Conformations of p53 response elements in solution
deduced using site-directed spin labeling and
Monte Carlo sampling
Xiaojun Zhang1, Ana Carolina Dantas Machado2, Yuan Ding1, Yongheng Chen2,
Yan Lu2, Yankun Duan2, Kenneth W. Tham1, Lin Chen1,2, Remo Rohs1,2,* and
Peter Z. Qin1,2,*
1Department of Chemistry, University of Southern California, Los Angeles, CA 90089, USA and2Department of
Biological Sciences, University of Southern California, Los Angeles, CA 90089, USA
Received June 8, 2013; Accepted November 4, 2013
recognizing diverse p53 response elements (REs).
Understanding the mechanisms of p53-DNA inter-
action requires structural information on p53 REs.
However, such information is limited as a 3D
structure of any RE in the unbound form is not avail-
able yet. Here, site-directed spin labeling was used
to probe the solution structures of REs involved in
p53 regulation of the p21 and Bax genes. Multiple
nanometer distances in the p21-RE and BAX-
resonance spectroscopy, were used to derive
molecular models of unbound REs from pools of
all-atom structures generated by Monte-Carlo simu-
lations, thus enabling analyses to reveal sequence-
dependent DNA shape features of unbound REs
in solution. The data revealed distinct RE con-
core domain, and support the hypothesis that
sequence-dependent properties encoded in REs
are exploited by p53 to achieve the energetically
quently enhancing binding specificity. This work
reveals mechanisms of p53-DNA recognition, and
approach for studying DNA shape in solution that
has far-reaching implications for studying protein–
The tumor suppressor protein p53 plays various essential
roles in maintaining the integrity of the human genome.
Sequence-specific binding of the p53 core DNA-binding
domain (DBD) to its response elements (REs) is a key
component of the regulation of a large number of signal-
ing pathways (1). The importance of DNA recognition by
p53 is highlighted by the fact that >80% of missense
located within the DBD (2), and many cancer hot-spot
mutants have been shown to impair recognition of target
The p53 REs are defined by two closely spaced
A,T; Y=C,T; n=spacer of length 0–20 base pairs
(bp)], and hundreds of them have been validated in
human and mouse (1). The mechanisms by which p53 spe-
cifically recognizes its REs have been a long-standing
question (1,3). p53 is known to recognize REs using
base readout in the major groove, as exemplified by the
bidentate hydrogen bonds between Arg280 and the
conserved guanines in the CWWG core (3). The import-
ance of DNA shape readout has also been noted: Arg248,
the most frequently mutated residue in human cancers,
recognizes its DNA target through readout of minor
groove geometry and electrostatic potential (4); similarly,
another cancer hot-spot residue, Arg273, plays a role in
maintaining DNA shape (5).
Intriguingly, several recent crystal structures of tetra-
meric p53DBDs bound to full REs have revealed
various deformations of the bound DNA (4,6–9), suggest-
ing a propensity of DNA conformational change upon
formation of the p53/RE complex. However, it remains
*To whom correspondence should be addressed. Tel: +1 213 821 2461; Fax: +1 213 740 0930; Email: firstname.lastname@example.org
Correspondence may also be addressed to Remo Rohs. Tel: +1 213 740 0552; Fax: +1 213 821 4257; Email: email@example.com
Published online 30 November 2013Nucleic Acids Research, 2014, Vol. 42, No. 42789–2797
? The Author(s) 2013. Published by Oxford University Press.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/
by-nc/3.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial
re-use, please contact firstname.lastname@example.org
at University of Southern California on March 18, 2014
unclear to what degree the observed DNA deformation
may be biased by crystal packing, and more importantly,
how inherent variations of the shape of REs impact the
mode of conformational change upon p53 binding.
Answering these important questions requires a compari-
son of the bound and unbound RE conformations in
solution. The latter is severely lacking—no atomic reso-
lution structure of any unbound p53 RE has yet been
The challenge of deducing DNA conformation in
solution is not limited to p53 REs. Current knowledge
on sequence-dependent DNA shape, particularly for free
DNA, is rather inadequate despite its important role in
protein–DNA recognition (10). For instance, sampling of
sequence versatility in structure databases is insufficient,
with the Dickerson dodecamer accounting for ?10% of all
DNA entries in the Nucleic Acid Database (11), and
analyses of short non-coding DNA sequences in several
eukaryotic genomes conclude that none of the abundant
sequences have had their structures determined (12). These
limited experimental data on intrinsic DNA shape is due,
in large part, to the difficulty of obtaining unbiased struc-
tural information of ‘naked’ DNA. Whereas high-reso-
lution structures of free DNA have been obtained by
X-ray crystallography and NMR spectroscopy, their
number is small compared with available data for
protein–DNA complexes (13). In addition, X-ray crystal-
lography studies are hindered by crystal-packing biases,
and NMR studies are constrained by the size of the DNA.
Here, we introduce a new experimental/computational
pipeline, in which the method of site-directed spin labeling
(SDSL) is combined with all-atom Monte Carlo (MC)
simulations to derive atomic resolution data representing
the sequence-dependent conformation of DNA duplexes
in solution. SDSL uses electron paramagnetic resonance
(EPR) spectroscopy to monitor nitroxide radicals (i.e. the
spin labels) attached at specific sites of biomolecules, and
has matured as a tool for studying the structure and
dynamics of proteins and nucleic acids (14,15). The MC
simulation technique was shown to enable efficient con-
formational sampling and was extensively validated using
massive experimental data from X-ray crystallography,
NMR spectroscopy and hydroxyl radical cleavage experi-
ments (13,16). In our new SDSL-MC scheme, a pulsed
EPR technique, double-electron-electron-resonance
(DEER) (17), is applied to measure distances between
nitroxide pairs attached to the target DNA duplex.
These distances are then used as constraints to query a
large pool of all-atom models generated by MC simula-
tions (18,19), thereby identifying those that conform to the
experimental measurements. The SDSL-MC approach is
not limited by the size of the DNA or the requirement of
crystalline samples, and provides a new method for
examining the sequence-dependent shape of DNA in
This work presents studies on two prototypic naturally
occurring p53 REs (1): the p21-RE with no spacer (0-bp)
between the two half-sites; and the BAX-RE with a 1-bp
insertion between the two half-sites (Figure 1A). Using the
SDSL-MC approach, conformations of the unbound p21-
RE andBAX-RE weredetermined insolution.
Comparing the unbound and bound DNA revealed con-
formational changes in the central region between the two
half-sites on protein binding, which allow formation of
key protein–DNA and protein–protein contacts. The
modes of conformational
between the two REs, could be linked to properties that
are encoded in the individual nucleotide sequences, sug-
gesting a possible means to achieve binding specificity
through sequence-dependent conformational changes.
MATERIALS AND METHODS
DNA spin labeling
DNA oligonucleotides were synthesized by solid-phase
chemical synthesis (Integrated
Coralville, IA, USA). Following previously reported
protocols, the R5 spin labels (1-oxyl-2,2,5,5-tetramethyl-
pyrroline, Figure 1B) were attached to specific DNA sites
using the phosphorothioate scheme, and the labeled DNA
was purified by HPLC (20). Concentrations of labeled
DNA were determined by absorbance at 260nm. In this
work, the Rp and Sp phosphorothioate diastereomers
present at each attachment site were not separated.
Previous studies using model systems have validated the
use of Rp/Spmixtures in DEER measurement and estab-
lished an appropriatemethod
measured inter-nitroxide distances (21–23).
EPR sample preparation and DEER spectroscopy
measurements of inter-spin distances
described in (9). Each DEER sample contains 50–
100mM of labeled DNA or p53DBD-DNA complex,
50mM HEPES (pH 7.5), 100mM NaCl, 5mM MgCl2
and 40% (v/v) glycerol. DEER measurements were
carried out at 80K on a Bruker ELEXSYS E580
X-band spectrometer equipped with a MD4 resonator.
Previously reported acquisition parameters and proced-
ures (24) were used. Inter-spin distance distributions
were computed from the resulting dipolar evolution data
using DEERanalysis2011 (25). From these distance distri-
bution profiles, the average distance (r0) and the width of
distance distribution (?) were calculated as reported pre-
viously (21). Repeated measurements indicated that errors
in the measured r0are less than 1A˚.
sampleswere prepared followingprocedures
Generation of DNA models using Monte Carlo
A previously published protocol was used to generate an
MC ensemble of all-atom DNA models with sequence-de-
pendent shape (18,26). Simulations started from idealized
B-DNA structures generated without any sequence-de-
pendent structural features. The MC sampling was based
on collective and internal degrees of freedom using the
AMBER force field implemented as previously described
combined with an implicit solvent description (27). MC
sampling was performed >1 million cycles with random
2790Nucleic Acids Research, 2014,Vol.42, No. 4
at University of Southern California on March 18, 2014
conformational changes of all degrees of freedom in each
cycle. All-atom structures were recorded every 100 MC
cycles, forming an MC ensemble of 10000 structures for
Computation of expected inter-R5 distances
Our previously validated NASNOX program (20,22) was
used to calculate expected inter-R5 distances (rmodel).
Briefly, with each DNA model, the program modeled R5
at the target site, then identified sterically allowed R5 con-
formers using the following search parameters [see (20) for
details on these parameters]: t1 steps: 3; t2 steps: 6; t3
steps: 6; fine search: on; t1 starting values: 180?; t2
starting values: 180?; t3 starting values: 180?; and no add-
itional conformer search criterion. Searches were carried
out separately for the Rpand Spdiastereomers (i.e. R5
attached to the O1P or O2P atom), and the results were
combined to yield the ensemble of allowable R5 confor-
mers at the given labeling site. The rmodelbetween two-
specific labeling sites was then calculated by averaging
all inter-R5 distances between the two corresponding R5
ensembles. Controls showed that varying the search par-
ameters resulted in <1A˚difference in rmodel.
For the bound REs, expected inter-R5 distances (rcrystal)
were computed based on the reported crystal structures:
PDB ID 3TS8 for the p21-RE (8); and 4HJE for the
BAX-RE (9). A modified version of the NASNOX
program was used to account for the presence of protein
Characterization of DNA duplex models
For a given model j of a DNA duplex, we defined a
scoring function Ptas:
where i designates a particular distance in the DNA
duplex, rmodel-j is the NASNOX computed expected
distance for the model j, r0 is the DEER measured
average distance and ? is the measured distance distribu-
tion width. Heavy atom root-mean-square-deviations
between DNA models (RMSDstruct) were calculated
using the program VMD (28). Calculations included,
unless otherwise stated, only the interior of the duplex
and excludedtwo base-pairs
DEER-measured distances reveal that p53DBD binding
induces RE conformational changes
To examine the conformation of REs, a pair of R5 probes
(Figure 1B) were attached at specific DNA sites using a
Figure 1. Experimental design and example DEER results. (A) Nucleotide sequences of the REs. The p21-RE has no spacer between the two half-
sites and is located 50upstream of the promoter of the CDKN1A gene involved in controlling cell cycle arrest. The BAX-RE has a 1-bp spacer
(indicated by the lower-case letter) and is present at the promoter of the Bax gene. For each RE, the numbering scheme of the phosphates is shown
next to each strand. The two half-sites are noted, with the CWWG core of each half-site shown in bold and the dotted box marking the central
region between the two half-sites. Colored dotted lines and corresponding phosphates designate, as examples, the measured distance sets shown in
(C). (B) Chemical configuration of the R5 nitroxide probe. (C) Examples of DEER data. Each data set is designated by the RE and the corres-
ponding labeling sites. DEER spectra measured in the absence (straight line) and presence (dashed line) of p53DBD are shown, with black dotted
lines included to aid comparison. See Supplementary Figures S3 and S5 for additional DEER data and analyses.
Nucleic Acids Research,2014, Vol.42, No. 4 2791
at University of Southern California on March 18, 2014
previously established phosphorothioate scheme (20,21),
and inter-R5 distances were measured by DEER. Each
measured distance was designated by the corresponding
labeling site numbers, for example, data sets [14; 34] in
p21-RE represents the distance measured with a pair of
R5 attached to the phosphorothioates of nucleotides C14
and C34(Figure 1A). For all double labeled REs, the
measured dipolar evolution traces showed an oscillating
decay pattern (see Figure 1C for examples), from which
the inter-R5 distances were determined. Control meas-
(Supplementary Figure S1), thus ensuring that the dis-
tances measured in the double-labeled DNA were not
biased by spin interactions due to undesired sample aggre-
gation. In addition, previous studies have demonstrated
that R5 did not significantly distort the DNA duplex,
and the measured distances accurately reported on the
native structure (21,22,30).
We measured multiple distances in REs in the absence
and presence of p53DBD to examine protein-induced
DNA conformational changes. In these measurements,
we chose labeling sites with minimal perturbation to p53
binding and we confirmed p53DBD-DNA complex for-
mation by gel shift assays (Supplementary Figure S2).
For the p21-RE, 2 of the 6 data sets, [14; 34]
(Figure 1C) and [13; 34] (Supplementary Figure S3D),
showed clear differences in the measured echo evolution
traces on p53DBD binding. The differences in r0between
the unbound and bound DNA were 5A˚and 3A˚, respect-
ively (Supplementary Table S1), which are well beyond the
error of r0 measurements (±1A˚). In addition, control
studies indicated that changes in R5 conformers on
p53DBD binding were minimal and were unlikely to
account for the observed r0 changes (Supplementary
Figure S4). As such, distance changes observed in data
sets [14; 34] and [13; 34] clearly revealed p21-RE conform-
ational changes on p53DBD binding. On the other hand,
the remaining four datasets, including [9; 24] shown in
Figure 1C, gave superimposable dipolar evolution traces
in the absence and presence of p53DBD, resulting in little
or no changes inr0
Supplementary Figure S3D).
For the BAX-RE, 10 sets of distances were measured in
unbound and bound DNA (Supplementary Table S1,
Supplementary Figure S5). Interestingly, the observed
distance changes were much smaller than those observed
for the p21-RE. The only noticeable distance changes were
from data sets [15; 36] (Figure 1C) and [14; 36]
(Supplementary Figure S5D), both at 2A˚, whereas the
remaining 8 data sets showed little or no changes in r0
(Supplementary Table S1, Supplementary Figure S5D).
Overall, in both the p21- and BAX-RE, p53 induced
distance changes were detected by DEER, thus unambigu-
ously revealing RE conformational changes on interacting
with p53. However, with the presence of both variable and
invariable distances, it was difficult to intuitively deduce
the mode of RE conformational changes, even though
crystal structures of both bound DNAs were available
(8,9). This motivated us to ‘solve’ the conformation of
the unbound REs in solution.
REs gaveflat traces
Conformations of unbound REs in solution
To obtain all-atom models of unbound REs, we imple-
mented a strategy that has been successfully used in
SDSL mapping of an RNA junction (24), namely, to use
the DEER measured distances to select sterically accept-
able models from a large pool of 3D structures. In our
approach, these structures were generated by unrestrained
MC simulations (18). For each structural model, the
NASNOX program (20) was used to select sterically
allowed R5 conformers at the respective labeling sites,
from which the relevant average inter-R5 distances
(rmodel) were obtained. We then computed a scoring
function Pt for each model (Equation 1, ‘Materials
measured and predicted average distances (r0
rmodel) and the width of the measured distance distribu-
tion (?). Pt represents, under the assumption of an
idealized normal distribution, the effective probability
of a given set of rmodel values that match the corres-
ponding r0values, with a perfect match resulting in a
maximum Ptscore of 1.
For the p21-RE, we used 16 sets of DEER-measured
distance (Supplementary Figure S3, Supplementary Table
S2) to compute Ptfor a pool of 10000 models obtained
from 1 million MC cycles. The top-ranked model had a Pt
score of 0.64 (Figure 2A, Supplementary Table S2), cor-
responding to on average a 97% probability of matching
each rmodelto the corresponding r0(0.9716&0.64). For this
between rmodel and r0 (RMSDdeer) was 0.81A˚, and the
largest difference between a corresponding set of r0and
rmodelwas 1.8A˚(Supplementary Table S2). Given that r0
and rmodeleach might incur errors of ±1A˚, differences
below 2A˚were deemed insignificant. The results indicated
that the top-ranked MC model satisfies all the measured
To characterize variations in EPR-derived p21-RE
models, we adopted a commonly used approach in
NMR studies and further analyzed the 20 MC models
with the highest Pt scores. The pairwise RMSDstruct
among this top-20 ensemble was (1.0±0.3) A˚
age±standard deviation, same below). This suggested
that models that conformed to the measured distances
were structurally similar (Supplementary Figure S6). In
addition, we carried out a search using only 14 of the 16
measured distances, which yielded the same top-ranked
model and a similar top-20 ensemble (Supplementary
Table S3). Overall, the data indicated that within the reso-
lution of the method, the 16 measured distances were suf-
ficient to identify a set of all-atom MC models with
We also examined models of generic B- and A-form
(Figure 2B and Supplementary Table S2), which therefore
exhibited uniform shapes without sequence-dependent
characteristics. As expected, a uniform 20-bp A-form
duplex, which drastically differed from the top-ranked
p21-RE model with an RMSDstruct of 5.0A˚, fitted
poorly to the measured distances (Pt=1.0?10?17;
RMSDdeer=8.5A˚) (Supplementary Table S2). More
2792 Nucleic Acids Research, 2014,Vol.42, No. 4
at University of Southern California on March 18, 2014
importantly, a uniform 20-bp B-form duplex (Figure 2B)
also yielded a low Pt score (3.3?10?6) and a high
RMSDdeer value of 3.9A˚
indicating that this generic B-DNA did not conform to
distances measured for the p21-RE. The generic B-DNA
had an RMSDstructof 2.9A˚when compared with the top-
ranked p21-RE model, which was larger than the variation
among the top-20 EPR-derived models (Figure 2C).
Therefore, the 16 sets of distances yielded a converged con-
formation of the unbound p21-RE with a sequence-de-
Using the SDSL-MC approach, we also obtained
all-atom models for the unbound BAX-RE with 18
sets of measured distances (Supplementary Figure S5,
Supplementary Table S4). The top-ranked BAX-RE
(Supplementary Table S4), i.e. an average 98% probabil-
ity of matching each rmodelto the corresponding r0. This
top-ranked MC model again differed from a 21-bp generic
B-form duplex (Figure 2E) and satisfied all the DEER
addition, the pair-wise RMSDstruct among the top-20
BAX-RE models was (1.0±0.3) A˚, indicating a high
degree of structural similarity (Supplementary Figure
S6). Furthermore, control studies showed that the use of
18 sets of distances was sufficient to identify a set of
(Supplementary Table S2),
all-atom models with convergent conformations for the
unbound BAX-RE (Supplementary Table S5).
Assessing p53DBD bound RE conformations in solution
To evaluate the bound RE conformation in solution, we
compared the expected average inter-R5 distances based
on the crystal structures (rcrystal) with the corresponding
DEER measured r0values (Supplemental Table S6). For
the p21-RE, the four distances spanning the central region
between the two CATG cores showed differences ranging
between ?1.4 and 1.0A˚, which were within the variability
range of 2A˚in our measurements. This indicated that
within the resolution accessible to our method, the
crystal structure of the bound p21-RE central region ac-
curately reflected its conformation in solution. The same
conclusion has also been previously drawn regarding the
central region of the bound BAX-RE (9) (see also
Supplementary Table S6).
Despite the general agreement observed at the central
region, four equivalent distances, each one measured
across one of the four CWWG cores presented in the
p21- and BAX-RE, showed that the measured r0values
(Supplementary Figure S2), although r0did not change
Figure 2. Characterization of the unbound p21-RE (A–C) and BAX-RE (D-F). (A) Top-ranked MC model of the unbound p21-RE. (B) Uniform
20-bp B-DNA model constructed with a standard set of base-pair parameters (Helix twist: 35.9?; X-displacement; ?0.66A˚; and C2’-endo sugar
pucker). (C) Data mining of MC-generated unbound p21-RE models using EPR-derived distances. X-axis: Ptcomputed based on 16 measured
distances in the unbound p21-RE duplex. Y-axis: RMSDstructcomputed against the top-ranked MC model. Data points corresponding to uniform
B-DNA (cyan) and bound DNA from PDB ID 3TS8 (red) are also included. (D) Top-ranked MC model of the unbound BAX-RE. (E) Uniform
21-bp B-DNA model. (F) Data mining of MC-generated unbound BAX-RE models using 18 EPR-measured distances. All color codes are the same
as those in (C), except that the bound DNA data point (red) was obtained using PDB ID 4HJE.
Nucleic Acids Research,2014, Vol.42, No. 4 2793
at University of Southern California on March 18, 2014
upon p53DBD binding (Supplementary Table S1). In
addition, these r0values were substantially larger than
the corresponding rcrystaldistances obtained from other
crystal structures (4–6). Taken together, the data indicated
that in solution the CWWG core conformation in the
bound REs likely differed from conformations captured
in the crystal (see ‘Discussion’). At this point, however,
extensive protein–DNA contacts involved in the CWWG
regions severely limited our ability to measure informative
distances to derive an EPR-based model, and bound con-
formations of the CWWG core regions of the REs could
not be deduced with certainty.
Distinct conformational changes at the central region of
the REs on p53DBD binding
The unbound p21-RE conformation derived from the
reported in the crystal structure with an RMSDstructof
2.6A˚, which was beyond variations among the top-20
models (Figure 2C, Supplementary Figure S7A). We
modeledthe unbound DNAs
complex by aligning half-site 1 (i.e. nucleotides A3-C10/
G31-T38; Figure 1A). This half-site could be aligned rea-
sonably well between the bound and unbound DNAs,
with an RMSD of 1.4A˚ between heavy atoms of the
aligned nucleotides (Figure 3A). However, with half-site
1 aligned, the bound and unbound DNAs deviated at half-
site 2 (i.e. C11–T18/A23–G30) with an RMSD of 8.4A˚
(Figure 3A). Apparently, if the protein tetramer were
maintained, the unbound DNA conformation would not
allow proper protein–DNA contacts (e.g. R280 to G7,
G17, G27and G37, Figure 3A) to form simultaneously at
both half-sites. This likely caused a deformation at the
central region of the bound p21-RE, resulting in a previ-
ously noted displacement between the helix axes of the two
half-sites (Supplementary Figure S8) (8).
For the BAX-RE, the RMSDstructbetween the unbound
and bound DNA was 2.1A˚ (Figure 2F). This is smaller
than that of the p21-RE, and indicated that the BAX-RE
underwent a more subtle p53-induced conformational
change (Figure 3). When we aligned the unbound and
bound BAX-RE based on half-site 1 (i.e. nucleotides
A3-A10/T33-T40), the bound DNA was superimposed
by theensembleof top-20
(Supplementary Figure S7B), and half-site 2 (i.e. nucleo-
tides A12-C19/G24-T31) of the top-ranked unbound BAX-
RE deviated from the corresponding segment of the
bound DNA with an RMSD of 3.9A˚(Figure 3B). These
observations were consistent with the small distance
changes observed at the central region of BAX-RE on
p53 binding (Supplementary Table S1), and might
suggest that the unbound BAX-RE was poised to
interact with the p53 tetramer due to its sequence-depend-
ent shape (Figure 3B). Most noticeably, for the 9-bp
(Figure 1A), the unbound BAX-RE was under-wound
by ?15?compared with a generic B-DNA. This facilitated
the transition into the p53 bound form, in which further
unwinding at the central region has been observed (9).
We established a new SDSL-MC approach to study con-
formations of two prototypic p53 REs in solution. The
unbound RE conformations, obtained using multiple
measured nanometer distances as constraints, were within
the B-DNA family while exhibiting sequence-dependent
structural properties distinct from a uniform B-DNA. In
both REs, p53-induced DNA deformations were detected
at the central region between the two half-sites. The results
indicate that sequence-dependent shapes of the unbound
RE influence the mode of DNA conformational changes
upon interacting with p53, which thereby may serve as a
mechanism to achieve p53-RE binding specificity.
Sequence-dependent conformational changes of REs on
Previous biochemical and computational studies have sug-
gested changes of RE conformations upon p53 binding
(31–33). However, the molecular details of the DNA con-
formational changes and their relationship to individual
RE sequences remained rather unclear. Earlier work sug-
gested that the bound REs undergo bending at the
CWWG region (31). However, recent structural (4–9)
and biochemical (33) studies of p53DBD bound REs
Figure 3. Conformational changes in REs upon p53 binding. In each
panel, shown on the left is the superimposition of the SDSL-MC-
derived unbound RE onto the corresponding co-crystal structure of
the complex, with the blue sticks representing the Arg280 residues;
and the CPK representation denoting the G7, G17, G27and G37nu-
cleotides in the RE. Shown on the right are schematic representations
of the bound and unbound REs. (A) p21-RE, with DNAs aligned at
nucleotides A3-C10/G31-T38(dashed box). In this work, the p53 con-
struct included only the wild-type DBD, whereas the co-crystal struc-
ture of the complex (PDB ID 3TS8) included the DBD covalently
linked to the oligomerization domain without the wild-type linker
present, but with mutations in both domains (7,8). (B) BAX-RE,
with the DNAs aligned at nucleotides A3-A10/T33-T40(dashed box).
2794 Nucleic Acids Research, 2014,Vol.42, No. 4
at University of Southern California on March 18, 2014
showed generally rather small bending. Instead, in a
number of crystal structures, deviations from canonical
B-DNA characteristics were noted at a confined location
between the two half-sites (i.e. the central region) (4,6–9).
In this work, SDSL measured distances unambiguously
demonstrated that in solution p53DBD binding induces
conformational changes at the central region of the p21-
and BAX-RE (Figure 1C, Supplementary Table S1).
Perhaps unexpectedly, the degree of p53 induced DNA
alteration was more subtle in the 1-bp-spacer BAX-RE as
compared with that in the 0-bp-spacer p21-RE (Figure 3),
whereas the p53DBD bound complexes exhibited a similar
tetrameric scaffold for both REs (9). This provides a hint
that sequence-dependent structural properties encoded in a
particular DNA target are exploited by p53 to achieve the
energetically most favorable mode of deformation. This hy-
pothesis is further supported by structural analyses of
unbound REs, which were enabled by the all-atom
models provided by the new SDSL-MC method. The
analyses revealed RE shape variations and suggested
tangible connections between structural features in the
unbound and bound DNA (Figure 4). For the BAX-RE,
larger positive Roll of the T9pA10 base pair step was
already apparent in the unbound DNA (Figure 4C). Such
intrinsic property of the TpA step facilitated widening of
the minor groove (10), which was observed in the bound
form (Figure 4A). In addition, the unbound BAX-RE was
under-wound at the central region (Figures 3B and 4), thus
facilitating further unwinding to accommodate the 9-bp
central region into the same volume occupied by 8 bp in
other REs with 0-bp spacers (9). On the other hand, in the
unbound p21-RE the relative positioning of the two
CWWG cores deviated significantly from the bound
form, necessitating a shift of the helix axis at a ‘hinge’
located at the interface between two half-sites (Figure 3).
Further analyses indicated that conformational changes at
the central region of the REs occurred to facilitate proper
protein–DNA interactions, while at the same time main-
taining the intra- and inter-dimer protein contacts (Figure
3). As the collective protein–DNA and protein–protein
contacts give rise to cooperative binding, sequence-depend-
ent conformational changes at the central region of REs
thus may modulate cooperativity in p53-RE interactions,
thereby contributing to specific RE recognition.
For both bound p21- and BAX-RE, the SDSL data
CWWG cores deviated to a certain degree compared
with the corresponding crystal structures despite the
good agreement in the central regions (Supplementary
Table S6). Although we cannot rule out the possibility
that this discrepancy might be due to the fact that the
CWWG cores and the central region were differentially
impacted by differences in experimental conditions (e.g.
frozen solution versus crystal; difference in constructs,
see Figure 3 caption), the SDSL data may also reflect an
intrinsic variability of the CWWG core as suggested by
previous studies (4–6,9,31,33). In particularly, the ApT
steps within the CWWG cores have been reported in
Figure 4. Analyses of p53 RE structures. The DNA shape parameters (A) minor groove width, (B) helix twist and (C) roll are shown for the p21-RE
(left panel) and the BAX-RE (right panel). Structural features were derived from the crystal structures of the complexes (red), the top-ranked MC
models (green) and the averages of the top-20 MC models (blue). The error bars indicate the standard deviations of structural parameters among the
top-20 models, demonstrating an efficient conformational sampling. The structural parameters indicate that the conformations observed in the crystal
structures of the bound forms were partially apparent in the intrinsic DNA shape of the unbound forms. Examples for this observation are the low
helix twist values at the C10pC11step of the p21-RE and the A10pG11step of the BAX-RE, as well as the negative Roll at the A12pA13step of the
p21-RE and the positive Roll at the T9pA10step of the BAX-RE.
Nucleic Acids Research,2014, Vol.42, No. 42795
at University of Southern California on March 18, 2014
either Watson–Crick (6) or Hoogsteen configuration (4).
geometry, which are associated with base flipping (34),
may account for the longer distances measured in
p53DBD binding induced a larger degree of deformation
at the central region of p21-RE as compared with that of
the BAX-RE, the p21-RE is known to bind tighter to p53
(35). One of the possible explanations for this apparently
puzzling observation is that the respective CWWG cores
responded differently to p53 binding. Further investiga-
tion of the CWWG core, particularly in the bound state,
is therefore required.
Finally, p53/DNA interactions can be impacted by
regions beyond the DBD and RE (3). Whereas biophysical
studies focusing on folded p53 fragments have provided a
wealth of information regarding p53 structure and
function, expanding beyond these ‘truncated’ systems is
highly desirable. The SDSL-MC approach, which is
non-crystalline complexes, is particularly suited for these
Mapping sequence-dependent DNA shape using the
Whereas early SDSL studies used DNA duplexes as model
systems (21,36,37), recent reports have emerged in which
SDSL measured distances were used to study DNA duplex
conformation in response to base lesion (38), mismatches
(39) and protein binding (40). In addition, SDSL has also
been used to study higher order DNA structures such as
quadruplexes (41) and four-way junctions (42). In this
work, using the R5 probe that can be attached to any nu-
cleotide within a target sequence, multiple distances were
readily measured, and they directly revealed conform-
ational changes between the bound and unbound DNA.
In addition, synergistic integration with MC sampling
allowed us to derive atomic models of the target DNA
with sequence-dependent shape. In each top-20 ensemble
of unbound REs, the models are: (i) structurally highly
similar; and (ii) clearly different from a uniform B-DNA
(Figures 2 and 4). This demonstrates that the SDSL-MC
pipeline has the ability to provide detailed structural infor-
mation of DNA duplexes. The bound p21-RE structure
differed from the top-ranked model of the unbound
DNA by an RMSDstructof 2.6A˚
DNA by only 1.8A˚. As such, obtaining the sequence-de-
pendent shape of the unbound DNA has a profound
impact on properly assessing protein induced deformations
of DNA targets. Furthermore, intrinsic DNA shape
features revealed by the SDSL-MC approach will benefit
a broad range of efforts, such as prediction of transcription
factor binding specificities based on regression models that
combine DNA sequence and shape (43–45).
Nevertheless, further studies are needed to explore the
utility and limitation of the SDSL-MC method. For
example, the ‘resolution’ that can be achieved by this
approach remains to be investigated. In addition, DEER
measures distances in a frozen solution state, whereas MC
simulations target solution-state equilibrium at ambient
and from uniform B-
temperature. It is not clear how unique aspects of each
methodology impact the interpretation of the resulting
In summary, results reported here clearly demonstrate
that the SDSL-MC approach reveals sequence-dependent
shapes of p53 REs that advance our understanding of p53/
DNA recognition. The method is not limited by the size of
the system and allows parallel examination of DNA shape
in both the unbound and protein-bound states. This is a
step forward toward uncovering the role of intrinsic
DNA shape on protein–DNA recognition on a general
Supplementary Data are available at NAR Online.
The authors thank I. Haworth for the continued develop-
ment of the NASNOX program. R.R. is an Alfred
P. Sloan Research Fellow. See Supplementary Data for
detailed author contributions.
U01GM103804 and R01HG003008 (in part to R.R.);
and R01GM064642 and 2U54RR022220 (to L.C.)]; and
NSF [MCB-0546529 and CHE-1213673 (to P.Z.Q.)].
Funding for open access charge: NSF, NIH.
[GM069557and RR028992(to P.Z.Q.);
Conflict of interest statement. None declared.
1. Riley,T., Sontag,E., Chen,P. and Levine,A. (2008) Transcriptional
control of human p53-regulated genes. Nat. Rev. Mol. Cell Biol.,
2. Petitjean,A., Mathe,E., Kato,S., Ishioka,C., Tavtigian,S.V.,
Hainaut,P. and Olivier,M. (2007) Impact of mutant p53
functional properties on TP53 mutation patterns and tumor
phenotype: lessons from recent developments in the IARC TP53
database. Hum. Mutat., 28, 622–629.
3. Joerger,A.C. and Fersht,A.R. (2010) The Tumor Suppressor p53:
From Structures to Drug Discovery. Cold Spring Harb. Perspect.
Biol., 2, a000919.
4. Kitayner,M., Rozenberg,H., Rohs,R., Suad,O., Rabinovich,D.,
Honig,B. and Shakked,Z. (2010) Diversity in DNA recognition by
p53 revealed by crystal structures with Hoogsteen base pairs. Nat.
Struct. Mol. Biol., 17, 423–429.
5. Eldar,A., Rozenberg,H., Diskin-Posner,Y., Rohs,R. and
Shakked,Z. (2013) Structural studies of p53 inactivation by DNA-
contact mutations and its rescue by suppressor mutations via
alternative protein-DNA interactions. Nucleic Acids Res., 41,
6. Chen,Y., Dey,R. and Chen,L. (2010) Crystal structure of the p53
core domain bound to a full consensus site as a self-assembled
tetramer. Structure, 18, 246–256.
7. Petty,T.J., Emamzadah,S., Costantino,L., Petkova,I., Stavridi,E.S.,
Saven,J.G., Vauthey,E. and Halazonetis,T.D. (2011) An induced
fit mechanism regulates p53 DNA binding kinetics to confer
sequence specificity. EMBO J., 30, 2167–2176.
8. Emamzadah,S., Tropia,L. and Halazonetis,T.D. (2011) Crystal
Structure of a Multidomain Human p53 Tetramer Bound to the
2796 Nucleic Acids Research, 2014,Vol.42, No. 4
at University of Southern California on March 18, 2014
Natural CDKN1A (p21) p53-Response Element. Mol. Cancer Download full-text
Res., 9, 1493–1499.
9. Chen,Y., Zhang,X., Dantas Machado,A.C., Ding,Y., Chen,Z.,
Qin,P.Z., Rohs,R. and Chen,L. (2013) Structure of p53 binding
to the BAX response element reveals DNA unwinding and
compression to accommodate base-pair insertion. Nucleic Acids
Res., 41, 8368–8376.
10. Rohs,R., West,S.M., Sosinsky,A., Liu,P., Mann,R.S. and
Honig,B. (2009) The role of DNA shape in protein-DNA
recognition. Nature, 461, 1248–1253.
11. Egli,M. and Pallan,P.S. (2010) The many twists and turns of
DNA: template, telomere, tool, and target. Curr. Opin. Struct.
Biol., 20, 262–275.
12. Subirana,J.A. and Messeguer,X. (2010) The most frequent short
sequences in non-coding DNA. Nucleic Acids Res., 38, 1172–1181.
13. Zhou,T., Yang,L., Lu,Y., Dror,I., Dantas Machado,A.C.,
Ghane,T., Di Felice,R. and Rohs,R. (2013) DNAshape: a method
for the high-throughput prediction of DNA structural features on
a genomic scale. Nucleic Acids Res., 41, W56–W62.
14. Fanucci,G.E. and Cafiso,D.S. (2006) Recent Advances and
applications of site-directed spin labeling. Curr. Opin. Struct.
Biol., 16, 644–653.
15. Sowa,G.Z. and Qin,P.Z. (2008) Site-directed spin labeling studies
on nucleic acid structure and dynamics. Prog. Nucleic Acids Res.
Mol. Biol., 82, 147–197.
16. Bishop,E.P., Rohs,R., Parker,S.C., West,S.M., Liu,P., Mann,R.S.,
Honig,B. and Tullius,T.D. (2011) A map of minor groove shape
and electrostatic potential from hydroxyl radical cleavage patterns
of DNA. ACS Chem. Biol., 6, 1314–1320.
17. Schiemann,O. and Prisner,T.F. (2007) Long-range distance
determinations in biomacromolecules by EPR spectroscopy. Q.
Rev. Biophys., 40, 1–53.
18. Rohs,R., Sklenar,H. and Shakked,Z. (2005) Structural and
energetic origins of sequence-specific DNA bending: Monte Carlo
simulations of papillomavirus E2-DNA binding sites. Structure,
19. Sklenar,H., Wu ¨ stner,D. and Rohs,R. (2006) Using internal and
collective variables in Monte Carlo simulations of nucleic acid
structures: chain breakage/closure algorithm and associated
Jacobians. J. Comput. Chem., 27, 309–315.
20. Qin,P.Z., Haworth,I.S., Cai,Q., Kusnetzow,A.K., Grant,G.P.G.,
Price,E.A., Sowa,G.Z., Popova,A., Herreros,B. and He,H. (2007)
Measuring nanometer distances in nucleic acids using a sequence-
independent nitroxide probe. Nat. Protoc., 2, 2354–2365.
21. Cai,Q., Kusnetzow,A.K., Hubbell,W.L., Haworth,I.S.,
Gacho,G.P.C., Van Eps,N., Hideg,K., Chambers,E.J. and
Qin,P.Z. (2006) Site-directed spin labeling measurements of
nanometer distances in nucleic acids using a sequence-independent
nitroxide probe. Nucleic Acids Res., 34, 4722–4734.
22. Price,E.A., Sutch,B.T., Cai,Q., Qin,P.Z. and Haworth,I.S. (2007)
Computation of nitroxide-nitroxide distances in spin-labeled DNA
duplexes. Biopolymers, 87, 40–50.
23. Cai,Q., Kusnetzow,A.K., Hideg,K., Price,E.A., Haworth,I.S. and
Qin,P.Z. (2007) Nanometer Distance Measurements in
RNA Using Site-Directed Spin Labeling. Biophys. J., 93,
24. Zhang,X., Tung,C.-S., Sowa,G.Z., Hatmal,M.M.M., Haworth,I.S.
and Qin,P.Z. (2012) Global structure of a three-way junction in a
phi29 packaging RNA dimer determined using site-directed spin
labeling. J. Am. Chem. Soc., 134, 2644–2652.
25. Jeschke,G., Chechik,V., Ionita,P., Godt,A., Zimmermann,H.,
Banham,J., Timmel,C., Hilger,D. and Jung,H. (2006)
DeerAnalysis2006—a comprehensive software package for
analyzing pulsed ELDOR data. Appl. Magn. Reson., 30, 473–498.
26. Rohs,R., Bloch,I., Sklenar,H. and Shakked,Z. (2005) Molecular
flexibility in ab initio drug docking to DNA: binding-site and
binding-mode transitions in all-atom Monte Carlo simulations.
Nucleic Acids Res., 33, 7048–7057.
27. Rohs,R., Etchebest,C. and Lavery,R. (1999) Unraveling proteins:
a molecular mechanics study. Biophys. J., 76, 2760–2768.
28. Humphrey,W., Dalke,A. and Schulten,K. (1996) VMD: visual
molecular dynamics. J. Mol. Graph., 14, 33–38, 27–38.
29. Lavery,R. and Sklenar,H. (1989) Defining the structure of
irregular nucleic acids: conventions and principles. J. Biomol.
Struct. Dyn., 6, 655–667.
30. Popova,A.M., Ka ´ lai,T., Hideg,K. and Qin,P.Z. (2009) Site-specific
DNA structural and dynamic features revealed by nucleotide-
independent nitroxide probes. Biochemistry, 48, 8540–8550.
31. Nagaich,A.K., Zhurkin,V.B., Durell,S.R., Jernigan,R.L.,
Appella,E. and Harrington,R.E. (1999) p53-induced DNA
bending and twisting: p53 tetramer binds on the outer side of a
DNA loop and increases DNA twisting. Proc. Natl Acad. Sci.
USA, 96, 1875–1880.
32. Pan,Y. and Nussinov,R. (2008) p53-Induced DNA bending: the
interplay between p53-DNA and p53-p53 interactions. J. Phys.
Chem. B, 112, 6716–6724.
33. Beno,I., Rosenthal,K., Levitine,M., Shaulov,L. and Haran,T.E.
(2011) Sequence-dependent cooperative binding of p53 to DNA
targets and its relationship to the structural properties of the
DNA targets. Nucleic Acids Res., 39, 1919–1932.
34. Honig,B. and Rohs,R. (2011) Biophysics: flipping Watson and
Crick. Nature, 470, 472–473.
35. Weinberg,R.L., Veprintsev,D.B., Bycroft,M. and Fersht,A.R.
(2005) Comparative binding of p53 to its promoter and DNA
recognition elements. J. Mol. Biol., 348, 589–596.
36. Ward,R., Keeble,D.J., El-Mkami,H. and Norman,D.G. (2007)
Distance determination in heterogeneous DNA model systems by
pulsed EPR. Chembiochem, 8, 1957–1964.
37. Schiemann,O., Cekan,P., Margraf,D., Prisner,T.F. and
Sigurdsson,S.T. (2009) Relative orientation of rigid nitroxides by
PELDOR: beyond distance measurements in nucleic acids. Angew.
Chem. Int. Ed. Engl., 48, 3292–3295.
38. Sicoli,G., Mathis,G., Aci-Seche,S., Saint-Pierre,C., Boulard,Y.,
Gasparutto,D. and Gambarelli,S. (2009) Lesion-induced DNA
weak structural changes detected by pulsed EPR spectroscopy
combined with site-directed spin labelling. Nucleic Acids Res., 37,
39. Wunnicke,D., Ding,P., Seela,F. and Steinhoff,H.-J. (2012) Site-
directed spin labeling of DNA reveals mismatch-induced
nanometer distance changes between flanking nucleotides. J. Phys.
Chem. B, 116, 4118–4123.
40. Reginsson,G.W., Shelke,S.A., Rouillon,C., White,M.F.,
Sigurdsson,S.T. and Schiemann,O. (2013) Protein-induced changes
in DNA structure and dynamics observed with noncovalent site-
directed spin labeling and PELDOR. Nucleic Acids Res., 41, e11.
41. Singh,V., Azarkh,M., Exner,T.E., Hartig,J.S. and Drescher,M.
(2009) Human telomeric quadruplex conformations studied by
pulsed EPR. Angew. Chem. Int. Ed. Engl., 48, 9728–9730.
42. Freeman,A.D.J., Ward,R., El Mkami,H., Lilley,D.M.J. and
Norman,D.G. (2011) Analysis of Conformational Changes in the
DNA Junction-Resolving Enzyme T7 Endonuclease I on Binding
a Four-Way Junction Using EPR. Biochemistry, 50, 9963–9972.
43. Gorda ˆ n,R., Shen,N., Dror,I., Zhou,T., Horton,J., Rohs,R. and
Bulyk,M.L. (2013) Genomic regions flanking E-box binding sites
influence DNA binding specificity of bHLH transcription factors
through DNA shape. Cell Rep, 3, 1093–1104.
44. Dror,I., Zhou,T., Mandel-Gutfreund,Y. and Rohs,R. (2014)
Covariation between homeodomain transcription factors and the
shape of their DNA binding sites. Nucleic Acids Res., 42,
45. Yang,L., Zhou,T., Dror,I., Mathelier,A., Wasserman,W.W.,
Gorda ˆ n,R. and Rohs,R. (2014) TFBSshape: a motif database for
DNA shape features of transcription factor binding sites. Nucleic
Acids Res., 42, D148–D155.
Nucleic Acids Research,2014, Vol.42, No. 42797
at University of Southern California on March 18, 2014