Research Article
Received: 19 May 2007
Revised: 18 September 2007
Accepted: 10 October 2007
Published online in Wiley Interscience:
(www.interscience.com) DOI 10.1002/mrc.2136
A solid-state 23Na NMR study of monovalent
cation binding to double-stranded DNA at low
relative humidity
Alan Wong,a† Zhimin Yan,b Yining Huangb and Gang Wua∗
We report a solid-state 23 Na NMR study of monovalent cation (Li+ , Na+ , K+ , Rb+ , Cs+ and NH4 + ) binding to double-stranded calf
thymus DNA (CT DNA) at low relative humidity, ca 0–10%. Results from 23 Na– 31 P rotational echo double resonance (REDOR)
NMR experiments firmly establish that, at low relative humidity, monovalent cations are directly bound to the phosphate group
of CT DNA and are partially dehydrated. On the basis of solid-state 23 Na NMR titration experiments, we obtain quantitative
thermodynamic parameters concerning the cation-binding affinity for the phosphate group of CT DNA. The free energy
difference (G◦ ) between M+ and Na+ ions is as follows: Li+ (−1.0 kcal mol−1 ), K+ (7.2 kcal mol−1 ), NH4 + (1.0 kcal mol−1 ), Rb+
(4.5 kcal mol−1 ) and Cs+ (1.5 kcal mol−1 ). These results suggest that, at low relative humidity, the binding affinity of monovalent
cations for the phosphate group of CT DNA follows the order: Li+ > Na+ > NH4 + > Cs+ > Rb+ > K+ . This sequence is drastically
different from that observed for CT DNA in solution. This discrepancy is attributed to the different modes of cation binding
in dry and wet states of DNA. In the wet state of DNA, cations are fully hydrated. Our results suggest that the free energy
balance between direct cation–phosphate contact and dehydration interactions is important. The reported experimental
results on relative ion-binding affinity for the DNA backbone may be used for testing theoretical treatment of cation-phosphate
c 2008 John Wiley & Sons, Ltd.
interactions in DNA. Copyright Keywords: solid-state 23 Na NMR; DNA; cation binding; REDOR; titration
Introduction
308
Negatively charged phosphate groups on the backbone of DNA
and RNA oligomers are often balanced by alkali metal cations such
as Na+ and K+ . This charge balance is necessary to reduce the
strong repulsion between closely packed phosphate groups in
DNA and RNA molecules. Because cation–phosphate interactions
are among the most important forces that can influence the
dynamics and structure of nucleic acids, they have been studied
extensively by both experimental and theoretical methods.[1 – 5]
The traditional view is that counterion condensation occurs at
the surface of DNA, primarily close to the phosphate group, and
cations are fully hydrated most of the time. This type of cation
condensation on DNA is not sequence specific.
If two different types of cations coexist in DNA solution, they
obviously would compete for the phosphate site. Factors that
control the relative binding affinity in this competition for DNA
backbone are, however, not fully understood. In addition, different
biophysical techniques often yield quite different pictures. For
example, Record and coworkers[6] used solution 23 Na relaxation
NMR method to monitor the cation competition in a doublestranded DNA and obtained the relative binding affinity as follows:
NH4 + > Cs+ > K+ > Li+ > Na+ . This sequence was confirmed
by a recent 23 Na magnetic relaxation dispersion (MRD) study.[7]
Ross and Scruggs[8] used DNA electrophoresis experiments to
determine the relative ion-binding affinities in the order of
Li+ > Na+ > K+ > TMA+ . A more recent DNA mobility study
by Stellwagen and coworkers[9] established that the free solution
mobility of single- and double-stranded DNA oligomers exhibits
the following order: Li+ < Na+ < NH4 + < K+ < Cs+ < Rb+ .
These results were interpreted as due to perturbation of hydrogen-
Magn. Reson. Chem. 2008; 46: 308–315
bonded structure of water molecules by the presence of cations
in solution, which in turn affects the friction that DNA molecules
experience during migration. An alternative interpretation was
proposed by Savelyev and Papoian.[10] They argued that lower
DNA mobility may indicate more efficient binding of cations
to DNA, which leads to charge neutralization of DNA chains.
Using this interpretation, the DNA mobility data would suggest
that the cation binding affinity takes the order: Li+ > Na+ >
NH4 + > K+ > Cs+ > Rb+ . Strauss and coworkers[11] observed a
similar ordering sequence by measuring the Donnan salt exclusion
coefficient (thermodynamic property) and incorporating a massaction treatment of specific ion binding. Using circular dichroism
(CD) data, Hald and Jacobsen[12] concluded that the order of
cation binding affinity on calf thymus DNA (CT DNA) is as follows:
NH4 + & Cs+ > Rb+ > Li+ > K+ > Na+ . Under molecular
crowding conditions, Zinchenko and Yoshikawa[13] found that
the potential of monovalent cations to promote DNA compaction
is Na+ > K+ > Li+ > Rb+ > Cs+ . If DNA compaction results
from cation condensation around DNA, their data would suggest
∗
Correspondence to: Gang Wu, Department of Chemistry, Queen’s University, 90
Bader lane, Kingston, Ontario, K7L 3N6, Canada.
E-mail: gang.wu@chem.queensu.ca
† Present address: Department of Physics, University of Warwick, Coventry, CV4
7AL, UK.
a Department of Chemistry, Queen’s University, 90 Bader lane, Kingston, Ontario,
K7L 3N6, Canada
b Department of Chemistry, The University of Western Ontario, London, Ontario,
N6A 5B7, Canada
c 2008 John Wiley & Sons, Ltd.
Copyright Solid-state 23 Na NMR of double-stranded DNA
that Na+ has the highest binding affinity to DNA. The authors
also argued that this effect is a result of partial dehydration of
cations when inducing DNA into the compact state. However,
Nordenskiöld and coworkers[14] studied competitive binding of
K+ , Na+ and Li+ to oriented DNA fibers in equilibrium with
ethanol/water solution and found that the cation binding affinity to
DNA fibers takes the order Na+ ≈ K+ > Li+ . All these experimental
results seem to be inconsistent, and sometimes even contradictory.
In addition to the aforementioned experimental studies, there
have also been many theoretical studies on cation binding to DNA
using methods ranging from Manning’s counterion condensation
(CC) hypothesis,[15] Poison–Boltzmann (PB) equation,[16 – 18] Monte
Carlo (MC) simulations,[19] to all-atom molecular dynamics (MD)
simulations.[10,20 – 24] These theoretical studies have yielded useful
insight into cation binding in DNA. However, it appears that
even the state-of-the-art MD simulations cannot reconcile all
aforementioned experimental results.
In addition to counterion condensation, which is primarily due
to cation–phosphate interactions in a sequence nonspecific way,
cations can also interact with DNA in a sequence-specific fashion,
for example, in the minor or major groove regions.[25] However,
because the cation occupancy at these binding sites is usually
low, it is believed that cation binding in these sequence-specific
sites does not contribute to the overall charge neutralization of
the bulk DNA chain. Therefore, in studying relative cation binding
affinity to DNA, it is important to make a distinction between these
different cation binding phenomena.
Despite numerous experimental and theoretical studies, cation
binding affinity of monovalent cations for a double-stranded CT
DNA at low relative humidity (<10%) has never been studied. At
very low relative humidity, cations are expected to be partially
dehydrated and perhaps directly coordinated to the oxygen atom
of the phosphate group (DNA backbone) in an inner-sphere
coordination mode. As we have successfully developed solidstate NMR methodologies to study site-specific cation binding
in organic compounds including DNA related systems,[26 – 37] we
decided to use solid-state 23 Na NMR to gain additional information
about sequence nonspecific cation condensation to CT DNA at
low relative humidity.
Results and Discussion
Dependence of 23 Na NMR spectra on relative humidity
Magn. Reson. Chem. 2008; 46: 308–315
those reported previously by Madeddu for CT DNA.[38] The general
picture for Na+ binding in the CT DNA sample shown in Fig. 1
can be summarized as follows. At high relative humidity, all Na+
ions are fully hydrated, giving rise to a sharp 23 Na NMR signal
close to 0 ppm. As the relative humidity is reduced, some Na+ ions
are localized most likely at the phosphate backbone of the DNA
and a very small fraction of Na+ ions form NaCl microcrystallites.
At very low relative humidity (dry state), a significant amount
of Na+ ions crystallize as NaCl and about 50% of Na+ ions are
attached to the phosphate group of the DNA. For this latter class
of Na+ ions, the 23 Na NMR signal exhibits a featureless broad peak
whose total line width arises primarily from second-order 23 Na
quadrupole broadening. It should be noted that the observed
partition between the two 23 Na NMR signals shown in Fig. 1 is
related to the actual ion composition of that particular CT DNA
sample (mixed Li+ /Na+ salt). For a Na+ salt of CT DNA, all Na+ ions
are localized at the surface of the DNA at very low relative humidity
(vide infra). The fact that two separate 23 Na NMR signals can be
observed for CT DNA in the dry state permits a direct measurement
of the amount of Na+ ions that are attached to the DNA backbone.
23 Na{31 P} REDOR experiment
To further investigate the mode of Na+ binding in double-stranded
DNA, we performed 23 Na{31 P} rotational echo double resonance
(REDOR) experiments for the CT DNA sample in both dehydrated
(dry) and hydrated (wet) states. In a 23 Na{31 P} REDOR experiment,
two sets of 23 Na MAS spectra are collected. One set of spectra
are the so-called reference (or control) signals, S0 , which are
recorded without the π pulses on the 31 P channel. This set of
signals is used to account for the intrinsic T2 decay. The other
set of signals, S, are recorded with the π pulses being applied.
This set of signals is known as the dipolar dephased signal. The
c 2008 John Wiley & Sons, Ltd.
Copyright www.interscience.wiley.com/journal/mrc
309
Figure 1 shows the one-dimensional (1D) 23 Na magic angle
spinning (MAS) spectra of a CT DNA sample in the form of mixed
Li+ /Na+ salt under various relative humidity conditions. The 23 Na
MAS NMR spectrum obtained at 93% relative humidity, which
corresponds to the CT DNA being in the wet state, exhibits only
a sharp signal centered at −0.9 ppm with a full width at the half
height (FWHH) of 128 Hz. At very low relative humidity of <10%,
which is denoted as the dry state in this study, the 23 Na MAS
spectrum of this CT DNA sample contains two signals. The sharp
signal (FWHH = 135 Hz) centered at 7.6 ppm is assigned to arise
from NaCl(s), and the broad signal (851 Hz line width) centered at
−4 ppm is assigned to Na+ ions that are in close proximity to the
singly charged phosphodiester group in DNA. At an intermediate
relative humidity of 60%, the spectrum also exhibits two signals,
similar to that obtained at low relative humidity. However, the
sharp signal (FWHH = 207 Hz) is less intense than that observed
at low relative humidity and the broad signal (FWHH = 546 Hz)
is centered at −0.7 ppm. These observations are quite similar to
Figure 1. Experimental 23 Na MAS spectra of the mixed Li+ /Na+ salt of
CT DNA under different relative humidity conditions. All spectra were
obtained at 11.75 T. The sample spinning frequency was 8500 Hz.
A. Wong et al.
of the CT DNA sample remained unchanged. It should be noted,
however, that a small degree of dehydration was indeed observed
when the CT DNA sample was prepared at an intermediate relative
humidity, e.g. 60%, and subject to a long period (ca 24 h) of sample
spinning at 10 kHz.
0.4
Dehydrated
Fully hydrated
0.3
intensity fraction
difference between S0 and S, S = S0 − S, reflects the true effect of
23 Na– 31 P dipolar dephasing. Usually, the so-called REDOR fraction,
S/S0 , is reported and then analyzed to yield internuclear distance
information.
Figure 2 shows some typical 23 Na{31 P} REDOR spectra for the
Na+ salt of CT DNA in both dry and wet states. It is clearly seen
from the data that, while strong REDOR signals were detected for
the dry CT DNA sample, no REDOR effect was observed for the
wet CT DNA sample. Figure 3 summarizes the 23 Na{31 P} REDOR
results obtained for several dephasing times. Considering the fact
that many factors are unknown for the CT DNA sample such as site
distribution of Na+ around the CT DNA and their corresponding
CQ (23 Na) values, it is not possible at this time to extract precise
geometric information about the Na–P distances in the dry CT DNA
sample. However, on the basis of the observed strong 23 Na{31 P}
REDOR effect, we hypothesize that the mode of Na+ binding in
the CT DNA at low relative humidity is inner-sphere coordination,
as illustrated in Fig. 4. As a consequence, the Na+ ion must be
partially dehydrated. For CT DNA in the wet state, each Na+ ion
is fully hydrated, and the 23 Na– 31 P distance is too far to produce
any REDOR effect even at −10 ◦ C.
To ensure that the CT DNA sample integrity did not change over
the course of the 23 Na{31 P} REDOR experiment, we monitored the
sample by recording 23 Na and 31 P MAS spectra before and after
the REDOR experiment. It is clear from Fig. 5 that the condition
0.1
0.0
1.2
1.6
2.0
dephasing time (ms)
Figure 3. The dependence of the 23 Na{31 P} REDOR fraction on dephasing
time for the Na+ salt of CT DNA.
Dehydrated (dry) DNA
(A)
(C)
S0
S0
S
S
∆S
∆S
X5
X5
100
50
0
-50
-100 -150
150
100
50
ppm
-50
0
-100 -150
ppm
(B)
(D)
S0
S0
S
S
X5
∆S
150
0.8
0.4
Hydrated (wet) DNA
150
0.2
100
50
0
ppm
-50
-100 -150
∆S
150
100
50
0
-50
-100 -150
ppm
310
Figure 2. Experimental 23 Na{31 P} REDOR results for the Na+ salt of CT DNA in (A) and (B) wet and dry (C) and (D) states. The dipolar diphase time is 0.4
and 2 ms for (A), (C) and (B), (D) respectively. All spectra were obtained at 9.4 T. The sample spinning frequency was 10 kHz.
www.interscience.wiley.com/journal/mrc
c 2008 John Wiley & Sons, Ltd.
Copyright Magn. Reson. Chem. 2008; 46: 308–315
Solid-state 23 Na NMR of double-stranded DNA
Figure 4. Schematic diagram illustrating the mode of Na+ binding in dry and wet DNA.
Hydrated (wet) DNA
Dehydrated (dry) DNA
Figure 5. 23 Na and 31 P MAS spectra of the Na+ salt of CT DNA before and after the REDOR experiments. All spectra were obtained at 9.4 T.
Cation titration experiment
To study cation competition for the phosphate group of CT
DNA, we used a solid-state 23 Na NMR methodology that we had
introduced a few years ago.[32] In the present study, the cation
competition process between Na+ and M+ at the phosphate site
of CT DNA can be expressed as follows:
Na+ · DNA + M+ (aq) M+ · DNA + Na+ (aq)
(1)
where Na+ · DNA and M+ · DNA describe Na+ and M+ bound to
DNA respectively. At equilibrium, the concentration of bound Na+
can be expressed as
[Na+ · DNA] = [Na+ · DNA]0 − x
(2)
where the subscript ‘0’ indicates the initial concentration and x
is amount of Na+ being replaced by M+ . From Eqn (1), x can be
expressed as a function of equilibrium constant (K),
K([Na+ · DNA]0 + [M+ ]0 )
+
− ([Na · DNA]0 − [M+ ]0 )2 K 2 + 4K[Na+ · DNA]0 [M+ ]0
x=
2(K − 1)
Magn. Reson. Chem. 2008; 46: 308–315
◦ (Na↔M)/RT
K = e−G
(4)
where R is the gas constant and T is absolute temperature, 298 K.
The cation binding affinity for CT DNA can be determined by
comparing G◦ (Na ↔ M) values for different M+ .
Figure 6 shows the 1D 23 Na MAS spectra from the titration
experiment for five different monovalent cations, M+ = Li+ , K+ ,
Rb+ , Cs+ and NH4 + . Before any M+ was added to the CT DNA
sample, i.e. [M]/[Na] = 0, the 23 Na MAS spectrum of CT DNA
exhibits a broad signal centered at −4 ppm, from which the
following 23 Na NMR parameters were estimated: δiso ≈ 1 ppm and
χ ≈ 1.0–1.5 MHz. These parameters are comparable to those of
Na+ ions found in several Na-nucleotide systems.[31]
The general trend of data shown in Fig. 6 is that, as M+
is added to the CT DNA, the Na+ ions originally bound
to the phosphate group are gradually replaced by M+ . As
c 2008 John Wiley & Sons, Ltd.
Copyright www.interscience.wiley.com/journal/mrc
311
(3)
Since the initial concentrations [Na+ · DNA]0 and [M+ ]0 are
known in the titration experiment, we can determine K for the
competition between Na+ and M+ by measuring [Na+ · DNA]
as a function of [M+ ]0 . The value of [Na+ · DNA] can be directly
monitored by the integrated area of the signal centered at −4 ppm
as shown in Fig. 1. Using Eqns (2) and (3), K can be determined
by a theoretical line-fitting of [Na+ · DNA] as a function of [M+ ]0 .
Once K is measured, the relative free energy values, G◦ (Na ↔ M),
for the M–Na cation-exchange process at the DNA surface can be
determined from the K values:
A. Wong et al.
Figure 6. 23 Na MAS spectra of CT DNA in cation titration experiments. All 23 Na MAS spectra were obtained at 11.75 T using an identical set of parameters:
256 transients, 2-s recycle delay, 8500 Hz sample spinning.
a consequence, the signal intensity of the 23 Na NMR signal
centered at ca 7 ppm is increased. It is quite interesting
that such a cation competition for the phosphate group of
DNA exhibits remarkable sensitivity on the nature of M+ .
For example, essentially all phosphate-bound Na+ ions are
replaced by Li+ ions at a [Li]/[Na] ratio of 2.5, whereas at a
[Rb]/[Na] ratio of 150 only 20% of the phosphate-bound Na+
ions are replaced by Rb+ . This immediately suggests that the
phosphate group of the CT DNA prefers Li+ over Na+ and Na+
over Rb+ .
It is also noted that, when Cs+ is added, the free ‘NaCl’
signal exhibits a typical second-order quadrupole line shape.
This indicates that the Na+ ion environment is noncubic in
the mixed NaCl/CsCl salt. To confirm this interpretation, we
examined 23 Na MAS spectra for a mixed salt sample containing
CsCl and NaCl in a ratio of Cs : Na = 10 : 1. The mixed salts
were first dissolved in water and then dried. The 23 Na MAS
spectrum (data not shown) for such a mixed salt sample is
identical to that of the free ‘NaCl’ signal shown in Fig. 6 for
the Cs/Na DNA. Clearly, the Na+ ion environment in other
mixed salts is cubic, therefore only a sharp 23 Na NMR signal is
observed.
Cation affinity for the phosphate group of CT DNA
312
The experimental NMR titration data and theoretical fits are
shown in Fig. 7. The K and G◦ (Na ↔ M) values obtained
from the analysis are reported in Table 1. For comparison,
www.interscience.wiley.com/journal/mrc
we have also listed the cation binding affinity determined in
solution by 23 Na NMR relaxation methods in Table 1. In this case,
Record and coworkers[6] used the concept of cation competition
coefficient, D, which is analogous to the equilibrium constant,
K. Our solid-state 23 Na NMR titration experiments suggest the
following cation binding affinity sequence for the phosphate
site of CT DNA:[6] Li+ > Na+ > NH4 + > Cs+ > Rb+ > K+ .
This cation binding affinity sequence is clearly different from
that determined in solution reported by Record and coworkers:
NH4 + > Cs+ > K+ > Li+ > Na+ . Because our measurement was
performed for a dry CT DNA, our data must reflect the interplay of
two key factors: cation–phosphate coordination and dehydration
cost. That is, in order to form a direct cation–phosphate contact,
at least one water molecule must be removed from the first
hydration shell of the cation. It is the balance between these two
factors that must have determined the relative binding affinity
between two different types of cations. The best-known example
to illustrate such a competition is the ion binding to the cavity
site of the G-quadruplex. The cavity site of G-quadruplex is known
to prefer K+ over Na+ . As Hud and coworkers[39] showed, this
affinity order does not result from a stronger K+ –O(carbonyl)
interaction than the Na+ –O(carbonyl) interaction. Rather, the
preference of the cavity site for K+ results from the fact that
it takes more free energy to fully dehydrate a Na+ ion than to
dehydrate a K+ ion. Consequently, the overall binding affinity
order for the cavity site of a G-quadruplex is K+ > Na+ . Similarly,
since we have established in the present case that the Na+
(or an M+ ) ion is partially dehydrated (losing at least one
c 2008 John Wiley & Sons, Ltd.
Copyright Magn. Reson. Chem. 2008; 46: 308–315
Solid-state 23 Na NMR of double-stranded DNA
Figure 7. Experimental solid-state 23 Na NMR results (data points) and theoretical fits (solid lines) for cation titration experiments. The color codes used in
the graph are as follows: DNA-bound Na+ (blue) and free NaCl (brown).
Table 1. Thermodynamic parameters for monovalent cation binding to the phosphate group of CT DNA
Dry DNAa
Cation
+
Li
Na+
K+
NH4 +
Rb+
Cs+
a
b
Ionic
radius (Å)
0.60
0.95
1.33
1.45
1.48
1.69
DNA solutionb
◦
Keq
G (Na ↔ M)
(kcal mol−1 )
D−1
G◦ (Na ↔ M)
(kcal mol−1 )
5.5 ± 0.5
1
5 × 10−6 ± 2 × 10−6
0.18 ± 0.02
5 × 10−4 ± 1 × 10−4
0.08 ± 0.04
−1.0 ± 0.1
0
7.2 ± 0.1
1.0 ± 0.2
4.5 ± 0.1
1.5 ± 0.2
1.12
1
1.19
1.92
–
1.56
−0.07
0
−0.10
−0.39
–
−0.26
This work.
Ref. [6].
Magn. Reson. Chem. 2008; 46: 308–315
shell of the cation) is unknown, it is difficult to assess the origin
of this larger variation. If we use the full dehydration free energy
as a guide, the difference between dehydration-free energies for
Li+ and K+ ions can be as large as 40 kcal mol−1 . It is reasonable
to conclude that the variations between direct cation–phosphate
contact free energies for different cations should be at least on
the same order of magnitude. On the basis of our experimental
results that Li+ is strongly favored by the phosphate group in
dry DNA, the Li+ –O (phosphate) interaction free energy must
considerably exceed the partial dehydration free energy of Li+ .
As also seen in Fig. 8, the overall cation binding affinity is similar
to that determined for cation binding to the phosphate group of
c 2008 John Wiley & Sons, Ltd.
Copyright www.interscience.wiley.com/journal/mrc
313
water molecule of hydration) when bound to the phosphate
group of a dry CT DNA, the difference in dehydration free
energy values among the monovalent cations must play an
important role.
Furthermore, as seen from Table 1, the cation binding affinity
values measured in DNA solution are all within 0.5 kcal mol−1 .
Record and coworkers found that, in general, the cation binding
affinity correlates with the hydrated radius (Stokes’ law radius)
of the cation. In contrast, the cation binding affinity values
determined for a dry DNA exhibit larger variations, ca 8 kcal mol−1 .
As the difference in partial dehydration free energy (e.g., the free
energy for losing just one water molecule from the first hydration
A. Wong et al.
Experimental
Sample preparation
Figure 8. Cation binding affinity for the phosphate groups of doublestranded DNA (in both dry and wet states) and 5 -GMP.
CT DNA in the form of sodium salt, KCl, RbCl and CsCl were
purchased from Sigma-Aldrich (Ontario, Canada). NH4 Cl was
obtained from Fisher Scientific (Canada). A CT DNA stock solution
was prepared by dissolving 25 mg of DNA in 10 ml of distilled
water. The CT DNA solution was shaken for 15 min at a high
frequency using a Brinkmann–Retsch shaker at room temperature
to ensure complete dissolution. CT DNA samples containing
mixed cations were prepared by simply adding MCl salt to DNA
solution. M/Na-DNA samples were dried by lyophilization to keep
a constant relative humidity of approximately 10%. The CT DNA
concentration, [P], was determined spectrophotometrically using
an extinction coefficient (ε) of 6600 M−1 cm−1 at 260 nm. The
total Na+ concentration, [Na]total , was measured by the 23 Na NMR
method. In particular, we inserted a capillary reference containing
a known concentration of Na+ ions and a shift reagent (10 mM NaCl
and 6 mM DyCl3 ) into the DNA sample tube. Careful calibration
of the signal intensity measurements yielded a [Na]total /[P] ratio
of ≈ 1.0 for the CT DNA solution. Elemental analysis (Canadian
Microanalytical Service, Delta, British Columbia) of the CT DNA
sample yielded an approximate Na/P ratio of 1.2.
Solid-state 23 Na NMR
314
5 -GMP. It is also interesting to note that all monovalent cations
studied here show a stronger binding affinity to the phosphate
group in 5 -GMP than in CT DNA, presumably due to the fact
that the phosphate group in 5 -GMP is doubly negatively charged,
which is compared with the singly charged phosphodiester group
in DNA. Nonetheless, in both titration NMR studies, we have
shown that thermodynamic properties of cation binding can be
determined by solid-state 23 Na NMR experiments. It would be
interesting to see whether theoretical calculations can reproduce
these results.
Finally, what is the relevance of the present study of CT DNA
in the dry state to the situation of DNA under physiological
conditions? To answer this question, one can imagine that the
direct cation–phosphate coordination, as illustrated in Fig. 4,
may be considered as a ‘snap shot’ of cation binding to DNA
in solution. That is, because cation binding to DNA in solution
is a dynamic process, cations may spend a short time at the
phosphate group in an inner-sphere coordination mode. Indeed,
MD simulations[20] suggest that two maxima are found in the
radial distribution plot for the Na+ –P distance: one is between
3.2 and 3.5 Å, while the second (more diffuse) peak appears
at around 5 Å. The first one definitely corresponds to a direct
Na+ –phosphate contact and the second is the water-separated
Na+ –phosphate contact (i.e. the Na+ ion is fully hydrated in the
latter case). More importantly, the MD simulations also found,
somewhat surprisingly, that about 40–50% of counterions are
in inner-sphere coordination mode. Although it is impossible
to say to what degree does this aspect of subnanosecond MD
simulations reflects the physical reality of cation condensation
to DNA, it is fair to conclude that inner-sphere coordination
is likely to occur in solution. As a result, our solid-state 23 Na
NMR results should also reflect the situation when this does
happen. It is also possible that such events actually play an
important role in determining the overall cation affinity to DNA in
solution.
www.interscience.wiley.com/journal/mrc
Solid-state 23 Na MAS NMR experiments were performed at 11.75 T
on a Bruker Avance-500 spectrometer operating at 132.26 MHz for
23 Na nuclei. The radio frequency (RF) field strength was 89 kHz. The
23
Na chemical shifts were referenced to NaCl (aq) by setting the
sharp 23 Na NMR signal of a solid NaCl sample at δiso = 7.21 ppm.
The 23 Na MAS spectra were recorded with a sample spinning
frequency of 8500 ± 2 Hz. A cooling gas was used to maintain
the sample temperature at 298 K. Single-pulse excitation with a
very small flip-angle pulse was used and typically a total of 256
transients were collected with a recycle delay of 2 s. The spectral
width was set to 50 kHz. The relative humidity of an air–water
mixture is defined as the ratio of the partial pressure of water
vapor in the mixture to the saturated vapor pressure of water at a
given temperature.: R. H. = 100%×p(H2 O)/p∗ (H2 O) where p(H2 O)
is the partial pressure of water vapor in the mixture and p∗ (H2 O)
is the saturation vapor pressure of water at the temperature of
the mixture. The relative humidity can be readily calibrated using
saturated salt solutions. The relative humidity of the DNA sample
was controlled by equilibrating the solid DNA sample (packed
into a 4-mm NMR rotor) with a proper saturated salt solution in
a closed container for several days. The saturated salt solutions
were NaBr, NH4 Cl4 and Na2 SO4 for relative humidity of 60, 80 and
93% respectively. The dry DNA sample was prepared by leaving
the DNA sample on a vacuum line for 24 h. 23 Na{31 P} REDOR
experiments using the original version of the pulse sequence[40]
were performed on a Varian/Chemagnetics Infinity-plus 400-WB
spectrometer operating at a magnetic field strength of 9.4 T.
The 31 P and 23 Na resonance frequencies at this field strength
were 161.72 and 105.67 MHz respectively. All MAS spectra were
acquired using a Varian/Chemagnetics T3 4-mm triple-tuned MAS
probe. Typical RF power levels corresponded to 180◦ pulse lengths
of 7.0 and 7.8 µs for 23 Na and 31 P nuclei respectively. A total of
512 transients were accumulated for each REDOR measurement.
The sample-spinning rate was kept constant at 10 000 ± 2 Hz. The
recycle delay was 0.2 s.
c 2008 John Wiley & Sons, Ltd.
Copyright Magn. Reson. Chem. 2008; 46: 308–315
Solid-state 23 Na NMR of double-stranded DNA
Conclusions
We have used 23 Na{31 P} REDOR NMR to establish that, in the
dry state of a double-stranded DNA, Na+ ions are in direct
contact with the phosphate group of the DNA backbone and
are partially dehydrated, losing at least one water molecule of
hydration. In the wet state of DNA, in contrast, all Na+ ions are
fully hydrated so that no 23 Na{31 P} REDOR effect was observed
even at −10 ◦ C. We have used a solid-state 23 Na NMR titration
method to determine the relative cation binding affinity to the
phosphate group of a double-stranded DNA in the dry state:
Li+ > Na+ > NH4 + > Cs+ > Rb+ > K+ . This cation binding
affinity sequence is drastically different from that determined for
CT DNA in solution: NH4 + > Cs+ > K+ > Li+ > Na+ . Our results
suggest that, in the dry state, the cation binding is determined by a
combination of cation–phosphate coordination and dehydration
cost. These experimental results will provide useful test data for
theoretical calculations to understand factors controlling cation
binding phenomena in DNA.
Acknowledgements
This work was supported by the Natural Sciences and Engineering
Research Council (NSERC) of Canada.
References
[1] Saenger W. Principles of Nucleic Acid Structure. Springer: New York,
1984.
[2] Sigel H. In Metal-DNA Chemistry, vol. 402, Tullius TD (ed). ACS:
Washington, DC, 1989; 159.
[3] Anderson CF, Record MT. Annu. Rev. Biophys. Biophys. Chem. 1990;
19: 423.
[4] Anderson CF, Record MT. Annu. Rev. Phys. Chem. 1995; 46: 657.
[5] Jayaram B, Beveridge DL. Annu. Rev. Biophys. Biomol. Struct. 1996;
25: 367.
[6] Bleam ML, Anderson CF, Record TR. Proc. Natl. Acad. Sci. U.S.A. 1980;
77: 3085.
[7] Denisov VP, Halle B. Proc. Natl. Acad. Sci. U.S.A. 2000; 97: 629.
[8] (a) Ross PD, Scruggs RL. Biopolymers 1964; 2: 79; (b) Ross PD,
Scruggs RL. Biopolymers 1964; 2: 231.
[9] Stellwagen E, Dong Q, Stellwagen NC. Biopolymers 2005; 78: 62.
[10] Savelyev A, Papoian GA. J. Am. Chem. Soc. 2006; 128: 14506.
[11] Strauss UP, Helfgott C, Pink H. J. Phys. Chem. 1967; 71: 2550.
[12] (a) Hald M, Jacobsen JP. Biophys. Chem. 1991; 41: 113; (b) Hald M,
Jacobsen JP. Chem. Phys. 1992; 159: 257.
[13] Zinchenko AA, Yoshikawa K. Biophys. J. 2005; 88: 4118.
[14] Korolev N, Lyubartsev AP, Ruppreecht A, Nordenskiöld L. J. Phys.
Chem. B 1999; 103: 9008.
[15] Manning GS. Q. Rev. Biophys. 1978; 11: 179.
[16] Wilson RW, Rau DC, Bloomfield VA. Biophys. J. 1978; 30: 317.
[17] Gonzales-Tovar E, Lozada-Cassou M, Henderson D. J. Chem. Phys.
1985; 83: 361.
[18] Shkel IA, Tsodikov OV, Record MT. Proc. Natl. Acad. Sci. U.S.A. 2002;
99: 2597.
[19] Paulsen MD, Anderson CF, Record MT. Biopolymers 1988; 27: 1249.
[20] Laughton CA, Luque FJ, Orozco M. J. Phys. Chem. 1995; 99: 11591.
[21] Cheatham TE, Miller JL, Fox T, Darden TA, Kollman PA. J. Am. Chem.
Soc. 1995; 117: 4193.
[22] York DM, Yang W, Lee H, Darden T, Pedersen LG. J. Am. Chem. Soc.
1995; 117: 5001.
[23] Young MA, Jayaram B, Beveridge DL. J. Am. Chem. Soc. 1997; 119:
59.
[24] Hamelberg D, Williams LD, Wilson WD. J. Am. Chem. Soc. 2001; 123:
7745.
[25] Hud NV, Polak M. Curr. Opin. Struct. Biol. 2001; 11: 293.
[26] Wong A, Wu G. J. Phys. Chem. A 2000; 104: 11844.
[27] Rovnyak D, Baldus M, Wu G, Hud NV, Feigon J, Griffin RG. J. Am.
Chem. Soc. 2000; 122: 11423.
[28] Wu G, Wong A. Chem. Commun. 2001; 2658.
[29] Wong A, Fettinger JC, Forman SL, Davis JT, Wu G. J. Am. Chem. Soc.
2002; 124: 742.
[30] Wu G, Wong A, Gan Z, Davis JT. J. Am. Chem. Soc. 2003; 125: 7182.
[31] Wong A, Wu G. J. Phys. Chem. A 2003; 107: 579.
[32] Wong A, Wu G. J. Am. Chem. Soc. 2003; 125: 13895.
[33] Wu G, Wong A. Biochem. Biophys. Res. Commun. 2004; 323: 1139.
[34] Wong A, Whitehead RD, Gan Z, Wu G. J. Phys. Chem. A 2004; 108:
10551.
[35] Ida R, Wu G. Chem. Commun. 2005; 4294.
[36] Liu X, Kwan ICM, Wang S, Wu G. Org. Lett. 2006; 8: 3685.
[37] Ida R, Kwan ICM, Wu G. Chem. Commun. 2007; 795.
[38] Madeddu M. Solid State Nucl. Magn. Reson. 2002; 22: 83.
[39] Hud NV, Smith FW, Anet FAL, Feigon J. Biochemistry 1996; 35: 15383.
[40] Gullion T, Vega AJ. Prog. Nucl. Magn. Reson. Spectrosc. 2005; 47: 123.
315
Magn. Reson. Chem. 2008; 46: 308–315
c 2008 John Wiley & Sons, Ltd.
Copyright www.interscience.wiley.com/journal/mrc