DNA Bending Is Essential for the Site-specific Recognition of DNA Response Elements by the DNA Binding Domain of the Tumor Suppressor Protein p53*

(Received for publication, December 20, 1996, and in revised form, March 25, 1997)

Akhilesh K. Nagaich Dagger , Ettore Appella § and Rodney E. Harrington Dagger

From the Dagger  Department of Biochemistry/330, University of Nevada Reno, Reno, Nevada 89557-0014 and the § Laboratory of Cell Biology, Building 37, NCI, National Institutes of Health, Bethesda, Maryland 20892

ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES


ABSTRACT

We have used circular permutation assays to determine the extent and location of the DNA bend induced by the DNA binding domain of human wild type p53 (p53DBD) upon binding to several naturally occurring DNA response elements. We have found that p53DBD binding induces axial bending in all of the response elements investigated. In particular, response elements having a d(CATG) sequence at the junction of two consensus pentamers in each half-site favor highly bent complexes (bending angle is ~50°), whereas response elements having d(CTTG) bases at this position are less bent (bending angles from ~37 to ~25°). Quantitative electrophoretic mobility shift assays of different complexes show a direct correlation between the DNA bending angle and the binding affinity of the p53DBD with the response elements, i.e. the greater the stability of the complex, the more the DNA is bent by p53DBD binding. The study provides evidence that the energetics of DNA bending, as determined by the presence or absence of flexible sites in the response elements, may contribute significantly to the overall binding affinity of the p53DBD for different sequences. The results therefore suggest that both the structure and the stability of the p53-DNA complex may vary with different response elements. This variability may be correlated with variability in p53 function.


INTRODUCTION

The wild type tumor suppressor protein p53 plays a critical role in many key cellular processes (1-3). It acts as a transcriptional activator for a number of DNA damage and growth arrest genes including mdm2, gadd45, and p21/waf1/cip1; the product of the last is directly involved in inhibiting Cdk complexes leading to cell cycle arrest at the G1/S phase checkpoint (4-7). In addition to this protective role, p53 also up-regulates the human bax gene, the product of which heterodimerizes with the survival factor Bcl-2 and directly controls the apoptotic process (8). p53 also negatively regulates the transcription of genes that have TATA box-initiated promoters by binding to the protein components of the basal transcription machinery and is thought to be directly involved in checking both viral and eukaryotic DNA replication (9-11).

Wild type p53 binds response elements through a sequence-specific DNA binding domain (p53DBD)1 extending from amino acid residues 96-308 (12). Studies of tumor-derived p53 mutants have shown that they are defective in sequence-specific DNA binding and consequently cannot activate transcription (13). These studies strongly suggest that sequence-specific DNA binding and transactivation are the key biochemical activities responsible for much of the biological function of p53. Mutations in the p53 protein have been associated with more than half of all forms of human cancers, and the p53 gene is thought to be the most frequently mutated gene in human cancer (13). Over 1000 tumor-derived mutations have been found in p53, and the vast majority of these mutations are located in the p53DBD and affect its sequence-specific DNA binding (14). This fact, supported by many biochemical and molecular genetic experiments, has suggested that p53 function is mediated by its DNA binding activity and transactivation properties (12, 15).

Wild type p53 binds over 100 different naturally occurring response elements associated with different specific functions; it has been estimated that the human genome may contain as many as 200-300 such sites (16). p53 response elements differ in details of specific base sequence, but all contain two tandem decameric elements or half-sites, each a pentameric inverted repeat. Most decamers follow the consensus sequence pattern RRRC(A/T)|(A/T)GYYY, where R and Y are purines and pyrimidines, respectively, and the vertical bar denotes the center of pseudodyad symmetry (17). One copy of these decameric half-sites is insufficient for the functional binding of p53, defined as the ability to transcriptionally activate a nearby reporter gene, but some binding is preserved when the two copies are separated by up to 21 bp (16).

A cocrystal structure of the p53DBD nucleoprotein complex (18) has provided valuable insights into the binding specificity of p53 by identifying specific binding contacts (3, 19, 20). However, since the asymmetric unit contained only a single normally bound p53DBD, many questions remain concerning the multisubunit nature of the full p53 nucleoprotein complex, the determinants of DNA binding specificity, and the overall organization of p53 tetramers bound to the DNA recognition site. A recent study using T4 ligase-mediated cyclization and analytical ultracentrifugation has shown that p53DBD binds cooperatively as a tetrapeptide to an important functional response element, p21/waf1/cip1, and induces substantial bending in the DNA (21). More recent studies from this laboratory have shown that the requirement of DNA bending is maintained in dipeptide p53DBD complexes with several half-sites, which form at lower binding affinities than tetrapeptide complexes with full response elements.2 From molecular modeling studies based upon high resolution chemical probe data, we have proposed a structural model for the complex of four human p53DBD peptides with the p21/waf1/cip1 response element. This model has provided a unique insight into the possible roles of DNA flexibility in the sequence specificity of p53 binding and has provided a rationale for the requirement of DNA bending both in the full tetrapeptide complex and in the individual decameric half-sites (22). However, none of these prior studies have addressed the relationship between DNA bending and binding affinity in p53 nucleoprotein complexes.

In the present work, to further explore the relationships between DNA bending and p53 function, we have used circular permutation gel retardation and quantitative gel band shift assays to study the gel mobility retardation pattern of five biologically important binding sites: the p53 consensus binding sequence, a symmetric 20-bp binding sequence, the p21/waf1/cip1 response element, the ribosomal gene cluster (RGC) sequence, and the SV40 replication origin sequence (23-27). The accuracy of bend angles estimated using circular permutation assays has been questioned recently in the case of several bZip proteins including Fos and Jun (28, 29) (reviewed by Hagerman (30)) whose elongated shape and leucine zipper region differentiates them from most globular proteins. On the other hand, circular permutation assays provide bending angles that are in excellent agreement with results obtained from phasing and cyclization experiments and from x-ray crystallography for many globular proteins such as CAP, Cro, and the TATA-binding protein, and this method has been widely used to estimate DNA bending in such systems (31, 32). Since the p53DBD peptide has a similar globular conformation (18), we believe that the circular permutation assay provides a satisfactory approximation to the true induced bending angles for this peptide bound to the various DNA response elements investigated in this work. Our data clearly show that the p53DBD binds with all of these binding sites cooperatively as a tetrapeptide and induces bending of the DNA. However, the bending angle varies considerably with different response elements over the range of 52-25°. Response elements having a d(CATG) sequence at the junctions of their consensus pentamers, i.e. the p53 consensus and symmetric sites and the p21/waf1/cip1 response element, are bent by ~50°, whereas bending is much less (~37-~25°) in the case of the RGC and SV40 response elements, both of which have a d(CTTG) sequence at the pentamer junctions.

Our results confirm an earlier report that p53DBD binds response elements cooperatively as a tetrapeptide and bends the DNA (21). In this earlier work, cyclization studies were used to estimate a bending angle of ~50° in the p21/waf1/cip1 response element; the excellent agreement with the present result for this binding site provides further justification for the use of circular permutation to quantitate p53DBD DNA bending. We also find a positive correlation between the observed bending angles and binding affinities, i.e. the response elements that show higher bending angles also show higher p53DBD binding affinities, while response elements showing lower bending angles show much less binding affinity. A direct correlation between the binding affinity of p53 with various response elements and transcription activation has been observed by Kern et al. (25). Furthermore, p53 consensus and RGC response elements have been characterized as distinctly different p53 binding elements; the former binds with certain p53 mutants and enhances transcriptional activity of the reporter genes, while the later does not bind with mutants (33-35). Thus, our findings suggest that DNA bending may be a ubiquitous feature of p53DBD-DNA complexes irrespective of the DNA sequence and imply that the energetics of DNA bending may contribute significantly to the binding affinity of p53. In this event, the change in free energy associated with DNA bending in the p53-DNA complex may fine tune the transcriptional activation of p53-regulated genes.


MATERIALS AND METHODS

Preparation and Purification of p53DBD

A human p53 cDNA clone encoding amino acid residues 96-308 was amplified by polymerase chain reaction using p53-specific primers 5'-ATATCATATGGTCCCTTCCCAGAAAACCTA-3' and 5'-ATATGGATCCTCACAGTGCTCGCTTAGTGCTC-3'. The amplified product was cloned in the pET12a expression vector (Novagen), and the core DNA binding domain was overproduced in Escherichia coli BL21 (DE3). The cells were incubated at 37 °C until an A600 of 0.6-1.0 was attained, and 0.25 mM isopropyl beta -D-thiogalactoside was added to induce the expression of the recombinant protein. Cells were harvested after 2 h by centrifugation, lysed in a French press, and sonicated for 2 min in 40 mM MES, pH 6.0, 100 mM NaCl, 5 mM dithiothreitol. The soluble fraction was loaded onto a Resource S column (Pharmacia Biotech Inc.) in 40 mM MES, pH 6.0, 5 mM dithiothreitol and was eluted in a 0-400 mM NaCl gradient. The pooled fractions were precipitated by ammonium sulfate addition to 80% saturation and purified further on a Sephadex 75 HR gel filtration column (Pharmacia) in 50 mM bis-Tris propane-HCl, pH 6.8, 100 mM NaCl, 1 mM dithiothreitol. The purified p53DBD was checked on an SDS-polyacrylamide gel for purity.

Oligonucleotide Synthesis, Plasmid Constructs, and Circular Permutation Gel Retardation Assay

All of the oligonucleotides used in this study contained a 20-bp consensus p53 binding site in a 30-mer oligonucleotide having XbaI and SalI cohesive termini (Fig. 1A). The oligonucleotides were synthesized using beta -cyanoethyl phosphoramidite chemistry and purified on polyacrylamide gel using standard procedures (36). These oligonucleotides were directionally subcloned at the XbaI and SalI restriction sites of DNA bending vector pBend3 (37). The recombinant plasmids were prepared by an alkaline lysis procedure and sequenced to confirm the size and orientation of the insert. The purified plasmids were digested with MluI, NheI, SpeI, EcoRV, StuI, NcoI, and BamHI restriction enzymes to generate DNA fragments having a circularly permuted p53 consensus binding site (Fig. 1B). The digested fragments were purified on a low melting agarose gel using a Geneclean kit (Bio101). The cohesive end fragments were labeled with [alpha -32P]dCTP and the Klenow fragment, whereas the blunt-ended fragments were dephosphorylated with calf intestinal alkaline phosphatase and then labeled with [gamma -32P]ATP and polynucleotide kinase. The labeled DNA fragments were again purified on a native polyacrylamide gel to eliminate traces of free label and other undesired fragments.


Fig. 1. A, 30-mer oligonucleotides with XbaI, SalI cohesive termini, containing 20-bp p53 response elements (boldface type, numbered and boxed): p53 consensus sequence (a), p21/waf1/cip1 (b), symmetric sequence (c), ribosomal gene cluster sequence (RGC) (d), SV40 promoter sequence (e). Each half-site is separated by a solid line. Each quarter-site is separated by a dashed line and is numbered. B, a schematic representation of the recombinant plasmids pCon30, pWaf30, pSS30, pRGC30, and pSV30 containing a p53 binding site (hatched box), flanked by tandemly duplicated DNA sequences. The plasmids were cleaved at the restriction sites indicated in the map. The DNA fragments obtained in this way (designated as MluI (M), NheI (Nh), SpeI (Sp), EcoRV (E), StuI (S), NcoI (Nc), and BamHI (B)) contained circular permutations of the same sequence of 145-149 bp.
[View Larger Version of this Image (31K GIF file)]

The labeled DNA fragments were mixed with poly(dI-dC) (200 ng) and incubated with the purified p53DBD in the DNA binding buffer (50 mM bis-Tris propane, HCl, pH 6.8, 1 mM dithiothreitol, and 50 mM NaCl) (10 µl) at 4 °C for 40 min. The amount of p53DBD was adjusted such that no more than 50% of the DNA was retarded due to complex formation. The reactions were mixed with 15% Ficoll (2 µl) and loaded onto 5% native polyacrylamide gel. The gels were run in 0.25 × TBE at 150 V for about 1.5 h, dried, and autoradiographed.

Calculation of p53DBD-induced DNA Bend Angles and Bend Center

p53DBD-induced DNA bending was measured by electrophoretic mobility shift assay using DNA fragments containing circularly permuted binding sites. The relative mobility (RF) of the p53DBD-DNA complexes was defined as follows,
R<SUB>F</SUB>=<FR><NU>Cm<SUB><UP>complex</UP></SUB></NU><DE>Cm<SUB><UP>free</UP></SUB></DE></FR> (Eq. 1)
where Cm represents the migration of the fragments in the gel in centimeters. The lengths of the different probes varied from 145 to 149 bp due to the sequence of the pBend3 polylinker and the difference in labeling procedures (end filling versus kinasing). To adjust for the slight differences in the probe length, the relative mobilities of the p53DBD-DNA complexes were normalized against highest mobility probe. The average relative mobility from three different experiments was plotted as a function of the fractional displacement, defined as the distance in base pairs from the center of the p53DBD response element to the 5'-end of the noncoding strand of the DNA fragment divided by the total length of the fragment. These data, when plotted as relative mobility versus fractional displacement, could be approximated by a least squares fit to a parabolic function of the form y = ax2 + bx + c. The bending center (y = minimum) was determined by setting dy/dx to 0 and solving for x. To determine bending angles, we have used the mathematical treatment derived by Ferrari et al., where the bend angle theta  is determined from the values for a, b, and c, taken from the least squares fit to the quadratic equation, using the following equation: a or (-b) = 2c(1 + costheta ). With this method, the values of theta  using a or -b should be identical, and their comparison offers a means for estimating the error in the measured bending angle. The angle of deviation from linearity, alpha , is related to theta  by the equation alpha  = 180° - theta  (38).

The magnitude of the protein-induced DNA bending was also calculated by a semiempirical equation described by Thompson and Landy (39).
<FR><NU>&mgr;<SUB>M</SUB></NU><DE>&mgr;<SUB>E</SUB></DE></FR>=<UP>cos</UP><FENCE>k <FR><NU>&agr;</NU><DE>2</DE></FR></FENCE> (Eq. 2)
where µM and µE are the mobility of the protein-DNA complex with the binding site at the middle and end of the DNA fragment, respectively, k is a coefficient that depends upon the electrophoresis conditions, and alpha  is the angle of the protein-induced DNA bend. A k value of ~1 was obtained for our gel electrophoretic conditions using a calibration oligonucleotide having an intrinsic bend of 54°. The magnitude of the p53DBD-induced bend angles were determined using this same value of k.

Electrophoretic Mobility Shift Assay and Determination of Apparent Equilibrium Dissociation Constants

DNA binding affinities of the p53DBD with different response elements were assayed using electrophoretic mobility shift assays. The 30-mer oligonucleotide duplexes were labeled with [gamma -32P]ATP and polynucleotide kinase, purified on a native polyacrylamide gel, and quantitated by UV spectrophotometry, using epsilon  values of 15,500 M-1 cm-1, 8,500 M-1 cm-1, 12,500 M-1 cm-1, and 7,500 M-1 cm-1 for the A, T, G, and C residues, respectively. The protein concentration was measured spectrophotometrically using a Bradford protein assay kit (Bio-Rad). The specific DNA binding activity of these preparations was found to be ~90%. Each labeled oligonucleotide (~15 pmol) was mixed with poly(dI-dC) (200 ng) and incubated with increasing amounts of the known concentrations of active p53DBD in DNA binding buffer (10 µl) for 40 min at 4 °C. The samples were mixed with 15% Ficoll (2 µl) to facilitate gel loading and electrophoresed on a pre-electrophoresed 7% polyacrylamide gel (29:1, acrylamide:bis) in 0.25 × TBE for 2 h at 8 V/cm at 4 °C. The loading was done rapidly to minimize the equilibration of the complex with the running buffer. The gels were then transferred to Whatman filter paper, dried under vacuum, and exposed to the x-ray film. The same gel was used to quantitate the free and the bound DNA using a PhosphorImager (Bio-Rad). The exposure was adjusted to achieve a linear response of radioactivity with measured band intensities. Radioactivity scattered between the bound and the free bands was counted with the free probe. The quantitation of the different bands was carried out using Molecular Analyst, a Bio-Rad gel analysis software using standard procedures (40).

The fraction of the free and bound molecules at each protein concentration was calculated by dividing the optical density of the band by the sum of the optical densities in all the bands in the same lane. Since the concentration of the DNA in these reaction mixtures was always <= 15 pmol, it was assumed that [p53DBD]total approx  [p53DBD]free, so the protein concentration required for half maximal binding is very close to Kd (app). The data were plotted in the form of a Bjerrum plot of the fraction of free DNA versus the log of protein concentration, and the dissociation constants (Kd) were determined as the protein concentration at which half of the free DNA was bound. In determining Kd (app), we chose the disappearance of the free DNA versus protein concentration rather than the appearance of the complex bands to account for any dissociation of the complex during the running of the gel. The free energy of binding was calculated using the relation Delta G0 =-RT ln K(app) at 4 °C, and the free energy of bending was calculated from the following equation (23).
&Dgr;G<SUB><UP>bend</UP></SUB>=<FR><NU>0.014</NU><DE>£(<UP>bp</UP>)</DE></FR> (&Dgr;&thgr;)<SUP>2</SUP>(<UP>kcal</UP>/<UP>mol</UP>) (Eq. 3)
In Equation 3, Delta theta is the bending angle in degrees induced by binding of protein to DNA, and £(bp) is the number of base pairs involved in the complex. The DNA is assumed to have a persistence length of 150 bp (41, 42).

We have described the cooperative binding of the p53DBD with the response elements using the method of Senear and Brenowitz (43) for the two-site cooperative binding of protein with the DNA as described below.

In Scheme 1, k1 and k2 are the intrinsic binding constants to sites 1 and 2, respectively, and k12 is the cooperativity coefficient. The fraction of the molecules [Theta ]i that are free, single, and doubly liganded can be written, respectively, as follows,
&THgr;<SUB>0</SUB>=<FR><NU>1</NU><DE>1+(k<SUB>1</SUB>+k<SUB>2</SUB>)[X]+(k<SUB>1</SUB>k<SUB>2</SUB>k<SUB>12</SUB>)[X<SUP>2</SUP>]</DE></FR> (Eq. 4)
&THgr;<SUB>1</SUB>=<FR><NU>(k<SUB>1</SUB>+k<SUB>2</SUB>)[X]<SUP>2</SUP></NU><DE>1+(k<SUB>1</SUB>+k<SUB>2</SUB>)[X]+(k<SUB>1</SUB>k<SUB>2</SUB>k<SUB>12</SUB>)[X<SUP>2</SUP>]</DE></FR> (Eq. 5)
&THgr;<SUB>2</SUB>=<FR><NU>(k<SUB>1</SUB>k<SUB>2</SUB>k<SUB>12</SUB>)[X]<SUP>2</SUP></NU><DE>1+(k<SUB>1</SUB>+k<SUB>2</SUB>)[X]+(k<SUB>1</SUB>k<SUB>2</SUB>k<SUB>12</SUB>)[X<SUP>2</SUP>]</DE></FR> (Eq. 6)
where X is the concentration of the free protein. Plots of the fraction of free and bound DNA versus the log of protein concentrations were fitted to Equations 4 and 6, respectively, and the cooperativity parameter was calculated as described by Senear and Brenowitz (43).


[View Larger Version of this Image (6K GIF file)]

Fig. S1.


RESULTS

Binding of p53DBD to the Response Elements Bends the DNA

Using T4 DNA ligase-mediated cyclization assays, our laboratory has previously shown that the p53DBD binds the response elements cooperatively as a tetrapeptide and bends the DNA (21). High resolution chemical footprinting has shown that four p53DBD peptides bind the full 20-bp p21/waf1/cip1 response element DNA in a staggered array, and molecular modeling based upon protein-DNA contacts identified by base-specific chemical probes has suggested that the bound p53DBD must bend the response element DNA to relieve two types of steric clashes among the bound peptides (22). In the present work, we have used the circular permutation polyacrylamide gel retardation assay (44) to confirm and amplify these findings and to examine the DNA bending relationships among different p53 response elements. The gel retardation assay is based on the assumption that static bending in DNA leads to a reduction in gel mobility as predicted by the Lumpkin-Zimm reptation model, in which the mobility of DNA through a gel is a quadratic function of its mean square end-to-end distance (45, 46). Uniform length DNA fragments show increasingly anomalous electrophoretic mobility on polyacrylamide gels, as the bending locus is moved from the ends to the center of the DNA fragment. Although the relationship between electrophoretic mobility and conformation for protein-DNA complexes is complex, a comparison of the mobilities of circularly permuted p53DBD-DNA complexes allows a relatively precise determination of the DNA bending locus and an estimate of the bending angle (38, 47).

Gel mobility retardation data for all circularly permuted fragments containing p53DBD bound to p53 consensus, p21/waf1/cip1, symmetric, RGC and SV40 response elements (Fig. 1) are shown in Fig. 2, A-E. Corresponding plots of relative gel mobility as functions of flexure displacements of different DNA sequences are shown in Fig. 2, F-J. The unbound fractions of the circularly permuted DNA fragments in all of the gels (bands marked as F in Fig. 2, A-E) show similar mobilities on the polyacrylamide gels, whereas the mobilities for bound fractions (bands marked as C) are clearly anomalous. The mobilities are maximum when the bound response elements are near the ends of the fragments (lanes marked as M and B) and minimum when they are near the centers (lanes marked as E). Intermediate mobilities are observed at intermediate positions (lanes marked as Nh, Sp, S, and Nc). In particular, the bound fractions for the p53 consensus, p21/waf1/cip1, and symmetric response elements (Fig. 2, F-H) show much higher migration anomalies compared with RGC and SV40 (Fig. 2, I and J), as is evident also from a comparison of calculated alpha  values for these sequences. Nevertheless, all of the response elements investigated show gel migration anomalies characteristic of protein-induced DNA bending, and the smallest bending angle observed, alpha  = 25° for the SV40 site (Fig. 4, E and J), still represents considerable DNA bending in this complex.


Fig. 2. A-E, electrophoretic mobility of the circularly permuted DNA fragments complexed to the human wild type p53DBD. The fragments were generated from the digestion of the plasmids pCon30 (A), pWaf30 (B), pSS30 (C), pRGC30 (D), and pSV30 (E) with restriction enzymes MluI (M), NheI (Nh), SpeI (Sp), EcoRV (E), StuI (S), NcoI (Nc), and BamHI (B). The bands marked as O indicate the origin of the wells; bands marked as C show the p53DBD-DNA complex; and bands labeled F show the free DNA. F-J, analysis of the induced DNA bends. Bending angles (alpha ) were calculated by the algorithm of Ferrari et al. (38). The second order equations fitted were as follows: y = 0.4165X2 - 4100X + 0.5479 (R2 = 0.9945) (F), y = 0.3367X2 - 0.3290X + 0.4863 (R2 = 0.9894) (G), y = 0.3437X2 - 0.3408X + 0.5306 (R2 = 0.9937) (H), y = 0.4142X2 - 0.2138X + 0.5406 (R2 = 0.9998) (I), and y = 0.0958X2 - 0.0940X + 0.5364 (R2 = 0.9814) (J).
[View Larger Version of this Image (25K GIF file)]


Fig. 4. Titration curves resolved from the gel mobility shift experiments of p53DBD with p53 consensus sequence (A), p21/waf1/cip1 (B), symmetric sequence (C), RGC sequence (D), and SV40 binding sequence (E). The fraction of free DNA averaged from three independent experiments versus the log of p53DBD concentrations has been plotted. The solid lines indicate the theoretical curve generated by fitting the data to Equation 4 (see "Materials and Methods").
[View Larger Version of this Image (18K GIF file)]

The data of Fig. 2 support the interpretation that the observed electrophoretic mobilities depend upon the position of the bound p53DBD within each DNA fragment and hence that the mobility anomalies must be due to DNA bending. However, to ensure that the gel conditions used were not a factor in the apparent differences in mobility, identical experiments were also carried out using 4-10% polyacrylamide gels and over different time intervals. In all cases, the most anomalous (retarded) mobilities occurred with the p53DBD bound to the fragment digested with EcoRV in which the p53 binding site was located in the center of the fragment. Displacement analyses of the gel data similar to Fig. 2, F-J, showed the bending loci to be centered in the p53 binding sites, i.e. at bases C9-G12 (Fig. 1A, a-e). In all of these cases, the calculated bending angles obtained using the geometrical model of Ferrari et al. (38) were identical to those in Table I and Fig. 2, F-J: 52° for the p53 consensus sequence; 48° for the symmetric sequence; and 50, 37, and 25° for the p21/waf1/cip1, RGC, and SV40 response elements, respectively. Bending angles were also calculated using a semiempirical equation developed by Thompson and Landy (39) and were found to be consistent within the error limits of the experiments (see "Materials and Methods").

Table I. Binding of p53DBD with various DNA response elements

Determination of apparent equilibrium dissociation constants Kd (app), free energy of binding (Delta G0), and free energy of bending (Delta Gbend) is shown.

p53 response elements Bending angles (Ref. 38) Equilibrium dissociation constants (Kd) Free energy of binding (Delta G0) Free energy of bending (Delta Gbend)

degrees M kcal M-1 kcal M-1
p53 consensus 52 3.7  × 10-9  -10.7 1.90
p21/waf1/cip1 50 7.0  × 10-9  -10.3 1.75
Symmetric 48 9.0  × 10-9  -10.2 1.61
RGC 37 1.8  × 10-8  -9.8 0.95
SV40 25 3.0  × 10-8  -9.5 0.43

Electrophoretic Mobility Shift Assays and Quantitation of Apparent Equilibrium Binding Constants

Electrophoretic mobility shift assays (EMSAs) on polyacrylamide gels represent a powerful approach to the analysis of both the qualitative and quantitative aspects of protein-nucleic acid interactions. A number of studies have indicated that accurate equilibrium binding constants of protein-DNA complexes can be obtained using EMSA under carefully controlled conditions (48, 49).

Fig. 3 shows EMSA results for the p53 binding sites studied here with increasing amounts of p53DBD. The p53 consensus, p21/waf1/cip1, and symmetric sequences show a highly cooperative tetrapeptide binding with the p53DBD starting at relatively low protein concentrations (DNA:protein molar ratio <= 5, bands marked as C, Fig. 3, A-C), and with no other species of lower stoichiometry. The RGC sequence under identical protein concentrations shows reduced p53DBD binding with much less cooperativity (bands marked as C, Fig. 3D). The tetramer complex appears to undergo weak dissociation under identical gel conditions. The SV40 binding sequence shows the weakest binding of all of the sequences investigated here. The tetramer complex starts to appear only at a 10 M excess of p53DBD and is not fully titrated even at an 80 M excess of p53DBD (bands marked as C, Fig. 3E). Despite their relatively weak binding affinities, the RGC and SV40 sequences do not produce appreciable species of intermediate stoichiometry, although a slight band smearing is visible between the free DNA and tetramer bands.


Fig. 3. Quantitative gel mobility retardation assay of the 30-mer oligonucleotide duplexes containing p53 response elements. A, p53 consensus sequence; B, p21/waf1/cip1 response element; C, symmetric sequence; D, RGC sequence; E, SV40 binding sequence. The bands marked as C show the tetrapeptide p53DBD complex, and the bands labeled F show the unbound DNA. The molar ratios of p53DBD:DNA are given as the abscissas of each gel.
[View Larger Version of this Image (35K GIF file)]

A quantitative analysis of the gel mobility shift data of three such independent experiments is presented in the form of Bjerrum plots of the fraction of free DNA versus the log of p53DBD concentration in Fig. 4, A-E. Estimates of equilibrium dissociation constants, obtained as described under "Materials and Methods," are given in Table I; p53DBD associates with the consensus, p21/waf1/cip, and symmetric sequences with mean equilibrium dissociation constants of 3.7 × 10-9 M, 7.0 × 10-9 M, and 9.0 × 10-9 M, respectively and with RGC and SV40 sequences with respective values of 1.8 × 10-8 M and 3.0 × 10-8 M. Thus, the binding affinities of p53DBD with the RGC and SV40 sequences are 5-12-fold reduced over those of the other binding sites. A visual estimation from the slope of the transition curves also indicates that the binding cooperativity of p53DBD with the consensus, p21/waf1/cip and symmetric sequences is significantly greater than with the RGC and SV40 sequences.

We have used the method of Brenowitz and co-workers (43, 49, 50) to estimate the cooperativity parameter, k12, from the EMSA data for p53DBD with the symmetric binding site, which has two identical half-sites. Since minimal p53DBD binding occurs only to the decameric half-site with negligible binding to single pentamers,2 we can rationalize our present results using a two-site cooperative model in which binding of pairs of p53DBD peptides occurs independently to each (decameric) site (see "Materials and Methods"). The fractions of bound and free DNA were plotted versus the log of protein concentration, and the data were fitted to Equations 4 and 6, respectively. For binding to two identical and noncooperative half-sites, k1 would equal k2, and k12 would be 1. For the full 20-bp symmetric site, bound and free DNA fractions from integrated band intensity data (Fig. 3C) were plotted in the form of Bjerrum plots and were fitted to Equations 4 and 6 (see "Materials and Methods") for various values of k12, constraining k1 = k2. The cooperativity parameter k12 approx  170 was calculated based on the best fit data. Our data suggest that p53DBD binding is much more cooperative to asymmetric sites such as p53 consensus and p21/waf1/cip1, but this additional cooperativity could not be quantitated for these binding sites since the two half-sites are different. The method of quantitating cooperativity used here is not as robust as the method based on DNase footprinting (51), but we nevertheless believe that the present results provide a meaningful comparison of cooperativity effects between the binding of p53DBD and other specific DNA binding proteins (50, 52).


DISCUSSION

We have summarized our data on the DNA bending and binding properties of p53DBD with different response elements in Table I. It is clear that the p53DBD binding induces considerable bending in all of the DNA binding sites studied. The propensity for DNA bending appears to be correlated with specific sequence elements. The p53 consensus, p21/waf1/cip1 and symmetric binding sites, which are significantly more sharply bent, have a d(CATG) sequence element at the pentamer junctions, whereas the RGC and SV40 sites have d(CTTG) elements at these same locations (Fig. 1). The equilibrium dissociation constants determined here for the p53 consensus, p21/waf1/cip1, and symmetric binding sites are in the range of 3.7-9 × 10-9 M, which is in general agreement with the result from analytical ultracentrifugation of Kd = (8.3 ± 1.4) × 10-8 M for the p21/waf1/cip1 response element (21). The binding to RGC and SV40 sequences is much weaker (Kd values are 1.8 and 3.0 × 10-8 M for RGC and SV40 sequences, respectively). The 5-10-fold discrepancy between these sets of values may be due to a lower dissociation rate of the complex in polyacrylamide gel due to caging effects (52). Although p53DBD is found to bind cooperatively to all of the investigated binding sites, the EMSA data strongly suggest that the binding of p53DBD with the latter two response elements is also much less cooperative (Fig. 3, D and E, and Fig. 4).

The data also point up a clear correlation between the observed DNA bending angles and the estimated free energies of binding (Table I). The response elements with higher bending angles in the nucleoprotein complex also show a tighter binding of p53DBD with the DNA, whereas the binding sites with the lower bending angles show much weaker binding. These results suggest that the energetics of DNA bending contribute significantly to the site-specific binding of p53DBD with the response elements. A number of studies have shown a correlation between the binding affinity of the p53 with response elements and the transcriptional activation of nearby reporter genes (20, 25, 53, 54). A similar correlation has also been observed in the case of TATA-binding protein, which also bends its target binding sites (32). It is possible that p53 differentially fine tunes the transcriptional activation of its target genes by its interaction with the responsive promoters, and the free energy of binding and DNA bending play significant discriminatory roles.

The results presented here are consistent with a structural model for the p53DBD-p21/waf1/cip response element complex developed by us from molecular modeling studies based upon the cocrystal structure of the p53DBD peptide (18) and high resolution specific protein-DNA contacts obtained using high resolution chemical probes (22). In this model, four p53DBD peptides bind to the full 20-bp response element in a staggered array, with each peptide binding to a single pentamer. The DNA is bent away from the complex to alleviate severe interpeptide clashes. These are fully relieved if the p21/waf1/cip1 DNA is bent by a 15° positive roll toward the major grooves in d(CA)·d(TG) dimers at the positions indicated by double arrows: GAAC Up-arrow  A|TUp-arrow GTCCCAACUp-arrow A|TUp-arrow GTTG. In the model, the roll in the d(CA)·d(TG) doublets at the pentamer junctions leads to an overall bend of 50° over the full response element. This is fully consistent with bending angles reported in this work for the p53 consensus, p21/waf1/cip1, and symmetric sequences, all of which have conserved d(CATG) sequence elements at the pentamer junctions. The putative rolls in the d(CA)·d(TG) steps are consistent with the flexible character of these dinucleotides in free B-DNA (55-57) and in protein-DNA complexes as shown by crystallography (58, 59) and gel electrophoresis (23, 60, 61). Energy calculations have also suggested that these dinucleotides can bend anisotropically toward the major groove (62). In the present comparison, however, the RGC and SV40 binding site sequences have d(CTTG) sequences at the pentamer junctions. These can presumably also roll toward the major groove at the d(TG) steps, but the overall bending would be expected to be less, and the conformation of the complex would be affected by the replacement of d(CA) by d(CT) dinucleotides at critical putative bending sites in the response elements. We propose that the higher bending and the binding affinity in the case of p53 consensus, p21/waf1/cip1, and symmetric sequences is mediated by the greater inherent flexibility associated with the critically positioned d(CATG) sequence elements.

It is known that d(CA)·d(TG) sequence elements are ubiquitous in promoter and other genomic regulatory sites and are particularly so in those sites for which DNA bending has been demonstrated or is thought to occur in regulatory nucleoprotein complexes (63). It is a reasonable postulate that binding specificity as well as complex stability may be enhanced by flexibility in binding site DNA, leading to more precise and perhaps more extensive protein-protein interactions and protein-DNA contacts (55, 56, 62, 64). Most functional p53 binding sites contain helically phased d(CA)·d(TG) sequence elements in approximately the same relative positions. This requirement may also provide a structural explanation for the fact that p53 response elements that contain spacers between the palindromic half-sites have generally been observed to be highly unstable and nonfunctional (16).

The data presented in this paper suggest that the sequence-dependent structural and dynamical properties of the DNA modulate both the stability and structure of p53DBD nucleoprotein complexes and may therefore have profound effects on the function of wild type and mutated p53. The present results also imply that both the sequence specificity of binding and the stability of p53-DNA nucleoprotein complexes are mediated by a complex interplay of protein-protein interactions and sequence-specific protein-DNA contacts, both of which are in turn regulated by sequence-specific flexibility in the structure of response element DNA.


FOOTNOTES

*   This work was supported by National Science Foundation Grant MCB 9117488, National Institutes of Health Grant CA70274, U.S. Department of Agriculture Hatch Project NEV032D through the Nevada Experiment Station (to R. E. H.), and the AIDS Targeted Anti-Viral Program of the Office of the Director of the National Institutes of Health (to E. A.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
   To whom correspondence and reprint requests should be addressed.
1   The abbreviations used are: p53DBD, p53 DNA binding domain; bp, base pair(s); RGC, ribosomal gene cluster; MES, 4-morpholineethanesulfonic acid; bis-Tris, 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)-propane-1,3-diol.
2   P. Balagurumoorthy, unpublished results.

ACKNOWLEDGEMENTS

We thank Dr. Sankar Adhya for a generous gift of the pbend3 circular permutation vector. We also thank Professor Ilga Winicov and Dr. Vaijayanti Pethe for many helpful suggestions.


REFERENCES

  1. Lane, D. P. (1992) Nature 358, 15-16 [CrossRef][Medline] [Order article via Infotrieve]
  2. Levine, A. J. (1993) Annu. Rev. Biochem. 62, 623-651 [CrossRef][Medline] [Order article via Infotrieve]
  3. Ko, L. J., and Prives, C. (1996) Genes & Dev. 10, 1054-1072 [CrossRef][Medline] [Order article via Infotrieve]
  4. Barak, Y., Juven, T., Haffner, R., and Oren, M. (1993) EMBO J. 12, 461-468 [Abstract]
  5. 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 [Medline] [Order article via Infotrieve]
  6. El-Deiry, W. S., Tokino, T., Velculescu, V. E., Levy, D. B., Parsons, R., Trent, J. M., Lin, D., Mercer, W. E., Kinzler, K. W., and Vogelstein, B. (1993) Cell 75, 817-825 [Medline] [Order article via Infotrieve]
  7. Harper, J. W., Adami, G. R., Wei, N., Keyomarsi, K., and Elledge, S. J. (1993) Cell 75, 805-816 [Medline] [Order article via Infotrieve]
  8. Miyashita, T., and Reed, J. C. (1995) Cell 80, 293-299 [Medline] [Order article via Infotrieve]
  9. Horikoshi, N., Usheva, A., Chen, J. D., Levine, A. J., Weinmann, R., and Shenk, T. (1995) Mol. Cell. Biol. 15, 227-234 [Abstract]
  10. Bargonetti, J., Friedman, P. N., Kern, S. E., Vogelstein, B., and Prives, C. (1991) Cell 65, 1083-1091 [Medline] [Order article via Infotrieve]
  11. Dutta, A., Ruppert, S. M., Aster, J. C., and Winchester, E. (1993) Nature 365, 79-82 [CrossRef][Medline] [Order article via Infotrieve]
  12. Bargonetti, J., Manfredi, J. J., Chen, X., Marshak, D. R., and Prives, C. (1993) Genes & Dev. 7, 2565-2574 [Abstract]
  13. Hollstein, M., Sidransky, D., Vogelstein, B., and Harris, C. C. (1991) Science 253, 49-53 [Medline] [Order article via Infotrieve]
  14. Harris, C. C. (1993) Science 262, 1980-1981 [Medline] [Order article via Infotrieve]
  15. Pavletich, N. P., Chambers, K. A., and Pabo, C. O. (1993) Genes & Dev. 7, 2556-2564 [Abstract]
  16. Tokino, T., Thiagalingam, S., El-Deiry, W. S., Waldman, T., Kinzler, K. W., and Vogelstein, B. (1994) Hum. Mol. Genet. 3, 1537-1542 [Abstract]
  17. El-Deiry, W. S., Kern, S. E., Pietenpol, J. A., Kinzler, K. W., and Vogelstein, B. (1992) Nat. Genet. 1, 45-49 [Medline] [Order article via Infotrieve]
  18. Cho, Y. J., Gorina, S., Jeffrey, P. D., and Pavletich, N. P. (1994) Science 265, 346-355 [Medline] [Order article via Infotrieve]
  19. Prives, C. (1994) Cell 78, 543-546 [Medline] [Order article via Infotrieve]
  20. Prives, C., Bargonetti, J., Farmer, G., Ferrari, E., Friedlander, P., Wang, Y., Jayaraman, L., Pavletich, N., and Hubscher, U. (1994) Cold Spring Harbor Symp. Quant. Biol. 59, 207-213 [Medline] [Order article via Infotrieve]
  21. 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 [Abstract]
  22. Nagaich, A. K., Zhurkin, V. B., Sakamoto, H., Gorin, A. A., Clore, G. M., Gronenborn, A. M., Appella, E., and Harrington, R. E. (1997) J. Biol. Chem. 272, 14830-14841 [Abstract/Free Full Text]
  23. Liu-Johnson, H.-N., Gartenberg, M. R., and Crothers, D. M. (1986) Cell 47, 995-1005 [Medline] [Order article via Infotrieve]
  24. Namba, H., Hara, T., Tukazaki, T., Migita, K., Ishikawa, N., Ito, K., Nagataki, S., and Yamashita, S. (1995) Cancer Res. 55, 2075-2080 [Abstract]
  25. Kern, S. E., Pietenpol, J. A., Thiagalingam, S., Seymour, A., Kinzler, K. W., and Vogelstein, B. (1992) Science 256, 827-830 [Medline] [Order article via Infotrieve]
  26. Funk, W. D., Pak, D. T., Karas, R. H., Wright, W. E., and Shay, J. W. (1992) Mol. Cell. Biol. 12, 2866-2871 [Abstract]
  27. Bargonetti, J., Reynisdottir, I., Friedman, P. N., and Prives, C. (1992) Genes & Dev. 6, 1886-1898 [Abstract]
  28. Kerppola, T. K., and Curran, T. (1993) Mol. Cell. Biol. 13, 5479-5489 [Abstract]
  29. Sitlani, A., and Crothers, D. M. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 3248-3252 [Abstract/Free Full Text]
  30. Hagerman, P. J. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 9993-9996 [Abstract/Free Full Text]
  31. VanderVliet, P. C., and Verrijzer, C. P. (1993) BioEssays 15, 25-32 [Medline] [Order article via Infotrieve]
  32. Starr, D. B., Hoopes, B. C., and Hawley, D. K. (1995) J. Mol. Biol. 250, 434-446 [CrossRef][Medline] [Order article via Infotrieve]
  33. Prives, C., and Manfredi, J. J. (1993) Genes & Dev. 7, 529-534 [CrossRef][Medline] [Order article via Infotrieve]
  34. McClure, W. R. (1985) Annu. Rev. Biochem. 54, 171-204 [CrossRef][Medline] [Order article via Infotrieve]
  35. Zhang, W., Shay, J. W., and Diesseroth, A. (1993) Cancer Res. 53, 4272-4775
  36. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Ed., pp. 13.78-13.95, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY
  37. Kim, J., Zweib, C., Wu, C., and Adhya, S. (1989) Gene (Amst.) 85, 15-23 [CrossRef][Medline] [Order article via Infotrieve]
  38. Ferrari, S., Harley, V. R., Pontiggia, A., Goodfellow, P. N., Lovell-Badge, R., and Bianchi, M. E. (1992) EMBO J. 11, 4497-4506 [Abstract]
  39. Thompson, J. F., and Landy, A. (1988) Nucleic Acids Res. 16, 9687-9705 [Abstract]
  40. Carey, J. (1991) Methods Enzymol. 208, 103-117 [Medline] [Order article via Infotrieve]
  41. Hagerman, P. J. (1981) Biopolymers 20, 1503-1535 [Medline] [Order article via Infotrieve]
  42. Cairney, K. L., and Harrington, R. E. (1982) Biopolymers 21, 923-934 [Medline] [Order article via Infotrieve]
  43. Senear, D. F., and Brenowitz, M. (1991) J. Biol. Chem. 266, 13661-13671 [Abstract/Free Full Text]
  44. Wu, H.-M., and Crothers, D. M. (1984) Nature 308, 509-513 [Medline] [Order article via Infotrieve]
  45. Lumpkin, O. J., and Zimm, B. H. (1982) Biopolymers 21, 2315-2316 [Medline] [Order article via Infotrieve]
  46. Levene, S. D., and Zimm, B. H. (1989) Science 245, 396-399 [Medline] [Order article via Infotrieve]
  47. Nardulli, A. M., and Shapiro, D. J. (1992) Mol. Cell. Biol. 12, 2037-2042 [Abstract]
  48. Carey, J. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 975-979 [Abstract]
  49. Brenowitz, M., Jamison, E., Majumdar, A., and Adhya, S. (1990) Biochemistry 29, 3374-3383 [Medline] [Order article via Infotrieve]
  50. Brenowitz, M., Senear, D. F., Shea, M. A., and Ackers, G. K. (1986) Methods Enzymol. 130, 132-181 [Medline] [Order article via Infotrieve]
  51. LeBowitz, J. H., Clerc, R. G., Brenowitz, M., and Sharp, P. A. (1989) Genes & Dev. 3, 1625-1638 [Abstract]
  52. Fried, M., and Crothers, D. M. (1981) Nucleic Acids Res. 9, 6505-6525 [Abstract]
  53. Farmer, G., Bargonetti, J., Zhu, H., Friedman, P., Prywes, R., and Prives, C. (1992) Nature 358, 83-86 [CrossRef][Medline] [Order article via Infotrieve]
  54. Kristjuhan, A., and Maimets, T. (1995) Eur. J. Biochem. 234, 827-831 [Abstract]
  55. McNamara, P. T., Bolshoy, A., Trifonov, E. N., and Harrington, R. E. (1990) J. Biomol. Struct. & Dyn. 8, 529-538 [Medline] [Order article via Infotrieve]
  56. Nagaich, A. K., Bhattacharyya, D., Brahmachari, S. K., and Bansal, M. (1994) J. Biol. Chem. 269, 7824-7833 [Abstract/Free Full Text]
  57. Gorin, A. A., Zhurkin, V. B., and Olson, W. K. (1995) J. Mol. Biol. 247, 34-48 [CrossRef][Medline] [Order article via Infotrieve]
  58. Steitz, T. A. (1990) Q. Rev. Biophys. 23, 205-280 [Medline] [Order article via Infotrieve]
  59. Schultz, S. C., Shields, G. C., and Steitz, T. A. (1991) Science 253, 1001-1007 [Medline] [Order article via Infotrieve]
  60. Gartenberg, M. R., and Crothers, D. M. (1988) Nature 333, 824-829 [CrossRef][Medline] [Order article via Infotrieve]
  61. Zinkel, S., and Crothers, D. M. (1991) J. Mol. Biol. 219, 201-216 [Medline] [Order article via Infotrieve]
  62. Zhurkin, V. B., Ulyanov, N. B., Gorin, A. A., and Jernigan, R. L. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 7046-7050 [Abstract]
  63. Harrington, R. E., and Winicov, I. (1994) in Progress in Nucleic Acids Research and Molecular Biology (Cohn, W., and Moldave, K., eds), Vol. 47, pp. 195-270, Academic Press, Inc., San Diego [Medline] [Order article via Infotrieve]
  64. Barber, A. M., and Zhurkin, V. B. (1990) J. Biomol. Struct. & Dyn. 8, 213-232 [Medline] [Order article via Infotrieve]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.