Functional Dissection Identifies a Conserved Noncoding Sequence-1 Core That Mediates IL13 and IL4 Transcriptional Enhancement*

Conserved noncoding sequence (CNS)-1 has been shown to coordinately regulate the expression of the Th2 cytokine genes IL4, IL13, and IL5. We have used the interaction between CNS-1 and the human IL13 and IL4 promoters as a model to pursue the molecular mechanisms underlying CNS-1-dependent regulation of Th2 cytokine gene transcription. CNS-1 potently enhanced the activity of IL13 and IL4 promoter reporter vectors upon full T cell activation. Analysis of CNS-1 deletion mutants mapped enhancer activity to a short core (CNS-1-(270–337)) that contains three closely spaced cyclic AMP-responsive elements (CRE). CRE site 2 bound CRE-binding protein (CREB) and activating transcription factor (ATF)-2 in vitro and was essential for CNS-1-dependent up-regulation of IL13 transcription. Cotransfection of an IL13 reporter construct with expression vectors for wild type or mutant CREB and ATF-2 showed that CREB, but not ATF-2, regulates CNS-1 enhancer activity. Notably, chromatin immunoprecipitation analysis showed T cell activation recruits CREB and the coactivator CREB-binding protein (CBP)/p300 to the endogenous CNS-1. Moreover, CBP/p300 activity was essential for CNS-1-mediated enhancement of IL13 transcription. Collectively, these data define the region within CNS-1 responsible for enhancement of IL13 and IL4 transcription and suggest CREB/CBP-dependent mechanisms play an important role in facilitating Th2 cytokine gene expression in response to T cell receptor signaling.

genes, closely arrayed within 150 kb of human chromosome 5q31 and the syntenic region of mouse chromosome 11, typically demonstrate coordinated expression (1, 2), a feature critical for the emergence of a bona fide allergic phenotype in experimental and clinical models. However, the molecular mechanisms underlying the concerted expression of Th2 cytokines remained elusive despite intense investigation.
A breakthrough came as a result of comparative genomics analyses to identify noncoding regions highly conserved (Ն70% identity) between humans and evolutionarily distant mammalian species (3). These elements, abundant in the human genome, display characteristics indicative of regulatory function. In particular, they tend to demonstrate higher selective constraint than genomic regions that encode translated or noncoding RNAs (4,5) and contain short, alternating stretches of sequence with high or low divergence, a pattern typical of protein-binding sites (5).
The search for highly conserved noncoding sequences in ϳ1 Mb of human chromosome 5q31 identified several elements (3). The largest of these, conserved noncoding sequence (CNS)-1, mapped within the IL4/IL13 intergenic region of the Th2 cytokine locus. Deletion of CNS-1, either from a transgene or the native murine locus, led to a marked decrease in the expression of all three cytokine genes (3,6), establishing CNS-1 as a vital regulatory element for coordinated Th2 cytokine expression. Consistent with this role, epigenetic changes in the endogenous CNS-1 chromatin, including changes in levels of histone acetylation (7) and DNA methylation (8,9) and appearance of DNase I-hypersensitive sites (10), were found to accompany Th2 cytokine expression.
More recently, analysis of long range intrachromosomal interactions within the murine Th2 cytokine locus highlighted events that accompany the coordinated transcriptional regulation of Th2 cytokine genes and provided clues about the role CNS-1 may play in this process (11). This element was found to come in close spatial proximity with all three Th2 cytokine promoters in both T and non-T cells, suggesting it may be important for the acquisition of the initial "pre-poised" chromatin configuration of the Th2 cytokine locus. Of note, the physical interaction between CNS-1 and the Th2 cytokine promoters persisted through the T cell-and Th2 cell-specific stages of Th2 locus reorganization, pointing to an involvement of this element throughout the regulatory process. The interactions between CNS-1 and the Th2 cytokine promoters need to be further dissected to understand how CNS-1 contributes to their transcriptional regulation.
We chose the interaction between CNS-1 and the IL13 and IL4 promoters as a model to characterize the molecular mechanisms by which CNS-1 regulates Th2 cytokine gene expression in human CD4 T cells. Among the genes targeted by CNS-1, IL13 is essential to mediate Th2 effector functions critical to the pathogenesis of allergic inflammation (12)(13)(14), and IL4 is critical to initiate Th2 cell differentiation (reviewed in Ref. 15). Furthermore, expression of IL-13 and IL-4 was strongly decreased in CNS-1 Ϫ/Ϫ mice (6). We show here that CNS-1 is a potent T cell activation-dependent enhancer of the human IL13 and IL4 promoters. CNS-1 enhancer activity mapped to a short (68 bp) core that bound cyclic AMP-responsive element-binding protein (CREB) and the coactivator CREB-binding protein (CBP)/p300 in activated T cells and required these factors to enhance Th2 cytokine gene transcription.

EXPERIMENTAL PROCEDURES
DNA Constructs-p2.7IL13luc was created by PCR amplification of a 2666-bp region encompassing the human IL13 promoter (Ϫ2672 to Ϫ6, relative to the IL13 ATG; GenBank TM accession numbers AC004041 and L42080) using genomic DNA as a template. We selected this region based on the analysis of a panel of human IL13 promoter reporter constructs. 3 The PCR primers (IL13pro2.7F and IL13proR; all primer sequences are provided in supplemental Table 1) contained KpnI and NheI sites that were used to clone the IL13 promoter fragment upstream of the firefly luciferase gene in pGL3Basic (Promega). p369IL13 was created by amplification of the Ϫ369 to Ϫ6 region using primers IL13pro369F and IL13proR and p2.7IL13luc as template, followed by cloning into pGL3 Basic. p800IL4 contains 800 bp of human IL4 promoter sequence (Ϫ800 to Ϫ1 relative to the IL4 ATG) amplified by PCR using primers IL4pro800F and IL4proR with the human P1 clone H11 (GenBank TM accession number AC004039) as template. The amplified fragment was cloned into the SacI and NheI restriction sites of pGL3 Basic.
To generate the CNS-1 constructs, we initially amplified a 965-bp fragment (ϩ5604 to ϩ6568 relative to the IL13 ATG) of the human IL4/IL13 intergenic region encompassing CNS-1, and we cloned it into the SalI site located downstream of the luciferase gene in p2.7IL13luc. The boundaries of the human CNS-1 element were defined based on a sequence alignment with the murine IL4/IL13 intergenic region (GenBank TM accession number AC005742). The full-length 372-bp CNS-1 element (GenBank TM accession number AC004039; nucleotides 42330 -42701) was amplified by PCR (primers CNS1 1F and CNS1 372R ) and cloned 3Ј of the luciferase gene in p2.7IL13luc in both genomic and reverse orientations. Additionally, fulllength CNS-1 was cloned downstream of the reporter gene in the SalI site of p369IL13luc and p800IL4luc.
Cell Culture and Transfections-Jurkat T cells (ATCC clone E6-1) were cultured in RPMI 1640 supplemented with fetal calf serum (10%, HyClone), penicillin (100 units/ml), streptomycin (100 g/ml), and L-glutamine (2 mM). Jurkat T cells (1 ϫ 10 7 ) in log phase of growth were transfected with endotoxin-free plasmid preparations by electroporation (1 pulse, 240 V, 50 ms). Cells were transfected with either p2.7IL13luc (20 g) or with equimolar amounts of the indicated reporter vectors along with pRL-TK (20 ng; Promega) to control for transfection efficiency. Following electroporation, cells (5 ϫ 10 6 ) were cultured in the presence or absence of phorbol 12-myristate 13-acetate (PMA; 20 ng/ml, Sigma) and ionomycin (1 M; Sigma) or plate-bound anti-CD3 (2.5 g/ml; R&D Systems) and soluble anti-CD28 antibody (1.25 g/ml; R&D Systems) for 16 h. When indicated, Jurkat cells were transfected with expression vectors for WT or mutant CREB (30 ng), ATF-2 (30 ng), and E1A 12S (1 g) or equimolar amounts of empty pcDNA3 to control for total DNA content. Firefly and Renilla luciferase activity was determined using the dual luciferase assay system (Promega). In addition, the protein concentration for each cell lysate was quantitated with a BCA protein assay (Pierce). The relative luciferase activity (RLA) for each sample represents luciferase counts corrected for transfection efficiency and total protein content. Fold induction represents the ratio of RLA values between stimulated and unstimulated cells.
Nuclear Extract Preparation-Nuclear extracts were prepared from Jurkat T cells (1.5 ϫ 10 7 ) cultured in the presence or absence of PMA (20 ng/ml) and ionomycin (1 M) for 3 h. Cells were resuspended in buffer A (3 mM MgCl 2 , 10 mM HEPES, 40 mM KCl, 5% glycerol, 0.2% Nonidet P-40) supplemented with protease and phosphatase inhibitors (1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, 10 g/ml antipain, 10 g/ml pepstatin, 5 mM ␤-glycerophosphate, 1 mM NaF, 1 mM NaV, and 1 mM benzamidine) and incubated on ice for 5 min. Following centrifugation, the nuclear pellets were resuspended in buffer C (1.5 mM MgCl 2 , 20 mM HEPES, 420 mM NaCl, 25% glycerol, 0.2 mM EDTA) supplemented with protease and phosphatase inhibitors as indicated above. After a 30-min incubation on ice, the nuclear lysis solution was centrifuged, and the supernatant fractions were flash-frozen in liquid nitrogen and stored at Ϫ80°C. The protein concentration for each preparation was quantitated with a BCA protein assay (Pierce).
Electrophoretic Mobility Shift Assay (EMSA)-Singlestranded complementary oligonucleotides were annealed and PAGE-purified. Annealed oligonucleotides were end-labeled with [␥-32 P]ATP with T4 polynucleotide kinase. EMSA were performed with 10 g of nuclear extract in binding buffer (100 mM NaCl, 10% glycerol, 200 ng/l bovine serum albumin, 50 ng/l poly(dI-dC), 10 mM HEPES (pH 7), 0.1 mM EDTA, 0.25 mM dithiothreitol, 0.6 mM MgCl 2 ). For competition or supershift assays, the indicated unlabeled oligonucleotide competitor (100-fold molar excess) or antibody (2 g) was added 30 min prior to addition of radiolabeled probe. Following addition of the radiolabeled probe, the samples were incubated for 30 min at room temperature and loaded onto a 5% (w/v) polyacrylamide gel. Electrophoresis was performed at a constant 19 mA for 6 h at 4°C, and the gels were dried prior to autoradiography. Antibodies used for supershift analysis included a polyclonal and a monoclonal anti-CREB (C-21 and X-12, respectively), a polyclonal anti-Jun (D), and a monoclonal anti-ATF-2 (F2BR-1), all from Santa Cruz Biotechnology. Normal rabbit IgG (Upstate Biotechnology, Inc.) or a monoclonal anti-STAT1 antibody (C-136, Santa Cruz Biotechnology) was used as control. The DNA sequences corresponding to oligonucleotide competitors and probes are provided in supplemental Table 3.
Chromatin Immunoprecipitation (ChIP)-Jurkat T cells, unstimulated or treated with PMA (20 ng/ml) and ionomycin (1 M) for 3 h, were incubated with formaldehyde (1%) for 10 min at 37°C. Glycine was added to a final concentration of 125 mM to halt the cross-linking. Cells were harvested, washed with 1ϫ PBS supplemented with protease inhibitors (1ϫ EDTA-free Complete protease inhibitor mixture (Roche Applied Science) and 1 mM phenylmethylsulfonyl fluoride), and lysed in ChIP lysis buffer (Upstate Biotechnology, Inc.) supplemented with protease inhibitors as above. Chromatin was sheared by sonication (five times with 10-s pulses, 30% maximum; Microson XL200, Misonix) and diluted 10-fold in ChIP dilution buffer (Upstate Biotechnology, Inc.) supplemented with protease inhibitors. An aliquot of chromatin (ϳ5-6 ϫ 10 6 whole-cell equivalents) was set aside and used as input DNA. The remaining sample was precleared with salmon sperm DNA/protein A-agarose slurry (Upstate Biotechnology, Inc.) and divided into aliquots (5 ϫ 10 7 whole-cell equivalents) for immunoprecipitation. Immunoprecipitation reactions were performed with 10 g of antibody specific for CREB (C-21), CBP (A-22), p300 (N-15) (Santa Cruz Biotechnology), or with normal rabbit IgG (Upstate Biotechnology, Inc.) overnight at 4°C with rotation.
The chromatin-antibody complexes were collected with salmon sperm DNA/protein A-agarose slurry and washed sequentially with low salt wash, high salt wash, LiCl wash, and TE (ChIP Assay kit; Upstate Biotechnology, Inc.). The chromatin-antibody complexes were eluted (1% SDS, 0.1 M NaHCO 3 ), and the DNA-protein cross-links were reversed at 65°C overnight. All samples were recovered by phenol/chloroform extraction and ethanol precipitation using glycogen (20 g) as a carrier. Real time PCR was performed with the QuantiTect SYBR Green PCR kit (Qiagen) on an ABI Prism 7900 sequence detection system. PCRs were performed in triplicate under the following cycling conditions: 15 min at 95°C followed by 40 cycles of 15 s at 95°C, 30 s at 57°C, and 30 s at 72°C. Dissociation curve analysis and agarose gel electrophoresis confirmed amplification of a single 152-bp product using primers (CNS1ChIPF, 5Ј-CACAGCGTCGTTCAGAAACAC-3Ј; CNS1ChIPR, 5Ј-CAGCCCCCGCACAGTTGT-3Ј) which target nucleotides 221-372 of CNS-1. For each experiment, serial dilutions of input DNA were used to generate a standard curve. Results were expressed as the ratio between the number of targets immunoprecipitated with specific antibodies and the number of targets immunoprecipitated with control IgG.
Comparative DNA Sequence Analysis-Genomic DNA sequence for the human IL13/IL4 locus and the syntenic regions in eight additional mammalian species were obtained either from current whole genome reference sequences (Pan troglodytes, Bos taurus, Canis familiaris, Rattus norvegicus, and Mus musculus) or the National Institutes of Health Intramural Sequencing Center (Papio anubis, Callithrix jacchus, and Otolemur garnettii; www.nisc.nih.gov). Accession numbers are provided in Supplemental Table 4. Multiple sequence alignments were generated using the MUltiple sequence Local AligNment and conservation visualization tool (MULAN) (17), which uses a local alignment strategy with the threaded block set aligner and utilizes the phylogenetic relationships of the sequences provided to build the multiple sequence alignment. The alignment was exported in FASTA format to GeneDoc for visualization.

CNS-1 Is a Potent Enhancer of IL13 and IL4 Promoter Activity-
CNS-1 is strategically positioned at the heart of the Th2 locus ( Fig. 1), consistent with its ability to physically interact with the Th2 cytokine gene promoters throughout the region (11). To start defining the molecular mechanisms underlying the CNS-1-dependent regulation of Th2 cytokine transcription, the CNS-1 element (372 bp) was cloned 3Ј of the luciferase reporter gene in a construct driven by a 2.7-kb fragment encompassing the promoter for human IL13. This construct (p2.7IL13/CNS-1-(1-372)) was transiently transfected into Jurkat T cells, and its activity was compared with that of an IL13 promoter construct lacking CNS-1 (p2.7IL13). Transfected cells were cultured for 16 h in the presence or absence of PMA (20 ng/ml) and ionomycin (1 M), a combination of stimuli that results in a Ͼ100-fold increase of IL13 mRNA levels in Jurkat T cells (data not shown). Fig. 2A shows that under basal conditions the IL13 promoter and CNS-1 were essentially inactive. Stimulation with PMA and ionomycin resulted in strong (16-fold) activa-tion of IL13 transcription, which was amplified dramatically when CNS-1 was linked to the IL13 promoter. Although the activity of a shorter, 369 bp, human IL13 promoter fragment was comparable with p2.7IL13, the CNS-1-dependent up-regulation of transcription was less than half compared with the full-length promoter (data not shown), indicating additional IL13 5Ј regulatory elements are required for optimal interactions with CNS-1. Fig. 2A also shows CNS-1 activity was orientation-independent, because CNS-1 enhanced IL13 transcrip-tion to comparable extents when cloned in the genomic and reverse orientation. Collectively, these results demonstrate in our model CNS-1 acts as a robust bona fide IL13 enhancer.
Of note, optimal synergism between CNS-1 and the IL13 promoter required both PMA-and ionomycin-delivered signals. Fig. 2B shows that although CNS-1 did enhance IL13 transcription in response to PMA alone, activity remained marginal compared with that induced by the two stimuli in combination. T cell activation with anti-CD3 and anti-CD28 antibodies resulted in CNS-1-driven enhancement of IL13 promoter activity similar to that observed with PMA and ionomycin (Fig. 2C). These data show the main signaling pathways that lead to full T cell activation converge on CNS-1.
Multiple lines of evidence support a role for CNS-1 in the regulation of IL4 expression (3,6,18,19). In particular, addition of a 2.7-kb region surrounding CNS-1 to a luciferase transgene driven by an 800-bp murine IL4 promoter markedly enhanced luciferase production (19). We tested whether CNS-1 could directly influence the activity of a human 800-bp IL4 promoter. Fig. 2D shows p800IL4 was modestly (8-fold) induced upon T cell activation but was strongly up-regulated by CNS-1. These results indicate that CNS-1 enhances transcription of two major genes within the Th2 cytokine locus.  Fig. 3B shows that remarkably the entire enhancer activity of CNS-1 mapped to a single fragment, CNS-1-(221-372), which encompasses the IL13 distal region of the element. We note that CNS-1-(221-372) was nearly twice as active as full-length CNS-1, suggesting the CNS-1 enhancer core may lie within a region that constrains its activity.

Identification of a 68-bp Enhancer Core within CNS-1-To
To define more closely the boundaries of the CNS-1 enhancer core and guide our subsequent analysis of DNA/pro-   Fig. 3C). Fig. 3D shows that virtually all of the IL13 enhancer activity resided within CNS-1-(270 -337). A highly similar pattern was observed when we studied the interactions between discrete CNS-1 domains and the IL4 promoter (Fig. 3D). These results define a 68-bp region of CNS-1, encompassing nucleotides 270 -337, as a potent enhancer for two distinct Th2 cytokine gene promoters.
The CNS-1 Enhancer Core Contains Binding Sites for CREB, ATF-2, and Jun Proteins-To identify the trans-acting factors involved in the enhancer activity of CNS-1-(270 -337) , we analyzed patterns of DNA/protein interactions by EMSA. Nuclear extracts prepared from Jurkat T cells cultured with or without PMA (20 ng/ml) and ionomycin (1 M) for 3 h were incubated with 32 P-labeled probes corresponding to nucleotides 270 -303, 287-320, or 318 -337 of CNS-1 (Fig. 4A). Fig. 4B shows that competition experiments with unlabeled self-related or unrelated oligonucleotides identified four specific nucleoprotein complexes (lanes 1-3 and 19 -21), three of which (complex I-III) bound the 270 -303-nucleotide region (Fig. 4B, left panel). Complex I and II, but not complex III, also bound the overlapping nucleotides 287-320. No additional interactions were detected in this region (Fig. 4B, center panel). Complex IV was detected using the 318 -337 probe (Fig. 4B, right panel).
Comparative analysis of the CNS-1-(270 -337) nucleotide sequence across distant species (Fig. 4C) and prediction of putative transcription factor-binding sites identified three motifs (CRE 1-3; Fig. 4B) each partially homologous to a CRE (TGA-CGTCA) (20) and the related AP-1 family consensus sequence (TGA-CTCA) (21). Antibody supershift experiments were therefore performed in order to test whether the complexes binding to CNS-1-(270 -337) contained CRE-interacting proteins (CREB and ATF) and/or AP-1 family members. Fig. 4B shows that complex I, a faint band up-regulated by stimulation, contained ATF-2 because it was supershifted by an ATF-2-specific antibody (lane 13) but not by an anti-CREB (lane 11) or an anti-Jun antibody (lane 12). The constitutively expressed complex II was formed by CREB because addition of a polyclonal anti-CREB antibody supershifted the complex completely (Fig. 4B, lanes 4 and 11). A monoclonal anti-CREB antibody, which does not cross-react with other CREB family members, also altered the mobility of this complex (data not shown). Complex III appeared to contain CREB because an anti-CREB antibody (Fig. 4B, lanes 4 and  11) reduced the intensity of this complex. Jun proteins may also be part of complex III, because the corresponding band became fainter upon addition of a Jun-specific antibody (lanes 5 and 12) and was competed by an AP-1 consensus oligonucleotide (data not shown). Migration of none of these complexes was affected by control IgG or STAT1 antibodies (Fig. 4B, lanes 7, 14, 15, and  24). Finally, the constitutive complex IV also contained CREB. Indeed, preincubation with the polyclonal or monoclonal anti-CREB antibody supershifted this complex completely (Fig. 4B,  lanes 22 and 23), whereas antibodies to STAT1 (lane 24), Jun, or ATF-2 (data not shown) did not. Collectively, these data demonstrate the CNS-1 enhancer core contains binding sites for CREB, ATF-2, and Jun proteins.
To position the observed complexes more closely within the CNS-1 enhancer core and test their interactions with the CRE motifs, each CRE site was mutated, alone or in combination (Fig. 5A). EMSA analysis revealed that each CRE site supported the formation of distinct CREB-containing complexes (Fig. 5B). Mutation of CRE 1 resulted in the loss of complex III (Fig. 5B,  lane 2). Both complex I and complex II were lost when the second CRE motif was mutated (Fig. 5B, lane 3). Despite the striking proximity of the first two CRE sites and their organization on opposite DNA strands, mutation of either one of these two sites did not appear to affect the formation of complexes on the other site. Notably, no residual binding was detected when both CRE 1 and CRE 2 were mutated (Fig. 5B, lane 4). Finally, mutation of CRE 3 abolished the binding of complex IV (Fig. 5B,  lane 6). Collectively, our results show CREB interacts with each of the three CRE sites in CNS-1-(270 -337), whereas Jun and ATF-2 bind selectively to CRE 1 and CRE 2, respectively (Fig. 5C).
To evaluate the contribution of each CRE motif to CNS-1 enhancer function, the CRE site mutations already characterized by EMSA were introduced in the p2.7IL13/CNS-1-(270 -337) reporter construct. Fig. 6 shows that mutation of the CRE 1 and CRE 3 sites reduced CNS-1 activity substantially but only partially (54.5 and 79%, respectively), whereas mutation of CRE 2 was sufficient to abolish CNS-1-dependent IL13 enhancement. In view of the topology of the CRE motifs, these results suggest that although each CRE site contributes to CNS-1 activity, CRE 2 may be required to coordinate the formation of a supramolecular complex critical for optimal CNS-1-induced enhancement of IL13 transcription.
CREB, but Not ATF-2, Regulates CNS-1 Enhancer Activity-Because CRE 2 and, to a lesser extent, CRE 3 were essential for CNS-1-dependent transcriptional enhancement, we then investigated the role of the CRE 2/3-binding proteins CREB and ATF-2 in the regulation of CNS-1 activity. To this purpose, the p2.7IL13/CNS-1-(270 -337) reporter construct was cotransfected with expression vectors encoding either WT or mutant forms of these factors. Specifically, pCMV-CREB133 encodes a dominant negative CREB variant containing a Ser 3 Ala substitution at position 133, a residue vital to CREB-mediated transactivation (22). pATF-2 ⌬2-107 encodes a truncated ATF-2 protein that lacks the N-terminal trans-activation domain (23) but retains the ability to bind CNS-1 CRE 2 (data not shown). Fig. 7 shows that expression of CREB133 markedly (47%) reduced T cell activation-dependent CNS-1 enhancer activity, whereas cotransfection of ATF-2, either WT or mutant, failed to affect IL13 transcription. CREB133-dependent inhibition was CNS-1-specific because IL13 promoter activity was virtually unaltered (data not shown). These results support a role for CREB, but not ATF-2, in the molecular events underlying CNS-1-mediated transcriptional enhancement.
T Cell Activation Recruits CREB and CBP/p300 to the Endogenous CNS-1-Our in vitro dissection identified CREB as a factor that binds functionally critical sites in the CNS-1 enhancer core and directs robust up-regulation of IL13 and IL4 transcription. Chromatin immunoprecipitation assays were therefore performed to test whether CREB-containing complexes dock onto the endogenous CNS-1. CREB protein-DNA complexes were immunoprecipitated from Jurkat T cells, resting or activated with PMA and ionomycin. Real time PCR was performed to detect a 152-bp region of CNS-1 (nucleotides 221-372) that spans the enhancer core. Although only low levels of target were immunoprecipitated with an anti-CREB antibody under basal conditions, Fig. 8 shows that T cell activation strongly increased CREB binding to CNS-1.
Activation-dependent CREB phosphorylation at serine 133 is known to foster the recruitment of the coactivator CBP (24) and its paralogue p300 (25), which augment CREB-mediated gene transcription. The inhibitory effect of the CREB133 mutant on CNS-1 activity (Fig. 7) raised the possibility these coactivators may contribute to CREB-mediated CNS-1 regulation. We there-  FEBRUARY 9, 2007 • VOLUME 282 • NUMBER 6 fore used chromatin immunoprecipitation to test whether CBP/p300 was recruited to CNS-1 in vivo. Fig.  8 shows that occupancy of CNS-1 by CBP-or p300-containing complexes was limited in unstimulated T cells. However, T cell activation resulted in robust recruitment of both factors to CNS-1.

DISCUSSION
Progression of naive CD4 T cells along the Th1 or Th2 differentiation pathway is a multistage process contingent upon antigenic T cell stimulation in an instructive cytokine milieu. In vivo studies defined CNS-1 as a regulatory element critical for optimal expression of Th2 cytokine genes and Th2 differentiation (3, 6) but did not dissect the molecular mechanisms underlying the role of CNS-1 in these processes. Our results characterize CNS-1 as a potent enhancer of human IL13 and IL4 transcription. Surprisingly, activity mapped to a discrete, short domain at the IL13 distal end of the element, raising the possibility that the regulatory properties of CNS-1 may be compartmentalized. CNS-1 appears to control several facets of Th2 cytokine locus expression, from modulation of chromatin accessibility (3,6) to positioning of the locus within a repressive nuclear domain in Th1 cells (28). It remains to be determined whether any of these specialized functions of CNS-1 also reside(s) in the enhancer core or if they map elsewhere, a possibility supported by the finding that the IL13 proximal region of CNS-1 is comparatively even more conserved across mammalian species than the enhancer core itself (data not shown). Surprisingly, the CNS-1 enhancer core does not encompass previously described binding sites for GATA3 (29) or Ikaros (28), even though ectopic overexpression of GATA3 in murine  Th1 cells was sufficient to establish DNase I hypersensitivity within the endogenous CNS-1 chromatin (29,30). The relationship between GATA3-induced chromatin remodeling and the activity of the CNS-1 enhancer core remains to be determined.
Our experiments identified CRE site 2 and 3 and the CRE 2/3-binding protein CREB as major regulators of CNS-1 enhancer activity. This finding was somewhat unexpected because a recent analysis of mice in which CREB and ATF-1 had been deleted selectively in T cells showed IL-4 expression to be preserved (31). However, a significant proportion of spleen and lymph node T cells (24 and 12%, respectively) in the knock-out mice still expressed CREB, raising the possibility that residual CREB activity was sufficient to support IL4 mRNA expression in that experimental model. Furthermore, and perhaps more importantly, deletion of CNS-1 delayed and reduced, but did not abrogate, expression of IL-13 and IL-4 (6), suggesting CNS-1 contributes to, but is not absolutely required for, Th2 cytokine expression.
CREB is a ubiquitous transcription factor that resides constitutively in the nucleus (20). CREB activation is known to be necessary for transcriptional regulation and contingent upon phosphorylation of Ser-133, which occurs typically, although not exclusively, in response to elevated intracellular cAMP (20). In line with these requirements, our experiments showed stimulation of T cells with cAMP resulted in vigorous (4.4-fold) CNS-1-dependent enhancement of IL13 promoter activity (data not shown) and expression of a dominant negative CREB variant, which cannot be phosphorylated on Ser-133, reduced activity by nearly half.
Although T cell receptor engagement induces CREB phosphorylation, maximal CREB trans-activation potential is   achieved only if the antigen-dependent signal is coupled with CD28-mediated costimulation (32,33). Consistent with the two-signal requirement for CREB trans-activation, our data revealed full T cell activation was essential for maximal CNS-1 activity. Signaling via the T cell receptor and the costimulatory pathway is required for a productive association between CREB and the transcriptional coactivator CBP (33), which recruits the RNA polymerase II holoenzyme (34). Fostering the assembly of this complex may represent a mechanism by which CREB contributes to CNS-1-dependent transcriptional enhancement in activated T cells. This possibility is strongly supported by our finding that CREB, CBP, and p300 were recruited to the endogenous CNS-1 in response to T cell stimulation, and CBP/p300 activity was necessary for CNS-1-mediated enhancement of IL13 transcription. In addition, CBP/p300 can mediate transcriptional activation through intrinsic histone acetyltransferase activity (35,36). Because the latter is required for CREBmediated transcriptional activation in several models (37)(38)(39), the role of CREB in CNS-1 function may also involve CBP/ p300-mediated histone modifications. Indeed, deletion of CNS-1 abrogated the basal acetylation of histone H3 at the IL4 and IL13 promoters in naive CD4 T cells (28), a permissive modification associated with rapid gene transcription (15).
Recent work on long range intrachromosomal interactions within the murine Th2 locus demonstrated CNS-1 behaves as a versatile element that interacts with distinct Th2 cytokine promoters in vivo (11). This essential property was captured by our experimental model, which revealed CNS-1-dependent enhancement of both IL13 and IL4 transcription in T cells. That the CNS-1 enhancer core relied on ubiquitously expressed proteins such as CREB and CBP/p300 for its activity suggests the role of this element in Th2 cytokine transcription may be a permissive one, which links Th2 locus regulation with antigenic T cell stimulation and complements Th2-specific regulatory mechanisms. The CRE sites clustered within the CNS-1 enhancer core could serve as a platform to recruit basal machinery and coactivators to confer transcriptional competence to and/or augment transcription from its target Th2 cytokine promoters. In vivo analysis of CNS-1 will be required to further elucidate the interactions between CNS-1 and Th2-specific transcriptional regulators and dissect the contribution of specific CNS-1 domains to the complex molecular events that orchestrate Th2 cytokine gene expression.