Induction of p53-dependent Activation of the Human Proliferating Cell Nuclear Antigen Gene in Chromatin by Ionizing Radiation*

A human fibroblast cell line with conditional p53 expression displayed a p53-dependent increase in both the protein and mRNA levels of proliferating cell nuclear antigen (PCNA) after exposure to ionizing radiation (IR). The combination of p53 induction and IR cooperated to activate a transiently expressed human PCNA promoter-reporter gene via a p53-responsive element. Chromatin immunoprecipitation assays with antibodies specific for p53 or p300/CREB-binding protein revealed specific p53-dependent enrichment of PCNA promoter sequences in immunoprecipitates of sheared chromatin prepared from irradiated cells. Maximal and specific association of acetylated histone H4 with the PCNA promoter also depended on p53 induction and exposure to IR. These data demonstrate p53 binding to a target site in the PCNA promoter, recruitment of p300/CREB-binding protein, and localized acetylation of histone H4 in an IR-dependent manner. These molecular events are likely to play a role in mediating activation of PCNA gene expression by p53 during the cellular response to DNA damage. The analyses indicate that the combination of p53 induction and IR activate the PCNA gene via mechanisms similar to that of p21/wild-type p53-activated factor but to a lesser extent. This differential regulation of PCNA and p21/wild-type p53-activated factor may establish the proper ratio of the two proteins to coordinate DNA repair with cell cycle arrest.

Cellular responses to DNA damage bear upon tumorigenesis and the development of anti-cancer therapies. The p53 tumor suppressor protein integrates many of the cellular responses to DNA damage (for recent reviews, see Refs. 1 and 2). p53 possesses sequence-specific DNA binding activity and primarily functions as a transcription factor that binds to degenerate response elements in a variety of target genes, including p21/ WAF1, 1 mouse double minute-2, and Bax (3). Activation of p53 target genes leads to either growth arrest and subsequent DNA repair or apoptosis depending on a number of variables including the extent of DNA damage (4). DNA-bound p53 recruits transcriptional co-activators that render localized remodeling of the chromatin structure to favor promoter access by the transcriptional machinery (5)(6)(7). To date, the best characterized p53 transcriptional co-activator is p300/CBP, which possesses intrinsic histone acetyltransferase activity and interacts with other proteins possessing histone acetyltransferase activity (8,9). C-terminal acetylation of p53 by p300/CBP along with phosphorylation in the N terminus converts inert p53 to a transcriptionally active form (10,11). p53-mediated transcriptional activation of the p21/WAF1 promoter correlates with p53 binding and p300/CBP-mediated acetylation of promoter-associated histones (12)(13)(14). The essential role of the TRRAP transcriptional co-activator in potentiating activation of the mouse double minute-2 promoter by p53 suggests that p53-associated transcriptional co-activators possess promoter selectivity (15).
The PCNA gene encodes a highly conserved protein that performs essential functions in DNA replication and DNA repair (16,17). PCNA forms a trimer that encircles DNA and interacts with a number of proteins involved in DNA metabolism (18) and regulators of both the cell cycle (19,20) and cell viability (21). Inhibition of PCNA expression in murine fibroblasts arrests cell growth (22), and PCNA overexpression in yeast blocks cell cycle progression (23) and in mammalian cells inhibits DNA repair (24,25). Growth factors (26 -28) and viral infection (29,30) stimulate PCNA synthesis, and PCNA protein and mRNA levels fluctuate with the cell cycle (31). Both transcriptional (32,33) and post-transcriptional (27,34) mechanisms can account for changes in cellular PCNA levels.
The PCNA gene displays complex regulation by p53. In cells exposed to genotoxic stress, p53 expression can be correlated both directly (35)(36)(37)(38) and inversely (39) with levels of PCNA. These seemingly conflicting results are supported by findings with transient expression assays, in which co-expression of p53 produces a variable response from a reporter construct fused downstream of the human PCNA promoter (see "Discussion"). Since assays of p53 function in transiently transfected cells may not correlate with p53-mediated regulation of a gene in native chromatin (40), more direct experiments are required to demonstrate regulation of the PCNA gene by p53. We showed previously that exposure of a rat fibroblast cell line to ionizing radiation (IR) increased endogenous PCNA mRNA levels and altered p53 binding to human PCNA promoter sequences in vitro (36). Moreover, induction of a transiently expressed human PCNA promoter-reporter construct by IR required an in-tact p53-binding site (36). To address mechanisms of PCNA promoter regulation in native chromatin by p53 and/or IR, we describe here results obtained with a human fibroblast cell line, TR9-7 cells, in which expression of wild-type p53 could be induced by withdrawal of tetracycline (41). This experimental model was used to establish regulation of the PCNA gene by p53 in vivo and, in conjunction with chromatin immunoprecipitation and real time PCR, to quantify association of p53, p300/ CBP, and acetylated histones with the PCNA promoter. By these parameters, p53-mediated activation of the PCNA gene is enhanced in irradiated cells and qualitatively similar to that of the p21/WAF1 gene, an established transcriptional target of p53.

EXPERIMENTAL PROCEDURES
Plasmids-The PCNA-CAT reporter constructs contained human PCNA promoter sequences Ϫ249 to ϩ62 or Ϫ213 to ϩ62 relative to the transcriptional initiation site (ϩ1) fused upstream of the reporter sequences in pBACAT as previously described (42). Plasmid pC53-SN3 expresses human wild type p53 from the cytomegalovirus immediate early promoter (43). pON260 expresses ␤-galactosidase from the cytomegalovirus immediate early promoter (44).
Cell Culture and Irradiation of Cells-TR9-7 cells (41) are a human fibroblast cell line that expresses human wild type p53 from a tetracycline-repressible promoter in a p53 null background. Cells were grown in Dulbecco's modified Eagle's medium (Sigma) with 10% FBS (v/v), 100 g/ml penicillin, 100 g/ml streptomycin, 50 g/ml hygromycin, 600 g/ml Geneticin (Invitrogen), and 1 g/ml tetracycline (Sigma) in a humidified incubator with 10% CO 2 . Lower concentrations of tetracycline in the medium (from 1 g/ml to 0 g/ml, final concentration) were used to induce p53 expression when needed. Cells were exposed to IR at the indicated doses at a rate of 1.25 Grays/min from a cesium-137 source in a Gammacell 40 low dose rate research irradiator (MDS Nordian, Inc., Kanata, Canada).
Antibodies-A sheep polyclonal antibody to p53 (Ab-7) and rabbit polyclonal antibody to p21/WAF1 (Ab-5) were purchased from Oncogene Research (Boston, MA). A mouse monoclonal antibody to PCNA (19F4) was provided by E. M. Tan (Scripps Clinic and Research Foundation, San Diego, CA). Mouse monoclonal antibodies to p53 (PAb1801 and PAb421) and a goat polyclonal to actin (I-19) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). A rabbit polyclonal antibody to an N-terminal epitope of p300 (N-15) was purchased from Santa Cruz Biotechnology. A rabbit polyclonal antibody to acetylated histone H4 was purchased from Upstate Biotechnology, Inc. (Charlottesville, VA). Control mouse IgG and normal rabbit serum were obtained from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA).
Transient Transfection Assays-The PCNA-CAT reporters were transfected into TR9-7 cells by the calcium phosphate method as previously described (36). Each transfection mixture contained 5 g of the reporter plasmid. After transfection, the TR9-7 cells were grown in Dulbecco's modified Eagle's medium with 10% FBS and 1 g/ml tetracycline for 18 h before transfer to fresh medium with different tetracycline concentrations (1, 0.1, and 0 g/ml). Immediately after changing the medium, the transfected cells were mock-exposed or exposed to the indicated doses of IR. In all transfection experiments, the cells were harvested at 48 h after transfection, and CAT activity was determined as previously described (33). pON-260 (1 g) was co-transfected as an internal control to monitor transfection efficiency. Assays of ␤-galactosidase activity were performed as described previously (45).
Western Blot Analysis-Subconfluent TR9-7 cells were kept in medium containing 0.5% FBS for 48 h before they were put into serum-rich medium (10% FBS) with different tetracycline concentrations (1, 0.1, and 0 g/ml) to induce p53 expression or not. Immediately after changing the medium, the cells were either exposed to 20-gray ionizing radiation or mock-irradiated as depicted in Fig. 1A. The cells were harvested at 0, 8, 24, and 32 h postirradiation, and samples for immunoblotting were prepared by resuspension of the cell pellets in 150 l of Laemmli buffer followed by boiling for 5 min. Equal amounts of protein from each sample (46) were fractionated in a 12% SDS-polyacrylamide gel, which was then transferred to a nitrocellulose membrane. The blots were blocked in Tween 20/Tris-buffered saline (100 mM Tris, pH 7.5, 0.9% NaCl, 0.1% Tween 20) containing 5% nonfat dry milk for 1 h at room temperature. Immunoblots for p53, PCNA, and p21/WAF1 were performed as previously described with the antibodies Ab-7, 19F4, and Ab-5, respectively (36).
Northern Blot Analysis-TR9-7 cells underwent the treatments depicted in Fig. 1A. Total cellular RNA was isolated with Ultraspec RNA (Biotecx, Houston, TX) following the instructions provided. Northern blots were performed essentially the same as described previously (36). Each RNA sample of 20 g was separated in a 1% agarose-formaldehyde gel, which was then blotted to an Immobilon-Nyϩ membrane (Millipore Corp., Bedford, MA). The RNA blot was probed by hybridization with gel-purified fragments of 1.3-kb human PCNA cDNA, 1.0-kb human p21/WAF1 cDNA, and 1.0-kb mouse 36B4 cDNA, radiolabeled by random priming as described previously (47). Hybridization was carried out at 68°C overnight for all three labeled probes in hybridization buffer (5ϫ SSPE: 0.75 M NaCl, 50 mM NaH 2 PO 4 , and 6 mM EDTA; 2% SDS; 5ϫ Denhardt's solution; 100 g/ml single-stranded DNA). Washing conditions were twice for 15 min at room temperature with 2ϫ SSPE, 0.1% SDS, followed by once at 68°C for 20 min and then two washes in 0.2ϫ SSPE, 0.1% SDS at 68°C for 20 min each. After washing, the blot was wrapped in plastic wrap and exposed to Biomax x-ray film (Eastman Kodak Co.) with an intensifying screen at Ϫ70°C. The intensity of each mRNA signal was quantified by phosphor imager analyses (Fuji Photo Film Co., Kanagawa, Japan).
Chromatin Immunoprecipitation Assays and Real Time PCR Analyses-TR9-7 cells were grown in Dulbecco's modified Eagle's medium containing 0.5% FBS and 1 g/ml tetracycline for 48 h before changing the medium to Dulbecco's modified Eagle's medium with 10% FBS with or without tetracycline. Immediately after changing the medium, the cells were exposed to 20-gray IR. Sixteen hours after the various treatments ( Fig. 1A), the cells were fixed in 1% formaldehyde and processed for chromatin immunoprecipitation as described elsewhere (48). Monoclonal antibodies specific to p53, PAb1801, and PAb421 (1 g); a polyclonal antibody to p300 (1 g); or 5 l of polyclonal rabbit serum to acetylated histone H4 were used for each ChIP assay, respectively. Equal amounts of mouse IgG or normal rabbit serum were used as controls. Recovery of p53 in the ChIP assays was monitored by immmunoblotting the immunoprecipitated chromatin (see above) and probing the blot with a sheep polyclonal antibody to p53. The amount of DNA in each chromatin immunoprecipitate was quantified by real time PCR on a Bio-Rad iCycler using Sybr Green PCR Core Reagents (Applied Biosystems, Foster City, CA). The sequences of the primers for the PCNAϩ1 amplicon (Fig. 4A) are as follows: upper primer, acacgattggcccta aagtcttccc; lower primer, gaccactctgctacgcctgcaaccg. The sequences of the primers for the PCNA-p53 amplicon (Fig. 4A) are as follows: upper primer, ccaccataaagctggggcttgacga; lower primer, gcgaggcccgccccctagagcatac. The sequences of the primer for the p21ϩ1 amplicon ( Fig. 4A) are as follows: upper primer, ttgtatatcagggccgcgct; lower primer, cgaatccgcgcccagctc. The sequences of the primer for the p21-p53 amplicon ( Fig. 4A) are as follows: upper primer, tccacctttcaccattcccctaccc; lower primer, cagcccaaggacaaaatagccacca. Reaction volumes were 50 l with primer concentrations at 15 M. Three-step PCR was performed for each amplicon with annealing temperatures as follows: 58°C for PCNA-p53, PCNAϩ1, and p21-53 amplicons; 60°C for p21ϩ1 amplicon. Temperatures for denaturing and extension were 94°C and 72°C, respectively. Each step lasted 30 s. A standard curve was established for each amplicon by inputting 4-or 6-fold serial dilutions of genomic DNA recovered from sheared chromatin with a starting input of 840 ng of DNA. The threshold cycle of each experimental sample was converted to the amount of immunoprecipitated DNA by comparison with the standard curve. The portion of DNA immunoprecipitated from each sample was determined after quantifying the corresponding input by real time PCR. The values from the group of serum-starved cells without p53 induced or exposure to IR, designated as 0 h, were set arbitrarily to 1. A -fold change of each ChIP assay was obtained by comparing the remaining groups to that of 0 h. Acid Extraction of Histone and Immunoblot for Acetylated Histone H4 -TR9-7 cells were cultured and treated as described for ChIP assays. To establish a positive control for histone H4 acetylation, asynchronous TR9-7 cells were treated with 5 mM sodium butyrate for 24 h. Cells were scraped from the culture dishes and collected by centrifugation at 200 ϫ g for 10 min. The cell pellets were resuspended in 5 packed cell volumes of lysis buffer (10 mM HEPES, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM dithiothreitol, 1.5 mM phenylmethylsulfonyl fluoride). Sulfuric acid was added to a final concentration of 0.2 M. The suspension was then incubated on ice for 30 min followed by centrifugation at 11,000 ϫ g for 10 min at 4°C. The supernatant was transferred to a fresh dialysis tube and dialyzed against 200 ml of 0.1 M acetic acid, twice for 1.5 h each. The supernatant was then dialyzed three more times against 200 ml of H 2 O for 1 h, 3 h, and overnight, respectively. The extracted histone was quantified and immunoblotted with an acetylated histone H4-specific antibody following the instructions of the provider (Upstate Biotechnologies).

Cooperative Induction of PCNA Expression by p53 and Ion-
izing Radiation-TR9-7 cells, which express wild-type p53 from a tetracycline-repressible promoter (41), were used as an experimental model to provide a means of independently modulating p53 expression level and activating post-translational modifications of p53 by IR. The independent and cooperative roles of p53 induction and irradiation upon regulation of two important cell cycle regulatory proteins, p21/WAF1 and PCNA, were examined in these cells. The TR9-7 cells were incubated in low serum (0.5%) and 1 g/ml tetracycline (no p53 induction) for 48 h (Fig. 1A). Then the cells were changed into fresh medium with serum (10%) containing 0, 0.1, or 1 g/ml tetracycline to induce different levels of p53 expression. Immunoblotting assays postinduction revealed a 4 -7-fold increase in p53 levels by 8 h in cells grown in the absence of tetracycline that was maintained through the 32-h time point (Fig. 1B). Immediately after changing the media, some cells were exposed to 20 grays of ␥-radiation. The effects of p53 expression and/or IR upon cellular levels of p21/WAF1 and PCNA were determined by immunoblotting total cell extracts from the untreated cells (0 h) and at 8, 24, and 32 h after the various treatments. In unirradiated cells without p53 induction, p21/WAF1 levels changed very little (Fig. 1C). Full p53 induction in unirradiated TR9-7 cells caused a gradual increase in levels of p21/WAF1, reaching a maximum at 24 h. Although exposure of the cells to IR alone produced an increase in p21/WAF1 levels (Fig. 1C, lanes 1, 2, 5, and 8), the combination of IR and p53 induction produced a larger increase (lane 10 versus lane 8). These data confirm p53-mediated regulation of a target gene in TR9-7 cells as shown previously (41) and indicate that IR and p53 cooperate to elevate cellular p21/WAF1 levels.
The effects of IR and p53 upon PCNA levels can be compared with the data for p21/WAF1 described above. In the absence of p53 induction, the levels of PCNA in unirradiated TR9-7 cells appeared higher at 24 h (Fig. 1C, lane 5), which probably reflects serum-stimulated entry into S phase. At the 24-and 32-h time points with p53 induced minus IR, PCNA levels in the serum-stimulated cells (lanes 7 and 10, respectively) appeared similar to that of serum-starved cells (lane 1). Independent of p53 induction, exposure to IR did not prevent the serum-induced increase in PCNA levels (lanes 5 and 8). In contrast to the effects of p53 induction in unirradiated cells, the combination of IR and p53 induction elevated PCNA levels at the 24-and 32-h time points (lanes 7 and 10). The augmented PCNA expression coincided with the increase in p21/WAF1 levels in the same cells. The increase of both p21/WAF1 and PCNA protein levels by p53 induction and IR is consistent with their physiological roles in the cellular response to DNA damage, participation in cell cycle arrest, and DNA repair. p53associated elevation of PCNA protein levels at 24 and 32 h required exposure to IR and appeared to be limited relative to that of p21/WAF1.
A Northern blot was performed to determine the p21/WAF1 and PCNA mRNA levels in TR9-7 cells subjected to the above experimental protocol (see Fig. 1A). Exposure to IR or p53 induction elevated p21/WAF1 mRNA levels at 16 h over that of the serum-starved control cells ϳ2and 6-fold, respectively (Fig. 2). Similar to the immunoblotting results, the combination of p53 induction and exposure to IR produced a maximal increase in p21/WAF1 mRNA levels, more than 14-fold. The The cells were lysed directly in Laemmli buffer, and equal amounts of protein from each sample were analyzed by immunoblotting with a p53-specific polyclonal antibody. C, p21/ WAF1 and PCNA protein levels in irradiated (IRϩ) and unirradiated (IRϪ) TR9-7 cells after p53 induction. Total cellular proteins were isolated at the indicated times. Protein from each sample (50 g; loaded on the gel as in B) was fractionated in a 12% polyacrylamide gel, followed by immunoblotting for p21/WAF1 (top two panels) or PCNA (middle two panels). An immunoblot for actin in each sample (bottom two panels) revealed comparable protein amounts in each lane. Densitometry was performed with correction to the actin loading control.
FIG. 2. The effect of p53 and/or IR upon PCNA and p21/WAF1 mRNA levels. Total cellular RNA was isolated from TR9-7 cells 16 h after serum stimulation with or without p53 induction and/or with or without exposure to 20 grays of ionizing radiation (see Fig. 1A). Equal amounts of RNA (20 g) from each preparation were fractionated in a formaldehyde-agarose gel that was subsequently transferred to a polyvinylidene difluoride membrane. The blot was probed by hybridization to radioactive cDNA probes specific for PCNA, p21/WAF1 and 36B4. The hybridized filter was exposed to x-ray film with an intensifying screen for 24 h. Changes in PCNA and p21 mRNA levels were quantified by phosphorimaging analyses after correction for 36B4. The 0-h sample (lane 1) was prepared from TR9-7 cells after 48 h in 0.5% serum. The remaining lanes display RNA from cells 16 h after the addition of 10% serum with or without p53 induction and/or IR. addition of serum alone did not affect p21/WAF1 mRNA levels.
Some similarities and differences in the regulation of p21/ WAF1 and PCNA mRNA levels were observed. Consistent with a serum response (26), the addition of serum alone elevated PCNA mRNA levels 2-fold over that of the serum-starved cells (Fig. 2). In the serum-stimulated cells, IR exposure or p53 induction elevated PCNA mRNA levels about 3-and 5-fold, respectively. Similar to p21/WAF1, the combination of p53 induction and IR maximally elevated PCNA mRNA levels, ϳ8-fold. With the exception of p53 induction alone, which did not increase PCNA protein levels but did increase PCNA mRNA levels, the Northern and immunoblotting assays revealed concordant regulation of the mRNAs and their respective proteins for p21/WAF1 and PCNA. Post-transcriptional regulation of PCNA expression, which may account for the disparity in PCNA mRNA and protein levels, has been previously reported (27,34). These data indicate that the combination of p53 and IR induces maximal expression of both the p21/WAF1 and the PCNA genes in serum-stimulated cells. The amplitude of induction of PCNA mRNA by the combination of p53 and IR is smaller relative to that of p21/WAF1 mRNA.
Cooperative Activation of the PCNA Promoter by p53 and Ionizing Radiation-We performed transient transfection assays with PCNA-CAT reporter constructs to investigate whether or not the cooperative roles of p53 and IR in induction of PCNA expression could be manifest at the transcriptional level (Fig. 3). Neither p53 expression nor IR independently had a major effect (less than 50% change) on expression from PCNA-CAT. Consistent with PCNA protein and mRNA expression, the combination of p53 induction and IR produced a 3-fold increase in PCNA promoter-directed CAT expression by the construct with an intact p53-binding site (Ϫ249 PCNA-CAT). The construct lacking the p53-binding site (Ϫ213 PCNA-CAT) failed to respond to the combination of p53 induction and irra-diation. These observations show that the cooperative effects of p53 and IR on induction of PCNA expression occur, at least in part, at the transcriptional level.
p53 Binding and Recruitment of p300 to the p21/WAF1 and PCNA Genes in Irradiated TR9-7 Cells-Previous gel shift assays demonstrated that complexes containing p53 and p300 formed with oligonucleotides corresponding to the p53-binding site in the PCNA promoter (36). To examine the recruitment of p300 by p53 to the PCNA gene in vivo, we performed ChIP assays using antibodies specific for p53 and p300. Partially cross-linked, sheared chromatin was prepared from TR9-7 cells released from low serum in the presence or absence of p53 induction with or without exposure to IR, as described above (Fig. 1A). Real time PCR assays of the amplicons diagrammed in Fig. 4A were employed to quantify the amount of the chromatin specifically immunoprecipitated with antibodies to p53 or p300. Restriction digestion (not shown) and melting curve analyses verified the identities of the p21/WAF1 (Fig. 4B) and PCNA (Fig. 4C) amplicons, thereby validating quantification of the PCR product by Sybr Green fluorescence. To quantify p53 binding, real time PCR amplification curves were established for the p21-p53 amplicon with the input sheared chromatin preparation and the material that was specifically immunoprecipitated with antibodies to p53 (Fig. 5A). Comparison of the threshold amplification value for each experimental sample to the threshold values for a titration of the input chromatin (data not shown) established the relative abundance of the immunoprecipitated material. The addition of serum or the combination of serum and IR had little effect on p53 binding to the p21/WAF gene at the 16-h time point (Fig. 5B). Only a slight increase, 1.6-fold, in p53 binding to the p21/WAF1 gene was observed upon p53 induction alone. The combination of p53 induction and exposure to IR dramatically enhanced p53 binding to the p21/WAF1 gene about 10-fold over that observed in cells without p53 induced.
Similar to the analyses of p21/WAF1, amplification curves were established for the PCNA-p53 amplicon with the input chromatin and the material immunoprecipitated with antibodies to p53 (Fig. 5C). A comparable, yet distinct, profile of p53 binding to the PCNA gene was observed in the ChIP assay. Exposure to IR or p53 induction produced a measurable increase in p53 binding to the PCNA gene, approximately a 2-fold increase over the control value for each one (Fig. 5D). As described for the p21/WAF1 gene, the combination of IR and p53 induction produced maximal p53 binding to the PCNA gene, about a 6-fold increase over the control value. Two control experiments indicated the specificity of the ChIP assays. First, ChIP assays of each sheared chromatin preparation with nonspecific mouse IgG produced threshold values comparable with those with the 0-h control sample and the anti-p53 antibody (Fig. 5, A and C). Second, threshold values of p53-specific ChIP assays for an amplicon corresponding to the ϩ1 site of the human ␤-globin promoter (no p53-binding site) and chromatin from the IR-p53 combination were above 32 cycles, similar to that of the mouse IgG ChIP results (data not shown). Taken together, these findings indicate that IR activates p53 binding to both the p21/WAF1 and PCNA genes with a magnitude of p53 recruitment to the p21/WAF gene greater than that for the PCNA gene.
The increase in p53 binding to both the PCNA and the p21 promoters in irradiated cells described above could be a consequence of the effects of IR upon the level of p53 protein or enhancement of the protein's DNA-binding activity. Exposure of cells to IR induces a number of post-translational modifications of p53, including phosphorylation and acetylation that both stabilize the protein and convert it to a transcriptionally active form (50). Immunoblots of total cell extracts from unirradiated and irradiated cells with p53 induced did not reveal detectable differences in p53 protein levels (Fig. 6A). Apparently, the potent p53 induction in TR9-7 cells overcame the protein's inherent instability mediated by human double minute-2 (51). Enhanced p53 binding to the p21/WAF1 and PCNA promoters in irradiated cells may be selective or a consequence of a general increase in DNA binding affinity. Immunoblots of protein recovered from chromatin immunoprecipitated with anti-p53 antibody from each experimental group did not show a general radiation-associated increase in p53 binding to chromatin (Fig. 6B). This finding indicates selective p53 binding to the p21/WAF1 and PCNA genes in irradiated cells.
Several reports have demonstrated that the p300/CBP potentiates transcriptional activation of p53-responsive promoters (5)(6)(7). ChIP assays were used to evaluate recruitment of p300 to the p53-binding sites in the PCNA and p21/WAF1 genes. The analyses were identical to the p53 binding assays described above with the exception that the specific antibody recognized p300. As described above, amplification curves for the p21-p53 amplicon were established, and comparison of threshold values to a standard curve of input chromatin served to quantify the amount of DNA immunoprecipitated with anti-p300 antibodies. Induction of p53 alone brought about a 4-fold increase in p300 binding to the p21/WAF1 gene (Fig. 7A). The combination of p53 induction and irradiation elicited the most p300 binding to the p21/WAF1 gene, about 7-fold higher than the 0-h control (Fig. 7A). ChIP analyses of p300 binding to the PCNA-p53 amplicon revealed similar results. p53 induction or the combination of p53 induction with irradiation increased the association of p300 with the PCNA promoter about 1.8-and 3-fold, respectively (Fig. 7B). Thus, the p53 and p300 ChIP results indicated that IR enhanced both p53 binding and p300 association to both the p21/WAF1 and the PCNA genes.
p53-dependent Acetylation of Promoter-associated Histone H4 in Irradiated TR9-7 Cells-The intrinsic histone acetyltransferase activity of p300 can promote chromatin decondensation via acetylation of core histones, which improves accessibility of the remodeled chromatin to the transcriptional machinery (8). Using antibodies specific for acetylated histones in the ChIP assay can evaluate the extent of histone acetylation (52). To this end, ChIP assays were performed on the sheared chromatin preparations described above with an antibody specific for acetylated histone H4. Real time PCR assays with primers flanking the transcription initiation sites of both the p21/WAF1 and the PCNA genes (p21ϩ1 and PCNAϩ1; Fig. 4A) were used to quantify the amount of DNA co-immunoprecipitated with the antibody to acetylated histone H4. As described above, the real time PCR products were identified by restriction analyses and melting curves (not shown). Amplification curves of chromatin immunoprecipitated with an antibody to acetylated histone H4 were established for the p21/WAF1ϩ1 and the PCNAϩ1 amplicons (data not shown). Again, real time PCR amplification of serial dilutions of sheared chromatin established a reference curve to quantify threshold values for both proximal promoter amplicons (data not shown). For p21ϩ1, p53 induction alone caused a moderate 2.5-fold increase of histone H4 acetylation over that of the 0-h control, whereas neither serum stimulation nor irradiation alone produced a change in H4 acetylation (Fig. 8A). Similar to the patterns of p53 binding and recruitment of p300, the combination of p53 induction and irradiation maximally increased acetylation of histone H4 associated with the p21/WAF1 proximal promoter, a 10-fold elevation over the 0-h control (Fig. 8A). p53 induction alone moderately augmented acetylation of histone H4 associated with the PCNA gene, a 2.6-fold increase over that of the 0-h control (Fig. 8B). Acetylation of PCNA promoter-associated histone H4 climbed 7.5-fold in the irradiated cells with p53 induced. Although there were differences in magnitude, p53 induction correlated with acetylation of promoter-associated histone H4 on both the p21/WAF1 and the PCNA promoters. In addition to this similarity, serum stimulation induced a 2-fold increase in promoter-associated histone H4 acetylation relative to that of the serum-starved 0-h control (Fig. 8B) that did not correlate with p300 recruitment to the p53-binding site (see Fig. 7B). Exposure of the cells to IR abrogated this modest serum-stimulated elevation in PCNA promoter-associated histone H4 acetylation (Fig. 8B). Low background levels in control experiments with normal rabbit serum suggested a basal level of H4 acetylation on both the p21/WAF1 and the PCNA-proximal promoters in the serumstarved cells. ChIP assays with the ␤-globin proximal promoter did not show correlations between acetylation of histone H4 and p53 induction and/or exposure to IR (data not shown). Immunoblotting of histones from cells in each experimental group demonstrated that association of acetylated histone H4 with the p21/WAF1 and PCNA promoters was not due to a general increase in the abundance of acetylated histone H4 (Fig. 8C). These data demonstrated maximal acetylation of histone H4 associated with both the p21/WAF1-and PCNAproximal promoters in irradiated cells with p53 induced.

DISCUSSION
The data shown here demonstrate that exposure of human fibroblasts to IR increases both PCNA and p21/WAF1 mRNA levels in a p53-dependent manner. Moderate increases of both p21/WAF1 and PCNA mRNA expression by either p53 induction or IR alone relative to the two in combination indicate cooperative effects of IR and p53 upon expression of the two target genes. Both transcriptional and post-transcriptional mechanisms account for regulation of PCNA expression (27,30), but the effects of p53 are primarily transcriptional. Regulation of the PCNA promoter by p53 appears to vary according to the assay system. The model used most frequently, transient co-expression of wild-type p53 with a PCNA promoter-reporter construct in p53 null cells, has demonstrated p53-mediated repression (53-55), no effect (56), or a variable concentrationdependent response (lower levels of p53 activate the PCNA promoter, whereas higher levels repress it) (42,57). In the experiments described here, the combination of IR and p53 induction activates a transiently expressed PCNA-CAT reimmunoprecipitated chromatin with a standard titration curve established the relative amount of DNA in each sample. The results were corrected for the amount of input chromatin (established by real time PCR) in each immunoprecipitation reaction. The graph shows the -fold change in p53 binding to p21/WAF1 chromatin for the various treatments (with or without 10% serum, with or without p53, and with or without IR) relative to the 0-h control sample (ϪFBS, Ϫp53, ϪIR), which was arbitrarily set to 1. The results shown are the average Ϯ S.E. of two separate experiments assayed in duplicate by real time PCR. C, real time PCR amplification curves of the PCNA-p53 amplicon with immunoprecipitated chromatin; same as A, except the results with the PCNA-p53 amplicon (see Fig. 4A) are shown. D, p53 binding to the PCNA promoter. Same as B, except that binding of p53 to the PCNA promoter was quantified by ChIP-real time PCR with the PCNA-p53 amplicon.  Fig. 1. Sixteen hours after the various treatments (with or without p53 induction, with or without IR), sheared chromatin was prepared from the cells. A pool of p53-specific monoclonal antibodies or nonspecific mouse IgG control (mIgG) was used to immunoprecipitate each sheared chromatin preparation. After reversal of the formaldehyde cross-linking, the amount of DNA in each immunoprecipitate was quantified by real time PCR. The graph shows the real time PCR amplification curves as fluorescence intensity versus cycle number for the p21-p53 amplicon (see Fig. 4A). Similar amplification profiles were generated for amplicons in each ChIP assay shown in Figs. 5, 7, and 8. B, p53 binding to the p21/WAF1 promoter. Comparison of the threshold values from real time amplification curves of FIG. 6. p53 protein levels with and without induction in the presence and absence of IR. A, levels of p53 in total cell lysates. TR9-7 cells were treated as shown in Fig. 1. Sixteen hours after the various treatments, the cells were harvested and lysed directly in Laemmli buffer. Cells with p53 induced were grown in the absence of tetracycline. Equal amounts of protein from each sample were analyzed by immunoblotting with a p53-specific polyclonal antibody. B, amounts of chromatin-bound p53. Equal amounts of sheared chromatin used in Fig. 3 were immunoprecipitated with pooled monoclonal antibodies (ϩ) or control mouse IgG (Ϫ). The immunoprecipitated proteins were dissolved in Laemmli buffer and fractionated in a 12% polyacrylamide denaturing gel. The amount of p53 in each sample was visualized by immunoblotting the gel with a p53-specific polyclonal antibody. Lane M contains a total cell lysate.
porter via a p53-responsive element (Fig. 3). However, transient reporter assays of p53 function may not be physiologically relevant (40,58). Chromatin immunoprecipitation assays address p53-dependent regulation of the endogenous PCNA gene. The combination of IR and p53 increases p53 binding to the endogenous PCNA promoter with enhanced recruitment of p300 and acetylation of histone H4. These molecular events at the chromatin level correlate with the profiles of PCNA and p21/WAF1 mRNA expression upon p53 induction and exposure to IR. Little variation in the levels of actin protein (Fig. 1C), 36B4 mRNA (Fig. 2), and ChIP assays of the ␤-globin promoter (not shown) confirmed gene-specific responses to the combination IR and p53 induction. Activation of the PCNA gene by p53 and IR in TR9-7 cells compares favorably with that of the p21/WAF1 gene, albeit to a lesser extent. The results demonstrate p53-mediated transcriptional activation of the PCNA gene during the cellular response to DNA damage. Activation of the PCNA gene by p53 agrees with previous findings that the two proteins are co-expressed in cells exposed to genotoxic stress (35,37,38,59,60). FIG. 7. Recruitment of p300 to the p21/WAF1 and PCNA promoters. A, binding of p300 to the p21 promoter; same as Fig. 4, except that association of p300 with the p21 gene was quantified by ChIP-real time PCR analyses with the p21-p53 amplicon after immunoprecipitation with a polyclonal antibody to p300. The graph shows the change in p300 recruitment associated with the indicated treatment (with or without p53 induction and with or without 20-gray IR) relative to that of untreated cells in low serum (Ϫserum, Ϫp53, ϪIR), which was arbitrarily set to 1. The results shown are the average Ϯ S.E. of two separate experiments assayed in duplicate by real time PCR. B, p300 binding to the PCNA promoter; same as in A, except that p300 recruitment to the PCNA promoter was assayed with the PCNA p53 amplicon (Fig. 3A).
FIG. 8. Acetylation of histone H4 associated with the p21/WAF1 and PCNA promoters. A, acetylation of histone H4 associated with the p21/WAF1 promoter. TR9-7 cells were treated as shown in Fig. 1. ChIP/real time PCR assays were used to assess acetylation of histone H4 associated with the p21/WAF1 promoter. The sheared chromatin was immunoprecipitated with an antibody specific for acetylated histone H4. The amount of DNA in each immunoprecipitate was quantified by real time PCR of the p21ϩ1 amplicon (see Fig. 4A). The graph shows the change in acetylated histone H4 associated with the p21/WAF1 promoter for the indicated treatment relative to the 0-h untreated control cells (Ϫserum, Ϫp53, ϪIR), which was arbitrarily set to 1. The results are the average Ϯ S.E. of two separate experiments assayed in duplicate by real time PCR. B, acetylation of histone H4 associated with the PCNA promoter; same as A, except that the amount of DNA immunoprecipitated with antibody to acetylated H4 from the sheared chromatin preparations was quantified by real time PCR with the PCNAϩ1 amplicon (see Fig. 4A). The graph shows the change in acetylation of PCNA promoter-associated histone H4 for the indicated treatment relative to the 0-h control cells (Ϫserum, Ϫp53, ϪIR), which was arbitrarily set to 1. The results are the average Ϯ S.E. from two separate experiments assayed in duplicate by real time PCR. C, alteration of histone H4 acetylation by IR and p53. TR9-7 cells were treated the same as for ChIP assays. Histones were acid-extracted, followed by immunoblotting with a specific antibody to acetylated histone H4. Equivalent loading of the samples was verified by Ponceau S staining of the transferred blot. Assays of p53 binding in vitro to short oligonucleotides support the model that p53 is expressed as a latent protein with the C-terminal domain inhibiting sequence-specific DNA binding (61). Deletion, post-translational modifications, binding of monoclonal antibody PAb421, or other perturbations of the C terminus increase the affinity of p53 binding to short oligonucleotides (61,62). Analyses of p53 binding to long DNA probes (12) or to chromatin, in vivo (14,63) or in vitro (12), have not supported the model that p53 exists in a latent form inactive for DNA binding. The results in Fig. 5 support the latent p53 model in that exposure of cells to IR enhances sequence-specific DNA binding by p53 in vivo. The IR-induced increase in sequence-specific binding by p53 was not due to an increase in protein levels or an increase in the general affinity of p53 for DNA, since immunoblotting revealed no detectable difference in either total or chromatin-bound p53 after exposure to IR (Fig. 6). Thus, expression of p53 that constitutively binds DNA at specific sites in vivo is not obligatory, and post-translational control of sequence-specific DNA binding activity to regulate p53 function can be achieved in cells. In cells exposed to genotoxic stress, those post-translational modifications that allow p53 accumulation may also activate DNA binding. In TR9-7 cells with inducible p53, protein accumulation can be achieved without activation. During embryogenesis, p53 is constitutively expressed at relatively high levels (64), and immunoblotting can readily detect p53 in rapidly proliferating mouse embryo fibroblasts grown in culture. 2 The latent p53 model provides an appealing hypothesis to explain how the rapidly proliferating embryonic cells tolerate abundant p53 expression.
p53-mediated regulation of PCNA expression displays differences in humans and rodents. Exposure of mice to IR fails to activate PCNA expression in a variety of tissues (65). Induction of the rat PCNA promoter in cells exposed to UV occurs in a p53-independent manner (66). No p53 binding site has been reported in the mouse PCNA promoter, and a site with only poor homology (50%) to the consensus p53 binding site has been identified in the rat PCNA promoter (55). In rat and mouse, p53-dependent repression, but not activation, of PCNA expression has been reported (55,66,67). In human cells exposed to genotoxic stress, p53 induces global genome repair, whereas most rodent tissues lack p53-inducible global genome repair (68). p53-mediated activation of PCNA for purposes of DNA repair in humans and the absence of this pathway in rodents appear to be consistent with the distinct responses to genotoxic stress in humans and rodents. Although rodents lack p53inducible global genome repair, the correlation between reduced PCNA expression and the enhanced radiosensitivity of "wasted mice" suggests that constitutive levels of PCNA may be critical to survival of rodent cells exposed to genotoxic stress (69,70).
The experiments described here show p53 binding in vivo to the p21/WAF1 and PCNA genes. However, binding of p53 to target sites in vivo can occur in the absence of transcriptional activation (14). Transient expression assays indicate that p53 cooperates with p300/CBP to activate the human PCNA promoter. Co-expression of the adenovirus E1A oncoprotein prevents p53-mediated activation of the PCNA promoter, and the N-terminal p300/CBP binding domain is essential for this activity of E1A (71,72). Furthermore, co-transfection of a dominant negative mutant p300 inhibits activation of the PCNA promoter by p53 (71). 3 In nuclear extracts prepared from irradiated cells, association of p300 with the PCNA p53-binding site in vitro correlates with activation of the PCNA gene by IR (36). Consistent with these findings, the data here show p53dependent association of p300 with the PCNA promoter in vivo. p300 interacts with the PCNA protein and may be involved in DNA repair (73). In keeping with this function, exposure of cells to UV increases p300 binding to chromatin. This type of p300 association with chromatin is distinct from the results described here, which show that p300 binding is dependent, in part, upon p53 induction and through a p53-responsive element. Moreover, acetylation of PCNA promoter-associated histone H4 correlates with p300 interaction with the PCNA promoter.
Assays of lymphoblastoid cell lines exposed to IR demonstrate p53-dependent post-translational regulation of the subcellular distribution of PCNA with little effect of IR upon PCNA mRNA levels (74). In contrast, similar to the findings here, irradiation of peripheral blood lymphocytes induces a 2-3-fold increase in PCNA mRNA levels by 24 h after exposure (59). One possible explanation for this discrepancy is that regulation of PCNA mRNA levels by IR differs in quiescent versus cycling cells. In accord with this postulate, IR induces a form of p53 in S phase cells that is transcriptionally inactive (75). The higher levels of PCNA in cycling cells may be sufficient for DNA repair. p53-dependent activation may become obvious in the experiments described here because of the low levels of PCNA after 48 h of serum starvation.
Although the PCNA and p21/WAF1 genes display similar p53-dependent regulation in irradiated cells, the magnitude of PCNA induction is reduced relative to that of p21/WAF1. Data from other experiments not shown here indicate that the low affinity p53 binding site in the PCNA promoter determines transient activation by p53 relative to the high affinity upstream p53 binding site in the p21/WAF1 promoter. 4 p21/ WAF1 has been shown to selectively inhibit PCNA-dependent DNA replication but not PCNA-dependent DNA repair (76). Since p21/WAF1 inhibits DNA replication through direct PCNA binding (77), the relative ratios of the two proteins are important determinants of cell cycle progression. Overexpression of PCNA induced by HTLV-1 Tax protein inhibits DNA repair, which can be rescued by overexpression of p21/WAF1, but not a p21/WAF1 mutant that fails to bind PCNA (24,25). The relatively reduced activation of the PCNA promoter by p53 probably serves to maintain the proper ratio of the two proteins to effect DNA repair and prevent cell cycle progression. Like p53-dependent activation of p53R2 (49), a ribonucleotide reductase subunit, activation of PCNA expression by p53 contributes to repair of DNA damage. The differential regulation of two DNA damage response effectors, p21/WAF1 and PCNA, by p53 may play a role in ensuring a prompt and appropriate growth arrest and DNA repair that is critical for maintenance of genomic stability in cells exposed to genotoxic stress.