Stabilization of p53 by Adenovirus E1A Occurs through Its Amino-terminal Region by Modification of the Ubiquitin-Proteasome Pathway*

The human epidermoid carcinoma-derived cell line MA1, established by introduction of the adenovirus E1A 12 S cDNA linked to the hormone-inducible promoter, elicits apoptosis after induction of E1A12 S in response to dexamethasone. E1A expression caused accumulation of wild type p53 more than 10-fold within 24 h after dexamethasone treatment. The cell lines that express E1A mutants containing a deletion either in the amino terminus or the conserved region 1 were unable to accumulate p53. p53 accumulated was degraded efficiently in vitro in the S10–0 extract (S10–0) prepared from MA1 cells in an ATP and ubiquitin-dependent manner, but not in S10–24 prepared after treatment with dexamethasone for 24 h. The p53 polyubiquitination activity in S100–0 was calcium-dependent and reduced greatly in S100–24. Ubiquitin affinity chromatography revealed that p53 ubiquitination activity in eluates thought to contain ubiquitin-conjugating enzymes decreased greatly in S100–24 as compared with S100–0. The accumulation of p53 was accompanied by the increase in the level of Mdm2, which has been shown to degrade p53 through binding to it. The high p53 level, however, was maintained until the late stage of the apoptotic process. These results indicate that the stabilization of p53 by E1A occurs through modification of a ubiquitin-specific enzyme(s) in the ubiquitin-proteasome pathway.

The p53 protein, one of the tumor suppressors is primarily involved with the maintenance of genomic integrity. The level of wild type p53 in cells is maintained at a low level due to a short protein half-life but is increased steeply by stabilization in response to external stress signals, e.g. ionizing radiation (1)(2)(3)(4)(5). p53 functions at G 1 cell cycle checkpoint and arrests the damaged cells before entering S phase until DNA damage is repaired (6 -8). Loss of p53 function therefore results in accumulation of mutations that lead the cells to malignancy (9). This p53-dependent growth arrest is mediated by the activation of the gene encoding cyclin-dependent kinase inhibitor p21 (10 -15).
The function of p53 is also involved in the signaling of apoptosis in response to external stress signals such as toxic insults that damage DNA (3, 5, 16 -18) and viral gene expression (19,20). In the case of the adenovirus E1A gene, the products E1A interact with multiple cellular factors and perturb the regulation of cell growth and differentiation, and in some cases they induce apoptosis in a wild type p53-dependent manner (19 -22). The E1A domain required for induction of apoptosis has been recently mapped in the amino terminus and conserved region 1 (CR1) (23,24) that correspond to p300 binding site (25). The region is also required for stabilization of p53 (20,26,27), although the precise mechanism of this stabilization has not yet been clarified.
The turnover of p53 has been shown to be regulated by the ubiquitin-dependent proteolysis system (28,29). Ubiquitin is first activated via thiolester formation with the ubiquitin-activating enzyme (E1) 1 and then transferred to members of the ubiquitin-conjugating enzyme family, E2s. E2s transfer ubiquitin to specific substrates and form the polyubiquitin chain at a particular Lys residue in collaboration with members of the ubiquitin ligase family, E3s. The polyubiquitin chain targets the substrate for degradation by the proteasome (30,31). Wildtype p53 accumulates markedly in a temperature-sensitive mutant of mouse BALB/3T3 cells at the restrictive temperature owing to a defect in E1 (32). The ubiquitin-mediated degradation of p53 is also shown by the action of human papillomavirus E6 protein. E6 protein associates with cellular E6-AP protein, a ubiquitin ligase (33), and targets p53 for degradation (34). It has been recently shown that Mdm2, a nuclear phosphoprotein first identified as an oncoprotein (35) binds to p53 and targets it for degradation by the ubiquitin-proteasome pathway (36,37). In reverse, the level of Mdm2 is regulated by p53, and the transcription of the mdm2 gene is stimulated by p53 (38,39). p53 is also a substrate for cleavage by the calcium-activated protease, calpain (40 -43). Calpain activity is regulated by autoproteolysis and the inhibitor protein calpastatin (44).
In the present study, the mechanism of wild type p53 stabilization in a human epidermoid carcinoma cell line KB was studied with its derivative cell line MA1, in which the expression of E1A can be regulated by dexamethasone (45). The role of the amino terminus and CR1 of E1A was studied by establishing the cell lines that express the mutant E1A containing a deletion in either of these regions in response to dexamethasone but fail to induce apoptosis. The results indicate that stabilization of p53 occurred exclusively by modification of the ubiquitin-proteasome pathway by E1A. The p53 polyubiquitination activity was much reduced in MA1 cells after E1A expression. Stabilization of p53 was accompanied by the elevation of Mdm2; however, the high level of p53 was maintained thereafter, indicating that the accumulated p53 is insensitive to the action of Mdm2 accumulated.

EXPERIMENTAL PROCEDURES
Materials-Bovine erythrocyte ubiquitin, yeast apyrase, creatine phosphokinase, and creatine phosphate were purchased from Sigma. Proteasome inhibitor, carbobenzoxy-L-leucyl-L-leucyl-L-norvalinal (Z-LLnV-OH), was obtained from Peptide Institute (Tokyo, Japan). Protein A-Sepharose FF beads were purchased from Amersham Pharmacia Biotech. Calpastatin peptide (46) was synthesized with Fmoc (N-(9fluorenyl)methoxycarbonyl) amino acid derivatives on a PerSeptive Biosystems 9050 synthesizer. After automatic synthesis, protecting groups were removed and peptides were released from supporting resins by treatment with trifluoroacetic acid (Kanto Chemicals, Japan) in the presence of reducing agents. Biotinylated ubiquitin was prepared from bovine erythrocyte ubiquitin by reacting with Sulfo-NHS-LC-Biotin (Pierce) as described (47). Mouse anti-adenovirus type 2 E1A monoclonal antibody (clone M73), anti-human Mdm2 monoclonal antibody (SMP14), anti-human p53 NH 2 -terminal region monoclonal antibody (clone Bp53-12), and rabbit polyclonal antibody for human p53 (FL-393) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Mouse monoclonal antibody to the COOH-terminal region of human p53 (pAb421) was purchased from Oncogene Science. Horseradish peroxidase (HRP)-conjugated goat anti-mouse immunoglobulin (Ig) was obtained from Cappel. HRP-conjugated donkey anti-rabbit Ig, which has reduced cross-reactivity with mouse, bovine, and human Igs, was purchased from Chemicon. The ECL detection system was obtained from Amersham Pharmacia Biotech.
Cell Lines-MA1 cells were established from human epidermoid carcinoma cell line KB by introducing the adenovirus E1A 12 S cDNA linked to the mouse mammary tumor virus long terminal repeat. MA1 cells express E1A protein (E1A 12 S ) in response to dexamethasone and elicit apoptosis (45). B12 and T5 cells were similarly established from KB cells by introducing mouse mammary tumor virus long terminal repeat-linked E1A 12 S cDNA containing a deletion either from codon 17 to 23 or 54 to 69, respectively (Fig. 1A). These cell lines were cultivated at 37°C in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. For the viability assay, both floating and adherent cells were pooled and assessed by trypan blue exclusion (45).
Construction of the Deletion Derivatives of E1A 12 S cDNA-For construction of E1A12S ⌬17-23 cDNA containing a deletion from codon 17 to 23, pBRE1A12S was cleaved at the PvuII site between codons 22 and 23. The DNA was shortened by successive digestion with ExoIII and mung bean nuclease and circularized by blunt end ligation. The extent of deletion and the joining of two ends in phase was confirmed by DNA sequencing. The plasmid DNA was then cleaved with HindIII and BglII, and the fragment containing the E1A12S ⌬17-23 cDNA was inserted between the HindIII-BglII site of pMTV-dhfr (48), displacing the dihydrofolate reductase cDNA to generate pMTVE1A12S ⌬17-23 . E1A12S ⌬54 -69 cDNA was constructed by using site-specific mutagenesis according to Kunkel et al. (49). M13E1A12S was constructed by insertion of the E1A 12 S cDNA at the KpnI-PstI site of M13mp18. The sense strand of M13E1A12S was annealed with the 31-base oligonucleotide lacking codons 54 -69, and the antisense strand was synthesized with T4 DNA polymerase and Escherichia coli DNA ligase. The doublestranded DNA was then transfected to E. coli ung ϩ cells (BMH71-18 mutS), and the recombinant plasmid, M13E1A12S ⌬54 -69 , was isolated. The DNA was cleaved with HindIII and BglII, and the fragment containing the E1A 12 S cDNA ⌬54 -69 was similarly inserted into pMTVdhfr to generate pMTVE1A12S ⌬54 -69 .
Preparation of Cell Extracts-Subconfluent cultures of MA1 cells, treated or untreated with 1 M dexamethasone, were washed twice in ice-cold phosphate-buffered saline (0.14 M NaCl, 0.01 M potassium phosphate, pH 7.4), once in 10 volumes of ice-cold hypotonic buffer (20 mM Tris⅐HCl, pH 7.4, 5 mM MgCl 2 , 8 mM KCl, and 1 mM dithiothreitol (DTT)), and then resuspended in 10 volumes of hypotonic buffer. After incubation on ice for 15 min, the swollen cells were collected in a 7.5-ml Dounce vessel and precipitated by centrifugation at 200 ϫ g for 2 min at 4°C to remove excess buffer. The cells were then disrupted by homogenization with a tight pestle 40 times on ice. The homogenate was centrifuged twice at 10,000 ϫ g for 5 min, and the turbid supernatant (S10 extract) was collected. Aliquots of the S10 extracts were centrifuged at 105,000 ϫ g for 6 h to prepare the S100 extracts, which lack most of the proteasome activity.
Western Blotting-25-50 g of protein in SDS-solubilized whole cell lysate (45) were electrophoresed on 9% polyacrylamide gels with Laemmli running buffer (25 mM Tris/glycine, pH 8.3, and 0.1% SDS) (50). Proteins were electrophoretically transferred to nitrocellulose membranes and incubated in TBS-T (25 mM Tris⅐HCl, pH 8.0, 137 mM NaCl, 2.7 mM KCl, 0.1% Tween 20) containing 5% skim milk and 0.1% thimerosal at room temperature for 1 h to minimize nonspecific binding of antibody. The membrane was incubated with primary antibody at an appropriate dilution as indicated in the figure legends at room temperature for 1 h and washed three times in TBS-T for 15 min. The membrane was then incubated with secondary antibody at a dilution of 1:10,000 at room temperature for 1 h and washed three times in TBS-T for 15 min. Immune complexes were detected by ECL by treating the membrane with the ECL detection system. Protein concentration was determined by a dye-binding assay (51).
Immunoprecipitation of p53-MA1 cells (approximately 2 ϫ 10 8 cells) treated with dexamethasone for 24 h were washed twice with ice-cold phosphate-buffered saline and lysed with 10 ml of RIPA buffer (2 mM Tris⅐HCl, pH 7.6, 5 mM EDTA, 150 mM NaCl, 1% Nonidet P-40, 1% sodium deoxycholate, 0.025% SDS, 0.1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride⅐HCl). The lysate was centrifuged at 1,700 ϫ g for 10 min to remove cell debris. p53 was immunoprecipitated from supernatant with 10 l of pAb421 and 20 l of protein A-Sepharose FF beads overnight. The immunoprecipitate was washed four times with 10 ml of RIPA buffer, resuspended in RIPA buffer to give a 10% suspension, and stored at 4°C. The immunoprecipitate was washed three times with hypotonic buffer and resuspended in the same buffer to a final concentration of 0.5% just before the degradation assay.
Ubiquitin Affinity Column-Ubiquitin-Sepharose beads (ϳ25 mg of ubiquitin/ml of swollen gel) were prepared as described by Ciechanover et al. (52). 4 ml of swollen ubiquitin-Sepharose beads were stuffed in a 16/10 column (Amersham Pharmacia Biotech) and equilibrated with 15 ml of buffer A (50 mM Tris-HCl, pH 7.2, 10 mM MgCl 2 , 0.2 mM DTT) containing 2 mM ATP. 15 mg of protein of S100 -0 and S100 -24 were incubated after the addition of 10 mM DTT on ice for 30 min to dissociate ubiquitin conjugates of E1 and E2s, and endogenous ubiquitin was removed by ultrafiltration with Macrosep-10k after 1:40 dilution with buffer A. The samples were concentrated to a final volume of 2 ml. The samples were mixed with an equal volume of buffer A containing 4 mM ATP, 20 mM phosphocreatine, and 10 units/ml inorganic pyrophosphatase and applied to the ubiquitin-Sepharose columns after the removal of insoluble materials by centrifugation at 15,000 ϫ g for 15 min. After unadsorbed fractions were collected, the columns were washed with 24 ml of buffer A containing 2 mM ATP, and adsorbed components were eluted with 12 ml of 1 M KCl in 50 mM Tris-HCl, pH 7.2 (KCl fraction), 12 ml of 50 mM Tris-HCl, pH 7.2, 12 ml of 2 mM AMP in 50 mM Tris-HCl, pH 7.2, containing 40 M sodium pyrophosphate (AMP fraction), 12 ml of 10 mM DTT in 50 mM Tris-HCl, pH 7.2 (DTT fraction) and 12 ml of 2 mM DTT in 50 mM Tris-HCl, pH 9.0 (pH 9 fraction). Fractions of 4 ml were collected at a flow rate of 0.2 ml/min. The AMP, DTT, and pH 9 fractions were combined and concentrated by ultrafiltration with Macrosep-10k (Pall Filtron). The salts in the eluates were removed by another ultrafiltration to a final volume of 0.4 ml with Microsep-1k (Pall Filtron) after 1:100 dilution with hypotonic buffer.
Degradation and Ubiquitination Assays of p53-Degradation of p53 was performed in either the S10 or the S100 extracts. The latter lacks most of the proteasome. Ubiquitination of p53 was carried out with the S100 extracts.
For degradation assays in the S10 extract, the reaction mixture (20 l) contained 40 mM Tris⅐HCl, pH 7.6, 5 mM MgCl 2 , 100 mM NaCl, 2 mM DTT, 2 mM ATP, 20 mM phosphocreatine, 0.2 unit/l creatine phosphokinase, 0.2 g/l bovine erythrocyte ubiquitin, and 150 g of protein in S10. The degradation in the S100 extract was performed in the reaction mixture (6 -30 l) containing 40 mM Tris⅐HCl, pH 7.6, 5 mM MgCl 2 , 100 mM NaCl, 2 mM DTT, and 30 -150 g of protein in S100. The reaction was carried out at 37°C for 20 min or at 30°C for 40 -80 min and terminated by boiling the mixture for 5 min after the addition of an equal volume of 2ϫ Laemmli sample buffer (280 mM Tris⅐HCl, pH 6.8, 200 mM DTT, 4% SDS, 12% glycerol, 0.01% bromphenol blue) (50). Aliquots of the mixture were subjected to Western blotting using anti-p53 polyclonal antibody FL-393 and HRP-conjugated secondary antibody to rabbit IgG.
For ubiquitination assays, the reaction mixture contained 40 mM Tris⅐HCl, pH 7.6, 5 mM MgCl 2 , 100 mM NaCl, 2 mM DTT, and 0.2 g/l biotinylated ubiquitin, 200 g of protein in S100 -24 as a source of p53 and 1 mg of protein in S100 -0. S100 -0 was prepared from MA1 cells, and S100 -24 was prepared from MA1 cells treated with dexamethasone for 24 h. The reaction was carried out at 30°C for 30 min and terminated on ice for 5 min. After the addition of 4-fold volumes of RIPA buffer to the reaction mixture, both ubiquitinated and unprocessed p53 were immunoprecipitated at 4°C for 3 h with 1 l each of pAb421 and Bp53-12 and 2.5 l of protein A-Sepharose FF beads. The immunoprecipitates were washed four times with 1 ml of RIPA buffer and then boiled after the addition of 40 l of 2ϫ Laemmli sample buffer for 5 min. Aliquots of 10 l of the sample were subjected to Western blot analysis. Ubiquitinated p53 was detected by HRP-conjugated streptavidin, and unprocessed p53 was detected by anti-p53 polyclonal antibody FL-393 and HRP-conjugated secondary antibody to rabbit IgG.

RESULTS
The Amino Terminus and CR1 of E1A Are Required for Accumulation of p53-The MA1 cell line, established from human epidermoid carcinoma KB cells by introducing the mouse mammary tumor virus long terminal repeat-linked E1A 12 S cDNA, elicited apoptosis following the induction of E1A 12 S in response to dexamethasone and began to lose viability after about 36 h (Fig. 1B). Western blot analysis showed that after induction of E1A expression, the level of p53 increased steeply and reached a maximal level of more than 10-fold higher than the original at 24 h (Fig. 1C), while the level of p53 mRNA was unchanged (45). Subsequently, topoisomerase II␣ began to be degraded via the ubiquitin-proteasome pathway reducing the level to 1 ⁄50 of the original as previously shown (45,47,53).
The E1A domain required for the induction of apoptosis has been mapped in the amino terminus and CR1 (23,24). These regions are required for binding to p300 (Fig. 1A) (25). To analyze the function of this domain, the expression vectors, pMTV-E1A⌬17-23 and pMTV-E1A⌬54 -69, that express E1A 12S lacking the codons between 17 and 23 and between 54 and 69, respectively in response to dexamethasone (dex) (Fig.  1A), were introduced into KB cells with pSV2neo (54), and the cell lines B12 and T5 were established. Upon treatment with dexamethasone, both cells expressed mutant E1A (Fig. 1C) but failed to induce apoptosis (Fig. 1B). Neither accumulation of p53 nor degradation of topoisomerase II␣ was induced significantly (Fig. 1C). These results indicate that the function of the amino terminus and the CR1 of E1A is required for stabilization of p53, and the increase in the p53 level correlates with the induction of topoisomerase II␣ degradation.
p53 Is Degraded by the Ubiquitin-Proteasome System in MA1 Cells-Although the mechanism of induction of p53 stabilization by E1A has not yet been clarified, it has been shown that p53 is degraded either by the ubiquitin-proteasome system or by calpain in a calcium-dependent manner. To determine the involvement of the ubiquitin-proteasome system and/or calpain in p53 degradation, the S10 extracts were prepared from MA1, B12, and T5 cells before and after treatment with dexamethasone, and p53 levels in these cell extracts were directly com-  (24), are shown by the thick lines below the map of E1A 12 S . The mutant cDNAs, E1A 12S⌬17-23 and E1A 12S⌬54 -69 , containing a deletion from codon 17 to 23 and from codon 54 to 69 are shown at the bottom. B, inability of E1A mutants to induce apoptosis. Sparse cultures of MA1, B12, and T5 cells were treated with 1 M dex and harvested at the times indicated. The latter two cells express E1A 12S⌬17-23 and E1A 12S⌬54 -69 , respectively, in response to dexamethasone. Cell viability was determined by trypan blue exclusion and expressed as the percentage of the total cell number. C, inability of E1A 12S⌬17-23 and E1A 12S⌬54 -69 to stabilize p53 and to induce degradation of topoisomerase II␣. The S10 extracts were prepared from MA1, B12, and T5 cells after treatment with dexamethasone. Aliquots of 50 g of protein/lane were electrophoresed, and E1A, p53, and topoisomerase II␣ were quantified by Western blotting. The levels of these proteins are shown as relative values. pared by Western blotting. p53 was scarcely detected in the extracts prepared from these cells before treatment with dexamethasone. Only the case of the MA1 0 h extract (S10 -0) is shown in Fig. 2A, lane 1. The p53 level in the MA1 24-h extract (S10 -24), prepared after treatment with dexamethasone for 24 h, was almost 50-fold higher than the original level (lane 2), while those in B12 and T5 48-h extracts, prepared from B12 and T5 cells treated with dexamethasone for 48 h, were very low and similar to the original level (lanes 3 and 4).
The degradation reaction of p53 in vitro was performed by incubating an aliquot of S10 -24 as a source of p53 in S10 -0 at 30°C for 80 min in the presence of 2 mM ATP, 0.4 g/l ubiquitin, and various protease inhibitors. After the reaction, the mixture was subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE), and the amount of p53 undegraded was analyzed by Western blotting (Fig. 2B). Under these conditions, most of p53 was degraded (lane 2). The presence of Z-LLnV-OH, a proteasome inhibitor or apyrase that destroys ATP, suppressed p53 degradation almost completely (lanes 3 and 4), while the presence of calpastatin peptide, an inhibitor of calpain or EGTA that chelates with calcium protected p53 from degradation only slightly (lanes 5 and 6).
To assay the calpain activity for degradation of p53 in the absence of the proteasome, the S100 extract (S100 -0), which lacks most of the proteasome, was prepared from MA1 cells, and p53 was prepared by immunoprecipitation with mouse anti-p53 monoclonal antibody pAb421 and protein A-Sepharose FF beads. The p53 conjugated beads were incubated in S100 -0 at 30°C for 80 min in the presence of 1 mM CaCl 2 but in the absence of ATP and ubiquitin. Degradation of p53 was analyzed by Western blotting using rabbit anti-p53 polyclonal antibody FL393 after electrophoresis (Fig. 2C). Under these conditions, a small but significant fraction of p53 was degraded (lane 2). Calpastatin peptide (lane 3) and EGTA (lane 5) protected p53 from degradation significantly, but the presence of apyrase (lane 4) and EDTA (lane 6) had no effect on the protection. These results suggested that the degradation of p53 in MA1 cells occurs primarily by the ubiquitin-proteasome system, although calpain is partly involved.
The Amino Terminus and CR1 Mutants of E1A Are Unable to Modify the Ubiquitin-Proteasome Pathway to Stabilize p53- The presence of a very small amount of p53 in B12 and T5 cells after treatment with dexamethasone suggested that the E1A mutants containing a deletion in either the amino terminus or CR1 are unable to modify the ubiquitin-proteasome pathway so astosuppresstheproteolyticactivityforp53.Toseetheubiquitindependent proteolytic activity for degradation of p53, an aliquot of S10 -24 (50 g of protein) containing accumulated p53 was incubated with increasing amounts of B12 or T5 48 h extract (B12 S10 -48 or T5 S10 -48) in the presence and absence of ATP and ubiquitin. As a control, S10 -0 was substituted for these extracts (Fig. 3). Degradation of p53 was similarly analyzed by Western blotting. p53 in S10 -24 was degraded slightly by the addition of 1 ⁄5 of the amount (10 g) of protein in S10 -0, but the fraction of p53 degraded increased along with the increase in the amount of S10 -0 added (Fig. 3A,  lanes 2 and 4). No significant degradation was observed in the absence of ATP and ubiquitin (lane 1) and in the presence of ubiquitin alone (lane 3). The degradation occurred in the presence of only ATP (lane 2), presumably due to residual ubiquitin present in the extracts. The degradation of p53 occurred similarly by the addition of B12 S10 -48 or T5 S10 -48 (Fig. 3B). The degradation proceeded even faster than that caused by S10 -0, and most of the p53 was degraded by the addition of a 5-fold greater amount of protein in these extracts in the presence of ATP (lanes 2 and 4), but not in the absence of ATP (lanes 1 and 3). These results suggest that a component(s) of the ubiquitin-proteasome pathway responsible for degradation of p53 is inactivated or much reduced through the function of the amino terminus and CR1 of E1A.
The Ubiquitination Activity for p53 Is Reduced by E1A-To analyze the intermediates of p53 degradation and its proteolytic cleavage products, an aliquot of 25 g of protein in S100 -24, prepared from the MA1 S10 -24 by high speed centrifugation, was used as a source of p53 and incubated with increasing amounts of S100 -0 in the absence of added ATP and ubiquitin. The reaction was first performed at 37°C for 20 min, and aliquots of the reaction mixture were analyzed by SDS-PAGE. p53 and its cleavage products were detected by Western blotting using anti-p53 antibody FL-393. As shown in Fig. 4A, most of p53 remained uncleaved in the absence of CaCl 2 (lane 1). Two or three thin bands that migrated slightly faster than p53

FIG. 2. Degradation of p53 in MA1 cells occurs predominantly by the ubiquitin-proteasome pathway.
A, the p53 levels in MA1, B12, and T5 cells. The p53 levels in the S10 extracts (50 g of protein) prepared from MA1 cells before (0 h) and after (24 h) treatment with dexamethasone (dex) (lanes 1 and 2) and from B12 and T5 cells treated with dexamethasone for 48 h (lanes 3 and 4) were analyzed by Western blotting with anti-p53 monoclonal antibody Bp53-12 and HRP-conjugated anti-mouse-Ig antibody. B, effects of various protease inhibitors on degradation of p53. Aliquots of the MA1 24-h extract (S10 -24, 50 g of protein) containing accumulated p53 were incubated in the MA1 0-h extract (S10 -0, 250 g of protein) in the presence of 2 mM ATP and 0.4 g/l ubiquitin. The concentrations of proteasome and calpain inhibitors added are as follows: Z-LLnV-OH, 50 M; apyrase, 0.125 units/l; calpastatin peptide, 100 M, and EGTA, 5 mM. The amounts of p53 undegraded were analyzed by Western blotting with anti-p53 polyclonal antibody FL-393 and HRP-conjugated anti-rabbit IgG antibody. C, degradation of p53 in S100 -0 in the presence of CaCl 2 . Aliquots of immunoprecipitated p53 were incubated in the S100 extract prepared from MA1 cells (S100 -0; 500 g of protein) at 30°C for 80 min in the presence of 1 mM CaCl 2 and inhibitors as indicated above each lane. No ATP and ubiquitin were added. The amounts of p53 undegraded were analyzed as stated above.
were observed, irrespective of the amounts of S100 -0 added in the presence of CaCl 2 (lane 2). These species of proteins seem to be the fragments of p53 cleaved by calpain, since the addition of EGTA or calpastatin peptide (lanes 3 and 4) abolished the generation of these species of proteins. Calpain usually degrades its substrates to limited extents (42). Unexpectedly, a large proportion of p53 was shifted to the high molecular weight region in the presence of CaCl 2 (lanes 2 and 4). The shift was dependent on calcium, since the presence of 5 mM EGTA abolished the formation of the high molecular weight form of p53 almost completely (lane 3). The presence of calpastatin peptide had no effect on this shift (lane 4). In the high molecular weight region shown by a caret, ladder-like bands were observed, suggesting that these are polyubiquitinated p53. The polyubiquitination might be caused by residual ATP and ubiquitin present in the extracts. The amount of high molecular weight form of p53 increased slightly along with the increase in the amount of S100 -0 added, presumably due to the increase in the amounts of ATP and ubiquitin supplied in addition to the increase in the ubiquitination enzymes. The S100 -0-dependent formation of the high molecular weight form of p53 became more evident when the reaction proceeded at a slower rate at 30°C (Fig. 4A). When 1 ⁄5 of the amount of S100 -0 was mixed with S100 -24, a very small amount of p53 was converted to the high molecular weight form, while the addition of a 5-fold greater amount of S100 -0 resulted in the conversion of most of p53 in the presence of CaCl 2 (lanes 6 and 8). Under these conditions, a very small amount of p53 was fragmented, again irrespective of the amounts of S100 -0 added. These results suggest that the formation of the high molecular weight form of p53 is much more dependent on the reaction temperature than the fragmentation of p53 by calpain and occurs in a dose-dependent manner. The requirement of calcium for the ubiquitindependent degradation of cyclin B has been reported in the Xenopus extract (55). The conversion of p53 to the high molecular weight form was also analyzed with B12 and T5 S100 -48 prepared from B12 and T5 cells treated with dexamethasone for 48 h in place of S100 -0. As shown in Fig. 4B, the addition of a 5-fold greater amount of B12 S100 -48 or T5-S100 -48 to MA1 S100 -24 in the presence of CaCl 2 resulted in the conversion of most of p53 to the high molecular weight form. These results indicate that E1A mutants containing a deletion in either the amino terminus or CR1 are unable to repress the conversion of p53 to the high molecular weight form.
To confirm that the high molecular weight form of p53 is polyubiquitinated p53, p53 in S100 -24 (200 g of protein) was incubated with a 5-fold greater amount of S100 -0 and biotinylated ubiquitin in the absence of added ATP at 30°C for 40 min (Fig. 5). Calpastatin peptide and Z-LLnV-OH were added to minimize p53 degradation. The biotinylated ubiquitin conjugates of p53 were immunoprecipitated using mouse anti-p53 monoclonal antibodies (pAb421 and Bp53-12) and protein A-Sepharose FF beads and subjected to SDS-PAGE. The biotinylated ubiquitin conjugates of p53 were analyzed by Western blotting using HRP-conjugated streptavidin, which binds to biotin. As shown in Fig. 5A, streptavidin revealed biotinylated ubiquitin conjugates of p53 in a wide range, and a series of ladder-like bands was observed (lane 2). The formation of these conjugates was stimulated by CaCl 2 (lane 3). Residual amounts of ATP and ubiquitin present in S100 -0 and S100 -24 seemed to be sufficient for polyubiquitination of p53. The presence of EGTA, which chelates with calcium, greatly reduced the amounts of polyubiquitinated p53 formed (lanes 4 and 5), but the presence of an excess of CaCl 2 partially overcame the effect of EGTA (lane 6). When the reaction was performed with S100 -24 alone in the absence of CaCl 2 (lane 7), p53 was not polyubiquitinated at all; however, in the presence of CaCl 2 a much smaller yet significant amount of polyubiquitinated p53 was formed (lane 8). These results indicate that the ubiquitination activity for p53 is still present in S100 -24, although the activity is much reduced and strictly dependent on calcium. The amounts of unubiquitinated p53 remained were inversely correlated with the amounts of polyubiquitinated p53 (Fig. 6B,  lanes 2, 3, and 6).
The Ubiquitin-specific Enzyme Activity Responsible for Degradation of p53 Is Reduced by E1A-The recognition of a target protein for degradation by the ubiquitin-proteasome system is usually made by the degradation signal present in a target protein and a ubiquitin-conjugating enzyme (E2), which transfer the activated ubiquitin to a Lys residue of a target protein (31). A ubiquitin ligase (E3) is often required for the ubiquitination of the Lys residue. Both the E2 and E3 enzyme families consist of multiple members, and each member recognizes a set of target proteins (56) .   FIG. 3. Inability of E1A mutants to reduce the ubiquitin-dependent proteolytic activity for p53. A, the activity for degradation of p53 in MA1 cell extracts. The S10 extracts were prepared from MA1 cells (0 h) and MA1 cells treated with dexamethasone (dex) for 24 h. Aliquots of 50 g of protein in the MA1 24-h extract (S10 -24) were incubated with 0.2-5-fold amounts of protein in the MA1 0-h extract (S10 -0) at 30°C for 90 min as shown on the left in the presence (ϩ) or absence (Ϫ) of ATP (2 mM) and ubiquitin (0.4 g/l). Aliquots of the reaction mixtures were electrophoresed followed by Western blotting as described in Fig. 2. B, inability of E1A 12S⌬17-23 and E1A 12S⌬54 -69 to suppress the proteolytic activity for p53 cells. The MA1 24 h extract (50 g of protein) was incubated with 0.2-5-fold amounts of protein in either the B12 48 h extract or the T5 48 h extract. p53 was similarly analyzed by Western blotting.
FIG. 4. Calcium-dependent formation of the high molecular weight form of p53 in MA1, B12, and T5 cell extracts. Aliquots of 25 g of protein in MA1 S100 -24 containing accumulated p53 were incubated with 0.2-5-fold amounts of protein in MA1 S100 -0 at 37°C for 20 min or at 30°C for 80 min in the presence and absence of 1 mM CaCl 2 . Inhibitors of calpain, EGTA, and calpastatin peptide were added to lanes 3 and 4, respectively. Aliquots of the mixture containing 7.5 g of protein from MA1 S100 -24 were electrophoresed, and p53 was analyzed by Western blotting using anti-p53 antibody (FL-393) and HRPconjugated anti-rabbit-IgG antibody. B, the same amounts of MA1 S100 -24 were incubated with 5-fold amounts of protein in either B12 S100 -48 or T5 S100 -48 prepared from B12 and T5 cells treated with dexamethasone (dex) for 48 h at 30°C for 80 min. p53 was similarly analyzed.
To see whether the activity of a ubiquitin-conjugating enzyme(s) responsible for polyubiquitination of p53 is reduced in MA1 cells following induction of E1A, S100 -0 and S100 -24 were applied to ubiquitin-Sepharose columns in the presence of ATP. The bound proteins were sequentially eluted with high salt (1 M KCl), 20 mM AMP, 10 mM DTT, and pH 9 solution as shown in Fig. 6A. The bulk of proteins were collected in the flow-through fraction. Most of the E3 enzymes are recovered in the flow-through fraction, and most of the E2 enzymes are eluted in the AMP, DTT, and pH 9 fractions (57). The ubiquitinactivating enzyme E1 is eluted mostly in the AMP fraction. The AMP, DTT, and pH 9 fractions were combined and concentrated by ultrafiltration. S100 -24 (25 g of protein) containing accumulated p53 was incubated with aliquots of the concentrated eluates Ub eluate-0 and Ub eluate-24; the latter was prepared from S100 -24 and would have reduced enzyme activity involved in polyubiquitination of p53. The reaction was performed at 30°C for 40 min in the presence and absence of 1 mM CaCl 2 , and the mixture was subjected to SDS-PAGE followed by Western blotting. As shown in Fig. 6B, the addition of both eluates stimulated polyubiquitination of p53 in the presence of CaCl 2 ; however, the activity was much reduced in Ub eluate-24 as compared with Ub eluate-0. The polyubiquitination was completely abolished by the presence of EGTA but not by the presence of calpastatin peptide. Polyubiquitination of p53 scarcely occurred in S100 -24 alone even in the presence of CaCl 2 . These results suggest that a component(s) of the ubiquitin-proteasome pathway, presumably a member of the E2 family, required for p53 polyubiquitination is altered by E1A either in the level of its expression or in its activity so as to stabilize p53.
Stabilization of p53 in MA1 Cells Accompanies the Increase in the Mdm2 Level-The stabilization of p53 has been shown to be regulated by Mdm2, a nuclear phosphoprotein that binds to p53 through its amino terminus. p53 bound by Mdm2 is targeted for degradation by the ubiquitin-proteasome pathway, and the pathway is continuously operating in cells to lower the levels of p53 (36,41). In reverse, the expression of Mdm2 is transcriptionally activated by p53 (38,39). To see the involvement of Mdm2 in the E1A-induced stabilization of p53, the levels of Mdm2 and E1A in MA1 cells were analyzed by Western blotting after treatment with dexamethasone. As shown in Fig. 7, A and C, the level of p53 increased linearly after induction of E1A, reaching a plateau at 16 -20 h. The level of Mdm2 began to increase after p53 reached its maximal level, suggesting that the transcriptional activation of the mdm2 gene by p53 also occurs in MA1 cells after p53 stabilization. The level of p53, however, did not decline after maximal elevation of Mdm2. The high level of p53 was maintained throughout the apoptotic process at least until 72 h as previously shown (45). The low level of p53 in untreated MA1 cells at 0 h (lane 1), however, was maintained primarily by the ubiquitin-proteasome system, since the addition of Z-LLnV-OH to the culture medium elevated the p53 level to almost the maximum (lane 8). The very low level of Mdm2 in the untreat cells suggests that Mdm2 may not be involved in lowering the p53 level and that E1A antagonizes the ubiquitin-proteasome pathway to reduce the activity to target p53 for degradation. DISCUSSION Stabilization of p53 seems to be one of the cellular responses to external stress signals that cause DNA damage such as ionizing radiation (1)(2)(3)(4)(5) and viral infection (58 -61). Elevation of the p53 level by DNA-damaging agents often results in the arrest of cell cycle progression in G 1 , preventing the damaged cells to enter S phase (7,8). p53 acts therefore as a guardian to maintain the genomic integrity. Alternatively, elevation of the wild type p53 level in rodent cells by the adenovirus E1A results in the induction of apoptosis (19,20), which might be a cellular defense mechanism to prevent virus growth.
The cell line MA1 used in the present study expresses adenovirus E1A 12 S in response to dexamethasone and elicits apoptosis. Following induction of E1A, the level of p53 increased more than 10-fold within 24 h. The accumulated p53 was degraded efficiently in vitro in the S10 -0 extract prepared from MA1 cells, but not in the S10 -24 extract prepared from MA1 cells after treatment with dexamethasone for 24 h. The degradation was protected almost completely by the presence of Z-LLnV-OH or apyrase, both of which inhibit ubiquitin-proteasomedependent proteolysis, but protected only weakly by the presence of EGTA or calpastatin peptide, which inhibits calpain protease activity (Fig. 2). The accumulation of p53 was inversely correlated with the extent of polyubiquitination. The ubiquitination activity for p53 in MA1 cells was much reduced after induction of E1A, but the activity in B12 and T5 cells, which express E1A mutants having a deletion in the amino terminus or CR1 in response to dexamethasone, was unchanged (Figs. 4 and 5). No p53 accumulation was observed in these cells. Since these regions correspond to the p300 binding site, interaction of E1A with p300 might be involved in a modification of the ubiquitin proteasome pathway so as to reduce p53 breakdown.
Polyubiquitination of p53 in S100 -0, prepared from the S10 -0 extract by high speed centrifugation and devoid of most of the proteasome, was enhanced greatly by CaCl 2 . Similar requirement of calcium for ubiquitin-dependent degradation of cyclin B has been reported in the Xenopus extract (55). In the extract prepared from metaphase II-arrested Xenopus eggs, FIG. 5. Calcium-dependent polyubiquitination of p53 in MA1 S100 extract. Aliquots of 0.2 mg of protein in S100 -24 were incubated with 1 mg of protein in S100 -0 and 0.2 g/l biotinylated ubiquitin at 30°C for 30 min in the presence and absence of CaCl 2 and EGTA as indicated above each lane. In lanes 7 and 8, S100 -24 alone was similarly incubated. After the reaction, p53 was immunoprecipitated with 1 l each of the anti-p53 monoclonal antibodies pAb421 and Bp53-12 and 2.5 l of protein A-Sepharose FF beads. The precipitates were analyzed by Western blotting. A, polyubiquitinated p53 was detected with HRPconjugated streptavidin. B, unprocessed p53 was detected with anti-p53 polyclonal antibody FL-393 and HRP-conjugated anti-rabbit IgG antibody. degradation of cyclin B occurs at micromolar free Ca 2ϩ concentration in the presence of calpastatin, an inhibitor of calpain. When free Ca 2ϩ is raised to millimolar range, cyclin B is also degraded by calpain. In the MA1 S10 -0 extract, however, the reduction in Ca 2ϩ concentration from 1 to 0.1 mM resulted in the decrease in the extent of p53 polyubiquitination to less than a half of the original in the presence of calpastatin peptide. A calcium-activated protease, calpain, is also involved in p53 degradation, but its contribution was minor, and the calpain activity was not altered after E1A expression.
The ubiquitin-specific enzyme activity involved in stabilization of p53 by E1A was analyzed by ubiquitin affinity chromatography. The enzymes in S100 -0 and S100 -24 prepared from MA1 cells before and after treatment with dexamethasone were applied to ubiquitin affinity columns, and bound enzymes were eluted by AMP, which competes with ATP, by increased concentration of a thiol compound, DTT, and by raising the pH. These fractions combined contain E1, most of the E2s and a portion of the E3s. The addition of the combined eluate to the in vitro p53 polyubiquitination reaction mixture (Fig. 6) revealed that the polyubiquitination was enhanced greatly when the eluate prepared from S100 -0 was added, but this activity was much reduced in the eluate prepared from S100 -24. This result suggests that ubiquitination of p53 might be suppressed by reduction either in the activity or in the expression level of a component in the ubiquitin-proteasome pathway, presumably a member of the ubiquitin-conjugating enzyme E2 family. In the case of ubiquitination of p53 by human papillomavirus E6, E6 first associates with E6-AP, a member of E3 ubiquitin ligases, and the resulting complex interacts with p53 to induce the ubiquitination (33). A cellular fraction thought to contain an E2 enzyme is required for this ubiquitination. A human E2, UbcH5 can function in the E6/E6-AP-mediated ubiquitination of p53, and this function could be replaced with Ubc8 from Arabidopsis thaliana (62). The expression level of UbcH5 in MA1 cells, so far tested by Northern blotting, was unchanged after treatment with dexamethasone (data not shown). A set of different members of E2s may be involved in polyubiquitination of p53, depending on a respective partner of the E3 family. We recently found that a species of E2 enzyme presumably involved in degradation of topoisomerase II␣ is induced in MA1 cells after E1A expression (47).
The Mdm2 oncoprotein is a potent inhibitor of p53 and inhibits the p53-dependent transcription of target genes through binding to the NH 2 -terminal transactivation domain (63). In addition, Mdm2 promotes the proteasome-dependent degradation of p53 through binding to it and thus terminates the p53 response to external signals (36,37). On the other hand, p53 activates the expression of the mdm2 gene (38,39). In MA1 cells, the level of p53 increased after expression of E1A, and the elevated p53, in turn, promotes the expression of Mdm2 (Fig.  7). A high level of p53, however, was maintained irreversibly until the late stage of the apoptotic process (45). These results suggest the possibility that either p53 or Mdm2 might be modified so as to make p53 insensitive to the ubiquitin-proteasome pathway. It has been recently reported that DNA-damaging FIG. 6. Fractionation of p53 ubiquitination activity in MA1 cells by affinity chromatography. A, ubiquitin-Sepharose chromatography of S100 -0 and S100 -24. Aliquots of 15 mg of protein were applied to the columns. The flow-through fraction and KCl, AMP, DTT, and pH 9 eluates were collected. The elution profile of S100 -0 proteins is shown. The profile of S100 -24 proteins is essentially the same. B, stimulation of p53 polyubiquitination by the eluates. The AMP, DTT, and pH 9 eluates were combined, and aliquots of 14 ng of protein prepared from S100 -0 and S100 -24 were added to S100 -24 (25 g of protein) containing accumulated p53 and incubated at 30°C for 40 min in the absence of added ATP and ubiquitin. CaCl 2 , EGTA, and calpastatin peptide were added as indicated above each lane. Aliquots of a third of the mixture were subjected to SDS-PAGE, and polyubiquitination of p53 was analyzed by Western blotting using anti-p53 antibody (FL-393) and HRP-conjugated anti-rabbit IgG antibody. agents induce stabilization of p53 through the activation of DNA-dependent protein kinase, which phosphorylates serine residues in the amino-terminal region of p53 and impairs the ability of Mdm2 to inhibit p53-dependent transactivation (64). However, the involvement of modification of p53 in its stabilization by E1A is unlikely, since p53 stabilized in MA1 cells was susceptible to degradation in MA1 0-h extract (Fig. 2), and the level of Mdm2 in MA1 cells was very low (Fig. 7).
The E1A-binding protein p300 and its related protein CBP have properties with a transcriptional coactivator (65). They do not bind to DNA but are recruited to promoters by association with transcription factors such as CREB, c-Jun, c-Fos, and c-Myb (66). E1A dissociates the complexes formed with p300/ CBP and the transcription factors and suppresses their negative regulation for cell cycle progression. The E1A mutants containing a deletion in the amino terminus or CR1 lose the ability to bind p300 and to suppress the negative role of p300/ CBP in cell proliferation. The transactivating function of p300/ CBP complexed with a transcription factor such as CREB or c-Jun might result in activation of the gene encoding an enzyme involved in the ubiquitin-proteasome pathway, which targets p53 for degradation, and E1A may inhibit this transactivation through binding to p300. , and E1A (C) were estimated by Western blotting using rabbit anti-p53 polyclonal antibody FL-393, mouse anti-Mdm2 monoclonal antibody SMP14, and mouse anti-E1A monoclonal antibody M73, respectively. HRP-conjugated anti-rabbit IgG and HRP-conjugated anti-mouse Ig antibodies were used as secondary antibodies. In lanes 8 and 9, 50 M Z-LLnV-OH was added to the culture medium 1 h before the cell harvest.