Triazole-linked dumbbell oligodeoxynucleotides with NF-κB binding ability as potential decoy molecules

Article abstract of DOI:10.1021/jo702459b

(Chemical Equation Presented) Triazole-cross-linked oligodeoxynucleotides were synthesized with use of the Cu(I) catalyzed alkyneazide cycloaddition (CuAAC) with oligodeoxynucleotides possessing N-3-(azidoethyl)thymidine and N-3-(propargyl)thymidine at the 3′- and 5′-termini. The newly synthesized oligodeoxynucleotides were thermally stable and their global structures retained those of non-cross-linked oligodeoxynucleotides. The newly synthesized dumbbell oligodeoxynucleotides showed excellent stability against snake venom phosphodiesterase (3′-exonuclease) and high thermal stability, which are necessary for decoy molecules to achieve biological responses leading to alteration of gene expression. Moreover, dumbbell oligodeoxynucleotides have the ability to bind to NF-κB p50 homodimer within a similar range to that of a control double-stranded decoy oligodeoxynucleotide. This strategy allows us to prepare triazole-linked dumbbell oligodeoxynucleotides with a range of loop lengths, and we found that the greater the number of the thymidine residues constituting the loop region, the higher the binding affinity of the dumbbell oligodeoxynucleotides to the nuclear factor κB. This means that a protein binding ability of the dumbbell oligodeoxynucleotides could be modulated by altering the loop size. This study clearly shows that cross-linking by the triazole structure does not prevent the dumbbell oligodeoxynucleotides from binding to the nuclear factor κB transcription factor. Therefore, the results obtained conclusively demonstrate that the triazole cross-linked dumbbell oligodeoxynucleotides could be proposed as powerful decoy molecules.

Full text of DOI:10.1021/jo702459b

Triazole-Linked Dumbbell Oligodeoxynucleotides with NF-KB
Binding Ability as Potential Decoy Molecules
Masanori Nakane, Satoshi Ichikawa,* and Akira Matsuda*
Faculty of Pharmaceutical Sciences, Hokkaido UniVersity, Sapporo 060-0812, Japan
matuda@pharm.hokudai.ac.jp; ichikawa@pharm.hokudai.ac.jp
ReceiVed NoVember 16, 2007
Triazole-cross-linked oligodeoxynucleotides were synthesized with use of the Cu(I) catalyzed alkyne-
azide cycloaddition (CuAAC) with oligodeoxynucleotides possessing N-3-(azidoethyl)thymidine and N-3-
(propargyl)thymidine at the 3′- and 5′-termini. The newly synthesized oligodeoxynucleotides were thermally
stable and their global structures retained those of non-cross-linked oligodeoxynucleotides. The newly
synthesized dumbbell oligodeoxynucleotides showed excellent stability against snake venom phosphodi-
esterase (3′-exonuclease) and high thermal stability, which are necessary for decoy molecules to achieve
biological responses leading to alteration of gene expression. Moreover, dumbbell oligodeoxynucleotides
have the ability to bind to NF-κB p50 homodimer within a similar range to that of a control double-
stranded decoy oligodeoxynucleotide. This strategy allows us to prepare triazole-linked dumbbell
oligodeoxynucleotides with a range of loop lengths, and we found that the greater the number of the
thymidine residues constituting the loop region, the higher the binding affinity of the dumbbell
oligodeoxynucleotides to the nuclear factor κB. This means that a protein binding ability of the dumbbell
oligodeoxynucleotides could be modulated by altering the loop size. This study clearly shows that cross-
linking by the triazole structure does not prevent the dumbbell oligodeoxynucleotides from binding to
the nuclear factor κB transcription factor. Therefore, the results obtained conclusively demonstrate that
the triazole cross-linked dumbbell oligodeoxynucleotides could be proposed as powerful decoy molecules.
Introduction
oping therapeutic agents as cis-elements, referred to as decoy
ODNs, to trap trans-activators, transcription factors.2-4 Thus
once delivered to cells, the decoy ODNs bearing a consensus
binding sequence of a specific transcription factor are specifi-
cally recognized by the target protein. Interaction with the decoy
ODNs prevents the target protein from binding to its cis-element.
As a result, activation and expression of the corresponding cis-
element are silenced.5,6 The transcriptional decoy strategy against
The accumulation of gene products caused by the aberrant
activation and expression of the corresponding cis-element is
involved in the initiation and progression of pathogenesis. The
altered activation of transcription factors is now known to be a
component of many pathways of disease pathogenesis, and the
development of strategies regulating transcription factors has
emerged as an attractive field of investigation.1 Recently much
effort has been dedicated to the application of the relatively
short double-stranded oligodeoxynucleotides (ODNs) in devel-
(2) Morishita, R.; Higaki, J.; Tomita, N.; Ogihara, T. Circ. Res. 1998,
82, 1023-1028.
(3) Cho-Chung, Y. S.; Park, Y. G.; Lee, Y. N. Curr. Opin. Mol. Ther.
1999, 1, 386-392.
(1) Papavassiliou, A. G. Mol. Med. Today 1998, 4, 358-366.
(4) Mann, M. J.; Dzau, T. J. J. Clin. InVest. 2000, 106, 1071-1075.
10.1021/jo702459b CCC: $40.75 © 2008 American Chemical Society
Published on Web 02/01/2008
1842
J. Org. Chem. 2008, 73, 1842-1851

Triazole-Linked Dumbbell Oligodeoxynucleotides
nuclear factor κB (NF-κB) is one of the most effective
approaches.7 Activation of NF-κB upregulates genes related to
inflammatory and immunological responses, such as TNF-R,
adhesion molecules, macrophage colony-stimulating factor (M-
CSF), granulocyte/macrophage colony-stimulating factor (GM-
CSF), and monocyte chemoattractant protein (MCP). Therefore,
transfection of NF-κB decoy ODNs leads to anti-inflammatory
and immuno-suppressive activities.8-15 However, several short-
comings of the decoy strategy modulating transcription factors
make its practical application to biological systems unfeasible
as yet. Generally, relatively short double-stranded ODNs are
used in the decoy strategy and possess less thermal stability
under physiological conditions. Another drawback is the pres-
ence of extra- and intracellular nucleases, which cleave ODNs.16,17
To achieve thermal stability and resistance to nucleases,
structural alteration or chemical modification of decoy ODNs
is necessary.18-26 Dumbbell ODNs, which are circular ODNs
consisting of double-stranded stem region and nucleotide loops
at both of their termini, possess increased exonuclease resistance
because they have no terminal nucleotide residues.27-29 Dumb-
bell decoy ODNs against NF-κB were shown to inhibit in vitro
and ex vivo transcription. These ODNs were prepared by
enzymatically ligating two identical molecules.27,30 However,
in some cases, it is difficult to provide large quantities of
modified ODNs by enzymatic preparation.31,32 An alternative
approach for cross-linking nucleic acids might be chemical
modification, where a chemically reactive functional group is
placed site-specifically in proximity to the opposing strands of
a duplex.33 Additionally, chemical modification at both terminal
positions of the double-stranded ODNs with covalent cross-
linking providing dumbbell-like ODNs is expected to resist
degradation catalyzed by enzymes such as exonucleases.34 One
of the strategies to cross-link double-stranded ODNs at the
termini is the oxidation of thiols to form a disulfide, as reported
by Glick et al.35,36 Terminal cross-linking of double-stranded
ODNs increases the thermal stability of the double-stranded
ODNs and does not cause significant structural changes to the
internal helix. Therefore this strategy has been used to inves-
tigate the structure of double-stranded ODNs. Though this
oxidation proceeds quickly and is effective, it is reversible and
the resulting disulfide bond might be easily cleaved by nucleo-
philes existing in the biological system. Other chemical
modifications of ODNs without cross-links have also been
studied including R-ODN/â-ODN duplexes,37 phosphorothioate-
modified ODNs,16,38 PNA-DNA chimera,39,40 and LNA-DNA
chimera,41 respectively, although they all have drawbacks.
(5) Bielinska, A.; Shivdasani, R. A.; Zhang, L.; Nabel, G. J. Science
1990, 250, 997-1000.
(6) Morishita, R.; Gibbons, G. H.; Horiuchi, M.; Ellison, K. E.; Nakajima,
M.; Zhang, L.; Kaneda, Y.; Ogihara, T.; Dzau, V. J. Proc. Natl. Sci. U.S.A.
1995, 92, 5855-5859.
The Hu¨isgen reaction is a 1,3-dipolar cycloaddition between
azide and alkyne groups to afford a chemically stable 1,2,3-
triazole. The usual thermal cycloaddition by heating gives both
the 1,4- and 1,5-disubstituted triazoles with diminished regi-
oselectivity; however, a long reaction time is necessary. On the
other hand, the Cu(I)-catalyzed alkyne-azide cycloaddition
(CuAAC),42 which is known as “Click Chemistry”,43 proceeds
quite rapidly in a variety of solvents including aqueous media
(7) Morishita, R.; Sugimoto, T.; Aoki, M.; Kida, I.; Tomita, N.;
Moriguchi, A.; Maeda, K.; Sawa, Y.; Kaneda, Y.; Higaki, J.; Ogihara, T.
Nature Med. 1997, 3, 894-899.
(8) Khaled, A. R.; Butfiloski, E. J.; Sobel, E. S.; Schiffenbauer, J. Cell.
Immunol. 1998, 185, 49-58.
(9) Ono, S.; Date, I.; Onoda, K.; Shiota, T.; Ohmoto, T.; Ninomiya, Y.;
Asari, S.; Morishita, R. Hum. Gene Ther. 1998, 9, 1003-1011.
(10) Sharma, H. W.; Perez, J. R.; Higgins-Sochaski, K.; Hsiao, R.;
Narayanan, R. Anticancer Res. 1996, 16, 61-69.
(11) Tomita, N.; Morishita, R.; Tomita, S.; Yamamoto, K.; Aoki, M.;
Matsushita, H.; Hayashi, S.; Higaki, J.; Ogihara, T. J. Hypertens. 1998,
16, 993-1000.
(12) Tomita, N.; Morishita, R.; Lan, H. Y.; Yamamoto, K.; Hashizume,
M.; Notake, M.; Toyosawa, K.; Fujitani, B.; Mu, W.; Nikolic-Paterson, D.
J.; Atkins, R. C.; Kaneda, Y.; Higaki, J.; Ogihara, T. J. Am. Soc. Nephrol.
2000, 11, 1244-1252.
(27) Hosoya, T.; Takeuchi, H.; Kanesaki, Y.; Yamakawa, H.; Miyano-
Kurosaki, N.; Takai, K.; Yamamoto, N.; Takaku, H. FEBS Lett. 1999, 461,
136-140.
(28) Erie, D.; Sinha, N.; Olson, W.; Jones, R.; Breslauer, K. Biochemistry
1987, 26, 7150-7159.
(13) Suzuki, J.; Morishita, R.; Amano, J.; Kaneda, Y.; Isobe, M. Gene
Ther. 2000, 7, 1847-1852.
(29) Chu, B. C. F.; Orgel, L. E. Nucleic Acids Res. 1992, 20, 5857-
5858.
(14) Tomita, N.; Morishita, R.; Tomita, S.; Gibbons, G. H.; Zhang, L.;
Horiuchi, M.; Kaneda, Y.; Higaki, J.; Ogihara, T.; Dzau, V. J. Gene Ther.
2000, 7, 1326-1332.
(30) Lee, I. K.; Ahn, J. D.; Kim, H. S.; Park, J. Y.; Lee, K. U. Curr.
Drug Targ. 2003, 4, 619-623.
(31) Erie, D. A.; Jones, R. A.; Olson, W. K.; Sinha, N. K.; Breslauer,
K. J. Biochemistry 1989, 28, 268 -273.
(15) Vos, I. H.; Govers, R.; Grone, H. J.; Kleij, L.; Schurink, M.; De
Weger, R. A.; Goldschmeding, R.; Rabelink, T. J. FASEB J. 2000, 14, 815-
822.
(32) Ashley, G. W.; Kushlan, D. M. Biochemistry 1991, 30, 2927 -2933.
(33) For examples, see: (a) Cowart, M.; Benkovic, S. J. Biochemistry
1991, 30, 788-796. (b) Milton, J.; Connolly, B. A.; Nikiforov, T. T.;
Cosstick, R. J. Chem. Soc., Chem. Commun. 1993, 779-780. (c) Ferentz,
A. E.; Keating, T. A.; Verdine, G. L. J. Am. Chem. Soc. 1993, 115, 9006-
9014 and references cited therein. (d) Erlanson, D. A.; Chen, L.; Verdine,
G. L. J. Am. Chem. Soc. 1993, 115, 12583-12584. (e) Kool, E. T. Annu.
ReV. Biophys. Biomol. Struct. 1996, 25, 1-28. (f) Kool, E. T. Acc. Chem.
Res. 1998, 31, 502-510.
(16) Uhlmann, E.; Peyman, A. Chem. ReV. 1990, 90, 543-583.
(17) Chu, B. C. F.; Orgel, L. E. Nucleic Acids Res. 1992, 20, 5857-
5858.
(18) Egholm, M.; Buchard, O.; Christensen, 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.
(19) Demidov, V. V.; Potaman, V. N.; Frank-Kamenetskii, M. D.;
Egholm, M.; Buchard, O.; So¨nnichsen, S. H.; Nielsen, P. E. Biochem.
Pharmacol. 1994, 48, 1310-1313.
(34) Murakami, A.; Yamamoto, Y.; Namba, M.; Iwase, R.; Yamaoka,
T. Bioorg. Chem. 2001, 29, 223-233.
(20) Romanelli, A.; Pedone, C.; Saviano, M.; Bianch, N.; Borgatti, M.;
Mischiati, C.; Gambari, R. Eur. J. Biochem. 2001, 268, 6066-6075.
(21) Petersen, M.; Nielsen, C. B.; Nielsen, K. E.; Jensen, G. A.;
Bondensgaard, K.; Singh, S. K.; Rajwanshi, V. K.; Koskin, A. A.; Dahl,
B. M.; Wengel, J.; Jacobsen, J. P. J. Mol. Recognit. 2000, 13, 44-53.
(22) Petersen, M.; Wengel, J. Trends Biotechnol. 2003, 21, 74-81.
(23) Crinelli, R.; Bianchi, M.; Gentilini, L.; Magnani, M. Nucleic Acids
Res. 2002, 30, 2435-2443.
(35) Glick, G. D. J. Org. Chem. 1991, 56, 6746-6747.
(36) Osborne, S. E.; Vo¨lker, J.; Stevens, S. Y.; Breslauer, K. J.; Glick,
G. D. J. Am. Chem. Soc. 1996, 118, 11993-12003.
(37) Tanaka, H.; Vickart, P.; Bertrand, J.-R.; Rayner, B.; Morvan, F.;
Imbach, J.-L.; Paulin, D.; Malvy, C. Nucleic Acids Res. 1994, 22, 3069-
3074.
(38) Stein, C. A.; Subasinghe, C.; Shinozuka, K.; Cohen, J. S. Nucleic
Acids Res. 1988, 16, 3209-3221.
(24) Hikishima, S.; Minakawa, N.; Kuramoto, K.; Ogata, S.; Matsuda,
A. ChemBioChem 2006, 7, 1970-1975.
(39) Mischiati, C.; Borgatti, M.; Bianch, N.; Rutigliano, C.; Tomassetti,
M.; Feriotto, G.; Gambari, R. J. Biol. Chem. 1999, 274, 33114-33122.
(40) Romanelli, A.; Pedone, C.; Saviano, M.; Bianchi, N.; Borgatti, M.;
Mischiati, C.; Gambari, R. Eur. J. Biochem. 2001, 268, 6066-6075.
(41) Crinelli, R.; Bianchi, M.; Gentilini, L.; Magnani, M. Nucleic Acids
Res. 2002, 30, 2435-2443.
(25) Bielenska, A.; Shivedasani, R. A.; Zhang, L.; Nabel, G. J. Science
1990, 250, 997-1000.
(26) Brown, D. A.; Kang, S.-H.; Gryaznov, S. M.; DeDionisio, L.:
Heidenreich, O.; Sullivan, S.; Xu, X.; Nerenberg, M. I. J. Biol. Chem. 1994,
269, 26801-26805.
J. Org. Chem, Vol. 73, No. 5, 2008 1843

Nakane et al.
FIGURE 1. Formation of triazole cross-linked dumbbell-like ODNs (tzODN and tz2ODN) by the cupper catalyzed alkyne-azide cycloaddition:
(a) cross-linking of hairpin ODN (hpODN) at a single site and (b) cross-linking of double-stranded ODNs (dsODN) at both termini.
at room temperature, and affords a 1,4-disubstituted triazole
selectively via a different reaction mechanism. Therefore, the
CuAAC is a powerful linking reaction, due to its high degree
of dependability, complete specificity, mild reaction conditions,
and the bio-compatibility of the reactants, and has been
increasingly used for in vivo and in vitro bioconjugation
applications.43-57 The CuAAC is also utilized in ODN biocon-
jugation and is suitable for terminal cross-linking of double-
stranded ODNs.43-54 Here we report the cross-linking of ODNs
using the CuAAC as “a stable staple” to develop a novel strategy
to synthesize decoy ODNs (Figure 1). To bridge the termini,
we have alkylated the N-3 position of 2′-deoxythymidine with
an azidoethyl and a propargyl group that permit the formation
of an intramolecular triazole bridge.
Results and Discussion
First, a single-stranded hairpin hpODN 1 (Figure 2) was
designed, where the azidoethyl and propargyl groups were
(42) (a) Tornøe, C. W.; Christensen, C.; Meldal, M. J. Org. Chem. 2002,
67, 3057-3064. (b) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless,
K. B. Angew. Chem., Int. Ed. 2002, 41, 2596-2599. (c) Kolb, H. C.;
Sharpless, K. B. Drug Disc. Today 2003, 8, 1128-1137.
(43) For review, see: (a) Kolb, H. C.; Finn, M. G.; Sharpless, K. B.
Angew. Chem., Int. Ed. 2001, 40, 2004-2021. (b) Kolb, H. C.; Sharpless,
K. B. Drug Disc. Today 2003, 8, 1128-1137.
FIGURE 2. Sequences of the ODNs used in this study. P ) N-3-
(propargyl)thymidine residue; Z ) N-3-(azidoethyl)thymidine residue;
T ) N-3-substituted thymidine residue. NF-κB recognition regions are
marked with a gray box.
(44) Seo, T. S.; Bai, X. P.; Ruparel, H.; Li, Z. M.; Turro, N. J.; Ju, J. Y.
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5488-5493.
(45) Punna, S.; Kuzelka, J.; Wang, Q.; Finn, M. G. Angew. Chem., Int.
Ed. 2005, 44, 2215-2220.
(46) Kanan, M. W.; Rozenman, M. M.; Sakurai, K.; Snyder, T. M.; Liu,
D. R. Nature 2004, 431, 545-549.
(47) Ustinov, A. V.; Korshun, V. A. Russ. Chem. Bull. 2006, 55, 1268-
1274.
(48) Kumar, R.; El-Sagheer, A.; Tumpane, J.; Lincoln, P.; Wilhelmsson,
L. M.; Brown, T. J. Am. Chem. Soc. 2007, 129, 6859-6864.
(49) Gierlich, J.; Burley, G. A.; Gramlich, P. M. E.; Hammond, D. M.;
Carell, T. Org. Lett. 2006, 8, 3639-3642.
(50) Weller, R. L.; Rajski, S. R. Org. Lett. 2005, 7, 2141-2144.
(51) Seo, T. S.; Li, Z.; Ruparel, H.; Ju, J. J. Org. Chem. 2003, 68, 609-
612.
(52) Byun, Y.; Al-Madhoun, A. S.; Johnsamuel, J.; Yang, W.; Barth, R.
F.; Eriksson, S.; Tjarks, W. J. Med. Chem. 2005, 48, 1188-1198.
(53) Chan, T. R.; Hilgraf, R.; Sharpless, K. B.; Fokin, V. V. Org. Lett.
2004, 6, 2853-2855.
(54) Bouillon, C.; Meyer, A.; Vidal, S.; Jochum, A.; Chevolot, Y.;
Cloarec, J.-P.; Praly, J.-P.; Vasseur, J.-J.; Morvan, F. J. Org. Chem. 2006,
71, 4700-4702.
attached at the N-3 positions of a thymine base at the 5′- and
3′-ends, to investigate the CuAAC at the single terminus as a
model study (Figure 1a). Its synthesis is described in Scheme
1. The N-3-(iodoethyl)thymidine 3′-phosphoramidite 6 was
devised as a precursor of the N-3-(azidoethyl)thymidine residue
because azido groups are known to be readily reduced by a
trivalent phosphorus atom, and thus the corresponding 3′-
phosphoramidite block is impossible to prepare.58 Once intro-
duced in solid-phase DNA synthesis via the general phosphor-
amiditemethod,theiodogroupwasconvertedtothecorresponding
azide group via the on-column application of NaN3. Treatment
of 5′-O-(4,4′-dimethoxytrityl)thymidine (1) with allyl bromide
and K2CO3 in DMF selectively gave the N-3-allyl derivative,
(55) Burley, G. A.; Gierlich, J.; Mofid, M. R.; Nir, H.; Tal, S.; Eichen,
Y.; Carell, T. J. Am. Chem. Soc. 2006, 128, 1398-1399.
(56) Chevolot, Y.; Bouillon, C.; Vidal, S.; Morvan, F.; Meyer, A.;
Cloarec, J.-P.; Jochum, A.; Praly, J.-P.; Vasseur, J.-J.; Souteyrand, E. Angew.
Chem., Int. Ed. 2007, 46, 2398-2402.
(57) Gogoi, K.; Mane, M. V.; Kunte, S. S.; Kumar, V. A. Nucleic Acids
Res. 2007, 35, e139.
(58) Wada, T.; Mochizuki, A.; Higashiya, S.; Tsuruoka, H.; Kawahara,
S.; Ishikawa, M.; Sekine, M. Tetrahedron Lett. 2001, 42, 9215-9219.
1844 J. Org. Chem., Vol. 73, No. 5, 2008

Triazole-Linked Dumbbell Oligodeoxynucleotides
SCHEME 1. Preparation of hpODN 1 Containing
N-3-(Propargyl)thymidine and N-3-(Azidoethyl)thymidine
Derivativesa
dihydroxylation of the terminal olefin with OsO4 and 4-meth-
ylmorpholine N-oxide in aqueous acetone, oxidative cleavage
of the resulting diol with NaIO4, and reduction of the resulting
aldehyde derivative with NaBH4 in MeOH to provide the N-3-
hydroxyethyl derivative 3 in 71% yield in three steps. After
the TBS group was removed (TBAF, THF, 87%), selective
iodination of the primary hydroxyl group was conducted to give
the corresponding N-3-iodoethyl derivative 5 (I2, PPh3, pyridine,
toluene, 85%), which was converted to the conventional
phosphoramidite block 6 (NCCH2CH2OPClNi-Pr2, i-Pr2NEt,
CH2Cl2, 82%). The N-3-propargylthymidine linking to the
controlled pore glass (CPG) 9 was also easily prepared by
selective N-propargylation (propargyl bromide, K2CO3, DMF,
93%) followed by succinylation of the 3′-hydroxyl group of 7
(succinic anhydride, DMAP, Et3N, MeCN, quant.) and attach-
ment to the CPG (amino-CPG, EDCI, DMF, 34.2 µmol/g).
Authentic triazole derivatives were also prepared as shown in
Scheme 2. Thus, mesylation of the primary alcohol of 3 at the
N-3 position and substitution with azide ion gave the corre-
sponding N-3-(azidoethyl)thymidine derivative, which was
sequentially deprotected to give N-3-(azidoethyl)thymidine (Z).
N-3-(Propargyl)thymidine (P) was obtained by deprotection of
the DMTr group in 7. When an equimolar amount of Z and P
was treated with CuSO4 and sodium ascorbate in H2O, 1,4-
bisthymidine substituted 1,2,3-triazole T-tz(1,4)-T was produced
selectively in 78% yield. On the other hand, the reaction of Z
and P in the absence of Cu(I) catalyst in boiling water gave a
mixture of T-tz(1,4)-T and its regioisomer T-tz(1,5)-T (62%
for T-tz(1,4)-T, 23% for T-tz(1,5)-T, T-tz(1,4)-T/T-tz(1,5)-T
) 2.7/1). The CPG-supported oligonucleotide 10 containing
N-3-(iodoethyl)thymidine at the 5′-terminus and N-3-(propar-
gyl)thymidine (P) at the 3′-terminus was synthesized on a DNA
synthesizer in the DMTr-off mode. The CPG linking to ODN
10 was treated with NaN3 in DMF (0.5 M) for 12 h at room
temperature to give 11. After azidation, the ODN was released
from the CPG and deprotected, and the resulting crude product
was purified by reverse-phase HPLC to give hpODN 1 (Figure
3a), which was characterized by analysis of the nucleoside
composition by enzymatic digestion using snake venom phos-
phodiesterase (SVPD) and calf intestine alkaline phosphodi-
esterase (Figure 4b). Two peaks corresponding to the authentic
P and Z were observed in the HPLC chart of the reaction
mixture, and the ratio of the peak intensity of P to that of Z
was 1.0, which was the theoretical value.
For the cross-linking reactions of hpODN 1, the Cu(I) catalyst
was prepared in situ from Cu(II) sulfate and sodium ascorbate.
After the annealing of the hpODN 1, the CuAAC was carried
out by adding 10 equiv of CuSO4 to a solution of hpODN 1
(0.1 M) in 0.1 M MOPS buffer (pH 7.0) containing 5 mM
sodium ascorbate and 1 M NaCl at room temperature for 12 h
and resulted in complete consumption of the hpODN 1.
However, extensive degradation of the ODN occurred under
these conditions, a problem that has been encountered previ-
ously.46 Since as was reported in the literature, the tris-
(benzyltriazolyl)amine (TBTA) ligand53 greatly reduces degra-
dation of ODNs,48 we therefore decided to investigate the use
of the TBTA ligand in our system. A premixed CuSO4-TBTA
complex (10 equiv) was added to a solution of hpODN 1 and
sodium ascorbate in the same buffer, giving the triazole cross-
linked tzODN 1 by HPLC analysis (93%, Figure 3b). The
a Reagents and conditions: (a) allyl bromide, K2CO3, DMF, 91%; (b)
TBSCl, imidazole, DMF, 90% in 2 steps; (c) OsO4, NMO, aq acetone, then
NaIO4; (d) NaBH4, MeOH, 71% in 2 steps; (e) TBAF, THF, 87%; (f) I2,
PPh3, pyridine, toluene, 85%; (g) NCCH2CH2OPClNi-Pr2, i-Pr2NEt, CH2Cl2,
82%; (h) propargyl bromide, K2CO3, DMF, 93%; (i) succinic anhydride,
DMAP, Et3N, MeCN, quant.; (j) EDCl, amino-CPG, DMF, 34.2 µmol/g;
(k) oligonucleotide synthesis; (l) NaN3, DMF; (m) NH4OH.
which was protected with a TBS protecting group (TBSCl,
imidazole, DMF) at the 3′-position to afford 259 in 90% yield
in two steps. The allyl group of 2 was transformed into the
desired hydroxyethyl group by a three-step sequence including
(59) Hayakawa, Y.; Hirose, M.; Noyori, R. J. Org. Chem. 1993, 58,
5551-5555.
J. Org. Chem, Vol. 73, No. 5, 2008 1845

Nakane et al.
SCHEME 2. Preparation of 1,4- and 1,5-Bisthymidine-Substituted Triazole Derivatives
a Reagents and conditions: (a) MsCl, pyridine; (b) NaN3, DMF; (c) TBAF, THF, 85% in 3 steps; (d) 80% aq AcOH, 85% for Z, 97% for P; (e) CuSO4,
sodium ascorbate, H2O, 78% (T-tz(1,4)-T); (f) H2O, reflux, 62% (T-tz(1,4)-T), 23% (T-tz(1,5-T).
progress of the reaction was also monitored by denaturing
polyacrylamide gel electrophoresis (Figure 5a). After addition
of the CuSO4-TBTA complex to a solution of the ODN and
sodium ascorbate, a new band was observed with higher
mobility than that of the starting hpODN 1, and the reaction
was complete in about 4 h according to electrophoretic analysis.
This clearly showed the formation of a cyclic product with
higher gel-mobility than the linear starting hpODN 1, as
generally observed for circular ODNs by denaturing gel elec-
trophoresis. The resulting tzODN 1 was purified by reverse-
phase HPLC, and enzymatic digestion of the ODN showed that
the 1,4-disubstituted triazole-bridged thymidine dimer T-tz-
(1,4)-T was contained in the ODN, which was identical with
the authentic material by HPLC analysis (Figure 4a,c). The
resulting nucleoside composition was in good accordance with
the desired one. In the absence of Cu(I) catalyst, the reaction
generally proceeds under thermal conditions or when both the
azide and acetylene groups would be in close proximity. In these
instances, a regioisomeric mixture of 1,4-disubstituted and 1,5-
disubstituted triazoles should be obtained. In our study, only
the 1,4-disubstituted triazole-bridged thymidine dimer T-tz-
(1,4)-T was detected in the enzymatic digestion experiment, and
the corresponding 1,5-disubstituted triazole derivative T-tz-
(1,5)-T was not seen at all. The triazole cross-linked ODN was
not observed during the annealing process, and no reaction was
noted without the presence of the Cu(I) catalyst when the
precursor ODN (such as hpODN 1) in the reaction buffer was
left for several weeks at room temperature. These results clearly
show that the cross-linking reaction between the 3′- and 5′-
positions of the ODN by the CuAAC successfully proceeded
on the addition of the Cu(I) catalyst.
Having established conditions for the terminal cross-linking
of hairpin ODNs by the CuAAC, we applied this method to
the assemblage of a double-stranded ODN with the covalent
triazole cross-linking providing dumbbell ODNs, which would
act as cis-elements against NF-κB. Several cis-element se-
quences of NF-κB consisting of 10 base pairs are known to
exist. To install stem sequences at both termini of the double-
stranded ODNs, it would be necessary to prepare relatively long
single-stranded ODNs, the preparation and isolation of which
are quite often troublesome when the required sequences are
relatively long. Another strategy for designing dumbbell ODNs
would be by terminal cross-linking, namely, the preparation of
relatively short complementary single-stranded ODNs, which
are annealed to form double-stranded ODNs, with terminal
cross-linking at both termini conducted simultaneously (Figure
1b). Owing to the facile preparation of each single-stranded
ODNs, this method allows us to prepare a relatively long
dumbbell ODNs. With this in mind, we decided to investigate
the triazole cross-linking of double-stranded ODNs at both
termini, although it proved to be challenging. Constraining the
double-stranded ODNs by terminal cross-linking could cause
structural deformation and/or limit conformational flexibility.
Protein-nucleic acid interactions sometimes involve an induced-
fit in their binding, and their affinity would be reduced with
double-stranded DNAs with a constrained conformation. How-
ever, increasing the size of the loop might reduce the constraint
within double-stranded ODNs. Therefore, we designed a series
of dsODNs 1-3 (Figure 2) that contained an 11-mer NF-κB
binding sequence with Z and P at the 3′- and 5′-termini,
respectively, while varying the number of the terminal thymidine
residues to observe the impact of the loop size on the cross-
linking reaction and the NF-κB binding ability. The nucleotide
sequence corresponding to a single asymmetric NF-κB binding
site of the HIV-1 RTL was chosen to minimize potential
problems related to self- and/or intrastrand hybridization as
previously investigated.27 In a manner similar to the preparation
of the model hpODN 1, dsODNs 1-3 were synthesized and
purified, and their structure was confirmed by analysis of
nucleoside composition by enzymatic digestion (data not
shown). With these dsODNs in hand, we then conducted the
triazole cross-linking at both termini. DsODNs 1-3 were
prepared by annealing equimolar amounts of the corresponding
single-stranded ODNs in a 0.1 M MOPS buffer (pH 7.0)
containing 1 M NaCl (each ODN concentration was 0.1 M).
The CuAAC of dsODN 1 was carried out under the same
conditions as for the one with hpODN 1 and monitored by
denaturing polyacrylamide gel electrophoresis (Figure 5b). After
the CuSO4-TBTA complex was added, a new band with lower
mobility than that of the starting dsODN 1 was observed, and
the reaction was complete within 4 h. The reaction proceeded
efficiently (Figure 3c,d), and the desired triazole cross-linked
ODN, tz2ODN 1, was obtained after HPLC purification. The
conversion was 95% by HPLC analysis. The structure of tz2-
ODN 1 was also confirmed by enzymatic digestion to be the
same as tzODN 1. Thus, no peaks corresponding to Z and P
were detected by HPLC analysis, and the 1,4-disubstituted
triazole-bridged thymidine dimer T-tz(1,4)-T was formed
instead (Figure 4d). This would imply that the triazole cross-
linking indeed occurred at both termini providing a dumbbell-
1846 J. Org. Chem., Vol. 73, No. 5, 2008

Triazole-Linked Dumbbell Oligodeoxynucleotides
FIGURE 3. Denaturing reverse-phase HPLC chromatograms (UV absorbance vs time) of the cross-linking reaction of hpODN 1 and dsODNs 1,
2, and 3: (a) hpODN 1 in the absence of Cu[I] catalysis; (b) after the CuAAC of hpODN 1; (c) dsODN 1 in the absence of Cu[I] catalysis; (d) after
the CuAAC of dsODN 1; (e) dsODN 2 in the absence of Cu[I] catalysis; (f) after the CuAAC of dsODN 2; (g) dsODN 3 in the absence of Cu[I]
catalysis; (h) after the CuAAC of dsODN 3. All the reactions were performed with 50 µM ODN in 100 mM NOPS-NaOH (pH 7, 1 M NaCl-
tBuOH) in the presence of 500 mM CuSO4-TBTA complex and 5 mM sodium ascorbate for 6 h to afford tzODN 1 and tz2ODN 1, 2, and 3
(indicated as arrows), respectively.
like ODN. In a manner similar to the cross-linking reaction of
dsODN 1, the CuAAC of the dsODNs 2 and 3, which possess
extra thymidine residues at both termini, proceeded smoothly
to provide tz2ODNs 2 (Figures 3e,f and 4e) and 3 (Figures 3g,h
and 4f), respectively.
CD spectroscopy was used to study the macroscopic helical
geometry of the chemically modified tz2ODNs 1-3 together
with the natural double-stranded control ODN (contODN). The
natural double-stranded 17-mer ODN (contODN) displayed a
characteristic B-DNA spectrum possessing a positive ellipticity
at 283 nm, a negative ellipticity at 252 nm, and a crossover
between 255 and 262 nm (Figure 6).60 Tz2ODNs 1-3 possessing
the triazole cross-linked thymidines at both ends of the helix
displayed an increase in positive amplitude at 283 nm and an
increase in the negative one without any other significant
changes in the overall CD spectrum. A similar CD characteristic
was observed in double-stranded ODNs possessing disulfide
bridges at the termini in a study by Glick et al.36 They speculated
that differences in the CD spectra between a natural ODN and
the cross-linked ODN may simply arise from inherent differ-
ences in the composition of the sequences, or the increase in
ellipticity at 283 nm may be related to an increased winding of
the helix induced by the presence of the thymidine bases. It
could also be speculated that, due to its analogous structure,
our triazole cross-linked ODNs would also possess similar
properties. Nevertheless, the general correspondence in the CD
curves suggests that the terminal base modifications do not
significantly distort the helical geometry of the duplex.
UV melting experiments were next conducted to characterize
the thermally induced denaturation of tz2ODNs with the goal
of elucidating the thermal consequences of modifying and
constraining DNA via the triazole cross-linking (Table 1). The
Tm values for tz2ODNs 1-3 were 89.8, 85.2, and 81.7 °C,
respectively. An increase in the number of thymidine residues
(60) Ivanov, V. I.; Minchenkova, L. E.; Schyolkina, A. K.; Poletayev,
A. I. Biopolymers 1973, 12, 89-110.
J. Org. Chem, Vol. 73, No. 5, 2008 1847

Nakane et al.
FIGURE 4. HPLC profiles (UV absorbance vs time) of enzymatic digestion of ODNs and their nucleoside compositions: (a) A mixture of authentic
nucleosides; (b) hpODN 1; (c) tz2ODN 1; (d) tz2ODN 1; (e) tz2ODN 2; and (f) tz2ODN 3.
TABLE 1. Thermal Stability of the Triazole Cross-Linked
Dumbbell ODNs
Tm (°C)a
∆Tm (°C)b
contODN
dsODN 1
dsODN 2
dsODN 3
tz2ODN 1
tz2ODN 2
tz2ODN 3
55.7
52.7
50.3
47.0
89.8
85.2
81.7
-3.0
-5.4
-8.7
+34.1
+29.5
+26.0
FIGURE 5. Denatured polyacrylamide gel electrophoresis of the cross-
linking reaction of (a) hpODN 1 and (b) dsODN 1. Arrows indicate
triazole-cross-linked ODNs.
a Conditions: 3 µM each ODN, 1 mM sodium cacodylate-HCI (pH
7.0), 1 mM NaC1. Average of three measurements. b Tm(modified) -
Tm(control).
the triazole cross-link, and it is estimated that the double triazole
cross-link results in an increase in thermal stability ranging from
26 to 34 °C, which is similar to that of a disulfide cross-linked
12-mer ODN (38.1 °C) reported by Glick et al., although the
sequence and length of the ODN were different.35,36
Cleavage by a 3′-exonuclease, the activity of which is
predominant in human serum,6 was next examined since an
enzyme recognition site is proximal to our modified thymidines.
Snake venom phosphodiesterase (SVPD) degrades DNA from
the 3′-end of a double-stranded DNA releasing 5′-dNMPs.61
Cross-linking on the terminal bases leaves the terminal hy-
droxyls available for [32P]-end-labeling by a standard procedure
used in ODN chemistry. SVPD degrades the control ODN
rapidly (t1/2 of 10.4 ( 2.4 min, Figure 7a). On the other hand,
FIGURE 6. Circular dichroism spectra (conditions: 2 µM of ODN in
sodium cacodylate-HCl buffer (1 mM, pH 7.0) containing 10 mM
NaCl) of contODN and tz2ODN 1-3.
in the loops resulted in a decrease in thermal stability.
Comparing the Tm values for cross-linked tz2ODNs (81.7-
89.8 °C) with those of the control double-stranded ODN
(55.7 °C) revealed a dramatic increase in thermal stability with
(61) Williams, E. J.; Sung, S.-C.; Laskowski, M., Sr. J. Biol. Chem. 1961,
236, 1130-1134.
1848 J. Org. Chem., Vol. 73, No. 5, 2008

Triazole-Linked Dumbbell Oligodeoxynucleotides
FIGURE 7. Susceptibility to snake venom phosphodiesterase degradation of triazole cross-linked dumbbell ODNs: (a) contODN; (b) tz2ODN 1;
(c) tz2ODN 2; and (d) tz2ODN 3. 5′-32P-labeled control contODN and tz2ODNs 1-3 were incubated for different lengths of time, as indicated, with
64 µU snake venom phosphodiesterase in 40 mM Tris-HCl buffer (pH 7.5) at 37 °C and then submitted to electrophoretic separation on 20%
denatured polyacrylamide gels.
tz2ODNs 1-3 were revealed to be totally resistant to SVPD
degradation, and the apparent half-lives are 500-fold greater
(t1/2 > 24 h, Figure 7b-d). These data suggest that the
nucleolytic activity may require the enzyme to partially expose
phosphodiester bonds near the 3′-side62 by unwinding or
“melting” the helix upon binding, which is limited by the
topological constraint imposed by the cross-link.36,63,64 In
conjunction with increased thermal stability, the newly synthe-
sized dumbbell ODNs possess suitable properties required for
biological application of decoy molecules.
With the synthetic strategy and the basic structural and
nuclease resistant properties of the ODNs established, the ability
of tz2ODNs 1-3 to bind with NF-κB was investigated by
electrophoretic mobility shift assays (EMSA) in a procedure
similar to that reported previouly.65 32P-labeled ODNs were
incubated with various amounts of NF-κB p50 homodimer at
0 °C for 30 min, and the mixture was analyzed by gel
electrophoresis. A single protein-ODN complex migrating with
the same electrophoretic mobility was observed with all probes
as well as with the control ODN, as shown in Figure 8. The
results indicated that formation of the protein-ODN complexes
had occurred. Apparent dissociation constants (Kd) of tz2ODNs
1-3 were determined to be 18 ( 4, 14 ( 1, and 9 ( 2 nM,
respectively, and these ODNs exhibited a similar binding affinity
to that of the control decoy ODN (12 ( 3 nM). Furthermore,
formation of the complex of 32P-labeled control ODN and the
NF-κB p50 homodimer was competitively inhibited by addition
of unlabeled tz2ODNs 1-3 in a dose-dependent fashion, and
their IC50 values were 3.9 ( 0.5 nM for tz2ODN 1, 2.6 ( 0.3
nM for tz2ODN 2, and 3.0 ( 0.5 nM for tz2ODN 3 (Figure 9).
FIGURE 8. Effects of the triazole cross-linked dumbbell ODNs on
This would indicate that the newly synthesized dumbbell ODNs
were recognized by NF-κB p50 homodimer in a manner similar
to that of the control decoy ODN and that they act as decoy
molecules. Of particular interest is the observation of different
binding of the synthesized modular dumbbell ODNs to the
protein. Comparison of the dumbbell ODNs with their binding
properties to NF-κB reveals some important differences. Tz2-
ODN 3 with six-base loops has the highest binding affinity
the interaction with the NF-κB p50 homodimer and the electromobility
shift assay with NF-κB. 32P-radiolabeled triazole cross-linked dumbbell
ODNs (40 fmol) was incubated at 0 °C for 30 min in 10 mM HEPES
buffer (pH 7.5) in the presence of the indicated concentration of NF-
κB p50 homodimer: (a) contODN; (b) tz2ODN 1; (c) tz2ODN 2; and
(d) tz2ODN 3. Protein-ODN complexes are marked with an arrow.
(e) Binding curves with NF-κB and triazole cross-linked dumbbell
ODNs. Each dissociation constant was calculated by three independent
experiments and shown as average (S.E.
among the three dumbbell ODNs, and the greater the number
of thymidine residues constituting the loop region, the higher
the binding affinity of the dumbbell ODNs to the NF-κB,
although the magnitude is subtle. Recently, several solution
structures of free DNAs containing NF-κB recognition se-
quences and their complexes with NF-κB dimer have been
(62) Tarko¨y, M.; Leumann, C. Angew. Chem., Int. Ed. Engl. 1993, 32,
1432-1434.
(63) Cowart, M.; Gibson, K. J.; Allen, D. J.; Benkovic, S. J. Biochemistry
1989, 28, 1975-1983.
(64) Catalano, C. E.; Benkovic, S. J. Biochemistry 1989, 28, 4374-4382.
(65) Murata, S.; Mizumura, Y.; Hino, K.; Ueno, Y.; Ichikawa, S.;
Matsuda, M. J. Am. Chem. Soc. 2007, 129, 10300-10301.
J. Org. Chem, Vol. 73, No. 5, 2008 1849

Nakane et al.
not prevent the dumbbell ODNs from binding to the NF-κB
transcription factor. Therefore, the results obtained conclusively
demonstrate that the triazole cross-linked dumbbell ODNs can
be proposed as powerful decoy molecules.
Experimental Section
1-(2-Deoxy-â-D-ribofuranosyl)-3-(2-azidoethyl)thymine (Z). A
solution of 1-[2-deoxy-5-O-(4,4′-dimethoxytrityl)-â-D-ribofurano-
syl]-3-(2-azidoethyl)thymine (97 mg, 0.16 mmol) in 80% aqueous
AcOH (10 mL) was stirred for 2 h at room temperature. The mixture
was concentrated in vacuo, and the resulting residue was co-
evapolated with EtOH (3×). The residue was purified by a silica
gel column (50-100% AcOEt-hexane) to afford Z (45 mg, 91%)
as a white powder. For analysis, a part of the material was
crystallized from AcOEt-hexane (mp 82 °C): 1H NMR (400 MHz,
CDCl3) δ 7.37 (s, 1H), 6.18 (dd, 1H, J ) 6.4, 7.3 Hz), 4.59 (m,
1H), 4.19 (t, 2H, J ) 6.2 Hz), 4.01 (m, 1H), 3.94 (dd, 1H, J ) 2.8,
11.9 Hz), 3.84 (dd, 1H, J ) 3.1, 11.9 Hz), 3.53 (t, 2H, J ) 6.2
Hz), 2.50, 2.29 (each s, each br s), 2.43 (ddd, 1H, J ) 3.8, 6.4,
13.7 Hz), 2.33 (ddd, 1H, J ) 6.4, 7.3, 13.7 Hz), 1.94 (s, 3H); 13C
NMR (100 MHz, CDCl3) δ 12.5, 40.0, 40.2, 48.4, 62.4, 71.4, 86.9,
87.0, 110.3, 135.1, 150.9, 163.2; UV (λmax ) 268 nm, ꢀ260 ) 7910
M-1 cm-1, ꢀ268 ) 9050 M-1 cm-1). Anal. Calcd for C12H17N5O5:
C, 46.30; H, 5.50; N, 22.50. Found: C, 46.56; H, 5.46; N, 22.08.
LRMS (FAB) m/z 312 (MH+); HRMS (FAB) calcd for C12H18N5O5
(MH+) 312.1308, found 312.1313.
1-(2-Deoxy-â-D-ribofuranosyl)-3-(2-propynyl)thymine (P). A
solution of 7 (400 mg, 0.68 mmol) in 80% aqueous AcOH (10
mL) was stirred for 2 h at room temperature. The mixture was
concentrated in vacuo, and the resulting residue was co-evapolated
with EtOH (3×). The residue was purified by a silica gel column
(0-4% MeOH-CHCl3) to afford P (190 mg, 97%) as a white
powder. For analysis, a part of the material was crystallized from
AcOEt-hexane (mp 87-88 °C). The data were identical in all
respects with the properties previously reported:71 1H NMR (400
MHz, CDCl3) δ 7.45 (s, 1H), 6.24 (t, 1H, J ) 6.8 Hz), 4.70 (d,
2H, J ) 1.8 Hz), 4.59 (m, 1H), 4.02 (m, 1H), 3.93 (dd, 1H, J )
2.8, 11.9 Hz), 3.85 (dd, 1H, J ) 2.4, 11.9 Hz), 2.55 (m, 2H), 2.38
(m, 2H), 2.18 (d, 1H, J ) 1.8 Hz), 1.95 (s, 3H); UV (λmax ) 268
nm, ꢀ260 ) 8090 M-1 cm-1, ꢀ268 ) 9060 M-1 cm-1). Anal. Calcd
for C13H16N2O5: C, 55.71; H, 5.75; N, 9.99. Found: C, 55.56; H,
5.75; N, 10.00.
FIGURE 9. Competition curves of triazole cross-linked dumbbell
ODNs for binding of NF-κB p50 homodimer to radiolabeled control
ODN. Competition experiments were performed by incubating NF-κB
p50 homodimer (40 fmol) together with the 32P-radiolabeled ODN (40
fmol) and the competitor ODNs. IC50 values for each type of competitor
were estimated by plotting these data as a competitor concentration
(M) and fitting to a dose-response curve. Each IC50 value was
calculated by three independent experiments and shown as average
(S.E.
reported.66-70 These structures allow investigation of the role
of DNA conformation on specific NF-κB dimer recognition and
binding. It is suggested that the NF-κB family discriminates
the subtle structural alteration of trans-elements including
hydration, bending, and unwinding. It can be speculated that
the terminal cross-link might result in some structural changes
even in internal double-stranded structures of the ODNs without
significantly changing CD spectra. ODNs with shorter loops
have more reduced conformational flexibility. As a result, their
binding affinity with the protein is reduced compared to the
parent unmodified decoy ODN because they cannot adopt a
structural alteration suitable for protein binding. Increasing the
size of the loop might reduce the constraint within double-
stranded ODNs to recover conformational flexibility.
Conclusion
Triazole-cross-linked ODNs were synthesized using the
CuAAC with hairpin DNA possessing N-3-(azidoethyl)thymi-
dine and N-3-(propargyl)thymidine at the 3′- and 5′-termini. The
newly synthesized ODNs were thermally stable and their
structures resembled those of non-cross-linked ODNs. This
strategy is effective for the preparation of cross-linked ODNs
because the precursor ODNs are stable and easy to prepare, and
the cross-link reactions proceed chemoselectively. The newly
synthesized dumbbell ODNs possess good properties, including
excellent stability against exonuclease and high thermal stability,
which are necessary for the decoy molecules in order to achieve
biological responses leading to alteration of gene expression. It
was also demonstrated that the dumbbell ODNs have the ability
to bind to NF-κB p50 homodimer within a range similar to that
of the control decoy ODN. Our methodology allows us to
prepare dumbbell ODNs with various loop lengths. This study
clearly shows that the cross-linking by the triazole structure does
Cycloaddition in the Presence of Cu(I) Catalyst. Sodium
ascorbate (5 mg, 50 µmol) and an aqueous solution of CuSO4 (10
mM, 1 mL, 10 µmol) were added to a solution of Z (31 mg, 0.10
mmol) and P (28 mg, 0.10 mmol) in H2O (1 mL), and the mixture
was stirred for 8 h at room temperature. The mixture was
concentrated in vacuo, and the residue was purified by a silica gel
column (10% MeOH-CHCl3) to afford T-tz(1,4)-T (46 mg, 78%)
as a white powder: 1H NMR (500 Hz, CD3OD) δ 7.88 (s, 1H),
7.85, 7.80 (each s, each 1H), 6.31, 6.12 (each m, each 1H), 5.17
(s, 2H), 4.64 (s, 2H), 4.35 (m, 4H), 3.91 (m, 2H), 3.82-3.70 (m,
4H), 2.28-2.14 (m, 4H), 1.92, 1.83 (each s, each 3H); 13C NMR
(100 MHz, CDCl3) δ 13.9, 14.0, 37.7, 42.2, 42.3, 42.5, 63.4, 63.5,
72.6, 72.8, 88.0, 88.1, 89.6, 111.2, 111.5, 126.7, 137.5, 145.5, 152.7,
152.8, 165.7, 165.8; UV (λmax ) 268 nm, ꢀ260 ) 13400 M-1 cm-1
,
ꢀ268 ) 15200 M-1 cm-1); LRMS (FAB) m/z 592 (MH+); HRMS
(FAB) calcd for C25H34N7O10 (MH+) 592.2367, found 592.2379.
Cycloaddition under Thermal Conditions. A solution of Z (31
mg, 0.10 mmol) and P (28 mg, 0.10 mmol) in H2O (1 mL) was
heated under reflux for 8 days. The mixture was concentrated in
vacuo, and the residue was purified by a silica gel column (1-
14% MeOH-CHCl3) to afford T-tz(1,4)-T (37 mg, 62%) as a white
powder and T-tz(1,5)-T (13 mg, 23%) as a white powder. Data
(66) Muller, C. W.; Harrison, S. C. FEBS Lett. 1995, 369, 113-117.
(67) Ghosh, G.; van Duyne, G.; Ghosh, S.; Sigler, P. B. Nature 1995,
373, 303-310.
(68) Chen, F. E.; Huang, D. B.; Chen, Y. Q.; Ghosh, G. Nature 1998,
391, 410-413.
(69) Chen, Y. Q.; Ghosh, S.; Ghosh, G.; Cramer, P.; Larson, C. J.;
Verdine, G. L.; Muller, C. W. EMBO J. 1997, 16, 7078-7090.
(70) Huang, D.-B.; Phelps, C. B.; Fusco, A. J.; Ghosh, G. J. Mol. Biol.
2005, 346, 147-160.
(71) Byun, Y.; Al-Madhoun, A. S.; Johnsamuel, J.; Yang, W.; Barth, R.
F.; Eriksson, S.; Tjarks, W. J. Med. Chem. 2005, 48, 1188-1198.
1850 J. Org. Chem., Vol. 73, No. 5, 2008

Triazole-Linked Dumbbell Oligodeoxynucleotides
for T-tz(1,5)-T: 1H NMR (500 Hz, CD3OD) δ 7.88, 7.81 (each s,
each 1H), 7.64 (s, 1H), 6.31, 6.11 (each m, each 1H), 5.33, 5.29
(each d, each 1H, J ) 15.1 Hz), 4.91 (s, 2H), 4.39 (m, 4H), 3.92
(m, 2H), 3.80-3.73 (m, 4H), 2.28-2.13 (m, 4H), 1.94, 1.85 (each
s, each 3H); 13C NMR (100 MHz, CDCl3) δ 13.8, 13.9, 34.4, 42.1,
42.2, 42.3, 47.5, 63.5, 72.6, 72.7, 88.1, 89.6, 89.7, 111.1, 111.4,
136.2, 136.7, 137.5, 137.8, 152.8, 165.7; LRMS (FAB) m/z 592
(MH+); HRMS (FAB) calcd for C25H34N7O10 (MH+) 592.2367,
found 592.2358.
UV Melting Experiments. Thermally induced transitions were
monitored at 260 nm on a Beckman DU 650 spectrophotpmeter.
A solution containing 3 µM samples in a buffer of 10 mM sodium
cacodylate (pH 7.0) containing 100 mM NaCl was heated at 90 °C
for 5 min, then cooled gradually to room temperature and used for
the thermal denaturation study. The sample temperature was
increased at a rate of 0.5 deg/min.
Labeling of the 5′-Ends of Oligodeoxynucleotides. A mixture
of oligodeoxynucleotides (100 pmol), [γ-32P]-ATP (1.7 nmol), and
T4 polynucleotide kinase (Takara Bio Inc., 1 U) in a buffer (20
µL) containing Tris-HCl (50 mM, pH 8.0), MgCl2 (10 mM), and
DTT (5 mM) was incubated for 45-120 min at 37 °C and heated
at 100 °C for 5 min. The resulting solution was purified by column
chromatography (C18, YMC dispo SPE, 100 mg/mL) to give 5′-
32P-labeled oligodeoxynucleotides.
Preparation of Cross-Linking Precursor Oligodeoxynucle-
otides. Oligonucleotides containing N-3-(iodoethyl)thymidine at the
5′-terminus and N-3-(propargylethyl)thymidine (P) at the 3′-
terminus were synthesized on a DNA synthesizer (Applied Bio-
systems Model 3400) in the DMTr-off mode, as we previously
reported.24 CPG supporting ODNs was treated with 0.5 M NaN3
in DMF (2 mL) at room temperature for 12 h. The CPG was
filtrated, washed successively with ethanol and acetone, and dried.
ODNs were released from CPG by treatment with 28% aqueous
ammonia at 55 °C for 16 h, and the resulting solution was
concentrated and purified by reverse-phase HPLC using a linear
gradient of 5-20.75% MeCN in 0.1 M TEAA buffer (pH 7.0).
CuAAC of ODNs. A solution of a duplex (25 nmol) composed
of two oligodeoxynucleotides was heated in 100 mM MOPS-
NaOH buffer (pH 7.0) containing 1 M NaCl at 90 °C for 1 min
and allowed to cool to room temperature for 4 h. A solution of
CuSO4-tris(benzyltriazolyl)amine53 (TBTA) in 50% aqueous t-
BuOH (28 mM, 9 µL, 250 nmol) and sodium ascorbate (2.5 µmol)
was sequentially added to a solution of the duplex, and the mixture
was incubated at 25 °C for 12 h. Hydrogen sulfide gas was passed
through the reaction mixture, which was washed with chloroform
(3×). The aqueous solution was desalted on a NAP-25 column (GE
Healthcare) and then lyophilized. The resulting cross-linking ODN
was analyzed and purified by reverse-phase HPLC using a linear
gradient of 5-25% MeCN in 0.1 M TEAA buffer (pH 7.0).
Enzymatic Digestion for Analyses of Nucleoside Composition
of ODNs. A mixture of oligodeoxynucleotides (0.2 OD), snake
venom phosphodiesterase (MP Biochemicals, 2.0 mg/mL; 6 µL),
nuclease P1 (Yamasa Co., 1 unit/µL; 5 µL) and calf intestine
alkaline phosphatase (Takara Bio Inc., 0.1 U/mL; 5 µL) in 0.1 M
Tris-HCl, 2 mM MgCl2 buffer (pH 7.7) in a total volume of 500
µL was incubated at 37 °C for 12 h. After the reaction mixture
was heated at 100 °C for 5 min, the enzymes were filtered off
through Micropure-EZ device (Millipore Co.). The filtrate was
concentrated to dryness, and the residue was dissolved in H2O (100
µL) and analyzed by reverse-phase HPLC with a 5-25% MeCN
in 0.1 M TEAA buffer (pH 7.0) linear gradient. Nucleoside
composition was quantified by dividing each peak area from the
HPLC profile, and the results were the following. dsODN 1: upper,
calcd for dC:dG:T:dA:2:3 ) 5:6:4:2:1:1, found 5.9:6.1:3.8:1.8:1.0:
1.1; lower, calcd for dC:dG:T:dA:2:3 ) 6:5:2:4:1:1, found 6.0:5.2:
2.1:3.7:1.0:1.2. dsODN 2: upper, calcd for dC:dG:T:dA:2:3 ) 5:6:
6:2:1:1, found 5.2:6.0:5.8:2.0:1.1:1.0; lower, calcd for dC:dG:T:
dA:2:3 ) 6:5:4:4:1:1, found 6.3:5.3:4.2:3.7:1.0:1.0. dsODN 3:
upper, calcd for dC:dG:T:dA:2:3 ) 5:6:8:2:1:1, found 5.3:5.9:7.8:
2.0:1.2:1.0; lower, calcd for dC:dG:T:dA:2:3 ) 6:5:6:4:1:1, found
5.7:5.3:6.2:3.9:0.9:1.2. tz2ODN 1: calcd for dC:dG:T:dA:12 ) 11:
11:6:6:2, found 11.3:11.2:6.1:5.7:1.9. tz2ODN 2: calcd for dC:
dG:T:dA:12 ) 11:11:8:6:2, found 11.2:11.3:7.9:5.7:1.9. tz2ODN
3: calcd for dC:dG:T:dA:12 ) 11:11:10:6:2, found 10.9:10.8:10.3:
5.8:1.9
Determination of Dissociation Constant (Kd) with NF-κB p50
Homodimer by Electromobility Sift Assay. 32P-Labeled oligo-
nucleotide duplexes (40 fmol) and NF-κB p50 homodimer (Promega
Co., various concentration: 0, 1, 5, 10, 15, 20, 50, 100 nM) in
HEPES buffer (10 mM, pH 7.5) containing MgCl2 (10 mM), LiCl
(50 mM), NaCl (100 mM), spermidine (1 mM), BSA (0.2 mg/
mL), poly(dI:dC) (1 nM), IGEPAL-CA630 (0.05%), and glycerol
(10%) were incubated at 4 °C for 40 min. The mixture was loaded
on 8% native polyacrylamide gel (39:1, acrylamide:bisacrylamide),
which was pre-cooled and pre-run (120 V, 1 h) at 4 °C. The gel
was electrophoresed at 120 V (4 °C) for 40 min, dried under reduced
pressure at 80 °C, and imaged on a IP plate (Fuji Film) for 1-5 h.
Quantitative measurements of bound and free oligodeoxynucleotides
were carried out by image-analyzer (BAS-2500, Fuji Film). The
data, which were plotted as a concentration of NF-κB versus binding
ratio, were fit to nonlinear regression according to the following
equation: y ) Bmaxx/(Kd + x).
Competition Assay. 32P-Labeled contODN (40 fmol), NF-κB
p50 homodimers (40 fmol), and dumbbell oligodeoxynucleotides
(various concentration: 0, 0.2, 1, 2, 4, 10, 20, 40, 200 nM) in
HEPES buffer (10 mM, pH 7.5) containing MgCl2 (10 mM), LiCl
(50 mM), NaCl (100 mM), spermidine (1 mM), BSA (0.2 mg/
mL), poly(dI:dC) (1 nM), IGEPAL-CA630 (0.05%), and glycerol
(10%) were incubated at 4 °C for 40 min. The mixture was loaded
on 8% native polyacrylamide gel (39:1, acrylamide:bisacrylamide),
which was pre-cooled and pre-run (120 V, 1 h) at 4 °C. The gel
was electrophoresed at 120 V (4 °C) for 40 min, dried under reduced
pressure at 80 °C, and imaged on a IP plate (Fuji Film) for 1-5 h.
The data were analyzed in a manner similar to that of the
dissociation constants.
Acknowledgment. This work was supported by grants-in-
aid for scientific research from the Ministry of Education. We
thank Ms. S. Oka (Center for Instrumental Analysis, Hokkaido
University) for measurement of the mass spectra.
Supporting Information Available: Experimental procedures
and 1H NMR and 13C NMR spectra for 3-9, P, and Z. This material
is available free of charge via the Internet at http://pubs.acs.org.
JO702459B
J. Org. Chem, Vol. 73, No. 5, 2008 1851