p53-dependent Transcriptional Regulation of theAPC Promoter in Colon Cancer Cells Treated with DNA Alkylating Agents*

The APC(adenomatous polyposis coli) gene product is involved in cell cycle arrest and in apoptosis. The loss of APC function is associated with the development of colorectal carcinogenesis. In previous studies, we have shown that theAPC gene is inducible and that the DNA damage-induced level of APC mRNA requires p53. In the present study, we examined the role of p53 in the transcriptional regulation ofAPC promoter and characterized two p53-binding sites on the cloned APC promoter (pAPCP). Results of electrophoretic mobility shift assay showed specific interactions of p53 protein with p53-binding site oligonucleotides. The DNA-protein complex formed in electrophoretic mobility shift assay was competed with unlabeled excess of p53-binding site oligonucleotide, unaffected with p53-binding site mutant or Sp1-binding site oligonucleotides, and supershifted with anti-p53 antibodies. In a transient transfection assay, the pAPCP promoter activity was lower in HCT-116(p53+/+) cellsversus HCT-116(p53−/−) cells. p53-dependent down-regulation was further confirmed after co-transfection of pAPCP plasmid with pCMV-p53 into HCT-116(p53−/−) and SAOS-2 (p53-negative) cells. However, the treatment of cells with DNA alkylating agents methylmethane sulfonate andN-methyl-N′-nitro-N-nitrosoguanidine, which cause phosphorylation of p53 at Ser15 and Ser392, induced pAPCP promoter activity in HCT-116(p53+/+) cells. Other than p53-binding sites, using deletion mutation constructs, we have shown thatN-methyl-N′-nitro-N-nitrosoguanidine-induced transcriptional activation of the pAPCP promoter in HCT-116(p53+/+) cells depended upon the Sp1-binding site and the E-box B site. From these results, we conclude that unphosphorylated p53 can down-regulate and phosphorylated p53 can up-regulate the pAPCP promoter activity involving the p53, Sp1, or E-box B elements. These studies are important to understanding the role of p53 and APC in DNA damage-induced cell cycle arrest and/or apoptosis of cancer cells.

The development and progression of colon cancer is a multistep process in which growth control is progressively impaired. Mutations of the APC 1 (adenomatous polyposis coli), Ki-ras, deleted in colorectal cancer (DCC), DNA mismatch repair (MMR), and p53 genes play important roles at different stages of colorectal carcinogenesis (for review, see Ref. 1). Mutation of the APC gene is an early event in familial adenomatous polyposis, a syndrome of inherited predisposition to colon cancer (2,3).
Mutations in the APC gene are also found in 60 -80% of sporadic colorectal cancers and adenomas (1). Patients with APC mutations are more prone to hundreds to thousands of colorectal adenomas and early onset carcinoma. Familial adenomatous polyposis patients are also prone to small intestinal adenomas (and carcinomas), intra-abdominal desmoids and osteomas tumors (Gardener syndrome), congenital hypertrophy of retinal pigment epithelium, fundic gland polyps in the stomach, pancreas, and thyroid, dental abnormalities, and epidermal cysts (for reviews, see Refs. 1 and 4). Mutations in APC are also associated with malignant brain tumors (Turcot's syndrome; Ref. 5), and the APC locus on chromosoyme 5q21 shows loss of heterozygosity in ϳ25% of breast cancers (6). Approximately 18% of somatic breast cancers carry APC gene mutations (7). Furthermore, loss of heterozygosity at the APC gene locus is prominent in the early stages of non-small cell lung cancers (8). An animal model for carcinogen-induced lung tumors was used to show association with a decrease in the APC gene expression instead of an increase in APC mutations (9).
Although APC is expressed constitutively within the normal colonic epithelium, little is known about how mutations of (or abnormal expression of) APC contribute to the development of colon cancer. The APC gene product is a 310-kDa homodimeric protein localized both in the cytoplasm and in the nucleus (Refs. 10 -13; for reviews, see Refs. 1 and 4). Previous studies indicate that the cellular level of wild-type APC is critical to cytoskeletal integrity, cellular adhesion, and Wingless/Wnt signaling (for reviews, see Refs. 1 and 4). In addition, APC may act as a negative regulator of ␤-catenin signaling in the transformation of colonic epithelial cells (14,15), and in melanoma progression (16).
In a simple model, Wingless/Wnt signaling regulates the assembly of a complex consisting of Axin (and its homolog Axil and conductin), APC, ␤-catenin, and glycogen synthase-3␤ kinase (GSK3␤). Axin (Axil/conductin) binds to APC, ␤-catenin, and GSK3␤ and thereby promotes ␤-catenin phosphorylation and subsequent ubiquitination and degradation in the proteasome (17,18). GSK3␤ regulates this process by phosphorylating components of the complex (19,20). Activation of the Wingless/Wnt signaling pathway inhibits GSK3␤ and stabilizes ␤-catenin (21)(22)(23). Stabilizing mutations in ␤-catenin or truncation in APC also occur both in colon cancer and melanoma cells and increase the stability of ␤-catenin (16,23). The stabilized pool of ␤-catenin associates with members of the Tcf-Lef family of transcription factors and regulates transcriptional expression of proto-oncogene and cell cycle regulator c-myc (24), the G 1 /S-regulating cyclin D1 (25), the gene encoding the matrix-degrading metalloproteinase, matrysin (26), the AP-1 transcription factors c-jun and fra-1, and the urokinasetype plasminogen activator receptor (27). Thus, the regulation by APC of the level of ␤-catenin plays a role in ␤-catenin/Tcf-Lef-mediated transcriptional regulation of genes. However, the mechanisms by which these transcriptional changes contribute to early stage colorectal carcinogenesis are still unclear.
Tumor suppressor p53, which plays a critical role in the cellular response to DNA damage, oxidative stress, and hypoxia, is also frequently mutated in colon cancer cells (for reviews, see Refs. 28 and 29). The target genes regulated by p53 include p21 WAF1/Cip1 , GADD45, mdm2, bax, and cyclin G (29). The transcriptional activation function of p53 requires post-translational modification of the p53 protein via phosphorylation, acetylation, or glycosylation (for reviews, see Refs. 29 and 30). Several reports show that p53 becomes phosphorylated at the serine and threonine residues after cells are treated with DNA-damaging agents (31)(32)(33)(34)(35)(36). In recent studies, we and others have reported that p53 is phosphorylated on Ser 15 and Ser 392 in HCT-116 cells treated with DNA alkylating agents N-methyl-N-nitrosourea, N-methyl-NЈ-nitro-N-nitrosoguanidine (MNNG), and methylmethane sulfonate (MMS) (37). 2 In our earlier studies, a link between APC and p53 was established in which APC mRNA levels were increased in a p53-dependent manner by DNA alkylating agents (38,39). To further understand the mechanisms of transcriptional regulation of APC, we recently cloned the APC promoter and identified transcriptional regulatory elements (40). The APC gene has a TATA-less promoter and contains consensus binding sites for octamer, AP2, Sp1, a CAAT-box, and three nucleotide sequences for E-box A, B, and M. In those previous studies, we showed that the APC gene is transcriptionally regulated by upstream stimulating factors 1 and 2 (USF1 and USF2; Ref. 40). In addition, we have identified two p53-binding elements on the APC promoter. The objective of the present study is to characterize the p53-binding element of the APC promoter and to examine the mechanisms by which p53 regulates DNA alkylation-induced APC gene transcription.

MATERIALS AND METHODS
Cell Lines and Treatments-Human colon cancer cell lines HCT-116(p53 Ϫ/Ϫ ) and HCT-116(p53 ϩ/ϩ ) and SAOS-2 cell lines were grown in McCoy's 5a and Dulbecco's modified Eagle's medium, respectively, and supplemented with 10% fetal bovine serum (Life Technologies, Inc.) at 37°C under a humidified atmosphere of 5% CO 2 . After cells reached 70% confluence, fresh medium without fetal bovine serum was added to each dish and then further incubated for an additional 20 h. Treatment regimens with DNA alkylating agents MNNG and MMS (Aldrich) are given in the figure legends.
Oligonucleotides and Plasmids-For use in electrophoretic mobility shift assay (EMSA), complimentary single-stranded oligonucleotides of p53-binding sites (indicated as P1 and P2) of the cloned APC promoter (pAPCP) were annealed to produce double-stranded oligonucleotides with the indicated sequence (see Fig. 1). Several other oligonucleotides were synthesized with XbaI and KpnI restriction endonuclease sites on the 5Ј and 3Ј ends of the P1 and P2 sites. Recently, we have shown that the APC promoter contains multiple transcriptional initiation sites, i.e. ϩ1, ϩ52, ϩ95, ϩ121, ϩ123, and ϩ134 (40). Three major initiation sites at ϩ1, ϩ52, and ϩ95 were identified by in vitro run-off transcription assay and by primer extension analysis (40). In the present study, the P1 and P2 oligonucleotides were joined with ϩ95 transcriptional initiation site oligonucleotides for cloning and gene expression studies. After annealing with the complementary oligonucleotides, DNA was digested with XbaI and KpnI restriction endonucleases and then subcloned into CAT reporter plasmids. The cloned plasmids and their oligonucleotide sequences are given in Fig. 1.
Electrophoretic Mobility Shift Assay-For EMSAs, the recombinant human p53 protein was an overexpressed and purified from baculovirus-infected insect cell. DNA-protein binding reactions were carried out in 20 l of final volume containing 20 mM HEPES, pH 7.9, 1 mM dithiothreitol, 3.5 mM MgCl 2 , 100 mM KCl, 0.03% (v/v) Nonidet P-40, 10% (v/v) glycerol, 1 g of poly(dI⅐dC), and different concentrations of purified p53 protein. Reactions were carried out for 10 min, followed by the addition of 1 ng of 32 P-labeled p53-binding oligonucleotide of the pAPCP or p21P promoter and incubated further for 20 min at 22°C. The entire reaction mixture was loaded directly on a native 4% polyacrylamide gel. The electrophoresis conditions were same as described by Gu and Roeder (41). After electrophoresis, the DNA-protein complexes were visualized by autoradiography. For competition experiments, a molar excess of unlabeled oligonucleotide was added to the reaction mixture 10 min before the addition of 32 P-labeled probe, as indicated in the figure legends. For supershift analysis, 1 g of anti-p53 antibody Ab-1 or DO-1 was added to the reaction mixture and incubated for 20 min before the addition of 32 P-labeled oligonucleotide.
Transient Transfection of Cells-HCT-116(p53 Ϫ/Ϫ ), HCT-116(p53 ϩ/ϩ ), and SAOS-2 cell lines were transfected with the indicated plasmids using LipofectAMINE reagent (Life Technologies, Inc.) as described earlier (39). We used p21 promoter (p21P, cloned in pGL2 reporter plasmid), which is well characterized for the mechanism of p53-dependent transcriptional regulation (42). For comparative purposes, the pAPCP wild-type promoter was also cloned into a pGL2 reporter plasmid while other constructs were cloned into the CAT reporter plasmid. Therefore, both luciferase and CAT reporter assays were performed to determine the promoter activity. Briefly, cells were grown to 60% confluence in 60-mm tissue culture dishes and transfected with 1.5 g/ml of APC promoter constructs, 1.5 g/ml of the p21P plasmid, 0.25 g/ml of the pCMV-␤-gal plasmid, and 14 g/ml of the LipofectAMINE reagent. pCMV-␤-gal served as an internal control to correct the differences in the transfection efficiency. A DNA-lipid complex was then added to the cells according to the recommended protocol of the manufacturer. After 24 h of post-transfection cells were starved in a serum-free medium for an additional 18 h and then treated with DNA alkylating agents for 50 h. After the treatment, cellular lysates were prepared for luciferase or CAT reporter assays. Quantitative analysis of the data was done by electronic autoradiography using Instant Imager from Packard Instrument Company (Meriden, CT). Luciferase activity was measured using Monolight TM 3010 illuminometer from Pharmingen (San Diego, CA)

Characterization of p53-binding Sites of APC Promoter Us-
ing EMSA-p53 binds to DNA in a sequence-specific manner and modulates the expression of target genes (28,29). To examine the role of p53 in transcriptional regulation of APC, we first characterized the p53-binding element of the pAPCP promoter. Based on sequence comparisons, two potential p53-binding sites P1 and P2 of 10 base pairs each were identified at Ϫ144 to Ϫ135 nucleotides (5Ј-GGGCATACCC-3Ј) and Ϫ122 to Ϫ113 nucleotides (5Ј-GGGCTAGGGC-3Ј), respectively. These sites are separated by 12 base pairs and have 90 and 80% homology to known consensus p53-binding sites (43,44). The consensus DNA-binding site for p53 is defined to contain two copies of a 10-base pair motif that are separated by 0 -13 base pairs. Based on this criterion, the P1 and P2 sites of the pAPCP plasmid appear to be genuine p53-binding sites. Because the purpose of this study is to elucidate the mechanisms by which the APC gene is transcriptionally regulated by p53, it was necessary to first determine whether the P1 and P2 sites of the pAPCP plasmid are functional cis-regulatory elements. To begin understanding the role of the P1 and P2 sites in the regulation of APC gene expression, the DNA binding activity of p53 was measured in EMSA with purified p53 and double-stranded P1 and P2 oligonucleotides from the pAPCP (Fig. 1). In these experiments, we also used p21P promoter oligonucleotide to serve as a positive control (44). The p53 protein formed a complex in a concentration-dependent manner with p21P, P1ϩ2, P1, and P2 oligonucleotides (Fig. 2). The migration of the DNA-p53 protein complex with p21P oligonucleotide was faster than P-oligonucleotides because of a difference in the size of their oligonucleotides. The DNA-p53 protein complexes formed with 32 P-labeled P1ϩ2 and p21P oligonucleotides were eliminated by competition with an excess of unlabeled P1ϩ2 and p21P oligonucleotides (Fig. 3A, lanes 2-4 and 5-7, respectively). On the other hand, the DNA-p53 protein complex formed with 32 P-labeled P1ϩ2 oligonucleotide was unaffected with a similar concentration of unlabeled Sp1 or P1ϩ2 mutant (Pmut) oligonucleotides (Fig. 3A, lanes 8 -10 and 11-13, respectively).
To further examine the specificity of p53 protein binding to p21P and P1ϩ2 oligonucleotides, we performed a similar EMSA in the presence of anti-p53 antibody Ab-1 or DO-1. A supershifted DNA-p53 protein complex was formed with these oligonucleotides (Fig. 3B, lanes 2 and 3 and lanes 5 and 6,  respectively). These results indicate that p53 protein binds to the pAPCP promoter at the P1ϩ2 sites whose binding affinity is similar to the p53-binding site of the p21P promoter. These results also indicate that p53 may influence pAPCP promoter activity after binding with the P1 and P2 sites.
Down-regulation of the pAPCP Promoter Activity in Untreated Cells containing Wild-type versus Knock-out p53 Gene-To investigate the possible role of p53 in p53-mediated transcriptional regulation of APC promoter, we used the HCT-116 cell line with the wild-type p53 gene (p53 ϩ/ϩ ) or a homozy-gous cell line with the p53 gene knock-out (p53 Ϫ/Ϫ ), and the SAOS-2 cell line with a deleted p53 gene. As a control, we always performed parallel experiments with p21 promoter, which is well known to exhibit p53-dependent transcriptional regulation (42). First, we determined the effect of pAPCP and p21P plasmid concentration on luciferase reporter activity in the HCT-116(p53 Ϫ/Ϫ ) and HCT-116(p53 ϩ/ϩ ) cell lines. Results showed a significantly lower pAPCP promoter activity in HCT-116(p53 ϩ/ϩ ) cells versus HCT-116(p53 Ϫ/Ϫ ) cells. On the other hand, the p21P promoter activity was significantly higher in HCT-116(p53 ϩ/ϩ ) cells versus HCT-116(p53 Ϫ/Ϫ ) cells (Fig. 4). Results with the p21P promoter are similar to those reported earlier (42).
Up-regulation of the pAPCP Promoter Activity in HCT-116(p53 Ϫ/Ϫ ) and HCT-116(p53 ϩ/ϩ ) Cell Lines Treated with DNA Alkylating Agents MNNG and MMS-In earlier studies, we found a p53-dependent increase of APC mRNA in cells treated with DNA alkylating agents (38,39). Nuclear run-on assays showed that nascent APC mRNA synthesis was higher in nuclei isolated from MNNG-treated versus untreated (control) cells, indicating that transcriptional mechanisms were involved in the regulation of APC after DNA alkylation damage (38). Therefore, to test that p53 can up-regulate APC promoter activity after stress, such as DNA alkylation damage, we transfected pAPCP plasmid into HCT-116(p53 Ϫ/Ϫ ) and HCT-116(p53 ϩ/ϩ ) cell lines. CAT reporter activity of the pAPCP promoter was determined after treatment of cells with different concentrations of MMS or MNNG. Results showed a concentration-dependent increase in the pAPCP promoter activity in both HCT-116(p53 Ϫ/Ϫ ) and HCT-116(p53 ϩ/ϩ ) cell lines. However, the extent of increase was higher in HCT-116(p53 ϩ/ϩ ) cells than in HCT-116(p53 Ϫ/Ϫ ) cells (Fig. 6). These results indicate that in response to DNA alkylation damage the pAPCP promoter activity is up-regulated in HCT-116 by both p53-dependent and independent mechanisms, although a higher increase in the pAPCP promoter activity was seen in the presence of p53.
Role of p53-binding Sites in MNNG-induced Transcriptional   FIG. 1. Structure of the oligonucleotides used in gel shift analysis and for cloning of p-53-binding sites of the pAPCP promoter into the CAT reporter plasmid. Plasmid p(P1ϩ2)P contains both p53-binding sites (Ϫ150 to Ϫ109 nucleotides) joined with transcriptional initiation site at ϩ95 (ϩ88 to ϩ105 nucleotides). Plasmids p(P1)P and p(P2)P contain distal and proximal p53-binding sites, respectively, with transcriptional initiation site at ϩ95 (ϩ88 to ϩ105 nucleotides). The bold letters of the oligonucleotide sequences indicate the p53-binding sites P1 and P2, and the lowercase letters indicate the mutations in these sites. The underlined letter indicates the transcriptional initiation site of the cloned promoters.
FIG. 2. Complex formation of p53 protein with the p53-binding site oligonucleotide of the APC promoter. A gel shift analysis of purified p53 protein and 32 P-labeled p53-binding site oligonucleotides (P1ϩ2, P1, and P2) of the APC promoter DNA was performed to determine protein-DNA complex formation. We used consensus p53-binding site oligonucleotide of the p21 promoter as a control (CON, lanes 1-4).
The arrows show DNA-protein complexes and free probes. The photograph of the autoradiograms is representative of three independent experiments.
Up-regulation of the pAPCP Promoter-Once we established that DNA alkylation-induced regulation of the APC promoter in HCT-116(p53 ϩ/ϩ ) cells is dependent upon p53, we examined whether p53 and other cis-elements play a role in transcriptional up-regulation of the APC gene. To test the function of p53-binding sites and the role of other elements, if any, in DNA alkylation-induced regulation of APC gene, we created nine deletion mutants of the pAPCP plasmid. These mutants were transfected into HCT-116(p53 ϩ/ϩ ) cells and tested for MNNG response in a CAT reporter assay (Fig. 7). Results showed a significant increase in the promoter activity of several deletion mutants including pAPCP(542), pAPCP(569), pAPCP(592), pAPCP(622), pAPCP(668), and pAPCP(737) plasmids. The pAPCP(737) plasmid containing E-box B and E-box M sites was the minimum promoter that responded significantly to MNNGinduced promoter activity. Further deletion of the E-box B site (i.e. pAPCP(752)) abolished the MNNG-induced transcriptional activity of the promoter that now contained only the E-box M site. This result suggested that MNNG treatment might activate APC gene expression through the APC promoter E-box B site. The maximum response of the MNNG treatment on the pAPCP promoter activity was found with pAPCP(592) plasmid, a construct containing the Sp1-binding site but lacking the p53-binding sites (Fig. 7). Because further deletion of the pAPCP plasmid retained similar promoter up-regulation until the E-box B site was deleted, the results indicated that the E-box A and the CAAT-box might not be important for MNNGinduced transcriptional up-regulation of the pAPCP promoter. From these results it can be concluded that DNA alkylationinduced expression of APC gene transcription may involve p53and Sp1-binding sites and the E-box B site.
Because the deletion mutation search did not firmly assign the role of p53-binding sites in DNA alkylation-induced APC gene expression, additional experimentation will be required to determine whether p53 was involved in this regulatory process. Given that DNA alkylation-induced pAPCP activity was higher in HCT-116(p53 ϩ/ϩ ) cells than in HCT-116(p53 Ϫ/Ϫ ) cells, the involvement of p53 cannot be ruled out (Fig. 6). Thus, to further determine whether the p53-binding sites of the pAPCP promoter can function independently in response to DNA alkylation damage or whether the presence of other cis-elements is required, synthetic oligonucleotides corresponding to the P1ϩ2, P1, or P2 sites and an oligonucleotide containing the ϩ95 transcriptional initiation site were cloned into CAT reporter plasmids (Fig. 1). First, we examined whether the p53- binding site mediated the down-regulation of the pAPCP promoter activity in the untreated cells. The p(P1ϩ2)P, p(P1)P, and p(P2)P plasmids were co-transfected with a pCMV-p53 overexpression plasmid into HCT-116(p53 Ϫ/Ϫ ) cells, and then the promoter activity was determined by CAT reporter assay. We found no effect of p53 overexpression on the promoter activity of p(P1ϩ2)P, p(P1)P, or p(P2)P plasmids, which indicates that the p53-binding site is not involved in down-regulation of the pAPCP promoter activity (Fig. 8A). Therefore, the down-regulation of the pAPCP promoter activity seen in the experiments described above (Figs. 4 and 5) must be mediated by other uncharacterized cis-elements.
To examine whether the plasmids containing the p53-binding site respond to MNNG treatment, we transfected pAPCP, p(P1ϩ2)P, p(P1)P, and p(P2)P plasmids into HCT-116(p53 ϩ/ϩ ) cells followed by treatment with MNNG. The results showed a significant increase in the activity of pAPCP, p(P1ϩ2), and p(P1)P promoters; however, the activity of the p(P2)P promoter was lower than that of p(P1)P (Fig. 8B). Also, when both P1 and P2 sites were together, the p(P1ϩ2)P promoter activity was lower than p(P1)P. These results suggest that the P2 site does play a role in p53-mediated up-regulation of the pAPCP promoter activity in response to DNA alkylation damage. DISCUSSION Colon cells can operate normally if they retain one normal APC allele and make sufficient functional APC protein (for review, see Ref. 1). However, once both alleles are damaged, the important brake on tumor suppression mediated by APC is lost. Previous findings suggest that the phenotype of a cell expressing a mutated APC gene can be reverted by increased expression of the remaining wild-type APC allele and provide evidence that overexpression of the APC wild-type gene alone can suppress tumorigenicity (46 -48). Thus, understanding the mechanisms by which APC gene expression can be induced in colon cancer cells may lead to the design of more effective chemopreventive drugs that block tumor cell progression. In previous studies, we have shown that the APC gene was transcriptionally activated in response to DNA alkylation damage and that this increase was dependent upon wild-type p53 (38,39). In the present study, we characterized the APC promoter for the p53-binding element and examined whether p53-dependent transcriptional up-regulation of the APC gene in re-sponse to DNA alkylation damage was mediated by p53 involving the p53-binding site of the APC promoter. The structure of the APC promoter has been previously described (40,49,50). Recently, we reported that APC gene could be regulated by USF1 and USF2 through E-box sites (40). Using electrophoretic mobility shift and supershift analysis, we found a specific binding of p53 protein with p53-binding sites of the APC promoter and showed that these sites are transcriptionally functional and respond to DNA alkylation damage. Currently, the mechanisms are unknown by which p53 interacts with the transcriptional machinery and regulates the rate of APC gene transcription in the presence or the absence of signals generated by DNA alkylating agents. p53 activates transcription from specific DNA-binding sites and represses transcription in a binding site-independent manner. Treatment of cells with DNA-damaging agents induces nuclear accumulation of p53, which trans-activates cell cycleand apoptosis-related genes (for reviews, see Refs. 28 -30). p53 is post-translationally modified by phosphorylation, acetylation, or glycosylation, which could be important in understanding how cells detect and respond to DNA damage (for reviews, see Refs. 28 -30). Because p53 is activated by a variety of DNA-damaging agents, it appears to be a universal signaling molecule for the DNA damage response (51). p53 is phosphorylated at multiple sites in both the amino-and carboxylterminal regions of the molecule (Refs. 31-36; for reviews, see Refs. 28 -30 and 52). The significance of the p53 phosphorylation is associated with sequence-specific DNA binding. Recently, we reported that the p53 of HCT-116 cells is phosphorylated at Ser 15 and Ser 392 after treatment with MMS. 2 Duckett et al. (37) reported a similar phosphorylation of p53 in HCT-116 cells after treatment with MNNG and N-methyl-N-nitrosourea.
In the present studies, a down-regulation of APC promoter activity might be a result of unphosphorylated form of p53, which is present in untreated (unstressed) cells. However, once the p53 was phosphorylated at Ser 15 and Ser 392 in response to DNA alkylating agents, it activated APC gene expression. Currently, it is unclear how unphosphorylated or phosphorylated p53 is involved in the transcriptional regulation of APC gene expression. In previous studies, it has been suggested that phosphorylation of p53 enhances its binding activity for the DNA consensus sequence that could strongly activate gene expression (53). On the other hand, phosphorylation of p53 did not affect its binding activity to DNA consensus sequence in the ribosomal gene cluster or muscle creatine kinase (54,55). Our recent studies showed that phosphorylation of p53 at Ser 15 and Ser 392 enhanced consensus DNA binding activity but decreased the p21 promoter activity. However, unphosphorylated p53 enhanced p21 promoter activity, but the opposite effect was seen with APC promoter, where the phosphorylated form of p53 up-regulated activity and the unphosphorylated form of p53 down-regulated APC gene expression. One of the differences in p21 and APC promoters is that the p21 promoter contains a TATA-box and the APC gene contains a TATA-less promoter (40,42). These studies indicate that differential phosphorylation of p53 may have differential effects on the regulation of target genes by modulating DNA binding activities and interacting with other transcription factors. For example, p53 represses the transcriptional activity of TATA-box-containing promoters but either has no effect or up-regulates TATA-less promoters (56).
The interaction of p53 with the transcriptional preinitiation complex of the TATA-less promoter of the APC gene could be through either p53-binding site or through protein-protein interaction involving the Sp1-binding site or the TFIIH complex. The results presented in this study showed that the APC promoter has functional p53-binding sites that respond to DNA alkylation damage. However, after DNA alkylation damage not only the wild-type or p53-binding site-containing plasmids, but also pAPCP(592), pAPCP(622), pAPCP(668), and pAPCP(737) plasmids showed increased promoter activities. These results indicate that p53 may regulate APC promoter activity through protein-protein interactions as well. The increased promoter activity observed with the pAPCP(592) plasmid in response to DNA alkylation damage can be attributed through one or more than one of the following elements: the Sp1-binding site, the E-box A, B, and M sites, and the CAAT-box site. The maximum response of MNNG treatment, observed with the Sp1-binding site-containing plasmid pAPCP(592), was decreased with the pAPCP(622) plasmid that lacks Sp1-binding site. Furthermore, a significantly increased promoter activity was seen with pAPCP(737) plasmid, which did not contain p53-and Sp1binding sites and E-box A and CAAT-box sites; instead it contained E-box B and E-box M sites. Further deletion of the E-box B site from the pAPCP(737) plasmid caused the MNNG re-sponse to be lost, indicating that E-box B site was important for MNNG-induced APC promoter activity. Thus, the Sp1-binding site and E-box B site may be involved in p53-mediated expression of APC gene transcription. Furthermore, from these results, we cannot rule out that the Sp1-binding site or the E-box B site is functioning independently of p53 in MNNG-induced expression of APC gene transcription. In fact, we have recently shown that the E-box B site and the USF1 and USF2 factors binding to this site are important for transcriptional regulation of the pAPCP promoter (40). Currently, we do not know how p53 interacts with the Sp1-binding site or the E-box B site of the APC promoter. However, from other studies we know that the interaction of Sp1 and p53 can occur through a GC-rich region (Sp1-binding site) as found with SV40 virus (57,58) or through protein-protein interaction as described with insulinlike growth factor-1 receptor gene (59). Our future studies will be focused to examine these possibilities.
TFIIH is another possible factor through which p53 can be involved in p53-binding site-independent transcriptional upregulation of APC gene expression. In our preliminary studies, an increased level of phosphorylated p53 has been detected in the TFIIH complex of HCT-116(p53 ϩ/ϩ ) cells after treatment with MNNG, 2 which suggests a role of TFIIH complex in p53dependent activation of APC gene expression. Among the general transcription factors, TFIIH is a versatile, multifunctional protein complex with transcriptional properties coupled to those of DNA repair (for review, see Ref. 60). If p53 of the TFIIH complex plays a role in DNA alkylation-induced transcriptional up-regulation of APC, then it is likely that increased expression of APC might be associated with cell cycle arrest and/or apoptosis of cancer cells. In fact, in previous studies a role for APC has been implicated in cell cycle arrest (61) and apoptosis (62). Thus, p53 is transcriptionally competent to stimulate APC gene expression by direct interaction through p53-binding sites or by indirect interaction through Sp1-binding site, E-box B site, or the TFIIH complex. Nonetheless, the collaboration between p53 and APC after DNA alkylation damage may indicate an important role for APC in the cell cycle arrest and/or apoptosis of cancer cells. Our results further suggest an alternative pathway for p53 to function as a tumor suppressor involving APC, rather than p21, which is generally associated with p53-induced G 1 arrest (42,63,64). In support of this hypothesis, our results suggest that p21 gene expression in HCT-116(p53 ϩ/ϩ ) cells is down-regulated in response to MMS, 2 indicating that the p53 3 p21 pathway might not play a significant role and suggesting the existence of an alternative pathway for MMS-induced signaling in the above cells.
Based on our present study, we suggest an alternative path- FIG. 7. Analysis of MNNG-responsive cis-element(s) of the pAPCP promoter. A describes the structure of the deletion mutant constructs of the pAPCP promoter. B shows the CAT reporter activity. HCT-116(p53 ϩ/ϩ ) cells were cotransfected with 0.25 g/ml of ␤-gal and 1.5 g/ml of different deletion mutant plasmids of pAPCP promoter as indicated. After 20 h of transfection, cells were grown in a serum-free medium for 18 h and then treated with 50 M of MNNG for additional 25 h. Cells were harvested and processed for determining the CAT reporter activity. The data were normalized to ␤-gal activity in the same experiment and are the means Ϯ S.E. of three different experiments. *, significantly different from untreated controls (p Ͻ 0.05). To study the role of p53-binding sites in the transcriptional regulation of pAPCP promoter, different oligonucleotides containing p53-binding sites (P1ϩ2, P1, or P2) were synthesized and cloned into a CAT reporter plasmid. In A, p(P1ϩ2)P, p(P1)P, and p(P2)P plasmids were co-transfected into HCT-116(p53 Ϫ/Ϫ ) cells with pCMV-p53 and ␤-gal overexpression plasmids for 48 h. In B, p(P1ϩ2)P, p(P1)P, and p(P2)P plasmids were co-transfected into HCT-116(p53 ϩ/ϩ ) cells with ␤-gal overexpression plasmid. After 20 h of transfection, cells were grown in a serum-free medium for 18 h and then treated with 50 M of MNNG for additional 25 h. Cells from A and B were harvested and processed for determining the CAT reporter activity. The data were normalized to ␤-gal activity in the same experiment and are the means Ϯ S.E. of three different experiments. *, significantly different from untreated controls (p Ͻ 0.05). way for p53 function involving APC. The wild-type APC downregulates the level of ␤-catenin, which controls the expression of ␤-catenin/Tcf-Lef-dependent cell cycle-related genes c-myc and cyclin D1 (24,25). At least in the cell lines, where ␤-catenin level is regulated by APC, the increase in the p53-dependent transcription of APC gene in response to DNA alkylating agents might block ␤-catenin/Tcf-Lef complex signaling and inhibit c-myc and/or cyclin D1 gene expression. The abnormal expression of c-myc and cyclin D1 has been shown in neoplastic transformation of cancer cells (24,25). In a recent study, we have reported an increased level of APC mRNA in the lung cancer cell lines with the decreased level of c-myc mRNA, which further establishes an inverse relationship between APC and c-Myc (65). Thus, the increased levels of APC and decreased levels of ␤-catenin might ultimately result in decreased levels of c-Myc and cyclin D1, resulting in cell cycle arrest and/or apoptosis of cancer cells. In summary, these studies suggest a possible link of p53 3 APC 3 ␤-catenin/Tcf-Lef pathway in DNA alkylation-induced cell cycle arrest and/or apoptosis of colon cancer or other cancer cells involving dysregulation of APC.