|Home | About | Journals | Submit | Contact Us | Français|
Integration of protein kinases into transcription activation complexes influences the magnitude of gene expression. The nuclear factor of activated T cells (NFAT) group of proteins are critical transcription factors that direct gene expression in immune and nonimmune cells. A balance of phosphotransferase activity is necessary for optimal NFAT activation. Activation of NFAT requires dephosphorylation by the calcium-mediated calcineurin phosphatase to promote NFAT nuclear accumulation, and the Ras-activated extracellular signal-regulated kinase (ERK) mitogen-activated protein (MAP) kinase, which targets NFAT partners, to potentiate transcription. Whether protein kinases operate on NFAT and contribute positively to transcription activation is not clear. Here, we coupled DNA affinity isolation with in-gel kinase assays to avidly pull down the activated NFAT and identify its associated protein kinases. We demonstrate that p90 ribosomal S6 kinase (RSK) is recruited to the NFAT-DNA transcription complex upon activation. The formation of RSK-NFATc4-DNA transcription complex is also apparent upon adipogenesis. Bound RSK phosphorylates Ser676 and potentiates NFATc4 DNA binding by escalating NFAT-DNA association. Ser676 is also targeted by the ERK MAP kinase, which interacts with NFAT at a distinct region than RSK. Thus, integration of the ERK/RSK signaling pathway provides a mechanism to modulate NFATc4 transcription activity.
Transcription machinery mediates the molecular basis of eukaryotic gene expression. Multiple transcription factors are assembled on critical cis-acting DNA enhancers to form an enhancer complex (enhanceosome) in order to control the spatial, temporal, and magnitude of gene expression (reviewed in references 2 and 7). Cooperative interactions within the enhanceosome are further promoted by transcription coregulators (activators or repressors) and architectural proteins (e.g., high-mobility-group DNA-binding proteins) to facilitate the formation of transcription activation complex. In addition, various extracellular stimuli elicit distinct enhanceosomes in specific cells, suggesting that signaling molecules are also involved in the making of the transcription activation complex. For example, recruitment of the mitogen-activated protein (MAP) kinase to the initiation mediator complex (31, 40) and transcriptional regulation by histone deacetylase through protein phosphatase 1 and CK2 protein kinase (6, 43) demonstrate integration of phosphotransferases in transcription activation. Thus, understanding the organization, regulation, and function of cis-acting DNA enhancers is important to elucidate the molecular basis of complex eukaryotic gene transcription.
The nuclear factor of activated T cells (NFAT) group of proteins were first characterized as transcription factors that bind to the antigen receptor response element (ARRE) of the interleukin-2 (IL-2) gene promoter (reviewed in references 14 and 22). Characterization of the ARRE indicates that a cytoplasmic (NFATc) and a nuclear (NFATn) component cooperate to regulate gene transcription. The NFATc component was subsequently identified as a family of transcription factors (NFATc1 to NFATc4) involve in multiple biological processes. The Fos and Jun group of transcription factors (AP1 proteins) were identified as the NFATn component in the IL-2 gene. Molecular and structural analyses further demonstrate intimate associations to promote cooperative interaction of NFAT and AP-1 (8, 29). Other partners have also been identified as critical NFATn components, including CCAAT/enhancer binding proteins (C/EBP) and transcription factor GATA (32, 47). Thus, formation of a composite enhancer complex is required for NFAT-mediated gene transcription.
Current models suggest that two signaling pathways are converged onto NFAT (14, 22). NFAT is located in the cytosol of resting cells. A calcium-mediated signaling pathway is involved to activate the calcineurin phosphatase, which binds to the conserved NH2-terminal NFAT homology domain and dephosphorylates NFAT. Presumably, dephosphorylation induces conformational changes, which then exposes nuclear localization sequences and promotes NFAT nuclear accumulation. Inducible (e.g., stress-activated MAP kinases) and constitutive-active (e.g., CK1 and GSK3β) protein kinases have been indicated to phosphorylate NFAT, in concert or in sequential manners, to oppose nuclear accumulation (4, 13, 34, 48, 49). In addition, differential regulation of subcellular distribution of specific NFAT members by distinct protein kinases has been proposed. Thus, phosphorylation at the conserved NH2-terminal NFAT homology domain impedes NFAT nuclear accumulation and contributes negatively on NFAT activation.
In addition to the calcium-mediated NFAT dephosphorylation, a second signaling pathway—the Ras-mediated ERK MAP kinase—is involved in phosphorylation and activation of the NFAT partners. Activation mechanisms include promoting nuclear accumulation, DNA binding, and transcription activation of the NFAT partners. Therefore, in contrast to the negative role of phosphorylation that opposes NFAT nuclear accumulation, phosphorylation promotes NFAT activation, in part, by targeting the NFAT partners. Hence, a balance of phosphotransferases activity is required for optimal NFAT function.
The purpose of the present study was to examine the molecular basis of NFAT transcription activation. We ask whether protein kinases are associated with the activation complex, and positively contribute to NFAT transcription. We report that ribosomal p90 S6 protein kinase (RSK) is recruited to the activated NFAT-DNA complex. RSK interacts with the COOH-terminal REL homology DNA-binding domain and phosphorylates Ser676 of NFATc4. Phosphorylation of Ser676 promotes NFATc4 DNA binding by escalating NFAT-DNA association. Ser676 of NFATc4 is also targeted by the ERK MAP kinase. Thus, integration of the ERK/RSK signaling pathway provides an additional means to modulate NFATc4 transcription activity.
The NFAT luciferase reporter plasmids and the expression vectors for MAP kinases, RSK, and NFATc4 have been described (20, 21, 37, 39, 48). NFATc4 mutants were generated by PCR. Recombinant ERK and NFATc4 proteins were purified by glutathione affinity chromatography. Antibodies to phospho-Ser676 NFATc4 were generated by using synthetic phospho-peptide [Lys-Arg-Ser(P)-Pro-Thr-Gln-Ser-Phe-Arg-Phe-Leu-Pro-Val-Ile-Cys] encoding human NFATc4 and conjugated to ovalbumin. The phospho-specific antibody was affinity purified from rabbit serum as described previously (12).
3T3/L1, COS, HEK293, and mouse embryonic fibroblasts were cultured in Dulbecco modified Eagle medium. Jurkat and BHK cells were cultured in RPMI 1640 and minimal essential medium, respectively. All media were supplemented with 10% fetal calf serum, 2 mM l-glutamine, penicillin (100 U/ml), and streptomycin (100 μg/ml) (Invitrogen). Cells were transfected by using Lipofectamine (Invitrogen). 3T3/L1 cells were differentiated into lipid-laden adipocytes as described previously (48).
Streptavidin-agarose-precleared cell extracts were incubated with double-stranded, biotinylated PPARγ2 proximal NFAT DNA binding element (10 pmol, biotin-ATTACAGGGAAAATATTGCCACACTGTCTC) at 4°C overnight in the presence of 1 μg of poly(dI-dC). Competition was performed by using 10-fold excess of nonbiotinylated NFAT binding element. NFAT-DNA complex was precipitated with 20 μl of streptavidin-agarose at 4°C for an additional 2 h. After three washes with Triton-lysis buffer, the bound proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and subjected to in-gel kinase assays as described previously (30) by using recombinant NFATc4 (0.5 mg/ml) and myelin basic protein (0.5 mg/ml) as substrates. In brief, recombinant proteins were copolymerized with acrylamide gel matrix, and the prepared gel were used to resolve protein kinases associated with the NFAT-DNA precipitates. Resolved gels were subjected to SDS removal (in 20% iso-propanol and 50 mM Tris for 1 h at room temperature), and the gel-bound protein kinases were denaturated (in 6 M guanidine-HCl in 50 mM Tris and 5 mM 2-mercaptoethanol for 1 h at room temperature) and renaturated (in 0.04% Tween 40 in 50 mM Tris and 5 mM 2-mercaptoethanol at 4°C overnight) before incubation in kinase reaction buffer (50 μM [γ-32P]ATP in 40 mM HEPES [pH 8.0], 2 mM dithiothreitol, 0.1 mM EGTA, and 5 mM MgCl2) for 1 h at room temperature. After multiple washes (in 5% trichloroacetic acid and 1% sodium pyrophosphate at room temperature), phosphorylation was examined by autoradiography and phosphorimaging. Biotinylated, mutated PPARγ2 proximal NFAT DNA-binding element (biotin-ATTACAGGCATTATATTGCCACACTGTCTC) was used as a control. Binding with biotinylated, wild-type (5′-TTCTTAAATGGAAAACTTAAATCTCTTGCT-3′) and mutated (5′-TTCTTAAATGCATTACTTAAATCTCTTGCT-3′) PPARγ2 distal NFAT DNA was also examined.
Cell extracts prepared by using Triton-lysis buffer (20 mM Tris [pH 7.4], 137 mM NaCl, 2 mM EDTA, 1% Triton X-100, 25 mM β-glycerophosphate, 1 mM sodium vanadate, 2 mM sodium pyrophosphate, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 10 μg of leupeptin/ml) were incubated (5 h at 4°C) with antibodies prebound to 20 μl of protein G-Sepharose. For binding assays with recombinant proteins (5 μg), 20 μl of glutathione-Sepharose were used. After three washes with Triton-lysis buffer, the bound proteins were separated by SDS-PAGE and electrotransferred to a polyvinylidene difluoride membrane (Millipore). Antibodies to NFATc4 (Santa Cruz catalog no. sc-13036), phospho-ERK (sc-7383), ERK (sc-94), phospho-RSK (sc-17033), and RSK (BD Transduction Laboratories catalog no. R23820) were used for immunoblot analysis. Enhanced chemiluminescence was performed to visualize NFAT, RSK, and ERK.
Hemagglutinin epitope (HA)-tagged ERK, JNK1, and p38α were coexpressed with or without constitutive-active MEK1 (MEK1 ΔN3+S218E+S222D), MLK3, and constitutive-active MKK6 (MKK6 S207E+T211E), respectively, in COS cells. HA-tagged RSK were expressed in COS cells and stimulated or not with 100 nM phorbol myristate acetate (PMA). Cell extracts were prepared with Triton-lysis buffer 48 h after transfection, and immunecomplex kinase assays were performed as described previously (48).
The NFAT luciferase reporter plasmid (0.3 μg) was cotransfected with the control plasmid pRSV β-galactosidase (0.1 μg) and NFAT (0.3 μg) or MEK1 (0.1 μg) expression plasmids as indicated. Luciferase and β-galactosidase activity were measured 48 h after transfection. Cells were stimulated with ionomycin (2 μM), serum (20%), and/or PMA (100 nM) as indicated. The data were presented as the relative luciferase activity, calculated as the ratio of the luciferase activity to the activity of β-galactosidase (mean ± the standard error [n = 4]).
Nuclear extracts were prepared from cultured cells as described previously (48). Double-stranded oligonucleotides for gel mobility shift assays were labeled with [α-32P]dCTP. The sequences for the PPARγ2 proximal and distal NFAT elements are described above. The sequence for the IL-2 ARRE NFAT was 5′-AGAAAGGAGGAAAAACTGTTTCATACAGAAGG-3′. The binding reactions were carried out at room temperature in gel shift buffer [1 mM CaCl2, 1 mM MgCl2, 10 mM HEPES (pH 7.9), 50 mM NaCl, 15 mM β-mercaptoethanol, 10% glycerol, 0.1 mg of bovine serum albumin/ml, and 1 mg of poly(dI-dC)/ml] for 30 min. The protein-DNA complexes were separated in 5% nondenaturing polyacrylamide gels in Tris-glycine-EDTA buffer (25 mM Tris, 200 mM glycine, 1 mM EDTA) and were visualized by autoradiography. For supershift analysis, antibody was preincubated with nuclear extract for 30 min at room temperature before addition of the labeled probe. For association analysis, NFAT-DNA complex was assembled for the indicated time and loaded into the running nondenaturing polyacrylamide gel. For dissociation analysis, an excess amount of unlabeled oligonucleotides (50 pmol) was added to the preassembled NFAT-DNA complex for the indicated times. The amount of NFAT-DNA complex was quantitated by phosphorimager analysis and plotted against the duration of association or dissociation. For saturation analysis, increasing amount of labeled probe (50, 100, 150, and 200 fmol) was incubated with a constant amount of nuclear extract. The amount of NFAT-DNA complexes and unbound probe were quantitated by phosphorimager analysis. The bound/free ratio was plotted against the amount of bound probe to assess relative binding affinity.
Nuclear factors that were associated with chromatin in differentiated and undifferentiated 3T3/L1 cells were cross-linked to DNA by using formaldehyde (1%). Cells were harvested and cross-linked chromatin was sheared by sonication. Sonicated cell lysate was immunoprecipitated by using NFAT or RSK antibodies. DNA present in the immunoprecipitated chromatin was isolated, after reversed cross-link and proteinase K digestion, and PCR was performed (5′-GAATTGGCTGGCACTGTCCT-3′; 5′-ATAGACTTGTTGAATAAATC-3′) to examine the presence of PPARγ2 gene promoter. The presence of GAPDH (for glyceraldehyde-3-phosphate dehydrogenase) promoter (5′-GGCTCTCTGCTCCTCCCTGTTCC-3′ and 5′-TCAATGAAGGGGTCGTTGATGGC-3′) was also examined.
To test whether protein kinases are involved in NFAT transcription activation, we performed DNA affinity isolation to avidly pull down activated NFAT and its associated proteins. We reasoned that associated proteins such as protein kinase exhibit enzymatic activity and, hence, we performed in-gel kinase assays to detect the presence of phosphotransferases in the NFAT-DNA precipitates. Coupled DNA-binding-in-gel kinase assays demonstrated that two protein kinases (70 and 90 kDa) are present in the NFAT-DNA precipitates (Fig. (Fig.1A).1A). Association of these protein kinases in the NFAT-DNA precipitates is dependent on stimulation with calcium ionophore ionomycin and phorbol ester (PMA). Competition with excess nonbiotinylated NFAT DNA binding element reduced the p70 and p90 protein kinases (70 and 90 kDa, respectively) in the NFAT-DNA precipitates, demonstrating the specificity of the interaction. Thus, these data demonstrate that protein kinases are recruited to the NFAT transcription complex upon activation.
We suspected that the p90 kinase might be RSK because of the similar mass and the reported function of RSK as a downstream effector kinase in the ERK signaling pathway (17). To test this hypothesis, we performed immunodepletion prior to DNA binding with an antibody to RSK. Coupled DNA-binding-in-gel kinase assays demonstrated that depletion of RSK reduced the p90 but not the p70 protein kinase in the NFAT-DNA precipitates (Fig. (Fig.1B).1B). These data indicate that RSK might be involved in the NFAT transcription activation complex.
To confirm the presence of RSK in the NFAT activation complex, we performed immunoblot analysis subsequent to the DNA-binding assays. Immunoblot analysis demonstrated the presence of endogenous RSK in the NFAT-DNA precipitates (Fig. (Fig.1C).1C). Importantly, phosphorylated RSK is preferentially recruited upon NFAT activation, and the interaction with NFAT-DNA is reduced by MEK inhibitor U0126. Moreover, recruitment of activated RSK is independent of NFAT partners, as demonstrated by the presence of RSK in both PPARγ2 proximal and distal NFAT DNA-binding elements, which form different NFAT complexes (48). Recruitment of RSK upon NFAT activation is further corroborated by coupled DNA-binding-immunoblot analysis with cell extracts expressing RSK2 (Fig. (Fig.1D).1D). Together, these data demonstrate that activated RSK is recruited to the NFAT-DNA complex.
Previous studies indicated that bound NFAT determines the formation of a ternary activation complex with NFAT partners and DNA (24, 25, 47). To further examine the recruitment of RSK to the NFAT activation complex, we performed mutagenesis to abolish the NFAT binding element. Mutational removal of the NFAT binding element reduced the presence of RSK in the NFAT-DNA precipitates (Fig. (Fig.2A).2A). These data indicate that NFAT is required for the recruitment of RSK.
We further examined whether lowered levels of total NFAT proteins affected RSK association. We performed coupled DNA-binding-immunoblot assays with extracts prepared from wild-type and Nfatc2−/− Nfatc4−/− cells (Fig. (Fig.2B).2B). Immunoblot analysis indicated that the loss of NFATc2 and NFATc4 expression in the Nfatc2−/− Nfatc4−/− cells, whereas the expression levels of NFATc1 and NFATc3 proteins were similar. The expression levels of RSK was also similar in wild-type and the Nfatc2−/− Nfatc4−/− cells. However, the amount of RSK in the NFAT-DNA precipitates was reduced in Nfatc2−/− Nfatc4−/− cells. Together, these data demonstrate that RSK targets activated NFAT in the DNA precipitates.
Next, we examined the interaction of RSK to NFAT. Coimmunoprecipitation assays demonstrated the presence of endogenous RSK in the NFATc4 precipitates (Fig. (Fig.3A).3A). The interaction between RSK and NFATc4 is further confirmed by coexpression and subsequent immunoprecipitation and immunoblot analysis with either RSK or NFATc4 antibody (Fig. (Fig.3B).3B). These data demonstrate that RSK binds to NFATc4.
To delineate the RSK binding site on NFAT, we examined the interaction between various NFATc4 deletion mutants and RSK. Immunoblot analysis demonstrated similar expression of various NFATc4 proteins (Fig. (Fig.3C).3C). Deletion of the most COOH-terminal end of NFATc4 (to residue 853) had a minimal effect on RSK binding. However, deletion to remove the COOH-terminal end of the NFATc4 REL homology domain (to residue 581) reduced RSK binding. Further deletion to remove the NH2-terminal end of the NFATc4 REL homology domain (to residue 450) abolished RSK binding. These data demonstrate that RSK binds to the NH2-terminal end of the NFATc4 REL homology domain.
Previous studies established that RSK is a prominent downstream effector of the ERK MAP kinase signaling pathway (17). In addition, both RSK and ERK have been demonstrated to phosphorylate the same substrate upon activation (5, 19, 44, 45). ERK binds to and phosphorylates multiple targets by initially docking to the conserved sequence motifs on its targets (reviewed in references 38 and 42). Two docking motifs—FxF and LxL sequences—have been indicated to facilitate ERK interaction. The FxF motif seems to be specific for the ERK substrates, whereas the LxL sequence provides a common recognition site for other MAP kinases, including the c-Jun NH2-terminal protein kinases (JNK) (23). Sequence analysis indicates that there is a potential FxF motif (Phe681,683) at the COOH-terminal end of the NFATc4 REL domain (Fig. (Fig.4A).4A). Similar sequence is also found in other NFAT members (FxY motifs in NFATc1, NFATc2, and NFATc3). Based on the recent structural analysis (8, 18, 26), the FxF motif on the NFAT REL domain is exposed and possibly allows interaction with ERK.
To investigate whether ERK binds to NFAT, we examined binding of endogenous ERK to NFATc4 by coimmunoprecipitation assays (Fig. (Fig.4A).4A). Immunoblot analysis indicated the presence of ERK in the NFATc4 precipitates. Importantly, NFATc4 preferentially interacts with activated ERK. These data demonstrate that NFATc4 interacts with ERK upon activation.
To delineate the ERK binding site on NFAT, we examined the interaction between various NFATc4 deletion mutants and ERK (Fig. (Fig.4B).4B). Deletion of the most COOH-terminal end of NFATc4 (to residue 853) reduced ERK binding. Further deletion to remove the COOH-terminal end of the NFATc4 REL homology domain (to residue 581) abolished ERK binding. These data indicate that ERK binds to the COOH terminus of the NFATc4 REL homology domain, which encompasses the conserved FxF motif.
Next, we tested whether the FxF motif in NFATc4 is important for interaction by using recombinant ERK (Fig. (Fig.4C).4C). Binding assays indicated that the NFATc4 REL domain (NFATc4 344-749) was found in the ERK precipitates. Deletion to residue 581 of NFATc4 (NFATc4 344-581) to eliminate the FxF motif or mutation removal of Phe681,683 of the FxF motif with Gly (NFATc4 344-749 FF681,683GG) abolished ERK binding. Similarly, with the wild-type and mutated [FF681,683GG] recombinant NFATc4 proteins in binding assays, replacement of Phe681,683 with Gly reduced interaction with ERK (Fig. (Fig.4D).4D). These data demonstrate that Phe681,683 of NFATc4 is important for ERK interaction. Together with the mapping of RSK interaction, these data demonstrate that RSK and ERK bind to different regions of the NFATc4 REL homology domain.
Binding of RSK and ERK may promote NFATc4 phosphorylation. Hence, we mapped the phosphorylation sites on NFATc4 targeted by RSK and/or ERK. Previous studies indicated that RSK phosphorylation sites (Ser or Thr) are frequently located adjacent to Arg or Lys residues (15, 16). On the other hand, Ser/Thr-Pro motifs are ERK phosphorylation sites, which are often located within 20 amino acids from the ERK binding motif (38). Sequence analysis indicates that there is a Ser-Pro motif (Ser676) locates near the NFATc4 FxF motif (Fig. (Fig.4A).4A). Thus, Ser676 is a potential ERK target. Interestingly, four consecutive Arg/Lys residues (Arg672Arg673Lys674Arg675) precede Ser676, suggesting that Ser676 may be an RSK phosphorylation site as well. In addition, a similar Ser residue is found in other NFAT members (Fig. (Fig.4A4A).
Next, we performed in vitro kinase assays to test whether RSK phosphorylates Ser676 of NFATc4 (Fig. (Fig.5A).5A). Immune complex kinase assays indicated that RSK phosphorylates NFATc4 upon activation. Replacement of Ser676 with Ala eliminated RSK phosphorylation. These data demonstrate that RSK phosphorylates Ser676.
We also tested whether ERK phosphorylates Ser676 (Fig. (Fig.5B).5B). Immune complex kinase assays indicated that ERK phosphorylates NFATc4 and, to a lesser extent, Ala676 NFATc4. These data demonstrate that ERK phosphorylates multiple sites, including residue Ser676, of the NFATc4 REL domain.
Similar to the ERK MAP kinase, JNK and p38 MAP kinases are also Pro-directed Ser/Thr kinase. To ascertain whether phosphorylation of NFATc4 REL homology domain is specific to the ERK/RSK signaling pathway, we examined phosphorylation of Ser676 by JNK and p38 MAP kinases (Fig. (Fig.5C).5C). Immune complex kinase assays indicated that activated ERK phosphorylated the NFATc4 REL domain. However, neither JNK nor p38 MAP kinases phosphorylated the NFATc4 REL domain. These data demonstrate that the ERK/RSK signaling pathway mediates phosphorylation at the NFATc4 REL homology domain.
We further examined phosphorylation of Ser676 of NFATc4 by generating phospho-Ser676 antibodies. Immunoblotting analysis indicated that in vitro ERK phosphorylated NFATc4 REL domain was detected by the phospho-Ser676 antibodies (Fig. (Fig.6A).6A). Preincubation of the antibodies with the phospho-Ser676 peptide abolished the recognition (data not shown). In addition, the phospho-Ser676 antibodies detected the wild-type NFATc4 proteins but not the Ser676 mutated NFATc4 proteins (Fig. (Fig.6B).6B). Importantly, detection of the wild-type NFATc4 by the phospho-Ser676 antibodies is dependent upon stimulation with PMA. Pretreatment with MEK inhibitor U0126 reduced NFATc4 Ser676 phosphorylation. These data confirm phosphorylation of Ser676 of NFATc4.
We also tested whether RSK phosphorylates Ser676. Constitutive-active RSK2 (36) was coexpressed with NFATc4, thus circumventing the requirement for activated ERK (Fig. (Fig.6C).6C). Immunoblot analysis with the phospho-Ser676 antibodies indicated that, similar to coexpression with MEK1, constitutive-active RSK2 phosphorylated NFATc4. These data demonstrate that both RSK and ERK phosphorylate NFATc4.
To ascertain phosphorylation of Ser676 in endogenous NFATc4, we performed immunoprecipitation with NFATc4 antibodies and subsequent immunoblotting analysis to identify phospho-Ser676 NFATc4. PMA stimulation increased NFATc4 phosphorylation at Ser676 (Fig. (Fig.6D).6D). Administration of MEK1 inhibitor U0126, however, reduced Ser676 phosphorylation. Furthermore, we performed coupled DNA-binding-immunoblot assays with phospho-Ser676 NFATc4 antibodies. Similar activation-dependent of NFATc4 phosphorylation of Ser676 was revealed (Fig. (Fig.6D).6D). Together, these data establish that Ser676 of NFATc4 is targeted by the MEK/ERK/RSK signaling pathway.
Previous studies indicated that mutational removal of Ser168,170 at the NH2-terminal of NFATc4 with Ala causes constitutive nuclear localization, which leads to increase in NFAT-mediated gene transcription (48). Thus, phosphorylation at the NH2-terminal of NFATc4 inhibits NFAT activation. We suspected that phosphorylation at Ser676 of the COOH-terminal NFAT REL homology domain, however, might positively regulate NFATc4 because of the activation-dependent recruitment and nuclear localization of RSK and ERK. To test this hypothesis, we examined whether MEK inhibitor U0126 blocked NFAT-mediated transcription (Fig. (Fig.7A).7A). Luciferase reporter gene assays indicated that pretreatment of U0126 reduced NFAT-mediated gene transcription. These data support the view that the ERK/RSK signaling pathway positively regulates NFAT.
We further examined whether MEK1 affects NFAT-mediated gene transcription by Ser676 phosphorylation (Fig. (Fig.7B).7B). Luciferase reporter gene assays indicated that expression of wild-type NFATc4 increased transcription activity. Expression of constitutive nuclear NFATc4 SS168,170AA led to an additional increase in transcription activity. In the absence of MEK1, NFATc4 SS168,170AA and NFATc4 SS168,170AA+S676A mediated transcription to a similar extent. Coexpression of MEK1 with either wild-type or SS168,170AA NFATc4 further potentiated NFATc4-mediated transcription. Mutational removal of Ser676, however, reduced MEK1-mediated transcription potentiation. Replacement of Ser676 with Glu could not mimic phospho-Ser676 and increase NFATc4-mediated transcription (data not shown). These data indicate that activation mediated by the ERK signaling pathway requires Ser676 phosphorylation, which further augments the elevated transcription upon dephosphorylation at Ser168,170 of NFATc4.
We further examined the effect of removal of ERK binding on MEK1-induced NFAT gene transcription (Fig. (Fig.7C).7C). Similar to the Ser676 phosphorylation-defective NFATc4, mutational removal of the ERK binding site (NFATc4 344-749 FF681,683GG) reduced MEK1-mediated transcription potentiation. These data demonstrate that binding and phosphorylation is required for the effect of ERK on NFAT activation.
Previous studies established that replacement of Arg474 and Asn475 with Ala and Thr541 with Gly of NFATc4 abolished cooperative interaction between NFAT and NFAT partners (8, 29, 47). To ascertain that MEK1-mediated NFAT activation channels to both NFAT and NFAT partners, we examined the effect of MEK1 on NFAT mutant (NFATc4 344-749 RN474,475AA, T541G) that is defective in cooperative interaction with AP-1 or C/EBP (Fig. (Fig.7D).7D). Similar to the Ser676 phosphorylation-defective NFATc4 and ERK binding-defective (NFATc4 344-749 FF681,683GG) NFATc4, mutational removal to abolish NFAT cooperative interaction (NFATc4 344-749 RN474,475AA, T541G) reduced MEK1-mediated transcriptionpotentiation. Compound mutation to abolish Ser676 phosphorylation and NFAT cooperative interaction (NFATc4 344-749 RN474,475AA, T541G + S676A) further reduced MEK1-mediated transcription. These data demonstrate that the MEK/ERK/RSK signaling pathway regulates both NFAT and NFAT partners to achieve optimal NFAT activity.
We also examined the effect of RSK on transcription mediated by the full-length, constitutive nuclear NFATc4 (NFATc4 SS168,170AA) (Fig. (Fig.7E).7E). Expression of constitutive-active RSK2, but not wild-type or dead RSK2, further increased NFATc4 SS168,170AA-mediated gene transcription. However, mutational removal of Ser676 reduced transcription potentiation mediated by constitutive-active RSK2. These data demonstrate that RSK also positively regulates NFATc4 activation.
NFATc4, ERK, and RSK are located in the nucleus upon activation (11, 14, 22, 28, 35). Since the interaction with NFATc4 is activation dependent, ERK and RSK may regulate NFAT nuclear function to potentiate gene transcription. Two nuclear events of NFAT—DNA binding and transcription activation—could be regulated by ERK and RSK. First, we tested whether activation of the ERK and RSK signaling pathway regulates NFAT transactivation. We performed Gal4-luciferase reporter gene assays with the NH2- and the COOH-terminal NFATc4 transactivation domains (46). Coexpression of MEK1 had minimal effect on the transcription mediated by the NH2- or the COOH-terminal NFATc4 transactivation domains (data not shown).
Next, we tested whether the DNA binding of NFAT is affected by ERK activation by using gel mobility shift assays. Stimulation with serum plus ionomycin increased NFAT DNA binding (Fig. (Fig.8A).8A). Administration of U0126, however, reduced NFAT DNA binding. Thus, these data demonstrate that the ERK/RSK signaling pathway regulates NFAT DNA binding.
We further examined whether phosphorylation at Ser676 affects DNA binding (Fig. (Fig.8B).8B). We prepared nuclear extracts from cells transfected with wild-type and mutated (S676A) NFATc4 REL domain for gel mobility shift assays. Immunoblot analysis indicated similar expression of NFATc4 proteins. Expression of the NFATc4 REL proteins promotes specific protein-DNA complexes as indicated by supershift analysis with antibody detects NFAT. Coexpression of MEK1 increased formation of NFATc4-DNA complexes. Saturation analysis indicated that there is approximately threefold increase in NFATc4 DNA binding upon ERK/RSK activation by MEK1 expression (Fig. (Fig.8C).8C). On the other hand, mutational removal of Ser676 with Ala, to prevent ERK/RSK phosphorylation, reduced NFATc4 DNA binding approximately sixfold. Importantly, phosphorylation at Ser676 of NFATc4 increased DNA association in three different NFAT DNA binding elements (Fig. 8B and C). Together, these data demonstrate that phosphorylation of Ser676 is critical for ERK/RSK-potentiated NFATc4 DNA binding.
Next, we examined the kinetic parameters of formation of NFATc4-DNA complexes upon Ser676 phosphorylation (Fig. (Fig.8D).8D). Phosphorylation at Ser676 increased the on-rate of NFATc4 DNA binding. In approximately 4 min, 50% of phosphorylated NFATc4 was DNA bound. On the contrary, approximately 7 min and >20 min were required for 50% DNA binding for wild-type and S676A NFATc4, respectively.
We also examined the dissociation of the NFATc4-DNA complexes (Fig. (Fig.8E).8E). Both wild-type and S676A NFATc4 were dissociated from the DNA at a similar rate. ERK/RSK activation had minimal effect on NFATc4-DNA dissociation. Together, these data demonstrate that Ser676 phosphorylation increases DNA binding by promoting association of NFATc4 with DNA.
Previous studies demonstrated that NFAT cooperates with C/EBP and regulates PPARγ2 gene transcription in adipocyte differentiation (47, 48). In addition, RSK phosphorylates NFAT (Fig. (Fig.5)5) and C/EBP (5, 19). Hence, we sought to determine whether RSK is associated with the NFAT transcription complex upon adipocyte differentiation. Chromatin immunoprecipitations indicated that NFATc4 is associated with the PPARγ2 gene promoter in differentiated, but not the undifferentiated, 3T3/L1 adipocytes (Fig. (Fig.9)9) (47). RSK is also recruited to the PPARγ2 promoter complex upon differentiation (Fig. (Fig.9).9). Importantly, phospho-RSK is present in the PPARγ2 promoter complex. Thus, these data demonstrate that active RSK is recruited to the NFAT transcription complex to mediate PPARγ2 gene expression upon adipocyte differentiation.
NFAT is located in the cytosol of resting cells. Dephosphorylation at the NH2-terminal NFAT homology domain promotes NFAT nuclear accumulation. Indeed, replacement of the conserved NH2-terminal phosphorylation sites with Ala, to mimic dephosphorylation, increases nuclear localization and NFAT-mediated transcription (3, 13, 33, 48). Plausibly, the mechanism of activation is by mass accumulation of nuclear NFAT. Thus, phosphorylation at the NH2-terminal end impedes NFAT activation. We demonstrate here that phosphorylation at the COOH-terminal REL homology domain by the ERK/RSK signaling pathway increases NFATc4 DNA binding and potentiates transcription activation. Hence, dephosphorylation at the NH2-terminal NFAT homology domain only partially activates NFAT. Further transcription potentiation by increasing DNA binding leads to optimal NFAT activation. Therefore, a balance of dephosphorylation and phosphorylation is required for NFAT function.
Requirement of a second signal to mediate NFAT phosphorylation by ERK/RSK suggests a stepwise activation mechanism to achieve optimal transcription. In addition to regulating DNA binding of nuclear NFAT, phosphorylation of Ser676 by ERK/RSK may indirectly act as a nuclear retention signal and prolong nuclear accumulation by NFAT association with DNA. Hence, nuclear NFAT is secluded, by DNA or DNA matrix, and escaped from rephosphorylation that leads to shuttling to the cytosol.
ERK/RSK phosphorylation may also modulate the duration of NFAT-DNA association and elicit various levels of activation, by recruiting diverse NFAT partners and/or coactivators, to attain a different threshold of NFAT-mediated transcription. For example, increased duration of NFAT-DNA association may favor recruitment of transcription coactivators and modulate the induction of NFAT targets, especially since NFATc4 interacts with CBP coactivator at two sites (46). Thus, increased duration of NFAT-DNA association may sustain the expression of specific NFAT targets.
Alternatively, ERK/RSK phosphorylation may modulate interaction of NFAT on the DNA (e.g., regulate formation of NFAT dimers versus NFAT partners) (8, 18, 26). Structural analysis indicates that the surrounding residues of Ser676 of NFAT are important for intramolecular interaction and/or binding to the minor groove of DNA. Phosphorylation at Ser676 may achieve specific conformations by providing additional salt bridges between NFAT and its interacting molecules, including DNA, NFAT-partners, and NFAT itself. In addition, phosphorylation at Ser676 may induce an S-switch type of conformational change, which has been demonstrated in the similar immunoglobulin fold Runt DNA-binding domain (1, 41), and promote NFAT DNA binding. Hence, a specific transcription profile mediated by distinct NFAT ternary complex may be accomplished upon ERK/RSK activation.
We show that both ERK and RSK interact with the NFAT REL domain. Although ERK and RSK target different regions of the REL domain, their interactions with NFAT are activation dependent. Current models suggest that elevated levels of intracellular calcium and subsequent dephosphorylation mediated by the calcineurin phosphatase are required to promote conformational changes and exposure of nuclear localization sequences for NFAT nuclear entry. On the contrary, mitogenic agonists promote dimerization and nuclear localization of the ERK kinase (28, 35), whereas RSK is disinhibited upon phosphorylation to release the autoinhibitory domain from the catalytic domain (17). Activated ERK/RSK signaling pathway targets NFAT REL domain and NFAT partners. Hence, the dual requirements of calcium and ERK activation may ensure that both calcineurin and ERK/RSK are in concerted to mediate expression of critical NFAT targets, such as IL-2 and tumor necrosis factor alpha cytokines in inflammation and PPARγ in terminated adipocyte differentiation.
Coupled DNA-binding-in-gel kinase assays indicate that RSK, but not ERK, is associated with the DNA-bound NFAT. Mapping of the ERK binding sites to the COOH-terminal end of the REL domain, which mediates NFAT-DNA binding, supports the release of ERK before DNA interaction. Since the ERK-NFAT interaction is activation dependent and ERK is translocated to the nucleus upon activation, ERK is likely to bind to NFAT after calcineurin-mediated dephosphorylation and ERK activation but before NFAT binds to DNA. If so, interaction with nuclear-bound activated ERK, which also process nuclear localization sequence, may further facilitate the nuclear entry of NFAT.
Similar to ERK, RSK is also translocated into nucleus upon activation. However, RSK remains associated with NFAT-DNA and thus may target additional proteins in the NFAT transcription activation complex. For example, analogous to the effect of JNK on phosphorylation of Jun protein complex (27), association of RSK in the NFAT transcription activation complex may trans-phosphorylate NFAT partners, such as C/EBP and Fos, upon activation (10, 19). trans-Phosphorylation of NFAT partners may contribute to the requirement of dual signals for optimal NFAT activity. trans-Phosphorylation of NFAT-associated transcription coactivators, such as CBP and chromatin-bound histones, may further modulate NFAT-mediated gene transcription.
We demonstrate here that two protein kinases (p70 and p90) are recruited to the NFAT transcription complex in an activation-dependent manner. We demonstrate that the p90 kinase is the ERK-activated RSK kinase. The identity of the p70 NFAT-associated kinase remains to be sought. Association of at least two distinct protein kinases upon NFAT activation suggests that, in addition to DNA binding, phosphorylation modulates additional functions of nuclear NFAT. Analogous to the role of protein kinase A on NF-κB activation (9), one function could be phosphorylation-dependent recruitment of CBP coactivator to the NFAT activation complex. Such phosphorylation may then provide additional means to modulate NFAT activity.
NFAT is phosphorylated under basal, unstimulated conditions. Termination of NFAT activity required nuclear export, which is mediated by multiple protein kinases, including GSK3β, CK1α, JNK, and p38 MAP kinases. These inhibiting protein kinases phosphorylate NFAT at the NH2-terminal end and oppose nuclear accumulation. We demonstrate here that ERK and RSK interact with activated NFAT and promote DNA binding. Hence, distinct protein kinases are associated with NFAT in the basal resting state and in the active form. Dynamic modulation of multiple protein kinases in the NFAT transcription complex suggests that NFAT may be the transducer in relaying activation signal from the plasma membrane and cytosol, as well as the regulator to initiate the expression of NFAT targets in the nucleus. If so, NFAT is acting as a recruitment platform to nucleate the assembly of a transcription complex. Thus, identification of other components of the composite enhancer complex will be essential to understand the molecular bases of NFAT-mediated transcription.
Assembly of the NFAT transcription activation complex is dependent on, in part, the interaction between NFAT, NFAT-partners, and DNA. Association and dissociation of NFAT-DNA contributes to the steady-state formation of the transcription activation complex. We demonstrate here that phosphorylation at Ser676 promotes NFATc4-DNA association, supporting the potential role of NFAT to nucleate assembly of a functional transcription activation complex. Phosphorylation mediated by the MEK/ERK signaling pathway also acts on NFAT partners and contributes to the final output. Our previous results demonstrated that dissociation of the NFATc4-DNA complex is regulated by the NFAT partners (47), suggesting that NFAT partners modulates termination of the transcription activation complex. Hence, in addition to understand NFAT proteins per se, elucidation of the NFAT partners is equally important to reveal the molecular mechanisms of NFAT-mediated transcription.
The coupled DNA-binding-in-gel kinase assays provide a resourceful avenue for identifying protein kinases present in the NFAT activation complex. This protocol could be extended to identify protein kinases in other transcription factor activation complexes by using distinct consensus DNA binding sequence. Furthermore, this protocol could be modified to allow identification of other NFAT-associated proteins that process enzymatic activity. For example, DNA-modifying enzymes, such as histone acetyltransferases, histone deacetylases, and histone methyltransferases, may also be present in the NFAT transcription complex. Biochemical assays to detect these DNA-modifying enzymes are established. A goal for further research will be to elucidate whether extracellular stimuli will recruit distinct protein kinases and/or DNA modifying enzymes to the NFAT transcription complex.
In conclusion, we have demonstrated that the ERK/RSK signaling pathway is recruited to the NFAT transcription complex upon activation. Phosphorylation by the ERK/RSK kinases promotes NFAT DNA binding. Identification and characterization of the organization, regulation, and function of components of the NFAT activation complex will provide the molecular bases to understand NFAT-mediated gene transcription.
We thank R. J. Davis, T. Hoey, J. A. Smith, T. R. Soderling, and T. W. Sturgill for providing reagents. We thank C. S. Rubin and Z.-Y. Zhang for critical reading of the manuscript. We also thank Paul Shore for his insightful discussion and the Diabetes Research and Training Center at AECOM for their support.
This research was supported, in part, by grants from the NIH/NIDDK and the American Heart Association.