The NFAT family transcription factors play crucial roles in immunological and other biological events; however, the functional differences among NFAT members have not been fully elucidated. This study investigated the relative contribution of NFATc2 and NFATc1 to the transactivation of cytokine genes in T cells. Ectopic expression of NFATc2 but not NFATc1, especially its short isoform, enhanced TNF-α synthesis in human T cells at the gene transcription level, whereas both NFATs augmented IL-2 expression. In addition, a reduction of the shortest NFATc1 isoform using RNA interference technology failed to suppress TNF-α expression. The promoter/enhancer activity of the NFAT-binding site in the TNF-α gene was up-regulated by NFATc2 but not by NFATc1, whereas both NFATs associated similarly with this region. A study of mRNA expression using NFATc2/NFATc1 chimeric molecules revealed that the enhancing activity of NFAT on the TNF-α gene was lost by truncation of its C-terminal transactivation domain. In addition, this domain derived from NFATc2 behaved as a dominant negative against the NFAT site in TNF-α promoter-dependent transcriptional activity in T cells. We conclude that the C-terminal transactivation domain in NFAT is crucial for TNF-α gene expression in human T cells.
The NFAT family of transcription factors consists of five members, NFATc1 to NFATc5, which are involved in the inducible expression of numerous genes concerned with immune responses as well as other biological events (1, 2, 3). NFATc1–c4 are dephosphorylated by a Ca2+-dependent serine/threonine phosphatase, calcineurin, and translocate into the nucleus where they associate with target DNA sequences. The immunosuppressive drugs FK506 and cyclosporin A depress the function of these NFATs to the same degree through the inhibition of calcineurin activity (4, 5, 6).
Studies using gene-targeted mice have suggested that each NFAT family member has a differential role in the synthesis of multiple cytokines. For example, IL-4 production by T cells is impaired in NFATc1 (NFATc, NFAT2)-deficient mice (7, 8), although IL-2 synthesis is relatively enhanced (7). Nevertheless, the representative phenotypes observed in NFATc2 (NFATp, NFAT1)-deficient mice are controversial. In the first report regarding NFATc2−/− mice by Hodge et al. (9), a striking defect in the early production of IL-4, IL-13, GM-CSF, and TNF-α by T cells in anti-CD3 Ab-treated mice was observed in vivo, whereas IL-2 and IFN-γ were minimally affected. However, Th2 development in these mice was enhanced at later time points along with increased IL-4 production both in vitro and in vivo. Up-regulation and contrary down-regulation of Th2 cytokines in NFATc2-deficient mice were also reported later by different groups (10, 11, 12, 13). These findings suggest at least a functional difference between NFATc2 and NFATc1 in the regulation of T cell cytokines.
However, the molecular mechanisms by which each NFAT family member plays a distinct role in cytokine synthesis still remain unclear. NFAT is composed of several functional domains (1). The DNA binding domain (DBD),3 which lies between amino acid residues ∼400 and ∼700, is highly conserved within the NFAT family and shows similarity to the DBD of the Rel family proteins (14). The Ca2+ regulatory domain (CRD) of ∼300 amino acids, which binds and is dephosphorylated by calcineurin, is located just N-terminal to the DBD. The CRD shows a lesser degree of pairwise sequence identity but strong conservation of several sequence motifs characteristic of the NFAT family (1). Both the N- and C-terminal ends of NFAT proteins contain a transactivation domain (TAD1 and TAD2, respectively) despite very limited sequence conservation (1). Amino acid homology between NFATc2 and NFATc1 is 23, 33, and 70% for TAD1, CRD, and DBD, respectively.
NFATc1 has multiple isoforms. The original form of NFATc1 was identified from a cDNA library of human peripheral blood lymphocytes and Jurkat cells (14). Thereafter, Park et al. (15) isolated a new NFATc1 isoform, NFATc.β, from a cDNA clone of the Raji B cell line. NFATc.β differs from NFATc.α (identical with the original NFATc1 that we mainly used in this study) in the first 29 N-terminal amino acid residues and contains an additional region of 142 residues at the C terminus. In addition, Lyakh et al. (16) identified two short forms of NFATc1 (82 and 86 kDa), which were strongly induced by stimulation, as well as two longer and relatively consistent isoforms (110 and 140 kDa). These two short isoforms resulted from the initiation of translation at two different AUG codons (16) and were predominantly induced upon cellular activation (16, 17). In addition, Chuvpilo et al. (17) reported that the longer NFATc1 isoforms were derived by alternative polyadenylation events. However, the physiological meaning of this heterogeneity of NFATc1 is unclear.
All NFAT members expressed in T cells can activate the IL-2 promoter (18), whereas their differential effects on the NFAT binding site in the TNF-α promoter (NFAT-TNF-α) have been reported (15, 18, 19). In addition, Wu et al. (20) recently demonstrated that promoter/enhancer activity of the NFAT-binding site in the IL-2 promoter (NFAT-IL-2), but not that of NFAT-TNF-α, was suppressed by FOXP3. Therefore, to investigate the molecular mechanisms underlying the differences among NFAT family members in activating distinct cytokines, in this study we comparatively examined the effects of NFATc2 and NFATc1 on the synthesis of IL-2 and TNF-α in human peripheral T cells as well as a Jurkat T cell line by using overexpression and knockdown systems as well as NFATc2/NFATc1 chimeric molecules. The results demonstrated a crucial role of the NFATc2-TAD2 domain in activation of the TNF-α promoter.
Materials and Methods
For the application of RNA interference (RNAi) technology, double-strand Stealth RNAi oligos (Invitrogen Life Technologies) designed using RNAi designer software (Invitrogen Life Technologies) were synthesized by Invitrogen Life Technologies. The target sequences used are as follows: NFATc2 no. 1, 5′-AUGGAUUCUGGAGCCGAGUUUCUCC-3′; NFATc2 no.2, 5′-UUAAGGAUCCGCUCAUCAGCUGUCC-3′; NFATc1 no.1, 5′-AAACUGGUUAUUGUUGUGGUACAGG-3′; NFATc1 no.2, 5′-GCCAACGGUAACGCCAUCUUUCUAA-3′; NFATc1 no.3, 5′-CACUGAUGAUUAUGAGCCUGCUCCA-3′; and NFATc1 no.4, 5′-GCAUGAGGACGGUAGUCCUAAUUUG. As the control, Stealth RNAi negative control duplexes (Invitrogen Life Technologies) were used. An anti-Xpress Ab was purchased from Invitrogen Life Technologies, anti-NFATc2 and anti-NFATc1 Abs were from Santa Cruz Biotechnology, and an anti-CD3 Ab came from Janssen Pharmaceutica. The anti-NFATc3 (NFATx, NFAT4) Ab is described elsewhere (21). All other reagents were from Sigma-Aldrich.
The PCR fragment of cDNA encoding full-length human NFATc2aa 1–925, NFATc11–716, and their chimeric molecules NFATc1-c1-c2 (NFATc11–716-NFATc2699–925), NFATc1-c2-c2 (NFATc11–419-NFATc2395–925), NFATc2-c1-c2 (NFATc21–394-NFATc1420–716-NFATc2699–925), NFATc1-c1-c1/C (NFATc11–697-NFATc1/C685–930), NFATc2-c2-c1/C (NFATc21–698-NFATc1/ C704–930), NFATc1-c2-c1/C (NFATc11–419-NFATc2395–698-NFATc1/C685–930), NFATc2/c2 (NFATc21–698), NFATc2/c1 (NFATc21–394-NFATc1420–716), and NFATc1/c2 (NFATc11–419-NFATc2395–698) or NFATc2-TAD2699–925 was subcloned in-frame into an appropriate site in the pEF6/His expression vector (Invitrogen Life Technologies). A mutant NFATc2 (NFATc2(KEF)) in which aa 914–916 (KEF) were exchanged for NDL and a mutant NFATc1-c1-c1/C (NFATc1-c1-c1/C(NDL)) in which aa 921–923 of NFATc1/C (NDL) were exchanged for KEF were also subcloned. In some experiments, the FLAG tag (MDYKDDDK) coding sequence was additionally subcloned in-frame into the resulting plasmid at the C-terminal end of the NFAT cDNA. The resulting NFATc2 and NFATc1 cDNA cassette, including the N-terminal Xpress-tag (DLYDDDDK), was cut out and subcloned into the CMV promoter-driven expression vector conjunct with an internal ribosomal entry site sequence (IRES) followed by the coding sequence of a luminescent protein, Venus, at the C-terminal end (22, 23). As reporter constructs, the 5′-flanking regions of human TNF-α (−670 to +147 relative to the transcription initiation site) and IL-2 (−418 to +2) genes were cloned into the pEGFP-1 vector (BD Bioscience Clontech). Six and five tandem repeats of the NFAT-binding site in the IL-2 promoter (−286 to −265) and the TNF-α promoter (−106 to −87), respectively, and three repeats of the AP-1 binding site in the metallothionein IIA gene (24, 25) and six repeats of the Jun/ATF2 binding site in the c-Jun promoter (AGCTAGCATTACCTCATCCCGATC) (26, 27) were subcloned into the pEGFP-1 vector in which the thymidine kinase minimum promoter sequence from the pRL-TK vector (Promega) was inserted. The correct sequences of all constructs were verified by sequencing.
Cells and transfection
With approval by the Ethical Review Committee of the Tokyo Metropolitan Institute of Medical Science (Tokyo, Japan), CD4+ T cells and CD4+CD45RO− naive T cells were prepared from the peripheral blood of healthy volunteers by positive selection using a magnetic cell sorting system (Miltenyi Biotec). The purity of the resulting cells was >95% as determined by flow cytometry. Then, NFAT-IRES-Venus expression vectors were transfected into the cells by electroporation using a human T cell Nucleofector kit (Amaxa). The transfected cells were cultured in RPMI 1640 medium supplemented with 10% FBS, 2 mM l-glutamine, 1 mM sodium pyruvate, 10 mM HEPES, 1 × MEM nonessential amino acid solution, 100 U/ml penicillin G, and 100 U/ml streptomycin.
SV40 T Ag-transfected human leukemic Jurkat T (Jurkat Tag) cells were grown in the same medium as described above. Cells in the logarithmic growth phase were transfected with various amounts of plasmid DNAs by electroporation as described previously (25). In each experiment, cells in different groups were transfected with the same total amount of plasmid DNA by supplementing expression vector DNA with the proper amount of the corresponding empty vector. To introduce Stealth RNAi oligos into the cells, the same transfection was repeated after a 48-h interval. Immunoblotting against expressed and endogenous protein in whole cell lysates was performed as described previously (28).
Messenger RNA expression
At 48 h after final transfection, cells were treated with 2.5 μg/ml anti-CD3 Ab plus 1 μg/ml anti-CD28 Ab in the presence of 5 μg/ml anti-mouse IgG cross-linking Ab or 5 nM PMA plus ionomycin for 6 h at 37°C. In some experiments, stimulation was performed after the purification of transfection-positive cells, as determined by the fluorescence derived from Venus, by the FACSAria cell sorting system (BD Biosciences). The purity of sorted cells was >98% (data not shown). Quantitative real-time RT-PCR for TNF-α, IL-2, IL-4, IL-13, and GM-CSF was performed using Assay-on-Demand gene expression products (TaqMan MGB probes) with an ABI prism 7900 sequence detection system (Applied Biosystems) as described previously (28).
Intracellular cytokine staining
At 48 h after transfection, CD4+ T cells were restimulated with 5 nM PMA plus 1 μM ionomycin for 6 h at 37°C in the presence of 2 μM monensin. The cells were fixed with 4% paraformaldehyde and permeabilized with 0.5% Triton X-100. After blocking with PBS containing 3% BSA, cells were stained with anti-IL-2 (clone MQ1-17H12)-PE and anti-TNF-α (clone Mab11)-PE Abs (BD Biosciences). Flow cytometric analysis was performed on a FACSCalibur device with CellQuest software (BD Bioscience).
After 16 h of stimulation, cytokine promoter-driven and transcription factor binding site-driven enhancer activity was assessed as the fluorescence of synthetic enhanced GFP (EGFP) detected by flow cytometry as described previously (28).
Chromatin immunoprecipitation (ChIP) assay
ChIP assay was performed using a ChIP assay kit (Upstate Biotechnology) with slight modifications. Briefly, FLAG-tagged, NFAT expression vector-transfected Jurkat cells (2 × 107) were cultured with 5 nM PMA plus 1 μM ionomycin for 30 min and then treated with 1% formaldehyde. After incubation for 30 min at room temperature, cells were washed twice, resuspended in 250 μl of lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl (pH 8.1), 1 mM PMSF, 1 μg/ml aprotinin, and 1 μg/ml pepstatin A), and incubated on ice for 10 min. The lysates were sonicated three times for 20 s each, and then the debris was removed by centrifugation. One percent of the lysate was used as the DNA input control. The remaining samples were precleaned with salmon sperm DNA/protein A agarose slurry and then incubated with anti-FLAG (M2) agarose beads for 16 h at 4°C. The resulting beads were washed five times according to the manufacturer’s protocol. The immunocomplexes were eluted and reverse cross-linked by incubation with 200 μl of elution buffer (1% SDS and 0.25 M NaCl) for 4 h at 65°C. The resulting DNA was subjected to analysis with a SYBR Green real-time PCR system (Takara) using primers specific for NFAT-IL-2 (5′ primer −281AAAAACTGTTTCATACAGAAGGCGTTA−255 and 3′ primer −141CTGATGACTCTTTGGAATTTCTTTA−165) and NFAT-TNF-α (5′ primer −228AGGATGGGGAGTGTGAGGG−210 and 3′ primer −81CCTTGGTGGAGAAACCCATGAGCTCATCT−109). The results were expressed as relative binding activity in comparison with the amount of input DNA. The detection limit of this assay, determined in the absence of anti-FLAG beads, was <1 × 10−5 of input DNA (data not shown).
Differential contribution of NFATc2 and NFATc1 to cytokine expression in T cells
To compare the roles of NFATc2 and NFATc1 in cytokine synthesis by T cells, each NFAT expression vector constructed with IRES-Venus at the C-terminal end was transfected into human peripheral naive CD4+ T cells. Twenty to 30% of the resulting cells were recognized to be transfection-positive, and the fluorescence levels were not significantly different among control vector-, NFATc2-IRES-Venus-, and NFATc1-IRES-Venus-transfected cells (Fig. 1,A). Upon activation through the TCR and a costimulatory molecule, the mRNA of IL-2 and TNF-α was clearly increased in the transfection-positive cells as determined by the fluorescence derived from Venus (Fig. 1,B). Inducible IL-2 expression was up-regulated by NFATc2 and NFATc1. In contrast, NFATc2 but not NFATc1 significantly enhanced TNF-α expression in naive CD4+ T cells (Fig. 1,B). Western blot analysis confirmed that NFATc2 and the short form NFATc1 proteins were overexpressed in each transfection-positive cell population (Fig. 1 C).
The production of IL-2 and TNF-α in the cells was further examined by intracellular staining. Forty to 50% of transfected cells produced IL-2 as well as TNF-α upon stimulation (Fig. 1,D), whereas <1% were cytokine positive without stimulation (data not shown). In the population of transfection-positive cells, IL-2 and TNF-α production was up-regulated by NFATc2. NFATc1 also augmented IL-2 production by CD4+ T cells, although the synthesis of TNF-α was not affected by the introduction of NFATc1 (Fig. 1 D). These findings suggest that NFATc1 is defective in the transactivation of TNF-α in human CD4+ T cells.
Next, the effects of NFATc2 and NFATc1 on the expression of several cytokines were comparatively analyzed. As shown in Fig. 2, Jurkat Tag cells expressed IL-4, IL-13, and GM-CSF along with IL-2 and TNF-α upon stimulation. Consistent with Fig. 1, increasing amounts of NFATc2 and NFATc1 overexpressed in the cells similarly up-regulated the expression of IL-2, as well as GM-CSF, in a dose-dependent manner. NFATc1 hardly affected the expression of not only TNF-α but also IL-13, although both were clearly enhanced by NFATc2. The up-regulation of IL-4 expression by NFATc1 was more potent than that by NFATc2 (Fig. 2). These findings suggest that the relative contribution of NFATc2 and NFATc1 differs among cytokines produced by T cells.
NFATc2-TAD2 is essential for transactivation of TNF-α
NFAT is composed of TAD1, followed by CRD, DBD, and TAD2. The shortest NFATc1 isoform is almost completely deficient in TAD2 (1, 17). To identify the responsible region causing the selective defect of NFATc1 in transactivation of the TNF-α gene, several NFAT chimeras in which functional domains were exchanged between NFATc2 and NFATc1 were used. For an accurate comparison of the potency of the expressed proteins, the density of the immunoblot bands was measured and the relative amount of expressed protein was determined in comparison with that of expressed NFATc2 as the control in a parallel experiment. At the same time, the relative potency of chimeric proteins as well as NFATc1 in the up-regulation of TNF-α expression was also assessed in comparison with that of NFATc2. Representative data in the case of NFATc1 are shown in Fig. 3,A. Through the introduction of increasing amounts of chimeric protein-expression vectors, expression level augmentation curves were plotted for each chimera (Fig. 3,B). Finally, the relative potency of chimeric proteins to transactivate the TNF-α gene at the same protein expression level as NFATc2 was calculated by extrapolation. The chimeras NFATc1-c1-c2, NFATc1-c2-c2, and NFATc2-c1-c2 substantially augmented TNF-α expression, whereas the effects of NFATc2-c2, NFATc2-c1, and NFATc1-c2 as well as wild-type NFATc1 lacking TAD2 were very weak, suggesting that TAD2 is required for NFAT-mediated transactivation of TNF-α. It was also suggested that the functional difference between NFATc2 and NFATc1 is not essentially caused by the heterogeneity of their TAD1, CRD, and DBD domains. NFATc1 is composed of several isoforms mainly derived by alternative translation initiation and polyadenylation events (15, 16, 17). Interestingly, the potency of the longest form, NFATc1/C (NFATc1-c1-c1/C), and the chimeras NFATc2-c2-c1/C and NFATc1-c2-c1/C, which carry TAD2 of NFATc1/C to activate TNF-α, were weaker than that of the NFATc2-TAD2-combined proteins (Fig. 3 B). These findings suggest that TAD2, especially that derived from NFATc2, is required for augmentation of TNF-α expression by NFAT.
It has been demonstrated that 15 amino acids in the C-terminal end of TAD2 are required for the maximum transactivation activity of NFAT (29). Therefore, we next examined the effects of mutants in which three amino acids in the corresponding region of NFATc2 (916KEF918) and NFATc1-c1-c1/C (921NDL923) were exchanged, on TNF-α expression. As shown in Fig. 3 B, the replacement of KEF by NDL in NFATc2 and that of NDL by KEF in NFATc1-c1-c1/C little repressed and augmented, respectively, their transactivation activity for TNF-α.
Knockdown effects of NFATs on cytokine expression
The roles of endogenously expressed NFATc2 and NFATc1 in cytokine expression were next examined by using RNAi technology. The expression of IL-2 and TNF-α in Jurkat Tag cells was similarly diminished by the introduction of two independent Stealth RNAi oligos against NFATc2, along with selective down-regulation of NFATc2 protein expression in the cells (Fig. 4). The expression of all NFATc1 isoforms in Jurkat Tag cells was down-regulated by the introduction of one NFATc1-Stealth RNAi oligo (no.1). In contrast another oligo (no.2), derived from a sequence close to the proximal poly(A) signal and not completely overlapping the sequences of longer isoforms, specifically diminished the expression of the shortest form of NFATc1 (Fig. 4, isoform A). The knockdown of all NFATc1 isoforms inhibited the expression of IL-2 as well as that of TNF-α. However, down-regulation of the shortest isoform alone resulted in a slight decrease in IL-2 but weak augmentation rather than suppression of TNF-α. In addition, successful knockdown of NFATc1/B plus NFATc1/C and NFATc1/C alone by the no. 3 and no.4 oligos induced a 50 ∼ 60% and 10 ∼ 20% reduction, respectively, of IL-2 and TNF-α expression. These findings are consistent with the results obtained by overexpression studies (Figs. 1–3) and suggest that the shortest NFATc1, like other isoforms and NFATc2, plays a positive role in IL-2 expression, although this isoform does not contribute as a transcription activator of TNF-α in physiological conditions.
Differential effects of NFATc2 and NFATc1 on cytokine promoters
To investigate the mechanisms by which NFATc2 and NFATc1 differentially affect the expression of IL-2 and TNF-α, a promoter reporter assay was performed. In agreement with the results of mRNA expression as shown in Fig. 2, inducible IL-2 promoter activity was enhanced by NFATc2 and NFATc1, whereas NFATc2 but not NFATc1 augmented TNF-α promoter activity (Fig. 5,A). Furthermore, a difference between NFATc2 and NFATc1 was also observed in the reporter assay using the NFAT binding sites in the IL-2 and TNF-α gene promoters (NFAT-IL-2 and NFAT-TNF-α, respectively). Thus, the enhancer activity of NFAT-IL-2 was up-regulated by NFATc2 and NFATc1, whereas that of NFAT-TNF-α was augmented by NFATc2 but not by NFATc1 (Fig. 5 A). Neither NFATc2 nor NFATc1 affected AP-1-derived activity, suggesting that the distinct effects of NFATc2 and NFATc1 on IL-2 and TNF-α expression were due, at least in part, to the effects on the corresponding binding region in their respective promoters.
To further investigate the role of TAD2 in TNF-α gene transcription, the effect of the TAD2 domain on NFAT-dependent transcriptional activity was examined. As shown in Fig. 5 B, ectopically expressed NFATc2-TAD2 behaved as a dominant negative in the transcriptional activity driven by the NFAT binding site in the TNF-α promoter. Thus, transfected NFATc2-TAD2 suppressed inducible NFAT-TNF-α activity in a dose-dependent manner, whereas NFAT-IL-2 activity was not affected. These results support the notion that TAD2 is important for transactivation of the TNF-α gene but not the IL-2 gene by NFAT.
It has been demonstrated that NFAT cooperates with the Jun/ATF2 heterodimer for the transactivation of TNF-α (30, 31, 32). Therefore, we next examined the effect of NFATc2-TAD2 on Jun/ATF2-dependent transcriptional activity. As shown in Fig. 5 B, inducible Jun/ATF2 activity was not affected by ectopically expressed NFATc2-TAD2, suggesting at least that the NFATc2-TAD2 domain, by itself, does not obstruct the association of Jun/ATF2 with their recognition sequence and/or the transactivation of Jun/ATF2-responsive genes.
NFATc2 and NFATc1 equivalently associate with NFAT-TNF-α
The findings of the reporter assay raise the possibility that interactions with NFAT sites in the IL-2 and TNF-α promoters differ between NFATc2 and NFATc1. A difference in the binding activity for cytokine promoters among NFAT family members has been suggested (18) although not fully elucidated, especially in physiological conditions. The next examination was, therefore, performed to compare the binding properties of NFATc2 and NFATc1 to the IL-2 and TNF-α promoters by using a ChIP assay. As shown in Fig. 6, almost the same amounts of the IL-2 and TNF-α promoter regions were coprecipitated with ectopically expressed NFATc2 as well as NFATc1 in Jurkat Tag cells upon stimulation. These results suggest that the selective defect of NFATc1 in up-regulation of TNF-α was not because of lack of binding activity for the TNF-α promoter/enhancer region.
The differential contribution of NFATc2 and NFATc1 to TNF-α gene transcription has been suggested by a reporter assay and an in vitro DNA binding assay (15, 18, 19). This study revealed that NFATc2, but not NFATc1, promoted TNF-α synthesis in human peripheral CD4+ T cells. This is generally consistent with previous reports that TNF-α synthesis by lymphocytes and/or T cells was impaired in NFATc2−/− mice (9) but not in NFATc1−/− mice (8). Furthermore, our present results demonstrating that the enhancement of TNF-α expression by NFAT was lost by the truncation of TAD2 and that NFATc2-TAD2 specifically suppressed NFAT-TNF-α activity clearly suggest that TAD2 is required for NFAT-mediated transactivation of the TNF-α gene but not the IL-2 gene.
It has been demonstrated that the full response at many NFAT sites requires concomitant activation of the AP-1 transcription factor family (1, 33). There is a discernible AP-1 site immediately downstream of the NFAT sites in the promoter region of IL-2 and other cytokine genes. Furthermore, the three-dimensional structures of NFATc2-DBD, AP-1 heterodimer (Jun/Fos), and the distal Ag-receptor response element in the IL-2 gene promoter, which we used as NFAT-IL-2 in this study, have been analyzed in detail (34). The importance of NFAT/AP-1 cooperativity in gene regulation was revealed by studies demonstrating that mutations of the AP-1-interacting domain of NFATc2 led to loss of transactivation of many cytokine genes, including IL-2 (35). Deletion of either the NFAT site or the AP-1 site is sufficient to destroy IL-2 promoter activity (36).
In contrast, the association of NFAT with the consensus sequence for a NFκB-binding site spanning −106 to −87 in the TNF-α gene promoter (κ3 site, which we used as NFAT-TNF-α in this study) is essential for TNF-α expression in T cells (37, 38, 39). Unlike the case of IL-2, Macian et al. (35) demonstrated that cooperation with AP-1 is not crucial for NFAT-dependent transactivation of the TNF-α gene. Cooperation with the Jun/ATF2 heterodimer, but not with AP-1 (Jun/Fos), is required for full induction of TNF-α expression by NFAT (30, 31, 32), although the Jun/ATF2-interacting region in NFAT has not been identified. In contrast to NFAT-IL-2 on which NFAT contacts AP-1 via its DBD, the following evidence suggests the possibility that NFAT interacts with Jun/ATF2 or other undefined coactivators on NFAT-TNF-α through TAD2. First, TAD2 was crucial for NFAT-dependent transactivation of the TNF-α gene (Fig. 3). Second, TAD2 behaved as a dominant negative against the transcriptional activity of NFAT-TNF-α, but not against that of NFAT-IL-2 (Fig. 5,B). In addition, NFATc2 and NFATc1 equivalently associated with NFAT-TNF-α in vivo (Fig. 6). Nevertheless, Jun/ATF2-dependent transcriptional activity was not affected by NFATc2-TAD2 (Fig. 5 B). Therefore, NFAT may cooperate with other factors than Jun/ATF2 in activating NFAT-TNF-α; otherwise, similar to the case of NFAT/AP-1 complex (34), strong interaction between NFAT and Jun/ATF2 may require their target DNA sequences. To elucidate additional details of the contribution of TAD2 to NFAT-TNF-α activation, analysis of the cocrystallized structure of the complex of transcription factors, including NFAT, on NFAT-TNF-α may be required as performed for the NFAT/AP-1/NFAT-IL-2 complex (34).
However, in disagreement with our present findings (Fig. 6), it has been reported that binding activity for NFAT-TNF-α was detectable in nuclear extracts of NFATc2- but not NFATc1- or NFATc3-transfected COS cells (18). In addition, Oum et al. (19), using in vitro DNA-protein binding assay and a reporter assay with HeLa cells, demonstrated that the functional disparity between NFATc2 and NFATc1 in the activation of TNF-α was due to the different binding specificity of NFATc2 and NFATc1 to NFAT-TNF-α. The reason for the discrepancy is unclear; our present data using NFATc2/NFATc1 chimeric molecules (Fig. 3) further support the notion that TAD2 rather than DBD predominantly contributes to the difference in TNF-α transactivation activity between NFATc2 and NFATc1. The condition in which the transcription complex was formed in vivo might not have been completely reproducible in the in vitro binding assay. Furthermore, interaction of transcription factors with their corresponding sequences may have differed among the cell types used. In fact, Monticelli et al. (40) reported that the association of NFATc1 with the murine IL-13 promoter was different between Th2 cells and mast cells. Therefore, our results showing that NFATc2 and NFATc1 equivalently associated with NFAT-TNF-α, demonstrated in T cells by ChIP assay, seem to be relatively reliable even though the physiological association between endogenous NFAT and the TNF-α promoter was too weak to further confirm this in our experimental conditions (data not shown).
In agreement with the notion that TAD2 is required for NFAT-TNF-α activation, positive participation of longer forms of NFATc1, which have TAD2, in TNF-α expression was suggested by the fact that reductions of all three NFATc1 isoforms by RNAi resulted in slight suppression of TNF-α, whereas the introduction of Stealth RNAi oligo, which down-regulated the shortest form of NFATc1 alone, failed to do so (Fig. 4). We also observed that TNF-α expression was enhanced by the longest NFATc1 isoform, although its potency was relatively weaker than that of NFATc2 (Fig. 3). These findings are partly consistent with a previous report demonstrating that the potency of NFATc.β, which has medium-length TAD2, to enhance NFAT-TNF-α activity was very low (15) and suggesting that the contribution of TAD2 derived from NFATc1 to TNF-α expression is smaller than that of NFATc2-TAD2. The essential domain in NFAT-TAD2 for its transactivation activity has not been fully elucidated. Imamura et al. (29) demonstrated that 15 amino acids, DITLDDVNEIIGRDM, in the C-terminal end of NFATx1 are required for its maximum transactivation activity in Jurkat T cells. In that region, the first 12 amino acids, DITLDDVNEIIG, are conserved in NFATc2904–915 as well as NFAT1c/C909–920. Exchange of the remaining three amino acids in NFATc2 (KEF) and NFATc1-c1-c1/C (NDL) little suppressed and enhanced, respectively, their transactivation activity, suggesting that the corresponding region is responsible for the differential participation of the TAD2 of NFATc2 and NFATc1/C origins. Because the effects of exchange were weak, the existence of another unknown region in TAD2 contributing to the transactivation activity is also suggested. Further examination will be required to elucidate the differential roles of TAD2 derived from NFATc2 and NFATc1/C in detail. Above all, NFATc1, especially its major and inducible shorter form, potentially contributes to the clonal expansion of activated T cells through IL-2 synthesis, but not to TNF-α-mediated exacerbation of inflammation.
Differential effects of NFATc2 and NFATc1 on the expression of IL-13 were also observed, although GM-CSF expression was enhanced by both NFATs (Fig. 2). These findings are consistent with previous reports demonstrating that the effect of NFATc1 on IL-13 gene transcription was weaker than that of NFATc2, even though this was shown in mast cells rather than Jurkat cells (40). Accordingly, IL-13 expression in mast cells was unchanged in NFATc1−/− mice (40). In the report of Macian et al. (35), augmentation of the promoter activity of IL-2 and GM-CSF by mutant NFATc2, which was unable to interact with AP-1, was weaker than that by wild-type NFATc2 even though the effects of both NFATc2s on TNF-α and IL-13 were equivalent. Similar to the case of IL-2, GM-CSF promoter activity was diminished by truncation of the AP-1 binding region neighboring the NFAT site (41). Taking these findings together with our present results, it is likely that NFATc1 activates GM-CSF in cooperation with AP-1, although TAD2 is required for transactivation of the IL-13 gene by NFAT through association with other cofactors. Because no sequence completely matching the NFAT-binding region in the TNF-α promoter was detectable in the 5′-flanking region of the IL-13 gene, the mechanisms by which NFATc1 fails to enhance IL-13 expression need to be further investigated.
In conclusion, TAD2 is required for NFAT-mediated transactivation of the TNF-α gene in T cells. NFATc1, especially its shortest isoform that lacks TAD2 and is predominantly induced upon cell activation, may play characteristic roles in immune responses through its specific activation of distinct cytokines. Our present findings will be useful for developing selective TNF-α inhibitors that will be promising drugs for treating autoimmune diseases, including rheumatoid arthritis.
We thank Drs. N. Arai and K. Arai for providing human NFAT cDNAs and M. Suzuki and Y. Fujiishi for technical assistance.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by grants-in-aid from the Japan Health Science Foundation (to M.S.), the Uehara Memorial Foundation (to K.O.), the Naito Memorial Foundation (to K.O.), the Pharmacological Research Foundation (to K.O.), and the Research Foundation for Pharmaceutical Sciences (to K.O.).
Abbreviations used in this paper: DBD, DNA-binding domain; ChIP, chromatin immunoprecipitation; CRD, Ca2+ regulatory domain; EGFP, enhanced GFP; IRES, internal ribosomal entry site; RNAi, RNA interference; TAD, transactivation domain.