p53 Homologue p63 Represses Epidermal Growth Factor Receptor Expression*

Tumor suppressor p53 has been shown to transactivate epidermal growth factor receptor (EGFR) expression through binding to a putative p53 responsive element in the EGFR promoter between nucleotides (cid:1) 265 and (cid:1) 239 (EGFRp53RE). Isotypes of p63 gene products, recently identified as p53 relatives, have a similar function to transactivate several p53 target gene promoters. However, our results indicate that TAp63 (cid:2) has a very low ability to bind to the EGFRp53RE and surprisingly represses both basal EGFR promoter activity and endogenous EGFR expression. Transient transfection assays show that the EGFR promoter region between (cid:1) 348 and (cid:1) 293, containing two Sp1 sites, is crucial for the repression of the EGFR expression by TAp63 (cid:2) . Mutations in these Sp1 sites in the reporter constructs result in loss of the TAp63 (cid:2) repression effect. We further show that TAp63 (cid:2) directly interacts with Sp1 by immunoprecipitation analysis and that TAp63 (cid:2) impairs Sp1 binding to the target DNA site in electrophoretic mobility

Tumor suppressor p53 has been shown to transactivate epidermal growth factor receptor (EGFR) expression through binding to a putative p53 responsive element in the EGFR promoter between nucleotides ؊265 and ؊239 (EGFRp53RE). Isotypes of p63 gene products, recently identified as p53 relatives, have a similar function to transactivate several p53 target gene promoters. However, our results indicate that TAp63␥ has a very low ability to bind to the EGFRp53RE and surprisingly represses both basal EGFR promoter activity and endogenous EGFR expression. Transient transfection assays show that the EGFR promoter region between ؊348 and ؊293, containing two Sp1 sites, is crucial for the repression of the EGFR expression by TAp63␥. Mutations in these Sp1 sites in the reporter constructs result in loss of the TAp63␥ repression effect. We further show that TAp63␥ directly interacts with Sp1 by immunoprecipitation analysis and that TAp63␥ impairs Sp1 binding to the target DNA site in electrophoretic mobility shift assays. These results suggest that TAp63␥ is involved in the regulation of the EGFR gene expression through interactions with basal transcription factors.
The epidermal growth factor receptor (EGFR) 1 plays an important role in cell growth and development (1)(2)(3). Overexpression of the EGFR can lead to epidermal growth factor-dependent transformation (4,5). High levels of the EGFR have been detected in several types of cancers such as glioblastomas due to gene amplification (6). Overexpression of EGFR transcripts in a variety of tumors such as ovarian, cervical, and kidney tumors results from transcriptional or post-transcriptional mechanisms (7). Also, a variety of agents have been shown to increase EGFR gene expression (8 -10). Repression of EGFR gene transcription by different agents has been also reported (11,12). Thus, transcriptional control plays a major role in regulation of EGFR gene expression.
The promoter of the EGFR gene lacks a TATA box and CAAT box but contains multiple GC boxes and multiple transcription initiation sites. A number of regions in the promoter have been identified that bind nuclear factors (13)(14)(15). Four repressor proteins, EGFR transcriptional repressor, GC-binding factor, GC-binding factor 2, and the Wilms' tumor suppressor bind to sites within the EGFR promoter (16 -19). Sp1, interferon-regulated factor-1, EGFR transcription factor, activator protein-1, and activator protein-2 have been shown to increase EGFR gene transcription (20 -24). Two groups have shown that p53 can transactivate EGFR via binding to an upstream site of the EGFR promoter between Ϫ265 and Ϫ239 base pairs (25,26).
p53 is the most frequently mutated tumor suppressor gene identified thus far in human cancers (27,28). p53 appears to induce cell cycle arrest or apoptosis in response to cellular stresses such as DNA damage, mitotic spindle misassembly, and hypoxia. The actions are mediated through transcriptional regulation of p53 target genes (29,30). Most functions of p53 involve its activity as a transcription factor. p53 binds to its consensus binding sequence (two copies of the 10-base pair motif 5Ј-PuPuPuC(A/T)(T/A)GPyPyPy-3Ј separated by 0 -13 base pairs) and transactivates expression of target genes (31). Several biologically significant genes were found to contain this consensus sequence and to be subjected to p53 regulation. Among those commonly studied are p21 Waf1/Cip1 (29), MDM2 (32), Gadd45 (33), BAX (34), proliferating cell nuclear antigen (35), and cyclin G (36). The p53 mutations found in many human cancers are clustered in the DNA-binding domain of the p53 molecule (37). This leads to an inactivation of p53 function through abolition of p53-specific DNA binding and transactivation. By this means, p53 acts as a tumor suppressor; its loss of function appears to confer selective advantages to cells through deregulated growth and resistance to cell death (38,39).
Two genes that are predicted to encode proteins with amino acid sequence homologous to p53 have been recently identified (40 -47). Each of the p53 amino acid residues that are implicated in sequence-specific DNA binding is conserved in these proteins (40 -47). One of these novel genes, termed p73, is mapped to chromosome 1p36.33 and encodes a protein with homology to p53 (40). Like p53, p73 activates the transcription of p21 Waf1/Cip1 and also induces apoptosis in a p53-independent manner (40,41). Another gene, termed p63/p51/p73L/p40/KET/ CUSP, maps to the long arm of chromosome 3 (42)(43)(44)(45)(46)(47). Although the amino acid sequences and the molecular weights were reported to be different, they have proven to be isotypes derived from a single gene (43). Thus far, there have been six reported isotypes of p63. Three of the variants are called TA (transactivation) type and encode proteins with transactivation, DNA binding, and oligomerization domains similar to p53. The three remaining variants, which lack the acidic N-terminal domain, are called dN (␦N terminus) type. Both types, TAp63 and dNp63, have different C termini that are described as ␣, ␤, and ␥ (43). Thus, the TA types are designated TAp63␣, TAp63␤, and TAp63␥ and the dN types are designated dNp63␣, dNp63␤, and dNp63␥. TAp63␥ has the shortest C terminus and the potential to induce apoptosis and growth suppression in a manner similar to p53. The mechanism is possibly through the p53 regulatory element and includes cellular responses similar to those induced by p53 (42,43). TAp63␥ transactivates several previously identified p53 target gene promoters such as p21 Waf1/Cip1 , BAX, and MDM2 (42,48). Thus, TAp63 is considered to be a tumor suppressor gene and it may serve as an alternative tumor suppressor gene whose expression is induced by the loss of p53 function (42).
On the other hand, variants that are truncated with the acidic N terminus (dN type) or encode the longest C terminus (␣ type) are thought to possess oncogenic properties (43). These variants also act in a dominant-negative manner toward both p53 and transactivating versions of p63 through the reporter construct containing multiple copies of p53-binding sequence termed PG13 (43). In the present study, we examined potential TAp63-dependent transactivation of the EGFR promoter using transfection assays and also TAp63 binding to the EGFR promoter fragments by electrophoretic mobility shift assays. Our results indicate a novel mechanism for regulation of EGFR gene expression through interaction of TAp63 with Sp1.

EXPERIMENTAL PROCEDURES
Cell Lines and DNA Plasmids-The human non-small cell lung carcinoma cell line H1299, which is p53 deficient, was maintained in RPMI 1640 medium (Life Technologies, Gaithersburg, MD) supplemented with 10% fetal bovine serum and antibiotics. The human osteosarcoma cell line Saos-2, which is also p53 deficient, was maintained in McCoy's 5A medium (Life Technologies) with 15% fetal bovine serum.
Transfections and Luciferase Assays-H1299 and Saos-2 cells were seeded at 2.5 ϫ 10 5 cells/35-mm dish and incubated overnight at 37°C in a 5% CO 2 incubator. For each transfection, 0.2-1.0 g of empty vector and/or expression vector plus 0.1 g of promoter-luciferase DNA were mixed in 3.0 ml of Opti-MEM (Life Technologies) and a precipitate was formed using LipofectAMINE (Life Technologies) according to the manufacturer's recommendations. The cells were washed with Opti-MEM and complexes were applied to the cells for 5 h. An equal volume of RPMI 1640 or McCoy's 5A medium containing 20% (30% for McCoy's 5A medium) fetal bovine serum was added, and cells were incubated for an additional 19 h. Cells were harvested and extracts were prepared with luciferase cell lysis buffer (Pharmingen, San Diego, CA). Luciferase activity was assayed in extracts in triplicate using the luciferase assay kit (Pharmingen).
Western Blot Analysis-H1299 cells were seeded at 2.5 ϫ 10 6 cells/ 150-mm dish and incubated overnight at 37°C and then transfected with 15 g of either pcDNA3 empty vector or constructs expressing TAp63␥ tagged at its N terminus with influenza hemagglutinin (HA) peptide by the LipofectAMINE method as described above. After 24 h, the media was changed to selective media containing 700 g/ml Geneticin (Life Technologies) to reduce the number of untransfected cells. Cells were harvested 4 days post-transfection and lysed on ice for 30 min in lysis buffer (10 mM Tris-HCl at pH 8.0, 1 mM EDTA, 400 mM NaCl, 10% glycerol, 0.5% Nonidet P-40, 5 mM sodium fluoride, 0.1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol), containing complete protease inhibitor mixture (Roche Molecular Biochemicals, Indianapolis, IN). The lysate was centrifuged at 14,000 rpm for 15 min and the soluble fraction was collected. Protein concentrations were measured with a Bio-Rad protein assay kit (Bio-Rad). Equal amounts of protein extract (40 g) were loaded onto a 4 -12% SDS-polyacrylamide gel and subjected to electrophoresis at 200 V for 50 min. The proteins were transferred onto a polyvinylidene difluoride membrane and probed with anti-EGFR antibodies (1005) (Santa Cruz Biotechnology, Santa Cruz, CA), anti-HA antibody (F-7) (Santa Cruz Biotechnology), and anti-actin antibody (C4) (Roche Molecular Biochemicals). The blot was probed with each antibody after stripping the membrane of the previous probe. EGFR was detected by horseradish peroxidase-conjugated secondary antibody coupled with enhanced chemiluminescence (ECL) Western blotting detection reagents (Amersham Pharmacia Biotech). The intensity of each EGFR band was normalized based on the intensity of the actin band.
RNA Isolation and Northern Analysis-H1299 cells were seeded at 2.5 ϫ 10 6 cells/150-mm dish and incubated overnight at 37°C and then transfected with 15 g of either pcDNA3 empty vector or constructs expressing TAp63␥ by the LipofectAMINE method as described above. After 24 h, the media was changed to selective media containing 700 g/ml Geneticin (Life Technologies) to reduce the number of untransfected cells. Cells were harvested 4 days post-transfection, and total cellular RNA was isolated using TRIzol reagent (Life Technologies) and quantified by A 260 /A 280 measurement using an Ultraspec 3000 (Amersham Pharmacia Biotech). Total RNA samples (20 g) were subjected to Northern blot analysis. After electrophoresis in MOPS electrophoresis buffer, the RNA was transferred to a nylon membrane in 10 ϫ SSC buffer by capillary action. The RNA was fixed to the nylon membrane by ultraviolet light exposure with a UV Stratalinker (Stratagene, La Jolla, CA). The membrane was hybridized with random-primed 32 P-labeled probes in ExpressHyb hybridization solution (CLONTECH) according to the manufacturer's recommendation. After hybridization at 68°C, the membrane was washed twice in 2 ϫ SSC containing 0.05% SDS at room temperature and then washed twice at 50°C using 0.1 ϫ SSC containing 0.1% SDS. The filters were autoradiographed with Kodak X-AR film for 24 -72 h at Ϫ80°C. The signal obtained from the Northern blots was normalized to the signal for ␤-actin.
In Vitro Transcription/Translation-The p53 and TAp63 cDNAs were subcloned into pcDNA3 and tagged at their N termini with HA peptide. To generate C-terminal truncated proteins, polymerase chain reaction was used to amplify the regions encoding HA and amino acids 1-363 for p53. This region was also subcloned into pcDNA3 and checked for fidelity of DNA sequence. Protein was synthesized in vitro in the presence of unlabeled amino acids with the coupled transcription/translation system (TNT) from Promega. Translated products were analyzed by Western blotting using anti-HA antibody (F-7) (Santa Cruz Biotechnology). For immunoprecipitation, 35 S-labeled p53 and TAp63␥ were synthesized in the presence of 40 Ci of 35 S-labeled methionine and other unlabeled amino acids with TNT system.
Electrophoretic Mobility Shift Assays-Electrophoretic mobility shift assays were performed as described previously (26). Briefly, a doublestranded oligonucleotide containing the p53 consensus DNA-binding site (PG) was prepared by annealing two complementary oligonucleotides, 5Ј-AGCTTAGACATGCCTAGACATGCCTA-3Ј and 5Ј-TAGGCAT-GTCTAGGCATGTCTAAGCT-3Ј, in a buffer containing 10 mM Tris-HCl (pH 8.0), 500 mM NaCl, and 1 mM EDTA. Equimolar amounts of the complementary oligonucleotides were mixed in a 1.5-ml microcentrifuge tube and placed in a heat block at 95°C. The heat block was allowed to cool to room temperature, and the sample was desalted on a G-25 microspin column (Amersham Pharmacia Biotech). The doublestranded oligonucleotide was end-labeled with 32 P using T4 polynucleotide kinase and [␥-32 P]ATP. For electrophoretic mobility shift analysis, the end-labeled double-stranded oligonucleotide 5000 cpm was incubated with 2 l of p53 and TAp63␥ at room temperature (22°C) for 30 min in the presence of a binding buffer (10% glycerol, 20 mM HEPES-KOH (pH 7.5), 25 mM KCl, 2 mM dithiothreitol, 2 mM MgCl 2 , 0.4% Nonidet P-40, and 1 g of salmon sperm DNA). When competition assays were performed, an unlabeled p53 consensus sequence oligonucleotide from p21 Waf1/Cip1 or the EGFR promoter was incubated with protein and buffer for 5 min prior to the addition of the labeled oligonucleotide. Each binding site oligonucleotide was purchased as two single-stranded DNAs from Genosys Biotechnologies (Woodlands, TX) and annealed as described above. Samples (20 l) were loaded onto a 5% nondenaturing polyacrylamide gel and subjected to electrophoresis at 150 V for 1 h using 0.33 ϫ TBE (1 ϫ TBE: 89 mM Tris-HCl, 8 mM boric acid, and 2 mM EDTA, pH 8.3) as running buffer.
For Sp1 binding, the Sp1 consensus oligonucleotide (Promega) was labeled with 32 P as described above. The end-labeled Sp1 consensus oligonucleotide and 5 g of HeLa nuclear extract (Promega) were incubated with or without the in vitro translated p53 or TAp63␥ for 30 min in binding buffer. For competition or supershift assays, 1.75 pmol of unlabeled Sp1 oligonucleotide or 0.2 g of anti-Sp1 antibody (PEP2) (Santa Cruz Biotechnology) was incubated with HeLa nuclear extract and the binding buffer for 5 min prior to the addition of labeled Sp1 oligonucleotide. Samples were subjected to electrophoresis with the same conditions as described above. After electrophoresis, gels were transferred to Whatman 3MM paper and exposed to Kodak X-AR film with intensifying screens at Ϫ80°C.
Immunoprecipitation-Sixty g of HeLa nuclear extract (Promega) and 35 S-labeled p53 or TAp63 were incubated in 500 l of Nonidet P-40 buffer (20 mM Tris-HCl, pH 7.8, 1% Nonidet P-40, 0.1 M NaCl, 10% glycerol, 1 mM EDTA) for 30 min. Two g of anti-Sp1 antibody (PEP2) was added and incubated for an additional 1 h. Subsequently, each immune complex was mixed with 20 l of protein A-agarose beads for 16 h and then washed seven times in Nonidet P-40 buffer and once in RIPA buffer (50 mM Tris-HCl, pH 7.8, 1% Triton X-100, 0.15 M NaCl, 0.1% SDS, 1% sodium deoxycholate). As a negative control, agarose-coupled normal rabbit antibody from Santa Cruz Biotechnology was used. All steps were carried out at 4°C. Immune complexes were boiled in sample buffer and separated on a SDS-polyacrylamide gel. The gels were washed and dried before being exposed to x-ray films at Ϫ80°C.

TAp63␥
Transrepresses EGFR Expression-In this study, we sought to determine whether p63 gene products, which are p53-related molecules, had a similar activating function on EGFR expression. To examine potential p53 and TAp63␥-dependent transactivation of the EGFR promoter, we co-transfected the EGFR luciferase reporters, pER1-luc, with wild-type p53, TAp63␥, or empty vector into p53-deficient H1299 and Saos-2 cells. Similar to the effect on PG13 (43), p53 activated pER1-luc reporter activity (Fig. 1, A and B). While TAp63␥ also transactivated PG13-luc, TAp63␥ surprisingly repressed pER1-luc reporter activity (Fig. 1, A and C). A 83 and 75% decrease in luciferase activity was observed when compared with the empty vector in Saos-2 and H1299 cells, respectively. Additional co-transfection analysis showed that TAp63␥, but not dNp63, repressed pER1-luc reporter activity in a dose-dependent manner (Fig. 1C).
To determine whether TAp63␥ represses endogenous EGFR expression, EGFR mRNA and protein levels were examined by introducing the exogenous TAp63␥ and parental vector plasmids, respectively, into H1299 cells, followed by Northern and Western blot analysis. EGFR mRNA level was decreased by 67% in the TAp63␥-transfected cells as compared with the vector-transfected cells (Fig. 2A). KB cells which have rela- p63 Represses EGFR Gene Expression tively high expression of EGFR was used as a positive control ( Fig. 2A). Also, the EGFR protein level was decreased by 64% in the TAp63␥-transfected cells as compared with the vectortransfected cells (Fig. 2B). The overexpression of TAp63␥ was confirmed by Western blot analysis using anti-HA antibody (Fig. 2B). These results indicate that TAp63␥ represses EGFR expression at the transcriptional level in an endogenous context. Thus, we focused on elucidating the mechanism(s) why TAp63␥, unlike p53, represses the EGFR expression.
TAp63␥ Does Not Bind to the EGFR p53 Response Element-Given the high degree of sequence homology within the DNAbinding domains of the p53 and p63 proteins (42)(43)(44)(45)(46)(47), it is likely that TAp63␥ can bind p53 DNA target sites. We generated a luciferase construct which has four tandem repeat of the p53-binding site identified in the EGFR promoter upstream of the thymidine kinase TATA region and named pERp53RE-luc (Fig. 3). To examine potential p53-dependent transactivation and TAp63␥-dependent transrepression of this construct, we co-transfected the pERp53RE-luc, with wild-type p53, TAp63␥, or an empty vector into H1299 cells. p53 was able to transactivate this construct by more than 70-fold (Fig. 3). Unexpectedly, TAp63␥ also transactivated pERp53RE-luc reporter ac-tivity (Fig. 3). A 10-fold increase in luciferase activity was observed when compared with the empty vector. Similar results were seen in Saos-2 cells (data not shown). As a control, a luciferase reporter containing the minimal thymidine kinase promoter was used. TAp63␥ or p53 did not activate this construct (data not shown). These results imply that TAp63␥ can transactivate using the p53-binding site of EGFR promoter when presented in an oligomeric context.
To address this issue further, we asked whether TAp63␥ could interact with p53 target DNA sites. We used HA-tagged p53 and TAp63␥ proteins prepared in vitro which were approximately equal in concentration as examined by Western blot (data not shown). Ludes-Meyers et al. (25) showed that p53 could bind the oligonucleotide consisting of the p53-binding site identified in the EGFR promoter (EGFRp53RE) (26). However, we could not detect binding of TAp63␥ to the EGFRp53RE (data not shown) but could confirm binding of p53 and TAp63␥ to the p53-binding site (PG) (Fig. 4, lanes 2 and 5). The binding of p53 and TAp63␥ to this element was sequence-specific since the binding was competed by 30-fold excess of the p53 response element of p21 (p21p53RE) (Fig. 4, lanes 3 and 6). Additionally, the binding could be supershifted by anti-HA antibody (data not shown). The binding to p53 was only slightly competed by 100-fold excess of cold EGFRp53RE (Fig. 4, lane 4), and the binding to TAp63␥ was not competed by cold EGFRp53RE oligonucleotide (Fig. 4, lane 7). The band intensities were quantified and the results of three independent experiments are plotted in Fig. 4. Taken together, these results indicate that p53 binds to the putative p53-binding site found in the EGFR promoter with relatively low affinity and p63 binding is not detectable by our assays.
Sp1-binding Sites in the EGFR Promoter Are Critical for TAp63␥-mediated Repression-We next aimed to determine the responsible element(s) in the EGFR promoter for TAp63mediated repression. For this purpose, we generated serial deletion mutants of the EGFR promoter region attached to the luciferase gene (Fig. 5A). Using these reporters, we co-transfected H1299 cells with the TAp63␥ expression vector. The results show that pER1-luc containing Ϫ1109 to Ϫ16, pER9-luc containing Ϫ381 to Ϫ16, and pER9A-luc containing Ϫ348 to Ϫ16 were repressed to the similar extent by co-transfection p63 Represses EGFR Gene Expression with TAp63␥, whereas pER9C-luc and pER10-luc that contain Ϫ292 to Ϫ16 and Ϫ150 to Ϫ16, respectively, were not repressed (Fig. 5B). A similar effect was also found in Saos-2 cells (data not shown). These results suggest that nucleotides between Ϫ348 and Ϫ293 within the EGFR promoter are critical for TAp63␥-mediated repression.
To confirm whether this EGFR promoter region was crucial for TAp63␥ repression, we generated a luciferase construct, which has this region upstream of a minimal thymidine kinase promoter. Co-transfection experiments in Saos-2 cells were performed using this reporter construct, pER348 -293-luc, and the results revealed that TAp63␥ repressed the activity by 67% (Fig. 6). A similar effect was also found using H1299 cells (data not shown).
In the EGFR promoter region between nucleotides Ϫ348 and Ϫ293, there are two Sp1-binding sites, Ϫ323 and Ϫ314 and Ϫ304 and Ϫ296 (13). It has been shown that Sp1 transactivates EGFR promoter activity (13) and that Sp1 interacts with p53 and may interact with p73 (49 -52). Since this region seems to be crucial for EGFR repression by TAp63␥, we tested whether TAp63␥ could bind to this region by electrophoretic mobility shift assay. Our results showed no interaction of TAp63␥ with this region (data not shown). We next examined the possible interaction between TAp63␥ and Sp1. To test whether the TAp63␥ repression activity could be mediated through Sp1binding sites, we mutated these sites in the pER348-293-luc. We transfected the pER348-293mt1-luc, pER348-293mt2-luc, or pER348-293mt1,2-luc into Saos-2 cells along with TAp63␥ or p53 expression plasmids. TAp63␥ repressed pER348-293-luc activity to a similar extent as pER1-luc and mutations in the Sp1-binding sites diminished this repression (Figs. 1 and 6). Also, p53 repressed the activity of this reporter construct. Neither TAp63␥ nor p53 was able to repress the activity when both Sp1 sites were mutated. Similar results were observed in H1299 cells (data not shown). These results suggest that the Sp1-binding sites are critical for TAp63␥-mediated repression of EGFR expression. To further investigate the role of this region in regulating EGFR promoter activity, we generated mutations in the p53-binding site and the Sp1-binding sites in the full promoter. These pER1 mutant constructs were tested for their ability to be regulated by p53 and TAp63␥ (Fig. 7). The pER1p53mt construct was not induced by p53 but was repressed by TAp63␥. The pER1Sp1mt construct was not repressed by TAp63␥ but induced by p53. The induction by p53 was not as strong as with pER1 luc. Mutation of both the p53 and Sp1 sites resulted in loss of p53 induction and TAp63␥ repression. These results confirm that nucleotides between Ϫ348 and Ϫ293 are critical for TAp63␥ mediated repression.
How does TAp63␥ repress EGFR gene expression? The hypothesis that we pursued was that TAp63␥ physically interacted with Sp1 and abrogated binding to the Sp1 elements in the EGFR promoter. To investigate whether there was proteinprotein interaction between Sp1 and TAp63␥, 35 S-labeled TAp63␥ was incubated with nuclear extracts from HeLa cells and Sp1 antibody used to immunoprecipitate complexes. As shown in Fig. 8, anti-Sp1 antibody precipitated TAp63␥⅐Sp1 and p53⅐Sp1 complexes. These results indicated that the TAp63␥ repressive activity on EGFR promoter may be mediated through the Sp1 sites and suggested a potential interference of Sp1-DNA binding by TAp63␥ through protein-protein interaction.
TAp63␥ Represses Several Promoters Containing Sp1-binding Sites-To examine the TAp63␥ effect on additional promoters that have Sp1-binding sites, co-transfection assays were performed using these reporter constructs and TAp63␥. In H1299 cells, co-transfection of TK-luc which has five Sp1-binding sites along with the empty vector pcDNA3 or TAp63 revealed that TAp63␥ could repress TK-luc more than 86%, whereas the minimal TK-luc construct, which contains only one Sp1-binding site, was repressed only 33% (Fig. 9, A and B). SV40-luc which has several Sp1-binding sites was repressed as much as 83%, whereas pCRE-luc and pNF-B-luc which have only one Sp1 site were not repressed or repressed less than 33% (Fig. 9A). These results indicate that TAp63␥ may generally compromise Sp1-mediated transcription.
TAp63␥ Impairs Sp1 Binding to DNA-It has been reported that p53 impairs Sp1 binding to its consensus sequence through protein-protein interaction (49 -51). To determine if TAp63␥ interferes with Sp1 binding to DNA, we performed EMSA experiments using HeLa nuclear extract and TAp63␥. As shown in Fig. 9, a 22-base pair oligonucleotide consisting of the Sp1-binding site, bound to Sp1 in HeLa nuclear extract (Fig. 10, lane 2). This binding was confirmed by supershifting with anti-Sp1 antibody (Fig. 10, lane 3). When TAp63␥ protein was added, the binding of Sp1 was reduced in a dose-dependent manner similar to the effect of p53 (Fig. 10, lanes 7-12). These results indicate that TAp63␥ impairs the binding of Sp1 to the promoter region through its direct interaction with Sp1.

DISCUSSION
In the present study, we investigated the roles of p53-related p63 gene products in the regulation of EGFR expression. We demonstrated reporter assays that showed TAp63␥-mediated repression of EGFR promoter activity (Fig. 1C). The results from Northern blot and Western blot analyses revealed that TAp63␥ repressed endogenous EGFR expression (Fig. 2, A and  B). This was surprising since TAp63␥ has been shown to possess p53-like transcriptional activating activity (43). To clarify such a discrepancy between p53 and TAp63␥ in the regulation of EGFR promoter, we demonstrated another set of reporter assays using pERp53RE-luc, which harbors four copies of the p53-responsible element found in the EGFR promoter region (Fig. 3). The results showed that TAp63␥ did not transactivate pERp53RE-luc activity to the same level as p53 (Fig. 3). EMSAs using EGFRp53RE as a probe showed that TAp63␥, in contrast to p53, did not bind to the EGFRp53RE (Fig. 4 and data not  shown). This is not surprising because p53 and TAp63␥ showed p63 Represses EGFR Gene Expression different profiles of DNA binding affinity to the same p53-responsible elements such as p21p53RE, mdm2p53RE, cyclinGp53RE, and BAXp53RE. 2 It is, therefore, possible to speculate that binding affinity of p53 and TAp63␥ to the EGFRp53RE is above and below detection at our level, respectively.
We determined by deletion mutants of pER1-luc that the responsible region of TAp63␥-mediated repression was between nucleotides Ϫ348 and Ϫ293 (Fig. 5B). Previous studies have shown that there are two Sp1-binding sites in this region (13). Mutation analysis in these two Sp1-binding sites showed that TAp63␥-mediated repression was largely dependent on the intact Sp1-binding sites (Figs. 6 and 7). Sp1 is a well known basal transcription factor that is involved in the regulation of several important genes (53). Therefore, we next examined possible association between Sp1 and TAp63␥ by immunoprecipitation analysis. The results show that TAp63␥ could physically associate with Sp1 as does p53 (Fig. 8). In this point of view, p53, as well as TAp63␥ in this study, might act as a negative regulator of certain gene promoters. In fact, p53 literally represses the transcription of HIV long terminal repeat, SV40 and telomerase reverse transcriptase, through forming a Sp1⅐p53 complex which no longer can mediate gene transcription (49 -51). Additionally, p73, another p53 family member, has been reported to repress vascular endothelial growth factor transcription through interference with Sp1 binding (52). These observations propelled us to investigate whether TAp63␥ compromises the ability of Sp1 to bind to the target sites in the EGFR promoter region. The results showed that TAp63␥, as well as p53, inhibited Sp1 binding to the target DNA fragment in a dose-dependent manner (Fig. 10). Altogether, it seems possible to generalize that p53/p73/TAp63␥ transactivates genes harboring their binding site(s) in the promoter region and that p53/p73/TAp63␥ can act as a negative regulator when the promoter lacks their binding site(s) and is regulated by Sp1. The results in Fig. 9, a finding that repression by TAp63␥ depended on the copy number of Sp1-binding site in the promoter region, also supports this speculation. Electrophoretic mobility shift assays were performed with the end-labeled Sp1 consensus oligonucleotide. Sp1 from HeLa nuclear extract (NE) bound to Sp1 consensus oligonucleotide. For the supershift assay, anti-Sp1 antibody (0.2 g) was added into binding reaction. For the competition assays, 100-fold molar excess of the unlabeled Sp1 consensus oligonucleotide was used. To examine the TAp63␥ effect on Sp1 binding to DNA, different amounts of TAp63␥ were added into each reaction. p53 was used as a positive control and rabbit reticulolysate (RL) devoid of in vitro expressed protein was used as a negative control.  35 S-Labeled p53 or TAp63␥ was analyzed in immunoprecipitation assays as described under "Experimental Procedures." No binding was observed using rabbit normal IgG instead of anti-Sp1 antibody. The input represents 10% of the amount of labeled protein used in each immunoprecipitation assay.
FIG. 9. Effect of TAp63␥ expression on promoter activity. A, H1299 cells were co-transfected with 1.0 g of pcDNA3 or the TAp63␥ expression plasmid and 0.1 g of the each promoter reporter plasmid. Luciferase assays were performed after transfection. B, transcriptional repression by TAp63␥ was decreased by deletion of Sp1 sites. The HSV-thymidine kinase promoter and minimal TK promoter were assayed for TAp63␥ repression. Error bars indicate standard deviation in triplicate assays.

p63 Represses EGFR Gene Expression
We have clearly demonstrated that TAp63␥ represses EGFR promoter activity through a mechanism that involves the Sp1binding sites. One initially could speculate that the interaction of Sp1 and TAp63␥ could result in a complex that inhibits the role of Sp1 in EGFR activation. However, p53, which interacts with Sp1 with a similar affinity, is able to activate EGFR promoter activity through direct binding to the promoter region. On the other hand, in the absence of a binding site, p53 leads to repression of promoter activity to a similar extent as TAp63␥ (Fig. 6). Thus, the mechanism(s) by which changes in EGFR promoter activity are mediated by TAp63␥ may involve its interaction with Sp1 and/or another as yet unidentified factor.
Overexpression of EGFR has been implicated in a number of malignant tumors (4 -7). Despite the extensive studies on the regulation of EGFR gene expression, little is known so far concerning the negative regulator(s) of its expression. Such negative regulators are potential explanations for the elevated level of EGFR gene expression, when they are inactivated, in human cancer cells. Our present results suggest that TAp63␥ might be a good candidate as a negative regulator of EGFR expression. p63 is necessary for limb and craniofacial development and a mutation has been found in an autosomal dominant disorder (54 -56). Up-regulation of the EGFR has been linked to defects in development (57). Although the p63 gene was originally isolated as a p53 relative, it is not frequently mutated in human cancers examined to date (58 -61). However, we and Park et al. (61) have reported that expression of TAp63 was frequently lost in certain epithelial cancer cells, but was constantly expressed in normal epithelial cells (61)(62)(63).
In summary, we have shown that TAp63␥ represses EGFR gene expression through a mechanism involving factors associated with the Sp1-binding site. These findings, showing that TAp63␥ can act as a transcriptional repressor of the EGFR promoter, may suggest an anti-oncogenic role of TAp63␥ in human epithelial cells. The ability of TAp63␥ to prevent malignant transformation of epithelial cells by repressing EGFR gene expression, and the correlation between TAp63␥ and EGFR expressions in human tumors are under investigation.