Conformationally restricted triplex-forming oligonucleotides (TFOs). Binding properties of α-l-LNA and introduction of the N7-glycosylated LNA-guanosine

Article abstract of DOI:10.1016/j.tet.2006.04.016

The method for scaled-up production of α-l-LNA phosphoramidite building blocks containing thymine and 5-methylcytosine nucleobases is described. Binding properties of pyrimidine TFOs modified with α-l-LNA are reported. In contrast to LNA TFOs, the fully m

Full text of DOI:10.1016/j.tet.2006.04.016

Tetrahedron 62 (2006) 5962–5972
Conformationally restricted triplex-forming oligonucleotides
(TFOs). Binding properties of a-L-LNA and introduction of
the N7-glycosylated LNA-guanosine
*
Alexei A. Koshkin
Exiqon A/S, Department of Chemistry, Bygstubben 9, DK-2950 Vedbæk, Denmark
Received 17 January 2006; revised 23 March 2006; accepted 6 April 2006
Available online 3 May 2006
Abstract—The method for scaled-up production of a-L-LNA phosphoramidite building blocks containing thymine and 5-methylcytosine
nucleobases is described. Binding properties of pyrimidine TFOs modified with a-L-LNA are reported. In contrast to LNA TFOs, the fully
modified a-L-LNA forms a stable triplex with a model DNA duplex. Pyrimidine DNA/LNA/a-L-LNA chimeras also efficiently hybridize
with a model DNA duplex in the parallel mode. LNA nucleoside containing unnatural N7-glycosylated guanine (LNA-7G) was synthesized
by a convergent method and incorporated into LNA oligonucleotides. The triplex-forming alternating DNA/LNA oligonucleotides containing
a single LNA-7G modification instead of internal LNA-mC demonstrate improved pH-dependent properties. The single LNA-7G modification
can also discriminatively reduce competitive binding of TFOs to natural nucleic acids in the antiparallel duplex mode.
Ó 2006 Elsevier Ltd. All rights reserved.
1. Introduction
major differences between chromosomes. Moreover, the
largest concentration of such sequences has been found in
regulatory regions, especially in promoter regions, suggest-
ing a tremendous potential for triplex technology.6
The design of molecules that can specifically recognize the
DNA double helix would provide a means to interfere with
gene expression at an early stage, thus giving tools for
a wide range of applications in gene-based biotechnology
and therapeutics. One approach for specific recognition
of predetermined DNA sequences is based on triple helix
formation, which is the phenomenon of third oligonucleotide
strand (triplex-forming oligonucleotide, TFO) binding to
a specific target site in duplex DNA.1–3 It has been found
that TFOs bind in the major groove of oligopyrimidine–
oligopurine sequences in duplex DNA by hydrogen bonding
with purine bases to produce base triplets in a sequence-
specific manner.4 Two main structural motifs have been
established in which a TFO binds either in a parallel or anti-
parallel fashion respective to the target purine strand. Parallel
triplexes contain C+$GC and T$AT triplets5 and have
attracted much attention due to their comparatively high sta-
bility under physiological conditions. The fact that TFOs can
only bind to oligopurine–oligopyrimidine target strands
results in a major limitation to the application of TFOs.
However, the recent in silico analysis of the human genome6
has revealed that regions likely to form triplexes are more
common than predicted by random models. The population
of TFO target sequences is large in all the genome, without
Much research on exploiting TFO binding in vivo has been
reviewed recently.7–9 However, there are still limitations to
the use of TFOs. These concern binding affinity and specific-
ity, uptake into cells and tissues, and in vivo stability. Major
efforts have been devoted to improving TFO characteristics
through chemical modifications. Thus, large variety of mod-
ifications to the TFO backbone has been attempted in order
to increase the general affinity of the TFO to duplex DNA
and its stability in vivo. The structures of the most promising
TFO analogues with modified backbone are depicted in
Figure 1. In the case of peptide nucleic acids (PNA), the
whole phosphodiester backbone is replaced by an uncharged
polyamide backbone.10,11 The triplex-forming feature of
PNAs has been described to lie in recognition of the DNA
duplex via a strand invasion mechanism.12,13 Cellular uptake
of PNAs, however, still raises major problems and needs to
be improved in order to fully exploit their properties.14
N30–P50 phosphoramidate TFOs (Fig. 1) have been demon-
strated to bind strongly to double-stranded DNA.15,16 It
appears that this class of modified oligonucleotides has a
favorable intracellular and tissue distribution, which makes
them good candidates for in vivo applications.16,17 The
replacement of the 30-oxygen atom by nitrogen, made in
N30–P50 phosphoramidates in comparison to DNA, results
in a substantial change of the sugar conformation. The
Keywords: LNA; a-L-LNA; N7-Glycosylated guanine; Triplex-forming
oligonucleotides.
* Tel.: +45 45650429; fax: +45 45661888; e-mail: koshkin@exiqon.com
0040–4020/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2006.04.016

A. A. Koshkin / Tetrahedron 62 (2006) 5962–5972
5963
dsDNA geometry.27 In other words, when incorporated
into the DNA duplex, they acted in a similar way to the fixed
S-type sugar conformers despite their unnatural stereochem-
istry. Based on these findings, we anticipated that TFOs
modified with a-L-LNAwould form, if any, very structurally
distinctive triplexes with dsDNA. The knowledge of partic-
ular properties of a-L-LNA TFOs can provide insights that
are useful for further progress in rational design of molecules
capable of specifically recognizing the dsDNA helix.
Base
Base
O
O
O
HN
N
HN
P
-
O
O
O
O
N3'-P5' phosphoramidites
PNA
Base
Base
O
O
O
Base
O
O
O
O
O
O
O
To improve the binding properties of LNA TFOs at higher
pH, we additionally introduce herein a new LNA nucleoside
containing N7-glycosylated guanine as a nucleobase. It has
been shown that N7-deoxyguanosine (d7G) mimics proton-
ated cytosine and specifically binds GC base pairs within
a parallel triple helix motif.28 The stabilities of d7G-contain-
ing triplexes are independent of pH but strongly dependent
on the sequence context. The TFOs containing d7G are par-
ticularly useful for targeting continuous GC base pairs.29
-
O
O
P
O
O
-
-
P
O
O
P
O
O
LNA
ENA
Figure 1. Molecular structures of selected TFOs with modified backbone.
resulting N-type conformation18 may give an explanation for
enhanced triplex-forming ability.
Modifications made to the ribose moiety in TFOs have also
been able to generate derivatives with advantageous proper-
ties. Thus, locked nucleic acid (LNA) and 20-O,40-C-ethyl-
ene-briged nucleic acid (ENA) are RNA derivatives in
which the ribose unit is constrained by a methylene or an
ethylene linkage between the 20-oxygen and the 40-carbon
(Fig. 1). This bridge reduces the conformational flexibility
of the ribose and thus results in a locked 30-endo (N-type)
sugar puckering.19–21 Along with N30–P50 phosphorami-
dates, the outstanding results on binding properties of LNA
and ENA have clearly demonstrated that such restriction of
the sugar conformation is a very promising approach for
TFOs design.21,22 Accordingly, TFOs consisting of alternat-
ing LNA and DNA residues have demonstrated significantly
enhanced affinity to dsDNAwith respect to the isosequential
DNATFOs.21 The NMR structure of an LNA$dsDNA triplex
has been determined recently.23 The triplex has a regular
seamless structure despite the fact that the TFO strand is
composed of DNA and LNA monomers with different sugar
puckers. The structure of the DNA duplex is changed to fit
TFO accommodation, and it adopts a geometry intermediate
between A and B types.
2. Results and discussion
The first synthesis of a-L-LNA 30-O-phosphoramidites con-
taining thymine, 5-methylcytosine, and adenine nucleobases
was reported by Wengel.25,26 However, to make a-L-LNA
practically available, a significantly improved and more
efficient route to a-L-LNA monomers has been developed
(Scheme 1). Analogous to commercially available LNA
O
MsO
MsO
O
i
O
O
OBn
OH
O
O
O
O
2
1
O
NH
O
MsO
N
ii
MsO
MsO
O
OBn
iv
O
OBn
OAc
MsO
OAc
OR
3
4:R = Ac
5:R = H
iii
It has been found that ENA TFOs (Fig. 1) form stable tri-
plexes with dsDNA similar to those partially modified
with LNA nucleosides.22 Furthermore, in contrast to LNA,
fully modified ENA oligonucleotides have high triplex for-
mation ability. Additionally, ENAs are generally more
nuclease resistant than LNAs and, as a result, have greater
potential for in vivo applications.20,22 However, the avail-
ability of ENA is still comparatively limited as its synthesis,
reported hitherto,20 is significantly more resource consum-
ing than that for LNA.24
O
O
NH
NH
O
O
N
N
O
O
O
vi
v
HO
MsO
BnO
O
BnO
6
7
O
N
O
NH
NH
O
O
N
Bearing in mind the achievements in constructing conforma-
tionally constrained TFOs, we report herein the results of our
investigations into the binding properties of TFOs modified
with a-L-ribo-configured locked nucleic acid (a-L-LNA;
Fig. 1). The first synthesis and some properties of a-L-
LNA have been recently reported.25,26 The binding studies
have demonstrated high affinity Watson–Crick hybridization
between a-L-LNA and single stranded DNA oligonucleo-
tides.25 Interestingly, NMR studies have revealed that, in
general, a-L-LNA nucleosides do not perturb native B-like
O
O
vii
DMTO
HO
O
O
O
HO
9a
O
P
8a
NC
N(iPr)2
Scheme 1. Reagentsandconditions:(i)Ac2O,concdH2SO4, AcOH,91%;(ii)
thymine, BSA, trimethylsilyl triflate, MeCN, quant.; (iii) HCl, MeOH, 99%;
(iv) (a) MsCl, pyridine; (b) NaOH, H2O, quant.; (v) (a) NaOBz, DMSO; (b)
NaOH, H2O, 74%; (vi) 20% Pd(OH)2/C, HCO2NH4, MeOH, 1,4-dioxane,
reflux, 93%; (vii) (a) DMT-Cl, pyridine; (b) 2-cyanoethyl-N,N,N0,N0 -tetra-
isopropylphosphordiamidite, 4,5-dicyanoimidazole, CH2Cl2, 92%, Ref. 31.

5964
A. A. Koshkin / Tetrahedron 62 (2006) 5962–5972
nucleosides,24 a convergent strategy was chosen for the pro-
duction of a-L-LNA. A starting material for the synthesis
of a-L-LNA-T phosphoramidite 9a (Scheme 1), the known
4-C-branched furanose 2,25 was obtained from commercial
1,2:5,6-di-O-isopropylidene-a-D-glucofuranose 1 as de-
scribed earlier. However, NMR data reported by Wengel25
for compound 2 was significantly different from that re-
ported here. Coupling sugar 3 was subsequently obtained
in 91% yield after one-pot acetolysis and acetylation of 2.
The anomeric mixture 3 was then used as a glycosyl donor
N
H
O
+
N
N
H
O
H
N
H
H
N
N
O
N
H
N
N
N
N
H
H
for Vorbrugen-type30 coupling reaction with silylated thy-
¨
mC+ - GC
mine to give a 40-C-branched nucleoside 4 after standard
aqueous work-up. Transformation of compound 4 into a-L-
ribo-configured LNA-T derivative 6 required epimerization
at the 20-position. To make this inversion, the 20-O-acetyl
group was first replaced with mesyl and the resulting com-
pound was then treated with aqueous sodium hydroxide.
Accordingly, treatment of 4 with methanolic hydrochloride
resulted in quantitative removal of the acetyl group to give
intermediate 5. In the same flask, the solvents were replaced
by anhydrous pyridine and 5 was consecutively treated with
mesyl chloride and sodium hydroxide to give protected a-L-
LNA-T nucleoside 6. Apparently, the transformation of 5
into 6 proceeded via the 2,20-anhydro intermediate, which
could be consecutively formed, hydrolyzed, and subjected
to ring-closure reaction in the same reaction mixture.26
Indeed, corresponding intermediates were detected by RP-
HPLC analyses of the reaction mixture (data not shown).
This reaction cascade proceeded very smoothly and was
completed overnight in nearly quantitative yield. The purity
of nucleoside 6 was about 96% (RP-HPLC, integration at
260 nm) after aqueous work-up and the compound was
used further without additional purification. Instead, 6 was
converted to crystalline compound 7 after deprotection of
the 50-hydroxy group. Thus, nucleoside 6 was reacted with
sodium benzoate in hot DMSO and the obtained 50-benzoate
was hydrolyzed by the excess of alkali to give 7 in 67% yield
(from 2) after crystallization from ethyl acetate. As ex-
pected, catalytic removal of the 30-O-benzyl group from 7
easily afforded diol 8a in 93% yield after crystallization
from water. Finally, a-L-LNA-T phosphoramidite building
block 9a was derived from 8a by means of the standard
methods.26,31 The synthetic route to phosphoramidite 9a de-
scribed above is significantly superior to the earlier reported
procedure,25,26 which makes the commercial production of
a-L-LNA-T nucleoside very feasible.
N
N
N
H
N
N
O
H
H
N
H
N
H
O
N
N
O
N
H
N
N
N
H
H
7G - GC
Figure 2. Schematic comparison of the base triplets involving 5-methyl-
cytosine (mC) and N7-glycosylated guanine (7G).28,29
derived by different methods from the corresponding thy-
min-1-yl diols 8a,b. The recent developments for a universal
(both for LNA and a-L-LNA) scale-up synthesis of 13a,b are
depicted in Scheme 2. The method has several advanced
points, which make it considerably more useful in compari-
son to already reported procedures. Thus, starting with a-L-
LNA-T diol 8a, fully carbohydride protected thymin-1-yl
derivative 10a was easily produced after one-pot consecutive
treatment of 8a with DMT-chloride and acetic anhydride in
anhydrous pyridine. The introduction of a hydrophobic
DMT group at this stage significantly facilitated handling
of nucleosides in later steps of the synthesis. The 30-O-acetyl
O
Seria a: α-L-ribo-LNA
Seria b: β-D-ribo-LNA
NH
O
N
DMTO
O
i
ii
8a,b
O
AcO
10a,b
An important limitation of the triplex technology is that par-
allel triplexes require conditions of low pH, which are neces-
sary for protonation of the nitrogen at position 3 of cytosines
in TFOs (Fig. 2). To address this problem, several derivations
of cytosine analogues have been introduced that provide less
pH-dependent triplex formation. Thus, 5-methylcytosine
(mC) replacement for cytosine in the DNA TFOs gave
more stable triplexes at neutral pH, because of the increase
in pKa for 5-methylcytosine (4.3) compared with cytosine
(4.15).32 A similar effect has recently been demonstrated
for LNA and ENA modified TFOs.22,33 For this reason, we
concentrated on studying a-L-LNA-modified TFOs contain-
ing thymine and 5-methylcytosine nucleobases.
Bz
N
N
HN
N
N
N
N
O
DMTO
O
O
iii
N
DMTO
O
O
RO
O
AcO
12a,b: R = H
13a,b: R =
iv
11a,b
P(O(CH2)2CN)N(iPr)2
Scheme 2. Reagents and conditions: (i) (a) DMT-Cl, pyridine; (b) Ac2O,
pyridine, 90% (10a) and 89% (10b); (ii) 1,2,4-triazole, POCl3, DIPEA,
MeCN, 97% (11a) and 90% (11b); (iii) (a) concd NH4OH, MeCN; (b)
Bz2O, pyridine; (c) NaOH, pyridine, MeOH, H2O, 81% (12a) and 76%
(12b); (iv) 2-cyanoethyl-N,N,N0,N0-tetraisopropylphosphordiamidite, 4,5-
dicyanoimidazole, CH2Cl2, 95%, Ref. 31.
It has already been reported25,34 that both a-L-LNA-mC and
LNA-mC phosphoramidites 13a,b (Scheme 2) could be

A. A. Koshkin / Tetrahedron 62 (2006) 5962–5972
5965
protection group was selected due to the simplicity of its
introduction and removal, avoiding undesired side reactions.
Second, nucleoside 10a was almost quantitatively converted
to a stable 4-(1,2,4-triazol-1-yl) derivative 11a by the method
of Reese.35 Accordingly, 10a was treated with freshly dis-
tilled phosphoryl chloride in the presence of 1,2,4-triazole
and excess of a base to give 11a after aqueous work-up.
Compound 11a can serve as a valuable intermediate in the
synthesis of different 4-substituted a-L-LNA-pyrimidine
nucleosides as it can be easily handled and stored. Thus, a
simple procedure was applied to convert 11a to 4-N-benzoyl-
cytosine-1-yl derivative 12a (Scheme 2). After compound
11a was allowed to react with ammonia overnight, the sol-
vents were thoroughly removed and the residue was treated
with benzoic anhydride in anhydrous pyridine to give a mix-
ture of mono- and di-benzoylated (MALDI-MS analysis)
nucleosides. Addition of aqueous sodium hydroxide to the
same reaction mixture led to very selective hydrolysis of
undesired 30-benzoate leaving the intact nucleoside 12a.
The nucleoside 12a was isolated in 81% yield by column
chromatography and phosphitylated using the method used
for the production of 9a.31 The above synthetic procedures
were also successfully applied for the preparation of phos-
phoramidite 13b (Scheme 2), which established them as a
reliable universal method for the production of both LNA
and a-L-LNA monomers 13a,b.
triflate as a catalyst in 1,2-dichloromethane; reflux), N7-gly-
cosylated compound 15 was formed as a minor product in
approximately 10% yield along with N9-isomeric counter-
part. However, when conditions favoring the formation of
N7-isomer28,29 were applied, 15 was produced as the sole
product after chromatographic purification albeit in moder-
ate 34% yield (Scheme 3). The correct structure of the ob-
tained isomer was confirmed by comparison of NMR data
for 15 with that published for N7- and N9-glycosylated gua-
nines.36,37 Thus, the most characteristic signal corresponding
to C5 positioned at 110.2 ppm in the 13C NMR spectrum of
15. The conversion of nucleoside 15 to diol 19 was performed
generally following the synthetic route developed for N9-
glycosylated LNA-G.24 Accordingly, the ring-closure reac-
tion promoted in 15 by sodium hydroxide afforded protected
LNA-7G nucleoside 16. The 50-methylsulfonate of 16 was
displaced by benzoate to give 17, which in turn was hydro-
lyzed to furnish nucleoside 18. To keep 2-N-isobutyryl intact,
debenzylation of 18 was carried out on Pd/C using formic
acid as a hydrogen supply24 to give diol 19. Monomeric
unit 20, finally derived from 19, was successfully applied
for automated oligonucleotide synthesis using the phosphor-
amidite method.38 However, the 2-N-isobutyryl group of the
LNA-7G mononucleotide turned out to be considerably more
stable in comparison to all the other commonly used protec-
tions. Therefore, modifications of deprotection procedure
were required for LNA-7G containing oligonucleotide after
the synthesis (vide infra). To facilitate the synthesis of oligo-
nucleotides containing multiple LNA-7G modifications,
(dimethylamino)methylidene39 protected phosphoramidite
23 was ultimately produced. Thus, the fully deprotected
nucleoside 21 was furnished in 86% yield after consecutive
catalytic hydrogenation and treatment with ammonium
hydroxide of 18. 2-N-(Dimethylamino)methylidene and
50-O-DMT groups were introduced into 21 by conventional
The unnatural N7-glycosylated deoxyguanosine was intro-
duced as another promising replacement for cytosine in
TFOs28 (Fig. 2). To investigate compatibility of this approach
with LNA technology, TFOs modified with LNA-7G were
synthesized. The synthesis of phosphoramidite monomers
20 and 23 was first attempted (Scheme 3). The coupling of
silylated 2-N-isobutyrylguanine with glycosyl donor 14 was
already reported.24 In the reaction conditions (trimethylsilyl
H
N
H
N
N
O
N
O
N
N
N
MsO
MsO
O
NH
O
v
N
H
N
O
RO
MsO
MsO
O
O
OAc
OAc
i
ii
BnO
16: R = Ms
17: R = Bz
18: R = H
BnO
iii
iv
O
BnO
OAc
14
15
vii
N
N
N
NH2
NH
HO
HO
O
H
N
N
O
N
N
O
21
NH
O
R1O
R2O
O
O
viii
O
19:
R1 = R2 = H
R1 = DMT; R2
vi
20:
=
P(O(CH2)2 CN)N(iPr)2
N
N
N
N
NH
DMTO
RO
O
O
O
22: R = H
23: R = P(O(CH2)2CN)N(iPr)2
ix
Scheme 3. Reagents and conditions: (i) 2-N-isobutyrylguanine, BSA, SnCl4, MeCN, 34%; (ii) NaOH,1,4-dioxane, H2O, 89%; (iii) NaOBz, DMF, 93%; (iv)
NaOH, pyridine, EtOH, H2O, 85%; (v) 10% Pd/C, HCO2H, MeOH, 81%; (vi) (a) DMT-Cl, pyridine; (b) 2-cyanoethyl-N,N,N0,N0-tetraisopropylphosphor-
diamidite, 4,5-dicyanoimidazole, CH2Cl2, 82%; (vii) (a) 10 % Pd/C, HCO2NH4, MeOH; (b) NH3, MeOH, 86%; (viii) (a) N,N-dimethylformamide dimethyl
acetal, DMF; (b) DMT-Cl, pyridine, 89%; (ix) 2-cyanoethyl-N,N,N0,N0-tetraisopropylphosphordiamidite, 4,5-dicyanoimidazole, CH2Cl2, MeCN, 92%.

5966
A. A. Koshkin / Tetrahedron 62 (2006) 5962–5972
methods36 and compound 22 was isolated in 89% yield after
column chromatography. Phosphoramidite 23 was derived
from 22 in 92% yield generally following the procedure
developed for other LNA nucleosides.31
group. This signal completely disappeared after additional
treatment of the reaction mixture with aqueous methylamine
for 2 h at 60 ꢀC. Analogous data demonstrating the increased
stability of the 2-N-isobutyryl group on N7-glycosylated
guanine have been previously reported by Dervan.29
Along with commercially available DNA and LNA phos-
phoramidites, the phosphoramidite building blocks 9a,
13a, and 23 were used for automated oligonucleotide syn-
thesis to produce the LNA oligonucleotides depicted in
Tables 1 and 2. The standard protocols of DNA synthesis
were applied for all LNA and a-L-LNA amidites, except
for coupling (extended to 500 s) and oxidation cycles (ex-
tended to 30 s). After synthesis, the oligonucleotides were
deprotected by treatment with concentrated ammonium
hydroxide for 6 h at 60 ꢀC and purified by RP-HPLC. To
synthesize LNA-7G modified oligonucleotide 46 (Table 2),
phosphoramidite 20 was used instead of 23. The above con-
ditions for the automated synthesis were also successfully
applied in this case. However, the deprotection procedure
was modified in order to complete the removal of the 2-N-
isobutyryl group from LNA-7G mononucleotide. Thus,
ammonium hydroxide treatment used during the synthesis of
all the other oligonucleotides was first applied. MALDI-MS
analysis of the mixture, however, revealed the presence of
a product (a signal of approximately equal intensity as for
the target product) containing intact isobutyryl protection
Triplex-forming properties of a-L-LNA-modified oligo-
nucleotides were assessed against target DNA duplex con-
taining an oligopurine–oligopyrimidine sequence of 14 bp
in the central core and presented in Table 1 as melting
temperatures of the corresponding triplexes. The recently
reported results on binding properties of LNA and ENA
TFOs22,33 have been obtained by studying the same oligo-
nucleotide model and can be directly compared with the
data reported herein. Thus, in good agreement with the data
reported by Imanishi,33 the sequences of alternating LNA
and DNA nucleotides 24 and 25 (Table 1, entries 1 and 2)
formed stable triplexes with the complementary duplex
40$41 in the pH range 6.5–7.5. It was found that stability
of triplex 25:40$41 slightly exceeds stability of 24:40$41
at pH 6.5. This observation clearly parallels the rules derived
recently for designing LNA-modified TFOs.40 Accordingly,
the sequence effect and the differential stabilization by
LNA-T and LNA-mC nucleotides have been pointed out.40
For LNA-modified TFOs, maximizing the number of
LNA50–30DNA steps (e.g., by the use of an LNA nucleotide
Table 1. Melting temperaturesa of the perfectly matched triplexes (ꢁ1 ꢀC) and duplexes (ꢁ0.5 ꢀC) containing b-D- and a-L-LNA nucleotides
Entry
Oligo structure (50–30)
(compound number)
Parallel triplexb at diff. pH
Antiparallel
duplexc
pH 6.5
pH 7.0
pH 7.5
1
2
3
4
5
6
7
8
tCtCtCtCcCtTtT (24)
TcTcTcTcCcTtTt (25)
tXtXtXtXcXtYtY (26)
56.5
58.5
45.0
36.0
20.0
57.5
47.0
37.0
52.0
nt
43.0
41.5
30.5
25.5
nt
43.0
29.0
21.0
38.0
nt
32.0
27.5
ntd
nt
nt
31.5
nt
nt
27.0
nt
nt
nt
71.8
71.7
68.4
90.5
83.4
70.6
68.4
66.5
69.3
nt
45.7
41.5
36.6
59.6
59.5
56.5
YXYXYXYXXXYYYY (27)
YCYCYCYCXCYTYT (28)
tCtCtCtXcCtTtT (29)
TcTcTcYcCcTtTt (30)
TcYcTcYcCcYtTt (31)
tCtXtCtXcCtYtT (32)
t7Gt7Gt7Gt7Gc7GtTtT (33)
t7GtCt7GtCc7GtTtT (34)
t7GtCt7Gtc7GctTtT (35)
t7Gtct7Gtc7GcttTt (36)
tCtCtCt7GcCtTtT (37)
TcTcTcTc7GcTtTt (38)
TcYcTcYc7GcYtTt (39)
9
10
11
12
13
14
15
16
nt
nt
41.5
38.0
57.0
>60
41.5
34.0
28.0
44.5
45.5
27.5
nt
34.5
31.0
nt
a
The melting temperatures (Tm values) were obtained as maxima of the first derivatives of the corresponding melting curves (optical density at 260 nm versus
temperature). Presented as the mean of three measurements. LNA residues are capitalized. 7G¼b-D-LNA-G (N7-isomer); X¼a-L-LNA-mC; Y¼a-L-LNA-T. C,
c¼5-methyl-cytosines.
b
c
d
The model complementary duplex (1 mM) was obtained by mixing purine strand 40: 50-gaacaaacagagagagggaaaatcccccta and pyrimidine strand 41: 50-
tagggggattttccctctctctgtttgttc in 0.01 M Na-phosphate buffer (pH 6.5, 7.0, or 7.5) containing 0.1 M NaCl and 1 mM EDTA. Concentration of TFOs: 1 mM.
Tms of the antiparallel duplexes were measured against complementary tetradecadeoxynucleotide 42: 50-aaaagggagagaga. Concentration of duplexes: 1 mM;
the same buffer as for triplex study (pH 7.4).
nt—No cooperative transition observed over 20 ꢀC.
Table 2. Melting temperatures (ꢁ1 ꢀC) of the perfectly matched and singly mismatched LNA$DNA duplexesa
Entry
Oligonucleotide structure (50–30)
Tms (ꢁ1 ꢀC) of the duplexes with complementary deoxynucleotides
30-ctgaatcc (47)
30-ctgtatcc (48)
30-ctggatcc (49)
30-ctgcatcc (50)
1
2
3
4
GACTTAGG (43)
GACATAGG (44)
GACGTAGG (45)
GAC7GTAGG (46)
61
38
32
nt
28
62
55
32
36
43
41
37
24
41
71
31
a
Concentration of duplexes: 2.5 mM. All other experimental conditions and abbreviations as in Table 1.

A. A. Koshkin / Tetrahedron 62 (2006) 5962–5972
5967
at 50-end as for 25) and maximizing the use of LNA-T is
thermodynamically preferable for triplex formation. On
the other hand, it has been reported41 that triplex formation
in the pyrimidine motif proceeds from the 50-end to the
30-end according to the nucleation–zipping mechanism.
Association rate is thus governed by the composition of
base triplets on the 50-end of the triplex, which are signifi-
cantly different for 24 and 25. According to this model,
oligonucleotide 25 containing two LNA-T nucleotides at
the 50-end might presumably bind to duplex 40$41 slightly
faster than its counterpart 24.
by pairing mode as suggested by Dervan28,29 (Fig. 2). The
most stable triplexes were detected for TFOs 37 and 38
(entries 14 and 15) containing a single LNA-7G replacement
for LNA-mC (entries 14 and 15). Accordingly, at pH 6.5, 37
bound to the model duplex with the same efficacy as parent
24. However, the melting temperature of 37:40$41 was
detectably higher than Tm of 24:40$41 at pH 7.5 (compare
entries 1 and 14). The melting temperature of the triplex
formed by 38 (entry 15) at pH 6.5 was close to Tm of duplex
40$41 and could not be measured by the used method.
Nevertheless, remarkable Tms¼45.5 and 31.0 ꢀC were
measured for 38:40$41 at pH 7.0 and 7.5, respectively.
Noteworthy, a lower pH sensitivity of Tm for triplex formed
by 37 in comparison to 38 (Table 1, entries 14 and 15) can be
explained by the higher content of LNA-mC nucleotides as
was already pointed out for TFOs 24 and 25 (vide supra).
The stabilizing effect of the single LNA-7G modification
was also shown in the case of DNA/LNA/a-L-LNA chimera
39 (compare entries 8 and 16).
The thermostability of triplex 24:40$41, however, demon-
strated a notably lower pH dependence than stability of
25:40$41 (Table 1, entries 1 and 2). The higher apparent
pKa of the LNA-modified 5-methylcytidines in the
dsDNA:LNA triplex has been attributed to lowered base-
pair dissociation rates.23 Therefore, due to the increased
number of LNA-mC nucleotides in 24 compared with iso-
sequential 25, the stability of triplex 24:40$41 at pH 7.5 is
already 4.5 ꢀC higher.
Binding properties of LNA-7G were further studied in the an-
tiparallel duplex context. It was shown that single LNA-7G
replacement for the internal LNA-mC nucleotide resulted
in significant destabilization(by 10–12 ꢀC) of the antiparallel
duplex with complementary oligodeoxynucleotide 42 (Table
1, entries 14–16). Furthermore, an additive destabilizing
effect of multiple LNA-7G modifications was shown for the
complementary duplexes involving DNA/LNA chimerical
oligonucleotides 33–36 (entries 10–13). Incompatibility of
LNA-7G with Watson–Crick duplex geometry was also
demonstrated for a fully modified LNA oligonucleotide of
random sequence. When incorporated into LNA octanucleo-
tide 46, LNA-7G was unable to efficiently pair with any nat-
ural DNA base (Table 2, entry 4). The melting temperatures
measured for duplexes formed by 46 against 47–50 werevery
low, close to the Tms of the most unstable mismatched
duplexes formed by unmodified 43–45 (entries 1–3). Thus,
replacement of LNA-G with unnatural LNA-7G led to desta-
bilization of the complementary duplex against 50 by 40 ꢀC
(compare duplexes 45$50 and 46$50). The discriminating
pairing properties of the single LNA-7G modification can
be very useful for in vivo application of alternating DNA/
LNA TFOs in order to minimize undesired competitive
biding to nucleic acids in the antiparallel duplex mode.
The triplex-forming ability of a-L-LNA-modified TFOs
turned out to be significantly reduced in comparison to LNA
TFOs. Thus, replacement of LNA nucleotides in 24 by
analogous a-L-LNA monomers resulted in a Tm decrease of
the corresponding triplex by 11.5 ꢀC at pH 6.5 (triplex
26:40$41; entry 3). The fully modified a-L-LNA oligonu-
cleotide 27 (Table 1, entry 4) was able to bind the comple-
mentary DNA duplex 40$41 at pH 6.5 and 7.0, albeit with
reduced affinity. These data are particularly interesting, con-
sidering the fact that no stable triplexes have been detected
for fully modified LNATFOs.33
The attempt to construct a potent TFO consisting of altering
LNA and a-L-LNA units failed as oligonucleotide 28 (Table
1, entry 5) did not form a stable triplex with 40$41. However,
LNA TFO 24 could tolerate a single a-L-LNA replacement
for LNA-mC monomer. Thus, a DNA/LNA/a-L-LNA
chimera 29 demonstrated the same binding properties as
the parent 24 (compare entries 1 and 6). In sharp contrast to
29, a single a-L-LNA-T replacement made for LNA-T in 25
resulted in significant destabilization of the triplex demon-
strating differential pairing abilities of a-L-LNA-T and a-L-
LNA-mC units in TFOs. This phenomenon was additionally
confirmed by triple a-L-LNA-T replacements for LNA-T.
The Tm value for the triplex formed by chimera 31 (entry 8)
containing three a-L-LNA-T units is already 21.5 ꢀC lower
than Tm of 25:40$41. At the same time, oligonucleotide 32
containing two a-L-LNA-mC and only a single a-L-LNA-T
modification bound to the model duplex with significantly
higher potency (compare entries 8 and 9; Table 1).
3. Conclusions
In conclusion, we developed an improved method for scaled-
up synthesis of the building blocks for oligonucleotides
containing a-L-LNA thymine and 5-methylcytosine units as
well as a method for synthesis of the novel N7-glycosylated
LNA-guanosine monomer. The compounds were success-
fully incorporated into triplex-forming oligonucleotides
and their binding properties were systematically studied.
Pyrimidine DNA/a-L-LNA chimeras demonstrate a reduced
binding ability compared with the analogous DNA/LNA
TFOs. However, in contrast to LNATFOs, the fully modified
a-L-LNA form a stable triplex with the model DNA duplex.
The alternating DNA/LNA TFOs containing a single
LNA-7G replacement for internal LNA-mC demonstrate
improved pH-dependent properties and can also be used to
discriminatively reduce competitive binding of TFOs to
Binding properties of TFOs modified with LNA-7G are pre-
sented in entries 10–16 (Table 1). Thus, TFOs 33 and 34 con-
taining five or three LNA-7G residues instead of LNA-mC
did not bind to 40$41 at any pH. However, a further experi-
ment demonstrated that DNA/LNA chimeras modified with
three LNA-7G could form triplexes when LNA-7G units
were separated by two DNA mononucleotides (TFOs 35
and 36; entries 12 and 13). Importantly, the improved pH
dependence observed for triplex stability in this case indi-
cates that LNA-7G units were involved in triplex formation

5968
A. A. Koshkin / Tetrahedron 62 (2006) 5962–5972
nucleic acids in the antiparallel duplex mode. Our findings
may be useful for further developments in the field of triplex
technology.
4.1.3. 1-(2-O-Acetyl-3-O-benzyl-4-C-methanesulfonoxy-
methyl-5-O-methanesulfonyl-a-L-threo-pentofurano-
syl)thymine (4). A mixture of furanose 3 (72.4 g, 0.14 mol)
and thymine (22.5 g, 0.18 mol) was suspended in anhydrous
MeCN (300 mL) and N,O-bis(trimethylsilyl)acetamide
(68 mL, 0.27 mol) was added. After refluxing for 1.5 h,
a clear solution was obtained and the mixture was cooled to
room temperature. Trimethylsilyl trifluoromethanesulfonate
(50 mL, 0.28 mol)was addedandrefluxingwas continuedfor
6 h. The reaction mixture was cooled to room temperature,
neutralized with saturated NaHCO3 (400 mL), and washed
with CH2Cl2 (2ꢂ300 mL). The combined organic layers
were washed with saturated NaHCO3 (2ꢂ300 mL), dried
(Na2SO4), and concentrated under reduced pressure to give
compound 4 (84.0 g, 103%) as a brownish solid residue,
which was used for the next step without additional purifica-
tion. RP-HPLC: Rt¼9.85 min, A290/260¼0.12, purity 97%
(integration at 260 nm). Analytical samplewas obtained after
silica gel column chromatography (1–3% v/v MeOH/
4. Experimental
4.1. General
All chemicals were obtained from commercial suppliers and
were used without additional purification. An atmosphere of
nitrogen was applied when reactions were conducted in
anhydrous solvents. Column chromatography was per-
formed using Silica gel 60 (0.063–0.200 mm) from Merck
and HPLC using a column Prep Nova-Pak HR Silica 60
(6 mm, 30ꢂ300 mm). Analytical reverse-phase chromato-
graphy (RP-HPLC) was performed on XTerra RP18 5 mm
(3.9ꢂ150 mm) column using Delivery System and Dual l
Absorbance Detector (all from Waters) in the following elu-
ent system: linear gradient 20–100% v/v of MeCN in 0.05 M
aqueous triethylammonium acetate (pH 7.2) during 15 min.
Flow rate: 1 mL/min. Rt—retention time. A290/260—the ratio
between absorbance at 290 and 260 mn measured at maxi-
mum of the corresponding peak. All 1H and 31P NMR spectra
were recorded at 400 and 161.9 MHz, respectively. 13C NMR
spectra for 10a, 10b, 16, and 18 were recorded at 62.9 MHz
and at 100.6 MHz for all other compounds. The values for
chemical shifts are reported in parts per million relativeto tet-
1
CH2Cl2) as a white solid foam. H NMR (CDCl3): d 9.34
(br s, 1H), 7.31–7.42 (m, 6H), 6.28 (d, J¼3.5 Hz, 1H), 5.36
(dd, J¼3.5 and 2.8 Hz, 1H), 4.78 (d, J¼11.4 Hz, 1H), 4.66
(d, J¼11.4 Hz, 1H), 4.55 (d, J¼11.2 Hz, 1H), 4.42 (d, J¼
11.2 Hz, 1H), 4.35 (d, J¼10.8 Hz, 1H), 4.26 (d, J¼10.8 Hz,
1H), 4.25 (d, J¼2.8 Hz, 1H), 3.05 (s, 3H), 3.02 (s, 3H),
2.14 (s, 3H), 1.80 (d, J¼1.1 Hz, 3H). 13C NMR (CDCl3):
d 169.6, 163.3, 150.3, 135.8, 134.9, 128.6, 128.5, 128.3,
112.3, 87.5, 84.6, 81.1, 79.3, 73.0, 66.9, 65.4, 37.6, 20.5,
12.2. MALDI-TOF MS: m/z 599.1 (M+Na+). Anal. Calcd
for C22H28N2O12S2: C, 45.83; H, 4.89; N, 4.86. Found: C,
45.97; H, 4.86; N, 4.80.
1
ramethylsilane (TMS) as internal standard for H and 13C,
and relative to 85% H3PO4 as external standard for 31P.
MALDI-TOF mass spectra of nucleosides were recorded
on a Voyager-DE PRO spectrometer in positive ion mode
using 2,5-dihydroxybenzoic acid as the matrix. Melting
4.1.4. 1-(3-O-Benzyl-5-O-methanesulfonyl-2-O,4-C-
methylene-a-L-ribofuranosyl)thymine (6). Compound 4
(83.7 g, 0.145 mol) was dissolved in 1 M HCl in MeOH
(450 mL) and stirred for 24 h. The solvents were removed
under reduced pressure to give a solid residue. The residue
was co-evaporated with anhydrous MeCN (250 mL) and
dried in vacuo overnight to give intermediate 5 (77.2 g,
99.6 %). RP-HPLC: Rt¼8.82 min, A290/260¼0.155. The in-
termediate was dissolved in anhydrous pyridine (200 mL),
cooled in an ice bath, and MsCl (14.5 mL, 0.183 mol) was
added under intensive stirring. The mixture was allowed to
warm to room temperature and stirred overnight, whereupon
1 M NaOH (1.5 L) was added. After stirring for 18 h, AcOH
(90 mL) was added and the mixture was washed with
CH2Cl2 (2ꢂ400 mL). The combined organic layers were
washed with saturated NaHCO3, dried (Na2SO4), and con-
centrated under reduced pressure to give an oily residue.
The residue was co-evaporated with anhydrous toluene
(2ꢂ300 mL) and dried in vacuo to give compound 6
(64.8 g, 102%) as a brown solid material. RP-HPLC:
Rt¼8.75 min, A290/260¼0.28, purity 96% (integration at
260 nm). Analytical sample was obtained after silica gel
column chromatography (1.5–3.5% v/v MeOH/CH2Cl2)
as a white solid foam. 1H NMR (DMSO-d6): d 11.40 (br s,
1H), 7.64 (d, J¼1.1 Hz, 1H), 7.40–7.30 (m, 5H), 5.96 (s,
1H), 4.71 (d, J¼12.1 Hz, 1H), 4.68 (d, J¼12.1 Hz, 1H),
4.60 (m, 3H), 4.50 (s, 1H), 4.12 (d, J¼8.6 Hz, 1H), 4.00
(d, J¼8.5 Hz, 1H), 3.25 (s, 3H), 1.84 (d, J¼1.1 Hz, 3H).
13C NMR (DMSO-d6): d 163.8, 150.3, 137.7, 135.8, 128.4,
127.8, 127.6, 108.3, 86.8, 86.5, 79.2, 76.5, 72.0, 71.3,
65.3, 37.0, 12.3. MALDI-TOF MS: m/z 460.8 (M+Na+).
points were measured on a Buchi Melting Point B-540 appa-
¨
ratus and uncorrected. The mp values are reported for all
crystalline compounds. All the other solid products were ob-
tained as amorphous solid foam after chromatographic puri-
fication and generally demonstrated broad phase transitions.
4.1.1. 3-O-Benzyl-4-C-methanesulfonoxymethyl-5-O-
methanesulfonyl-1,2-O-isopropylidene-b-L-threo-pento-
furanose (2). Colorless viscous oil; obtained as described
earlier.25 1H NMR (CDCl3): d 7.40–7.31 (m, 5H), 5.98 (d, J¼
4.0 Hz, 1H), 4.71 (d, J¼11.5 Hz, 1H), 4.70 (d, J¼3.9 Hz,
1H), 4.55 (d, J¼11.5 Hz, 1H), 4.39 (d, J¼10.4 Hz, 1H),
4.38 (d, J¼10.8 Hz, 1H), 4.33 (d, J¼10.6 Hz, 2H), 4.10 (s,
1H), 3.02 (s, 3H), 3.00 (s, 3H), 1.57 (s, 3H), 1.33 (s, 3H).
13C NMR (CDCl3) d 136.1, 128.5, 128.3, 127.9, 113.2,
105.7, 85.9, 84.0, 82.8, 72.7, 67.3, 67.0, 37.5, 37.3, 26.4,
25.8. MALDI-TOF MS: m/z 489.0 (M+Na+).
4.1.2. 1,2-Di-O-acetyl-3-O-benzyl-4-C-methanesulfonoxy-
methyl-5-O-methanesulfonyl-a,b-L-threo-pentofuranose
(3). To a solution of furanose 2 (73.4 g, 0.157 mol) in
AcOH (250 mL) were added Ac2O (40 mL, 0.424 mol) and
concd H2SO4 (1.1 mL). The solution was stirred overnight,
cooled in an ice bath, and 2 M NaOH (400 mL) was added
under intensive stirring. The mixture was washed with
CH2Cl2 (3ꢂ200 mL). The combined organic layers were
washed with saturated NaHCO3 (2ꢂ400 mL) and brine
(400 mL), dried (Na2SO4), and concentrated under reduced
pressure to give 3 (72.6 g, 91%) as a brownish viscous
oil.25

A. A. Koshkin / Tetrahedron 62 (2006) 5962–5972
5969
4.1.5. 1-(3-O-Benzyl-2-O,4-C-methylene-a-L-ribofurano-
syl)thymine (7). A mixture of 6 (64.0 g, 0.146 mol) and
NaOBz (41 g, 0.285 mol) suspended in anhydrous DMSO
(450 mL) was stirred for 3 h at 120 ꢀC. The mixture was
cooled to room temperature and 1 M NaOH (500 mL) was
added under intensive stirring. The obtained clear solution
was stirred for 1 h, neutralized with AcOH (30 mL), and
washed with CH2Cl2 (3ꢂ300 mL). The combined organic
layers were filtered through a silica gel pad (300 cm3); silica
gel was washed with 4% MeOH/CH2Cl2 (v/v, 500 mL). The
combined filtrates were dried (Na2SO4) and concentrated
under reduced pressure at 70 ꢀC until complete removal of
DMSO. The solid residue was crystallized from EtOAc to
give nucleoside 7 (30.5 g) as a white powder. The mother
liquor was concentrated to a solid residue and applied to sil-
ica gel column chromatography (1–4% v/v MeOH/CH2Cl2).
The fractions containing 7 were pooled, concentrated under
reduced pressure, and the residue was crystallized from
EtOAc to give another 8.4 g of 7 (total yield 74%). Mp
169–170 ꢀC. MALDI-TOF MS: m/z 382.7 (M+Na+). Anal.
Calcd for C18H20N2O6: C, 55.99; H, 5.59; N, 7.77. Found:
C, 60.02; H, 5.47; N, 7.79. NMR data were identical with
those reported earlier.26
113.4, 108.4, 88.0, 86.6, 85.7, 77.1, 73.6, 72.5, 58.8, 55.1,
20.6, 12.4. MALDI-TOF MS: m/z 636.5 (M+Na+).
4.1.8. Compound 10b. White solid foam; prepared by the
method described for 10a (yield 89%). H NMR (DMSO-
1
d6): d 11.50 (br s, 1H), 7.59 (s, 1H), 7.43–7.23 (m, 9H),
6.93–6.90 (m, 4H), 5.56 (s, 1H), 5.15 (s, 1H), 4.51 (s, 1H),
3.81 (d, J¼8.4 Hz, 1H), 3.75 (d, J¼8.4 Hz, 1H), 3.74 (s,
6H), 3.46 (d, J¼11.5 Hz, 1H), 3.38 (d, J¼11.5 Hz, 1H),
2.02 (s, 3H), 1.63 (d, J¼0.9 Hz, 3H). 13C NMR (DMSO-
d6): d 169.4, 163.8, 158.3, 150.0, 144.5, 135.1, 135.0,
133.8, 129.8, 129.7, 128.1, 127.6, 127.0, 113.4, 109.2,
86.6, 86.5, 86.0, 77.4, 71.7, 70.6, 57.8, 55.1, 20.5, 12.5.
MALDI-TOF MS: m/z 635.9 (M+Na+). Anal. Calcd for
C34H34N2O9$1/4H2O: C, 65.96; H, 5.62; N, 4.52. Found:
C, 65.89; H, 5.55; N, 4.47.
4.1.9. 1-[3-O-Acetyl-5-O-(4,40-dimethoxytrityl)-2-O,4-C-
methylene-a-L-ribofuranosyl]-5-methyl-4-(1,2,4-triazol-
1-yl)-2-oxypyrimidine (11a). A mixture of compound 10a
(18.5 g, 29.2 mmol) and 1,2,4-triazole (18.6 g, 263 mmol)
was suspended in anhydrous acetonitrile (300 mL) and
freshly distilled phosphoryl chloride (4.1 mL, 50 mmol)
was added. The mixture was cooled in an ice bath and diiso-
propylethylamine (51 mL, 292.3 mmol) was slowly added
under intensive stirring. The mixture was stirred for 2 h at
room temperature, diluted with ethyl acetate (300 mL),
washed with saturated NaHCO3 (3ꢂ300 mL) and brine
(250 mL), dried (Na2SO4), and concentrated under reduced
pressure to give after drying compound 11a (18.9 g, 97%)
as a brownish solid material. RP-HPLC: Rt¼12.76 min,
A290/260¼0.235, purity 98% (integration at 260 nm).
Analytical sample was obtained after silica gel column chro-
matography (50–80% v/v EtOAc/hexane, containing 1% of
4.1.6. 1-(2-O,4-C-Methylene-a-L-ribofuranosyl)thymine
(8a). To a solution of compound 7 (30.4 g, 84 mmol)
in MeOH/1,4-dioxane (1:1 v/v, 200 mL) were added
Pd(OH)2/C (6 g) and HCO2NH4 (15 g). The mixture was
stirred under reflux for 1 h. More HCO2NH4 was added in
portions of 5 g at an interval of 30 min (total amount 25 g).
The mixture was cooled to room temperature, diluted with
30% aqueous NH4OH (50 mL), and filtered through
a Celite pad. The Celite was washed with 30% aqueous
NH4OH (75 mL). The filtrates were combined, concentrated
under reduced pressure to ca. half of its volume and kept at
4 ꢀC overnight. The precipitate formed was collected by
filtration, washed with cold H2O, and dried in vacuo to give
compound 8a (21.1 g, 93%) as a white crystalline powder.
Mp 239–242 ꢀC (dec.); starts sintering at 234 ꢀC. MALDI-
TOF MS: m/z 292.9 (M+Na+). Anal. Calcd for
C11H14N2O6$1/5H2O: C, 48.25; H, 5.30; N, 10.23. Found:
C, 48.18; H, 5.16; N, 10.23. NMR data were identical with
those reported earlier.26
1
Et3N) as a white solid foam. H NMR (CDCl3): d 9.29 (s,
1H), 8.12 (s, 1H), 8.10 (s, 1H), 7.45–7.41 (m, 2H), 7.34–
7.24 (m, 7H), 6.88–6.83 (m, 4H), 6.18 (s, 1H), 5.34 (s, 1H),
4.95 (1H, s), 4.07 (d, J¼9.0 Hz, 1H), 4.01 (d, J¼9.0 Hz,
1H), 3.80 (s, 6H), 3.57 (d, J¼11.0 Hz, 1H), 3.46 (d,
J¼11.0 Hz, 1H), 2.55 (s, 3H), 2.05 (s, 3H). 13C NMR
(CDCl3): d 169.1, 158.5, 158.2, 153.5, 153.3, 146.1, 144.9,
144.0, 135.0, 129.8, 127.8, 127.6, 126.9, 113.1, 105.0,
89.05, 88.98, 86.3, 76.8, 73.7, 73.3, 59.1, 55.1, 20.5, 17.4.
MALDI-TOF MS: m/z 686.7 (M+Na+).
4.1.7. 1-[3-O-Acetyl-5-O-(4,40-dimethoxytrityl)-2-O,4-C-
methylene-a-L-ribofuranosyl]thymine (10a). To a solution
of compound 8a (9.13 g, 33.8 mmol) in anhydrous pyridine
(60 mL) was added DMT-Cl (14.2 g, 42.2 mmol). The mix-
ture was stirred for 4 h and Ac2O (4.5 mL, 43.9 mmol) was
added. After stirring overnight, the mixture was diluted
with EtOAc (250 mL), washed with saturated NaHCO3
(2ꢂ300 mL), dried (Na2SO4), and concentrated under re-
duced pressure to an oily residue. The residue was co-evapo-
rated with toluene (2ꢂ200 mL) and purified by silica gel
column chromatography (0.5–2% v/v MeOH/CH2Cl2, con-
taining 1% of Et3N) to give compound 10a (18.7 g, 90%)
4.1.10. Compound 11b. White solid foam; prepared in 90%
yield as described for 11a. 1H NMR (CDCl3): d 9.29 (s, 1H),
8.27 (s, 1H), 8.12 (s, 1H), 7.46–7.24 (m, 9H), 6.89–6.84 (m,
4H), 5.85 (s, 1H), 5.15 (s, 1H), 4.82 (1H, s), 3.91 (d,
J¼8.2 Hz, 1H,), 3.88 (d, J¼8.2 Hz, 1H), 3.80 (s, 6H),
3.64 (d, J¼11.2 Hz, 1H), 3.38 (d, J¼11.2 Hz, 1H), 2.22
(s, 3H), 2.03 (s, 3H,). 13C NMR (CDCl3): d 169.1, 158.6,
158.4, 153.3, 153.2, 145.4, 145.0, 143.9, 134.8, 129.9,
129.8, 127.9, 127.8, 127.0, 113.2, 106.1, 88.3, 87.4, 86.7,
77.1, 72.1, 70.2, 57.4, 55.1, 20.5, 17.2. MALDI-TOF MS:
m/z 687.5 (M+Ma+). Anal. Calcd for C36H35N5O8: C,
64.95; H, 5.30; N, 10.52. Found: C, 64.69; H, 5.26; N,
10.45.
1
as a white solid foam. H NMR (DMSO-d6): d 11.46 (br s,
1H), 7.62 (d, J¼1.1 Hz, 1H), 7.41–7.24 (m, 9H), 6.91 (m,
4H), 6.12 (s, 1H), 5.45 (s, 1H), 4.55 (s, 1H), 4.18 (d, J¼
8.8 Hz, 1H), 3.86 (d, J¼8.8 Hz, 1H), 3.74 (s, 6H), 3.32 (m,
2H; overlaps with H2O), 2.04 (s, 3H), 1.84 (d, J¼1.1 Hz,
3H). 13C NMR (DMSO-d6): d 169.5, 163.9, 158.3, 150.4,
144.6, 135.6, 135.2, 135.1, 129.9, 128.0, 127.7, 126.9,
4.1.11. 4-N-Benzoyl-1-[5-O-(4,40-dimethoxytrityl)-2-O,4-
C-methylene-a-L-ribofuranosyl]-5-methylcytosine (12a).
To a solution of compound 11a (18.5 g, 27.8 mmol) in
MeCN (250 mL) was added concentrated NH4OH

5970
A. A. Koshkin / Tetrahedron 62 (2006) 5962–5972
(200 mL). The mixture was kept overnight at room temper-
ature and the solvents were removed under reduced pressure
to give a solid residue. The residue was co-evaporated with
anhydrous pyridine (2ꢂ200 mL), dissolved in anhydrous
pyridine (250 mL), and Bz2O (19.0 g, 84.0 mmol) was
added. After stirring overnight, to the mixture were added
MeOH (150 mL) and 2 M NaOH (150 mL). The obtained
clear solution was stirred for 1 h and washed with CH2Cl2
(2ꢂ300 mL). The combined organic phases were exces-
sively washed with saturated NaHCO3 and brine, dried
(Na2SO4), and concentrated under reduced pressure to
a solid residue. Purification by silica gel column chro-
matography (5–15% v/v EtOAc/CH2Cl2, containing 1% of
Et3N) gave compound 12a (15.2 g, 81%) as a white solid
foam.25
4.1.14. 7-(3-O-Benzyl-5-O-methanesulfonyl-2-O,4-C-
methylene-b-D-ribofuranosyl)-2-N-isobutyrylguanine
(16). White solid foam; prepared using the method described
earlier24 for the N9-glycosylated counterpart (yield 89%). 1H
NMR (CDCl3): d 10.17 (br s, 1H), 8.12 (s, 1H), 7.28–7.20
(m, 5H), 6.20 (s, 1H), 4.82 (s, 1H), 4.67 (d, J¼11.9 Hz,
1H), 4.60 (d, J¼11.9 Hz, 1H), 4.58 (d, J¼11.5 Hz, 1H),
4.53 (d, J¼11.5 Hz, 1H), 4.21 (s, 1H), 4.15 (d, J¼7.9 Hz,
1H), 3.97 (d, J¼7.9 Hz, 1H), 3.11 (s, 3H), 2.90 (m, 1H),
1.28 (d, J¼6.8 Hz, 3H), 1.22 (d, J¼7.0 Hz, 3H). 13C NMR
(CDCl3): d 179.7, 157.4, 152.8, 147.8, 140.0, 136.5, 128.4,
128.1, 127.7, 110.4, 88.3, 85.7, 78.0, 76.1, 72.3, 71.8,
64.2, 37.7, 35.9, 19.0, 18.8. MALDI-TOF MS: m/z 533.6
(M+H+). Anal. Calcd for C23H27N5O8S$1/6H2O: C, 51.49;
H, 5.13; N, 13.05. Found: C, 51.46; H, 5.02; N, 12.64.
4.1.12. Compound 12b. White solid foam; prepared as de-
scribed for 10a (yield 76%). 1H NMR (DMSO-d6): d 13.14
(br s, 1H), 8.20 (m, 2H), 7.79 (s, 1H), 7.57–7.24 (m, 12H),
6.93 (m, 4H), 5.87 (d, J¼3.9 Hz, 1H), 5.54 (s, 1H), 4.31 (s,
1H), 4.22 (d, J¼4.4 Hz, 1H), 3.81 (d, J¼7.9 Hz, 1H), 3.75
(s, 7H), 3.51 (d, J¼11.0 Hz, 1H), 3.39 (d, J¼11.0 Hz, 1H),
1.80 (s, 3H). 13C NMR (DMSO-d6): d 178.3, 159.3, 158.3,
147.0, 144.8, 137.0, 135.4, 135.4, 135.2, 132.6, 129.9,
129.5, 128.3, 128.1, 127.8, 126.9, 113.4, 109.6, 87.8, 87.1,
85.9, 78.7, 71.5, 69.3, 58.6, 55.2, 13.3. MALDI-TOF MS:
m/z 697.3 (M+Na+).
4.1.15. 7-(5-O-Benzoyl-3-O-benzyl-2-O,4-C-methylene-
b-D-ribofuranosyl)-2-N-isobutyrylguanine (17). White
solid foam; prepared using the method described earlier24
1
for the N9-glycosylated counterpart (yield 93%). H NMR
(CDCl3): d 10.39 (br s, 1H), 7.99–7.97 (m, 3H), 7.61–7.57
(m, 1H), 7.47–7.43 (m, 2H), 7.23–7.18 (m, 5H), 6.21
(s,1H), 4.84 (s, 1H), 4.80 (d, J¼12.7 Hz, 1H), 4.70 (d,
J¼12.8 Hz, 1H), 4.62 (d, J¼11.7 Hz, 1H), 4.48 (d,
J¼11.7 Hz, 1H), 4.25 (d, J¼7.9 Hz, 1H), 4.15 (s, 1H),
4.07 (d, J¼7.9 Hz, 1H), 2.96 (m, 1H), 1.25 (d, J¼6.8 Hz,
3H), 1.21 (d, J¼7.0 Hz, 3H). 13C NMR (CDCl3): d 179.8,
165.7, 157.5, 152.7, 148.0, 139.7, 136.5, 133.4, 129.4,
129.0, 128.5, 128.3, 128.0, 127.5, 110.5, 88.3, 86.1, 78.0,
76.2, 72.3, 72.1, 59.8, 35.9, 19.0, 18.9. MALDI-TOF MS:
m/z 560.0 (M+H+). Anal. Calcd for C29H29N5O7$1/5H2O:
C, 61.85; H, 5.26; N, 12.44. Found: C, 61.91; H, 5.25; N,
11.95.
4.1.13. 7-(2-O-Acetyl-3-O-benzyl-4-C-methanesulfoxy-
methyl-5-O-methanesulfonyl-b-D-erythro-pentofurano-
syl)-2-N-isobutyrylguanine (15). A mixture of furanose
1424 (6.0 g, 11.7 mmol) and 2-N-isobutyrylguanine (3.1 g,
14.1 mmol) was suspended in anhydrous MeCN (60 mL)
and N,O-bis(trimethylsilyl)acetamide (5.7 mL, 23.1 mmol)
was added. The mixture was stirred overnight whereupon
the presence of insoluble residue was still observed. The
solvents were removed under reduced pressure and the resi-
due was co-evaporated with anhydrous MeCN (2ꢂ50 mL)
and suspended in anhydrous MeCN (50 mL). To the mix-
ture was added N,O-bis(trimethylsilyl)acetamide (2.8 mL,
11.3 mmol). After stirring for 1 h, a clear solution was ob-
tained and SnCl4 (1.5 mL, 12.8 mmol) was added. After stir-
ring for 20 h, the mixture was diluted with EtOAc (100 mL)
and saturated NaHCO3 (50 mL). The mixture was filtrated
through Celite pad and the Celite was additionally washed
with 10% v/v MeOH/EtOAc (50 mL). Organic layer was
separated, washed with saturated NaHCO3 (2ꢂ100 mL)
and brine (50 mL), dried (Na2SO4), and concentrated to a
solid residue. Purification with silica gel HPLC (0–4% v/v
MeOH/CH2Cl2) gave compound 15 (2.7 g, 34%) as a white
4.1.16. 7-(3-O-Benzyl-2-O,4-C-methylene-b-D-ribofura-
nosyl)-2-N-isobutyrylguanine (18). White crystalline pow-
der; synthesized using the method described earlier24 for the
N9-glycosylated counterpart; crystallized from MeOH/H2O
(yield 85%). Mp 190–191.5 ꢀC. 1H NMR (DMSO-d6):
d 12.19 (br s, 1H), 11.57 (br s, 1H), 8.29 (s, 1H), 7.31–
7.22 (m, 5H), 6.11 (s,1H), 5.26 (t, J¼5.8 Hz, 1H), 4.64 (s,
1H), 4.57 (s, 2H), 4.13 (s, 1H), 3.93 (d, J¼7.9 Hz, 1H),
3.85 (d, J¼5.8 Hz, 2H), 3.79 (d, J¼7.9 Hz, 1H), 2.76 (m,
1H), 1.14–1.11 (m, 6H). 13C NMR (DMSO-d6): d 180.1,
157.7, 152.5, 147.4, 140.6, 137.9, 128.3, 127.6, 127.5,
110.5, 88.6, 87.3, 77.8, 76.0, 71.8, 71.3, 56.5, 34.8, 19.0,
18.9. MALDI-TOF MS: m/z 457.1 (M+H+). Anal. Calcd
for C29H29N5O7$H2O: C, 55.81; H, 5.75; N, 14.79. Found:
C, 55.63; H, 5.59; N, 14.69.
1
solid foam. H NMR (CDCl3): d 12.07 (br s, 1H), 9.48 (br
4.1.17. 7-(2-O,4-C-Methylene-b-D-ribofuranosyl)guanine
(21). To a solution of compound 18 (540 mg, 1.19 mmol) in
MeOH (10 mL) were added Pd/C (10%, 250 mg) and
HCO2NH4 (0.5 g). The mixture was refluxed until complete
consumption of 18 had occurred whereupon formation of
insoluble product(s) was observed. Saturated NH3 in MeOH
was added (20 mL) under intensive stirring and the catalyst
was filtered off and washed with saturated NH3 in MeOH
(100 mL). The combined filtrates were stirred for 72 h and
concentrated to a solid residue. The residue was suspended in
MeOH/EtOAc (1:1 v/v, 100 mL) under reflux for 30 min.
The mixture was concentrated to ca. half of its volume, and
cooled in an ice bath. Insoluble material was filtered off,
s, 1H), 7.97 (s, 1H), 7.30 (m, 5H), 6.62 (d, J¼5.7 Hz, 1H),
6.32 (s, 1H), 5.33 (d, J¼5.8 Hz, 1H), 4.76 (d, J¼11.4 Hz,
1H), 4.74 (d, J¼11.5 Hz, 1H), 4.68 (d, J¼11.4 Hz, 1H),
4.57 (d, J¼10.6 Hz, 1H), 4.46 (d, J¼11.5 Hz, 1H), 4.44 (d,
J¼10.5 Hz, 1H), 3.05 (s, 3H), 3.00 (s, 3H), 2.77 (m, 1H),
2.27 (s, 3H), 1.39 (d, J¼7.0 Hz, 3H), 1.30 (d, J¼6.8 Hz,
3H). 13C NMR (CDCl3): d 179.7, 170.3, 157.4, 152.2,
147.4, 141.0, 137.1, 128.4, 128.3, 128.1, 110.2, 90.8, 84.7,
(77.2, 76.9, 76.6 overlap with CDCl3), 74.1, 68.0, 37.6,
36.7, 36.3, 21.0, 19.0, 18.6. MALDI-TOF MS: m/z 671.8
(M+H+). Anal. Calcd for C26H33N5O12S2: C, 46.49; H,
4.95; N, 10.43. Found: C, 46.45; H, 4.93; N, 9.90.

A. A. Koshkin / Tetrahedron 62 (2006) 5962–5972
5971
washed with EtOAc and Et2O, and dried in vacuo to give 21
(302 mg, 86%) as a white powder. 1H NMR (DMSO-d6; low
solubility): d 10.87 (br s, 1H), 8.08 (s, 1H), 6.18 (br s, 2H),
5.94 (s, 1H), 5.59 (br s, 1H), 5.03 (br s, 1H), 4.29 (s, 1H),
4.04 (s, 1H), 3.88 (d, J¼7.8 Hz, 1H), 3.79 (m, 2H), 3.70 (d,
J¼7.7 Hz, 1H). 13C NMR (DMSO-d6; sugar part): d 88.7,
87.0, 80.1, 71.1, 68.8, 56.4. MALDI-TOF MS: m/z 296.0
(M+H+).
amidites coupling times and times of oxidation were ex-
tended to 500 and 30 s, respectively. After oligomerization,
the solid support bound oligonucleotides were transferred
into 1.5 mL reaction tubes and treated with concentrated
NH4OH for 6 h at 60 ꢀC. Oligonucleotide 46 (phosphor-
amidite 20 used) was subsequently treated with 30%
MeNH2 for 2 h at 60 ꢀC. The solvents were removed under
reduced pressure and target oligonucleotides were purified
by RP-HPLC. Structures of the oligonucleotides were veri-
fied by MALDI-TOF mass spectra recorded in negative
ion mode using picolinic acid as the matrix (results shown
in Table 3).
4.1.18. 7-[5-O-(4,40-Dimethoxytrityl)-2-O,4-C-methyl-
ene-b-D-ribofuranosyl]-2-N-dimethylaminomethylene-
guanine (22). Compound 21 (250 mg, 0.85 mmol) was
dissolved in hot anhydrous DMF (8 mL) and N,N-dimethyl-
formamide dimethyl acetal (300 mL, 2.26 mmol) was added.
The mixture was stirred overnight and H2O (1 mL) was
added. The solvents were removed under reduced pressure
to give an oily residue. The residue was co-evaporated with
anhydrous pyridine (2ꢂ10 mL), dissolved in anhydrous pyr-
idine (5 mL), and DMT-Cl (310 mg, 0.92 mmol) was added.
The mixture was stirred for 4 h, diluted with EtOAc (20 mL),
washed with saturated NaHCO3 (2ꢂ20 mL), dried (Na2SO4),
and concentrated to a solid residue. Purification by silica gel
HPLC (1–4% v/v MeOH/CH2Cl2, containing 0.1% of pyri-
dine) gave compound 22 (490 mg, 89%) as a white solid
Table 3. The results of MALDI-TOF analysis
Oligonucleotide
(MꢃH)ꢃ
Calcd
Found
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
4385.8
4385.8
4385.8
4581.8
4581.8
4385.8
4385.8
4385.8
4385.8
4515.7
4463.7
4435.7
4379.7
4411.8
4411.8
4411.8
9259.0
9150.8
4417.7
2672.7
2681.8
2697.5
2697.5
4380.1
4381.3
4385.2
4576.6
4583.0
4386.0
4383.9
4384.0
4384.7
4513.1
4459.4
4435.4
4376.8
4405.1
4407.8
4410.6
9264.0
9147.0
4414.8
2972.8
2681.9
2698.9
2697.9
1
foam. H NMR (DMSO-d6): d 11.51 (br s, 1H), 8.62, (s,
1H), 8.11 (s, 1H), 7.46–7.28 (m, 9H), 6.91 (m, 4H), 6.10
(s, 1H), 5.71 (d, J¼4.4 Hz, 1H), 4.41 (s, 1H), 4.20 (d, J¼
4.4 Hz, 1H), 3.89 (d, J¼7.9 Hz, 1H), 3.86 (d, J¼7.9 Hz,
1H), 3.75 (s, 6H), 3.56 (d, J¼11.0 Hz, 1H), 3.33 (d,
J¼11.0 Hz, 1H), 3.15 (s, 3H), 3.02 (s, 3H). 13C NMR
(DMSO-d6): d 159.5, 158.2, 157.9, 156.9, 155.2, 144.8,
139.6, 135.5, 135.3, 129.9, 129.8, 128.0, 127.8, 126.9,
113.4, 109.9, 87.3, 85.6, 80.2, 71.6, 69.5, 59.6, 55.2, 40.6,
34.7. MALDI-TOF MS: m/z 675.7 (M+Na+).
4.1.19. 7-[3-O-Cyanoethoxy(diisopropylamino)phosphin-
oxy)-5-O-(4,40-dimethoxytrityl)-2-O,4-C-methylene-b-D-
ribofuranosyl]-2-N-dimethylaminomethyleneguanine
(23). To a solution of compound 22 (415 mg, 0.64 mmol) in
anhydrous CH2Cl2 (3 mL) were added 0.75 M solution of
4,5-dicyanoimidazole (700 mL, 0.53 mmol) in CH3CN and
2-cyanoethyl-N,N,N0,N0-tetraisopropyl phosphordiamidite
(230 mL, 0.72 mmol). The mixture was stirred overnight,
diluted with EtOAc (25 mL), and washed with saturated
NaHCO3 (2ꢂ25 mL) and brine (25 mL). The organic layer
was dried (Na2SO4) and concentrated to a white solid resi-
due. The residue was dissolved in anhydrous CH2Cl2
(5 mL) and pentane (100 mL) was added under intensive
stirring. The precipitate formed was collected by filtration,
washed with pentane, and dried in vacuo to give compound
23 (498 mg, 92%) as a white amorphous powder consisting
4.3. Thermal denaturation studies
The samples of model triplexes and duplexes were obtained
after mixing the equal molar amounts of complementary
oligonucleotides and buffers (final volumes 0.5 mL). The
buffer and concentration conditions are described in the
footnotes to Tables 1 and 2. To ensure triplex formation,
the samples were kept for 12 h at 4 ꢀC prior to melting exper-
iments. Melting temperature measurements were carried out
on a Perkin–Elmer UV-spectrometer equipped with a PTP-6
Peltier temperature controller. The optical density of sam-
ples was monitored at 260 nm while raising the temperature
from 20 to 95 ꢀC at a rate of 1 ꢀC/min. The Tm values were
determined as the maxima of the first derivatives of the
melting curves obtained.
of two stereoisomers. RP-HPLC: Rt¼15.05 min, A290/260
¼
2.54. MALDI-TOF MS: m/z 791.2 (MꢃNiPr2+OH+Na+).
31P NMR (DMSO-d6): d 149.20, 147.98.
Compound 20. 31P NMR (DMSO-d6): d 148.96, 148.70.
References and notes
4.2. Oligonucleotide synthesis
1. Felsenfeld, G.; Davies, D. R.; Rich, A. J. Am. Chem. Soc. 1957,
79, 2023–2024.
All oligonucleotides were synthesized in 0.2 mmol scales on
Expedite DNA synthesizer using the phosphoramidite
method. Standard protocols recommended by the manufac-
turer were used for commercial DNA amidites. For all LNA
2. Giovannangeli, C.; Helene, C. Curr. Opin. Mol. Ther. 2000, 2,
288–296.
3. Uil, T. G.; Haisma, H. J.; Rots, M. G. Nucleic Acids Res. 2003,
31, 6603–6614.

5972
A. A. Koshkin / Tetrahedron 62 (2006) 5962–5972
4. Soyfer, V. N.; Potoman, V. N. Triple-helical Nucleic Acids;
Springer: New York, NY, 1996.
23. Sørensen, J. J.; Nielsen, J. T.; Pedersen, M. Nucleic Acids Res.
2004, 32, 6078–6085.
5. Beal, P. A.; Dervan, P. B. Science 1991, 251, 1360–1363.
6. Goni, J. R.; de la Cruz, X.; Orozco, M. Nucleic Acids Res. 2004,
32, 354–360.
24. Koshkin, A. A.; Fensholdt, J.; Pfundheller, H. M.; Lomholt, C.
J. Org. Chem. 2001, 66, 8504–8512.
˚
25. Sørensen, M. D.; Kværnø, L.; Bryld, T.; Hakansson, A. E.;
7. Kuan, J. Y.; Glazer, P. M. Methods Mol. Biol. 2004, 262,
173–194.
Verbeure, B.; Gaubert, G.; Herdewijn, P.; Wengel, J. J. Am.
Chem. Soc. 2002, 124, 2164–2176.
˚
8. Faria, M.; Giovannangeli, C. J. Gene Med. 2001, 3, 299–310.
9. Buchini, S.; Leumann, C. J. Curr. Opin. Chem. Biol. 2003, 7,
717–726.
26. Hakansson, A. E.; Koshkin, A. A.; Sørensen, M. D.; Wengel, J.
J. Org. Chem. 2000, 65, 5161–5166.
˚
27. Nielsen, K. M. E.; Pedersen, M.; Hakansson, A. E.; Wengel, J.;
10. Nielsen, P. E.; Egholm, M.; Berg, R. H.; Buchardt, O. Science
1991, 254, 1497–1500.
11. Egholm, M.; Buchardt, O.; Chistensen, L.; Behrens, C.; Freier,
S. M.; Driver, D. A.; Berg, R. H.; Kim, S. K.; Norden, B.;
Nielsen, P. E. Nature 1993, 365, 566–568.
Jacobsen, J. P. Chem.—Eur. J. 2002, 8, 3001–3009.
28. Hunziker, J.; Priestley, E. S.; Brunar, H.; Dervan, P. B. J. Am.
Chem. Soc. 1995, 117, 2661–2662.
29. Brunar, H.; Dervan, P. B. Nucleic Acids Res. 1996, 24, 1987–
1991.
12. Nielsen, P. E. Curr. Med. Chem. 2001, 8, 545–550.
13. Cherny, D. Y.; Belotserkovskii, B. P.; Frank-Kamenetskii,
M. D.; Egholm, M.; Buchardt, O.; Berg, R. H.; Nielsen, P. E.
Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 1667–1670.
14. Cutrona, G.; Carpaneto, E. M.; Ulivi, M.; Roncella, S.; Landt,
O.; Ferrarini, M.; Boffa, L. C. Nature Biotechnol. 2000, 18,
300–303.
30. Vorbruggen, H.; Hofle, G. Chem. Ber. 1981, 114, 1256–
¨
¨
1268.
31. Pedersen, D. S.; Rosenbohm, C.; Koch, T. Synthesis 2002,
802–808.
32. Lee, J. S.; Woodsworth, M. L.; Latimer, L. J.; Morgan, A. R.
Nucleic Acids Res. 1984, 12, 6603–6614.
33. Obika, S.; Uneda, T.; Sugimoto, T.; Nanbu, D.; Minami, T.;
Doi, T.; Imanishi, T. Bioorg. Med. Chem. 2001, 9, 1001–
1011.
15. Gryaznov, S. M.; Chen, J.-K. J. Am. Chem. Soc. 1994, 116,
3143–3144.
16. Escude, C.; Giovannangeli, C.; Sun, J. S.; Lloyd, D. H.; Chen,
J. K.; Gryaznov, S. M.; Garestier, T.; Helene, C. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 4365–4369.
34. Koshkin, A. A.; Singh, S. K.; Nielsen, P.; Rajwanshi, V. K.;
Kumar, R.; Meldgaard, M.; Olsen, C. E.; Wengel, J.
Tetrahedron 1998, 54, 3607–3630.
17. Skorski, T.; Perrotti, D.; Nieborowska-Skorska, M.; Gryaznov,
S.; Calabretta, B. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 3966–
3971.
18. Tereshko, V.; Gryaznov, S.; Egli, M. J. Am. Chem. Soc. 1998,
120, 269–283.
19. Obika, S.; Nanbu, D.; Hari, Y.; Morio, K.; In, Y.; Ishida, T.;
Imanishi, T. Tetrahedron Lett. 1997, 38, 8735–8738.
20. Morita, K.; Takagi, M.; Hasegawa, C.; Masakatsu, K.;
Tsutsumi, S.; Sone, J.; Ishikawa, T.; Imanishi, T.; Koizumi,
M. Bioorg. Med. Chem. 2003, 11, 2211–2226.
21. Obika, S. Chem. Pharm. Bull. 2004, 52, 1399–1404.
22. Koizumi, M.; Morita, K.; Daigo, M.; Tsutsumi, S.; Koji, A.;
Obika, S.; Imanishi, T. Nucleic Acids Res. 2003, 31, 3267–
3273.
35. Divakar, K. J.; Reese, C. B. J. Chem. Soc., Perkin Trans. 1
1982, 1171–1176.
36. Seela, F.; Leonard, P. Helv. Chim. Acta 1996, 79, 477–487.
37. Rosenbohm, C.; Pedersen, D. S.; Frieden, M.; Jensen, F. R.;
Arent, S.; Larsen, S.; Koch, T. Bioorg. Med. Chem. 2004, 12,
2385–2396.
38. Caruthers, M. H. Acc. Chem. Res. 1991, 24, 278–284.
39. McBride, L. J.; Kierzek, R.; Beaucage, S. L.; Caruthers, M. H.
J. Am. Chem. Soc. 1986, 108, 2040–2048.
40. Sun, B.-W.; Babu, B. R.; Sørensen, M. D.; Zakrzewska, K.;
Wengel, J.; Sun, J.-S. Biochemistry 2004, 43, 4160–4169.
41. Alberti, P.; Arimondo, P. B.; Mergny, J.-L.; Garestier, T.;
´ `
Helene, C.; Sun, J.-S. Nucleic Acids Res. 2002, 30, 5407–
5415.