Impaired NFAT Transcriptional Activity in Antigen-Stimulated CD8 T Cells Linked to Defective Phosphorylation of NFAT Transactivation Domain¹

Among the different NFAT isoforms, NFATc2 (NFAT1, NFATp), NFATc1 (NFAT2, NFATc), and NFATc3 (NFAT4) are constitutively expressed in T cells. Mostly studied in CD4 T cells, NFAT has been implicated in the regulation of numerous cytokines, regulatory molecules such as Fas ligand and CD40 ligand, and cell cycle regulators (1, 2, 3). Studies of compound-deficient animals indicate that NFATc3 primarily regulates T cell differentiation in the thymus, whereas NFATc1 and NFATc2 are more specifically involved in regulating mature T cell activation and differentiation (4). Mice deficient for NFATc2 develop mild splenomegaly that correlates with enhanced CD4 T cell response with Th2 characteristics (5, 6, 7, 8). Enhanced proliferation of NFATc2-deficient CD4 T cells has been linked to reduced apoptosis and increased cell cycle entry and progression, the latter resulting from increased expression of cyclin A2, E, and F and cyclin-dependent kinase 4 (2, 6, 9). IL-2 and IFN-γ expression were moderately reduced in NFATc2-deficient CD4 T cells (7). Likewise, IFN-γ production is partially reduced in NFATc2-deficient CD8 T cells (10). Apart from a marked reduction in IL-4 production, NFATc1-deficient CD4 T cells did not demonstrate other major dysfunctions (11, 12). These observations, together with the observation that doubly deficient CD4 T cells are severely impaired in the production of multiple cytokines, led to the conclusion that NFATc2 negatively regulates T cell proliferation while promoting Th1 cell differentiation, whereas NFATc1 would support type 2 immune response (13). Whether NFAT proteins similarly regulate CD8 T cell activation and differentiation is, however, unclear.

The NFAT proteins are phosphorylated and reside in the cytoplasm of unstimulated T cells. An increase in intracellular Ca²⁺ induced by TCR signaling will activate the serine phosphatase calcineurin that dephosphorylates NFAT, leading to exposure of the nuclear localization sequence and NFAT translocation to the nucleus. An in depth study of NFATc2 regulation showed that NFATc2 is phosphorylated on 14 conserved serine residues in its regulatory domains, 13 of which are dephosphorylated in a calcineurin-dependent manner (14). Although dephosphorylation of the 13 phosphoserines is necessary for nuclear translocation, it does not by itself permit NFAT-dependent transactivation in Jurkat T cells. Additional phosphorylation of the ⁵³SSPS⁵⁶ motif in the transactivation domain by protein kinase C (PKC)⁴ -dependent, calcineurin-independent kinase(s) is important for transcriptional activity (14). Although all serine motifs of the regulatory domains are highly conserved among the different NFAT family members, the ⁵³SSPS⁵⁶ motif is only present in NFATc2. Several kinases are known to rephosphorylate NFAT on different serine residues and thus contribute to fine regulation of NFAT subcellular localization, DNA binding affinity, and transcriptional activity (15). Recent studies showed that the kinases DYRK1A and 2 (dual-specificity tyrosine-phosphorylated regulated kinases 1A and 2) rephosphorylate the SP3 motif of the NFAT regulatory domain, thus priming for further phosphorylation of the conserved SP2 and SRR-1 domains by casein kinase 1 and glycogen synthase kinase 3 (16, 17, 18, 19). These concerted phosphorylations promote NFAT nuclear export and reduce DNA binding affinity (16, 17, 18). Likewise, the MAPKs JNK1 and p38 promote NFAT nuclear export through phosphorylation of the SRR-1 domain and inhibit NFATc2-mediated activity in HeLa cells (20, 21, 22, 23). In parallel to regulating NFAT nuclear accumulation, some kinases such as Cot1, Pim1, and p38 may control NFAT transcriptional activity. The Pim1 kinase was shown to directly interact with and phosphorylate NFATc1 on serine residues of the regulatory domain that are not dephosphorylated by calcineurin (24). It is, however, unclear whether Pim1 phosphorylates the NFATc2 ⁵³SSPS⁵⁶ motif. In contrast, the Cot1 kinase cooperates with PKCζ to phosphorylate Ser⁵³ and Ser⁵⁶ of human NFATc2 (25, 26). Apart from its effect on NFAT nuclear export in HeLa cells, the p38 MAPK was also reported to regulate NFATc2 transcriptional activity in the CD8 T cell hybridoma through the phosphorylation of Ser⁵⁴ (27). Although these different studies support the contention that NFAT phosphorylation may regulate its transcriptional activity in different cell lines, a critical but still unresolved issue is whether and how these kinases may orchestrate NFAT-mediated transcription during naive T cell activation.

In this study we examined the regulation of NFAT during naive CD8 T activation. Surprisingly, we found that Ag stimulation did not induce significant NFAT transcriptional activity in naive CD8 T cells. Defective NFAT activity was not due to impaired nuclear accumulation but instead to impaired phosphorylation of the ⁵³SSPS⁵⁶ motif in the NFAT transactivation domain. In sharp contrast, TCR stimulation induces in CD4 T cells phosphorylation of this regulatory motif and consequently NFAT transcriptional activity. These studies therefore reveal the critical role of the ⁵³SSPS⁵⁶ motif in regulating NFAT-mediated gene expression in primary T cells. They further disclose one important signaling pathway that is induced specifically in CD4 but not CD8 T cells, thus providing a mechanism that explains some of the functional differences between these two subsets.

The Désiré (Des) (28) and P14 (29) mice express a transgenic TCR specific for the alloantigen H-2K^b and the lymphocytic choriomeningitis virus epitope presented by H-2K^b, respectively. The AP1- and NFAT-luciferase (NFAT-luc) transgenic mice express the luciferase gene controlled by AP1 and NFAT regulatory sequences, respectively (30, 31). The different transgenic mice were maintained on a C57BL6 or B10.BR background.

Lymph node CD4 and CD8 T cells were purified by negative selection using a mixture of Abs containing RA36B2 (anti-B220), M1/70.15.11.5HL (anti-CD11b), 24G2 (anti-FcRII/III), PK136 (anti-NK1.1), M5/114.15.2 (pan anti-MHC class II), and either H59.101.2 (anti-CD8) or H129.19.6 (anti-CD4) followed by incubation with sheep anti-rat IgG and sheep anti-mouse IgG magnetic beads (Dynal; Biotech). After magnetic depletion, the selected population was >95% enriched in CD4 or CD8 T cells. Purified T cells were stimulated with 1 μg/ml soluble anti-CD3 (145.2C11) and anti-CD28 (37.51) Abs with or without 1 ng/ml PMA or 400 ng/ml ionomycin in the presence of 2 × 10⁶ irradiated, T cell-depleted, syngeneic splenocytes as APCs. For Western blotting experiments and immunofluorescence staining, cells were stimulated with 10 μg/ml plate-bound anti-CD3 and 1 μg/ml soluble anti-CD28. The Des and P14 transgenic T cells were stimulated with 2 × 10⁶ irradiated B6 APCs alone or together with the gp33 aa 33–41 peptide.

The different NFATc2 forms were subcloned into the bicistronic pMX-enhanced GFP retroviral vector that contains an internal ribosome entry site with an enhanced GFP (32). Retroviral particles were generated using the PlatE packaging cell line (33). Purified CD8 T cells were stimulated for 20–22 h with soluble anti-CD3 and anti-CD28 Abs as described above and then infected with retrovirus-containing supernatants as previously described (34). Infected cells were maintained in IL-2 (50 U/ml)-containing medium. Where indicated, transduced GFP⁺ and control GFP⁻ cells were FACS-sorted at 24 h postinfection based on GFP expression.

Luciferase activity was developed using a luciferase reagent (Promega) as described (35). All measurements were done in duplicate and correspond to the luciferase activity of 5 × 10⁵ activated cells integrated for 10 s. Experimental values expressed as relative luminescence units were calculated by subtracting the value of unstimulated samples from each sample. IL-2 production was measured by bioassay using the CTL.L-2 cell line.

Whole cell extracts or cytoplasmic and nuclear extracts were prepared, of which 8–10 μg were analyzed by Western blotting as previously described (35, 36). Abs used were the NFATc2-specific Ab 25A10.D6.D2, the NFATc1-specific Ab 7A6, the phospho-Ser⁵⁴-NFATc2-specific Ab from Affinity Bioreagents, the p38 and phospho-p38 Abs (Cell Signaling Technology), and anti-hemagglutinin (HA) and anti-actin Abs. Blots were developed with a HRP-conjugated anti-rabbit or anti-mouse Ab and the enhanced chemiluminescent system (Sigma-Aldrich).

For gel shift assays, 4 μg of nuclear extracts were incubated with 10⁵ cpm of ³²P end-labeled, double-stranded NFAT oligonucleotides derived from the distal NFAT-AP1 composite site of the mouse IL2 promoter (5′-GCCCAAAGAGGAAAATTTGTTTCATACAG-3′). A consensus AP1 probe was used as a cold competitor (5′-GTCGACGTGAGCGCGC-3′). The binding reaction was conducted in 20 mM HEPES (pH 7.6), 100 mM NaCl, 10 mM KCl, 1 mM EDTA, 1 mM EGTA, and 0.03 μg/μl polydeoxyinosinic-polydeoxycytidylic acid.

Cells were incubated onto poly-l-lysine-coated microscope slides, fixed in 4% paraformaldehyde for 20 min at room temperature, and permeabilized with ice-cold methanol for 20 min at −20°C. Staining was performed in PBS supplemented with 2% BSA using anti-NFATc2 or anti-NFATc1 Abs and anti-mouse IgG1-Alexa Fluor 555 (Invitrogen). Nuclei were stained with TO-PRO-3 (Invitrogen) and the samples were postfixed in 4% paraformaldehyde. The z-scan sections were acquired with a Zeiss LSM 510 confocal microscope. Images were analyzed and quantified using NIH ImageJ software. In brief, naive unactivated T cells were used to define a threshold above background nuclear staining using ImageJ software, and the fraction of cells showing nuclear staining above this threshold was then calculated by dividing the number of cells with nuclear staining by the total number of cells analyzed. A total of at least 50–100 cells in more than five fields were analyzed in each experiment.

Total RNA was extracted using the High Pure RNA isolation kit (Qiagen) and cDNA was synthesized as previously described (37). Real-time PCR was performed on cDNA samples using the SYBR Green JumpStart system (Sigma-Aldrich). The primer sequences and cycling conditions are available upon request. The relative quantitation value is expressed as 2^−ΔC_T, where ΔC_T is the difference between the mean C_T value of duplicates of the sample and of the endogenous hypoxanthine phosphoribosyltransferase control (where C_T is cycle threshold).

Although NFAT has been extensively studied in CD4 T cells, the regulation and function of this transcription factor in CD8 T cells remain unclear. We examined NFAT-mediated transcription in response to Ag-specific stimulation in CD8 T cells using the previously described NFAT-luc reporter transgenic mice (31) crossed with Des TCR transgenic mice expressing an alloreactive, H-2K^b-specific TCR (Des-NFAT). Luciferase activity was read at different time points following the activation of CD8 T cells from Des-NFAT mice with H-2K^b-expressing APC. No significant NFAT transcriptional activity was observed at any of the periods of time examined (Fig. 1,A). To determine whether the lack of NFAT activity observed for Des-CD8 T cells may result from the low affinity of this TCR for its cognate ligand, we tested the response of CD8 T cells from P14 TCR transgenic mice expressing a high-affinity TCR for H-2K^b/gp33 complexes (29). Thus, NFAT-luc mice were crossed with P14 TCR transgenic mice and CD8 T cells from these mice activated with different concentrations of gp33 peptide and APCs. Marginal, close to undetectable luciferase activity could only be observed in response to a high gp33 peptide concentration (Fig. 1 B). Similar results were observed when LPS-stimulated APCs were used (data not shown), indicating that low costimulatory signals are not responsible for the lack of NFAT-mediated transcription observed in Ag-stimulated CD8 T cells.

We further examined NFAT transcriptional activity in CD8 T cells in response to polyclonal TCR stimulation. CD8 T cells and, as control, CD4 T cells from NFAT-luc mice were stimulated with anti-CD3 and anti-CD28 (anti-CD3/28) Abs. In agreement with previous reports, such stimulation induced significant levels of NFAT-mediated activity in CD4 T cells (Fig. 1,C and Refs. 38 and 39). In CD8 T cells, however, such stimulation did not induce detectable NFAT transcriptional activity, confirming our preceding observation with Ag stimulation (Fig. 1,C). To ensure that TCR triggering can induce in CD8 T cells transcription by other factors, we examined AP1 transcriptional activity in response to anti-CD3/28 Ab stimulation using CD8 T cells from the previously described AP1 luciferase reporter transgenic mice (30). Comparable levels of AP1 activity were induced in both CD8 and CD4 T cells (Fig. 1 D). Thus, in CD8 T cells, TCR-mediated signals can induce transcription mediated by AP1 but not by NFAT.

To determine whether TCR-independent stimuli can induce NFAT activity in CD8 T cells, we activated CD8 T cells and, as control, CD4 T cells with PMA and the calcium ionophore ionomycin. Interestingly, the combination of PMA and ionomycin induced NFAT transcriptional activity in CD8 T cells that was even higher than the activity induced in CD4 T cells (Fig. 1,E). We therefore examined which of these two stimuli could complement the signals triggered by TCR and CD28 engagement to induce NFAT transactivation in CD8 T cells. Thus, CD8 T cells were isolated from NFAT-luc mice and activated with anti-CD3 and anti-CD28 Abs alone or in combination with either ionomycin or PMA. The addition of ionomycin had no effect on NFAT activity induced by anti-CD3 and anti-CD28 Abs stimulation (Fig. 1,F). The addition of PMA, however, substantially enhanced NFAT activity (Fig. 1,F). As expected, the NFAT activity detected under these conditions of stimulation was inhibited by cyclosporin A (Fig. 1 G). No NFAT transcriptional activity was detected in CD8 T cells stimulated with PMA alone or with ionomycin (not shown).

Collectively these results indicate that, in CD8 T cells, TCR ligation through either Ag or a polyclonal stimulation is unable to induce NFAT-mediated transcription and that this defect is not due to impaired calcium mobilization but rather through an impaired pathway that can be compensated by the presence of PMA.

To determine whether impaired NFAT transcriptional activity in TCR-stimulated CD8 T cells may result from defective nuclear translocation, we examined the regulation of NFATc2 and NFATc1 subcellular localization by anti-CD3 and anti-CD28 Abs stimulation in the presence or absence of PMA. Both NFATc2 and NFATc1 were detected in the nucleus of anti-CD3/28 Abs-stimulated T cells with a kinetic that was comparable whether PMA was present or not during initial activation (Fig. 2,A). Immunofluorescence confocal microscopy confirmed these results and further showed that the kinetic and the level of nuclear accumulation of NFATc2 are comparable in anti-CD3/28 Ab-stimulated CD8 and CD4 T cells (Fig. 2, B and C). In addition, regardless of the activation condition and cell type, maximal nuclear accumulation of NFATc2 was achieved by 4 h after the initial activation and sustained thereafter (Fig. 2, B and C). In agreement with published studies, we did not detect significant nuclear translocation of NFATc1 4 h after anti-CD3/28 Abs stimulation of either CD4 or CD8 T cells even in the presence of PMA (Fig. 2, A and C, and Ref. 40). At 12 h postactivation, although significant, the frequency of cells showing nuclear NFATc1 remained lower than that observed for NFATc2 but comparable in CD4 and CD8 T cells stimulated with anti-CD3/28 Abs alone or in the presence of PMA (Fig. 2,C). Similarly, we showed by gel shift experiments that NFAT-DNA complex formation was comparable in CD8 T cells stimulated with anti-CD3/28 Abs alone or in combination with PMA and in similarly stimulated CD4 T cells (Fig. 2 D). Collectively, these different studies show that TCR-stimulated CD8 T cells do not show defective NFAT nuclear translocation.

To further demonstrate that the inability of NFAT to mediate transcription in CD8 T cells was not due to inefficient nuclear accumulation, we overexpressed in CD8 T cells the previously described NFATc2 mutant with 12 serine-to-alanine substitutions in the conserved sequence motifs of the regulatory domain (Nuc-NFATc2) that mimics an almost completely dephosphorylated form of NFATc2 (14). Mutant Nuc-NFATc2 has a constitutive nuclear localization and is not subjected to known regulatory export kinases such as glycogen synthase kinase 3. CD8 T cells were activated in vitro with anti-CD3/28 Abs, transduced with a bicistronic GFP retroviral vector expressing either wild-type (WT) NFATc2 (WT-NFATc2), Nuc-NFATc2, or empty vector. Western blotting experiments showed that WT-NFATc2 was expressed at high levels in transduced CD8 T cells leading to a global increase in total NFATc2 expression level (Fig. 3,A). As expected, Nuc-NFATc2 was mainly nuclear but expressed at lower levels than WT-NFATc2 in transduced T cells (Fig. 3,A and not shown). Neither WT-NFATc2 nor Nuc-NFATc2 had a significant incidence on expression of endogenous NFATc1 (Fig. 3,A). Overexpression of constitutively nuclear NFATc2 did not restore NFAT transactivation in anti-CD3/28 Ab-stimulated CD8 T cells (Fig. 3,B). In addition, the synergistic effect of PMA with anti-CD3/28 Abs was observed even in the presence of Nuc-NFATc2 (Fig. 3 B).

Together, these results show that impaired dephosphorylation of the 12 phosphoserines that regulate NFATc2 nuclear accumulation and DNA binding does not account for the failure of NFAT to mediate transcription in CD8 T cells.

Several studies in different cell lines indicate that phosphorylation of the ⁵³SSPS⁵⁶ motif in the NFAT transactivation domain by PKC-dependent, calcineurin-independent kinase(s) is important for transcriptional activity of this transcription factor (14). We therefore determined whether the positive effect of PMA on NFAT-mediated transcription in anti-CD3/28 Ab-stimulated cells likewise results from phosphorylation of the ⁵³SSPS⁵⁶ motif in the NFATc2 transactivation domain. To test this possibility, we overexpressed in CD8 T cells a transcriptionally inactive NFATc2 molecule with serine-to-alanine substitutions in the ⁵³SSPS⁵⁶ motif (AAPA-NFATc2) and, as control, WT-NFATc2 (14). Transduced CD8 T cells express comparable levels of WT or mutant NFATc2 and comparable levels of endogenous NFATc1 (Fig. 3,A). To directly analyze the transcriptional activity of overexpressed NFAT molecules, we compared the luciferase activity of FACS-sorted GFP⁺ and GFP⁻ cells. As observed for WT-NFATc2 and Nuc-NFATc2, overexpression of AAPA-NFATc2 itself did not affect NFAT transactivation triggered by anti-CD3/28 Abs (data not shown). However, although the presence of PMA stimulated transcription from WT-NFATc2, it had minimal effect on AAPA-NFATc2-mediated transcription (Fig. 4 A). Indeed, the luciferase activity of GFP⁺ and GFP⁻ cells was, in this later case, comparable. These results indicate that PMA restores TCR-induced NFAT activity directly by promoting phosphorylation of the ⁵³SSPS⁵⁶ motif in the NFATc2 transactivation domain.

To further evaluate the role of ⁵³SSPS⁵⁶ motif in regulating NFAT-mediated transcription in activated CD4 and CD8 T cells, we directly examined phosphorylation on this motif by a Western blotting experiment using an Ab that specifically recognizes phosphorylated Ser⁵⁴ in the NFATc2 transactivation domain. As anticipated from our functional studies, activation of CD4 T cells with anti-CD3/28 Abs induces a rapid and sustained phosphorylation of NFATc2 Ser⁵⁴ that peaks at 6 h poststimulation (Fig. 4,B, lane 1). In sharp contrast, anti-CD3/28 Ab stimulation induces in CD8 T cells only minimal phosphorylation of NFATc2 Ser⁵⁴, which was significantly increased by the addition of PMA (Fig. 4 B, lanes 2 and 3).

Collectively, these results show that phosphorylation of the NFATc2 ⁵³SSPS⁵⁶ motif, which is critical for NFATc2-mediated transcription, is induced by TCR stimulation in CD4 but not in CD8 T cells.

The above results indicate that defective NFAT transactivation in Ag-stimulated CD8 T cells results from impaired phosphorylation of Ser⁵⁴ in the NFAT transactivation domain. Although the different kinases involved in phosphorylation of this motif are not yet well characterized, several studies suggest that Pim1 and Cot/Tpl2 may do so (24, 25, 41). Because these kinases are primarily regulated at the transcriptional level, we evaluated the expression levels of these two kinases in naive CD8 T cells, in CD8 T cells stimulated with anti-CD3/28 Abs alone or in the presence of PMA, and in stimulated CD4 T cells, but no significant difference was found (Fig. 5,A). A recent study has shown that p38 MAP kinase directly interacts with NFATc2 and phosphorylates it at Ser⁵⁴ (27). We therefore examined p38 MAPK activation in TCR-stimulated CD8 and CD4 T cells by Western blot analysis. Phosphorylation of p38 MAPK on Thr¹⁸⁰ and Tyr¹⁸² was clearly detectable in CD4 T cells stimulated with anti-CD3/28 Abs and slightly increased by PMA addition (Fig. 5,B, lanes 1 and 3). In CD8 T cells stimulated with anti-CD3/28 Abs, we observed only a modest phosphorylation of p38 MAPK on Thr¹⁸⁰ and Tyr¹⁸² that was significantly increased by PMA addition (Fig. 5,B, lanes 5 and 7). As the level of p38 MAPK correlated with the level of NFAT phosphorylation on Ser⁵⁴ in CD4 and CD8 T cells, we evaluated whether blocking p38 MAPK activity with SB203580 may prevent NFAT phosphorylation on Ser⁵⁴. Regardless of the activating stimuli, the addition of SB203580 almost completely blocked p38 phosphorylation on Thr¹⁸⁰ and Tyr¹⁸² in both CD4 and CD8 T cells (Fig. 5,B, lanes 2, 4, 6, and 8). Thus, in agreement with published studies, TCR stimulation primarily activates p38 MAPK by the alternative pathway (42). Blocking p38 activity with SB203580 had, however, no effect on NFAT phosphorylation in anti-CD3/28 Ab-stimulated CD4 and CD8 T cells (Fig. 5 C and data not shown). Importantly, under the experimental conditions used, SB203580 prevented phosphorylation of MAP-KAPK-2, a direct substrate of p38, further demonstrating efficient inhibition of the p38 MAPK (data not shown).

Collectively, these results show that TCR stimulates NFAT transcriptional activity through phosphorylation of the ⁵³SSPS⁵⁶ motif in the NFAT transactivation domain by a pathway that does not involve Pim1, Cot/Tpl2, or p38 MAPK. Although this pathway is strongly induced in CD4 T cells upon TCR ligation, it is only mildly induced in CD8 T cells.

We further determined whether the difference in NFAT activity between CD4 and CD8 T cells may explain part of the different behavior of these two subsets in response to TCR stimulation. In contrast to naive CD4 T cells, naive CD8 T cells produce a limited amount of IL-2 upon Ag stimulation. Among different regulatory sequences, the IL-2 promoter contains NFAT binding sites that are critical for a high level of IL-2 production. We therefore determined whether the signals that increase NFAT activity may increase IL-2 production by activated CD8 T cells. In agreement with published data, anti-CD3/28 Ab-stimulated CD8 T cells produce low levels of IL-2, ∼8.9-fold less than similarly stimulated CD4 T cells (Fig. 6). The addition of PMA to anti-CD3/28 Ab stimulation, but not ionomycin, significantly increased the production of IL-2 by both CD8 and CD4 T cells. Importantly, under these activation conditions the level of IL-2 produced by CD8 T cells was comparable to that of CD4 T cells (CD4:CD8 ratio of 1.5). Thus, experimental conditions that lead to a significant increase in NFAT transcriptional activity in anti-CD3/28 Ab-stimulated CD8 T cells also increased IL-2 production.

The NFAT transcription factors play a critical role in the regulation of transcription during immune responses. Largely studied during CD4 T cell activation and differentiation, very little is known about NFAT regulation during CD8 T cell activation. We show in this study, that Ag stimulation does not stimulate NFAT-mediated transcription in naive CD8 T cells. We further show that impaired NFAT activity in TCR-stimulated CD8 T cells does not result from inefficient NFAT nuclear translocation or accumulation but instead from defective phosphorylation of the ⁵³SSPS⁵⁶ motif of the NFAT transactivation domain. These results reveal the critical role of this motif in regulating NFAT transcriptional activity in primary T cells. In addition, we show in this study that TCR engagement induces massive phosphorylation of Ser⁵⁴ in the NFAT transactivation domain in CD4 T cells whereas it does so only minimally in CD8 T cells. Hence, our study discloses an unsuspected difference between CD4 and CD8 T cells in the TCR downstream signaling pathway, thus providing a mechanism that explains some of the functional differences of these two subsets.

Our study points to an essential role of the ⁵³SSPS⁵⁶ motif in regulating NFATc2 transcriptional activity in primary T cells. Mouse NFATc1 lacks this regulatory motif and could therefore contribute to NFAT-mediated transcription in Ag-stimulated CD8 T cells. In agreement with published studies (40), we show in the current study that TCR stimulation induces only limited nuclear translocation of NFATc1. Delayed and reduced nuclear accumulation explains why NFATc1 does not compensate for defective NFATc2-mediated transcription in Ag-stimulated CD8 T cells.

The kinase or kinases that may phosphorylate the ⁵³SSPS⁵⁶ motif of NFAT transactivation domain in T cells are not well characterized. Studies in cell lines have suggested that Pim1, Cot/Tpl2, and p38 MAPK regulate NFAT transcriptional activity, possibly through phosphorylation of this regulatory motif (24, 25, 26, 27). We show in this study that neither of these kinases is responsible for the failure of NFAT to mediate transcription in primary CD8 T cells. These results contrast with work from Round and colleagues who provided indirect experimental support for p38 MAPK-dependent phosphorylation NFATc2 on Ser⁵⁴ (27). This pathway was, however, only demonstrated for multiply activated CD8 T cells and for CD8 T cell hybridomas, whereas our study focuses on the response of naive CD8 T cells. It is therefore possible that naive and effector CD8 T cells use different pathways to phosphorylate Ser⁵⁴ in the NFAT transactivation domain.

We found that, in CD8 T cells, PMA can compensate for defective phosphorylation of the ⁵³SSPS⁵⁶ motif of NFAT transactivation domain and NFAT-mediated transcription induced by anti-CD3/28 Ab stimulation. This observation suggests that, in CD8 T cells, TCR engagement may induce suboptimal PKC activation. Activation of the ERK and JNK MAPKs (data not shown) and AP1 transcriptional activity were comparable in anti-CD3/28 Ab-stimulated CD4 and CD8 T cells, suggesting that PKC activation is not significantly impaired in TCR-stimulated CD8 T cells. In addition, anti-CD3/28 Ab stimulation induces in CD8 T cells recruitment of PKCθ and phospho-PKCθ, the PKC isoform critically involved in T cell activation, at the immunological synapse, further suggesting that activation of the PKC signaling pathway is not impaired in TCR-stimulated CD8 T cells (43). It remains possible, however, that other PMA-dependent conventional or novel PKC isotypes are activated upon TCR engagement in CD4 but not in CD8 T cells and regulate phosphorylation of the NFAT transactivation domain. Alternatively, PMA may trigger an alternative pathway that directly or indirectly may phosphorylate the NFAT transactivation domain. Further characterization of the signaling pathway and kinase leading to phosphorylation of the NFAT ⁵³SSPS⁵⁶ motif will be required to resolve this issue.

Although phosphorylation of the ⁵³SSPS⁵⁶ motif of the NFATc2 transactivation domain is critical for NFATc2-mediated transcription (Ref. 14 and this study), how it may do so remains unclear. Expression of a gene relies on the multiprotein assemblage of different transcription factors and chromatin modifiers at the promoter sequence and ultimately coupling with the polymerase complex. Phosphorylation of transcription factors may modify the composition of an enhanceosome by favoring or inhibiting the recruitment of different partners, themselves having positive or negative effects on transcription. Multicomplex formation therefore allows for fine tuning of gene expression. To examine the biological significance of NFAT phosphorylation, we analyzed the expression levels of different cytokines in CD8 T cells stimulated under conditions inducing or not inducing NFAT phosphorylation. NFATc2 is critically required for both IL-2 and IFN-γ production by CD8 T cells, as expression of these cytokines is impaired in NFATc2-deficient mice (7, 10, 44). We show in this study that phosphorylation of the ⁵³SSPS⁵⁶ motif strongly enhances IL-2 gene expression but has minimal effect on IFN-γ expression (not shown), which is expressed at high levels by anti-CD3/28 Ab-stimulated CD8 T cells (Fig. 1, data not shown, and Ref. 10). Collectively, these different observations strongly support a model in which phosphorylation of the ⁵³SSPS⁵⁶ motif is essential for proper organization of the transcriptional complex at the IL-2 promoter, but dispensable at the IFN-γ promoter.

Following Ag stimulation, CD4 and CD8 T cells will undergo an autonomous program of division and differentiation, leading to effector and memory cells. Although they share several similarities, the responses of CD4 and CD8 T cells also present several important differences. As discussed above, the difference in IL-2 production between Ag-stimulated CD4 and CD8 T cells may result from impaired phosphorylation of the NFAT ⁵³SSPS⁵⁶ motif in CD8 T cells. In addition, CD8 T cells divide faster and sooner after Ag stimulation than do CD4 T cells. Interestingly, NFATc2-deficient T cells show increase cell cycle entry and progression, a phenotype that correlates with increased expression of cyclin A2, E, and F and cyclin-dependent kinase 4 (5, 8, 9, 44). Thus, NFATc2 represses, by yet unknown mechanisms, cyclins and cyclin-dependent kinase 4 gene expression (3, 9). Further analysis of cell cycle progression will be required to determine whether a difference in NFATc2 phosphorylation contributes to increased proliferation of CD8 T cells as compared with CD4 T cells. Nonetheless, our study reveals an unsuspected signaling pathway that explains some of the functional differences between CD4 and CD8 T cell responses to Ag stimulation.

We thank A. Rao for the different NFATc2 constructs, N. Arai for the pMX-EGFP retroviral vector, and T. Kitamura for the PlatE cells. We thank A-M Schmidt-Verhuslt for a helpful discussion and A. Rao for critical reading of the manuscript.

The authors have no financial conflict of interest.