Anergic T cells display a marked decrease in their ability to produce IL-2 even in the presence of optimal TCR and costimulatory signals. Using IL-2 enhancer/promoter-driven reporter constructs, we have previously identified a region that appears to be a target for cis transcriptional repression in anergy. This region of the promoter, which shares partial homology with a consensus AP-1-binding sequence, is located about −180 bp from the transcriptional start site. In the present study, we demonstrate that cAMP response element-binding protein/cAMP response element modulator (CREB/CREM), activating transcription factor-2/c-Jun, and Jun-Jun/Oct complexes bind to this site. However, the induction of anergy by prolonged stimulation through the TCR led to an increase in binding of only the CREB/CREM complex. Furthermore, the level of binding of this complex appeared to be up-regulated in both resting and restimulated anergic T cells. Finally, an IL-2 promoter-driven reporter construct that contained a mutation that specifically reduced the binding of the CREB/CREM complex displayed a decreased ability to be affected by anergy, while a construct that contained a mutation that decreased the binding of the Jun-Jun/Oct complex was still susceptible to anergy. These findings suggest that the −180 region of the IL-2 promoter is the target of a CREB/CREM transcriptional inhibitor that contributes to the repression of IL-2 production in T cell anergy.
The production of IL-2 and subsequent proliferation by T lymphocytes require two signaling events (1). Signal one refers to engagement of the TCR. The second signal or costimulation is believed to be predominantly mediated by signaling through CD28 and results in increased transcription and stabilization of IL-2 mRNA. T cells that receive signal 1 in the absence of costimulation not only fail to produce IL-2 and proliferate, but they do not proliferate to subsequent rechallenge, a state known as T cell clonal anergy (2, 3). A hallmark of anergic T cells is that their ability to produce IL-2 even in the presence of optimal TCR and costimulatory signals is significantly reduced. On the other hand, anergic T cell clones proliferate fully to exogenous IL-2, and in fact, incubating them in IL-2 reverses the anergic state, restoring their ability to produce IL-2 on rechallenge (4, 5, 6).
Examination of TCR-mediated signal transduction in anergic cells has revealed that there is normal CD3 ζ-chain phosphorylation and serine/threonine phosphorylation of the TCR γ-chain (7, 8). Signaling through the calcium pathway appears to be intact in that the translocation of NF-AT (1) from the cytoplasm to the nucleus remains unaffected (9). On the other hand, Li et al. (10) were able to demonstrate decreased Jun N-terminal kinase and early response kinase activation in anergic cells. Fields et al. also found a decrease in early response kinase activity and, in addition, showed that the decrease in mitogen-activated protein (MAP)2 kinase activity was a result of decreased p21ras activation (11). Indeed, using multimerized 12-O-tetradecanoylphorbol-13-acetate-responsive element (TRE)-driven reporter constructs, it has been shown that anergic T cells have a marked decrease in the induction of TRE-mediated transcription upon stimulation (12). In anergy, presumably, a block in the MAP kinase pathway would result in decreased Jun and Fos induction/activation, and subsequently a decrease in IL-2 transcription.
Our laboratory, and recently others, have suggested that the profound block in IL-2 production seen in T cell anergy is not only the result of a decrease in the production of positive transcription factors as a result of the block in the MAP kinase pathway, but also involves active negative regulation of IL-2 transcription (13, 14). We have identified a region of the IL-2 promoter that appears to be a target of cis-negative transcriptional regulation in anergic cells (13). Using T cell clones stably transfected with IL-2 promoter-driven reporter constructs, it was shown that the region located about −180 bp upstream of the transcription start site was necessary for the reporter to be susceptible to anergy. Thus, a mutation of this site in a construct that contained the native IL-2 promoter abrogated the ability of the reporter to be down-regulated after the induction of anergy. These observations suggest that the down-regulation of IL-2 production in anergic T cells involves more than just a failure to induce and activate Jun and Fos proteins.
The −180 region of the promoter lies between an NF-κB binding site and the CD28 response element (CD28RE), and is known as the distal AP-1 site. This designation is based on the fact that the sequence at this site differs from the consensus AP-1 site by 1 bp and proteins from nuclear extracts that had been affinity purified using consensus AP-1 double-stranded oligonucleotides could bind to this region (15). Jain et al., in their analysis of AP-1 sites of the IL-2 promoter, failed to demonstrate any binding of proteins to this site in EMSA (16). On the other hand, some groups have been able to demonstrate binding at this site, but by unidentified proteins that did not appear to be AP-1 (12, 15, 17).
In light of our functional data demonstrating a role for the −180 site in the prevention of IL-2 transcription during anergy, we decided to reexamine this region as a possible target for transcription factor binding. Our approach was 2-fold. First, we utilized nuclear extracts from the thymoma EL-4 as well as recombinant proteins in EMSAs. This provided us with an abundant source of transcription factors and facilitated the optimization of binding conditions as well as the identification of factors that could bind to this site. This approach yielded the identification of four protein-DNA-binding complexes, three of which were determined to consist of AP-1 and cAMP-responsive element-binding protein (CREB) family members. Second, we examined extracts from anergic T cell clones to determine which of these factors might play a role in the transcriptional block seen in anergy. The binding of only one of these complexes, which consists of CREB-1 and cAMP-responsive element modulator (CREM) proteins, was up-regulated in anergic extracts. Most importantly, using IL-2 promoter-driven reporter constructs containing mutations that selectively reduced the binding of the CREB-1/CREM complex, we are able to demonstrate a role for the binding of this complex in promoting the inhibition of IL-2 transcription in anergy.
Materials and Methods
Cell lines and culture conditions
EL-4 is a murine T cell lymphoma and A.E7 is a CD4+, Th1, pigeon cytochrome c (PCC)-specific T cell clone (18, 19). Both EL-4 cells and the A.E7 T cell clone were maintained in medium consisting of equal volumes of RPMI 1640 and Eagle’s Hank’s amino acids (Biofluids, Rockville, MD) supplemented with 10% FCS, 5 × 10−5 M 2-ME, 4 mM glutamine, and antibiotics, at 37°C in a 5% CO2 humidified incubator. The A.E7 T cell clone was expanded as previously described (6). In general, the cells were stimulated with Ag for 48 h and then expanded in 10 U/ml of IL-2, approximately every 3 wk. The T cells were not utilized in experiments for at least 12 days after expansion to allow them to rest down.
Anergy was induced as previously described (13). Briefly, 100 × 106 cells were incubated overnight in T 162-cm2 tissue culture flasks that had been precoated with 10 μg/ml of anti-TCR Ab (anti-TCR-β Ab H57-597) (20). The cells were then harvested using a cell scraper, washed, and rested for a minimum of 5 days in fresh medium that did not contain IL-2. Proliferation to Ag was performed by incubating 20 × 103 cells with 500 × 103 splenocytes (irradiated with 3000 rad) from B10.A mice as APCs with varying doses of PCC, in 0.2 ml of complete medium in 96-well plates. The cultures were pulsed with [3H]thymidine at 48 h and harvested at 64 h. Exogenous rIL-2 (R&D Systems, Minneapolis, MN) was added to a triplicate of each sample in the absence of Ag as a positive control for functional viability. The cells were tested for their ability to produce IL-2 by incubating 500 × 103 cells in anti-TCR-coated 24-well plates (Costar, Cambridge, MA) with 1/5000 dilution of ascitic fluid containing anti-CD28 mAb 37.51 (a gift of Dr. J. Allison, University of California, Berkeley, CA) in a total volume of 0.5 ml (21). Supernatant fluid was collected after 16 h, and IL-2 activity was determined by measuring proliferation of the IL-2-dependent CTLL cell line (American Type Culture Collection, Manassas, VA), as previously described (13).
Preparation of nuclear extracts
Nuclear extracts from EL-4 cells and the A.E7 T cell clone were prepared by modification of the procedure of Dignam et al. (22). The cells were first incubated on ice for 15 min with 10 mM HEPES, pH 8, 10 mM KCL, 0.1 mM EDTA, 0.1 mM EGTA, 3.3 μg/ml aprotinin, 10 μg/ml leupeptin, 2.5 μM 4-(2-aminoethyl)-benzenesulfonylfluoride hydrochloride, and 1 mM DTT. Next, an equal volume of the same solution with 2% Triton-X was added, and the cells were mixed for 15 s and spun in a microcentrifuge at 10,000 rpm for 30 s. The supernatant fluid was discarded and the nuclear pellet was resuspended in a solution containing 20 mM HEPES, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, protease inhibitors, and 1 mM DTT. The final concentration of the nuclear extract was adjusted to 400 × 103 cells/μl for the EL-4 extracts and 1 × 106 cells/μl for the A.E7 extracts. The resuspended nuclear pellet was incubated rocking for 30 min at 4°C and then was spun at 12,000 rpm in a microcentrifuge to remove insoluble material. The extracts were frozen at −70°C until they were assayed.
Electrophoretic mobility shift assays
EMSA were conducted using 4% polyacrylamide gels, as previously described, with some modifications (13). Nuclear extracts (1–3 μl) were incubated with 30,000 cpm of 32P end-labeled, double-stranded oligonucleotide probe (10–50 pg) with 0.1–1 μg of poly(dG)·poly(dC) (Sigma, St. Louis, MO) in 4.5 mM Tris, 32.5 mM KCl, 2 mM DTT, 1 mM EDTA, 10% glycerol, 0.03% Nonidet P-40, and 100 μg/ml BSA at 4°C for 30 min. Gel electrophoresis was then run at 4°C. Where noted, poly(dI · dC) was used in place of poly(dG)·poly(dC) as the nonspecific competitor. It was found that the dose of poly(dI · dC) and poly(dG)·poly(dC) needed for optimal binding varied between preparations. Thus, each new batch was titrated for optimal results. In some assays, recombinant proteins were utilized in EMSA. Where indicated, 0.01–1 protein footprint units (optimum binding was batch dependent) of recombinant c-Jun (Promega, Madison, WI; catalogue E3061) were added to the binding reaction, while 25 ng of rCREM (Santa Cruz Biotechnology, Santa Cruz, CA; catalogue sc-4005) was used. The dsDNA probes of the −180 site utilized were the following: SLD-2, AATCCATTCAGTCAGTGTATGGGGGT; LD, AATCCATTCAGTCAGTGTATGGGGGGTTTAAA; LDM-1, AATCCATTCttgCAGTGTATGGGGGTTTAAA; and SLD-12, CCATTCAGTCAGTGTATGGGGGT.
Additionally, dsDNA probes of the following sequences were used for cold competition analysis: NF-IL-2a, GAAAATATGTGTAATATGTAAAACATCGT (15); TRE proximal (−150 AP-1 site), GAAATTCCAGAGAGTCATCAGAAGA (15); NF-AT, TCGACCAAAGAGGAAAATTTGTTTCATACAGAG (15); consensus AP-1, CGCTTGATGAGTCAGCCGGAA (Promega; E3201); consensus octamer (Oct), TGTCGAATGCAAATCACTAGAA (Santa Cruz Biotechnology; sc-2506); consensus EGR, GGATCCAGCGGGGGCGAGCGGGGGCGA (Santa Cruz Biotechnology; sc-2530); and consensus CREB, AGAGATTGCCTGACGTCAGAGAGCTAG. (Santa Cruz Biotechnology; sc-2504).
All blots were developed by the STORM PhosphorImager (Molecular Dynamics, Sunnyvale, CA). All comparisons concerning band density were made for each individual gel, as there was great variation in background between gels. Quantification of each of the bands was determined by the PhosphorImager using a fixed area and by using the object average program (Imagequant; Molecular Dynamics) for determining the background. In this way, the background was determined for each individual lane and subtracted from the band density to account for interlane background variation.
Supershift analysis and immunodepletion
Nuclear extracts were preincubated with 1 μg of the indicated Abs for 30 min before the addition of the labeled probe. The following Abs were obtained from Santa Cruz Biotechnology: antiactivating transcription factor-2 (ATF-2) (sc-6233x), broadly reactive anti-c-Jun (sc-044x), specific anti-c-Jun (sc-045x), broadly reactive anti-CREB-1 (sc186x), specific anti-CREB-1 (specific sc-240x), anti-CREM (sc-440x), anti-Oct-1 (sc-232x), anti-Oct-2 (sc-233x), and anti-c-Rel (sc-272x). For immunodepletions, 30 μl of protein A/G beads (Santa Cruz Biotechnology) were preincubated with 10 μg of Ab for 1 h, while rocking at 4°C. The beads were washed three times with PBS and then incubated for 2 h with 12 μl of extract and 48 μl of H20, while rocking at 4°C. The beads were spun down, and 10 μl of the supernatant fluid was assayed by EMSA.
Western blot analysis
Nuclear extracts utilized in EMSA were subjected to 10% polyacrylamide gel electrophoresis, transferred to nitrocellulose, and blotted with the following Abs: Anti-CREM (sc-440x, 1 μg/ml), anti-CREB (sc-186x, 1 μg/ml), and anti-phospho-CREB (1 μg/ml; Upstate Biotechnology, Lake Placid, NY). The secondary Ab consisted of an alkaline phosphatase-labeled anti-rabbit (Sigma), and the blot was developed with Vistra ECF substrate (Amersham, Arlington Heights, IL) using the blue fluorescence mode of the STORM PhosphorImager.
Site-directed mutagenesis and transfections
Reporter assays were performed using a luciferase reporter gene (pGL-2; Promega) driven by 353 bp of the 5′ IL-2 promoter. Specific mutations were made using the Quick-Change site-directed mutagenesis kit (Stratagene, La Jolla, CA). The following primers were used to generate the mutations: M-1, 5′-CCGACCAAGAGGGATTTCACCTAAATCCATCCATTCTTGCAGTGTATGG-3′ (forward), 5′-CCATACACTGACTGAACAGAACAGATTTAGGTGAAATCCCTCTTGGTC-3′ (reverse); M-6, 5′-GAGGGATTTCACCTAAATCCATTCAGTGCGTGTATGGGGGGTTTAAAG-3′ (forward), 5′-CTTTAAACCCCCATACACGCACTGAATGGATTTAGGTGAAATCCCTC-3′ (reverse); M-12, 5′-CCATACACTGCAAGAATGGATTTAGGTGAAATCCCTCTTGGTCGG-3′ (forward), 5′-GACCAAGAGGGATTTCACCTAAATCTGTTCAGTCAGTGTATGG-3′ (reverse). The fidelity of the mutations was checked by sequencing on an ABI 377 automated sequencer.
Transfections were performed using plasmid-coated gold particles and a gene gun (Bio-Rad, Richmond, CA), as previously described, with some modification (23). Briefly, anti-CD44-coated 35-mm dishes were layered with 10 million A.E7 cells. The gold beads were coated with 10 μg of DNA. To eliminate the variation in transfection efficiency between transfections, each condition was derived by pooling cells from multiple transfections. Cotransfection with SV40-driven Renella luciferase (Promega) revealed that both normal A.E7 and anergic A.E7 cells could be transfected with equal efficiency; however, the presence of the second plasmid appeared to affect the activity of the mutated reporter plasmids (data not shown). Therefore, in all of the experiments shown, the cells were transfected with only one reporter construct. The cells were stimulated overnight in six-well plates (10 million cells/well) that had been previously coated with anti-TCR Ab (H57, 10 μg/ml). Luciferase activity was determined using a luciferase reporter kit (Promega) and a femtomaster FB-12 luminometer (Zylux, Maryville, TN). Data were evaluated for statistical significance by a Student’s t test.
Multiple protein complexes can bind to the −180 site
In light of our functional data defining the −180 site as a target of negative transcriptional regulation in anergy, our initial goal was to determine what, if any, transcription factors bind to this region of the IL-2 promoter. To this end, we first performed a series of experiments using nuclear extracts from the immortal T cell lymphoma EL-4, which provided us with an abundant source of nuclear extracts. As seen in Fig. 1 A (lane 1), once EMSA conditions were optimized, we were able to observe the binding of four protein-DNA complexes at this site. Of note, when the EMSA were performed with poly(dI · dC) as the nonspecific DNA in the binding buffer, the formation of these bands was inhibited (data not shown). Poly(dI · dC) is known to inhibit the binding of transcription factors that bind to the minor groove of DNA. We hypothesize that this is the reason that Jain et al. (16) were unable to demonstrate binding at this site.
Identification of bands I, II, and III
A series of cold competition assays was performed as a screening method to gain insight into which proteins were involved in the formation of each of the four protein-DNA complexes. Nuclear extracts were prepared from unstimulated EL-4 cells and run in an EMSA with the labeled SLD-2 probe in the presence and absence of various unlabeled probes as cold competitors. As seen in Fig. 1 A, all four bands are competed by excess (50×) cold SLD-2 (lane 2), but not by an excess of an unlabeled irrelevant (EGR) dsDNA probe (lane 5). An unlabeled probe with the consensus metallothionein AP-1 sequence completely eliminates bands I, II, and IV and decreases band III (lane 3). This suggests that all four bands contain proteins that have some affinity for the consensus AP-1 binding site. Additionally, both bands I and II were completely eliminated by an unlabeled consensus CREB probe (lane 4) and band III was diminished to a much lesser degree. This result raised the possibility that CREB family member proteins might be binding to this site. Interestingly, both the cold consensus AP-1 and consensus CREB probes failed to completely inhibit band III. This may indicate that this protein-DNA complex has higher affinity for the native AP-1-like sequence contained in the SLD-2 probe than the consensus sequences contained in the cold competitors.
In light of the cold competition data, supershift analyses were performed utilizing Abs against various CREB and AP-1 family members. Fig. 1,B demonstrates that anti-ATF-2 Abs completely supershift band I while having no effect on bands II and III (lane 2). Anti-c-Jun also supershifts band I to some extent (lane 3), suggesting that band I contains at least an ATF-2/c-Jun heterodimer. Anti-CREB-1 (lane 4), anti-ATF-1, and anti-JunB (data not shown) had no effect on this band. Band II, on the other hand, is completely eliminated by an anti-CREB-1 Ab (lane 4). This suggests that band II contains CREB-1, although this particular Ab also cross-reacts with other CREB family members. Anti-ATF-1 Ab (data not shown) and anti-ATF-2 (lane 2) did not supershift this band. Thus, although band II was competed by both cold consensus cAMP response element and AP-1 DNA probes, the supershift data suggest that this complex consists of CREB family members. Such a finding is not without precedence. Masquilier et al. (24) have demonstrated the ability of CREB and CREM to bind to AP-1 binding sites and in fact inhibit AP-1-mediated transcription at these sites. Finally, in spite of the cold competition data demonstrating that unlabeled AP-1 and CREB probes could partially cold compete band III (Fig. 1 A), none of the Abs supershifted this band.
To identify the proteins that comprise band III, we screened multiple dsDNA consensus probes in cold competition EMSA using a labeled LD probe (this probe enhances the binding of band III). A representative experiment is shown in Fig. 1,C. As seen in lane 2, band III is reduced by the presence of excess unlabeled self probe. The specificity of the protein-DNA interaction is demonstrated by the fact that the band is not competed by the unlabeled mutated probe LDM-1 (lane 3). Band III is partially inhibited by the presence of unlabeled NF-IL-2a (lane 4). This latter sequence, an element of the proximal IL-2 promoter, has been shown to bind NF-AT, AP-1, and Oct proteins (15). An unlabeled consensus NF-AT double-stranded probe did not cold compete (lane 5); however, consensus AP-1 and Oct sequences did partially inhibit binding (lanes 6 and 7). These data suggest a possible role for AP-1 and Oct proteins in the formation of band III. Anti-Oct-1 and anti-Oct-2 Abs, as well as anti-Jun B Abs sometimes resulted in the partial inhibition of band III formation (data not shown). Furthermore, band III formation was repeatedly reduced by immunodepletion of the extracts with anti-JunB and anti-Oct Abs. Fig. 1 D shows a representative experiment of this immunodepletion. Extracts depleted with an anti-pan Jun Ab (lane 3), anti-Jun B (lane 4), or anti-Oct 1 and anti-Oct 2 (lane 5) showed decreased formation of band III. The specificity of the immunodepletion is demonstrated by the fact that the depletion of multiple other factors including ATF-2 (lane 2) and CREB-1 (lane 6), shown to play a role in bands I and II, had no effect on band III. Of note, multiple anti-Fos antisera failed to supershift or immunodeplete band III (data not shown).
Thus, these data and the cold competition experiments suggest that Jun and Oct proteins participate in the formation of band III. However, the inability to completely eliminate band III may indicate that other proteins also bind at this site. The binding of Oct, CREB, or AP-1 family member proteins to a region of a promoter that does not contain consensus sequences for their binding has precedent in several promoters, including the proximal IFN-γ promoter (25). Furthermore, Sterling and Bresnick have identified a negative regulatory element of the liver-specific CYP1A1 promoter that shares partial homology with the −180 site and binds Oct proteins (26).
Nuclear extracts from anergic T cell clones
Having utilized nuclear extracts from the T cell lymphoma EL-4 to characterize three protein-DNA complexes that can bind to the −180 site, we next wanted to determine which, if any, of these complexes might be involved in anergy. A.E7 is a PCC-specific Th1-type murine T cell clone maintained in our laboratory. When these cells are stimulated overnight with signal 1 alone, in the form of plate-bound anti-TCR Abs, then washed and rested, such cells are hyporesponsive to subsequent full rechallenge with anti-TCR and anti-CD28 (signal 1 + 2). As seen in Fig. 2,A, the anergic A.E7 cells show a marked decrease in their ability to proliferate to PCC. They proliferate, however, as well as nonanergic cells to exogenous IL-2 (see legend). Even more striking is their inability to produce IL-2 under optimal stimulatory conditions (Fig. 2 B).
Initial experiments were designed to examine binding to the −180 site by components of nuclear extracts derived from T cells immediately following anergy induction. As shown in Fig. 3 A, EMSA was performed on nuclear extracts obtained from A.E7 cells that were mock stimulated or stimulated with anti-TCR for 16 h to induce anergy. Lane 1 shows the binding pattern of EL-4 nuclear extracts for comparison. The EL-4 extracts in this particular experiment using the probe SLD-2 demonstrate prominent band I and band III binding and minimal band II binding. In contrast, extracts from resting A.E7 cells demonstrate slight binding of all three complexes (lane 2). Interestingly, the extract from the A.E7 cells stimulated to induce anergy shows increased binding only of band II (lane 3). The Jun containing band III, which is so prominent in the extracts from the immortalized EL-4 cells, is barely detectable in the extracts from the T cell clones. The fourth, faster migrating band seen with the EL-4 extracts is also present in the A.E7 extracts and was unaffected by the activation status of the cell (data not shown). Thus, under the stimulatory conditions that lead to the induction of anergy, the level of binding of band II (the CREB complex) appears to be selectively up-regulated in the A.E7 T cell clones.
Next, we examined nuclear extracts from fully rested and stimulated anergic cells. A.E7 cells were anergized by stimulation with anti-TCR overnight, and the cells were then washed and rested for 5 days. These are the same cells that were tested functionally for anergy in Fig. 2. Nuclear extracts were made from the resting anergic cells and their nonanergic counterparts as well as anergic and nonanergic cells that had been stimulated for just 3 h with anti-TCR and anti-CD28. As seen in Fig. 3,A, there is enhanced binding of band II in the extract from the resting anergic T cells compared with their nonanergic counterpart (20.9 vs 7.7 arbitrary units, respectively, by PhosphorImager analysis). In both the nonanergic and anergic extracts, stimulation up-regulated band II: lanes 4 and 5, 7.7–17.5, and lanes 6 and 7, 20.9–35.6, respectively. Note that all numbers were derived by subtracting the background from each lane to account for the interlane background differences. The faint band I in the resting anergic extracts and the small increase in band III in the stimulated anergic extracts were not always observed in other experiments (Fig. 3 B). In contrast, the increased formation of band II using extracts from anergic cells compared with extracts from nonanergic cells was always observed, although the magnitude of the increase was variable. Among multiple experiments (>10), band II binding from the anergic extracts was anywhere from 1.5–6-fold increased compared with the extracts from the nonanergic cells.
From the experiments utilizing the EL-4 nuclear extracts, we determined that band II contained CREB family member proteins. To determine whether this was also the case for the anergic extracts, as well as to use more specific Abs to further characterize this band, a supershift experiment was conducted using nuclear extracts from stimulated anergic cells (Fig. 3 B). In this experiment, once again the most prominent (only) complex observed is band II. This band is eliminated with the broadly reactive CREB antiserum (lane 2). Furthermore, band II is supershifted by a CREB-1-specific antiserum (lane 3), as well as by a CREM-specific antiserum (lane 4). No effect was seen with an anti-ATF-1 antiserum (lane 5). These data suggest that band II contains a CREB/CREM heterodimer. CREM, which is known to form heterodimers with CREB, has a very similar DNA binding domain as that of CREB; however, most of the CREM isoforms lack a transactivating domain, and thus, CREB-CREM heterodimers and CREM-CREM homodimers have been shown to be inhibitors of transcription (27). Thus, band II, the complex whose binding is up-regulated in the anergic state, consists of transcription factors known to participate in the inhibition of transcription.
To determine whether the increased binding of CREB/CREM at the −180 site in anergy was due to an increase in the amount of CREB and CREM found in the extracts derived from the anergic cells, A.E7 cells were either mock stimulated (denoted: M), or stimulated with plate-bound anti-TCR for 16 h to induce anergy (denoted: A), and nuclear extracts were obtained. As seen in Fig. 4,A, there is an increase in band II in the extracts derived from the anergic cells. Western blot analysis of the same nuclear extracts (Fig. 4 B) does not reveal any differences in total CREM or CREB. The lower band in the CREB and phospho-CREB blots is CREM, as both Abs cross-react with CREM. There is a slight increase, however, in the levels of phosphorylated CREB and CREM (upper and lower bands, respectively). Thus, it does not appear that the differences in binding of CREB and CREM to the −180 site are due to quantitative differences in their expression, but rather may be due to qualitative differences such as their phosphorylation status.
Differential binding properties of the CREB/CREM and Jun-Jun/Oct complexes
To determine whether we could further dissect the binding properties of each of the protein complexes and in so doing define their role in either promoting or inhibiting transcription, we used the SLD-12 probe as a backbone (it is the minimal probe able to bind all three complexes). By using a series of deletion mutants, we had determined that the CREB/CREM complex appeared to bind the −180 site and the region immediately 5′ of this site, while the Jun-Jun/Oct complex appeared to bind to the −180 site and the region immediately 3′ of this site (data not shown). Based on these data, mutations were made to try to selectively inhibit the binding of the CREB/CREM complex or the Jun-Jun/Oct complex. M-6 is a 2-bp mutation of the last 2 bases of the 3′ end of the −180 site. Likewise, a 2-bp mutation (M-12) was made in the CCA sequence immediately 5′ of the −180 site. The consequences of these mutations on the binding of each of the complexes are depicted in Fig. 5. In A, using extracts from stimulated anergic A.E7 cells, we see that the M-6 mutation results in enhancement of both band I and band II, while the M-12 mutation leads to a diminution of both of these bands. In addition, band III, barely detectable using the wild-type probe, is more prominent when band II is inhibited by the M-12 mutation. Using EL-4 extracts, we see the same pattern of binding for each of the mutations (B). Note that whereas there is very little band II when using the EL-4 extracts with the WT probe (B, lane 1), when the M-6 mutation is present, not only does this result in a decrease in the dominant band III, but it also results in a dramatic up-regulation of band II binding (B, lane 2). This suggests that the lack of substantial band II binding observed for the EL-4 extracts is not secondary to a deficiency in CREB and CREM proteins, but rather the ability of the Jun-Jun/Oct complex to outcompete the CREB/CREM complex for binding to the central part of the −180 site.
Consistent with these data are the effects of these mutations on the binding of recombinant proteins. As seen in Fig. 5,C, the M-6 mutation has no effect on recombinant CREM binding, but the M-12 mutation results in a decrease in CREM binding. On the other hand (as seen in Fig. 5 D), the M-6 mutation results in the diminution of recombinant Jun binding, while the M-12 mutation actually results in an increase in Jun binding. Note that we have observed that both the recombinant CREM and c-Jun proteins appear as doublets. The exact nature of these two bands is unclear, but may be due to degradation of the bacterially produced protein.
To further demonstrate the differential binding properties of the recombinant CREB and Jun, we performed cold competition experiments using the mutated probes. As seen in Fig. 5,E, the binding of recombinant CREM to labeled SLD-12 is nearly completely competed away by 100X unlabeled SLD-12 M-6 (which contains the 3′ mutation). In contrast, the addition of 100X unlabeled SLD-12 M-12 (which contains the 5′ mutation) was less effective in competing with the labeled SLD-12 probe. Alternatively (Fig. 5,F), recombinant Jun is preferentially cold competed by SLD-12 M-12 and less so by the SLD-12 M-6. Taken together, the binding data presented in Fig. 5 support the notion that CREM and band II rely upon the 5′ region of the −180 site for binding, while Jun and band III rely upon the 3′ region of the −180 site for binding.
Correlating protein-DNA binding with function
Although we have been able to demonstrate the binding of four protein-DNA complexes to the −180 site, only the formation of the CREB/CREM complex appears to be up-regulated in nuclear extracts from anergic T cells. This provided circumstantial evidence that the CREB/CREM complex was involved in promoting anergy. We wanted to address this hypothesis functionally. Based on our binding data, we would predict that the M-12 mutation that disrupts the binding of CREB/CREM would also impair the ability of a reporter construct to be affected by anergy. Conversely, because inhibiting the binding of the Jun-Jun/Oct complex did not impair CREB/CREM binding, we would predict that the M-6 mutation would not inhibit an anergic effect. To this end, luciferase reporter constructs driven by the 5′ 353 bp of the IL-2 promoter were made containing the M-1, M-6, and M-12 mutations. These constructs were examined in transient transfection assays using a gene gun that facilitated the transfection of our nonimmortalized A.E7 T cell clones. A.E7 or anergic A.E7 cells were transfected with each of the plasmids and stimulated overnight in anti-TCR-coated six-well plates. Preliminary experiments using the gene gun determined that overnight stimulation was optimal, and that the luciferase activity of unstimulated transfected A.E7 cells was typically less than 200 relative light units/s (data not shown).
Fig. 6,A depicts the luciferase activity of a representative experiment, while Fig. 6,B depicts the fold difference between the anergic and nonanergic cells for this experiment. This pattern of inhibition is representative of multiple experiments, as seen in Fig. 6,C, in which the data are presented as the geometric mean fold differences between luciferase values from the nonanergic and anergic cells. As seen in Fig. 6 A, the constructs that contained mutations in the −180 site demonstrate decreased luciferase activity in nonanergic cells when compared with the WT constructs. This is consistent with the data of Jain et al. (16), who observed 2–4-fold decreases in reporter activity by mutating the −180 site.
There was a significant decrease in IL-2 promoter-driven luciferase activity in the anergic cells transfected with the WT promoter, when compared with their nonanergic counterparts (Fig. 6 C). In contrast, there was no significant difference between the luciferase activity of the anergic and nonanergic cells transfected with the construct that contained the M-1 mutation, which inhibits the binding of all transcription factors to the −180 site. Similar results were obtained utilizing reporter constructs containing the M-12 mutation, which impairs the binding of the CREB/CREM complex (as well as the ATF-2/c-Jun complex), but leaves the Jun-Jun/Oct binding intact. The M-6 mutation, on the other hand, which impairs the binding of the Jun-Jun/Oct complex and favors the binding of the CREB/CREM complex, retained the ability to be suppressed under anergic conditions. This pattern of results has also been obtained in two experiments using A.E7 cells stably transfected with WT, M-1, and M-12 reporter constructs (data not shown). Thus, only mutations that impair the binding of the CREB/CREM complex decrease the susceptibility of the reporter constructs to anergy, supporting a role for this complex in maintaining the anergic state. Because the binding of the ATF-2/c-Jun complex is also impaired by the M-12 mutation, we cannot rule out a role for this complex as well in anergy.
Although it is referred to as the distal AP-1 site, the −180 site does not share precise homology with the consensus AP-1 sequence (15). Indeed, previously, neither our group nor others have been able to demonstrate the binding of Jun/Fos heterodimers to this region of the IL-2 promoter (12, 16). Furthermore, the functional role, if any, that Jun/Fos heterodimers play at this site was undefined. The present study not only demonstrates an increase in CREB/CREM binding at this site in extracts derived from anergic cells, but also functionally correlates the binding of the CREB/CREM complex to anergy using mutated IL-2 promoter-driven reporter constructs. Because there was less of a difference in the reporter activity between the anergic and nonanergic cells transfected with the M-1 and M-12 constructs, these data suggest that the profound decrease in IL-2 production seen in anergy is not exclusively due to a decrease in activated Jun and Fos production, but also involves cis dominant negative regulation of transcription.
The differential binding properties of Jun and CREB family proteins to the −180 site present a potential mechanism for regulation at this site. For the proximal IFN-γ promoter, Penix et al. (25) have shown that the binding of ATF-1/CREB can inhibit transcription, while the binding of ATF-2/Jun (which is favored during activation) can enhance transcription. Interestingly, band II often appeared as a minor band in extracts from the immortalized, nonanergizable EL-4 cells (Fig. 5 B). When Jun-Jun/Oct binding was inhibited by the M-6 mutation, however, the CREM band became prominent, suggesting that the Jun complex (band III) might be inhibiting CREB/CREM binding.
CREB/CREM protein binding at the −180 site is up-regulated upon the induction of anergy. This is consistent with the findings of Feuerstein et al. (28), who were able to demonstrate increased cAMP response element-binding activity in nuclear extracts derived from cells stimulated with anti-CD3. Interestingly, the up-regulation of band II formation was also observed in the nuclear extracts from resting anergic cells. The precise mechanism of this increased binding is unclear. Western blot analysis has not revealed any differences in the amount of CREB or CREM in anergic vs nonanergic T cells (Fig. 4). It might be that prolonged TCR engagement leads to the modification of the CREB/CREM proteins that favor the binding of this complex to the −180 site. Consistent with this hypothesis is our finding that CREM and CREB are phosphorylated in anergic cells. Phosphorylation at serine 133, which is detected by the anti-phospho-CREB Ab in Fig. 4, does not appear to be involved in CREB or CREM binding to DNA (29). However, it has been shown that in vitro phosphorylation of CREM τ can enhance its affinity for DNA (29), and thus we are currently investigating the phosphorylation of CREB and CREM at a number of different sites. Along these lines, ongoing studies in our laboratory have demonstrated the binding of recombinant high mobility group (I) Y (HMG(I) Y) to the −180 site (unpublished data). This family of proteins usually binds to A-T-rich regions in the minor groove of the DNA and can serve to enhance or inhibit the binding of various transcription factors (30, 31). Preliminary data suggest that recombinant HMG(I) Y can inhibit recombinant Jun binding while stabilizing recombinant CREM binding. It remains to be determined whether HMG(I) Y proteins play such a role in anergic T cells.
The potential role of CREB family proteins in inhibiting IL-2 production is consistent with the observation that increases in cAMP result in decreases in IL-2 production by lymphocytes and that persistent elevation of cAMP in Th1 clones leads to a hyporesponsive state (32, 33). In this regard, it is of interest that band II, which is up-regulated in anergic cells, consists of a CREM-containing complex; many of the CREM isoforms lack a transactivating domain, and these isoforms have been shown to act as inhibitors of transcription. Similarly, Bodor and Habener have demonstrated that the inducible cAMP early repressor is able to inhibit transcription of a number of cytokine genes, including IL-2 (34). It is of further interest that there is a profound decrease in IL-2 production by T cells derived from mice that express a transgenic dominant negative CREB under the CD2 promoter (35). Additionally, others have demonstrated the ability of overexpression of dominant negative CREB to inhibit IL-2 promoter-driven luciferase activity (36). That the overexpression of the dominant negative CREB (that can bind to DNA, but cannot transactivate, and thus behaves like an inhibitory CREM isoform) inhibits IL-2 transcription, serves to support our contention that CREM promotes IL-2 repression in anergy. This observation, however, does not prove that CREM binding at the −180 site inhibits IL-2 transcription, because Butscher et al. (36) have argued that the decrease in IL-2 production is due to the essential role of CREB family member binding at the −150 site. If indeed the overexpression of the dominant negative CREB had failed to inhibit IL-2 production, it would have challenged the veracity of our model. Thus, the data obtained in various model systems are compatible with the idea that CREB/CREM family member proteins can negatively regulate IL-2 transcription.
Several investigators have begun to explore the composite CD28RE-AP-1 (−150) site as a model for the role of coordinate transcription factor binding in promoting transcription (36, 37, 38). Based on the present characterization of protein-DNA binding and our functional data, we propose that perhaps the scope of this composite site might be expanded to include the −180 site. More importantly, we propose that the −180 site might prove to be a useful model for the investigation of the coordinate inhibition of transcription. The proximity of the −180 site to the CD28RE-AP-1 site, the requirement for the −150 site in addition to the −180 site to manifest an anergic effect (13), and the fact that the proteins that bind to the two sites are from the same transcription factor families is intriguing and lends itself to speculation concerning the mechanism of transcriptional inhibition. First, the CREB/CREM complex may disrupt the efficient formation of a single large IL-2 promoter enhanceosome (39). It may be that the interposition of CREM into the transcription factor complex renders it inactive. Alternatively, the binding of CREB-1/CREM and perhaps even HMG(I) Y might bend the DNA in such a way that the enhanceosome cannot be efficiently formed. Second, it might be that CREB/CREM proteins actually form an inhibitory complex. Both CREM/CREM homodimers and CREB/CREM heterodimers have been shown to antagonize both CREB and AP-1-mediated transcription (40). Indeed, it has been shown that CREM can inhibit AP-1-mediated transcription by outcompeting Jun for binding at conventional AP-1-binding sequences (24). In this regard, it might be that under anergic conditions, the CREB/CREM complex outcompetes a Jun/Jun-Oct complex.
A third possibility is that the CREB/CREM complex inhibits IL-2 production at the level of coactivation of transcription. It has been shown that the amount of the coactivators’ CREB-binding protein (CBP) and p300 are tightly maintained within the cell (40). Kamei et al. showed that in conjunction with a number of nuclear receptors, overexpression of CREB protein can inhibit AP-1-mediated transcription (41). Their proposed mechanism for this repression is the ability of the CREB protein to recruit away limiting amounts of coactivator proteins. Similarly, Parry et al. (42) suggested that phosphorylated CREB can inhibit NF-κB transcription by competing with p65 for limiting amounts of CBP. Along these lines, the binding of CREB/CREM to the −180 site could serve to recruit CBP/p300 into a nonfunctional complex. Alternatively, it might be that the binding of the CREB/CREM complex at these sites facilitates the binding of a corepressor molecule. The coactivators CBP and p300 both possess histone acetyltransferase activity, which is believed to play a role in their ability to promote transcription (43). Recently, it has been shown that corepressor molecules, which possess deacetylase activity, can also be recruited into protein-DNA complexes and act to inhibit transcription (44). Finally, it has recently been shown that p300 and CBP have distinct, nonoverlapping roles in coactivating transcription (45). It might be that in anergy the formation of the CREB/CREM complex favors the recruitment of the wrong coactivator, ultimately leading to repression of transcription.
We thank Dr. Kevin Gardner and Dr. Luciano D’Adamio for their helpful suggestions, as well as the members of Laboratory of Cellular and Molecular Immunology for all of their help and support.
Abbreviations used in this paper: MAP, mitogen-activated protein; ATF, activating transcription factor; CBP, CREB-binding protein; CD28RE, CD28 response element; CREB, cAMP response element-binding protein; CREM, cAMP response element modulator; EGR, early gene response; HMG(I) Y, high mobility group (I) Y; PCC, pigeon cytochrome c; TRE, 12-O-tetradecanoylphorbol-13-acetate response element.