THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 281, NO. 31, pp. 21934 –21941, August 4, 2006 © 2006 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

Effects of Common Cancer Mutations on Stability and DNA Binding of Full-length p53 Compared with Isolated Core Domains* Received for publication, May 2, 2006, and in revised form, June 5, 2006 Published, JBC Papers in Press, June 5, 2006, DOI 10.1074/jbc.M604209200

Hwee Ching Ang1, Andreas C. Joerger, Sebastian Mayer, and Alan R. Fersht2 From the Centre for Protein Engineering, Medical Research Council Centre, Hills Road, Cambridge CB2 2QH, United Kingdom

The tumor suppressor protein p53 is a transcription factor that plays a critical role in the network of signals that control the fate of a cell (1). In about half of human tumors, p53 is inactivated as a result of point missense mutations in the sequence-specific DNA binding core domain of the protein (2, 3). Six “hot spots” are most frequently associated with cancer: Arg175, Gly245, Arg248, Arg249, Arg273, and Arg282 (3) (Fig. 1). From quantitative folding and DNA binding studies of numerous cancer-associated core domain mutants, three phenotypes have been broadly categorized: (i) DNA contact mutations with only minor effects on folding and stability, such as R273H; (ii) mutations that disrupt the local structure, mainly in proximity to the DNA binding surface, which destabilize the protein by ⱕ2 kcal/ mol relative to wild type, such as G245S; and (iii) highly destabilizing mutations, such as R175H, which destabilize the protein by ⬎3 kcal/mol (4 – 6). The structural effects of the

* This

work was supported by the Medical Research Council and Cancer Research UK. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 Supported by the Singapore Agency for Science, Technology and Research (A*STAR). 2 To whom correspondence should be addressed: Centre for Protein Engineering, Medical Research Council, Hills Road, Cambridge CB2 2QH, UK. Tel.: 44-1223-402-136; Fax: 44-1223-402-140; E-mail: [email protected].

21934 JOURNAL OF BIOLOGICAL CHEMISTRY

mutations R249S and R273H have recently been elucidated by x-ray crystallography (7). p53 has several domains: the N-terminal transactivation domain (residues 1– 63) (8), the proline-rich regulatory domain (residues 64 –92) (9, 10), the DNA binding core domain (residues 94 –312) (11), the tetramerization domain (residues 324 – 355) (12), and the C-terminal domain (residues 360 –393) (13). The core domain and the C-terminal domain bind DNA. The isolated core domain (p53C)3 binds specifically to a doublestranded DNA consensus site containing two copies of the 10-base pair “half-site” motif 5⬘-PuPuPuC(A/T)(T/A)GPyPyPy-3⬘ (Pu ⫽ A/G, Py ⫽ T/C) that can be separated by up to 13 bases (14). p53C monomers bind specific DNA to give a 4:1 complex (15, 16). Binding to a minimal tetrameric p53 construct comprising the core and tetramerization domains is cooperative with a stoichiometry of two dimers per DNA molecule (16). The C-terminal domain, its post-translational modifications and its role as a regulatory domain have been extensively studied and debated (13, 17–24). This domain has been shown to bind nucleic acid in a sequence-nonspecific manner (13, 25) with a strong electrostatic component (26). A new approach in cancer therapy is to use drugs that can rescue the activity of mutant p53 (6, 27–29). To assess further the feasibility of rescue drug therapy, we have quantitatively analyzed the effect of p53 cancer mutations on the stability of full-length protein and the effect of the contact mutation R273H on p53-DNA interactions. We used an engineered thermostable mutant of p53 core domain (T-p53C), containing the mutations M133L/V203A/N239Y/ N268D (30, 31) to produce p53 mutant proteins suitable for accurate biophysical measurements.

EXPERIMENTAL PROCEDURES Protein Expression and Purification—T-p53C, T-p53CT, and mutants of these constructs were purified as previously described (16, 31). The plasmid for human p53 (residues 1–393) subcloned into the polylinker region of vector pET24a(⫹) (Novagen) using the NdeI and EcoRI restriction sites was kindly provided by C. Blair. Additional point mutations 3

The abbreviations used are: p53C, p53 core domain (residues 94 –312); T-p53C, thermostable variant of p53 core domain (residues 94 –312) containing the four point mutations M133L, V203A, N239Y, and N268D; T-p53CT, thermostable p53 truncation mutant (residues 94 –360, comprising the core and tetramerization domains); T-p53FL, thermostable variant of p53 full-length protein; DTT, dithiothreitol; KD, dissociation constant; AUC, analytical ultracentrifugation; DSC, differential scanning calorimetry.

VOLUME 281 • NUMBER 31 • AUGUST 4, 2006

Downloaded from www.jbc.org at Karolinska institutet library on February 20, 2007

Common cancer mutations of p53 tend either to lower the stability or distort the core domain of the protein or weaken its DNA binding affinity. We have previously analyzed in vitro the effects of mutations on the core domain of p53. Here, we extend those measurements to full-length p53, using either the wild-type protein or a biologically active superstable construct that is more amenable to accurate biophysical measurements to assess the possibilities of rescuing different types of mutations by anticancer drugs. The tetrameric full-length proteins had similar apparent melting temperatures to those of the individual domains, and the structural mutations lowered the melting temperature by similar amounts. The thermodynamic stability of tetrameric p53 is thus dictated by its core domain. We determined that the common contact mutation R273H weakened binding to the gadd45 recognition sequence by ⬃700 –1000 times. Many mutants that have lowered melting temperatures should be good drug targets, although the common R273H mutant binds response elements too weakly for simple rescue.

Stability and DNA Binding Affinity of Full-length p53 Mutants

were introduced using the QuikChange site-directed mutagenesis kit (Stratagene). The expression vector was transformed into Escherichia coli BL21 for overexpression. Expression cultures were incubated at 37 °C at 250 rpm until OD600 reached ⬃0.8. Zinc sulfate was added to give a final concentration of 100 ␮M, and protein expression was induced with 500 ␮M isopropyl ␤-D-thiogalactoside. Cells were harvested 16 h later by centrifugation. The cell pellet from 3 liters of culture was suspended in cell-cracking buffer of 50 mM NaPi, pH 7.5, 150 mM NaCl, 5 mM DTT, and 15% glycerol and cracked using an Emulsiflex C5 high pressure homogenizer (Glen Creston). The soluble fraction was loaded onto a Heparin HP column and eluted with a 0 –1 M NaCl gradient over 20 column volumes. The pooled fractions from this column were diluted 10-fold with 25 mM Tris/Bis-Tris propane buffer pH 9, 5 mM DTT, and 15% glycerol, loaded onto a Poros 20HQ anionic exchange column and eluted with a 0 –1 M NaCl gradient over 20 column volumes. The pooled fractions were purified further on a Superdex 200 26/60 preparative gel filtration column (Amersham Biosciences) in 25 mM NaPi, pH 7.2, 300 mM NaCl, 5 mM DTT, and 10% glycerol. The purified p53 proteins were greater than 95% pure as judged by SDS/polyacrylamide gel. Protein samples were flash-frozen and stored in liquid nitrogen for further use. DNA Duplex Assembly and Purification—Double-stranded DNAs were assembled from HPLC-purified oligonucleotides AUGUST 4, 2006 • VOLUME 281 • NUMBER 31

JOURNAL OF BIOLOGICAL CHEMISTRY

21935

Downloaded from www.jbc.org at Karolinska institutet library on February 20, 2007

FIGURE 1. Wire frame model of p53 core domain bound to gadd45 consensus DNA (PDB ID code 1TSR, molecule B). Secondary structural elements are highlighted by semitransparent ribbons and cylinders. A ␤-sandwich provides the scaffold for the DNA binding surface, which is rich in basic amino acids and contains a zinc ion (yellow sphere). The six residues that are most frequently mutated in human cancer are shown in orange. The small blue spheres indicate the location of the mutation sites in the superstable quadruple mutant T-p53C. Trp146 (highlighted in green) was used as a fluorescence probe in this study. The figure was generated using MOLSCRIPT (46) and RASTER3D (47).

and purified according to procedures already described (16, 32). The oligonucleotides were labeled on the 5⬘-end of the forward strand with fluorescein. Only the forward DNA strand was labeled to avoid energy transfer between fluorophores. The gadd45 promoter sequence was verified against the published recognition element (33), and encoded by the oligonucleotides 5⬘-fluorescein-GTACAGAACATGTCTAAGCATGCTgGGGAC and 5⬘-GTCCCcAGCATGCTTAGACATGTTCTGTAC. Bases in capital letters agree with the published consensus sequence (14). Nonspecific DNA (random DNA) that did not contain a p53 recognition element was encoded by the oligonucleotides 5⬘-fluorescein-AATATGGTTTGAAATAAAGAGTAAAGATTTG-3⬘ and 5⬘-CAAATCTTTACTCTTTATTCAAACCATATT-3⬘. Equilibrium Denaturation—Samples for urea denaturation experiments were prepared using a Hamilton MicrolabM dispenser from stock solutions of urea, buffer, and protein to contain 1 ␮M protein in 25 mM sodium phosphate buffer, pH 7.2, 150 mM KCl, 5 mM DTT, and increasing concentrations of urea. Samples were incubated at 10 °C for 14 h prior to measurement. The intrinsic fluorescence spectra of p53, excited at 280 nm, were recorded in the range of 300 – 400 nm on a PerkinElmer Life Sciences LS50B spectrofluorometer equipped with a Waters 2700 sample manager and controlled by laboratory software. The data were analyzed as previously described (4). Differential Scanning Calorimetry (DSC)—DSC experiments were performed using a Microcal VP-Capillary DSC instrument (Microcal, Amherst, MA) with an active cell volume of ⬃125 ␮l. Temperatures from 10 to 85 °C were scanned at a rate of 250 °C/h. Protein samples were buffer-exchanged into a buffer of 25 mM sodium phosphate, pH 7.2, 150 mM NaCl, 5 mM DTT. This buffer was also used for baseline scans. For core domain, 100 ␮M protein was used. For full-length protein, 15 ␮M (in monomeric units) protein was used. A pressure of 2.5 bars (nitrogen) was applied to the cell. The data were analyzed with ORIGIN software (Microcal). The average apparent Tm value is presented in parentheses in Fig. 2. Analytical Centrifugation—All analytical centrifugation (AUC) experiments were performed at 10 °C using a Beckman Optima XL-I centrifuge and a 60Ti rotor. To determine DNA binding affinities, samples of 5 ␮M 5⬘-fluorescein-labeled gadd45 DNA and 100 ␮M T-p53 monomers were made up in 25 mM sodium phosphate, pH 7.2, 150 mM KCl, and 5 mM DTT. To measure the stoichiometry and dissociation constant for protein binding to fluorescein-tagged DNA, absorbance values at 495 nm and interference patterns at 675 nm were recorded, and the data were analyzed using UltraSpin software. The apparent dissociation constant per monomer was calculated as KD ⫽ ([free DNA]/[complex]) ⫻ [free protein] and presented in Table 2. This value is equivalent to P50. Fluorescence Anisotropy—Fluorescence anisotropy measurements were recorded on a PerkinElmer Life Sciences LS55 Luminescence Spectrometer equipped with a Hamilton Microlab titrator and controlled by laboratory software. The excitation (␭ex) and emission (␭em) wavelengths used were 480 nm and 530 nm, respectively, and the slit widths for excitation and emission were 15 nm and 20 nm. The photomultiplier voltage

Stability and DNA Binding Affinity of Full-length p53 Mutants TABLE 1 Changes in free energy on urea-induced unfolding of p53 core domain mutants Data were collected at 10 °C in 25 mM sodium phosphate, pH 7.2, 150 mM KCl, 5 mM DTT. The mean m value of 13 urea denaturation curves at 10 °C is 2.89 (⫾0.11) kcal mol⫺1 M ⫺1, which was used to calculate the difference between 关D兴50% 关D兴50% wild type and mutant proteins, ⌬⌬GD-N , from the equation, ⌬⌬GD-N ⫽ ⬍m⬎ ⌬ 关D兴50%. Mutation

V143A R175H G245S R249S R273H R282W a b

关D兴50% ⌬⌬GD-N

T-p53C

Wild type (p53C)a

kcal/mol

kcal/mol

3.7 ⫾ 0.15 2.5 ⫾ 0.10 0.8 ⫾ 0.04 2.0 ⫾ 0.12b 0.1 ⫾ 0.02b 3.0 ⫾ 0.20

3.5 ⫾ 0.06 3.5 ⫾ 0.06 1.2 ⫾ 0.03 2.0 ⫾ 0.05 0.3 ⫾ 0.04 3.3 ⫾ 0.10

Data for mutations in the wild-type context are taken from Refs. 4 and 45. Data are taken from Ref. 7.

RESULTS Common Cancer Mutations Have a Similar Effect on the Stability of T-p53C, Wild-type p53C, and Full-length p53 We measured the effects of structural hot spot mutations (R175H, G245S, and R282W) and the classic temperaturesensitive mutation V143A on the stability of T-p53C by urea

21936 JOURNAL OF BIOLOGICAL CHEMISTRY

FIGURE 2. Stability measurements of T-p53 and hot spot variants. A, ureainduced unfolding of T-p53C and hot spot variants, represented as fraction of unfolded protein versus concentration of denaturant. Unfolding was monitored by fluorescence, at 10 °C in 25 mM sodium phosphate, pH 7.2, 150 mM KCl, 5 mM DTT. The main probe for denaturation is Trp146, which is located away from the DNA binding surface, at the far end of the ␤-sandwich. B, thermal denaturation curves are shown for core domain (T-p53C) and its hot spot variants (lower set of 7 curves), and for the full-length protein (T-p53FL) and its hot spot variants (upper set of 5 curves). Average apparent melting temperatures (Tm) are given in parentheses. Representative melting curves (raw data) are shown and are offset for clarity. The apparent Tm for wild-type p53C and p53FL are 44.1 °C and 45.1 °C, respectively.

denaturation (Table 1 and Fig. 2A). Overall, the trends followed the wild-type protein (4, 5). G245S destabilized the core domain by 1.2 kcal/mol in wild type and 0.8 kcal/mol in T-p53C. The structural mutations V143A, R175H, and R282W that substantially destabilize the wild type by 3.3–3.5 kcal/mol also caused major stability loss in T-p53C, ranging from 2.5–3.7 kcal/mol. In the full-length protein, the signal from Trp146, which was used to monitor the unfolding of the core domain by urea denaturation, is masked by the relatively strong fluorescence signal from 3 tryptophan residues in the natively unfolded N-terminal region. We therefore studied the effects of mutations on the stability of full-length p53 by differential scanning calorimetry VOLUME 281 • NUMBER 31 • AUGUST 4, 2006

Downloaded from www.jbc.org at Karolinska institutet library on February 20, 2007

used was 950 with an integration time of 5 s for each measurement. The concentrations of tetrameric constructs were 1–50 ␮M (dimers). The initial concentrations of DNA were 4 –10 nM. The experiments were performed at 10 °C in 25 mM imidazole, pH 7.2, 213.4 mM NaCl, 5 mM DTT, with a total ionic strength of 225 mM, as described by Weinberg et al. (16) to minimize artifacts caused by nonspecific binding effects. For T-p53CT-R273H, experiments were performed in the presence of 10% glycerol to prevent aggregation. Control experiments comparing T-p53CT binding gadd45 in 0, 5, and 10% glycerol showed minimal differences in binding affinities. Proteins were buffer exchanged into the reaction buffer using NAP-10 columns (Amersham Biosciences), and protein concentrations were measured spectrophotometrically immediately prior to use using molar extinction coefficients calculated according to the method of Gill and von Hippel (48). p53 was titrated into a cuvette containing fluorescein-labeled DNA, and the solution was stirred for 30 s. After 60 s, the fluorescence and fluorescence polarization values were measured, using an integration time of 5 s. Different data analyses were performed depending on the protein concentrations used. (i) At concentrations of p53 far below Ktd: When the affinity for DNA is very high, the measurements are made at concentrations of p53 that are much below the dissociation constant for tetramers into dimers (Ktd). It can be assumed that unbound protein is dimeric under the reaction conditions and does not tetramerize until bound to DNA. The binding isotherm simplifies to that of the simple sequential two-site model, which was previously used to describe the sequence-specific DNA binding of p53CT (15, 16). (ii) At concentrations of p53 approaching or exceeding Ktd: At the other extreme, for weak binding to DNA, measurements are made at concentrations of p53 that approach or exceed the value of Ktd. The concentration of tetramer at each titration point was calculated and plotted to give the KDNA value directly.

Stability and DNA Binding Affinity of Full-length p53 Mutants TABLE 2 Binding of monomeric p53 to specific gadd45 DNA, as measured by analytical ultracentrifugation at total ionic strength of 210 mM KD values were determined by AUC at 10 °C in 25 mM sodium phosphate, pH 7.2, 150 mM KCl, and 5 mM DTT, at a total ionic strength of 210 mM. Measurements were performed with 5 ␮M 5⬘-fluorescein-labeled double-stranded DNA that contained the 20-base pair, specific recognition element from the gadd45 promoter, and 100 ␮M protein. Data were analyzed as described under “Experimental Procedures.” Data (except for T-p53C-G245S) have been published (7). Protein

KD

p53C T-p53C T-p53C-G245S T-p53C-R249S T-p53C-R273H

14 ⫾ 1 10 ⫾ 3 35 ⫾ 2 56 ⫾ 3 51 ⫾ 4

␮M

Binding of DNA to Full-length p53 and Shorter Constructs We have previously reported the DNA binding affinities of p53 core domains by analytical ultracentrifugation studies (7). The affinity of T-p53C-G245S using the same technique under similar conditions was 35 ⫾ 2 ␮M (Table 2). We analyzed the binding to tetrameric p53 using two constructs to distinguish between specific DNA binding to the core and nonspecific binding to the C terminus: (i) a minimal tetrameric construct consisting of the core and tetramerization domains (T-p53CT), and (ii) the full-length protein (T-p53FL) 4

S. Mayer and A. R. Fersht, unpublished data.

AUGUST 4, 2006 • VOLUME 281 • NUMBER 31

FIGURE 3. Hill plots for the binding of T-p53CT and T-p53FL to gadd45 DNA. Binding was cooperative in each case. Hill coefficients for T-p53CT and T-p53FL were 1.95 and 1.99, respectively.

FIGURE 4. p53 in a dimer-tetramer equilibrium. p53 can bind DNA as a dimer or tetramer. Where p53 exists in a dimer-tetramer equilibrium over the range of measurement of DNA affinity, this two-component model is applied, ignoring the monomers.

(24, 26). The DNA binding experiments were conducted using fluorescence anisotropy at an ionic strength of 225 mM to reduce nonspecific electrostatic interactions at the DNA binding interface, which is positively charged and contains numerous arginine residues that are involved in DNA binding (16). Two types of 5⬘-fluorescein-labeled 30-mer double-stranded DNA were tested: one that contained the 20-base pair specific recognition element from the gadd45 promoter (gadd45); another that did not contain a p53 recognition element (random DNA) (16, 32). We analyzed the binding data for T-p53CT and T-p53FL using the Hill equation, as previously described (16) (Fig. 3). The Hill coefficient for binding to gadd45 promoter DNA was 1.95 for T-p53CT and 1.99 for T-p53FL (Fig. 3), consistent with previous results for p53CT binding p21 and Mdm2 promoter DNA (16), and indicates a cooperative binding event for T-p53CT and T-p53FL binding gadd45 promoter DNA. DNA Binding of the Minimal Tetrameric Construct Comprising Core and Tetramerization Domains—Tetrameric p53 is in a dynamic tetramer-dimer equilibrium (Fig. 4). The equilibrium constant for the minimal tetrameric construct dissociating into dimers is 400 nM (per p53 dimer at 10 °C) (34). p53 mutants that bind DNA weakly are present as tetJOURNAL OF BIOLOGICAL CHEMISTRY

21937

Downloaded from www.jbc.org at Karolinska institutet library on February 20, 2007

(DSC). The observed Tm is not the true melting temperature but an apparent one because p53 does not denature reversibly with increasing temperature (5). Although the data are qualitative, they reflect the relative destabilizing effects of the mutations; the apparent Tm values of T-p53C paralleled the reversible urea-induced thermodynamic data (Fig. 2, A and B). The apparent Tm values are also useful for assessing the stability of p53 in vivo.4 The apparent Tm of T-p53C-R273H was virtually the same as T-p53C at 49.9 °C. The apparent Tm of T-p53C-G245S was reduced to 47.3 °C. The apparent Tm of T-p53C-R249S was further reduced to 45.0 °C. The highly destabilizing mutations, R175H and V143A, greatly reduced the Tm to 42.5 °C and 43.0 °C, respectively. Experiments for full-length p53 were done with ⬃4 ␮M protein. Because the equilibrium constant for fulllength p53 tetramers dissociating into dimers is 300 nM (per p53 dimer at 10 °C) (34), the protein should exist as tetramers at this concentration. Control experiments (not shown here) showed no large effects of concentration on aggregation, so the apparent Tm values were qualitative but relative as the observed values depend on heating rate. The apparent Tm of wild-type full-length p53 was 45.1 °C, close to that of wild-type core domain 44.1 °C. In full-length T-p53FL, the cancer mutations G245S, R249S, and R273H showed the same relative effect on stability as in core domain T-p53C (Fig. 2B). Introduction of the R273H mutation had little effect on the stability, and the apparent Tm shifted marginally to 48.2 °C. The structural mutation G245S reduced the apparent Tm from 49.7 °C to 47.3 °C. The mutation R249S, which was more destabilizing than the G245S mutation, further reduced the apparent Tm to 43.8 °C.

Stability and DNA Binding Affinity of Full-length p53 Mutants ramers at the high protein concentrations used. Measurements of the concentration of p53 (in dimers) for 50% saturation (P50) for different constructs and mutants of p53 were made. T-p53CT bound the gadd45 recognition sequence with P50 of 21 ⫾ 2 nM (Table 3, Fig. 5A) and random DNA about 30 times more weakly with P50 of 680 ⫾ 20 nM (Table 3 and Fig. 5B). In contrast, the minimal tetrameric T-p53CTR273H mutant bound gadd45 about 1000 times more weakly than did T-p53CT (see “Discussion”). Nevertheless, it could still discriminate between the two types of DNA used, and bound gadd45 with P50 of 3000 ⫾ 400 nM, several times more tightly than nonspecific random DNA (P50 of ⬎10,000 nM) (Fig. 5, C and D).

DISCUSSION Destabilizing Effects of Common Cancer Mutations in p53 Binding affinities are expressed in terms of P50 values. P50 is the protein concentraCore Domain and Full-length Protein Follow the Same Trend— tion (in dimers) at 50% saturation. The specific and nonspecific DNA used are Common cancer mutations had the same relative effects on the 5⬘-fluorescein-tagged gadd45 and random DNA, respectively. The oligonucleotide sequences are described under “Experimental Procedures.” stability of wild-type p53 core domain and the stabilized core P50 domain variant T-p53C. For example, R249S destabilizes both Protein Specific DNA Nonspecific DNA core domain variants by ⬃2 kcal/mol whereas the introduction of nM nM the R273H mutation had virtually no effect on stability (7). The use T-p53CT 21 ⫾ 2 680 ⫾ 200 of this stabilized T-p53 construct enabled us to produce full-length T-p53CT-R273H 3000 ⫾ 400 ⬎10,000 T-p53FL 4⫾2 360 ⫾ 60 p53 mutants for accurate biophysical measurements. Interest1500 ⫾ 200a T-p53FL-R273H 1400 ⫾ 200a ingly, all cancer hot spot mutations tested had the same relative T-p53FL-G245S 55 ⫾ 7 250 ⫾ 50 T-p53FL-R249S 1200 ⫾ 200 850 ⫾ 100 effects on the stability of the full-length protein as measured in a Likely to be nonspecific binding to the C terminus. isolated core domain (Fig. 8, A and B). This indicates that the core domain is pivotal in governing the overall stability of p53, and that all mutation-induced stability changes observed at core domain level directly translate into similar relative stability changes in the full-length protein. Effects of R273H Mutation on Specific DNA Binding of p53—The R273H contact mutation reduced the binding of the gadd45 response element by raising the P50 for the wild-type minimal tetrameric core and tetramerization domain construct from 21 to 3000 nM (Table 3). The contact mutation did not completely abrogate DNA binding, consistent with previous mammalian cell-based studies in which reduced but detectable transactivation function is observed for the mutant R273H (35–37) and mixed tetramers of the R273H mutant with 2 or more wild-type monomers are transcriptionally active (38). These observations are also in good agreement with crystallographic studies, FIGURE 5. DNA binding isotherms for minimal tetrameric T-p53CT and T-p53CT-R273H binding to 5ⴕ-fluorescein-labeled double-stranded 30-mers that contained either the recognition element from the which show that while a critical gadd45 promoter or a randomly generated sequence of 30 bases (16, 32). A, T-p53CT binding gadd45 DNA; DNA contact is lost, the overall B, T-p53CT binding random DNA; C, T-p53CT-R273H binding gadd45 DNA; D, T-p53CT-R273H binding random DNA. Experiments were done at 10 °C in 25 mM imidazole, 213.4 mM sodium chloride (total ionic strength of 225 architecture of the DNA binding surface is preserved (7). mM), 5 mM DTT, and 10% glycerol.

21938 JOURNAL OF BIOLOGICAL CHEMISTRY

VOLUME 281 • NUMBER 31 • AUGUST 4, 2006

Downloaded from www.jbc.org at Karolinska institutet library on February 20, 2007

TABLE 3 Binding of tetrameric p53 to specific gadd45 DNA and nonspecific random DNA, as measured by fluorescence anisotropy at a total ionic strength of 225 mM

DNA Binding of the Full-length Proteins—The full-length protein T-p53FL bound gadd45 DNA with P50 of 4 ⫾ 2 nM, about 5 times more tightly than did T-p53CT (Table 3 and Fig. 6A). Like T-p53CT, T-p53FL also showed strong selectivity for gadd45 over random DNA, and bound about 90 times more tightly to gadd45 than random DNA (Table 3 and Fig. 6B). The R273H mutation reduced the binding of gadd45 to the fulllength protein to P50 of 1400 ⫾ 200 nM (Table 3 and Fig. 6C). But, this binding is mainly attributable to nonspecific binding to the C terminus (24, 26), because the dissociation constant for nonspecific DNA was 1500 ⫾ 200 nM (Table 3 and Fig. 6D). The same is true for T-p53FL-R249S, which is a weakly binding mutant (P50 of 1200 ⫾ 200 and 850 ⫾ 100 nM, respectively, for binding to specific and nonspecific DNA) (Fig. 7).

Stability and DNA Binding Affinity of Full-length p53 Mutants

FIGURE 7. DNA binding isotherms for full-length mutants T-p53FL-G245S and T-p53FL-R249S binding to 5ⴕ-fluorescein-labeled double-stranded 30-mers that contained either the recognition element from the gadd45 promoter or a randomly generated sequence of 30 bases (16, 32). A, T-p53FL-G245S binding gadd45 DNA; B, T-p53FL-G245S binding random DNA; C, T-p53FL-R249S binding gadd45 DNA; D, T-p53FLR249S binding random DNA. Experiments were done at 10 °C in 25 mM imidazole, 213.4 mM sodium chloride (total ionic strength of 225 mM), and 5 mM DTT.

AUGUST 4, 2006 • VOLUME 281 • NUMBER 31

JOURNAL OF BIOLOGICAL CHEMISTRY

21939

Downloaded from www.jbc.org at Karolinska institutet library on February 20, 2007

FIGURE 6. DNA binding isotherms for full-length T-p53FL and T-p53FL-R273H binding to 5ⴕ-fluoresceinlabeled double-stranded 30-mers that contained either the recognition element from the gadd45 promoter or a randomly generated sequence of 30 bases (16, 32). A, T-p53FL binding gadd45 DNA; B, T-p53FL binding random DNA; C, T-p53FL-R273H binding gadd45 DNA; D, T-p53FL-R273H binding random DNA. Experiments were done at 10 °C in 25 mM imidazole, 213.4 mM sodium chloride (total ionic strength of 225 mM), and 5 mM DTT.

Comparing Binding to Monomeric, Dimeric, and Tetrameric Constructs—There are various complications in attempting to compare directly the binding of DNA to isolated core domains with oligomeric p53 that is in equilibrium between dimers and tetramers. One problem is that as the dissociation constant of tetrameric p53 into dimers is about 400 nm, the binding of response elements with affinities in the nanomolar region are measured in a p53 concentration range in which it is mainly dimeric. Hence, the Hill constant of close to 2 for the binding of gadd45 DNA to fulllength constructs, showing that the p53 dimers have effectively to associate to tetramers to bind to the response elements that have 4 sites. It may be shown that for tight binding response elements such as gadd45 to wild-type proteins at concentrations ⬍⬍400 nM, the P50 value is the square root of the product of the dissociation constants of the tetramer into dimer (Ktd) and the DNA from the tetramer (KDNA), i.e. P50 ⫽ 公Ktd䡠KDNA (Fig. 4). For T-p53CT, substituting the measured value of the dissociation constant for the tetramer to dimer (400 nM) into the equation, together with the P50 value obtained from DNA binding experiments, gives a value of KDNA for gadd45 and T-p53CT of 1 nM. For weakly binding DNA-protein complexes, such as T-p53CTR273H, we can obtain the true dissociation constant of DNA from the tetramer, KDNA, by calculating the amount of tetrameric p53 in solution using the 400 nM dissociation constant value and plotting the observed anisotropy against the tetrameric p53 concentration. The KDNA for gadd45 and T-p53CTR273H was found to be 1000 ⫾ 200 nM. By comparing the KDNA values, it can be seen that the intrinsic binding to the tetramer is weakened by ⬃1000 times on the mutation of R273H. The absolute values of binding of DNA to tetrameric constructs cannot be quantitatively compared with the binding to core domains

Stability and DNA Binding Affinity of Full-length p53 Mutants

because of differences in molecularity, which lead to vastly different entropies of binding. The effects of mutation can be compared, however. The P50 values for monomeric core domains binding gadd45 are 10 ⫾ 3 ␮M for T-p53C and 51 ⫾ 4 ␮M for T-p53C-R273H (Table 2) (7). At first sight, it seems that the mutation of R273H in the monomeric core domain weakens binding by only a factor of five. It can be shown that the observed P50 for four monomers binding to DNA with four binding sites is the fourth root of the product of the four individual dissociation constants for each binding event (⫽ (K1K2K3K4)1/4).5 The true relative binding affinity of “tetrameric” wild-type core domains (i.e. 4 domains to the four-site response element) to the R273H mutant is, therefore, (K1K2K3K4)R273H/(K1K2K3K4)wt, that is (51/10)4, ⬃700, and not 5.1, which is within experimental error of that calculated from the tetrameric constructs. Consequences of Mutations in Vivo—The similarity of the thermodynamic stability of full-length p53 to that of core domain and the parallel effects of destabilizing mutations indicate that the studies on rescuing the stability of core domain provide a valid framework for the design of drugs to rescue destabilized conformations. The rescue of destabilized mutants is a feasible drug 5

A. R. Fersht, unpublished data.

21940 JOURNAL OF BIOLOGICAL CHEMISTRY

Acknowledgments—We thank Dr. Dmitry Veprintsev for help with analytical ultracentrifugation, Dr. Chris Johnson for help with DSC experiments, Dr. Maria Rosario Fernandez-Fernandez and Dr. Stacey Rutledge for helpful advice, and Caroline Blair for assistance with protein purification. Wild-type full-length p53 was a kind gift from Dr. Heather Peto. REFERENCES 1. 2. 3. 4.

5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

Vogelstein, B., Lane, D., and Levine, A. J. (2000) Nature 408, 307–310 Beroud, C., and Soussi, T. (1998) Nucleic Acids Res. 26, 200 –204 Hainaut, P., and Hollstein, M. (2000) Adv. Cancer Res. 77, 81–137 Bullock, A. N., Henckel, J., DeDecker, B. S., Johnson, C. M., Nikolova, P. V., Proctor, M. R., Lane, D. P., and Fersht, A. R. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 14338 –14342 Bullock, A. N., Henckel, J., and Fersht, A. R. (2000) Oncogene 19, 1245–1256 Bullock, A. N., and Fersht, A. R. (2001) Nat. Rev. Cancer 1, 68 –76 Joerger, A. C., Ang, H. C., Veprintsev, D. B., Blair, C. M., and Fersht, A. R. (2005) J. Biol. Chem. 280, 16030 –16037 Fields, S., and Jang, S. K. (1990) Science 249, 1046 –1049 Mu¨ller-Tiemann, B. F., Halazonetis, T. D., and Elting, J. J. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 6079 – 6084 Walker, K. K., and Levine, A. J. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 15335–15340 Cho, Y., Gorina, S., Jeffrey, P. D., and Pavletich, N. P. (1994) Science 265, 346 –355 Jeffrey, P. D., Gorina, S., and Pavletich, N. P. (1995) Science 267, 1498 –1502 Ahn, J., and Prives, C. (2001) Nat. Struct. Biol. 8, 730 –732 El-Deiry, W. S., Kern, S. E., Pietenpol, J. A., Kinzler, K. W., and Vogelstein,

VOLUME 281 • NUMBER 31 • AUGUST 4, 2006

Downloaded from www.jbc.org at Karolinska institutet library on February 20, 2007

FIGURE 8. Stability of full-length p53 and its mutants compared with isolated core domains. Apparent melting temperatures for full-length p53 and its mutants are plotted against (A) changes in free energy upon urea-induced [D]50% unfolding of p53 core domain mutants (⌬⌬GD-N ) and (B) against the apparent melting temperature for p53 core domain and its mutants.

strategy. But, our quantitative data on the binding affinity of R273H indicate that it is too weak for functional transactivation of normal p53-responsive genes, as we can see from comparing the apparent KD values for recognition elements that p53 is known to bind weakly (32) with the experimental KD values reported here. For example, wild-type p53CT binds a low affinity promoter sequence PUMA BS1 with P50 of 260 nM per dimer (32), about 10 times more tightly than T-p53CT-R273H binding to a high affinity promoter sequence gadd45 with P50 of 3000 nM. This suggests that the binding of the R273H mutant to low affinity promoter sequences would be even weaker and would be insufficient for functional transactivation. Cell lines containing R273H have been used to test rescue drugs. We suggest that this is an inappropriate mutant as it does not lose stability on mutation but activity by the high loss of affinity. It is difficult to see how its activity could be revived by small molecules. Mutations in the DNA binding core domain have different, sometimes intermediate, effects on p53 transactivation, cell cycle arrest, and apoptotic functions in various cell types (39, 40). Possibly, many p53 mutants retain residual binding affinity for DNA, which may not be sufficient for normal function, but still allows them to bind to promoters with high affinity p53 binding sites, as in the p21 promoter, but not to promoters with low affinity sites, as in the Bax promoter (41). Given the complex nature of interactions and feedback loops in the p53 pathway (42), the differences in binding affinity could result in different levels of cell cycle control and different phenotypes (43, 44). But, massive up-regulation of a contact mutant could lead to its higher concentration offsetting its weaker binding affinity, especially noticeable for higher affinity response elements.

Stability and DNA Binding Affinity of Full-length p53 Mutants

AUGUST 4, 2006 • VOLUME 281 • NUMBER 31

1291–1296 32. Weinberg, R. L., Veprintsev, D. B., Bycroft, M., and Fersht, A. R. (2005) J. Mol. Biol. 348, 589 –596 33. Kastan, M. B., Zhan, Q., el-Deiry, W. S., Carrier, F., Jacks, T., Walsh, W. V., Plunkett, B. S., Vogelstein, B., and Fornace, A. J., Jr. (1992) Cell 71, 587–597 34. Veprintsev, D. B., Freund, S. M., Andreeva, A., Rutledge, S. E., Tidow, H., Canadillas, J. M., Blair, C. M., and Fersht, A. R. (2006) Proc. Natl. Acad. Sci. U. S. A. 103, 2115–2119 35. Chen, J. Y., Funk, W. D., Wright, W. E., Shay, J. W., and Minna, J. D. (1993) Oncogene 8, 2159 –2166 36. Chumakov, A. M., Miller, C. W., Chen, D. L., and Koeffler, H. P. (1993) Oncogene 8, 3005–3011 37. Pietenpol, J. A., Tokino, T., Thiagalingam, S., el-Deiry, W. S., Kinzler, K. W., and Vogelstein, B. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 1998 –2002 38. Chan, W. M., Siu, W. Y., Lau, A., and Poon, R. Y. (2004) Mol. Cell Biol. 24, 3536 –3551 39. Rowan, S., Ludwig, R. L., Haupt, Y., Bates, S., Lu, X., Oren, M., and Vousden, K. H. (1996) EMBO J. 15, 827– 838 40. Ryan, K. M., and Vousden, K. H. (1998) Mol. Cell Biol. 18, 3692–3698 41. Campomenosi, P., Monti, P., Aprile, A., Abbondandolo, A., Frebourg, T., Gold, B., Crook, T., Inga, A., Resnick, M. A., Iggo, R., and Fronza, G. (2001) Oncogene 20, 3573–3579 42. Harris, S. L., and Levine, A. J. (2005) Oncogene 24, 2899 –2908 43. Menendez, D., Inga, A., and Resnick, M. A. (2006) Mol. Cell Biol. 26, 2297–2308 44. Resnick, M. A., and Inga, A. (2003) Proc. Natl. Acad. Sci. U. S. A. 100, 9934 –9939 45. Nikolova, P. V., Wong, K. B., DeDecker, B., Henckel, J., and Fersht, A. R. (2000) EMBO J. 19, 370 –378 46. Kraulis, P. J. (1991) J. Appl. Crystallogr. 24, 946 –950 47. Merritt, E. A., and Bacon, D. J. (1997) Methods Enzymol. 277, 505–524 48. Gill, S. C., and von Hippel, P. H. (1989) Anal. Biochem. 182, 319 –326

JOURNAL OF BIOLOGICAL CHEMISTRY

21941

Downloaded from www.jbc.org at Karolinska institutet library on February 20, 2007

B. (1992) Nat. Genet. 1, 45– 49 15. Balagurumoorthy, P., Sakamoto, H., Lewis, M. S., Zambrano, N., Clore, G. M., Gronenborn, A. M., Appella, E., and Harrington, R. E. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 8591– 8595 16. Weinberg, R. L., Veprintsev, D. B., and Fersht, A. R. (2004) J. Mol. Biol. 341, 1145–1159 17. Hupp, T. R., Meek, D. W., Midgley, C. A., and Lane, D. P. (1992) Cell 71, 875– 886 18. Espinosa, J. M., and Emerson, B. M. (2001) Mol. Cell 8, 57– 69 19. Anderson, M. E., Woelker, B., Reed, M., Wang, P., and Tegtmeyer, P. (1997) Mol. Cell Biol. 17, 6255– 6264 20. Bayle, J. H., Elenbaas, B., and Levine, A. J. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 5729 –5733 21. Prives, C., and Manley, J. L. (2001) Cell 107, 815– 818 22. Feng, L., Lin, T., Uranishi, H., Gu, W., and Xu, Y. (2005) Mol. Cell Biol. 25, 5389 –5395 23. McKinney, K., Mattia, M., Gottifredi, V., and Prives, C. (2004) Mol. Cell 16, 413– 424 24. Weinberg, R. L., Freund, S. M., Veprintsev, D. B., Bycroft, M., and Fersht, A. R. (2004) J. Mol. Biol. 342, 801– 811 25. Bakalkin, G., Selivanova, G., Yakovleva, T., Kiseleva, E., Kashuba, E., Magnusson, K. P., Szekely, L., Klein, G., Terenius, L., and Wiman, K. G. (1995) Nucleic Acids Res. 23, 362–369 26. Friedler, A., Veprintsev, D. B., Freund, S. M., von Glos, K. I., and Fersht, A. R. (2005) Structure 13, 629 – 636 27. Friedler, A., Hansson, L. O., Veprintsev, D. B., Freund, S. M., Rippin, T. M., Nikolova, P. V., Proctor, M. R., Ru¨diger, S., and Fersht, A. R. (2002) Proc. Natl. Acad. Sci. U. S. A. 99, 937–942 28. Wiman, K. G. (2006) Cell Death Differ. 13, 921–926 29. Bykov, V. J., Selivanova, G., and Wiman, K. G. (2003) Eur. J. Cancer 39, 1828 –1834 30. Nikolova, P. V., Henckel, J., Lane, D. P., and Fersht, A. R. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 14675–14680 31. Joerger, A. C., Allen, M. D., and Fersht, A. R. (2004) J. Biol. Chem. 279,