Research Critique Paper

Cell, Vol. 90, 595–606, August 22, 1997, Copyright 1997 by Cell Press

Activation of p53 Sequence-Specific DNA Binding by Acetylation of the p53 C-Terminal Domain

Wei Gu and Robert G. Roeder* seems critical for normal cell growth and development, as the activity of p53 is dramatically elevated in theLaboratory of Biochemistry and Molecular Biology cells following DNA damage or during spermatogenesisThe Rockefeller University (Gottlieb and Oren, 1996). The precise mechanism byNew York, New York 10021 which p53 is activated by cellular stress is of intense interest. Although it could involve an increase in the p53 protein level, it is generally thought to involve mainlySummary posttranslational modifications of p53 (Kastan et al., 1991; Ko and Prives, 1996). For example, at low doses ofThe tumor suppressor p53 exerts antiproliferation UV irradiation, p53-dependent transcription is activatedeffects through its ability to function as a sequence- without any detectable increase in the p53 protein levelspecific DNA-binding transcription factor. Here, we (Hupp et al., 1995). Posttranslational modifications ofdemonstrate that p53 can be modified by acetylation the carboxyl terminus of p53 have been shown to play anboth in vivo and in vitro. Remarkably, the site of p53 important role in controlling p53-specific DNA bindingthat is acetylated by its coactivator, p300, resides in a (Hupp et al., 1992; Meek, 1994; Hansen and Oren, 1997;C-terminal domain known to be critical for the regula- Levine, 1997). Thus, modification of a highly basic regiontion of p53 DNA binding. Furthermore, the acetylation within the carboxy-terminal 30 amino acids of the p53of p53 can dramatically stimulate its sequence-spe- by phosporylation, antibody binding, or deletion of thiscific DNA-binding activity, possibly as a result of an region can convert p53 from an inert to an active formacetylation-induced conformational change. These for DNA binding (Ko and Prives, 1996). Furthermore, p53observations clearly indicate a novel pathway for p53 C-terminal peptides stimulate DNA binding of full-lengthactivation and, importantly, provide an example of an p53 in trans (Hupp et al., 1995; Jayaraman and Prives,acetylation-mediated change in the function of a 1995; Lee et al., 1995; Selivanova et al., 1997). Thesenonhistone regulatory protein. These results have sig- observations, which have led to a model for allostericnificant implications regarding the molecular mecha- regulation of p53, are consistent with the existence ofnisms of various acetyltransferase-containing tran- distinct cellular signaling pathways capable of modulat-scriptional coactivators whose primary targets have ing the conversion between latent and activated formsbeen presumed to be histones. of p53 (Hupp and Lane, 1994).

Recently, we and others have demonstrated that CBP/ p300, which also is implicated in cell proliferation andIntroduction differentiation (reviewed in Shikama et al., 1997), acts as a coactivator for p53 and potentiates its transcriptionalThe p53 tumor suppressor protein exerts antiprolifera- activity in vivo (Gu et al., 1997; Lill et al., 1997). Thetion effects, including growth arrest and apoptosis, in N-terminal activation domain of p53 physically interacts

response to various types of stress (Gottlieb and Oren, with the carboxy-terminal portion of the CBP protein

1996; Ko and Prives, 1996; Levine, 1997). The biochemi- both in vitro and in vivo. In transfected SaoS-2 cells,

cal activity of p53 that is required for tumor suppression, CBP/p300 dramatically potentiates activation of a p53

and presumably the cellular response to DNA damage, target gene. Since sequence-specific transcriptional ac-

involves the ability of the protein to bind to specific DNA tivation by p53 correlates well with the ability of p53 to

sequences and to function as a transcription factor (El- suppress cell growth (Crook et al., 1994), the cell growth

Deiry et al., 1992; Funk et al., 1992; Pietenpol et al., suppression function of CBP/p300 may act, at least in 1994). The importance of the DNA-binding property of

part, through synergistic activation with p53 on its target p53 is underscored by the fact that the vast majority of

genes. p53 mutations derived from tumors usually map within

CBP and p300 exhibit strong sequence similarity and the domain for sequence-specific DNA binding (Holl- similar functions, such as binding to common DNA- stein et al., 1991; Ko and Prives, 1996). Overexpression binding transcription activators that include nuclear hor- of p53 in cells clearly activates, through consensus p53 mone receptors (Chakravarti et al., 1996; Kamei et al., binding sites, a number of genes that have been impli- 1996), CREB (Chrivia et al., 1993; Kwok et al., 1994; cated as functional targets in p53-induced cell growth Arany et al., 1995; Lundblad et al., 1995), c-Jun/v-Jun repression or apoptosis. These include GADD45 (Kastan (Arias et al., 1994; Bannister and Kouzarides, 1995), et al., 1992), mdm2 (Barak et al., 1993; Wu et al., 1993), c-Myb/v-Myb (Dai et al., 1996; Oelgeschlager et al., WAF/p21/CIP1 (El-Deiry et al., 1993), cyclin G (Okamoto 1996), Stat-1 (Zhang et al., 1996), Stat-2 (Bhattacharya and Beach, 1994), IGF-BP3 (Buckbinder et al., 1995), et al., 1996), c-Fos (Bannister and Kouzarides, 1995), and Bax (Miyashita and Reed, 1995). MyoD (Yuan et al., 1996), and NFkB p65 (Gerristen et

Regulation of p53 function remains less well under- al., 1997). Although the precise mechanisms by which stood. p53 is a short-lived protein that is maintained at these activators stimulate the transcriptional machinery low, often undetectable levels in normal cells (Ko and through CBP/p300 remain unclear, recent discoveries Prives, 1996; Levine, 1997). Tight regulation of p53 that p300/CBP and an interacting factor (P/CAF) both

have histone acetyltransferase activities (Bannister and Kouzarides, 1996; Ogryzko et al., 1996; Yang et al., 1996)*To whom correspondence should be addressed.

Cell 596

suggest that CBP/p300 may play a distinct role in tran- a nonspecific interaction between p53 and [14C]-acetyl- CoA or p53 autoacetylation.scriptional regulation through histone acetylation.

Acetylated histones are a characteristic feature of Furthermore, by measuring [14C]-acetate incorpora- tion in a filter binding assay (Ogryzko et al., 1996), wetranscriptionally active chromatin. Hyperacetylated his-

tones accumulate within particular active chromatin do- not only confirmed that p53 can be acetylated by p300, but also that the acetylation of p53 is as strong as themains, whereas hypoacetylated histones accumulate

within transcriptionally silenced domains (reviewed in acetylation of core histones by p300 under the same conditions (Figure 1C). Thus, we have demonstratedTurner, 1993; Brownell and Allis, 1996; Wolffe and Pruss,

1996). Thus far, all core histone proteins can be variably that p53 is a novel, and genuine, substrate for the p300 acetyltransferase in vitro.acetylated by a number of proteins, designated histone

acetyltransferases (HATs), that include GCN5, P/CAF, p300/CBP, and TAFII250 (Bannister and Kouzarides, The p53 Site Acetylated by p300 Is Located at the 1996; Kuo et al., 1996; Mizzen et al., 1996; Ogryzko et al., C-Terminal DNA-Binding Regulatory Domain 1996; Yang et al., 1996). Importantly, a global increase in The p53 protein can be roughly divided into three distinct core histone acetylation does not necessarily induce functional domains (Ko and Prives, 1996): an amino- widespread transcription. In some instances, an in- terminal domain that contains the transcriptional activa- crease in histone acetylation correlated with a decrease tion domain (residues 1–43), a central core that contains in transcriptional activity (reviewed in Pazin and Kado- the sequence-specific DNA-binding domain (residues naga, 1997), suggesting that the relation between acetyl- 100–300), and the multifunctional carboxy-terminal do- transferase function and transcriptional activity may be main (residues 300–393) (Figure 2A). To map the more complex than a matter of histone acetylation and p300-dependent acetylation site of p53 in vitro, GST- consequent transcriptional activation. Indeed, the pos- p53(1–73), GST-p53(100–290), and GST-p53(290–393) sibility that these acetyltransferase activities might have fusion proteins were expressed in bacteria and purified other protein substrates has not been reported. tonear homogeneity on glutathione-agarose beads (Fig-

In support of this possibility, we demonstrate that ure 2D). As indicated in Figure 2B, only GST-p53(290– p300 can directly modify the C-terminal domain of p53 393) was detectably acetylated by p300, suggesting that by acetylation and, remarkably, that this modification the acetylation site is located in the C-terminal domain. not only is position-specific, but also dramatically stimu- The p53 carboxyl terminus can be further subdivided lates the sequence-specific DNA-binding activity of p53. into three regions, which include (Figure 2A) a flexible This discovery has important implications for current linker (residues 300–320) that connects the DNA-binding views regarding the molecular mechanisms by which domain to the tetramerization domain, the tetrameriza- acetyltransferase-containing coactivators effect tran- tion domain itself (residues 320–360), and, at the ex- scriptional activation. treme carboxyl terminus, a 30 amino acid stretch that

plays a very important role in the regulation of p53 DNA binding (residues 363–393). Purified GST fusion proteinsResults containing these subdomains (Figure 2E) were tested for acetylation by p300. Although each subdomain containsIdentification of p53 as a Bona Fide Substrate several lysines, which are potential acetylation sites forfor p300 Acetyltransferase acetyltransferases, only the C-terminal 30 amino acidTo test whether p53 could be specifically acetylated domain was acetylated (Figure 2C).by p300, the following proteins were expressed with

N-terminal Flag epitopes in bacteria and purified to near homogeneity on M2-agarose affinity columns: a region p53 Is Acetylated In Vivo

p53 is a short-lived protein that is maintained at low,of p300 (residues 1195–1810) containing the HAT do- main (Ogryzko et al., 1996), full-length p53 (Gu et al., often undetectable, levels in normal cells (Ko and Prives,

1996). Compared with histones, it is extremely difficult1997), Max (Blackwood and Eisenman, 1991), USF (USF-1) (Gregor et al., 1990), and the p300/CBP-inter- to purify sufficient p53 protein from normal cells to test

whether acetylated p53 exists in vivo. Furthermore,acting CREB (Chrivia et al., 1993). These highly purified recombinant proteins were used in the protein acetyl- overexpression of wild-type p53 induces apoptosis in

many cell types, thus making it impossible to obtaintransferase assay to avoid possible contamination by either histones or other acetyltransferases. p53–stably transfected cell lines for this purpose.

Here, we tested whether p53 is acetylated in tran-Individual reactions containing 1 mg of each substrate were incubated with [14C]-acetyl-CoA and 100 ng of re- siently transfected cells. As indicated in Figure 3A, p53

can be transiently overexpressed shortly after transfec-combinant p300 for 1 hr at 308C. The products were analyzed by autoradiography of acetylated reaction tion (Figure 3A, lane 2). These cells were immediately

labeled by [3H]-sodium acetate (1 mci/ml) for only 1 hrproducts resolved on SDS–PAGE gels. As shown in Fig- ure 1A, the p53 protein was specifically labeled by to limit the protein labeling toposttranslational modifica-

tions. Figure 3B shows that the p53 protein immunopuri-[14C]-acetyl-CoA, whereas no signals were detected in reactions containing USF, Max, or CREB. Labeling of fied under very high stringency conditions (see Experi-

mental Procedures) by anti-p53 monoclonal antibodyp53 was completely dependent on the presence of both [14C]-acetyl-CoA and p300 (Figure 1B), thus excluding (D0-1) from the p53-transfected cell extract (Figure 3B,

lane 2) was specifically labeledby [3H]-acetate.No signalthe possibility that the labeled p53 resulted from either

Activation of p53 DNA Binding by Aceytlation 597

Figure 1. In Vitro Acetylation of p53 by p300 Acetyltransferase

(A) Acetylation of p53 relative to other activators. Following incubation of purified activators (1 mg) with [14C]-acetyl-CoA and 100 ng of recombinant p300(1195–1810) for 1 hr at 308C, reaction products were separated by SDS–PAGE and visualized by autoradiography. Lane 1, p53; lane 2, CREB; lane 3, USF; lane 4, Max. (B) Substrate and enzyme requirements. Reaction products were separated by SDS-PAGE and visualized by autoradiography following incubation of various combinations of p53 (1 mg), p300(1195–1810) (100 ng), and [14C]-acetyl-CoA. Lane 1, p53; lane 2, p53 and [14C]-acetyl- CoA; lane 3, all components; lane 4, p300(1195–1810) and [14C]-acetyl-CoA. (C) Comparative acetylation of p53 and core histones. Amounts (1 mg) of human core histones, p53, Max, USF, and CREB were incubated with [14C]-acetyl-CoA in the presence or absence of p300(1195–1810). The [14C]-acetate incorporation of each protein was measured by the filter binding assay.

was detected either in proteins affinity-purified from the residues in vivo (Hendzel et al., 1991; Kijima et al., 1993; Vettese-Dadey et al., 1996) and, to date, has limitedsame extract by anti-hemagglutinin (HA) control anti-

body (Figure 3B, lane 1) or in proteins immunoprecipi- further identification of in vivo acetylation positions. tated with the same p53 antibody from nontransfected SaoS-2 cells that were similarly labeled with [3H]-sodium The Acetylation Site of p53 Is

Evolutionarily Conservedacetate (Figure 3B, lane 4). Interestingly, a C-terminal truncated p53 (p53D370), which lacks the last 24 amino Recent studies have demonstrated position-specific

acetylation of histone H4 by distinct acetyltransferasesacids of the protein (Crook et al., 1994; Marston et al., 1994), is expressed at comparable levels in the cells by (Brownell and Allis, 1996). Thus, while cytoplasmic ace-

tyltransferases for histone deposition and chromatintransient transfection (Figure 3A, lane 3); however, the signal of acetylated p53D370, purified by the same assembly modify positions 5 and 12, the transcriptional

coactivator GCN5 preferentially modifies positions 8method as acetylated wild-type p53, is significantly re- duced (Figure 3B, lane 3), suggesting that theC terminus and 16 (Kuo et al., 1996). To similarly determine the

positions of p53 acetylated by p300, a synthetic peptideof p53 may be mainly responsible for p53 acetylation in vivo. This also provides evidence indicative of similar that contains the carboxy-terminal 26 amino acids was

incubated with p300 and [14C]-acetyl-CoA. The ace-sites of acetylation in vivo and in vitro. Taken together, these results indicate that p53 is in- tylated peptide was purified and subjected to amino-

terminal peptide sequencing. The amount of [14C]-ace-deed acetylated in vivo. However, under the conditions employed, the proportion of acetylated p53 remains low. tate incorporated in each residue indicated the degree

of acetylation of this residue. As shown in Figure 4A, theThis may reflect the action of strong deacetylase activi- ties during the purification or rapid turnover of the acetyl [14C]-acetate incorporation levels for various positions

Cell 598

Figure 2. The p53 Acetylation Site Is Located at the C-Terminal DNA-Binding Regulatory Domain

(A) shows schematic representations of p53 and its functional domains. The p53 fragments indicated (A) were expressed as GST fusion proteins, purified to near homogeneity, and analyzed by resolution on SDS–PAGE gel and Coomassie blue staining (D and E). Corresponding GST fusion proteins (2.5 mg) were assayed for acetylation by p300 as described in Figure 1A, with reaction products analyzed by SDS–PAGE and autoradiography (B and C).

indicate two peaks located at K373 and K382, respec- due to the specific experimental approach (Ogryzko et al., 1996). We find that five lysines of p53 (K370, K372,tively, suggesting that these two lysines are preferen-

tially acetylated. Lysines K370, K372, and K381 also K373, K381, and K382) can beacetylated by p300 (Figure 4B),and that K373and K382are preferentially acetylatedshowed significant, but lower, levels of acetylation,

whereas K386 appeared completely unacetylated in on the basis of the incorporated [14C]-acetate activities. In comparison with the acetylated sites identified fromcomparison with the background. These results indicate

that p53 acetylation is highly position-specific. the GCN5-mediated acetylation of histone H4 by a simi- lar approach (Kuo et al., 1996), we note that both thep53 is an evolutionarily conserved protein, particularly

among vertebrate species (Soussi et al., 1990). There is pattern and the position of the preferentially acetylated lysines are very similar (Figure 4C). Taken together,one conserved region in the amino-terminal activation

domain, and there are four conserved regions in the these results indicate that the acetylation site of p53 is not only specific, but also evolutionarily conserved.sequence-specific DNA-binding domain, while no con-

served region has been found in the C terminus (Soussi et al., 1990; Ko and Prives, 1996). Surprisingly, a com- parison of the acetylation site of human p53 with the Acetylation of p53 Activates Its Sequence-Specific

DNA Bindingsame region from other vertebrates reveals that the positions of all of the acetylated lysines are highly con- A number of studies indicate that the C-terminal 30

amino acids of p53 play a key role in the regulation ofserved, including K370, K372, K373, K381, and K382 (Figure 4B), whereas the nonacetylated K386 is not con- sequence-specific DNA binding by p53 (Hupp et al.,

1992; Hupp and Lane, 1994; Meek, 1994; Jayaramanserved (Figure 4B). These results suggest that p53 modi- fications by acetylation may remain highly conserved and Prives, 1995). Since the acetylation site was mapped

to this domain, it was important to examine the effectthroughout evolution. Interestingly, an arginine (R379) and its position are also highly conserved, indicating of acetylation on site-specific DNA binding by p53. To

exclude the possibility that other modifications, such asthat besides lysine, arginine may also play a role in acetylation. phosphorylation (Meek, 1994) or glycosylation (Shaw et

al., 1996), may interfere with this effect, we used highlyp300 can acetylate all four N-terminal lysines (K5, K8, K12, and K16) in histone H4. However, differential ace- purified, bacterially produced full-length human p53 in

this assay. As reported previously (Hupp et al., 1992),tylation of these sites has not been observed, perhaps

Activation of p53 DNA Binding by Aceytlation 599

Figure 3. In Vivo Acetylation of p53

(A) Ectopic expression of p53 and p53D370 in Saos-2 cells. Extracts from control cells (lane 1), p53-transfected cells (lane 2), or p53D370- transfected cells (lane 3) were subjected to SDS–PAGE and a West-

Figure 4. Identification and Evolutionary Conservation of the Ace-ern blot analysis with anti-p53 monoclonal antibody DO-1. tylation Site of p53(B) Radiolabeling of ectopic p53 and p53D370 by [3H]-acetate. Ex- (A) A synthetic peptide (5 mg) corresponding to the carboxy-terminaltracts from control cells (lane 4), p53-transfected cells (lanes 1–2), 26 amino acids of p53 was incubated with 500 ng of p300(1195–or p53D370-transfected cells (lane 3) pulsed with [3H]-acetate were 1810) and [14C]-acetyl-CoA for 2 hr at 308C. The peptide was purifiedsubjected to immunoprecipitation with anti-p53 monoclonal anti- and subjected to N-terminal peptide sequencing (X axis; single-body (DO-1) (lanes 2–4) or anti-HA monoclonal antibody (lane 1), letter amino acid code). The amount of [14C] incorporation in eachand immunoprecipitated proteins were analyzed by SDS–PAGE and residue (CPM) was determined.autoradiography. (B) Alignment of the C-terminal regions of p53 from human (residues 367–388), monkey (residues 367–388), mouse (residues 361–382), rat (residues 365–386), Xenopus laevis gene A (residues 334–355),bacterially produced p53 is virtually inactive in site-spe- and gene B (residues 333–354), rainbow trout (residues 367–388),cific DNA binding in an electrophoretic mobility shift and chicken (residues 345–365).assay (Figure 5A, lane 2 versus lane 1), but the binding (C) Comparison of the acetylation sites in human p53 with those in

can be strongly and selectively activated by anti-p53 yeast histone H3 (residues 7–17), yeast histone H4 (residues 1–12), monoclonal antibody PAb421 (Figure 5A, lane 5) relative and Tetrahymena (Tet.) histone H4 (residues 1–10 in the upper pep- to anti-p53 monoclonal antibody DO-1 (Figure 5A, lane tide and residues 9–18 in the lower peptide). The asterisk indicates

the preferentially acetylated residue.6), which nonetheless recognizes the whole p53/DNA complex, and other control monoclonal antibodies (Fig- ure 5A, lanes 3 and 4).

Surprisingly, a dose response analysis showed that a 200-fold molar excess of an unlabeled oligonucleotide containing the wild-type p53 binding site but not byacetylation by p300 in vitro could dramatically increase

the DNA-binding activity of fixed amounts of p53 (Figure comparable amounts of an oligonucleotide containing a mutated binding site (Figure 5D, lanes 4–6).5B). As indicated in Figure 5C, activation of the DNA

binding by acetylation reached levels as high as 20- to Interestingly, the acetylated form of p53 was super- shifted both by anti-p53 monoclonal antibody PAb42130-fold over untreated p53 binding. No change in DNA-

binding activity was found after treatment of p53 with (epitope in the C-terminal domain of p53) and by anti- p53 monoclonal antibody DO-1 (epitope in the N-termi-acetyl-CoA or p300(1195–1810) alone (Figure 5D, lanes

1–3), indicating that the stimulation of DNA binding is nal domain of p53) (Figure 4D, lanes 8 and 9), similar to what was observed with the unmodified form (Figurespecifically due to the acetylation of p53. Furthermore,

the strongly enhanced DNA-binding activity of p53 is 4A, lanes 5 and 6). Whereas DO-1 did not enhance bind- ing either of acetylated p53 (Figure 4D) or unmodifiedsequence-specific since the binding was competed by

Cell 600

Figure 5. The Latent Sequence-Specific DNA-Binding Activity of p53 Is Activated by Acetylation

(A) Activation of bacterially produced p53 DNA binding by anti-p53 monoclonal anti- body PAb421. Human p53 (50 ng, lanes 2–6) was incubated with 0.2 ng of a 32P-labeled p53-binding oligonucleotide (GADD45, Kas- tan et al., 1992) in the absence of other pro- teins (lane 2) or with the addition of mono- clonal antibodies specific for c-Myc (lane 3), the hemagglutinin antigen (lane 4), or p53 (PAb421 in lane 5, and DO-1 in lane 6). The reaction mixtures were analyzed by EMSA. Lane 1 shows an analysis of the DNA probe without protein. The arrows indicatep53-anti- body/DNA complex (upper) and p53/DNA complex (lower), respectively. (B and C) DNA binding of unmodified versus acetylated forms of p53. Indicated quantities of unmodified p53 (lanes 2–5) and in vitro- acetylated p53 (lanes 6–9) were incubated in reaction mixtures containing 32P-labeled wild- type GADD45 oligonucleotides (0.2 ng), re- solved by electrophoreses, and visualized by autoradiography (B). The resulting protein– DNA complexes were quantitated (C) by phos- phorimaging using Image-Quant software. (D) Specificity of acetylated p53 binding to DNA. Recombinant p53 (25 ng) was incu- bated with 32P-labeled wild-type GADD45 oligonucleotides (0.2 ng) following prior incu- bation with no factors (lane 1), or1 mM acetyl- CoA (lane 2), or 5 ng p300 (lane 3). Acetylated p53 (25 ng) was incubated with 32P-labeled wild-type GADD45 oligonucleotide (0.2 ng) following prior incubation with no factors

(lane 4), a 200-fold excess of unlabeled wild-type (lane 5) or mutant (lane 6) GADD45 oligonucleotides, anti-HA monoclonal antibody (lane 7), anti-p53 monoclonal antibody PAb421 (lane 8), or anti-p53 monoclonal antibody (DO-1) (lane 9). Lane 10 contained the DNA probe without protein. The arrows indicate the p53-antibody/DNA complex (upper) and the p53/DNA complex (lower), respectively.

p53 (Figure 4A), PAb421 did enhance binding of ace- activate p53 DNA binding. To further confirm this hy- pothesis, we tested theeffect of acetylation on theabilitytylated p53 (Figure 4D). However, this latter effect was

minimal (z2-fold) compared to the very large effect (over of the C-terminal peptide to activate p53 in trans. As shown in Figure 6A, the unmodified form of the synthetic20-fold) of PAb421 on DNA binding by the unmodified

p53 (Figure 4A). This suggests that activation by anti- peptide derived from the C terminus strongly activated (up to 7.8-fold) the DNA-binding capability of the full-p53 antibody PAb421 and activation by acetylation may

involve similar molecular mechanisms. length p53 (Figure 6B), whereas the acetylation of the C-terminal peptide significantly (z2-fold) reduced its ability to activate p53 in the same assay (Figures 6A andInsight into the Mechanism of

Acetylation-Mediated Activation 6B). This result suggests that the acetylated C-terminal peptide has a reduced ability to interact with the coreThe specificity of the effect of the p53 C terminus on p53

sequence-specific DNA binding suggested that there domain and thus is less active in stimulating DNA bind- ing of the full-length p53 in trans.might be intramolecular interactions between the C ter-

minus and the central core domain (Hupp et al., 1995; Interestingly, previous studies indicated that substitu- tion of any of the five lysines (K370, K372, K373, K381,Jayaraman and Prives, 1995). It has been proposed that

this interaction locks the core into a conformation that and K382) with alanine dramatically reduces the ability of the synthetic C-terminal peptide to activate p53 inis inactive for DNA binding and that the core is then able

to adopt the active form when this tail–core interaction trans (Hupp et al., 1995). This is completely consistent with our discovery that acetylation of these five lysinesis disrupted by phosphorylation, antibody binding, or

deletion (Hupp and Lane, 1994; Ko and Prives, 1996). reduces the ability of the C-terminal peptide to activate p53 in trans.In addition, there appears to be a competitive disruption

of this interaction by synthetic C-terminal peptides that can activate p53 in trans (Hupp et al., 1995; Jayaraman Discussion and Prives, 1995).

The acetylation by p300 of these lysines in the C termi- p300/CBP has emerged as a transcriptional coactivator for a broad group of cellular DNA-binding regulatorynus of p53 leads to a neutralization of positive charges

that in turn could disrupt the tail–core interaction and proteins (reviewed in Shikama et al., 1997), including the

Activation of p53 DNA Binding by Aceytlation 601

suppress cell proliferation, to the associated effect on cell growth. Our discovery is also of special interest in relation to current models regarding the mechanism of transcriptional activation by a number of coactivators (including p300) with histone acetyltransferase activity, the assumption having been that histone modifications play the major (if not exclusive) role in the function of these coactivators.

Activation of p53 by Acetylation Transactivation by p53 is tightly regulated in vivo and correlates well with its binding to cognate DNA se- quence elements (El-Deiry et al., 1992; Funk et al., 1992; Hansen and Oren, 1997). Most oncogenic mutants of p53 have lost their ability to function effectively in a sequence-specific DNA-binding manner, suggesting that these activities of p53 are critical for its tumor suppres- sor function. Furthermore, considerable evidence indi- cates that posttranslational modifications of the car- boxyl terminus of p53 play a key role in converting p53 from an inert to an active form for DNA binding (Ko and Prives, 1996).

The specific mechanisms involved in p53 latency and activation after posttranslational modification remain unclear. According to the allosteric model (Hupp et al., 1992, 1995; Jayaraman and Prives, 1995), the C-terminal tail of p53 interacts with the core DNA-binding domain and locks the DNA-binding domain into an inactive con- formation. Significant in this regard is our demonstration of p300-mediated acetylation of specific lysine residues in the C terminus of p53, with a consequent activation of DNA binding. The acetylation-mediated neutralization of positive charge could disrupt interactions between the C-terminal domain and the core domain, thus allowing the DNA-binding domain to adopt an activeFigure 6. Acetylation of p53 C-Terminal Peptide (364–389) Inhibits

Its Ability To Stimulate DNA Binding by Full-Length p53 conformation (Figure 7). This model is further substanti- (A) Effects of unmodified versus acetylated peptides on p53 binding ated by our observation that acetylation of the C-termi- to DNA. Human p53 (25 ng) was incubated with 32P-labeled wild- nal peptide abrogates its ability to activate DNA binding type GADD45 oligonucleotides (0.2 ng) in the absence (lane 1) or in by the full-length p53 in trans. As suggested in Figure the presence of the indicated amounts of unmodified (lanes 2, 3,

7, the unmodified C-terminal peptide could activate p53and 6) oracetylated (lanes 4 and 5) p53 C-terminal peptide (364–389). simply by competitive disruption of the core–tail interac-In lane 6, 25 ng of p53 was incubated with the probe in the presence tion, whereas the acetylated form that lacks the positiveof 0.50 nmol of the p53 C-terminal peptide (364–389) and 10 ng

of p300(1195–1810), indicating that contaminating p300(1195–1810) charges would also lack the ability to fully activate p53 from acetylated p53(364–389) has no effect on p53 DNA binding. in trans. Interestingly, the proposed role of acetylation Lane 7 contained the oligonucleotide probe and 0.50 nmol of p53 in reversing the p53 autoinhibitory mechanism is analo- C-terminal peptide (364–389) but no human p53 protein. The arrow gous to the role of phosphorylation in blocking the func- indicates the p53/DNA complex.

tion of autoinhibitory domains in several kinases (Engel(B) Quantitation of the p53 DNA binding data from panel A. Data et al., 1995; Mohammadi et al., 1996).were quantitated by phosphorimaging using Image-Quant software,

and representive results depict the average of three experiments The ability to activate the p53-mediated transcrip- with standard deviations indicated. tional function both in vitro and in vivo, by interaction

either with monoclonal antibody PAb421 (Hupp et al., 1995) or with C-terminal peptides (Selivanova et al.,tumor suppressor p53 (Gu et al., 1997; Lill et al., 1997). 1997), underscores the rate-limiting nature of the se-Following our earlier demonstration that p53 functions quence-specific DNA binding in p53 activation. Thesesynergistically with p300/CBP, we show here: (i) that findings also suggest the presence of a cellular poolp53 is a substrate for the acetyltransferase activity of of latent p53 that can be activated posttranslationally.p300; (ii) that acetylation of p53 is restricted to the C Serine phosphorylation of the carboxyl terminus hasterminus and activates the latent sequence-specific also been shown to stimulate the DNA binding of p53DNA-binding activity of p53; (iii) that p53 is acetylated in vitro (Hupp et al., 1992; Meek, 1994). However, elimi-in vivo, as well as in vitro; and (iv) that residues in the nation of one phosphorylation site (serine 386 in mouseacetylation site are evolutionary conserved. These re- p53 or serine 392 in human p53) by alanine substitutionsults are especially relevant to the in vivo transcription had no significant influence on p53-dependent trans-function of p53 and, since sequence-specific transcrip-

tional activation by p53 correlates well with its ability to activation in cells (Fiscella et al., 1994; Hall et al., 1996),

Cell 602

Figure 7. A Model for Activation of the Latent DNA-Binding Activity of p53 by Acetylation

Full-length p53 exists in a latent DNA-binding form as result of the tail–core interactions, as suggested by several studies (Hupp et al., 1992; Hupp et al., 1995; Jayaraman and Prives, 1995). Acetylation of lysine residues in the C-terminal region results in neutraliza- tion of positive charges and leads to the dis- ruption of interactions between the C-termi- nal domain and the core domain, thus allowing the DNA-binding domain to adopt an active conformation. Competitive disrup- tion of this tail–core interaction by synthetic C-terminal peptides can also activate the DNA binding of the full-length p53 (Hupp et al., 1995; Jayaraman and Prives, 1995). How- ever, when the positive charges in the C-ter- minal peptides are neutralized by acetylation, the acetylated peptide loses the ability to ac- tivate p53 DNA binding in trans.

raising the question of whether this type of phosporyla- and Kouzarides, 1996; Ogryzko et al., 1996), P/CAF (Yang et al., 1996), and TAFII250 (Mizzen et al., 1996)tion is functional in vivo. It is possible that the effects

of other activation pathways (e.g., acetylation) may were shown to possess intrinsic HAT activities. Similarly, mask this effect in vivo. various yeast and mammalian corepressors were impli-

Compared with antibody PAb421 binding, addition of cated as, or shown to interact with, proteins that have C-terminal peptides, and phosphorylation, activation of intrinsic histone deacetylase activities (Alland et al., p53 by acetylation may bemore physiologically relevant. 1997; Hassig et al., 1997; Heinzel et al., 1997; Kadosh This is supported by our previous observations that and Struhl, 1997; Laherty et al., 1997; Nagy et al., 1997; p300/CBP strongly interacts with p53, both in vivo and Zhang et al., 1997). Histone acetylation can result in in vitro, and potentiates p53-mediated transcriptional nucleosome destabilization and chromatin remodeling activation in cells (Gu et al., 1997; Lill et al., 1997). Fur- (Turner, 1993; Brownell and Allis, 1996; Wolffe and thermore, the present study shows that residues in the Pruss, 1996), which is likely important for accessibility acetylation site are evolutionarily conserved and, impor- and/or stable binding of transcription factors to cognate tantly, that p53 is also specifically acetylated in vivo. DNA-binding sites within the chromatin templates, and Future studies must involve the development of meth- in some cases, histone acetylation has been correlated ods to confirm the in vivo-acetylated lysine positions with increased levels of transcription (Van Lint et al., and to analyze the regulation of p53 acetylation by CBP/ 1996; Chen et al., 1997). Hence, it has been proposed p300 in the cell. Interestingly, many deacetylase inhibi- that transcriptional activation by sequence-specific ac- tors, including sodium butyrate, trichostatin A, and tra- tivators may involve recruitment and function of histone poxin, induce cell growth repression, transformation re- acetyltransferases, while transcriptional repression by version, and apoptosis (Kijima et al., 1993; Yoshida et sequence-specific repressors may be mediated by the al., 1995; Buckley et al., 1996). These phenotypes are recruitment of histone deacetylases (Pazin and Kado- very similar to the effects of p53 activation in the cell, naga, 1997). Consistent with this view, mutations of the making it interesting to find out whether these histone HAT domain have been shown to inactivate thecoactiva- deacetylase inhibitors can also inhibit p53 deacetylation tor function of GCN5 in yeast (Candau et al., 1997). and whether some histone deacetylases can directly However, there is no direct evidence that histone acety- deacetylate p53. lation and deacetylation events are either causally re-

In addition, the proportion of the whole p53 protein lated to, or the exclusive function of, the cognate coacti- population that is acetylated in normal cells and how vators and corepressors. the acetylation of p53 is regulated, particularly in re- In support of the possibility of alternate functions for sponse to specific stimuli (e.g., UV), remain unknown. coactivators with acetyltransferase (or deacetylase) Since an antitumor agent, epoposide, can activate p53 activities is our demonstration that p300 has a potent and induce apoptosis in human testicular tumors by transcription factor acetyltransferase (FAT) activity in switching the latent form into an active form of p53 addition to the HAT activity, and that p300-mediated (Chresta and Hickman, 1996; Lutzker and Levine, 1996), acetylation of p53 can dramatically increase the latent the present studies suggest that it may be useful to DNA-binding activity. The direct effect of the p300 FAT develop an activation strategy based on the stimulation function is clear, since the binding studies were done of p53 acetylation. in vitro with recombinant proteins, and the observations

that p53 is acetylated in vivo and that sequence-specific DNA binding of p53 is limiting for p53-mediated tran-p300/CBP: HAT versus FAT Functions

As first reported for the yeast coactivator GCN5 (Kuo et scriptional activation suggest that a similar p300 FAT function is operative in vivo. The possibility of cofactoral., 1996), mammalian coactivators p300/CBP (Bannister

Activation of p53 DNA Binding by Aceytlation 603

FAT activities is further supported by the observation be that acetylation by p300 leads to joint activator/ that other general transcriptional factors and cofactors coactivator recruitment to the promoter, both through can be acetylated by p300 in vitro (W. G. and R. G. R., enhanced p53 DNA-binding capability and through per- unpublished data). sistent p53–p300 interactions, and that subsequent

Our present study adds an important new perspective functions of both p300 (e.g., histone acetylation) and to our current understanding of the molecular mecha- p53 (e.g., TAF interactions) are then possible. nism of acetyltransferase/coactivator functions in tran- In conclusion, we have provided an example of a non- scriptional activation by implicating a FAT function that histone regulatory protein, p53, which can be acetylated may operate in place of or, more likely, in addition to both in vitro and invivo, and we have shown that acetyla- the HAT function. Thus, acetyltransferase-containing tion of p53 by its coactivator p300 dramatically activates coactivators may be recruited by (interact with) se- its biochemical function. This study clearly suggests a quence-specific activators for acetylation of either the potential pathway for activation of p53 in vivo and pre- activators themselves (e.g., p53) or the general tran- dicts that other cellular regulatory proteins can be modi- scription machinery (e.g., TFIIE, TFIIF, and PC4; W. G. fied and functionally regulated in a similar manner. It and R. G. R., unpublished data), or both, with conse- further supports the views that protein acetylation and quent transcriptional activation. At the same time, the deacetylation are fundamental regulatory mechanisms recruitment of acetyltransferases to the promoter region in transcriptional regulation. could also induce local nucleosome modifications by histone acetylation, facilitating stable binding of the acti- Experimental Procedures vators or general transcriptional factors to the nucleo-

Plasmids and Recombinant Proteinssome template prior to transcription initiation. These For production of GST fusion proteins, DNA sequences correspond-mechanisms are especially applicable to HAT/FAT co- ing to the indicated regions of human p53 were amplified by PCRactivators that are targeted to promoters either by direct and subcloned into pGEX-2T (Pharmacia). GST fusion proteins were

interactions with site-specific DNA-binding proteins expressed in E. coli, extracted with buffer BC500 (20 mM Tris–HCl (p300/CBP, TAFII250) or through interactions with asso- [pH 8.0], 0.5 mM EDTA, 500 mM KCl, 20% glycerol, 1 mM DTT, and ciated factors (P/CAF, GCN5) that in turn recognize site- 0.5 mM PMSF) containing 1% NP-40, and purified on glutathione-

Sepharose (Pharmacia). The Flag-p53 and Flag-p300(1195–1810)specific DNA-binding proteins. proteins wereexpressed in BL21 (Lys)cells at room temperature andInterestingly, several studies have indicated that tran- purified on an M2 agarose affinity column as indicated previouslysiently transfected templates are not efficiently pack- (Chiang and Roeder, 1993; Gu et al., 1997). All proteins were dialyzed

aged into chromatin (Archer et al., 1992; Van Lint et al., in buffer BC100 (20 mM Tris–HCl [pH 8.0], 0.5 mM EDTA, 100 mM 1996; Pazin and Kadonaga, 1997), and that upregulation KCl, 20% glycerol, 0.5 mM DTT, and 0.5 mM PMSF) before storage of histone acetylation with deacetylase inhibitors can at 2808C. result in a marked increase in transcription of integrated

Protein Acetyltransferase Assaystemplates with little or no effect on transiently trans- Protein acetyltransferase assays were performed essentially as de-fected templates (Van Lint et al., 1996; Chen et al., 1997). scribed (Ogryzko et al., 1996) with some modifications. In the stan-In this regard, synergism between p300/CBP and inter- dard assay, 30 ml reactions contained 50 mM HEPES (pH 8.0), 10%

acting activators (including p53) is mainly evident on glycerol, 1 mM DTT, 1 mM PMSF, 10 mM sodium butyrate, 1 ml of transiently transfected templates (Gu et al., 1997; Lill et [14C]-acetyl-CoA (55 mci/mmol, Amersham), 1 mg of highly purified al., 1997; reviewed in Shikama et al., 1997). Along with substrate proteins or 2.5 mg of GST fusion proteins, and 100 ng of

p300(1195–1810) and were incubated at 308C for 1 hr. For filterour more direct demonstration of functional FAT activi- binding assays, the reaction mixture was spotted onto Whatmanties, these results further support our hypothesis that P-81 phosphocellulose filter paper. The filter paper was air-driedcoactivator-mediated acetylation of nonhistone regula- for 2–5 min and washed with 0.2 M sodium carbonate buffer (pH

tory proteins may be important in transcriptional acti- 9.2) at room temperature with five changes of the buffer for a total vation. of 30 min. For electrophoretic assays, the reaction mixture was

Furthermore, although a major function of many se- subjected to SDS–PAGE gels and autoradiography. Gels containing [14C]-acetate-labeled proteins were fixed with 10% glacial aceticquence-specific activators may beto recruitacetyltrans- acid and 40% methanol for 1 hr and were enhanced by impregnatingferase-containing coactivators for utilization of the ace- with a commercial fluorography enhancing solution (Amplify, Amer-tyltransferase activities per se, other functions, such sham) for 30 min. Gels were then dried and autoradiography wasas direct interactions that result in recruitment of the performed at 2708C for 1–3 days. For the preparation of acetyl-p53

general transcriptional machinery, are also possible for and acetyl-p53(364–389) for EMSA, [14C]-acetyl-CoA was replaced both activators (reviewed in Ptashne and Gann, 1997) with unlabeled acetyl-CoA (Pharmacia) in the standard assay. and acetyltransferase-containing coactivators (Kwok et

Mapping p53 Acetylation Sitesal., 1994; Nakajima et al., 1997; reviewed in Shikama Synthetic peptides corresponding to the C terminus of p53 (residueset al., 1997). In the case of p53, secondary activator 364–389) were synthesized by the protein/DNA technology centerfunctions are indicated by the synergistic effect of p53 of the Rockefeller University and purified to 95% purity by HPLC.

on activation by CBP, even when the latter is tethered Peptides were incubated with [14C]-acetyl-CoA and p300 at 308C for to the promoter by an alternative mechanism (Gu et al., 2 hr. After incubation, the acetylated peptides were separated from 1997), and may reflect the p53–TAF interactions impli- contaminant p300 after the reaction by passage through a Microcon

10 (Amicon). The peptides were then subjected to standard N-termi-cated in target promoter activation by p53 in vitro (Lu nal peptide microsequenceing analyses with 50% of each cycleand Levine, 1995; Thut et al., 1995). Taken together, used for amino acid identification and 25% for radioactivity determi-these studies and the work reported here support a nation by liquid scintillation counting.model in which both DNA-binding activators and inter-

acting acetyltransferase/coactivators may have multiple Cell Transfection, Labeling, and Immunopurification functions in a multistep transcriptional activation path- SaoS-2 cells (2 3 106) were transfected by calcium phosphate pre-

cipitation on a 10-cm plate essentially as previously described (Guway. A speculative comprehensive model for p53 would

Cell 604

et al., 1994) with minor modifications. p53 or p53D370 plasmids (5 Archer, T.K., Lefebvre, P., Wolford, R.G., and Hager, G.L. (1992). Transcription factor loading on the MMTV promoter: a bimodalmg) (Crook et al., 1994) with carrier plasmid pcDNA3 (15 mg) were mechanism for promoter activation. Science 255, 1573–1576.used for transfections on each plate. Thirty hours after transfection,

the cells were transferred to the same DMEM medium containing Arany, Z., Newsome, D., Oldread, E., Livingston, D.M., and Eckner, 1 mci/ml [3H]-sodium acetate (2–10 Ci/mmol, about 100 mCi total) R. (1995). A family of transcriptional adaptor proteins targeted by (Amersham) for 1 hr before lysis. All steps of the immunoprecipitation the E1a oncoprotein. Nature, 374, 81–84. were carried out on ice or in the cold room. Cells were washed twice Arias, J., Alberts, A., Brindle, P., Claret, F., Smeal, T., Karin, M., with cold phosphate-buffered saline and lysed in lysis buffer with Feramisco, J., and Montiminy, M. (1994). Activation of cAMP and freshly added protease inhibitors (50 mM HEPES-KOH [pH 8.0], 150 mitogen responsive gene relies on a common nuclear factor. Nature mM NaCl, 1% Triton X-100, 0.2% Sarkosyl, 10% glycerol, 0.5 mM 370, 226–229. DTT, 1 mM NaF, 1 mM Na3VO4, 10 mM sodium butyrate, 0.5 mg/ml

Bannister, A.J., and Kouzarides, T. (1995). CBP-induced stimulationleupeptin, 1 mg/ml aprotinin, 1 mg/ml pepstatin, and 0.5 mM PMSF). of c-Fos activity is abrogated by E1a. EMBO J. 14, 4758–4762.The lysate was centrifuged twice at 17,000 g for 30 min at 48C. Bannister, A.J., and Kouzarides, T. (1996). The CBP coactivator isSupernatants collected from 10 plates were immunoprecipitated a histone acetyltransferase. Nature 384, 641–643.with the indicated antibodies, which had been cross-linked to aga-

rose beads, for 4 hr at 48C. The beads were washed five times with Barak, Y.T., Juven, T., Haffner, R., and Oren, M. (1993). mdm2 ex- 1 ml of lysis buffer. Immunopurified proteins were solubilized with pression is induced by wild type p53 activity. EMBO J. 12, 461–468. SDS–PAGE sample buffer and resolved on 10% SDS–PAGE gels. Bhattacharya, S., Eckner, R., Grossman, S., Oldread, E., Arany, Z., Gels containing [3H]-acetate-labeled p53 were fixed with 10% glacial D’Andre, A., and Livingston, D.M. (1996). Cooperation of Stat2 and acetic acid and 40% methanol for 1 hr and enhanced by impregnat- p300/CBP in signalling induced by interferon-a. Nature 383, ing with a commercial fluorography enhancing solution (Amplify, 344–347. Amersham) for 30 min. Gels were then dried and subjected to autora-

Blackwood, E.M., and Eisenman, R.N. (1991). Max: a helix-loop- diography at 2708C for 10–14 days. helix zipper protein that forms a sequence-specific DNA binding

complex with Myc. Science 251, 1211–1217. Immunoblotting

Brownell, J.E., and Allis, C.D. (1996). Special HATs for special occa-Proteins were resolved on 10% SDS–PAGE gels, followed by semi- sions: linking histone acetylation to chromatin assembly and genedry transfer to nitrocellulose members for 1 hr at room temperature. activation. Curr. Opin. Genet. Dev. 6, 176–184.After incubation with the first antibody for 2 hr at room temperature, Buckbinder, L., Talbott, R., Velasco-Miguel, S., Takenaka, I., Faha,blots were subsequently incubated with secondary antibody and B., Seizinger, B.R., and Kley, N. (1995). Induction of the growthvisualized by ECL as suggested by the manufacturer (Amersham). inhibitor IGF-binding protein 3 by p53. Nature 377, 646–649.

Electrophoretic Mobility Shift Assay Buckley, A., Leff, M.A., Buckley, D.J., Magnuson, N.S., de Jong, G., EMSA was carried out essentially as described (Jayaraman and and Gout, P.W. (1996). Alteration in pim-1 and c-myc expression Prives, 1995) with some modifications. Sequences of the oligonu- associated with sodium butyrate-induced growth factor depen-

dency in autonomous rat Nb2 lymphoma cells. Cell Growth Differ.cleotides containing either wild-type or mutant p53 DNA-binding 7, 1713–1721.sequence are as follows: GADD45 wild type, TACAGAACATGTC

TAAGCATGCTGGGG; GADD45 mutant, TACAGAATCGCTCTAAG Candau, R., Zhou, J.X., Allis, C.D., and Berger, S.L. (1997). Histone CATGCTGGGG. The DNA binding reactions (20 ml) contained 20 acetyltransferase activity and interaction with ADA2 are critical for mM Tris–HCl (pH 7.5), 50 mM KCl, 5 mM MgCl2, 1 mM DTT, 0.5 mM GCN5 function in vivo. EMBO J. 16, 555–565. EDTA, 10% glycerol, 10 mM sodium butyrate, 0.5 mg/ml BSA, 100 Chakravarti, D., LaMorte, V.J., Nelson, M.C., Nakajima, T., Schulman, ng poly(dI-dC), and proteins as indicated. Reaction mixtures were I.G., Juguilon, H., Motminy, M., and Evans, R.M. (1996). Roles of preincubated at room temperature for 20 min before a 32P-labeled CBP/p300 in nuclear receptor signaling. Nature 383, 99–103. probe DNA (0.2 ng) was added and further incubated at room tem-

Chen, W.Y., Bailey, E.C., McCune, S.L., Dong, J-Y., and Townes,perature for 20 min. Each reaction mixture was then loaded onto a T.M. (1997). Reaction of silenced, virally transduced genes by inhibi-native 4% polyacrylamide gel (acrylamide:Bis, 50:1) containing 0.53 tors of histone deacetylase. Proc. Natl. Acad. Sci. USA 94, 5798–TBE and electrophoresed in 0.25 3 TBE at 180–220 V for 3 hr at 5803.48C. All procedures were basically the same for the analyses of Chiang, C.M., and Roeder, R.G. (1993). Expression and purificationactivation of p53 DNA binding by synthetic peptides, except that of general transcriptional factors by Flag epitope-tagging and pep-the preincubation was performed at 308C for 30 min before the tide elution. Peptide Res. 6, 62–64.probe was added. In the case of supershift assays, the indicated Chresta, C.M., and Hickman, J.A. (1996). Oddball p53 in testicularmonoclonal antibodies (200 ng) were added to reaction mixtures tumors. Nat. Med. 2, 745–746.during preincubation. Chrivia, J.C., Kwok, R.P., Lamb, N., Hagiwara, M., Montiminy, M.R.,

Acknowledgments and Goodman, R.H. (1993). Phosphorylated CREB binds specifically to the nuclear protein CBP. Nature 365, 855–859.

We thank K. H. Vousden and Y. Nakatani for plasmids; X. L. Shi for Crook, T., Marston, N.J., Sara, E.A., and Vousden, K.H. (1994). Tran- highly purified Flag-CREB protein; T. K. Kundu for highly purified scriptional activation by p53 correlates with suppression of growth human core histones; and S. Stevens, D. Tamasauskas, and other but not transformation. Cell 79, 817–827. members of the laboratory for critical comments on the manuscript.

Dai, P., Akimaru, H., Tanaka, Y., Hou, D.X., Yasukawa, T., Kanei-Peptide synthesis and microsequencing wereperformed by the Pro- Ishii, C., Takahashi, T., and Ishii, S. (1996). CBP as a transcriptional

tein/DNA TechnologyCenter of the Rockefeller University; we partic- coactivaor of c-Myb. Genes Dev. 10, 528–540.

ularly thank J. Fernandez for helpful discussions and excellent tech- El-Deiry, W.S., Kern, S.E., Pietenpol, J.A., Kinzler, K.W., and Vo-nical assistance. This work was supported by a postdoctoral gelstein, B. (1992). Definition of a consensus binding site for p53.fellowship from the Life Science Foundation for Advanced Cancer Nat. Genet. 1, 45–49.Studies to W. G., and by grants from the National Institutes of Health El-Deiry, W.S., Tokino, T., Velculescu, V.E., Levy, D.B., Parson, R.,to R. G. R. Trent, J.M., Lin, D., Mercer, W.E., Kinzler, K.W., and Vogelstein B. (1993). WAF1, a potential mediator of p53 tumor suppression. CellReceived July 16, 1997; revised July 29, 1997. 75, 817–825.

References Engel, K., Schultz, H., Martin, F., Kotyarov, A., Plath, K., Hahn, M., Heinemann, U., and Gaestel, M. (1995). Constitutive activation of

Alland, L., Muhle, R., Hou, H., Potes, J., Chin, L., Schreiber-Agus, mitogen-activated protein kinase-activated protein kinase 2 by mu- N., and DePino, R.A. (1997). Role for N-CoR and histone deacetylase tation of phophorylation sites and a-helix motif. J. Biol. Chem. 270,

27213–27221.in Sin3-mediated transcriptional repression. Nature 387, 49–55.

Activation of p53 DNA Binding by Aceytlation 605

Fiscella, M., Zambrano, N., Ullrich, S.J., Unger, T., Lin, D., Cho, B., Kuo, M.H., Brownell, J.E., Sobel, R.E., Ranalli, T.A., Cook, R.G., Edmondson, D.G., Roth, S.Y., and Allis, C.D. (1996). Transcription-Mercer, W.E., Anderson, C.W., and Appella, E. (1994). The carboxy- linked acetylation by GCN5p of histone H3 andH4 at specific lysines.terminal serine 392 phosphorylation site of human p53 is not re- Nature 383, 269–272.quired for wild-type activities. Oncogene 9, 3249–3257.

Kwok, R., Lundblad, J., Chriva, J., Richards, J., Bachinger, H., Bren-Funk, W.D., Pak, D.T., Karas, R.H., Wright, W.E., and Shay, J.W. nan, R., Roberts, S., Green, M., and Goodman, R. (1994). Nuclear(1992). A transcriptionally active DNA-binding site for human p53 protein CBP is a coactivator for the transcription factor CREB. Na-protein complexes. Mol. Cell. Biol. 12, 2866–2871. ture 370, 223–226.Gerristen, M., William, A., Neish, A., Moore, S., Shi, Y., and Collins, Laherty, C.D., Yang, W.M., Sun, J.M., Davie, J.R., Seto, E., andT. (1997). CREB-binding protein CBP/p300 are transcriptional co- Eisenman, R.N. (1997). Histone deacetylases associated with theactivators of p65. Proc. Natl. Acad. Sci. USA 94, 2927–2932. mSin3 corepressor mediate Mad transcriptional repression. Cell 89,Gottlieb, T.M., and Oren, M. (1996). p53 in growth control and neo- 349–356.plasia. Biochim. Biophys. Acta 1287, 77–102. Lee, S., Elenbaas, B., Levine, A., and Griffith, J. (1995). p53 and itsGregor, P.D., Sawadogo, M., and Roeder, R.G. (1990). The adenovi- 14 kDa C-terminal domain recognize primary DNA damage in therus major late transcription factor USF is a member of the helix- form of insertion/deletion mismatches. Cell 81, 1013–1020.loop-helix group of regulatory proteins and binds to DNA as a dimer. Levine, A.J. (1997). p53, the cellular gatekeeper for growth and divi-Genes Dev. 4, 1730–1740. sion. Cell 88, 323–331.Gu, W., Shi, X.L., and Roeder, R.G. (1997). Synergistic activation of Lill, N.L., Grossman, S.R., Ginsberg, D., DeCaprio, J., and Livingston,transcription by CBP and p53. Nature 387, 819–823. D.M. (1997). Binding and modulation of p53 by p300/CBP coactiva-Gu, W., Bhatia, K., Magrath, I.T., and Dalla-Favera, R. (1994). Binding tors. Nature 387, 823–827.and suppression of the Myc transcriptional activation domain by Lu, H., and Levine, A.J. (1995). Human TAF31 protein is transcrip-p107. Science 264, 251–254. tional activation coactivator of p53 protein. Proc. Natl. Acad. Sci.

Hall, S.R., Campbell, L.E., and Meek, D.W. (1996). Phosphorylation USA 92, 5154–5158.of p53 at the casein kinase site selectively regulates p53-dependent Lundblad, J.R., Kwok, R.P., Laurance, M.E., Harter, M.L., and Good-transcriptional repression but not transactivation. Nucleic Acids man, R.H. (1995). Adenoviral E1a-associated protein p300 as a func-Res. 24, 1119–1126. tional homologue of the transcriptional coactivator CBP. Nature 374,

Hansen, R., and Oren, M. (1997). p53; from inductive signal to cellular 85–88. effect. Curr. Opin. Genet. Dev. 7, 46–51.

Lutzker, S.G., and Levine, A.J. (1996). A functionally inactive p53 Hassig, C.A., Fleischer, T.C., Billin, A.N., Schreiber, S.L., and Ayer, protein in teracarcinoma cells is activated by either DNA damage D.E. (1997). Histonedeacetylase activity is required for full transcrip- or cellular differentiation. Nat. Med. 2, 804–810. tional repression by mSin3A. Cell 89, 341–347.

Marston, N.J., Crook, T., and Vousden, K.H. (1994). Interaction of Heinzel, T., Lavinsky, R.M., Mullen, T., Soderstom, M., Laherty, C., p53 with MDM2 is independent of E6 and does not mediate wild Torchia, J., Yang, W., Brard, G., Ngo, S., Davie, J., et al. (1997). A type transformation suppressor function. Oncogene 9, 2707–2716. complex containing N-CoR, mSin3 and histone deacetylase medi-

Meek, D. (1994). Post-translational modification of p53. Semin. Can-ates transcriptional repression. Nature 387, 43–48. cer Biol. 5, 203–210.

Hendzel, M.J., Delcuve, G.P., and Davie, J.R. (1991). Histone deace- Miyashita, T., and Reed, J.C. (1995). Tumor suppressor p53 is atylase is a component of the internal nuclear matrix. J. Biol. Chem. direct transcriptional activator of the human bax gene. Cell 80,15, 21936–21942. 293–299.

Hollstein, M., Sidransky, D., Vogelstein, B., and Harris, C.C. (1991). Mizzen, C.A., Yang, X.J., Kokubo, T., Brownell, J.E., Bannister, A.J.,p53 mutations in human cancers. Science 253, 49–53. Owen-Hughes, T., Workman, J., Wang, L., Berger, S.L., Kouzarides,

Hupp, T.R., Sparks, A., and Lane, D.P. (1995). Small peptides acti- T., et al. (1996). The TAFII250 subunit of TFIID has histone acetyl- vate the latent sequence-specific DNA binding function of p53. Cell transferase activity. Cell 87, 1261–1270. 83, 237–245.

Mohammadi, M., Schlessinger, J., and Hubbard, S.R. (1996). Struc- Hupp, T.R., and Lane, D.P. (1994). Allosteric activation of latent p53 ture of the FGF receptor tyrosine kinase domain reveals a novel tetramers. Curr. Biol. 4, 865–875. autoinhibitory mechanism. Cell 86, 577–587. Hupp, T.R., Meek, D.W., Midgley, C.A., and Lane, D.P. (1992). Regu- Nagy, L., Kao, H.Y., Chakravarti, D., Lin,R.J., Hassig, C.A., Ayer, D.E., lation of specific DNA binding function of p53. Cell 71, 875–886. Schreiber, S.L., and Evans, R.M. (1997). Nuclear receptor repression Jayaraman, L., and Prives, C. (1995). Activation of p53 sequence- mediated by a complex containing SMRT, mSin3A, and histone specific DNA binding by short single strands of DNA requires the deacetylase. Cell 89, 373–380. p53 C terminus. Cell 81, 1021–1029. Nakajima, T., Uchida, C., Anderson, S., Pavin, J., and Montiminy,

M. (1997). Analysis of a cAMP-responsive activator reveals a two-Kadosh, D., and Struhl, K. (1997). Repression by Ume6 involves component mechanism for transcriptional induction via signal-recruitment of a complex containing Sin3 corepressor and Rpd3 dependent factors. Genes Dev. 11, 738–747.histone deacetylase to target promoters. Cell 89, 365–371.

Oelgeschlager, M., Janknecht, R., Krieg, J., Schreek, S., andKamei, Y., Xu, L., Heinzel, T., Torchia, J., Purokawa, R., Gloss, B., Luscher, B. (1996). Interaction of the co-activator CBP with MybLin, S., Heyman, R., Rose, D., Glass, C.K., and Rosenfeld, M.G. proteins: effects on Myb-specific transactivation and on the cooper-(1996). A CBPintegrator complex mediates transcriptional activation ativity with NF-M. EMBO J. 15, 2771–2780.and AP-1 inhibition by nuclear receptors. Cell 85, 403–414. Ogryzko, V.V., Schiltz, R.L., Russanova, V., Howard, B.H., and Naka-Kastan, M.B., Onyekwere, O., Sidransky, D., Vogelstein, B., and tani, Y. (1996). The transcriptional coactivator p300 and CBP areCraig, R.W. (1991). Participation of p53 protein in the cellular re- histone acetyltransferases Cell 87, 953–959.sponse to DNA damage. Cancer Res. 51, 6304–6311. Okamoto, K., and Beach, D. (1994). Cyclin G is a transcriptionalKastan, M.B., Zhan, Q., El-Deiry, W.S., Carrier, F., Jacks, T., Walsh, target of the p53 tumor suppressor protein. EMBO J. 13, 4816–4822.W.V., Plunkett, B.S., Vogelstein, B., and Fornace, A.J., Jr. (1992). A

mammalian cell cycle checkpoint pathway utilizing p53 and GADD45 Pazin, M.J., and Kadonaga, J.T. (1997). What’s up and down with is defective in ataxia-telangiectasis. Cell 71, 587–597. histone deacetylation and transcription. Cell 89, 325–328. Kijima, M., Yoshida, M., Sugita, K., Horinouchi, S., and Beppu, T. Pietenpol, J.A., Tokino, T., Thiagalingam, S., El-Deiry, W.S., Kinzler, (1993). Trapoxin, an antitumor cyclic tetrapeptide, is an irreversible K.W., and Vogelstein, B. (1994). Sequence-specific transcriptional inhibitor of mammalian histone deacetylase. J. Biol. Chem. 268, activation is essential for growth suppression by p53. Proc. Natl. 22429–22435. Acad. Sci. USA 91, 1998–2002.

Ko, L.J., and Prives, C. (1996). p53: puzzle and paradigm. Genes Ptashne, M., and Gann, A. (1997). Transcriptional activation by re- cruitment. Nature 386, 569–577.Dev. 10, 1054–1072.

Cell 606

Selivanova, G., Iotsova, V., Okan, I., Fritsche, M., Storm, M., Groner, B., Crafstrom, R.C., and Wiman, K.G. (1997). Restoration of the growth suppression function of mutant p53 by a synthetic peptide derived from the p53 C-terminal domain. Nat. Med. 3, 632–638.

Shaw, P., Freeman, J., Bovey, R., and Iggo, R. (1996). Regulation of specific DNA binding by p53: evidence for a role for O-glycosyla- tion and charged residues at the carboxy-terminus. Oncogene 12, 921–930.

Soussi, T., de Fromentel, C.C., andMay, P. (1990). Structural aspects of the p53 protein in relation to gene evolution. Oncogene 5, 945–952.

Shikama, N., Lyon, J., and La Thangue, N.B. (1997). The p300/CBP family: integrating signals with transcription factors and chromatin. Trends Cell Biol. 7, 230–236.

Thut, C., Chen, J.L., Klemm, R., and Tjian, R. (1995). p53 transcrip- tional activation mediated by coactivators TAFII40 and TAFII60. Sci- ence 267, 100–104.

Turner, B.M. (1993). Decoding the nucleosome. Cell 75, 5–8.

Van Lint, C., Emiliani, S., Ott, M., and Verdin, E. (1996). Transcrip- tional activation and chromatin remodeling of the HIV-1 promoter in response to histone acetylation. EMBO J. 15, 1112–1120.

Vettese-Dadey, M., Grant, P.A., Hebbes, T.R., Crane-Robinson, C., Allis, C.D., and Workman, J.L. (1996). Acetylation of histone H4 plays a primary role in enhancing transcription factor binding to nucleosomal DNA in vitro. EMBO J. 15, 2508–2518.

Wolffe, A.P., and Pruss, D. (1996). Targeting chromatin disruption: transcription regulators that acetylate histones. Cell 86, 817–819.

Wu, X., Bayle J.H., Olsen D., and Levine A.J. (1993). The p53-mdm-2 autoregulatory feedback loop. Genes Dev. 7, 1126–1132.

Yang, X.J., Ogryzko, V., Nishikawa, J., Howard, B., and Nakatani, Y. (1996). A p300/CBP-associated factor that competes with the adenoviral oncoprotein E1a. Nature 382, 319–324.

Yoshida, M., Horinouchi, S., and Beppu, T. (1995). Trichostatin A and trapoxin: novel chemical probes for the role of histone acetylation in chromatin structure and function. BioEssays 5, 423–430.

Yuan, W., Condorelli, G., Caruso, M., Felsani, A., and Giordano, A. (1996). Human p300 protein is a coactivator for the transcription factor MyoD. J. Biol. Chem. 271, 9009–9013.

Zhang, J.J., Vinkemeier, U., Gu, W., Chakravarti, D., Horvath, C.M., Darnell, J.E., Jr. (1996). Two contact regions between Stat1 and CBP/p300 in interferon gamma signaling. Proc. Natl. Acad. Sci. USA 93, 15092–15096.

Zhang, Y., Iratni, R., Erdjument-Bromage, H., Tempst, P., and Rein- berg, D. (1997). Histone deacetylases and SAP18, a novel polypep- tide, are components of a human Sin3 complex. Cell 89, 357–364.