Identification and Characterization of a Critical CP2-binding Element in the Human Interleukin-4 Promoter*

Expression of cytokine genes in T cells is thought to result from a complex network of antigen- and mitogen-activated transcriptional regulators. CP2, a factor homologous to Drosophila Elf-1 and previously found to be a critical regulator of several viral and cellular genes in response to developmental signals, is rapidly activated in T helper (Th) cells in response to mitogenic stimulation. Here we show that overexpression of CP2 enhances interleukin (IL)-4 promoter-driven chloramphenicol acetyltransferase expression, while repressing IL-2 promoter activity, in transiently transfected Jurkat cells. A CP2-protected element, partially overlapping the nuclear factor of activated T cell-binding P2 sequence, was required for IL-4 promoter activation in CP2-overexpressing Jurkat cells. This CP2-response element is the site of a cooperative interaction between CP2 and an inducible heteromeric co-factor(s). Mutation of conserved nucleotide contacts within the CP2-response element prevented CP2 binding and significantly reduced constitutive and induced IL-4 promoter activity. Expression of a CP2 mutant lacking the Elf-1-homology region of the DNA-binding domain inhibited IL-4 promoter activity in a dominant negative fashion in transiently transfected Jurkat cells. Moreover, overexpressed CP2 markedly enhanced, while its dominant negative mutant consistently suppressed, expression of the endogenous IL-4 gene in the murine Th2 cell line D10. Taken together, these findings point to CP2 as a critical IL-4 transactivator in Th cells.

Interleukin (IL) 1 -4 is a pleiotropic cytokine that modulates the differentiation and the biologic activities of virtually all cells of hematopoietic origin (1). IL-4 is typically expressed in (and, at the same time, promotes the differentiation of) the T helper 2 (Th2) functional subset of CD4 ϩ T cells, thereby playing a pivotal role in the regulation of humoral and allergic responses. Conversely, Th1 cell-associated cytokines, such as IL-2 and interferon-␥, are central to the development of cellmediated, delayed type, and autoimmune responses (2). The biochemical and molecular mechanisms accounting for polarized expression of cytokine genes in T cells have been the focus of a number of studies over the past few years (3). Calcineurinmediated activation of members of the nuclear factor of activated T cell (NFAT) family of transcription factors has been associated with antigen-dependent cytokine gene expression in both Th1 and Th2 cells (4). However, NFAT-directed IL-4 transcription is preferentially induced in Th2 cells, presumably due to the involvement of Th2-restricted co-factors (5,6).
It is likely that the commitment toward an IL-4-producing, Th2 phenotype is the outcome of a complex, dynamic interaction of multiple transcriptional regulators rather than the effect of a single protein. Irrespective of the cytokine milieu leading to preferential Th1 or Th2 differentiation (2), priming of T cells by antigen or mitogens is a requirement for IL-4 but not IL-2 production (7). In fact, the IL-4 gene is only transcribed in actively dividing T cells, while IL-2 expression occurs independently of the cell cycle (8). T cell priming affects the expression and/or activation of a number of factors presumably involved in the regulation of cytokine genes. For example, primed T cells express higher levels of NFAT-1 and support greater NFAT-directed transcription than naive T cells (9). An earlier study also showed rapid and marked up-regulation of the DNA binding activity of the transcription factor CP2 following mitogenic stimulation of peripheral blood naive T cells, suggesting the involvement of this protein in the activation of immediate and/or early genes in these cells (10).
CP2, also known as leader-binding protein (LBP)-1c (11) or late simian virus 40 factor (12), is the prototypical member of a novel family of mammalian proteins sharing a high degree of similarity to Elf-1, a Drosophila melanogaster tissue-specific factor encoded at the embryonic lethal locus Grainyhead (11,13). By interacting with a hyphenated sequence composed of two directly repeated 4-base pair (bp) motifs separated by a 6-bp linker (CNRG-N 6 -CNR(G/C)) (11,14), CP2 has been reported to regulate transcription of a number of viral and cellular genes in response to developmental signals, including those encoding for rat ␥-fibrinogen (15), mouse ␣-globin (16), simian virus 40 (12), human immunodeficiency virus-1 (11), herpes simplex virus-1 (17), ␥and ⑀-globin, and the human ␤-like globin gene cluster (18), major histocompatibility complex class II Ea (19), and human c-fos and mouse thymidylate synthase (20). However, its contribution to cytokine gene expression in T cells has not been explored to date. In this study, we investigated the contribution of CP2 to the mechanisms regulating expression of the two pivotal T cell-restricted cytokines IL-2 and IL-4.

EXPERIMENTAL PROCEDURES
Cells-A line of Jurkat T cells, which constitutively expresses IL-4 and produces IL-2 and interferon-␥ following activation, was a gift by Dr. Jack L. Strominger (Harvard University). Cells were cultured in RPMI (Mediatech, Herndon, VA), containing 2 mM L-glutamine, 10% heat-inactivated fetal bovine serum (Gemini Bio-Products, Calabasas, CA), and 50 g/ml gentamicin (complete medium). Aliquots of cells frozen at early passages were recovered from liquid nitrogen and used for experiments between 1 and 6 weeks after thawing. The murine conalbumin-specific Th2 clone D10.G4.1 (D10; American Type Culture Collection, Manassas, VA) was maintained in culture by biweekly restimulation with antigen and irradiated (2,000 rads) syngeneic splenocytes as antigen-presenting cells in RPMI supplemented with 10% heat-inactivated fetal bovine serum, 2 mM L-glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, and 50 M 2-mercaptoethanol. IL-4 expression in these cultures was verified using a sensitive commercial enzyme-linked immunosorbent assay (Cytoscreen-US, BIOSOURCE International, Camarillo, CA).
Plasmids-The IL-4 promoter fragments used in this study were generated by polymerase chain reaction using human genomic DNA as a template and cloned into the HindIII and XbaI sites of a Bluescript vector (Stratagene Cloning Systems, La Jolla, CA). An XbaI-tailed oligonucleotide, corresponding to bp ϩ36 to ϩ55 of the human IL-4 gene, and one of eight HindIII-tailed oligonucleotides, corresponding to bp Ϫ741 to Ϫ722, Ϫ311 to Ϫ292, Ϫ265 to Ϫ246, Ϫ225 to Ϫ206, Ϫ175 to Ϫ156, Ϫ145 to Ϫ126, Ϫ95 to Ϫ76, or Ϫ65 to Ϫ46, were used to introduce the appropriate restriction sites at the invariant 3Ј-and at the 5Ј-ends, respectively. After sequence verification, each fragment was inserted into the HindIII and XbaI sites of the chloramphenicol acetyltransferase (CAT)-expressing reporter vector, pCAT-Basic (Promega, Madison, WI) to construct the corresponding reporter plasmids IL-4.741, , and IL-4.65 (21). An IL-4.225 plasmid carrying two point mutations within the CP2-response element (CPRE) was generated by site-directed mutagenesis (QuikChange; Stratagene). The following mutagenic primer (mutations are underlined) and its complement were synthesized (Genosys Biotechnologies, The Woodlands, TX): 5Ј-CATTTTCCTATTGGTAT-TATTTCACAGGAACATTTTACCTG-3Ј. The CP2 expression plasmid was generated by insertion of the murine CP2 cDNA into the HindIII and NotI sites of a pRc/CMV vector (Invitrogen, San Diego, CA) (14). A CP2 mutant lacking the Elf-1-homologous region of the DNA-binding domain (⌬Elf-1) was generated from a full-length template as described (22) and inserted into the same vector. The IL-2.15⌬CX reporter plasmid, bearing bp Ϫ319 to ϩ52 of the human IL-2 gene in pCAT-Basic (23), has been kindly donated by Dr. Gerald R. Crabtree (Stanford University, Stanford, CA).
Transfections and Reporter Gene Analysis-Jurkat cells (0.5-1 ϫ 10 6 /ml) were transfected with 1 g of reporter plasmids and 2 g of expression plasmids by 48-h culture in 5 ml of complete medium containing 5.2 mg/ml LipofectAMINE (Life Technologies, Inc.) as described (24). Equal amounts of the corresponding noncoding vectors were added to control samples to yield a constant amount (3 g) of DNA in each transfection. Where indicated, cells were stimulated 16 -18 h before harvest. 40 -80 g of proteins extracted from each sample were analyzed for CAT concentration using a commercial enzyme-linked immunosorbent assay (Roche Molecular Biochemicals). In all experiments, CAT concentrations were normalized by total protein content, measured in cell lysates using the Bradford reagent (Bio-Rad). D10 cells (5 ϫ 10 7 cells/ml) were separated from irradiated splenocytes by centrifugation onto a Ficoll-Isopaque gradient (Amersham Pharmacia Biotech) and then electroporated (960 microfarads, 270 V) with 45 g of expression or control plasmids in 0.4 ml of serum-free RPMI using a Gene Pulser II System (Bio-Rad). Following 20-h recovery in complete medium, D10 cells were seeded on plates coated with 1 g/ml anti-CD3 antibody (clone 145-2C11; BD PharMingen, San Diego, CA) and then incubated for 16 h prior to RNA extraction.
RNase Protection Analysis-Total RNA was extracted from D10 cells using RNAzol B (Tel-Test, Inc., Friendswood, TX). Specific antisense DNA templates for murine IL-4 and for the internal control glyceral-dehyde-3-phosphate dehydrogenase (GAPDH) were purchased (BD PharMingen). These were used to synthesize the ␣-[ 32 P]UTP (3,000 Ci/mmol, 10 mCi/ml; Amersham Pharmacia Biotech)-labeled probes in the presence of a GACU pool using a T7 RNA polymerase (Promega). Riboprobes (500,000 cpm) were hybridized with 5-15 g of target RNA at 45°C, followed by digestion with RNase A (Promega) for 30 min at 30°C as described (25). The samples were loaded on an acrylamide-urea sequencing gel next to labeled DNA molecular weight markers and next to the labeled probes and run at 50 watts in 45 mM Tris, pH 8.2, 45 mM boric acid, and 1 mM EDTA. The gel was adsorbed to filter paper, dried under vacuum, and exposed to film (X-Omat TM AR; Eastman Kodak Co.) at Ϫ70°C with an intensifying screen. Autoradiographs were quantified by computerized densitometry using a Kodak Digital Science Electrophoresis Documentation and Analysis System. Results are shown as the ratio of IL-4/GAPDH densitometric units.
Copper-Phenanthroline Footprinting-Footprinting experiments were performed as described (27). To identify discrete nucleotide contacts for rCP2 within the CP2-responsive IL-4 promoter region, a 195-146 oligonucleotide was 5Ј-end-labeled on the coding or noncoding strand and then incubated (30 min, 25°C) with 100 ng of rCP2 in 25 l of EMSA buffer (described above). Samples were electrophoresed on a 4% native polyacrylamide gel, which was then immersed into a solution containing 10 mM Tris, pH 8.0, 0.2 mM 1,10-phenanthroline, and 0.045 mM CuSO 4 (Sigma). The chemical nuclease reaction was started by the addition of mercaptopropionic acid (Sigma) to 0.05%, allowed to proceed for 12 min at 25°C, and then quenched in 2 mM 2,9-dimethyl-1,10phenanthroline (Sigma). Free DNA and DNA-protein complexes, visualized by autoradiography, were eluted from the gel matrix (18 h at 37°C) in 0.5 M ammonium acetate, pH 7.5, 1 mM EDTA, and 0.1% sodium dodecyl sulfate. Equivalent amounts of DNA from each sample and of a Maxam-Gilbert GϩA ladder of the same probe (27) were resolved by electrophoresis on an 8% acrylamide, 7 M urea gel.

Differential Effect of CP2 Overexpression on the IL-4 and IL-2
Promoters-Several potential CP2 elements are located within the proximal 300 bp of the human IL-4 promoter. We analyzed the effect of transient CP2 overexpression on CAT gene expression driven by an IL-4 promoter region included between bp Ϫ311 and ϩ55 in the human T cell line Jurkat. In agreement with previous studies (24,28,29), we found that these cells are in a preactivated state with respect to IL-4 expression, characterized by constitutively elevated IL-4 mRNA accumulation (data not shown) and promoter activity (Fig. 1). Constitutive IL-4 promoter activity was enhanced up to 5-fold in Jurkat cells transiently transfected with a CP2 expression plasmid (Fig. 1). CP2 overexpression resulted in less noticeable increase of IL-4 promoter activity in cells stimulated with PMA (10 ng/ml) and the Ca 2ϩ ionophore A23187 (1 M), presumably due to the contribution of endogenous CP2 to IL-4 activation in these cells (see below). In striking contrast, PMA-and A23187-induced IL-2 promoter activity was reduced by up to 80% in CP2overexpressing cells (Fig. 1).
Identification of a CPRE in the IL-4 Promoter-To map the IL-4 promoter element(s) necessary for transactivation by CP2, we generated a panel of deletional mutants for use in reporter studies ( Fig. 2A) (21). Transiently overexpressed CP2 was more effective on promoter constructs truncated at bp Ϫ265 or Ϫ225 than on IL-4.311 or IL-4.741 (Fig. 2B). However, removal of the region spanning bp Ϫ225 to Ϫ176 decreased transcriptional activation in CP2-overexpressing cells by almost 4-fold, while constructs truncated at bp Ϫ145 through Ϫ65 were unresponsive to the factor (Fig. 2B). This indicated that a CPRE is located between bp Ϫ175 and Ϫ146, although additional nucleotide contacts upstream of bp Ϫ175 might be required for full promoter inducibility by CP2.
The IL-4 promoter region included between bp Ϫ225 and Ϫ146, shown in Fig. 2C, harbors a number of elements contributing to a varying degree to IL-4 transcriptional activation (21,30,31). To verify whether CP2 can directly bind to elements located within this IL-4 promoter region, we generated three oligonucleotides spanning partially overlapping sequences for use in EMSA (Fig. 2C). rCP2, used at a concentration (100 ng/15 l) sufficient for saturation of an ␣-globin consensus oligonucleotide (Fig. 2D, lane 1), did not bind to oligonucleotides 175-146 (lane 2) and 225-176 (lane 4). Since nucleotide contacts in both subregions might be required for CP2 binding, we also used as a probe an oligonucleotide (195-146) including the proximal 20 bp of the 225-176 segment in addition to 175-146 (Fig. 2C). This resulted in consistent formation of a CP2 complex of apparently lower affinity than that formed on the ␣-globin probe (Fig. 2D, lane 3). A complex of faster mobility also formed on the 195-146 oligonucleotide, which was accounted for by binding of monomeric CP2, as assessed by coincubation with CP2-specific antibodies (data not shown).
To define the nucleotide contacts necessary for CP2 binding to this region of the IL-4 promoter, we analyzed the pattern of protection from chemical cleavage of a 195-146 oligonucleotide following EMSA with rCP2. In experiments using copper-phenanthroline as the cleaving agent (27), rCP2 (100 ng) protected a sequence extending from bp Ϫ177 to Ϫ158 (Fig. 2E). A CPRE ( Ϫ174 CTGATTTCACAGG Ϫ162 ) is recognizable within this sequence. Its homology to CP2 elements identified within other cellular and viral promoters is shown in Fig. 2F. The IL-4 CPRE displays a conserved ( Ϫ165 CAGG Ϫ162 ) and an imperfect CNRG motif ( Ϫ174 CTGA Ϫ171 ) separated by a 5-bp linker.
Binding of Endogenous CP2 to the IL-4 CPRE-A constitutive CP2-immunoreactive complex was formed in EMSA using a consensus ␣-globin CP2 oligonucleotide and nuclear extracts from Jurkat cells (Fig. 3, A and B). CP2 nuclear expression and/or DNA binding activity in these cells was not affected by stimulation with PMA and A23187 (Fig. 3A, lanes 2 and 3). Four complexes were detectable in EMSA with Jurkat nuclear extracts and a similar length oligonucleotide probe centered on the IL-4 promoter CPRE (lanes 4 -6). Differently from the ␣-globin complex, Ca 2ϩ -mediated stimulation did affect formation of the slower mobility complexes I and II, with consistently increased formation of complex II and diminished complex I formation (lane 5). Both complexes appeared to contain immunoreactive CP2, as indicated by the neutralizing effect of rabbit anti-CP2 antibodies (Fig. 3B, lane 6), while neither complex reacted to CBF-A-, NFAT-1-, or STAT6-specific antibodies (lanes 7 and 8 and data not shown). Complexes III and IV apparently consisted of unrelated constitutive nuclear protein(s) of unclear sequence specificity.
To elucidate the contribution of endogenously expressed CP2 to constitutive and inducible activity of the IL-4 promoter, we analyzed CAT expression in Jurkat cells transiently transfected with plasmids carrying point mutations within the IL-4 CPRE that would selectively affect the formation of the CP2containing complexes. Disruption of the distal CNRG motif ( Ϫ174 ATTATTTCACAGG Ϫ162 ) was sufficient to markedly decrease CP2 binding in EMSA using Jurkat nuclear extracts (Fig. 3C) or rCP2 (not shown). This was accompanied by an almost 50% decrease in constitutive and induced CAT expression in Jurkat cells transiently transfected with an IL-4.225 plasmid carrying the same mutation within the CPRE (Fig. 3D).

IL-4 Transcriptional Repression in T Cells Expressing a CP2 Dominant
Negative-To further assess the contribution of endogenous CP2 to IL-4 promoter activity, we generated, for use in transient transfection experiments, a plasmid encoding a CP2 polypeptide lacking the Elf-1 homology region of the DNAbinding domain (13) and referred to as ⌬Elf-1 mutant. In a previous study, this mutant did not bind to DNA and specifically inhibited binding of full-length CP2 in a dominant negative fashion (22). Its overexpression inhibited by Ͼ75% constitutive IL-4 promoter activity in Jurkat cells (Fig. 4A). A similar  (30); CCAAT, binding elements for CBF (31). Octamer-associated protein (OAP) and c-Maf response element (MARE) bind activation protein-1 family members (i.e. JunB) and the related proto-oncoprotein c-Maf, respectively (5,6). The negative regulatory elements (NRE-I and -II) bind as yet unidentified transcriptional repressor(s) (29). Interferon stimulation-response element (ISRE) is a negative regulatory interferon-responsive factor-2-binding element (31). Box II, A/T-rich region contributing to repression by high mobility group protein I(Y) (36). B, Jurkat cells were transiently transfected with 1 g each of the indicated human IL-4 promoter constructs and 2 g of a pRc/CMV-CP2 expression vector or its corresponding noncoding control (dashed line). Shown is mean Ϯ S.E. -fold induction of control CAT expression in three independent experiments. C, sequence of the region required for IL-4 promoter activation by overexpressed CP2. The relevant elements and their cognate transcription factors are indicated. Shown is, among other elements also indicated in a, the recently identified P2 negative regulatory element (NRE), binding the putative transcriptional repressor Rep-1 (21). Also shown is a schematic representation of three oligonucleotides (225-176, 195-146, and 175-146), spanning promoter subregions included between the indicated bp, that were used as probes in EMSA using rCP2. HMG I(Y), high mobility group protein I(Y); IRF-2, interferon-responsive factor-2; ISRE, interferon stimulation-response element. D, shown for comparison (lane 1) is the binding of rCP2 (100 ng) to an oligonucleotide probe including the proximal ␣-globin CP2 consensus site. The formation of a complex with CP2 homodimers is indicated. E, copper-phenanthroline footprinting of a 195-146 oligonucleotide, degree of IL-4 promoter inhibition was seen in cells stimulated with A23187, while IL-2 promoter activity was fundamentally unaffected (data not shown).
Up-regulation of IL-4 promoter activity in CP2-overexpressing Jurkat cells was often paralleled by markedly increased IL-4 secretion, suggesting direct activation of the endogenous IL-4 gene (data not shown). However, due to low level expression of IL-4 and the poor transfection efficiency in these cells, the effect of overexpressed ⌬Elf-1 CP2 was far less apparent (not shown). To elucidate the role of CP2 in transcriptional activation of the endogenous IL-4 gene, we therefore assessed the effect of overexpressed CP2 polypeptides on IL-4 transcript accumulation in the murine Th2 cell line D10. As shown in Fig.  4, B and C, D10 cells transfected to higher than 50% efficiency with a full-length CP2-encoding plasmid expressed, upon CD3 ligation, markedly higher levels of IL-4 mRNA than cells transfected with the corresponding noncoding plasmid. In contrast, and in clear agreement with our findings shown in Fig. 4A, overexpression of the dominant negative ⌬Elf-1 mutant consistently interfered with CD3-dependent activation of the IL-4 gene in D10 cells.

DISCUSSION
In this study, we provide the first evidence of the involvement of CP2, the prototypical member of a novel family of transcription factors related to the Drosophila developmental protein Elf-1, in the regulation of cytokine genes expressed in T cells. CP2 has been involved in gene regulation in a variety of cell lineages, where it can act as both a transcriptional activator and a repressor (14,22,32,33). We confirm here the dual nature of this factor, as we found that CP2 can repress IL-2 promoter activity, while its expression and function are required for IL-4 promoter-driven transcription in the human T cell line Jurkat (Figs. 1 and 4A) and for activation of the 5Ј-end-labeled on the coding strand. Following EMSA with rCP2, free (F) and bound DNA (CP2) were cleaved within the gel matrix and then eluted and electrophoresed on an 8% sequencing gel. The alignment with a Maxam-Gilbert ladder (GϩA) of the same probe is shown. Positions relative to the IL-4 transcription initiation site are indicated to the left. A hypersensitive site at bp Ϫ151 and a footprint extending from bp Ϫ177 to Ϫ158 are indicated to the right as an asterisk and a dashed line, respectively. F, the CP2-protected sequence is aligned with a series of CPREs identified in several cellular and viral promoters. The CNRG repeats (see "Discussion") are indicated in boldface letters. Shown are the distal and proximal CPRE from the mouse ␣-globin promoter, the high affinity sites from the rat ␥-fibrinogen promoter and the human immunodeficiency virus-1 long-terminal repeat, and a low affinity site from the major histocompatibility complex class II Ea proximal promoter. Also shown is an element within the mouse steroid 16␣-hydroxylase gene (Cyp 2d-9) that interacts with the CP2-related protein LBP-1a (41).

FIG. 3. Binding of endogenous CP2
to the IL-4 CPRE and its contribution to promoter activity. A, EMSA using nuclear proteins (5 g/lane) extracted from Jurkat cells treated as indicated and oligonucleotide probes spanning the proximal ␣-globin (lanes 1-3) or the IL-4 CPRE (lanes 4 -6). The ␣-globin CP2 complex is indicated to the left. endogenous IL-4 gene in the established murine Th2 line D10 (Fig. 4, B and C). These findings point to CP2 as a specific activator of the IL-4 gene in T cells.
We show that a CPRE is located between bp Ϫ177 and Ϫ158 of the human IL-4 promoter, a region previously shown to be required for maximal promoter activation in Th cells (30,34). The IL-4 CPRE ( Ϫ174 CTGATTTCACAGG Ϫ162 ) only diverges by 1 bp from the reported CP2 consensus (Fig. 2F), which was defined by assessing the effect of in-frame clustered mutations within high affinity CP2 sites, such as the proximal ␣-globin site or the ␥-fibrinogen site, on rCP2 binding in EMSA (14). While the nucleotide composition of the linker sequence was found to only marginally affect CP2 binding, spacing of the two CNRG motifs was critical for the stability of the CP2-DNA complex (14). Thus, CNRG repeats separated by 5 bp, such as the ␣-globin distal CPRE (Fig. 2F), can bind CP2 with about 4-fold lower affinity than elements containing a linker of 6 bp (14). Consistent with this view, the IL-4 CPRE, featuring a 5-bp linker, bound rCP2 with apparently lower affinity than the proximal ␣-globin site (Fig. 2D).
CP2 represents a novel family of homo-oligomeric transcription factors binding direct DNA repeats (13,22). However, heteromeric interaction of CP2 with Elf-1-related and nonrelated factors, depending on the promoter context, has also been described (13,35). In a previous study, two classes of CPRE have been defined (35). Type I CPRE, including the high affinity ␣-globin and ␥-fibrinogen sites, preferentially bind CP2 as a homodimer or homotetramer (20), while type II CPRE, including sites within the ␥and ⑀-globin promoters, bind CP2 as an obligate heterodimer (35). Our findings show that the IL-4 CPRE behaves as a type II site. In fact, while rCP2 homodimers bind to this site with relatively low affinity, native CP2 accounts for the formation of two complexes having noticeably slower mobility than the complex formed on an ␣-globin type I site (Fig. 3, A and B) or in EMSA using rCP2 (Fig. 2D). This suggests that additional factor(s) might contribute to CP2 interaction with an otherwise low affinity site. Consistent with this view, Ca 2ϩ -mediated stimulation of Jurkat cells, while not affecting CP2 binding to the ␣-globin site, led to increased formation of a major IL-4 complex, with decreased formation of an additional, slower mobility complex (Fig. 3A). This suggests the involvement of a Ca 2ϩ -regulated, as yet unknown heteromeric co-factor(s) in CP2 interaction with the IL-4 promoter.
The IL-4 CPRE is surrounded by binding elements for the factors NFAT, activation protein-1, CBF, interferon response factor-2, and high mobility group protein I(Y) (30,31,36) (Fig.  2C). A CCAAT box, binding the constitutive factor CBF, lies immediately upstream of its 5Ј-end (31). Mutation of a 5-bp cluster spanning bp Ϫ178 to Ϫ174, thereby disrupting this CCAAT box and the distal CNRG motif of the CPRE, resulted in 40% reduction of overall IL-4 promoter activity in Jurkat cells (31). We show (Fig. 3D) that the isolated disruption of the IL-4 CPRE is sufficient to produce a similar or greater decrease of promoter activity (ϳ50%) in the same cellular host. Although CP2 and CBF, also known as CP1 or nuclear factor-Y, also bind to adjacent elements in the ␣-globin promoter (37), the two factors form mutually exclusive complexes on oligonucleotides including these elements (22). The 3Ј-half of the CPRE partially overlaps the NFAT-binding P2 sequence and a consensus STAT6 element (38). However, CP2 does not apparently interact with NFAT or STAT6 in EMSA using oligonucleotides including both the CPRE and the P2 sequence (data not shown).
Our finding of reduced transcriptional activity of constructs carrying a mutated CPRE demonstrated the critical involvement of endogenous CP2 in IL-4 promoter activity in Jurkat cells. This observation was confirmed using cells transiently expressing a CP2 mutant having disrupted DNA binding but intact dimerization domains. The DNA-binding domain of CP2 includes a region, featuring the highest Elf-1 homology, that has been reported to be essential for interaction of both CP2 and Elf-1 with their cognate sites (13,14). This region, encoded by exon 6 of the CP2 gene and spanning amino acids 189 -239, is missing in several non-DNA-binding CP2 variants, such as FIG. 4. Dominant negative effect of the CP2 ⌬Elf-1 mutant. A, full-length CP2 (CP2) or the ⌬Elf-1 CP2 variant (see "Results") was cloned into pRc/CMV and transiently transfected in Jurkat cells together with an IL-4.311 construct. Mean Ϯ S.E. -fold induction of CAT expression relative to noncoding controls (dashed horizontal line) in three independent experiments is shown. B, a full-length (CP2) or ⌬Elf-1 expression plasmid or the corresponding noncoding vector (Control) was transfected (45 g each) in D10 cells as indicated. IL-4 and GAPDH transcripts were detected by RNase protection of total RNA extracted from these cells 16 h after stimulation with 1 g/ml platebound anti-CD3 antibodies (see "Experimental Procedures"). Shown is a typical experiment out of three performed. C, the relative amounts of IL-4 and GAPDH RNA in these experiments were measured by densitometry (see "Experimental Procedures"). The ratios of IL-4 to GAPDH RNA densitometric units were calculated for each sample and expressed as -fold induction over samples transfected with control vector (dashed horizontal line). Bars represent the mean Ϯ S.E. of three independent experiments. LBP-1d, generated in several cell lines by alternative splicing of CP2 transcripts (11)(12)(13). Some studies have documented the ability of these polypeptides to function as CP2 antagonists, specifically inhibiting CP2-DNA interactions in vitro (13,22). However, the expression in vivo of sufficiently high levels of these mutants to act as dominant negatives has not been demonstrated. Here we show that overexpression of the ⌬Elf-1 mutant markedly represses IL-4 promoter activity in Jurkat cells as well as IL-4 gene expression in activated D10 cells (Fig.  4), stressing the role of CP2 in the activation of the IL-4 gene in T cells. Use of this dominant negative in vivo should provide a valuable tool to understand the role of CP2 in differential cytokine gene expression in Th cell clones.
CP2 is constitutively expressed in the nuclei of Jurkat (Fig.  3A) and D10 cells (not shown), and stimulation did not affect its expression and/or binding to an ␣-globin consensus in EMSA. Our data suggest that inducible co-factor(s) stabilize CP2 binding to the IL-4 CPRE and contribute to IL-4 activation in stimulated cells. Jurkat cells are in a preactivated state with respect to IL-4 transcription and the nuclear expression of NFAT or other IL-4 promoter-binding factors (29,39). Interestingly, both IL-4 expression and CP2 binding activity are induced by mitogenic stimulation of circulating naive Th cells in a cell cycle-dependent fashion (8,10). Consistent with these findings, histone deacetylase inhibitors, such as butyrate, can supplant the cell cycle requirements for both IL-4 transcription and CP2 activation (8,40). In conclusion, our findings support the hypothesis that CP2 expression and DNA-binding activity in T cells are correlated with IL-4 gene expression and point to CP2 as an important participant in the mechanisms regulating Th cell differentiation in the periphery.