Hybridization-promoted and cytidine-selective activation for cross-linking with the use of 2-amino-6-vinylpurine derivatives

Article abstract of DOI:10.1021/jo048298p

(Chemical Equation Presented). Recently, we have proposed a new concept for cross-linking agents with inducible reactivity, in which the highly reactive cross-linking agent, the 2-amino-6-vinylpurine nucleoside analogue (1), can be regenerated in situ fro

Full text of DOI:10.1021/jo048298p

Hybridization-Promoted and Cytidine-Selective Activation for
Cross-Linking with the Use of 2-Amino-6-vinylpurine Derivatives
Takeshi Kawasaki,† Fumi Nagatsugi,†,‡ Md. Monsur Ali,†,‡ Minoru Maeda,†
Kumiko Sugiyama,§ Kenji Hori,§ and Shigeki Sasaki*, †,‡
Graduate School of Pharmaceutical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku,
Fukuoka 812-8582, Japan, CREST, Japan Science and Technology Agency, Kawaguchi Center Building,
4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan, and Faculty of Engineering, Yamaguchi University,
Tokiwadai, Ube 755-8611, Japan
sasaki@phar.kyushu-u.ac.jp
Received September 24, 2004
Recently, we have proposed a new concept for cross-linking agents with inducible reactivity, in
which the highly reactive cross-linking agent, the 2-amino-6-vinylpurine nucleoside analogue (1),
can be regenerated in situ from its stable precursors, the phenylsulfide (4) and the phenylsulfoxide
(3) derivatives, by a hybridization-promoted activation process with selectivity to cytidine. The
phenylsulfide precursor (4) exhibited cross-linking ability despite its high stability toward strong
nucleophiles such as amines and thiols. In this study, we investigated the substituent effects of
the phenylsulfide group on the cross-linking reaction, and determined the 2-carboxy substituent of
the phenylsulfide derivative (11k) as an efficient cross-linking agent with inducible reactivity.
Detailed investigations have shown that the phenylsulfoxide (3) and phenylsulfide (4) precursors
produce the 2-amino-6-vinylpurine nucleoside (1) as the common reactive species. It has been
concluded that the nature of the inducible reactivity of the precursors (3 and 4) is acceleration of
their elimination to the 2-amino-6-vinylpurine nucleoside (1) through the selective process in the
duplex with the ODN having cytidine at the target site.
Introduction
directed mutagenesis (or site-directed manipulation of
genes) toward the target duplex has been proposed as a
new biological tool with the use of a triplex-forming
oligonucleotide.6,7 Strand invasion has been recently
demonstrated by the use of ODNs attaching a cross-
linking agent to triplex-forming guide sequences.8 Inter-
est in a variety of potential applications of reactive ODNs
has encouraged us to generate new cross-linking agents
that might be applicable to both in vitro and in vivo
studies.
Oligodeoxynucleotides (ODN) and their analogues have
been widely used for the inhibition of gene expression in
antisense or antigene methods, etc.1 ODNs attaching a
functional reagent may cause irreversible change to the
target sequence.2 Inhibition of the elongation step has
been achieved by the use of reactive antisense ODNs
conjugating a group for alkylation.3-5 Furthermore, site-
* Address correspondence to this author. Phone: 81-92-642-6615.
Fax: 81-92-642-6876.
There are several types of alkylating agents that have
† Kyushu University.
been conjugated to the ODNs.9 Halocarbonyl10-16 or
‡ CREST, Japan Science and Technology Agency.
§ Yamaguchi University.
(1) A comprehensive review: Murray, J. A., Ed. Antisense RNA and
DNA; John Wiley & Sons, INC.: New York, 1992.
(2) Beaucage, S. L.; Iyer, R. P. Tetrahedron 1993, 49, 1295-1963.
(3) Kean, J. M.; Murakami, A.; Blake, K.; Cushman, C.; Miller, P.
S. Biochemistry 1988, 27, 9113-9121.
(5) Boiziau, C.; Kurfurst, R.; Cacenave, C.; Roig, V.; Thuong, N. T.;
Toulme, J.-J. Nucleic Acids Res. 1991, 19, 1113-1119.
(6) Woolf, T. M. Nature Biotechnol. 1988, 16, 341-344.
(7) Chan, P. P.; Lin, M.; Faruqi, A. F.; Powell, J.; Seidman, M. M.;
Glazer, P. M. J. Biol. Chem. 1999, 274, 11541-11548.
(4) Lee, B. L.; Murakami, A.; Blake, K. R.; Lin, S.-B.; Miller, P. S.
Biochemistry 1988, 27, 3197-3203.
(8) Gamper, H. B., Jr.; Hou, Ya-M.; Stamm, M. R.; Podyminogin,
M. A.; Meyer, R. B. J. Am. Chem. Soc. 1998, 120, 2182-2183.
10.1021/jo048298p CCC: $30.25 © 2005 American Chemical Society
Published on Web 12/02/2004
14
J. Org. Chem. 2005, 70, 14-23

Cytidine-Selective Activation for Cross-Linking
FIGURE 1. The new concept of cross-linking agent that can be activated in situ by duplex formation as the trigger.
aziridine17-20 groups alkylated bases in an SN2 manner
(SN2 reagents). Highly reactive functional groups such
as haloacetyl have been used for alkylation at N7 of
guanosine, whereas their high reactivities might suffer
disadvantages for biological applications. As the alkylat-
ing group of ODNs should receive nucleophilic attack by
a number of chemical entities in the living system such
as water, amines, and thiols, etc., achievement of both
high reactivity and stability of the alkylating group falls
into a dilemma. A promising approach to overcome such
problems is the use of agents that have inducible reactiv-
ity by triggering signals such as UV-irradiation, enzy-
matic, or chemical reactions. Some compounds, such as
ET-743 and CC-1065, are furnished with inducible
reactivity that is activated in the microenvironments of
the DNA.21-26 Psoralen and its analogues have enjoyed
remarkable successes as T-selective photo-cross-linking
agents that are applied to a wide variety of in vitro as
SCHEME 1. Selective Cross-Linking of Cytosine by
2-Amino-6-vinylpurine (1)
well as in vivo studies.27-31 However, as UV irradiation
might cause damages to other sites of DNA and UV light
might not be suitable for in vivo study, researchers have
been seeking other types of alkylating agents with
inducible reactivity. Rokita’s group developed the inter-
mediates of a quinone methide as inducible alkylating
agents that can be activated by enzymatic reduction32,33
or chemical reaction.34-36 Quinone methide derivatives
have been used to propose a general strategy for target-
promoted alkylation.37 An interesting cross-linking reac-
tion has been proposed in the hybridization-activated
conversion of the manondialdehyde M1G adduct to ring-
opened oxopropenyl-dG adduct.38
(9) A recent review: Sasaki, S. Eur. J. Pharm. Sci 2001, 13, 43-
51.
(10) Coleman, R. S.; Pires. R. M. Nucleic Acids Res. 1997, 25, 4771-
4777.
Our group has developed 2-amino-6-vinylpurine nu-
cleoside (1) as a selective cross-linking agent to cytidine
(Scheme 1).39,40 In our original design, a close proximity
(11) Coleman, R. S.; Kesicki, E. A. J. Org. Chem. 1995, 60, 6252-
6253.
(12) Tabone, J. C.; Stamm, M. R.; Gamper, H. B.; Meyer, R. B., Jr.
Biochemistry 1994, 33, 375-383.
(13) Meyer, R. B., Jr.; Tabone, J. C.; Hurst, G. D.; Smith, T. M.;
Gamper, H. J. Am. Chem. Soc. 1989 111, 8517-8519.
(14) Grant, K. B.; Dervan, P. B. Biochemistry 1996, 35, 12313-
12319.
(27) Murakami A.; Yamayoshi, A.; Iwase, R.; Nishida, J.; Yamaoka,
T.; Wake, N. Eur. J. Pharm. Sci. 2001, 13, 25-34.
(28) Takasugi, M.; Guendouz, A.; Chassigonl, M.; Decout, J. L.;
Lhomme, J.; Thuong, N. T.; Helene, C. Proc. Natl. Acad. Sci. U.S.A.
1991, 88, 5602-5606.
(15) Povsic, T. J.; Strobel, S. A.; Dervan, P. B. J. Am. Chem. Soc.
1992, 114, 5934-5941.
(16) Povsic, T. J.; Dervan, P. B. J. Am. Chem. Soc. 1990, 112, 9428-
9430.
(29) Puri, N.; Majumdar, A.; Cuenoud, B.; Natt, F.; Martin, P.; Boyd,
A.; Miller, P. S.; Seidman, M. M. J. Biol. Chem. 2001, 276, 28991-
28998.
(17) Webb, T. R.; Matteucci, M. D. J. Am. Chem. Soc. 1986, 108,
2764-2765.
(30) Musso, M.; Wang, J. C.; Vandyke, M. W. Nucleic Acids Res.
1996, 24, 492-493.
(18) Webb, T. R.; Matteucci, M. D. Nucleic Acids Res. 1986, 14,
7661-7674.
(31) Grigoriev, M.; Praseuth, D.; Guieysse, A. L.; Robin, P.; Thuong,
N. T.; Helene, C.; Bellan-H. A. Proc. Natl. Acad. Sci. U.S.A. 1993, 90,
3501-3505
(19) Cowart, M.; Benkovic, S. J. Biochemistry 1991, 30, 788-796.
(20) Shaw, J.-P.; Milligan, J. F.; Krawczyk, S. H.; Matteucci, M. D.
J. Am. Chem. Soc. 1991, 113, 7765-7666.
(32) Chatterjee, M.; Rokita, S. E. J. Am. Chem. Soc. 1994, 116,
1690-1697.
(21) Zewail-Foote, M.; Hurley, M. H. J. Med. Chem. 1999, 42, 2493-
2497.
(33) Chatterjee, M.; Rokita, S. E. J. Am. Chem. Soc. 1991, 113,
5116-5117.
(22) Afonina, I.; Zivarts, M.; Kutyavin, I. V.; Lukhtanov, E. A.;
Gamper, H.; Meyer, R. B. J. Am. Chem. Soc. 1997, 119, 6214-6225.
(23) Lukhtanov, E. A.; Podyminogin, M. A.; Kutyavin, I. V.; Meyer,
R. B.; Gamper, H. B. Nucleic Acids Res. 1996, 24, 683-687.
(24) Dempcy, R. O.; Kutyavin, I. V.; Mills, A. G.; Lukhtanov, E. A.;
Meyer, R. B. Nucleic Acids Res. 1999, 27, 2931-2937.
(25) Reed, M. W.; Wald, A.; Meyer, R. B. J. Am. Chem. Soc. 1998,
120, 9729-9734.
(34) Pande, P.; Shearer, J.; Yang, J.; Greenberg, G. A.; Rokita, S.
E. J. Am. Chem. Soc. 1999, 121, 6775-6779.
(35) Zeng, Q.; Rokita, S. E. J. Org. Chem. 1996, 61, 9080-9081.
(36) Li, T.; Zeng, Q.; Rokita, S. E. Bioconj. Chem. 1994, 5, 497-
500.
(37) Zhou, Q. B.; Rokita, S. E. Proc. Natl. Acad. Sci. U.S.A. 2003,
100, 15452-15457.
(38) Mao, H.; Schnetz-Boutaud, N. C.; Weisenseel, J. P.; Marnett,
L. J.; Stone, M. P. Proc. Natl, Acad. Sci. U.S.A. 1999, 96, 6615-6620.
(26) Lukhtanov, E. A.; Mills, A. G.; Kutyavin, I. V.; Gorn, V. V.; Reed,
M. W.; Meyer, R. B. Nucleic Acids Res. 1997, 25, 5077-5084
J. Org. Chem, Vol. 70, No. 1, 2005 15

Kawasaki et al.
SCHEME 2. The Synthesis of the Oligodeoxynucleotides Incorporating the Cross-Linking Agenta
a Reagents and conditions: (a) MeSNa, CH3CN, (b) PhOCH2COCl, 1-HBT, pyridine, CH3CN, (c) TBAF, THF, (d) (1) DMTrCl, pyridine,
(2) iPr2NP(Cl)OC2H4CN, iPr2NEt, CH2Cl2, (e) (1) automated DNA synthesizer, (2) 0.1 N NaOH, (3) neutralized AcOH, (4) 10% AcOH, (f)
(1) 3.0 equiv of MMPP, pH 10, (2) 470 mM NaOH, (g) X-PhSH, (h) 3.0 equiv of MMPP, pH 10.
FIGURE 2. Gel electrophresis of the cross-linking reaction. The reaction was performed with 7 µM ODNs (11 or 12) and 3 µM
target ODN (13) including a labeled one with 32P at the 5′ end as a tracer in 0.1 M NaCl, 50 mM MES, pH 5.0, 30 °C.
of the 4-amino group of cytidine to the vinyl group of 1
was expected in the complex between the protonated form
of 1 and cytidine to work effectively for covalent bond
formation.39
The structure of this alkylating agent is unique in that
both the reactive site and the recognition site are built
in a single molecule. Furthermore, the reactivity of the
vinyl group could be adjusted by substitution on the vinyl
group with an electron-donating group for slower rate40
or an electron-withdrawing group for faster rate under
neutral conditions.41 In a new application, site-directed
mutagenesis has been achieved by the use of the triplex-
forming oligonucleotides incorporating the 2-amino-6-
(39) Nagatsugi, F.; Uemura, K.; Nakashima, S.; Maeda, M.; Sasaki,
S. Tetrahedron Lett. 1995, 36, 421-424.
(40) Nagatsugi, F.; Uemura, K.; Nakashima, S.; Maeda, M.; Sasaki,
S. Tetrahedron 1997, 53, 3035-3044.
(41) Nagatsugi, F.; Tokuda, N.; Maeda, M.; Sasaki, S. Bioorg. Med.
Chem. Lett. 2001, 11, 2577-2579.
16 J. Org. Chem., Vol. 70, No. 1, 2005

Cytidine-Selective Activation for Cross-Linking
TABLE 1. Substituent Effect on the Cross-Linking
Reactiona
yield (%)
compd
X
11 (sulfide)
12 (sulfoxide)b
a
b
c
d
e
f
g
h
i
4-NO2
9
14
25
30
30
37
44
28
37
42
62
18
66
56
82
81
78
83
-
4-Br
4-H
4-OMe
4-OH
4-Me
4-NH2
2,4-Me2
3,4-(MeO)2
2,4,6-Me3
2-COOH
4-COOH
-
-
-
73
-
j
k
m
a Yield was determined by gel electrophoresis at 24 h. Reaction
was performed as described in the footnote of Figure 2. b The
reaction was not performed in the column marked with “-”.
FIGURE 3. Analysis of the reaction of 10 (A) and 11k (B)
with 1 mM Glutathione. A mixture of 10 µM ODN and 1 mM
glutathione was kept at 37 °C in the buffer containing 0.1 M
NaCl,10 mM KH2PO4, pH 5.0. HPLC conditions: ODS column,
1.0 mL/min; solvent A ) 0.1 M TEAA; solvent B ) CH3CN,
linear gradient from 10% to 30% over 20 min, 30% to 100%
over 30 min, monitored at 254 nm.
hybridization with a complementary strand to produce
efficient and selective cross-linking to cytidine (Figure
1).43 An interesting finding from a practical viewpoint
was that the phenylsulfide precursor (4) produced cross-
linked products despite its high stability toward strong
nucleophiles such as amines and thiols. The ODN having
the phenylsulfide precursor (4) are currently under
evaluation for antisense effects in biological studies. In
this study, we investigated the substituent effects of the
phenylsulfide group on the cross-linking reaction, and
determined the 2-amino-6-[2-(2-carboxyphenylthio)ethyl]-
purine precursor (11k) as an efficient cross-linking agent
vinylpurine derivative.42 While studying new cross-
linking agents, we proposed a new strategy of inducible
reactivity, in which a reactive species is generated in situ
from its stable precursors in close proximity with a target
base in the duplex. The validity of this concept has been
successfully demonstrated in the previous study by the
fact that the phenylsulfoxide (3) is activated in situ by
FIGURE 4. HPLC analysis of the cross-linking reaction with 10 (A), 11k (B), and 12c (C) and comparison of the reaction yield
(D). The peak labeled with 14 corresponds to the cross-linked product. A 0.3 mM sample of each oligomer was used in 300 µL of
H2O buffer containing 0.1 M NaCl and 50 mM MES at pH 5.0. HPLC conditions: ODS column, 1.0 mL/min; solvent A ) 0.1 M
TEAA; solvent B ) CH3CN, linear gradient from 0% to 30% over 20 min, 30% to 100% over 30 min, monitored at 254 nm. The
peak marked with an asterisk (*) in B indicates 2-carboxyphenyl disulfide.
J. Org. Chem, Vol. 70, No. 1, 2005 17

Kawasaki et al.
with inducible reactivity. Here we wish to describe a
detailed study on the mechanism of the inducible reactiv-
ity of the phenylsulfide (4) and phenylsulfoxide precur-
sors (3) to prove that they produce the 2-amino-6-
vinylpurine nucleoside analogue (1) as the common
reactive species, and that the nature of their inducible
reactivity is selective acceleration of their elimination to
the vinyl structure in the duplex having cytidine at the
target site.
Results
Synthesis of ODNs Incorporating 2-Amino-6-
vinylpurine Derivatives. We have previously used the
phenylsulfide (-SPh) derivative of the 2-amino-6-
vinylpurine nucleoside (1) for incorporation into the
ODNs.43 However, it turned out that the phenylsulfide
group was not suitable for the synthesis of some DNA
sequences such as G-rich ones. Therefore, we utilized the
methylsulfide (-SMe) group as a more stable protecting
group of the vinyl group of 1 in this study. The synthesis
of the ODN incorporating the 2-amino-6-vinylpurine
derivatives (9-12) is summarized in Scheme 2. The 2′-
deoxynucleoside derivative of 2-amino-6-vinylpurine (1)
was synthesized from 2′-deoxyguanosine as described
previously.40 Treatment of 1 with sodium methanesulfide
(MeSNa) in CH3CN gave a 6-(2-ethylthiomethyl)purine
derivative (5), followed by protection of the 2-amino group
with a phenoxyacetyl group (-COCH2OPh) and subse-
quent deprotection of the TBDMS group with n-tetra-
butylammoniumfluoride (TBAF) to afford the diol product
(7). The amidite precursor (8) was synthesized by a
standard procedure with 2-cyanoethyl-N,N-diisopropyl-
chlorophosphoramidite and applied to an automated
DNA synthesizer. Synthesized ODN was cleaved from the
solid support with 0.1 M NaOH solution and purified by
HPLC. The terminal 5′-dimethoxytrityl group was depro-
tected with 10% acetic acid to give ODN (9), which was
further purified by HPLC. The subsequent treatment
with MMPP and aqueous NaOH solution produced 10
having a vinyl motif. To search for the proper phenyl-
sulfide structure for in situ activation, we synthesized a
series of substituted phenylsulfide derivatives (11) by the
addition of the corresponding thiophenol derivatives to
10. Some of the phenylsulfide compounds were oxidized
with MMPP to give 12. Structures of the substituent
groups are shown in Table 1.
FIGURE 5. 2D-DQF COSY spectra of the dulexes: (A)
reaction mixture at 0 h and (B) at 18 h; (C) natural duplex 15
with 13; (D) isolated 14. A 1.3 mM sample of each oligomer
was reacted in 300 µL of D2O buffer containing 0.1 M NaCl
and 5 mM KH2PO4 at pH 5.0. Spectrum D was taken at pH
7.0. The 2D-DQF COSY spectra were recorded at 600 MHz at
30 °C with sweep widths of 6500 Hz in both dimension. The
residual HDO peak was suppressed by presaturation. The
DQF-COSY spectrum consisted of 2048 datapoints in t2 and
256 increments in t1.
yields at 24 h are summarized in Table 1. Regardless of
the substituent X of the phenyl ring, the sulfoxide
precursors (12) produced the efficient cross-linking reac-
tion. On the other hand, substituent effects were clearly
observed in the reaction with the sulfide precursors (11).
Introduction of the electron-withdrawing group on the
phenyl ring decreased the yield (11a and 11b vs 11c),
whereas the yield was increased with the electron-
donating group (11d-j vs 11c). Interestingly, 11k (X )
2-COOH) exhibited the highest yield (62% yield). High
selectivity of 11k to cytidine was clearly observed (Figure
2, lanes 15-18). It is also apparently shown that the
ortho position of carboxylic acid of 11k is important for
the efficient reaction, because para-substituted sulfide
derivative (11m) gave only low yield. Under the condi-
tions where 10 produced the conjugate adduct with
glutathione in excellent yield, 11k exhibited high stability
(Figure 3B). As 11k gradually produces 10 under acidic
conditions as explained later (Figure 4D), the results of
Figure 3B imply that such transformation is inhibited
by glutathione.
Cross-Linking Study with Substituted Phenyl-
sulfide Derivatives. The cross-linking reaction was
investigated by using the reactive ODNs (10, 11, and 12)
and the complementary strand (13). Reactions were
performed for 24 h at 30 °C with 7.0 µM of the reactive
ODNs and 3.0 µM of the target strand (13) containing
the 5′-32P-labeled target strand as a tracer at pH 5.0, and
were analyzed by gel electrophoresis with denatured
polyacrylamide gel, and some representative examples
are shown in Figure 2.
The cross-linked product was identified as the less
mobile bands, and the cross-linking yields were obtained
by quantification of the radioactivity of each band; the
(42) Nagatsugi, F.; Sasaki, S.; Miller, P. S.; Seidman, M. M. Nucleic
Acids Res. 2003, 31, e31.
(43) Nagatsugi, F.; Kawasaki, T.; Usui, D.; Maeda, M.; Sasaki, S.
J. Am. Chem. Soc. 1999, 121, 6753-6754.
Comparison of the Reaction Rates. The cross-
linking reaction was also followed by HPLC analysis, and
18 J. Org. Chem., Vol. 70, No. 1, 2005

Cytidine-Selective Activation for Cross-Linking
FIGURE 6. HPLC analysis of the cross-linking reaction between 16 and 15a (A and B), and the adducts from the enzymatic
hydrolysate (C and D). For parts A and B the cross-linking reaction conditions are the same as the footnote to Figure 5, and the
HPLC conditions are the same as the footnote to Figure 3. The peak 17 was isolated and digested with snake venom
phosphodiesterase I (0.6 u/µL) and bacterial alkaline phosphatase (0.4 u/µL) in 1 µL of buffer (50 mM Tris-HCl, 5 mM MgCl2).
HPLC condtions for parts C and D: ODS column, 1 mL/min; solvent A ) 50 mM ammonium formate, solvent B ) CH3CN, linear
gradient from 5% to 30% over 20 min, 30% to 100% over 30 min, monitored at 254 nm. (E) NMR spectra of the adducts obtained
from the enzymatic hydrolyzate.
the chromatogram changes are summarized in Figure 4.
The reaction was performed at 30 °C in the buffer
containing 0.1 M NaCl and 50 mM MES at pH 5.0 by
using 0.3 mM each ODNs (10, 11k, and 12c) and the
target ODN (13) with cytidine at the reaction site. The
peaks corresponding to 14 in Figure 4, panels A, B, and
C, were isolated, and proven to be the cross-linked
adducts by MALDI-TOF MS measurements. These ad-
ducts showed the identical 1H NMR spectra as described
below. Three reactive ODNs exhibited efficient cross-
linking with the reaction rate in the order of 10, 12c,
and 11k (Figure 4D). We reported previously that the
reaction with the ODN (10) at micromolar concentrations
gave only moderate yield (<50%).43 We observed that
decomposition of 10 competed with the cross-linking
reaction under low concentrations. In contrast, almost
no decomposition of 10 was detected at the high concen-
trations, probably because the vinyl group of 10 is
protected in the duplex. Thus, instability of the vinyl
group of 10 may account for the lower yield at low
concentrations.
It was observed by HPLC and MALDI-TOF MS that
10 was formed in the reaction mixture of the phenylsul-
fide-ODN (11k) or the phenylsulfoxide-ODN (l2c) with
the nontarget ODN (13: Z ) T), clearly showing that 10
is the active intermediate in the cross-linking with 11k
and 12c. The single strands of 11k and 12c are quite
stable under neutral conditions, but slowly form 10
together with unknown decomposed materials under the
acidic conditions used for the cross-linking reactions.
These formation rates of 10 are slower than those
observed within the nontarget double strand (Z ) T),
showing that the reactivity of the cross-linking agents is
selectively induced in the duplex having cytidine at the
target site. Such sequence-selective activation mecha-
nisms also must be responsible for high selectivity to
cytidine (Figure 2). In the reaction with 11k, 2-carboxy-
phenyl disulfide was released (the peak marked with an
J. Org. Chem, Vol. 70, No. 1, 2005 19

Kawasaki et al.
FIGURE 7. Calculated pathway of cross-linking with the 2-amino-6-vinylpurine (1) and estimated activation energy of the
transition states by ab initio calculation at B3LYP/6-31G* level of theory. The superscript a indicates the activation energy obtained
by including solvent effects in the transition state.
asterisk in Figure 4B), suggesting that 10 was formed
directly from 11k without forming the corresponding
phenylsulfoxide intermediate.
at the 5′-end and its complementary strand (16) (Figure
6). The corresponding cross-linked adduct (17) was
completely hydrolyzed with SVPD (snake venom phos-
phodiesterase) and BAP (bacterial alkaline phosphatase)
to produce the nucleosides (Figure 6C). The peak 18 was
proven to be identical with the authentic alkylated adduct
that was prepared from the 2′-deoxynucleoside deriva-
tives of 2-amino-6-vinylpurine and cytidine by HPLC, 1H
NMR, and mass measurements (Figure 6D,E). The same
adducts were also obtained from the cross-linking reac-
tions with the other two ODNs (15b and 15c). These
results have clearly indicated that the three cross-linking
agents, 2-amino-6-vinyl-, 2-amino-6-(2-phenylsulfinyl)-
ethyl-, and 2-amino-6-(2-o-carboxyphenythio)ethylpurine
derivatives, produce the same cross-linked adducts with
the 4-amino group of the target cytidine.
Structure Determination of the Cross-Linked
Adducts. It has been shown that both the phenylsulfide
and phenylsulfoxide precursors produce the 2-amino-6-
vinylpurine analogue as the common reactive intermedi-
ate. We next investigated the adduct structure to confirm
that the cross-linking reactions take place with the
4-amino group of the target cytidine. The 1H NMR
spectrum was measured to follow the cross-linking reac-
tion between the ODNs (10, 11k, and 12c) and the target
ODN (13) with cytidine at the reaction site. Reaction was
performed at 30 °C in the D2O buffer containing 0.1 M
NaCl and 5 mM KH2PO4 at pH 5.0 by using 1.3 mM each
ODNs. Figure 5A shows selected regions of the H-H
COSY spectrum of the reaction between 10 and 13, where
all six correlations of H5-H6 of cytidines were observed,
and is similar to that of the natural duplex (15:13)
(Figure 5C). According to the progress of the cross-linking
reaction, the signal marked “a” was decreased, and a new
signal marked “b” appeared (Figure 5B). These results
have suggested that the signal marked “a” is due to
cytidine, and the signal marked “b” corresponds to the
cross-linked cytidine. Downfield shift of H5 and H6
caused by alkylation of 4-NH2 of cytidine accords with
the results of the monomer adduct between the cytidine
and the 2-amino-6-vinylpurine nucleoside derivative
(Figure 6, E). Its COSY spectra showed a new signal “c”
at pH 7.0 (Figure 5D) in addition to “b”. Since the signal
“b” increased as pH was lowered, the signals “c” and “b”
probably correspond to nonprotonated and protonated
Discussion
A Plausible Mechanism of in Situ Activation. We
designed the cross-linking reaction initially based on
semiempirical MO calculations (PM3) on N1-protonated
2-amino-6-vinylpurine (1) (pKa ) 3.7).40 However, ab
initio MO calculations have indicated that this process
associates with high activation energy of Ea ) 53.3 kcal/
mol for direct proton transfer from 20 to 21 (Figure 7).
Then, MO calculations were performed on an alternative
mechanism with the imino tautomer of cytosine (22) to
produce a transition state (23) with a lower activation
energy of 20.9 kcal/mol (Figure 7). Since the imino
tautomer is in equilibrium with the amino tautomer, the
pathway from 22 to 24 may account for the rapid
alkylation rate of the 2-amino-6-vinylpurine derivative
with the target cytidine.
1
cytidine of the adduct (14), respectively. H NMR mea-
surements of the cross-linking reaction with 11k and 12c
produced the same change, and have proven that the
adducts from the three experiments are identical (peak
14 of Figure 4, panels A, B, and C). Attempts to isolate
and determine the adduct structure by degradation of 14
were not successful, because enzymatic digestion was
inhibited near the cross-linked sites. Therefore, we
carried out the cross-linking reaction with the ODNs
(15a) that contains the 2-amino-6-vinylpurine derivative
The in situ activation mechanism was designed so as
to generate the reactive vinyl group of 2-aminopurine (1)
from its phenylsulfoxide precursor (3) (Figure 1). To check
this design concept, transformation of the phenylsulfoxide
derivative (12c) was followed by HPLC and MALDI-TOF
MS, and it has been proven that the ODN (10) was
formed under the conditions of cross-linking (Figure 4D).
It is surprising that 10 is formed from 12c in the buffer,
because the corresponding monomer phenylsulfoxide (3)
20 J. Org. Chem., Vol. 70, No. 1, 2005

Cytidine-Selective Activation for Cross-Linking
FIGURE 8. Plausible pathways of in situ activation.
is much faster than the formation of 10 in the nontarget
duplex (Figure 4D) clearly indicates that this activation
is also promoted selectively by hybridization with the
target ODN. The results that 2-COOH of 11k is more
effective than 4-COOH of 11m and that electron-donating
groups give higher yields (Table 1) suggest that proto-
nation of the sulfur atom of the phenylsulfide might
accelerate elimination to the vinyl structure. However,
this hypothesis was ruled out by MO calculations,
because the activation energy for elimination of the
phenylsulfide group was not lowered by protonation on
the sulfur atom. On the other hand, when the cross-
linking yields of 11 (Table 1) are plotted against the
HOMO energy of substituted thioanisoles as their struc-
tural components, a linear correlation is obtained (Figure
9). Figure 9 clearly suggests that an activation process
of the sulfide derivative of 11 may contain an HOMO
energy-related one. The 2-COOH-substituted derivative
(11k) produced a higher cross-linking yield than that
expected from its HOMO energy, suggesting that the
2-COOH group provides additional rate enhancement
effects, although its mechanism is uncertain. Further
study is now ongoing to reveal the activation mechanism
of the phenylsulfide precursors.
FIGURE 9. Relationship between HOMO energy of substi-
tuted thioanisole and cross-linking yield (%). Data were taken
from Table 1. The HOMO ab initio calculation was performed
on the substituted thioanisole at the 6-31G** level.
Conclusion
is very stable in the organic solvents.43 It should be noted
that complexation with cytosine effects to reduce the
activation energy of elimination of the phenylsulfoxide
from Ea ) 31.2 kcal/mol in the monomer to Ea ) 22.2
kcal/mol in the complex with cytosine (25 to 19). The
experimental facts and ab initio calculations have indi-
cated that the nature of the target-promoted activation
of the phenylsulfoxide derivative is lowering the activa-
tion energy of the elimination by complexation with the
target cytosine.
Detection of 2-carboxyphenyl disulfide in the cross-
linking reaction with 11k (X ) 2-COOH) (the peak
marked with an asterisk in Figure 4B) suggests the direct
formation of 10 (27 to 19, Figure 8) not through the
corresponding sulfoxide derivative (25). The experimental
fact that the cross-linking of 11k with the target ODN
We have established a new target-promoted cross-
linking reaction with the use of 2-amino-6-vinylpurine
derivatives. The validity of the new concept has been
successfully demonstrated by the fact that the stable
phenylsulfide precursors are activated selectively in the
target duplex to form cross-linked products with cytidine.
The new cross-linking agents are characteristic in that
no chemical or biological reagents are needed for the
activation except for the duplex formation between the
reactive ODN and the target strand. It has been con-
cluded that the nature of the inducible reactivity of the
phenylsulfoxide- and the phenylsulfide-nucleoside de-
rivative is acceleration of their elimination to the 2-amino-
6-vinylpurine through the selective processes in the
duplex with the ODN having cytidine at the target site.
J. Org. Chem, Vol. 70, No. 1, 2005 21

Kawasaki et al.
2-Phenoxyacetylamino-9-[5′-O-dimethoxytrityl-3-O-
(N,N-diisopropyl-2-cyanoethylphosphoramidyl)-2′-deoxy-
D-ribofranosyl]-6-(2-methylthioethyl)purine (8). A solu-
tion of 7 (310 mg, 0.67 mmol) in dry pyridine (3.0 mL) was
stirred for 1 h in the presence of 4 Å molecular sieves at 0 °C.
Then dimethoxytrityl chloride (343 mg, 1.0 mmol) was added
into the solution and the mixture was stirred for 2 h. The
resulting mixture was diluted with chloroform (30 mL) and
washed with saturated aqueous NaHCO3, H2O, and brine. The
organic layer was dried over anhydrous Na2SO4, filtered, and
concentrated under reduced pressure. The residue was purified
by silica gel chromatography (chloroform-MeOH 95:5, 0.5%
pyridine, v/v) to afford 2-phenoxyacetylamino-9-(5-O-dimethoxy-
trityl-2′-deoxy-D-ribofranosyl)-6-(2-methylthioethyl)purine as
a colorless oil (500 mg, 0.61 mmol, 90%): 1H NMR (CDCl3,
400 MHz) δ 8.93 (1H, br s), 8.11 (1H, s), 7.68-7.00 (14H, m),
6.79-6.73 (4H, m), 6.56 (1H, t, J ) 6.3 Hz), 4.87 (1H, dt, J )
5.9, 3.0 Hz), 4.68 (2H, br s), 4.17 (1H, m), 3.75 (6H, s), 3.48-
3.33 (4H, m), 3.06 (2H, t, J ) 6.9 Hz), 2.87-2.57 (2H, m), 2.18
(3H, s); FABMS (m/z) 761 (M)+.
To a solution of the above product (110 mg, 0.14 mmol) in
dry dichloromethane (3.0 mL) was added diisopropylethyl-
amine (0.15 mL, 0.86 mmol) at 0 °C in the presence of 4 Å
molecular sieves. After the resulting mixture was stirred for
30 min, 2-cyanoethyl N,N-diisopropylchlorophosphoramidite
(0.11 mL, 0.43 mmol) was added. The reaction mixture was
stirred for an additional 30 min at the same temperature. The
resulting mixture was diluted with ethyl acetate (30 mL) and
washed with saturated aqueous NH4Cl, H2O, and brine. The
organic layer was dried over anhydrous Na2SO4, filtered, and
concentrated under reduced pressure. The residue was purified
by silica gel chromatography (hexanes-ethyl acetate 1:1, v/v)
to afford 8 as a colorless oil (123 mg, 0.13 mmol, 93%). 1H NMR
(CDCl3, 400 MHz) δ 8.81 (1H, br s), 8.14 (1H, s), 7.39-7.01
(14H, m), 6.76 (4H, dd, J ) 6.8, 2.0 Hz), 6.47 (1H, t, J ) 6.3
Hz), 4.78 (3H, m), 4.29 (1H, m), 3.88-3.78 (2H, m), 3.76 (6H,
s), 3.74-3.56 (4H, m), 3.47-3.32 (2H, m), 3.37 (2H, t, J ) 6.3
Hz), 2.88-2.60 (2H, m), 2.64 (2H, t, J ) 6.3 Hz), 2.18 (3H, s),
1.19 (6H, d, J ) 6.9 Hz), 1.17 (6H, d, J ) 6.9 Hz); 31P NMR
(CDCl3, 162 MHz) δ 148.8; FABMS (m/z) 962 (M)+.
Although detailed mechanisms of the activation process
are still uncertain, stability, selectivity, and reactivity
of the new cross-linking agents will be beneficial to
antisense and antigene methods, and their biological
applications are now in progress.
Experimental Section
2-Amino-9-(3′,5′-di-O-tert-butyldimethylsilyl-2′-deoxy-
D-ribofranosyl)-6-(2-methylthioethyl)purine (5). To a so-
lution of 2-amino-9-(3,5-di-O-tert-butyldimethylsilyl- 2′-deoxy-
D-ribofranosyl)-6-vinylpurine (1) (1.1 g, 2.2 mmol) in CH3CN
(10 mL) was added methylmercaptane sodium salt (15% in
water) (1.2 mL, 2.6 mmol) at room temperature. After being
stirred for 20 min, the reaction mixture was diluted with ethyl
acetate (50 mL), then washed with H2O and brine. The organic
layer was dried over anhydrous Na2SO4, filtered, and concen-
trated under reduced pressure. The residue was purified by
silica gel chromatography (chloroform-methanol 97:3, v/v) to
afford 5 as a pale yellow oil (1.1 g, 2.03 mmol, 93%): IR (neat)
1
3400, 1600 cm-1; H NMR (CDCl3, 400 MHz) δ 8.02 (1H, s),
6.32 (1H, t, J ) 6.3 Hz), 5.21 (2H, br s), 4.58 (1H, dt, J ) 5.6,
3.0 Hz), 3.99 (1H, dd, J ) 6.9, 3.3 Hz), 3.77 (2H, m), 3.33 (2H,
t, J ) 6.6 Hz), 3.03 (2H, t, J ) 6.6 Hz), 2.65-2.32 (2H, m),
2.18 (3H, s), 0.90 (18H, s), 0.10 (12H, s); 13C NMR (125 MHz,
CDCl3) δ -5.5, -5.4, -4.8, -4.7, 15.4, 18.0, 18.4, 25.7, 25.9,
32.0, 32.9, 40.6, 62.8, 72.1, 83.5, 87.6, 127.3, 139.3, 152.4, 159.5,
161.3; FABMS (m/z) 554 (M + 1)+; HRFABMS calcd for
C25H48N5O3Si2S 554.3016, found 554.3066.
2-Phenoxyacetylamino-9-(3′,5′-di-O-tert-butyldimeth-
ylsilyl-2′-deoxy-D-ribofranosyl)-6-(2-methylthioethyl)pu-
rine (6). To a solution of 5 (800 mg, 1.45 mmol) in dry CH3CN
(3.0 mL) was added 1-hydroxybenzotriazole (490 mg, 4.3 mmol)
in dry pyridine solution (3.0 mL) in the presence of 4 Å
molecular sieves at room temperature. After the resulting
mixture was stirred for 1 h, freshly distilled phenoxyacetyl
chloride (0.62 mL, 4.3 mmol) was added slowly to the reaction
mixture. The reaction mixture was stirred for an additional 3
h, diluted with ethyl acetate (50 mL), and washed with H2O
and brine. The organic layer was dried over anhydrous Na2-
SO4, filtered, and concentrated under reduced pressure. The
residue was purified by silica gel chromatography (hexanes-
ethyl acetate 7:3, v/v) to give 6 as a pale yellow oil (747 mg,
Oligonucleotide Synthesis. All oligonucleotides were
synthesized at a 1 µmol scale on Cyclon Plus DNA-synthesizer
(Milligen/Biosearch) with standard â-cyanoethyl phosphora-
midite chemistry. The 5′-terminal dimethoxytrityl-bearing
ODNs were removed from the solid support by treatment with
0.1 M NaOH (2.0 mL) and the solution was neutralized with
CH3COOH instantly. The crude product was purified by
reverse-phase HPLC with C-18 column (nacalai tesque: COS-
MOSIL 5C18-AR-II, 10 × 250 mm) by a linear gradient of 10-
40%/20 min of acetonitrile in 0.1 M TEA buffer at a flow rate
of 4 mL/min. The dimethoxytrityl group of the purified ODN
was cleaved with 10% AcOH and the mixture was additionally
purified by HPLC with the same elution method. MALDI-TOF
MS (m/z) 9: calcd 4812.83, found 4809.78
1
1.09 mmol, 75%): IR (neat) 3400, 1720, 1600 cm-1; H NMR
(CDCl3, 400 MHz) δ 8.90 (1H, br s), 8.25 (1H, s), 7.38-7.31
(2H, m), 7.08-7.01 (3H, m), 6.47 (1H, t, J ) 6.6 Hz), 4.79 (2H,
br s), 4.64 (1H, dt, J ) 5.6, 3.0 Hz), 4.00 (1H, dd, J ) 7.3, 3.3
Hz), 3.88 (1H, dd, J ) 11.2, 3.3 Hz), 3.77 (1H, dd, J ) 11.2,
3.3 Hz), 3.44 (2H, t, J ) 7.3 Hz), 3.07 (2H, t, J ) 7.3 Hz), 2.73-
2.44 (2H, m), 2.18 (3H, s), 0.91 (18H, s), 0.10 (12H, s); 13C NMR
(125 MHz, CDCl3) δ -5.5, -5.4, -4.8, -4.7, 15.4, 18.0, 18.4,
25.7, 25.9, 31.8, 32.7, 40.9, 62.8, 68.0, 72.0, 84.2, 88.0, 114.9,
122.2, 129.7, 130.5, 142.4, 151.3, 151.5, 157.1, 161.5, 166.3;
FABMS (m/z) 688 (M + 1)+.
Preparation of the ODN 10. To a solution of ODN 9 (0.97
µmol, 1 mL) was added a solution of magnesium monoper-
phthalate (MMPP) (2.9 µmol) in carbonate buffer (50 µL)
adjusted to pH 10 at room temperature. After 30 min, NaOH
(4 mol/L, 150 µL) was added, and the mixture was left for an
additional 30 min. The resulting mixture was dialyzed with a
buffer (0.1 M NaCl, 5 mM potassium phosphate, pH 7.0) to
give ODN (9) (0.75 µmol). MALDI-TOF MS (m/z) 10: calcd
4764.83, found 4762.97.
Preparation of the ODN 11a-l. To a solution of ODN 10
(0.41 µmol, 1.0 mL) was added a solution of substituted
thiophenol in CH3CN (2.1 µmol, 20 µL) at room temperature.
After 30 min, the resulting mixture was purified by HPLC to
give ODN 11a-l. The yields and MALDI-TOF MS (m/z) date
of 11a-l are summarized in Table 2.
2-Phenoxyacetylamino-9-(2′-deoxy-D-ribofranosyl)-6-
(2-methylthioethyl)purine (7). To a solution of 6 (600 mg,
0.87 mmol) in dry THF solution (1.0 mL) was added n-Bu4NF
in THF solution (1.0 M solution, 2.62 mmol, 2.62 mL) at 0 °C.
After 1 h, the reaction mixture was concentrated under
reduced pressure. The residue was purified by silica gel
chromatography (CHCl3-MeOH 95:5, v/v) to give the diol 7
as a colorless crystal (392 mg, 0.85 mmol, 98%): mp 158 °C;
IR (neat) 3400, 1720, 1600 cm-1; 1H NMR (d-DMSO, 400 MHz)
δ 9.10 (1H, br s), 8.03 (1H, s), 7.38-7.31 (2H, m), 7.08-7.01
(3H, m), 6.31 (1H, t, J ) 6.3 Hz), 5.10 (1H, m), 4.70 (2H, s),
4.15 (1H, dd, J ) 5.6, 2.6 Hz), 3.90 (1H, m), 3.43 (2H, t, J )
7.3 Hz), 3.07 (2H, t, J ) 7.3 Hz), 3.03 (1H, m), 2.41 (1H, m),
2.18 (3H, s); 13C NMR (125 MHz, d-DMSO) δ 14.4, 31.0, 31.0,
32.1, 61.6, 67.2, 70.6, 83.2, 87.9, 90.1, 99.4, 114.4, 120.8, 129.3,
143.2, 151.1, 152.0, 157.9, 160.4, 167.3; FABMS (m/z) 460 (M
+ 1)+. Anal. Calcd for C21H25N5O5: C, 54.89; H, 5.48; N, 15.24.
Found: C, 54.66; H, 5.49; N, 15.09.
Preparation of the ODN 12a-g,k. To a solution of ODN
11 (0.97 µmol, 1 mL) was added a solution of magnesium
monoperphthalate (MMPP) (2.9 µmol) in carbonate buffer (50
22 J. Org. Chem., Vol. 70, No. 1, 2005

Cytidine-Selective Activation for Cross-Linking
TABLE 2.
Isolation of the Cross-Linking Adducts. The cross-
linking reaction was performed with the reactive ODN (0.3
mM) and the complementary ODN (0.3 mM) in a buffer (100
mM NaCl, 50 mM MES buffer, pH 5.0) at 30 °C. The total
volume of the reaction mixture was 1.5 mL. After 24 h, the
cross-linking adduct was purified by HPLC with an ODS
column (nacalai tesque: COSMOSIL 5C18-AR-II, 10 × 250
mm), using a linear gradient of 5-30%/20 min acetonitrile in
50 mM ammonium acetate.
NMR Measurements. NMR samples were lyophilized from
99.96% D2O and dissolved in 300 µL of the buffer containing
0.1 M NaCl, 5 mM KH2PO4 at pH 7.0 in D2O. The total
concentration of the duplex was 1.3 mM. 1H NMR spectra were
recorded on a Varian Inova 600 spectrometer at 600 MHz. The
1D spectra were 65536 complex points with 64 scan averaged.
The 2D-DQF COSY spectra were recorded at 30 °C with sweep
widths of 6492.5 Hz in both dimensions, and consisted of 2048
data points in t2 and 512 increments in t1. The residual HDO
peak was suppressed by presaturation.
Molecular Orbital Calculations. Ab initio MO calcula-
tions were performed by using the GAUSSIAN94 program to
optimize stable and transition state (TS) structure at the
B3LYP/6-31G* level of theory for Figure 7. Vibrational fre-
quency calculations showed all the transition states to have
only one imaginary frequency. HOMO energies of the substi-
tuted thioanisoles were calculated at the RHF/6-31G* level of
theory.
MALD-TOF MS
ODNs
yield (%)
cacld
found
11a
11b
11c
11d
11e
11f
11g
11h
11i
11j
11k
11l
12a
12b
12c
12d
12e
12f
38
62
60
53
63
57
57
55
47
42
55
50
4919.8
4952.7
4874.8
4904.8
4890.8
4888.8
4889.0
4902.9
4934.9
4916.9
4918.8
4918.8
4935.8
4968.7
4890.8
4920.8
4906.8
4904.8
4905.0
4934.8
4919.1
4951.9
4874.6
4903.8
4891.8
4888.8
4886.7
4900.3
4934.9
4919.8
4919.2
4915.7
4933.4
4967.3
4888.9
4917.7
4907.8
4906.3
4902.5
4936.7
12g
12k
µL) adjusted to pH 10 at room temperature. After 30 min, the
reaction mixture was dialyzed with a buffer (0.1 M NaCl, 5
mM potassium phosphate, pH 7.0) to give ODN (12) (0.75
µmol), and was used without further purification. The MALDI-
TOF MS (m/z) data of 12a-g,k are summarized in Table 2.
Electrophoresis Assay of the Cross-Linking Reaction.
The complementary ODN was 5′-end-labeled by treatment
with T4 polynucleotide kinase and [γ-32P] ATP under standard
conditions. The reaction was performed with 7 µM of the
sulfide ODN, 3 µM of the complementary unlabeled ODN, and
labeled ODN (1 × 104 cpm) as a tracer in a buffer of 100 mM
NaCl and 50 mM MES buffer at pH 5.0. The reaction mixture
was incubated at 37 °C for 24 h. The reaction was stopped by
the addition of the loading dye (95% formamide, 20 mM EDTA,
0.05% xylene cyanol, and 0.05% bromophenol blue) and
heating at 90 °C for 10 min. The cross-linked products were
analyzed by a denaturing 20% polyacrylamide gel electro-
phoresis containing urea (7 M) with TBE buffer at 300 V for
1.5 h. The labeled bands were visualized by autoradiography
and quantified with use of a BAS 2500 Phosphor Imager.
Acknowledgment. This work has been supported
by a Grant-in-Aid for Scientific Research (B) and (C)
from Japan Society for the Promotion of Science (JSPS)
and Priority Area (A)(2) (Molecular Synchronization)
from the Ministry of Education, Culture, Sports, Science
and Technology (MEXT), CREST from the Japan Sci-
ence and Technology Agency (JST), Kyushu University
P&P “Green Chemistry” Program, and the Research
Consortium from Kyushu Bureau of Economy, Trade
and Industry.
Supporting Information Available: Experimental de-
tails and 1H, 13C, and 31P NMR spectra of nucleoside analogues
in Scheme 2. This material is available free of charge via the
Internet at http://pubs.acs.org.
JO048298P
J. Org. Chem, Vol. 70, No. 1, 2005 23