Article
pubs.acs.org/JPCB
DNA Structural Correlation in Short and Long Ranges
Chan Gu,†,‡ Jun Zhang,†,‡ Y. Isaac Yang,‡ Xi Chen,†,‡ Hao Ge,† Yujie Sun,† Xiaodong Su,† Lijiang Yang,†,‡
Sunney Xie,*,†,§ and Yi Qin Gao*,†,‡
†
Biodynamic Optical Imaging Center (BIOPIC), School of Life Sciences and ‡Institute of Theoretical and Computational Chemistry,
College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China
§
Department of Chemistry and Chemical Biology, Harvard University, Cambridge, Massachusetts 02138, United States
S Supporting Information
*
ABSTRACT: Recent single-molecule measurements have revealed the DNA allostery in
protein/DNA binding. MD simulations showed that this allosteric effect is associated with the
deformation properties of DNA. In this study, we used MD simulations to further investigate
the mechanism of DNA structural correlation, its dependence on DNA sequence, and the
chemical modification of the bases. Besides a random sequence, poly d(AT) and poly d(GC)
are also used as simpler model systems, which show the different bending and twisting
flexibilities. The base-stacking interactions and the methyl group on the 5-carbon site of
thymine causes local structures and flexibility to be very different for the two model systems,
which further lead to obviously different tendencies of the conformational deformations,
including the long-range allosteric effects.
■
INTRODUCTION
The double-helix structure of DNA is maintained by a number
of inter- and intramolecular interactions. Among them, the base
stacking of base-pair steps and the Watson−Crick hydrogen
bonds between complementary bases play prominent roles.
The backbone of DNA is made up of phosphodiester bonds,
formed between adjacent nucleotides. In addition, there exist
major and minor grooves along the helical B-form DNA. DNA
is not a rigid cylinder but presents flexibility in both short and
long length scales, making its structure deviate from the
canonical B form in a sequence-dependent manner.1,2 For short
DNAs, one systematic way to describe the sequence-dependent
conformation is to use the six rigid body parameters (shift,
slide, rise, tilt, roll, and twist, Figure 1) recommended in the
Cambridge Convention in 1988.3 These parameters can be
used to characterize the local geometry of base-pair steps. On
the other hand, for long DNAs, the wormlike chain (WLC)
model4−6 is one of the most widely used theoretical models.
Bending, twisting, and stretching are the most common
motions of WLC, which are closely related to the six basestep parameters.
Different types of motions can be used to represent the
structural responses of DNA to the binding with other
molecules, such as drugs,8 chromophoric dyes,9 and proteins.10−15 Many of these binding events take place at the major
grooves of DNA. In general, the binding of other molecules
such as proteins induces the conformation changes of DNA.
These structural changes include the increase of the major
groove width,10 the breakage of hydrogen bonds (like base
flipping),14,16 the change of the backbone dihedral angles,15 or
even the transformation of B to Z forms.17,18 It should be
© 2015 American Chemical Society
Figure 1. Illustration of the six base-step parameters. Parameter values
were calculated using 3DNA software.7
emphasized that these structural properties of DNA show
sequence dependence, which can be strong enough that
changes among only two or three base pairs demolish the
original structural features such as the DNA curvature.19−22
Thus, disparate sequences of DNA can have physiological
responses far from resemblance merely due to the intrinsic
structural motions of DNA, despite the concomitant changes of
complex interactions between DNA and DNA-binding
proteins, like histone.23 Such responses are supported by
bioinformatics evidence, showing general sequence preferences
Received: June 29, 2015
Revised: October 5, 2015
Published: October 6, 2015
13980
DOI: 10.1021/acs.jpcb.5b06217
J. Phys. Chem. B 2015, 119, 13980−13990
Article
The Journal of Physical Chemistry B
tural changes of DNA in vivo/vitro are too demanding.
Molecular dynamics (MD) simulations can serve as a useful
tool to provide detailed descriptions of DNA structures and
motions. Over the last decades, MD simulation studies have
provided valuable information on the sequence-dependent
DNA flexibility including the structural characteristics of
different base steps.21,41−44 One of the most comprehensive
studies is the ABC (Ascona B-DNA Consortium) data set.45−48
By ABC data set, the base-pair and base-step conformations of
the 136 unique tetranucleotides sequences were analyzed
systematically. The results suggest that the sequence dependence of dinucleotides is significant and that the intrinsic
flexibility of YR steps is higher than RR (YY) and RY steps (R
represents purine and Y represent pyrimidine). Moreover, the
dinucleotide conformation can be strongly affected by the
flanking bases. The roll and tilt angles are not so sensitive to the
flanking sequences as the other base-step parameters, while the
twist angles exhibit significant variations, especially the YR
steps. In addition, the structures of AT step are relatively rigid
as all base-step parameters are weakly perturbed by the flanking
sequences, while the CG step shows apparent flexibility on rise
and twist. It remains interesting to examine how molecular
interactions and structures determine the conformational
deformations of DNA. In order to further understand the
molecular details that affect both local and more long-ranged
structural properties of DNA helices including allostery and its
sequence dependence, a series of all-atom MD simulations was
performed in this study. Since the examination of the longrange structure properties of DNA involves a large number of
base pairs and thus many possible sequences, we use mainly
ploy-d(AT) and poly d(GC) as model systems to illustrate the
sequence dependence in allostery, although other less repetitive
sequences are also studied. The paper is organized as follows:
we first explore the possible connections between DNA
allostery and the various structural changes. Next, we
investigate the DNA structures for different sequences and
show that the change of sequence can give rise to alterations of
long-range behaviors, which then arouse varied DNA allostery.
for specific functions, for instance, the prevalent existence of
AT-rich sequences in promoters which determines the correct
transcription starting site,24,25 the construction of epigenetic
modification elements by GC-rich sequences26,27(e.g., CpG
islands), and the nucleosome-disfavoring sequences consisting
of poly d(AT) tracts which facilitate transcription.28
It is well documented and acknowledged that AT-rich BDNA behaves dissimilarly with GC-rich ones in a wide range of
aspects. For instance, AT-rich sequences have an unusually
short helical repeat, a distinctively narrow and deep minor
groove, as well as a high propeller twist angle.29,30 Such features
are not found in GC-rich sequences. In addition, AT-rich DNA
shows an inclination for bending, as seen in the crystal
structures of DNA-interacting proteins like RNA polymerase
and TATA box-binding protein (TBP), which is a general
transcription factor binding specifically to a highly kinked ATrich DNA sequence (TATA box).31,32 In contrast, instead of
bending, poly d(GC) sequences are prone to twisting into
negative supercoils and can even “flip over” to transform from
B- to Z-form DNA. These characteristics are very likely to have
connections to previous physiological studies, which showed
that in most cases the local bending deformation of DNA is
essential for the affinity of DNA-binding proteins, while
twisting deformation is responsible for the formation of ZDNA and negative supercoils.17,33,34
Base modification also has a large impact on the structure
and functions of DNA, even the bases are unnatural.35,36
Cytosine-5-methylation at CpG islands catalyzed by DNA
methyltransferases is related to one of the most important
established mechanisms for transcriptional regulation, which
normally results in the reduction of gene expression.26,27,37 5Methylcytosine (5mC) can lead to distinct changes of DNA
structures like narrowed minor grooves, altered conformations
of sugar rings, and increased steric hindrance between the
major groove and the binding molecules.38,39 All of the factors
can influence the DNA deformation and alter the binding
affinities of protein.
Recently, new features of DNA allostery were observed in the
single-molecule measurements.40 Kim et al. found that the
binding affinity of one specific binding protein was measured
under the influence of another protein bound to the same DNA
at systematically varied distances. It was found that the free
energy of the ternary complex oscillates as a function of the
separation distance between the two proteins with a periodicity
of ∼10 base pairs, the helical pitch of B-form DNA, and a decay
length of ∼15 base pairs. The experiments also showed that
allostery became conspicuous when an AT-rich or mixed
sequence DNA linker was used. When the linkers were replaced
by the GC-rich counterparts, the allosteric effect became much
weaker. The in vivo measurement of the expression level of
RNA also exhibited the characteristically periodic oscillation,
indicating the possible importance of DNA allostery in gene
regulation.
Therefore, it is important and interesting to investigate the
underlying correlation between the structure and the sequence
of DNA, which can help to understand the contributions of
such correlations to corresponding biological processes. The
structural changes including bending and twisting are intrinsic
properties of DNA, which can be triggered by interactions with
other molecules. In fact, most experimental studies focused on
the latter, whereas the intrinsic motions of DNA are not fully
understood, given the fact that the experimental conditions and
instrumental resolution required for detecting delicate struc-
■
SIMULATION DETAILS
Five different DNA sequences were used in our simulations. In
each system, the initial structure was constructed to be in a
canonical B form. The DNA was then immersed into a cubic
box containing water and counterions (Na+). The layer of water
surrounding the DNAs is at least 10 Å thick in our simulations.
All MD simulations were conducted using the AMBER11 suite
of programs.49 The SPC/E model50 was used for the water.
Nucleic acid parameters were taken from the AMBER ff10
parameter set.51 For each system, the simulation procedure
included the energy minimization, heating up, and a production
equilibrium simulation using the NPT ensemble. The system
was first prepared through 500 steps of steepest descent
minimization and a following 500 steps of conjugate gradient
minimization with DNA being fixed using harmonic restraints.
Then the restraints on DNA were released, and the system was
further minimized using 1000 steps of steepest descent and
then 1500 steps of conjugate gradient minimization. For further
relaxation, the systems were heated to different temperatures
(Table 1) and equilibrated at that temperature for 300 ps
before being cooled down to 300 K. Finally, production runs
were performed and the temperature of all systems was
maintained at 300 K using the Langevin dynamics with a
friction coefficient of 5 ps−1. The pressure of the system was
13981
DOI: 10.1021/acs.jpcb.5b06217
J. Phys. Chem. B 2015, 119, 13980−13990
Article
The Journal of Physical Chemistry B
Table 1. Parameters for Individual Simulations
sequence
poly d(AT)
poly d(GC)
poly d(G5mC)
poly d(AU)
trajectory
heating-up temperature (K)
no. of waters
1
2
3
1
2
3
1
2
1
2
360
320
300
360
320
300
360
300
360
300
11 673
10 348
11 673
8417
10 385
8417
11 398
11 398
11 228
11 228
Figure 2. Definitions of the (A) diameter of helix and (B) major
groove width of the ith base pair. Red solid lines in A are the diameters
of helix defined as the distances between C3′ atoms of two nucleotides
composing one base pair. Blue solid line in B is the Z axis obtained by
CURVES+.58 Black dashed lines and red solid line are the schematic
vectors from ith nucleotide to the i + 3th to i + 9th nucleotides on the
complementary strand. The i + 6th vector (red solid line) in this
diagram has the smallest angle with the helical axis, and its module is
chosen to represent the ith major groove width.
adjusted to 1 atm by the Berendsen weak-coupling algorithm52
with a relaxation time constant of 2.0 ps. In all simulations, the
SHAKE algorithm53 was used to restrain all covalent bonds
involving hydrogen. All dynamic runs employed an integral
time step of 2 fs. Periodic boundary conditions were used, and
the particle mesh Ewald (PME)54 with a direct space cutoff of
10.0 Å was utilized to treat long-range electrostatic interactions.
A 33-bp DNA segment with the sequence of 5′GAGATGCTAACCCTGATCGCTGATTCCTTGGAC-3′
was used as the first model system. The ratios of G/C and A/T
contents are 51.5% and 48.5%, respectively. We performed
three independent simulations using different heating up times
(200, 400, and 600 ps) at 360 K. The short heating times (<1
ns) were performed here to maintain the DNA in their B form,
and equilibration was mainly executed at room temperature.
There are 9929 water and 64 counterions in these systems, and
the simulation times are all 200 ns. One additional simulation
using NVE ensemble was also performed to calculate dynamics
information.
To understand how the base steps affect DNA structures,
four 30bp homogeneous oligonucleotide, poly d(AT), poly
d(GC), poly d(G5mC), and poly d(AU), are simulated using
NPT ensemble. The heating-up times of these four systems are
all 200 ps, and the simulation times are 200 ns. Fifty eight Na+
were added to each system as counterions. The force field of 5methylcytosine was obtained from the work of SantaLucia,55
and that of uracil was established in reference to the standard
nucleotide library in AMBER after first assigning the uracil
fragment with restricted electrostatic potential (RESP) charges
resorting to R.E.D. toolkit.56,57 The other parameters for
individual simulations are listed in Table 1.
major groove width with respect to the ith base pair. The
identification of the Z axis is realized by using CURVES+.58
To investigate the correlation of a pair of structural
parameters, we calculated the time-averaged correlation
coefficients and cross-correlation coefficients by
ρX , Y =
E(XY ) − E(X )E(Y )
2
E(X ) − E(X )2 E(Y 2) − E(Y )2
(1)
where ρX,Y is the correlation coefficient of interest and E is the
expectation value operator.
For the 33-bp DNA segment, there are 24 × 24 correlation
coefficients when each pair of major groove widths is used
(Figure 3A). For a clearer representation, we averaged them by
interval m between base pair i and i + m. The averaged
correlation coefficients of the major groove width−major
groove width and those of the diameter of helix−diameter of
helix are shown in Figure 3B and 3C, respectively. It is obvious
from Figure 3A that the correlation coefficients of the major
groove width show an oscillation with a periodicity of about 10
base pairs, resembling the result of binding free energy
measured in experiments.40 On the other hand, such a longrange and oscillating correlation is not observed for the
diameter of helix (Figure 3C), which is a measure for the
hydrogen bonding between a base pair.
Furthermore, we also calculated time autocorrelation
functions for the major groove width and the diameter of the
helix. We give the results in Figure 4. The time autocorrelation
coefficient between time t and time t + k was calculated as
■
RESULTS AND DISCUSSION
Long-Range Correlations in DNA Deformation. As
mentioned earlier, protein binding is expected to affect the
major/minor groove width10 and/or hydrogen bonding.14,16
The diameter of the helix, defined as the distance between the
C3′ atoms of two nucleotides composing the ith base pair, was
selected to reflect the variation of hydrogen bonding (Figure
2A). In the experimental work,40 the major groove width
corresponding to the ith base pair was measured by the distance
between C3′ atoms of the ith nucleotide and the i + 6th
nucleotide on the complementary strand (i′ + 6th). Here, we
use a more robust measure of the major groove width (Figure
2B), which is calculated as follows: we first obtain the vectors
from the C5′ atom of the ith nucleotide to the C5′ and C3′
atoms of the i + 3th to i + 9th nucleotides on the
complementary strand, identify among the 14 vectors the one
that is parallel to the Z axis, and then calculate its module as the
n−k
R (k ) =
∑t = 1 (X t − μ)(X t + k − μ)
n−k
∑t = 1 σ 2
(2)
where μ and σ are the mean value and variance of the
observations X, respectively. As can be seen in Figure 4, the
decay of the time correlation function of the major groove
width (solid) is significantly slower than that of the helix
diameter (dot), indicating that the changes of major groove
widths are maintained for a longer time. The fast decay of the
helix diameter autocorrelation indicates that the perturbation to
a local hydrogen-bond length dissipates quickly and makes
hydrogen bonding less likely the origin of the DNA long-range
allostery.
13982
DOI: 10.1021/acs.jpcb.5b06217
J. Phys. Chem. B 2015, 119, 13980−13990
Article
The Journal of Physical Chemistry B
between the major/minor groove widths and the fundamental
modes of DNA structure changes listed in Figure 1. As
discussed earlier, according to WLC,4−6 bending, twisting, and
stretching are the three most important modes of conformational changes. Base-stacking interactions make large stretching
movements rare, and thus, they are expected to play a minor
role on changing the major groove length. Twisting can be
described by the twist angle, and bending magnitude is a
commonly used reaction coordinate to measure the DNA
bending,59 which is represented as a vectorial sum of roll (θR)
and tilt angles (θT)
θR2 + θT2
(3)
To identify the correlation between the major/minor groove
widths and the bending/twisting, averaged cross-correlation
coefficients between major groove widths and the various
structural parameters mentioned before were calculated using
eq 1. The corresponding base-step parameters were obtained
from 3DNA.7 For bending, the cross correlation between major
groove widths and bending magnitudes (Figure 5A, solid line)
shows a more apparent periodic oscillation pattern in
comparison with that between major groove widths and twist
angles (Figure 5A, dotted line), indicating a connection
between the DNA allostery in major groove changes and its
bending deformation. Such a long-distance correlation is largely
absent between the major groove widths and the twist angles.
In addition, bending can be caused by the changes of both roll
and tilt angles. To analyze whether they have different
correlations with major groove widths, we calculated the
cross-correlation coefficients between major groove widths and
the two base-step parameters, respectively (Figure 5B). The
cross correlation between major groove widths and roll angles
(Figure 5B, solid line) is more pronounced compared with that
between major groove widths and tilt angles (Figure 5B, dotted
line). Such a comparison indicates that the change of major
groove widths is related more closely to the roll than tilt angles.
Furthermore, we also calculated the correlation coefficients for
roll (Figure 6A) and twist angles (Figure 6B). In contrast to the
twist angle (Figure 6B), a stronger oscillating pattern appears
for the roll angles (Figure 6A). These results again show that
the twist deformation is less likely a reason that the roll motion
for the long-range DNA allostery observed in the experiment.
On the basis of the results given above, we conclude that the
bending of a DNA is most likely to cause a periodically
oscillating DNA structure correlation. In fact, when a B-form
DNA molecule bends along its Z axis, one does expect the
major groove widths to alternatively expand and narrow with a
periodicity of ∼10 base pairs. As illustrated (and probably
exaggerated) in Figure 7, bending causes the major grooves at
the side with a positive curvature to be wider and those at the
opposite side narrower. A local bending of a DNA structure can
be caused, for example, by protein binding, which would in turn
affect the major groove length. When such a bending
propagates along the DNA helix, periodically widening and
narrowing of the major grooves (compared to the naked DNA)
would be observed along the DNA and away from the local
binding site. The binding affinity of a second protein on the
same DNA then would be enhanced at positions with a
favorable major groove width and decreased at positions with
unfavorable ones, with the ∼10 bp periodicity described above.
Such a mechanism provides a natural explanation of the
periodic oscillation of Koff measured by experiments. It is worth
Figure 3. (A) Two-dimensional map of correlation coefficients of
major groove widths of the 33-bp DNA. High- and low-correlation
coefficients are shown in purple and orange, respectively. (B)
Averaged correlation coefficients with respect to the major groove
width, and (C) correlation function of the diameter of helix.
Figure 4. Time autocorrelation coefficients of major groove widths
(solid line) and diameters of helix (dotted line) with respect to the 33bp sequence.
Next, to find the possible relation between local and longrange structure properties, we try to identify the correlation
13983
DOI: 10.1021/acs.jpcb.5b06217
J. Phys. Chem. B 2015, 119, 13980−13990
Article
The Journal of Physical Chemistry B
Figure 5. Averaged cross-correlation coefficients between major groove widths and various structural parameters of the 33-bp sequence. (A) Crosscorrelation coefficients between the major groove and the bending magnitude (solid line) as well as that between the major groove width and the
twist angle (dotted line). (B) Cross-correlation coefficient between the major groove and the roll angle (solid line) as well as between the major
groove and the tilt angle (dotted line).
Figure 6. Averaged correlation coefficients of (A) roll angles and (B) twist angles for the 33-bp DNA.
bp periodicity of the increase and decrease of protein binding.
However, it is worth pointing out that bending and twisting can
be coupled, especially at short length scales,60 and the coupling
could affect the periodic oscillation shown in DNA allostery,
although it would not change the physical picture described
above.
For a random sequence, the DNA is expected to bend
nonuniformly along the helix. The base-step parameters are
different among RY, RR (YY), and YR steps, and the flanking
base also has a large impact on the inter base step.48 We
calculated the average of each roll (Figure 8A) and twist angle
(Figure 8B) along the helix as well as their standard deviations.
The results are in agreement with the previous research48 that
the RY steps have smaller roll and larger twist values than the
YR steps and the deviations of RY steps are smaller than that of
YR steps, indicating that YR steps are more flexible in terms of
roll and twist. As discussed earlier, the roll angles are used in
calculations of bending magnitudes. The larger deviations of
bending magnitudes and twist angles corresponding to YR
steps indicated that YR steps have larger flexibility on DNA
local bending and twisting deformations. Meanwhile, the same
kind of base step such as A16T17 and G19C20 shows distinct
structural parameters. The average roll, twist, and bending
magnitudes are all different for these two sequences, and the
G19C20 step showed a higher flexibility of twisting. On the other
hand, the values of roll and twist angles of the base steps are
also affected by their flanking bases. For instance, the TG step
Figure 7. Schematic and exaggerated diagram of bending along Z axis.
Bending causes the major grooves at the same side (blue arrows) to be
wider and those at the opposite side (black arrow) narrower.
noting that even when protein binding does not cause a longrange bending of the DNA, the major groove widths can also be
affected systematically at a relatively long distance away from
the direct binding site (Figure S1). In Figure S2, we compare
the averaged major groove widths of BamHI-DNA and
GRDBD-DNA complexes with the corresponding naked
DNAs (corresponding simulation details are listed in the
Supporting Information). Although BamHI and GRDBD are
considered as straight binders, both protein-bound DNAs have
significantly different major groove widths compared with
naked DNAs, showing the non-negligible impact of protein
binding on the DNA structure. In fact the binding of GRDBD
and BamHI also exhibited a periodic oscillation as a function of
their separation distance.40 On the other hand, twisting of the
DNA helix does not cause an alternative change of major/
minor groove lengths and is thus not expected to yield the ∼1013984
DOI: 10.1021/acs.jpcb.5b06217
J. Phys. Chem. B 2015, 119, 13980−13990
Article
The Journal of Physical Chemistry B
paper, and we leave it to future studies. On the basis of the
current simulations, we calculated the averaged correlation
coefficients of major groove widths, and the results are shown
in Figure 9. Indeed, a weaker periodic oscillation was observed
for poly d(GC) (red) than for poly d(AT), although the
difference is relatively small.
Figure 9. Averaged correlation coefficients of poly d(AT) (black line)
and poly d(GC) (red line).
Next, we examine how poly d(AT) and poly d(GC) differ in
terms of bending and twisting deformations. We also use the
bending magnitude as the parameter to quantify the degree of
bending. The calculated local curvature distribution functions
for poly d(AT) and poly d(GC) are compared in Figure 10.
Figure 8. Average (A) roll angle, (B) twist angle, and (C) bending
magnitude along the 33-bp DNA. Error bars are also shown.
in C-T21G22-A tetranucleotide fragment exhibited apparently
different roll and twist angles compared with that in the TT29G30-G fragment. As mentioned before, the local structural
properties are correlated with DNA long-range effects such as
allostery, and sequence dependency was indeed observed in
experiments.40
Structural Origin of Sequence Dependence. Thus, we
try to examine the sequence dependence of the structure
correlation. Experiments40 have shown that the allosteric effect
of DNA varies with the DNA sequence, and an AT-rich
sequence showed a stronger allosteric effect than a GC-rich
one. To quantify the structural differences between AT-rich and
GC-rich sequences and investigate the mechanism behind such
a sequence specificity shown in allostery, we performed
simulations for DNAs of poly d(AT) and poly d(GC)
sequences as two simpler model systems to avoid the
complicated impacts from flanking bases. A full study of the
complexity of sequence dependence is beyond the scope of this
Figure 10. Distribution of bending magnitudes with respect to RY
(AT and GC, solid lines) steps and YR (TA and CG, dotted lines)
steps.
Since in poly d(AT) and poly d(GC) there are only RY steps
and YR steps, meanwhile YR steps are more flexible on bending
according to the results above, we calculated the distributions of
bending magnitude relative to RY and YR steps, respectively.
Apparently, the YR steps (TA and CG step, dotted lines) show
wider distributions and their average values are larger than the
RY steps (AT and GC steps, solid lines), confirming that the
degree of bending is more significant for YR steps than for RY
steps again.
Since the roll and twist angles are correlated with bending
and twisting, respectively, the calculated roll and twist angle
13985
DOI: 10.1021/acs.jpcb.5b06217
J. Phys. Chem. B 2015, 119, 13980−13990
Article
The Journal of Physical Chemistry B
Figure 11. Distributions of (A) roll angles and (B) twist angles for poly d(AT) (black lines) and poly d(GC) (red lines).
twisting conformational changes and the nearest-neighbor or
base-step effects.63 Base-step effects stress on the inhomogeneity in the interactions between neighbor pairs, leading to
sequence-dependent stacking enthalpy, entropy, and melting
point. The published free energy parameters of base steps63
exhibit significant variation among base steps. The ranking
order of the base steps by their contribution to stability was
summarized to be 5′-GC-3′/3′-CG-5′ = 5′-CG-3′/3′-GC-5′ >
5′-GG-3′/3′-CC-5′ > 5′-CA-3′/3′-GT-5′ = 5′-GT-3′/3′-CA-5′
= 5′-GA-3′/3′-CT-5′ > 5′-CT-3′/3′-GA-5′ > 5′-AA-3′/3′-TT5′ > 5′-AT-3′/3′-TA-5′ > 5′-TA-3′/3′-AT-5′. From these data
one sees that the 5′-GC-3′/3′-CG-5′ step has a much higher
stacking energy than 5′-AT-3′/3′-TA-5′. Meanwhile, the free
energy of 5′-AT-3′/3′-TA-5′ is higher than that of 5′-TA-3′/3′AT-5′, as a result of the breaking of symmetry due to the DNA
double helix, namely, a base step running in the 5′ to 3′
direction is different from that runs from 3′ to 5′.
To further examine the effects of the base steps, the overlap
areas of bases between adjacent base pairs were calculated using
3DNA 7 to quantify the detailed structural differences
corresponding to stacking interactions. To avoid the ambiguity
caused by the side chains, only the rings of the purine and the
pyrimidine were included in the overlap area calculation. For
poly d(AT) and poly d(GC), a 5′-RY-3′ step is always followed
by a 5′-YR-3′ step. The average overlap area of each base step
in our two poly d(RY) systems (Figure 12) shows alternate
large and small values, exhibiting a large overlap between the
two base pairs in a 5′-RY-3′ step and a small one between the
two base pairs of a 5′-YR-3′ step. Such a result demonstrates
that the 5′-RY-3′ step is stacked better than the 5′-YR-3′ step.
distribution functions for RY (ApT or GpC, solid line) and YR
base steps (TpA and CpG, dotted line) are given in Figure 11.
It can be seen that for the roll angle both AT and GC steps
have a distribution centered at ∼0°, in clear contrast to the TA
and CG steps, which possess positive roll angles. These results
indicate that the local DNA bending occurs mainly through the
YR steps. Meanwhile, the roll angle distribution of the TA step
is broader, and its average value is larger than that of the CG
step, demonstrating that poly d(AT) has a higher probability to
exhibit large roll angles than poly d(GC) does, consistent with
the larger curvature and higher bending flexibility of poly
d(AT). On the other hand, the YR steps show broader twist
angle distributions than RY steps do, indicating a higher
flexibility of twisting between YR steps. In addition, poly
d(GC) is more prone to twisting since compared to AT and
TA steps both GC and CG steps have, on average, larger twist
angles with broader distributions.
As mentioned earlier, the observation that the AT-rich
sequence is more prone to bending while the GC-rich sequence
prefers twisting is consistent with experimental observations.
Crystal structures of DNA-interacting proteins like RNA
polymerase and TATA box-binding protein (TBP) indicated
that some AT-rich DNAs have an inclination for bending.10,31,32 However, not all AT-rich DNAs are prone to
bending. By hydroxyl radical cleavage experiments, the
(CGAAAATTTT)n sequence is shown to be easy to bend
whereas the (CGTTTTAAAA)n sequence dose not display the
cleavage pattern associated with a bent DNA.61 In addition,
(T5N4A5N7)n exhibit a more apparent bending cleavage pattern
than the (T5A5N11)n sequence,62 indicating that the elastic
properties of AT-rich DNA are not only decided by the
abundance of AT base pairs but also influenced by the detailed
sequence arrangement mode and flanking bases. In contrast,
instead of bending, GC-rich sequences are often found to twist
into negative supercoils and even to transform from B- to Zform DNA.17,33,34 Additional evidence has also emerged from
MD simulations that the change of G:C sites into A:T in a
given sequence dramatically enhances the bending propensity
of DNA.19 Furthermore, in this study, we only discussed the
structural properties of RY and YR steps using simple model
systems, and the RR steps were not studied in detail. From
Figure 8B and 8C it seems that the flexibilities of bending and
twisting of the RR steps are between that of the RY and YR
steps; thus, its elastic properties could also be strongly
influenced by the flanking sequence.
Next, we try to make a connection between the different
tendencies of poly d(AT) and poly d(GC) in bending and
Figure 12. Average overlap area for base steps of poly d(AT) (black
line) and poly d(GC) (red line).
13986
DOI: 10.1021/acs.jpcb.5b06217
J. Phys. Chem. B 2015, 119, 13980−13990
Article
The Journal of Physical Chemistry B
functions of twist (Figure 14A) and roll angles (Figure 14B) for
poly d(G5mC) and poly d(AU) and compared the results with
Therefore, the poly d(RY) structure becomes less uniform,
composed of alternatively well-stacked (5′-RY-3′ step) and
poorly stacked base pairs (5′-YR-3′ step), consistent with the
formation of “twin pairs”.64 The tendency of twin pairing is
significantly higher for poly d(AT) than poly d(GC).
Furthermore, the large overlap of the 5′-RY-3′ step and small
overlap of the 5′-YR-3′ step are not only in our model systems
(poly d(AT) and poly d(GC)). For different nearest bases (N’s
in tetranucleotides 5′-NRYN-3′ and 5′-NRYN-3′, PY = AT or
GC, YP = TA or CG, N = A, C, T, or G), the overlaps of RY
steps (Figure S3, black lines) are larger than that of YR steps
(Figure S3, red lines) and the overlaps of 5′-NATN-3′ (Figure
S3A, black line) are generally higher than 5′-NGCN-3′ (Figure
S3B, black line), indicating that even the flanking bases are
various; the formation of “twin pairs” persists with the change
of flanking bases.
In terms of molecular structures, a significant distinction
between thymine and cytosine is that the former but not the
latter has a methyl group on its C5. In fact, the methylation on
this site of cytosine (5mC) represents an important chemical
modification in epigenetics. We performed MD simulations for
DNA of a poly d(G5mC) sequence to test the effect of the
methyl group on the twin pair formation. It can be seen from
Figure 12 that the overlap area of poly d(G5mC) exhibits an
apparent alternative pattern (dotted blue line) similar to poly
d(AT) (solid black line), namely, by adding methyl groups on
cytosines, poly d(G5mC) shows a much stronger trend of twin
pair formation than poly d(GC). This result suggests that the
methyl group on carbon 5 of pyrimidines has a significant
impact on the formation of twin pairs. As a second test, another
chemically modified DNA, poly d(AU), was simulated. In
contrast to cytosine methylation, deoxyuridine (dU) is the
product of deoxythymidine demethylation. One would expect
poly d(AU) to be similar to poly d(GC) since both of them are
lacking the methyl group on carbon 5 of pyrimidines. Indeed, as
shown in Figure 12, twin pair formation is largely demolished
in poly d(AU), dramatically different from poly d(AT). Thus,
methylation/demethylation can affect strongly the formation of
twin pairs between RY steps.
Effects of methylation and demethylation are not limited to
the formation of twin pairs. We calculated the distribution
Figure 14. Distribution of (A) roll angles, (B) twist angles, and (C)
bending magnitude with respect to the poly d(AT) (black line), poly
d(GC) (red line), poly d(G5mC) (blue line), and poly d(AU) (green
line).
those of poly d(AT) and poly d(GC). From Figure 14A it can
be seen that the distribution of the 5mCG step is narrower than
that of the CG step, whereas the distribution of the UA step is
broader than that of the TA step. The distribution of the 5mCG
step’s roll angles more closely resembles that of the TA step
than that of the CG step, and the peak of the UA step
distribution function is shifted to the positive values compared
to the TA step, approaching the distribution pattern for the CG
step. For the bending magnitude (Figure 14C), the results are
similar for roll angles in that after methylation the 5mCG step
exhibits a broader distribution than that of the CG step,
whereas the distribution of the UA step is narrower than the
TA steps, namely, the addition of a methyl group to the C5 of
cytosines makes poly d(G5mC) resemble better poly d(AT),
whereas removal of the methyl group of thymines makes the
Figure 13. Average of overlap area with respect to each base step of
poly d(AT) (black line), poly d(GC) (red line), poly d(G5mC) (blue
line), and poly d(AU) (green line).
13987
DOI: 10.1021/acs.jpcb.5b06217
J. Phys. Chem. B 2015, 119, 13980−13990
Article
The Journal of Physical Chemistry B
range properties, especially the major groove length, which was
known to be important in protein/DNA binding. Long-range
structural correlations were found in the simulations for major
groove widths, which exhibited periodic oscillation. Such an
oscillation in major groove widths may explain the
experimentally observed DNA allostery observed in protein/
DNA binding affinity measurement. Furthermore, analyses of
the various structure parameters suggested that the DNA
bending deformation is closely related to the periodic
oscillating allosteric effect. The bending deformation is not
uniform in each DNA base step. The YR steps are more flexible
than RY and RR steps in terms of bending and twisting.
Two homogeneous oligonucleotides, poly d(AT) and poly
d(GC), were used as simple models to investigate the sequence
specificity of DNA allostery. The results, consistent with
experiments,40 showed that AT-rich sequences possess a
stronger allosteric effect than the GC-rich ones. The analyses
of roll and twist angles indicated that poly d(AT) has a high
tendency to bend while poly d(GC) is easy to twist. This
finding is also consistent with experimental observations40 and
reflects the fact that the base stacking of 5′-GC-3′ is
considerably stronger than 5′-AT-3′. By comparing the average
overlap area between base steps, poly d(AT) is found to be
more prone to forming twin pairs than poly d(GC). The weak
interaction between the 5′-TA-3′ base step renders the high
bending flexibility of poly d(AT). The methylation of cytosine
in poly d(GC) makes it behave more closely to poly d(AT) in
the sense of twin pair formation. It is important to point out
here that the sequence dependence on the allosteric effect of
DNA structure is complex. Its study requires a large number of
different sequences due to the long-range characteristics and
involvement of many base pairs. A more complete study with
an effort similar to ABC consortium and for real biologically
related sequences is needed to understand better the
interactions in DNAs and the allosteric effect. The poly
d(RY) systems were used as examples to show that changes on
DNA structures in short ranges can arouse changes on DNA
properties in long ranges. Poly d(GC) and poly d(AT) were
chosen to restrict their neighbors to the single RyrY or YryR to
avoid complications of changing flanking bases. Random
sequences were also used to validate the rules found in poly
d(RY) systems on the base-step dependence of twist/roll angles
and overlap areas (Figures 8 and S3). These results indicate
that methylation on C5 of pyrimidines has a significant
influence on the formation of twin pairs. Consistently, the
distributions of roll and twist angles of poly d(G5mC) and poly
d(AU) are consistent with the notion that the former is easier
to twist whereas the latter bends more easily. Therefore, the
base modifications can also alter the conformational transformation of DNA, which further affect DNA allostery. The
major groove correlations were found in our simulation to be
much stronger for methylated poly d(GC) than for the
nonmethylated one. Such an observation further supports our
speculation that the long-range DNA allostery is linked to its
bending flexibility, which is affected by the existence of weak
points of base stacking along the helix. Due to the high
complexity of DNA sequences, it is difficult to study thoroughly
the sequence and chemical modification dependence of DNA
structural properties such as those related to allostery.
However, the current study on simple model sequences does
show their strong influence, which calls for more systematic
studies. Furthermore, since both the AT/GC-rich sequences
and DNA methylation play a specific role in biology, the
resulted poly(AU) behave like poly d(GC). These results also
suggest poly d(G5mC) is easy to bend, and poly d(AU) prefers
twisting rather than bending.
The alteration in the fundamental modes of DNA
conformation changes caused by the addition or removal of
the methyl group in the pyrimidine is also expected to affect the
DNA long-range structural properties, e.g., allostery. In fact,
from Figure 15 one sees that the averaged correlation
Figure 15. Averaged correlation coefficient of poly d(G5mC) and poly
d(AU) comparing with poly d(AT) and poly d(GC).
coefficients of poly d(G5mC) show a much more pronounced
periodic oscillation pattern than poly d(GC) does. In contrast,
the oscillating behavior of the correlation coefficients for poly
d(AU) is significantly damped compared to poly d(AT). Figure
15 thus again indicates that the tendency of conformational
deformation of a DNA can be affected by its sequence and
probably local structures.
Finally, we note that the formation of twin pairs can be used
to interpret a number of experimental data on protein−DNA
recognition. For example, DNaseI is an endonuclease that binds
to the minor groove of DNA.11 The cleavage takes place at the
O3′-P bond, leaving a 5′-terminal phosphate. When poly
d(AT) is treated with DNaseI, almost all of the 5′-ends of
product oligonucleotides are deoxythymidines. If poly d(AT) is
replaced by poly d(GC), the most common products still have
a cytosine on the 5′-ends but the ratio is smaller than that of
poly d(AT).64 Whereas if the cytosine of the CpG is
methylated, the selectivity of products produced by DNase I
digestion is enhanced.39 These results are in agreement with the
different tendencies of twin pair formation among poly d(AT),
poly d(GC), and poly d(G5mC). In addition, cytosine
methylation can also change the protein binding affinity. For
the binding site of BamHI, 5′-CACGTG-3′, the binding affinity
decreases when the two cytosines in the middle CG step are
methylated, whereas methylation on the cytosines on the two
sides of the binding site have little impact on the binding
affinity (data not published). Both experiments and our
simulations suggest that the variation of the DNA local
structure can cause significant changes of DNA properties such
as the formation of twin pairs and overall bending.
■
CONCLUSION
In this study, MD simulations were performed to examine in
detail how the DNA sequence affects its various structural
properties. In particular, we examine the relation between
short-range structural properties such as twist and roll and long13988
DOI: 10.1021/acs.jpcb.5b06217
J. Phys. Chem. B 2015, 119, 13980−13990
Article
The Journal of Physical Chemistry B
(12) Pabo, C. O.; Sauer, R. T. Transcription Factors - Structural
Families and Principles of DNA Recognition. Annu. Rev. Biochem.
1992, 61, 1053−1095.
(13) Pabo, C. O.; Nekludova, L. Geometric Analysis and Comparison
of Protein-DNA Interfaces: Why Is There No Simple Code for
Recognition? J. Mol. Biol. 2000, 301, 597−624.
(14) Cheng, X. D.; Roberts, R. J. Adomet-Dependent Methylation,
DNA Methyltransferases and Base Flipping. Nucleic Acids Res. 2001,
29, 3784−3795.
(15) Pingoud, A.; Jeltsch, A. Structure and Function of Type Ii
Restriction Endonucleases. Nucleic Acids Res. 2001, 29, 3705−3727.
(16) Klimasauskas, S.; Kumar, S.; Roberts, R. J.; Cheng, X. D. Hhal
Methyltransferase Flips Its Target Base out of the DNA Helix. Cell
1994, 76, 357−369.
(17) Rich, A.; Zhang, S. G. Z-DNA: The Long Road to Biological
Function. Nat. Rev. Genet. 2003, 4, 566−572.
(18) Wolfl, S.; Wittig, B.; Rich, A. Identification of Transcriptionally
Induced Z-DNA Segments in the Human C-Myc Gene. Biochim.
Biophys. Acta, Gene Struct. Expression 1995, 1264, 294−302.
(19) Rohs, R.; Sklenar, H.; Shakked, Z. Structural and Energetic
Origins of Sequence-Specific DNA Bending: Monte Carlo Simulations
of Papillomavirus E2-DNA Binding Sites. Structure 2005, 13, 1499−
1509.
(20) Hancock, S. P.; Ghane, T.; Cascio, D.; Rohs, R.; Di Felice, R.;
Johnson, R. C. Control of DNA Minor Groove Width and Fis Protein
Binding by the Purine 2-Amino Group. Nucleic Acids Res. 2013, 41,
6750−6760.
(21) Fujii, S.; Kono, H.; Takenaka, S.; Go, N.; Sarai, A. SequenceDependent DNA Deformability Studied Using Molecular Dynamics
Simulations. Nucleic Acids Res. 2007, 35, 6063−6074.
(22) Dans, P. D.; Perez, A.; Faustino, I.; Lavery, R.; Orozco, M.
Exploring Polymorphisms in B-DNA Helical Conformations. Nucleic
Acids Res. 2012, 40, 10668−10678.
(23) Widlund, H. R.; Kuduvalli, P. N.; Bengtsson, M.; Cao, H.;
Tullius, T. D.; Kubista, M. Nucleosome Structural Features and
Intrinsic Properties of the Tataaacgcc Repeat Sequence. J. Biol. Chem.
1999, 274, 31847−31852.
(24) Schaller, H.; Gray, C.; Herrmann, K. Nucleotide-Sequence of an
Rna-Polymerase Binding-Site from DNA of Bacteriophage-Fd. Proc.
Natl. Acad. Sci. U. S. A. 1975, 72, 737−741.
(25) Pribnow, D. Nucleotide-Sequence of an Rna-Polymerase
Binding-Site at an Early T7 Promoter. Proc. Natl. Acad. Sci. U. S. A.
1975, 72, 784−788.
(26) Huidobro, C.; Fernandez, A. F.; Fraga, M. F. Aging Epigenetics:
Causes and Consequences. Mol. Aspects Med. 2013, 34, 765−781.
(27) Holliday, R.; Pugh, J. E. DNA Modification Mechanisms and
Gene Activity During Development. Science 1975, 187, 226−232.
(28) Lee, W.; Tillo, D.; Bray, N.; Morse, R. H.; Davis, R. W.; Hughes,
T. R.; Nislow, C. A High- Resolution Atlas of Nucleosome Occupancy
in Yeast. Nat. Genet. 2007, 39, 1235−1244.
(29) Wing, R.; Drew, H.; Takano, T.; Broka, C.; Tanaka, S.; Itakura,
K.; Dickerson, R. E. Crystal-Structure Analysis of a Complete Turn of
B-DNA. Nature 1980, 287, 755−758.
(30) Dickerson, R. E.; Drew, H. R. Structure of a B-DNA Dodecamer
0.2. Influence of Base Sequence on Helix Structure. J. Mol. Biol. 1981,
149, 761−786.
(31) Weisblum, B.; Dehaseth, P. L. Quinacrine, a Chromosome Stain
Specific for Deoxyadenylate-Deoxythymidylate-Rich Regions in DNA.
Proc. Natl. Acad. Sci. U. S. A. 1972, 69, 629−632.
(32) McAteer, K.; Aceves-Gaona, A.; Michalczyk, R.; Buchko, G. W.;
Isern, N. G.; Silks, L. A.; Miller, J. H.; Kennedy, M. A. Compensating
Bends in a 16-Base-Pair DNA Oligomer Containing a T(3)a(3)
Segment: A Nmr Study of Global DNA Curvature. Biopolymers 2004,
75, 497−511.
(33) Schroth, G. P.; Chou, P. J.; Ho, P. S. Mapping Z-DNA in the
Human Genome-Computer-Aided Mapping Reveals a Nonrandom
Distribution of Potential Z-DNA-Forming Sequences in Human
Genes. J. Biol. Chem. 1992, 267, 11846−11855.
sequence dependence of allostery through DNA may have real
biological significance. Therefore, the allostery through DNA
can be an important mechanism in affecting DNA/protein
interactions, and DNA acts as more than a simple docking site
but participates actively in gene regulation through shape
changing as a result of chemical modification.
■
ASSOCIATED CONTENT
S Supporting Information
*
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jpcb.5b06217.
Schematic diagram of the structural impact on DNA
from protein binding, average of major groove widths of
two protein−DNA complex systems, and average overlap
areas of RY or YR steps in the tetranucleotides (PDF)
■
AUTHOR INFORMATION
Corresponding Authors
*E-mail: [email protected].
*E-mail: [email protected].
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
We thank the National Program on Key Basic Research Project
(973 Programs, no. 2012CB917304 to Y.Q.G.) and the Natural
Science Foundation of China (21233002, 91027044, and
21125311to Y.Q.G., 21373016 to L.J.Y., 21373021 to H.G.)
for financial support.
■
REFERENCES
(1) Peters, J. P.; Maher, L. J. DNA Curvature and Flexibility in Vitro
and in Vivo. Q. Rev. Biophys. 2010, 43, 23−63.
(2) Prevost, C.; Takahashi, M.; Lavery, R. Deforming DNA: From
Physics to Biology. ChemPhysChem 2009, 10, 1399−1404.
(3) Dickerson, R. E. Definitions and Nomenclature of Nucleic-Acid
Structure Components. Nucleic Acids Res. 1989, 17, 1797−1803.
(4) Kratky, O.; Porod, G. Rontgenuntersuchung Geloster Fadenmolekule. Recl. Trav. Chim. Pays-Bas-J. R. Neth. Chem. Soc. 1949, 68,
1106−1122.
(5) Shimada, J.; Yamakawa, H. Ring-Closure Probabilities for
Twisted Wormlike Chains - Application to DNA. Macromolecules
1984, 17, 689−698.
(6) Noy, A.; Golestanian, R. Length Scale Dependence of DNA
Mechanical Properties. Phys. Rev. Lett. 2012, 109, 109.
(7) Lu, X. J.; Olson, W. K. 3dna: A Versatile, Integrated Software
System for the Analysis, Rebuilding and Visualization of ThreeDimensional Nucleic-Acid Structures. Nat. Protoc. 2008, 3, 1213−
1227.
(8) Sherman, S. E.; Lippard, S. J. Structural Aspects of Platinum
Anticancer Drug-Interactions with DNA. Chem. Rev. 1987, 87, 1153−
1181.
(9) Furstenberg, A.; Julliard, M. D.; Deligeorgiev, T. G.; Gadjev, N. I.;
Vasilev, A. A.; Vauthey, E. Ultrafast Excited-State Dynamics of DNA
Fluorescent Intercalators: New Insight into the Fluorescence
Enhancement Mechanism. J. Am. Chem. Soc. 2006, 128, 7661−7669.
(10) Rohs, R.; Jin, X. S.; West, S. M.; Joshi, R.; Honig, B.; Mann, R. S.
Origins of Specificity in Protein-DNA Recognition. In Annual Review
Of Biochemistry; Kornberg, R. D., Raetz, C. R. H., Rothman, J. E.,
Thorner, J. W., Eds.; Annual Reviews: Palo Alto, 2010; Vol. 79, pp
233−269.
(11) Wang, Y. T.; Yang, W. J.; Li, C. L.; Doudeva, L. G.; Yuan, H. S.
Structural Basis for Sequence-Dependent DNA Cleavage by Nonspecific Endonucleases. Nucleic Acids Res. 2006, 35, 584−594.
13989
DOI: 10.1021/acs.jpcb.5b06217
J. Phys. Chem. B 2015, 119, 13980−13990
Article
The Journal of Physical Chemistry B
(34) Pulleyblank, D. E.; Shure, M.; Tang, D.; Vinograd, J.; Vosberg,
H. P. Action of Nicking-Closing Enzyme on Supercoiled and
Nonsupercoiled Closed Circular DNA - Formation of a Boltzmann
Distribution of Topological Isomers. Proc. Natl. Acad. Sci. U. S. A.
1975, 72, 4280−4284.
(35) Jissy, A. K.; Datta, A. Design and Applications of Noncanonical
DNA Base Pairs. J. Phys. Chem. Lett. 2014, 5, 154−166.
(36) Jahiruddin, S.; Datta, A. What Sustains the Unnatural Base Pairs
(Ubps) with No Hydrogen Bonds. J. Phys. Chem. B 2015, 119, 5839−
5845.
(37) Jurkowska, R. Z.; Jurkowski, T. P.; Jeltsch, A. Structure and
Function of Mammalian DNA Methyltransferases. ChemBioChem
2011, 12, 206−222.
(38) Perez, A.; Castellazzi, C. L.; Battistini, F.; Collinet, K.; Flores,
O.; Deniz, O.; Ruiz, M. L.; Torrents, D.; Eritja, R.; Soler-Lopez, M.;
et al. Impact of Methylation on the Physical Properties of DNA.
Biophys. J. 2012, 102, 2140−2148.
(39) Lazarovici, A.; Zhou, T. Y.; Shafer, A.; Machado, A. C. D.; Riley,
T. R.; Sandstrom, R.; Sabo, P. J.; Lu, Y.; Rohs, R.;
Stamatoyannopoulos, J. A.; et al. Probing DNA Shape and Methylation
State on a Genomic Scale with Dnase I. Proc. Natl. Acad. Sci. U. S. A.
2013, 110, 6376−6381.
(40) Kim, S.; Brostromer, E.; Xing, D.; Jin, J. S.; Chong, S. S.; Ge, H.;
Wang, S. Y.; Gu, C.; Yang, L. J.; Gao, Y. Q.; et al. Probing Allostery
through DNA. Science 2013, 339, 816−819.
(41) Orozco, M.; Noy, A.; Perez, A. Recent Advances in the Study of
Nucleic Acid Flexibility by Molecular Dynamics. Curr. Opin. Struct.
Biol. 2008, 18, 185−193.
(42) Zacharias, M. Comparison of Molecular Dynamics and
Harmonic Mode Calculations on Rna. Biopolymers 2000, 54, 547−560.
(43) Cheatham, T. E.; Young, M. A. Molecular Dynamics Simulation
of Nucleic Acids: Successes, Limitations, and Promise. Biopolymers
2000, 56, 232−256.
(44) Perez, A.; Luque, F. J.; Orozco, M. Frontiers in Molecular
Dynamics Simulations of DNA. Acc. Chem. Res. 2012, 45, 196−205.
(45) Beveridge, D. L.; Barreiro, G.; Byun, K. S.; Case, D. A.;
Cheatham, T. E.; Dixit, S. B.; Giudice, E.; Lankas, F.; Lavery, R.;
Maddocks, J. H.; et al. Molecular Dynamics Simulations of the 136
Unique Tetranucleotide Sequences of DNA Oligonucleotides. I.
Research Design and Results on D(C(P)G) Steps. Biophys. J. 2004,
87, 3799−3813.
(46) Dixit, S. B.; Beveridge, D. L.; Case, D. A.; Cheatham, T. E.;
Giudice, E.; Lankas, F.; Lavery, R.; Maddocks, J. H.; Osman, R.;
Sklenar, H.; et al. Molecular Dynamics Simulations of the 136 Unique
Tetranucleotide Sequences of DNA Oligonucleotides. Ii: Sequence
Context Effects on the Dynamical Structures of the 10 Unique
Dinucleotide Steps. Biophys. J. 2005, 89, 3721−3740.
(47) Lavery, R.; Zakrzewska, K.; Beveridge, D.; Bishop, T. C.; Case,
D. A.; Cheatham, T.; Dixit, S.; Jayaram, B.; Lankas, F.; Laughton, C.;
et al. A Systematic Molecular Dynamics Study of Nearest-Neighbor
Effects on Base Pair and Base Pair Step Conformations and
Fluctuations in B-DNA. Nucleic Acids Res. 2010, 38, 299−313.
(48) Pasi, M.; Maddocks, J. H.; Beveridge, D.; Bishop, T. C.; Case, D.
A.; Cheatham, T. C.; Dans, P. D.; Jayaram, B.; Lankas, F.; Laughton,
C.; et al. Mu Abc: A Systematic Microsecond Molecular Dynamics
Study of Tetranucleotide Sequence Effects in B-DNA. Nucleic Acids
Res. 2014, 42, 12272−12283.
(49) Case, D. A.; Darden, T. A.; Cheatham, T. E.; Simmerling, C. L.,
III; Wang, J.; Duke, R. E.; Luo, R.; Walker, R. C.; Zhang, W.; Merz, K.
M., et al. Amber 11; University of California: San Francisco, 2010.
(50) Berendsen, H. J. C.; Grigera, J. R.; Straatsma, T. P. The Missing
Term in Effective Pair Potentials. J. Phys. Chem. 1987, 91, 6269−6271.
(51) Perez, A.; Marchan, I.; Svozil, D.; Sponer, J.; Cheatham, T. E.;
Laughton, C. A.; Orozco, M. Refinenement of the Amber Force Field
for Nucleic Acids: Improving the Description of Alpha/Gamma
Conformers. Biophys. J. 2007, 92, 3817−3829.
(52) Berendsen, H. J. C.; Postma, J. P. M.; Vangunsteren, W. F.;
Dinola, A.; Haak, J. R. Molecular-Dynamics with Coupling to an
External Bath. J. Chem. Phys. 1984, 81, 3684−3690.
(53) Ryckaert, J. P.; Ciccotti, G.; Berendsen, H. J. C. NumericalIntegration of Cartesian Equations of Motion of a System with
Constraints - Molecular-Dynamics of N-Alkanes. J. Comput. Phys.
1977, 23, 327−341.
(54) Darden, T.; York, D.; Pedersen, L. Particle Mesh Ewald - an
N.Log(N) Method for Ewald Sums in Large Systems. J. Chem. Phys.
1993, 98, 10089−10092.
(55) Aduri, R.; Psciuk, B. T.; Saro, P.; Taniga, H.; Schlegel, H. B.;
SantaLucia, J. Amber Force Field Parameters for the Naturally
Occurring Modified Nucleosides in Rna. J. Chem. Theory Comput.
2007, 3, 1464−1475.
(56) Dupradeau, F. Y.; Pigache, A.; Zaffran, T.; Savineau, C.; Lelong,
R.; Grivel, N.; Lelong, D.; Rosanski, W.; Cieplak, P. The R.Ed. Tools:
Advances in Resp and Esp Charge Derivation and Force Field Library
Building. Phys. Chem. Chem. Phys. 2010, 12, 7821−7839.
(57) Pigache, A.; Cieplak, P.; Dupradeau, F. Y. Automatic and Highly
Reproducible Resp and Esp Charge Derivation: Application to the
Development of Programs Red and X Red. Abstr. Pap. Am. Chem. Soc.
2004, 227, U1011−U1011.
(58) Blanchet, C.; Pasi, M.; Zakrzewska, K.; Lavery, R. Curves Plus
Web Server for Analyzing and Visualizing the Helical, Backbone and
Groove Parameters of Nucleic Acid Structures. Nucleic Acids Res. 2011,
39, W68−W73.
(59) Young, M. A.; Ravishanker, G.; Beveridge, D. L. A 5-ns
Molecular Dynamics Trajectory for B-DNA: Analysis of Structure,
Motions, and Solvation. Biophys. J. 1997, 73, 2313−2336.
(60) Shen, H. J.; Hricko, P. J.; Colina, C. M.; Milner, S. T. Chiral
Elasticity of DNA. Soft Matter 2012, 8, 10090−10094.
(61) Burkhoff, A. M.; Tullius, T. D. Structural Details of an Adenine
Tract That Does Not Cause DNA to Bend. Nature 1988, 331, 455−
457.
(62) Price, M. A.; Tullius, T. D. How the Structure of an Adenine
Tract Depends on Sequence Context - a New Model for the Structure
of T(N)a(N) DNA-Sequences. Biochemistry 1993, 32, 127−136.
(63) SantaLucia, J. A Unified View of Polymer, Dumbbell, and
Oligonucleotide DNA Nearest-Neighbor Thermodynamics. Proc. Natl.
Acad. Sci. U. S. A. 1998, 95, 1460−1465.
(64) Klug, A.; Jack, A.; Viswamitra, M. A.; Kennard, O.; Shakked, Z.;
Steitz, T. A. Hypothesis on a Specific Sequence-Dependent
Conformation of DNA and Its Relation to the Binding of the LacRepressor Protein. J. Mol. Biol. 1979, 131, 669−680.
13990
DOI: 10.1021/acs.jpcb.5b06217
J. Phys. Chem. B 2015, 119, 13980−13990