Yin-Yang 1 activates interleukin-4 gene expression in T cells.

Interleukin-4 (IL-4) is a multifunctional cytokine that plays an important role in immune and inflammatory responses. Expression of the IL-4 gene is tightly controlled at the level of gene transcription by both positive and negative regulatory elements in the IL-4 promoter. Several constitutive nuclear factors have been identified that can interact with IL-4 promoter elements in DNA binding assays. Here we report that the zinc-finger protein YY-1 (Yin-Yang 1) can bind to multiple elements within the human IL-4 promoter. Cotransfection of Jurkat T cells with different IL-4 promoter/reporter constructs together with expression vectors encoding antisense, wild-type, or zinc finger-deleted mutant YY-1 suggested that YY-1 enhanced IL-4 promoter activity in a DNA-binding domain-dependent manner. Site-directed mutagenesis revealed that a proximal YY-1-binding site, termed Y0 ((-59)TCATTTT(-53)), was essential for YY-1-driven IL-4 promoter activity. In addition, cotransfected YY-1 enhanced both IL-4 promoter activity and endogenous IL-4 gene expression in nontransformed peripheral blood T cells. Thus, YY-1 positively regulates IL-4 gene expression in lymphocytes.

Interleukin-4 (IL-4), 1 a pleiotropic cytokine produced by activated T cells and basophils, plays a critical role in cellular and humoral immune responses (1). Dysregulated expression of IL-4 has been linked with autoimmune and allergic diseases (2,3). In T cells, IL-4 gene expression is regulated at the transcriptional level by both ubiquitous and cell type-restricted factors (4 -11). These factors interact with a proximal promoter region composed of multiple regulatory elements and can both positively and negatively influence transcriptional activation (see Fig. 1). Other regions have been identified outside of the proximal IL-4 promoter that can regulate IL-4 gene expression, including the IL-4/IL-13 intergenic region (12), the IL-4 second intron (13), and downstream of the IL-4 gene (14). Some of these elements appear to coordinately regulate IL-4 gene expression at the chromatin level (15).
Major insights into the regulation of IL-4 gene expression came from studies using transgenic or knockout approaches to investigate the molecular basis of Th2 differentiation in mice (reviewed in Ref. 16). Using these approaches, several transcription factors have been identified that are critical for this process, including Stat6 (17), NFATc (18), c-Maf (19), GATA-3 (20), and JunB (21). The mechanisms by which these factors influence T cell differentiation is under active investigation. To date, direct binding to and/or activation of the IL-4 promoter has been demonstrated for NFATc (22), c-Maf (19), and JunB (21), but not for Stat6 (23) or GATA-3 (24). A two-step model has recently been proposed to explain IL-4 gene expression in Th2 cells (25). Based on the appearance of DNase-hypersensitive sites in the IL-4 gene locus (26,27), this is thought to involve an initial chromatin remodeling step followed by cytokine gene transcription in response to T cell receptor-activated transcription factors. Although chromatin remodeling is likely an important regulatory step during Th2 differentiation, the proximal IL-4 promoter confers a high degree of tissue specificity when linked to reporter genes in transgenic mice (15,28,29). Thus, an analysis of the factors that regulate the activity of the proximal IL-4 promoter will enhance our understanding of IL-4 gene expression.
We recently performed a detailed deletional analysis of the human proximal IL-4 promoter and discovered novel binding sites for several transcription factors, including NFAT (nuclear factor of activated T cells) (9), CP-2 (30), and an uncharacterized repressor factor (10). This analysis also uncovered binding sites for several other constitutive nuclear factors, in keeping with previous studies on the IL-4 promoter in other systems (4,5,8,31). Unlike the inducible IL-4 promoter-binding factors that have been intensively studied, only a few of these constitutive factors have been well characterized. Here we report that YY-1 (Yin-Yang 1), a constitutive nuclear member of the GLI-Krü ppel family of zinc-finger transcription factors, can interact with four binding sites in the human proximal IL-4 promoter. We use site-directed mutagenesis together with cotransfection assays to define the role of YY-1 in regulating the transcriptional activation of the IL-4 promoter.

EXPERIMENTAL PROCEDURES
Plasmid Construction-A luciferase-based human IL-4 promoter construct containing 270 bp upstream from the transcription start site and ending at ϩ65 (termed 270luc) was synthesized as described (9). A construct containing 235 bp upstream from the transcription start site was amplified by PCR from 270luc and ligated into the pCR-2.1-TOPO vector (Invitrogen), followed by ligation into the XhoI site of pGL3-Basic (Promega). A construct containing 145 bp upstream from the transcription start site was amplified by PCR from genomic DNA and ligated into the KpnI and SacI sites of pGL3. Mutations were introduced into the Y0 YY-1-binding site within 270luc using site-directed mutagenesis with * This work was supported by NHLBI Grant HLAI61875 from the National Institutes of Health and by research grants from the American Lung Association, the American Heart Association, and the Johns Hopkins University School of Medicine Clinician Scientist Program. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. the QuikChange kit (Stratagene) to generate construct 270lucY0mut. The following mutations specifically disrupted the Y0 site (see Fig. 4): wild-type 270luc, Ϫ59 TCATTTT Ϫ53 ; to 270lucY0mut, Ϫ59 TgtaTTT Ϫ53 (mutations in lowercase). All products were sequenced to confirm accurate replication. An SV40 promoter-driven YY-1 expression vector was synthesized by ligating the EcoRI insert from pcDNA1-YY-1 (described in Refs. 34 and 35) into the EcoRI sites of pSG5 (Stratagene). An expression vector encoding a YY-1 mutant with a deletion of amino acids 333-408 within the zinc-finger domain (ZFDmut) was synthesized by restricting pSG5-YY-1 with HindIII and BamHI and religation using the following linkers: HindIII, 5Ј-AGCTTCCAACAACCAGTGA-G-3Ј; and BamHI, 5Ј-GATCCTCACTGGTTGTTGGA-3Ј. The antisense YY-1 vector was a kind gift of Dr. Michael Atchison (University of Pennsylvania) and has been described (32). Expression vectors encoding full-length NFATp (pREP4-NFATp) and NFATc (pSH107-NFATc) were gifts of Dr. Timothy Hoey (Tularik Inc.) and Dr. Gerald Crabtree (Stanford University), respectively.
Cell Lines and Transfections-Jurkat T cells (courtesy of Dr. Jack Strohminger, Harvard University) were maintained and transfected using 2.5 ϫ 10 6 cells, 1 g of reporter, and 3 l of Superfect ® (QIAGEN Inc.) per g of plasmid DNA as previously reported (22). COS-7 cells were obtained from American Type Culture Collection (Manassas, VA). Jurkat cells and a Jurkat subline stably transfected with the SV40 large T antigen (JTAg cells, kindly provided by Dr. Ron Wange, NIA, National Institutes of Health) were transfected using electroporation as follows. Cells were resuspended at a concentration of 8 ϫ 10 6 cells/300 l of Opti-MEM (Invitrogen) together with the indicated amounts of reporter plasmid and then electroporated at 300 V and 960 microfarads in cuvettes with a 0.4-cm gap width (Bio-Rad Gene-Pulser II). Transfection efficiencies of standard Jurkat and JTAg cells using electroporation averaged 40 and 85%, respectively, as determined by analyzing green fluorescence of cells cotransfected with a GFP expression vector (data not shown). Eighteen hours after transfection, cells were lysed and analyzed by luminometry using a Monolight 3010C luminometer and luciferase assay kit (Analytical Bioluminescence, Gaithersburg, MD). In some experiments, cells were incubated with the calcium ionophore A23187 (Calbiochem) for 18 h prior to analysis of reporter gene expression as indicated. Two strategies were taken to control for transfection efficiency. First, at least three different reporter plasmid and expression vector preparations were used, and cells of similar passage number were transfected under identical reaction conditions. In cotransfection experiments, cells were also transfected with empty vector to keep the total amount of DNA constant in a given condition. Second, the efficiency of transfection was monitored in some experiments by cotransfecting a second internal control plasmid. In these experiments, cells were cotransfected with pSEAP2-Control (1 g; CLONTECH), and firefly luciferase activity was normalized to secreted alkaline phosphatase measured by chemiluminescence (Great EscAPe SEAP kit, CLONTECH). pSEAP2 was chosen because, compared with other standard internal control vectors, its expression was affected the least by cell stimulation, and it did not interfere with YY-1 overexpression in pilot cotransfection experiments (data not shown). In experiments with the antisense YY-1 vector, cell growth was analyzed by (i) counting viable cells using trypan blue exclusion and (ii) comparing total protein content in lysates of control and antisense-transfected cells.
Peripheral Blood Cells and Transfection-Highly enriched peripheral blood T cells were obtained by countercurrent elutriation of leukapheresis packs (Johns Hopkins Oncology Center). This procedure yielded ϳ70% pure CD3 ϩ cells as assessed by flow cytometry (data not shown). Cells were transfected by electroporation according to the method of Cron et al. (33) as follows. Cells were first incubated for 18 h in the presence of 1 g/ml phytohemagglutinin (Calbiochem). Subsequently, 10 ϫ 10 6 cells were resuspended in a volume of 300 l Opti-MEM with 10 g of reporter together with the indicated amounts of expression vector or empty vector to keep total DNA constant. Cells were electroporated using a 0.4-cm gap width cuvette and a 1-s duration pulse at 300 V and 960 microfarads (Bio-Rad Gene-Pulser II). These settings were determined in pilot experiments to yield maximal luciferase expression (data not shown). Omission of phytohemagglutinin from the overnight culture significantly reduced transfection efficiency (data not shown). Cells were stimulated with calcium ionophore (1 M A23187) plus phorbol 12-myristate 13-acetate (20 ng/ml) or Me 2 SO control for 18 h prior to analysis of reporter gene activity by luminometry as described above. IL-4 protein secretion was measured in supernatants using a sensitive enzyme-linked immunosorbent assay (Ultrasensitive IL-4 kit, detection limit of 0.27 pg/ml, BIOSOURCE) according to the manufacturer's instructions. All measured values fell within the standard curve.
Western Blot Analysis-Whole cell, cytoplasmic, and nuclear lysates were obtained from Jurkat cells. Nuclear extracts from mouse Th1 (AE7) and Th2 (D10) clones were kindly provided by Dr. Anuradha Ray (Yale University). Cell extracts were separated by 6% SDS-PAGE and then transferred to Trans-Blot transfer medium polyvinylidene difluoride membrane (Bio-Rad). After blocking in phosphate-buffered saline containing 5% bovine serum albumin and 0.1% Tween 20 for 1 h, membranes were probed with mouse anti-YY-1 monoclonal antibody H-10 (1:500 dilution; Santa Cruz Biotechnology) for 10 h at 4°C, followed by incubation for 1 h with horseradish peroxidase-conjugated goat anti-mouse secondary antibody (1:3000 dilution; Amersham Biosciences, Inc.). After a final washing step, immunoreactive bands were visualized by enhanced chemiluminescence and autoradiography using the ECL Western blotting detection kit (Amersham Biosciences, Inc.) according to the manufacturer's directions.
Electrophoretic Mobility Shift Assays (EMSAs)-The following 30-bp oligonucleotides and their complements were synthesized: 5Ј-ATTGC- Mutations were introduced into the P0 and P2 consensus YY-1-binding sites (underlined) as indicated in Figs. 2 and 4. A recombinant fragment of murine NFATp (including 298 amino acids of the DNA-binding domain, highly conserved among different NFAT family members) was expressed as a hexahistidine-tagged protein and extracted as described (34). Recombinant YY-1 was extracted from bacteria transformed with a histidine-tagged YY-1 expression vector. Nuclear extracts were obtained from 5 ϫ 10 6 Jurkat cells using the method of Schreiber et al. (35). EMSAs were performed using 5 g of nuclear protein and ␥-32 P-end-labeled probe with 0.8 g of poly(dG⅐dC) (Amersham Biosciences, Inc.) in a final volume of 10 l. Experiments with recombinant proteins were performed with 1-l aliquots of the recombinant NFATp DNA-binding domain with or without 1 l of YY-1, and the final buffer composition in all samples was adjusted to contain 84 mM KCl, 34 mM NaCl, 7% glycerol, 20 mM HEPES (pH 7.5), 1 mM dithiothreitol, and 0.1% Nonidet P-40. Free probes and protein-DNA complexes were resolved by 5% PAGE with 0.5ϫ Tris borate/EDTA. In some experiments, extracts were incubated with 1 l of the following antisera for 30 min at 4°C prior to addition of probe: rabbit anti-NFATp polyclonal antibody (Upstate Biotechnology, Inc.), rabbit anti-YY-1 polyclonal antibody C-20 (Santa Cruz Biotechnology), rabbit anti-Stat6 polyclonal antibody S-20 (Santa Cruz Biotechnology), and isotype-and species-matched control IgG (Sigma).

Multiple YY-1-binding Sites in the Human IL-4 Promoter-
The IL-4 promoter contains multiple binding sites for members of the NFAT family of transcription factors, termed the P elements P0 -P5 (4, 5, 9) (Fig. 1). During a detailed deletional analysis of the human IL-4 promoter, we identified a factor (termed complex IV) that bound downstream of NFAT to the P2 element (10). As shown in Fig. 2 (A-C), multiple nuclear factors from Jurkat cells recognized this element in EMSAs. These included a slowly migrating factor with apparent repressor properties (termed Rep-1 (10)), a Ca 2ϩ -induced factor containing NFATp, and the constitutive complexes III-V. We previously reported that complex IV binds in a sequence-specific manner and that its formation is not competed for by a panel of consensus oligonucleotides, including those containing high affinity NFAT and AP-1 sites (10). Mutational analysis revealed that the formation of complex IV required sequences located just downstream of the NFAT consensus sequence Ϫ163 GGAACA Ϫ158 (Fig. 2, A and B). Inspection of this region revealed a sequence ( Ϫ160 ACATTTT Ϫ154 ) highly homologous to a consensus YY-1-binding site, which we termed Y2 (Fig. 2B). The most conserved nucleotide in the YY-1 consensus sequence is the second cytosine (5Ј-CCATNTT-3Ј), which appears to be required for high affinity DNA binding (36). When we mutated this nucleotide within the P2 element, we did not detect complex IV binding in EMSA using Jurkat nuclear extracts (data YY-1 Enhances IL-4 Gene Expression not shown). Using anti-transcription factor antibodies and nuclear extracts from Jurkat T cells in EMSA, we found that complex IV indeed contained immunoreactive YY-1 (Fig. 2C,  lanes 3 and 6). Fig. 2D also shows that factors previously shown to bind to the P2 region, such as NFATp and Stat6 (23), did not affect complex IV. YY-1 appeared to bind in a competitive manner with complex III, which bound just upstream of the NFAT site ( Fig. 2A, compare lanes 1 and 7). The identities of the constitutive complexes III and V are currently unknown.
Because the IL-4 promoter P elements are homologous even outside of the NFAT core consensus sequence, we speculated that YY-1 would interact with additional sequences within the IL-4 promoter. Fig. 3 shows that a YY-1-immunoreactive complex formed on oligonucleotide probes encompassing the P3 and P4 (but not P1 and P5) elements using Jurkat nuclear extracts in EMSA. Importantly, sequence inspection revealed five of six base pair matches for the YY-1 consensus sequence (including the critical second cytosine) within the P3 and P4 (but not P1 and P5) oligonucleotides (Fig. 3).
The most proximal IL-4 NFAT site (termed P0) has been shown to bind both constitutive and inducible nuclear factors and to contribute to promoter activity in T cell lines and Th2 cells (29,37). Sequence inspection revealed a potential YY-1binding site located immediately adjacent to the P0 NFAT consensus sequence (Fig. 4). Fig. 4A shows that a YY-1-immunoreactive complex formed on the P0 element using Jurkat nuclear extracts in EMSA. To precisely map the YY-1-binding site in this region, we introduced mutations into the YY-1 consensus sequence and studied the ability of recombinant YY-1 to bind to wild-type and mutant oligonucleotides. As shown in Fig. 4B, mutation of two nucleotides within the YY-1 consensus sequence (including the second cytosine, C Ϫ59 ) drastically impaired the ability of YY-1 to bind to this region (compare lanes 4 and 6). As a control, we also found that recombinant YY-1 readily bound to the P2 oligonucleotide. These experiments mapped the proximal IL-4 promoter YY-1-binding site to sequence Ϫ59 TCATTTT Ϫ53 , which we termed Y0. The introduced mutations also inhibited the binding of native YY-1 to the Y0 element (Fig. 4C). Note that a constitutive complex that was partially inhibited by the anti-YY-1 antibody (Fig. 4A, asterisk) bound equally well to the mutant probe (Fig. 4C). The identity of this complex, which does not represent sequencespecific binding by YY-1, is currently unknown. We did not detect NFATp binding to the P0 element in these experiments (data not shown), in keeping with prior studies showing that the P0 element binds NFAT with lower affinity than other P elements (37).
An Antisense YY-1 Vector Inhibits Basal IL-4 Promoter Activity in Jurkat T Cells-Our biochemical data showed that YY-1 can interact with four binding sites in the proximal IL-4 promoter. Sequence inspection and computer analysis did not detect additional YY-1-binding sites within ϳ1000 bp surrounding the IL-4 transcription start site. To test the functional significance of YY-1 in regulating IL-4 promoter activity, we next studied the effect of cotransfecting an antisense YY-1 expression vector with different promoter constructs into Jurkat T cells. These experiments were prompted by the observations that (i) the IL-4 promoter is constitutively active in transiently transfected Jurkat cells (30) and (ii) YY-1 is a constitutive nuclear protein in these cells (Fig. 2). We used both a full-length promoter construct (270luc) and a minimal construct containing only the Y0 site (145luc; see "Experimental Procedures") together with an antisense YY-1 vector that has been shown to inhibit glucocorticoid receptor expression by ϳ50% in other systems (32). Interestingly, cotransfection of antisense YY-1 inhibited transcription driven by the full-length IL-4 promoter by ϳ80% (Fig. 5). The activity of the minimal construct 145luc was also reduced by antisense YY-1, although this result did not reach statistical significance (p ϭ 0.06). Transfection of antisense YY-1 did not appear to affect cell growth (see "Experimental Procedures"). These results suggest that YY-1 contributes to constitutive IL-4 promoter activity in Jurkat cells.

YY-1 Activates the IL-4 Promoter in Two Different Jurkat
Sublines, but Not in COS-7 Cells-We next overexpressed YY-1 and reasoned that increasing the cellular concentration of this factor would further enhance IL-4 promoter activity. In these experiments, Jurkat cells were cotransfected by electroporation with a series of IL-4 promoter constructs and a wild-type YY-1 expression vector. Cotransfection of YY-1 resulted in a concentration-dependent enhancement of transcription driven by the full-length IL-4 promoter (data not shown) as well as of transcription driven by 145luc (Fig. 6A). Western blot analysis confirmed that total cellular YY-1 content was also increased in a concentration-dependent manner in these cells (Fig. 6A). Overexpressed YY-1 also enhanced IL-4 promoter activity in transiently transfected JTAg cells, a subline of Jurkat cells stably transfected with the SV40 T antigen ( Fig. 6B; see "Experimental Procedures"). In both JTAg and standard Jurkat cells, the expression of YY-1 was confined largely to the cell nucleus, indicating that the subcellular localization of overexpressed YY-1 was faithfully regulated (Fig. 6B and data not shown). The detection of some YY-1 within the cytoplasmic fraction likely reflects ongoing expression of YY-1 from the transfected construct. In parallel experiments, we examined the effects of overexpressing YY-1 on IL-4 promoter activity in non-lymphoid COS-7 cells. IL-4 145luc was constitutively active in these cells, as was reported for the human IL-4 promoter in non-lymphoid HeLa cells (38). Interestingly, cotransfection of YY-1 sufficient to increase total cellular YY-1 expression as determined by Western blotting down-regulated the IL-4 promoter in COS-7 cells (ϳ70% inhibition; n ϭ 4) (data not shown). Thus, IL-4 promoter enhancement by YY-1 appears to be cell type-specific.
The Y0 Element Is Critical for YY-1-dependent IL-4 Promoter Activity in Jurkat T Cells-The observation that 145luc, which contains only the Y0 element, can be transactivated by overexpressed YY-1 suggests that the Y0 element is critical for YY-1-dependent IL-4 promoter activity. To test this hypothesis, we mutated Y0 in the context of the full-length promoter to generate the construct 270lucY0mut and studied promoter activity under a variety of experimental conditions. Strikingly, the constitutive activity of 270lucY0mut was significantly reduced compared with that of the wild-type construct (Fig. 7A). In addition, whereas overexpression of YY-1 transactivated the wild-type IL-4 promoter, the activity of 270lucY0mut was not enhanced under similar conditions (Fig. 7A). To ensure that these observations were not due to subtle differences in transfection efficiency, we normalized the activity of 270lucY0mut and the wild-type reporter to an internal control plasmid cotransfected in parallel. When analyzed in this fashion, the constitutive activity of the Y0 mutant construct was still significantly reduced and was not enhanced by overexpressed  YY-1 (Fig. 7C). This suggests that in the context of the fulllength promoter, the Y0 site is critical for constitutive YY-1dependent promoter activity. To exclude the possibility that disruption of the Y0 site affected the adjacent NFAT-binding site, we used a recombinant fragment of the NFATp DNAbinding domain in EMSA. In these experiments, we found that NFAT bound equally well to the wild-type and Y0 mutant oligonucleotide probes (data not shown). This is supported by the observation that calcium inducibility of 270lucY0mut was not appreciably hindered compared with that of wild-type 270luc (Fig. 7C). In addition, to exclude the possibility that the Ϫ59 TgtaTTT Ϫ53 mutation used in 270lucY0mut generated a new protein-binding site, we used an oligonucleotide encompassing this sequence in additional EMSAs with Jurkat nuclear extracts. In these experiments, only the YY-1-specific band was found to be inhibited, and no new binding activities were observed (data not shown). Therefore, the reduced activity of the Y0 mutant is due to specific inhibition of YY-1 binding to bind its cognate site in this region.
An Intact YY-1 DNA-binding Domain Is Required for IL-4 Promoter Transactivation-To determine whether DNA binding by YY-1 was required for IL-4 promoter activity, we synthesized a mutant YY-1 expression vector disrupting three of the four zinc-finger domains (ZFDmut; see "Experimental Procedures") known to be required for DNA binding. We studied the ability of both the wild-type and ZFDmut constructs to transactivate different IL-4 promoter constructs in transient transfection assays. Unlike wild-type YY-1, which enhanced the activity of each construct examined, cotransfection of ZFDmut did not transactivate any of the reporter constructs (Fig.   7B). Similar results were obtained using transiently transfected JTAg cells (data not shown) and when normalizing data to an internal control reporter plasmid (Fig. 7D). Taken together with the results obtained from the site-directed mutagenesis experiments, these data suggest that YY-1 transactivates the IL-4 promoter by directly binding to the proximal promoter.

YY-1 Enhances IL-4 Promoter Activity and Endogenous IL-4 Gene Expression in Nontransformed
Human T Cells-We next studied the ability of YY-1 to regulate IL-4 promoter activity in nontransformed human peripheral blood T cells. IL-4 promoter activity was undetectable in unstimulated cells, but was inducible following activation of Ca 2ϩ -and protein kinase C-sensitive signaling pathways (Fig. 8). Interestingly, cotransfection of YY-1 resulted in a concentration-dependent, 2-8-fold enhancement of promoter activity (Fig. 8A). Similar results were obtained when normalizing luciferase activity to a cotransfected internal control vector (ϳ5-fold enhancement) (data not shown). We compared the transactivating ability of YY-1 with that of two factors known to enhance IL-4 gene expression in T cells, viz. NFATp and NFATc, by cotransfecting expression vectors encoding each of these factors (see "Experimental Procedures"). YY-1 activated the IL-4 promoter to a similar degree compared with the NFAT proteins. Importantly, the secretion of IL-4 protein from YY-1-cotransfected peripheral blood T cells was significantly increased (Fig. 8B), indicating that overexpressed YY-1 enhances endogenous IL-4 gene expression.

DISCUSSION
Because of its importance in allergic and inflammatory immune reactions, the molecular regulation of IL-4 gene expression has come under intense scrutiny (15,16,39). In activated T cells, IL-4 expression is controlled at the transcriptional level by the coordinated actions of multiple transcription factors interacting with a proximal promoter region (40). Here we report the novel observations that the transcription factor YY-1 bound to multiple elements within the IL-4 promoter and that YY-1 enhanced IL-4 gene expression in T cells. Using sitedirected mutagenesis, we found that a proximal YY-1 element (termed Y0) was critical for both constitutive and YY-1-dependent IL-4 promoter transactivation in Jurkat cells. Support for the notion that DNA binding by YY-1 contributes to IL-4 promoter transactivation is provided by the observation that overexpressed wild-type YY-1 (but not a zinc-finger mutant) enhanced IL-4 promoter activity in different Jurkat sublines. Importantly, cotransfected YY-1 enhanced IL-4 promoter activity and gene expression in nontransformed human T cells.
YY-1 belongs to the GLI-krü ppel family of transcription factors and can initiate, activate, or repress transcription depending on the promoter context (36,(41)(42)(43). YY-1 is a constitutive nuclear phosphoprotein that appears to be expressed in most cell types (44). Targeted deletion of the YY-1 gene in mice results in peri-implantation lethality secondary to neural and other defects (45). Since its original isolation as a repressor of the P5 promoter of adeno-associated virus (46), the number of cellular and viral genes now known to be regulated by YY-1 has grown at a remarkable pace (reviewed in Refs. 42 and 43). The pleiotropic actions of YY-1 are explained in large part by its ability to interact with a diverse array of other sequence-specific factors (47)(48)(49)(50), viral proteins (51), cofactors (52,53), immunophilins (54), and components of the basal transcriptional machinery (41,44,55). DNA bending may influence the ability of YY-1 to regulate gene expression (56). Further levels of complexity are provided by the observations that YY-1 resides within multiple nuclear compartments (57) and that its associations with cofactors can be modulated by acetylation and deacetylation (58).
Different domains of YY-1 interact in a complex fashion to transactivate gene expression. Experiments using heterologous constructs fused to the Gal4 DNA-binding domain uncovered a bipartite N-terminal transactivation domain including two acidic regions (44,59). The observation that the interaction of YY-1 with TATA-binding protein, TAF II 55, and transcription factor IIB does not require the N-terminal transactivation domain (44,59) suggests that it functions in a non-canonical manner. In addition, the zinc-finger domain is also required for transactivation, in part by mediating nuclear translocation and/or DNA binding (44). Certain carboxyl-terminal sequences appear to mask activation domains of YY-1 (57,59). Growing data suggest that recruitment of a coactivator is an important mechanism by which YY-1 enhances gene expression (43). Candidates in this regard include both CBP (cAMP response element-binding protein-binding protein) (44) and p300 (60). Interestingly, strong interaction with these factors requires an intact zinc-finger domain (44,51,60). Our observation that deletion of residues within the C-terminal zinc-finger domain inhibited the ability of YY-1 to enhance IL-4 promoter activity (ZFDmut) (Fig. 8) is consistent with a role for CBP/p300 in this process. Alternatively, YY-1 could interact with other factors to activate IL-4 gene expression, such as the Elf-related factor CP-2. This is suggested by our recent observation that CP-2 is a critical transactivator of the IL-4 promoter (30) and is supported by the detection of direct interactions between YY-1 and CP-2 in Jurkat cells (50). Studies aimed at identifying the role of cofactors in YY-1-dependent IL-4 gene expression are underway.
IL-4 gene expression is regulated at the transcriptional level by a proximal promoter and other regulatory elements. Although essential for tissue-restricted gene expression (29), the IL-4 promoter alone is a relatively weak activator of transcription. We (61) and others (62) previously showed that the promoter is inducible ϳ3-fold with a calcium signal alone, reflecting the need for NFAT activation. These observations underscore the importance of carefully controlling for the efficiency of transient transfections, as was done throughout this work. Based on cotransfection experiments and site-directed mutagenesis of the IL-4 promoter, we have shown that the Y0 element is required for YY-1-dependent promoter activation. Support for an enhancing role for the Y0 element is provided by a previous study of the mouse IL-4 promoter by Hodge et al. (37), who found that mutation of nucleotides within the murine counterpart of the Y0 sequence ( Ϫ59 TCATTT Ϫ54 to Ϫ59 gacgTT Ϫ54 ) in the context of a 157-bp reporter construct abrogates anti-CD3 antibody-induced IL-4 promoter activity in D10 cells. Although not addressed in that report, our finding that D10 cells constitutively express nuclear YY-1 suggests a role for YY-1 in mouse IL-4 promoter activity in D10 cells. Because the P0 NFAT and Y0 YY-1 sites are intimately associated, precise nucleotide substitutions were needed to distinguish between the cognate sites in each element (Fig. 4). Thus, previous mutational analyses in which the ability of YY-1 to bind to the IL-4 promoter was not accounted for (see, for example, P0 mutants in Refs. 11 and 63) need to be reevaluated. The functional role(s) of the other YY-1-binding sites identified in this study remain to be determined. However, the observations that (i) mutation of Y0 alone in the context of the fulllength promoter abrogated YY-1-dependent transactivation and (ii) overexpressed YY-1 was able to induce transcription in a construct containing only Y0 (145luc) suggest that this proximal element alone plays a critical role in regulating IL-4 transcriptional activation.
YY-1 has recently been shown to regulate the expression of several T cell cytokines. The interferon-␥ promoter contains at least two YY-1-binding sites, and YY-1 was initially shown to inhibit IFN-␥ promoter activity in Jurkat cells in part by competing for AP-1 binding (64). However, in a more recent report, Sweetser et al. (65) concluded that YY-1 acts in a more complex fashion to regulate the IFN-␥ promoter. They found that YY-1 cooperates with NFAT to activate the IFN-␥ promoter in primary mouse splenocytes, but not in Jurkat cells. YY-1 was also shown to down-regulate the expression of the IL-5 promoter in a specific human T cell line (66). Repressor activity appears to involve YY-1 together with Oct-1 and octamer-like proteins (66). These results underscore the complex nature of transcriptional regulation by YY-1 and suggest that the effect of YY-1 on cytokine gene expression will be highly dependent on the promoter context.
Taken together with our findings that YY-1 enhances IL-4 promoter-driven transcription, the apparent ability of YY-1 to differentially regulate IFN-␥ and IL-5 promoter activities (65,66) suggests that YY-1 does not play a dominant role in the polarization of cytokine gene expression in Th cells. In support of this notion, when we analyzed nuclear extracts from established Th1 (AE7) and Th2 (D10) clones by Western blotting, we found that YY-1 was equally expressed in both cell types (data not shown). However, YY-1 could interact with lineage-restricted factors (e.g. T-bet (67) or GATA-3 (20)) to regulate expression of their target genes in Th subsets. YY-1 could also regulate cytokine gene expression under less polarizing conditions, in which, for example, IL-4 and IFN-␥ are coexpressed (68). Alternatively, chromatin remodeling could restrict the accessibility of YY-1 to the IL-4 locus only in activated Th2 cells. Identifying the precise role of YY-1 in the differentiation of T cells will require the creation of both transgenic and YY-1-deficient mice in which YY-1 can be regulated at discrete stages of T cell development.
In summary, we have identified a novel role for the pleiotropic transcription factor YY-1 in enhancing IL-4 gene expression in T cells. YY-1 is generally considered a constitutively expressed factor, although emerging data in other systems suggest that the level of YY-1 protein can be physiologically regulated (69,70). It will be interesting to determine whether altered YY-1 expression contributes to dysregulated IL-4 gene expression in human diseases.