The Acetylase Activity of p300 Is Dispensable for MDM2 Stabilization*

It has been shown that p300 binds to MDM2 and leads to down-regulation of the p53 function. However, it remains unclear whether the acetylase activity of p300 is necessary for regulating MDM2 stability. In this study, we address this issue. First, p300 did not acetylate MDM2 in solution and in cells. Second, overexpression of p300 in cells increased the level of the MDM2 protein but not its mRNA. Similarly, the acetylase-defective p300 AT2 mutant stabilized the MDM2 protein as well. Consistently, the deacetylase inhibitor, trichostatin A, did not significantly affect the half-life of the endogenous MDM2 protein, whereas p300 enhanced the half-life of MDM2. Finally, both wild type and acetylase-defective mutant p300 proteins associated with MDM2 in nuclear body-like structures where MDM2 might be protected from proteasomal degradation. Thus, these results suggest that p300 appears to stabilize MDM2 by retaining this protein in a specific nuclear structure rather than by acetylating it.

p300 and CBP 1 are the two acetylase-possessing transcriptional co-activators that have been shown to mediate transcription by many different transcriptional activators (1). They also regulate the function of the tumor suppressor p53 protein and its homologs (2)(3)(4)(5)(6)(7)(8). It may be this regulation that contributes to the role of p300/CBP in cell growth control and neoplasia (Ref. 12 and references therein). Although p300 and CBP, encoded by two different genes (9 -11), share a significant homology in their functional domains with similar biochemical functions, their roles in development and neoplasia are not redundant (12,13), suggesting that each of them is essential for cell growth control. Importantly, the intrinsic acetylase activity of these co-activators is crucial for their regulatory function (14,15). For instance, by acetylating p53, p300/CBP stimulates its ability to bind to DNA in a sequence-specific fashion in vitro (14), stabilizes its protein (16), and enhances its transcription in cells (2-4, 17, 18). Indeed, DNA damage signals stimulate p53 acetylation by p300 (15). This effect can be reversed by deacetylases such as histone deacetylase 1 (HDAC1) and Sir2␣ (19 -22). Thus, up-regulating p53 activity is an important function for p300 and CBP. p300 has also been shown to interact with the p53 negative regulator MDM2 (5,23), whose oncogenic activity is mainly attributed to its ability to down-regulate p53 function (24). MDM2 possesses an E3-like ubiquitin ligase activity and thus ubiquitinates p53 (25), leading to its degradation (26,27). MDM2 also binds to p53 when this transcriptional activator resides at the promoter regions of its target genes (28,29), blocking p53-dependent transcription (24,30). Because MDM2 itself is transcriptionally induced by p53, it serves as a negative feedback regulator of this activator (31,32). Intriguingly, by binding to MDM2, p300 assists this protein in degrading p53 (23). Though seemingly contradictory to the fact that p53 is activated by p300, as discussed above, the enhancement of MDM2-mediated p53 degradation by p300 appears to be correlated with the observation that MDM2 forms a ternary complex with p300 and p53 (33) and inhibits p300-catalyzed p53 acetylation (16,33). Reasonably, inhibition of p300-catalyzed p53 acetylation would favor MDM2-mediated p53 ubiquitination and, thus, degradation, because these enzymes target the same set of lysine residues at the C-terminal domain of p53 (14,15,34). Hence, the MDM2-p300 interaction leads to down-regulation of p53 function.
However, it is unclear whether p300 requires its acetylase activity to assist MDM2 in degrading p53. We recently found that overexpression of p300 increased the level of MDM2 (Fig.  2). This phenomenon was also reported by others (35). Furthermore, our study as presented here shows that p300 does not need its acetylase activity to stabilize MDM2. First, p300 apparently did not acetylate MDM2 in solution and in cells. Second, overexpression of p300 in cells increased the level of the MDM2 protein, but not its mRNA, and led to elongation of the half-life of MDM2. Similarly, the acetylase-defective p300 AT2 mutant stabilized the MDM2 protein as well. Consistently, the deacetylase inhibitor, trichostatin A (TSA), did not significantly affect the half-life of the endogenous MDM2 protein. Moreover, both wild-type and acetylase-defective mutant p300 proteins recruited MDM2 to nuclear body-like structures where MDM2 might be protected from targeting by nuclear proteosomes. Thus, these results demonstrate that p300 stabilizes MDM2 by retaining this protein in a specific nuclear structure rather than by acetylating it.

EXPERIMENTAL PROCEDURES
Cell Culture-Human lung small cell carcinoma H1299 cells, human embryonic kidney epithelial 293 cells, human astrocytoma SJSA cells (American Type Culture Collection) and mouse p53 null embryonic fibroblast (MEF) cells were cultured as previously described (28,52). * This work was supported by National Institutes of Health Grants CA095441 and CA93614 (to H. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
¶ Supported by National Institutes of Health Grant DK42748 and a National Institutes of Health postdoctoral fellowship.
Transient Transfection and Western Blot Analysis-H1299 cells (60% confluence in a 60-mm plate) were transfected with pCDNA3-HA-MDM2 (0.5 g) alone or together with pCDNA3-FLAG-p300 or pCDNA3-FLAG-p300AT2 (see the legends for Figs. 1 and 2 for the amount of p300 plasmids used). 36 h post transfection, cells were harvested for preparation of whole cell lysates. Whole cell lysates containing 100 g of protein were directly loaded onto an SDS gel, and proteins were detected by ECL reagents (Bio-Rad) after Western blotting using antibodies, as indicated in the legends for Figs. 1-4.
In Vivo Acetylation Labeling-The acetylation in cells (60% confluence in a 10-cm plate) was performed according to the published method with minor modifications (53). Human H1299 cells were transfected with the MDM2 expression plasmid alone or together with the p300 or p300-AT2 plasmid. 30 h after transfection, the cells were pretreated with cycloheximide (CHX) (5 g/ml, Sigma) for 1 h and then incubated in Dulbecco's modified Eagle's medium containing 0.5 mCi/ml [ 3 H]sodium acetate (Amersham Biosciences) and 5 g/ml CHX for 1 h at 37°C before cell lysates were prepared. Immunoprecipitation with the lysates and antibodies against MDM2 was carried out, and immunoprecipitated proteins were analyzed by SDS-PAGE and detected by autoradiography for 2 months. MDM2 was detected by Western blotting.
Northern Blot Analysis-Northern blot analysis was conducted as described (54). Total RNA was isolated from transfected H1299 cells using the Trizol reagent (Invitrogen). 15 g of RNA were loaded onto a 1.5% agarose gel and transferred to a nitrocellulose membrane. The membrane was exposed to UV light in a UV cross-linker (Fisher) and incubated with 32 P-labeled cDNA probes encoding human MDM2 at 42°C overnight. After washing with 4ϫ SSC once and 1ϫ SSC twice, the blot was exposed to x-ray film overnight.
Analysis of MDM2 Half-life in Cells-H1299 cells were transfected with the MDM2 plasmid alone or together with the p300 or p300-AT2 expression plasmid as described above. 30 h after transfection, the cells were treated with TSA or Me 2 SO for 6 h. Separately, mouse p53 null embryonic fibroblast cells, human HEK, and H1299 cells were also treated with TSA or Me 2 SO for 6 h. Then, all the cells were treated with 50 g/ml CHX and harvested at different time points afterward for the preparation of cell lysates. Lysates containing 100 g of proteins were loaded directly onto an SDS gel for Western blot analysis with antibodies against MDM2. The level of MDM2 from the transfected cells was quantified by scanning the blot and using the Adobe Photoshop pro-gram and plotted with the Cricket Graph (Fig. 3A). These experiments were repeated twice.
Construction and Use of p300 Adenoviruses-A wild-type human p300 cDNA and the AT2 mutant lacking histone acetyltransferase activity (7,37) were used to generate recombinant adenoviruses with the adenase Vector Kit (Quantum Biotechnologies). Each p300 cDNA was first cloned into a modified shuttle plasmid containing a tetracycline-inducible promoter using standard molecular biological manipulations. Each pShuttle:TetR:p300 recombinant plasmid was linearized and recombined with the pedally plasmid in Escherichia coli strain BJ5183 as described by the manufacturer. Five micrograms of purified pAdEasy:TetR:p300 and pAdEasy:TetR:p300AT2 plasmids were then digested with PacI endonuclease and transfected into low passage HEK293 cells using FuGENE-6 at a ratio of 1 g of plasmid DNA per 5 l of FuGENE 6. After 2 days, the transfected cells were replated in 6-well dishes, and growth medium containing 1.25% Seaplaque Agarose was added 1 day later to promote the formation of recombinant viral plaques, as described in the manufacturer's protocol. Approximately 17 days later, individual plaques were picked, amplified in the HEK293 cells, and purified over a discontinuous CsCl gradient as outlined in the supplier's protocol.
Cells were co-infected with approximately equal amounts of two recombinant adenoviruses, one encoding a tetracycline-inhibited transactivator (Ad:tTA), and the other either Ad:GFP, Ad:FLAG-p300, or Ad:FLAG-p300-AT2. Cells at ϳ50% confluent density were washed once with PBS and then incubated for 3 h at 37°C in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and an equal mixture of two viruses at a multiplicity of infection (m.o.i.) that depends upon cell type. The medium was then replaced with fresh growth medium, and the cells were incubated until confluence before being used for experiments.
Immunofluorescent Staining and Fluorescent Microscopic Analysis-Human astrocytoma SJSA cells containing detectable MDM2 were coinfected with the recombinant adenovirus encoding a tetracycline-inhibited transactivator (Ad:tTA) and the recombinant adenovirus harboring GFP, FLAG-p300, or FLAG-p300AT2 at an m.o.i. of 1000. 48 h after infection, cells were fixed for immunofluorescent staining with monoclonal anti-FLAG antibodies and polyclonal anti-MDM2 antibodies as well as DNA staining with 4Ј,6-diamidino-2-phenylindole (DAPI). The Alexa Fluor 488 (green) goat anti-mouse antibody and the Alexa Fluor 546 (red) goat anti-rabbit antibody (Molecular Probes, Eugene, Oregon) were used for FLAG and MDM2, respectively. Stained cells were analyzed under the Zeiss Axiovert 25 fluorescent microscope. The infected cells were also harvested for Western blot analysis using antibodies against p300, MDM2, and ␣-tubulin.

RESULTS AND DISCUSSION
p300 Does Not Acetylate MDM2 in Vitro and in Cells-The interaction between p300 and MDM2 (33) prompted us to examine whether MDM2 is also a substrate for this acetylase. To this end, purified p300, MDM2, and GST-MDM2 were used in a set of in vitro acetylation assays with purified PCAF-HAT fragment and p53 as controls. [ 3 H]acetyl-CoA was used as a substrate, and acetylated proteins were detected by autoradiography. As shown in Fig. 1A, both p300 and the PCAF-HAT fragment acetylated p53 (lanes 2 and 3) as expected (33,36). However, none of these acetylases were able to acetylate either MDM2 or GST-MDM2 (lanes 4 -9). To further confirm this result, MDM2 was titrated in the same acetylation reaction setting with increasing amounts of p53 as positive controls. Again, p53 was acetylated by p300, but not by the p300 AT2 mutant (7), in a dose-dependent fashion (lanes 1-5 of Fig. 1B). Although a significant amount of the MDM2 protein was used in the reaction, no acetylated MDM2 was detected even after an extended exposure (lanes 6 -10 of Fig. 1B). Thus, these results demonstrate that p300, like PCAF (28), does not acetylate MDM2 in solution.
To determine whether this observation is also true in cells, we performed an in vivo acetylation assay by co-introducing p300 with MDM2 into human small cell lung carcinoma H1299 cells that are free of p53 proteins. Cells were labeled with [ 3 H]acetic acid 36 h after transfection and harvested 2 h after labeling in the presence of the deacetylase inhibitor TSA. Al-p300 Stabilizes MDM2 without Its Acetylase Activity though both MDM2 and p300 were well expressed, we were unable to detect acetylated MDM2 in the presence of the exogenous p300 (Fig. 1C). Taken together, these data demonstrate that p300 is unable to acetylate MDM2 in solution and in cells.
Wild Type as Well as Acetylase-defective Mutant p300 Stabilize MDM2 in Cells-We next tested whether p300 affects the level of MDM2 in cells. To this end, H1299 cells were transfected with plasmids encoding HA-MDM2 and p300 alone or together. Protein levels were detected by Western blot analysis using antibodies against p300 and HA, respectively. As shown in Fig. 2A, the MDM2 level increased dramatically when MDM2 was co-expressed with p300. This increase was in a p300 dose-dependent manner, but it was not due to the in-crease of the mdm2 gene transcription by p300, as the mRNA level of the mdm2 gene did not change significantly (two bottom panels of Fig. 2A) no matter whether the p300 level was high or low (top panel). Surprisingly, the acetylase-defective p300AT2 mutant (37) was also able to induce the level of MDM2 in a dose-dependent manner (Fig. 2B). Consistent with this result, the deacetylase inhibitor TSA did not significantly affect the half-life of either exogenous (Figs. 2C or 3) or endogenous MDM2 in H1299 cells (Figs. 2D). This finding was also true in HEK293, p53 null MEF, and Saos2 cells (data not shown). In addition, the NAD-dependent deacetylase inhibitor vitamin B3 (nicotinic acid, 5 mM) was without effect on the half-life of MDM2 either (data not shown). Although MDM2 level increased in the presence of TSA, this increase was inconsistent, as it was not reproducible in H1299 and several other cell lines including HEK293 and p53 null MEF cells (Fig. 2D, and data not shown). The stabilization of MDM2 by p300 appeared to require a direct association between these proteins, as the MDM2 level did not significantly increase when two p300 deletion mutants that lack the MDM2-binding domain (amino acids 350 -450) (5, 23) were independently co-introduced with MDM2 into H1299 cells (Fig. 2E). These results indicate that p300 does not need its acetylase activity to increase the level of MDM2.
P300 Extends the Half-life of MDM2, Which Is Not Affected by a Deacetylase Inhibitor-To test whether the increase of MDM2 level is due to its stabilization by p300, the turnover rate of exogenous MDM2 proteins was determined. H1299 cells were transfected with the MDM2 expression plasmid alone or together with the p300 expression vector. 30 h post transfection, cells were treated with TSA or 50% ethanol (used for preparation of TSA). 3 h after TSA treatment, cells were treated with the translation inhibitor cycloheximide to stop the synthesis of nascent MDM2 proteins and harvested at different time points immediately after treatment. As shown in Fig. 3, the half-life of MDM2 was about 15 min in the absence of exogenous p300, whereas it extended to ϳ25 min in the presence of p300. Again, acetylation may not be involved in this extension, as the deacetylase inhibitor TSA did not extend the turnover rate of MDM2 further (Fig. 3). Although the half-life of the endogenous MDM2 was a little bit longer than that of the exogenous MDM2 (compare Fig. 2, panel C with panel D), overall it was no longer than 20 min, as tested in three different cell lines, including H1299, HEK293, and MEF cells (Fig. 2, C and D). 2 These results suggest that p300 increases the MDM2 level by stabilizing this protein and that this stabilization is not through acetylation modification of MDM2.
Both Wild Type and Mutant p300s Can Associate with MDM2 in Nuclear Body-like Structures-Because p300 does not acetylate MDM2, how may p300 stabilize MDM2? There are two possibilities. First, p300 may retain MDM2 in certain subcellular structures in order to prevent this protein from being targeted by the proteosome machinery. Alternatively, p300 may inhibit MDM2 auto-ubiquitination by binding to this protein. We tested the first hypothesis by infecting human astrocytoma SJSA cells, which contain a high level of MDM2 (Fig. 4), with the recombinant adenovirus encoding GFP, FLAG-p300, or FLAG-p300AT2. 24 h after infection, cells were either harvested for Western blot to detect p300 and MDM2 or analyzed by immunofluorescent staining with antibodies against MDM2 or FLAG. Again, overexpression of both the wild type and mutant p300 proteins, but not GFP, induced the level of the endogenous MDM2 (Fig. 4A). Immunofluorescent staining revealed that p300 and p300-AT2 proteins co-localized 2 S. X. Zeng, Y. Jin, and H. Lu, unpublished observations. FIG. 1. p300 does not acetylate MDM2. A, the in vitro acetylation assay. 50 ng of p53 and 300 ng of MDM2 and GST-MDM2 or 200 ng of p300 and PCAF-HAT as well as 0.2 Ci of [ 3 H]acetyl-CoA were used in this assay. This autoradiogram was exposed for 1 week. B, titration of substrates in acetylation reaction. The same acetylation as that described above was conducted except that different amounts of p53 (10, 30, and 60 ng) and MDM2 (50, 100, and 200 ng) were used here as substrates. Acetylated proteins were detected by autoradiography (top panel), and p53 and MDM2 were detected by Western blot (WB; middle and bottom panels). C, in vivo acetylation labeling. In vivo acetylation labeling was carried out as described under "Experimental Procedures." MDM2 was immunoprecipitated with the anti-MDM2 antibody and analyzed by SDS-PAGE. The bottom panel presents the result of autoradiography after 2 months of exposure of the gel to x-ray film, and the middle panel shows the MDM2 level. The top panel shows the levels of FLAG-p300 and FLAG-p300AT2.
p300 Stabilizes MDM2 without Its Acetylase Activity with endogenous MDM2 in the nucleus, whereas GFP defused throughout the whole cell (Fig. 4B). Interestingly, when examining individual cells under a higher magnification lens (ϫ32), we found that MDM2 co-localized exactly with p300 in nuclear body-like structures, forming focus-like spots (Fig. 4C). By striking contrast, MDM2 in the absence of exogenous p300 apparently did not display such patterns and was instead evenly distributed in the nucleus (Fig. 4C, see arrows and also see the GFP-MDM2 image). Also, the focus-like spots of MDM2 were not seen when SJSA cells were exposed to ␥-irradiation (Fig. 4D), suggesting that the stabilization of MDM2 by p300 and the co-localization of MDM2 with p300 were not due to a nonspecific and stressful effect caused by overexpressing p300. Although we observed a marginal increase of MDM2 3 h after ␥-irradiation of SJSA cells, this increase was caused by endogenous and activated p53 in response to the physical stress (Fig.  4D, top). Thus, in combination with the above data, these results suggest that p300 may protect MDM2 from proteosomemediated degradation by keeping MDM2 in nuclear body-like subnuclear structures.
The study we described here demonstrates that p300 can stabilize MDM2 by binding directly to this protein and confining it in the focus-like structure, thus protecting it from targeting by proteosomes. This protection does not appear to require acetylation, as the acetylase-defective p300AT2 mutant displayed the same function as the wild type p300 to stabilize with TSA. E, wild type p300 but not the MDM2 binding-defective deletion mutant p300 induces MDM2 level. H1299 cells were transfected with the MDM2 expression vector (0.5 g) alone or together with pCDNA3 plasmids encoding p300 (2 g), Gal-⌬242-1737 (1ϫ ϭ 1 g; 2ϫ ϭ 2 g) or Gal-1945-2424 (1ϫ ϭ 1 g; 2ϫ ϭ 2 g). Lysates containing 100 g of proteins were loaded onto an SDS gel, and WB was carried out to detect MDM2. Bottom panel shows a protein loading control.
MDM2. The latter observation is not in favor of the idea that p300 might inhibit MDM2 auto-ubiquitination by chemically modifying the lysines of MDM2 that are potential target residues for auto-ubiquitination. Consistently, others have shown that p300 still enhances MDM2-mediated p53 ubiquitination (35) and degradation (23,35), suggesting that p300 may not inhibit the E3-like ubiquitin ligase activity of MDM2. Nevertheless, it still remains to be clarified whether p300 may prevent MDM2 auto-ubiquitination by associating with this protein and causing a steric hindrance without affecting ubiquitination of other substrates. Noticeably, p300 only ex-tended the turnover rate of MDM2 by ϳ10 min. Two possible interpretations of this observation are either that the association of p300 with MDM2 may be weak and dynamic so that the latter undergoes active self-destruction in cells, or that p300 may require other proteins to further stabilize MDM2.
There are two other mammalian cellular proteins that have been shown to stabilize MDM2 by direct association. They are p14arf (p19arf for mouse) (38 -41) and MDMx (42,43). Similar to p300, they all are stable proteins and not directly regulated by p53. However, the outcomes of the associations of these proteins with MDM2 and the mechanisms by which they reg- FIG. 4. p300 associates with MDM2 in nuclear focus-like structures. A, induction of endogenous MDM2 by p300 and p300AT2 but not by GFP. SJSA cells were infected with recombinant adenovirus encoding GFP, p300, or p300AT2. 48 h after infection, cells (100 g protein) were harvested for Western blot (WB) analysis of p300 and MDM2 with ␣-tubulin as a control. B and C, immunofluorescent staining analyses. SJSA cells were infected as described for panel A. Cells were fixed for immunofluorescent staining as described under "Experimental Procedures." Viruses encoding GFP, p300, and p300AT2 used for infection are indicated at the top. Antibodies used for staining are indicated at the left. Images were taken under a fluorescent microscope with a lens of either 10ϫ (B) or 32ϫ magnification. C, arrows indicate a comparison between the cells that expressed high and low levels of exogenous p300 proteins. D, SJSA cells were irradiated with 7 gray of ␥-ray and harvested for Western blot analysis with antibodies against MDM2, p53, and ␣-tubulin (top) or fixed for immunofluorescent staining with anti-MDM2 antibodies and 4Ј,6-diamidino-2-phenylindole (DAPI; bottom) at 0, 1, and 3 h intervals, respectively. No single focus-like spot of MDM2 was identified in the nucleus after irradiation. Representative images taken under the 32ϫ magnification lens is shown in this figure. ulate MDM2 are distinct. P14arf binds to MDM2, retains it in the nucleolus (39,40), and inhibits its ubiquitin ligase activity (44). Thus, p14arf prevents MDM2-mediated p53 degradation and subsequently activates p53, although it also stabilizes MDM2. In contrast, MDMx binds to MDM2 through their Ring finger domains (45,46) and inhibits its auto-ubiquitination without affecting its ubiquitin ligase activity on p53, leading to p53 inactivation partially by stabilizing MDM2 (38,42,45). Normally, MDMx locates in the cytoplasm and is transported to the nucleus with the help of MDM2 (47,48). Thus, these studies suggest that MDM2 and MDMx act cooperatively to inactivate p53. This notion is substantially supported by the MDM2-p53 and the recent MDMx-p53 double knockout studies (49 -51). Similarly to MDMx, but unlike p19arf, p300 appears to enhance p53 degradation when co-expressed with MDM2 (23,48). However, p300 alone can also stabilize p53 by acetylating this protein and possibly by preventing MDM2-mediated p53 ubiquitination, as p300 and MDM2 target the similar lysine residues of p53 (14,16,34). How p300 regulates p53 in two opposite ways may depend upon the status or existence of MDM2. It is possible that p300 activates p53 at the very early stage of cellular response to DNA damage when MDM2 level is low, whereas at the late stage of this response it may stabilize newly synthesized MDM2 that is transcriptionally activated by p53 and thus assists this ubiquitin ligase to fulfill its negative feedback regulation of p53. This hypothesis is in agreement with the fact that the half-life (10 -15 min) of MDM2 is shorter than that (30 min) of p53; thus the former needs to be stabilized, such as by p300, in order to control the latter. Whether and how p14arf and MDMx may be involved in these aspects by interplaying with p300 to modulate p53 stability and function would be an interesting and important question to address for a better understanding of the p53 regulatory network. Also, whether CBP, similarly to p300, functions to stabilize MDM2 remains to be investigated