Protein Kinase D1 Phosphorylates HDAC7 and Induces Its Nuclear Export after T-cell Receptor Activation*

HDAC7, a class II histone deacetylase that is highly expressed in thymocytes, inhibits both transcription of the orphan steroid nuclear receptor Nur77 and induction of apoptosis in response to activation of the T-cell receptor (TCR). Here, we report that HDAC7 is exported to the cytoplasm by a calcium-independent signaling pathway after TCR activation. Protein kinase D1 (PKD1) was activated after TCR engagement, interacted with HDAC7, and phosphorylated three serines (Ser155, Ser318, and Ser448) at its N terminus, leading to its export from the nucleus. Mutation of Ser155, Ser318, and Ser448 blocked the nucleocytoplasmic shuttling of HDAC7 in response to TCR activation, as did overexpression of a kinase-inactive form of PKD1. Consistent with the regulatory role of HDAC7 in Nur77 expression, PKD1 activation led to the transcriptional activation of Nur77 via myocyte enhancer factor 2-binding sites in its promoter. In a mouse model of negative selection, PKD1 was activated during thymocyte activation. These observations indicate that PKD1 regulates the expression of Nur77 during thymocyte activation at least in part by phosphorylating HDAC7.

Histone acetylation and deacetylation are important in modifying chromatin structure and in regulating gene expression in eukaryotes. Mammalian histone deacetylases (HDACs) 1 are divided into three classes based on their homology to yeast proteins. Class I HDACs are homologous to Rpd3; class II HDACs are related to Hda1; and class III HDACs are homologous to Sir2. Class II HDACs are further subdivided into class IIa (HDAC4, HDAC5, HDAC7, and HDAC9 and the HDAC9 splice variant MITR) and class IIb (HDAC6 and HDAC10) (reviewed in Ref. 1).
The class IIa HDACs possess a conserved C-terminal catalytic HDAC domain and interact with myocyte enhancer factor 2 (MEF2) transcription factors through an N-terminal 17-amino acid motif. This interaction leads to the recruitment of class IIa HDACs to select promoters, where MEF2 is bound, resulting in the repression of its transcriptional activity. The repressive activity of class IIa HDACs is tightly regulated by nucleocytoplasmic shuttling and by their phosphorylation-dependent association with the intracellular 14-3-3 proteins (2)(3)(4)(5). Phosphorylation of conserved residues in the N-terminal regions of HDAC4, HDAC5, HDAC7, and HDAC9 in response to cellular signals leads to interaction with 14-3-3 proteins, dissociation of the class IIa HDAC⅐MEF2 complexes, and a conformational change culminating in export from the nucleus (6,7).
Given the central role of phosphorylation in the regulation of class IIa HDAC activity, there is considerable interest in the identification of the responsible kinase(s). It has been reported that the calcium/calmodulin-dependent kinase (CaMK) phosphorylates HDAC4, HDAC5, and HDAC9, leading to their export from the nucleus and to gene activation in skeletal and cardiac myocytes (6 -11). However, the endogenous HDAC kinase activity in cardiac myocytes is resistant to pharmacological inhibitors of CaMK, suggesting that other kinases may be responsible for HDAC phosphorylation (12).
We recently reported that HDAC7 is expressed at high levels in thymocytes during the CD4 ϩ CD8 ϩ double-positive stage (14). In resting thymocytes, HDAC7 is localized in the cell nucleus and represses the expression of the orphan steroid nuclear receptor Nur77 by interacting with the transcription factor MEF2D, which binds constitutively to the nur77 promoter (13,14). nur77, an immediate-early gene, is up-regulated in response to T-cell receptor (TCR) activation and has been implicated in the negative selection (apoptosis) of T cells (15,16). After TCR activation, HDAC7 is exported to the cytoplasm, leading to the derepression of the nur77 promoter and the induction of apoptosis. Interestingly, mutation of three serine residues at the N terminus of HDAC7 inhibits its nucleocytoplasmic shuttling in response to TCR activation and also suppresses TCR-induced apoptosis (14). These observations indicate that an intracellular signaling pathway, originating at the TCR, is involved in the phosphorylation and subcellular localization of HDAC7 in thymocytes and that this signal controls the apoptosis of thymocytes in response to TCR activation.
T cells use a complex array of signal transduction pathways to control their proliferation, differentiation, survival, and apoptosis. TCR engagement leads to the activation of phospholipase C␥1, which in turn initiates the activation of two well characterized major signaling pathways. Phospholipase C␥1 induces the cleavage of phosphatidylinositol in the plasma membrane, producing inositol triphosphate and diacylglycerol. Inositol polyphosphates increase intracellular calcium levels by binding to specific receptors in the endoplasmic reticulum, resulting in the activation of the calcineurin/NF-AT cascade. Diacylglycerol (DAG) induces a calcium-independent signal transduction pathway, which comprises the activation of protein kinase C (PKC) and protein kinase D (PKD) signaling modules (reviewed in Ref. 17). PKD1, also called PKC, is a serine/threonine kinase that belongs to a new family of protein kinases with two other members, PKD2 and PKD3/PKC. PKD1, the main isoform expressed in T cells (18), is mainly activated by a phospholipase C-DAG-PKC signal transduction pathway (reviewed in Ref. 19). PKD1 activation is mediated by the PKC-dependent phosphorylation of Ser 744 and Ser 748 in the activation loop of its catalytic domain (20,21). PKD1 is activated in B-and T-cells after engagement of their respective receptors (18,(22)(23)(24). However, the specific substrates for PKD and its exact function in lymphocytes remain unclear.
Here, we show that a calcium-independent signaling pathway is responsible for the nucleocytoplasmic shuttling of HDAC7 in a thymocyte hybridoma cell line (DO11.10) after TCR activation. Moreover, PKD1, which is activated after TCR engagement, interacts with and phosphorylates HDAC7, leading to its nuclear export and to the activation of Nur77 transcription. Finally, we show that PKD1 is activated by TCR activation in vivo in a mouse model of negative selection.

EXPERIMENTAL PROCEDURES
Plasmids-The pcDNA3.1-based expression vector for FLAG-tagged human HDAC7 has been described (25). C-terminal green fluorescent protein (GFP) fusions were constructed in pEGFP-N1 (Clontech). Deletion constructs of HDAC7 were generated by PCR and the cloning procedures described previously (25). Site-directed mutagenesis was performed with a QuikChange kit (Stratagene, La Jolla, CA). All mutations were verified by DNA sequencing. The glutathione S-transferase (GST) fusion proteins containing the N or C terminus of HDAC7 have been described (25). The luciferase reporter plasmid driven by the nur77 promoter (pNur77-Luc) was generated by cloning the Ϫ3800 to ϩ87 genomic sequences of the nur77 promoter (a kind gift from Astar Winoto, University of California, Berkeley, CA) (13) into pGL2-Basic (Promega). Minimal wild-type and mutant MEF2 reporter constructs (pMEF2wt-Luc and pMEF2mt-Luc) were as described (13). PKD1 expression vectors were provided by Dr. Alex Toker (Beth Israel Deaconess Medical Center, Boston, MA).
Cell Culture, Transfections, and Reporter Assay-DO11.10 T-cell hybridomas and 293T cells were grown at 37°C in RPMI 1640 medium and Dulbecco's modified Eagle's medium, respectively, supplemented with 10% fetal bovine serum, 2 mM glutamine, and 50 units/ml streptomycin/penicillin. DO11.10 cells were transfected by the DEAE-dextran/chloroquine method. The DNA concentration was kept constant in different samples by using the corresponding empty vector. In some cases, cells were treated with phorbol 12-myristate 13-acetate (PMA; 10 ng/ml) and ionomycin (0.5 M) for 4 h, beginning 36 h after transfection, and harvested for reporter assays. All transfections represent the result of at least three independent experiments performed in triplicate. Luciferase reporter assays were performed with the Dual-Luciferase reporter assay system (Promega) using an elongation factor 1␣ promoterdriven Renilla luciferase expression vector as an internal control. 293T cells were transfected by the standard calcium phosphate precipitation method. For anti-CD3/CD28 antibody stimulation, monoclonal antibodies 500A2 and 37.51 were bound to the culture flask by incubating a 1:2000 dilution in phosphate-buffered saline overnight at 4°C, followed by three rinses with phosphate-buffered saline.
Protein Kinase Assays-Immunoprecipitated PKD1 was incubated with myelin basic protein or purified GST-HDAC7 fusion proteins. Phosphorylation reactions were performed in 30 l of PKD1 kinase buffer supplemented with 20 M ATP and 5 Ci of [␥-32 P]ATP at 30°C for 30 min. Reactions were stopped by the addition of 4ϫ Laemmli sample buffer and resolved by SDS-PAGE on 8% gels.
SDS-PAGE and Western Blotting-SDS-PAGE and Western blot analysis were performed according to standard procedures (27). Western blots were developed with an ECL detection kit (Amersham Biosciences). Anti-PKD1 and anti-hemagglutinin (HA) antibodies were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). Anti-tubulin antibody was from Sigma. Anti-mouse Nur77 antibody was from Pharmingen (San Diego, CA). The antibody specific for PKD1 phosphorylated at Ser 744 and Ser 748 was from Cell Signaling (Beverly, MA).
GST Fusion Proteins and Pull-down Assays-These assays were performed as described (28).
Peptide Injection of Mice-DO11. 10 transgenic mice have been described previously (29). Balb/c wild-type mice and DO11.10 transgenic mice were injected intraperitoneally with 250 l of a sterile 100 M solution of ovalbumin peptide (ISQAVHAHAEINEAGR). Mice were sacrificed at the indicated times after injection. (14) has documented that signals emanating from the TCR lead to the phosphorylation of HDAC7 and to its export from the nucleus. Accordingly, treatment of the thymocyte hybridoma cell line DO11.10 with phorbol esters and calcium ionophores (PMA/ ionomycin), a treatment that mimics the two main pathways activated by the TCR, led to the nuclear export of an HDAC7-GFP fusion protein. In contrast, an HDAC7 mutant in which phosphorylation was prevented by substituting Ser 155 , Ser 318 , and Ser 448 with alanines (HDAC7⌬P-GFP) remained in the nucleus (Fig. 1).

A Calcium-independent Pathway Regulates Nucleocytoplasmic Shuttling of HDAC7-Our previous work
Surprisingly, treatment with PMA alone (but not ionomycin) led to the nuclear export of HDAC7 ( Fig. 1), suggesting that a calcium-independent pathway is responsible for the nucleocytoplasmic shuttling of HDAC7 after TCR activation. The HDAC7 mutant was also unresponsive to PMA treatment, indicating that the conserved serines in HDAC7 are potential phosphorylation targets for this intracellular signaling cascade.
A PKC/PKD-dependent Pathway Regulates Nucleocytoplasmic Shuttling of HDAC7 after TCR Activation-To further define the factors involved in the nucleocytoplasmic shuttling of HDAC7, we used a number of specific inhibitors of distinct signaling pathways. Before PMA was added, DO11.10 cells expressing HDAC7-GFP were treated for 30 min with Gö6976, an inhibitor that targets both calcium-dependent PKC isoforms and PKD; with Gö6983 and GF109203X, two general inhibitors of PKCs; with KN62, a specific inhibitor of CaMK; or with cyclosporin A (CsA), a specific inhibitor of calcineurin. Pretreatment with Gö6976, Gö6983, and GF109203X completely inhibited HDAC7 nuclear export mediated by PMA (Fig. 2). In contrast, KN62 and CsA, which inhibit calcium-dependent pathways, had no effect on the cellular distribution of HDAC7 in response to PMA. Importantly, similar results were obtained after cross-linking the T-cell antigen receptor with anti-CD3/ CD28 antibodies, a more physiologically relevant stimulus (Fig. 2). These observations indicate that CaMK and calcineurin are not involved in the nucleocytoplasmic shuttling of HDAC7 after TCR activation and also that a PKC/PKD-dependent mechanism is responsible for the nucleocytoplasmic shuttling of HDAC7 after TCR activation.
A PKC/PKD-dependent Pathway Controls Nur77 Expression and Transcriptional Activation after TCR Engagement-In view of the key role of HDAC7 in regulating Nur77 expression after TCR activation, we tested the role of a PKC/PKD signaling pathway in Nur77 induction after TCR activation. In agreement with the results shown in Fig. 2, pretreatment with the inhibitors Gö6976, Gö6983, and GF109203X suppressed the induction of Nur77 protein expression in response to PMA, whereas KN62 and CsA surprisingly potentiated the effect of PMA (Fig. 3A). Similar results were obtained in cells transfected with a construct containing the nur77 promoter driving a luciferase reporter. The PMA-mediated transcriptional activation of the nur77 promoter was inhibited by the PKC inhibitors and potentiated by KN62 and CsA (Fig. 3B).
The effects of the same inhibitors were tested after the induction of Nur77 expression by anti-CD3/CD28 antibodies. Pretreatment of DO11.10 cells with the PKC/PKD inhibitor (Gö6976) abolished the induction of Nur77 at the level of both protein expression and promoter induction (Fig. 3, C and D). The other inhibitors (Gö6983, GF109203X, and CsA) partially suppressed Nur77 expression and its promoter (Fig. 3, C and  D). The calmodulin inhibitor (KN62) had no effect on the induction of Nur77 after TCR activation, further confirming that CaMKs are not involved in the HDAC7-dependent regulation of Nur77 by PMA or the TCR (Fig. 3, C and D).
PKD1 Is Activated in Thymocyte Hybridoma Cells after TCR Engagement-Next, we analyzed the amino acid sequences surrounding the conserved serines in HDAC7 for homology to known consensus protein kinase sites. The three conserved serines in HDAC7 (Ser 155 , Ser 318 , and Ser 448 ) and other class IIa HDACs are consensus sites for CaMK ((R/K)XX(S/T)). However, as shown above, chemical blockage of CaMK did not suppress HDAC7 nuclear export after TCR activation. Interestingly, leucine was present at position Ϫ5 relative to each serine (Fig. 4A), a requirement for substrates of PKD1 (30).
To determine whether PKD1 is activated by antigen receptor signals, we used an antibody specific for the active/phosphoryl-  (Fig. 4B). PMA alone (but not ionomycin) had a similar effect (Fig. 4B), demonstrating that a calcium-independent mechanism is involved. Cell treatment with anti-CD3/CD28 antibody also activated PKD1 (Fig. 4B). To further confirm the activation of PKD1 in DO11.10 cells, we performed an in vitro kinase assay. The cells were treated with PMA/ionomycin, and endogenous PKD1 was immunoprecipitated with anti-PKD1 antibody, followed by a kinase assay using myelin basic protein as substrate. Endogenous PKD1 was activated after TCR activation, as indicated by the PKD1 autophosphorylation band and the phosphorylation of myelin basic protein (Fig. 4C). These results demonstrate that PKD1 is  (HDAC4, HDAC5, HDAC7, and HDAC9) are aligned. B, DO11.10 cells were treated with PMA/ionomycin, PMA, ionomycin, or anti-CD3/CD28 antibodies. Total cell lysates were prepared 0, 10, 30, and 60 min after activation and analyzed by Western blotting with an antibody specific for PKD1 phosphorylated at Ser 744 and Ser 748 (Phospho-PKD1) and an antibody specific for total PKD1. Tubulin levels were measured by Western blotting of the same samples. C, DO11.10 cells were treated with PMA/ionomycin, and total cell lysates were prepared 0, 10, 30, and 60 min after activation. PKD1 activity was determined by immunoprecipitation of endogenous PKD1 from the cell lysates, followed by a kinase assay using the substrate myelin basic protein (MBP). activated in DO11.10 cells by TCR activation independently of calcium signaling.
PKD1 Interacts with and Phosphorylates HDAC7-To test whether HDAC7 interacts with PKD1, we incubated GST fusion proteins containing the N terminus (GST-HDAC7(N-ter)) or the C terminus (GST-HDAC7(C-ter)) of HDAC7 after expression in bacteria with extracts of DO11.10 cells, either untreated or treated with PMA/ionomycin. Immunoblotting with an antibody specific for PKD1 revealed that endogenous PKD1 bound to GST-HDAC7(N-ter), but not to GST-HDAC7(C-ter) (Fig. 5A). HDAC7 bound to PKD1 present in resting or activated extracts (PMA/ionomycin) (Fig. 5A), suggesting that enzymatic activation of PKD1 is not necessary for HDAC7 binding. To further confirm the interaction of HDAC7 with PKD1, FLAG-HDAC7 and HA-PKD1 expression plasmids were coexpressed in 293T cells. FLAG-HDAC7 was immunoprecipitated, and the immunoprecipitate was probed for the presence of PKD1. Both proteins interacted, regardless of the order of immunoprecipitation (Fig. 5B).
To investigate whether PKD1 phosphorylates HDAC7 directly, we immunoprecipitated endogenous PKD1 from DO11.10 cells, untreated or treated with PMA/ionomycin, and used it for kinase activity assay with GST-HDAC7(N-ter) or GST-HDAC7(C-ter) as substrate. PKD1 was activated after PMA/ionomycin treatment, as indicated by the PKD1 autophosphorylation band (Fig. 5C). As predicted, only the N-terminal region of HDAC7, which bears the conserved residues Ser 155 , Ser 318 , and Ser 448 , was phosphorylated by PKD1. To confirm that the predicted PKD1 phosphorylation sites were indeed phosphorylated, we performed an in vitro kinase assay using a GST-HDAC7 construct in which the conserved residues Ser 155 , Ser 318 , and Ser 448 were mutated to alanines (GST-HDAC7⌬P) as substrate. The HDAC7 triple phosphorylation mutant was not phosphorylated by PKD1 after cell treatment with PMA/ionomycin (Fig. 5D). These observations demonstrate the ability of PKD1 to bind and phosphorylate HDAC7 and the requirement for the conserved HDAC7 residues Ser 155 , Ser 318 , and Ser 448 .
PKD1 Activity Modulates HDAC7 Nucleocytoplasmic Shuttling-Because HDAC7 Ser 155 , Ser 318 , and Ser 448 were required for nuclear export after TCR activation, we tested whether PKD1 could regulate the subcellular distribution of HDAC7. DO11.10 cells were cotransfected with HDAC7-GFP or the phosphorylation mutant HDAC7⌬P-GFP fusion construct as well as an expression vector for wild-type, constitutively active (PKD1 SS/EE), or kinase-inactive (PKD1 SS/AA) PKD1. HDAC7 was exported to the cytoplasm after PMA treat-FIG. 6. PKD1 induces nucleocytoplasmic shuttling of HDAC7. DO11.10 cells were transfected with the HDAC7-GFP or HDAC7⌬P-GFP expression vector and, where indicated, with a plasmid expressing the wild-type (PKD1 wt), constitutively active (PKD1 SS/EE), or kinaseinactive (PKD1 SS/AA) form of PKD1. Thirty-six hours after transfection, cells were left untreated or treated with PMA, and the intracellular location of GFP fusion proteins was monitored by confocal immunofluorescence microscopy after 1 h of treatment. One or two cells are shown for each experimental condition. They are representative of Ͼ90% of the total population.

FIG. 5. PKD1 interacts with and phosphorylates HDAC7.
A, binding of PKD1 to GST-HDAC7(N-ter) or GST-HDAC7(C-ter) was measured in pulldown assays with endogenous PKD1 from DO11.10 cells that were untreated or treated with PMA/ionomycin for 30 min. B, total cellular extracts were prepared from DO11.10 cells transfected with empty vector, FLAG-HDAC7, HA-PKD1, or both FLAG-HDAC7 and HA-PKD1. These extracts were subjected to immunoprecipitation (IP) with anti-FLAG antibody and analyzed by immunoblotting (IB) with anti-HA antibody. Cell lysates were analyzed by Western blotting with anti-FLAG and anti-HA antibodies. C, DO11.10 cells were treated with PMA/ ionomycin, and total cell lysates were prepared 0, 10, 30, 60, and 120 min after activation. HDAC7 phosphorylation was analyzed by immunoprecipitating endogenous PKD1, followed by a kinase assay using GST-HDAC7(N-ter) or GST-HDAC7(Cter) as substrate. D, DO11.10 cells were treated as described for C. PKD1 kinase assays were carried out using GST-HDAC7(N-ter) or GST-HDAC7⌬P(N-ter) as substrate. ment in the absence or presence of wild-type PKD1 (Fig. 6). In contrast, overexpression of the constitutively active form of PKD1 (PKD1 SS/EE) led to export of HDAC7 from the nucleus to the cytoplasm, even in untreated cells (Fig. 6). Importantly, the kinase-inactive form of PKD1 (PKD1 SS/AA) inhibited HDAC7 nuclear export induced by PMA (Fig. 6). The subcellular localization of HDAC7⌬P-GFP was not affected by any stimulus, and the protein was observed in the nucleus under all experimental conditions (Fig. 6). These results demonstrate that PKD1 activity is both necessary and sufficient for the regulation of the nucleocytoplasmic shuttling of HDAC7.
PKD1 Overexpression Activates Nur77 Expression through MEF2-To determine whether PKD1-induced HDAC7 nuclear export is sufficient to activate the nur77 promoter, we cotransfected DO11.10 cells with a nur77 promoter-reporter construct and a vector encoding constitutively active PKD1 (PKD1 SS/ EE). PKD1 SS/EE increased the basal transcriptional activity of the nur77 promoter by ϳ5-fold; activation was more modest in the presence of PMA/ionomycin (Fig. 7A). We reported previously that HDAC7 overexpression can block the TCR-mediated activation of the nur77 promoter (14). To test whether PKD1 overexpression can block the repressive activity of HDAC7, we transfected DO11.10 cells with the HDAC7 or HDAC7⌬P expression vector. As expected, overexpression of HDAC7 inhibited the transcriptional activation of Nur77 (75%) induced by PMA/ionomycin (Fig. 7B). Overexpression of constitutively active PKD1 reversed the repressive effect of wild-type HDAC7 (Fig. 7B). Importantly, PKD1 SS/EE had no effect on the transcriptional repression mediated by the HDAC7⌬P phosphorylation mutant (Fig. 7B). To confirm that MEF2 is a target of the PKD1-dependent activation, we cotransfected DO11.10 cells with a reporter construct containing four MEF2D-binding sites upstream of a minimal promoter (pMEF2wt-Luc) and the PKD1 SS/EE expression vector. A construct containing mutant MEF2D-binding sites served as a control (pMEF2mt-Luc). PMA/ionomycin increased the activity of the pMEF2wt-Luc reporter construct by 18-fold, but had no effect on the pMEF2mt-Luc construct (Fig. 7C). Overexpression of PKD1 SS/EE caused a 12-fold activation of pMEF2wt-Luc under basal conditions and further increased its activity in response to PMA/ionomycin (Fig. 7C). As predicted, active PKD1 had no effect on pMEF2mt-Luc (Fig. 7C).

FIG. 8. PKD1 is activated during negative selection of T cells.
A, wild-type (wt) and DO11.10 transgenic (Tg) mice were intraperitoneally injected with 250 l of 100 M ovalbumin peptide. Mice were sacrificed at the indicated time points post-injection, and thymocytes were harvested. Total thymocyte extracts were analyzed by Western blotting with antisera against Nur77 and tubulin. B, thymocyte lysates were analyzed by Western blotting with an antibody specific for PKD1 phosphorylated at Ser 744 and Ser 748 (Phospho-PKD1) and with an antibody specific for total PKD1 antibody. C, HDAC7 phosphorylation by PKD1 was analyzed using a kinase assay. Endogenous PKD1 was immunoprecipitated from thymocyte lysates, and a kinase assay was performed with GST-HDAC7(N-ter) as substrate.
major histocompatibility complex-restricted TCR specific for a peptide sequence derived from ovalbumin. CD4 ϩ T-cells carrying this TCR undergo massive apoptosis (negative selection) after injection of the ovalbumin peptide in mice (29).
To assess the activation status of PKD1 in vivo, we injected wild-type or DO11.10 transgenic mice with the ovalbumin peptide. Nur77 expression was rapidly and specifically induced in DO11.10 mice, but not in wild-type controls (Fig. 8A). PKD1 was activated in thymocytes from DO11.10 transgenic mice after ovalbumin peptide injection, but not in control mice (Fig.  8B). The activation of PKD1 followed kinetics similar to that of Nur77 expression. Moreover, endogenous PKD1 immunoprecipitated from in vivo activated DO11.10 transgenic thymocytes could phosphorylate recombinant GST-HDAC7. These results demonstrate that PKD1 is activated during negative selection and could play a significant role in the observed derepression of Nur77 expression during this process. DISCUSSION This study shows that the export of HDAC7 from the nucleus is induced by a calcium-independent mechanism after TCR activation. We have identified the serine/threonine kinase PKD1 as a key player in the nuclear-cytoplasmic transfer of HDAC7. PKD1 is activated in DO11.10 cells in a calciumindependent manner after TCR activation or PMA treatment. PKD1 interacts with and phosphorylates HDAC7, leading to its nuclear export and to the transcriptional activation of Nur77 (Fig. 9). In agreement with our previous observation (14) that HDAC7 plays a critical role in the transcriptional repression of the orphan receptor Nur77, PKD1 plays a critical role in the transcriptional regulation of Nur77. Importantly, a kinaseinactive form of PKD1 suppresses the nucleocytoplasmic transfer of HDAC7 in response to TCR activation. In agreement with these observations, we observed that PKD1 is activated during an in vivo model of negative selection of T cells, a process driven in part by Nur77.
Strong TCR activation leads to the induction of Nur77 in T-cell hybridomas and thymocytes, leading to apoptosis (15,16). In vivo, Nur77 plays a major role in negative selection, an apoptotic process responsible for the elimination of developing self-reactive thymocytes (reviewed in Ref. 31). Two MEF2binding sites in the nur77 promoter are critical for its regulation of Nur77 expression (13,32). The transcriptional corepressor Cabin-1 represses MEF2-dependent Nur77 transcriptional activity by recruiting mSin3, HDAC1, and HDAC2 and by competing with the coactivator p300 for MEF2 binding (32). However, a mutant transgenic mouse expressing Cabin-1 lacking the MEF2-binding domain shows normal thymocyte development and apoptosis (33). This observation suggests that other proteins participate in the repression of Nur77 during thymocyte development and that the role of Cabin-1 might be restricted to a subpopulation of lymphocytes. Our observations are consistent with the model that HDAC7 contributes to the repression of Nur77 in developing thymocytes, either in cooperation with or independently of Cabin-1.
We recently reported that HDAC7 is the predominant class IIa HDAC in thymocytes (14). HDAC7 interacts directly with MEF2 and inhibits the expression of Nur77 through MEF2D in resting T cells. The repressive activity of HDAC7 is regulated at the level of its subcellular distribution between the nucleus and the cytoplasm. While in the nucleus of resting thymocytes, HDAC7 is presumably bound to MEF2 at promoter sites and participates in the repression of their transcriptional activity. TCR activation leads to the nuclear export of HDAC7 and the derepression of promoters previously repressed by HDAC7. Importantly, the subcellular localization of HDAC7 mutated at three serine residues (Ser 155 , Ser 318 , and Ser 448 ) is not modified in response to TCR activation. This mutant remains in the nucleus after TCR activation and inhibits TCR-mediated apoptosis. These three serine residues are highly conserved among all class IIa HDACs, undergo reversible phosphorylation, and control the nucleocytoplasmic shuttling of these factors. Phosphorylation of these residues unmasks a nuclear export signal and leads to cytoplasmic localization of class IIa HDACs. Phosphorylated class IIa HDACs are bound to 14-3-3 proteins and are present in the cytoplasm dissociated from the N-Cor⅐SMRT⅐HDAC3 corepressor complex. These observations therefore identify the phosphorylation of HDAC7 and the kinase(s) that mediate this event as key in the transcriptional activation of the nur77 promoter and other promoters controlled by HDAC7.
CaMK can phosphorylate HDAC4 and HDAC5, and the expression of constitutively active CaMK induces nucleocytoplasmic shuttling in myocytes (6 -11). However, the endogenous HDAC kinase activity in cardiac myocytes is resistant to pharmacological inhibitors of CaMK, suggesting that other kinases are involved in the phosphorylation and regulation of the subcellular distribution of HDACs in these cells (12). While this manuscript was in preparation, Vega et al. (34) reported that PKC and PKD are involved in the nuclear export of HDAC5 in cardiac myocytes in response to hypertrophic agonists. In that experimental system, both ionomycin and PMA independently induce HDAC5 nucleocytoplasmic shuttling. In contrast, we have reported that PMA alone (and not ionomycin) induced HDAC7 nucleocytoplasmic shuttling. These observations indicate that PKD1 is a novel regulator of the nucleocytoplasmic shuttling of class IIa HDACs and reveal an interesting divergence in the regulatory pathways controlling the nucleocytoplasmic shuttling of different class IIa HDACs. Other differences in nucleocytoplasmic shuttling between class IIa HDACs and between cell types have been noted. In transfected fibroblasts, exogenously overexpressed HDAC5 is exclusively nuclear, whereas HDAC4 is cytoplasmic in 50 -80% of cells, suggesting that HDAC4 and HDAC5 are differentially phos- FIG. 9. The signaling pathway responsible for nucleocytoplasmic shuttling of HDAC7 after TCR activation. In resting T cells, HDAC7 represses the transcriptional activation of Nur77 by binding to the transcription factor MEF2D. After TCR activation, PKD1 is activated, presumably by a phospholipase C␥ (PLC␥)-DAG-PKC signal transduction pathway. PKD1 phosphorylates HDAC7, leading to its interaction with 14-3-3 proteins and to its export from the nucleus, resulting in the activation of Nur77 and the induction of apoptosis. phorylated in vivo (10). HDAC7 is predominantly nuclear in CV-1 and HeLa cells (4,35), whereas it is distributed in both the nucleus and the cytoplasm of T cells (14).
PKD1 is a member of a new family of DAG-stimulated serine/ threonine protein kinases that consists of three members: PKD1, PKD2, and PKD3. Based on the sequence similarity of their kinase domains, PKD family members are classified as a subgroup of the CaMKs (reviewed in Ref. 19). PKD family members share a common mechanism of activation, but vary in their tissue distribution and subcellular localizations. PKD1 is most abundant in hematopoietic cells and is the predominant PKD isoform expressed in T cells. PKD1 is activated by the TCR and is localized either at the plasma membrane or in the cytosol of lymphocytes (18). PKD1 targeted to either the membrane or the cytosol induces differentiation in transgenic mice reminiscent of ␤-selection: down-regulation of CD25 and upregulation of CD2 and CD5 (18). Strikingly, membrane (but not cytosolic) PKD1 can induce the expression of CD8 and CD4 in RAG recombinase-null mice, and cytosolic (but not membrane) PKD1 suppresses V␤-to-DJ␤ rearrangements of the TCR ␤-chain locus in wild-type T cells, suggesting their involvement in the signal mediating TCR allelic exclusion. The fact that HDAC7 is regulated by PKD1 suggests the intriguing possibility that the nucleocytoplasmic shuttling of HDAC7 could affect not only late events in thymocyte differentiation, such as negative and positive selection, but also earlier events, such as ␤-selection. Our observations suggest the existence of a nuclear form of PKD1 in activated thymocytes. Although this has not been formally demonstrated, we were able to co-immunoprecipitate HDAC7 and PKD1, indicating that the two proteins interact. In addition, PKD translocates to the nucleus under defined physiological conditions, e.g. in response to G proteincoupled signaling (36).
TCR engagement leads to the activation of a wide array of signaling events that mediate the transcriptional activation or repression of many target genes. Key in this process is the activation of phospholipase C␥, which induces the production of inositol triphosphate and DAG, resulting in an increase in intracellular calcium levels and in the activation of a number of kinases, including PKC. Accordingly, many of the events induced by TCR activation can be recapitulated by treating target cells with phorbol esters (a DAG mimic) and ionomycin (a calcium ionophore). Considerable experimental evidence suggests an important role for calcium signaling in Nur77 induction after TCR activation (13,32). This activation is mediated in part through the recruitment of NF-AT and in part through the dissociation of Cabin-1 from MEF2D. In DO11.10 cells, the addition of CsA, an inhibitor of the calcium-dependent phosphatase calcineurin, partially inhibited the late expression of Nur77 in response to PMA/ionomycin or anti-CD3 antibody and also prevented subsequent cell death. However, it is also clear that calcium-independent processes play a critical role in activating the nur77 promoter. For example, treatment of DO11.10 cells with PMA alone induced early Nur77 expression. Calcium-independent signals activate the nur77 promoter in part by recruiting the extracellular signal-regulated kinase ERK5 to the promoter (37). The induction of HDAC7 nucleocytoplasmic shuttling by PKD1 phosphorylation in response to TCR activation represents a novel calcium-independent signal participating in the induction of the nur77 promoter.
Future efforts will be directed toward dissecting the molecular events linking TCR cross-linking and the activation of PKD1. It has been reported that PKD1 is a downstream target of PKC in COS-7 cells and Jurkat T-lymphocytes (23). PKC activation has been linked to the induction of apoptosis in double-positive (CD4 ϩ CD8 ϩ ) thymocytes in response to TCR cross-linking (38). A phospholipase C␥-PKC axis could be a critical link between the TCR and PKD1. The involvement of a PKC in the nucleocytoplasmic shuttling of HDAC7 could also explain why general PKC inhibitors, such as Gö6983 and GF109203X, suppressed the cytoplasmic translocation of HDAC7 in response to TCR cross-linking (Fig. 2).
Our observations that PKD1 regulates Nur77 induction by phosphorylating HDAC7 indicate that PKD1 could be important in the negative selection of T cells. Indeed, PKD1 was activated in thymocytes from DO11.10 transgenic mice after the induction of apoptosis with a peptide specific for their TCR. This activation correlated with the induction of Nur77 and further supports its potential role in the negative selection of T cells. Future experiments examining the biological consequences of knocking down either the Pkd1 or Hdac7 gene will test these predictions and further define the role of PKD1 and HDAC7 in thymocyte development.