DNA-dependent acetylation of p53 by the transcription coactivator p300.

Reconstitution of the stages in the assembly of the p300.p53 transcription complex has identified a novel type of DNA-dependent regulation of p300-catalyzed acetylation. Phosphorylation at the CHK2 site (Ser(20)) in the N-terminal activation domain of p53 stabilized p300 binding. The phosphopeptide binding activity of p300 was mapped in vitro to two domains: the C-terminal IBiD domain and the N-terminal IHD domain (IBiD homology domain). The IHD or IBiD minidomains can bind to the p53 activation domain in vivo as determined using the mammalian two-hybrid VP16-GAL4 luciferase reporter assay. The IHD and IBiD minidomains of p300 also functioned as dominant negative inhibitors of p53-dependent transcription in vivo. Upon examining the affects of p300 binding on substrate acetylation, we found that the p53 consensus site DNA promotes a striking increase in p53 acetylation in vitro. Co-transfection into cells of the p53 gene and plasmid DNA containing the consensus DNA binding site of p53 activated DNA-dependent acetylation of p53 in vivo. The phosphopeptide binding activity of p300 is critical for DNA-dependent acetylation, as p53 acetylation was inhibited by phospho-Ser(20) peptides. Consensus site DNA-dependent acetylation of p53 stabilized the p300.p53 protein complex, whereas basal acetylation of p53 by p300 in the presence of nonspecific DNA resulted in p300 dissociation. These data identify at least three distinct stages in the assembly of a p300.p53 complex: 1) p300 docking to the activation domain of p53 via the IBiD and/or IHD domains; 2) DNA-dependent acetylation of p53; and 3) stabilization of the p300.p53(AC) complex after acetylation. The ability of DNA to act as an allosteric ligand to activate substrate acetylation identifies a conformational constraint that can be placed on the p300-acetylation reaction that is likely to be an amplification signal and influence protein-protein contacts at a promoter.

Regulation of eukaryotic gene expression is a complex process regulating growth control in normal cells and dysregulation of gene expression contributes to diseases including cancer.
Characterizing the protein-protein interactions that assemble transcription complexes will provide a molecular basis for combinatorial specificity at the promoter level and provide an understanding of how transcriptosomes affect disease progression. One of the most studied regulators of eukaryotic gene transcriptional activation is the stress-activated tumor suppressor protein p53 (1). The vast array of genes that p53 induces and represses in response to cell injury provides an excellent model system with which to identify the complex and transient associations with co-activators/repressors and the core promoter transcriptional apparatus that drive promoter activation.
The tumor suppressor activity of p53 is linked to its activity as a stress-activated transcription factor. p53 is subject to multiple post-translational modifications such as phosphorylation, acetylation, and ubiquitination that modulate the function of p53 as a transcription factor. p53 protein is tetrameric (2) and the oligomeric nature of p53 provides the basis for complex intra/interdomain control mechanisms that modulate its activity (3,4). The most widely studied post-translational modification of p53 is phosphorylation via protein kinases that target specific domains on the assembled tetramer that can activate its function by a two-step process (5). The first identified involves phosphorylation of the C-terminal regulatory domain of p53 by CK2/FACT at Ser 392 and cyclin-dependent protein kinases at Ser 315 , which play a role in activating its latent sequence-specific DNA binding activity in vitro (6,7) and in vivo (8,9). The second step involves phosphorylation of specific N-terminal residues of p53 that reside within the activation domain by an ATM-dependent signal transduction cascade at Ser 15 (10), which stimulates p300 binding to p53 in vitro (11) and in vivo (12). Additionally, an adjacent Ser 20 phosphorylation event by a CHK2-dependent pathway activates p53 (13) by stabilizing the transcription coactivator p300⅐p53 complex formation (14). The specific interaction of p300 in the p53-dependent transactivation pathway became apparent when it was demonstrated that ectopically expressed p300 stimulated p53-dependent gene expression and that adenoviral E1A protein inhibited p53-dependent transcription by virtue of binding to p300 (15)(16)(17). Although this initial p300-p53 interaction was mapped to the N-terminal BOX-I domain of p53, an additional role for p300 in the control of p53 activity came from the observation that acetylation of p53 by p300 in the C-terminal negative regulatory domain adjacent to the CK2/FACT phosphorylation site activated the specific DNA binding function of p53 to short oligonucleotides containing the consensus binding site for p53 (18). Together, these data suggest the existence of a phosphorylation-acetylation cascade that targets p53 in response to genotoxic stress (19).
The acetyltransferase activity of p300 and its homolog CBP on histone and nonhistone substrates has been well documented and is important for co-activator function (20,21). The * 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  role for acetylation on the N-terminal tails of histones is somewhat more widely appreciated, with the acetylation of lysine residues potentially weakening the nucleosome-DNA interaction, thereby allowing promoter access for the core transcriptional machinery such as TFIID and RNA polymerase II. Acetylation of transcription factors is now a widely observed phenomenon (22)(23)(24)(25) but the role and mechanism of acetylation of transcription factors remains relatively elusive (26). However, more recent studies have also suggested that acetylation may function to stabilize coactivator complexes at a promoter (27,28), presumably in part through the Bromo-homology domain of p300/CBP that has the potential to bind to acetylated residues (29).
The molecular mechanism underlying how p53 tetramers interact with p300 and how this modulates acetylation has not been defined biochemically. Such analysis will likely give novel insights into how p300 protein interacts with an oligomeric substrate as well as information into how p300 links transcription to cell cycle control. We have started to reconstitute the p300⅐p53 protein complex in vitro by first examining how fulllength p300 interacts with various phosphorylated regions in the N-terminal BOX-I activation domain of p53 (14). This study has indicated that of three N-terminal phosphorylation sites in the BOX-I activation domain, phosphorylation by CHK2 at Ser 20 has the most stabilizing affect on p300⅐p53 complex formation. In this report, we have used biochemical assays to develop a consensus phosphopeptide-binding site for p300 and to map the phosphopeptide-binding domain of p300 to two distinct domains. One of the domains maps to a previously undefined N-terminal region on p300 we name IBiD homology domain (IHD), which has homology to the established p53binding domain in the C terminus of p300 previously called IBiD or SRC-1 (30,31). Further characterization of how p300 docking to the activation domain of p53 influences acetylation has revealed that p300 docking is essential for sequence-specific DNA-dependent acetylation of p53 that in turn stabilizes the p300⅐p53 AC complex. These data reveal three distinct stages in the assembly of the p300⅐p53 complex: (i) p300 docking, (ii) p300 catalyzed DNA-dependent acetylation, and (iii) acetyl-CoA-dependent stabilization of the p300⅐p53 complex. These data further suggest that p53 protein conformation when DNA bound at promoters may play a role in modulating the rate of its acetylation and recruitment of chromatin remodeling factors.
Immunochemical Assays-The peptide or p53 tetramer binding activity of p300 was examined by enzyme-linked immunosorbent assay (ELISA), 1 as described previously (3,14). Essentially, 96-well plates (Dynex Microlite 2) were first coated with p53, antibody-captured p300, or streptavidin and the indicated biotinylated peptide for 16 h, as described previously (14). Nonreactive sites were blocked in 3% bovine serum albumin in PBS/Tween 20 (0.02% v/v) to reduce the nonspecific binding. This was followed by titrating increasing amounts of p53, p300, MDM2, peptides, without or with RNA (poly(G); data now shown)) (as in Ref. 34) in 3% bovine serum albumin in PBS/Tween 20 (0.02% v/v) for 1 h, followed by an extensive wash, and incubation with the indicated IgG. All reactions were carried out at 4°C and detected by the appropriate secondary antibody linked to horseradish peroxidase from DAKO. The signal detection by enhanced chemiluminescence was developed using Fluoroscan Ascent FL. The BOX-I transactivation domain of p53 or its Ser 20 -phosphorylated derivative contains amino acids 14 -27, as described previously (14). All synthetic peptides were obtained from Chiron Mimitopes.
Cell Culture, Transfections, ELISAs, and Western Blots-A375 and Saos-2 cells were maintained in Dulbecco's modified Eagle's medium (Invitrogen), whereas HCT116 cells were in McCoys 5A medium, both supplemented with 10% fetal bovine serum and incubated at 37°C with an atmosphere of 10% CO 2 . Transient transfections and ELISAs were carried out as previously described (14). Full-length p300 and His-p300infected Sf9 cells were harvested 72 h postinfection and purified as described previously (14). Sf9-expressed wtp53 tetramers were purified by heparin-Sepharose chromatography as described previously (3). Transfected lysates were run on a 12% SDS-PAGE, transferred to nitrocellulose membrane, and protein loading was confirmed by Ponceau S (Sigma) staining. The p300 antibody used was N-15 (Stressgen). Primary antibodies anti-p53 (DO-1 or ICA-9 (3)) were used and the appropriate secondary antibody conjugated to horseradish peroxidase. The signal detected by enhanced chemiluminescence was developed using autoradiography film (Amersham Biosciences).
ELISA-Microtiter wells were coated with 100 ng of streptavidin and incubated overnight at room temperature. To prevent nonspecific binding, 180 l of 50 l/well of 3% bovine serum albumin ϩ PBST (3BPBST) was added and incubated for 1 h at 4°C. Wells were washed 3 times with 180 l of PBST and biotinylated peptides, resuspended in 3BPBST with 50 mM NaF, 5 mM ␤-phosphoglycerate, and incubated for 1 h at 4°C. Nonspecific binding sites were blocked by adding 180 l of 5% milk ϩ PBS (5MPBST)/NaF/␤-phosphoglycerate for 30 min at 4°C before adding specific protein, resuspended in 50 l/well of Sf9LB or PBST for p300 and MDM2, respectively. The proteins were incubated for 1 h at 4°C and wells were rigorously washed 3 times with 180 l of PBST before incubating for 1 h with specific antibodies in 5MPBST/NaF/␤phosphoglycerate. Wells were given another 3 washes with PBST to remove unbound antibody and specific binding was detected with secondary antibody (anti-mouse or anti-rabbit) conjugated to horseradish peroxidase. The binding was detected by ECL and quantified using a luminometer (Fluoroskan Ascent FL). The protocol for the streptavidin capture ELISA was followed, except increasing amounts of peptide was incubated with the protein for 30 min at 4°C before incubating with the substrate peptide.
In Vitro Acetylation Reactions-Purified p53 (50 -400 ng) from Sf9 cells was incubated with 300 -400 ng of purified His-p300 in 30 -100 l of AT buffer (50 mM Tris⅐HCl (pH 8), 10% glycerol, 0.1 mM EDTA, 1 mM dithiothreitol, 5 M trichostatin A, and 2 M acetyl-CoA) for 4 -10 min at 30°C where the enzymatic reaction was linear. The specific oligonucleotides (pG DNA) and nonspecific oligonucleotides (NS-DNA) were described previously (35). Reactions were incubated on ice for 10 min and started with the addition of p300. Acetylation of p53 was detected by the antibody-capture ELISA technique using anti-p53 (ICA9) or by direct Western blot using anti-acetyl p53 (AcK373/382) and normalizing with anti-p53 (19.1). Histone acetylation was carried out in a similar manner but with the use of 1 g of purified histone H4 (Upstate Biotechnology) as the substrate for p300 and using anti-histone (Roche Molecular Biochemicals) or anti-acetyllysine (Upstate Biotechnology) to detect reaction products. Quantification of acetylation was carried out using bioluminescence (Genegnome (Syngene Bioimaging System)).
In Vivo DNA-dependent p53 Acetylation Assay-HCT116 (p53 Ϫ/Ϫ ) cells were transfected with pcDNA3.1-p53 (or p53-6KR acetylation mutants as indicated (56)) and either pG13-CAT DNA (p53 consensussites) or pMG13-CAT DNA (mutated p53 consensus-sites; vectors described as indicated (36)) for 48 h before harvesting. After washing twice with 5 ml of ice-cold PBS, lysates were incubated with acetylation immunoprecipitation lysis buffer (50 mM Hepes, pH 7.8, 200 mM NaCl, 1% (v/v) Triton X-100, 10 mM EDTA, 5 mM dithiothreitol, 1ϫ PIM, and 5 M trichostatin A) for 15 min on ice before scraping and passing through a 21-gauge syringe needle 10 times and then transferring the supernatant to fresh microcentrifuge tubes. The resultant lysates were then immunoprecipitated using a p53-specific antibody (DO-1 or ICA-9) or p300 antibody (N-15) and then probed for acetyl-p53. The specific acetylation was normalized to the quantity of total p53 captured assessed by probing with anti-p53 (DO-1 or CM1/5). 20 -phosphorylated BOX-I Activation Domain of p53-The overlapping docking site for MDM2 and p300 on p53 (Fig. 1E) creates a negative and positive effect, respectively, on its ability to function as a tumor suppressor protein. The notable differences in the affinity of p300 and MDM2 toward p53-derived Thr 18 and Ser 20 phosphopeptides suggested an intrinsic difference in their binding contacts in the BOX-I activation domain (14). To define essential amino acid contacts, a series of alanine-substituted Ser 20 -phosphorylated peptides were synthesized and the protein-ligand binding assay was employed using full-length p300 and full-length MDM2 proteins (Fig. 1, A and B, respectively).

Developing a Consensus p300-binding Motif by Mapping Its Contact Site in the Ser
Ser 20 phosphorylation of p53 is required for maximal fulllength p300 binding to this domain, whereas MDM2 is not inhibited (14) thus allowing the same phosphodomain of p53 to be used for consensus site mapping of MDM2 and p300. If Gln 16 , Trp 23 , Lys 24 , Leu 25 , or Leu 26 are individually mutated to alanine, then this abrogates the ability of the p300 protein to bind to the Ser 20 phosphodomain (Fig. 1A). Certain alanine substitutions in the Ser 20 phosphodomain including Ser 15 , Glu 17 , Thr 18 , Phe 19 , and Asp 21 stabilized the binding of p300 to the phospho-motif (Fig. 1A). In contrast to p300, a Thr 18 , Phe 19 , Leu 22 , Trp 23 , or Leu 26 substitution to alanine on the BOX-Iderived Ser 20 -phosphorylated domain inhibits MDM2-binding ( Fig. 1B), correlating with the essential residues required for MDM2 protein binding to p53 identified by the crystal structure of the MDM2-p53 peptide complex (37). This evidence that MDM2 and p300 have distinct binding contacts explains, in part, their differential binding to phospho-Thr 18 or phospho-Ser 20 p53 (14).
Interestingly, a search for proteins in the data base that have significant homology to the consensus BOX-I p300-binding motif on p53, QXXXS PO3 XXWKLL (Fig. 1C), identified the HMG box architectural factor, UBF1 (upstream binding factor 1; Fig.  1C). UBF necessitates major chromatin remodeling and the SXXWKLL region of UBF is within a domain that is already known to interact with the p300 homologue, CBP (CREB-binding protein) (38). To determine whether p300 can bind to this region of UBF1 we employed the protein-ligand binding assay containing two putative S PO3 XXWKLL consensus peptides from UBF1 (amino acids 231-246 and 321-336) with or without a phosphate moiety at the predicted serine residue. Notably, when a phosphate group was attached to Ser 329 in UBF1-(321-336), which contains strict homology to the S PO3 XXWKLL motif, p300 bound with an affinity slightly less than the p53 BOX-I-Ser 20 phosphopeptide (Fig. 1D), suggesting that this may indeed be a consensus contact region for p300. In contrast, UBF1-(231-246) did not bind p300 with or without a phosphate substitution at Ser 239 (Fig. 1D) nor did p300 bind the BOX-I domain of p53 without a phosphate (Fig. 1D). These differences may be because of the absence of a tryptophan residue at position 242 (or the presence of a destabilizing lysine) (Fig. 1C), which is required for p300 binding to the p53 BOX-I-Ser 20 phosphopeptide alanine scan (Fig. 1A).
Identification of Two p53-activation Motif Binding Domains on p300 -Various techniques have been utilized to originally define the p53 activation domain-binding sites in p300 to within the C/H1, KIX, and C/H3 minidomains of p300 (Fig. 2). However, the most recent mapping of the p53-binding domain of p300 was localized to the C-terminal interferon-binding do-main (IBiD, Fig. 2A), which is known to bind to peptides of apparently unrelated primary amino acid homology including phosphorylated peptides derived from interferon regulatory FIG. 1. Developing a consensus phosphopeptide binding motif for p300 protein. The binding of: A, full-length p300 protein, and B, full-length MDM2 protein, to biotinylated-Ser 20 phosphopeptides (amino acids 14 -27 of human p53), as described previously (14), substituted with alanine at the indicated positions were analyzed by ELISA. Increasing amounts of biotinylated peptide (0, 0.01, 0.1, or 1 ng) were titrated onto streptavidin-coated surfaces as indicated (14). The amount of p300 or MDM2 protein bound is quantitated as luciferase activity (RLU) using a peroxidase-linked secondary antibody coupled to anti-p300 or anti-MDM2 antibodies. C, the alanine scan of p53 derived the putative consensus site for p300 in blue (underlined) as QXXXS PO3 XXWKLL. An e-motif search of a protein data base with these amino acids identified the transcription factor UBF as containing BOX-I homology domains and are shaded in green. D, the binding of p300 to UBF or the p53-BOX-I peptide with or without a phosphate substitution at the highlighted serine residue was analyzed by ELISA as indicated in A. The amount of biotinylated peptide titrated onto streptavidin-coated surfaces is as indicated. The amount of p300 protein bound is quantitated as luciferase activity (RLU) using a peroxidase-linked secondary antibody coupled to an anti-p300 antibody. UBF-1 has the strongest homology to the p300 consensus site in the C-terminal hydrophobic region (S PO3 XXWKLL). E, domain structure of human p53. The conserved domains I through V, the tetramerization motif, the polyproline domain, and C-terminal negative regulatory domains (NRD1 and NRD2) are as indicated. The position of the p300-and MDM2-binding sites in BOX-I, the p300 acetylation sites in NRD1, and the ATM and CHK2 phosphorylation sites in BOX-I are as indicated.
factor-3 (30). This motif on IRF-3 has homology to the phosphorylated p53-activation domain (data not shown) and in a recent study, a p53-activation domain-binding region in p300 was mapped to the IBiD domain and independently called SRC-1 ( Fig. 2A) (31). An in vitro assay was employed to localize the p53 peptide-binding domains of full-length p300 to either previously known or unmapped regions of p300 protein by assaying fragments of p300 for Ser 20 phospho-BOX-I activation domain binding. Various N-and C-terminal fragments of p300 that overlap, which are distinct from, or that contain the CH/1, KIX, CRD1, Bromo, C/H2, C/H3, and IBiD minidomains were tagged and tested for p53 binding activity. These include ( Cell lysates obtained from HCT116 (p53 Ϫ/Ϫ ) cells transfected with the indicated GAL4-p300 fusion constructs were incubated with an anti-GAL4 antibody in ELISA wells to capture and purify the p300 miniproteins. The corresponding biotinylated peptide derived from p53 (BOX-I, Ser 20 -phospho-BOX-I, or no peptide) was added, and p300-peptide complex stability was verified and quantitated using streptavidin-horseradish peroxidase coupled to luminography (Fig. 2B, GAL4 versus GAL4-full-length p300, 50 -65 RLUs background versus 120 RLUs, respectively). The minimal p300 fragments that bound to the phospho-BOX-I motif were from 1757 to 2414 in the C terminus (147 RLUs) and from 192 to 504 in the N terminus (93 RLUs) (Fig. 2B). These p300 domains contain the C/H1, part of the C/H3 (C/H3␦) domain, and the IBiD domain, which were previously reported to bind to p53 (20). Further localization of the phosphopeptide-binding region of p300 was carried out using portions of p300 from 1757 to 2414 in the C terminus and from 192 to 504 in the N terminus (  Fig. 2A) and defines a previously unreported region on p300 that binds to p53. The minimal p300 fragment from the C terminus that bound to the phospho-BOX-I motif mapped out the C-terminal C/H3 domain and encoded a 51-kDa miniprotein ( Fig. 2C; GAL4-C2, 94 RLUs) that displayed specificity for the BOX-I Ser 20 -phospho motif of p53, relative to the negative control GAL4 alone (50 -60 RLUs) or the positive control GAL4-full-length p300 (122 RLUs). The C-terminal p53-binding domain in our studies maps out the C/H3 domain, which is consistent with recent studies showing that the p53-binding domain in the C terminus of p300 resides within a domain called IBiD or SRC-1 (30,31). An alignment of the minimal C-terminal IBiD domain with the N-terminal GAL4-N3 fragment (amino acids 406 -566) demonstrated a significant degree of identity (Fig. 3C). We have named the N-terminal domain of p300 that binds the phosphorylated activation domain of p53 as IHD ( Figs. 2A and 3C).
To determine whether the N-terminal IHD domain of p300 exhibited similar p53-binding characteristics to the C-terminal FIG. 2. Identification of two phosphopeptide-binding domains on p300 using an in vitro p300-miniprotein binding assay. A, a summary of the well characterized peptide-binding domains of p300: C/H1, KIX, C/H2, Bromo, C/H3, IBiD, and (see Fig. 3) the N-terminal phosphopeptide-binding domain with homology to IBiD (IHD). B, mapping of the phosphopeptide-binding domain to two large regions of p300. HCT116 p53 Ϫ/Ϫ cells were transfected with 5 g of GAL4, GAL4-p300, or the indicated GAL4-p300 miniproteins and the p300 from lysates was captured on ELISA wells with an anti-GAL4 antibody to purify the indicated miniproteins. The indicated biotinylated peptide (BOX-I, Ser 20phospho-BOX-I, or no peptide) was added and the amount of p300 protein bound is quantitated as luciferase activity (RLU) using a peroxidaselinked secondary antibody coupled to an anti-p300 antibody. C, small fragments of p300 used to fine map the phosphopeptide-binding domain to two regions flanking the C/H1 and C/H3 domains of p300. HCT116 p53 Ϫ/Ϫ cells were transfected with 5 g of GAL4-p300 miniproteins (C1-C3 and N1-N3) and p300 miniproteins lysates were captured on ELISA wells with an anti-GAL4 antibody to purify the indicated polypeptides. The indicated biotinylated peptide (BOX-I, Ser 20 -phospho-BOX-I, or no peptide) was added and the amount of p300 protein bound is quantitated as luciferase activity (RLU) using a peroxidase-linked secondary antibody coupled to an anti-p300 antibody.
IBiD domain of p300, we used the VP16-GAL4 two-hybrid system that was used originally to fine map the C-terminal IBiD domain on p300 that binds to the activation domain of p53 when fused to GAL4 (30,31). The co-transfection of pACT-VP16-IBiD or pACT-VP16-IHD along with p53-TAD:GAL4 both supported GAL4-dependent transcription (Fig. 3A, right 5 panels; 1063 RLUs for IBiD and 877 RLUs for IHD), relative to the negative control using pACT-VP16 only (180 RLUs). The pACT-VP16-C/H3␦ domain gave rise to p53-binding activity above the background (324 RLUs), consistent with some studies showing that the C/H3 domain can bind to p53. Control transfection of the same pACT constructs without the p53-TAD-GAL4 vector did not support transcription (Fig. 3A, left  5 panels).
As an independent test of the ability of IHD and IBiD as important effectors of p300-coactivated p53-dependent transcription, the ability of pACT-IBiD and pACT-IHD were tested as dominant negative inhibitors of p53-dependent transcription from the p21-Luc reporter (Fig. 3B). The maximal p53-dependent transcription upon co-transfection with p300 (Fig. 3B, 2483 RLUs) was significantly attenuated using either pACT-IBiD (1002 RLUs) or pACT-IHD (1146 RLUs), relative to pACT-C/H3␦ (1918 RLUs) or pACT only (2538 RLUs). Together, these data identify IBiD as the major p53-binding domain of p300 and the novel N-terminal domain IHD as an equally effective p53-binding domain. We cannot rule out the previously identified p53-binding domains including the C/H3, the KIX , or the C/H1 domains as effective p53-binding domains of p300, as the folding and assembly of the relatively large p300 protein with its various repressor and activator motifs compli-cates a clear understanding of how p300 really binds to p53. For example, there is some weak, but significant in vivo p53binding activity associated with the region of the C/H3␦ domain used in this study (Fig. 3, A and B). However, IBID and IHD are the predominate p53-binding domains with similar specific activities as defined in: (i) in vitro p53-binding assays (Fig. 2B); (ii) in vivo p53-binding assays (Fig. 3A); (iii) dominant negative inhibitors in p53-dependent transcription assays (Fig. 3B).
Consensus Site DNA-dependent Acetylation of p53-To establish a function for the IBiD and IHD domains of p300, the contribution of the phosphopeptide-binding activity of p300 to p53 acetylation was examined. Specifically, does binding of p300 to the BOX-I domain of p53 affect acetylation in the C-terminal NRD1 domain of p53 (see p53 domain structure in Fig. 1E). Before setting up an in vitro p53 acetylation assay, the purified p300 protein was first characterized enzymatically using a well known substrate, histone H4. The kinetics of the acetylation reaction between p300 and histone H4 were determined (Fig. 4, A and B). The data indicate that p300 is behaving like a classical histone acetyltransferase by displaying monophasic kinetics on a histone substrate (39,40) and this preparation of p300 was tested for the ability to acetylate native p53 tetramers expressed in Sf9 cells (41). Using an antibody against acetylated Lys 373/382 of human p53 that has been previously described as being specific for acetyl-p53 (27) and normalizing to levels of p53 protein with an anti-p53 antibody, acetyl-CoA-dependent acetylation of p53 was observed (Fig. 4C, lane 3 versus lane 2).
In a further characterization of p53 acetylation, we found that including consensus site DNA in the acetylation buffer In the left 5 panels, the indicated constructs were co-transfected (1 g) into HCT116 p53 Ϫ/Ϫ cells with a ␤-galactosidase control vector (pCMV-␤-gal) and lysed 24 h later to test for the ability to affect transcription from a GAL4-luciferase reporter construct (pG5-luc). In the right 5 panels, the indicated pACT-VP16 fusion constructs (1 g) were co-transfected with a construct encoding GAL4 fused to the N-terminal transactivation domain of p53 (amino acids 10 -40) and a ␤-galactosidase control vector (pCMV-␤-gal). The activity is summarized as relative light units normalized to ␤-galactosidase activity. B, the IBiD and IHD domains attenuate p53-dependent transcription in cells. HCT116 p53 Ϫ/Ϫ cells were co-transfected with the indicated constructs (1 g; pCMV-p53, pCMV-p53 ϩ pCMV-p300, alone, or with the indicated pACT fusion constructs), pCMV-␤-gal (1 g), and the p53-responsive p21-luc reporter vector. The activity is summarized as relative light units normalized to ␤-galactosidase activity. C, the identity between the C-terminal IBiD and N-terminal IHD domains is summarized. resulted in an unexpected stimulation in p53 acetylation (see below). Using the basal p300-acetylation reaction where acetyl-CoA promotes weak p53 acetylation (Fig. 4C, lane 3 3 and 9). A titration of consensus site DNA demonstrated that at a ratio of DNA to p53 of ϳ1:1, the acetylation of p53 approached saturation (Fig. 5C, lanes 4 and 5 versus lane  1). If consensus site DNA were stimulating general p300 activity by an allosteric mechanism via a sequence-specific DNAbinding domain on p300, then basal histone acetylation (Fig. 4,  A and B) would be expected to be stimulated by DNA. However, histone acetylation by p300 was neither inhibited nor stimulated by the addition of consensus site DNA (Fig. 5D, lanes  1-6). These data suggest that the acetylation of p53 was unique to a change in the conformation of p53 when bound to consensus site DNA.
The kinetics of p300 acetylation on histones has been published recently (40) and displays a ping-pong mechanism, via an ordered reaction involving first the formation of a stable p300-acetyl intermediate followed by binding of the histone substrate and transfer of the acetyl moiety to a lysine residue. As a final control to define the integrity of the acetyltransferase reaction, we determined whether the acetylation of DNA-bound p53 by p300 displays similar kinetics to that of histone substrates. The linearity of the p53-acetylation reaction over time (Fig. 6A) and as a function of DNA concentration (Fig. 6B) was optimized using bioluminescent quantification of reaction products. By changing the concentration of acetyl-CoA at a fixed p300 concentration and consensus site DNA concentration, varying the p53 levels resulted in linear changes in the levels of acetylation. After quantification using bioluminescence, a double reciprocal plot of the initial velocities was drawn using Prism kinetics software (Fig. 6C). p300 acts in a similar manner to p53⅐DNA complexes as a substrate compared with histones suggesting that the p53-p300 acetylation reaction similarly displays a ping-pong mechanism as opposed to a sequential (ternary complex) mechanism. Kinetics could not be performed on p53 acetylation without DNA as the stoichiometry of acetylation was too low to acquire statistically significant data. 2 Although acetylation can stimulate latent p53 binding to small consensus site DNA oligonucleotides (18), acetylation does not change the affinity of p53 for its consensus site in large fragments of DNA (28). As such, the role of acetylation in the control of p53 function has remained elusive (26). However, it is very difficult to acquire pure isoforms of recombinant latent phosphorylated p53, latent unphosphorylated p53, or kinaseactivated p53 (41), especially in insect cell expression systems where dozens of p53 isoforms exist (42). This heterogeneity in p53 modifications possibly explains, in part, the apparent controversy on the role of acetylation in modulating the specific activity of p53 using in vitro DNA binding assays. It therefore is important to determine whether consensus site DNA stimulation of p53 acetylation also could be promoted in large plasmid of the DNA rather than small oligonucleotides.
Using a supercoiled plasmid DNA containing multiple copies of the p53 consensus site (pG13) (43), a DNA-dependent stimulation of p53 acetylation by p300 was observed, compared with parallel reactions containing pMG13 nonconsensus site control DNA (Fig. 7A, lane 4 versus lane 3). These data indicate that DNA-dependent acetylation was not confined to small oligonucleotides and indicate that although acetylation can stimulate unphosphorylated and latent p53-DNA binding (18), our use of in vivo kinase-activated, purified p53 protein isoforms (as in Ref. 41) demonstrates that acetylation is a post-DNA binding event.
It was next important to determine whether sequence-specific DNA-dependent acetylation could occur in vivo. Two p53 constructs (p53 and p53-6KR (the nonacetylatable p53 (56))) were co-transfected with consensus site plasmid DNA (or plasmid control containing the mutated consensus site) and the expressed p53 protein was assayed for both binding to endogenous p300 protein and in vivo acetylation (Fig. 7, B-E). The transfection of the p53 gene with either pMG13 DNA or pG13 DNA, followed by immunoprecipitation with a p300 antibody resulted in the same level of p53 protein being co-precipitated with p300 as detected by blotting with a p53 antibody (Fig. 7C,  lanes 1 and 2). However, only consensus site pG13 DNA promoted a substantial acetylation of p53 in the p300 immune complex as detected by blotting with a p53-acetylation site antibody (Fig. 7B, lane 2 versus lane 1). This is in contrast to the control transfection of the p53 gene with the mutant consensus site supercoiled plasmid pMG13, which did not produce a stable complex between p300 and acetylated p53 (Fig. 7B,  lane 1). Thus, because significant amounts of nonacetylated p53 protein bound to p300 upon co-transfection with the p53 gene and pMG13 plasmid DNA (Fig. 7, B and C, lanes 1), these data indicate that acetylation of p53, rather than p53 binding to p300, is promoted by consensus site plasmid DNA. As a further control for total p53 acetylation, co-transfection of the 2 D. Dornan and T. R. Hupp, unpublished data. 4. p300 acetylation of p53 and histone H4. A, monophasic kinetics of acetylation using purified p300, p53, and the histone H4 substrate. 100 ng of purified p300 was incubated over the indicated times (0 to 30 min) with 1 g of histone H4 and acetyl-CoA. The relative acetylation over time was determined by immunoblotting the reaction products onto nitrocellulose and detecting total histone protein (H4) or acetylated histone (Ac-H4) using either histone antibody or an antiacetylation antibody. B, quantitation of the acetylation using bioluminescence with a Genegnome-Syngene Bioimaging System. C, acetylation of p53 by p300. 100 ng of p300 and 400 ng of p53 were incubated for 10 min and relative acetylation was determined by Western blot using an anti-p53 acetylation antibody: p53 alone (lane 1), p53 ϩ p300 (lane 2), and p53 ϩ p300 ϩ acetyl-CoA (lane 3). p53 gene with pG13 plasmid (versus the pMG13 plasmid) stimulated total acetylation of p53 after precipitation of protein complexes with an anti-p53 antibody and immunoblotting with an acetylation-specific p53 antibody (Fig. 7D, lane 2 versus lane  1). This is similar to the in vitro stimulation of p53 acetylation by p300 with the inclusion of pG13 DNA (Fig. 7A, lane 4 versus  lane 3). Together, these data indicate that consensus site plasmid DNA-dependent acetylation can be observed in vitro or vivo.

FIG.
Acetyl-CoA Stabilizes the p300⅐p53 AC Complex-With the DNA-dependent p53 acetylation assay biochemically characterized, we could finally access the role of the phospho-Ser 20 peptide-binding domains of p300 to both DNA-dependent p53 acetylation and DNA-independent histone acetylation (Fig. 8).
In the absence of peptide, acetyl-CoA was required for DNAdependent acetylation of p53 by p300 (Fig. 8A, from left, lane 2  versus lane 1). A titration of BOX-I peptide displayed no effect on the p53 acetylation reaction (Fig. 8A, from left, lanes 3-6  versus lane 2). However, a titration of the BOX-I phospho-Ser 20 peptide inhibited p53 acetylation with an IC 50 of ϳ150 M in this assay (Fig. 8A, from left, lanes 7-10 versus lane 2). These data indicate that the phosphopeptide-binding domain(s) of p300 interacts with the N-terminal domain of p53 to facilitate acetylation in the C-terminal domain of p53. As a control, histone acetylation was not affected by the BOX-I phospho-Ser 20 peptide (Fig. 8B) indicating that the acetylation reaction is docking-and DNA-independent with respect to histone acetylation. Using the p53 acetylation assay in an ELISA format where monoclonal-captured p53 could be probed with an anti-acetylation polyclonal antibody, the phospho-Ser 20 peptide similarly prevented DNA-dependent acetylation of full-length p53 (Fig. 8C).
In summary, our reconstitution of the p300⅐p53 complex has indicated that acetylation is DNA-dependent and requires p300 docking to the N terminus of p53 presumably via the IBiD and/or IHD domains of p300. This data provides an alternate   5-8). Relative acetylation was quantitated by Western blot using an anti-p53 acetylation antibody and normalized to total p53 protein using the monoclonal antibody DO-1. B, p53 acetylation by p300 is not stimulated by the addition of nonspecific DNA. Acetylation reactions were carried out with p300 and p53 for 10 min in the presence of acetyl-CoA and increasing amounts of nonspecific DNA (NS-DNA; lanes 5-8 versus lanes [1][2][3][4]. A control including consensus site DNA (pG-DNA) is in lane 9. Relative acetylation was quantitated by Western blot using an anti-p53 acetylation antibody and normalized to total p53 protein using the monoclonal antibody DO-1. C, p53 acetylation by p300 is stimulated by equimolar amounts of p53 protein to consensus site DNA. Acetylation reactions were carried out with p300 and p53 for 10 min in the presence of acetyl-CoA without DNA (lane 1) and increasing amounts of consensus site oligonucleotide DNA (pG-DNA, from 1 to 200 ng, lanes [2][3][4][5][6][7][8][9][10][11][12]. From 10 to 15 ng of 26-mer oligonucleotide DNA gave rise to maximal acetylation of p53 that is at the 1:1 molar ratio of DNA⅐p53 protein tetramer in the reaction. Relative acetylation was quantitated by Western blot using an anti-p53 acetylation antibody and normalized to total p53 protein using the monoclonal antibody DO-1. D, DNA does not stimulate histone acetylation by p300. Acetylation reactions were carried out as described in C but with the inclusion of histone H4 in place of p53 and without DNA (lane 1) or with the addition of p53 consensus site DNA (lanes 2-6). The top panel measures acetylated histone and the bottom panel measures total histone protein levels.
model for acetylation compared with other studies showing that acetylation can either stimulate the latent DNA binding activity of p53 (18), have no effect on p53 DNA binding (28), or block MDM2-dependent ubiquitination of p53 (32). What then is the role of the acetylation of DNA-bound p53? One previous study has suggested a clue, as acetylation of MyoD by CBP (in the absence of DNA) can promote a high salt-resistant stable complex between the two proteins presumably through the bromo-homology domain (29). The stability of the p300⅐p53 complex was therefore quantitated without or with acetyl-CoA using the ELISA solid phase assay to determine whether p300 dissociates from p53 after acetylation in the absence or presence of consensus site DNA. In contrast to the immunoprecipitation of p300⅐p53 complexes that is a very sensitive, but rapid assay not allowing significant time for dissociation of weak protein-protein complexes (Fig. 7), the ELISA measures very stable complex formation because 20 -30-fold longer wash times permit weakly bound proteins to dissociate. Using purified p300 and p53 proteins, we show that small oligonucleotides containing the p53 consensus site (Fig. 9A) or supercoiled plasmid DNA containing the p53 consensus site (Fig.  9B), can promote a striking stabilization of the p300⅐p53 complex after the addition of acetyl-CoA (Fig. 9, A and B, lane  12), relative to nonspecific DNA (Fig. 9, A and B, lane 11  versus 12). As controls for this quantitation, acetyl-CoA-de-  2-4) was purified from insect cells and incubated with p300 and supercoiled plasmid containing the p53-consensus site (pG13, lanes 4) or nonspecific DNA (pMG13, lanes 3). Relative acetylation was quantitated by Western blot using an anti-p53 acetylation antibody and normalized to total p53 protein using the monoclonal antibody DO-1. B-E, in vivo DNAdependent acetylation of p53 by p300. pCMV-P53 (lanes 1 and 2) or pCMV-p53-6KR (lanes 3 and 4; mutated acetylation sites) were cotransfected into p53 Ϫ/Ϫ cells in the presence of consensus site plasmid DNA (pG13-CAT, lanes 2 and 4) or plasmid DNA without the consensus site (pMG13-CAT, lanes 1 and 3). The amount of total p53 in each transfection was quantified by direct immunoblotting with the monoclonal antibody DO-1. B and C, IP p300 panel: endogenous p300 protein was immunoprecipitated with an anti-p300 antibody and levels of total p53 protein (C) or acetylated p53 bound to p300 (B) were quantitated by immunoblotting, as indicated, with either the acetylation-specific antibody or a p53 protein antibody to normalize to the total p53 protein. D and E, IP p53 panel: the total p53 protein was immunoprecipitated with a mixture of ICA-9 and DO-1 monoclonal antibodies and the total levels of acetylated p53 (D) or p53 protein (E) were quantitated by immunoblotting, as indicated, with either the acetylation-specific antibody or a p53 protein antibody to normalize to the total p53 protein. protein levels. B, histone acetylation is not inhibited by the Ser 20 phosphopeptide derived from p53. Acetylation reactions were carried out as in A but with 1 g of histone H4 as the substrate for p300. The insensitivity of histone acetylation to the p300 Ser 20 phosphopeptide derived from p53 indicates that p53 acetylation inhibition by this peptide is not necessarily allosteric. C, p53 acetylation is inhibited using the ELISA p53 capture assay. Reaction products from A were incubated in ELISA wells with monoclonal antibody ICA-9 to capture p53 protein (30). The amount of acetylated p53 captured was quantitated by incubating the wells with polyclonal antibody to acetyl-p53 or to total p53 protein and depicted in relative light units (ac-p53/total p53). pendent p53 acetylation was stimulated by consensus site DNA in oligonucleotide or plasmid form (Fig. 9, A and B, lane  9 versus 8 and lane 2 versus 3). Furthermore, p300 binding to p53 (Fig. 9, lane 5) was de-stabilized by acetyl-CoA in reactions with nonspecific DNA (Fig. 9, A and B, lane 11 versus 5). By contrast, acetyl-CoA stabilized the p300⅐p53 complex when p53 is pre-bound to consensus site DNA (Fig. 9, A and  B, lane 12 versus 11). The docking-dependent acetylation of DNA-bound p53 that results in subsequent stabilization of the p300⅐p53 complex is consistent with the data suggesting that one role of acetylation is to allow p300 to nucleate transcriptional coactivators like TRAPP or BRG-1 and to promote chromatin remodeling. DISCUSSION The acetylation of histone and nonhistone chromosomal proteins is critical for the fine-control of gene expression and the characterization of how p300 recognizes and acetylates these classes of proteins will help to further define how gene expression is controlled. Histone acetylation reduces the extent of nucleosome assembly presumably by minimizing electrostatic DNA-protein interactions thus assisting in stabilizing the RNA polymerase preinitiation and transcription complex. By contrast, acetylation of sequence-specific transcription factors including E2F, p53, and MyoD have been shown to enhance their affinity for specific sequences in promoter regions (44). The molecular mechanism of p300 binding to and acetylation of sequence-specific transcription factors has not been defined. The relatively large size of p300 suggests that an intrinsically complex intradomain and/or interdomain communication exists between itself and the variety of target proteins at a promoter. The interferon-responsive enhanceosome is one such example of multiprotein nucleation at a promoter (45). Given that molecular reactions in vivo likely involve large protein complexes, or protein machines (46), it is likely that p300 will be regulated and function in a complex fashion. Furthermore, because acetylation may prove to be a covalent modification as important as phosphorylation mechanistic insight into acetylation function and regulation is of fundamental importance. In this report, we have assembled in vitro the p300⅐p53 complex, defined co-factors that modulate the acetylation of p53, and reconstituted a docking-dependent and DNA-dependent acetylation reaction required to clamp the p300⅐p53 complex.
The oligomeric nature of p53 provides a unique model with which to define conformational elements that modulate the binding and acetylation of a target protein by the transcriptional co-activator p300. The N-terminal BOX-I domain of p53 was originally shown to contain a p300-binding site, as mutation of this region produces a transcriptionally inert protein.
The first evidence for a multidomain component for the interaction between p53 and p300 came from data showing that phosphorylation of p53 at Ser 15 by DNA-PK stimulates acetylation in the C terminus (11). Subsequent studies have shown that phosphorylation of p53 in the BOX-I domain at the CHK2 Thr 18 and Ser 20 phosphorylation sites can stabilize the p300⅐p53 protein complex (14). As different class of p53-activating kinases target the Ser 15 , Thr 18 , or Ser 20 residues (ATM, CK1, and CHK2, respectively), these phosphorylation events provide a method for kinase signaling networks to regulate gene expression by altering the stability of the p300⅐p53 complex (14). Developing a consensus phosphate-binding motif for p300 that is distinct from the contact sites involved in the MDM2⅐p53 complex (Fig. 1) explains why kinase phosphorylation can differentially affect p300⅐p53 and MDM2⅐p53 complex stability and led to the identification of phosphorylated-UBF as a p300-binding sites (Fig. 1, C and D).
Our mapping of one of the p53-binding domains of p300 to IBiD is consistent with recent studies (30), whereas our identification of a second domain with homology to IBiD (IHD) was previously undefined. The original studies on mapping the p53-p300 interactions have focused on the use of glutathione S-transferase fused N-and C-terminal fragments of p53 in pull-down and immunoprecipitation experiments. For example, the C-terminal region of p300 (amino acids 1990 -2414) containing the Gln-rich domain and IBiD/SRC-1 binds p53 (amino acids 1-72) in vitro; consistent with our findings that p300-GAL4-C2 binds p53 (amino acids 1945-2414). Other groups have shown interactions between p53 and the KIX domain and C/H1 or C/H3 domain exclusively (47), which are not observed in vitro using our p300-phospho-ligand binding assays. However, weak binding of the C/H3␦ domain of p300 to the N-terminal activation domain of p53 was observed in cells FIG. 9. Acetyl-CoA stabilizes the p300⅐p53 protein complex only in the presence of consensus-site DNA. A, acetyl-CoA-dependent p300⅐p53 protein complex stabilization by consensus site oligonucleotides. Insect cell-purified human recombinant p53 was captured in a microtiter well with the ICA-9 anti-p53 monoclonal antibody (3) in the presence of consensus site pG-DNA or NS-DNA oligonucleotides. The captured p53 isoforms were incubated with buffer containing insect cell-purified p300 protein in the absence (lanes 1-6) or presence of 1 mM acetyl- CoA (lanes 7-12). The level of p53 acetylation: (Ϫ/ϩ)-acetyl-CoA, (Ϫ/ϩ)-NS-DNA, or (Ϫ/ϩ)-pG-DNA was quantitated using the antiacetylation antibody (lanes 1-3 and 7-9). The amount of p300 protein stably bound to p53 protein: (Ϫ/ϩ)-acetyl-CoA, (Ϫ/ϩ)-NS-DNA, or (Ϫ/ϩ)-pG-DNA was determined using an antibody to p300 and quantitated as p300 bound/total p53 using enhanced chemiluminescence expressed as relative light units (RLU). B, acetyl-CoA-dependent p300⅐p53 protein complex stabilization by consensus site supercoiled plasmid DNA. Insect cell-purified human recombinant P53 was captured in a microtiter well with the ICA-9 anti-p53 monoclonal antibody (3) in the presence of pG13-CAT plasmid DNA or pMG13-CAT plasmid DNA. The captured p53 isoforms were incubated with buffer containing insect cell-purified p300 protein in the absence (lanes 1-6) or presence of 1 mM acetyl-CoA (lanes 7-12). The level of p53 acetylation: (Ϫ/ϩ)acetyl-CoA, (Ϫ/ϩ)-NS-DNA, or (Ϫ/ϩ)-pG-DNA was quantitated using the anti-acetylation antibody (lanes 1-3 and 7-9). The amount of p300 protein stably bound to p53 protein, (Ϫ/ϩ)-acetyl-CoA, (Ϫ/ϩ)-NS-DNA, or (Ϫ/ϩ)-pG-DNA was determined using an antibody to p300 and quantitated as p300 bound/total p53 using enhanced chemiluminescence expressed as relative light units. The notable effects include the dissociation of p300 from p53 after acetylation in reactions with nonconsensus site DNA (lanes 11 versus 5), whereas a stabilization of the p300⅐p53 complex occurs after acetylation in reactions with consensus site DNA (lanes 12 versus 11). (Fig. 3A), but not in vitro (Fig. 2). Additionally, a more recent study has shown that the minimal IBiD domain (or SRC-1) is another peptide-binding domain within p300 that can also interact with p53 (31). Such differences may be explained by the fact that previous p300-p53 interactions were mapped by protein binding assays, or nonphosphorylated p53 protein domains, whereas our assays utilize a combination of phosphopeptide binding assays and full-length p300 protein to measure p53⅐p300 complex stability. The use of truncations of p300 protein may also complicate studies, as we have previously shown that the C-terminal acetyltransferase domain of p300 displays significant binding activity toward the Ser 15 -BOX-I activation domain but little activity toward Ser 20 -BOX-I domain fragments (14).
The observation that DNA promotes p53 acetylation brings to mind a recent discussion of allosteric regulation of transcription regulators through specific DNA binding (48). In the latter report, it was proposed that DNA response elements contain information that is interpreted by bound transcription factors. In particular, DNA could act as an "allosteric ligand whose binding alters the regulator's affinity for other ligands, such as co-activators or co-repressors." Presumably, this allosteric control of p53 acetylation is brought about by a conformational change in p53 after DNA binding. The conformational changes in p53 that occur upon specific DNA binding remain to be determined. Additionally, after acetylation of p53, p300 binding is further stabilized by consensus site DNA (Fig. 9) identifying a role for acetylation as a stabilizer of protein-protein interactions. The docking dependence in p300 acetylation of a substrate further addresses the poorly understood nature of modularity in a native transcription regulator. Transactivation domains are defined, in part, as minimal motifs that can drive transcription when fused to a heterologous DNA-binding protein, like GAL4, and such domain swap experiments are usually used as arguments for a modular or scaffold nature of a transcription factor. However, such domain swap experiments underestimate the complexity of a transcription factor DNAbinding domain and transactivation domain and do not rule out intramolecular or allosteric associations on the native protein.
For example, the USF2 transactivation domain activates transcription when fused to its native DNA-binding domain, but when the USF2 transactivation domain is fused to GAL4 it is inactive or even repressive (49). ATF-2 and OCT-4 transactivation domains are thought to function through an interaction with their DNA-binding domains, as defined using the OCT-4 transactivation domain and the chimera of the DNA-binding domain (50,51). The N terminus of p53 has an activation motif that drives transcription when fused to the GAL4 DNA-binding domain in the yeast transactivation system (52). However, when the p53 transactivation domain is fused to a p53 DNAbinding domain mutant, this blocks transcription (58), further indicating that the p53 activation domain is allosterically regulated by its DNA-binding domain. Although such widespread evidence that the allosteric control of transactivation by DNAbinding domains is lacking, there is growing data showing that transcription factors change their conformation when DNAbound, based on proteolytic fingerprints, CD, and NMR studies. The observation that DNA binding by p53 affects its acetylation by p300 indicates that the DNA-binding domain is pivotal in linking the activation domain to acetylation and coactivator recruitment. A recent study has confirmed using NMR that the conformation of p53 changes upon DNA binding; presumably this conformational change in the core domain and surface of p53 also creates a conformationally stable binding site for p300 or releases the acetylation motif into the acetyltransferase active site.
In summary, the use of biochemical techniques to reconstitute the p300⅐p53 complex using active forms of full-length p300 has identified specific stages in the assembly reaction that include a docking-dependent and consensus site DNAdependent acetylation of p53. The docking dependence in the DNA-dependent acetylation of p53 identifies a conformational constraint imposed on the p300-acetylation reaction at a promoter. Identification of the mechanism of conformational regulation of p53 acetylation may reveal novel factors that modulate acetylation of the DNA-bound transcription factor.