Transcriptional regulation of Fas gene expression by GA-binding protein and AP-1 in T cell antigen receptor.CD3 complex-stimulated T cells.

Fas (CD95 or APO-1), a transmembrane cell surface receptor of the tumor necrosis factor receptor family, is up-regulated in activated T lymphocytes. Our present study identified an upstream enhancer element (between nucleotide positions -862 and -682) containing a GA-binding protein (GABP) site and a low affinity activating protein-1 (AP-1)-binding site. T cell activation increased the DNA binding of GABP and AP-1 to this enhancer site. The specificity of GABP and AP-1 binding was demonstrated by competition electrophoretic mobility shift assay and supershift electrophoretic mobility shift assay with antibodies against GABP and AP-1, respectively. Mutational analysis of Fas promoter revealed that both GABP- and AP-1-binding sites were required for initiating Fas gene transcription. We further show that anti-CD3 mAb, phorbol 12-myristate 13-acetate, and phorbol 12-myristate 13-acetate/ionomycin strongly activated promoters carrying multiple copies of the Fas enhancer, and mutation of either the GABP or AP-1 binding site severely reduced transcriptional activity. Taken together, these results suggest that the transcription factors GABP and AP-1 play a critical role in the induction of Fas gene expression in T cell antigen receptor.CD3-stimulated Jurkat cells.

Fas (CD95 or APO-1), a type I transmembrane surface receptor of the tumor necrosis factor receptor family, is ubiquitously expressed in a variety of tissues and with particular abundance in the thymus, liver, and kidney. In contrast, peripheral naive T cells (CD45RA ϩ ) express little or no cell surface Fas, whereas activated memory T cells (CD45RO ϩ ) and cultured T cells express relatively high levels of cell surface Fas (1). Stimulation of T cell antigen receptor (TCR) 1 ⅐CD3 complex by antibody cross-linking, peptide antigen, or superantigen up-regulates the expression of both Fas and its ligand (FasL) (2)(3)(4). Fas binding to FasL activates a cascade of proteases known as caspases that subsequently induce activation-induced cell death (apoptosis). Fas-mediated apoptosis is used not only as an effector mechanisms by T cells to induce apoptosis of Fas expressing cells but is also implicated in clonal downsizing of the immune cell population. Fas-deficient MRLlpr/lpr mice and FasL-deficient gld mice develop a lymphoproliferative and autoimmune disease, indicating that Fas plays a critical role in immune homeostasis (5,6).
Several downstream targets of these MAP kinases have been identified that could be involved in Fas gene expression, including c-Jun, c-Fos, ATF-2, Elk, and Sap-1 (22,(27)(28)(29)(30)(31)(32). The homoor heterodimerization of Fos and Jun forms the AP-1 complex, a transcription factor that is intimately involved in T cell activation (9,10). The activity of AP-1 is regulated both at the level of jun/fos gene transcription and post-translational modifications of the Jun/Fos heterodimer. The AP-1 complex by itself can bind directly to AP-1-responsive elements in the promoters of many genes, such as the IL-2 promoter. In addition, AP-1 from heterodimers with transcription factors of the CREB-ATF family, resulting in distinct DNA binding specificities (9,10,33).
Recent studies have identified the GA-binding protein (GABP) as a target of MAP kinases (34,35). GABP, an Ets-like transcription factor, is constitutively expressed in many cell types. GABP is a heterodimer composed of ␣ and ␤ subunit (36,37). The 52-kDa ␣ subunit shows significant homology to the DNA binding domains of transcription factors of the Ets family and facilitates weak binding to DNA (38). The 42-kDa ␤ subunit of GABP contains four repeats of a sequence present in several transmembrane proteins, including Drosophila Notch and ankyrin. The N-terminal repeats of the GABP ␤ subunit do not bind DNA but can interact with and stabilize the binding of the ␣ subunit (39). GABP has been implicated in regulating the expression of several viral genes, including the ICP-4 gene of human herpes simplex virus-1 (37,40), E1A gene of adenovirus (40), and the long terminal repeat of human immunodeficiency virus (34). GABP also controls the basal expression of the TATA-less cytochrome c-oxidase subunit IV gene promoter (41). Recent studies have identified a dyad symmetry element in the distal enhancer of the IL-2 promoter, which is transactivated by GABP (35). GABP is predominantly phosphorylated and activated by JNK and weakly activated by ERK. Blocking JNK pathway by the expression of dominant-negative (DN) SEK and SAPK impairs expression of IL-2, whereas blocking both JNK and ERK pathways completely abrogates IL-2 gene expression (35).
To characterize Fas promoter activation in T cells, we investigated the signaling pathways and transcription factors required for TCR⅐CD3-stimulated Fas gene expression in human Jurkat T cell line. We report here that a transcriptional regulatory element containing two ERE binding sites and an AP-1 binding site was located between the nucleotide positions from Ϫ866 to Ϫ682. Mutation of the GABP-and AP-1-binding sites markedly impaired enhancer function in TCR⅐CD3-, PMA-, and PMA/ionomycin-stimulated Jurkat cells, indicating that GABP and AP-1 play a critical role in regulating Fas gene expression in the TCR⅐CD3-activated T cells.
Transfection-Jurkat cells (1 ϫ 10 6 /sample) were transfected with various plasmids (2 g/sample) by using DMRIE-C (2 l/sample) following the manufacturer's instructions. Plasmid pCMV/SPORT (0.5 g/sample) carrying the gene of ␤-galactosidase was co-transfected as internal control. In some experiments, protein concentrations in cell lysates were measured to confirm that an equal number of cells had been harvested. Following transfection, cells were immediately stimulated with plastic-coated anti-CD3 mAb, OKT3 (10 g/ml in phosphatebuffered saline, 4°C overnight), PMA (50 ng/ml), and/or ionomycin (0.5 M). After incubation for 24 h at 37°C in 5% CO 2 , cells were harvested, and luciferase activity was quantitated by using luciferase substrate kit (Promega, Madison, WI) and reading in a Packard luminometer (Packard Instrument Company, Inc., Meriden, CT).
Preparation of Nuclear Extracts and EMSA-Jurkat cells (1 ϫ 10 7 / ml) were stimulated with PMA (50 ng/ml) and/or ionomycin (1 M) at 37°C for 30 min or with 4 g/ml of anti-CD3 mAb (OKT3) for 2 min followed by cross-linking with goat anti-mouse IgG (16 g/ml) for 30 min. Nuclear extracts were prepared following a standard protocol as described previously (34) Double-stranded oligonucleotides were endlabeled with T4 polynucleotide kinase, and the labeled probes were separated by NucTRap probe purification columns. The sequences of AP-1 and NF-B probes are CGCTTGATGAGTCAGCCGGAA and AGT-TGAGGGGACTTTCCCAGGC, respectively. Labeled oligonucleotide (20,000 -50,000 cpm) was incubated at room temperature for 20 min with 2 g of protein of nuclear extract. The DNA binding buffer used in all experiments except in Fig. 8B contains 4% glycerol, 1 mM MgCl 2 , 0.5 mM EDTA, 0.5 mM dithiothreitol, 50 mM NaCl, 10 mM Tris-HCl, pH 7.5, 50 g/ml poly(dI-dC)⅐poly(dI-dC). The buffer used in Fig. 8B contains less poly(dI-dC)⅐poly(dI-dC) (10 g/ml) and more dithiothreitol (5 mM). For supershift or competition EMSA, nuclear extract (2-5 g/reaction) was preincubated with 1 g of antibodies against AP-1 or 5-200X unlabeled oligonucleotide at 4°C for 30 min, and then hot probes were added. The reaction was separated on 5% nondenaturing polyacrylamide gel. Dried gel was exposed to X-Omat film or analyzed in a STORM 860 Image System (Molecular Dynamics).
Plasmid Constructs-Human Fas promoter containing 1007 bp upstream of the translation initiation site was PCR-amplified with a forward oligonucleotide GGCGGAGGTACCAGTAATGATGTCATTAT-CCAAAC (the KpnI cleavage site is in boldface type) and a reverse oligonucleotide GTTCCGAAGCTTGGTTGTTGAGCAATCCTCCGA-AGT (the HindIII cleavage site is in boldface type). The genomic DNA extracted from peripheral blood lymphocytes of healthy donor was used as template. PCR reaction was conducted with 2.5 units of proofreading Pfu DNA polymerase/reaction for 30 cycles with the following parameters: 94°C denature for 1 min, 48°C annealing for 1 min, and 72°C extension for 2 min. PCR product was extracted with phenol and then digested by HindIII and KpnI. The resultant fragment was extracted and ligated to pGL3/Basic that had been digested with KpnI and Hin-dIII. The ligation reaction was used to transform competent DH5␣. Plasmid DNA was purified using QIAGEN Midi Kit and sequenced before transfection. We further generated two constructs with truncated Fas promoter, pGL3/Fas-866, pGL3/Fas-682, and pGL3/Fas-425, which contains the 866, 682, and 425 bp upstream of the translation initiation ATG site. Three forward oligonucleotides with the sequences 5Ј-TGGCCAGGAAATAATGAGTAACGAAGGACAGGAAGTAATTGT-3Ј, 5Ј-GCCATTCCAGAAACGTCTGTG-3Ј and 5Ј-CTGCAGGAACGCC-CCGGGACAC-3Ј, and a universal reverse oligonucleotide with the sequence GTTCCGAAGCTTGGTTGTTGAGCAATCCTCCGAAGT (the HindIII cleavage site is in boldface type) were synthesized and used in PCR reaction for construction of pGL3/Fas-866, pGL3/Fas-682, and pGL3/Fas-425, respectively. pGL3/Fas-1007 plasmid DNA was used as the template. PCR reaction was performed according to the following parameters: 94°C denature for 1 min, 48°C annealing for 1 min, and 72°C extension for 2 min, 30 cycles. PCR product was extracted by phenol, digested with HindIII, and then ligated to pGL3/Basic plasmid that had been digested with SmaI and HindIII. The truncation was confirmed by restriction enzyme digestion and sequencing.
pGL3/2xGABP was constructed by inserting a synthesized doublestranded DNA with a cleaved KpnI site at one end and a NheI site at another end into pGL3/promoter plasmid. The sequences of two complementary strands are 5Ј-cTGGCCAGGAAATAATGAGTAACGAAGG-ACAGGAAGTAAGTGGCCAGGAAATAATGAGTAACGAAGGACAGG-AAGTAAg-3Ј and 5Ј-ctagcTTACTTCCTGTCCTTCGTTACTCATTATT-TCCTGGCCATTACTTCCTGTCCTTCGTTACTCATTATTTCCTGGC-CAGgtac-3Ј. pGL3/2xGABP/M1 carries mutation at all ERE sites, whereas pGL3/2xGABP/M2 has mutation at two AP-1 sites. The annealed double-stranded DNA was directly ligated into pGL3/promoter plasmid that had been digested with NheI and KpnI. The ligation reaction was used to transform competent DH5␣ Escherichia coli. Plasmid DNA was purified and sequenced. To generate pGL3/4xGABP, a two-copy GABP insert containing a cleaved NheI site at one end and a cleaved BglII site at another site was inserted into the pGL3/2xGABP that had been digested with NheI and BglII. Identification of the Signaling Pathway Required for Anti-CD3 mAb Activation of the Fas Promoter-Several inhibitors of the TCR⅐CD3-mediated signaling pathways were used to determine which pathway was required for anti-CD3 mAb-induced Fas promoter activity. As shown in Fig. 2, cyclosporin A, a specific inhibitor of the NF-AT pathway, and wortmannin, a specific inhibitor of PI 3-kinase, did not inhibit anti-CD3 stimulated luciferase activity, suggesting that anti-CD3-stimulated Fas expression is not mediated by the transcription factor NF-AT or by the PI 3-kinase pathway. PD98059 (50 M), a specific inhibitor of ERK pathway, and SB202190 (2 M), a specific inhibitor of the p38 kinase pathway, did not affect anti-CD3 mAb-stimulated luciferase expression. H7, an inhibitor of PKC, modestly inhibited luciferase activity by 62.6%, as calculated after the subtraction of luciferase activity in unstimulated Jurkat cells. These data, in combination with recent finding that H7 inhibits anti-CD3 and PMA-induced Fas mRNA expression (45), suggest that activation of the Fas promoter via the T cell receptor complex is largely mediated through the PKC activation pathway.

Identification of a Fas Enhancer
Identification of GABP and AP-1 Binding Sites in the Fas Enhancer-Activated JNK and ERK kinases phosphorylate and activate several downstream substrates, such as c-Fos, c-Jun, and GABP. Recent studies by Rapp and his colleagues (35,46) demonstrate that GABP is able to bind the distal enhancer region of the IL-2 gene and activate gene transcription. Sequence analysis of the Fas enhancer revealed a putative GABP binding site (Ϫ863 to Ϫ820) consisting of two ERE repeats with sequences identical to those of the GABP binding site in the IL-2 promoter (Fig. 3A). Interestingly, a putative AP-1 binding site with the sequence TGAGTAA is located between two EREs. We assessed whether stimulation of Jurkat cells with anti-CD3 mAb as well as PMA and/or ionomycin was able to activate GABP and/or AP-1 to bind the 32 P-labeled Fas/GABP probe by EMSA. As shown in Fig. 3B, the nuclear extract prepared from unstimulated Jurkat cells generated four gel shift complexes with the 32 P-labeled oligonucleotides spanning the region from nucleotide positions Ϫ863 to Ϫ820 of the Fas promoter. Stimulation of Jurkat cells with anti-CD3 antibody, PMA, or PMA plus ionomycin but not ionomycin alone increased the binding activity on each complex.
We next examined the effect of mutations of the GABP and AP-1 binding sites (Ϫ820 to Ϫ863) on the binding complexes. As shown in Fig. 4, mutation of both ERE sites (M1) markedly decreased the formation of the C1-C3 complexes, whereas mutation of the putative AP-1 site (M2) essentially had no effect on the formation of the C1-C4 complexes. Deletion of either ERE-A or ERE-B site abolished the formation of the C4 complex and significantly reduced the formation of the C2 and C3 complexes. It appears that the ERE-B oligonucleotide had slightly higher affinity for GABP to form the C1 complex than the ERE-A oligonucleotide. In addition, no detectable complex was formed with the F/AP-1 oligonucleotide. These results suggest that both ERE sites can bind GABP, whereas the putative AP-1 site, although its sequence falls into the definition of the consensus AP-1 binding site TGA(G/C)T(C/A)A (47), cannot bind AP-1 at detectable levels under the experimental conditions we used here.
Specificity of GABP Binding to the Fas Enhancer-To con-firm the specificity of GABP in the complexes formed with the 32 P-labeled Fas/GABP oligonucleotide, competition EMSA was conducted using the oligonucleotides listed in Fig. 4A to compete with the 32 P-labeled Fas/GABP (Fig. 5A), ERE-A (Fig. 5B), and ERE-B (Fig. 5C) oligonucleotides. As shown in Fig. 5A, the M1 oligonucleotide, which has mutations in both ERE sites, did not affect the formation of the C1-C3 complexes but did inhibit the formation of the C4 complex. The M2 oligonucleotide, which has mutations in the putative AP-1 binding site, reduced the formation of all the C1-C4 complexes, whereas both unlabeled ERE-A and ERE-B oligonucleotides reduced the formation of the C1-C3 complexes but had no effect on the formation of the C4 complex. The F/AP-1 oligonucleotide, which contains the AP-1 binding site, did not affect the formation of any of the complexes. The consensus NF-B site, included as negative control, also did not affect the formation of any of the complexes. Consistently, the Fas/GABP, M2, ERE-A, and ERE-B oligonucleotides were able to compete with the ERE-A (Fig. 5B) and ERE-B (Fig. 5C) oligonucleotides to bind GABP, whereas M1, F/AP-1, and the consensus NF-B probes had no effect on GABP binding to the ERE-A and ERE-B probes.
Characterization of the C1-C4 Complexes-To determine the composition of the C1-C4 complexes, we conducted supershift EMSA with antibodies against the ␣ and ␤ subunits of GABP to determine whether the gel shift complexes contained GABP. In these supershift experiments, the antibodies were preincubated with the nuclear extracts. Under this circumstance, the antibody binding to transcription factor may prevent the DNA/ protein interaction, this type of supershift may sometimes appear as an inhibition EMSA. As shown in Fig. 6A, both antibodies inhibited the formation of the C1-C3 complexes, suggesting that the C1-C3 complexes contain both the ␣ and ␤ subunits. The ␣ subunit of GABP by itself is able to bind the GGAA sequence but with low affinity; our supershift analyses show that the C1-C3 complexes were supershifted by both anti-GABP␣ and anti-GABP␤ antibody, indicating that GABP␣ alone is unable to form the complex with either ERE site of the Fas enhancer. It is noteworthy that removal of GABP␣ and GABP␤ by addition of specific antibodies dramatically increased the formation of the C4 complex, indicating that the nuclear protein in the C4 complex may bind the sequence overlapping with the ERE-A or ERE-B site. We also performed ml) were treated with anti-CD3 mAb (OKT3, 4 g/sample) for 2 min followed by cross-linking with goat anti-mouse IgG (16 g) for 30 min or treated with PMA (50 ng/ml) and/or ionomycin (1 M) for 30 min. Nulcear extracts were prepared, and protein concentrations were measured. A doublestranded oligonucleotide encompassing the sequence of Fas promoter from positions Ϫ863 to Ϫ820 was chemically synthesized. The sequence of this oligonucleotide, 5Ј-TGGCCAGGAAATAATGAGTAACGAAG-GACAGGAAGTAATTGT-3Ј, contains two putative ERE sites (underlined) and a putative AP-1 binding site (boldfaced). The 32 P end-labeled probe (30,000 cpm/sample) was incubated with 2 g of protein of nuclear extract at room temperature for 20 min, and the reactions were then separated on a 5% nondenaturing polyacrylamide gel. Dried gel was exposed to X-Omat film overnight.

FIG. 4. Binding of nuclear proteins from Jurkat cells to the mutated or truncated Fas/GABP probe.
A, sequence of the oligonucleotides used for EMSA. B, double-stranded oligonucleotides were endlabeled with [␥-32 P]ATP, and probe (30,000 cpm) was incubated at room temperature with 2 g of nuclear proteins prepared from anti-CD3stimulated Jurkat cells. The reactions were then separated on a 5% nondenaturing polyacrylamide gel. Dried gel was exposed to X-Omat film overnight.
supershift analysis with the ERE-B oligonucleotide. As shown in Fig. 6B, the upper complex was supershifted by both anti-GABP␣ and GABP␤ antibodies, whereas a lower unidentified band was not affected by either antibody.
Functional Analysis of the Fas Enhancer-To test whether the GABP and putative AP-1 sites in the Fas enhancer play a role in initiating the transcription of Fas gene expression, Fas promoter constructs with mutation of the ERE and/or AP-1 sites in the enhancer (Fig. 7A) were assessed for their ability to induce the expression of the linked luciferase reporter gene. pGL3/Fas-866 reporter plasmid contains a 866-bp insert of the wild-type Fas promoter, pGL3/Fas-866/M1 has mutations in both ERE sites, pGL3/Fas-866/M2 has mutations in the putative AP-1 site, and pGL3/Fas-866/M3 has mutations at both the AP-1 and ERE sites. As shown in Fig. 7B, mutation of the GABP binding site markedly reduced the induction of luciferase activity in response to anti-CD3 mAb, PMA, or PMA/ionomycin. Surprisingly, mutation of the putative AP-1 site also significantly reduced the luciferase activity in activated Jurkat cells. Mutation of both the GABP and AP-1 sites did not further reduce luciferase activity, compared with mutation of the GABP site or AP-1 site alone. These results suggest that the GABP and putative AP-1 binding sites may cooperate to activate Fas gene expression in TCR⅐CD3-stimulated T cells.  Fig.  7 revealed that the putative AP-1 site is indispensable for initiating the Fas gene transcription. We hypothesized that this putative AP-1 site may be able to bind AP-1, but its affinity is low. To assess this possibility, we performed competition EMSA using various oligonucleotides listed in Fig. 4A to com- , and ERE-B (C) oligonucleotide (30,000 cpm/sample) was added, and the reaction was allowed to incubate at room temperature for another 20 min. The DNAprotein interaction was resolved on a 5% native polyacrylamide gel, and dried gels were exposed to X-Omat films.
pete with 32 P-labeled consensus AP-1 probe. As shown in Fig.  8A, F/AP-1, F/GABP, and M1, all of which contain the intact Fas AP-1 binding site, competed with the consensus AP-1 site, whereas M2 and ERE-A, which contain mutated or deleted AP-1 sites, respectively, failed to compete with the consensus AP-1 site. In contrast, the consensus AP-1 site, included as positive control, strongly competed with itself. Similar results were obtained in a competition EMSA with less stringent binding conditions when the 32 P-labeled Fas/AP-1 site was used as hot probe (data not shown). These results suggest that the putative AP-1 site in the Fas promoter is able to bind AP-1 but with an affinity that is about 10-fold lower than the consensus AP-1 site.
We then conducted EMSA and supershift EMSA with reduced concentration of poly(dI-dC)⅐poly(dI-dC) and increased concentration of dithiothreitol in the binding reaction. We hoped that under these mild binding conditions, the binding of AP-1 to the putative Fas AP-1 site might be detectable. As shown in Fig. 8B, stimulation of Jurkat cells with anti-CD3 mAb induced the formation of the AP-1 complex whose identity was confirmed in supershift with antibodies against c-Fos, c-Jun, or JunD. These results suggest that AP-1 is indeed able to interact with the putative AP-1 site in the Fas promoter.
Induction of Luciferase Gene Expression Driven by Multiple Copies of the Fas GABP Site-To further examine the functional role of the GABP and AP-1 sites, two double-stranded oligonucleotides containing two or four copies of the Fas/GABP site (from Ϫ863 to Ϫ820) were inserted into the reporter plasmid pGL3/promoter to generate pGL3/2xGABP and pGL3/ 4xGABP, respectively. As shown in Fig. 9, anti-CD3 mAb stimulated luciferase activity 8 -10-fold in pGL3/4xGABP and pGL3/2xGABP-transfected Jurkat cells. Interestingly, PMA plus ionomycin appeared to be more effective in stimulating 2xGABP-and 4xGABP-driven luciferase expression than PMA alone. Mutation of the two ERE sites (pGL3/2xGABP/M1) or the AP-1 binding site (pGL3/2xGABP/M2) significantly reduced the induction of luciferase activity in response to stimulation by anti-CD3 mAb, PMA, or PMA plus ionomycin. These results further suggest that both the GABP and AP-1 sites are required for optimal induction of Fas gene transcription.
Identification of the Signaling Pathway Required for GABPdependent Luciferase Reporter Gene Expression-We next examined whether 2xGABP-driven luciferase reporter gene expression in response to anti-CD3 mAb, PMA, or PMA plus ionomycin was also mediated by PKC. As shown in Fig. 10A, inhibition of PKC by H7 (50 M) abolished the induction of luciferase activity in Jurkat cells stimulated by all stimuli. In contrast, the inhibitor of ERK (PD98059) and p38 MAP kinase cascades (SB202190) had no effect on promoter induction. These results, in combination with previous observations (45), suggest that PKC may play an essential role in regulating the activity of GABP and AP-1 in inducing Fas gene expression.
Previous studies have demonstrated that activation of JNK kinase in Jurkat cells by a novel Tet-On/Tet-Off system induces Fas expression (48). To examine whether inhibition of the JNK kinase cascade affected anti-CD3-stimulated luciferase activity, we co-transfected Jurkat cells with pGL3/2xGABP and DN JNK, DN Raf, or DN p38. As shown in Fig. 10B, co-transfection of pGL3/2xGABP with DN JNK inhibited the induction of luciferase activity in a dose-dependent manner, whereas cotransfection of Jurkat cells with pGL3/2xGABP plus DN Raf or DN p38 did not significantly inhibit anti-CD3-stimulated luciferase activity. However, co-transfection of pGL3/2xGABP with DN JNK plus DN Raf further decreased promoter induction. FIG. 8. Evidence that AP-1 is capable of complexing with the putative AP-1 site in the Fas enhancer. A, the ability of oligonucleotides containing the intact Fas/AP-1 site to compete with consensus AP-1 probe to bind AP-1. Nuclear extracts (2 g/samples) prepared from unstimulated (first lane) or anti-CD3 mAb-stimulated Jurkat cells (remaining lanes) were preincubated with 10ϫ, 50ϫ, or 200ϫ molar excess of indicated cold oligonucleotides on ice for 20 min. 32 P-Labeled consensus AP-1 (30,000 cpm) was added and then incubated at room temperature for another 20 min. The DNA-protein interactions were resolved on a 5% native polyacrylamide gel, and dried gel was exposed to X-Omat film. B, interaction of AP-1 with the putative AP-1 site in the Fas enhancer. 2 g of protein of nuclear extracts from unstimulated (first lane) or stimulated with anti-CD3 mAb (remaining lanes) was preincubated on ice with indicated antibodies for 30 min in the buffer containing 4% glycerol, 1 mM MgCl 2 , 0.5 mM EDTA, 5 mM dithiothreitol, 50 mM NaCl, 10 mM Tris-HCl, pH 7.5, 10 g/ml poly(dI-dC)⅐poly(dI-dC). 32 P-Labeled Fas/AP-1 oligonucleotide (30,000 cpm/reaction) was added and incubated at room temperature for 20 min. The DNA-protein interactions were resolved on a 5% native polyacrylamide gel, and dried gel was exposed to X-Omat film. These data indicate that JNK activation plays a critical role in mediating anti-CD3-stimulated Fas gene expression, whereas ERK plays only a minor or complementary role. These results are consistent with previous studies showing that GABP is predominantly phosphorylated by JNK and only weakly phosphorylated by ERK and not at all by p38 kinase (35). DISCUSSION The signaling pathways and transcription factors involved in TCR⅐CD3-stimulated Fas expression are largely unknown. In this study, we characterized a newly identified upstream enhancer element important for transmitting T cell activation signals to the Fas promoter. Transient transfection and DNA binding studies demonstrated that the Fas enhancer was com-posed of multiple sequence elements that functionally cooperated to activate the Fas promoter in response TCR⅐CD3 stimulation. Using signal transduction pathway blockers, we also show that PKC and JNK are important kinase intermediates involved in the activation of the Fas enhancer. Because H7 is the only pharmacological inhibitor used to elucidate the role of PKC, we could not rule out the possibility that other signaling molecules nonspecifically inhibited by H7 may also be involved in regulating TCR⅐CD3-mediated Fas expression. Nevertheless, our observations are consistent with a recent report showing that PKC is involved in TCR⅐CD3-and PMA-stimulated Fas expression in a T cell hybridoma cell line (45). In addition, we found that cyclosporin A and wortmannin, which inhibits calcineurin and PI 3-kinase, respectively, do not inhibit Fas enhancer activity, suggesting that Fas up-regulation is not mediated by PI 3-kinase and the transcription factor NF-AT. The inability of cyclosporin A to inhibit Fas enhancer activity is consistent with a previous report showing that cyclosporin does not inhibit TCR⅐CD3-triggered Fas mRNA production (49). These results further suggest that induction of Fas gene expression in TCR⅐CD3-stimulated T cells is mediated via a signaling pathway and a set of transcription factors distinct from that required for the induction of FasL gene expression in which cyclosporin A-sensitive NF-AT and Egr-3 play a dominant role (50 -55).
The downstream signaling molecules of PKC have been partially elucidated. It is believed that activation of PKC results in the activation of Ras-ERK MAP kinase cascade and the Rac⅐CDC42-JNK MAP kinase cascade (56 -58). Our inhibitor studies suggest that the ERK MAP kinase cascade does not play a significant role in regulating TCR⅐CD3-triggered Fas expression. In addition, our functional studies demonstrate that dominant-negative mutants of ERK did not inhibit Fas enhancer activity. In contrast, a dominant-negative mutant of JNK markedly inhibited Fas enhancer activity, indicating that activation of JNK plays a critical role in regulating the transcription of Fas gene expression. These results are in agreement with previous observations by Faris et al. (48,59) showing that activation of JNK kinase in Jurkat cells controlled by a Tet/On-Tet/Off system strongly induces Fas expression.
GABP is comprised of two subunits; the ␣ subunit binds weakly to its target sequence GCCGGAAGT but binds with high affinity when complexed with the ␤ subunit (36,37). Both ␣ and ␤ subunits are phosphorylated at serine/threonine residues by JNK1 and SAPK␣/␤ and weakly phosphorylated by ERK1/2 kinase (35). It is not clear so far whether other serine/ threonine kinases such as PKC are also able to directly phosphorylate GABP␣/␤. Our DNA binding studies show that anti-CD3 mAb and PMA, which predominantly activate ERK kinase and weakly JNK kinase in Jurkat cells (13), markedly increased GABP binding activity. In contrast, ionomycin alone, which does not activate ERK or JNK kinases, also did not increase the binding activity of GABP. These observations suggest that phosphorylation of GABP␣/␤ enhances its DNA binding activity. Because significant binding activity can be detected in unstimulated cells, as demonstrated in this studies as well as in several previous reports (34,35), it appears that phosphorylation of GABP is not obligatory for GABP binding to its target site. In support of this notion, Hoffmeyer et al. (35) show that bacterially expressed, nonphosphorylated GABP is able to bind to a GABP site derived from the IL-2 promoter.
In addition to GABP, our results suggest that AP-1 is also involved in the activation of the Fas enhancer. Sequence analysis revealed that an AP-1 site residing between the two EREs of the Fas enhancer has only one nucleotide difference from the consensus AP-1 binding site. The ability of AP-1 to interact FIG. 10. Inhibition of JNK but not ERK and p38 MAP kinase pathways blocks the Fas promoter-driven luciferase activity. A, inhibition of 2xGABP-driven luciferase expression by the specific inhibitor of PKC, H7, but not inhibited by the inhibitors of ERK kinase and p38 kinase pathways, PD98059 and SB202190, respectively. Jurkat cells were transfected with pGL3/2xGABP plasmid DNA and then incubated in the absence or presence of plastic-coated anti-CD3 mAb, PMA (50 ng/ml), and/or ionomycin (0.5 M). Cells were harvested, and cell lysates were prepared. Luciferase activity was quantitated in a Packard luminometer. The results shown here are the means Ϯ S.D. of a representative experiment in triplicate from two independent experiments with similar results. B, co-transfection of pGL3/2xGABP with dominant-negative mutants of MAP kinase cascades. Jurkat cells were transfected with pGL3/2xGABP or co-transfected with pGL3/2xGABP (1 g) plus various amounts of DN Raf, DN JNK, and DN p38 as indicated (g). Transfected cells were then incubated for 20 h in the absence or presence of plastic-coated anti-CD3 mAb. Cells were harvested, and cell lysates were prepared. Luciferase activity was quantitated in a Packard luminometer. The fold induction equals the luciferase activity in anti-CD3 mAb-stimulated Jurkat cells divided by the luciferase activity in unstimulated Jurkat cells. The results are the means of a representative experiment in triplicate from two independent experiments with similar results. with this AP-1 site was undetectable under normal binding conditions (Figs. 4 and 5) but could be detected under less stringent conditions (Fig. 8B), indicating that the putative Fas AP-1 site may form complex with AP-1 with low affinity. Consistent with this notion, our competition binding studies show that high molar excess of the Fas/GABP, M1, and F/AP-1 probes, all of which contain the intact putative AP-1 binding site, was able to compete for binding with the consensus AP-1 site (Fig. 8A). In contrast, the M2 and F/AP-1 oligonucleotides, which have mutation or deletion of the putative AP-1 site, failed to compete for binding with the consensus AP-1 site (Fig.  8A). Functional studies show that mutation of the Fas/AP-1 site greatly reduced enhancer activity, demonstrating that the Fas AP-1 site is functional and important for Fas gene expression even though it binds AP-1 with low affinity (Fig. 7). Moreover, because both the AP-1 and GABP sites were required for Fas enhancer function, it appears that the two sites cooperate, acting in concert to synergistically activate Fas gene expression. Transcription factors in the Ets family such as Ets-1 have been reported to physically interact with AP-1 and function cooperatively to activate gene expression in T (60).
In contrast to our studies, Chan et al. (61) recently reported that the PMA/ionomycin response of the Fas promoter in Jurkat cells is mediated by promoter sequences within 460 bp upstream of the ATG site (61). The reason for this discrepancy is not clear. In their study, the Fas promoter is induced about 17-fold by PMA/ionomycin, whereas we only detected an 8-fold increase. These apparent differences may be due to differences in the Jurkat cells themselves, the plasmid vectors the Fas promoter was cloned into, or the methods used to transfect the plasmids into the cells. Nevertheless, despite these apparent differences, our functional data are reproducible and consistent with our DNA binding and signal transduction studies, demonstrating that the majority of the PMA/ionomycin response is mediated by the upstream Fas enhancer through the cooperative interaction of GABP and AP-1.