-SMRT-(751-1495), and pSG5GAL4-AD-RAR␣, pSG5GAL4-AD-RXR␣, or pSG5GAL4-AD-T 3 R␣ in the presence or absence of 5 M siRNA corresponding to nucleotides 802-822 of PKC or a nonspecific scrambled sequence. After 12 h, the cells were treated with anti-CD3 and-CD28 antibodies and incubated for an additional 24 h. Fig. 6B demonstrates that expression of siRNA specific for PKC attenuated both basal as well as CD3/CD28-induced interaction between SMRT and all three receptors. Expression of a nonspecific scrambled RNA duplex did not significantly alter the SMRT-receptor interaction. Together these results indicate that PKC expression is essential for the basal as well as T cell activation-induced SMRT-receptor interaction.

DISCUSSION
Nuclear hormone receptors RAR, RXR, and TR play an important role in the development and homeostasis both by liganddependent activation and active repression of target genes. SMRT and a closely related corepressor N-CoR are involved in repression by binding to the receptors and forming complexes with various members of histone deacetylases, a key event in the deacetylation of histones and subsequent transcriptional repression (25). The role of nuclear receptor-driven active re- pression in T cell function is not known. A recent report (30) that T lymphocytes in N-CoR Ϫ/Ϫ mouse embryos are arrested at the double-negative stage emphasizes the importance of corepressors and active repression in T cell development.
In this paper we have provided evidence that T cell crosslinking by CD3 and CD28 surface molecules induces loss of RAR-, RXR-, and TR-driven transcriptional activation, indicating that T cell activation signals mediate active repression by these nuclear receptors. In an attempt to investigate the mechanism that would explain this inhibition, we have identified the role of SMRT in the activation-induced transcriptional repression and demonstrated that CD3/CD28 ligation induced SMRT interaction with all the three (RAR, RXR, and TR) receptors. CD3/CD28 ligation of Jurkat cells did not increase the levels of SMRT protein, excluding the possibility that an induced expression of SMRT may account for an enhanced SMRT-receptor interaction. However, this raised the possibility that T cell activation signals induce interaction indirectly by post-translational modifications of either the components of the repressional machinery or the factors that regulate their activity.
Phosphorylation and dephosphorylation are key post-translational events that are known to regulate the functional outcome of T cell activation and IL-2 production. Recently, PKC, due its highly T lymphocyte-specific expression, has emerged as one of the foremost protein kinases in the study of T cell function. We have reported previously (32) that PKC enhances calcineurin-dependent transactivation of RXRE-dependent promoters suggesting the role of this PKC isoform in nuclear hormone receptor-dependent signal transduction. By utilizing a number of different experimental approaches, we have identified PKC as a key regulator of SMRT-receptor interaction during T cell activation. One of these was a "reverse genetics" approach in which PKC expression was silenced in Jurkat cells by inducing a novel PKC-specific RNAi pathway. RNAi, a recently described phenomenon, is emerging as a powerful tool to study the function of genes. Together, these observations support a model in which CD3/CD28 ligation, or signals that activate PKC, induce a strong interaction between SMRT and nuclear receptors, a phenomenon that may in part be responsible for the observed transcriptional repression following T cell activation. Although we have not studied the effect of T cell ligation on N-CoR-nuclear receptor interactions, a similar increase may be speculated because of a close resemblance between SMRT and N-CoR proteins in structure and function.
The mechanism that explains CD3/CD28-induced and PKCdependent SMRT-receptor interaction remains unknown. Direct phosphorylation of SMRT or the receptor protein by PKC, if it occurs, may stabilize interactions between the interacting partners. Alternatively, an unknown auxiliary protein may act as a substrate for PKC and directly or indirectly activate or stabilize the SMRT-receptor interaction. In an earlier study (31), we reported that activation of the JNK pathway resulted in the attenuation of RXRE-driven transcription in Jurkat cells. T cell activation by CD3/CD28 cross-linking or expression of DA-PKC are known to activate JNK pathway in T cells. Is the induction of PKC-mediated SMRT-receptor interaction a direct consequence of PKC-dependent JNK activation? Earlier data have shown that activation of JNK pathways inhibits SMRT-receptor interaction in CV-1 cells (45). Expression of DA-PKC or CD3/CD28 ligation in T cells may induce SMRTreceptor interaction involving alternate pathways. Whatever is the mechanism of PKC-mediated induction of SMRT function following T cells activation, it is noteworthy that this PKC isoform uniquely functions in T cells in modulating SMRT function. The failure of DA-PKC to induce SMRT-nuclear receptor interaction in CV-1 cells emphasizes this specificity and also indicates that PKC may not function alone in T cells but needs active participation of other unknown T cell-specific effectors. A robust basal level SMRT-receptor interaction observed in CV-1 cells that does not respond to PKC also indicates that PKC-independent regulation of receptor-corepressor interaction is operative in cells that may not express PKC.
Our data suggest that T cell activation-induced inhibition of RAR-, RXR-, and TR-dependent transactivation may involve the participation of active repressional machinery in which SMRT-receptor interaction is strongly enhanced by PKC-dependent signals. In such a scenario, an increased interaction between receptors and the SMRT may stabilize the components of repressional apparatus on the receptor-bound promoters and induce repression. SMRT and N-CoR, although initially known to mediate repression by RAR, RXR, and TR, have also been found to inhibit transcription driven by a number of other unrelated transcription factors (25). Of particular interest is SMRT-mediated suppression of AP-1 and NFB (46), factors known to be involved in the activation of the IL-2 promoter, and interaction with BCL-6, a mediator of apoptosis in T cells (47). Whether PKC plays a role in modulating SMRT interaction with these factors is not known but remains a possibility. It is clear that active repression mediated by corepressors like SMRT and N-CoR may have important physiological consequences not only in the previously described T cell development (30), but also in antigen-driven T cell activation function. Characterization of corepressor function in determining T cell responses to various signals is an area that remains open for investigation.
To summarize, the studies reported in this paper represent an attempt to bridge the gap between our understanding of the phenomenon of nuclear receptor-mediated active transcriptional repression and its role in T cell activation function. We have shown a direct link between the two processes and also identified PKC, a T cell-specific PKC isoform, as a key modulator of nuclear receptor-mediated active repression. The data presented should help to unravel the role of active repressional machinery in T cell activation function.