Constitutive Expression of the Cyclin-dependent Kinase Inhibitor p21 Is Transcriptionally Regulated by the Tumor Suppressor Protein p53*

The tumor suppressor protein p53 has been implicated in the response of cells to DNA damage. Studies to date have demonstrated a role for p53 in the transcriptional activation of target genes in the cellular response to DNA damage that results in either growth arrest or apoptosis. In contrast, here is demonstrated a role for p53 in regulating the basal level of expression of the cyclin-dependent kinase inhibitor p21 in the absence of treatment with DNA-damaging agents. Wild-type p53-expressing MCF10F cells had detectable levels of p21 mRNA and protein, whereas the p53-negative Saos-2 cells did not. Saos-2 cells were infected with recombinant retrovirus to establish a proliferating pool of cells with a comparable constitutive level of expression of wild-type p53 protein to that seen in untreated MCF10F cells. Restoration of wild-type but not mutant p53 expression recovered a basal level of expression of p21 in these cells. Constitutive expression of luciferase reporter constructs containing the p21 promoter was inhibited by co-transfection with the human MDM2 protein or a dominant-negative p53 protein and was dependent on the presence of p53 response elements in the reporter constructs. Furthermore, p53 in nuclear extracts of untreated cells was capable of binding to DNA in a sequence-specific manner. These results implicate a role for p53 in regulating constitutive levels of expression of p21 and demonstrate that the p53 protein is capable of sequence-specific DNA binding and transcriptional activation in untreated, proliferating cells.

The tumor suppressor protein p53 has been implicated in the response of cells to DNA damage. Studies to date have demonstrated a role for p53 in the transcriptional activation of target genes in the cellular response to DNA damage that results in either growth arrest or apoptosis. In contrast, here is demonstrated a role for p53 in regulating the basal level of expression of the cyclin-dependent kinase inhibitor p21 in the absence of treatment with DNA-damaging agents. Wild-type p53expressing MCF10F cells had detectable levels of p21 mRNA and protein, whereas the p53-negative Saos-2 cells did not. Saos-2 cells were infected with recombinant retrovirus to establish a proliferating pool of cells with a comparable constitutive level of expression of wild-type p53 protein to that seen in untreated MCF10F cells. Restoration of wild-type but not mutant p53 expression recovered a basal level of expression of p21 in these cells. Constitutive expression of luciferase reporter constructs containing the p21 promoter was inhibited by co-transfection with the human MDM2 protein or a dominant-negative p53 protein and was dependent on the presence of p53 response elements in the reporter constructs. Furthermore, p53 in nuclear extracts of untreated cells was capable of binding to DNA in a sequence-specific manner. These results implicate a role for p53 in regulating constitutive levels of expression of p21 and demonstrate that the p53 protein is capable of sequence-specific DNA binding and transcriptional activation in untreated, proliferating cells.
The tumor suppressor protein p53 is a transcription factor that binds to DNA in a sequence-specific manner, has been implicated in the cellular response to DNA damage, and appears to play a role in a variety of cellular responses including growth arrest, apoptosis, differentiation, and senescence (1)(2)(3)(4). Studies to date have documented a role for p53 in transcriptional activation of target genes in response to extracellular stimuli including DNA damage leading to a cellular response involving either growth arrest or apoptosis. DNA-damaging agents trigger an increase in p53 expression leading to activa-tion of particular target genes most notably that of the cyclindependent kinase inhibitor, p21 (5). Consistent with this, cells that lack p21 expression have an impaired p53-dependent response to DNA damage (6,7). This transcriptional activation of p21 expression is mediated by the interaction of p53 with two response elements located in the p21 promoter (8).
The DNA binding activity of p53 appears to be regulated by the terminal 30 amino acids of the protein. Phosphorylation by either casein kinase II or protein kinase C, acetylation by p300, and binding by a monoclonal antibody 421, or the bacterial dnaK protein all occur within this region of p53 and will activate the ability of p53 to bind to DNA in a sequence-specific manner in vitro (9 -15). There have been several reports that the ability of p53 in nuclear extracts to bind to DNA requires the presence of antibody 421, leading to the notion that p53 exists in a latent form prior to DNA damage (10,12). Consistent with this idea, microinjection of the antibody 421 into cells activates p53-dependent expression from reporter constructs (13,16). Thus, it has been proposed that in untreated cells, the p53 protein exists in a latent state that is unable to bind to DNA and that the ability of p53 to activate target gene expression is not merely dependent on the increase in protein level but also requires post-translational modification of p53 to convert this latent form into a form that is active for DNA binding (12,17). This notion is supported by studies demonstrating that p53 becomes phosphorylated at particular sites after treatment of cells with DNA-damaging agents (18,19).
Prior to the cloning of the gene, it was noted that p21 was absent from cyclin/cyclin-dependent kinase complexes in cells lacking functional p53 (20). Other studies have noted that the level of p21 mRNA was much lower in fibroblasts and keratinocytes derived from mice containing a homozygous deletion of p53 as compared with the corresponding cells from mice expressing wild-type p53 (21)(22)(23)(24). This suggests that p53 may play a role in the level of p21 expression in untreated, proliferating cells. The experiments presented here tested this idea directly and demonstrate that constitutive expression of the p21 protein in untreated cells is, indeed, dependent on p53 and thus implicate a role for p53 not only in the increased expression of p21 in response to DNA damage leading to either growth arrest or apoptosis but also in the basal level of expression of p21 in normally proliferating cells. vector pGL2 (Promega). The plasmid p21D2.1 has 2.1 kb at the 5Ј end of the promoter sequence removed and lacks the two p53 response elements of the p21 promoter (25). The plasmid pRL-SV40 contains the SV40 enhancer and early promoter upstream of a Renilla luciferase reporter gene (Promega). The plasmid pCMV-hdm2 encodes the human MDM2 protein under control of the cytomegalovirus (CMV) promoter and the plasmid pCMV-p53Ala-143 encodes the tumor-derived mutant human p53 protein containing a missense mutation of valine to alanine at residue 143 (26).
Antibodies and Cells-Saos-2 and WI38 cells were obtained from the American Type Culture Collection and were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS). MCF7 cells were maintained in RPMI medium containing 10% heat-inactivated FBS and 5 g/ml insulin. MCF10F cells were grown in 50% DMEM and 50% Ham's F12 medium containing 5% horse serum, 20 ng/ml epidermal growth factor, 100 ng/ml cholera toxin, 10 g/ml insulin, and 500 ng/ml hydrocortisone. PA12-p53BN and PA12-p53EN are cell lines that produce recombinant retrovirus encoding human wild-type p53 or the mutant p53 His-273 , respectively (27). These cell lines were grown in 10% FBS in DMEM containing high glucose and 400 g/ml G418 sulfate. The hybridoma cell line producing the mouse monoclonal antibody 1801 was grown in DMEM containing 10% FBS. Hybridoma cell lines expressing the mouse monoclonal antibodies 421 and 419 were grown in 50% DMEM and 50% Fischer's medium containing 10% FBS. Monoclonal antibody 1801 specifically reacts with human p53 (28), 421 recognizes p53 from a variety of species, and 419 recognizes an epitope on the SV40 large T antigen (29). All cell lines were grown at 37°C in a humid incubator containing 5% CO 2 . Antibody against p21 WAF1, CIP1 was obtained commercially (Ab-1/clone EA10, Calbiochem). For treatment with ultraviolet light, the medium was removed, and the cells were exposed to ultraviolet light using a UV Stratalinker (Stratagene).
Northern Analysis-Total RNA was extracted from 5 ϫ 10 6 cells using RNAzol (Tel-test), and Northern blot analysis was performed following conventional procedures, using a 2.1-kb full-length human p21 cDNA or human glyceraldehyde-3-phosphate dehydrogenase cDNA (Ambion) as probes.
Immunoblotting-Cells were lysed in a buffer containing 0.5% sodium deoxycholate, 2% Nonidet P-40, 0.2% sodium dodecyl sulfate (SDS), 50 mM NaCl, 25 mM Tris-HCl, pH 7.5, and the protease inhibitors, phenylmethylsulfonyl fluoride (1 mM), aprotinin (50 g/ml), and leupeptin (50 g/ml) for 10 min on ice. Lysates were spun at 15,000 rpm for 10 min, and the supernatant was saved. Protein levels were determined by the bicinchoninic acid protein assay (Pierce). Appropriate amounts of total cellular protein were loaded on 10% SDS-polyacrylamide gels and electrophoresed at 150 V constant voltage for 3 h. Samples were transferred to nitrocellulose paper and probed with the appropriate antibody. Second antibody was a horseradish peroxidaseconjugated goat anti-mouse IgG, and the signal was detected by the enhanced chemiluminescence method (Amersham Pharmacia Biotech).
Establishment of Retrovirally Infected Cells Expressing Ectopic p53-The retrovirus-producing cell lines PA12-BN and PA12-EN were grown to 75% confluence and fed with fresh DMEM containing 10% FBS. After incubation at 37°C for 16 h, the supernatant was harvested and filtered through a 0.2-m filter. Old medium was removed from a subconfluent 60-mm dish of Saos-2 cells and replaced with 1 ml of filtered supernatant containing 8 g/ml Polybrene. Dishes were rocked for 2 h at 37°C in a humid incubator containing 5% CO 2 and then 3 ml of DMEM containing 10% FBS was added to the dish, and it was further incubated for 48 h. The cells were then trypsinized and replated in a 100-mm dish using DMEM containing 10% FBS and 400 g/ml G418 sulfate. Cells were fed every 3 days with this same medium. After 2 weeks, the resulting drug-resistant colonies were pooled and passaged.
Incorporation of Bromodeoxyuridine-For detecting replicative DNA synthesis, cells were incubated with 10 M bromodeoxyuridine for 30 min prior to fixation. The proportion of cells actively synthesizing DNA was quantitated by anti-bromodeoxyuridine immunoanalysis, and the total DNA content was analyzed by staining with propidium iodide as described previously (30). Cells were fixed with 70% ethanol for at least 2 h, resuspended in the following solutions in order: 0.25% paraformaldehyde in phosphate-buffered saline (PBS), 0.5 mg/ml ribonuclease A in PBS, 0.5% Triton X-100 in 0.1 N HCl, and finally distilled water. Samples were then heated at 97°C for 10 min, immediately placed on ice for additional 10 min, and washed with 0.5% Tween 20 in PBS. The incorporation of bromodeoxyuridine was detected by monoclonal antibromodeoxyuridine antibody conjugated to fluorescein isothiocyanate (Becton Dickinson). Flow cytometric analysis was performed using a FACScan flow cytometer (Becton Dickinson).
Transfection of Reporter Constructs-MCF7, MCF10F, or Saos-2 cells were transfected using the N- [1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium salts liposomal transfection reagent (DOTAP, Boehringer Mannheim). One confluent 100-mm dish of cells was split into three 6-well dishes and incubated for 24 h. Cells were fed with complete medium containing serum and incubated for an additional 3 h. N-[1-(2,3-Dioleoyloxy)propyl]-N,N,N-trimethylammonium salts/DNA mixtures containing 2 g of the relevant reporter plasmid plus 50 ng of the p53 expression plasmid or an equal amount of an empty vector plasmid were prepared according to the manufacturer's instructions and incubated at room temperature for 15 min. Serum-free medium was then added to the mixtures and used to replace the media in the wells. The dishes were incubated at 37°C for 3 h, after which the transfection mix was removed and replaced with complete medium containing serum. After 48 h, the 6-well plates were placed on ice and washed once with PBS. The cells were then lysed by scraping into 120 l of Reporter Buffer (Promega Luciferase Assay System), and samples were spun for 1 min at 14,000 rpm at 4°C. Total protein concentration was determined using a commercially available assay (Bio-Rad). 40 l of each sample was warmed to room temperature and mixed with luciferase assay substrate that was reconstituted with Luciferase Assay Buffer (Promega). Light emission was determined in a TD-20e luminometer (Turner).
Preparation of Nuclear and Cytosolic Extracts-Nuclear and cytosolic extracts were performed as described by Graeber et al. (31). Cells were homogenized in 10 mM Tris-HCl, pH 7.4, containing 10 mM NaCl, 6 mM MgCl 2 , 1 mM dithiothreitol, 0.4 mM phenylmethylsulfonyl fluoride, and 100 M Na 3 VO 4 and spun at 10,000 rpm for 30 s. The supernatant was saved as the cytosolic extract. The pellet was repacked by spinning at 14,000 rpm for 1 min and then nuclei were suspended in a nuclear extraction buffer (20 mM Hepes, pH 7.5, containing 20% glycerol, 500 mM NaCl, 1.5 mM MgCl 2, , 0.2 mM EDTA, 0.1% Triton X-100, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 50 M leupeptin, and 50 g/ml aprotinin), incubated at 4°C for 1 h, and spun at 14,000 rpm for 10 min. This supernatant was saved as the nuclear extract. Lactate dehydrogenase activity was assayed according to Ramirez et al. (32), and histone levels were determined by immunoblotting using an anti-histone antibody that reacts with an epitope that is present on all five histone proteins (H11-4, Boehringer Mannheim). Such assays showed less than 10% cross-contamination between cytosolic and nuclear extracts.
Electrophoretic Mobility Shift Assays-The specific probe that was used for binding, TCGAGCCGGGCATGTCCGGGCATGTCCGGGCAT-GTC, contains the high affinity binding sequence identified by Halazonetis et al. (11) named by them BC or BB.9. In the competition experiments, the nonspecific oligonucleotide (referred to as Sens-1), TCGAAGAAGACGTGCAGGGACCC, was used. Complementary single-stranded oligonucleotides were annealed by incubation at 95°C for 4 min, 65°C for 10 min, and then gradually brought to room temperature. Ends were filled using the Klenow fragment of DNA polymerase to produce a labeled double-stranded oligonucleotide. Appropriate amounts of extracts (1-7 l) were mixed with 1 ng of labeled doublestranded oligonucleotide in a total reaction mixture of 30 l containing 6 l of 5ϫ electrophoretic mobility shift assay buffer (100 mM Hepes, pH 7.9, 0.5 mM EDTA, 50% glycerol, 10 mM MgCl 2 ), 1.5 l of 40 mM spermidine, 1.5 l of 10 mM dithiothreitol, 1 l of 500 g/ml doublestranded poly(dI/dC), and 5-13 l of water with a final salt concentration of 85 mM. The amount of total protein per reaction was normalized, and the reactions were carried out at room temperature for 30 min. For antibody supershift analysis, 2 l of the appropriate undiluted hybridoma supernatant was added. His-tagged human p53 was produced by infection of insect cells with a recombinant baculovirus and purified by nickel-agarose chromatography and used as a positive control (52). Samples were electrophoretically separated on a native 4% polyacrylamide gel at 4°C at 200 V for 2 h. After drying, gels were exposed to Kodak XAR film at Ϫ70°C with an intensifying screen.

MCF10F Cells Express Detectable Levels of p21 mRNA and Protein, Whereas p53-Negative Saos-2 Cells Do Not-Previous
studies have noted that either fibroblasts or keratinocytes from mice that were homozygously deleted for p53 expressed lower basal levels of p21 mRNA as compared with fibroblasts or keratinocytes from mice expressing both alleles of the wild-type p53 gene (21)(22)(23)(24). To characterize further a role for p53 in the basal level of expression of p21, the p53-negative cell line

p53 Regulates Constitutive Expression of p21
Saos-2 was compared with the wild-type p53-expressing cell line MCF10F. Total RNA was extracted from each cell line, and Northern analysis was performed. The p53-negative Saos-2 cell line expressed low levels of p21 mRNA as compared with the wild-type p53-expressing MCF10F cells (Fig. 1A). Total cellular extracts of each cell line were subjected to SDS-polyacrylamide gel electrophoresis and subsequent immunoblotting with an anti-p21-specific antibody (Fig. 1B). MCF10F cells expressed a detectable level of p21, whereas the level of p21 expression in Saos-2 cells was undetectable. To confirm that Saos-2 cells retained the ability to synthesize p21, both MCF10F and Saos-2 cells were treated with 10 mM N,NЈ-hexamethylenebisacetamide (HMBA). HMBA is a non-retinoid, differentiating agent that has previously been shown to induce p21 expression in a p53-independent manner (24). Treatment of Saos-2 cells with HMBA induced expression of p21 demonstrating that Saos-2 cells retained the ability to synthesize p21. Thus, both the level of protein and messenger RNA for p21 were much higher in the p53-expressing MCF10F cells than in the p53negative Saos-2 cells.
Retroviral Infection of Saos-2 Cells Restores Expression of p53 and p21-Previous studies have shown that restoration of wild-type p53 expression through transfection of a suitable expression plasmid did not allow for establishment of stable cell lines expressing wild-type p53 (33)(34)(35)(36)(37). This was presumably due to the fact that plasmid transfection results in a high level of expression of p53 which is incompatible with cell proliferation. Chen et al. (27) utilized recombinant retroviral infection to restore a level of wild-type p53 expression in Saos-2 cells that was comparable to that seen in normal cells and that was compatible with continued proliferation of these cells. To that end, Saos-2 cells were infected with recombinant retroviruses expressing either wild-type human p53 or the mutant human p53 His-273 , and pools of G418 sulfate-resistant cells were established. Immunoblotting of whole cell extracts from these drug-resistant pools demonstrated that both wild-type ( Fig. 2A, lane 4) and mutant ( Fig. 2A, lane 3) p53 expression could be detected in comparison to the parent cells which are p53-negative ( Fig. 2A, lane 2). Furthermore, the pool of Saos-2 cells expressing wild-type p53 expressed a level that is comparable to the endogenous p53 level in MCF10F cells (Fig. 2A,  lane 5). Consistent with previous observations, this level of expression of wild-type p53 that was obtained using recombinant retroviral infection was sufficiently low to allow the cells to continue to grow (Table I). These drug-resistant pools were labeled with bromodeoxyuridine and subjected to flow cytometric analysis to demonstrate that they were actively incorporating DNA. Indeed, the pools expressing wild-type p53 had a similar percentage of bromodeoxyuridine-positive cells as the parent cell line, the pool expressing mutant p53, or the wildtype p53 expressing cell lines WI38, MCF10F, or MCF7 (Table  I). These pools were then examined for the level of p21 expression. Immunoblotting of whole cell extracts demonstrated that Saos-2 cells expressing wild-type but not mutant p53 expressed a level of p21 that was comparable to that of WI38 or MCF7 cells and, in fact, was greater than that seen with MCF10F cells (Fig. 2B). Thus, restoration of expression of wild-type p53 in a p53-negative cell line also restored a basal level of expression of p21.
Constitutive Expression of Luciferase Reporter Constructs Containing the p21 Promoter Is p53-dependent-The observation that reintroduction of p53 expression in Saos-2 cells restored a basal level of p21 expression (Fig. 2) suggests that in the absence of DNA damage, p53 regulates expression of p21. To test directly this notion and to confirm that such regulation is at the level of transcription, wild-type p53-expressing MCF7 cells were transfected with a luciferase reporter construct containing 2.4 kb of the human p21 promoter. To determine whether the basal level of expression that is observed was p53-dependent, an expression plasmid for the human MDM2 protein was co-transfected with the reporter. Mdm2 binds to p53 and inhibits its transcriptional activity, apparently by targeting the p53 protein for degradation (19, 38 -41). Transfection of MCF7 cells with a luciferase reporter construct under control of the p21 promoter, p21P, confirmed a basal level of activation of the p21 promoter (Figs. 3 and 4A). Co-transfection of an expression plasmid encoding human MDM2 protein caused repression of that basal level of expression ( Fig. 3 and  4A). Deletion of the p53-binding sites from this reporter (p21P 2.1) resulted in a complete loss of basal luciferase activity (Fig.  3). In contrast, co-transfection of the plasmid encoding Mdm2 into the p53-negative Saos-2 cells had no effect on the low level of luciferase activity seen in these cells from the same reporter construct (Figs. 3 and 4C). Treatment of MCF7 cells with ultraviolet light induced expression of the full-length p21 promoter construct but not the construct that lacks the p53-binding sites (Fig. 3). Furthermore, treatment of Saos-2 cells with ultraviolet light had no effect on the expression of luciferase from the full-length p21 promoter reporter construct (Fig. 3). These latter results are consistent with the fact that MCF7 cells express a functional wild-type p53 protein (42)(43)(44). 2 The ability of the Mdm2 protein to repress basal expression from a luciferase reporter containing the full-length p21 promoter was confirmed in the wild-type p53-expressing cell line MCF10F (Fig. 4B). Similar to MCF7 cells (Fig. 4A) and in contrast to the p53-negative Saos-2 cells (Fig. 4C), co-transfection of the expression plasmid for Mdm2 protein inhibited the constitutive level of expression that is seen with the reporter p21P. The apparently low level of basal expression seen in Saos-2 cells further strengthens the notion that the basal level of expression is p53-dependent. To confirm that this is indeed the case and not a reflection of different transfection efficiencies by the various cell lines, MCF7, MCF10F, and Saos-2 cells were co-transfected with p21P and an additional reporter that contains Renilla luciferase under the control of the SV40 enhancer and early promoter. This latter reporter construct was used to normalize for transfection efficiency. The results of this analysis demonstrated that, indeed, both MCF7 and MCF10F have a much higher basal level of expression of p21P than the p53-negative Saos-2 cells (Fig. 4D).
To provide further evidence that the basal expression that is seen upon transfection of MCF7 cells with p21P is p53-dependent, MCF7 and Saos-2 cells were co-transfected with p21P and increasing amounts of an expression plasmid encoding a dominant-negative mutant p53 Ala-143 (45). Increasing amounts of the p53 Ala-143 expression plasmid repressed the basal level of expression of p21P in MCF7 cells but not in Saos-2 cells (Fig. 5). Thus, co-transfection of either a dominant-negative p53 or the human MDM2 protein, both of which are capable of inhibiting the endogenous wild-type p53, caused repression of the basal level of expression from the reporter construct containing the p21 promoter. In contrast, co-transfection of the dominantnegative p53 or the human MDM2 protein into the p53-negative Saos-2 cells had no effect on the low level of luciferase activity seen in these cells from the same reporter construct.
p53 in Nuclear Extracts of Untreated Cells Is Capable of Binding to DNA in a Sequence-specific Manner-The data presented thus far implicate a role for endogenous p53 in transcriptional activation of the p21 promoter in untreated proliferating cells. If this is the case, then this endogenous p53 should be capable of binding to DNA. To examine this, electrophoretic mobility shift assays utilized nuclear extracts from three different wild-type p53-expressing cell lines, WI38, MCF10F, and MCF7 to demonstrate that, indeed, the endogenous p53 was capable of binding to DNA prior to DNA damage. All three cell lines were either untreated or treated with 50 J/m 2 of ultraviolet light and then were fractionated into nuclear and cytosolic extracts. Immunoblotting for p53 demonstrated that prior to DNA damage, WI38 and MCF10F cells express a p53 that was primarily localized to the nucleus (Fig.  6, lanes 1 and 2, and 5 and 6), whereas the p53 in untreated MCF7 cells was present primarily in the cytoplasm with a low level detectable in the nuclear fraction (Fig. 6, lanes 9 and 10). After treatment with ultraviolet light, the p53 levels increased substantially in all three cell lines (Fig. 6, lanes 3-4, 7 and 8, and 11-12). Extracts were assayed for a cytoplasmic marker, lactose dehydrogenase, 3 as described under "Experimental Procedures" and were immunoblotted for a nuclear marker, histone H1 (Fig. 6, lower panel). It is estimated that there was less than 10% cross-contamination between the cytoplasmic and nuclear extracts using these markers.
Nuclear and cytoplasmic extracts from untreated and UVtreated MCF cells were normalized for level of p53 protein and used in an electrophoretic mobility shift assay using a consensus p53-binding site as radiolabeled probe (Fig. 7). Both the nuclear and cytoplasmic extracts from UV-treated cells demonstrate a shifted complex with a similar mobility as that of purified p53 (Fig. 7, lanes 11 and 14). The p53-specific antibody 1801 efficiently supershifted this complex, whereas the nonspecific antibody 419 did not (Fig. 7, lanes 12-13 and 15-16). The extracts from untreated cells contained a shifted complex with a similar mobility as purified p53 (Fig. 7, lanes 5 and 8). Incubation with the p53-specific antibody 1801 produced a slower migrating complex but did not substantially affect the original protein-DNA complex (Fig. 7, lanes 6 and 9). The nonspecific antibody 419 had no such effect (Fig. 7, lanes 7 and  10). This result suggests that there is p53 in these extracts which is capable of binding to DNA, but there is also an addi-   p53 Regulates Constitutive Expression of p21 tional DNA-binding protein that is distinct from p53 which produces a shifted complex of similar mobility as purified p53. To test this, nuclear extracts of untreated MCF7 cells were immunoprecipitated with an anti-p53 antibody to clear all detectable p53 protein from the extract as determined by immunoblotting. 4 This extract was compared in an electrophoretic mobility shift assay with a comparable extract that had been immunoprecipitated with the nonspecific antibody 419 as a control as well as a nuclear extract for UV-treated cells that had similarly been immunoprecipitated with 419 (Fig. 8A). Incubation of the p53-specific antibodies 1801 or 421 with untreated nuclear extracts resulted in the detection of slow migrating DNA-protein complexes that were not present in the absence of antibody (Fig. 8A, lanes 2 and 3). These slower migrating complexes were not seen in an extract that had been cleared of p53 by immunoprecipitation but were present in extract that had been immunoprecipitated with a nonspecific antibody (Fig. 8A, lanes 5 and 6 and 8 and 9). Clearing of p53 from the extract had no effect on the protein-DNA complex that migrated to a similar mobility as the p53-DNA complex, confirming that there is a DNA-binding protein in the extract which is distinct from p53. Extracts from UV-treated cells were used to identify the p53-DNA complex that was confirmed by its ability to be efficiently supershifted by both 1801 and 421 (Fig. 8A, lanes 11 and 12). To determine that the binding that was seen was sequence-specific, competition experiments were performed (Fig. 8B). Nuclear extract from untreated MCF7 cells was used in an electrophoretic mobility shift assay in the presence of increasing amounts of either specific probe, BB.9, or a nonspecific probe, Sens-1. Since it was difficult to detect the p53-DNA complex in the absence of antibody, the competition was also performed in the presence of the p53-specific antibody 1801. Increasing amounts of unlabeled BB.9 (Fig. 8B, lanes 9 -11) competed well for the binding to 1801-supershifted complexes, whereas increasing amounts of Sens-1 (Fig. 8B,  lanes 12-14) did not. The faster migrating complex that did not appear to contain p53 was similarly competed suggested that the binding of this protein is also sequence-specific (Fig. 8B).
The untreated MCF7 cells used in these experiments expressed an endogenous p53 that is localized primarily in the cytoplasm (Fig. 6). The DNA binding results were subsequently confirmed in WI38 and MCF10F cells in which the p53 is primarily nuclear prior to DNA damage (Fig. 6). Nuclear and cytoplasmic extracts from untreated and UV-treated WI38 and MCF10F cells were normalized for level of p53 protein and used in similar electrophoretic mobility shift assays (Fig. 9). Both the nuclear and cytoplasmic extracts from UV-treated cells from both cell lines demonstrated a shifted complex with a similar mobility as that of purified p53 (Fig. 9, A and B, lanes  11 and 14). The p53-specific antibody 1801 efficiently supershifted this complex, whereas the nonspecific antibody 419 did not (Fig. 9, A and B, lanes 12 and 13, and 15 and 16). As seen with extracts from MCF7 cells, the extracts from untreated WI38 or MCF10F cells contained a shifted complex with a similar mobility as purified p53 (Fig. 9A, lane 5, and 9B, lanes  5 and 8). Incubation with the p53-specific antibody 1801 pro-  MCF10F cells (B), or Saos-2 cells (C) were transfected as described under "Experimental Procedures" with 2 g of p21P and either 50 ng of pCMV or pCMV-hdm2 as indicated. 48 h later cells were washed, lysed, and assayed for luciferase activity and total protein levels as described under "Experimental Procedures." 18 h prior to lysis, a set of wells containing only pCMV were treated with 50 J/m 2 of ultraviolet light, indicated by ϩUV. The bars represent the average of three independent experiments that had been performed in duplicate. D, MCF7 cells, MCF10F cells, or Saos-2 cells were transfected as described under "Experimental Procedures" with 2 g of p21P and 50 ng of pRL-SV40. 48 h later cells were washed, lysed, and assayed for luciferase activity and total protein levels as described under "Experimental Procedures." The bars represent the average of two independent experiments that had been performed in duplicate. duced a slower migrating complex but did not substantially affect the original protein-DNA complex (Fig. 9, A and B, lanes  6 and 9). The nonspecific antibody 419 had no such effect (Fig.  9, A and B, lanes 7 and 10). These results are consistent with those seen with extracts of MCF7 cells demonstrating that there is p53 in these extracts that is capable of binding to DNA but there is also an additional DNA-binding protein that is distinct from p53 that produces a shifted protein-DNA complex of a similar mobility as p53. The presence of a slower migrating shifted complex that was induced by the p53-specific antibody 1801 but not the nonspecific antibody 419 in untreated extracts from all three cell lines is consistent with the notion that the endogenous p53 in these cells is capable of binding to DNA in untreated, proliferating cells. DISCUSSION Previous studies have suggested that p53 exists in a latent or inactive form in untreated cells and that upon DNA damage not only does the p53 level increase but the p53 itself is modified in some way to activate it for DNA binding and transcriptional activation. Reintroduction of p53 into the p53-negative cell line Saos-2 restored a constitutive level of expression of the cyclin-dependent kinase inhibitor p21 (Fig. 2). A similar result has been reported upon similar retroviral infection of a p53negative peripheral neuroepithelioma cell line (46). These results suggest that the endogenous p53 in untreated, proliferating cells may be capable of transcriptionally regulating p21 expression. Indeed, experiments involving transfection of wildtype p53-expressing MCF7 cells with a luciferase reporter construct containing the p21 promoter have confirmed that this is the case (Figs. 3-5). A reporter containing the full-length p21 promoter but not a promoter construct in which the p53 response elements have been deleted demonstrated a basal level of expression in MCF7 or MCF10F cells but not in p53-negative Saos-2 cells (Fig. 4). This basal level of expression was inhibited by coexpression of either a dominant-negative p53 or the human MDM2 protein (Figs. 3-5). These results imply that the endogenous p53 in MCF7 or MCF10F cells is capable of binding to DNA. This was directly tested through the use of electro- FIG. 6. Biochemical fractionation demonstrates that nuclear p53 levels increase upon UV treatment in WI38, MCF10F, and MCF7 cells. WI38, MCF10F, or MCF7 cells were untreated or treated with 50 J/m 2 of ultraviolet light (ϩUV) and then incubated at 37°C for 20 h prior to fractionation into cytosolic (C) or nuclear extracts (N) as described under "Experimental Procedures." Samples were electrophoresed on a 10% SDS-polyacrylamide gel, transferred to nitrocellulose, and immunoblotted with the anti-p53 monoclonal antibody 1801 or a polyclonal antibody directed against histone H1 as indicated. The protein level for the cytosolic samples ranged from 60 to 120 g. For each cell line, the cytosolic and nuclear samples were normalized so that the loaded samples were obtained from an equivalent number of cells.
FIG. 7. p53 in nuclear extracts of untreated MCF7 cells binds to DNA in an electrophoretic mobility shift assay. Nuclear and cytosolic extracts were prepared from MCF7 cells that were untreated (lanes 5-10) or 24 h after treatment with 50 J/m 2 of ultraviolet light (lanes [11][12][13][14][15][16]. Electrophoretic mobility shift assay was performed as described under "Experimental Procedures." 1 ng of radiolabeled probe (BB.9) was incubated in the absence of protein (lane 1), or in the presence of 0.5 g of human purified human p53 (lanes 2-4) or the appropriate amount of nuclear (lanes 5-7 and 11-13) or cytosolic (lanes 8 -10 and 14 -16) extracts as indicated. The level of p53 loaded in cellular extracts was normalized by adjusting the total protein. Incubations were performed either with no addition (lanes 1, 2, 5, 8, 10, and 14) or in the presence of either 1801 (anti-p53 antibody, lanes 3, 6, 9, 12, and 15) or 419 (anti-SV40 large T antigen, lanes 4, 7, 10, 13, and 16). phoretic mobility shift assays (Figs. 7-9). Although untreated extracts from either MCF7, MCF10F, or WI38 cells contain a DNA-binding protein other than p53 which is capable of shifting a specific radiolabeled probe representing a p53 consensus binding site, experiments in the presence of a p53-specific monoclonal antibody 1801 clearly demonstrated that p53 in these extracts could bind to DNA in a sequence-specific manner . Taken together, these results demonstrate that endogenous p53 in untreated, proliferating cells is capable of binding DNA and activating transcription. Thus, p53 is implicated as playing a role in constitutive expression of a particular target gene, that of the cyclin-dependent kinase inhibitor p21, in proliferating cells in the absence of treatment with DNAdamaging agents.
Attempts to demonstrate the ability of p53 from nuclear extracts to bind DNA have often relied on the use of monoclonal antibody 421. The epitope for 421 is located in the carboxyl end of p53, a region that has been suggested to have a negative effect on the sequence-specific DNA binding of p53 (10 -15). Studies in vitro have demonstrated that 421 can stimulate the binding of wild-type p53 and in some cases can activate select tumor-derived mutant p53 proteins that are incapable of binding to DNA in the absence of antibody (10 -12, 14, 47). In the electrophoretic mobility shift analyses performed here, care was taken to avoid the use of monoclonal antibody 421 for these reasons. The presence of a BB.9-binding protein that is not p53 in the untreated extracts made it necessary to supershift gel shift complexes containing p53 in order to detect the complex of p53 with the probe (Figs. 7-9). Use was made of the antibody 1801 that has an epitope on p53 near the amino-terminal end of the protein (28). Studies have demonstrated that in contrast to 421, 1801 does not restore DNA binding activity to mutant p53 proteins (48). It does, however, exert an enhancing effect on the ability of p53 to bind to DNA, but this is due to the ability of 1801 to stabilize p53 against thermal denaturation that occurs during the incubations that are performed to detect specific DNA binding (48,49). Hence, it is unlikely that 1801 is conferring on the p53 in untreated cell extracts an ability to bind to DNA that this p53 would not otherwise have. Thus, the supershifted complexes produced by incubation with 1801 do indeed reflect the ability of endogenous p53 in the cell to interact in a specific manner with DNA.
Studies utilizing mice that have been homozygously deleted for p53 have shown that the majority of tissues express p21 in a p53-independent manner. Only in the spleen was there substantial differences in p21 expression between p53-null and p53-expressing animals (24). These results indicate that in addition to the p53-dependent mechanism demonstrated here, there must also be p53-independent mechanisms for the regulation of basal levels of p21 expression.
Nevertheless, treatment of cells with DNA-damaging agents clearly inhibits cellular proliferation and involves an increase  7-12). The resulting supernatants were used in an electrophoretic mobility shift assay using 1 ng of radiolabeled probe (BB.9). Incubations were performed either in the absence (lanes 1, 4, 7, and 10) or presence of 1801 (anti-p53 antibody, lanes 2, 5, 8, and 11) or presence of 421 (anti-p534 antibody, lanes 3, 6, 9, and 12). The arrow to the left indicates the position of the p53-DNA complex, and the bracket to the right indicates the position of the supershifted p53-DNA-antibody complex. B, 1 ng of radiolabeled probe (BB.9) was incubated with an appropriate amount of nuclear extract alone (lanes 1 and 8), or in the presence of increasing amounts of unlabeled BB.9 oligonucleotide (lanes 2-4 and 9 -11) or increasing amounts of a nonspecific oligonucleotide, Sens-1 (lanes 5-7 and 12-14). Incubations were performed either in the absence (lanes 1-7) or presence of 1801 (anti-p53 antibody, lanes 8 -14). The bracket to the right indicates the position of the supershifted p53-DNA-antibody complex.
FIG. 9. p53 in nuclear extracts of untreated WI38 or MCF10F cells binds to DNA in an electrophoretic mobility shift assay. Nuclear and cytosolic extracts were prepared from WI38 (A) or MCF10F (B) cells that were untreated (lanes 5-10) or 24 h after treatment with 50 J/m 2 of ultraviolet light (lanes [11][12][13][14][15][16]. Electrophoretic mobility shift assay was performed as described under "Experimental Procedures." 1 ng of radiolabeled probe (BB.9) was incubated in the absence of protein (lane 1), or in the presence of 0.5 g of human purified human p53 (lanes 2-4) or the appropriate amount of nuclear (lanes 5-7 and 11-13) or cytosolic (lanes 8 -10 and 14 -16) extracts as indicated. The level of p53 loaded in cellular extracts was normalized by adjusting the total protein. Incubations were performed either with no addition (lanes 1, 2, 5, 8, 10, and 14), or in the presence of either 1801 (anti-p53 antibody, lanes 3, 6, 9, 12, and 15) or 419 (anti-SV40 large T antigen, lanes 4, 7, 10, 13, and 16). The arrow to the left indicates the position of the p53-DNA complex, and the bracket to the right indicates the position of the supershifted p53-DNA-antibody complex.
p53 Regulates Constitutive Expression of p21 in p21 expression that is p53-dependent (1)(2)(3). It is reasonable to expect that cells growing in vitro experience a low level of oxidative DNA damage, and there may be damage resulting from errors during DNA synthesis. This low level of DNA damage may be responsible for activation of a subset of the p53 protein in the cell leading to transcriptional activation of particular target genes at a low level. Indeed it is likely that cells in vivo are subjected to similar low levels of DNA damage. Thus, the results presented here do not necessarily contradict the notion that upon DNA damage, p53 may, in fact, be modified in some way to increase its ability to bind DNA and transcriptionally activate target genes. Post-translational modification of p53 upon DNA damage of cells has been documented, and some studies suggest that this modification may be necessary to achieve the full induction of p53 target gene expression that is seen after treatment with DNA-damaging agents (18,19).
The human gene for thrombospondin-1 has previously been identified as a target for transcriptional activation by p53 (50). Studies leading to this observation demonstrated that fibroblasts from early passage cells obtained from Li-Fraumeni patients constitutively expressed thrombospondin-1, but later passage cells that had lost expression of p53 no longer secreted thrombospondin-1. Transfection studies demonstrated that the thrombospondin-1 promoter was a target for transcriptional activation by p53, although a specific binding site for p53 in this promoter has yet to be identified (50). As thrombospondin-1 has anti-angiogenic activity, the observation that its constitutive expression is p53-dependent is consistent with the role of p53 as a tumor suppressor. The experiments reported here with the p21 promoter confirm the ability of p53 to transcriptionally regulate constitutive expression of particular target genes in proliferating cells, thereby suggesting a mechanism that is consistent with the report of p53-dependent expression of thrombospondin-1 in proliferating human fibroblasts (50).
Chen et al. (27) demonstrated that expression of wild-type but not mutant p53 in Saos-2 cells by retroviral infection will inhibit the ability of these cells to grow in soft agar and grow as tumors in nude mice. Similar results were obtained by retroviral infection of a p53 null peripheral neuroepithelioma cell line (51). As with the studies reported here ( Fig. 2 and Table I), the level of p53 that was expressed in the cells in both these studies was sufficiently low to allow the cells to continue to proliferate albeit at a slower rate than the parent cell lines (27,51). The implication of these observations is that this low level of p53 is capable of suppressing the oncogenic phenotype in these cells suggesting that the ability of p53 to transcriptionally regulate constitutive expression of select target genes may, therefore, play a role in its ability to function as a tumor suppressor. The increased tumorigenicity that results from the loss of basal expression of these p53-dependent targets would then contribute to the selective pressure for the loss of wild-type p53 function in human tumors.