Report

Polyubiquitination of p53 by a Ubiquitin Ligase Activity of p300

See allHide authors and affiliations

Science  11 Apr 2003:
Vol. 300, Issue 5617, pp. 342-344
DOI: 10.1126/science.1080386

Abstract

Rapid turnover of the tumor suppressor protein p53 requires the MDM2 ubiquitin ligase, and both interact with p300–CREB-binding protein transcriptional coactivator proteins. p53 is stabilized by the binding of p300 to the oncoprotein E1A, suggesting that p300 regulates p53 degradation. Purified p300 exhibited intrinsic ubiquitin ligase activity that was inhibited by E1A. In vitro, p300 with MDM2 catalyzed p53 polyubiquitination, whereas MDM2 catalyzed p53 monoubiquitination. E1A expression caused a decrease in polyubiquitinated but not monoubiquitinated p53 in cells. Thus, generation of the polyubiquitinated forms of p53 that are targeted for proteasome degradation requires the intrinsic ubiquitin ligase activities of MDM2 and p300.

Degradation of the tumor suppressor p53 by the proteasome is essential for its physiologic regulation and requires the ubiquitin ligase MDM2 (1–3). In vitro, MDM2 catalyzes the addition of single ubiquitin (Ub) moieties (monoubiquitination) to a cluster of six COOH-terminal lysines in p53 (4, 5). Polyubiquitinated forms of p53, bearing polymerized Ub chains, are a likely signal for proteasome degradation (6) and are detected in cells when proteasome activity is inhibited (7). MDM2 does not polyubiquitinate p53 in vitro (4), suggesting that the activity of another ubiquitin ligase, e.g., an E4 subtype (8), could generate this modification. A role for the transcriptional coactivator p300 in p53 degradation is suggested by the observation that adenovirus E1A protein stabilizes p53, possibly by affecting its ubiquitination. This effect may depend partly on an E1A interaction with p300 (9–11). p300 and its homolog CREB-binding protein (CBP) are coactivators that acetylate and coactivate p53 after DNA damage (12). In unstressed cells, p300 interacts with MDM2 and is an important element in MDM2-dependent p53 turnover (13).

To determine whether the effect of p300 on p53 turnover is mediated by the modulation of p53 ubiquitination, the intrinsic ubiquitin ligase (E3) activity (14) of p300 was assessed. Purified, insect cell–derived p300 (15) was incubated with ubiquitin reaction components (URC) Ub, E1 (Ub-activating enzyme), and ubch5a (Ub-conjugating enzyme) (15). The ability of p300 to catalyze the formation of multi-Ub containing chains in such an autoubiquitination reaction (16), a hallmark of E3 activity (17), was determined by Western blotting of the reaction products with antibody to Ub (anti-Ub antibody) (Fig. 1A). E3 activity was observed when all reaction components were present, but not when p300, E1, or ubch5a was absent. Identification of polymerized Ub in the anti-Ub immunoblot was confirmed by its absence when the Ub derivative methyl Ub (Me-Ub) or the mutated Ub protein Ub-K48R (where K is Lys and R is Arg), both of which are defective in forming polyubiquitin chains (6), were used in the reaction (fig. S1). Thus, purified p300 exhibits E3 activity.

Figure 1

Purified recombinant p300 is a ubiquitin ligase. (A) (Left) Silver-stained gel containing purified p300 (15). (Right) Purified p300 or a control preparation (mock) were incubated in ubiquitination reactions containing the indicated components (15), and products were analyzed by Western blot with anti-Ub antibody. (B) Schematic of p300 domains and truncated polypeptides used in this study. (C) Purified p300N (residues 1 to 595), p300BD (residues 965 to 1810), or p300CT (residues 1135 to 2414) proteins or a control protein preparation (mock) (15) were separated by SDS-PAGE. Each gel lane was sliced into 5-mm segments subjected to in situ renaturation (15) followed by incubation with URC and analysis of reaction products by Western blot with anti-Ub antibody (top). A portion of each gel slice was also analyzed by Western blot with anti-epitope tag antibody (bottom). (A to C) Molecular weight markers in kD are at the left of blots and gels. Ubn indicates the migration position of polyubiquitin conjugates. Asterisk indicates p300BD degradation product. (D) p300, p300N, or p300M (residues 671 to 1196) were incubated with buffer, GST, GST-E1A, or GST-E1AΔ(2-36) (15) before incubation with URC and analysis of reaction products by Western blot with anti-Ub antibody.

Truncated, overlapping polypeptides spanning much of the p300 sequence were generated to map the domain(s) responsible for E3 activity (15) (Fig. 1B; fig. S2). Potent E3 activity was exhibited by the NH2-terminal 595 residues (p300N), and weaker activity was displayed by a polypeptide spanning residues 671 to 1196 (p300M) (fig. S2). Neither p300N nor p300M displays clear homology to known E3 motifs, such as RING, U-box, PHD, or hect domains (17, 18). p300N does contain a cysteine-histidine rich sequence (C/H1) that interacts with MDM2 and has been implicated in p53 stability control (13).

To determine whether E3 activity is intrinsic to p300, the p300N polypeptide was further gel-purified and renatured (15) (Fig. 1C). E3 activity was observed only in those gel slices where p300N protein was present and was not detected in slices from the control lane. Control polypeptides that lack E3 activity [p300BD, residues 965 to 1810; p300CT, residues 1135 to 2414 (fig. S2)] lacked activity after renaturation, as well (Fig. 1C). As a separate test of the intrinsic nature of p300N E3 activity, p300N was purified from Escherichia coli (15), which do not contain ubiquitination enzymes (19). Bacterial p300N displayed E3 activity, which was eliminated by boiling (fig. S3), implying that E3 activity is conformation-dependent. Thus, gel-purified insect and affinity-purified bacterial p300N are both intrinsically active E3 enzymes. This result implies that full-length p300 is also likely to be an intrinsic E3.

To investigate the effect of E1A on p300 E3 activity, we individually incubated purified p300, p300N, and p300M with purified glutathione S-transferase (GST) or GST fusion proteins containing either E1A (GST-E1A) or E1AΔ2-36, a p300 binding-defective mutant (GST-E1AΔ2-36) (15). URC were then added, and E3 activity was measured (Fig. 1D). Increasing quantities of GST-E1A progressively inhibited the E3 activity of p300 and p300N, both of which bind E1A (20). GST and GST-E1AΔ2-36 were inactive. The activity of p300M, which lacks an E1A-binding domain (20), was unaffected. Thus, E1A-mediated inhibition of p300 E3 activity correlates with the ability of p300 to interact with E1A, suggesting that certain cellular functions of E1A, such as p53 stabilization, may result from inhibition of p300 E3 function.

To determine whether p300 is an E3 in vivo, we incubated anti-p300 immunoprecipitates (IPs) of U2OS human osteosarcoma or mouse embryo fibroblast (MEF) lysates with URC and analyzed reaction products by Western blot with anti-Ub antibody. Human (Fig. 2A) and murine (Fig. 2B) p300 both displayed E3 activity, whereas control IPs lacked activity. An anti-CBP IP likewise demonstrated E3 activity (21), suggesting that CBP and p300 are both active E3s. p300 E3 activity was not explained by cross-reactivity of the anti-p300 antibody with an irrelevant E3, because anti-p300 IPs from p300–/– MEFs (15) were devoid of activity (Fig. 2B). Moreover, the E3 activity of p300-CBP was independent of MDM2, because p300-CBP immunoprecipitated from MDM2-null MEF lysate demonstrated robust E3 activity (fig. S4).

Figure 2

Cell-derived p300 is a ubiquitin ligase. Cell lysates or buffer were immunoprecipitated with irrelevant (anti-ras) or anti-p300 antibodies, followed by incubation of IPs with URC and analysis of reactions by Western blot with anti-Ub antibody. (A) U2OS cell lysate. (B) Wild-type or p300–/– MEF (15) lysates. A portion of each IP in (B) was also analyzed by Western blot with anti-p300 antibody (right). Molecular weight markers in kD are at the left of blots. Ubn indicates the migration position of polyubiquitin conjugates.

Human p300 E3 activity, like that of insect cell–derived p300, was efficiently inhibited by E1A, but not an E1A mutant defective for p300 binding (fig. S5A) (15). Moreover, p300 lacking the COOH-terminal C/H3 E1A interaction domain (p300ΔC/H3), such that it interacts less avidly but still specifically [due to interaction with the C/H1 domain (20)] with E1A (fig. S5C), was less susceptible to E1A inhibition than wild-type p300 (fig. S5B) (15). Thus, cell-derived p300 is susceptible to E1A inhibition, and the amplitude of E1A inhibition correlates with the avidity of the physical interaction between E1A and p300.

The ability of E1A to both stabilize p53 and inhibit p300 E3 activity suggests that p300 may play a direct role in p53 ubiquitination. Given that MDM2 catalyzes p53 monoubiquitination (4), the possibility that p300 acts as an E4 (8) to polyubiquitinate monoubiquitinated p53 was investigated. Purified p53 or p53-MDM2 complexes (15) were incubated with URC. Purified p300 or buffer was then added with additional URC, and ubiquitination of p53 was assessed. The expected ladder of monoubiquitinated p53 species was detected when p53-MDM2 complexes were exposed to URC, followed by a second control incubation with buffer and additional URC (Fig. 3A, lanes 3 and 4, arrows). These laddered forms were not observed in the absence of MDM2 (Fig. 3A, lane 2), and MDM2 protein by itself generated no signal with Western blotting (21). The pattern of p53-Ub conjugates did not change when Me-Ub was substituted for Ub, demonstrating that each p53 band represents monoubiquitination at relevant p53 lysines (Fig. 3A, lane 4). Incubation with p300 instead of buffer before the second ubiquitin reaction resulted in the production of high molecular weight forms of p53 suggestive of polyubiquitination (Fig. 3A, lane 5). Substitution of Me-Ub for unmodified Ub blocked the formation of these forms of p53 (Fig. 3A, lane 6), consistent with the suggestion that they represent polyubiquitinated p53. Thus, p300 exhibited the properties of an E4, which is capable of generating polyubiquitin conjugates of p53 from monoubiquitinated precursors.

Figure 3

p300 is an E4 ubiquitin ligase for p53. (A) Purified p53 or p53-MDM2 complexes (15) were incubated with URC. Buffer or FLAG-p300 were then added as indicated, followed by additional incubation. Further incubation with additional URC was followed by Western blot with anti-p53 antibody. Lane 1 represents p53 alone, without reaction. Ub or Me-Ub were included in reactions 1 and 2, as indicated. Arrowheads indicate monoubiquitinated p53 species (4). Asterisk indicates a covalent dimer form of p53 seen when purified p53 is incubated at 37°C (21). (B) p53-MDM2, p53-MDM2-p300N, or p53-MDM2ΔR-p300N complexes (15) were incubated with URC and reactions analyzed by Western blot with anti-p53 antibody. Input lanes indicate control reactions that were performed without URC. Molecular weight markers in kD are at the left of blots. Ubn indicates the migration position of polyubiquitin conjugates.

To assess whether p300 could ubiquitinate p53 in the absence of MDM2, we incubated purified p53 or p53-p300 complexes with URC and increasing quantities of MDM2 (15) (fig. S6). p300 alone caused no discernable change in p53 migration. However, addition of full-length MDM2 to p53-p300 complexes resulted in the formation of polyubiquitinated p53. MDM2 lacking its E3 domain (MDM2ΔR) (15) was inactive in this regard, confirming a necessary role for MDM2 E3 function in p53 polyubiquitination and further supporting the notion that p300 recognizes MDM2-produced monoubiquitinated species of p53 as a substrate for its E4 activity.

To determine whether p300N alone can provide E4 function for p53, we purified complexes of p53-MDM2, p53-MDM2-p300N, or p53-MDM2ΔR-p300N from insect cells (15). After their incubation with URC, polyubiquitinated p53 was generated by the p53-MDM2-p300N complex only (Fig. 3B). The absence of p300N or inclusion of MDM2ΔR or Me-Ub in the reactions abrogated the production of polyubiquitinated p53. Thus, p300N exhibited E3 autoubiquitination activity (Fig. 1) and the ability to cooperate with MDM2 and act as a specific E4 for p53 in vitro (Fig. 3B).

To gauge the contribution of p300-CBP E4 activity to p53 polyubiquitination in vivo, we analyzed the ubiquitination of endogenous wild-type (wt) p53 in cells synthesizing E1A by immunoprecipitation of transfected cell lysates with anti-p53 antibody followed by Western blot with either anti-p53 or anti-Ub antibody (15) (Fig. 4, A and B). E1A expression was confirmed by Western blot analysis (Fig. 4C). High molecular weight polyubiquitin conjugates of p53 were present in lysates from control and E1AΔ2-36–expressing cells but were nearly absent from lysates of wt E1A-expressing cells (Fig. 4A). Contrary to the results seen with polyubiquitinated p53, the abundance and gel-migration properties of monoubiquitinated p53 species were not affected by E1A expression (Fig. 4B, arrows). These results suggest that p300-CBP participates in the in vivo polyubiquitination, but not monoubiquitination, of p53.

Figure 4

E1A inhibits p53 polyubiquitination in vivo. U2OS cells were transfected with expression vectors encoding E1A, E1AΔ2-36, or control vectors. After incubation in proteasome inhibitor, cells were lysed and immunoprecipitated with immunoglobulin G (IgG) or anti-p53 antibody, followed by Western blot (15). (A) Western blot with anti-Ub antibody. Molecular weight markers in kD are at the left of blots. Ubn indicates the migration position of polyubiquitin conjugates. (B) Western blot with anti-p53 antibody. Arrows indicate the migration positions of monoubiquitinated p53 species. (C) Western blot of cell lysates with anti-E1A antibody.

We have shown that the requirement for p300 in p53 turnover is explained by a p300 E4 activity that causes p53 polyubiquitination. An additional role for p300 in p53 turnover is suggested by its interaction with the proteasome adaptor hHR23A (22). Whether p300 E4 function is regulated by DNA damage or other stressors is unknown. The effect of p300 acetylase activity on its E4 function is also unknown. Intriguingly, the same p53 residues can be acetylated or ubiquitinated, and MDM2 promotes p53 deacetylation along with degradation (23). Ubiquitination of other transcription factors by p300-CBP may also occur, and this activity may be relevant to coactivation by p300-CBP (24). In addition, inhibition of p300 E4 function might be of value in treating tumors in which p53 turnover is constitutively rapid, such as those in which the MDM2 gene is amplified or overexpressed.

Supporting Online Material

www.sciencemag.org/cgi/content/full/300/5617/342/DC1

Materials and Methods

Figs. S1 to S6

References

  • * Present address: Department of Cancer Biology, University of Massachusetts Medical School, 364 Plantation Street, Worcester, MA 01605, USA.

  • Present address: Department of Molecular and Cell Biology, University of California, Berkeley, Berkeley, CA 94720, USA.

  • To whom correspondence should be addressed. E-mail: david_livingston{at}dfci.harvard.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article