W-shape nucleic acid (WNA) for selective formation of non-natural anti-parallel triplex including a TA interrupting site

Article abstract of DOI:10.1016/S0040-4039(01)01446-0

Novel nucleoside analogs have been designed for selective formation of anti-parallel triplexes including a TA or a CG interrupting site. The new compounds are constructed of a W-shape bicyclic nucleic acid (WNA) bearing an aromatic ring as a stacking motif and a guanine for the formation of Hoogesteen hydrogen bonds, and are expected to effect triplex stabilization by both stacking and complementary hydrogen bonds. Purine-rich triplex-forming oligodeoxynucleotide (TFO) incorporating the new analog, WNA-7βG, formed a stable triplex with high selectivity to the TA site.

Full text of DOI:10.1016/S0040-4039(01)01446-0

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 6915–6918
W-shape nucleic acid (WNA) for selective formation of
non-natural anti-parallel triplex including a TA interrupting site
Shigeki Sasaki,* Hiroyuki Yamauchi, Fumi Nagatsugi, Ryo Takahashi, Yosuke Taniguchi and
Minoru Maeda
Graduate School of Pharmaceutical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan
Received 13 July 2001; accepted 3 August 2001
Abstract—Novel nucleoside analogs have been designed for selective formation of anti-parallel triplexes including a TA or a CG
interrupting site. The new compounds are constructed of a W-shape bicyclic nucleic acid (WNA) bearing an aromatic ring as a
stacking motif and a guanine for the formation of Hoogesteen hydrogen bonds, and are expected to effect triplex stabilization by
both stacking and complementary hydrogen bonds. Purine-rich triplex-forming oligodeoxynucleotide (TFO) incorporating the
new analog, WNA-7bG, formed a stable triplex with high selectivity to the TA site. © 2001 Elsevier Science Ltd. All rights
reserved.
Since triplex formation within the major groove of
duplex DNA has been proposed as a selective method
for specific inhibition of gene expression at a predeter-
mined sequence, a major concern has been to overcome
the intrinsic limitation that triplexes are formed only
toward homopurine–homopyrimidine sequences of the
duplex.1 In triplexes, pyrimidine oligonucleotides adopt
parallel orientation and purine oligonucleotides tend to
enjoy an anti-parallel orientation. In both cases, pyrim-
idine bases within the homopurine strand of the duplex
inhibit triplex formation; therefore, efforts have been
focused on the development of a non-natural base
structure to stabilize triplexes at such interrupting
sites.2 Recently, interesting nucleoside analogs have
been shown to stabilize triplexes by forming one hydro-
gen bond to the pyrimidine base at the interrupting
site.3 Nevertheless, triplex formation at any DNA
sequence has remained a challenging theme.
In this approach, we focused on an anti-parallel triplex
formed with purine-rich triplex-forming oligonucle-
otides (TFO’s), because the anti-parallel triplexes are
formed under physiological conditions with higher sta-
bility than the parallel ones with pyrimidine TFO’s. A
disadvantageous point of the anti-parallel triplex is that
their formation is inhibited by physiological ionic con-
Figure 1. Schematic structure of the new bicyclic nucleoside analog, W-shape nucleoside analog (WNA); A, and expected role in
the anti-parallel triplex; B.
* Corresponding author. Tel.: +81-92-642-6651; fax: +81-92-642-6654; e-mail: sasaki@phar.kyushu-u.ac.jp
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)01446-0

6916
S. Sasaki et al. / Tetrahedron Letters 42 (2001) 6915–6918
ditions, especially the presence of K+;4 nevertheless,
purine-rich TFO’s have been applied to a number of in
vivo studies.5
We designed a new general structure of nucleoside
analogs bearing an aromatic part for stacking and a
purine base for Hoogesteen hydrogen bonds (Fig. 1A).
The bicyclic structure might contribute to fix conforma-
tion of these two components, and was named the
W-shaped nucleic acid (WNA) after its shape. In the
anti-parallel orientation, the purine base is expected to
form two Hoogesteen hydrogen bonds toward the
target purine base within a pyrimidine strand of the
duplex, and an aromatic may play a role to maintain
the stacking interaction of the TFO continuously
through the new nucleoside analog (Fig. 1B).
Figure 2. Gel shift assay for determination of triplex forma-
tion. Triplex formation was done for 12 h at 22°C in the
buffer containing 20 mM Tris–HCl, 20 mM MgCl2, 2.5 mM
spermidine and 10% sucrose at pH 7.5. Electrophoresis was
done at 10°C with 15% non-denatured polyacrylamide gel.
A–D: 10 nM TFO containing the 32P-labeled one as the tracer
was used. The concentration of duplex was increased from
lane 1 to 12 (nM); 10, 20, 40, 60, 80, 100, 10, 20, 40, 60, 80,
100. The combination of X-Y-Z is shown such as 7bG-CG.
As the first concrete examples, we synthesized bicyclic
nucleoside analogs bearing a benzene ring for the stack-
ing part and guanine for the hydrogen bonding site
(Scheme 1). The compound connecting a guanine at
9-N or 7-N with b-configuration is named WNA-9bG
or WNA-7bG, respectively (4). The synthesis started
with
D-ribose (Scheme 1). Protection of the 2,3-dihy-
droxyl group with acetonide, acetylation of the residual
1,5-hydroxyl groups, selective deacetylation at the 1-
position, oxidation of the 1-hydroxyl group to car-
bonyl, followed by addition of phenyllithium furnished
introduction of the benzene ring (1). Protection of the
5-hydroxyl group and the following allylation at the
1-position gave 2 in a ratio of a:b=7:6, from which the
desired a-isomer was easily isolated. Subsequently, oxi-
dative cleavage of the vinyl group and deprotection of
2,3-O-acetonide spontaneously provided hemiketal,
then the formed hydroxyl groups were acetylated to
afford 3. N-Glycosidation with 2N-isobutyrylguanine
HO
O
HO
O
OH
b
OH
a
O
O
O
HO
OH
1
D-ribose
TBDPSO
TBDPSO
O
d
produced a mixture of isomers in a ratio of
c
9a:7a:9b:7b=17:11:44:28, which were easily isolated by
flash chromatography.6 The stereochemistry of 2 and
WNA-9bG (4) was determined by 1H COSY and
NOESY as illustrated in Scheme 1.
OAc
O
O
O
AcO
3
2
4: R1=TBDPS, R2=Ac
The structures of other isomers were similarly deter-
mined. The phosphoramidite intermediates (5) were
derived from 4 (WNA-9bG and WNA-7bG) by the
conventional method, and applied to an automated
DNA synthesizer. The oligodeoxynucleotide incorpo-
rating the WNA was cleaved from the resin with 30%
NH4OH and the DMTr-protected oligomer was
purified by HPLC with an ODS column. As an ODS
column is not suitable for purification of the non-pro-
tected TFO’s used in this study, deprotection of DMTr
and further purification was done by HPLC with a
Poros column.7 Incorporation of the WNA into the
TFO was confirmed by MALDI-TOF MASS
measurement.
NHCOiPr
NH
(9α: 15%, 7α: 10%,
9β: 38%, 7β: 25%)
H
H
N
O
R1O
H
H
O
e
N
7
N
9
H
5: R1=DMTr,
R2=P(NiPr2)OC2H4CN
H
O
H
OR2
H
Scheme 1. (a) 1. Acetone, H+, 2. Ac2O, pyridine, 3. pipe-
ridine, THF (55%), 4. PCC, CH2Cl2, 5. PhLi, THF (53%); (b)
1. TBDPSCl, triazole, pyridine, 2. allyltrimethylsilane, ZnBr2,
CH3NO2, (44%, a-isomer in a:b=7:6); (c) 1. OsO4, NaIO4,
pyridine (45%), 2. 5% H2SO4, THF, 60°C, 3. Ac2O, pyridine
(57%); (d) 2N-isobutyryllguanine, BSA, TMSOTf, CH3CN,
50°C; (e) 1. TBAF, THF, 2. 0.2 M NaOH, MeOH, 3.
DMTrCl, pyridine, 4. iPr2NEt, CH2Cl2, ClP(NiPr2)-
OC2H4CN, (57%).
The triplex-forming ability of TFO’s incorporating the
new WNA-9bG or 7bG was evaluated by gel shift assay

S. Sasaki et al. / Tetrahedron Letters 42 (2001) 6915–6918
6917
with non-denatured polyacrylamide gel (Fig. 2). As the
purine-rich TFO tends to form disordered-aggregation,8
the gel-shift assay with the labeled-duplex bands, in
which relatively high concentration of the TFO is
needed, did not produce reliable results. On the other
hand, the gel-shift assay with the use of the labeled
TFO at low concentrations gave reproducible results
with better resolution between the single strand and the
triplex.8 The less-mobile bands were confirmed to those
corresponding to triplexes. In this assay, selectivity of
WNA-7bG and 9bG has been clearly demonstrated
(Fig. 2A–D). The WNA-7bG exhibited selectivity to the
TA pair, and WNA-9bG showed selective affinity to
the AT pair. Affinity constants between the TFO and
the duplex were obtained approximately by quantifica-
tion of each band in Fig. 2 as follows; Kd (nM):
7bG-CG, >1000; 7bG-TA, 25, 7bG-GC, >1000; 7bG-
AT, 65, in the series of WNA-7bG, and of 9bG-CG,
>1000; 9bG-TA, >1000; 9bG-GC, >1000; 9bG-AT, 55,
in the series of the 9b-isomer. As the dissociation
constant of Kd=1 nM was obtained in the same gel
shift assay with the natural triplex containing a G-GC
combination at the XYZ site,9 the triplex in the combi-
nation of 7bG-TA is slightly less stable than the natural
triplex with G-GC. Thus, it has been demonstrated that
the new nucleoside analog WNA-7bG exhibited selec-
tive triplex stabilization at a TA interrupting site with
relatively high affinity.
It has been suggested from a molecular modeling by
MM and MD that the guanine base of WNA-7bG may
form two hydrogen bonds with the adenine of the
homopyrimidine strand within the duplex, and that the
benzene ring may face the thymine as shown in Fig. 3A.
In such a triplex structure, benzene was shown to be
stacked by the two purine bases in the TFO to effect
stacking interaction (Fig. 3C).10 Stability and selectivity
obtained with WNA-7bG may be due to the effect of
both the hydrogen bonds and the stacking interaction.
In a similar molecular modeling, the regioisomer
WNA-9bG may form hydrogen bonds with adenine in
the purine strand of the duplex (Fig. 3B). An MD
calculation of the triplex containing WNA-9bG indi-
cated that the guanine base rather than the benzene
ring might be stacked between the other purine bases
(Fig. 3D). The fact that different selectivity was
obtained with the same WNA skeleton has suggested
that the WNA skeleton would be useful as a common
structure for the design of new nucleoside analogs.
In conclusion, we have designed new W-shaped nucleic
acid derivatives (WNA) and discovered that WNA-7bG
is a new candidate for the formation of non-natural
type triplexes with high selectivity and affinity to a TA
interrupting site. As the new structure of WNA would
maintain stacking interaction with the benzene ring, its
structure would provide a useful platform for the
design of new nucleic acid analogs for the formation of
triplexes at any sequence of the duplex.
Acknowledgements
This work was supported by Grant-in-Aid for Scientific
Research from Japan Society for the Promotion of
Science (JSPS, C13672218).
References
1. (a) Moser, H. E.; Dervan, P. B. Science 1987, 238,
645–650; (b) Soyfer, V. N.; Potaman, V. N. Triple-Helical
Nucleic Acids; Springer-Verlag: New York, 1996 and
references cited therein.
2. Selected reviews: (a) Maher, L.; Wold, B.; Dervan, P. B.
Antisense Res. Develop. 1991, 1, 277–81; (b) Sasaki, S. J.
Synth. Org. Chem. 1997, 55, 590–599; (c) Giovannangeli,
C.; He´le`ne, C. Antisense Nucleic Acids Drug Develop.
1997, 7, 413–421; (d) Gowers, D. M.; Fox, K. R. Nucleic
Acids Res. 1999, 27, 1569–1577. Recent articles: (e) Zim-
merman, S. C.; Schmitt, P. J. Am. Chem. Soc. 1995, 117,
10769–10770; (f) Sasaki, S.; Nakashima, S.; Nagatsugi,
F.; Tanaka, Y.; Hisatome, M.; Maeda, M. Tetrahedron
Lett. 1995, 36, 9521–9524; (g) Cho, Y. L.; Jeong, K.-S.
Tetrahedron Lett. 1997, 38, 8337–8340; (h) Lecubin, F.;
Benhida, R.; Fourrey, J.-L.; Sun, J.-S. Tetrahedron Lett.
1999, 40, 8085–8088; (i) Saito, A.; Kuroda, R. Tetra-
hedron Lett. 1999, 40, 4837–4840; (j) Lo´pez de la Paz, M.;
Gonza´lez, C.; Vicent, C. Chem. Commun. 2000, 411–412;
(k) Mertz, E.; Mattei, S.; Zimmerman, S. C. Org. Lett.
2000, 2, 2931–2934; (l) Lengeler, D.; Weisz, K. Tetra-
hedron Lett. 2001, 42, 1479–1481.
Figure 3. Plausible base triplet structures with WNA-7bG (A)
and WNA-9bG (B). The limited region of triplex includine
the WNA was subjected to MD calculation, and only the
TFO is shown with Space-Filling Model in C and D. The
benzene ring (in C) and the guanine (in D) are shown in
green.

6918
S. Sasaki et al. / Tetrahedron Letters 42 (2001) 6915–6918
3. (a) Prevot-Halter, I.; Leumann, C. J. Bioorg. Med. Chem.
14.3, 6.6 Hz), 2.59 (1H, pseudoquintet, J=6.9 Hz), 2.02
(3H, s), 1.29 (6H, d, J=6.9 Hz), 1.01 (9H, s), IR (cm−1
Lett. 1999, 9, 2657–2660; (b) Obika, S.; Hari, Y.; Sugi-
moto, T.; Sekiguchi, M.; Imanishi, T. Tetrahedron Lett.
2000, 41, 8923–8927; (c) Obika, S.; Hari, Y.; Sekiguchi,
M.; Imanishi, T. Angew. Chem., Int. Ed. 2001, 40, 2079–
2081.
,
CHCl3), 1740, 1690, HR-FABMS (m/z): calcd for
C40H46N5O7Si (M+H)+ 736.3175, found 736.3166. WNA-
7bG: mp 122–125°C, 1H NMR l (ppm) 12.2 (1H, bs),
9.25 (1H, bs), 8.01 (1H, s), 7.73–7.63 (6H, m), 7.45–7.26
(9H, m), 6.39 (1H, dd, J=8.2, 5.8 Hz), 5.34 (1H, d,
J=3.6 Hz), 5.11 (1H, dd, J=9.1, 3.6 Hz), 4.29 (1H, ddd,
J=8.8, 3.9, 3.6 Hz), 4.00 (1H, dd, J=11.5, 3.3 Hz), 3.78
(1H, dd, J=11.5, 4.1 Hz), 3.16 (1H, dd, J=13.7, 8.5 Hz),
2.94 (1H, dd, J=13.7, 5.8 Hz), 2.69 (1H, pseudoquintet,
J=6.9 Hz), 1.99 (3H, s), 1.25 (6H, d, J=6.9 Hz), 1.01
(9H, s), IR (cm−1, CHCl3), 1740, 1690, FABMS (m/z):
736.4 (M+H)+.
4. See discussion in: Olivas, W. M.; Maher, III, L. J. Nucleic
Acids Res. 1995, 23, 1936–1941 and references cited
therein.
5. For example: (a) Nakanishi, M.; Weber, K. T.; Guntaka,
R. V. Nucleic Acids Res. 1998, 26, 5218–5222; (b) Vasquz,
K. M.; Narayanan, L.; Glazer, P. M. Science 2000, 290,
530–533 and references cited therein.
6. N7- or N9-Alkylation was determined based on different
chemical shift of guanine, in which the proton of C8 of
the N7-isomer shows lower chemical shift than that of the
N9-isomer: Masko¨y, M.; Bolli, M.; Schweizer, B.; Leu-
mann, C. J. Helv. Chim. Acta 1993, 76, 481–510. The
chemical shift of the proton of C8 of each isomer (ppm
from TMS in CDCl3 at 500 MHz): WNA-7bG; 8.01,
WNA-9bG; 7.82, WNA-7aG; 8.43, WNA-9aG; 8.20.
Selected data for compound 4: WNA-9bG: mp 120–
125°C, 1H NMR l (ppm) 11.9 (1H, bs), 7.88 (1H, bs),
7.82 (1H, s), 7.66–7.60 (6H, m), 7.45–7.27 (9H, m), 6.30
(1H, dd, J=7.4, 6.9 Hz), 5.22 (1H, d, J=3.6 Hz), 5.14
(1H, dd, J=9.1, 3.8 Hz), 4.31 (1H, dt, J=9.1, 3.3 Hz),
4.01 (1H, dd, J=11.5, 3.0 Hz), 3.75 (1H, dd, J=11.5, 3.6
Hz), 3.08 (1H, dd, J=14.3, 7.7 Hz), 3.01 (1H, dd, J=
7. Only a broad peak or many peaks were observed by
HPLC under normal condition with ODS. Poros R3
column was obtained from Perseptive Biosystems.
8. A similar gel-shift assay was reported with the use of
labeled TFO’s: Arimondo, P. B.; Barcelo, F.; Sun, J.-S.;
Maurizot, J.-C.; Garestier, T.; He´le`ne, C. Biochemistry
1998, 37, 16627–16635 and in Ref. 3a.
9. This value is comparable to the dissociation constant of
Kd=0.25 nM measured with 26mer triplex with the
homopurine motif: Durland, R. H.; Rao, T. S.; Bodpudi,
V.; Seth, D. M.; Jayaraman, K.; Revankar, G. R. Nucleic
Acids Res. 1995, 23, 647–653.
10. MM and MD calculation were performed by CAChe
with augmented MM3 parameters.