Direct demonstration of NFATp dephosphorylation and nuclear localization in activated HT-2 cells using a specific NFATp polyclonal antibody.

Nuclear factor of activated T cells (NFAT) regulates transcription of a number of cytokine genes, and NFAT DNA binding activity is stimulated following T cell activation. Several lines of evidence have suggested that NFAT is a substrate for calcineurin, a serine/threonine phosphatase. Using a polyclonal antibody to murine NFATpp, Western blot analysis of various mouse tissues demonstrated that the 110-130-kDa NFATp protein was highly expressed in thymus and spleen. Treatment of immunoprecipitated NFATp from untreated HT-2 cells with calcineurin resulted in the dephosphorylation of NFATp, demonstrating that NFATp is an in vitro substrate for calcineurin. NFATp immunoprecipitated from 32P-labeled HT-2 cells migrated as an approximately 120-kDa protein that was localized to the cytosol of the cells. Treatment of the cells with ionomycin resulted in a decrease in the molecular weight of NFATp and a loss of 32P, consistent with NFATp dephosphorylation. The dephosphorylation of NFATp was accompanied by localization of the protein to the nuclear fraction. Both of these events were blocked by preincubation of the cells with FK506, a calcineurin inhibitor, consistent with the hypothesis that NFATp is a calcineurin substrate in cells.

P, consistent with NFAT p dephosphorylation. The dephosphorylation of NFAT p was accompanied by localization of the protein to the nuclear fraction. Both of these events were blocked by preincubation of the cells with FK506, a calcineurin inhibitor, consistent with the hypothesis that NFAT p is a calcineurin substrate in cells.
The immunosuppressant drugs cyclosporin A (CsA) 1 and FK506 inhibit the early steps of antigen-stimulated T cell activation. These drugs prevent activation of a variety of cytokine genes, including IL-2, granulocyte-macrophage colony-stimulating factor, IL-4, and tumor necrosis factor ␣ (for review, see Refs. 1 and 2). IL-2 plays a key role in controlling T cell proliferation, and consequently, numerous studies have focussed on understanding IL-2 gene regulation. The results of these studies demonstrated that transcriptional activation requires nuclear factor of activated T cells (NFAT), a DNA binding protein, which binds to specific sites on the regulatory regions of the cytokine genes (3)(4)(5)(6).
Molecular cloning and biochemical studies have provided insights into the actions of NFAT. The approximately 120-kDa NFAT is a member of a gene family whose members appear to contain a Rel homology region, which is important for DNA binding activity. Two members were identified: NFAT c was cloned from a human T cell library (7), while NFAT p was purified and cloned from murine T cells (8). In addition, several alternatively spliced forms of NFAT p were identified (2). NFAT c and NFAT p are approximately 73% identical in the Rel homology region, although they share little similarity outside that domain. In unactivated T cells, NFAT DNA binding activity is localized primarily in the cytosol, and following T cell activation, activity is detected in the nucleus. Nuclear NFAT forms a complex with Fos and Jun, and mutations in NFAT that inhibit binding to these proteins eliminate NFAT-mediated gene transcription. Furthermore, treatment of cells with either CsA or FK506 inhibits the appearance of NFAT binding activity in the nucleus (3, 9 -11).
An understanding of the link between the drugs FK506 and CsA and NFAT has recently started to emerge with the identification of calcineurin (CaN) as a cellular target of the drugs. FK506 and CsA bind to their respective immunophilins, FKBP-12 and cyclophilin A, and the drug-immunophilin complexes bind to and inhibit the activity of CaN, a calmodulin-dependent phosphatase (12,13). Overexpression of a constitutively active form of CaN in T cells renders the cells more resistant to the effects of the drugs, and the ability of a number of FK506 and CsA analogues to inhibit the phosphatase activity of CaN correlates with their ability to inhibit T cell activation (14).
Evidence suggests that CaN is not just involved in mediating the actions of FK506 and CsA but that the enzyme plays a role in the physiological pathway of T cell activation. O'Keefe, et al. (15) demonstrated that in cells transfected with CaN, PMAand ionophore-stimulated IL-2-driven gene expression was stimulated approximately 2-fold. Furthermore, in cells overexpressing a constitutively active form of CaN, IL-2-driven gene expression was stimulated approximately 150-fold in the presence of PMA alone, indicating that active CaN can replace the usual requirement for ionophore. Similar results were obtained by Clipstone and Crabtree (16), who demonstrated reporter gene activity in CaN-transfected cells at suboptimal concentrations of ionomycin that elicited little or no activity in control cells. Taken together, these results demonstrate that CaN overexpression augments T cell activation resulting from both calcium and PKC stimulation and suggest that CaN plays a key role in this process.
The studies with CsA and FK506 led to the suggestion that NFAT is a CaN substrate, either direct or indirect, in T cells. First, indirect evidence indicates that NFAT p is a phosphoprotein. Fractionation of Jurkat cell cytosol demonstrated that NFAT DNA binding activity is detected with proteins in a molecular mass range of 94 -116 kDa, suggesting heterogeneity in the size of the NFAT protein (17). To determine whether NFAT phosphorylation could explain the heterogeneity, Mc-Caffrey, et al. (18) treated cell extracts enriched for NFAT with alkaline phosphatase. This enzymatic treatment resulted in a shift in NFAT binding activity from the 127-143-kDa molecular weight fraction to a 101-113-kDa fraction, which is consistent with the hypothesis that NFAT is a phosphoprotein. Second, treatment of purified NFAT with purified bovine brain CaN also resulted in a shift in the molecular weight of NFAT, suggesting that NFAT is an in vitro substrate for CaN (19). Whether NFAT is a direct CaN substrate in intact cells has not yet been determined, however.
In the majority of reported NFAT studies, NFAT DNA binding activity has been the end point usually measured, since relatively little is known about the NFAT protein. However, the recent cloning and expression of NFAT genes has yielded valuable information about the proteins, and direct protein analyses should now be possible. Towards that end, antibodies to NFAT will be important reagents to develop for probing protein structure and function. In the studies reported here, we developed a peptide antibody to NFAT p , and we used it to characterize NFAT p in cells. We demonstrate directly in 32 P labeling experiments that NFAT p is a phosphoprotein and that T cell activation results in NFAT p dephosphorylation that can be blocked by pretreatment with FK506. Furthermore, we show that dephosphorylation is accompanied by translocation of NFAT p protein from the cytosol to the nucleus of cells.  (22) was synthesized at The Upjohn Company.
Cell Culture-The human T lymphoblastoid cell lines J5D9 and HSB and the murine T lymphoma cell line EL-4 were grown in RPMI 1640 medium containing 10% fetal calf serum, 2 mM glutamine, 100 units/ml penicillin G, and 100 g/ml streptomycin. The murine helper T cell line HT-2 was grown in the medium just described with the addition of 12% rat T-STIM with concanavalin A as a source of IL-2 (Collaborative Biomedical Products), 50 M ␤-mercaptoethanol, and 10 mM HEPES, pH 7.4. Human peripheral blood lymphocytes were prepared from leukophoresis packs. Cells were centrifuged over Ficoll, and cells at the interface were collected, diluted with RPMI, and pelleted. The cell pellet was suspended in RPMI containing 10% fetal calf serum and incubated for 1 h at 37°C. Nonadherent cells were collected, pelleted, and stored at Ϫ70°C.
Antibody Production-The amino acids CGG were added to the amino terminus of the sequence LSPGAYPTVIQQQTAPSQR corresponding to peptide 25 of NFAT p (8). The cysteine residue was coupled to maleimide-activated KLH using the Imject Immunogen Conjugation Kit (Pierce), and rabbits were injected. Sera was tested against purified peptide, and bleeds determined to have the highest titer were affinitypurified by using the Pierce Ag/Ab immobilization kit with Sulfolink coupling gel. The antibody preparation used in the experiments shown here recognized 1 ng of purified peptide at a 1:1000 dilution.
Tissue Homogenates-Homogenates were prepared from organs dissected from BALB/C mice as described in Ref. 20 using a Polytron homogenizer. The tissues were homogenized at 0.3 g wet weight/ml in 10 mM Tris, pH 8.0, containing 50 mM NaCl, 1 mM EDTA, 100 g/ml aprotinin and soybean trypsin inhibitor, 250 M leupeptin, 10 mM iodoacetamide, and 2 mM phenylmethylsulfonyl fluoride. SDS was added to bring the final concentration to 2%. The sample was heated at 100°C for 10 min and then centrifuged for 10 min in a microfuge. The supernatant was removed, and protein was determined by the microassay method of Bradford (21) using reagents from Bio-Rad.
Cellular Fractionation-Treatment of HT-2 cells (1 ϫ 10 6 cells/condition) was carried out as described in the figure legends. The cells then were washed in ice-cold PBS and resuspended (2.5 ϫ 10 6 cells/ml) in buffer (10 mM Tris-Cl, pH 7.5, containing 10 mM NaCl, 3 mM MgCl 2 , 1 mM phenylmethylsulfonyl fluoride, 0.5 mM dithiothreitol, 0.1 mM EGTA, 2 M leupeptin, 1 g/ml aprotinin, and 0.05% Nonidet P-40). Cells were centrifuged at 650 ϫ g to pellet the nuclei, and the supernatant was removed. The nuclear pellet was washed in the above buffer minus detergent, and the final pellet was resuspended in Laemmli buffer. The supernatant was recentrifuged to remove any residual nuclei. Equal cell equivalents of soluble and nuclear fractions were electrophoresed on 6% SDS-polyacrylamide gels, and Western blotting was carried out as described below.
Western Blots-Cells were treated (2.5 ϫ 10 5 cells/condition) as indicated in the figure legends, and control cells received Me 2 SO as a vehicle. Following treatment, the cells were washed with ice-cold PBS and resuspended in 40 mM Tris, pH 7.5, containing 10 mM EDTA, and 60 mM sodium pyrophosphate. SDS was added to bring the final concentration to 5%. The sample was boiled for 15 min and electrophoresed on a 6% SDS-polyacrylamide gel. Proteins were transferred to nitrocellulose, and nonspecific sites were blocked with 5% nonfat dry milk in PBS. The blot was developed with a 1:500 dilution of affinity-purified NFAT p antibody in 20 mM Tris, pH 7.5, containing 500 mM NaCl and 3% bovine serum albumin for 1 h at room temperature followed by incubation with a 1:5000 dilution of donkey anti-rabbit IgG linked to horseradish peroxidase for 1 h (Amersham Corp.). Immunoreactive proteins were detected with the ECL system. For the phosphotyrosine blot ( Fig.  3B), lysate from untreated HT-2 cells or epidermal growth factortreated A431 cells (300 ng/ml epidermal growth factor for 15 min) were probed with 1 g/ml 4G10 antibody (UBI), and the blot was developed by ECL. 32 P Immunoprecipitation-HT-2 cells at a density of 1 ϫ 10 5 cells/ml were grown overnight in growth medium as described above except that the medium contained 2% of the usual phosphate concentration and 10% dialyzed fetal calf serum. The cells were resuspended in fresh medium containing 100 Ci/ml [ 32 P]orthophosphate. Cells were aliquoted at 1.5 ϫ 10 6 cells/condition, and treatments were carried out as described in the figure legends. The cells were allowed to incorporate radiolabel for a total time of 4 h. After 2.8 h of incubation, FK506 (500 nM) was added for 1 h, followed by ionomycin (2 M) for an additional 10 min. The cells were washed in ice-cold PBS and lysed in radioimmune precipitation buffer (10 mM Tris, pH 7.5, containing 1% Triton X-100, 1% deoxycholate, 0.1% SDS, and 150 mM NaCl) containing 60 mM sodium pyrophosphate and 10 mM EDTA. The sample was precleared with Protein A-Sepharose CL-4B beads for 30 min at 4°C. SDS was added (0.3% final concentration), and 10 l of affinity-purified antibody was added to the sample for 2 h at 4°C followed by Protein A for 1 h. The Protein A beads were washed three times in radioimmune precipitation buffer containing 500 mM NaCl and 2 mg/ml bovine serum albumin, followed by three washes in radioimmune precipitation buffer containing 0.3% SDS, followed by a final wash in radioimmune precipitation buffer. The sample was resuspended in Laemmli buffer and electrophoresed on a 6% SDS-polyacrylamide gel followed by autoradiography. The autorad was quantitated on a Molecular Dynamics PhosphorImager using ImageQuant Software.
Immunocomplex Assays-Unlabeled cells (1.5 ϫ 10 6 cells/condition) were untreated or stimulated for 10 min with 2 M ionomycin, immunoprecipitated as described above, and washed five times in assay buffer (40 mM Tris, pH 7.5, containing 100 mM NaCl, 0.5 mM CaCl 2 , 0.5 mM dithiothreitol, 0.1 mg/ml bovine serum albumin, 1 mM MgCl 2 , and 0.1 mM MnCl 2 ). The immunoprecipitate from untreated cells was incubated in assay buffer containing the enzymes and inhibitors as indicated in the Fig. 3A legend in a final volume of 100 l for 30 min at 30°C. 50 l of Laemmli buffer was added, the sample was boiled, and the Sepharose beads were pelleted. The supernatant was electrophoresed on a 6% SDS-polyacrylamide gel, and Western blotting was carried out as above.
Phosphatase Assays-The phosphatase activity of CaN and PP2A was measured in a total volume of 100 l containing 40 mM Tris, pH 8.6, 100 mM NaCl, 0.5 mM CaCl 2 , 0.5 mM dithiothreitol, 0.1 mg/ml bovine serum albumin, 10 mM MgCl 2 , 600 M 4-methylumbelliferryl phosphate and either 25 nM CaN or 0.1 units of PP2A. To measure CaN activity, assays also included 100 nM calmodulin and the presence or absence of 100 M inhibitory peptide (22). The mixture was incubated for 1 h at 37°C followed by the addition of 50 l of stop buffer (300 mM glycine, pH 11.2, containing 15 mM EDTA), and the fluorescence was read at 365/405 nm.
[ 35 S]Methionine Pulse-Chase-Cells were resuspended at 3 ϫ 10 5 cells/ml for 30 min in growth medium containing 2% of the usual methionine plus 10% dialyzed fetal calf serum, in the presence or absence of 2 M ionomycin. [ 35 S]Methionine was added at 175 Ci/ml for 5 min, and 1.5 ϫ 10 6 cells were removed for the zero time point. The remainder of the cells were washed in PBS and resuspended in normal growth medium in the presence or absence of 2 M ionomycin. At the indicated times, aliquots (1.5 ϫ 10 6 cells) were removed, lysed, and immunoprecipitated with the NFAT p antibody as described above. The autorad was quantitated as described above.
[ 35 S]Methionine Degradation-Cells were resuspended at 6 ϫ 10 5 cells/ml in low methionine-containing medium plus [ 35 S]methionine (50 Ci/ml) and incubated for 18 h. The cells were washed in PBS and resuspended at 3 ϫ 10 5 cells/ml in normal growth medium containing Me 2 SO (control), 2 M ionomycin, or 100 nM FK506 plus 2 M ionomycin. Cells were pretreated with FK506 5 min prior to the addition of ionomycin. Cells (1.5 ϫ 10 6 ) were removed immediately (zero time point) or at the indicated times, and the lysates were immunoprecipitated as described above. Autorads were quantitated as described above and analyzed by linear regression to determine the NFAT p half-life. RESULTS A polyclonal antibody was developed against a peptide from the COOH-terminal domain of NFAT p . This peptide is outside the Rel homology domain and is from a unique region of NFAT p that is not present in NFAT c (7). Therefore, this antibody is a specific reagent for characterizing the NFAT p protein.
Western blot analysis was carried out in order to determine the expression of NFAT p in various mouse tissues (Fig. 1A). The highest level of NFAT p expression was observed in spleen and thymus tissue. In thymus, the antibody strongly hybridized with proteins in the 110 -130-kDa molecular mass range, suggesting protein heterogeneity, while in spleen tissue, reactivity was predominantly against a 110-kDa protein. In both these tissues, reactivity with smaller proteins of 72, 66, and 59 kDa was observed. It is unknown whether these proteins are proteolytic fragments of the larger NFAT p protein. However, in competition experiments in which the NFAT p peptide was included in the antibody incubation, no reactivity was observed, indicating the specificity of the interactions (lanes 8 and 9). In kidney and liver, there was little or no antibody reactivity against proteins in the 120-kDa range. Cross-reactivity against proteins of 66 and 72 kDa was shared among kidney, spleen, and thymus, although interestingly, in spleen and thymus, reactivity against the 66-kDa protein was predominant, while in kidney, antibody reactivity was greater against the 72-kDa protein, compared with the 66-kDa protein. Expression of NFAT p was very low in brain, while in heart no NFAT p was detected.
The high level of NFAT p expression in spleen and thymus is consistent with previous reports demonstrating NFAT DNA binding activity in T cells. Various T cell lines were profiled to investigate whether NFAT p levels differ among T cells (Fig.  1B). Peripheral blood lymphocytes had the lowest level of NFAT p expression (lane 4). These results could not simply be explained by a lack of reactivity of the murine peptide-derived antibody against human protein, since human T cell Jurkat cells showed significant reactivity with the antibody (lane 1). HSB cells, a human T cell line, also expressed the approximately 140-kDa NFAT p (lane 2). In EL-4 cells, a mouse thymoma cell line, NFAT p was expressed, but the protein was more heterogeneous, since proteins from 120 to 140 kDa reacted with the antibody (lane 3). HT-2 cells, an IL-2-dependent mouse T cell line, had the highest level of NFAT p expression compared with the other cell lines (lane 5). In the experiment shown in Fig. 1B, lysate from 2.5 ϫ 10 5 HT-2 cells resulted in a comparable or greater signal, compared with the other cell lines, in which 4-fold more cell equivalents (1 ϫ 10 6 ) were used. In all of the cell lines, cross-reactivity with a 66-kDa protein was seen, suggesting, as in the tissue samples, that other NFAT p protein forms may be expressed. Competition experiments using NFAT p peptide and EL-4 or HT-2 cell extracts ( lanes 6 and 7) demonstrated that all the reactivity could be competed by excess peptide. No NFAT protein was detected in MOLT-4 cells, a human T cell leukemia line or in NIH 3T3 cell fibroblasts (data not shown).
Activation of T cells by treatment with PMA and calcium ionophore stimulates NFAT DNA binding activity (2). The NFAT p antibody was used to investigate directly the effect of T cell activation on NFAT protein levels ( Fig. 2A). For these and subsequent experiments, HT-2 cells were used, due to their high level of expression of NFAT p . NFAT p migrated as an approximately 120-kDa protein in untreated cells. Treatment of the cells with either ionomycin alone (lane 2) or the combination of PMA plus ionomycin (data not shown), resulted in a decrease in the molecular weight of NFAT p , suggestive of a proteolytic and/or dephosphorylation event. Identical shifts in the NFAT p protein were observed when the cell treatments and lysate preparation were carried out in the presence of a panel of protease inhibitors, including leupeptin, phenylmethylsulfonyl fluoride, and 1-chloro-3-tosylamido-7-amino-2-heptanone (data not shown), suggesting that the mobility shift was not the result of proteolysis. The predominant NFAT p protein band in the ionomycin-treated cells was approximately 110 kDa, although the protein was heterogenous in size. Pretreatment of the cells with 500 nM FK506 blocked the ionomycin-induced shift in NFAT p (lane 3), suggesting a role for CaN. As a control, cells were pretreated with rapamycin prior to ionomycin treat- ment. Rapamycin binds to FKBP-12, but unlike FK506, it does not inhibit CaN, nor does it inhibit NFAT DNA binding activity. Rapamycin pretreatment also did not block the NFAT p mobility shift induced by ionomycin (lane 4).
To determine whether the changes in NFAT p mobility were the result of ionomycin-induced dephosphorylation, 32 P labeling experiments were carried out (Fig. 2B). Cells were labeled with [ 32 P]orthophosphate for 4 h followed by treatments similar to those shown in Fig. 2A, and immunoprecipitation was carried out with the NFAT p antibody. A similar decrease in molecular weight was observed in the 32 P-labeled NFAT p following ionomycin treatment as was observed in the Western blot experiments. Quantitation of the 32 P autoradiograms showed that ionomycin treatment resulted in an approximately 65% loss in 32 P from NFAT p (data not shown). Interestingly, ionomycin treatment never resulted in a complete loss of 32 P from NFAT p , indicating that complete dephosphorylation did not occur. In addition, FK506 inhibited the loss of 32 P, while rapamycin had no effect. The results of these experiments demonstrate directly that the shift in molecular weight following ionomycin treatment was the result of NFAT p dephosphorylation and that FK506 pretreatment blocked the dephosphorylation.
A role for CaN in the ionomycin-stimulated dephosphorylation of NFAT p was shown directly in immune complex assays (Fig. 3). NFAT p was immunoprecipitated from untreated HT-2 cells, and the immunoprecipitates were washed and incubated with CaN. CaN dephosphorylated NFAT p in vitro, as shown by a decrease in the NFAT p molecular weight, which was comparable to the molecular weight shift resulting from treatment of cells with ionomycin ( compare lanes 2 and 3). CaN contains an autoinhibitory domain in its COOH-terminal domain, and a peptide from this region inhibits CaN dephosphorylation of a cAMP-dependent protein kinase peptide in vitro (Ref. 22 and data not shown). Addition of the autoinhibitory peptide to the immune complex assay blocked the dephosphorylation of NFAT p by CaN (lane 4). In contrast, treatment of the precipitated NFAT p with another serine/threonine phosphatase, PP2A, resulted in little or no dephosphorylation of NFAT p (lane 5), although the enzyme readily dephosphorylated 4-methylumbelliferryl phosphate (data not shown). The tyrosine phosphatase PTP1 also did not dephosphorylate NFAT p (lane 7), suggesting that NFAT p may not contain phosphorylated tyrosine residues. Consistent with this hypothesis was the observation that there was no reactivity on Western blots of antiphosphotyrosine antibodies with NFAT p (Fig. 3B). Taken together, these results indicate that NFAT p readily serves as an in vitro substrate for CaN.
Previous studies demonstrated that activation of T cells is accompanied by an increase in nuclear NFAT DNA binding activity. Using the NFAT p antibody, we showed directly the presence of the NFAT p protein in the nucleus of stimulated cells (Fig. 4A). In untreated cells, all of the NFAT p protein was detected in the low speed supernatant (lane 1). Following a 10-min ionomycin treatment, however, the lower molecular weight, dephosphorylated form of NFAT p was predominantly localized to the nuclear fraction (lane 5). Pretreatment of cells with FK506 prior to ionomycin stimulation blocked the dephosphorylation, as previously observed, and also blocked the translocation of NFAT p to the nucleus (compare lanes 3 and 6).
The time course for NFAT p dephosphorylation and nuclear localization was examined in more detail (Fig. 4B). Fractionation of cells immediately following the addition of ionomycin indicated that NFAT p dephosphorylation was rapid (lane 4), as shown by the ladder of protein bands. After ionomycin treatment for 10 min, the dephosphorylated NFAT p was present in both the low speed supernatant and the nuclear fractions (lane 5), suggesting that NFAT p is dephosphorylated in the cytosol, followed by localization of the dephosphorylated form to the nucleus. With increasing time of incubation with ionomycin, dephosphorylated NFAT p was present almost exclusively in the nuclear fraction, although a small level of dephosphorylated NFAT p was detected in the supernatant throughout the 4-h time course. These results suggest that NFAT p is rapidly dephosphorylated and localized to the nucleus following ionomycin treatment and that the dephosphorylated NFAT p remains in the nucleus in the continuing presence of ionomycin.
A similar localization experiment was carried out with cells pretreated with ionomycin, followed by removal of the ionomycin (Fig. 4C). At the start of the experiment, following a 10 min treatment with ionomycin, essentially all of the NFAT p was localized to the nucleus, and the nuclear NFAT p was in the lower molecular weight, dephosphorylated form. After 10 min following washout of the ionomycin, NFAT p was still in the nuclear fraction, although a low level of the higher molecular weight, phosphorylated form of NFAT p was detected in the low speed supernatant. However, at the 30-min time point and beyond, NFAT p was localized in the low speed supernatant fraction, with no detectable NFAT p in the nucleus. These results suggest that following ionomycin washout, NFAT p is recycled from the nucleus back into the cytosol. Furthermore, the NFAT p in the low speed supernatant was exclusively in the higher molecular weight, phosphorylated form, suggesting that migration from the nucleus was accompanied by rapid rephosphorylation of NFAT p .
Loss of NFAT p from the nuclear fraction could result from NFAT p degradation. To investigate this possibility, [ 35 S]methi-onine labeling experiments were carried out to measure the half-life of the NFAT p protein. Cells were labeled overnight with [ 35 S]methionine and then chased in medium containing unlabeled methionine. The rate of loss of immunoprecipitated NFAT p was determined for untreated, ionomycin-treated, and FK506-and ionomycin-treated cells, and is shown as a semi-log plot (Fig. 5A). In untreated cells the t1 ⁄2 for NFAT p was 16.9 Ϯ 0.86 h (mean Ϯ S.E., n ϭ 3), which was similar to the value obtained from cells treated with the combination of FK506 and ionomycin (19.2 Ϯ 3.9 h). Treatment with ionomycin, however, caused a slight decrease in the time required for loss of 50% of the prelabeled NFAT p , to 11.9 Ϯ 2.7 h. These results indicate that the loss of NFAT p from the nucleus was not the result of degradation of the protein and support the hypothesis that nuclear NFAT p reappears in the cytosol following ionomycin removal.
During short labeling times, the major NFAT p was in the dephosphorylated form (Fig. 5B). However, this form was rapidly chased into the higher molecular weight, phosphorylated form of NFAT p , suggesting that phosphorylation of NFAT p occurs very rapidly following synthesis. DISCUSSION Much of the available information concerning NFAT is the result of studies of NFAT DNA binding activity (1-3, 9, 11). More recently, several forms of NFAT have been cloned, and Northern analysis demonstrated that NFAT mRNA levels vary among different tissues and under stimulus conditions (7). Investigations with a specific NFAT p antibody allowed expansion of these studies to include direct analysis of the NFAT p protein. The peptide sequence used to generate the antibody described here is derived from the COOH-terminal region of NFAT p , which is not present in NFAT c , and thus the antibody differentiates between the two NFAT forms.
The NFAT p protein was highly expressed in thymus and spleen tissue, which is consistent with mRNA analysis (7). However, Northrup et al. (7) reported that NFAT p mRNA levels were similar in brain, heart, thymus, and spleen, while our immunoblotting results (Fig. 1) showed that no NFAT p protein was detected in heart tissue, and only very low levels in brain. These results suggest that the presence of NFAT p mRNA may not predict the expression of the protein in all tissues or cells. In addition, lower molecular weight proteins reacted with the NFAT p antibody, suggesting that other forms of NFAT p may exist. Consistent with these results is the report that COOHterminal splicing variants of NFAT p have been cloned (2). Peptide mapping or purification of the lower molecular weight, cross-reactive proteins is required to determine whether they are related to NFAT p .
Ionomycin treatment of the HT-2 cells resulted in a significant shift in the molecular weight of NFAT p , as shown by immunoblotting analysis. The 32 P labeling experiments provided a direct demonstration that NFAT p is a phosphoprotein and that the shift in molecular weight resulted from dephosphorylation of NFAT p . The dephosphorylation-induced molecular weight mass was approximately 10 kDa and represented an approximately 65% loss in phosphate from the protein. The extent of phosphorylation of NFAT p is not known, and the results shown here indicate that dephosphorylation results in a significant change in the mobility of the protein on SDS-polyacrylamide gels. This result has practical applications, for it allows detection of the phosphorylated and dephosphorylated forms of NFAT p using Western blot analysis.
The results of the localization experiments suggest that the phosphorylation state of NFAT p may be a determinant of its localization within the cell. In untreated cells, NFAT p was in the phosphorylated state, and was localized to the low speed supernatant. However, in ionomycin-treated cells, dephosphorylated NFAT p was found in the nuclear fraction. Kinetic analysis, shown in Fig. 4, suggests that dephosphorylation occurs prior to nuclear localization, since at the 10-min time point, dephosphorylated NFAT p was present in both the cytosolic and nuclear fractions, while at later time points the dephosphorylated form was exclusively localized to the nucleus. As long as ionomycin was present, NFAT p remained both dephosphorylated and in the nucleus. Previous studies demonstrated that NFAT p DNA binding activity is localized to the cytosol of untreated T cells, and to the nuclear fraction in stimulated cells (3). The results shown here demonstrate directly that the shift in DNA binding activity seen by others reflects a change in the cellular location of the NFAT p protein.
Following ionomycin removal, NFAT p reappeared in the cell supernatant fraction, in the higher molecular weight, phosphorylated form. The [ 35 S]methionine labeling experiments demonstrated that NFAT p is a relatively stable protein, with a half-life of approximately 17 h in untreated cells and approximately 12 h in ionomycin-treated cells. Thus, the loss of signal from the nucleus following ionomycin removal cannot be explained by degradation of NFAT p . These results suggest that NFAT p may shuttle in and out of the nucleus, depending on the activation state of the cell as well as the phosphorylation state of the protein. A potential nuclear localization sequence has been identified within the Rel homology region of NFAT p (23). There are a number of potential phosphorylation sites in NFAT p , including one within the nuclear localization sequence. One possibility is that activation-induced dephosphorylation of NFAT p results in an unmasking of the nuclear localization sequence, leading to nuclear localization of the protein.
The immunocomplex assay demonstrated directly that NFAT p is a CaN substrate in vitro. Furthermore, the CaNinduced molecular weight shift in NFAT p was similar to that resulting from ionomycin treatment of cells. Treatment of intact cells with FK506, a CaN inhibitor, blocks the ionomycininduced dephosphorylation, which further supports but does not prove the hypothesis that NFAT p is a CaN substrate in intact cells.
In general, no consensus sequence for dephosphorylation of substrates by CaN has been identified. However, Donella-Deana et al. (24) have demonstrated in phosphopeptide studies that basic residues on the NH 2 -terminal side of the phosphoamino acid, especially at the Ϫ3 position, are positive determinants for dephosphorylation by CaN, while acidic residues on the COOH-terminal side are negative determinants. Because of a lack of a strong consensus sequence, these results and the results of others (25,26) have led to the suggestion that more complex protein structural determinants also play a role in defining CaN substrate specificity. Taken together, these results suggest that a number of sites within NFAT p may be sites for CaN dephosphorylation. Mapping these sites in vitro and comparing them with the sites of dephosphorylation in intact cells is required to establish whether NFAT p is a direct substrate for CaN in intact cells. The results of such experiments will define in more detail the exact role of CaN in T cell activation.