Effects of 8-methyl-2′-deoxyadenosine incorporation into quadruplex forming oligodeoxyribonucleotides

Article abstract of DOI:10.1016/j.bmc.2004.11.037

In this paper we report the synthesis and the structural characterization of two modified oligodeoxyribonucleotides (ODNs), namely d(A^8MeGGGT) and d(TA^8MeGGGT), where A^8Me represents a 8-methyl-2′-deoxyadenosine. Both ODNs have been studied by ¹H NMR, CD spectroscopy and molecular modelling and shown to form fourfolds symmetric G-quadruplex structures, with all strands parallel and equivalent to each other. The complexes are characterized by thermal stabilities comparable to that of their natural counterparts. NOE patterns involving 8-methyl group in A^8Me residues allowed us to define the main structural features at the 5′-end of the complexes. Particularly, inter- and intrastrand NOEs show a syn-orientation and a symmetrical arrangement of A^8Me bases stacking on the adjacent G-tetrad.

Full text of DOI:10.1016/j.bmc.2004.11.037

Bioorganic & Medicinal Chemistry 13 (2005) 1037–1044
Effects of 8-methyl-20-deoxyadenosine incorporation into
quadruplex forming oligodeoxyribonucleotides
Antonella Virgilio, Veronica Esposito, Antonio Randazzo,
Luciano Mayol and Aldo Galeone*
`
Dipartimento di Chimica delle Sostanze Naturali, Universita degli Studi di Napoli ‘Federico II’, via D. Montesano 49,
I-80131 Napoli, Italy
Received 3 August 2004; accepted 19 November 2004
Available online 15 December 2004
Abstract—In this paper we report the synthesis and the structural characterization of two modified oligodeoxyribonucleotides
(ODNs), namely d(A8MeGGGT) and d(TA8MeGGGT), where A8Me represents a 8-methyl-20-deoxyadenosine. Both ODNs have
been studied by 1H NMR, CD spectroscopy and molecular modelling and shown to form fourfolds symmetric G-quadruplex struc-
tures, with all strands parallel and equivalent to each other. The complexes are characterized by thermal stabilities comparable to
that of their natural counterparts. NOE patterns involving 8-methyl group in A8Me residues allowed us to define the main structural
features at the 50-end of the complexes. Particularly, inter- and intrastrand NOEs show a syn-orientation and a symmetrical arrange-
ment of A8Me bases stacking on the adjacent G-tetrad.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
sidic conformation of the base moiety: in parallel quad-
ruplexes all G-residues prefer to adopt an anti-
conformation, while in antiparallel complexes, for each
tetrad, two anti and two syn G-residues are present.4
However, in a recent publication of some of us5 the
probable occurrence of an all syn-tetrad formed by 8-
bromo-dG (GBr) residues in the parallel quadruplexes
was pointed out. In fact, the presence of a bulky substi-
tuent such as the bromine atom at the C8 position,
destabilizing the usual anti-glycosidic conformation,
constrains the residue to adopt a syn-orientation of the
base. Unexpectedly, one of the resulting tetramolecular
quadruplexes, namely [d(TGBrGGT)]4, showed a con-
siderable increase of the thermal stability (about
16 ꢁC) compared to its natural counterpart.
G-quadruplex structures are formed in vitro by DNA
and RNA oligonucleotides containing runs of guanine
bases able to form a variable number of planar arrays
of four Gs, named G-tetrads. The interest level in these
structures has markedly increased in the past decade as
evidence for possible functional roles in vivo have been
accumulated. In fact, G-rich sequences occur in several
biologically important regions such as chromosomal
telomeres, gene promoter regions, recombination sites,
RNA packaging sites and RNA dimerization do-
mains.1–3 Of particular interest is the structural versatil-
ity of G-quadruplexes that are able to assume a
surprising amount of arrangements differing for the
strands stoichiometry and their relative orientation (par-
allel or antiparallel), as well as the glycosidic angle of
guanosine residues (syn or anti).
A further contribution to the structural variability of G-
quadruplexes comes from the discovery of unusual tetr-
ads such A-tetrads,6–8 T-tetrads,9 C-tetrads,10 mixed
tetrads11 and modified bases containing tetrads5,12 that
increase the number of sequences potentially able to
form quadruplexes. Recently, our attention has been ad-
dressed to A-tetrads since they are present in several se-
quences forming tetramolecular parallel quadruplexes
(among which the human telomere sequence), where
both an all syn A-tetrad (AGGGT)6 and all anti A-tetr-
ads (TTAGGGT, TGGAGGC)7,8 have been found.
A crucial structural feature in G-quadruplex complexes
is the relationship between strand orientation and glyco-
Keywords: Quadruplex; 8-Methyl-20-deoxyadenosine; A-tetrad; Gly-
cosidic conformation.
*
Corresponding author. Tel.: +39 081678542; fax: +39 081678552;
e-mail: galeone@unina.it
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.11.037

1038
A. Virgilio et al. / Bioorg. Med. Chem. 13 (2005) 1037–1044
In view of the above considerations, we have undertaken
the synthesis of 20-deoxy-8-methyladenosine containing
ODNs potentially able to form stable quadruplex struc-
tures in order to find a possible connection among
relative strand orientation, glycosidic conformation in A-
tetrads and thermal stability. In fact, the methyl group
at the 8-position, besides promoting the syn-glycosidic
conformation, should also make easier the structural
studies thanks to information possibly arising from
NOE contacts, which the methyl proton may establish.
Furthermore, 8-methyl-20-deoxyadenosine could be
potentially useful as alternative to 20-deoxyadenosine
in SELEX13 technology, thus expanding the structural
features that could be used for the search of new
aptamers.
was treated with 5,50-dimethoxytritylchloride in dry pyr-
idine17 for the final protection of the 50-OH group and
the derived compound (6) was, in turn, transformed into
the corresponding phosphoramidite monomer (7) by 2-
cyanoethyl-N,N-diisopropyl chlorophosphoramidite.18
Compound 7 was used for the preparation of the ODNs
I and II following the usual protocols. During the auto-
mated syntheses, 7 exhibited similar coupling efficiencies
as the common, commercially available, phosphorami-
dite derivatives.
The ODNs were purified by HPLC, and the fractions
containing only full-length products were pooled and
used to prepare NMR samples at a concentration of ap-
prox 5 mM (0.6 mL, 90% H2O/10% D2O), in a 10 mM
potassium phosphate, 70 mM KCl, 0.2 mM EDTA
(pH 7.0) buffer. 1H NMR spectra of I and II were
recorded using pulsed-field gradient watergate19 for
H2O suppression.
In this paper, we report the synthesis, CD experiments,
NMR and molecular modelling studies of ODNs
A
8MeGGGT (I) and TA8MeGGGT (II) (A8Me-ODNs),
where A8Me represents 20-deoxy-8-methyladenosine.
The relatively simple appearance of one-dimensional
spectra of I and II indicates that, in the conditions used
here, both the modified oligomers mainly form a single
well-defined hydrogen-bonded conformation, confi-
dently identified as quadruplexes as described below,
having in mind the structures of the unmodified ODNs
counterparts. Nevertheless, some weaker resonances
2. Results and discussion
The fundamental step for the synthesis of oligon-
ucleotides containing A8Me is the preparation of the
suitably protected A8Me phosphoramidite monomer
required for the site-specific incorporation into DNA
oligomers. The synthetic strategy of the fully prote-
cted A8Me monomer is outlined in Scheme 1.
1
attributable to minor forms are present in both the H
NMR spectra, whose appearance is not significantly im-
proved by altering the buffer or changing the tempera-
ture. Since some of those signals were also present
between 11–12 ppm, that is a diagnostic region to iden-
tify quadruplex structures, the weaker resonances were
attributed to minor quadruplex conformations.
20-Deoxyadenosine (1) was brominated by treatment
with bromine in CH3COOH/CH3COONa aqueous buf-
fer according to Eason et al.14 thus affording 8-bromo-
20-deoxyadenosine (2). Compound 2 was protected at
30 and 50 hydroxyl functions by HMDS to yield 3 and
subsequently transformed into the intermediate 4 fol-
lowing the Van Aershot et al.15 procedure. Since a par-
tial detachment of protecting groups at 30 and 50
functions occurred during the chromatographic purifica-
tion of 4, the successive reaction was the well-known
transient protection16 carried out on the partial depro-
tected intermediates collected from the column. The so
obtained N-benzoyl-20-deoxy-8-methyladenosine (5)
At the temperature of 80 ꢁC, the oligomers result
unstructured and only the resonances corresponding to
the single strand are present.
As far as 1D proton spectrum of I (T = 30 ꢁC) is con-
cerned, resonances corresponding to three G-H8 and
to T-H6 and A8Me–H2 protons in the aromatic region
and two methyl resonances around 1.6 ppm for T-CH3
and 2.3 ppm for A8Me-CH3 are observed. Analogously,
NH2
NH2
NHBz
NHBz
N
N
N
N
N
N
N
N
N
N
N
N
R1
H3C
H3C
N
N
d, e, f
a
g
N
HO
N
R2O
HO
DMTrO
O
O
O
O
OH
OR3
OH
OR3
OCH2CH2CN
1
5
2: R1 = Br; R2 = H; R3 = H
b
c
6: R3 = H
7: R3 =
h
3: R1 = Br; R2 = TMS; R3 = TMS
4: R1 = CH3; R2 = TMS; R3 = TMS
P
N(iPr)2
TMS = trimethylsilyl
Bz = benzoyl
DMTr = dimetoxytriphenylmethyl
HMDS = hexamethyldisilazane
Scheme 1. Synthesis of phosphoramidite derivative 7. Reagents and conditions: (a) Br2 in acetate buffer; (b) hexamethyldisilazane, (NH4)2SO4 in dry
dioxane; (c) Pd(PPh)4, (CH3)4Sn in N-methylpirrolidinone; (d) trimethylsilyl chloride in dry pyridine; (e) benzoyl chloride in dry pyridine; (f) H2O/
NH3; (g) dimethoxytriphenylmethyl chloride in dry pyridine; (h) 2-cyanoethyl-N,N-diisopropyl chlorophosphoramidite in dry CH2Cl2.

A. Virgilio et al. / Bioorg. Med. Chem. 13 (2005) 1037–1044
1039
in the case of proton spectrum of II, there are only six
signals in the aromatic region, whereas, as expected,
three methyl resonances are present in the region
between 1.5 and 2.5 ppm, where the chemical shift
of the methyl group of A8Me is ca. 0.7 ppm further
downfield shifted than in thymine. Since resonances
from only one strand are observed for both molecules,
should the ODNs be structured in multistranded comp-
lexes, these must be symmetric.
(T = 30 ꢁC) for both molecules (see experimental sec-
tion), showed well dispersed cross peaks, so both
exchangeable and nonexchangeable protons could be
nearly completely assigned following the standard pro-
cedures.25 The observed NOEs among G-H8/T-H6 and
their own H10, H20 and H200 sugar protons and the
H10, H20 and H200 protons of the preceding residue
(Fig. 2) suggested that both quadruplexes adopt a right
handed helical winding. Moreover, PE-COSY spectra
analysis indicates that H10/H20 coupling constants are
reasonably large, thus suggesting that the sugar geome-
tries are predominantly S-type. Therefore, the structure
of each strand may be taken to be similar to B-DNA
form.
Furthermore, proton NMR spectra of both I and II in
K+ containing aqueous solution at a temperature of
30 ꢁC also show three signals in the region 10.5–
12 ppm attributable to three exchangeable guanine N1
imino protons. These signals persist at temperature
>40 ꢁC and slowly exchange after dilution of the sample
into D2O solution, consistently with the high kinetic sta-
bility and low solvent accessibility of quadruplex
structures.20
As for the glycosidic torsion angles, useful information
could be obtained comparing the relative intensities of
NOEs between H8/H6 and H20 and H8/H6 and H10 pro-
tons of the same residue. All Gs and Ts resulted to pos-
sess an anti-glycosidic conformation, while the modified
adenosines (A8Me) adopt a syn-conformation, showing
intense NOEs between methyl group in 8-position and
H10 sugar proton and more weak crosspeaks between
methyl and H20.
Thus, the whole of data suggest that the guanines from
each strand form three G-tetrads and the observed num-
ber of resonances is consistent with symmetrical four-
stranded quadruplexes with all strands equivalent to
each other.
It is interesting to note that the normal sequential con-
nectivities path is broken at 50-T A8Me-30 step since the
syn-orientation of the modified base places the protons
of methyl group in 8-position to a distance greater than
6 A away from the sugar protons on the neighboring 5
nucleotide.26
The proton decoupled phosphorus spectra of both I and
II in D2O at 30 ꢁC show that all 31P signals are clustered
within ꢀ0.8 and ꢀ2.2 ppm region, characteristic of
unperturbed backbone phosphates of parallel stranded
quadruplexes.21–23
0
˚
Moreover, circular dichroism (CD) data (Fig. 1), ac-
quired for both samples at 30 ꢁC, further inferred the
formation of parallel quadruplexes. In fact, the presence
of a maximum and minimum Cotton effects at 262–
264 nm and 243 nm, respectively, are typical of quadru-
plex involving four parallel strands.24
Furthermore, NMR data of both I and II show a num-
ber of NOE connectivities involving A8Me bases. Partic-
ularly, the presence of interstrand NOEs between the
methyl group of an A8Me residue and the H2 proton
of the modified base on the adjacent strand (Fig. 2),
and intrastrand NOEs between A8Me protons nucleo-
tides and the protons of the adjacent Gs on the same
strand suggest that A8Me residues are not randomly ori-
ented and are in mutual close proximity arranging in a
symmetrical fashion and stacking on the top of G-tetr-
ads in both [d(A8MeGGGT)]4 and [d(TA8MeGGGT)]4.
Unfortunately, no clearly resolved resonances for the
6-NH2 group (indicating the formation of H-bonds) of
A8Me bases could be observed.
In spite of the little structural heterogeneity observed for
both complexes, we were able to perform both reso-
nance assignments (Table 1) and structural studies of
the main species present in solution. Particularly,
NOESY and TOCSY spectra, obtained at 500 MHz
Crosspeaks in the NOESY spectra (D2O, mixing
time = 180 ms) of quadruplexes I and II are reasonably
well dispersed, allowing the quantization of the experi-
mental NMR data. The distance restraints, deduced
from the Overhauser effect intensities by the tool CALI-
BA of the program CYANA,27 were used at the various
stages of the calculations (see experimental section) to
determine the three-dimensional structure of the com-
plexes formed by I and II. Hydrogen-bond distance re-
straints about three layers of G-tetrad were also
incorporated during the computations. Particularly,
336 distance constraints for I and 320 for II were derived
from NOE peak intensities and reduced to 206 and 160,
respectively, after removal of the irrelevant ones, and 24
H-bonds (48 distance restraints: HN1–O6, N1–O6,
HN2–N7, N2–N7) were took into account during the
Figure 1. CD spectra of I (—–) and II (---).

1040
A. Virgilio et al. / Bioorg. Med. Chem. 13 (2005) 1037–1044
Table 1. Nonexchangeable and imino protons chemical shifts (500 MHz) for [d(A8MeGGGT)]4 and [d(TA8MeGGGT)]4 quadruplexes in 10 mM
KH2PO4, 70 mM KCl, 0.2 mM EDTA (pH 7.0, T = 30 ꢁC)
Base (50–30)
H8/H6
H10
H20/H200
H30
H40
H50/H500
H2/Me
NH
A8Me
G
6.09
6.02
5.95
6.25
6.08
2.42–2.54
2.81–3.03
2.64
4.89
5.04
4.98
4.88
4.47
4.213.58–3.73
4.47
4.50
7.65–2.28
8.13
7.74
7.67
7.35
4.14
4.32
4.26
4.22
11.57
11.36
11.05
G
G
2.51–2.66
2.17
4.49
4.06
T
1.61
1.65
T
7.20
5.88
6.03
6.01
5.95
6.24
6.07
1.75–2.22
2.55
4.60
4.90
5.03
4.99
4.90
4.46
3.94
3.48
A8Me
G
4.513.93
4.45
4.46
7.78–2.45
8.05
7.72
7.67
7.35
2.49–2.77
2.64
4.17
4.28
4.26
4.20
11.56
11.27
10.97
G
G
2.52–2.67
2.16
4.49
4.04
T
1.60
Figure 2. Expanded region of a NOESY spectrum (500 MHz, 180 ms mixing time, T = 30 ꢁC) containing aromatic and H20/H200 sequential
crosspeaks and NOEs involving methyl groups of I.
calculation for both quadruplex structures. Further-
more, in accordance to the observed 31P chemical
shifts,21–23 backbone torsion angles were restricted to
be in a range of 20ꢁ of helical values of the natural
quadruplexes [d(AGGGT)]4 and [d(TAGGGT)]4.6
According to NMR data, glycosidic torsion angles for
all guanines and thymines were fixed in the anti-domain
(ꢀ157ꢁ/ꢀ117ꢁ), while a range of 0ꢁ/90ꢁ (syn-conforma-
tion) was used for v angles of A8Me residues.
RMSD values of 1.96 0.50 and 2.14 0.46 were calcu-
lated for II, showing that the NOE restraints are largely
satisfied for both complexes.
As expected, each strand of the complex formed by I
possesses a right-handed helical backbone geometry
and the same guanines from each of the four strands
align in a plane to form three G-tetrads. As for A8Me
arrangement, they assume a almost planar arrangement
and adopt a syn-conformation around the glycosidic
bonds. Interestingly, while in the unmodified
The structure calculations on the two molecules I and II
were undertaken by using restrained distance geometry
calculations (CYANA).27 The 10 out of 100 structures
with the lowest CYANA target functions resulting from
van der Waals and restraints violations were analysed in
both cases and subjected to restrained energy minimiza-
tion by using the force field CVFF (see experimental sec-
tion). Particularly, average RMSD values of 1.30 0.43
and 1.35 0.37 for the backbone and all heavy atoms,
respectively, were obtained from the superimposition
of the 10 minimized structures obtained for I, while
6
[d(AGGGT)]4 two different patterns of H-bonds could
be observed between H6 and N1(named N61) and alter-
natively between H6 and N7 (named N67) (Fig. 3), the
A8Me residues seem to be characterized by neither pat-
tern N61nor N67, even though the mean structure sug-
gests the formation of the N61pattern of H-bond ( Fig.
4).
In order to better understand the different behaviour of
the 50 end of I in comparison to the one of the unmod-

A. Virgilio et al. / Bioorg. Med. Chem. 13 (2005) 1037–1044
1041
Figure 3. N61and N67 H-bond patterns for an A-tetrad. H-bonds are indicated by dashed lines.
As for the structures obtained for II, they are character-
ized by a remarkable conformational disorder at 50-
TA8Me-30 step. Even if the resulting mean structure
shows the possibility of formation of an A8Me-tetrad
with all residues in syn-conformation and held together
by H-bond between A8Me-amino hydrogen of one base
and nitrogen N6 of the adjacent one, the A8Me residues
in the single structures deviate a lot from planarity and
the stacking with the adjacent G-quartet is rather poor.
By contrast, the 50-GGGT-30 tracks of the molecule
resemble perfectly the structures of [d(A8MeGGGT)]4
and of their natural counterparts.6 Molecular modelling
studies were performed in order to better understand
this disorder. As in I, two models were built taking into
account all structural features deduced from NMR stud-
ies and imposing in the calculations either the H-bond
pattern N61or N67. The two models generated for II re-
veals that methyl groups of A8Me residues seem to suffer
of steric effects due to their interaction with deoxyribose
rings of the previous residue (T) of each strand. Partic-
ularly, this is mainly evident when the formation of
H-bond pattern N61is forced ( Fig. 5), whereas the pres-
ence of the methyl groups of A8Me makes the form-
ation of the hydrogen bonds of the type N67 possible
only on condition that the planarity for the four modi-
fied adenines is lost. These results may suggest a possible
explanation for the absence of a good definition of the
50-edge of the complex, indicating that the terminal
thymine bases at 50-edge exert a very relevant influence
on the arrangement of the underneath residues.
Figure 4. A8Me-tetrad observed in the mean structure of I. Heavy
atoms are depicted in colored ÔstickÕ (carbons, green; nitrogens, blue;
oxygens, red; hydrogens, white). Dashed yellow lines represent
hydrogen bonds.
ified quadruplex, we performed further molecular mod-
elling studies on I. Particularly, we have generated two
models taking into account all the experimental con-
straints (interproton distances) and several structural
features deduced from NMR studies, such as (i) the 4-
fold symmetry and (ii) right-handed helicity of the over-
all structures, (iii) the anti-glycosidic conformation of all
G and T residues, (iv) the syn-glycosidic torsion angle
adopted by A8Me nucleosides, (v) the presence of three
G-quartets in a planar hydrogen bonded arrangement.
The two models only differ for the H-bonds pattern,
namely N61and N67. The models clearly showed that
the formation of both N61and N67 patterns of H-bond
are equally possible, thus suggesting that the slight line
broadening suffered by the signals belonging to A8Me
residues and the consequent lack of a sufficient number
of constraints at the 50-end of the molecule might have
prevented the right definition of that part of the
molecule.
In order to estimate the effects of the substitution of a
regular A residue with an A8Me one on the thermal sta-
bility of the resulting quadruplex structures were further
analysed by CD thermal denaturation experiments. The
melting profile of I, II and their natural counterparts re-
corded at 264 nm give well-shaped sigmoid curves (Fig.
6). The Tm values, calculated as the maximum of the first
derivative plots of molar ellipticity versus temperature,
are listed in Table 2. These data show that thermal sta-
bility of both I and II are quite similar to that observed
for the reference structures, thus suggesting that the
incorporation of a methyl group does not affect the

1042
A. Virgilio et al. / Bioorg. Med. Chem. 13 (2005) 1037–1044
Table 2. Melting temperature of I, II and their natural counterparts
Tm (ꢁC)
[d(AGGGT)]4
[d(A8MeGGGT)]4 (I)
[d(TAGGGT)]4
[d(TA8MeGGGT)]4 (II)
63
61
67
68
complexes. Particularly inter- and intra-strand NOEs
show a syn-orientation and a symmetrical arrangement
of A8Me bases stacking on the adjacent G-tetrad for
both quadruplex structures. Unfortunately, we were
not able to detect signals attributable to hydrogen
bonded 6-NH2 group of A8Me residues as Patel et al.6
have observed just in the case of the natural counterpart
of I. However, Searle et al.8 in a study concerning the
parallel quadruplex [d(TGGAGGC)]4, suggest that the
presence of a strong hydrogen bond might not be a
feature of A-tetrad stabilization. Regarding oligonucl-
eotide II, we also found a syn-orientation for A8Me
bases, in contrast with the above cited report6 where
the natural counterpart of II adopts an anti-glycosidic
conformation. These data suggest that A8Me can be
regarded as a convenient and useful probe for glycosidic
conformational preferences. On the other hand the
molecular modelling studies suggest probable steric
effects due to the methyl group that perturb the planar-
ity of the modified adenines.
Figure 5. Side view of average structure of the best 10 structures of II.
Heavy atoms are shown with different colors (carbons, green;
nitrogens, blue; oxygens, red; hydrogens, white). One A8Me residue
and the dT residue of the same strand are reported in CPK. Steric effect
between methyl of A8Me and the dT residue is plainly observable.
quadruplex stability. Data acquired for I and II differ
from the ones obtained by some of us for ODNs with
the same sequences and containing 8-propynyl-20-deox-
yadenosine (APr-ODNs),28,29 whose thermal stability re-
sulted to be higher than the natural counterparts. In
order to find a possible explanation of the different
behaviour of the A modified containing ODNs, detailed
thermodynamic analyses of APr-ODNs have been car-
ried out and will be published elsewhere,30 while similar
analyses of AMe-ODNs are currently in progress.
Thermal stabilities of both I and II are relatively similar
to those detected for their unmodified counterparts sug-
gesting that the presence of A8Me bases does not signif-
icantly affect the quadruplex stability.
The incorporation of A8Me into a DNA synthetic se-
quence might be of interest at the level of ON-based
therapeutics, a research area which yielded, so far, vari-
ous type of pharmacologically active molecules.31
Among these, aptamers are of particular importance
since they can be generated against a wide variety of bio-
logically significant targets (both proteins and small
molecules) through iterative in vitro selection techniques
(SELEX).13 The scaffold of several aptamers is based on
G-quadruplex structures and the presence of modified
bases could, in principle, improve the biological activity.
In this frame, A8Me appears to be a useful vehicle to
introduce an alkyl group into the aptamer grooves,
potentially able to establish hydrophobic contacts with
the target molecule, thus improving the interaction.
3. Conclusions
In this paper we report the synthesis and a structural
study by NMR and molecular mechanics and dynamics
calculation of the ODNs A8Me-GGGT (I) and TA8Me
-
GGGT (II), where A8Me represents 8-methyl-20-deoxy-
adenosine. Our results indicate that both oligonucleo-
tides I and II form a parallel four stranded quadruplex
in solution. As expected, NOE patterns involving
8-methyl group in A8Me residues allowed us to define
the main structural features of the 50-end of the
Figure 6. CD thermal denaturation profile of [d(A8MeGGGT)]4 (I) (—–) and [d(AGGGT)]4 (---) (A); [d(TA8Me GGGT)]4 (II) (—–) and
[d(TAGGGT)]4 (---) (B).

A. Virgilio et al. / Bioorg. Med. Chem. 13 (2005) 1037–1044
1043
The behaviour of A8Me residues in sequences potentially
able to form A-tetrads in other parallel and/or antipar-
allel quadruplex structures is currently in progress in
our laboratory.
in the NOESY spectra (with 180 ms mixing time) with
the CALIBA tool of the program CYANA.27 Pseudo-
atoms were introduced where needed. 336 distance con-
straints for I and 320 for II were derived from NOE
peak intensities and reduced to 206 and 160, respectively,
after removal of the irrelevant ones. Hydrogen bonds
constraints (16 upper and 16 lower limit constraints/
G-tetrad: HN1–O6, N1–O6, HN2–N7, N2–N7) were
incorporated with upper and lower distance limits of
4. Material and methods
4.1. Synthesis of the oligonucleotides
˚
˚
2.0 A and 1.7 A for the hydrogen-acceptor distance
˚
˚
Oligonucleotides I, II and their natural counterparts
were synthesized on a Millipore Cyclon Plus DNA syn-
thesizer, using solid phase b-cyanoethyl phosphorami-
dite chemistry at 15 lmol scale. The oligomers were
detached from the support and deprotected by treat-
ment with concd aqueous ammonia at 55 ꢁC for 12 h.
The combined filtrates and washings were concentrated
under reduced pressure, redissolved in H2O and ana-
lysed and purified by HPLC on a Nucleogel SAX col-
and 3.0 A and 2.7 A for the donor–acceptor distance,
respectively. These constraints for H-bonds did not lead
to an increase in residual constraints violation. In accor-
dance to the observed 31P chemical shifts,21–23 backbone
torsion angle constraints were restricted to be in a range
of 20ꢁ with the respect of the angles of the unmodified
quadruplexes. Glycosidic torsion angles for all guanines
were kept in a range of ꢀ190ꢁ/ꢀ140ꢁ (anti-conforma-
tion), whereas a range of ꢀ157ꢁ/ꢀ117ꢁ (anti-conforma-
tion) was used for thymine residues. The 10 structures
with the lowest CYANA target functions were subjected
to energy minimization (with no angle constraints) using
the conjugate gradient method and the CVFF force field
as implemented in the program DISCOVER (Accelrys,
San Diego, USA). During energy minimization, inter-
proton distances and H-bond constraints involving G-
tetrads were used with a force constant of 20 and
umn (1000-8/46, Macherey-Nagel, Duren, Germany);
¨
using buffer A: 20 mM KH2PO4 aq solution, pH 7.0,
containing 20% (v/v) CH3CN; buffer B: 1M KCl,
20 mM KH2PO4 aq solution, pH 7.0, containing 20%
(v/v) CH3CN; a linear gradient from 0% to 100% B in
30 min. and flow rate 1mL/min. were used.
All oligomers resulted to be more than 98% pure
(NMR).
100 kcal molꢀ1 A-2, respectively. Illustrations of struc-
˚
tures were generated with the INSIGHT II program
(Accelrys, San Diego, USA). All the calculations have
been performed on a SGI Octane workstation.
4.2. Nuclear magnetic resonance
The NMR samples had a concentration of approx
5 mM, in 0.6 mL (H2O/D2O 9:1) buffer solution having
1 0 mM KHPO4, 70 mM KCl, 0.2 mM EDTA, pH 7.0.
4.4. Circular dichroism and CD melting experiments
2
For D2O experiments, the H2O was replaced by drying
down the sample, lyophilization and redissolution in
D2O. NMR spectra were recorded with a Varian UnityI-
NOVA 500 MHz spectrometer at 30 ꢁC. 1H chemical
shifts were referenced relative to external sodium 2,2-di-
methyl-2-silapentane-5-sulfonate (DSS), whereas 31P
chemical shifts were referenced to external phosphoric
acid (H3PO4 85% v/v). 1D proton spectra of samples
in H2O were recorded using the watergate sequence.18
Phase sensitive NOESY spectra32 were recorded with
mixing times of 100 and 180 ms (T = 30 ꢁC). The water-
gate technique was also used for acquiring NOESY
spectra in H2O. TOCSY33 spectra with mixing times of
120 ms were recorded with D2O solution. NOESY and
TOCSY were recorded using the TPPI34 procedure for
quadrature detection. In all 2D experiments the time do-
main data consisted of 2048 complex points in t2 and
400-512 FIDs in t1 dimension. The relaxation delay
was kept at 1.2 s for all experiments. The NMR data
were processed on a SGI Octane workstation using
FELIX 98 software (Accelrys, San Diego, USA).
CD samples of I, II and their natural counterparts
[d(AGGGT)]4 and [d(TAGGGT)]4 were prepared at a
concentration of 2.5 · 10ꢀ5 M, by using the buffer solu-
tion used for NMR experiments: 10 mM KH2PO4,
70 mM KCl, 0.2 mM EDTA, pH 7.0. CD spectra of
all quadruplexes and CD melting curves were registered
on a Jasco 715 circular dichroism spectrophotometer in
a 0.1cm pathlength cuvette. For the CD spectra, the
wavelength was varied from 220 to 340 nm at
5 nm minꢀ1 and the spectra recorded with a response
of 16 s, at 2.0 nm bandwidth and normalized by subtrac-
tion of the background scan with buffer. The tempera-
ture was kept constant at 20 ꢁC with
a
thermoelectrically controlled cell holder (Jasco PTC-
348).
CD melting curves were registered as a function of tem-
perature from 20 to 90 ꢁC at 264 nm for all quadru-
plexes. The CD data were recorded in the same buffer
as used for NMR experiments in a 0.1cm pathlength
cuvette with a scan rate of 1 ꢁC mꢀ1
.
4.3. Structure calculations
The structure calculations were performed with the pro-
gram CYANA27 starting from 200 random conforma-
tions. Upper limit distance constraints for both
exchangeable and nonexchangeable hydrogens were
classified according to the intensity of the cross peaks
Acknowledgements
This work is supported by Italian M.U.R.S.T.
(P.R.I.N. 2002 and 2003) and Regione Campania
(L.41, L.5). The authors are grateful to ÔCentro Ricerche

1044
A. Virgilio et al. / Bioorg. Med. Chem. 13 (2005) 1037–1044
17. Gait, M. J. In Oligonucleotide Synthesis; IRL: Oxford,
1984; pp 27–28.
18. Gait, M. J. In Oligonucleotide Synthesis; IRL: Oxford,
1984; pp 41–45.
Interdipartimentale di Analisi StrumentaleÕ, C.R.I.A.S.,
for supplying NMR facilities.
19. Sklenar, V.; Piotto, M.; Leppik, R.; Saudek, V. J. Magn.
Reson. 1993, 102, 241–245.
References and notes
20. Wang, K. Y.; Mc Curdy, S.; Shea, R. G.; Swaminathan,
S.; Bolton, P. H. Biochemistry 1993, 32, 1899–1904.
21. Wijmenga, S. S.; van Buuren, B. N. M. Prog. Nucl. Mag.
Reson. Spectrosc. 1998, 32, 287–387.
22. Gorenstein, D. G. In Phosphorus-31 NMR: Principles and
Applications; Gorenstein, D. G., Ed.; Academic: New-
York, 1984.
1. Smirnov, I.; Shafer, R. H. Biopolymers 2001, 56, 209–227.
2. Simonsson, T. Biol. Chem. 2001, 382, 621–628.
3. Arthanari, H.; Bolton, P. H. Chem. Biol. 2001, 8, 221–230.
4. For a general recent review see: Davis, J. T. Angew. Chem.,
Int. Ed. 2004, 43, 668–698.
5. Esposito, V.; Randazzo, A.; Piccialli, G.; Petraccone, L.;
Giancola, C.; Mayol, L. Org. Biomol. Chem. 2004, 2, 313–
318.
23. Roongta, V. A.; Jones, C. R.; Gorenstein, D. G.
Biochemistry 1990, 29, 5245–5258.
6. Patel, P. K.; Koti, A. S. R.; Hosur, R. V. Nucleic Acids
Res. 1999, 27, 3836–3843.
7. Gavathiotis, E.; Searle, M. S. Org. Biomol. Chem. 2003, 1,
1650–1656.
8. Searle, M. S.; Williams, H. E. L.; Gallagher, C. T.; Grant,
R. J.; Stevens, M. F. G. Org. Biomol. Chem. 2004, 2, 810–
812.
24. Jin, R.; Gaffney, B. L.; Wang, C.; Jones, R. A.; Breslauer,
K. J. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 8832–
8836.
25. Wuthrich, K. NMR of Proteins and Nucleic Acids; Wiley:
¨
New York, 1986.
26. Smith, F. W.; Feigon, J. Biochemistry 1993, 32, 8683–
8692.
9. Patel, P. K.; Hosur, R. V. Nucleic Acids Res. 1999, 27,
2457–2464.
27. Guntert, P.; Mumenthaler, C.; Wuthrich, K. J. Mol. Biol.
1997, 273, 283–298.
10. Patel, P. K.; Bhavesh, N. S.; Hosur, R. V. Biochem.
Biophys. Res. Commun. 2000, 270, 967–971.
11. Meyer, M.; Schneider, C.; Brandl, M.; Suehnel, J. J. Phys.
Chem. A 2001, 105, 11560–11573, and references cited
therein.
12. Chen, J.; Zhang, L. R.; Min, J. M.; Zhang, L. H. Nucleic
Acids Res. 2002, 30, 3005–3014.
13. For a recent review see: Rimmele, M. Chem. Bio. Chem.
2003, 4, 963–971.
14. Eason, R. G.; Burkhardt, D. M.; Phillips, S. J.; Smith, D.
P.; David, S. S. Nucleic Acids Res. 1996, 24, 890–897.
15. Van Aerschot, A. A.; Mamos, P.; Weyns, N. J.; Ikeda, S.;
De Clercq, E.; Herdewijn, P. A. J. Med. Chem. 1993, 36,
2938–2942.
28. Catalanotti, B.; Galeone, A.; Gomez-Paloma, L.; Mayol,
L.; Pepe, A. Bioorg. Med. Chem. Lett. 2000, 10, 2005–
2009.
29. Esposito, V.; Randazzo, A.; Galeone, A.; Varra, M.;
Mayol, L. Bioorg. Med. Chem. 2004, 12, 1191–1197.
30. Petraccone, L.; Erra, V.; Esposito, V.; Randazzo, A.;
Galeone, A.; Barone, G.; Giancola, C. Biopolymers, in
press.
31. Opalinska, J. B.; Gewirtz, A. M. Nat. Rev. Drug Discov.
2002, 1, 503.
32. Jeener, J.; Meier, B. H.; Bachmann, P.; Ernst, R. R. J.
Chem. Phys. 1979, 71, 4546–4553.
33. Braunschweiler, L.; Ernst, R. R. J. Magn. Reson. 1983, 53,
521–528.
16. Gait, M. J. In Oligonucleotide Synthesis; IRL: Oxford,
1984; pp 25–27.
34. Marion, D.; Wuthrich, K. Biochem. Biophys. Res. Com-
mun. 1983, 113, 874–967.