Sugar-oligoamides: Bound-state conformation and DNA minor-groove-binding description by TR-NOESY and differential-frequency saturation-transfer- difference experiments

Article abstract of DOI:10.1002/chem.200701103

Selective-frequency saturation-transfer-difference (STD) spectra allow the description of complexes established between minor-groove binders and long tracts of calf thymus DNA (ct-DNA). Two sets of experiments with selective saturation of either the H1′ or H4′/H5′/H5″ proton NMR regions of deoxyribose allow the description of the ligand residues close to the inner (H1′) and outer regions (H4′/H5′/H5″) of the minor groove of double-helical DNA. A series of complexes of sugar-oligoamides (2-6) with ct-DNA have been studied by both TR-NOESY and STD experiments. The binding mode of the complexes is similar to that of netropsin (1) and allows us to define a general binding mode for this family of ligands, in which an NH rim points towards the internal area (inner region) and a CH₃ rim points towards the external part (outer region) of the minor groove of DNA. Also by means of both TR-NOESY and STD experiments, a description of the asymmetric centers of the sugar residue close to the inner and outer regions of the groove has been achieved. These results confirm that the sugar is responsible for the differences previously found in binding energetics.

Full text of DOI:10.1002/chem.200701103

FULL PAPER
DOI: 10.1002/chem.200701103
Sugar–Oligoamides: Bound-State Conformation and DNA Minor-Groove-
Binding Description by TR-NOESY and Differential-Frequency Saturation-
Transfer-Difference Experiments
[
a]
[a]
[a]
[a]
Florence Souard, Eva MuÇoz, Pablo PeÇalver, Concepción Badía,
[b]
[
a]
[a]
Rafael del Villar-Guerra, Juan Luis Asensio, Jesffls JimØnez-Barbero, and
[
a]
Cristina Vicent*
Abstract: Selective-frequency satura-
tion-transfer-difference (STD) spectra
allow the description of complexes es-
tablished between minor-groove bind-
ers and long tracts of calf thymus DNA
sugar–oligoamides (2–6) with ct-DNA
have been studied by both TR-NOESY
and STD experiments. The binding
mode of the complexes is similar to
that of netropsin (1) and allows us to
define a general binding mode for this
family of ligands, in which an NH rim
points towards the internal area (inner
region) and a CH rim points towards
3
the external part (outer region) of the
minor groove of DNA. Also by means
of both TR-NOESY and STD experi-
ments, a description of the asymmetric
centers of the sugar residue close to
the inner and outer regions of the
groove has been achieved. These re-
sults confirm that the sugar is responsi-
ble for the differences previously found
in binding energetics.
(
ct-DNA). Two sets of experiments
with selective saturation of either the
H1’ or H4’/H5’/H5’’ proton NMR re-
gions of deoxyribose allow the descrip-
tion of the ligand residues close to the
inner (H1’) and outer regions (H4’/H5’/
H5’’) of the minor groove of double-
helical DNA. A series of complexes of
Keywords:
conformation
carbohydrates
analysis
·
·
DNA recognition · glycoconjugates ·
NMR spectroscopy
Introduction
ROESY experiments focus on the easily detected NMR sig-
[1–4]
nals of the free ligand to gain this information.
Addition-
Molecular recognition is at the heart of key biological
events. The 3D structures of the molecular complexes
formed by ligands and receptors are key factors for the bio-
logical response to take place. NMR spectroscopy has made
possible the study of the bound-state conformation in solu-
tion of a wide range of small ligands bound to biologically
relevant macromolecules. In particular, TR-NOESY and
ally, description of the epitope of the ligand directly in-
volved in complexation is also possible by using saturation-
[1,5,6]
transfer-difference (STD) experiments.
These two ex-
periments, which are based on intra- or intermolecular mag-
netization transfer between the ligand and receptor mole-
cules, have been widely used for the description of ligand
[6–14]
binding in protein and RNA complexes.
These macro-
molecules share the characteristic of highly efficient satura-
tion diffusion, which is typically observed in globular macro-
molecules.
[
a] F. Souard, E. MuÇoz, P. PeÇalver, C. Badía, R. del Villar-Guerra,
Dr. J. L. Asensio, Dr. C. Vicent
Departamento de Química Orgµnica Biológica
Instituto de Química Orgµnica general, CSIC
c/Juan de la Cierva 3, 28006 Madrid (Spain)
Fax : (+34)91-564-4853
The use of DNA polymers (calf thymus DNA (ct-DNA),
poly
A
H
R
U
G
[15–19]
ity of well-known natural DNA binders
signed ligands
E-mail: iqocv18@iqog.csic.es
ing thermodynamics and the kinetic parameters of the inter-
action by different techniques, such as fluorescence and mi-
[
b] Prof. J. JimØnez-Barbero
Dept. of Protein Science
Centro de Investigaciones Biológicas
c/Ramiro de Maetzu 9, 28040 Madrid (Spain)
[26]
crocalorimetry. These DNA polymers of either random or
fixed oligonucleotide sequences are helpful models to addi-
tionally define the binding selectivity of the ligands for spe-
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
[15,27–30]
cific DNA sequences.
Thus, these interaction studies
Chem. Eur. J. 2008, 14, 2435 – 2442
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2435

might help in the proper design of short oligonucleotide se-
quences and, moreover, in obtaining the structural details of
the binding mode of the ligand in solution by NMR spec-
troscopy.
tion of the molecular-recognition process at the structural
level has been elusive, due to the intrinsic difficulties of
using the proper techniques (such as NMR experiments) in
these complex systems.
Within a project aimed at the characterization of the basic
requirements of the interaction between glycoconjugates
and nucleic acids, we have recently designed a vector mole-
cule, the structure of which is made of an oligoamide dista-
mycin-type g-aminobutyric-linked covalent dimer (Py–g–Py–
Ind; Py: N-methylpyrrole; g: g-aminobutyric linker; Ind:
indole), to study its binding ability with DNA by using fluo-
On this basis, we herein present nonambiguous NMR evi-
dence of the molecular-recognition ability of the previously
[32]
studied (2, 4, and 5) and new sugar–oligoamides (3 and 6)
towards DNA sequences and provide additional tools, also
based on recently proposed STD-based NMR experi-
[33]
ments, to study this interaction at atomic resolution.
rescence spectroscopy. DNA polymers (ct-DNA, poly
dT), and poly(dC–dG)) were employed. This method al-
lowed the selectivity of the vector molecule for ATAT se-
A
H
R
U
G
A
C
H
T
R
E
U
N
G
Results and Discussion
[31]
quences to be determined.
The understanding of the recognition features of 2–6 (see
Scheme 1) with respect to DNA requires a proper under-
standing of the conformational properties of these molecules
in both the free and bound states. Thus, we aimed to get ex-
perimental evidence that could be related to the sugar prox-
imity to the DNA grooves. Additionally, knowledge of the
disposition of the pyranose moiety relative to either the
inner or the outer region of the DNA minor groove could
be used for further design within the sugar and/or the vector
part of the ligand.
Additionally, we have also shown, again by using fluores-
cence methods, that the vector can be used as a sugar carrier
[
32]
to the minor groove of DNA (Scheme 1). Initially NMR
and fluorescence studies
[31]
suggested that the previously
synthesized sugar–oligoamides 2, 4, and 5 presented a no-
ticeable percentage of hairpin conformation in the free
state, which was kept upon DNA binding. Also, competition
NMR experiments suggested that the designed ligands are
[32]
minor-groove binders.
Thus, at least for the initial de-
signed molecules, the sugar residue seems to modulate the
binding for the different DNA polymers studied and, even
more relevantly, the sugar contributes to the selectivity of
binding.
From a detailed structural perspective, and with regard to
the description of the complexes formed by long tracts of
helical DNA (DNA polymers) with small ligands, the defini-
On this basis, STD experiments were employed to get in-
formation on the ligand epitope with long tracts of helical
DNA. Recently, Gomez-Paloma and co-workers have de-
scribed the differential-frequency STD experiment as a tool
to distinguish between the three main different DNA–ligand
binding modes: minor groove, external electrostatic, and in-
tercalative binding. For this particular recognition process,
the difference in the vertical
and horizontal dimensions of
the macromolecule generates
an anisotropy that renders dif-
ferent efficiencies of saturation
diffusion along the two axes,
thereby making the STD ex-
periment different to that used
for other biomolecules. This im-
portant anisotropic effect in sat-
uration diffusion also generates
different experimental out-
comes depending on the chosen
[33]
saturation
frequency
(and,
therefore, the DNA region).
The structure of the minor
groove of B-DNA is character-
ized by the presence of the
sugar deoxyribose in the phos-
phodiester backbone, and the
attachment of the bases to the
backbone sugars through the
glycosidic bonds is asymmetri-
cal. This fact results in the for-
mation of two different grooves,
Scheme 1. Netropsin (1) and sugar–oligoamides (2–6). Xyl: xylose; Glc: glucose; Gal: galactose; Fuc: fucose. the major and minor grooves,
2436
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 2435 – 2442

Binding of Sugar–Oligoamides to DNA
FULL PAPER
on opposite sides of the base pairs. The floors of the grooves
serve as recognition sites for interacting molecules through
hydrogen-bonding centers. The binding of small molecules is
expected to give close contacts with the internal and exter-
nal regions of the groove, always depending on the structure
of the ligand and on the width of the groove. Nevertheless,
this last feature varies depending on the specific oligonu-
cleotide sequence at the DNA binding site. With regard to
the location of the deoxyribose hydrogen atoms attached to
the asymmetric carbon atoms within the groove structure,
inspection of a B-DNA model (see Figure 1) indicates that
hydrogen atom H1’ is closer to the internal region of the
groove, while hydrogen atoms H4’, H5’, and H5’’ are direct-
ed towards the external area.
tion will be not efficient enough and the signal/noise ratio
will be too small to clearly identify the bound epitopes. On
the other hand, if it is too long, the magnetization could be
transferred throughout the entire bound molecule by spin
diffusion. Thus, after carefully checking for spin-diffusion ef-
fects, we decided to employ a relatively short saturation
time of 400 ms for the study of all DNA-bound sugar–oli-
goamides.
[35]
As a first step, netropsin (1) was used to validate the
method as its binding mode to ct-DNA has been very well
established. The two sets of STD experiments for the 1–ct-
DNA complex by selective saturation on H1’ (in) or H4’/
[36]
H5’/H5’’ (out) gave a general clear-cut trend that could be
related to the location of the different protons of the ligands
in the DNA binding site (see Figure 4). While H1’ saturation
gave a significant enhancement of the resonances assigned
to Py3A and Py3B (in red), the transfer of saturation from
H4’/H5’/H5’’ rendered a major enhancement on CH A,
3
CH B, P5A, and P5B (in blue). For netropsin labeling and
3
STD results, see Table 1 in the Supporting Information.
These observations are in agreement with a crescent confor-
mation of netropsin in the bound state, for which the methyl
groups of the pyrrole amino acids and Py5 are directed to-
wards the outer region of the groove, while Py3A and Py3B
are pointing towards the inner region (Figure 2).
Figure 1. X-ray crystal structure of B-DNA, with identification of the po-
sitions of H1’, H4’, H5’, and H5’’ of the deoxyribose ring within the
double helix.
Figure 2. STD results for the netropsin–ct-DNA complex.
Thus, we decided to explore the use of the anisotropic
effect in saturation diffusion found in the STD experiment
for our minor-groove ligands (2–6).
Hence, two parallel sets of STD experiments were per-
formed under the same experimental conditions, with only
variations in the position at which saturation was elicited. Ir-
radiations were selected by choosing suitable frequencies
from the specific DNA regions, either H1’ at d=5.6 ppm
Interestingly, this conformation is completely in accord-
ance with that described for the bound state of netropsin in
the 1:1 complexes formed with several dodecanucleotides in
[37,38]
[39–45]
solution
and in the solid state.
Hence, the differen-
tial-frequency STD experiment is providing valuable geo-
metric information on the bound state of this minor-groove
binder. In this particular case, the NH rim is directed
toward the inner side of the minor groove, while the CH3
rim of the ligand is pointing towards the outer side of the
minor groove.
(
4
internal-minor-groove region) or H4’, H5’, and H5’’ at d=
[34]
.5 ppm (external-minor-groove region, Figure 2).
In all
cases, off-resonance irradiation was set at d=50 ppm. A
careful comparative analysis of the observed effects for the
two STD sets was then performed.
Thus, once the method was validated, our target sugar–
oligoamides (2–6; Scheme 1) were studied.
[6]
It is well known that the percentage of saturation trans-
fer from the DNA protons to the ligand protons essentially
depends on the saturation time, the dissociation rate, and
the molar fractions of the bound and free ligand. In general,
if the time of saturation is too short, the transfer of satura-
However, if the STD method is to be used, obviously, the
careful evaluation of the STD-based structure information
for the complexes between 2–6 and ct-DNA needs to be sus-
tained by previous knowledge of the free- and bound-state
conformations of these sugar–oligoamides with ct-DNA.
Chem. Eur. J. 2008, 14, 2435 – 2442
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2437

C. Vicent et al.
With regard to the free state, we have previously reported
a mixture of conformers in solution around the glycosidic
linkage. As a matter of fact, no NOE (either NH5–H1 or
NH5–H2) was detected that could support a major confor-
mer in solution. Thus, at least at the glycosidic linkage level,
3 behaves differently to 2 and 4–6, for which major anti con-
formers were present.
that 2, 4, and 5 adopt major hairpin-like conformations in
[32]
the free state. For this geometry, an NH rim and a CH3
rim (see Figure 4) can be safely defined. For the new mole-
cules 3 (a anomer of xylose) and 6 (b-l-Fuc), Figure 3a
shows the key interstrand NOEs found. The results show
The hairpin-like conforma-
tion of the oligoamides is de-
fined by a set of NOEs between
the indole-strand and sugar-
strand moieties joined by the g-
aminobutyric fragment: in par-
ticular, the CH A/Py5B and
3
Py3A/Py3B proton pairs. This
geometry is further clarified by
a set of additional NOEs be-
tween some pyranose and
indole protons.
Indeed, for the b-l-Fuc ana-
logue, 6, both sets of CH A/
3
Py5B and Py3A/Py3B were
found (Figure 3) and, moreover,
an interstrand NOE between
the H3 atom of the fucopyra-
nose ring and the In3 atom of
the indole allowed us to con-
clude that, in this case, the a
face of the sugar is pointing to-
Figure 3. a) Interstrand NOEs for sugar–oligoamides 3 and 6 in the free state. b) Interstrand NOEs for sugar– wards the indole ring in the
oligoamides 3 and 6 in the bound state.
free state (Figure 3). A differ-
ent trend was observed for 3. I n
this case, the interstrand NOEs
that neither the a/b nor the d/l series seems to alter the
presence of a crescent hairpin conformation in the free state
in a significant manner(See the discussion below and the
Supporting Information).
corresponding to the oligoamide fragment (Py5A/CH B and
3
Py5B/CH A) were detected, but no NOEs between the
3
indole and sugar proton resonances were found. Although
negative NOE data should not be taken as a major source
for conclusions, this observation is in agreement with the
coupling-constant data supporting a different shape for 3 to
that for the rest of the analogues (Figure 3a). The a-anome-
ric geometry places the sugar out of the plane of the oligoa-
mide, while, for instance, for the b-d-Xyl derivative 2, the b
Just as additional experimental data, and for sake of clari-
ty, a general discussion of the coupling constant values
3
(
J
NH5–H1) related to the conformation around the glycosidic
linkages (see Table 3 in the Supporting Information) will be
given here. For 2 and 4–6, the observed values are between
[32]
9.2–8.5 Hz. There is not any particular Karplus equation for
face of the xylopyranose ring is close to the indole ring.
this particular arrangement of H1–C1–N–H5, in which the
C1 atom is substituted by both oxygen and carbon atoms
and the nitrogen atom holds a carbonyl group. The most
similar equation has been recently reported, but it only ac-
Nevertheless, in both 2 and 3, the crescent-shaped hairpin
conformation is present in aqueous solution, as deduced
from the Py5A/CH B and Py5B/CH A NOEs.
Thus, ligands 2–6 adopt a significant percentage of a cres-
cent-shaped folded conformation in the free state, for which
3
3
[46]
counts for acetamide sugars at position C2 and not at C1.
Nevertheless, according to previous observations in glycosyl
NH and CH rims can be defined (see Figure 4).
3
[47]
amides, these values are in agreement with a major trans
orientation of NH5 and the H1 atom of the sugar pyranose
ring. Moreover, an NH5–H2 NOE was detected in these
cases, clearly supporting the predominant contribution of
the anti conformer, which allows a quasiparallel orientation
between the plane defined by the sugar pyranose and the
indole ring.
With regard to the bound-state conformation in the ct-
DNA complexes, as mentioned above, this knowledge has
been previously deduced for 2, 4, and 5 based on TR-
NOESY data. Thus, on this basis, we carried out the first
TR-NOESY experiments for the new sugar–oligoamides (3
and 6) in the presence of ct-DNA. For the b-l-Fuc analogue
6, TR-NOESY experiments were carried out in H O and
2
By contrast, for the a-d-Xyl analogue, 3, the JNH5–H1 value
D O to obtain NOEs from both exchangeable (amide) and
2
(
7.7 Hz) is significantly smaller and indicates the presence of
nonexchangeable protons (see the Supporting Information).
2438
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 2435 – 2442

Binding of Sugar–Oligoamides to DNA
FULL PAPER
The representation in Figure 3b exemplifies the interstrand
NOEs detected in the TR-NOESY experiments for 3 and 6
in the presence of ct-DNA. The complexes of both 3 and 6
Figure 4 shows, in blue, the resonances of the ligands that
are enhanced by irradiating the H4’/H5’/H5’’ spectral region.
As mentioned above, these sets of experiments allow the
ligand epitope close to the outer region of the DNA groove
to be described. On the other hand, the resonances of the
ligand in proximity with the inner region of the minor
groove (enhanced by irradiating H1’) are shown in red. The
obtained results suggest that the binding mode of these g-
linked covalent dimers (2–6) places their NH-rim areas
close to the inner region of the ct-DNA minor groove
(Py3A and Py3B are enhanced when H1’ is saturated).
with ct-DNA present interstrand NOEs between CH A/
3
Py5B and CH B/Py5A. The complex of 6 also shows the
3
P3A/P3B NOE and, more importantly, the H5ax proton of
the fucopyranose moiety shows NOEs with the resonances
of In3, In4 and In5 on the indole ring. Hence, for 6, the TR-
NOESY spectrum of the complex with ct-DNA does not
only show the interstrand contacts for the vector-oligoamide
fragment (crescent hairpin conformation) but additionally
permits confirmation of the spatial proximity between the a
face of the sugar ring and the indole residue in the bound
state. In the case of the a-d-Xyl derivative 3, as observed in
the free state, no NOEs were found between the sugar and
the indole moiety in the bound state. In both cases, all of
these negative TR-NOE cross-peaks confirmed the presence
of the hairpin conformation in the bound state.
The observed inter- and intrastrand NOEs (see the Sup-
porting Information) conclusively prove that the change
from a b to an a linkage or from the d to l series of the
sugar residues does not strongly affect the presence of a
crescent-shaped folded conformation in the ct-DNA bound
state. However, the relative orientation of the sugar with re-
spect to the rest of the oligoamide chain is clearly perturbed
when passing from equatorial (2, 4–6) to axial (3) anomers.
Nevertheless, according to the hairpin-like conformation,
Moreover, the CH rim is close to the outer region of the
3
groove (CH A, CH B, Py5A, and Py5B are enhanced when
3
3
H4’/H5’/H5’’ are saturated). These results, which are in per-
fect accordance with those obtained for netropsin (1), re-
semble the hairpin conformation with a crescent structure,
which was also determined by NMR spectroscopy, in com-
plex aromatic oligoamide ligands bound to short DNA frag-
[48–50]
ments.
Enhancement on the sugar region could only be interpret-
ed on a qualitative basis since strong signal overlap, as well
as spin diffusion along the pyranose protons, precluded the
derivation of clear-cut structure information for the relative
orientation of these protons in most of the cases. Neverthe-
less, this constitutes a significant result. Saturation is trans-
ferred from the ct-DNA to the sugar residues for all of the
sugar–oligoamide/ct-DNA complexes. Thus, even under
these conditions, the STD experiment constitutes a relevant
experimental demonstration of the proximity of the sugar to
the minor groove and confirms that the designed vector-oli-
goamide molecule, Py–g–Py–Ind, allows sugars to be carried
to the minor groove of B-DNA.
NH and CH rims can be defined (see Figure 4).
3
Once the bound-state geometry had been deduced, the
next step was to perform the differential-frequency STD ex-
periments for the complexes formed by 2–6 and ct-DNA.
Here, as for the 1–ct-DNA complex described above, two
parallel sets of STD experiments were performed for each
complex under the same experimental conditions.
In any case, the consideration of the bound-state confor-
mation (as deduced from the TR-NOESY data) with the
Figure 4. STD results for the complexes of netropsin (1) and sugar–oligoamides 2–6 with ct-DNA.
Chem. Eur. J. 2008, 14, 2435 – 2442
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2439

C. Vicent et al.
proximity of either the NH or CH rim to the groove of ct-
helpful to us in the rational design of potent multivalent
DNA binders.
3
DNA (as determined from STD data) permitted the forma-
tion of a general idea of the global orientation of the sugar–
oligoamide in the DNA. Indeed, although only the STD re-
sults obtained for the b-l-Fuc analogue 6 (Figure 4) could
be clearly assessed, the small observed overlapping for the
key Fuc protons allowed a well-defined binding mode to be
proposed. The differential-frequency STD enhancements
Experimental Section
Solvents were purified according to standard procedures. NMR spectro-
scopic measurements were recorded on a Bruker AVANCE 500 MHz
spectrometer. Calf thymus DNA (ct-DNA) was purchased from Sigma
permitted us to deduce that C5 (CH , H5) and C4 (H4) are
3
pointing toward the outer region of the groove, with the a
face of the fucose moiety pointing toward the indole ring in
the hairpin structure, as already characterized by the TR-
NOESY results.
Thus, the results obtained herein have shown a general
trend regarding the orientation of the oligoamide fragment
of all of the sugar–oligoamides studied (2–6). The most sig-
nificant result is that neither the a or b linkage of the sugar
to the oligoamide or the intrinsic features of the sugar resi-
due (different axial or equatorial substituents, d or l series)
strongly modify the position of the oligoamide Py–g–Py–Ind
vector within the groove of the B-DNA.
and used without further purification. H O for NMR studies was freshly
filtered milli-Q water. Netropsin was purchased from Fluka (lot
no. 405625-1) and used without further purification. Sugar–oligoamides
2
(
2–6) were synthesized according to the procedure previously described
for b-Gal-, b-Glc-, and b-Xyl-Py-g-Py-Ind (see Figure 23 in the Support-
[32]
ing Information).
Structural studies:
1
H NMR experiments with ligands 1–6: All spectra in aqueous solution
were recorded with presaturation of the water signal. The chemical shifts
were reported in ppm relative to 2,2-dimethyl-2-silapentane-5-sulfonic
acid or trimethylsilylpropionic acid (0.00 ppm) when D
relative to residual acetone (2.04 ppm) when D O/[D ]acetone or H
]acetone were used in the experiment. NMR structural studies of
2
O was used and
2
6
2
O/
[
D
6
compounds 1–6 were based on monodimensional and bidimensional
Therefore, based on TR-NOESY and STD experiments,
the sugar–oligoamides 2–6 when bound to ct-DNA adopted
a crescent-shaped folded conformation in which the NH rim
is directed toward the inner region of the minor groove,
(TOCSY, HSQC, NOESY, ROESY) experiments and were recorded at
4
00 or 500 MHz and 268C in a Varian instrument. Sample solutions were
prepared at concentrations ranging between 2 and 0.1 mm depending on
the solubility of the compounds.
Bound-state NMR spectroscopic experiments (TR-NOESY and differen-
tial-frequency STD): These experiments were carried out in phosphate
buffer (10 mm, pH 7). Ligand samples were prepared at a constant con-
centration of 1 mm. The ct-DNA titrant sample (stock solution) was pre-
pared by dissolving ct-DNA (2 mg) in a 1 mm solution of ligand (1 mL).
The concentration of ct-DNA was calculated by UV/Vis spectroscopy
while the CH rim is directed toward the outer region, a
3
binding mode similar to that of the netropsin–ct-DNA com-
plex.
À3
À1
À1 [51,52]
(
c=2.51”10 m, e=13200m cm ).
The NMR spectroscopic
Conclusion
sample was prepared by titration of the ligand solution (0.6 mL) with in-
creasing amounts of the titrant DNA solution, with the ligand concentra-
tion kept constant. A 1D NMR spectrum was recorded in the same “ac-
quisition mode” (no. of scans =256, T=268C) after each addition of ct-
DNA. A progressive broadening and disappearance of the proton signals
from the ligand was observed; this indicates binding of the ligand to ct-
DNA. The bound-state NMR experiments were carried out once the
spectra of the free ligand was clear but slightly broadened (addition of
ct-DNA: 100 mL for b-d-Glc and b-d-Gal derivatives 4 and 5, 140 mL for
b-l-Fuc analogue 6, 170 mL for netropsin (1), 425 mL for a-d-Xyl deriva-
tive 3). The same NMR tube was used for both the TR-NOESY and
STD experiments in every case.
Differential-frequency STD experiments have been used as
a major tool to characterize the key structural features of
small minor-groove binders with long tracts of DNA in solu-
tion. Selective saturation at different spectral regions can be
used as a fingerprint trace to locate the residues from the
bound ligand close to the inside or outside areas of the
groove. Thus, the combination of STD and TR-NOESY ex-
periments has allowed the determination of the conforma-
tion of 2–6 in the bound state, as well as their binding
modes in the groove of ct-DNA. Moreover, this method has
unequivocally assessed the spatial proximity between the
sugar residue and the groove. We feel that this experiment,
TR-NOESY experiments: TR-NOESY experiments for the bound ligand
were performed on 500 MHz spectrometers (Varian or Bruker) with satu-
ration of the residual H
TR-NOESY experiments were recorded at 268C and performed with
mixing times of 200, 300, and 400 ms. The experiments in H O/15% D
were recorded at 188C to minimize the hydrogen/deuterium exchange.
2
O signal or with the Watergate pulse sequence.
2
2
O
[33]
as originally proposed by Gomez-Paloma and co-workers,
is indeed an adequate means to study the binding ability of
ligands to DNA chains.
Negative and intense NOEs (which are indicative of binding to the mac-
romolecule) were found in all of the experiments; these were in contrast
with the weak NOEs (typical of small molecules) observed in the free-
state conformation for all of our oligoamides (2–6).
Thus, with all this new structure information at hand, it is
now possible to strongly support the theory that the previ-
[31]
STD experiments: STD experiments for the bound ligands were per-
formed on a 500 MHz AVANCE Bruker spectrometer. In the case of the
samples in H O, the Watergate pulse sequence was used. In the case of
2
ously reported differences in binding energetics
for the
sugar–oligoamides (2–6) with ct-DNA depend on the chemi-
cal nature of the sugar residue (DDG8 5 (poly
A
H
R
U
G
the samples in D O, either no water suppression or the Watergate se-
quence was used depending on the obtained signal/noise ratio without
2
À1
DNA)=À4.0 kcalmol ; DDG8 2 (poly
A
H
R
U
G
À1
solvent suppression. A ligand/receptor molar excess of up to 2.8:1 for b-
d-Glc and b-d-Gal analogues 4 and 5 was used for the best STD effects.
The STD effects of the individual protons were calculated for each com-
pound relative to a reference spectrum with off-resonance saturation at
d=50 ppm. 128 scans were recorded for the reference STD spectrum.
À1.5 kcalmol ).
On this basis, it is shown that the different sugars interact
differently with the DNA, possibly due to the different am-
phiphilic character of their surfaces. This new knowledge is
2440
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 2435 – 2442

Binding of Sugar–Oligoamides to DNA
FULL PAPER
The best duration of the saturation pulse, the power of the selective
Gaussian pulse, and the recycling delay were optimized to get the best
possible signal/noise ratio with no spin diffusion within a reasonable ex-
perimental time. The final experiments were performed at 298 K with a
recycling delay of 4 s, an acquisition time of around 1.3 s, and a saturation
time of 400 ms. The saturation was accomplished by using 8 Gaussian
shaped pulses of 49 ms each, separated by 1 ms, with an approximate
power of gB=20 Hz. Two saturation frequencies were selected: d=
[14] G. K. Jarori, N. Murali, B. D. N. Rao, Biochemistry 1994, 33, 6784–
6791.
[15] S. Neidle, Nat. Prod. Rep. 2001, 18, 291–309.
[16] J. B. Chaires, Curr. Opin. Struct. Biol. 1998, 8, 314–320.
[17] D. E. Wemmer, P. B. Dervan, Curr. Opin. Struct. Biol. 1997, 7, 355–
361.
[18] J. Hunziker, Chimia 1996, 50, 248–256.
[19] D. Kahne, Chem. Biol. 1995, 2, 7–12.
4
.50 ppm (to hit the H
4
’, H
5
’, and H
5
’’ region of deoxyribose in DNA) and
[20] U. Pindur, M. Jansen, T. Lemster, Curr. Med. Chem. 2005, 12, 2805–
2847.
[21] F. X. Han, N. Taulier, T. V. Chalikian, Biochemistry 2005, 44, 9785–
9794.
[22] P. B. Dervan, B. S. Edelson, Curr. Opin. Struct. Biol. 2003, 13, 284–
299.
[23] D. Renneberg, P. B. Dervan, J. Am. Chem. Soc. 2003, 125, 5707–
5716.
[24] P. B. Dervan, Bioorg. Med. Chem. 2001, 9, 2215–2235.
[25] L. Wang, C. Bailly, A. Kumar, D. Ding, M. Bajic, D. W. Boykin,
W. D. Wilson, Proc. Natl. Acad. Sci. USA 2000, 97, 12–16.
[26] F. G. Loontiens, P. Regenfuss, A. Zechel, L. Dumortier, R. M.
Clegg, Biochemistry 1990, 29, 9029–9039.
[27] A. Abudaya, P. M. Brown, K. R. Fox, Nucleic Acids Res. 1995, 23,
3385–3392.
[28] S. Eriksson, S. K. Kim, M. Kubista, B. Norden, Biochemistry 1993,
32, 2987–2998.
d=5.60 ppm (to hit the H1’ region in DNA). The closest proton resonan-
ces of the ligand were the anomeric proton of 6, at d=4.75 ppm, and the
Py3A hydrogen atom at d=6.46 ppm. Blank experiments performed
without the DNA receptor showed that no effect was observed in the
ligand protons when the H1’ region (d=5.60 ppm) in DNA was chosen,
while less than 5% saturation was observed in the anomeric H1 reso-
nance of compound 6 for on-resonance saturation at d=4.50 ppm. Al-
though this minor saturation does not imply any major concern in the
actual STD experiments and does not affect the vicinal protons through
spin diffusion, no STD effects in the sugar moiety of 6 were considered.
The intensity of each peak in the STD spectra was standardized in rela-
tion to a proton resonance the intensity of which was assigned to be 100.
The normalization of the signal intensity was done relative to Py3 and
Py5. Both sets of data treatment rendered identical conclusions. Such
conclusions were obtained by comparing the normalized value of each
proton resonance intensity upon irradiation of the outer or inner region
of the DNA groove. (A proton is thought to be closer to the irradiated
region of DNA when a bigger normalized intensity value is obtained if
the difference in percentage between both regions of DNA is at least 10–
[29] K. J. Breslauer, D. P. Remeta, W. Y. Chou, R. Ferrante, J. Curry, D.
Zaunczkowski, J. G. Snyder, L. A. Marky, Proc. Natl. Acad. Sci.
USA 1987, 84, 8922–8926.
1
5%.)
[30] L. A. Marky, K. J. Breslauer, Proc. Natl. Acad. Sci. USA 1987, 84,
4
359–4363.
[
31] P. PeÇalver, A. Samadi, P. Bosch, C. A. Hunter, C. Vicent, 2007, un-
published results.
Acknowledgements
[32] J. N. Martin, E. M. Munoz, C. Schwergold, F. Souard, J. L. Asensio,
J. Jimenez-Barbero, J. Canada, C. Vicent, J. Am. Chem. Soc. 2005,
127, 9518–9533.
[33] S. Di Micco, C. Bassarello, G. Bifulco, R. Riccio, L. Gomez-Paloma,
Angew. Chem. 2006, 118, 230–234; Angew. Chem. Int. Ed. 2006, 45,
224–228.
[34] Since the chemical shifts of these key DNA proton resonances show
rather distinct values to those of netropsin (1) and the sugar–oligoa-
mide ligands 2–6, selective irradiation of the internal or external
region of DNA was possible (see Figure 16 in the Supporting Infor-
mation). On this basis, and with the assumption of almost negligible
spin diffusion, it should be possible to verify which part of the
ligand is close to either area.
Financial support for this work was provided by the MEC (BQU2003-
3550-C03-01 and -02 and CTQ2006-10874-C02-01 and -02) and a Euro-
pean RTN project (HPRN-CT-2002-00190). P.P. is grateful to the Consejo
superior de Investigaciones Científicas for an I3P predoctoral fellowship.
C.B. and R.V.-G. thank the MEC for FPIand FPU predoctoral fellow-
ships, respectively, while F.S. thanks RTN for a postdoctoral position.
0
[
[
1] M. Vogtherr, T. Peters, J. Am. Chem. Soc. 2000, 122, 6093–6099.
2] J. L. Asensio, J. F. Espinosa, H. Dietrich, F. J. Canada, R. R.
Schmidt, M. Martin-Lomas, S. Andre, H. J. Gabius, J. Jimenez-Bar-
bero, J. Am. Chem. Soc. 1999, 121, 8995–9000.
3] T. Haselhorst, J. F. Espinosa, J. Jimenez-Barbero, T. Sokolowski, P.
Kosma, H. Brade, L. Brade, T. Peters, Biochemistry 1999, 38, 6449–
459.
4] J. Klein, R. Meinecke, M. Mayer, B. Meyer, J. Am. Chem. Soc. 1999,
21, 5336–5337.
5] V. V. Krishnan, Curr. Anal. Chem. 2005, 1, 307–320.
6] B. Meyer, T. Peters, Angew. Chem. 2003, 115, 890–918; Angew.
Chem. Int. Ed. 2003, 42, 864–890.
[
35] With regard to netropsin (1), we first carried out a NOESY experi-
1
2
ment in D O that allowed us to assign the H NMR resonances in
[
the free state. There is an incorrect assignment in reference [33].
Our assignment is in accordance with reference [38] and D. J. Patel,
Proc. Natl. Acad. Sci. USA 1982, 79, 6424–6428).
6
[
1
[36] The normalization/standardization of the signal intensity was done
relative to Py3 and Py5; both sets of data treatment render identical
conclusions.
[37] M. W. Freyer, R. Buscaglia, D. Cashman, S. Hyslop, W. D. Wilson,
J. B. Chaires, E. A. Lewis, Biophys. Chem. 2007, 126, 186–196.
[38] D. J. Patel, L. Shapiro, J. Biol. Chem. 1986, 261, 1230–1240.
[39] D. S. Goodsell, M. L. Kopka, R. E. Dickerson, Biochemistry 1995,
34, 4983–4993.
[
[
[
7] M. Mayer, P. T. Lang, S. Gerber, P. B. Madrid, I. G. Pinto, R. K.
Guy, T. L. James, Chem. Biol. 2006, 13, 993–1000.
8] M. Mayer, T. L. James, J. Am. Chem. Soc. 2004, 126, 4453–4460.
9] M. Mayer, T. L. James, J. Am. Chem. Soc. 2002, 124, 13376–13377.
[
[
[
10] T. Maltseva, E. Usova, A. Eriksson, J. Milecki, A. Foldesi, J. Chatto-
padhayaya, J. Chem. Soc. Perkin Trans. 2 2000, 2199–2207.
[40] K. Van Hecke, P. C. Nam, M. T. Nguyen, L. Van Meervelt, FEBSJ.
2005, 272, 3531–3541.
[
[
11] M. Vogtherr, S. Limmer, Magn. Reson. Chem. 1998, 36, 901–906.
12] S. Song, D. V. Velde, C. W. Gunn, R. H. Himes, Biochemistry 1994,
[41] N. G. A. Abrescia, L. Malinina, J. A. Subirana, J. Mol. Biol. 1999,
294, 657–666.
3
3, 693–698.
13] Determination of solution conformation of receptor-bound ligands
by NMR spectroscopy: transferred nuclear-Overhauser-effect
[42] X. Chen, S. N. Mitra, S. T. Rao, K. Sekar, M. Sundaralingam, Nucleic
Acids Res. 1998, 26, 5464–5471.
[43] L. Tabernero, N. Verdaguer, M. Coll, I. Fita, G. A. Vandermarel,
J. H. Vanboom, A. Rich, J. Aymami, Biochemistry 1993, 32, 8403–
8410.
[
A
study of cyclophilin and a model substrate: L. T. Kakalis, I. M. Ar-
mitage in Molecular Modeling: From Virtual Tools to Real Problems
(
ACS Symposium Series, Vol. 576) (Eds.: T. F. Kumosinski, M. N.
[44] M. Coll, J. Aymami, G. A. Vandermarel, J. H. Vanboom, A. Rich,
A. H. J. Wang, Biochemistry 1989, 28, 310–320.
Liebman), ACS, Washington D.C., 1998, pp. 29–44.
Chem. Eur. J. 2008, 14, 2435 – 2442
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2441

C. Vicent et al.
[
45] M. L. Kopka, C. Yoon, D. Goodsell, P. Pjura, R. E. Dickerson, J.
Mol. Biol. 1985, 183, 553–563.
[50] D. S. Pilch, N. Poklar, C. A. Gelfand, S. M. Law, K. J. Breslauer,
E. E. Baird, P. B. Dervan, Proc. Natl. Acad. Sci. USA 1996, 93,
8306–8311.
[51] M. Gopal, M. S. Shahabuddin, S. R. Inamdar, Proc. Indian Acad.
Sci. Chem. Sci. 2002, 114, 687–696.
[
46] M. Mobli, A. Almond, Org. Biomol. Chem. 2007, 5, 2243–2251.
47] M. Avalos, R. Babiano, M. J. Carretero, P. Cintas, F. J. Higes, J. L. Ji-
menez, J. C. Palacios, Tetrahedron 1998, 54, 615–628.
[
[
[
48] C. A. Hawkins, R. P. de Clairac, R. N. Dominey, E. E. Baird, S.
White, P. B. Dervan, D. E. Wemmer, J. Am. Chem. Soc. 2000, 122,
[52] R. D. Wells, R. C. Grant, B. E. Shortle, C. R. Cantor, J. Mol. Biol.
1970, 54, 465–497.
5
235–5243.
Received: July 18, 2007
Revised: November 28, 2007
49] R. P. L. de Clairac, B. H. Geierstanger, M. Mrksich, P. B. Dervan,
D. E. Wemmer, J. Am. Chem. Soc. 1997, 119, 7909–7916.
Published online: January 18, 2008
2442
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 2435 – 2442