c-Jun N-terminal Kinase (JNK) Positively Regulates NFATc2Transactivation through Phosphorylation within the N-terminalRegulatory Domain*□S

Received for publication, February 18, 2005Published, JBC Papers in Press, March 2, 2005, DOI 10.1074/jbc.M501898200

Inmaculada Ortega-Perez‡§, Eva Cano¶�, Felipe Were‡, Margarita Villar¶, Jesus Vazquez¶,and Juan Miguel Redondo‡¶**

From the ‡Centro Nacional de Investigaciones Cardiovasculares (CNIC), Ronda de Poniente 5, Tres Cantos, Madrid 28760and the ¶Centro de Biologıa Molecular Severo Ochoa, Consejo Superior de Investigaciones Cientıficas, UniversidadAutonoma de Madrid, Facultad de Ciencias, Madrid 28049, Spain

The nuclear factor of activated T cells (NFAT) familyof transcription factors regulates the transcription ofcytokine genes and other genes involved in the regula-tion and function of the immune system. NFAT activityis regulated by the phosphatase calcineurin, whichbinds and dephosphorylates the NFAT N-terminal regu-latory domain, a critical step required for nuclear trans-location and transcriptional activity. Here we show thatthe mitogen-activated protein kinase (MAPK) JNK acti-vates NFATc2-dependent transcription. Mass spectrom-etry revealed that JNK phosphorylates at least six resi-dues within the NFATc2 regulatory domain in vitro.Transfection of cells with a chimeric construct encodingthe GAL-4 DNA binding domain linked to wild-typeNFATc2 showed that JNK stimulates the NFATc2 trans-activation domain in activated Jurkat T lymphocytes,an effect that is inhibited by dominant-negative ver-sions of JNK. Likewise, the mutation of the phosphoryl-ation sites identified revealed that Thr116 and Ser170 arecritical for the transactivation of NFATc2 by JNK. Inaddition, clustered mutation of the SP-conserved motifsof NFATc2 showed that SP1 and SP2, but not SP3, arealso important for the inducible transactivation ofNFATc2. Furthermore, mass spectrometry analysis ofNFATc2-transfected cells indicated that the activationof the JNK pathway results in the in vivo phosphoryla-tion of Thr116. Our results indicate that, unlike otherNFAT members, the transcriptional activity of NFATc2is up-regulated by JNK. JNK-mediated phosphorylationof NFATs thus appears to play a differential physiolog-ical role among NFAT family members.

Reversible phosphorylation of proteins is a common processin the regulation of most cellular functions. The mitogen-acti-vated protein kinase (MAPK)1 family is a set of protein kinases

that are activated by extracellular stimuli, whose function andregulation has been conserved during evolution from yeast tomammals (1). The major role of these enzymes is to amplify andintegrate signals from a wide variety of extracellular stimuli,thereby allowing cells to respond and adapt to changes in theirenvironment. MAPK signal transduction pathways controlmany processes, including gene expression, protein synthesis,the cell cycle, differentiation, transformation, and programmedcell death. The MAPK family comprises the extracellular sig-nal-regulated kinases (ERK), the p38 MAPKs, and the stress-activated protein kinases (SAPK/JNK). All MAPKs are acti-vated through the dual phosphorylation of conserved TXYmotifs by upstream kinases termed MAPK kinases (MAPKK),which are themselves activated by Ser/Thr phosphorylation bythe third component of the MAPK system, the MAPKKkinases (MAPKKK).

Members of the SAPK/JNK family phosphorylate the N-terminal transactivation domain of c-Jun and other transcrip-tion partners of AP-1 (2, 3). There are at least three JNKMAPKs, called JNK-1, JNK-2, and JNK-3, and each is encodedby a different gene. Although the MAPKKs MKK4 and MKK7have been established as specific activators of JNK, more thana dozen additional MAPKKKs are also known to activate thispathway (reviewed in Ref. 4). This huge diversity of activatorsallows a wide range of stimuli to activate JNK, although JNKsare more commonly activated in response to environmentalstress or signaling through GTPases of the rho family, whichcan trigger the MAPKKK cascade pathway. Once activated,JNKs phosphorylate an array of different substrates rangingfrom transcription factors to apoptosis regulatory molecules,involving these kinases in regulation of cell viability, cellularstress, induction of apoptosis, and cell proliferation (reviewedin Ref. 4).

Unlike JNK-3, whose expression is restricted to brain, heart,and testis, isoforms 1 and 2 are ubiquitously expressed (re-viewed in Ref. 5). Much attention has been paid, however, totheir roles in T cell development and function (6). In these cells,complete activation of JNK requires signals from both the Tcell receptor (TCR)-CD3 complex and the costimulatory factorCD28. These signals, which can be replaced by combined treat-ment with the pharmacological agents phorbol 12-myristate

* This work was supported by grants from Ministerio de Ciencia yTecnologıa (SAF 2003-02920), from the Fundacio La Marato de TV3,and from the Ministerio de Sanidad y Consumo from Spain (RECAVAnetwork). The costs of publication of this article were defrayed in partby the payment of page charges. This article must therefore be herebymarked “advertisement” in accordance with 18 U.S.C. Section 1734solely to indicate this fact.

□S The on-line version of this article (available at http://www.jbc.org)contains supplemental data.

§ Recipient of a CNIC BANCAJA fellowship.� Recipient of a Ramon y Cajal contract from the Ministerio de Ciencia

y Tecnologıa of Spain.** To whom correspondence should be addressed. Tel.: 34-91-497-

8270; Fax: 34-91-497-8087; E-mail: jmredondo@cbm.uam.es.1 The abbreviations used are: MAPK, mitogen-activated protein ki-

nase; PMA, phorbol 12-myristate 13-acetate; wt, wild type; HA, hemag-glutinin; JNK, c-Jun N-terminal kinase; Io, calcium ionophore A23187;GST, glutathione S-transferase; RIPA, radioimmune precipitation as-say buffer; IL, interleukin; ERK, extracellular signal-regulated kinase;NFAT, nuclear factor of activated T cells; HEK, human embryonickidney cells; RP-HPLC, reverse phase high pressure liquid chromatog-raphy; NLS, nuclear localization signal.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 280, No. 21, Issue of May 27, pp. 20867–20878, 2005© 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org 20867

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from

13-acetate (PMA) and calcium ionophore A23187 (Io), lead to Tcell activation (7, 8). The transcription factors c-Jun and NFATare JNK substrates, implicated in T cell function as effectormolecules of JNK signaling in these cells.

NFAT was initially identified in T cells as a transcriptionfactor responsible for mediating Ca2�/calcineurin-dependenttranscription of genes including IL-2 and other important cy-tokines involved in T cell activation (9–11). Further studiesidentified new members of the family that participate in manyphysiological processes in different cell types inside and outsideof the immune system. The NFAT family of transcription fac-tors contains five members, which are related to the Rel/NF�Bfamily. Four of them NFATc1 (NFAT2/c), NFATc2 (NFAT1/p),NFATc3 (NFAT4/x), and NFATc4 (NFAT3) are regulated bycalcineurin and share a highly homologous region located attheir regulatory N-terminal domain (reviewed in Ref. 12). Ac-tivation of these members is primarily regulated through thecontrol of their subcellular localization. In the absence of Ca2�

signals, NFAT proteins are highly phosphorylated and remainin the cytoplasm of resting cells. When intracellular calciumlevels rise, the phosphatase calcineurin is activated and de-phosphorylates NFATs. This favors the exposure of a nuclearlocalization signal (NLS) and the masking of a nuclear exportsignal (NES), thus promoting the translocation of NFATs fromthe cytosol to the nucleus, their binding to target sequences,and therefore the transcription of different NFAT-dependentgenes (12, 13). Calcineurin-dependent nuclear accumulation ofNFAT is subject to further regulation by the activity of anumber of serine/threonine kinases (14). These either act in thenucleus to promote the export of nuclear NFAT, or, alterna-tively, in the cytosol to inhibit NFAT import. Many proteinkinases, including the MAPK group (p38, JNK, and ERK),GSK3�, PKA, MEKK1, and CK1�, have been reported to phos-phorylate different members of NFAT at different serine resi-dues (15–24). In most cases rephosphorylation is believed tofacilitate NFAT inactivation and export to cytoplasm, there aredata indicating that phosphorylation can also positively regu-late NFAT transactivation. For example, Okamura et al. re-ported inducible phosphorylation within a functional site of theNFATc2 transactivation domain upon stimulation of T cellswith PMA and Io (13). In addition, Pim kinase 1 (25) and Cotkinase (26) have been shown to phosphorylate NFATc1 andNFATc2, respectively, and to enhance their transactivationproperties.

Despite the importance of phosphorylation in the regulationof NFAT activity, the detailed molecular mechanisms underly-ing this process are little known. Particularly, it is still notclear which kinases phosphorylate NFATs under basal condi-tions and which ones mediate rephosphorylation and nuclearexclusion of the transcription factor after activation. Althoughproteomic analysis has allowed the identification of NFAT sitesphosphorylated under basal conditions and after stimulationwith PMA and Io (13), the exact identification of sites dynam-ically phosphorylated by particular kinases upon cell activationusing these techniques, are currently a technical challengebecause of the implication of different kinases and phosphata-ses and the low stoichiometry of transient phosphorylation. Wehave previously shown that p38 and JNK MAPKs phosphoryl-ate NFATc2 in vitro and interact with it in vivo. However,activation of the p38 pathway (but not JNK), results in theinhibition of NFATc2-driven transcription and nuclear accu-mulation (27). In this work we analyze the specific role of JNKpathway in the dynamic regulation of NFATc2 by phosphoryl-ation, identifying by mass spectrometry amino acid residues ofNFATc2 phosphorylated both in vitro and in vivo by JNK, andsystematically addressing the physiological relevance of these

sites. Our results show that activation of this pathway en-hances the transcriptional activity of this factor, and that mu-tation of identified residues, located in conserved regions (Fig.1) results in the inhibition of JNK-mediated NFAT transcrip-tional activity in Jurkat T cells.

EXPERIMENTAL PROCEDURES

Cell Culture and Reagents—HeLa and Jurkat T cells were culturedin Dulbecco’s modified Eagle’s medium or RPMI medium respectively,supplemented with 10% fetal calf serum, 2 mM glutamine, and a peni-cillin-streptomycin mixture. The JNK inhibitor SP600125 (46) waspurchased from Biomol, and PMA, Io, calcineurin, and calmodulin werepurchased from Sigma.

Expression Constructs and Transfections—Expression constructs forwild-type (wt) HA-tagged mouse NFATc2 (full-length), NFATLuc,MEKK1, JNK1, MKK6b(E), and p38� were as previously described (27).GAL-DBD Luc was kindly provided by Rosario Perona, and JIP1, JBD,and JNK KR were kind gifts from Pura Munoz. The empty vectorpcDNA 3.0 was from Invitrogen.

pAB GAL4-mouse NFATc2-(3–385) constructs were created by PCRfrom wt full-length HA-NFAT and subcloned into the pAB empty vector(47) between the SalI and PvuII sites. Mutations were conducted withthe QuikChange site-directed mutagenesis kit from Stratagene, in ac-cordance with the manufacturer’s indications. Western blot experi-ments using extracts of HeLa cells transfected with 1 �g of each con-struct confirmed comparable expression levels (data not shown).

Transfections in HeLa cells were performed with Lipofectamine Plusreagent (Invitrogen). Cells were cotransfected at 60–70% confluence inOPTIMEM medium (Invitrogen). The transfection mixture consisted of0.2 �g of reporter plasmid (one among NFAT Luc, IL-13 Luc, Ref. 32, orIL-4 Luc, Ref. 33), the HA-mouse NFATc2 full-length expression plas-mid (0.3 �g), and a kinase combination (0.3 �g each of constitutivelyactive MKK6b (E) plus p38�, or constitutively active MEKK1 plusJNK1). In control experiments the empty vector pcDNA 3.0 (0.6 �g),from Invitrogen, was substituted for the kinase combination. Transfec-tion took place over 3 h, and afterward cells were maintained in Dul-becco’s modified Eagle’s medium, 10% fetal calf serum. 24 h later, cellswere split, and after one night of recovery were stimulated with PMAand Io. Luciferase was assayed with the luciferase kit from Promega.

Jurkat T cells were also transfected with Lipofectamine reagent. Inthese experiments, 6 � 106 cells were cotransfected with 3 �g of totalDNA distributed as follows: 0.1 �g of expression vector pAB GAL4NFATc2 (wt or mutant); 1.2 �g of GAL-DBD Luc reporter plasmid; andeither 1.8 �g of JIP1, JBD, or JNK KR, or 1.8 �g of pcDNA 3.0 emptyvector, as indicated. For the JIP1 dose dependence experiment, theamounts of plasmid transfected were 1.8, 0.9, 0.45, and 0.2 �g, made upto 1.8 �g with pcDNA 3.0 empty vector as required. Transfections wereperformed for 4 h, and after that, cells were recultured in RPMI, 10%fetal calf serum at a concentration of 106 cells/ml. Stimuli were added16 h post-transfection. Luciferase determination was as described forHeLa cells.

When required, cells were stimulated with PMA and calcium iono-phore at a concentration of 20 ng/ml and 1 �M, respectively, for 3–5 h.In the case of HeLa cells, we supplemented stimulation medium with 3mM Ca2� to obtain a strong NFAT dephosphorylation and activation.The JNK inhibitor SP600125 was added 1 h before stimulation, at aconcentration of 20 �M.

For in vivo phosphorylation experiments, HEK cells were transfectedby the calcium phosphate method (48), with the constructs encoding thefull-length FLAG-tagged NFATc2 (10 �g), MEKK1 (10 �g), and JNK1(5 �g). In control experiments pcDNA 3.0 substituted the kinase com-bination. The total amount of DNA was completed up to 30 �g withpGL3 basic vector. Activation of JNK was determined by Western blotanalysis using an antiphospho-JNK antibody purchased from Cell Sig-naling Technology (cat. 9251).

In Vitro Calcineurin Binding Assay and Western Blot Experiments—The mouse NFAT N-terminal domain (amino acids 4–385) was HA-tagged and fused to GST. Mutants were based on this construct andwere generated by using the Stratagene mutation kit.

1 �g of GST-NFAT (either wt or the indicated mutant), adsorbed toglutathione-Sepharose beads (from Amersham Biosciences), was incu-bated with 20 nM calcineurin and 600 nM calmodulin in a binding buffercontaining 20 mM Tris, pH 8.0, 100 mM NaCl, 6 mM MgCl2, 1.5 mM

CaCl2, 0.2% Triton X-100, and a protease and phosphatase inhibitorsmixture, in a final reaction volume of 60 �l. The reaction was incubatedon a rocking platform at 4 °C for 30 min. Beads were then washed fivetimes with binding buffer and prepared for SDS-PAGE and Western

JNK-mediated Activation of NFATc220868

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from

blot analysis. Immunodetection was done with purified mouse mono-clonal anti-calcineurin antibody, acquired from Pharmingen (Pharmin-gen Europe), diluted 1:1000 in phosphate-buffered saline-0.1% Tween20.

Western blot assays to check MAPK activation were performed ontotal extracts from 106 JK cells, lysed in a buffer containing 20 mM

HEPES pH 8, 5 mM EDTA, 10 mM EGTA, 5 mM NaF, 10% glycerol, 1 mM

dithiothreitol, 400 mM KCl, 0.4% Triton X-100, 20 mM �-glycerophos-phate, and inhibitors (0.2 �M okadaic acid, 0.1 mM sodium orthovana-date, 2 �g/ml aprotinin, 5 �g/ml leupeptin, and 1 mM phenylmethylsul-fonyl fluoride). Lysates were rotated at 4 °C for 30 min and thencentrifuged at maximum speed to sediment cellular debris. Superna-tants were subjected to SDS-PAGE and immunodetection. JNK activa-tion was verified by c-Jun phosphorylation, using the monoclonal an-tiphospho-c-Jun antibody SC-822 purchased from Santa CruzBiotechnology diluted 1:1000 in TBS-0.1% Tween. Phospho-Erk andphospho-p38 were detected with anti-ACTIVETM MAPK from Promega(V6671), and antiphospho-p38 from Cell Signaling. Antibody bindingwas visualized in all cases using the enhanced chemiluminiscence(ECL) system from Amersham Biosciences.

In Vitro Kinase Assay and Proteolytic Digestions—1 �g of His-JNK2,kindly provided by Dr. J. Han (The Scripps Research Institute), and 2�g of GST-HA-mouse NFATc2-(4–385) were incubated at 37 °C for 30min in a kinase buffer containing 20 mM HEPES pH 7.5, 20 mM

�-glycerophosphate, 20 mM MgCl2, 0.1 mM sodium orthovanadate, and2 mM dithiothreitol. The reaction took place in the presence of 250 �M

ATP and 10 �Ci of [�32-P]ATP. Assays were stopped with Laemmlisample buffer and loaded onto a 7% SDS-PAGE gel.

Protein Analysis by Mass Spectrometry and Phosphopeptide Map-ping—The bands corresponding to phosphorylated proteins, were ex-cised and digested with trypsin (Promega) or V8 (Sigma) proteases, asdescribed (49). Peptides extracted from radiolabeled bands were sepa-rated by reverse phase HPLC as described (50), and analyzed by nano-spray off-line with ion trap mass spectrometry as described (50, 51),using an LCQ Classic model (Finnigan, ThermoQuest, San Jose, CA).

Peptides derived from the digestion of unlabeled NFAT bands, werealternatively analyzed by RP-HPLC-MS/MS using a 0.18 mm � 150mm BioBasic 18 RP column (ThermoHypersil-Keystone), operating at�1.5 �l/min, connected to a Surveyor HPLC system on-line with aLCQ-DECA XP ion trap mass spectrometer (Thermo Finnigan, SanJose, CA). Peptides were eluted using a 90-min gradient from 5 to 60%solvent B (solvent A: 0.5% acetic acid; solvent B: 0.5% acetic acid, 80%acetonitrile). Peptides were detected in survey scans from 400 to 1600amu (8 �scans), followed by a ZoomScan and a MS/MS analysis, usingan isolation width of 3 amu, a normalized collision energy of 35%.

Phosphopeptides were mapped by two-dimensional separation on20 � 20 thin layer cellulose plates purchased from Sigma. Electrophore-sis was carried out in a Hunter HTLE-7200 thin layer peptide mappingelectrophoresis system (CBS Scientific Company, Del Mar, California).Gels were run for 45 min in pH 8.9 buffer containing 10% w/v ammo-nium carbonate. The chromatography dimension was performed over-night in standard phosphochromatography buffer composed of 3.7%n-butyl alcohol (v/v), 2.5% pyridine (v/v), and 0.75% (v/v) glacial aceticacid (52). Phosphopeptide signals were visualized by autoradiography.

In Vivo Analysis of Phosphorylation—A minimum of 300 � 106 HEK

cells were transfected as described above. Forty-eight hours after trans-fection, every 5 � 106 cells were lysed in 100 �l of RIPA buffer (50 mM

Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 0.25%sodium deoxycholate, and a combination of proteases and phosphatasesinhibitors: leupeptin, aprotinin, pepstatin, phenylmethylsulfonyl fluo-ride, Na3VO4, and NaF). The lysates were incubated at 4 °C on arocking platform for 20 min. Lysates were cleared by centrifugation for15 min at 14,000 � g in a precooled centrifuge. Supernatants wereincubated for 1 h with 10 �l of RIPA-equilibrated anti-FLAG M2 aga-rose beads to immunoprecipitate the FLAG-NFATc2 protein.

Inmunoprecipitates were washed three times with RIPA and loadedonto a 6% gel for SDS-PAGE. The gel was stained with Sypro Ruby, andthe band corresponding to NFATc2 was excised and trypsin digested.The mixture was then analyzed by RP-HPLC on-line with an LTQlinear ion trap mass spectrometer, using the same HPLC setup andconditions as described above. Single ion reaction monitoring was pro-grammed over selected ions along the entire gradient.

RESULTS

JNK Activation Enhances NFAT-dependent Transcription—We have previously shown that activation of the p38, but notthe JNK/MAPK pathway reduces NFATc2 nuclear accumula-tion upon cell stimulation and results in inhibition of NFAT-driven transcription. However, p38 and JNK both interact withNFATc2 in vivo (27). For this reason, we decided to analyze thepotential role of the JNK/MAPK pathway in the regulation ofNFATc2. Fig. 1 shows the main regulatory sites in the regula-tory domain of NFATc2 that are conserved among other NFATfamily members, and indicates the positions of potential targetsites for JNK-mediated phosphorylation analyzed in the exper-iments described here. We initially carried out transient trans-fection experiments in HeLa cells, in which NFATc2 has beenshown to translocate to the nucleus in response to Ca2�/cal-cineurin stimulation (28, 29). We cotransfected cells with JNKand constitutively active MEKK1 to achieve complete activa-tion of the JNK pathway, together with full-length NFATc2and a luciferase reporter plasmid directed by a triple repeat ofthe NFAT/AP1 composite site from the interleukin-2 promoter(NFATLuc) (30). Activation of the JNK pathway led to anincrease in NFAT-dependent transcription and further poten-tiated the activation provoked by pharmacological stimulationof cells with PMA and Io (Fig. 2A). In contrast, parallel cotrans-fections using constitutively active MKK6 and p38, to activatethe p38 pathway, resulted in the predicted inhibition of NFAT-driven transcription (27).

Because the reporter construct used in these first experi-ments contains an NFAT/AP-1 composite site, the possibilityexists that AP-1, not NFAT, could be the target of the JNKMAPK pathway (31). Therefore, in a new set of experiments weexamined the ability of JNK to activate NFAT-driven luciferase

FIG. 1. Schematic representation of the N-terminal regulatory domain of NFATc2. The conserved regions among NFAT members aredetailed: SRR1 and SRR2 are the serine-rich regions. SP1, SP2, and SP3 are serine-proline motifs. CnBSA and CnBSB correspond, respectively,to the PxIxIT and LxVP motifs involved in calcineurin binding, and NLS is the nuclear localization signal. The residues analyzed in this study areindicated by dots (●).

JNK-mediated Activation of NFATc2 20869

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from

reporter plasmids, lacking NFAT:AP1 composite sites, such asthose directed by the IL13 or the IL4 promoters (32, 33). In bothcases, the luciferase reporter activity was activated by JNK,which also potentiated the action of PMA and Io, and thisactivation was similar to that displayed by the compositeNFAT/AP-1 (NFATLuc) reporter construct (Fig. 2B). These re-sults suggested that the JNK/MAPK pathway plays a role onthe regulation of NFATc2.

Mapping of the NFATc2 Residues Phosphorylated by JNK inVitro.—To determine which of the NFATc2 residues phospho-rylated by JNK could be responsible for the increased NFATtranscriptional activity we observed, we first performed in vitrophosphorylation experiments. We generated a recombinantconstruct containing the N-terminal part of NFATc2 protein(amino acids 4–385) fused to GST, and this construct wassubjected to in vitro phosphorylation by recombinant His-JNKin the presence of radiolabeled ATP. The phosphorylated frac-tion of the protein could be recognized on SDS-PAGE analysisbecause its migration is slightly retarded compared with theunphosphorylated protein (Fig. 3A). The phosphoprotein wasexcised from the gel and digested with trypsin, and the pep-tides were extracted as described under “Experimental Proce-dures.” The peptide extract was then fractionated by RP-HPLC, and the radioactivity of the fractions measured.Radioactivity was found in fractions 15, 16, 28, 31, and 35.Analysis of fractions 31 and 35 by nanospray-ion trap massspectrometry failed to detect the presence of any phosphopep-tide. However, a peptide ion of 492.5 Da was detected in frac-tion 15, whose tandem mass spectrometry fragmentation pat-tern agreed very well with that expected for the oxidized formof the peptide TSPIM(ox)SPR, phosphorylated in one of the twopossible serines that constitute consensus sequences for MAPK

phosphorylation (Ser followed by Pro). A detailed analysis ofthe fragment spectra revealed that the two possible phos-phopeptides, corresponding to Ser219 and Ser223 in the NFATc2sequence (Fig. 1), were in fact both present in the same 492.5Da peptide ion peak (Fig. 3B). Phosphopeptide analysis by thinlayer chromatography produced a map containing two of themost prominent spots detected in the non-fractionated sample(Fig. 3C, spots 1 and 2). When fraction 16 was analyzed by massspectrometry, an ion was detected whose mass, 732.7 Da, wasin agreement with that of the double-charged ion from peptideTSPDPTPVSTAPSK phosphorylated in one position. Analysisof the fragmentation spectra confirmed the sequence of thepeptide and allowed us to locate the phosphorylation in thesixth residue, corresponding to Thr332 (Figs. 3B and 1).The 492.5 Da ion was also detected in this fraction, though itsfragment spectra indicated that only Ser223 was phosphoryl-ated (not shown). These results are in good agreement withthose obtained by phosphopeptide analysis of this fraction (Fig.3, f.16) by thin layer chromatography, where one of the spots ofthe former map (Fig. 3C, spot 1) was also detected, in additionto a new one (Fig. 3C, spot 3). This allowed the assignment ofthe phosphopeptide corresponding to each one of the threespots. To corroborate all these results, we made single NAFTc2mutants where residues Ser219, Ser223, or Thr332 were substi-tuted by alanines, and subjected these mutants to in vitrophosphorylation with JNK. As predicted, the correspondingspots were found to disappear from the phosphopeptide mapsin each of the three cases (not shown).

In subsequent experiments, we tried to identify additional,less abundant phosphorylation sites by a more comprehensivemass spectrometry analysis, without the use of radioactivity.To fragment as many peptides as possible, the tryptic peptide

FIG. 2. The JNK/MAPK pathway activates NFATc2-driven transcription. A, JNK pathway activates transcription from the NFAT-Lucreporter, and potentiates the response to PMA and Io, whereas the p38 pathway suppresses it. HeLa cells were cotransfected with an expressionvector encoding full-length HA-NFATc2, the NFAT-responsive luciferase reporter plasmid NFAT-Luc (which contains a triple repeat of thecomposite NFAT:AP1 binding site from IL-2), and one of the following combinations of constructs: pcDNA 3.1 empty vector (MOCK); expressionvectors for p38 MAPK and constitutively active MKK6 (p38); or expression vectors for JNK1 and constitutively active MEKK1 (JNK). After 48 h,cells were left unstimulated (open bars) or stimulated for 5 h with 20 ng/ml PMA and 1 �M Io (solid bars). Luciferase activities are presented asthe means � S.D. from a representative experiment in which each measurement was performed in triplicate. B, transcription from promoters inwhich the NFAT site is not composite with AP1 shows a sensitivity to activation by JNK similar to that of NFAT-Luc. Hela cells were cotransfectedas in A, or with the NFAT/AP1-Luc construct replaced by luciferase reporter constructs driven by the promoter regions of the IL-4 or IL-13 genes,as indicated. Cells were stimulated as in A, and luciferase activities are presented as the means � S.D. from a representative experiment in whicheach measurement was performed in triplicate. R.L.U., relative light units.

JNK-mediated Activation of NFATc220870

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from

extracts from JNK-phosphorylated GST-NFATc2 were sub-jected to RP-HPLC on-line with microspray ion trap mass spec-trometry analysis. As expected, most fragment spectra corre-sponded to unmodified peptides from GST or NFATc2. Theremaining non-assigned spectra were inspected closely to de-tect potential phosphopeptides. Using this approach, we wereable to detect the three phosphorylation sites characterizedpreviously, plus two new monophosphorylated peptide ions of947.1 and 960.9 Da. The fragment spectra of these ions, respec-tively, corresponded to peptide FGEPDSIGFQNFLSPVKPA-GASGPsPR (Fig. 4A) and to the oxidized (ox) form of peptideIEItPSHELM(ox)QAGGALR (Fig. 4B), phosphorylated at theresidues indicated in lowercase letters, and corresponding toSer110 and Thr116 (Fig. 1). Finally, to identify possible addi-tional phosphorylation sites corresponding to very large pep-

tides, whose identification could pass unnoticed when usingtrypsin, we repeated the same experiment with V8 as theprotease. This approach revealed a further peptide ion at 909.8Da, corresponding to the carboxamidomethyl (ca) form of pep-tide GYREPLC(ca)LsPASSGSSA, phosphorylated at Ser170

(Figs. 4C and 1).Role of Ser110 and Thr116 in the Interaction of NFATc2 with

Calcineurin—Ser110 and Thr116, two of the residues phospho-rylated by JNK, mapped within the conserved SPRIEIT motifof NFATc2, which serves as calcineurin docking site (28) (Fig.1). We generated NFATc2 mutants by replacing Ser110 orThr116 with alanine in constructs encoding GST-NFATc2(amino acids 4–385), and analyzed the contribution of theseresidues to the in vitro NFATc2 binding to calcineurin. Asshown in lanes 1 and 3 of Fig. 5, binding of GST-NFATc2 S110A

FIG. 3. Identification of the NFATc2 residues phosphorylated by JNK in vitro. A, SDS-PAGE analysis of GST-NFATc2 phosphorylatedin vitro with JNK in the presence of radiolabeled ATP. Protein bands were visualized by Coomassie Blue staining. P and C indicate the positionof the phosphorylated and non-modified proteins, respectively; radioactivity was only found in the P band. This band was in-gel-digested, and theresulting peptides separated by RP-HPLC. B, MS/MS analysis by nanospray ion trap mass spectrometry of ions at m/z 492.5 in HPLC fraction 15(upper panel) and at m/z 732.7 in fraction 16 (lower panel). Neutral loss of phosphate from the ions is indicated by �P. The peptide sequence andthe assignment of the fragmentation series are also indicated, according to the nomenclature of Roepstorff and Fohlman (53). The phosphorylatedSer and Thr residues are indicated by lowercase letters. Superscripts 1 and 2 denote fragments phosphorylated at each of the two possible residues.C, two dimensional peptide mapping analysis by thin layer chromatography of HPLC fractions 15 and 16 (f.15, f.16). Spots 1, 2, and 3, respectively,correspond to: peptide 218TSPIMSPR225 phosphorylated at Ser223; to the same peptide phosphorylated at Ser219; and to peptide 327TSPDPTPV-STAPSK340 phosphorylated in residue Thr332. Spots were visualized by autoradiography.

JNK-mediated Activation of NFATc2 20871

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from

FIG. 4. Identification of phosphopep-tides by RP-HPLC online with ion trapmass spectrometry. Fragmentation spec-tra are shown of the ions at m/z 947.1 (A),960.9 (B), and 909.8 (C). The ions are labeledas in Fig. 3, except that the ‚ prefix indi-cates ions with 98 Da less, originated fromthe fragmentation of the ions produced afterneutral phosphate loss of the precursors.

JNK-mediated Activation of NFATc220872

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from

to purified calcineurin was comparable to that of the wild-typeprotein. But the T116A mutation completely inhibited bindingto calcineurin (Fig. 3, lane 4). This inhibition was even moreacute than that produced when the wild-type construct wasincubated with the VIVIT peptide, a high affinity competitor ofNFAT/calcineurin interaction obtained by combinatorial chem-istry (29). These results are consistent with previous studiesshowing that mutation of Thr116 impairs NFATc2 translo-cation in activated cells (28), and indicate that Thr116 plays animportant role in the NFATc2-calcineurin interaction.

Mutation of the JNK-targeted NFATc2 Residues Ser170 andThr116 Specifically Inhibits the Transactivation Activity ofNFATc2 in Vivo—We next applied site-directed mutagenesis toanalyze the role of each of the identified target sites of JNK inthe transactivation of NFATc2. Since mutation of sites such asThr116 can affect NFAT translocation, thus obscuring the in-terpretation of results, we used an experimental approachwhere translocation of NFAT was not dependent on any one ofthese sites. We fused the NFATc2 transactivation domain(amino acids 3–385) to the GAL4 DNA binding domain (GAL4-DBD) to generate a GAL4-NFATc2 chimeric construct. GAL4-DBD contains a NLS that confers constitutive nuclear localiza-tion (34). By using this chimera we also avoided potentialproblems that could arise from the activation of AP-1 or othertranscriptional partners that influence the activity of compos-ite NFAT reporter vectors. The GAL4-NFATc2 construct wasused as the parental vector to introduce single-site mutationsat each residue phosphorylated by JNK. In each case the res-idue was replaced by alanine, thus generating the GAL-4NFATc2 S110A, T116A, S170A, S219A, S223A, and T332Aconstructs (see Fig. 1 for location of residues). The relativecontribution of each site to NFATc2 transactivation activitywas assayed by contransfection of Jurkat T cells with theGAL4-NFATc2 construct (wild type or mutant) and a luciferasereporter gene containing downstream of five copies of a GAL4binding site. The activation of endogenous JNK in these cellswas achieved by treatment with PMA and Io, a combination ofstimuli known to elicit complete activation of JNK in T lym-phocytes (7). As observed in HeLa cells cotransfected withMEKK1 and JNK (Fig. 2A), the combined treatment of trans-fected Jurkat cells with PMA and Io increased GAL4-NFATc2-driven transcription, which was substantially inhibited in thepresence of the JNK inhibitor SP600125 (Fig. 6A). Similarresults were obtained when the transcriptional activities of theGAL4-NFATc2 mutants S110A, S219A, S223A, or T332A wereanalyzed in resting or activated Jurkat cells. In clear contrast,the GAL-4 NFATc2 S170A or T116A mutations markedly re-

duced the response to PMA and Io, strongly suggesting thatthese residues play an important role in JNK-mediated NFATactivation (Fig. 6A).

In another approach, we cotransfected Jurkat T lymphocyteswith constitutively active MEKK1 plus JNK. However, thiswas unsuccessful because this treatment invariably produced asustained activation of the JNK pathway and subsequent celldeath (data not shown), in good agreement with previous find-ings (35).

Simultaneous Mutation of the Serine Residues in the NFATc2Regulatory Domain SP1 or SP2 Motifs, but Not the SP3 Motif,Inhibits NFATc2 Transcriptional Activity—The N-terminalregulatory domains of NFATs contain two serine-rich regions(SRR1 and SRR2) and three serine/proline repeat motifs (SP1,SP2, and SP3). These domains are important for phosphoryla-tion and subcellular localization of NFAT proteins, and areconserved among all NFATc members (12, 36). All the phos-phorylation sites we have identified in this work, with theexception of Thr332, are located within conserved regulatorymotifs in NFATc2 (Fig. 1). In particular, Ser219 and Ser223 arelocated in the SP2 region. We were unable, however, to identifyby mass spectrometry peptides phosphorylated by JNK in anyof the residues within the SP1 (Ser186, Ser190, Ser194) or SP3(Ser276, Ser278, Ser282) regions. This observation does not con-firm that these sites are not phosphorylated by JNK, since nocomplete protein sequence coverage is to be expected by thisapproach, and neither were the unmodified tryptic peptidesspanning SP1 or SP3 regions detected by this technique (notshown). The SP motifs contain clusters of serines followed byprolines, representing consensus targets for MAPKs. So despitethe negative mass spectrometry findings, we decided to evalu-ate the contribution of the SP motifs to JNK-mediated NFAT-dependent transcription. For this we generated a set of mutantprotein constructs: GAL4-NFATc2 -SP1A, -SP2A, and -SP3A.These contain clustered serine-to-alanine substitutions in theSP consensus sites (Fig. 1). The activity of these mutants wasassayed in JNK-activated Jurkat cells cotransfected with theGAL-4 reporter plasmid. Remarkably, whereas individual mu-tation of Ser219 or Ser223 had failed to affect NFATc2 transac-tivation (as described above), the simultaneous mutation of thetwo sites produced a partial inhibition of NFATc2 transactiva-tion (data not shown). And when all three serine residues inSP2 were mutated this activity was almost completely inhib-ited. Simultaneous mutation of the three serines in the SP1 siteproduced a similar reduction in NFATc2 transactivation. Inclear contrast, the combined mutation of Ser276, Ser278, andSer282 from the SP3 motif had no impact (Fig. 6B).

Inhibition of JNK MAPK Signaling Blocks the Transcrip-tional Activity of NFATc2 Induced by PMA and Io in Jurkat TCells—To confirm that the JNK/MAPK pathway mediates theactivation of NFATc2 by PMA and Io we examined the effect ofthe JNK inhibitor SP600125 on the phosphorylation state of itssubstrate c-Jun. As shown in Fig. 6C, the phosphorylation ofc-Jun induced by PMA and Io was efficiently inhibited bySP600125, whereas the phosphorylation state of the ERK andp38 MAPKs, and of the p38 substrate ATF-2, were unaffected(Fig. 6C and data not shown).

SP600125 has been shown to target kinases other than JNK(54). As a general test for nonspecificity, we examined theactivation of one of these, the ribosomal protein S6 kinase 1(p70S6K1). We found that SP600125 had no effect on the acti-vation of p70S6K1 in our experiments; Western blot analysiswith an antibody specific for activated p70S6K1 showed thatthis kinase was efficiently activated by PMA and Io, but thatthis stimulation was not affected by SP600125. Inhibition ofJNK in this experiment was confirmed by blockade of c-Jun

FIG. 5. Mutation of the NFATc2 Thr116 residue, but not theSer110 residue, blocks in vitro binding of GST-NFATc2 to cal-cineurin. Calcineurin pull-down assays were performed with wild-typeGST-NFATc2 (wt) (lane 1), or the GST-NFATc2 mutants S110A (lane 3)or T116A (lane 4) as bait. Calcineurin union was detected by Westernblotting with a monoclonal anti-calcineurin antibody. Calcineurin bind-ing to wild-type GST-NFAT was competed with 20 �M VIVIT peptide(lane 2). The locations of mutations within the SPRIEIT domain areindicated above the blot.

JNK-mediated Activation of NFATc2 20873

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from

phosphorylation in the same cell extracts (Fig. 1, Supplemen-tary Data).

Further evidence for the role of JNK in NFATc2 activationwas obtained with different dominant-negative inhibitors ofJNK activity. JIP1 (JNK-interacting protein 1 full-length) is ascaffold protein of the JNK activation complex, and JBD is theJIP1 JNK binding domain. Overexpression of full-length JIP1or JBD causes cytosolic retention of JNK, thereby inhibitingJNK signaling (37). JNK1 KR is a catalytically inactive mutantof JNK1 (38).

Jurkat T cells were cotransfected with an expression vectorfor one of these dominant-negative proteins, GAL4-NFATc2,and the GAL-Luc reporter. As shown in Fig. 7A, expression ofJIP1, JBD, or JNK KR significantly decreased the NFATc2-de-pendent transactivation induced by PMA and Io in Jurkat Tcells. The inhibitory effect of JIP1 was dose-dependent, anddeclined progressively as the amounts of JIP1 plasmid usedwere decreased (Fig. 7B).

JNK Phosphorylates Thr116 in Vivo—Our in vitro experi-ments identified serine and threonine residues in NFATc2 thatwere phosphorylated by JNK (Fig. 3). To determine whetherphosphorylation by JNK could be detected in vivo, we moni-tored single ion reactions of selected phosphopeptides in tryp-sin digests of NFATc2 immunoprecipitates from cells trans-fected with FLAG-tagged NFATc2 expression vectors. To attainthe highest degree of sensitivity and specificity, we focused onthe phosphorylated tryptic peptides containing Thr116, one ofthe two major functional residues of NFATc2 identified in thein vitro studies. Because PMA plus ionophore treatment acti-vates other kinases besides JNK (including ERK and p38), cellswere cotransfected with a combination of MEKK1 and JNKconstructs to achieve a complete and specific activation of JNK

pathway. HEK cells were used for these experiments becausethis treatment provoked cell death in Jurkat cells, as observedby other authors (35). The efficient activation of JNK byMEKK1 in transfected cells was confirmed in control immuno-blot experiments using a specific antiphospho-JNK antibody(Fig. 8A). As shown in Fig. 8, D and E the control, non-phos-phorylated peptide (113IEITPSHELMQAGGALR129) wasclearly and specifically detected in either mock-transfected orJNK-activated cells, at similar levels and retention times.However, the peptide phosphorylated at Thr116 could only bedetected in JNK-activated cells (compare Fig. 8, F and G); theexact sequence and modification site corresponding to the mon-itored ion was evident from the fragment spectra obtained atthe peak apex (Fig. 8G, inset). The base peak chromatogram,showing the intensity of several NFATc2-derived peptides (Fig.8, B and C), provided a further internal control, demonstratingthat the NFATc2 concentration was identical in both cases.However, the intensity of phosphorylated peptide was alwaysgreater in JNK-activated cells. Therefore, these experimentsprovide evidence that JNK phosphorylates NFATc2 in vivo atthe Thr116 site.

DISCUSSION

The data presented in this study provide the first full char-acterization of the positive regulation of NFATc2 by JNK, akinase previously considered an exclusively negative regulatorof the nuclear accumulation and transcriptional activity ofNFATc1 and NFATc3. Moreover, our finding that JNK targetsboth in vitro and in vivo Thr116, a residue located within thedocking site for calcineurin in the NFATc2 regulatory domain,suggests an important role for JNK in modulating the cal-cineurin-NFAT interaction. The observation that simultaneous

FIG. 6. Contribution of JNK-phosphorylated sites to NFATc2-driven transcription. Jurkat T cells were cotransfected with a GAL-Lucreporter plasmid and the GAL4-NFATc2 WT or one of the following mutant constructs: S110A, T116A, S170A, S219A, S223A, or T332A (A); SP1,SP2, or SP3 (B). 24 h post-transfection, cells were left unstimulated (Control, open bars) or stimulated with 20 ng/ml PMA and 1 �M Io (PMA�Io,solid bars) to activate JNK. Pretreatment with SP600125 was for 1 h before JNK activation (gray bars, SP/PMA�Io). Data are presented as thefold-induction of treated cells (PMA�Io or SP/PMA�Io) versus nontreated cells (Control) (mean � S.D.; n � 3). C, specificity of the effect ofSP600125 on the activation of MAPK pathways was assessed in Jurkat cells pretreated for 30 min with 20 �M SP600125 and afterward stimulatedwith PMA�Io for 1 h. Activation of MAPKs was detected by Western blotting. JNK activation was determined by c-Jun phosphorylation; ERK andp38 activation was detected by the presence of phospho-ERK or phospho-p38, using antiphospho-ERK and antiphospho-p38 antibodies, respec-tively. Phosphorylated proteins are indicated by arrows.

JNK-mediated Activation of NFATc220874

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from

mutation of JNK-targeted residues in the NFATc2 SP1 andSP2 motifs inhibits NFATc2 transactivation reveals hithertounidentified roles for these regions in NFAT activity.

The established view is that NFAT is activated by dephos-phorylation of residues within a conserved regulatory domainlocated at the N-terminal region. In the case of NFATc2, massspectrometry has revealed that 13 conserved serine residuesare desphosphorylated upon stimulation (13). Dephosphoryl-ation has been suggested to result in NFATc2 activationthrough a conformational switch to the active form of the factor(nuclear localization sequence exposed, nuclear export se-quence masked). The residues that are dephosphorylated arelocated in diverse sequence contexts within the N-terminalregulatory domain, and it is thought that different constitutiveand inducible kinases are involved in the rephosphorylation ofNFATc2. In addition, although the presence of conserved mo-tifs in the N-terminal domain could suggest a common regula-tory mechanism for all NFATs, the overall homology in thisregion is relatively low among the members, which could ac-count for member-specific regulatory mechanisms. A number ofstudies have identified different amino acid residues targetedby several NFAT kinases involved in the nuclear shuttling ofdifferent NFAT members (reviewed in Refs. 12 and 39). How-ever, gaps remain in current understanding of the cycle ofdephosphorylation/rephosphorylation of NFATs.

Among the protein kinases implicated in the negative regu-lation of NFAT, CKI�, and GSK3� have been identified asconstitutive NFAT kinases that phosphorylate conserved mo-tifs in different members. CKI directly phosphorylates NFATc3and NFATc2, suppressing NFAT nuclear import in synergywith MEKK1 (17) or GSK3 (24), respectively. Phosphorylation

by GSK3, in contrast, promotes the nuclear export of NFATc4(18) and NFATc1 (16). GSK3-mediated phosphorylation re-quires the activity of a priming kinase, and PKA has beenshown to phosphorylate NFAT and to function as a primingkinase for GSK3 (23). The inducible MAPKs p38 and JNK havealso been reported to counteract NFAT nuclear accumulation.p38 has been shown to target NFATc2 (27), NFATc4 (40), andNFATc1 (41), whereas JNK opposes the nuclear accumulationof NFATc3 (15), NFATc1 (20), and NFATc2 (42).

In contrast to these reports, which describe NFAT phospho-rylation opposing nuclear accumulation and inhibiting NFAT-dependent transcription, recent studies provide evidence thatphosphorylation can in some cases enhance NFAT transactiva-tion. This is the case for the inducible Ser phosphorylationwithin residues 53 to 56 (53SSPS56) of NFATc2, since mutationof these serines to alanines effectively abolishes the transcrip-tional activity of the transactivation domain in response toPMA and Ionomycin (13). Similarly, it has been reported thatCot kinase potentiates the transactivation function of NFATc2by acting directly on its regulatory domain in a calcineurin-independent fashion (26). Pim-1 kinase has also been shown tophysically interact with and phosphorylate NFATc1, and toactivate NFAT-dependent transcription and IL-2 production inT cells (25). However, despite these studies, the residues inNFATc2 or NFATc1 that are phosphorylated by such kinases toachieve this transactivation have not been identified. Morerecently, the ERK/RSK signaling pathway has been implicatedin the activation of both the DNA binding and transcriptionalactivity of NFATc4, through direct phosphorylation of its DNAbinding domain (19).

We had previously reported that JNK phosphorylatesNFATc2 and can interact with it in vivo (27). In order to furtheranalyze the regulation of NFATc2 by JNK at the molecularlevel, we tried to conduct functional experiments in Jurkat Tcells, which express endogenous NFATc2, by cotransfectionwith MEKK1 and JNK1. However, the potent, sustained acti-vation of JNK in these cells led to programmed cell death ((35)and our own data not shown). We therefore employed an indi-rect approach to evaluate the role of the JNK pathway in thetransactivation of NFATc2, based on the combined stimuluswith PMA and Io. This mimicks signals from both the T cellreceptor (TCR)-CD3 complex and the costimulatory factorCD28 (7, 8), and elicits a complete and synergistic activation ofJNK in T lymphocytes. This approach was possible since theJNK inhibitor SP600125 was found to inhibit most of the in-ducible activity of the GAL4 constructs, and did not affect otherkinases activated by PMA and Io such as ERK and p38 MAPKs,or p70S6K1. Furthermore, expression of dominant-negativesuppressors of JNK signaling (JIP1, JBD, and JNK KR) inhib-ited the transcriptional activation of NFATc2 in Jurkat cells,and this effect was dose-dependent for JIP1. This providescomplementary evidence that JNK signaling is required forNFATc2 transcription induced by PMA plus Io.

Our in vitro data show that NFATc2 can be phosphorylatedby JNK at specific conserved residues. And at least in the caseof residue Thr116, this can also occur in vivo. We have alsoshown that activation of the JNK pathway enhances the tran-scriptional activity of NFATc2, and that mutation of some ofthe identified target residues for phosphorylation, preventsboth NFAT binding to calcineurin and JNK-activated NFATtranscriptional activity in Jurkat T cells. Some of the conservedmotifs of NFATc2 we have identified as targets of JNK havebeen shown previously to be phosphorylated by other kinasesand/or to play roles in the subcellular localizalization of otherNFAT members. CK1 targets the SP1 domain of NFATc3, andits deletion results in constitutive nuclear localization of

FIG. 7. Dominant-negative JNK inhibits NFATc2 transactiva-tion. Jurkat T cells were cotransfected with GAL4-NFATc2, GAL-Lucand plasmids encoding (A) one of three dominant-negative versions ofJNK: JIP1, JBD, or JNK KR, or (B) with decreasing amounts of JIP1.Twenty-four hours after transfection, cells were left unstimulated (openbars) or stimulated with PMA and Io (solid bars), and luciferase activitywas measured after 4 h. Data are presented as the fold-induction oftreated cells (PMA�Io) versus nontreated cells (Control) (mean � S.D.;n � 3).

JNK-mediated Activation of NFATc2 20875

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from

NFATc3, whereas the deletion of the SRR1 region, proposed asthe docking site for CK1, renders NFATc3 import highly re-sponsive to calcineurin activation (17). CKI has also more re-cently been shown to target the SRR1 region of NFATc2, and topromote NFATc2 nuclear export (24). However, the specificresidues phosphorylated by CKI have not been identified. Adifferent study suggests that JNK targets serines 163 and 165in NFATc3 (equivalent to Ser170 in NFATc2), which are alsowithin the SRR1 region (15). Mutation of these residues causednuclear accumulation of NFATc3. JNK similarly regulates thesubcellular localization of NFATc1 by targeting residue Ser170

in the SRR1 region (Ser170 in NFATc2) and Ser117 (Ser110 in

NFATc2), which is adjacent to the PXIXIT calcineurin bindingmotif (20).

Our data suggest roles for JNK-mediated phosphorylation ofNFAT different from those suggested by the data obtained byother groups. First, we have not observed any significant effectof JNK on the subcellular localization of NFATc2 upon cellstimulation with calcium ionophore. Similar to others, we haveidentified Ser170 as a JNK site within the SRR1 of NFATc2.However, unlike NFATc3 and NFATc1, the overall effect ofJNK activation in NFATc2 was to increase NFAT-dependenttranscription and transactivation, and the mutation of Ser170

resulted in inhibition of the inducible transcriptional activity of

FIG. 8. NFATc2 is specifically phosphorylated by JNK at Thr116 in vivo. NFATc2 was immunoprecipitated from control (MOCK) orJNK-activated (MEKK1�JNK) NFATc2-transfected HEK cells. Control of activation of JNK by MEKK1 in transfected cells was performed byimmunoblot analysis using an antiphospho-JNK-specific antiserum (A). After trypsin digestion, peptides containing Thr116 were monitored byRP-HPLC on line with a LTQ linear ion trap mass spectrometer. B and C, base-peak chromatograms showing that the intensity of someNFATc2-derived peptides (m/z indicated on the peaks) is identical in control and JNK-activated cells. E and G, chromatogram traces of fragmention y13

� (m/z 1366.5) obtained from the fragmentation of the doubly charged precursor ions at m/z 912.0, corresponding to peptide 113IEITP-SHELMQAGGALR129 (D and E), or at m/z 952.0, corresponding to the monophosphorylated form of the same peptide (F and G). Intensities areexpressed in arbitrary units, taking the intensity of the unmodified peptide as a reference. Numbers indicate the retention times, and insets showthe MS/MS spectra obtained at the peak apex. Note that the MS/MS inset in G confirms peptide phosphorylation at Thr116.

JNK-mediated Activation of NFATc220876

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from

the transactivation domain by PMA and Io. These results in-dicate that the same kinase may target the same region withdifferent functional effects depending on the member, and thatdifferent kinases can also act on the same regions. This is thecase with p38 MAPK, which has been shown to phosphorylateSer168 and Ser170 (Ser170 in NFAc2 and NFATc1) of the SRR1region of NFATc4 promoting its nuclear exclusion. Interest-ingly, this phosphorylation is selective for NFAT member andkinase, since p38 phosphorylates NFATc4 very efficiently butNFATc3 only weakly, and NFATc4 is not phosphorylated byJNK (40). Similarly, we have found that Ser110 in NFATc2(homolog of Ser117 in NFATc1) is phosphorylated by JNK.However, whereas phosphorylation of Ser117 by JNK inhibitscalcineurin targeting to NFATc1, our work shows that themutation of Ser110 failed to interfere with the binding to cal-cineurin or to affect the transcriptional activity of NFATc2.Despite these differences among NFAT members, the PXIXITmotif appears to be highly regulated by kinases, not only viathe important role of Ser117 in the subcellular localization ofNFATc1 (20), but also through the requirement for Thr116

(located within the PXIXIT motif of NFATc2) for binding cal-cineurin, and for the inducible transcriptional activity ofNFATc2 that we report here.

An additional difference among different members ariseswhen the mechanisms of masking/unmasking the NLS arecompared. Thus, the NFATc3 NLS has been shown to remainmasked under basal conditions due to SP1 phosphorylation(14), whereas the SP1 region in NFATc2 has been reported tobe unphosphorylated in resting T cells (13). Although we havenot addressed the role of the SP1 region in the subcellularlocalization of NFATc2, our analysis of transactivation with theGAL4 mutant constructs indicates that this region plays animportant role in the inducible transactivation of NFATc2. Inaddition, we have observed a reproducible increment in thebasal activity of the SP1A mutant when compared with theother constructs (data not shown). This result could be attrib-uted to the role of SP1 in NLS masking/unmasking. AnotherNFATc2 residue previously shown to be un-phosphorylated inresting cells is Thr116 (13). These data are in agreement withour mass spectrometry results, which show that phosphoryl-ated Thr116 is undetectable in mock-transfected cells. ButThr116 was phosphorylated in vivo upon activation of the JNK/MAPK pathway. Our results also show that individual muta-tion of this residue results in complete blockade of the in vitrobinding of NFATc2 to calcineurin, and also dramatically affectsthe transactivation activity of NFATc2 induced by PMA and Io.Previous studies have shown that although NFATc2 nuclearlocalization is sufficient for NFAT-dependent transcription, ex-tensive dephosphorylation is required for maximal activation(13). For this reason, we do not exclude the possibility that thedefect in transactivation observed with the T116A mutantcould in part be caused by its inability to bind calcineurin, andhence undergo complete dephosphorylation.

We have shown that SP2 is also a region phosphorylated byJNK in vitro, and involved in NFATc2 transactivation inducedby PMA and Io in vivo. This region has not been implicatedpreviously in the regulation of nuclear localization or transac-tivation of other NFAT members, but we have seen that al-though single mutants at residues 219 or 223 have no effect onNFAT-dependent transcription, the triple mutant 215-219-223(mutant SP2A) and the double mutant 219 and 223 reduced theinduced NFAT activity. These results suggest that some or allthe residues phosphorylated may not have a role by them-selves, but may act in cooperation with other residues to func-tionally contribute to the transcriptional activity of theNFATc2 transactivation domain.

Our findings show that JNK regulates not only NFAT sub-cellular localization, but can also actively cooperate in enhanc-ing NFATc2-dependent transcription, by phosphorylating spe-cific conserved residues that appear to play different rolesdepending on the member analyzed. The regulation of NFATactivation and subcellular localization is further complicatedby the fact that different kinases operate in these processes andcan even target the same homologous residues in differentmembers. This is the case of Ser168 and Ser170, which arephosphorylated by p38 in NFATc4, and the equivalents Ser163

and Ser165 in NFATc3, which are phosphorylated by JNK(Ser170 in NFATc2). An additional degree of complexity appearsto be related with cell-type specific regulation of NFATs, as thekinases could display different activity depending of the celltype. This seems to be the case with JNK, which promotescytoplasmic accumulation of NFATc2 in cardiomiocytes, aneffect that is mediated by p38 (but not JNK) in HeLa cells (27).Likewise, it seems clear that more kinases might be implicatedin the regulation and activation of NFATc2, such as thoseinvolved in the phosphorylation of Ser53 and Ser56 of SSPS,which have previously been shown to play a functional role inthe inducible transactivation of NFATc2 in Jurkat cells (13)

Finally, the role of JNK in the activation of NFAT has beenanalyzed in Jnk1- and Jnk2-deficient mice, and these studieshave yielded different results, with both a positive and negativeroles reported for JNK in NFAT regulation in lymphocytes(43–45). However, as yet there has been no detailed study ofthe effects of JNK isoforms on the different NFAT members.Given the redundant expression of NFATc1, c2, and c3 mem-bers in lymphocytes, and the different roles of JNK in theactivation of these members, it is difficult to establish theimpact of JNK on NFATc2 in these studies.

Our results are consistent with a model in which calcineurinbinds and dephosphorylates specific NFATc2 residues, causingthe exposure of the NLS and the nuclear translocation of thefactor. JNK would phosphorylate residue 116 and probably170, and other residues within SP2 and SP1, promoting NFAT-driven transcription. NFATc2-induced transcription would alsobe modulated by other kinases such as Cot, Pim, or the SSPSkinase, which would also be involved in the transcription proc-ess. Later, calcineurin would dephosphorylate these residues,and other kinases would act to return NFATc2 to its basalphosphorylation state and induce nuclear export of the factor.

Acknowledgments—We thank Pablo Gomez del Arco for his help inthe initiation of this project in the laboratories of Jiahuai Han andJuan Miguel Redondo. We also thank Jiahuai Han for providing thedifferent kinase constructs and Anjana Rao for the HA-NFATc2 con-struct; Jose Aramburu and Miguel Angel Iniguez for helpful discussionand comments on the manuscript; Simon Bartlett for editing work; andSamuel Ogueta for excellent technical support.

REFERENCES

1. Widmann, C., Gibson, S., Jarpe, M. B., and Johnson, G. L. (1999) Physiol. Rev.79, 143–180

2. Hibi, M., Lin, A., Smeal, T., Minden, A., and Karin, M. (1993) Genes Dev. 7,2135–2148

3. Derijard, B., Raingeaud, J., Barrett, T., Wu, I. H., Han, J., Ulevitch, R. J., andDavis, R. J. (1995) Science 267, 682–685

4. Manning, A. M., and Davis, R. J. (2003) Nat. Rev. Drug. Discov. 2, 554–5655. Davis, R. J. (2000) Cell 103, 239–2526. Crabtree, G. R. (1989) Science 243, 355–3617. Su, B., Jacinto, E., Hibi, M., Kallunki, T., Karin, M., and Ben-Neriah, Y. (1994)

Cell 77, 727–7368. Rincon, M., and Flavell, R. A. (1994) EMBO J. 13, 4370–43819. Rao, A., Luo, C., and Hogan, P. G. (1997) Annu. Rev. Immunol. 15, 707–747

10. Jain, J., Loh, C., and Rao, A. (1995) Curr. Opin. Immunol. 7, 333–34211. Crabtree, G. R., and Clipstone, N. A. (1994) Annu. Rev. Biochem. 63,

1045–108312. Hogan, P. G., Chen, L., Nardone, J., and Rao, A. (2003) Genes Dev. 17,

2205–223213. Okamura, H., Aramburu, J., Garcia-Rodriguez, C., Viola, J. P., Raghavan, A.,

Tahiliani, M., Zhang, X., Qin, J., Hogan, P. G., and Rao, A. (2000) Mol. Cell6, 539–550

14. Shibasaki, F., Price, E. R., Milan, D., and McKeon, F. (1996) Nature 382,

JNK-mediated Activation of NFATc2 20877

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from

370–37315. Chow, C. W., Rincon, M., Cavanagh, J., Dickens, M., and Davis, R. J. (1997)

Science 278, 1638–164116. Beals, C. R., Sheridan, C. M., Turck, C. W., Gardner, P., and Crabtree, G. R.

(1997) Science 275, 1930–193417. Zhu, J., Shibasaki, F., Price, R., Guillemot, J. C., Yano, T., Dotsch, V., Wagner,

G., Ferrara, P., and McKeon, F. (1998) Cell 93, 851–86118. Graef, I. A., Mermelstein, P. G., Stankunas, K., Neilson, J. R., Deisseroth, K.,

Tsien, R. W., and Crabtree, G. R. (1999) Nature 401, 703–70819. Yang, T. T., Xiong, Q., Graef, I. A., Crabtree, G. R., and Chow, C. W. (2005)

Mol. Cell. Biol. 25, 907–92020. Chow, C. W., Dong, C., Flavell, R. A., and Davis, R. J. (2000) Mol. Cell. Biol. 20,

5227–523421. Porter, C. M., Havens, M. A., and Clipstone, N. A. (2000) J. Biol. Chem. 275,

3543–355122. Neal, J. W., and Clipstone, N. A. (2001) J. Biol. Chem. 276, 3666–367323. Sheridan, C. M., Heist, E. K., Beals, C. R., Crabtree, G. R., and Gardner, P.

(2002) J. Biol. Chem. 277, 48664–4867624. Okamura, H., Garcia-Rodriguez, C., Martinson, H., Qin, J., Virshup, D. M.,

and Rao, A. (2004) Mol. Cell. Biol. 24, 4184–419525. Rainio, E. M., Sandholm, J., and Koskinen, P. J. (2002) J. Immunol. 168,

1524–152726. de Gregorio, R., Iniguez, M. A., Fresno, M., and Alemany, S. (2001) J. Biol.

Chem. 276, 27003–2700927. Gomez del Arco, P., Martinez-Martinez, S., Maldonado, J. L., Ortega-Perez, I.,

and Redondo, J. M. (2000) J. Biol. Chem. 275, 13872–1387828. Aramburu, J., Garcia-Cozar, F., Raghavan, A., Okamura, H., Rao, A., and

Hogan, P. G. (1998) Mol. Cell 1, 627–63729. Aramburu, J., Yaffe, M. B., Lopez-Rodriguez, C., Cantley, L. C., Hogan, P. G.,

and Rao, A. (1999) Science 285, 2129–213330. Rooney, J. W., Sun, Y. L., Glimcher, L. H., and Hoey, T. (1995) Mol. Cell. Biol.

15, 6299–631031. Derijard, B., Hibi, M., Wu, I. H., Barrett, T., Su, B., Deng, T., Karin, M., and

Davis, R. J. (1994) Cell 76, 1025–103732. Macian, F., Garcia-Rodriguez, C., and Rao, A. (2000) EMBO J. 19, 4783–479533. Szabo, S. J., Gold, J. S., Murphy, T. L., and Murphy, K. M. (1993) Mol. Cell.

Biol. 13, 4793–480534. Silver, P. A., Keegan, L. P., and Ptashine, M. (1984) Proc. Natl. Acad. Sci.

U. S. A. 81, 5951–5955

35. Chen, Y., and Lai, M. Z. (2001) J. Biol. Chem. 276, 8350–835736. Kiani, A., Garcia-Cozar, F. J., Habermann, I., Laforsch, S., Aebischer, T.,

Ehninger, G., and Rao, A. (2001) Blood 98, 1480–148837. Yasuda, J., Whitmarsh, A. J., Cavanagh, J., Sharma, M., and Davis, R. J.

(1999) Mol. Cell. Biol. 19, 7245–725438. Li, Y. S., Shyy, J. Y., Li, S., Lee, J., Su, B., Karin, M., and Chien, S. (1996) Mol.

Cell. Biol. 16, 5947–595439. Crabtree, G. R., and Olson, E. N. (2002) Cell 109, (suppl.) S67–S7940. Yang, T. T., Xiong, Q., Enslen, H., Davis, R. J., and Chow, C. W. (2002) Mol.

Cell. Biol. 22, 3892–390441. Wu, C. C., Hsu, S. C., Shih, H. M., and Lai, M. Z. (2003) Mol. Cell. Biol. 23,

6442–645442. Liang, Q., Bueno, O. F., Wilkins, B. J., Kuan, C. Y., Xia, Y., and Molkentin,

J. D. (2003) EMBO J. 22, 5079–508943. Dong, C., Yang, D. D., Wysk, M., Whitmarsh, A. J., Davis, R. J., and Flavell,

R. A. (1998) Science 282, 2092–209544. Behrens, A., Sabapathy, K., Graef, I., Cleary, M., Crabtree, G. R., and Wagner,

E. F. (2001) Proc. Natl. Acad. Sci. U. S. A. 98, 1769–177445. Sabapathy, K., Kallunki, T., David, J. P., Graef, I., Karin, M., and Wagner,

E. F. (2001) J. Exp. Med. 193, 317–32846. Bennett, B. L., Sasaki, D. T., Murray, B. W., O’Leary, E. C., Sakata, S. T., Xu,

W., Leisten, J. C., Motiwala, A., Pierce, S., Satoh, Y., Bhagwat, S. S.,Manning, A. M., and Anderson, D. W. (2001) Proc. Natl. Acad. Sci. U. S. A.98, 13681–13686

47. Baniahmad, A., Kohne, A. C., and Renkawitz, R. (1992) EMBO J. 11,1015–1023

48. Rodriguez, A., and Flemington, E. K. (1999) Anal. Biochem. 272, 171–18149. Shevchenko, A., Jensen, O. N., Podtelejnikov, A. V., Sagliocco, F., Wilm, M.,

Vorm, O., Mortensen, P., Boucherie, H., and Mann, M. (1996) Proc. Natl.Acad. Sci. U. S. A. 93, 14440–14445

50. Ogueta, S., Rogado, R., Marina, A., Moreno, F., Redondo, J. M., and Vazquez,J. (2000) J. Mass Spectr. 35, 556–565

51. Marina, A., Garcia, M. A., Albar, J. P., Yague, J., Lopez de Castro, J. A., andVazquez, J. (1999) J. Mass Spectr. 34, 17–27

52. Boyle, W. J., van der Geer, P., and Hunter, T. (1991) Methods Enzymol. 201,110–149

53. Roepstorff, P., and Fohlman, J. (1984) Biomed. Mass Spectr. 11, 60154. Bain, J., McLauchlan, H., Elliott, M., and Cohen, P. (2003) Biochem. J. 371,

199–204

JNK-mediated Activation of NFATc220878

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from

Juan Miguel RedondoInmaculada Ortega-Pérez, Eva Cano, Felipe Were, Margarita Villar, Jesús Vázquez and

through Phosphorylation within the N-terminal Regulatory Domainc-Jun N-terminal Kinase (JNK) Positively Regulates NFATc2 Transactivation

doi: 10.1074/jbc.M501898200 originally published online March 2, 20052005, 280:20867-20878.J. Biol. Chem. 

  10.1074/jbc.M501898200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

Supplemental material:

  http://www.jbc.org/content/suppl/2005/03/10/M501898200.DC1

  http://www.jbc.org/content/280/21/20867.full.html#ref-list-1

This article cites 54 references, 32 of which can be accessed free at

by guest on February 13, 2018http://w

ww

.jbc.org/D

ownloaded from