A chromatin-associated and transcriptionally inactive p53-Mdm2 complex occurs in mdm2 SNP309 homozygous cells.

In cancer cells, the function of the tumor suppressor protein p53 is usually blocked. Impairment of the p53 pathway results in tumor cells with endogenous overexpression of Mdm2 via a naturally occurring single nucleotide polymorphism (SNP) in the mdm2 gene at position 309. Here we report that in mdm2 SNP309 cells, inactivation of p53 results in a chromatin-associated Mdm2-p53 complex without clearance of p53 by protein degradation. Nuclear accumulation of p53 protein in mdm2 SNP309 cells results after 6 h of camptothecin, etoposide, or mitomycin C treatment, with the p53 protein phosphorylated at Ser15. Chromatin immunoprecipitation demonstrated p53 and Mdm2 bound to p53 responsive elements. Interestingly, although the p53 protein was able to bind to DNA, quantitative PCR showed compromised transcription of endogenous target genes. Additionally, exogenously introduced p53 was incapable of activating transcription from p53 responsive elements in SNP309 cells, confirming the trans-acting nature of the inhibitor. Inhibition of Mdm2 by siRNA resulted in transcriptional activation of these p53 targets. Our data suggest that overproduction of Mdm2, resulting from a naturally occurring SNP, inhibits chromatin-bound p53 from activating the transcription of its target genes.

In cancer cells, the function of the tumor suppressor protein p53 is usually blocked. Impairment of the p53 pathway results in tumor cells with endogenous overexpression of Mdm2 via a naturally occurring single nucleotide polymorphism (SNP) in the mdm2 gene at position 309. Here we report that in mdm2 SNP309 cells, inactivation of p53 results in a chromatin-associated Mdm2-p53 complex without clearance of p53 by protein degradation. Nuclear accumulation of p53 protein in mdm2 SNP309 cells results after 6 h of camptothecin, etoposide, or mitomycin C treatment, with the p53 protein phosphorylated at Ser 15 . Chromatin immunoprecipitation demonstrated p53 and Mdm2 bound to p53 responsive elements. Interestingly, although the p53 protein was able to bind to DNA, quantitative PCR showed compromised transcription of endogenous target genes. Additionally, exogenously introduced p53 was incapable of activating transcription from p53 responsive elements in SNP309 cells, confirming the trans-acting nature of the inhibitor. Inhibition of Mdm2 by siRNA resulted in transcriptional activation of these p53 targets. Our data suggest that overproduction of Mdm2, resulting from a naturally occurring SNP, inhibits chromatin-bound p53 from activating the transcription of its target genes.
The p53 tumor suppressor protein plays a central role in the prevention of cancer development by causing growth arrest and/or apoptosis during stress (1). When the p53 pathway is inhibited, tumorigenesis is accelerated. Often p53 is inhibited by mutations within the DNA binding domain of p53 that lead to loss of the p53 tumor suppressor activity (2,3). Additionally, p53 protein function is inhibited when oncoproteins bind to p53. The Mdm2 oncoprotein binds to p53 and inactivates the p53 tumor suppressor activity (4). Interestingly, the Mdm2 protein is part of a negative feedback loop with p53; p53 activates mdm2 transcription, and then the Mdm2 protein inhibits p53 function. The Mdm2 protein functions to inactivate p53 in at least two ways. Mdm2 is an E3 ubiquitin ligase for p53 and targets the tumor suppressor for degradation by the ubiquitin proteolysis pathway (5,6). The p53 protein is maintained at low levels in normally dividing cells in part through its interaction with Mdm2 (7). This protein-protein interaction also blocks the p53 trans-activation domain and thereby inhibits p53 transcriptional activity (8)(9)(10). In addition, under certain situations the overexpression of Mdm2 results in p53-Mdm2 complexes that fail to bind tightly to DNA (11) and thus causes subsequent inhibition of p53 activity.
Many cancer cells have high levels of the oncogenic Mdm2 protein because of either increased expression (12) or amplification of the mdm2 gene (13). Mdm2 overexpression also results from the P2 promoter when a single nucleotide polymorphism (SNP) 1 at position 309 in the first intron of the mdm2 gene causes increased affinity for the ubiquitous transcription factor SP1 (14). In cells homozygous for SNP309 the p53 pathway is compromised.
In an effort to closely examine the compromised signaling from p53 in cells homozygous for mdm2 SNP309, DNA damage was used to activate the checkpoint pathway and p53 functional activities were monitored. In doing so we found that p53 in cell lines homozygous for SNP309 could be significantly stabilized after 6 h of DNA damage treatment with different drugs. This p53 remained associated with the cellular Mdm2 protein in the nucleus and was not sequestered to the nucleolus. Importantly, this p53 bound to chromatin in conjunction with Mdm2 and was unable to activate transcription of downstream target genes. As discussed herein, we addressed the inhibitory actions of Mdm2 by showing that a DNA-bound form of p53 is blocked for trans-activation activity in the presence of high levels of Mdm2.
Cell Culture-ML-1 cells were a generous gift from Michael Kastan, and the MANCA cell line was a generous gift from Andrew Koff. K562 cells were purchased from ATCC and do not contain p53. A875 cells were a generous gift from the Levine laboratory. H1299 cells were a generous gift from C. Prives laboratory. All cells were maintained in 10% fetal bovine serum (FBS, Gemini), RPMI 1640 medium (Mediatech) and 5% CO 2 . The medium was supplemented with 100 g of penicillin per ml and 100 g of streptomycin per ml. Cells were seeded at 2.5 ϫ 10 5 , and exponentially growing cells were used in all experiments.
DNA-damaging Agents-DNA-damaging agents and the proteasome inhibitor LLnL were added to the medium of exponentially growing cell cultures. The cells were treated at a concentration of 5 ϫ 10 5 to 6 ϫ 10 5 cells per ml. The following concentrations were used throughout the study: CPT, 0. Flow Cytometry-FACS analysis was carried out on a BD Biosciences FACS scan. Cells were spun down at 2000 rpm for 7 min, washed twice with phosphate-buffered saline (PBS) containing 2% bovine serum albumin and 0.1% NaN 3 . Cells were then fixed in 30% ethanol. Propidium iodide staining and RNase treatment were carried out at 37°C for 30 min prior to analysis.
Nuclear Extracts-Cells were pelleted at 2000 rpm for 7 min at 4°C and washed 2ϫ with ice-cold 1ϫ PBS (pH 7.3). Cells were then washed 1ϫ in 5 packed cell volumes with Buffer A (Hepes pH 7.9, 10 mM; MgCl 2 , 1.5 mM; KCl, 10 mM; PMSF, 0.5 mM; DTT, 0.5 mM; leupeptin, 2 g/ml; and phosphatase inhibitor mixture I (Sigma)). After washing, cells were resuspended in 2 packed cell volumes of Buffer A and incubated on ice for 10 min. After centrifugation (10 min at 12,000 rpm at 4°C) the supernatant was removed to give the cytoplasmic fraction. The pellet was then resuspended in Buffer C (Hepes pH 7.9 20 mM; glycerol, 25%; NaCl, 420 mM; MgCl 2 , 1.5 mM; EDTA, 0.2 mM; PMSF, 0.5 mM; DTT, 0.5 mM; leupeptin, 2 g/ml; and phosphatase inhibitor mixture I. Cells were resuspended with a 20-gauge needle to a concentration of 40 million cells/120 l. The cell suspension was rocked for 30 min at 4°C and then centrifuged for 30 min at 13,000 rpm at 4°C. The supernatant was the nuclear fraction, which was stored at Ϫ80°C. Western Blotting Analysis-Nuclear extracts, whole cell extracts, or immunoprecipitated samples were separated on a 10% SDS-PAGE followed by electrotransfer to a nitrocellulose membrane. The following p53 antibodies were used: 421, 1801, and 240 (supernatant antibodies), Ser 15 , and the p53 polyclonal antibody. The following Mdm2 antibodies were used: 2A10 (supernatant antibody), SMP14, and D7. For p21/Waf1 detection, we used Ab-1 from Oncogene Research Science. For PARP cleavage we used anti-mouse monoclonal antibody from PharMingen. For nucleolin detection we used the antibody MS-3 from Santa Cruz Biotechnology. For actin detection we used a rabbit polyclonal purchased from Sigma. The signals were visualized after incubation with goat anti-mouse or goat anti-rabbit secondary antibody (Sigma) using ECL solutions.
Immunofluorescence-Cells were plated onto glass coverslips using poly-L-lysine (Sigma), fixed with 4% paraformaldehyde for 15 min at room temperature, and then permeabilized with Triton X-100 1% for 5 min at Ϫ20°C. After washing three times with PBS-FBS, 1%, fixed cells were incubated for 1 h at room temperature with mouse anti-human p53 Ab6 from Calbiochem (ML-1 cells) or rabbit anti-human polyclonal p53 antibody from Santa Cruz Biotechnology (MANCA and A875 cells) 1:200 dilution in PBS-FBS, 1%. Cells were washed three times with PBS-FBS, 1% and then incubated for 30 min with secondary Texas Red-conjugated donkey anti-mouse antibody (Santa Cruz Biotechnology) or fluorescein isothiocyanate-conjugated goat anti-rabbit antibody (Santa Cruz Biotechnology) 1:400 dilution in PBS-FBS, 1%. Coverslips were then mounted using Vectashield mounting medium with 4Ј,6diamidino-2-phenylindole (DAPI, Vector Laboratories) and observed under a fluorescence microscope.
Quantitative RT-PCR-For each sample, 5 g of RNA were obtained using the Qiagen RNeasy kit, per the manufacturer's protocol. Each sample was then diluted up to 50 l with water. The samples were then incubated with reagents from the cDNA Archive Kit (Applied Biosystems). The mixture contained RT buffer, dNTP mix, random primers, and multiscribe RT. The 2ϫ RT Master Mix, along with the cDNA, was incubated at room temperature for 10 min and then at 37°C for 2 h. cDNA was stored at Ϫ20°C. For PCR, the program was as follows: one cycle of 50°C UNG incubation for 2 min, and one cycle of 95°C priming for 10 min, followed by 40 cycles of 95°C denaturation for 15 s and 60°C annealing for 1 min. This reaction was carried out in an Applied Biosystems 5700 Prism Spectrofluorometric Thermal Cycler. Fluorescence was measured during the annealing step and plotted automatically for each sample. Assays on demand mdm2 (order: HS00242813_M1), fas (order: HS00538709_M1), p21/waf1 (order: HS00355782_M1), and gadd45 (order: HS00169255_M1) were purchased from Applied Biosystems, as was the PDAR for GAPDH (order: 4333764F). Sequences are the copyright of the company and will not be offered upon request. Samples were normalized to the GAPDH values.
Transient Transfection and Luciferase Activity Assay-The plasmids used in this study were the SN3 plasmid containing the wild-type p53 gene (17), and a plasmid containing the mdm2 (14) or the p21/waf1 (18) promoter with p53 responsive elements adjacent to a luciferase reporter (14). 3 ϫ 10 6 cells were transferred from RPMI media to OptiMEM media (Invitrogen) and co-transfected with 2 g of a plasmid containing the mdm2 or p21/waf1 p53 binding site adjacent to a luciferase reporter and increasing amounts (50 -400 ng) of SN3 plasmid, containing the wild-type p53 gene. The total amount of DNA transfected into each sample was normalized with a carrier plasmid, pGL2. The transient transfections were performed using the Lipofectamine 2000 reagent (Invitrogen), according to the manufacturer's instructions. After 24 h, cells were harvested and analyzed using the Luciferase Assay system (Promega) according to the manufacturer's indications. Luciferase activity was read using a Luminoskan reader, and the results were normalized to total protein values.
Chromatin Immunoprecipitation-Cells were cross-linked with 1% formaldehyde at 37°C for 30 min then quenched with glycine to 125 mM. The cells were washed with phosphate-buffered saline and collected into 100 mM Tris-Cl, pH 9.4, 10 mM dithiothreitol. The cell pellet was resuspended (10 mM Tris-Cl, pH 8.0, 0.25% Triton X-100, 0.5% Nonidet P-40, 10 mM EDTA, 0.5 mM EGTA, 1 mM PMSF) and incubated on ice for 10 min. Nuclei were collected by centrifugation, washed (10 mM Tris-Cl, pH 8.0, 0.2 M NaCl, 1 mM EDTA, 0.5 mM EGTA, 1 mM PMSF) and resuspended in the same buffer without NaCl. Samples were sonicated (10ϫ, 10 s each), centrifuged, and 0.10 volume of 10ϫ precipitation buffer (10% Triton X-100, 1% sodium deoxycholate, 1.4 M NaCl) was added to the supernatants. One-fifth of each sample was saved and designated as "Input." Immunoprecipitations were carried out overnight with 2 g of Mdm2-specific antibody (SMP14), 2 g of p53-specific antibody Ab6, or 1:200 dilution of p53-Ser 15 phosphospecific antibody. The next day protein A plus G-Sepharose beads (Amersham Biosciences) were added for 2 h with rocking at 4°C. The subsequent washes were as described (19). The washed resin was resuspended in 100 l of elution buffer (1% SDS, 0.1 M NaHCO 3 ) and reverse cross-linked at 65°C overnight. DNA fragments were purified (QiaQuick spin kit) and PCR-amplified using primers designed to amplify the p53 binding sites in the mdm2 gene (forward primer: 5Ј-C-GGGAGTTCAGGGTAAAGGT-3Ј, reverse primer: 5Ј-AGCAAGTCGGT-GCTTACCTG-3Ј) or p21/waf1 gene (forward primer: 5Ј-GTGGCTCTG-ATTGGCTTTCTG-3Ј, reverse primer: 5Ј-CTGAAAACAGGCAGCCCA-AG-3Ј). [ 32 P]dCTP (PerkinElmer Life Sciences) was added to the PCR reaction. Amplified products were analyzed by acrylamide gel electrophoresis. Gels were dried for 1 h at 55°C, and autoradiography was performed. Quantitative PCR using ChIP samples was carried out on a PE5700 PCR machine using a TaqMan Master Mix (Applied Biosystems). The PCR cycles were: 50°C for 2 min, 95°C for 10 min, then cycles of 95°C for 15 s, and 60°C for 1 min repeated 40 times. The fold change in the specific binding was normalized to GAPDH and Mock IP values. The probe and primer sequences are given below. Primers are in italics, TaqMan probe is in bold, and the p53 binding site is underlined: Electrophoretic Mobility Shift Assay (EMSA)-Custom oligonucleotides for DNA binding analysis were ordered from Operon Technologies and for SCS and RGC mutant as described (20). The superconsensus site (SCS) contained three adjacent p53 half sites. The sequence of the oligonucleotide was: Top, 5Ј-TCGAGCCGGGCATGTCCGGGCATGTC-CGGGCATGTC-3Ј; and Bottom, 5Ј-TCGAGACATGCCCGGACATGCC-CGGACATGCCCGGC-3Ј. Labeling of the oligo was performed using the large fragment of DNA polymerase and [ 32 P]dCTP. Electrophoretic mobility shift assays were carried out in reaction mixtures with 150 pmol of 32 P oligonucleotide. 10 g of nuclear extract was added, and the reaction was incubated for 20 min at room temperature in a reaction buffer containing 20 mM Hepes pH 7.8, 100 mM KCl, 1 mM EDTA pH 8.0, 1 mM DTT, 1 g of sheared salmon sperm DNA, and 10% glycerol. In the case of competition, unlabeled competitor in 50 -100ϫ fold-excess was added into the incubation. mdm2, fas, and a mutant oligo were used for competition studies. The sequences of the oligos were as follows: mdm2: Top, 5Ј-CCGGGCTGGTCAAGTTCAGACACGTTCCGAA-ACTGCAGTAAAAGGAGTTAAGTCCTGACTTGTCTCCC-3Ј; and Bottom, 5Ј-CCGGGGGAGACAAGTCAGGACTTAACTCCTTTTACTGCA-GTTTCGGAACGTGTCTGAACTTGACCAGC-3Ј; fas: Top, 5Ј-CCGGGC-TCCTGGACAAGCCCTGACAAGCCAAGCCAC-3Ј, and Bottom, 5Ј-CC-GGGTGGCTTGGCTTGTCAGGGCTTGTCCAGGAGC-3Ј; mutant: Top, 5Ј-TCGAGTTTAATGGACTTTAATGGCCTTTAATTTTC-3Ј, and Bottom: 5Ј-TCGAGAAAATTAAAGGCCATTAAAGTCCATTAAAC-3Ј.
Reactions were carried out in the presence or absence of 421 antibody as indicated. Samples were separated by 4% polyacrylamide gel electrophoresis (gels were prerun at 100 V for 15 min at 4°C) at 200 V for 3-3.5 h. Gels were dried for 1 h at 55°C and autoradiography was performed.
siRNA-Mdm2 expression was lowered using mdm2 siRNA (Darmacon, SmartPool-a mixture of 4 different mdm2 siRNA, catalog M-003279-02). 200-pmol siRNA was transfected into cells at 70 -80% confluency using oligofectamine reagent (Invitrogen) according to the manufacturer's instructions. Control siRNA (Darmacon, SmartPool, catalog D-001206-13-05) has no known target in mammalian genomes. Cells were lysed 48 h after transfection, and protein levels of Mdm2, p53, p21/Waf1, and actin were analyzed. Total RNA was also extracted from transfected cells using the RNeasy kit, per the manufacturer's protocol and analyzed by quantitative PCR as described above. Cells were also retransfected with 2 g of the reporter constructs mentioned above 24 h after siRNA transfection and grown for an additional 24 h. Lysates were prepared for luciferase analysis as described previously.

Endogenous Overexpression of Mdm2 via a Naturally Occurring SNP Inhibits Apoptosis following Chemotherapeutic Drug
Treatment-We previously documented induced p53 stabilization and subsequent apoptosis in ML-1 cells treated for 6 h with CPT, ETOP, or MC, and no apoptosis induction in the p53deficient human cell line K562 treated with the same drugs (21). A naturally occurring SNP at position 309 in the promoter 2 region of the mdm2 gene (Fig. 1A) results in overexpression of Mdm2 protein, and inhibition of drug-induced apoptosis occurs in homozygous mdm2 SNP309-containing cell lines MANCA and A875 (14). Both cell lines overexpress Mdm2 at the protein and RNA levels, which results in the attenuation of p53 stabilization in the first 3 h after ETOP treatment (14).
To closely examine the compromised molecular signaling, we compared p53 and Mdm2 protein levels in ML-1, MANCA, and A875 cell lines after 6 h of drug treatment with expression of the proteins in the documented wild-type p53-expressing cell line ML-1 (which is homozygous wild-type for mdm2) (14). Cells were treated with chemotherapeutic DNA-damaging drugs or were left untreated, and extracts were analyzed by Western blot for p53 and Mdm2 proteins (Fig. 1B). The drugs CPT, ETOP, and MC increased the level of p53 in ML-1 cells as well as in MANCA and A875 cells (Fig. 1B, lanes 1-12), whereas no p53 was expressed in the p53-negative control cell line K562 (Fig. 1B, lanes 13-16). We reproducibly noted that the level of p53 protein in MANCA cells before drug treatment was high relative to other cell lines examined, and we do not have an explanation for this increased basal p53 in MANCA cells at this time. The p53 in MANCA and A875 is wild-type (14), as seen by DNA sequence analysis of p53 exons 1-11 described under "Materials and Methods." An examination of apoptosis indicators (PARP cleavage and FACS analysis with propidium iodide staining to score sub-G 1 DNA content) was used to compare drug-induced apoptosis. Western blot analysis demonstrated PARP cleavage in ML-1 cells after 6 h of CPT and ETOP treatment and slight cleavage of PARP after 6 h of MC treatment but no PARP cleavage was detected in the other cell lines when they were treated with the drugs (Fig. 2A). FACS analysis demonstrated an increased sub-G 1 population (indicated by the M1 gate) up to 55% only in the ML-1 cells treated with drugs and not in any of the other drug-treated cell lines (Fig. 2, B and C).
p53 Protein Is Phosphorylated at Ser 15 in Cells with mdm2 SNP309 -The signal transduction pathway toward p53 involves a critical phosphorylation event at Ser 15 , which helps not only to stabilize the p53 but also to activate the transcriptional activity of the protein (22). To examine kinase signaling to the p53 protein in the cell lines examined, Western blot analysis with antibody specifically recognizing Ser 15 -phosphorylated p53 was carried out. Phosphorylation of p53 at Ser 15 in the MANCA and A875 cells was reproducibly detected (Fig.  3A). We examined the ability of the p53 protein to be phosphorylated in MANCA and A875 cells after 6 h of chemotherapeutic treatment because we knew p53 activation in ML-1 cells occurred at this time point (21). Interestingly, we observed significant p53 phosphorylation in cell lines homozygous for the mdm2 SNP309 after CPT, ETOP, and MC treatment (Fig. 3A,  lanes 5-12). In response to drug treatments the level of p53 phosphorylation in the ML-1 cell line increased as previously described (21), and MANCA and A875 cells demonstrated p53 phosphorylation greater than ML-1 cells (Fig. 3A, compare  lanes 1-4 to lanes 5-12). The p53 stabilization was not greater in MANCA and A875 cells, as shown in Fig. 1.
p53 localized to the nucleus in both wild-type and SNP309 homozygous cells, as seen in immunofluorescence experiments (Fig. 4). Additionally, immunofluorescence studies indicated that the nuclear p53 levels increased following drug treatment of ML-1, MANCA, and A875 cells (Fig. 4).
The p53 Protein Is Compromised for Activating Downstream Target Genes in MANCA or A875 Cells-The p53 protein induces the cyclin-dependent protein kinase inhibitor p21/Waf1 (23), and its increase is part of the cell cycle checkpoint (24). As one indicator of checkpoint activation after DNA damage, we examined the levels of p21/Waf1 protein. As expected, treatment of the ML-1 cells with CPT, ETOP, and MC resulted in an increase in p21/Waf1 protein, whereas in K562 cells the absence of p53 resulted in no p21/Waf1 increase (Fig. 3B, lanes  1-4 and lanes 13-16). p21/Waf1 protein levels in MANCA and A875 cells showed great variation with severely attenuated DNA damage induction by CPT, ETOP, and MC treatment. Densitometric analysis demonstrated a 2.25-fold induction of p21/Waf1 levels in ML-1 cells treated with CPT and MC and a 7.3-fold induction after ETOP treatment. In MANCA and A875 cells, CPT and MC treatments did not produce any increase in p21/Waf1 levels. ETOP treatment gave 1.56-fold induction of the checkpoint protein in MANCA cells and 1.74-fold increase in A875 cells (data not shown).
We examined the ability of p53 target genes to be activated after the DNA damage treatments that resulted in p53 stabilization. Examination of p53 downstream target genes activation by quantitative RT-PCR revealed compromised activation in the MANCA and A875 cells, even though p53 stabilization was achieved (Fig. 5). Whereas the stabilized p53 in ML-1 cells activated p21/waf1, gadd45, fas, and mdm2, we saw compromised activation of the genes in MANCA and A875 cells (Fig.  5). Comparison of activation of these p53 targets in the mdm2 SNP309 cell lines to the activation in the p53-deficient cell line K562 demonstrated transcription slightly higher indicating some, albeit limited, p53 activity. Additionally, in the absence of p53 in the K562 cells slight activation of p21/waf1 transcription occurred. Overall, activation of the downstream target genes in MANCA and A875 cells resembled the results seen in the cell line that did not express p53 (K562) more than those seen in the ML-1 cell line (Fig. 5), indicating that the stabilized p53 protein in the two homozygous mdm2 SNP309 cell lines was not functioning correctly.
We asked if exogenously introduced p53 was capable of activating transcription from constructs with either mdm2 or p21/waf1 p53 responsive elements in the MANCA and A875 cells to closely examine the ability of wild-type p53 to be inhibited in the SNP309 cell lines. Transient transfection experiments were carried out with p53 protein provided from the plasmid SN3 (a generous gift from Bert Vogelstein). Both reporters were activated by the exogenously introduced p53 in the K562 and H1299 cells, neither of which is homozygous for mdm2 SNP309. These two cell lines do not express endogenous p53. The H1299 adherent cells were compared with A875 whereas the K562 suspension cells were compared with MANCA. We used p53-null cell lines for comparison in the transfection study because our quantitative PCR data demonstrated that p53 target gene activation in SNP309 homozygous cells resembled the activation seen in cells with no p53 expression. Activation from both reporter constructs occurred with increasing p53 plasmid in the H1299 and K562 cell lines (Fig.  6, A-D), whereas severely compromised activation was evident in both MANCA and A875 cell lines (Fig. 6, A-D). Western blot analysis indicated that this compromised activation occurred in the presence of a significant increase in exogenous p53 expression in the transient transfected cells (Fig. 6E and data not  shown). These results point toward the presence of a transacting inhibitor of the p53 function.
Stabilized p53 in MANCA and A875 Cells Binds to DNA-Inhibition of the DNA binding ability is one way in which Mdm2 inhibits p53 (11). Immunofluorescence images demonstrated that p53 in SNP309 cell lines was present in the nu-  4, 8, 12, and 16) for 6 h, and nuclear extracts were then prepared from these samples. 50 g of nuclear protein were subjected to SDS-PAGE (10%) and Western blot analysis. The nitrocellulose membranes were probed with the p53-specific polyclonal antibody or Mdm2-specific monoclonal antibody SMP14.  1, 5, 9, and 13) or were treated with 0.5 M CPT (lanes 2, 6, 10, and 14), 8 M ETOP (lanes 3, 7, 11, and 15), or 5 M MC ( lanes  4, 8, 12, and 16) for 6 h, and nuclear extracts were then prepared from these samples. 50 g of nuclear protein were subjected to SDS-PAGE (10%) and Western blot analysis. The nitrocellulose membranes were probed with the p53-Ser 15 phosphospecific antibody (A), the p21/Waf1-specific monoclonal antibody (Ab-1, B), anti-actin (C). cleoplasm (Fig. 4), and therefore we examined if this p53 was associated with the DNA. To determine if the p53 protein that was compromised for activating transcription was able to bind to p53 responsive elements in chromatin, we compared p53 ChIP in ML-1, MANCA, K562, and A875 cells (Fig. 7A). Increased mdm2 and p21/waf1 chromatin was immunoprecipi- tated with a p53-specific antibody in ML-1 cells treated with ETOP (Fig. 7, lane 4) and some p53 localized on the chromatin of these genes prior to DNA damage (Fig. 7, lane 3). The same experiment was carried out in MANCA, A875, and K562 cells. Comparison of the input chromatin and mock immunoprecipitation for all samples demonstrated that the sets were normalized and gave barely detectable background (Fig. 7A, lanes 1, 2 , 5, 6, 9, 10, 13, and 14). In MANCA cells p53 protein was associated with the mdm2 gene with no evident increase after DNA damage (Fig. 7A, lanes 7 and 8, mdm2 gene). Barely detectable p21/waf1 chromatin was immunoprecipitated in MANCA cells using an antibody recognizing the N terminus of p53 and was only evident by phosphorimager analysis (Fig. 7A,  lanes 7 and 8, p21/waf1 gene and data not shown). It appears that this epitope was masked in the MANCA cell chromatinassociated p53, as an antibody to phosphorylated p53 indicated the association of the protein with p21/waf1 chromatin (Fig.  8B, compare lanes 7 and 8, p21/waf1 gene). Increased mdm2 and p21/waf1 chromatin were immunoprecipitated with the p53-specific antibody in A875 after ETOP treatment (Fig. 7A,  compare lanes 15 and 16). In the p53-null cell line K562, p53 antibody did not precipitate any chromatin-containing p53 re-sponsive elements before or after ETOP treatment (Fig. 7A,  lanes 11 and 12).
We compared the ability of ML-1, MANCA, and A875 p53 protein in nuclear extracts to bind to the SCS using an EMSA. The p53 DNA binding activity in the presence of the p53specific antibody 421 is known to be activated in EMSA (25), and such activated binding was assayed using nuclear extracts from CPT-, ETOP-, and MC-treated cells. In the presence of the p53 antibody 421, we observed an induced p53 shift in the ML-1, MANCA, and A875 nuclear extract samples (Fig. 7B, compare odd lanes that contain no 421 antibody to even lanes where the 421 antibody is present and Fig. 7C, compare lanes  1-2 and 3-4). Interestingly, whereas the p53 binding activity was activated in the ML-1 cells after drug treatment, p53 binding activity was present in the MANCA nuclear extract prior to drug treatment, and no further increase was evident after drug treatment (correlating with the Western blot and immunofluorescence data showing high basal levels of p53, as well as with ChIP results). In A875 cells the p53 binding activity was activated after ETOP treatment, and this binding was specific as demonstrated by competition with oligonucleotides containing p53 responsive element sequence but not with FIG. 5. The MANCA and A875 p53 protein is transcriptionally compromised. Quantitative RT-PCR was used to detect fold induction of the p21/waf1, gadd45, fas, and mdm2 transcripts in ML-1, MANCA, A875, and K562 cells after DNA damage. All results were normalized to untreated samples and GAPDH values. Standard deviations represent results from two independent experiments. a mutant oligonucleotide sequence (Fig. 7C). This DNA binding specificity for p53 in EMSA was also seen for MANCA cell p53 (data not shown).
Increased Mdm2 Protein Binds to Chromatin with p53 Responsive Elements in MANCA and A875 Cells-Although a bimodal mechanism for the inhibition of p53 by Mdm2 has been FIG. 6. The p53 protein provided in trans is transcriptionally inactive in the mdm2 SNP309 homozygous cells. Transient transfection in K562 and MANCA cell lines (A and B) or H1299 and A875 cells (C and D) was carried out with increasing amounts (50 -400 ng) of a plasmid for expression of the wild-type p53 cDNA (SN3) and 2 g of the plasmid containing the human mdm2 (A and C) or p21/waf1 (B and D) p53 binding site adjacent to a luciferase reporter. The amount of DNA co-transfected in each sample was normalized using a carrier plasmid, pGL2. Fold induction in luciferase activity was measured and normalized to total protein concentration. Results are representative of three independent experiments. Western blot analysis (E and F) was performed to assess the transfection efficiency and p53 expression in H1299 and A875 cells. Whole cell protein extract was prepared from the co-transfected cells, and 50 g of extract were subjected to SDS-PAGE (10%) and Western blot analysis. The nitrocellulose membrane was probed with a mixture of the p53-specific monoclonal antibodies (240, 421, 1801; E) or anti-actin (F). Lanes 1 and 5 represent non-transfected samples; lanes 2 and 6 are protein extracts from a 50-ng SN3 transfection, lanes 3 and 7 from a 100-ng SN3 transfection, and lanes 4 and 8 from a 400-ng SN3 transfection. described, the capacity of Mdm2 to target p53 for degradation often overshadows the capacity of Mdm2 to inhibit p53 transcription activity. Recently however it has been seen that Mdm2 protein is present at p53 binding sites in chromatin in the Mdm2-overexpressing cell line SJSA-1 (26). This Mdm2chromatin association was p53-mediated and caused gene silencing because of histone ubiquitylation. Our transient transfection experiments pointed toward a trans-acting inhibitory factor of the DNA-bound p53. We carried out ChIP studies in wild-type or mdm2 SNP309 homozygous cells to examine if Mdm2 was associated with p53 bound to chromatin in MANCA and A875, but not in ML-1 cells. In ML-1 cells we were unable to identify Mdm2 bound to p53 responsive elements of the mdm2 or p21/waf1 genes in both untreated samples or after 3 h of ETOP treatment (Fig. 8A, lanes 3 and 4). In MANCA and A875 cells there was a small amount of Mdm2 bound to mdm2 and p21/waf1 genes chromatin in untreated cells (Fig. 8, A,  lane 7 and C, lane 3) and a striking increase in Mdm2 localized at these sites after ETOP treatment (Fig. 8, A, lane 8 and C,  lane 4). Quantitative PCR using the precipitated chromatin showed a 2.5-fold increase above background in untreated samples and went up to a 8-fold increase in ETOP-treated cells (data not shown; fold increase represents the average of three independent experiments). No Mdm2 protein was present at the mdm2 or p21/waf1 responsive elements in the p53-null cell line K562 (Fig. 8A, lanes 11 and 12). We examined if the recruitment of Mdm2 to the p53 responsive elements was mediated by the p53 protein. ChIP with p53-Ser 15 phosphospecific antibody showed enhanced binding of p53 protein after DNA damage in ML-1 cells (Fig. 8B, compare lanes 3 and 4) and also in MANCA and A875 cells (Fig. 8B, compare lanes 7 and 8 and  Fig. 8C, compare lanes 5 and 6). As expected, p53-Ser 15 ChIP did not demonstrate any signal in the K562 cell line (Fig. 8B,  lanes 11 and 12). The ChIP data argue that Mdm2 can localize to p53 responsive elements in the MANCA and A875 cells and increases with increased p53 localization.
We examined the soluble p53-Mdm2 complex present in the nucleus of the tested cell lines before and after DNA damage. Co-immunoprecipitation studies were carried out using nuclear extracts derived from the cells treated with the DNA-damaging agents to confirm that Mdm2 was associated with the stabilized p53 that had attenuated function. ML-1, MANCA, and A875 nuclear extracts were compared for p53-Mdm2 complex formation using the Mdm2-specific antibody 2A10 to pull-down the complex, and the blot was analyzed for Mdm2 and p53 proteins (Fig. 8D). In the ML-1 cells no p53 was detected when Mdm2 was immunoprecipitated. However in MANCA and A875 cells, p53 was pulled-down in complex with Mdm2 when  4, 8, 12, and 16) for 3 h. PCR was carried out using primers specific for the p53 responsive elements in the mdm2 and p21/waf1 genes. In Mock IP samples (lanes 2, 6, 10, and 14), no antibody was added to the immunoprecipitation reaction. One-fifth of the input DNA from each sample was also amplified and designated as Input (lanes 1, 5, 9, and 13). B, EMSA with the 32 P-labeled SCS containing three p53 binding sites and 10 g of nuclear extracts from ML-1 and MANCA untreated samples (lanes 1 and 2 and 9 and 10) or samples treated with 0.5 M CPT (lanes 3 and 4 and lanes 11 and 12), 8 M ETOP (lanes 5 and 6 and lanes 13 and 14), or 5 M MC (lanes 7 and 8 and lanes 15 and 16) for 6 h. Reactions contained the p53-specific monoclonal antibody 421 where indicated. C, EMSA with the 32 P-labeled SCS containing three p53 binding sites and 10 g of nuclear extracts from A875 untreated samples (lanes 1 and 2) or treated with 8 M ETOP (lanes 3 and 4). Competition of the 421-induced gel shift from ETOP-treated A875 nuclear samples was carried out using 50ϫ or 100ϫ fold-excess unlabeled oligonucleotide corresponding to the unlabeled SCS (lanes 5 and 6), mdm2 (lanes 7 and 8), or fas (lanes 9 and 10) p53 binding sites, as well as 100ϫ nonspecific mutant oligonucleotide (lane 11). cells were treated with ETOP but not before treatment (Fig.  8D, lanes 4 -7). These co-immunoprecipitation results indicate that the increase in chromatin-associated Mdm2 correlates with an increased Mdm2-p53 protein association and strengthens the argument that Mdm2 is recruited to chromatin only in complex with p53. The reciprocal co-immunoprecipitation experiment was done using an antibody specific for the p53 protein, and this analysis demonstrated a greater interaction of p53 with Mdm2 in MANCA and A875 cells than in ML-1 cells (data not shown).
We performed mdm2 siRNA experiments to ensure that the Mdm2-elevated protein levels were required to attenuate the p53-mediated transcriptional activation in cells homozygous for SNP309. Down-regulation of Mdm2 (Fig. 8E, lane 3, Mdm2 Western blot) did not lead to any increase in the p53 protein levels (Fig. 8E, lane 3, p53 Western blot). However, p21/Waf1 protein levels were increased when Mdm2 was down-regulated (Fig. 8E, lane 3, p21/Waf1 Western blot). Mdm2 has been shown to be a p21/Waf1-negative regulator independently of p53, by facilitating the interaction of p21/Waf1 with the pro-  4, 8, and 12) for 3 h. PCR was carried out using primers specific for the p53 responsive elements in the mdm2 and p21/waf1 genes. In Mock IP samples (lanes 2, 6, and 10), no antibody was added to the immunoprecipitation reaction. One-fifth of the input DNA from each sample was also amplified and designated as Input (lanes 1, 5, and 9). B, p53-Ser 15 chromatin immunoprecipitation in ML-1 (lanes 1-4), MANCA (lanes [5][6][7][8], and K562 cells (lanes 9 -12). Cells were either left untreated (lanes 3, 7, and 11) or treated with 8 M ETOP (lanes 4, 8, and 12) for 3 h. PCR was carried out using primers specific for the p53 responsive elements in the mdm2 and p21/waf1 genes. In Mock IP samples (lanes 2, 6, and 10) no antibody was added to the immunoprecipitation reaction. One-fifth of the input DNA from each sample was also amplified and designated as Input (lanes 1, 5, and 9). C, Mdm2 (lanes 3 and 4) or p53-Ser 15 (lanes 5 and 6) chromatin immunoprecipitation in A875 cells. Cells were either left untreated (lanes 3 and 5) or treated with 8 M ETOP (lanes 4 and 6) for 3 h. PCR was carried out using primers specific for the p53 responsive elements in the mdm2 and p21/waf1 genes. In Mock IP sample (lane 2) no antibody was added to the immunoprecipitation reaction. One-fifth of the input DNA from each sample was also amplified and designated as Input (lane 1). D, co-immunoprecipitation using the Mdm2 antibody 2A10. Lane 1, mock IP using MANCA cell extract, no antibody was added to the immunoprecipitation reaction; lanes 2, 4, and 6: immunoprecipitation in untreated ML-1, MANCA, and A875 samples; lanes 3, 5, 7: immunoprecipitation in ETOP-treated ML-1, MANCA, and A875 samples. Upper panel was probed with Mdm2-specific antibody 2A10, bottom panel was probed with a mix of p53-specific antibodies (421, 1801, and 240). E, Western blot analysis of A875 cells after mdm2 siRNA treatment. Cells were transfected with mdm2 or control siRNA and harvested 24 h later. 50 g of total protein were subjected to SDS-PAGE (10%) and Western blot analysis. The nitrocellulose membranes were probed with the Mdm2-specific antibody SMP14, the p53-specific polyclonal antibody, the p21/Waf1-specific monoclonal antibody (Ab-1), and anti-actin. F, quantification of mdm2 and p21/waf1 transcripts following mdm2 siRNA treatment. RNA was extracted 24 h following siRNA transfection in A875 cells. Quantitative PCR was used to detect fold change of the mdm2 and p21/waf1 transcripts. All results were normalized to untreated samples and GAPDH values. Standard deviations represent results from three independent experiments. G, transient transfection in A875 cells. Cells were transfected with mdm2 or control siRNA and retransfected 24 h later with a plasmid containing the p21/waf1 or mdm2 p53 binding site adjacent to a luciferase reporter. Protein extracts were prepared and fold induction in luciferase activity was measured and normalized to total protein concentration. teasomal C8 subunit (27). To make certain that the increase in p21/Waf1 protein was caused by reactivation of the p53 transcriptional activity, we looked at p21/waf1 mRNA levels after mdm2 siRNA treatment. Mdm2 down-regulation led on average to a 3-fold increase in p21/waf1 mRNA levels (Fig. 8F), illustrating reactivation of the p53 transcription factor ability. Re-transfection of cells with two p53-dependent reporter constructs after mdm2 siRNA treatment also showed enhanced p53 function in the absence of Mdm2, as illustrated by the fold change in the luciferase readings (Fig. 8G). DISCUSSION The interaction between soluble Mdm2 and p53 has been well described in experimental systems with forced overexpression of both proteins. However the chromatin-bound associations between p53 and Mdm2 proteins have only begun to be addressed (26). It has been shown that Mdm2 regulates p53 by at least two mechanisms. The interaction of Mdm2 with p53 blocks the trans-activation domain of p53 and inhibits the protein transcriptional activity (8). Additionally, Mdm2 is an E3 ubiquitin ligase for p53, helping to target p53 for degradation by the ubiquitin proteolysis pathway (5), (6). Mdm2-mediated inhibition of p53 trans-activation activity contributes to the impairment of the p53 pathway in cancer cell lines endogenously overexpressing Mdm2. Here we report that endogenous overexpression of Mdm2 via a naturally occurring SNP results in the formation of an inactive complex on and off chromatin. We determined that two cancer cell lines (MANCA and A875) expressing wild-type p53 have stabilized and phosphorylated p53 after DNA damage, demonstrating limited clearance of p53 protein by degradation. Mdm2 down-regulation did not lead to any increase in the p53 protein levels, suggesting also that p53 is not excessively degraded in SNP309 homozygous cells. This p53 bound site-specifically to DNA and chromatin. Although the amount of p53 bound to the mdm2 gene in MANCA cells remained unchanged after treatment in the total p53 antibody ChIP (Fig. 7A, lanes 7 and 8, mdm2 gene) and very little p53 was detected at the p21/waf1 gene (Fig. 7A, lanes 7 and 8, p21/waf1 gene), we reason that the antibody epitope (which is located at the N terminus) is masked by p53 phosphorylation and association with its Mdm2 partner. We provide evidence that Mdm2 localized with p53 at the p53 responsive elements, suggesting that Mdm2 blocked the transactivation activity of DNA-bound p53. This directly agrees with the recent work by Minsky and Oren (26) indicating that Mdm2 can be recruited to chromatin resulting in histone ubiquitylation and repression of transcription. Both p53 and Mdm2 bound to chromatin in MANCA and A875 cells after DNAdamaging drug treatment, and Mdm2 protein recruitment to the chromatin was increased after such drug treatment (Fig. 8, A and C). Our data support the findings that stabilized p53 can, at times, maintain an interaction with Mdm2 (22). The colocalization of Mdm2 with p53 on chromatin correlates with attenuated p53 function. This mechanism for inhibition of p53 transcriptional activity most likely applies in all cell types that overexpress Mdm2 and would result in a latent form of p53 that could be reactivated in the absence of Mdm2 protein.
Using siRNA, we demonstrated that Mdm2 down-regulation resulted in reactivation of p53 transcriptional activity (Fig. 8, E-G).
We saw that exogenously introduced p53 was incapable of activating transcription from p53 response element reporter constructs in either MANCA or A875 cells, whereas exogenously introduced p53 was able to activate transcription from the reporter constructs in two other cell lines not homozygous for mdm2 SNP309 (H1299 and K562). These data further support the argument that a trans-acting factor, i.e. Mdm2, inhibited the activity of the exogenously introduced p53 in the naturally occurring mdm2 SNP309 cell lines. The inactivation of mdm2 by siRNA reactivated the endogenous p53 so that it was able to activate the reporter constructs.
We present a model for how Mdm2 blocks the ability of p53 to activate target genes (Fig. 9). In cells that have wild-type p53 we see stabilization of p53 protein after DNA damage. In the ML-1 cell line, which does not have a SNP at position 309 of the mdm2 gene, we see dissociation of p53-Ser 15 from Mdm2. This allows for the activated p53 protein to behave as a viable FIG. 9. A model illustrating that in cells homozygous for the mdm2 SNP309, p53 that is induced after DNA damage does not dissociate from the excessively expressed Mdm2 protein. This association does not interfere with the ability of p53 to bind to DNA but does impair p53 transcriptional activity.
transcription factor and turn on downstream target genes. In cells overexpressing Mdm2, such as cells homozygous for mdm2 SNP309 (MANCA and A875), there is increased association of the p53-Ser 15 -Mdm2 complex on chromatin after DNA damage. This p53, in the presence of Mdm2 on chromatin, does not act as a viable transcription factor.
We have begun to look for other proteins that might be part of the chromatin-associated complex together with p53 and Mdm2. In vivo footprinting results have suggested that a large complex protects the p53 responsive element regions (28). Coimmunoprecipitation with the p53 antibody 421 and subsequent Coomassie Blue staining demonstrated that a number of high mobility bands were specifically co-immunoprecipitated in SNP309 cells. Mass spectrometry analysis revealed that the nucleolin protein was one of the interacting proteins in MANCA cells, whereas this same band was not detected in the ML-1 samples (data not shown). Western blot analysis of coimmunoprecipitated samples to compare the interaction between p53, Mdm2, and nucleolin indicated an interaction in the MANCA and A875 cells and none in the ML-1 cells (data not shown). In Fig. 4 we demonstrated no p53 nucleolar sequestration despite its association with the nucleolar protein. Nucleolin has previously been described to interact with p53 and translocate into the nucleoplasm after DNA damage (29). It is possible that nucleolin is part of the chromatin-associated complex. Nucleolin plays a role in transcription, being a functional B-cell-specific transcription factor (30) and may have a yet to be determined role in p53 function.
Recent work in cancer prevention has demonstrated the effectiveness of small molecule inhibitors that disrupt the Mdm2-p53 interaction (31,32). Cancer cells that demonstrate inhibition of wild-type p53 function due to the overexpression of Mdm2 protein are perfect targets for small molecule inhibitor therapy. Our work suggests that the p53 in such cells exists in a latent state associated with chromatin, ready to activate transcription. By releasing the chromatin-associated Mdm2, the latent p53 could be reactivated in the presence of DNAdamaging agents. A type of combination therapy coupling the release of Mdm2 from p53 with the activation of p53 by DNA damage might prove effective. In the future we would like to investigate if such small molecule inhibitors are able to inhibit the interaction of p53-Ser 15 with Mdm2 in the MANCA and A875 cells. The release of Mdm2 from chromatin-bound p53 protein would allow reactivation of downstream target genes (as we saw with siRNA) and would result in cellular growth arrest or apoptosis.