Regulation of the Murine Nfatc1 Gene by NFATc2*

NFAT proteins play a key role in the inducible expression of cytokine genes in T lymphocytes. NFATc1 and NFATc2 are the predominant NFAT family members in the peripheral immune system. NFATc2 is found abundantly in the cytoplasm of resting T cells, whereasNfatc1 expression is induced during T cell activation. To investigate Nfatc1 regulation, we characterized the structure of the murine Nfatc1 gene and its 5′-flanking region. A 290-bp sequence proximal to the transcription start site is highly conserved between mouse and human and possesses both basal and inducible promoter activities. Multiple binding sites for transcription factors were identified within this region, including a consensus NFAT-binding site. This promoter segment was cyclosporin A-sensitive, and mutation of the NFAT site abrogated inducible promoter activity and inhibited formation of an inducible DNA·protein complex containing NFATc2 in primary T cells. Overexpression of NFATc2 increased inducibleNfatc1 promoter activity, whereas this inducibility was attenuated in NFATc2−/− splenocytes. This study suggests that pre-existing NFATc2 contributes to the subsequent induction ofNfatc1 during T cell activation.

The regulation of NFAT transcriptional activity during T cell activation is controlled at several levels. NFAT proteins exist mainly in an inactive phosphorylated state in the cytoplasm of resting T cells. Activation requires a sustained increase in intracellular Ca 2ϩ induced by T cell receptor engagement or the calcium ionophore ionomycin, which in turn activates the Ca 2ϩ -dependent phosphatase calcineurin (11). Calcineurin dephosphorylates NFAT proteins and induces their nuclear translocation (4,12,13). The immunosuppressants cyclosporin A (CsA) and FK506, which block the phosphatase activity of calcineurin (14,15), inhibit T cell activation by preventing the nuclear translocation of NFAT (16) and subsequent activation of cytokine gene transcription. The nuclear import of NFAT is also blocked by MEKK1 and casein kinase-1␣, which mask the nuclear import signal (17). Export of NFAT out of the nucleus is also highly regulated. Glycogen synthase kinase-3, a nuclear kinase, phosphorylates NFATc1 at the same sites that are dephosphorylated by calcineurin and thus promotes NFATc1 nuclear export (18). Similarly, NFATc3 is exported from the nucleus after phosphorylation by JNK (19). A critical balance between the import and export of NFAT in and out of the nucleus is a well established mechanism of NFAT regulation (20).
Several lines of evidence suggest that increased NFAT transcriptional activity in T cells is not solely due to the nuclear translocation of NFAT proteins. For example, the DNA-binding affinity of NFAT proteins is greatly enhanced by cooperative interaction with AP-1 proteins (9). AP-1 protein synthesis is induced by the activation of protein kinase C and Ras following T cell receptor engagement. This can be mimicked in vitro by treatment with the phorbol ester phorbol 12-myristate 13-acetate (PMA). In addition, the transcriptional activity of NFAT proteins has recently been demonstrated to depend on the phosphorylation status of critical amino acid residues in the transcriptional activation domain (21). Finally, levels of NFAT expression may contribute to overall NFAT activity. NFATc2 is constitutively expressed in resting human peripheral blood T cells (22)(23)(24), whereas expression of NFATc1 is greatly induced upon T cell stimulation (2, 22, 24 -26). Moreover, the inducible expression of NFATc1 in activated T cells is CsA-sensitive (22), indicating that the calcium-dependent calcineurin pathway may regulate NFATc1 induction. However, the factors that directly regulate NFAT expression levels are unknown.
To better understand the regulation of NFATc1 expression, the murine Nfatc1 gene and upstream regulatory regions were cloned. Herein, we provide the genomic organization of the murine Nfatc1 gene and initial characterization of a 6-kb genomic fragment located upstream of the transcription start site (TSS). A CsA-sensitive and inducible minimal promoter region containing a consensus NFAT motif was identified within this 6-kb element. Mutagenesis studies indicated that this NFAT site is critical for optimal inducible promoter activ-ity in primary human T cells. Moreover, this NFAT site bound the NFATc2 protein present in T cell nuclear extracts, and increased NFATc2 expression in primary T cells augmented Nfatc1 minimal promoter activity. In contrast, the inducible Nfatc1 promoter activity was diminished in T cells deficient in NFATc2. These data suggest a role for NFATc2 in the regulation of the Nfatc1 gene during the immune response.

MATERIALS AND METHODS
Isolation and Characterization of Murine Nfatc1 Genomic DNA Clones-A mouse Bacterial Artificial Chromosome (BAC) library (mouse 129/SvJ, Research Genetics, Huntsville, AL) was screened by PCR and Southern blotting for Nfatc1 gene-specific clones according to the manufacturer's recommendation. A single clone was determined to contain the entire coding region of the murine Nfatc1 gene by Southern blot analysis. To determine the genomic organization of murine Nfatc1, the intron/exon borders and gene flanking regions were sequenced on the purified BAC clone. A range of sequencing oligonucleotide primers flanking the putative intron/exon boundaries were initially chosen by the cDNA sequences of murine Nfatc1 isoforms ␣ and ␤ (GenBank TM / EBI accession numbers AF087434 and AF049606, respectively) and subsequently by derived sequences.
Rapid Amplification of cDNA Ends (RACE)-The 5Ј-RACE method was carried out using the Sure-RACE kit (Origene, Rockville, MD), which contains full-length cDNAs from multiple mouse tissues, following the manufacturer's instructions. Briefly, the 5Ј-end of murine Nfatc1 cDNA was amplified by PCR through two rounds of 30 cycles each at 94°C for 20 s, followed by 68°C for 60 s. The primers used in the PCR included the 5Ј-adapter primers provided by the manufacturer as the sense primers and the gene-specific primers GSP1 (first round, positions ϩ55 to ϩ74 relative to the translation start site) and GSP2 (second round, positions ϩ30 to ϩ53) as the antisense primers (see Fig.  1). The resulting amplification products were subcloned into pCRII (Invitrogen, Carlsbad, CA) and sequenced.
5Ј-Primer Extension Analysis-To confirm the TSS defined by 5Ј-RACE, 5Ј-primer extension analysis was carried out using the avian myeloblastosis virus reverse transcriptase primer extension system (Promega, Madison, WI). The primer (positions ϩ14 to ϩ36 bp relative to the putative 5Ј-end of murine NFATc1 mRNA; see Fig. 1) was end-labeled with [␥-32 P]ATP and incubated with 50 g of total RNA from the heart, lung, liver, and spleen of a 3-month-old BALB/c mouse. The extension reaction was carried out according to the manufacturer's protocol, and the extended products were resolved on a 7 M urea and 6% acrylamide gel in TBE along with 10-bp DNA size markers.
Plasmids and Reporter Constructions-The mammalian expression vectors for NFATc1 and NFATc2 have been described elsewhere (5,24). The firefly luciferase reporter vector (pGL2-Basic) and the Renilla luciferase control vector used for transfection efficiency (pRL-null) were purchased from Promega. The parent murine Nfatc1 reporter (p6.2) was constructed by insertion of a 6233-bp NheI-XhoI DNA fragment of murine Nfatc1 (positions Ϫ5861 to ϩ372 relative to the TSS) into the multiple cloning site of the pGL2-Basic vector. A series of deletion constructs (pX, where X represents the insert size in kilobases) were generated by consecutive deletions of the 5Ј-end of the 6233-bp NheI-XhoI insert in p6.2 by restriction enzyme digestions and religations of the vector (see Fig. 1). For example, the p0.7 plasmid was generated by insertion of a 0.7-kb KpnI-XhoI DNA fragment (positions Ϫ351 to ϩ372) upstream of the luciferase gene into pGL2-Basic. The 5Ј-end sequences of all deletion constructs were confirmed by sequencing. The luciferase reporter genes pIL-2-Luc, pIL-4-Luc, pCMV-Luc, and pNFAT-Luc, which are driven by the human IL-2 promoter, the human IL-4 promoter, the immediate-early cytomegalovirus promoter, and a multimerized NFAT-binding site, respectively, have been previously described (24,27).
PCR Mutagenesis-The p0.7m plasmid with a mutated NFAT-binding site was generated by PCR-based mutagenesis as previously described (28). Briefly, two overlapping PCR products carrying the mutation were generated using p0.7 as a template and two sets of primers. The first set of primers was the 5Ј-forward primer (TGTATCTTATGG-TACTGTAACTG) and the 3Ј-mutation reverse primer (AGCTGaagAA-CACCTCCCCCGGCTCCCG). The mutated NFAT site is underlined, with its mutated core residues indicated by lowercase letters. The wild-type sequence of the NFAT-binding site on the noncoding strand is TGGAAAA. The second set of primers was the 5Ј-mutation forward primer (GTGTTcttCAGCTTTAAAAAGGCAGAAG) and 3Ј-reverse primer (CTTTATGTTTTTGGCGTCTTCC). After the first round of PCR, the two overlapping PCR products were used as the template for the second round of PCR with the 5Ј-forward and 3Ј-reverse primers. The final PCR product was then digested with KpnI-XhoI before insertion into the multiple cloning sites of pGL2-Basic. A clone with the desired mutation and no other sequence alterations was confirmed by sequence analysis.
Isolation of Primary T Cells-Wild-type BALB/c mice were purchased from the Jackson Laboratory (Bar Harbor, ME). NFATc2 Ϫ/Ϫ mice on the BALB/c background were kindly provided by Dr. Laurie H. Glimcher (Harvard University, Cambridge, MA) (29). The isolation of human whole mononuclear cells from peripheral blood of healthy adult donors (30) and of whole mononuclear cells from spleens of 3-4-monthold mice (31) has been described previously. Human CD4 ϩ T cells were purified by negative selection using T-helper Lympho-Kwik antibody plus complement reagents (One Lambda, Inc., Canoga Park, CA) according to the manufacturer's protocol. For some experiments, CD4 ϩ T cells were expanded in vitro as described (24) Reporter Gene Assays-The protocols used for transient transfection of nontransformed human and murine CD4 ϩ T cells have been described previously (31). In brief, after submitogenic stimulation overnight with concanavalin A (murine) or phytohemagglutinin (human), 5 ϫ 10 6 cells were electroporated in 250 l of medium at 250 V with 950 microfarads of resistance in 0.4-cm gap cuvettes (Bio-Rad). Firefly luciferase reporter constructs (5-10 g) were co-transfected with or without 10 g of expression plasmids. All transfections contained 1 g of the pRL-null reporter gene (Renilla luciferase) as a control for transfection efficiency. Following transient transfection with plasmid DNA, the T cells were activated with 25 ng/ml PMA (Sigma) and 1.5 M ionomycin (Calbiochem) for 6 h. For experiments involving CsA, 100 ng/ml CsA (Novartis, East Hanover, NJ) was added to T cell cultures 30 min prior to polyclonal activation. After 6 h in culture at 37°C, cells were lysed and analyzed for luciferase activity as previously described (32) using the Dual-Luciferase Reporter 1000 assay system (Promega). All samples were run in duplicate, and firefly luciferase values were corrected for transfection efficiency based on Renilla luciferase readings.
Preparation of Nuclear Proteins and Electrophoretic Mobility Shift Assays (EMSAs)-In vitro primed human CD4 ϩ T cells were either left unstimulated or were treated with 25 ng/ml PMA and 1.5 M ionomycin for 2 h before nuclei were harvested. Nuclear extracts were prepared using a standard protocol. Briefly, cells were washed twice with ice-cold phosphate-buffered saline and once with cold hypotonic buffer (10 mM Hepes, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM dithiothreitol, and 0.2 mM phenylmethylsulfonyl fluoride). The cell pellets were resuspended in hypotonic buffer to three times the volume of the cell pellets, incubated for 10 min on ice, and homogenized with 10 strokes in a glass Dounce homogenizer using a type B pestle. Nuclei were pelleted at 3300 ϫ g for 15 min and resuspended in ice-cold low-salt buffer (20 mM Hepes, pH 7.9, 25% glycerol, 200 mM KCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, 0.5 mM dithiothreitol, and 0.2 mM phenylmethylsulfonyl fluoride). Nuclear proteins were then extracted by adding an equal volume of ice-cold high-salt buffer (identical to low-salt buffer except containing 1.2 M KCl) and incubated on ice for 30 min. Debris was removed by centrifugation at 14,000 rpm for 1 h at 4°C. The supernatants were dialyzed against 50 volumes of binding buffer (10 mM Tris-HCl, pH 7.5, 100 mM KCl, 10% glycerol, 50 ng/ml poly(dI-dC), 1.5 mM MgCl 2 , 0.2 mM EDTA, 1 mM dithiothreitol, and 0.2 mM phenylmethylsulfonyl fluoride) overnight at 4°C, aliquoted, and frozen at Ϫ80°C. The protein concentration was estimated using the bicinchoninic acid protein assay kit (Pierce). Single-stranded oligonucleotides were 5Ј-end-labeled with [␥-32 P]ATP using T4 polynucleotide kinase, annealed, and purified. The sequences of the oligonucleotide probes used are listed in Fig. 4. For each binding reaction, radiolabeled probe (1 ϫ 10 4 cpm, 0.1 ng) was incubated with 1 g of nuclear extract in 10 l of binding buffer for 10 min at room temperature. Samples were then analyzed by electrophoresis on pre-cast 6% polyacrylamide gels in 0.5ϫ TBE (Novex, San Diego, CA). When indicated, the nuclear extracts were preincubated with anti-NFAT antibodies on ice for 20 min before addition of the radiolabeled probes. The anti-NFATc2 antibody (MA1-025) was obtained from Affinity Bioreagents (Golden, CO), and the anti-NFATc1 (7A6) and anti-HMG (I)Y (FL-95) antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA).
Northern and Western Blot Analyses-Northern and Western blot analyses were carried out using standard protocols. 10 g of total RNAs or cell lysates from each treatment was used for Northern and Western blot analyses, respectively. The probe for Northern blot analysis was generated by random priming full-length cDNA of Nfatc1 in the presence of [␣-32 P]dCTP. The antibody for Western blot analysis of NFATc1 was 7A6, a mouse monoclonal antibody, which was used at 1:1000 dilution, followed by horseradish peroxidase-conjugated rabbit antimouse IgG used at 1:5000 dilution.

Determination of Murine Nfatc1 Genomic Structure and
Transcription Start Site-To study the transcriptional regulation of NFATc1, we isolated the murine Nfatc1 gene from a mouse BAC library. The murine Nfatc1 gene contains nine exons (Fig. 1A). The first two exons encode the unique N termini of isoforms ␣ and ␤, respectively, and translation of isoform ␣ or ␤ starts from two alternative ATG start sites (33,34). It has been postulated that the human NFATc1 gene contains 11 exons (10), and the inducible isoform A is transcribed from the first exon and stops at the proximal poly(A) site located at exon 9, whereas the two constitutively expressed isoforms B and C are transcribed from the second exon and end at the distal poly(A) site located at exon 11 (35,36). However, we were unable to locate the mouse orthologs of human exons 10 and 11 by PCR using various degenerate oligonucleotide primers based on the human NFATc1 cDNA sequence. Moreover, by Northern analysis, only two murine NFATc1 messages of similar sizes (ϳ4.5 kb) have been detected in murine splenocytes by us (see Fig. 5) and in murine T cells by others (33,34).
We next determined the TSS for exon 1 by RACE using a commercial mouse panel that contains full-length cDNA from multiple tissues. A single PCR product was observed in PCR amplification using cDNAs from the heart, liver, and lung as templates (data not shown). Sequencing of two RACE products from two separate PCR amplifications using liver cDNA at two different concentrations (1ϫ and 5ϫ) indicated that they both started at the same nucleotide, a C (indicated by the black dot in Fig. 1B), located 160 bases upstream of the translation start site (uppercase). Primer extension analysis with total RNA extracted from 3-month-mouse heart, lung, liver, and spleen ( Fig. 1C) was then performed to confirm the 5Ј-RACE result. A major band of the same size was observed in all reactions, which stopped at the C nucleotide (indicated by the asterisk) that is 10 bases upstream of the tentative TSS defined by 5Ј-RACE. Thus, the GC-rich motif located 170 to 160 bp upstream of the translation start site (underlined in Fig. 1B) appears to be a major transcription initiation site for the murine Nfatc1 gene. The inconsistent results between the 5Ј-RACE and primer extension analysis and the observation that no TATA-box like motif is present near the TSS suggest that this 10-bp GC motif may represent a weak binding site for a transcription initiation complex. For convenience, we refer to the C nucleotide at the 5Ј-end of the GC island as the TSS (position ϩ1).
Identification of an Inducible Nfatc1 Minimal Promoter-To determine the sequences essential for efficient transcription of murine Nfatc1 in T cells, a DNA segment extending from Ϫ5861 to ϩ372 was inserted 5Ј to a luciferase gene in a reporter construct. The construct was designated p6.2 according to its insert size in kb. The 6.2-kb inserted DNA contained the putative promoter region, 199 bp of exon 1, and 173 bp of the 5Ј-end of the first intron. The promoter activity of p6.2 was initially tested by transient transfection assays in primary cultures of mouse splenocytes and thymocytes as well as in human peripheral whole mononuclear cells in the presence or absence of PMA plus ionomycin for 6 h. In all three cells, the 6.2-kb fragment yielded a weak promoter activity (2-fold that of the control promoterless vector) in the absence of stimulation and a 2-fold induction in promoter activity after stimulation with PMA and ionomycin (data not shown). No changes were observed when the cells were transfected with the control vector after stimulation. These results suggest that the 6.2-kb DNA fragment acts as a weak promoter in T cells. Alternatively, a repressor may exist within the 6.2-kb segment.
To further dissect this regulatory region, a series of 5Ј-end deletions of p6.2 were generated ( Fig. 2A). The resultant constructs are referred to as pX, where X represents the insert size in kb. Purified human CD4 ϩ T cells were analyzed for the remainder of the transcriptional assays. The deletion analysis revealed that a region from Ϫ351 to ϩ372 (p0.7) was sufficient to confer optimal basal as well as inducible promoter activity in primary CD4 ϩ T cells. Similar results were obtained using a reporter gene construct containing a region from Ϫ351 to ϩ14 that lacks the ATG translation site (data not shown). Further 5Ј-truncation (p0.4) resulted in a loss of both activities. Thus, this DNA fragment (Ϫ351 to Ϫ61) in p0.7 was defined as the minimal murine Nfatc1 promoter. Further comparison between p6.2 and p0.7 in three separate experiments in primary CD4 ϩ cells showed that p0.7 had an average of 4-fold inducible promoter activity compared with p6.2 (Fig. 2B), indicating that p0.7 indeed serves as a minimal promoter. In addition, a drop in transcriptional activity in T cells transfected with p6.2 compared with p3.6 suggests that a repressor element may exist between Ϫ5861 and Ϫ3268. Similarly, a potential repressor element may be located between Ϫ1723 and Ϫ352.
Alignment of mouse and human sequences revealed that Ϫ297 to ϩ1 of the minimal promoter region is highly conserved (85% identity at the nucleotide level) between the two species (Fig. 2C), suggesting that this region serves as a key regulatory element conserved in mouse and human. Moreover, several potential binding sites for various T cell transcription factors are located within the region as determined by screening using the TRANSFAC transcription factor-binding site data base (37). Notably, a consensus NFAT-binding site (TGGAAAA on the noncoding strand) is located at Ϫ177 to Ϫ171 bp upstream of the TSS. Flanking the NFAT-binding site immediately 5Ј is a G-rich combinatory binding site for Sp1/Egr transcription factors (38,39). Just 3Ј of the NFAT site is a short AT-rich region reminiscent of an HMG I(Y) protein-binding site (40).
The NFAT-binding Site Is Critical for Optimal Transcriptional Activity of the Murine Nfatc1 Minimal Promoter-It has previously been shown that NFATc1 is notably up-regulated at ϳ6 h after T cell activation (22,33,36). To study this, the p0.7 construct was transiently transfected into primary human CD4 ϩ T cells. Cells were then stimulated with PMA plus ionomycin and analyzed for minimal promoter activity at sequential time points. A steady linear increase in luciferase activity was observed between 3 and 9 h of treatment (Fig. 3A). This time course of promoter activation by PMA/ionomycin mimics the expression of endogenous NFATc1 protein induced by PMA/ ionomycin (22) or 12-O-tetradecanoylphorbol-13-acetate/ionomycin (36) in normal human peripheral T cells. Moreover, similar to the induction of NFATc1 protein synthesis in activated T cells (22), this inducible promoter activity was also completely CsA-sensitive, as documented by the loss of induction of reporter activity in the presence of CsA (Fig. 3B). This suggested that the potential NFAT-binding site in the minimal promoter might be involved in regulating NFATc1 transcription. To test this directly, a NFATc2 or NFATc1 expression vector was co-transfected along with the p0.7 reporter construct. Exogenous expression of NFATc2 (and to some degree NFATc1 as well) further augmented the induction of Nfatc1 promoter activity by PMA plus ionomycin stimulation in T cells (Fig. 3C). Thus, the NFAT protein was able to up-regulate the inducible murine Nfatc1 minimal promoter, and the NFATbinding site within the minimal promoter was likely involved in this activity. It is unclear why there appears to be differential activation by NFATc2 and NFATc1, but this may relate to differences in interactions with neighboring transcription factors, differential modification of the transactivation domains, or differences in intrinsic binding affinities for unique AG-rich regions.
To confirm the role of the NFAT site, a mutation was introduced into the NFAT-binding site in p0.7 by a PCR-based strategy to generate p0.7m. The core (boldface) of the NFAT site in p0.7 (TGGAAAA) was altered in p0.7m (TGAAGAA, with changes underlined) to maintain the same A/G ratio, but to disrupt the critical GGA core sequence required for DNA binding (41). As predicted, mutation of the NFAT site had a deleterious effect on the inducible activity of the murine Nfatc1   FIG. 2. Defining the murine Nfatc1 proximal promoter. A, constructs consisting of various lengths of serial deletions of the murine Nfatc1 5Ј-flanking region (lines) were inserted upstream of the firefly luciferase gene (Luc; boxes) in the reporter plasmid pGL2-Basic. Numbers above the constructs indicate nucleotide positions relative to the TSS (position ϩ1). The firefly luciferase activity of each construct is presented relative to that of the control plasmid (p0.4) after normalization to Renilla luciferase activity. White bars represent unstimulated freshly isolated human peripheral blood whole mononuclear cells, and black bars indicate treatment of the cells with PMA (at a final concentration of 25 ng/ml) and ionomycin (1.5 M) (PI) for 6 h. B, shown is the Nfatc1 promoter activity in primary CD4 ϩ T cells. Human peripheral blood CD4 ϩ T cells were isolated and transiently transfected with Nfatc1 promoter reporter constructs. -Fold induction of the p6.2 and p0.7 plasmids relative to the control plasmid (p0.4) is displayed along the y axis. The data are presented as means Ϯ S.D. for three independent experiments. C, shown is a sequence comparison between the murine Nfatc1 (GenBank™ accession no. AF480838) and human NFATc1 proximal promoters. Putative DNA-binding sites are underlined. Notably, a consensus NFAT-binding site is located between combinatory sites for Sp1/Egr and HMG. ATF, activating transcription factor; CREB, cAMP response element-binding protein.
minimal promoter (Fig. 3D). Little if any induction of luciferase activity by PMA/ionomycin was observed in cells transfected with p0.7m. Surprisingly, despite the observation that mutation of the NFAT site also reduced the basal promoter activity to 40% of the wild-type promoter activity (Fig. 3D), the basal promoter activity was not CsA-sensitive (data not shown). Taken together, we conclude that the NFAT-binding site within the murine Nfatc1 minimal promoter is required for maintaining its basal and inducible activities, but that a transcription factor(s) other than NFAT also participates in the basal activity.
NFATc2 Binds to the NFAT Site within the Murine Nfatc1 Minimal Promoter-To determine whether NFAT proteins interact with the NFAT-binding site within murine Nfatc1 min-imal promoter, we performed the EMSA with three different sets of probes containing various combinations of putative transcription factor-binding sites flanking the NFAT site (Fig. 4A). Previously, NFAT proteins have been shown to bind coordinately with AP-1 (4) or other transcription factors or coactivators, including c-Maf (42), NIP-45 (43), CBP/p300 (44), GATA-4 (45), Sp1 (46), Ets (47), nuclear factor-B (48), octamer (49), and HMG I(Y) (50), to the promoter regions of numerous cytokine genes (8 -10). Therefore, to evaluate NFAT binding in the context of neighboring transcription factors, we generated the oligonucleotide probe (SNH) identical to the sequence found in the murine Nfatc1 minimal promoter, consisting of a 5Ј-combinatory site for Sp1/Egr, the central NFAT-binding site, and a 3Ј-site for the HMG (I) protein. Two shorter probes were also made by removing the Sp1/Egr or HMG site and termed NH and SN, respectively.
Nuclear protein extracts from primary human CD4 ϩ T cells (Fig. 4B) were used for EMSAs. An inducible binding complex "i" was seen after PMA/ionomycin stimulation. This inducible complex was supershifted and/or disrupted by antibodies against either NFATc2 or NFATc1. Moreover, formation of this inducible binding complex was absent when the NFAT site was mutated (data not shown). Thus, this NFAT site harbors both NFATc2 and NFATc1 in activated T cells. Because the inducible NFAT complex was detected in CD4 ϩ T cells with all three probes, neither the Sp1 nor HMG site appears to be independently critical for NFAT binding in CD4 ϩ cells. Luciferase (Luc) activity is presented relative to the activity of the control plasmid (p0.4) under identical conditions and normalized to Renilla luciferase activity. C, the p0.7 reporter was cotransfected with expression vectors pREP4 (as a control), pREP4/NFATc2, and pREP4/ NFATc1 following PMA plus ionomycin treatment for 6 h. Luciferase activity is presented relative to the activity of p0.7 cotransfected with the empty parent plasmid (pREP4). The data are presented as means Ϯ S.E. for three independent experiments. D, mutation of the core NFATbinding site in p0.7 (p0.7m) reduced the basal activity of the Nfatc1 minimal promoter and abrogated its inducible activity. Luciferase activity is presented relative to the activity of p0.7 without any treatment (control (c)). Data are representative of two independent experiments. Several other binding complexes were also observed in EM-SAs, and their identities remain to be determined. Of these, the fast migrating complex "b" resembles the HMG I(Y) protein in that it was reduced by PMA plus ionomycin stimulation and was dependent on the presence of the 3Ј-HMG site for binding (50). Despite this, the EMSA carried out with an antibody specific to HMG I(Y) was unable to identify complex b (data not shown). The slow migrating complex "a" was constitutively present in nuclear extracts of primary human CD4 ϩ T cells, but it was not inducible upon activation when analyzed with the complete SNH probe. By supershift analysis with NFAT-specific antibodies, it appeared that complex a contained NFATc2 and NFATc1 because anti-NFAT antibodies partially blocked its formation. Because the consensus NFAT sequence is the binding site for both NFATc1 and NFATc2, and neither NFATc1 nor NFATc2 is known to form a dimer or heterodimer, it is likely that NFATc1 and NFATc2 in complex a bound independently to DNA with other partners. Further experiments will be required to determine the complete nature of this interaction.
Attenuation of Inducibility of the Murine Nfatc1 Minimal Promoter in NFATc2 Ϫ/Ϫ T Cells-To further determine whether NFATc2 plays an essential role in the induction of Nfatc1 promoter activity, the p0.7 construct was transfected into NFATc2 Ϫ/Ϫ splenocytes following overnight stimulation with concanavalin A. This pre-stimulus allows for T cellspecific transfection of plasmid DNA (31). The NFATresponsive IL-2 and IL-4 minimal promoter reporter con-structs (pIL-2-Luc and pIL-4-Luc, respectively) and an NFATdependent reporter in which multiple NFAT sites have been inserted upstream of luciferase (pNFAT-Luc) were studied in parallel. Compared with normal splenic T cells, a decrease in basal promoter activity was observed in NFATc2 Ϫ/Ϫ T cells transfected with the pIL-2-Luc, pIL-4-Luc, and pNFAT-Luc reporter constructs, whereas p0.7 reporter activity was modestly increased (Fig. 5A, white bars). Similarly, both Northern (Fig. 5C) and Western (Fig. 5D) blot analyses showed that the basal level of NFATc1 in NFATc2 Ϫ/Ϫ T cells (lanes 3) was increased compared with that in normal T cells (lanes 1). The three protein bands between molecular masses of 79 and 122 kDa on the Western blot (Fig. 5D) represent the NFATc1 protein recognized by anti-NFATc1 antibody 7A6. As described previously (22), the predominant band (ϳ90 kDa, indicated by the arrow) is the inducible form of NFATc1, whereas the other two bands represent the uninducible form and/or phosphorylated NFATc1. Upon ionomycin and PMA stimulation, a reduction in activities of all four reporters, including p0.7, was found in NFATc2-deficient splenic T cells (Fig. 5A, black bars). The inducibility of all four reporter genes was attenuated in NFATc2 Ϫ/Ϫ splenic T cells (Fig. 5B, hatched bars) compared with that in normal cells (shaded bars). By contrast, a comparable increase in the induction of NFATc1 expression was found in both normal (Fig. 5, C and  D, lanes 2) and NFATc2 Ϫ/Ϫ (lanes 4) splenocytes after 6 h of polyclonal stimulation in vitro. These data argue that although NFATc2 is required for optimal induction of Nfatc1 FIG. 5. Inducible Nfatc1 minimal promoter activity is attenuated in NFATc2 ؊/؊ T cells. Splenic T cells from wild-type or NFATc2 Ϫ/Ϫ mice were transiently transfected with the pIL-2-Luc, pIL-4-Luc, p0.7, and pNFAT-Luc reporters (representative of the IL-2, IL-4, and murine Nfatc1 promoters and NFAT-dependent transcriptional responses, respectively). A, reporter activity in NFATc2 Ϫ/Ϫ cells compared with that in wild-type cells without (white bars) or with (black bars) PMA plus ionomycin (PI) stimulation for 6 h. B, -fold inducible promoter activity in wild-type (shaded bars) and NFATc2 Ϫ/Ϫ (hatched bars) cells after ionomycin and PMA stimulation. Luciferase (Luc) activity was normalized between plasmids using Renilla luciferase activity and to the activity in the absence of stimulation between mouse cell types using the activity of the pCMV-Luc reporter construct as a measure of transfection efficiency as previously described (24). The data are representative of two independent experiments. C and D, Northern and Western blot analyses, respectively, of NFATc1 expression in normal (lanes 1 and 2) and NFATc2 Ϫ/Ϫ (lanes 3 and 4) splenocytes that were unstimulated (lanes 1 and 3) or were activated by ionomycin plus PMA (lanes 2 and 4). The arrows indicate the inducible expression of NFATc1 mRNA (C) or protein (D). The ethidium bromide staining of 28 S and 18 S ribosomal RNAs was used as a loading control for the Northern blot analysis. The Coomassie Blue staining of protein after transfer was the loading control for Western blot analysis. The blots represent two independent experiments. minimal promoter activity in T cells, in vivo it is dispensable for the overall transcriptional activity of NFATc1. DISCUSSION The NFAT family members are critically important for immune regulation by effector CD4 ϩ T cells, as most if not all cytokines produced by these cells are influenced by NFAT activity (8 -10). The importance of the individual NFAT family members has been demonstrated by the immune phenotypes present in NFAT-deficient mice (29,(51)(52)(53)(54)(55)(56)(57)(58)(59)(60). NFATc1 Ϫ/Ϫ T cells display an increased propensity for the generation of Th1 cytokines, arguing that NFATc1 may normally play a role in up-regulating the opposing Th2 cytokine response (55,56). By contrast, NFATc2 Ϫ/Ϫ mice overexpress Th2 cytokines upon repeated antigenic stimulation and develop allergic/asthmatic phenotypes (29,(51)(52)(53)(54). The regulation of cytokine immune responses and the subsequent immune responses are thus notably influenced by the interplay of individual NFAT family members. Therefore, understanding the regulation of this family of T cell transcription factors is of utmost importance.
NFATc1 expression in peripheral T cells is predominantly regulated by T cell activation through T cell receptor engagement, which can be mimicked in vitro via stimulation with ionomycin and PMA. Following T cell activation, the level of NFATc1 mRNA is detectable within 1 h after activation (6) and peaks around 3-6 h (33,36), resembling the induction of many T cell-derived cytokine genes. Furthermore, CsA, a powerful cytokine inhibitor, eliminates inducible Nfatc1 gene expression (22). The similarity in the regulation of gene expression between Nfatc1 and cytokine genes supports the contention that NFAT contributes to Nfatc1 expression.
Here, we provide evidence that the transcriptional activity of Nfatc1 in T cells is indeed regulated by NFATc2, the most abundantly expressed NFAT family member in human peripheral T cells (22,24,31). First, the minimal promoter of murine Nfatc1 was defined in primary T cells that were activated by ionomycin plus PMA stimulation. This induction was completely blocked by CsA. Second, a proximal consensus NFATbinding motif located within the murine Nfatc1 minimal promoter was identified, and the sequence of the NFAT site and surrounding sequences were highly conserved between mouse and human. Mutational analysis of the NFAT-binding site demonstrated that this DNA element contributed to optimal basal and inducible activities of the minimal promoter. Third, the NFAT site was required for binding of NFATc2 upon T cell activation, and this binding correlated with increased promoter activity. Last, the inducible Nfatc1 promoter activity was increased by overexpression of NFAT, particularly NFATc2. Conversely, this induction was notably diminished in NFATc2 Ϫ/Ϫ T cells. Together, these studies demonstrate the importance of NFATc2 in increasing NFATc1 expression.
The requirement of the NFAT-binding site for NFATc2 interaction and the minimal promoter activity of Nfatc1 does not, however, establish that NFATc2 is essential for constitutive or even inducible Nfatc1 expression in vivo. In fact, no alteration in the level of endogenous NFATc1 protein in DNA⅐NFAT complexes was shown by EMSA using nuclear protein extracts from NFATc2 Ϫ/Ϫ mouse splenocytes (54). Furthermore, constitutive expression of NFATc1 has been reported in mice doubly deficient in NFATc2 and NFATc3 (58). Similarly, we found, by Northern and Western blot analyses, an increase in the basal level of NFATc1 in NFATc2 Ϫ/Ϫ splenocytes (Fig. 5, C and D). In addition, the induction of NFATc1 expression in NFATc2 Ϫ/Ϫ splenocytes after 6 h of polyclonal stimulation was comparable to that in normal splenocytes. Thus, dysregulation of NFATc1 coexists with alterations in the expression of multiple cytokine genes present in Nfatc2 gene-disrupted mice.
One possible explanation for the discrepancy of reporter gene assays and endogenous NFATc1 expression is that although the optimal Nfatc1 minimal promoter activity is dependent on NFATc2, its role is compensated by other transcription factors in the context of the entire gene in the NFATc2 Ϫ/Ϫ splenocytes. Along these lines, the activity of a potent transcription factor (AP-1) was markedly increased in NFATc2 Ϫ/Ϫ splenocytes, especially upon PMA and ionomycin activation (Ͼ5-fold) (data not shown). Alternatively, the dysregulated Nfatc1 expression in NFATc2 Ϫ/Ϫ splenocytes may be secondary to functional redundancy or perturbed thymic development in the knockout mice. Moreover, T cell function is not entirely normal in NFATc2 Ϫ/Ϫ mice, as the T cells express a preactivated phenotype and are less susceptible to activation-induced apoptosis (54). It is therefore possible that NFATc2 does contribute to the expression of endogenous NFATc1 in T cells. To more directly address this, it may be necessary to examine the role of NFATc2 in Nfatc1 transcription at the chromatin level (e.g. chromatin immunoprecipitation assays) or by a dominant-negative strategy similar to what has been used to document the requirement for NFAT in IL-2 expression in vivo (61).
In addition to NFAT, other transcription factors must contribute to Nfatc1 gene regulation. The gel shift assays suggested that other protein(s) bind adjacent to the NFAT site in the murine Nfatc1 promoter prior to T cell activation because deletion either 5Ј or 3Ј of the NFAT site from the composite elements abolished the fast migrating complex b (Fig. 4B). Although the formation of binding complex b was characteristic of HMG (I)Y described for the NFAT-containing composite of the murine IL-4 promoter (50), an anti-HMG (I)Y antibody was unable to confirm the protein identity within this complex b. Further experiments will be directed at examining the possible involvement of members of the HMG family other than HMG (I)Y.
The complexity of transcription complexes forming at the overlapping motifs was further manifested by the slow migrating complex a in Fig. 4B. The site immediately 5Ј of the NFAT motif potentially serves as an overlapping binding site for the constitutively expressed transcription factor Sp1 and inducible Egr-1, which has also been found in the distal NFAT/AP-1binding site of the human IL-2 gene (38,39). This cooperation of NFAT proteins with AP-1 factors at the NFAT/AP-1 element has been the prototype of synergistic induction of cytokine gene transcription (9). Sp1 or Egr factors have also been shown to mediate IL-2, CD95 ligand, and CD40 ligand transcription by functioning as potent coactivators for NFAT proteins (39,61,62). Thus, the close proximity of the Sp1, NFAT, and HMG sites in the Nfatc1 proximal promoter and the complete conservation of sequences here between mouse and human suggest an intricate balance in the control of NFATc1 transcription by several transcription factors. Therefore, identification of NFAT cofactors and further characterization of their interactions at the NFAT site will elucidate the higher order assembly of transcription factors operating in a temporal, cell type-specific, and stimulus-dependent manner required for NFATc1 transcription during the immune response.
The transcriptional regulation of transcription factors themselves is only just beginning to be explored. Recently, the Oct-1 transcription factor promoter was found to possess multiple octamer-binding sites (63). It is similarly interesting that NFAT proteins are able to positively regulate self-expression. Transcriptional regulation of transcription factors will most certainly play a role during embryogenesis and organ development. In the thymus, for example, expression of NFAT family members is differentially regulated during the development of T cells (26,64). NFATc3 is preferentially expressed at a high level in immature double-positive CD4 ϩ /CD8 thymocytes. Conversely, NFATc2 is dominantly expressed in mature singlepositive CD4 ϩ T cells. NFATc1, however, is present at low levels in all subsets of T cells, but is strongly induced upon treatment with phorbol ester and calcium ionophore (26). Finally, NFATc1 is both temporary and spatially restricted to a subset of cardiac endothelial cells and is required for cardiac valve formation (65,66). Thus, the Nfatc1 minimal promoter characterized here will be useful to delineate the transcriptional regulatory mechanism of NFATc1 in the heart that is important in understanding the genetic pathway controlling normal cardiac valve development and the subsequent disease process.