New base pairing motifs. The synthesis and thermal stability of oligodeoxynucleotides containing imidazopyridopyrimidine nucleosides with the ability to form four hydrogen bonds

Article abstract of DOI:10.1021/ja0347686

The synthesis and thermal stability of oligodeoxynucleotides (ODNs) containing imidazo[5′,4′: 4,5]pyrido[2,3-d]pyrimidine nucleosides 1-4 (NN, O^O, N^O, and O^N, respectively) with the aim of developing two sets of new base pairing motifs consisting of four hydrogen bonds (H-bonds) is described. The proposed four tricyclic nucleosides 1-4 were synthesized through the Stille coupling reaction of a 5-iodoimidazole nucleoside with an appropriate 5-stannylpyrimidine derivative, followed by an intramolecular cyclization. These nucleosides were incorporated into ODNs to investigate the H-bonding ability. When one molecule of the tricyclic nucleosides was incorporated into the center of each ODN (ODN I and II, each 17mer), no apparent specificity of base pairing was observed, and all duplexes were less stable than the duplexes containing natural G:C and A:T pairs. On the other hand, when three molecules of the tricyclic nucleosides were consecutively incorporated into the center of each ODN (ODN III and IV, each 17mer), thermal and thermodynamic stabilization of the duplexes due to the specific base pairings was observed. The melting temperature (T_m) of the duplex containing the N^O:O^N pairs showed the highest T_m of 84.0 °C, which was 18.2 and 23.5 °C higher than that of the duplexes containing G:C and A:T pairs, respectively. This result implies that N^O and O^N form base pairs with four H-bonds when they are incorporated into ODNs. The duplex containing N^O:O ^N pairs was markedly stabilized by the assistance of the stacking ability of the imidazopyridopyrimidine bases. Thus, we developed a thermally stable new base pairing motif, which should be useful for the stabilization and regulation of a variety of DNA structures.

Full text of DOI:10.1021/ja0347686

Published on Web 07/24/2003
New Base Pairing Motifs. The Synthesis and Thermal Stability
of Oligodeoxynucleotides Containing
Imidazopyridopyrimidine Nucleosides with the Ability to Form
Four Hydrogen Bonds
Noriaki Minakawa,* Naoshi Kojima, Sadao Hikishima, Takashi Sasaki,
Arihiro Kiyosue, Naoko Atsumi, Yoshihito Ueno, and Akira Matsuda*
Contribution from the Graduate School of Pharmaceutical Sciences, Hokkaido UniVersity,
Kita-12, Nishi-6, Kita-ku, Sapporo 060-0812, Japan
Received February 20, 2003; E-mail: noriaki@pharm.hokudai.ac.jp; matuda@pharm.hokudai.ac.jp
Abstract: The synthesis and thermal stability of oligodeoxynucleotides (ODNs) containing imidazo[5′,4′:
4,5]pyrido[2,3-d]pyrimidine nucleosides 1-4 (NN, OO, NO, and ON, respectively) with the aim of developing
two sets of new base pairing motifs consisting of four hydrogen bonds (H-bonds) is described. The proposed
four tricyclic nucleosides 1-4 were synthesized through the Stille coupling reaction of a 5-iodoimidazole
nucleoside with an appropriate 5-stannylpyrimidine derivative, followed by an intramolecular cyclization.
These nucleosides were incorporated into ODNs to investigate the H-bonding ability. When one molecule
of the tricyclic nucleosides was incorporated into the center of each ODN (ODN I and II, each 17mer), no
apparent specificity of base pairing was observed, and all duplexes were less stable than the duplexes
containing natural G:C and A:T pairs. On the other hand, when three molecules of the tricyclic nucleosides
were consecutively incorporated into the center of each ODN (ODN III and IV, each 17mer), thermal and
thermodynamic stabilization of the duplexes due to the specific base pairings was observed. The melting
temperature (Tm) of the duplex containing the NO:ON pairs showed the highest Tm of 84.0 °C, which was
18.2 and 23.5 °C higher than that of the duplexes containing G:C and A:T pairs, respectively. This result
implies that NO and ON form base pairs with four H-bonds when they are incorporated into ODNs. The
duplex containing NO:ON pairs was markedly stabilized by the assistance of the stacking ability of the
imidazopyridopyrimidine bases. Thus, we developed a thermally stable new base pairing motif, which should
be useful for the stabilization and regulation of a variety of DNA structures.
Introduction
for biological processes in cells. Consequently, much attention
has been devoted to the synthesis of artificial oligodeoxynucleo-
DNA, as is well-known, is the storage and carrier of genetic
information in all living organisms and generally forms a right-
handed double helix with a complementary DNA molecule. The
structure of the DNA double helix is based on the Watson-
Crick hydrogen bonding (H-bonding) of adenine:thymine (two
H-bonds) and guanine:cytosine (three H-bonds) base pairings.
H-bonding interaction plays a critical role in not only conserving
and transmitting genetic information but also in double helix
stability. In addition to the general right-handed double helix
with a straight axis, DNA is also known to adopt a variety of
structures including left-handed Z conformations, hairpins,
cruciforms, branched junctions, and quadruplexes depending on
their sequences and environmental conditions.1 The foundation
of the structures in these intra- and intermolecular interactions
is not only Watson-Crick base pairing but also Hoogsteen type
interactions.2 Thus, H-bonding between nucleobases, as well
as sequences, is thought to play a critical role in the confor-
mational diversity of DNA, each of which would be essential
tides (ODNs) including unnatural nucleobases, which could form
a more stable, higher-ordered structure with DNA, RNA and
protein, or DNA itself, with application to biochemistry,
biotechnology, and medicinal chemistry.3 Their therapeutic
application, for example, in antisense and antigene methodolo-
gies is an area of intense study, and a large number of ODNs
have been synthesized and evaluated for the thermal stability
of their duplex and triplex structures.4 In contrast to the research
directed toward thermally stable base pairing between natural
and unnatural nucleobases, few studies have been reported on
the possibility of alternative base pairing consisting of new
H-bonding motifs.3d,3f,5 For example, Benner et al. proposed new
base pairing, that is, isoguanine:isocytosine (isoG:isoC) and
2,6-diaminopyrimidine:1-methyl-pyrazolo[4,3-d]pyrimidine-
5,7(4H,6H)-dione base pairs, and investigated their enzymatic
incorporation into DNA and RNA with the aim of “extending
the genetic alphabet”.6 Although such effort has now been
replaced by the development of a new base pairing motif through
(1) Recent reviews of conformation diversity, see (a) Lilley, D. M. J. In
Bioorganic Chemistry: Nucleic Acids; Hecht, S. M., Ed.; Oxford University
Press: Oxford, 1996; Ch. 7, pp 186-215. (b) Belmont, P.; Constant, J.-F.;
Demeunynck, M. Chem. Soc. ReV. 2001, 30, 70-81.
(2) (a) Hoogsteen, K. Acta Crystallogr. 1963, 16, 907-916. (b) Quigley, G.
J.; Ughetto, G.; van der Marel, G.; van Boom, J. H.; Wang, A. H.-J.; Rich,
A. Science 1986, 232, 1255-1258.
9
9970
J. AM. CHEM. SOC. 2003, 125, 9970-9982
10.1021/ja0347686 CCC: $25.00 © 2003 American Chemical Society

Synthesis and Thermal Stability of Oligodeoxynucleotides
A R T I C L E S
hydrophobic interactions,7 rather than H-bonding interactions,
the results showing that the isoG:isoC bearing duplex has a
thermal stability comparable to that of the corresponding
guanine:cytosine base pair and is considerably greater than that
of the adenine:uracil base pair8 warrant consideration. Prompted
by these results, we envisioned a new base pairing motif, which
would be more thermally stable than the Watson-Crick
H-bonding. To this end, we propose two sets of new base pair
combinations consisting of four H-bonds, as shown in Figure
1.9 Thus far, several investigations of ODNs, including the
extended nucleobases, have been reported.3e,j,n,4g,10 However,
to our knowledge, there have been no reports with the goal of
stabilizing and controlling not only helical structures but also
other possible structures. Therefore, it seems worthwhile to
examine the proposed new base pairing motif when these
nucleobases are incorporated into complementary positions of
ODNs. Since the tricyclic nucleobases have extended flat
aromatic surfaces, enhancement of the stacking interaction is
also expected.11 In addition, the tricyclic skeletons also have a
Hoogsteen type interaction ability as do natural purine bases,
Figure 1. Proposed new base pairs consisting of four H-bonds.
Scheme 1 a
(3) For examples except for antisense and antigene methodologies: (a) Goeddel,
D. V.; Yansura, D. G.; Caruthers, M. H. Proc. Natl. Acad. Sci. U.S.A. 1978,
75, 3578-3582. (b) Taboury, J. A.; Adam, S.; Taillandier, E.; Neumann,
J.-M.; Tran-Dinh, S.; Huynh-Dinh, T.; Langlois d'Estaintot, B.; Conti, M.;
Igolen, J. Nucleic Acids Res. 1984, 12, 6291-6305. (c) Chollet, A.;
Kawashima, E. Nucleic Acids Res. 1988, 16, 305-317. (d) Solomon, M.
S.; Hopkins, P. B. J. Org. Chem. 1993, 58, 2232-2243. (e) Langouet, S.;
Muller, M.; Guengerich, F. P. Biochemistry 1997, 36, 6069-6079. (f)
Beaussire, J.-J.; Pochet, S. Tetrahedron 1998, 54, 13547-13556. (g) Yang,
X.; Sugiyama, H.; Ikeda, S.; Saito, I.; Wang, A. H.-J. Biophys. J. 1998,
75, 1163-1171. (h) He, X.; Krawczyk, S. H.; Swaminathan, S.; Shea, R.
G.; Dougherty, J. P.; Terhorst, T.; Law, V. S.; Griffin, L. C.; Coutre, S.;
Bischofberger, N. J. Med. Chem. 1998, 41, 2234-2242. (i) Sherer, E. C.;
Harris, S. A.; Soliva, R.; Orozco, M.; Laughton, C. A. J. Am. Chem. Soc.
1999, 121, 5981-5991. (j) Wang, Z.; Rizzo, C. J. Org. Lett. 2000, 2, 227-
230. (k) Sugiyama, T.; Schweinberger, E.; Kazimierczuk, Z.; Ranmzaeva,
N.; Rosemeyer, H.; Seela, F. Chem. Eur. J. 2000, 6, 369-378. (l) Sessler,
J. L.; Sathiosatham, M.; Doerr, K.; Lynch, V.; Abboud, K. A. Angew.
Chem., Int. Ed. 2000, 39, 1300-1303. (m) Seela, F.; Debelak, H. Nucleic
Acids Res. 2000, 28, 3224-3232. (n) Wilhelmsson, L. M.; Holmen, A.;
Lincoln, P.; Nielsen, P. E.; Norden, B. J. Am. Chem. Soc. 2001, 123, 2434-
2435.
(4) For examples: (a) Sanghvi, Y. S. In Antisense Research and Applications;
Crooke, S. T., and Lebleu, B., Eds.; CRC Press: Boca Raton, 1993; Ch.
15, pp 273-288. (b) Miller, P. S. In Bioorganic Chemistry: Nucleic Acids;
Hecht, S. M., Ed.; Oxford University Press: Oxford, 1996; Ch. 12, pp
347-374. (c) Doronina, S. O.; Behr, J.-P. Chem. Soc. ReV. 1997, 26, 63-
71. (d) Haginoya, N.; Ono, A.; Nomura, Y.; Ueno, Y.; Matsuda, A.
Bioconjugate Chem. 1997, 8, 271-280. (e) Doronina, S. O.; Behr, J.-P.
Tetrahedron Lett. 1998, 39, 547-550. (f) Heystek, L. E.; Zhou, H.; Dande,
P.; Gold, B. J. Am. Chem. Soc. 1998, 120, 12165-12166. (g) Lin, K.;
Matteucci, M. D. J. Am. Chem. Soc. 1998, 120, 8531-8532.
a Reagents: (a) dba3Pd2‚CHCl3, DMF, 100 °C; (b) EtOH(aq) Na2CO3,
80 °C; (c) 1,4-dioxane-NH4OH, 100 °C; and (d) TBAF, THF.
(5) Voegel, J. J.; Benner, S. A. HelV. Chim. Acta 1996, 79, 1881-1898.
(6) (a) Switzer, C.; Moroney, S. E.; Benner, S. A. J. Am. Chem. Soc. 1989,
111, 8322-8323. (b) Piccirilli, J. A.; Krauch, T.; Moroney, S. E.; Benner,
S. A. Nature 1990, 343, 33-37. (c) Roberts, C.; Bandaru, R.; Switzer, C.
J. Am. Chem. Soc. 1997, 119, 4640-4649.
which may contribute to adopting the three-dimensional struc-
tures of ODNs.
In this paper, we describe the synthesis and incorporation of
these tricyclic imidazopyridopyrimidine nucleosides into ODNs
and their effects on thermal stabilities of the ODNs, including
the newly designed base paring motifs.
(7) (a) Ogawa, A. K.; Wu, Y.; McMinn, D. L.; Liu, J.; Schultz, P. G.;
Romesberg, F. E. J. Am. Chem. Soc. 2000, 122, 3274-3287. (b) Wu, Y.;
Ogawa, A. K.; Berger, M.; McMinn, D. L.; Schultz, P. G.; Romesberg, F.
E. J. Am. Chem. Soc. 2000, 122, 7621-7632. (c) Ogawa, A. K.; Wu, Y.;
Berger, M.; Schultz, P. G.; Romesberg, F. E. J. Am. Chem. Soc. 2000,
122, 8803-8804.
Results and Discussion
(8) Roberts, C.; Bandaru, R.; Switzer, C. Tetrahedron Lett. 1995, 36, 3601-
3604.
1. Chemistry. The most straightforward synthesis of the
desired nucleosides was thought to be through intramolecular
cyclization of the 5-pyrimidinylimidazole nucleosides, which
would be prepared from the Stille coupling reaction of a
5-iodoimidazole nucleoside with an appropriate tributylstannyl-
pyrimidine. The 5-iodoimidazole nucleosides 5 (Scheme 1) and
6 (Scheme 3) can be synthesized from 2′-deoxyinosine.12 On
the other hand, the tributylstannylpyrimidine derivatives 7 and
8 can be synthesized from 2,4-dichloropyrimidine.13 These
chemical conversions are presented in the Supporting Informa-
tion.14
(9) For the sake of simplicity, the aglycon of 1 (and 11) shall be subsequently
referred to as diamino. Similarly, those of 2, 3, and 4 (and each of the
protected derivatives) shall be referred to as dioxo, amino-oxo, and oxo-
amino, respectively. When these nucleoside units were incorporated into
ODNs, the imidazopyridopyrimidine bases shall be referred to as NN, OO,
NO, and ON, respectively.
(10) (a) Leonard, N. J.; Hiremath, S. P. Tetrahedron 1986, 42, 1917-1961. (b)
Baus, A. K.; Niedernhofer, L. J.; Essigmann, J. M. Biochemistry 1987, 26,
5626-5635. (c) Lin, K.; Jones, R. J.; Matteucci, M. J. Am. Chem. Soc.
1995, 117, 3873-3874. (d) Godde, F.; Toulme, J.-J.; Moreau, S. Nucleic
Acids Res. 2000, 28, 2977-2985. (e) Hirao, I.; Ohtsuki, T.; Fujiwara, T.;
Mitsui, T.; Yokogawa, T.; Okuni, T.; Nakayama, H.; Takio, K.; Yabuki,
T.; Kigawa, T.; Kodama, K.; Yokogawa, T.; Nishikawa, K.; Yokoyama,
S. Nature Biotech. 2002, 20, 177-182.
(11) Gukian, K. M.; Schweitzer, B. A.; Ren, R. X.-F.; Sheils, C. J.; Tahmassebi,
D. C.; Kool, E. T. J. Am. Chem. Soc. 2000, 122, 2213-2222.
9
J. AM. CHEM. SOC. VOL. 125, NO. 33, 2003 9971

A R T I C L E S
Minakawa et al.
Scheme 2 a
Scheme 3 a
a Reagents: (a) (Boc)2O, Et3N, DMAP, CH2Cl2; (b) NaOMe, MeOH;
(c) 7, dba3Pd2‚CHCl3, DMF, 100 °C; (d) 1,4-dioxane-NH4OH, 100 °C; and
(e) TBAF, THF.
a Reagents: (a) dba3Pd2‚CHCl3, DMF, 100 °C; (b) H2O2, NH4OH-
MeOH; (c) isoamyl nitrite, THF(aq), 60 °C; (d) NH3/MeOH, 120 °C; and
(e) TBAF, THF.
yield. Deprotection of the silyl groups with tetrabutylammonium
fluoride (TBAF) gave the free nucleoside 1.
Preparation of the dioxo derivative 2 and the amino-oxo
derivative 3 was achieved via the Stille coupling reaction
between 5 and 8, as shown in Scheme 2. When a mixture of 5
and 8 was heated with dba3Pd2‚CHCl3, 12 was obtained in 81%
yield. To convert 12 into the dioxo derivative 2, the 4-cyano
group of 12 was hydrolyzed with hydrogen peroxide in the
First, the synthesis of the diamino derivative 1 was done by
the reaction of 5 and 7 (Scheme 1). When a mixture of 5 and
7 was heated in DMF in the presence of tris(dibenzylidene-
acetone)dipalladium(0)-chloroform adduct (dba3Pd2‚CHCl3), two
fluorescent spots were detected by TLC analysis. The UV
spectra of both spots showed absorption maxima at around 340
nm;15 thus, the new spots were presumed to be the coupling
products between 5 and 7. In addition, the less polar spot on
TLC was converted to the more polar spot when the reaction
mixture was treated under basic conditions (see Experimental
Procedures). As a result, the less polar spot was thought to be
the coupling product 9 (not isolated), and spontaneous cycliza-
tion giving a tricyclic product 10 took place partially during
the Stille coupling reaction. Treatment of 10 with a mixture of
1,4-dioxane and NH4OH gave the diamino derivative 11 in 86%
presence of NH4OH to give the 4-carboxamide derivative 13
in 92% yield. Although 13 was observed as a single spot by
1
TLC analysis, the H NMR spectrum of 13 was somewhat
complicated. For example, four sets of aromatic proton signals
were observed at 8.17, 8.05, 7.87, and 7.86 ppm each as a singlet
(in a ratio of 2:1:1:2) in CDCl3. Interestingly, when the 1H NMR
measurement of 13 was carried out at 60 °C, the four sets of
proton signals converged into two sets of proton signals at 8.12
and 7.81 ppm in a ratio of 1:1. Other proton signals were also
simplified under the same conditions. Consequently, compound
13 is considered to exist as a mixture of rotamers, arising from
steric repulsion between the spatially close amide and the
methoxy substituents. Hydrolytic deamination of 13 by treatment
with isoamyl nitrite in THF containing a small amount of H2O
gave 14 in 51% yield.16 The existence of rotamers was also
(12) (a) Minakawa, N.; Sasabuchi, Y.; Kiyosue, A.; Kojima, N.; Matsuda, A.
Chem. Pharm. Bull. 1996, 44, 288-295. (b) Napoli, L. D.; Messere, A.;
Montesarchio, D.; Piccialli, G.; Varra, M. J. Chem. Soc., Perkin Trans. 1
1997, 2079-2082.
(13) 2,4-Diaminopyrimidine and 2,4-dimethoxypyrimidine were also examined
as starting materials to prepare the tributylstannylpyrimidine derivatives.
In the former case, introduction of the tributylstannyl group did not work
well, while in the latter, it was very difficult to carry out the intramolecular
cyclization to give the desired tricyclic skeletons after the Stille coupling
reaction.
(14) Additional references for the synthesis of 5-8: (a) Minakawa, N.; Takeda,
T.; Sasaki, T.; Matsuda, A.; Ueda, T. J. Med. Chem. 1991, 34, 778-786.
(b) Harris, M. G.; Stewart, R. Can. J. Chem. 1977, 55, 3800-3806. (c)
Das, B.; Kundu, N. G. Synthetic Commun. 1988, 18, 855-867.
(15) The maximal absorption of 5 was observed at 237 nm, while those of 7
were observed at 244 and 280 nm, respectively.
1
observed in the H NMR spectrum of 14. When the resulting
14 was heated with methanolic ammonia in a steel container,
the dioxo derivative 15 was obtained in 73% yield. The amino-
oxo derivative 17 was synthesized also starting from 12 using
(16) This reaction was a modification of reductive deamination reported by Nair
and Chamberlain, see Nair, V.; Chamberlain, S. D. Synthesis 1984, 401-
403.
9
9972 J. AM. CHEM. SOC. VOL. 125, NO. 33, 2003

Synthesis and Thermal Stability of Oligodeoxynucleotides
A R T I C L E S
the same procedure, that is, hydrolytic deamination, followed
by the intramolecular cyclization. Both of the protected nucleo-
sides 15 and 17 were deprotected with TBAF to give the free
nucleosides 2 and 3, respectively.
assume that if irradiation of H-8 in purine 2′-deoxyribonucleo-
sides exhibits a strong NOE at H-1′ and small NOEs at H-2′â
and H-3′, then the syn conformation is preferred. On the other
hand, if irradiation of H-8 exhibits strong NOEs at H-2′â and
H-3′ and a smaller NOE at H-1′, then the anti conformation is
dominant. The NOE experiments were carried out using 4 by
irradiation of H-2 (corresponding to the H-8 position of the
purine base) and H-9 (the proton at the pyrimidine ring).
Compound 4 exhibited a small NOE (1.3%) value at H-1′ and
larger ones at H-2′â and H-3′ (5.6 and 2.3%) upon irradiation
of H-2. In addition, a strong NOE (13.1%) was observed at
H-1′ upon irradiation of H-9. The same NOE tendency was
observed when the experiment was carried out at 80 °C (data
not shown). As a result, 4 is thought to be a highly anti-
constrained compound, which is also expected for 1-3. In the
DNA-DNA double helix, the nucleoside unit generally adopts
the anti conformation to form the Watson-Crick H-bonding
except for the unusual left-handed Z-DNA.19 Consequently, it
was concluded that the tricyclic nucleosides 1-4 would not
structurally prevent double helix formation when these are
incorporated into the ODNs.
The synthesis of the remaining oxo-amino derivative 4 was
next carried out, which was rather troublesome relative to those
of the other tricyclic nucleosides 1-3. The desired 22 is
presumed to be obtained via the intramolecular cyclization of
13, using the same method as for 15. However, when 13 was
treated with methanolic ammonia at 120 °C for 48 h in a steel
container, the desired product 22 was not obtained, and 13 was
recovered in quantitative yield. As compared with the intra-
molecular cyclization of 14, that of 13 would be more difficult
because of higher electron density of the pyrimidine moiety.17
After several attempts, 22 was obtained efficiently as illustrated
in Scheme 3. First, the 4-carboxamide group of 6 was converted
to the methyl ester according to our previous method.18 Thus,
treatment of 6 with di-tert-butyl dicarbonate in the presence of
Et3N and DMAP gave 18 in 95% yield, which was then treated
with NaOMe in MeOH to give 19 in 98% yield. After the Stille
coupling reaction of 19 with 7 in the presence of dba3Pd2‚CHCl3,
the resulting crude brown syrup (probably including the
intermediates 20 and 21) was treated with a mixture of NH4OH
and 1,4-dioxane at 100 °C for 60 h in a steel container to give
the oxo-amino derivative 22 in 57% yield in three steps.
Deprotection of the silyl groups with TBAF gave the free
nucleoside 4 in 81% yield.
1
3. Characterization of H-Bonding Motif Using H NMR
Spectroscopy. We designed the tricyclic nucleosides 1-4 with
the aim of forming stable four H-bonds in synthetic ODNs. Since
H-bonding interactions in monomers have provided considerable
insight into those of the interior of polymeric units,22 charac-
terization of the H-bonding motifs of the tricyclic nucleosides
was next investigated with 1H NMR studies over a wide
temperature range. The silylated derivatives (i.e., 11, 15, 17,
and 22) were used for 1H NMR measurements in the nonpolar
solvent CDCl3, and monitoring of the H-bonds was done to note
downfield shifts of their amino and amide proton signals.22,23
As shown in Figure 1, the tricyclic nucleosides would form
a heterodimer with four H-bonds between the diamino 11 and
the dioxo 15 derivatives and the amino-oxo 17 and the oxo-
amino 22 derivatives in a parallel direction. In addition to the
heterodimer formation, 17 and 22 are both expected to form a
homodimer with four H-bonds in an antiparallel direction as
shown in Figure 2B. This type of interaction has already been
observed in the NMR spectrum of 17, measured at room
temperature, that is, the amide proton at the 6 position was
observed at 13.77 ppm, and the 4-amino protons were observed
as distinct signals at 10.30 and 6.44 ppm (see Experimental
Procedures). In contrast, the 4-amino protons were observed at
7.80 ppm as a single coalesced signal when the NMR spectrum
of 17 was measured in DMSO-d6 at room temperature. This
result seemed to arise from an intermolecular interaction of 17
in CDCl3. Consequently, the temperature-dependence studies
were first carried out for the homodimers of 17 and 22,
respectively. As an example, the stacked partial 1H NMR spectra
of 22 at variable temperatures in CDCl3 (at 25 mM) are shown
in Figure 2A. No chemical shift changes of H-2, H-9, and H-1′
were observed between (60 °C, while those of the amide and
amino protons were dramatic. In the spectrum at 30 °C, the
2. Conformational Analysis Using NMR and NOE Ex-
periments. The conformation of a nucleoside is critical for
adopting a double helix structure when the nucleoside is
incorporated into ODNs. Generally, a 2′-deoxyribonucleoside
prefers a C2′-endo conformation of the sugar, which is also
maintained in the B-DNA double helix structure.19 Since the
synthetic tricyclic nucleosides possess novel structures, con-
formational properties, such as the sugar pucker mode and the
syn/anti equilibrium around the glycosyl bond, were first
investigated on the nucleoside level. The sugar pucker mode
1
of a nucleoside can be easily predicted from the H NMR
coupling constants of the ribose moiety.20 The coupling
constants (J1′,2′â and J3′,4′) of 1-4 in DMSO-d6 were determined.
In 1, the J1′,2′â was determined as 6.0 Hz, while the J3′,4′ was
3.1 Hz; thus, the C2′-endo conformer was calculated to have a
population of 66%. Similar values for both of the coupling
constants were obtained in compounds 2-4 (see Experimental
Procedures); thus, all of the novel tricyclic nucleosides prefer a
C2′-endo conformation just as the natural deoxyribonucleosides
(68% for 2, 65% for 3, and 64% for 4).
The syn/anti equilibrium around the glycosyl bond was next
examined using NOE experiments.21 As is well-known, one can
(17) As in the case of 13, treatment of 12 with methanolic ammonia did not
give the diamino derivative 11.
(18) Kojima, N.; Minakawa, N.; Matsuda, A. Tetrahedron 2000, 56, 7909-
7914.
(19) Saenger, W. Principles of Nucleic Acid Structure; Springer-Verlag: New
York, 1984; pp 253-282.
(20) For examples: (a) Altona, C.; Sundaralingam, M. J. Am. Chem. Soc. 1973,
95, 2333-2344. (b) Davies, D. B.; Danyluk, S. S. Biochemistry 1974, 13,
4417-4434.
(22) For examples: (a) Katz, L.; Penman, S. J. Mol. Biol. 1966, 15, 220-231.
(b) Shoup, R. R.; Miles, H. T.; Becker, E. D. Biochem. Biophys. Res.
Commun. 1966, 23, 194-201. (c) Newmark, R. A.; Cantor, C. R. J. Am.
Chem. Soc. 1968, 90, 5010-5017.
(23) For examples: (a) Williams, L. D.; Shaw, B. R. Proc. Natl. Acad. Sci.
U.S.A. 1987, 84, 1779-1783. (b) Williams, N. G.; Williams, L. D.; Shaw,
B. R. J. Am. Chem. Soc. 1989, 111, 7205-7209. (c) Williams, L. D.;
Williams, N. G.; Shaw, B. R. J. Am. Chem. Soc. 1990, 112, 829-833. (d)
Sessler, J. L.; Wang, R. J. Org. Chem. 1998, 63, 4079-4091.
(21) For examples: (a) Hart, P. A.; Davis, J. P. J. Am. Chem. Soc. 1969, 91,
512-513. (b) Davis, J. P.; Hart, P. A. Tetrahedron 1972, 28, 2883-2891.
(c) Rosemeyer, H.; Toth, G.; Golankiewicz, B.; Kazimierczuk, Z.;
Bourgeois, W.; Kretschmer, U.; Muth, H. P.; Seela, F. J. Org. Chem. 1990,
55, 5784-5790. (d) Minakawa, N.; Kojima, N.; Matsuda, A. Heterocycles
1996, 42, 149-154. (e) Minakawa, N.; Kojima, N.; Matsuda, A. J. Org.
Chem. 1999, 64, 7158-7172.
9
J. AM. CHEM. SOC. VOL. 125, NO. 33, 2003 9973

A R T I C L E S
Minakawa et al.
on temperature.23b As can be seen in the spectra at 0, -30, and
-60 °C, no significant downfield shift was observed in the
frequency of the Hb signal (less than 0.4 ppm). Therefore, the
equilibrium predominance between the monomer 22 and the
homodimer 22‚22 with four H-bonds was demonstrated, and
formation of the more intricate H-bonded complexes was
excluded in the temperature range from +60 to -60 °C. A
similar phenomenon was also observed with 17 showing
formation of a homodimer (Supporting Information).
To determine the heterodimer formation between 17 and 22,
1H NMR measurements were carried out by mixing 1 equiv of
17 (25 mM) with 1 equiv of 22 (25 mM) in CDCl3. At 60 °C,
two sets of amide protons were observed at 13.41 and 12.91
ppm, respectively, thus indicating that 17 and 22 seem to exist
as monomers at that temperature. As the temperature was
decreased, the spectra became complicated; two sets of ad-
ditional amide protons were observed around 13.5 ppm at -30
°C. In addition, broad singlets assignable to the amino protons
appeared. Such complicated spectra seem to be the result of
formation of both homodimers (17‚17 and 22‚22) and the
heterodimer (17‚22). At -60 °C, the spectrum became clear,
that is, four sets of amide protons ranging from 13.0 to 14.0
ppm and at least seven signals of the amino protons from 5.5
to 10.5 ppm were observed (Supporting Information). Proton
signals of the heterodimer 17‚22 could be assigned by com-
parison with those of the homodimers 17‚17 and 22‚22 at the
same temperature. In Figure 3, the 1H NMR spectrum of 17‚22
at -60 °C is shown along with those of the homodimers. As a
result, the proton signals, except for those of the homodimers,
were easily detected. The proton signals indicated by arrows
would correspond to a new structure formed at low temperature,
and these are assigned to the aglycon protons of the heterodimer
(17‚22) with four H-bonds. Thus, the proton signals at 13.58
and 13.37 ppm are likely to be two sets of amide protons. The
amino protons involved in H-bonding can be seen at 10.24 and
9.49 ppm, while those at 6.78 and 5.68 ppm are assigned as
non-H-bonded amino protons. In addition, four aromatic protons
were also detected between 9.3 and 8.0 ppm. Unfortunately,
the H-bonding motif could not be confirmed by NOE experi-
ments because each signal was too close to irradiate selectively.
However, integral values of the signals corresponding to the
heterodimer were essentially the same as those of the homo-
dimers. Therefore, this also appears to support our findings that
the heterodimer 17‚22, like the homodimers, consists of four
H-bonds, as expected.
1
Figure 2. (A) Partial H NMR spectra obtained in the course of carrying
out a temperature-dependence study of 22 (270 MHz, 25 mM in CDCl3).
(B) Proposed aglycon structure of homodimer (22‚22).
5-amide proton (Hc) was observed at 12.88 ppm, while the
amino protons at the 7 position were not detected. When the
measurement was done at 60 °C, the amide proton signal was
shifted upfield, and the amino protons were observed at 7.16
ppm as a single coalesced signal. These results show that 22
prefers to exist as a monomer at that temperature. On the other
hand, the amide proton signal was shifted downfield with
decreasing temperatures. In addition, the amino protons were
split into two sets of broad singlets, clearly observed at 9.33
ppm (Ha) and 5.79 ppm (Hb) at -60 °C in the spectrum.
Therefore, the proton signals at 13.14 and 9.33 ppm would be
involved in H-bonding, while that at 5.79 ppm is not involved
in the H-bonding. On the basis of these results, 22 is considered
to form the homodimer 22‚22 in an antiparallel direction with
four H-bonds at low temperature as illustrated in Figure 2B.
This consideration was further confirmed by NOE experiments
at -60 °C. Thus, a strong NOE (40.4%) was observed at the
H-bonded amino proton (Ha) upon irradiation of the amide
proton (Hc).
In contrast to the experiments using 17 and 22, no obvious
interaction between 11 and 15 was detected in the temperature-
dependence studies. Neither amide nor amino proton signals
were observed at 30 °C. These proton signals were still
undetectable, except for one broad signal at 7.06 ppm, at 60
°C. Although several broad signals were observed when the
temperature decreased, the spectrum was too complicated and
broad to assign them (Supporting Information). However, 11
and 15 were thought to interact with each other through
H-bonding because the insoluble dioxo derivative 15 in CDCl3
was clearly dissolved upon addition of an equimolar amount of
11. In addition, this type of interaction was suggested to occur
even at 60 °C. In Figure 4, two partial spectra of the mixed
sample (11:15 ) 1:1) and 11 at 60 °C are shown. Two sets of
the amino proton signals were observed at 5.90 and 5.12 ppm,
Since the resulting homodimer 22‚22 still has H-bonding
abilities on its N-3 and O-4 positions as a proton acceptor and
the 7-amino group (Hb) as a proton donor, formation of more
intricate H-bonded complexes such as tetramers may be possible.
However, this is unlikely at least above -60 °C. This being
the case, the Hb of 22 would also be involved in the H-bonding;
thus, the Hb signal is expected to move downfield, depending
9
9974 J. AM. CHEM. SOC. VOL. 125, NO. 33, 2003

Synthesis and Thermal Stability of Oligodeoxynucleotides
A R T I C L E S
parallel and antiparallel type, would be expected between them.
In addition, formation of some other complexes (i.e., tetramer,
oligomer, etc.) might occur in the monomers. These interactions
may have caused the complicated spectra.
As is demonstrated, the tricyclic nucleosides 17 and 22
formed base pairs with four H-bonds. In contrast, no obvious
evidence of base pairing with four H-bonds was obtained
between 11 and 15 in monomers. However, the geometry and
direction of the base pairing should be controllable when these
nucleosides are introduced into ODNs. Therefore, further
investigation of dioxo and diamino base pairing in polymers
would be worth evaluating. The base pairing abilities of the
tricyclic nucleosides in ODNs including comparison with natural
Watson-Crick base pairing was next examined.
4. Synthesis of Oligodeoxynucleotides. After conversion of
1-4 into the corresponding phosphoramidites with appropriate
protecting groups on each nucleobase (Supporting Informa-
tion),24 ODNs were synthesized on a DNA/RNA synthesizer
by the phosphoramidite method.25 Two sets of complementary
17mers (ODN I and II and ODN III and IV) were designed to
investigate the H-bonding abilities of the tricyclic nucleosides,
where the NN, OO, NO, and ON were incorporated in the X and
Y positions (see the sequences in Figures 5, 6, and 8). The fully
protected ODNs (1 µmol) linked to the solid support were
treated with concentrated NH4OH at 55 °C for 16 h, followed
by C-18 column chromatography. The ODNs were detritylated
with 80% aqueous AcOH, and each ODN obtained showed a
single peak on reverse-phase HPLC. Each ODN was completely
hydrolyzed to the corresponding nucleosides by a mixture of
nuclease P1, snake venom phosphodiesterase, and alkaline
phosphatese, and the nucleoside composition was analyzed by
HPLC. The peak corresponding to the tricyclic nucleosides NN,
OO, NO, or ON, confirmed by coelution with authentic samples,
was clearly observed, and the nucleoside composition calculated
from the areas of the peaks supported the structure of each
(Supporting Information).
1
Figure 3. (A) Partial H NMR spectra of 17 (top), 22 (bottom), and 1:1
mixtures of 17 and 22 (middle) at -60 °C (270 MHz, 25 mM in CDCl3).
(B) Proposed aglycon structure of heterodimer (17‚22).
5. Thermal Stability of the DNA-DNA Duplexes Contain-
ing Base Pairings between Imidazopyridopyrimidine Bases.
Thermal stability of duplexes formed by ODN I and comple-
mentary ODN II, which contained one molecule of the NN, OO,
NO, or ON in their X or Y position, was first studied by thermal
denaturation in a buffer of 0.01 M sodium cacodylate (pH 7.0)
containing 0.1 M NaCl. Each profile of the thermal denaturation
showed a single transition corresponding to a helix-to-coil
transition to give melting temperatures (Tm). The thermodynamic
parameters (∆H°, ∆S°, and ∆G°) of the duplexes were
determined by calculations based on the slope of a 1/Tm versus
ln(CT/4) plot,26 where CT is total concentration of single strands
(Supporting Information). The melting temperature (Tm) values
represented by bar graph and negative free energy (-∆G°298
)
Figure 4. (A) Partial 1H NMR spectra of 1:1 mixtures of 11 and 15 (top)
and 11 (bottom) at 60 °C (270 MHz, 10 mM in CDCl3).
are shown in Figure 5. Differences in the Tm values of the
duplexes containing the tricyclic nucleosides were quite small
(5.6 °C; maximum, 59.6 °C; minimum, 54.0 °C), and no
each with an integral of two, in the spectrum of the diamino
nucleoside 11. On the other hand, a broad signal was observed
at 7.06 ppm instead of these amino proton signals in that of the
mixed sample. Although there is no direct evidence for the
heterodimer formation between 11 and 15 consisting of four
H-bonds, interactions between 11 and 15 through H-bonding
are indicated by the above results. Unlike 17 and 22, formation
of stable homodimers are unlikely for both 11 and 15. However,
two sets of base pairing motifs with four H-bonds, namely, the
(24) References for the synthesis of phosphoramidites: (a) Kamaike, K.;
Hasegawa, Y.; Ishido, Y. Nucleosides Nucleotides 1988, 7, 37-43. (b)
Froehler, B. C.; Matteucci, M. D. Nucleic Acids Res. 1983, 11, 8031-
8036. (c) Caruthers, M. H.; McBride, L. J.; Bracco, L. P.; Dubendorff, J.
W. Nucleosides Nucleotides 1985, 4, 95-105.
(25) Sinha, N. D.; Biernat, J.; Koster, H. Tetrahedron Lett. 1983, 24, 5843-
5846.
(26) Breslauer, K. J. In Methods in Molecular Biology, Volume 26, Protocols
for Oligonucleotide Conjugates; Agrawal, S., Ed.; Humana Press Inc.:
Totowa, New Jersey, 1994; Ch. 14, pp 347-372.
9
J. AM. CHEM. SOC. VOL. 125, NO. 33, 2003 9975

A R T I C L E S
Minakawa et al.
Figure 5. Hybridization data of ODNs I:II (X and Y ) NN, OO, NO, or
ON). Experimental conditions are described in Experimental Procedures.
Figure 7. Possible base pairing motifs between imidazopyridopyrimidine
bases.
Tm of 70.6 °C. The duplex containing the NO:NO pairs (Tm
)
Figure 6. Hybridization data of ODNs III:IV (X and Y ) NN, OO, NO, or
ON). Experimental conditions are described in Experimental Procedures.
70.0 °C) was also stabilized to the same degree as that of the
NN:OO pairs. These Tm values are distinctly lower than that of
the NO:ON pair and rather close to the G:C pair. Consequently,
we propose their possible base pairing motifs to be those shown
in Figure 7. As is supported in the temperature dependent NMR
study, NO and ON can form base pairs with four H-bonds even
when they are incorporated into ODNs to give the thermally
and thermodynamically stable duplex. In the case of the NN:
OO pair, a reasonable explanation for the above result would
be that this base pair has fewer H-bonds than the expected four
(Figure 1). A tautomeric form of OO, represented as tOO, can
be considered, and this could form a base pair with NN by three
H-bonds, as shown in Figure 7 (NN:tOO). The Tm value of the
duplex containing the NO:NO pair predicts that this pair also
has three H-bonds. These Tm values are in good agreement with
those of the G:C pair, which has three H-bonds, implying that
the proposed base pairing motifs are reasonable. Except for these
three pairs, the ON:ON and OO:ON (represented as tOO:ON in
Figure 7) pairs also stabilized the duplexes relative to the A:T
pair, and the proposed motifs are illustrated. The remaining five
combinations were less stable than the duplex containing the
A:T pair; however, all of them were more stable than the
duplexes containing the A:G and A:A mismatched pairs. Since
the Tm of the duplex containing NO:ON went up dramatically
and the ∆Tm [Tm (NO:ON) - Tm (NN:NO)] was quite large (34.6
°C), it can be concluded that the H-bonding abilities between
the imidazopyridopyrimidine bases are essential and affect the
thermal stability of the duplex. The attached -∆G°298 data also
apparent specificity of base pairing was observed. Although the
duplex with the NO:ON pair had the highest Tm value (59.6 °C),
this value was lower than not only that of the duplex with the
G:C pair (Tm ) 61.9 °C) but also that with the A:T pair (Tm )
60.5 °C). Thus, contrary to our expectation, when one molecule
of the tricyclic nucleosides was incorporated into each strand,
the thermal and thermodynamic stabilities of the duplexes did
not increase. The Tm values of the duplexes formed by ODN I
and II were all slightly lower than those of the duplexes
containing a natural G:C or A:T pair at the corresponding
positions.
On the other hand, when three molecules of the tricyclic
nucleosides were consecutively incorporated into the center of
each ODN (ODNs III and IV), the thermal and thermodynamic
stabilization of the duplexes due to the specific base pairings
was observed. As shown in Figure 6, the best result was obtained
in the case of the duplex containing the NO:ON pair (Tm ) 84.0
°C). The value was between 18.2 and 23.5 °C higher than those
of the duplexes containing three consecutive G:C (Tm ) 65.8
°C) and A:T pairs (Tm ) 60.5 °C), respectively. Consequently,
it was found that the NO:ON pair stabilized the duplex by about
+6 and +8 °C per modification as compared with those of the
duplexes containing the G:C and A:T pairs, respectively. As
described in section 3, no obvious H-bonding motif was
observed between NN and OO in the monomers. In this
experiment, however, the NN:OO pair showed a rather higher
9
9976 J. AM. CHEM. SOC. VOL. 125, NO. 33, 2003

Synthesis and Thermal Stability of Oligodeoxynucleotides
A R T I C L E S
Figure 8. Hybridization data of ODNs III:IV (X ) NN, OO, NO, or ON, Y
) A, C, G, or T). Experimental conditions are described in Experimental
Procedures.
support this conclusion. A stacking interaction influence, arising
from the extended aromatic surface of the imidazopyrido-
pyrimidine bases, may also contribute to their stabilities.27,28
This will be discussed in section 7 of this paper.
Figure 9. Possible base pairing motifs between imidazopyridopyrimidine
6. Thermal Stability of the DNA-DNA Duplexes Contain-
ing Base Pairings between the Imidazopyridopyrimidine
Bases and the Natural Bases. In the previous section, we
showed that consecutive incorporation of NO:ON pairs into the
duplexes markedly enhanced the thermal and thermodynamic
stability of the duplex by four H-bonds. Since the imidazo-
pyridopyrimidine bases possess multiple types of H-bonding
ability, formation of stable base pairs with a natural base was
predicted. Thus, an investigation of base pairing with natural
bases was next undertaken using ODNs I:II and III:IV, where
NN, OO, NO, or ON was introduced at the X position, and A, C,
G, or T was introduced at the Y position. When ODNs I and II
were used, as in the preceding results, no base pair was found
more stable than the G:C pair (data not shown). In contrast, a
striking change was observed in the investigations of ODNs
III:IV. The Tm values represented by bar graph and negative
free energy (-∆G°298) are shown in Figure 8. Among the 16
combinations, the duplex containing the NN:G pair, which had
the highest Tm (72.8 °C), was stabilized by +7.0 °C relative to
the duplex containing the G:C pair (+2.3 °C per modification).
The OO also preferred G to give a thermally stable duplex equal
to the duplex containing the G:C pair (65.6 vs 65.8 °C), while
NO and ON showed different preferences in base pairing. In the
case of NO, the most stable pair was NO:T (Tm ) 62.0 °C). The
ON preferred both A and T to give the Tm values of 60.9 °C
(ON:A) and 63.0 °C (ON:T), respectively. Interestingly, no stable
pair with C was observed. If the OO took the form shown in
base and natural base.
Figure 1, this form should make a stable base pair with C. In
fact, however, G was the best partner of OO but not of C. These
results strongly suggest that the tautomeric form tOO is
preferable, as discussed in the previous section. The possible
base pairing motifs of the imidazopyridopyrimidine base with
the natural bases, which had higher Tm values than the duplex
containing A:T pair, are illustrated in Figure 9. Consequently,
it was found that the imidazopyridopyrimidine base also forms
base pairs with the natural bases with some sequence specifici-
ties.
7. Measurement of Stacking Abilities of the Imidazo-
pyridopyrimidine Bases. From the results described in sections
5 and 6, it was shown that the H-bonding abilities of the
imidazopyridopyrimidine base are quite essential for the stability
of the duplex. In this section, we discuss the effects of the
stacking abilities of the imidazopyridopyrimidine bases. As
mentioned above, these nucleobases have extended aromatic
surfaces, and this would also contribute to the duplex stabil-
ity.27,28 For example, one can see stabilization of the duplex by
the NN:G pair relative to the G:C pair, despite each pair being
predicted to have the same number of H-bonds (Figure 9). This
result undoubtedly implies that the stacking interaction would
also be an important contributing factor for the stability. For
this reason, evaluation of the stacking abilities was next
examined. According to the method reported by Guckian et al.,28
a series of duplexes, where Z was added at the end of each
paired duplex (dangling end), were prepared, and the Tm values
and the thermodynamic parameters of the duplexes were
determined. Results of the thermodynamic parameters made in
a buffer of 0.01 M sodium phosphate (pH 7.0) containing 1 M
NaCl are presented in Table 1. As a comparison, the Tm data
(27) For examples: (a) Lin, K.-Y.; Jones, R. J.; Matteucci, M. J. Am. Chem.
Soc. 1995, 117, 3873-3874. (b) Brotschi, C.; Haberli, A.; Leumann, C. J.
Angew. Chem., Int. Ed. 2001, 40, 3012-3014. (c) Hashmi, S. A. N.; Hu,
X.; Immoos, C. E.; Lee, S. J.; Grinstaff, M. W. Org. Lett. 2002, 4, 4571-
4574.
(28) Guckian, K. M.; Schweitzer, B. A.; Ren, R. X.-F.; Sheils, C. J.; Paris, P.
L.; Tahmassebi, D. C.; Kool, E. T. J. Am. Chem. Soc. 1996, 118, 8182-
8183.
9
J. AM. CHEM. SOC. VOL. 125, NO. 33, 2003 9977

A R T I C L E S
Minakawa et al.
Table 1. Measurement of Aromatic Stacking Abilities
Table 2. Sequences of ODNs and Hybridization Data of ODN
V:VI-XII
a
−∆G°298
Tm (°C)a
∆Tm (°C)b
(kcal/mol)
∆∆G°298
b
Z
none
42.7
47.3
47.5
52.2
52.4
57.0
52.5
58.8
57.4
10.4
11.4
11.7
12.8
12.9
14.6
12.8
16.0
14.2
thymine (T)
cytosine (C)
adenine (A)
guanine (G)
NN
4.6
4.8
9.5
9.7
14.3
9.8
1.0
1.3
2.4
2.5
4.2
2.4
5.6
3.8
OO
NO
16.1
14.7
a Experimental conditions are described in Experimental Procedures.
ON
b Sequences are same as ODN VI except for NN.
a Experimental conditions are described in Experimental Procedures.
b Stacking parameters (∆Tm, ∆∆G°) were obtained by subtracting data for
the core hexamer duplex from that for duplexes possessing the unpaired
(Z) unit.
positions, respectively, was carried out. The duplex containing
the IQ:ON pair (Tm ) 47.2 °C) was drastically destabilized by
-36.8 °C relative to the duplex containing the NO:ON pair (Tm
) 84.0 °C; Figure 6). In contrast, the stacking ability of IQ,
estimated by the Tm of the two unpaired IQ at the Z position
(Tm ) 61.5 °C), was higher than that of the NO (Tm ) 58.8 °C;
Table 1). Consequently, in our substrates, it can be concluded
that the H-bonding ability is the essential factor in the high
duplex stability, for example, with the NO:ON pair, and the
stacking ability is a minor component of the stability.
8. Why Consecutive Introduction of the Imidazopyrido-
pyrimidine Bases Markedly Enhances the Thermal and
Thermodynamic Stability of the Duplexes. As described
above, introduction of one molecule of the imidazopyrido-
pyrimidine base slightly destabilizes the duplex despite prefer-
able base pairings such as NO:ON and NN:G is introduced. In
contrast, three consecutive introductions of these base pairs
markedly stabilizes the duplexes. To elucidate these results, we
further examined the sequence dependence of Tm using the
duplexes containing the NN:G pair. Thus, setting the Tm of the
duplex between ODNs V and VI as a control (68.3 °C), the
sequence dependent variation was compared by substituting the
G:C pair with the G:NN pair at various positions (ODNs V:VII-
XII). The sequences of ODNs and resulted Tm values are shown
in Table 2. It was found that the duplexes became thermally
less stable as the numbers of nonconsecutive isolated introduc-
tions of the G:NN pairs increased (ODNs VII-X). On the other
hand, the duplexes containing the consecutive G:NN pairs were
more stable than the control duplexes (ODNs XI and XII).
Through comparison of the results between ODNs V:IX (Tm )
63.0 °C) and V:XI (Tm ) 72.8 °C), it was revealed that
consecutive introductions of a preferable base pairing motif is
essential in contributing to the thermal stability of the duplex.
We considered these results, and the hypothetical structures
are illustrated in Figure 11. The base pairs such as NO:ON and
NN:G form stable H-bonds in the duplex. However, the pairs
would spread the width of the helix around the strand where
they were introduced. Thus, it is known that the average
intrastrand C1′-C1′ distance in a canonical Watson-Crick base
pair is 10.5 ((0.2) Å, while the distances in purine(anti)-
purine(anti) base pairs are typically much longer.31 For example,
Prive´ et al. reported the C1′-C1′ distance in the G(anti)-A(anti)
mismatch to be 12.5 Å in an X-ray structure.32 The intrastrand
Figure 10. Structure of IQ.
for nondangling and natural bases (T, C, A, or G) at the dangling
end are listed. The two unpaired Ts added +4.6 °C of Tm and
+1.0 kcal/mol of -∆G°298 relative to the self-complementary
duplex (Z ) none), and the Cs also added a similar stability.
The unpaired purines (Z ) A and G) afforded about 2-fold the
duplex stability relative to the pyrimidines arising from expan-
sion of the aromatic surface. These results agreed with those
reported by Guckian et al. Among the tested compounds, NO
showed the highest Tm increase over the unsubstituted duplex
of 16.1 °C (8.0 °C per base) and a total stacking energy of 5.6
kcal/mol (2.8 kcal/mol per base). The NN and ON also showed
almost 1.5× higher stacking ability than those of the natural
purines and 3× those of pyrimidines. Even though the efficacy
of OO was slightly low,29 it was almost equal to those of adenine
and guanine. From these results, it was revealed that the stacking
abilities of the imidazopyridopyrimidine bases also increase the
duplex stability, and this would be enhanced by three consecu-
tive introductions of these bases into the ODNs.
To clarify whether the H-bonding ability or the stacking
ability of the imidazopyridopyrimidine bases is the essential
factor for the duplex stability, we prepared the imidazoquinoline
derivative 23 (IQ; Figure 10).30 Compound 23 was designed to
eliminate the H-bonding ability except for one position but still
have the same size of aromatic surface. Using this nucleoside
unit, the Tm measurement of the duplex composed of ODNs III
and IV, where IQ and ON were introduced at the X and Y
(29) Although relationships between stacking abilities of the imidazopyrido-
pyrimidine bases and their physical properties such as log P, dipole moment,
and surface area were examined, no obvious result was obtained to explain
the low stacking ability of OO; see ref 11.
(30) The synthesis of IQ was done via the Stille coupling reaction of 1-(2-
deoxy-3,5-di-O-triisopropylsilyl-â-D-ribofuranosyl)-5-iodoimidazole-4-car-
bonitrile with N-acetyl-2-tributylstannylaniline; see Supporting Information.
(31) Cote´, M. L.; Georgiadis, M. M. Acta Crystallogr. 2001, D57, 1238-1250.
(32) Prive´, G. G.; Heinemann, U.; Chandrasegaran, S.; Kan, L.-S.; Kopka, M.
L.; Dickerson, R. E. Science 1987, 238, 498-504.
9
9978 J. AM. CHEM. SOC. VOL. 125, NO. 33, 2003

Synthesis and Thermal Stability of Oligodeoxynucleotides
A R T I C L E S
when the favorable base pairs were introduced consecutively
into the duplex. Of course, our conclusion will not be applicable
in all cases. In the case of an unfavorable base pair such as
NN:NO, flip-out of either base from the duplex may minimize
its destabilization.
In conclusion, the synthesis of four imidazo[5′,4′:4,5]pyrido-
[2,3-d]pyrimidine nucleosides 1-4 has been accomplished via
the Stille coupling of 5-iodoimidazole nucleosides with an
appropriate 5-stannylpyrimidine derivative. Using ODNs con-
taining these tricyclic nucleosides, preferable base pairing motifs
were investigated. When ODNs containing one residue of the
nucleoside were used, no base pairing motif was found to be
more stable than either the G:C or the A:T pair. On the other
hand, three consecutive introductions of the nucleosides were
quite effective, and the NO:ON pair markedly stabilized the
duplex. The ∆Tm as compared with the G:C pair was +18.2
°C, and this pair stabilized the duplex by about +6 °C per
modification. We presented the results that formation of four
H-bonds of the NO:ON pair and the stacking ability also
contributed to the increased stabilization. We expected that the
NO:ON pair will find use in the stabilization and regulation of
a variety of DNA structures depending on positions where it is
incorporated. The use of the imidazopyridopyrimidine nucleo-
sides in nucleic acid chemistry should be the focus of further
study.
Figure 11. Hypothetical structures of our consideration.
C1′-C1′ distance in NO:ON and NN:G would be similar to the
above distance. Therefore, we speculated that thermal and
thermodynamic destabilization of the duplex occurs at both sides
of these pairs. When one pair of the tricyclic nucleoside was
incorporated into the duplex, the destabilization factor arising
from disruption of the Watson-Crick base pairs next to the
pair would be greater than the stabilization arising from the
stable H-bonds. Consequently, the duplex would become less
stable (Figure 11A). This undesirable effect would be enhanced
as the numbers of the nonconsecutive tricyclic nucleosides
increased (Figure 11B). On the other hand, when the tricyclic
nucleosides were consecutively incorporated into the duplex,
the duplex would be thermally and thermodynamically stabilized
to a great extent since the base pairs between the tricyclic
nucleosides has stable H-bonds and a strong stacking ability
with the adjacent bases. These stabilization factors would be
superior to the conformational destabilization around the bound-
ary of the base pairs (Figure 11C). To prove these consider-
ations, an ODN, 5′-GCACCGTNONONOTACCACG-3′, was
prepared, and a Tm with ODN IV (Y ) ON), 3′-CGTGGC-
TONONONTTGGTGC-5′, which contains two T-T mismatches
next to the tricyclic nucleosides, was measured. If the above
considerations were correct, this duplex would show a similar
Tm relative to the complementary duplex (ODNs III:IV; X )
NO, Y ) ON) because the H-bonding interaction of the base
pair adjacent to the NO:ON pair would be negligible. In fact,
the Tm obtained from this duplex was 76.2 °C, which was 7.8
°C less stable than that of the complementary duplex (84.0 °C).
However, the degree of destabilization of this duplex seems
rather low despite the duplex containing the two T:T mis-
matches.33 This result implied that the H-bonding interaction
of the base pair adjacent to the favorable base pair such as NO:
ON and NN:G became weaker, thereby acting as a destabilizing
factor in the duplex. It can be concluded that the thermal stability
of the duplex would depend on a balance of stabilization and
destabilization factors and that the duplex is markedly stabilized
Experimental Procedures
General Methods. Physical data were measured as follows; melting
points are uncorrected. 1H and 13C NMR spectra were recorded at 270,
400, or 500 MHz and 67.5, 100, or 125 MHz in CDCl3 or DMSO-d6
as the solvent with tetramethylsilane as an internal standard. Chemical
shifts are reported in parts per million (δ), and signals are expressed
as s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), or br
(broad). All exchangeable protons were detected by addition of D2O.
1
Assignment of H signals was based on two-dimensional NMR and
NOE experiments. Mass spectra were measured on a JEOL JMS-D300
spectrometer. TLC was done on Merck Kieselgel F254 precoated plates.
Silica gel used for column chromatography was Merck silica gel 5715.
4-Amino-7-chloro-1-(2-deoxy-3,5-di-O-triisopropylsilyl-â-D-ribo-
furanosyl)imidazo[5′,4′:4,5]pyrido[2,3-d]pyrimidine (10). To a mix-
ture of 5 (2.20 g, 3.4 mmol) and 7 (2.13 g, 5.1 mmol) in DMF (25
mL) was added dba3Pd2‚CHCl3 (280 mg, 0.27 mmol), and the whole
mixture was heated at 100 °C for 11 h. The reaction mixture was filtered
through a Celite pad and washed with AcOEt. The combined filtrate
and washings were diluted with AcOEt and washed with H2O (3×),
followed by brine. The organic layer was dried (Na2SO4) and
concentrated in vacuo to give a crude mixture of 9 and 10. The resulting
mixture was dissolved in a mixture of EtOH-5% aqueous Na2CO3, and
the reaction mixture was heated at 80 °C for 1 h. The solvent was
removed in vacuo, and the residue was partitioned between CHCl3 and
H2O. The separated organic layer was washed with H2O, followed by
brine. The organic layer was dried (Na2SO4) and concentrated in vacuo.
The residue was purified by a silica gel column and eluted with 0-5%
EtOH in CHCl3, to give 10 (1.53 g, 69%, crystallized from acetone-
hexane): mp 262-264 °C; FAB-LRMS m/z 649 (MH+); 1H NMR
(CDCl3) δ: 9.36 (s, 1 H), 8.21 (s, 1 H), 6.60 (br s, 2 H), 6.52 (dd, 1
H, J ) 7.8, 5.2 Hz), 4.83 (ddd, 1 H, J ) 5.0, 2.7, 1.4 Hz), 4.24 (ddd,
1 H, J ) 1.4, 2.8, 3.6 Hz), 3.88 (dd, 1 H, J ) 2.8, 11.3 Hz), 3.84 (dd,
1 H, J ) 3.6, 11.3 Hz), 2.71 (ddd, 1 H, J ) 7.8, 12.5, 5.0 Hz), 2.65
(ddd, 1 H, J ) 5.2, 12.5, 2.7 Hz), 1.11-0.98 (m, 42 H); 13C NMR
(CDCl3) δ: 159.60, 159.05, 157.46, 153.91, 139.31, 131.75, 128.85,
107.06, 89.32, 86.30, 72.59, 63.28, 41.21, 18.02, 17.89, 12.12, 11.84.
Anal. Calcd. for C31H53ClN6O3Si2: C, 57.33; H, 8.23; N, 12.94.
Found: C, 57.49; H, 8.31; N, 12.73.
(33) Insertion of one T:T mismatch in ODNs destabilized the duplex by -15
°C, see Peyret, N.; Seneviratne, A.; Allawi, H. T.; SantaLucia, J., Jr.
Biochemistry 1999, 38, 3468-3477.
9
J. AM. CHEM. SOC. VOL. 125, NO. 33, 2003 9979

A R T I C L E S
Minakawa et al.
4,7-Diamino-1-(2-deoxy-3,5-di-O-triisopropylsilyl-â-D-ribo-
furanosyl)imidazo[5′,4′:4,5]pyrido[2,3-d]pyrimidine (11). To a solu-
tion of 10 (410 mg, 0.63 mmol) in 1,4-dioxane (20 mL) was added
NH4OH (28%, 20 mL), and the whole was heated at 100 °C for 24 h
in a steel container. The solvent was removed in vacuo, and the residue
was partitioned between CHCl3 and H2O. The separated organic layer
was washed with H2O, followed by brine. The organic layer was dried
(Na2SO4) and concentrated in vacuo. The residue was purified by a
silica gel column and eluted with 0-16% EtOH in CHCl3, to give 11
(338 mg, 86%, crystallized from acetone): mp >270 °C (dec); FAB-
LRMS m/z 630 (MH+); 1H NMR (CDCl3) δ: 9.13 (s, 1 H), 8.06 (s, 1
H), 6.48 (dd, 1 H, J ) 7.7, 5.3 Hz), 5.91 and 5.14 (each br s, each 2
H), 4.81 (ddd, 1 H, J ) 5.0, 3.0, 1.6 Hz), 4.21 (ddd, 1 H, J ) 1.6, 2.9,
3.8 Hz), 3.90 (dd, 1 H, J ) 2.9, 11.2 Hz), 3.84 (dd, 1 H, J ) 3.8, 11.2
Hz), 2.65 (ddd, 1 H, J ) 7.7, 12.4, 5.0 Hz), 2.60 (ddd, 1 H, J ) 5.3,
12.4, 3.0 Hz), 1.11-1.00 (m, 42 H); 13C NMR (CDCl3) δ: 162.01,
159.93, 156.81, 153.89, 137.52, 133.17, 126.19, 101.50, 88.90, 86.08,
72.32, 63.32, 41.13, 18.02, 17.91, 12.13, 11.86. Anal. Calcd. for
C31H55N7O3Si2: C, 59.10; H, 8.80; N, 15.56. Found: C, 59.00; H, 8.75;
N, 15.37.
organic layer was washed with H2O, followed by brine. The organic
layer was dried (Na2SO4) and concentrated in vacuo. The residue was
purified by a silica gel column and eluted with 0-5% MeOH in CHCl3,
to give 13 (1.89 g, 92% as a white foam): FAB-LRMS m/z 663 (MH+);
1H NMR (CDCl3) δ: 8.17 (s, 2/3 H), 8.05 (s, 1/3 H), 7.87 (s, 1/3 H),
7.86 (s, 2/3 H), 7.01 and 5.24 (each br s, each 1 H), 5.77 (dd, 1/3 H,
J ) 7.4, 6.0 Hz), 5.60 (dd, 2/3 H, J ) 8.0, 5.6 Hz), 5.12 (br s, 4/3 H),
5.10 (br s, 2/3 H), 4.70 (m, 2/3 H), 4.58 (m, 1/3 H), 4.03 (m, 1/3 H),
3.93 (m, 2/3 H), 3.89-3.82 (m, 2 H), 3.87 (s, 1 H), 3.84 (s, 2 H), 2.41
(ddd, 2/3 H, J ) 8.0, 12.9, 5.3 Hz), 2.34 (ddd, 2/3 H, J ) 5.6, 12.9,
2.3 Hz), 2.08-2.06 (m, 2/3 H), 1.09-1.00 (m, 42 H). Anal. Calcd. for
C32H58N6O5Si2: C, 57.97; H, 8.82; N, 12.68. Found: C, 57.92; H, 8.88;
N, 12.55.
5-[4-Methoxypyrimidin-2(1H)-on-5-yl]-1-(2-deoxy-3,5-di-O-triiso-
propylsilyl-â-D-ribofuranosyl)imidazole-4-carboxamide (14). To a
solution of 13 (1.58 g, 2.39 mmol) in THF-H2O (70-0.7 mL) was
added isoamyl nitrite (3.2 mL, 23.9 mmol) at 60 °C, and the mixture
was stirred at the same temperature. After 1.5 h, an additional isoamyl
nitrite (0.8 mL, 6.0 mmol) was added to the reaction mixture, and the
mixture was stirred for additional 1.5 h. The reaction mixture was
neutralized by addition of saturated aqueous NaHCO3, and the solvent
was removed in vacuo. The residue was partitioned between AcOEt
and H2O, and the separated organic layer was washed with H2O,
followed by brine. The organic layer was dried (Na2SO4) and
concentrated in vacuo. The residue was purified by a silica gel column
and eluted with 0-10% EtOH in CHCl3, to give 14 (804 mg, 51% as
a white foam): FAB-LRMS m/z 664 (MH+); 1H NMR (CDCl3) δ:
12.07 (br s, 1/3 H), 11.75 (br s, 2/3 H), 7.91 (s, 1/3 H), 7.88 and 7.80
(each s, each 2/3 H), 7.72 (s, 1/3 H), 7.11 and 5.89 (each br s, each 1/3
H), 7.07 and 5.80 (each br s, each 2/3 H), 5.64 (m, 1 H), 4.72 (m, 2/3
H), 4.64 (m, 1/3 H), 4.05-3.82 (m, 3 H), 3.96 (s, 1 H), 3.94 (s, 2 H),
2.48 (ddd, 2/3 H, J ) 7.3, 13.1, 5.5 Hz), 2.34 (m, 2/3 H), 2.17 (m, 1/3
H), 2.11 (m, 1/3 H), 1.09-1.03 (m, 42 H). Anal. Calcd. for C32H57N5O6-
Si2: C, 57.88; H, 8.65; N, 10.55. Found: C, 57.62; H, 8.60; N, 10.44.
4,7-Diamino-1-(2-deoxy-â-D-ribofuranosyl)imidazo[5′,4′:4,5]-
pyrido[2,3-d]pyrimidine (1). To a solution of 11 (100 mg, 0.16 mmol)
in THF (15 mL) was added TBAF (1 M, 0.4 mL, 0.4 mmol) at 0 °C,
and the reaction mixture was stirred at room temperature for 1.5 h.
After addition of AcOH (23 µL, 0.4 mmol), the solvent was removed
in vacuo, and the residue was partitioned between H2O and CHCl3.
The separated aqueous layer was washed with AcOEt and concentrated
in vacuo. The residue was coevaporated with EtOH, and a mixture of
acetone-EtOH (5:1, 5 mL) was added to the resulting solid. The
suspension was heated in a water bath and then cooled to room
temperature. The precipitate was collected and washed with EtOH to
give 1 (45 mg, 88% as a white powder): FAB-LRMS m/z 318 (MH+);
UV λmax (H2O) 341 nm (ꢀ ) 12 800), 254 nm (ꢀ ) 28 700), 232 nm
(ꢀ ) 37 600); λmax (0.5 N HCl) 353 nm (ꢀ ) 10 800), 249 nm (ꢀ )
42 400); λmax (0.5 N NaOH) 341 nm (ꢀ ) 12 600), 255 nm (ꢀ )
31 400), 231 nm (ꢀ ) 41 400); 1H NMR (DMSO-d6) δ: 8.95 (s, 1 H),
8.38 (s, 1 H), 7.19 and 6.41 (each br s, each 2 H), 6.53 (t, 1 H, J ) 6.0
Hz), 5.40 (d, 1 H, J ) 6.9 Hz), 4.90 (t, 1 H, J ) 5.1 Hz), 4.39 (dddd,
1 H, J ) 6.9, 6.6, 5.2, 3.1 Hz), 3.94 (ddd, 1 H, J ) 3.1, 4.3, 4.2 Hz),
3.51 (ddd, 1 H, J ) 5.1, 4.3, 11.8 Hz), 3.47 (ddd, 1 H, J ) 5.1, 4.2,
11.8 Hz), 2.72 (ddd, 1 H, J ) 6.0, 13.2, 6.6 Hz), 2.47 (ddd, 1 H, J )
6.0, 13.2, 5.2 Hz); 13C NMR (DMSO-d6) δ: 162.2, 159.6, 156.7, 153.4,
138.3, 132.4, 125.2, 100.0, 87.9, 85.4, 69.9, 61.0. Anal. Calcd. for
C13H15N7O3‚3/5H2O‚1/10EtOH: C, 47.65; H, 5.09; N, 29.47. Found:
C, 47.83; H, 4.85; N, 29.35.
1-(2-Deoxy-3,5-di-O-triisopropylsilyl-â-D-ribofuranosyl)imidazo-
[5′,4′:4,5]pyrido[2,3-d]pyrimidine-4,7(5H,6H)-dione (15). A solution
of 14 (800 mg, 1.20 mmol) in methanolic ammonia (saturated at 0 °C,
40 mL) was heated at 120 °C for 48 h in a steel container. The solvent
was removed in vacuo, and the residue was crystallized from EtOH-
CHCl3 to give 15 (553 mg, 73%): mp >300 °C; FAB-LRMS m/z 632
1
(MH+); H NMR (DMSO-d6) δ: 12.16 and 11.68 (each br s, each 1
H), 8.45 (s, 1 H), 8.21 (s, 1 H), 6.41 (dd, 1 H, J ) 5.4, 4.9 Hz), 4.70
(ddd, 1 H, J ) 5.0, 7.3, 4.3 Hz), 3.95 (ddd, 1 H, J ) 4.3, 3.8, 4.1 Hz),
3.74 (dd, 1 H, J ) 3.8, 11.4 Hz), 3.59 (dd, 1 H, J ) 4.1, 11.4 Hz),
3.04 (ddd, 1 H, J ) 5.4, 13.1, 5.0 Hz), 2.50 (ddd, 1 H, J ) 4.9, 13.1,
7.3 Hz), 1.05-0.84 (m, 42 H); 13C NMR (DMSO-d6) δ: 158.1, 138.2,
132.0, 129.4, 87.9, 84.5, 70.8, 62.2, 38.6, 17.6, 17.3, 11.6, 11.2. Anal.
Calcd. for C31H53N5O5Si2: C, 58.92; H, 8.45; N, 11.08. Found: C,
59.02; H, 8.45; N, 11.16.
5-(2-Amino-4-methoxypyrimidin-5-yl)-1-(2-deoxy-3,5-di-O-triiso-
propylsilyl-â-D-ribofuranosyl)imidazole-4-carbonitrile (12). In the
similar manner as described for 10, a mixture of 5 (3.40 g, 5.25 mmol)
and 8 (4.35 g, 10.5 mmol) was treated with dba3Pd2‚CHCl3 (540 mg,
0.52 mmol) to give 12 (2.75 g, 81% as a white foam): FAB-LRMS
1-(2-Deoxy-â-D-ribofuranosyl)imidazo[5′,4′:4,5]pyrido[2,3-d]-
pyrimidine-4,7(5H,6H)-dione (2). To a solution of 15 (115 mg, 0.18
mmol) in THF (15 mL) was added TBAF (1 M, 0.45 mL, 0.45 mmol)
at 0 °C, and the reaction mixture was stirred at room temperature for
1.5 h. After addition of AcOH (26 µL, 0.45 mmol), the solvent was
removed in vacuo, and the residue was partitioned between H2O and
CHCl3. The separated aqueous layer was washed with AcOEt and
concentrated in vacuo. The residue was coevaporated with EtOH, and
a mixture of EtOH-H2O (4:1, 5 mL) was added to the resulting solid.
The suspension was heated in a water bath and then cooled to room
temperature. The precipitate was collected and washed with EtOH to
give 2 (51 mg, 89% as a white powder): FAB-LRMS m/z 320 (MH+);
UV λmax (H2O) 338 nm (ꢀ ) 9700), 243 nm (ꢀ ) 35 000); λmax (0.5
N HCl) 330 nm (ꢀ ) 10 000), 241 nm (ꢀ ) 29 200); λmax (0.5 N
NaOH) 339 nm (ꢀ ) 14 000), 254 nm (ꢀ ) 26 900); 1H NMR (DMSO-
d6) δ: 12.09 and 11.61 (each br s, each 1 H), 8.45 (s, 1 H), 8.29 (s, 1
H), 6.35 (dd, 1 H, J ) 6.0, 5.7 Hz), 5.36 (d, 1 H, J ) 4.7 Hz), 4.78 (t,
1
m/z 645 (MH+); H NMR (CDCl3) δ: 8.12 and 7.99 (each s, each 1
H), 5.69 (dd, 1 H, J ) 7.7, 5.9 Hz), 5.21 (br s, 2 H), 4.68 (ddd, 1 H,
J ) 5.2, 2.7, 1.6 Hz), 3.99 (m, 1 H), 3.91 (s, 3 H), 3.87 (m, 2 H), 2.36
(ddd, 1 H, J ) 7.7, 13.1, 5.2 Hz), 2.31 (ddd, 1 H, J ) 5.9, 13.1, 2.7
Hz), 1.08-1.04 (m, 42 H); 13C NMR (CDCl3) δ: 167.43, 163.92,
160.00, 137.00, 134.16, 114.95, 113.98, 97.62, 88.74, 85.76, 72.61,
63.50, 53.80, 43.55, 18.00, 17.95, 12.06, 11.92. Anal. Calcd. for
C32H56N6O4Si2: C, 59.59; H, 8.75; N, 13.03. Found: C, 59.73; H, 8.77;
N, 13.11.
5-(2-Amino-4-methoxypyrimidin-5-yl)-1-(2-deoxy-3,5-di-O-triiso-
propylsilyl-â-D-ribofuranosyl)imidazole-4-carboxamide (13). To a
solution of 12 (1.99 g, 3.55 mmol) in MeOH (80 mL) was added H2O2
(35%, 4 mL, 35.5 mmol) and NH4OH (28%, 40 mL), and the whole
mixture was stirred at room temperature for 13 h. After addition of
saturated aqueous Na2S2O3 (5 mL), the solvent was removed in vacuo,
and the residue was partitioned between CHCl3 and H2O. The separated
9
9980 J. AM. CHEM. SOC. VOL. 125, NO. 33, 2003

Synthesis and Thermal Stability of Oligodeoxynucleotides
A R T I C L E S
1 H, J ) 5.0 Hz), 4.38 (dddd, 1 H, J ) 4.7, 6.3, 5.3, 2.8 Hz), 3.92
(ddd, 1 H, J ) 2.8, 4.7, 4.8 Hz), 3.46 (ddd, 1 H, J ) 5.0, 4.7, 11.8
Hz), 3.39 (ddd, 1 H, J ) 5.0, 4.8, 11.8 Hz), 2.80 (ddd, 1 H, J ) 6.0,
13.2, 6.3 Hz), 2.40 (ddd, 1 H, J ) 5.7, 13.2, 5.3 Hz); 13C NMR (DMSO-
d6) δ: 158.4, 139.0, 132.0, 129.3, 88.2, 84.9, 69.9, 61.0, 38.4. Anal.
Calcd. for C13H13N5O5: C, 48.91; H, 4.10; N, 21.93. Found: C, 48.86;
H, 4.15; N, 21.88.
(ddd, 1 H, J ) 8.1, 12.8, 5.2 Hz), 1.10-1.05 (m, 42 H); 13C NMR
(CDCl3) δ: 163.05, 149.79, 137.79, 88.81, 88.68, 83.29, 81.67, 76.86,
72.36, 63.14, 43.38, 27.84. 27.50, 17.79, 11.88, 11.66.
Methyl 1-(2-Deoxy-3,5-di-O-triisopropylsilyl-â-D-ribofuranosyl)-
5-iodoimidazole-4-carboxylate (19). To a solution of 18 (4.2 g, 4.86
mmol) in MeOH (70 mL) was added 28% NaOMe (2 mL, 9.8 mmol),
and the whole mixture was stirred at room temperature for 3 h. The
reaction was quenched by addition of saturated aqueous NH4Cl, and
the solvent was removed in vacuo. The residue was partitioned between
AcOEt and H2O. The separated organic layer was washed with H2O,
followed by brine. The organic layer was dried (Na2SO4), and
concentrated in vacuo. The residue was purified by a silica gel column
and eluted with hexane/AcOEt (9:1-2:1), to give 19 (3.24 g, 98%,
crystallized from hexane): mp 90-91 °C; FAB-LRMS m/z 681 (MH+);
1H NMR (CDCl3) δ: 8.10 (s, 1 H), 6.14 (dd, 1 H, J ) 5.5, 7.9 Hz),
4.69 (ddd, 1 H, J ) 2.4, 5.2, 1.6 Hz), 4.11 (m, 1 H), 3.92 (m, 2 H),
2.50 (ddd, 1 H, J ) 5.5, 12.9, 2.4 Hz), 2.22 (ddd, 1 H, J ) 7.9, 12.9,
5.2 Hz), 1.10-1.05 (m, 42 H); 13C NMR (CDCl3) δ: 162.75, 138.58,
135.65, 89.11, 88.93, 76.43, 72.72, 63.38, 51.77, 43.54, 17.95, 12.06,
11.84. Anal. Calcd. for C28H53IN2O5Si2: C, 49.40; H, 7.85; N, 4.11.
Found: C, 49.19; H, 7.82; N, 4.04.
5-[4-Methoxypyrimidin-2(1H)-on-5-yl]-1-(2-deoxy-3,5-di-O-triiso-
propylsilyl-â-D-ribofuranosyl)imidazole-4-carbonitrile (16). In the
similar manner as described for 14, 12 (1.68 g, 2.61 mmol) was treated
with isoamyl nitrite (3.5 mL, 26.1 mmol) to give 16 (967 mg, 57% as
a white foam): FAB-LRMS m/z 646 (MH+); 1H NMR (CDCl3) δ:
12.31 (br s, 1 H), 8.01 and 7.79 (each s, each 1 H), 5.70 (dd, 1 H, J )
7.8, 5.7 Hz), 4.72 (dt, 1 H, J ) 5.4, 2.3 Hz), 4.03 (m, 4 H,), 3.87 (m,
2 H), 2.44 (ddd, 1 H, J ) 7.8, 12.8, 5.4 Hz), 2.36 (ddd, 1 H, J ) 5.7,
12.8, 2.3 Hz), 1.08-1.04 (m, 42 H); 13C NMR (CDCl3) δ: 169.85,
158.92, 146.90, 137.74, 131.84, 115.11, 114.52, 95.77, 89.18, 85.99,
72.79, 63.64, 55.38, 43.47, 18.19, 18.15, 12.26, 12.10. Anal. Calcd.
for C32H55N5O5Si2: C, 59.50; H, 8.58; N, 10.84. Found: C, 59.72; H,
8.66; N, 10.68.
4-Amino-1-(2-deoxy-3,5-di-O-triisopropylsilyl-â-D-ribofuranosyl)-
imidazo[5′,4′:4,5]pyrido[2,3-d]pyrimidin-7(6H)-one (17). In the
similar manner as described for 15, 16 (960 mg, 1.49 mmol) was treated
with methanolic ammonia (saturated at 0 °C, 40 mL) to give 17 (882
mg, 94%, crystallized from acetone-CHCl3): mp 287-289 °C; FAB-
LRMS m/z 631 (MH+); 1H NMR (CDCl3) δ: 13.77 (br s, 1 H), 10.30
and 6.44 (each br s, each 1 H), 9.33 (s, 1 H), 8.02 (s, 1 H), 6.42 (dd,
1 H, J ) 7.9, 5.2 Hz), 4.82 (ddd, 1 H, J ) 5.0, 2.8, 1.6 Hz), 4.22 (ddd,
1 H, J ) 1.6, 2.6, 3.7 Hz), 3.88 (dd, 1 H, J ) 2.6, 11.3 Hz), 3.83 (dd,
1 H, J ) 3.7, 11.3 Hz), 2.70 (ddd, 1 H, J ) 7.9, 12.7, 5.0 Hz), 2.62
(ddd, 1 H, J ) 5.2, 12.7, 2.8 Hz), 1.12-0.99 (m, 42 H); 13C NMR
(CDCl3) δ: 160.38, 159.06, 157.45, 153.53, 137.92, 133.40, 125.15,
96.31, 89.07, 85.95, 72.35, 63.15, 40.99, 18.00, 17.86, 12.10, 11.81.
Anal. Calcd. for C31H54N6O4Si2: C, 59.01; H, 8.63; N, 13.32. Found:
C, 58.92; H, 8.56; N, 13.30.
7-Amino-1-(2-deoxy-3,5-di-O-triisopropylsilyl-â-D-ribofuranosyl)-
imidazo[5′,4′:4,5]pyrido[2,3-d]pyrimidin-4(5H)-one (22). In the simi-
lar manner as described for 11, a mixture of 19 (3.38 g, 4.97 mmol)
and 7 (3.12 g, 7.46 mmol) was treated with dba3Pd2‚CHCl3 (414 mg,
0.40 mmol) to give a brown syrup. The resulting crude mixture was
then dissolved in 1,4-dioxane (30 mL), and NH4OH (28%, 30 mL)
was added to the mixture. The whole was heated at 100 °C for 60 h in
a steel container. The solvent was removed in vacuo, and the residue
was partitioned between CHCl3 and H2O. The separated organic layer
was washed with H2O, followed by brine. The organic layer was dried
(Na2SO4) and concentrated in vacuo. The residue was purified by a
silica gel column and eluted with 0-5% EtOH in CHCl3, to give 22
(1.78 g, 57%, crystallized from acetone-CHCl3): mp 187-188 °C; FAB-
1
LRMS m/z 631 (MH+); H NMR (CDCl3) δ: 12.90 (br s, 1 H), 9.05
(s, 1 H), 7.96 (s, 1 H), 6.41 (dd, 1 H, J ) 7.7, 5.2 Hz), 4.82 (ddd, 1 H,
J ) 5.1, 2.9, 1.0 Hz), 4.20 (ddd, 1 H, J ) 1.0, 2.9, 4.0 Hz), 3.86 (dd,
1 H, J ) 2.9, 11.2 Hz), 3.80 (dd, 1 H, J ) 4.0, 11.2 Hz), 2.73 (ddd, 1
H, J ) 7.7, 12.4, 5.1 Hz), 2.63 (ddd, 1 H, J ) 5.2, 12.4, 2.9 Hz),
1.12-0.98 (m, 42 H); 13C NMR (CDCl3) δ: 162.18, 160.13, 154.56,
154.44, 137.93, 133.88, 129.81, 98.03, 89.09, 85.79, 72.48, 63.29, 40.76,
18.05, 17.90, 12.15, 11.85. Anal. Calcd. for C31H54N6O4Si2: C, 59.01;
H, 8.63; N, 13.32. Found: C, 59.06; H, 8.59; N, 13.24.
4-Amino-1-(2-deoxy-â-D-ribofuranosyl)imidazo[5′,4′:4,5]pyrido-
[2,3-d]pyrimidin-7(6H)-one (3). In the similar manner as described
for 2, 17 (220 mg, 0.35 mmol) was treated with TBAF (1 M, 0.88 mL,
0.88 mmol) to give 3 (83 mg, 74% as a white powder): FAB-LRMS
m/z 319 (MH+); UV λmax (H2O) 348 nm (ꢀ ) 17 900), 252 nm (ꢀ )
24 000); λmax (0.5 N HCl) 356 nm (ꢀ ) 17 400), 304 nm (ꢀ ) 8300),
244 nm (ꢀ ) 18 700); λmax (0.5 N NaOH) 344 nm (ꢀ ) 12 600), 252
1
nm (ꢀ ) 29 700); H NMR (DMSO-d6) δ: 11.74 (br s, 1 H), 9.09 (s,
7-Amino-1-(2-deoxy-â-D-ribofuranosyl)imidazo[5′,4′:4,5]pyrido-
[2,3-d]pyrimidin-4(5H)-one (4). In the similar manner as described
for 2, 22 (255 mg, 0.40 mmol) was treated with TBAF (1 M, 1.0 mL,
1.0 mmol) to give 4 (103 mg, 81% as a white powder): FAB-LRMS
m/z 319 (MH+); UV λmax (H2O) 333 nm (ꢀ ) 14 400), 281 nm (ꢀ )
10 900), 240 nm (ꢀ ) 28 700); λmax (0.5 N HCl) 330 nm (ꢀ ) 8500),
244 nm (ꢀ ) 33 300); λmax (0.5 N NaOH) 337 nm (ꢀ ) 13 900), 257
1 H), 8.39 (s, 1 H), 7.78 (br s, 2 H), 6.52 (dd, 1 H, J ) 6.1, 6.0 Hz),
5.37 (d, 1 H, J ) 4.5 Hz), 4.85 (t, 1 H, J ) 4.9 Hz), 4.39 (dddd, 1 H,
J ) 4.5, 6.6, 5.0, 3.4 Hz), 3.95 (ddd, 1 H, J ) 3.4, 4.0, 4.8 Hz), 3.51
(ddd, 1 H, J ) 4.9, 4.0, 11.8 Hz), 3.46 (ddd, 1 H, J ) 4.9, 4.8, 11.8
Hz), 2.75 (ddd, 1 H, J ) 6.1, 13.2, 6.6 Hz), 2.48 (ddd, 1 H, J ) 6.0,
13.2, 5.0 Hz); 13C NMR (DMSO-d6) δ: 159.5, 156.8, 156.3, 152.8,
138.8, 133.0, 123.9, 94.7, 87.9, 85.4, 69.9, 61.0. Anal. Calcd. for
C13H14N6O4‚1/3H2O: C, 48.15; H, 4.56; N, 25.91. Found: C, 48.26;
H, 4.49; N, 25.67.
1
nm (ꢀ ) 26 200); H NMR (DMSO-d6) δ: 11.52 (br s, 1 H), 8.85 (s,
1 H), 8.30 (s, 1 H), 6.89 (br s, 2 H), 6.47 (t, 1 H, J ) 6.0 Hz), 5.37 (d,
1 H, J ) 5.0 Hz), 4.86 (t, 1 H, J ) 5.0 Hz), 4.38 (dddd, 1 H, J ) 5.0,
6.3, 4.7, 3.4 Hz), 3.93 (ddd, 1 H, J ) 3.4, 4.3, 4.4 Hz), 3.51 (ddd, 1 H,
J ) 5.0, 4.3, 11.8 Hz), 3.45 (ddd, 1 H, J ) 5.0, 4.4, 11.8 Hz), 2.72
(ddd, 1 H, J ) 6.0, 13.2, 6.3 Hz), 2.45 (ddd, 1 H, J ) 6.0, 13.2, 4.7
Hz); 13C NMR (DMSO-d6) δ: 161.7, 158.3, 154.3, 153.3, 138.6, 132.7,
129.3, 96.9, 87.9, 85.3, 69.8, 61.0, 39.0. Anal. Calcd. for C13H14N6O4‚
3/10H2O‚1/5EtOH: C, 48.35; H, 4.78; N, 25.24. Found: C, 48.42; H,
4.73; N, 25.28.
1-(2-Deoxy-3,5-di-O-triisopropylsilyl-â-D-ribofuranosyl)-5-iodo-
imidazole-4-[N,N-di-(tert-butoxycarbonyl)]carboxamide (18). To a
solution of 6 (3.5 g, 5.26 mmol) in CH2Cl2 (60 mL) containing Et3N
(2.2 mL, 15.8 mmol) and DMAP (640 mg, 5.26 mmol) was added
di-tert-butyl dicarbonate (3.6 mL, 15.8 mmol), and the whole mixture
was stirred at room temperature. After 10 h, an additional di-tert-butyl
dicarbonate (1.8 mL, 7.9 mmol) was added to the reaction mixture.
After an additional 10 h, the solvent was removed in vacuo, and the
residue was purified by a silica gel column and eluted with hexane/
AcOEt (20:1-6:1), to give 18 (4.32 g, 95% as a colorless oil): FAB-
LRMS m/z 866 (MH+); FAB-HRMS calcd. for C37H69IN3O8Si2
Synthesis of ODNs. ODNs were synthesized on a DNA/RNA
synthesizer (Applied Biosystem Model 392) by the phosphoramidite
method.34 For the incorporation of the tricyclic nucleosides into the
ODNs, a 0.12 M solution of each tricyclic nucleoside phosphoramidite
1
866.3669, found: 866.3685. H NMR (CDCl3) δ: 8.04 (s, 1 H), 6.13
(dd, 1 H, J ) 5.4, 8.1 Hz), 4.68 (ddd, 1 H, J ) 2.3, 5.2, 2.6 Hz), 4.10
(m, 1 H), 3.91 (m, 2 H), 2.50 (ddd, 1 H, J ) 5.4, 12.8, 2.3 Hz), 2.22
(34) Beaucage, S. L.; Caruthers, M. H. Tetrahedron Lett. 1981, 22, 1859-1862.
9
J. AM. CHEM. SOC. VOL. 125, NO. 33, 2003 9981

A R T I C L E S
Minakawa et al.
in CH3CN and a coupling time of 120 s was used. The fully protected
ODNs were then deblocked and purified by the same procedure as for
the purification of normal ODNs.35 Thus, each ODN linked to the resin
(1 µmol) was treated with concentrated NH4OH at 55 °C for 16 h, and
the released ODN protected by a DMTr group at the 5′-end was
chromatographed on a C-18 silica gel column (1 × 12 cm, Waters)
with a linear gradient of CH3CN from 0 to 40% in 0.1 M TEAA buffer
(pH 7.0). The fractions were concentrated, the residue was treated with
aqueous 80% AcOH at room temperature for 20 min, then the solution
was concentrated, and the residue was coevaporated with H2O. The
residue was dissolved in H2O, the solution was washed with Et2O, and
then the H2O layer was concentrated to give the deprotected ODN.
The ODN was further purified by reverse-phase HPLC, using a J’sphere
ODN M80 column (4.6 × 150 mm, YMC) with a linear gradient of
CH3CN (from 10 to 25% over 30 min) in 0.01 M TEAA buffer (pH
7.0) to give highly purified ODNs.
dA, 15 400; dC, 7300; dG, 11 700; T, 8800. The extinction coefficients
for the tricyclic nucleosides at 260 nm were determined to be the
following: NN, 26 200; OO, 11 900; NO, 18 100; ON, 9400.
Thermal Denaturation. Each sample that contained appropriate
ODNs in a buffer of either 0.01 M sodium cacodylate (pH 7.0)
containing 0.1 M NaCl or 0.01 M sodium phosphate (pH 7.0) containing
1.0 M NaCl was heated at 95 °C for 5 min, cooled gradually to an
appropriate temperature, and used for the thermal denaturation study.
Thermal-induced transitions of each mixture of ODNs were monitored
at 260 nm on a Perkin-Elmer Lambda 2S. The sample temperature
was increased 0.5 °C/min. Thermodynamic parameters were derived
by the reported method.36
Acknowledgment. This work was supported in part by Grant-
in-Aid for Scientific Research on Priority Areas and Encourage-
ment of Young Scientists from the Ministry of Education,
Science, Sports, and Culture of Japan. We would like to thank
Ms. H. Matsumoto and Ms. A. Maeda (Center for Instrumental
Analysis, Hokkaido University) for elemental analysis. We also
would like to thank Ms. S. Oka (Center for Instrumental
Analysis, Hokkaido University) for measurement of mass
spectra. This paper constitutes Part 222 of Nucleosides and
Nucleotides. Part 221 is ref 37.
Complete Hydrolysis of the ODNs. Each ODN (0.2 OD units at
260 nm) was incubated with snake venom phosphodiesterase (10 µg),
nuclease P1 (10 µg), and alkaline phosphatase (0.4 units) in a buffer
containing 100 mM Tris-HCl (pH 7.7) and 2 mM MgCl2 (total 125
µL) at 37 °C for 12 h. After this was heated in boiling water for 5 min,
cold EtOH (320 µL) was added to the reaction mixture, and the mixture
was kept at -30 °C for 1 h. The cold solution was centrifuged for 20
min (12 000 rpm), and then the supernatant was separated and
concentrated. The residue was analyzed by reverse-phase HPLC, using
a J’sphere ODN M80 column (4.6 × 150 mm, YMC) with a linear
gradient of CH3CN (from 2 to 12% over 40 min) in 0.01 M TEAA
buffer (pH 7.0).
Supporting Information Available: Experimental details of
synthesis of 5 and 6 (SI-Scheme 1), 7 and 8 (SI-Scheme 2),
phosporamidites (SI-Scheme 3), 23 (SI-Scheme 4), SI-Figure
1 (1H NMR spectra of homodimer 17‚17), SI-Figure 2 (1H
NMR spectra of a mixture of 17 and 22), SI-Figure 3 (1H NMR
spectra of a mixture of 11 and 15), enzymatic digestion of ODNs
(SI-Figure 4 and SI-Table 1), thermodynamic parameters of
the duplexes containing the tricyclic nucleosides (SI-Tables 2
and 3). This material is available free of charge via the Internet
at http://pubs.acs.org.
Hyperchromicities and Extinction Coefficients of the ODNs. Each
ODN (0.25 OD units at 260 nm) was enzymatically hydrolyzed under
the conditions described above. Hyperchromicity of each ODN was
determined by comparing UV absorbances at 260 nm of the solutions
before and after hydrolyses. The extinction coefficient (at 260 nm) of
each ODN was determined using the following equation: ꢀODN ) the
sum of ꢀnucleoside/hyperchromicity. The extinction coefficients (at 260
nm) of the natural nucleosides used for calculations were as follows:
JA0347686
(36) Marky, L. A.; Breslaure, K. L. Biopolymer 1987, 26, 1601-1620.
(37) Itoh, T.; Ueno, Y.; Komatsu, Y.; Matsuda, A. Nucleic Acids Res. 2003,
31, 2514-2523.
(35) Gait, M. L., Ed. Oligonucleotides synthesis: a practical approach; IRL
Press: Oxford, UK, 1984.
9
9982 J. AM. CHEM. SOC. VOL. 125, NO. 33, 2003