Targeting of p300 to the interleukin-2 promoter via CREB-Rel cross-talk during mitogen and oncogenic molecular signaling in activated T-cells.

In this report, we explore the mechanisms of targeting of p300 to the interleukin-2 (IL-2) promoter in response to mitogenic and oncogenic molecular signals. Recruitment of p300 by cAMP-responsive element-binding protein-Rel cross-talk at the composite CD28 response element (CD28RE)-TRE element of the IL-2 promoter is essential for promoter inducibility during T-cell activation, and CD28RE-TRE is the exclusive target of the human T-cell lymphotropic virus type I oncoprotein Tax. The intrinsic histone acetyltransferase activity of p300 is dispensable for activation of the IL-2 promoter, and the N-terminal 743 residues contain the minimal structural requirements for synergistic transactivation of the CD28RE-TRE, the IL-2 promoter, and endogenous IL-2 gene expression. Mutational analysis of p300 reveals differential structural requirements for the N-terminal p300 module by individual cis-elements within the IL-2 promoter. These findings provide evidence that p300 assembles at the IL-2 promoter to form an enhanceosome-like signal transduction target that is centrally integrated at the CD28RE-TRE element of the IL-2 promoter through specific protein module-targeted associations in activated T-cells.

In this report, we explore the mechanisms of targeting of p300 to the interleukin-2 (IL-2) promoter in response to mitogenic and oncogenic molecular signals. Recruitment of p300 by cAMP-responsive element-binding protein-Rel cross-talk at the composite CD28 response element (CD28RE)-TRE element of the IL-2 promoter is essential for promoter inducibility during T-cell activation, and CD28RE-TRE is the exclusive target of the human T-cell lymphotropic virus type I oncoprotein Tax. The intrinsic histone acetyltransferase activity of p300 is dispensable for activation of the IL-2 promoter, and the N-terminal 743 residues contain the minimal structural requirements for synergistic transactivation of the CD28RE-TRE, the IL-2 promoter, and endogenous IL-2 gene expression. Mutational analysis of p300 reveals differential structural requirements for the N-terminal p300 module by individual cis-elements within the IL-2 promoter. These findings provide evidence that p300 assembles at the IL-2 promoter to form an enhanceosome-like signal transduction target that is centrally integrated at the CD28RE-TRE element of the IL-2 promoter through specific protein module-targeted associations in activated T-cells.
The role of nuclear transcription factors and coactivators as primary targets of molecular signaling in eukaryotic cells has emerged as a fundamental concept in signal transduction biology (1)(2)(3). Detailed functional characterization of the molecular correlates that underlie the control of these events has been essential to our understanding of their importance in cellular homeostasis and has provided key insights into the manner in which they become deranged during oncogenic transformation. The study of molecular signaling events involved in the control of cytokine gene expression in activated T-cells has led to several important paradigms with relevant application to the understanding of signal transduction and gene regulation in various cell types (4 -7).
The interleukin-2 (IL-2) 1 gene is a key molecular signaling target in activated T-cells and plays a pivotal role in the control of the immune response. As such it has become a model system through which the fundamental mechanisms of stimulusevoked responses can be studied. Although we are far from a precise understanding of the "fine tuned" regulation of IL-2, there is general agreement that many of the molecular signaling events, set in motion during T-cell activation, converge on a 300-bse pair region just upstream of the transcription start site (4,8,9). This region is often referred to as the interleukin-2 proximal promoter. A common theme that has emerged from the study of the IL-2 promoter and other Ras-controlled genes is that they are modulated by the activity of one or more composite gene regulatory elements. These composite sites are targeted cooperatively by the interaction of two or more separate transcription factors, in a linked fashion, to provide both increased specificity and stability.
Recently it has been shown that the CD28 response element (CD28RE), one of the major Rel-B-controlled sites in the IL-2 promoter, functions as a composite site in conjunction with a 3Ј-sequence that interacts with members of the basic leucine zipper (B-Zip) family of transcription factors (10,11). This composite site, termed the CD28RE-TRE (CD28RE-AP1), is located at position Ϫ174 to Ϫ146 within the proximal IL-2 promoter and provides a sequence-specific regulatory interface where Rel and CREB-targeted molecular signaling events converge in synergy with p300 (10 -12).
Human T-cell lymphotropic virus type I is the etiologic agent of adult T-cell leukemia/lymphoma and the demyelinating syndrome, tropical spastic paraparesis (13,14). The genome of human T-cell lymphotropic virus type I encodes a 40-kDa transactivator protein, Tax, which drives oncogenic transformation of human T-lymphocytes and can act, in conjunction with Ras, to promote the transformation of rodent fibroblasts (15). In infected cells, Tax can act to increase the activation of target cellular genes containing nuclear factor (NF)-B and CREB/ATF sites by up-regulating the activity of both CREB/ ATF and Rel/NF-B transcription factors (16,17). IL-2 and the IL-2 receptor-␣ are well known targets of Tax activation in T-cells (15, 18 -20). Although the ability of Tax to up-regulate Rel/B function is thought to be the sole underlying mechanism for its action at the IL-2 promoter, many details of its true mode of regulation at the CD28RE-TRE remain debated (20,21). As a result, a precise accounting of the role of Tax at the IL-2 promoter continues to be elusive.
In this report we analyze the mechanism through which cross-talk between the Rel/B and ATF/CREB signaling pathways together with p300 control the transactivation of the IL-2 promoter at its central composite enhancers when challenged by both mitogens and the T-cell oncogene tax. We present a model predicting that a key target for the signal transduction pathways generated during T-cell activation is the assembly of a CD28RE-TRE directed "enhanceosome-like" complex at the IL-2 proximal promoter. This assembly is centrally linked to distinct domains of p300 through multiple, yet specific, combinatorial interactions with factors coordinately bound to its composite gene regulatory elements.

Plasmids
CD28RE-TRE-CAT, IL-2 CAT, NF-AT-CAT, HIVB-CAT, AP1-CAT, and ⌬56-CAT parental control reporter plasmids have been described (10,22). The NFIL-2A-CAT reporter was constructed by inserting two copies of the sequence AGCTTCACGATGTTTTACATATTACACATAT-TTTCAAAGA into the HindIII site upstream of the minimal fos promoter of ⌬56-CAT. The CRE-CAT was a generous gift from Dr. Maria Laura Avantaggiati. The UAS-CAT reporter plasmids contains five tandem copies of the UAS Gal4 recognition site linked to an E1B minimal promoter and was a generous gift from Dr. Irene Collins (NCI). The Tax expression vector has been described previously (23). The p300 expression vectors used in this study have been described previously (24). The Gal4-p300 truncation mutants have been described previously (25). The dominant interfering expression vector of Ras (N17Ras) was a generous gift from Dr. Silvio Gutkind (NIDCR). The CREB dominant negative vector is highly selective for CREB/ATF transactivation in T-cells and represses CREB-dependent transactivation without affecting AP-1-dependent transcriptional activation (10,26). The B-specific superdominant vector is a non-hydrolyzable I-B expression vector that selectively inhibits B-dependent transactivation without inhibiting transactivation from other cis-elements including the NF-AT site and has been described (27). Western blot analysis of lysates from cells transfected with cytomegalovirus-p300 expression vectors shows that coexpression with either Ras, CREB, or B dominant inhibitors does not alter the level of protein produced from the exogenous genes (not shown). The mutant Tax expression vectors pcTax M47, pcTax M22, and the pcTax wt control were a gift from Warner C. Greene (J. David Gladstone Institutes) and have been described extensively (16).
A simple screen for functional domains within the N-terminal 743 amino acids of p300 in the Gal4-p300 fusion constructs was designed by mutating 4 of 10 N-terminal sequences that were predicted to assume an ␣helical structure by the method of Chou and Fasman. The targeted regions were arbitrarily referred to as: helix I, residues 21-32; helix II, residues 72-85; helix VIII, residues 600 -616; and helix IX, residues 633-663 (see Fig. 5B). The ␣-helical structure was disrupted in either single or double mutants by the substitution of three proline residues for three consecutive residues in the center of each of the putative ␣-helical segments described above. The mutations were introduced by site-directed mutagenesis using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla CA). The substitution of alanine for Tyr-638 within the p300 KIX domain (Y638A mutant) was also accomplished by site directed mutagenesis, as described above, using the primers (TCTAGTTCTTTCTGGATCTT GGCGATTTTCTCAGCTAGAAGGT) and (ACCTTCTAGCTGAGAAAA-TCGCCAAGATCCAGAAAGAACTAGA). The IL-2 mutant TRE CAT reporter plasmid contains the IL-2 flanking region from Ϫ575 to ϩ47 with a 3-base substitution at the TRE portions of the CD28RE-TRE composite site (see Fig. 3A). This reporter mutant was constructed by site-directed mutagenesis of the IL-2-CAT plasmid using the primers GTTTAAAGAAATTCCAAAGCAACATCAGAAGAGGAAAAATG and CATTTTTCCTCTTCTGATGTTGCTTTGGAATTTCTTTAAAC and was provided by Dr. V. Doseeva. 2

Cell Culture and Transfection Assays
Jurkat T-cells were cultured in RPMI medium (Life Technologies, Inc.), 10% fetal calf serum, and 100 units/ml penicillin/streptomycin at 37°C in 5% CO 2 . Cells (1-2 ϫ 10 7 ) were resuspended in 250 l of ice-cold phosphate buffer with 2.5-8 g of reporter plasmid and 1-8 g of the indicated expression vectors followed by electroporation (200 V, 1180 microfarads). The cells were then resuspended in 5 ml of complete RPMI at 37°C. After transfection and resuspension, the cells were either left unstimulated or were stimulated with the indicated combinations of 1 M ionomycin (Calbiochem), 50 ng/ml phorbol 12-myristate 13-acetate (Sigma Chemical Co.), 2 g/ml phytohemagglutinin (Murex), a 1:1,000 dilution of monoclonal antibody 9.3 ascites (anti-CD28) (Bristol-Myers Squibb Research Institute, Seattle), and a 1:1000 dilution of monoclonal antibody OKT3 ascites (ATCC). The construction of Jurkat cells showing stable expression of the Gal4-p300 (amino acids 1-743) was established by selection of cells transfected with a pcDNA3 vector carrying the Gal4-p300 cDNA and a neomycin resistance gene. Cells were selected by growth in medium containing 1.6 mg/ml G418. Resistant pools of cells were screened for the G4-p300 expression by transient transfection with UAS-CAT and subsequent CAT assay. Positive pools of cells were cloned by limiting dilution in G418 and screened a second time by UAS-CAT reporter assay. All transient transfections studies were performed in duplicate or triplicate and, in all cases, are representative of at least two independent experiments. CAT activity in cell extracts was determined as described (10). Quantitation of the CAT activity was performed using a Molecular Dynamics PhosphorImager. Results are expressed as the mean Ϯ S.E. 300 Protein-DNA Binding Assays 0.3-1.5 g of affinity-purified ATF-1/CREB and c-Rel (10) was incubated for 15 min at 25°C in binding buffer containing 10 mM Hepes, pH 7.5, 60 mM NaCl, 0.05% Tween 20, 20% glycerol, 5 mM ␤-mercaptoethanol with 500 ng of biotinylated duplex oligonucleotides encoding the CD28RE-TRE sequence 5Ј-TGGGGGTTTAAAGAAATTCCAAAGAGT-CATCAGAAGAGG-3Ј and 5Ј-CCTCTTCTGATGACTCTTTGGAATTT-CTTTAAACCCCCA-3Ј at a final volume of 50 l. The reaction mix was then placed on ice and incubated with 2 l of [ 35 S]methionine-labeled Gal4-p300 truncation mutants for an additional 15 min. 20 l of a 1:1 slurry of streptavidin-agarose beads (Amersham Pharmacia Biotech) was added to the reaction and incubated for a final 15 min with shaking at 4°C. The samples were then loaded onto 200-l 10% sucrose cushions containing 1 ϫ binding buffer and pelleted by centrifugation at 500 ϫ g for 5 min The sucrose cushion and supernatant were removed, and the pellet was resuspended in 2.5 ϫ SDS-polyacrylamide gel buffer and separated on 7.5% polyacrylamide gels by SDS-polyacrylamide gel electrophoresis. Bound protein was visualized by autoradiography, and the percent of total bound labeled protein was determined on a Molecular Dynamics PhosphorImager.

Antibodies and Western Blots and Nuclear Extracts
Western blots were performed as described previously using anti-Gal4-DNA binding domain (CLONTECH and Santa Cruz) (10,29). Nuclear extracts were prepared from resting or stimulated Ficoll-purified Jurkat T-cells following electroporation as described previously (30).

Determination of IL-2
IL-2 levels were determined by enzyme-linked immunoabsorbent analysis using the Quantikine immunoassay kit (R&D Systems) according to the manufacturer's instructions.

Mitogen Induction of CREB-Rel Targeting Pathways Is an
Essential Component of Molecular Signaling at the IL-2 Proximal Promoter and Acts in Synergy with p300 -As reported previously (10), overexpression of p300 leads to synergistic increases in the transactivation of the IL-2 CD28RE-TRE gene regulatory element (Fig. 1A) in response to the well characterized T-cell mitogens phorbol ester and ionomycin (31,32). The dependence of this synergy on intact Ras-responsive CREB and B molecular signaling pathways is demonstrated by the disruption of the p300-controlled CD28RE-TRE transactivation by the activity of three previously described pathway-specific inhibitors: a dominant negative CREB/ATF (CREB leucine zipper with acidic extension) expression vector, a superdominant I-B (hydrolysis-resistant I-B-␣ mutant, Rel/B inhibitor) expression vector, and a dominant negative vector for Ras (Asn-17 substitution; see "Materials and Methods"). As shown in Fig. 1, B-D, blockade of Ras, Rel/B, or CREB/ATF effector signaling abrogates mitogen-induced transactivation of the CD28RE-TRE, in both the presence and absence of exogenous p300.

Tax Targets both B and CREB-dependent of Elements of CD28RE-TRE to Transactivate the IL-2 Promoter in Synergy
with p300 -The Tax oncoprotein has pleiotropic effects on a variety of immediate early and cytokine genes in activated T-cells. Several earlier studies have shown that the IL-2 promoter is a major target of the human T-cell lymphotropic virus type I Tax oncoprotein (18,20). Although it has been generally reported that the B pathway is the exclusive target of Tax action at the IL-2 reporter through its ability to increase the phosphorylation-dependent degradation of the I-B inhibitor, other laboratories have suggested roles for factors distinct from NF-B (20,21). Our observation that the TRE half of the IL-2 CD28RE-TRE element associates with ATF/CREB B-Zip factors (10) suggests that CREB signaling at this site may be a major mechanism through which Tax acts at the IL-2 promoter. To test this possibility, the action of Tax at an IL-2-CAT reporter construct, in which the TRE portion of the CD28RE-TRE composite element within the promoter was altered by sitedirected mutagenesis (IL2-mutTRE), was compared with that of the wild type promoter sequence (wt IL2). As demonstrated in Fig. 2A, induction of the IL-2 promoter by tax in either stimulated or costimulated cells is completely dependent on the integrity of the IL-2 CD28RE-TRE 3Ј-site. Reporter studies utilizing a single copy of the CD28RE-TRE show that Tax directly up-regulates the CD28RE-TRE, and this response is increased dramatically in the presence of exogenous p300 (Fig.  2B). Tax-dependent transactivation is repressed effectively by dominant negative Ras, the superdominant I-B, and the dominant negative CREB expression vectors (Fig. 2B), thus demonstrating that Tax targeting of the CD28RE-TRE and its synergy with p300 are dependent on both an intact B and CREB/ATF signaling pathway. This point is demonstrated further by using mutants of Tax which are deficient in transactivating either CREB/ATF or Rel/B-dependent genes (Fig. 2C). The Tax mutant M47 can activate Rel/B pathways but not the CREB/ATF pathways. Similarly, the M22 Tax mutant is capable of activating the CREB/ATF pathways but not the Rel/B pathways (16). As demonstrated in Fig. 2C, coexpression of either pathway-defective Tax mutant produces negligible activation of CD28RE-TRE in the presence of p300. This evidence, in combination with the dramatic effects of both the CREB and Rel/B-specific dominant negative vectors on Tax induction of CD28RE-TRE, shows that the primary mechanism of action of Tax at this element involves a dual targeting of CREB-Rel cross-talk in activated T-cells.
Tax Selectively Targets the IL-2 CD28RE-TRE and Not the NF-AT, NFIL-2A, or an AP-1 Consensus Sites in Activated T-cells-A comparison of the effect of Tax on the composite elements of the IL-2 promoters demonstrates that the CD28RE-TRE composite element of the IL-2 promoter is exclusively targeted by the Tax oncogene in activated T-cells. In ionomycin-and phorbol ester-stimulated cells, Tax coexpression significantly increases the transactivation of the CD28RE-TRE (Fig. 3A). In contrast, transactivation of the NF-AT site is actually repressed in the presence of Tax (Fig. 3B). NFIL-2A shows negligible influence by Tax (Fig. 3C), and like the NF-AT element, a consensus AP-1 site is repressed by Tax (Fig. 3D). Prior studies have shown that the mechanism of action of Tax at cellular CRE sites is via the enhanced recruitment of p300/ CBP to the KID domain of CREB bound to the CRE element (49). Fig. 3E is presented as a positive control to demonstrate that CREB molecular signaling is targeted effectively by Tax in mitogen-activated T-cells. In this experiment, low levels of the transfected Gal4-CREB fusion are synergistically transactivated when coexpressed with Tax. Notably, the level of expression is maximal with phorbol ester and ionomycin alone, as it shows little increase by CD28 costimulation.

The N-terminal 743 amino acids of p300 Are Necessary and Sufficient for the Mediation of CREB-Rel Cross-talk at the CD28RE-TRE and Potently
Transactivates the CD28RE-TRE, the IL-2 Promoter, and Endogenous IL-2 Expression-The transcriptional activity of p300 at individual promoters occurs in the absence of any intrinsic DNA binding activity. The main mechanism of p300 action occurs through its recruitment to individual cis-elements and promoters through protein-protein interactions with specific transcription factors and components of the basal transcriptional apparatus (34). The CREB/ATF binding domain has been localized previously to the KIX domain of p300/CBP, which encompasses amino acids 566 -647 in p300 (35, 36). Two separate Rel/B binding domains have been identified recently: one in the N-terminal 594 amino acids of p300, and a portion spanning residues 1572-2414 of the C terminus (37,38). To test which portions of p300 are necessary for effective recruitment to the CD28RE-TRE, we determined the in vitro domain requirements of p300 to be effectively recruited to the CD28RE-TRE. In this assay, biotinylated duplex oligonucleotides encoding a single copy of the CD28RE-TRE were incubated with in vitro translated Gal4 fusion constructs containing different portions of p300 (see upper panel Fig. 4A). Incubations were carried out in the presence or absence of affinity-purified CREB-Rel complexes obtained from phorbol ester-and ionomycin-activated Jurkat T-cells (10). DNA-bound versus free complexes were then separated by centrifugation with avidin-coated beads. In Fig. 4B, the portion of p300 containing the N-terminal 743 amino acids (and thus both Rel and CREB/ATF-1 binding domains) was effectively recruited to the CD28RE-TRE with increasing concentrations of purified CREB-Rel but did not associate with CD28RE-TRE in their absence.
A screen of the CREB-Rel-dependent recruitment of different p300 truncation mutants (described in Fig. 4A) to the CD28RE-TRE shows that the p300 mutant encompassing residues 1-743 is the most efficiently recruited (see Fig. 4C, bottom panel). Thus the portion of p300 which contains both the CREB/ATF-1 binding KIX domain and the N-terminal Rel binding domain is the most efficiently targeted to the CD28RE-TRE by purified CREB-Rel complexes. In addition, this portion of p300 (residues 1-743) is the most effective at inducing transcriptional activation of the CD28RE-TRE-CAT reporter when stimulated by phorbol ester and ionomycin (see Fig. 4D). The small amount of recruitment and transactivation potential demonstrated by the C-terminal 1945-2414 residues of p300 (see Fig.  4, C and D) is consistent with a single Rel-interacting domain identified previously in that region (38). The N-terminal 743 amino acids of p300 is also sufficient to superinduce transactivation of the entire proximal IL-2 promoter (Fig. 4E). Moreover, cells that stably express the N-terminal p300 residues exhibit increased expression of endogenous IL-2 (Fig. 4F).
Cis-element Recruitment and the Transactivation Potential of p300 Show Differential Dependence on Distinct Structural Domains within p300 -To determine the domain dependence of p300 for its ability to interact with the transcriptional activators and other cofactors that regulate the IL-2 promoter, a series of domain specific mutants was generated within the N-terminal 743 residues of p300 (see "Materials and Methods"). One class of mutations was a 1-amino acid substitution of alanine for tyrosine at position 638 within the KIX domain of p300 (Y638A p300). This mutation has been characterized previously to be critical for the interaction of CREB/ATF with the KIX domain (35). As shown in Fig. 5A, the N-terminal 743 amino acid residues of p300 are competent to activate the NF-AT and NFIL-2A sites of the IL-2 promoter, in addition to consensus B and AP-1 sites. In contrast, the KIX mutation impairs the ability of the p300 N terminus to induce the CD28RE-TRE, B, AP-1, and the NF-AT sites (Fig. 5, A and C  top). Interestingly, the induction of the NFIL-2A element by the KIX mutant is equivalent to wild type (Fig. 5A).
To identify other candidate domains of p300 which contribute to the regulation of the CD28RE-TRE and the IL-2 promoter, a set of p300 N-terminal mutants was derived by substituting 3 consecutive proline residues in the center of 9 predicted ␣-helical regions within the p300 743-amino acid N terminus (see Fig. 5B). With the exception of the double mutant II/IX (not shown), each of the p300 mutants showed expression comparable to that of the wild type fusion protein in activated Jurkat T-cells (see Fig. 5B, inset). The activity of these mutants, including the p300 Y638A substitution, was assayed for their function as Gal4 fusions when cotransfected with either the recruitment-dependent CD28RE-TRE or the recruitmentindependent (recruitment bypass) UAS-CAT reporters (Fig.  5C, top and bottom, respectively). As shown in Fig. 5C, although the KIX Y638A mutant is deficient in transactivating the CD28RE-TRE, it has near wild type activity in the recruitment bypass assay. This functional dissociation between the intrinsic transactivation potential and recruitment is also observed to a lesser extent with the helix IX mutants. Interestingly, both of these mutants contain substitutions in the ␣-helical regions of the KIX domain as determined by its solution structure. Of note is that mutation of the helix II domain of p300 results in a gain-of-function mutant that has above wild type inducibility in both the CD28RE-TRE recruitment assay and the recruitment bypass experiments, even in the absence of stimulation. This observation could partially reflect a slight increased stability of this protein over the wild type as suggested by the Western blot analysis (Fig. 5B, inset); however, truncation mutants of p300 (see "Materials and Methods"). TA, transactivation domains. C/H, cysteine/histidine-rich regions. Panel B, biotinylated CD28RE-TRE oligonucleotides (see "Materials and Methods") were incubated with 35 S-labeled p300 (amino acids 1-743) in the absence or presence of increasing amounts (0.03-1.5 g) of DNA affinity-purified ATF-1/CREB and Rel derived from mitogen-activated Jurkat T-cell nuclear extracts (10) (see "Materials and Methods"). After incubation, protein-DNA complexes were precipitated with avidin-coated beads. The pellets were washed, eluted into SDS-polyacrylamide gel buffer, and analyzed on a 7.5% Laemmli gel. Panel C, top, the indicated 35 S-labeled p300 truncation mutants were incubated with 0.75 g of affinity-purified ATF-1/CREB-Rel and biotinylated CD28RE-TRE nucleotides (see "Materials and Methods"). DNA-associated p300 complexes were visualized as described in panel B. Bottom, quantitation of the percent bound 35 S-labeled truncation mutant/total added (see "Materials and Methods"). Panel D, Jurkat T-cells were cotransfected in the absence or presence of 3 g of the expression vectors encoding the p300 truncation mutants described in panel A and 4 g of the CD28RE-TRE-CAT reporter and assayed for CAT activity as described under "Materials and Methods" following either no treatment or stimulation with phorbol 12-myristate 13-acetate and ionomycin (P/I). Panel E, Jurkat T-cells were transfected with 4 g of the IL-2-CAT reporter plasmid in the presence or absence of 3 g of p300 (amino acids 1-743) expression vectors and were either untreated or stimulated with phorbol ester and ionomycin prior to assaying for reporter activity. Panel F, parent Jurkat T-cells and Jurkat cells selected for stable expression of the N-terminal amino acids (1-743) p300 sequences (see "Materials and Methods") were evaluated for endogenous IL-2 expression and secretion in response to treatment with either no additions or the presence of the indicated concentrations of phorbol ester (PMA) with or without 1 M ionomycin. the helix II domain is also known to contain a nuclear hormone receptor binding motif, (LXXLL), which is a suspected target for steroid hormone-mediated repression of p300/CBP-controlled genes (39,40). DISCUSSION p300 plays a quintessential role as an adaptor by providing the necessary scaffolding to link diverse factors in an active transcriptional complex at the IL-2 promoter (Fig. 6). A prominent feature of the IL-2 enhanceosome which distinguishes it from others, such as the prototypical interferon-␤ promoter (41), is the capacity of the N-terminal module of p300 (743 amino acids) alone to assemble the enhanceosome and activate the IL-2 promoter. This difference is particularly notable because the N-terminal module not only lacks histone acetyltransferase activity but is missing interaction domains for polymerase II, TFIIB, and other basal factors (1,2). These observation suggests that the interaction of p300 with polymerase II (as shown in Fig. 6) may be a redundant or sequential feature of the enhanceosome which can be compensated for by interactions with other coactivator complexes such as mediator (for discussions, see Ref. 50). It is also possible that p300 may interact with promoters as a dimer, although no evidence of p300 oligomerization has ever been reported. Prior studies with isolated enhancer elements have shown differential requirements for the intrinsic histone acetyl-transferase activity of p300/CBP (47). However, few of these studies have extended this observation to whole functional promoters in vitro or in vivo, and none has shown the selective modular requirement of p300 domains as demonstrated in this work. To the contrary, some promoters, such as the interferon-␤ gene, are repressed by the N terminus of the p300/CBP module (48). Understanding the differential requirements of promoters for specific regulatory and enzymatic modules of p300 will be an important challenge in the future for determining how they are targeted selectively by molecular signaling events in activated T-cells. This is particularly true given that the N terminus of p300 is the region with least homology to CBP. In this regard, the use of the proline mutations to dissect the functional domains within the p300 N-terminal module (see Fig. 5) will be an effective tool in determining the structural requirements for FIG. 5. Cis-element recruitment and transactivation potential of p300 show differential dependence on structural domains within the N terminus of p300. Panel A, Jurkat cells were transfected as indicated with 4 g of human immunodeficiency virus-B, 4 g of 3X-AP-1, 4 g of 3X NF-AT, or 4 g of NFIL-2A reporter plasmids in the presence or absence of p300 1-743 expression vectors containing either the wild type or Y638A mutation in the p300 KIX domain (see "Materials and Methods"). Cells were either untreated or stimulated with phorbol ester (PMA) and ionomycin prior to harvest and assay for reporter plasmid activity. Panel B, schematic diagram of the proline-disrupted predicted ␣-helical segments of the p300 N terminus. Each putative helix was scored by its percent sequence identity to analogous regions in CBP. Helical wheel diagrams show the positions of the proline substitutions for helix II and helix IX (see "Materials and Methods"). The inset shows an immunoblot analysis of Gal4-p300 mutant expression relative to wild type, in stimulated Jurkat T-cells. Panel C, Jurkat T-cells were transfected with 4 g of CD28RE-TRE (top) or 2 g of UAS-CAT (bottom) reporter plasmids in the presence of 3 g of the indicated Gal4-p300 wild type and mutant expression vectors. The cells were either untreated or stimulated with phorbol ester and ionomycin prior to harvest and assay for CAT expression. the recruitment of this module at different promoters. Similar analyses will have to be performed for the domain structure of CBP N terminus.
A repetitive feature of the combinatorial nature of the IL-2 elements is the pairing of weak binding sites for ubiquitous factors (3)(4)(5). A fundamental trait that explains these differences lies in their coordinated ability to recruit p300. The recruitment of p300 is the primary factor that accounts for the dramatic synergy between the elements of the IL-2 promoter. It is also the main explanation for superinduced transactivation that is routinely elicited from multiple copy enhancer elements. This observation also accounts for the superinduction of the naked CD28RE-TRE element transactivation observed upon overexpression of p300 (10) (see Fig. 1). The single CD28RE-TRE and its bound elements have a much lower affinity for p300; therefore, enforced expression of p300 is able to drive transactivation from the CD28RE-TRE by mass action. Simply put, the IL-2 promoter is a tandem array of non-identical ciselements (a "composite of composites") that act together as a multivalent binding interface for specific p300 modules. By this thinking, different p300-controlled promoters should fall into categories that can be distinguished by the number and types of cis-elements that are necessary to recruit specific p300 modules to achieve transcriptional activation. The mutational analysis of the N terminus (Fig. 5) will help define the specificity of these binding interfaces at different promoters.
Although previous studies have shown a direct link between CREB/ATF activity at the CD28RE-TRE and CREB dominant negative transgenic studies show inhibition of IL-2 gene expression (10,42), the identification of the B-Zip components that interact with the CD28RE-TRE remains controversial (11). Other studies have implied that a Fos/Jun AP-1 complex is the active partner that interacts with the TRE (NFIL-2B) portion of the composite CD28RE-TRE (11). However, in addition to the fact that the presence of AP-1 has never been identified in an in vivo complex that binds the CD28RE-TRE, prior work by others has suggested that Fos/Jun does not bind to the CD28RE-TRE (43). Dominant negative CREB expression vectors repress the CD28RE-TRE without repressing a consensus AP-1 element (10), and past studies show that dominant negative Jun expression does not repress the CD28RE-TRE even though it potently inhibits transactivation of both NF-AT and consensus AP-1 enhancers (44). In another study, the CD28RE-TRE was shown to be repressed in T-cells by activation of ICER, an endogenous CREB dominant negative induced by elevations in intracellular cAMP (40). Finally, consistent with recent observations that Tax expression represses AP-1 transactivation by competing for AP-1 binding to p300/CBP (45), work in this report confirms that both a consensus AP-1 enhancer and NF-AT are repressed rather than stimulated by Tax. An explanation for the discrepancy between these findings and those published by Sun and co-workers (21), who suggest that Tax up-regulates NF-AT transactivation, may lie in the fact that the element in their study contained multiple tandem copies of only the CD28RE half of the CD28RE-TRE site in binding and transactivation studies with overexpressed NF-AT. Our findings do not exclude a role for AP-1 family members at the IL-2 promoter because there is overwhelming evidence that Fos/Jun plays a significant role as a congener at the distal NF-AT site. In fact, the B-Zip dimer composition at the CD28RE-TRE is not clear. An AP-1/CREB heterodimer (e.g. Jun/ATF-2 or CREB2/Fra) could be the true active B-Zip dimer at the CD28RE-TRE. In all likelihood, there is probably a stochastic distribution of B-Zip heterodimers that target the CD28RE-TRE in any given activated T-cell. Certainly the current evidence presented in this paper does more to rule in a role for CREB/ATF rather than rule out one for fos/jun at the CD28RE-TRE.
It was suggested previously that AP-1 may be a molecular target for T-cell anergy. In this prior study, a reporter plasmid driven by six tandem copies of the entire CD28RE-TRE sequence was shown to be down-regulated in anergic cells (46). Although the site was referred to as the IL-2 AP-1 element, this current work provides dramatic evidence in support of an expanded role for CREB family members and p300/CBP modules in T-cell anergy via the CD28RE-TRE.
Finally, the difference shown by the p300 N-terminal mutants in the recruitment and bypass transcriptional assays (Fig. 5C) suggests a prominent role for protein-protein interaction between p300 and other non-DNA-binding regulatory complexes. This is likely to be a promoter-specific process that distinguishes among various coregulator complexes serving as coactivators or corepressors at the IL-2 gene through a dynamic exchange linked to different modules of p300 (see 51 and references cited therein). A major imperative in the future will be to define the molecular rules of interaction which determine what type of p300-containing regulatory complexes will assemble at various target genes during T-cell activation and how these complexes are influenced by upstream signaling cascades.