Anthraquinone-sensitized photooxidation of 5-methylcytosine in DNA leading to piperidine-induced efficient strand cleavage

Article abstract of DOI:10.1002/chem.201001884

One-electron photooxidations of 5-methyl-2′-deoxycytidine (d ^mC) and 5-trideuteriomethyl-2′-deoxycytidine ([D ₃]d^mC) by sensitization with anthraquinone (AQ) derivatives were investigated. Photoirradiation of an aerated aq

Full text of DOI:10.1002/chem.201001884

FULL PAPER
DOI: 10.1002/chem.201001884
Anthraquinone-Sensitized Photooxidation of 5-Methylcytosine in DNA
Leading to Piperidine-Induced Efficient Strand Cleavage
Hisatsugu Yamada,*[a] Yuya Kitauchi,[b] Kazuhito Tanabe,[b] Takeo Ito,[b] and
Sei-ichi Nishimoto*[b]
Abstract: One-electron photooxida-
tions of 5-methyl-2’-deoxycytidine
transfer of the C5-hydrogen atom of a
dmC-tetroxide intermediate to produce
dfC and dhmC. In the case of a 5-meth-
treatment. In accordance with the sup-
pression of the strand cleavage at 5-tri-
deuterio-methylcytosine observed in a
similar AQ photosensitization, it is sug-
gested that deprotonation at the C5-
methyl group of an intermediate mC
radical cation may occur as a key ele-
mentary reaction in the photooxidative
strand cleavage at the mC site. Incorpo-
ration of an AQ sensitizer into the in-
terior of a strand of the duplex en-
hanced the one-electron photooxida-
tion of mC, presumably because of an
increased intersystem crossing efficien-
cy that may lead to efficient piperi-
dine-induced strand cleavage at an mC
site in a DNA duplex.
(dmC) and 5-trideuteriomethyl-2’-de-
oxycytidine ([D3]dmC) by sensitization
with anthraquinone (AQ) derivatives
were investigated. Photoirradiation of
an aerated aqueous solution containing
dmC and anthraquinone 2-sulfonate
(AQS) afforded 5-formyl-2’-deoxycyti-
dine (dfC) and 5-hydroxymethyl-2’-de-
oxycytidine (dhmC) in good yield
through an initial one-electron oxida-
tion process. The deuterium isotope
effect on the AQS-sensitized photooxi-
dation of dmC suggests that the rate-de-
termining step in the photosensitized
oxidation of dmC involves internal
ylcytosine
(mC)-containing
duplex
DNA with an AQ chromophore that is
incorporated into the backbone of the
DNA strand so as to be immobilized at
a specific position, mC underwent effi-
cient direct one-electron oxidation by
the photoexcited AQ, which resulted in
an exclusive DNA strand cleavage at
the target mC site upon hot piperidine
Keywords: isotope effects · DNA
cleavage
· DNA methylation ·
AHCTUNGTRENNUNG
Introduction
cific cytosine residues in genomes. Among the various chem-
ical methods for identifying and analyzing cytosine methyla-
tion explored so far,[3–7] the selective chemical modification
of methylated cytosines is recognized as a useful reaction
that can be applied to identify correctly the methylation
site.[6] In particular, with consideration of the evidence that
5-methylcytosine (mC) has a slightly lower oxidation poten-
tial than normal cytosine (C) and thymine, much effort has
recently been made to get a discriminating method by which
oxidation of a specific target mC base in DNA could be se-
lectively induced for sequence-selective DNA methylation
analysis.[6–7]
Photosensitized oxidation of DNA has been studied ex-
tensively in relation to positive-charge (hole) transfer
through a DNA duplex.[8–11] In general, photosensitized one-
electron oxidation of a DNA base forms the primary inter-
mediate of a base radical cation, which can migrate through
the p stacking between paired DNA bases and thereby pro-
duces strand cleavage at the lower-oxidation-potential sites,
such as consecutive guanine (G) sites.[10] More recent studies
have shown that one-electron oxidation of DNA without G
bases also induces selective strand cleavage at thymine sites
through transfer of a photosensitizer-injected hole through
the DNA duplex.[11] These studies have been performed
with a broad variety of organic photosensitizers that are co-
valently linked to oligonucleotides.[8–11] Our group has also
reported that one-electron oxidation and site-selective
Cytosine methylation, recognized to be a physiological
modification of DNA in mammalian genomes, is believed to
play a key function in the epigenetic regulation of genetic
information and has been implicated as an important factor
in the control of many cellular processes.[1,2] Consistent with
such a crucial function, various types of human diseases are
associated with misplaced cytosine methylation.[1,2] The
identification and analysis of methylated cytosine sites in
human genomes has become important for diagnosis of
these diseases, and there is an emerging necessity to develop
an effective method to assess the methylation status of spe-
[a] Dr. H. Yamada
Advanced Biomedical Engineering Research Unit
Kyoto University, Katsura Campus
Nishikyo-ku, Kyoto 615-8510 (Japan)
Fax : (+81)75-383-2805
E-mail: hisatsugu@t03.mbox.media.kyoto-u.ac.jp
[b] Y. Kitauchi, Dr. K. Tanabe, Dr. T. Ito, Prof. Dr. S.-I. Nishimoto
Department of Energy and Hydrocarbon Chemistry
Graduate School of Engineering
Kyoto University, Katsura Campus
Nishikyo-ku, Kyoto 615-8510 (Japan)
Fax : (+81)75-383-2501
E-mail: nishimot@scl.kyoto-u.ac.jp
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201001884.
Chem. Eur. J. 2011, 17, 2225 – 2235
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2225

strand cleavage at an mC base in DNA is caused upon pho-
toirradiation of 2-methyl-1,4-naphthoquinone (NQ)-tethered
oligodeoxynucleotides (ODNs).[12] In contrast to oxidative
strand cleavage occurring at the target mC site with an oppo-
sitely located NQ photosensitizer, the corresponding normal
C site, arranged instead of mC in the duplex, underwent no
such NQ-photosensitized strand cleavage. In addition, com-
petitive strand cleavage was virtually suppressed at a se-
quence of adjacent G bases, which have a lower oxidation
potential than mC. This striking photooxidative reactivity has
been applied to the fluorometric detection of a methylation
site in a partial sequence of naturally occurring genomic
DNA, by using a combination of mC-selective strand cleav-
age and an invasive cleavage reaction.[13] In this case, the
photosensitized reaction that directly oxidizes mC, but not C,
has potential for the convenient analysis of cytosine methyl-
ation at a specific site in a given DNA. To establish an opti-
mized protocol of photosensitized oxidation and cleavage at
the target mC site in DNA, which can discriminate between
C and mC more efficiently and selectively, the mechanism of
mC photooxidation and the structure–activity relationship of
photosensitizers in strand cleavage at the mC site in DNA
are still the subjects for further studies.
Anthraquinone (AQ) derivatives are among the most
potent oxidizing photosensitizers; they undergo efficient in-
tersystem crossing to form the long-lived triplet excited
state (3AQ*), which is known to induce hole injection fol-
lowed by positive-charge transfer through DNA.[11,14] These
photochemical properties prompted us to characterize the
AQ-photosensitized oxidation of monomeric 5-methyl-2’-de-
oxycytidine (dmC) and to compare it with the photooxidative
strand cleavage reaction at mC in DNA. Upon photoirradia-
tion of an aqueous solution in the presence of anthraqui-
none 2-sulfonate (AQS) and dmC, efficient formation of 5-
formyl-2’-deoxycytidine (dfC) and 5-hydroxymethyl-2’-de-
oxycytidine (dhmC) was observed. A similar AQS-sensitized
photooxidation of 5-trideuteriomethyl-2’-deoxycytidine in-
stead of dmC suggested that the rate-determining step in the
AQ-photosensitized oxidation of dmC involves formation of
a tetroxide intermediate and subsequent internal hydrogen
transfer at the C5 position to produce the final oxidation
products, dfC and dhmC. We also investigated AQ-photosen-
sitized oxidative strand cleavage at an mC site in DNA with
the AQ chromophore incorporated into the DNA backbone
so as to be immobilized at a specific position. Thus, we ob-
tained evidence that the AQ chromophore in the triplet ex-
cited state formed due to efficient intersystem crossing en-
hances oxidative DNA strand cleavage at the mC site. These
findings may provide a new guide for the design of highly
sensitive photochemical methods to identify and analyze
methylated modification of cytosine in DNA.
of dmC sensitized by AQS. Aerobic solutions of dmC
(200 mm) and AQS (200 mm) in 2 mm sodium cacodylate
buffer containing 20 mm NaCl at pH 7.0 were photoirradiat-
ed with 312 nm UV light. Figure 1 shows a representative
Figure 1. HPLC profiles of the photooxidation of dmC (200 mm) sensitized
by AQS (200 mm) in 10 mm sodium cacodylate buffer containing 100 mm
NaCl (pH 7.0) upon 312 nm irradiation at 08C. The eluents were moni-
tored with UV absorbance at 260 nm.
time course of HPLC profiles observed in the AQS-photo-
sensitized oxidation of dmC. The column eluents were moni-
tored by UV absorbance analysis at 260 nm. A major photo-
oxidation product, dfC, was produced, along with the degra-
dation of dmC, as identified from the characteristic HPLC
peaks by reference to the respective authentic samples. In
addition to the formation of dfC, a minor product peak as-
signed to dhmC was also observed. The yields of dfC and
dhmC were quantified to be 22% and 14%, respectively,
after 1 h of photoirradiation. We further characterized the
photooxidation products dhmC and dfC by using mass spec-
troscopy with reference to the authentic samples. The LC–
ESI mass analyses of the characteristic LC peaks indicated
the formation of dhm
C
([MÀH]À =256.1) and dfC
([MÀH]À =254.1), along with several minor products of the
dmC degradation (Figure S1 in the Supporting Information).
The formation of the final oxidation products, such as dfC
and dhmC, has been characterized previously by detailed
product analysis of the photosensitized oxidation of dmC
with an NQ chromophore.[15–17] The one-electron reduction
potential of AQS in the triplet excited state (3AQS*; Erdn
=
2.30 V vs. a normal hydrogen electrode (NHE))[18] has a
much higher positive value relative to that of the mC base
(Eox =1.73 V vs. NHE).[19] It is therefore most likely that the
3
triplet excited-state chromophore AQS* may one-electron
oxidize dmC to produce the final products of dfC and dhmC,
as in the case of the NQ-photosensitized oxidation of dmC.
To obtain mechanistic insight into the AQS-photosensi-
tized one-electron oxidation of dmC, we also performed
laser flash photolysis of dmC in aqueous acetonitrile solution
(acetonitrile/water, 9:1) containing AQS sensitizer. In ac-
cordance with the previous reports,[20] laser flash excitation
at 355 nm of sensitizer AQS in deoxygenated aqueous aceto-
nitrile solution without the substrate dmC showed the build-
up and subsequent exponential decay of a characteristic
transient absorption band at around 380 nm that is assigned
Results and Discussion
Photooxidation of dmC derivatives sensitized by an AQ
chromophore: We initially investigated the photooxidation
2226
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2225 – 2235

Piperidine-Induced DNA Strand Cleavage at 5-Methylcytosine
FULL PAPER
m
into a radical cation intermediate (d C +), along with forma-
tion of the AQS species.
C
CÀ
An attempt was also made to identify the transient ab-
sorption of the possible counterpart radical cation inter-
m
mediate d C +, which may be generated by one-electron
C
transfer from dmC to 3AQS*. Unfortunately, however, the
absorption of the d C species could not be detected di-
rectly in the laser flash photolysis experiment, presumably
m
+
C
m
+
C
because of the smaller extinction coefficient of the d C
and because of overlap with the intense absorptions of
3
CÀ
AQS* and AQS in the wavelength range of 300–600 nm.
In view of the evidence that the growth-time constant of the
CÀ
AQS species was well matched with the decay-time con-
3
stant of AQS* in the presence of dmC, it is reasonable to
conclude that the photoinduced one-electron transfer from
3
dmC to AQS* is a key reaction on the pathway leading to
m
+
the final oxidation products dfC and dhmC via a d C inter-
C
mediate.
A further attempt was made to evaluate the deuterium ki-
netic isotope effect at the C5-methyl group on the AQS-
photosensitized oxidation of dmC to produce dfC. The C5-
methyl hydrogen atoms of dmC, as the target site of photo-
sensitized oxidation, were fully deuterated to obtain 5-tri-
deuteriomethyl-2’-deoxycytidine ([D3]dmC), as prepared
from 3’,5’-O-bis(tert-butyldimethylsilyl)-5-(methyl-d3)-2ꢂ-de-
oxyuridine (1) as outlined in Scheme 1.[22]
An aerobic aqueous solution of [D3]dmC (200 mm) contain-
ing AQS (200 mm) was irradiated with 312 nm UV light
under the conditions described above and analyzed by
HPLC. As shown in Figure 3, the yield of 5-deuterioformyl-
2’-deoxycytidine ([D1]dfC) was suppressed in comparison
Figure 2. Transient absorption spectra of the intermediates as observed
~
^
*
0.3 ( ), 3 ( ), and 20 ms ( ) after 355 nm laser flash photolysis of AQS
(50 mm) in the a) absence or b) presence of dmC (500 mm) in acetonitrile/
water (9:1).
to the triplet excited intermedi-
3
ate AQS* (Figure 2a and Fig-
ure S2 in the Supporting Infor-
mation). Upon similar 355 nm
laser flash excitation of AQS in
the presence of dmC in deoxy-
genated aqueous acetonitrile
solution, an intense transient
absorption of 3AQS* was also
Scheme 1. Reagents and conditions: a) 1,2,4-Triazole, POCl3, Et3N, CH3CN, 72%; b) 25% ammonium deuter-
oxide, CH3CN, quant.; c) TBAF, AcOH, THF, 63%. TBAF: tetrabutylammonium fluoride; TBS: tert-butyldi-
methylsilyl; THF: tetrahydrofuran.
observed 0.3 ms after the laser
flash (Figure 2b). In contrast to
the transient spectral behavior
in the absence of dmC (Fig-
ure 2a), the decay of AQS* in the presence of dmC accom-
panied the buildup of a new transient species with an ab-
sorption maximum at 520 nm (Figure 2b). By reference to
the previous study,[21] this transient species is assigned to an
with that of normal dfC, which indicated that deuterium-iso-
tope substitution at the C5-methyl hydrogen atoms of dmC
significantly decreased the formation of dfC. The kinetic iso-
tope effects on the AQS-photosensitized oxidation of dmC
were quantified by the ratios of the initial rates, to give kH/
kD =1.3 for the dmC degradation and kH/kD =5.3 for the dfC
3
CÀ
CÀ
AQS radical anion (AQS ). The AQS species formation
gave a pseudo-first-order rate constant of 1.38ꢁ106 sÀ1, as
characterized by the buildup of transient absorption at
520 nm, a value showing good agreement with the decay-
formation. As discussed previously,[15,16,23] the d C inter-
mediate generated by the one-electron oxidation undergoes
deprotonation at the C5-methyl group followed by addition
of molecular oxygen to produce the final oxidation products
dfC and dhmC via a stable tetroxide intermediate, according
to the Russell mechanism[24] (Scheme 2). Recently, Prado
m
+
C
rate constant (1.37ꢁ106 sÀ1) of AQS* at 380 nm (see Fig-
3
ure S2 in the Supporting Information). This provides strong
evidence suggesting that the long-lived triplet excited spe-
cies AQS* could induce the one-electron oxidation of dmC
3
Chem. Eur. J. 2011, 17, 2225 – 2235
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2227

H. Yamada, S.-I Nishimoto et al.
primary intermediate, d C +, may undergo rapid competi-
tive deprotonation from the methyl group and the N4-exo-
cyclic amino group, in what is called a proton-coupled elec-
tron-transfer process.[16] Although the dmC-tetroxide inter-
mediate has not yet been identified, the rate-determining
step in the photosensitizied oxidation of dmC is presumed to
involve the formation of a tetroxide intermediate followed
by C5-hydrogen transfer to produce dfC and dhmC. In addi-
tion, the yields of dfC and dhmC as indicated in Figure 1 may
be partly accounted for by the occurrence of further oxida-
tion from dhmC to dfC during AQS-photosensitization.[23]
In a separate experiment, we also examined similar AQS-
photosensitized degradation of dmC and [D3]dmC in D2O so-
lution, under which conditions the N4-hydrogen atoms of
dmC and [D3]dmC are fully deuterated (Figure S3 in the Sup-
porting Information). If the competitive deprotonation from
m
C
Figure 3. Deuterium isotope effects on the AQS-photosensitized oxida-
tions of dmC and [D3]dmC. Aerobic aqueous solutions of dmC (200 mm)
and [D3]dmC (200 mm) were photoirradiated at 312 nm at 08C in the pres-
ence of AQS (200 mm) in 2 mm sodium cacodylate containing 20 mm
m
+
C
m
m
the N4-exocyclic amino group of the d C species is in-
volved in this reaction mechanism,[16] the rate of dmC degra-
dation as a result of the deprotonation from the C5-methyl
*
NaCl (pH 7.0). Degradation of d C ( ) and [D3]d C (&) and formation
f
f
&
*
of d C ( ) and deuterated d C ( ) were monitored by HPLC and quanti-
fied by reference to authentic samples. Deuterium isotope effects (kH/kD)
were estimated from the ratios of the initial rates of the dmC and
[D3]dmC degradation and those of the dfC and deuterated dfC formation.
m
+
C
group of the d C species is expected to be enhanced upon
replacement of the solvent H2O by D2O. The inverse solvent
isotope effects on the initial rates of the dmC and [D3]dmC
degradations were evaluated as kH O/kD O =0.90 and 0.54, re-
and co-workers suggested the formation of a transient tetr-
oxide by recombination of 5-(hydroperoxymethyl)-2’-deoxy-
uridinyl radicals that underwent decomposition with the re-
lease of singlet oxygen (1O2) by the Russell mechanism.[25]
This mechanism was partly supported by spectroscopic
measurements indicating the bimolecular and monomolecu-
2
2
spectively. Although the difference in the inverse isotope
effect between dmC and [D3]dmC is not fully understood at
present, it is presumable that the enhanced degradation in
D2O solution is attributable to the occurrence of competi-
tive deprotonation from the exocyclic amino group of the
1
m
+
m
+
lar decays of O2.[26] In this light, the observed kinetic iso-
d C or [D3]d C species.
C
C
tope effects may be associated with internal transfer of an
a-hydrogen atom in the dmC-tetroxide intermediate and/or
Photooxidation and strand cleavage at mC derivatives in
DNA sensitized by AQ-tethered ODNs: The photosensi-
tized one-electron oxidation and strand cleavage at mC in
m
+
C
deprotonation of the d C
species into a C5-methyl
carbon-centered radical. It has also been proposed that the
Scheme 2. A plausible reaction mechanism for the AQ-photosensitized oxidation of dmC.
2228
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2225 – 2235

Piperidine-Induced DNA Strand Cleavage at 5-Methylcytosine
FULL PAPER
DNA was characterized by using a modified ODN with an
AQ sensitizer in the interior of a DNA strand (ODN 1
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
cinimidyl ester[27] with ODN (N), which possessed an amino
linker in the middle of sequential bases.[28] In this study, we
used a partial sequence of the human p53 gene correspond-
ing to codons 280–285 of exon 8 and targeted the mC at
codon 282.[29] The sequences and structures of ODNs used
in this study are summarized in Scheme 3.
Figure 4. Photosensitized oxidation of AQ-tethered ODNs. A representa-
tive autoradiogram of denaturing gel electrophoresis for ODN 1(X) and
32P-5’-end-labeled ODN 2(Y) upon 312 nm photoirradiation for 0 h
(lanes 1, 5, and 8), 1 h (lanes 2, 6, and 9), and 2 h (lanes 3, 4, 7, and 10) in
10 mm sodium cacodylate buffer containing 100 mm NaCl (pH 7.0) at
208C. After treatment with hot piperidine (908C, 20 min), the samples
were electrophoresed through denaturing 20% polyacrylamide/7m urea:
Lanes 1–4: ODN 1ACHTUNTRGENNUG ACHTUNGTRENNUNG(AQ1)/
(AQ1)/ODN 2(mC) duplex; lanes 5–7: ODN 1
ODN 2(C) duplex; lanes 8–10: ODN 1(G)/ODN 2(mC) duplex; lane 11:
Maxam–Gilbert G+A sequencing lanes. The sample in lane 4 was not
treated with hot piperidine.
Scheme 3. Sequences and structures of ODNs used in this study.
photoproducts and their Dewar valence isomers.[31] The pre-
vious ESI mass experiments showed that mC residue in
DNA is converted into 5-formylcytosine by photosensitiza-
tion with ODNs possessing an NQ chromophore at the 5’-
end.[12a] In addition, 5-formyluracil, one of the thymidine ox-
idation products, has been shown to form alkali-labile le-
sions in DNA.[11,32] In this context, the present results clearly
indicate that photoexcitation of the AQ sensitizer in the in-
terior of the strand can one-electron oxidize mC to produce
potentially alkali-labile oxidation products such as 5-formyl-
cytosine and can thus result in strand cleavage at the mC
site. In contrast, substantially less strand cleavage at the un-
methylated C site is attributable to the less efficient AQ-
photosensitized one-electron oxidation of C into the corre-
sponding radical cation intermediate than that of mC, proba-
bly because of the considerably smaller free-energy change
of charge separation (ÀDGCS =0.02 eV) between C and AQ
in the triplet excited state (3AQ*).[33,34]
Photoirradiation at lex =312 nm in air of a duplex of
ODN 1ACHTUNGTRENNUNG
(AQ1) with 32P-5’-end-labeled ODN 2(C) or
ODN 2(mC) was performed in 10 mm sodium cacodylate
buffer (pH 7.0) containing 100 mm NaCl at 208C.[30] The
photoirradiated solution was treated with hot piperidine and
then analyzed by polyacrylamide gel electrophoresis
(PAGE). As shown in Figure 4, specific strand cleavage oc-
ACTHNUTRGNEUNG
curred at the target mC site in the ODN 1(AQ1)/ODN 2(mC)
duplex after photoirradiation, whereas only a background
level of strand cleavage was observed at the corresponding
unmethylated
C site for the ODN 1ACHTNUTGNRUE(GN AQ1)/ODN 2(C)
duplex. When the hot piperidine treatment was not em-
ployed after AQ-sensitized photooxidation, the strand cleav-
age bands were substantially suppressed. In a control photo-
reaction of a similar duplex comprising ODN 1(G) and 32P-
5’-end-labeled ODN 2(mC) with a complementary G base in-
stead of the AQ sensitizer, no strand cleavage band was ob-
served at the target mC site. This suggests that the present
photoirradiation conditions could not produce possible
alkali-labile dimeric photoproducts, such as pyrimidine (6-4)
In conjunction with the intense cleavage at the target mC
site, possible strand cleavages at the adjacent G sites, which
are one-electron oxidized the most readily in the DNA
duplex,[10] were negligible in the present AQ-sensitized pho-
Chem. Eur. J. 2011, 17, 2225 – 2235
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2229

H. Yamada, S.-I Nishimoto et al.
tooxidation. Such negligible strand cleavage at the adjacent
guanine bases is consistent with the NQ-sensitized photooxi-
dation of mC in DNA.[12] Thus, the secondary intermediate,
dine.[36] To confirm whether the one-electron transfer pro-
cess may be involved in the selective strand cleavage at the
mC site, we conducted photoreactions of duplexes with a dif-
ferent structure of AQ analogue. Breslin and Schuster re-
ported that a 2-aminoanthraquinone chromophore (AQ2)
has the lowest pp* triplet excited state and can react only
the guanine radical cation (G +), formed through hole trans-
C
m
+
C
fer from the primary intermediate
C
mostly undergoes
CÀ
rapid charge recombination with the neighboring AQ to
suppress net strand cleavage at adjacent G sites.[14c] In con-
by electron transfer.[37] A duplex of ODN 1
ACHTNGUTRENNU(G AQ2), bearing
+
trast to the behavior of the more stable radical cation G
,
AQ2 instead of AQ1, with its complementary ODN 2(mC)
was prepared and photoirradiated at 312 nm in a similar
manner. As illustrated in Figure 5, we could detect specific
C
m
+
C
CÀ
the charge recombination of
cient, probably because the
C
with AQ may be ineffi-
species undergoes kineti-
m
+
C
C
cally more favorable deprotonation at the C5-methyl
group.[35] As a consequence, selective strand cleavage pro-
ceeds almost exclusively at the target mC site without unfav-
orable cleavage at the adjacent G sites upon AQ-sensitized
photooxidation. In view of the evidence that mC is usually
generated in CpG-rich regions of DNA,[1,2,29] the suppres-
sion of a competitive strand cleavage at adjacent G doublets
should be a strong advantage for distinct detection of the
mC site on a sequencing gel.
To gain further insight into the suppression of photooxida-
tive strand cleavage at G sites, the photoreactivity of a
ODN 1(G)/32P-labeled ODN 2(mC) duplex was examined in
the presence of exogenous AQS sensitizer instead of ODN 1
ACHTUNGTRENNUNG(AQ1). Efficient strand cleavage at the 5’-G site of a G dou-
blet was observed after 2 h photoirradiation followed by
treatment with hot piperidine (Figure S4 in the Supporting
Information), which indicated that a positive charge injected
on a DNA base from the exogenous 3AQS* migrated
through the DNA duplex until it was trapped by a G dou-
blet. In this case, longer separation of the positive and nega-
tive charges occurs between the exogenous AQS chromo-
phore and the G doublet: the charge recombination process
+
C
CÀ
between the G and exogenous AQS species may become
more difficult and, therefore, preferential strand cleavage at
the G site can proceed. Thus, arrangement of the AQ sensi-
tizer in the vicinity of the target mC residue would be essen-
tial for selective strand cleavage at mC residues without un-
favorable cleavage at adjacent G residues.
Figure 5. A representative autoradiogram of denaturing gel electrophore-
sis for ODN 1ACHTNUGRTENUNG(AQ2) and ODN 2(Y) upon 312 nm photoirradiation for
0 h (lanes 2 and 5), 1 h (lanes 3 and 6), and 2 h (lanes 4 and 7) in 10 mm
sodium cacodylate buffer containing 100 mm NaCl (pH 7.0) at 208C.
After treatment with hot piperidine (908C, 20 min), the samples were
electrophoresed through denaturing 20% polyacrylamide/7m urea:
lanes 2–4: ODN 1
ODN 2(C) duplex; lane 1: Maxam–Gilbert G+A sequencing lanes.
The thermal stability of duplexes was evaluated by moni-
toring the melting temperatures (Tm) in 10 mm sodium caco-
dylate buffer (pH 7.0) containing 100 mm NaCl (Figure S5 in
the Supporting Information). The lower Tm value of the
ODN 1(N)/ODN 2(mC) duplex (Tm =55.18C) relative to a
reference duplex of ODN 1(G)/ODN 2(mC) (Tm =59.68C)
suggests that the introduction of an amino linker moiety
into the DNA backbone reduces the thermal stability of the
duplex, probably because of the partial absence of hydrogen
bonding. In contrast, slight stabilization was observed for
(AQ2)/ODN 2(mC) duplex; lanes 5–7: ODN 1
ACHUTGTNRENNUG ACHTUNGTRENNUNG(AQ2)/
strand cleavage at the mC site in the ODN 1
ACHTUNGTRENNUNG(AQ2)/
ODN 2(mC) duplex after 2 h photoirradiation, whereas only
a background level of cleavage at the corresponding C resi-
the AQ-modified duplex of ODN 1
A
due was observed for the ODN 1ACTHUNGTERNNU(G AQ2)/ODN 2(C) duplex.
62.78C). These results suggest that the observed thermal sta-
bilization is responsible for the intercalation of the AQ
chromophore between flanking base pairs.
In addition, strand cleavages at adjacent G sites were sub-
stantially suppressed. These results strongly suggest that the
m
+
C
C
species formed by eventual one-electron transfer to
3AQ2* is attributable for the photosensitized oxidative
strand cleavage at mC in the ODN 1(AQ2)/ODN 2(mC)
One-electron transfer from the intercalated AQ deriva-
tives with np* configuration can also occur to form a base
AHCTUNGTRENNUNG
CÀ
radical cation and the AQ species, due to orbital overlap
duplex. It is thus concluded that selective strand cleavage at
between nucleobases and the AQ chromophore; this leads
to oxidative strand cleavage upon treatment with piperi-
the mC site is likely to occur through one-electron transfer
3
m
+
oxidation of mC by intercalated AQ* into
C
C
CÀ
and AQ
2230
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2225 – 2235

Piperidine-Induced DNA Strand Cleavage at 5-Methylcytosine
FULL PAPER
species, but not through direct hydrogen atom abstraction
from mC by 3AQ*.
ODN, a duplex of ODN 1ACTHUNGTERNUN(G NQ1), bearing NQ instead of AQ,
with the complementary ODN 2(mC) containing mC was also
prepared and photoirradiated at 312 nm, by reference to the
AQ1 photosensitization. As shown in Figure 7, whereas an
intense spot corresponding to strand cleavage at the mC site
was clearly observed in the AQ-sensitized photooxidation, a
smaller amount of strand cleavage occurred at the mC site in
the NQ-sensitized photooxidation: the apparent efficiencies
(relative values) of strand cleavage at the mC site were 12%
In light of the kinetic isotope effects observed in the
AQS-photosensitized oxidation of dmC in aqueous solution,
we further examined the AQ-photosensitized oxidation re-
activity of the DNA duplex containing 5-trideuteriomethyl-
cytosine ([D3]mC) instead of mC at the target site. The syn-
thesis of the ODN containing [D3]mC is outlined in
Scheme 4. N-Benzoyl-5-([D3]methyl)-2’-deoxycytidine (5)
in the ODN
ACHTUNGTRENNUNG
4% in the ODNACHTUNGTRENNNUG
More quantitatively, by using a
phenylglyoxylic acid chemical
actinometer,[38] the relative
quantum yields of AQ- and
NQ-photosensitized
strand
cleavage at the mC site were es-
timated to be 0.9ꢁ10À5 and
0.6ꢁ10À5, respectively, which in-
dicated that the AQ photosensi-
tization shows 1.5-times higher
efficiency in the one-electron
oxidation of mC than the NQ
Scheme 4. Reagents and conditions: a) Benzoyl chloride, 4-(dimethylamino)pyridine, CH2Cl2, 94%; b) triethyl-
amine trihydrogenfluoride, THF, quant.; c) 4,4’-dimethoxytritylchroride, pyridine, 92%; d) N,N-diisopropyl-
methyl phosphonamidic chloride, diisopropylethylamine, CH3CN, quant.; e) automated DNA synthesis. Bz:
benzoyl; DMTr: 4,4’-dimethoxytrityl.
was prepared from 3 by standard benzamide protection of
the exocyclic amine and subsequent desilylation of the hy-
droxy groups. The resulting N4-protected 2’-deoxycytidine
analogue 5 was converted into 4,4’-dimethoxytrityl deriva-
tive 6 and then converted into the corresponding phosphora-
midite derivative 7; this was followed by incorporation into
DNA by using a DNA synthesizer and conventional b-cya-
nophosphoramidite chemistry. Photoirradiation at 312 nm of
the ODN 1ACHTUNGTRENNUNG
(AQ1)/ODN 2([D3]mC) duplex and subsequent
piperidine treatment induced selective strand cleavage at
the [D3]mC site, as in a similar photooxidation of mC in the
ODN 1ACHTUNGTRENNUNG
(AQ1)/ODN 2(mC) duplex (Figure S6 in the Support-
ing Information). Figure 6 shows the quantitative analysis of
AQ-photosensitized strand cleavages at the mC and [D3]mC
sites in the ODN 1
ODN 1ACHTUNGTRENNUNG
(AQ1)/ODN 2([D3]mC) duplex, respectively. The effi-
(AQ1)/ODN 2(mC) duplex and the
ACHTUNGTRENNUNG
Figure 6. Photooxidative strand cleavages at the mC and [D3]mC sites as
determined by polyacrylamide gel electrophoresis. The ODN 1ACHTUNGTRENNUNG(AQ1)/
ciency of strand cleavage at [D3]mC was slightly suppressed
in comparison with that at mC. Whereas the kinetic isotope
effect on the net strand cleavage was small, it is presumable
ODN 2(Y) duplexes (0.5 mm) were photoirradiated (312 nm, 0–3 h) in
10 mm sodium cacodylate buffer containing 100 mm NaCl (pH 7.0) at
08C. After treatment with hot piperidine (908C, 20 min), the samples
were electrophoresed through denaturing 20% polyacrylamide/7m urea.
The cleavage efficiencies were calculated from the intensities of the re-
spective strand cleavages at the mC (*) and [D3]mC (&) sites relative to
the total intensities. Each error bar represents the standard deviation cal-
culated from three experimental results.
m
+
C
that deprotonation of the
C
intermediate into a C5-
methyl carbon-centered radical is involved in the formation
of alkali-labile oxidation products, such as 5-(hydroxyme-
thyl)cytosine and/or 5-formylcytosine,[12,15] which lead to se-
lective DNA strand cleavage at the mC site. The occurrence
of bimolecular reaction processes such as internal a-hydro-
gen transfer in an mC-tetroxide intermediate, as in the case
of the AQS-photosensitized oxidation of dmC, may be ruled
out in the AQ-photosensitized oxidation of the mC site in
the restricted structure of the DNA duplex. Similar reaction
pathways were recently proposed to explain strand cleavage
at consecutive T residues during hole transfer in DNA.[11]
To compare the one-electron oxidizing abilities between
photoirradiated AQ1 and NQ1 sensitizers as tethered to
photosensitization. The relatively more efficient strand
ACTHNUTRGNEUNG
cleavage at the mC site in the ODN 1(AQ1)/ODN 2(mC)
duplex is attributable to a considerably larger quantum yield
of intersystem crossing for AQ (Fisc =0.9)[39] relative to NQ
(Fisc =0.66).[40] Recently, Lewis, Wasielewski and co-workers
reported that one-electron oxidation of DNA bases by sin-
glet excited AQ (1AQ*) and charge recombination of the re-
sulting singlet radical ion pair is more efficient than the for-
Chem. Eur. J. 2011, 17, 2225 – 2235
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2231

H. Yamada, S.-I Nishimoto et al.
porated an AQ sensitizer into the backbone of the DNA
strand. The AQ photosensitization of the DNA duplex in-
duced efficient one-electron oxidation of mC and thus result-
ed in exclusive strand cleavage at the target mC site upon
treatment with hot piperidine. In contrast to the efficient
strand cleavage at the mC site, the adjacent GG site, which
would normally be a hotspot for photosensitized oxidation,
underwent a smaller amount of such oxidative strand cleav-
age, presumably due to the occurrence of rapid charge re-
+
C
CÀ
combination processes between the G and AQ species.
Evidence was also obtained that similar AQ-photosensitized
strand cleavage at the [D3]mC site was slightly suppressed
relative to cleavage at the mC site, which suggests that the
AQ-photosensitized strand cleavage at the mC site potential-
ly involves a pathway of deprotonation at the C5-methyl
m
+
C
group on the
C
species into a methyl carbon-centered
radical intermediate. The more efficient AQ-sensitized pho-
tooxidative strand cleavage at the mC site is probably attrib-
utable to an enhanced intersystem crossing of photoexcited
AQ to generate a larger yield of the primary radical cation
m
intermediate,
+. The results presented herein may pro-
C
C
vide a guide for the molecular design of highly sensitive
photochemical methods for the identification and analysis of
methylated cytosine in DNA.
Figure 7. Comparison of AQ- and NQ-photosensitized oxidations of
DNA duplexes bearing mC. A representative autoradiogram of denatur-
ing gel electrophoresis for 32P-5’-end-labeled AQ- or NQ-tethered du-
plexes possessing an mC site. The ODN 1(X)/ODN 2(mC) duplexes
(0.5 mm) were photoirradiated (312 nm, 0–2 h) in 10 mm sodium cacody-
late buffer containing 100 mm NaCl (pH 7.0) at 208C. After treatment
with hot piperidine (908C, 20 min), the samples were electrophoresed
through denaturing 20% polyacrylamide/7m urea: lane 1: Maxam–Gil-
Experimental Section
General methods: 9,10-Anthraquinone 2-sulfonate (AQS) was purchased
from Tokyo Chemical Industry and was used without further purification.
5-Methyl-2’-deoxycytidine (dmC) was obtained commercially from MP
Biomedicals. 5-Formyl-2’-deoxycytidine (dfC)[42] and 5-(hydroxymethyl)-
2’-deoxycytidine (dhmC)[43] were prepared as described previously. An-
thraquinone 2-carboxylic acid N-hydroxysuccinimidyl ester (AQ-NHS)
and 2-(3-bromopropionamido)anthraquinone were prepared following re-
ported methods.[27,37] The reagents for the DNA synthesizer were pur-
chased from Glen Research. Oligodeoxynucleotides (ODNs) containing
an amino linker were synthesized on an Applied Biosystems Model 3400
DNA/RNA synthesizer with BD Uni-Link AminoModifier (BD Biosi-
cences Clontech). NQ-tethered ODNs were prepared by conjugation of
3-(N-hydroxysuccinimidylethyl)-NQ with ODNs containing an amino
linker.[12a,16] All complementary ODNs were purchased from Invitrogen.
Mass spectrometry analyses of ODNs were performed with a matrix-as-
sisted laser desorption ionization time-of-flight (MALDI-TOF) mass
spectrometer (Perseptive Voyager Elite, acceleration voltage 21 kV, nega-
tive mode) with 2’,3’,4’-trihydroxyacetophenone as the matrix, by using
T8 ([MÀH]À 2370.61), T17 ([MÀH]À 5108.37), and T27 ([MÀH]À 8150.33)
as the internal standards. Calf intestinal alkaline phosphatase (AP), nu-
clease P1 (P1), and phosphodiesterase I were purchased from PROME-
GA, YAMASA, and ICN, respectively. [g-32P]Adenosine triphosphate
bert G+A sequencing lanes; lanes 2–4: ODN 1
(AQ1)/ODN 2(mC)
ACHTUNGTRENNUNG
duplex; lanes 5–7: ODN 1
AHCTUNGTRENNUNG
mation of the long-lived triplet charge separated state that is
responsible for the oxidative DNA strand cleavage.[41] In
this light, the considerably low quantum yield of strand
cleavage at the mC base in the present AQ-photosensitized
oxidation may be accounted for by the concomitant occur-
rence of fast singlet charge separation and recombination
1
processes between AQ* and mC. Another possibility cannot
be ruled out that hole transfer from the primary radical
m
cation intermediate,
+, to the neighboring G sites and/or
C
C
m
+
CÀ
C
back electron transfer from AQ to
C
may occur to
lower the efficiency of strand cleavage at the mC site.[9g,14c]
Conclusion
(ATP; 6000 CimmolÀ1 and T4 polynucleotide kinase (10 unitsmLÀ1
) )
were obtained from Perkin–Elmer and Nippon Gene, respectively. All
aqueous solutions were prepared by using purified water (YAMATO,
WR600A).
In this study, we have demonstrated that photosensitized
3
one-electron oxidation of dmC by AQS* affords the charac-
4-(N-1-Triazoyl)-3’,5’-O-bis(tert-butyldimethylsilyl)-5-([D3]methyl)-2’-de-
oxyuridine (2): 3’,5’-O-Bis(tert-butyldimethylsilyl)-5-([D3]methyl)-2’-deox-
yuridine (1) was prepared as reported previously.[44] 1,2,4-Triazole
(848 mg, 12.3 mmol) was suspended in acetonitrile (20 mL) cooled to
08C, to which POCl3 (0.27 mL, 2.96 mmol) was slowly added. Triethyla-
mine was then added dropwise and the suspension was stirred for 20 min.
Compound 1 (290 mg, 0.6 mmol) was dissolved in acetonitrile (8 mL) and
added to the solution, and the solution was continuously stirred over-
night. The reaction was quenched with water and extracted with ethyl
teristic final oxidation products of dfC and dhmC. The deute-
rium isotope effects on the AQS-photosensitized one-elec-
tron oxidation of [D3]dmC suggest that internal a-hydrogen
transfer in an mC-tetroxide intermediate is a possible rate-
determining step to produce dfC and dhmC. Photooxidative
strand cleavage at an mC site in a DNA duplex was also in-
vestigated by using a modified ODN that covalently incor-
2232
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2225 – 2235

Piperidine-Induced DNA Strand Cleavage at 5-Methylcytosine
FULL PAPER
À
À
acetate. The organic layer was washed with brine, dried over Na2SO4, fil-
tered, and concentrated. The crude product was purified by column chro-
matography (SiO2, 30% ethyl acetate/hexane) to give 2 (229 mg 72%) as
a colorless oil: 1H NMR (CDCl3, 400 MHz): d=9.26 (s, 1H; azole), 8.22
(t, 1H, J=6.6 Hz; C1’ H), 5.27 (d, 1H, J=4.3 Hz; C3’ OH), 5.11 (t, 1H,
À
À
À
J=5.1 Hz; C5’ OH) 4.28–4.24 (1H; C4’ H), 3.81 (1H; C3’ H), 3.67–3.55
(2H; C5’ H), 2.31–2.06 ppm (2H; C2’ H); 13C NMR ([D6]DMSO,
100 MHz): d=178.1, 159.2, 147.4, 138.9, 136.7, 132.5, 129.3, 128.3, 109.7,
104.8, 87.7, 84.9, 70.0, 53.5, 20.5 ppm; FABMS (matrix: glycerol): m/e:
349 [(M+H)+]; HRMS: calcd for C17H17D3N3O5: 349.1591; found:
349.1591.
À
À
À
À
(s, 1H; azole), 8.08 (s, 1H; C6 H), 6.27 (t, 1H, J=6.2 Hz; C1’ H), 4.39–
À
À
À
4.35 (1H; C4’ H), 4.04 (1H; C3’ H), 3.94 (dd, 1H, J=11.6, 2.6 Hz; C5’
À
H), 3.78 (dd, 1H, J=11.6, 2.6 Hz; C5’ H), 2.62 (ddd, 1H, J=13.5, 6.2,
À
À
3.7 Hz; C2’ H), 2.09–2.02 (1H; C2’ H), 0.89 (s, 9H; TBS), 0.88 (s, 9H;
TBS), 0.09 (d, 6H, J=5.1 Hz; TBS), 0.06 ppm (d, 6H, J=4.9 Hz; TBS);
13C NMR (CDCl3, 100 MHz): d=158.1, 153.9, 153.4, 146.6, 145.0, 105.1,
88.8, 87.8, 71.6, 62.6, 42.6, 25.9, 25.7, 18.4, 18.0, À4.5, À4.9, À5.3,
À5.4 ppm; FABMS (matrix: 3-nitrobenzyl alcohol): m/e: 525 [(M+H)+];
HRMS: calcd for C24H41D3N5O4Si2: 525.3120; found: 525.3132.
5’-Dimethoxytrityl-N-benzoyl-5-([D3]methyl)-2’-deoxycytidine (6): 4,4’-
Dimethoxytritylchroride (420 mg, 1.24 mmol) was added to a solution of
5 (300 mg, 0.86 mmol) in anhydrous pyridine and the mixture was stirred
at room temperature for 2.5 h. The reaction was quenched with methanol
(1 mL) and the solvent was evaporated under reduced pressure. The
crude product was purified by column chromatography (SiO2, 0–5%
methanol/dichloromethane with 1% triethylamine) to give 6 (547 mg,
92%) as a foam: 1H NMR ([D6]DMSO, 400 MHz): d=8.17 (d, 2H, J=
7.1 Hz; Bz), 7.79 (s, 1H; C6 H), 7.60–7.22 (13H; DMTr, Bz), 6.90 (dd,
4H J=8.8, 2.0 Hz; DMTr), 6.20 (t, 1H, J=6.5 Hz; C1’ H), 5.36 (d, 1H,
J=4.6 Hz; C3’ OH), 4.34 (1H; C4’ H), 3.94–3.93 (1H; C3’ H), 3.72 (s,
6H; OCH3), 3.31–3.23 (2H; C5’ H), 2.32–2.24 ppm (2H; C2’ H);
13C NMR ([D6]DMSO, 100 MHz): d=178.1, 161.9, 158.9, 158.6, 158.1,
149.6, 147.4, 144.7, 138.3, 137.3, 136.6, 135.4, 135.2, 132.5, 129.7, 129.3,
128.9, 128.3, 128.2, 127.9, 127.7, 126.8, 125.3, 123.9, 113.2, 110.0, 85.9,
84.9, 70.2, 63.5, 55.0, 21.0 ppm; FABMS (matrix: 3-nitrobenzyl alcohol):
m/e: 651 [(M+H)+]; HRMS: calcd for C38H35D3N3O7: 651.2898; found:
651.2888.
3’,5’-O-Bis(tert-butyldimethylsilyl)-5-([D3]methyl)-2’-deoxycytidine (3):
25% Ammonium deuteroxide (10 mL) was added to a solution of 2
(732 mg, 1.4 mmol) in acetonitrile (20 mL) and the mixture was stirred at
room temperature for 4 h. The reaction mixture was concentrated in
vacuo, then diluted with water and extracted with ethyl acetate. The or-
ganic layer was washed with brine, dried over Na2SO4, filtered, and con-
centrated. The crude product was purified by column chromatography
À
À
À
À
À
À
À
(SiO2, 33% acetone/chloroform) to give 3 (730 mg, quant.) as a colorless
1
À
oil: H NMR (CDCl3, 400 MHz): d=7.57 (s, 1H; C6 H), 6.30 (t, 1H, J=
À
À
À
À
6.5 Hz; C1’ H), 4.35–4.32 (1H; C4’ H), 3.92–3.84 (2H; C3’ H, C5’ H),
3.74 (dd, 1H, J=11.3, 2.6 Hz; C5’ H), 2.36 (1H; C2’ H), 1.95 (dd, 1H,
À
À
À
J=13.4, 6.6 Hz; C2’ H), 0.88 (18H; TBS), 0.06 ppm (12H; TBS);
13C NMR (CDCl3, 100 MHz): d=165.2, 155.6, 138.4, 101.2, 87.7, 85.9,
71.8, 62.7, 42.2, 29.3, 25.9, 25.7, 18.4, 18.0, À4.6, À4.9, À5.4 ppm; FABMS
(matrix: 3-nitrobenzyl alcohol): m/e: 473 [(M+H)+]; HRMS: calcd for
C22H41D3N3O4Si2: 473.3059; found: 473.3045.
3’-(N,N-Diisopropylmethylphosphonamidite)-5’-dimethoxytrityl-N-benzo-
yl-5-([D3]methyl)-2’-deoxycytidine (7): Compound 6 was coevaporated
twice with anhydrous acetonitrile (1 mL). N,N-Diisopropylmethyl phos-
phonamidic chloride (16 mL, 0.082 mmol) was added to a solution of 6
(50 mg, 0.077 mmol) and diisopropylethylamine (39 mL, 0.23 mmol) in an-
hydrous acetonitrile (0.75 mL) and the mixture was stirred at room tem-
perature for 1.5 h. The reaction mixture was filtered and then placed on
a DNA synthesizer.
5-[D3]Methyl-2’-deoxycytidine ([D3]dmC): Acetic acid (32.5 mL) and tet-
rabutylammonium fluoride (1.2 mL, 1.2 mmol) were added to a solution
of 3 (179 mg, 0.38 mmol) in anhydrous tetrahydrofuran (3.8 mL) and the
mixture was stirred at room temperature for 7 h. The reaction mixture
was concentrated. The crude product was roughly purified by column
chromatography (SiO2, 20% methanol/chloroform with 2% triethyl-
amine) and then purified by reversed-phase HPLC (7% methanol/water)
to give [D3]dmC (58 mg, 63%) as a white solid: 1H NMR (CD3OD,
Synthesis of ODNs containing a C5-deuterated 5-methylcytosine base:
ODNs containing 5-[D3]methyl-2’-deoxycytidine were synthesized on an
Applied Biosystems Model 3400 DNA/RNA synthesizer by using stan-
dard b-cyanoethylphosphramidite chemistry. Synthesized ODNs were pu-
rified by reversed-phase HPLC on an Inertsil ODS-3 column (10ꢁ
250 mm, elution with a solvent mixture of 0.1m triethylammonium ace-
tate (TEAA) at pH 7.0, linear gradient over 60 min from 0% to 30%
acetonitrile at a flow rate of 3.0 mLminÀ1). The purity and concentration
of the modified ODN was determined by complete digestion with AP,
P1, and phosphodiesterase I. The synthesized ODN ([D3]mC) was identi-
fied by MALDI-TOF mass spectrometry ([MÀH]À; calcd: 5583.67;
found: 5585.5).
À
À
400 MHz): d=7.76 (s, 1H; C6 H), 6.16 (t, 1H, J=6.5 Hz; C1’ H), 4.30–
À
À
À
4.26 (1H; C4’ H), 3.82 (1H; C3’ H), 3.74–3.62 (2H; C5’ H), 2.23 (ddd,
1H, J=13.5, 6.2, 3.7 Hz; C2’ H), 2.07–2.01 ppm (1H; C2’ H); 13C NMR
(CD3OD, 100 MHz): d=167.1, 158.0, 140.2, 104.2, 88.8, 87.3, 71.9, 62.7,
42.0, 30.7 ppm; FABMS (matrix: glycerol): m/e: 245 [(M+H)+]; HRMS:
calcd for C10H13D3N3O4: 245.1329; found: 245.1329.
À
À
3’,5’-O-Bis(tert-butyldimethylsilyl)-N-benzoyl-5-([D3]methyl)-2’-deoxycy-
tidine (4): Benzoyl chloride (40 mL, 0.33 mmol) was added to a solution
Synthesis of ODNs possessing 9,10-anthraquinone (AQ) sensitizer:
of
3
(100 mg, 0.21 mmol) and 4-dimethylaminopyridine (55 mg,
0.45 mmol) in anhydrous dichloromethane (9.5 mL). The reaction was
stirred at room temperature overnight. The reaction was quenched with
methanol (1 mL) and concentrated under reduced pressure. The crude
product was purified by column chromatography (SiO2, 20% ethyl ace-
Synthesis of ODN 1ACTHNGUTERNNU(G AQ1): 10 mm AQ-NHS in acetonitrile solution con-
taining 20 vol% dimethylformamide (45 mL) and saturated NaHCO3
(30 mL) were added to a 1.2 mm solution (total volume: 75 mL) of ODNs
possessing an aminohexyl linker internally and incubated at 378C over-
night. The reaction mixture was first subjected to gel filtration by using
MicroBio-spin 6 columns (BioRad) and then purified by reversed-phase
HPLC with a 0–30% linear gradient (over 30 min) of acetonitrile/0.1m
TEAA buffer solution at pH 7.0. The purity and concentration of the
AQ-modified ODN were determined by complete digestion with AP, P1,
1
tate/hexane) to give 4 (114 mg, 94%) as a white foam: H NMR (CDCl3,
À
400 MHz): d=8.30 (2H; Bz), 7.67 (s, 1H; C6 H), 7.49 (3H; Bz), 6.32
À
À
À
(dd, 1H, J=7.6, 6.0 Hz; C1’ H), 4.39 (1H; C4’ H), 3.96 (1H; C3’ H),
À
3.88 (dd, 1H, J=11.4, 2.4 Hz; C5’ H), 3.76 (dd, 1H, J=11.4, 2.4 Hz;
À
À
À
C5’ H), 2.34–2.29 (1H; C2’ H), 2.01 (1H; C2’ H), 0.91 (18H; TBS),
0.11 ppm (12H; TBS); 13C NMR (CDCl3, 100 MHz): d=179.6, 159.8,
147.9, 137.3, 136.8, 132.3, 129.9, 128.1, 111.5, 88.1, 85.5, 72.3, 63.0, 41.7,
27.6, 25.9, 25.7, 18.4, 18.0, À4.6, À4.8, À5.4 ppm; FABMS (matrix: 3-ni-
trobenzyl alcohol): m/e: 577 [(M+H)+]; HRMS: calcd for
C29H45D3N3O5Si2: 577.3321; found: 577.3310.
and phosphodiesterase I. The synthesized ODN 1ACHTNUTGRNEUNG(AQ1) was identified by
MALDI-TOF mass spectrometry ([MÀH]À; calcd: 5546.68; found:
5546.84).
Synthesis of ODN 1ACTHNUTRGNEUNG(AQ2): Saturated 2-(3-bromopropionamido)anthra-
quinone (approximately 5 mm) in ethanol solution (250 mL) and saturated
NaHCO3 (30 mL) were added to a 0.6 mm solution (total volume: 480 mL)
of ODNs possessing an aminohexyl linker internally and incubated at
608C for 20 h. The reaction mixture was purified by reversed-phase
N-Benzoyl-5-([D3]methyl)-2’-deoxycytidine (5): Triethylamine trihydro-
genfluoride (43 mL, 0.26 mmol) was added to a solution of 4 (50 mg,
0.087 mmol) in anhydrous tetrahydrofuran (0.9 mL). The reaction mix-
ture was stirred at room temperature overnight. The solvent was re-
moved under reduced pressure. The crude product was purified by
column chromatography (SiO2, 0–10% methanol/chloroform) to give 5
(30 mg, quant.) as a white foam: 1H NMR ([D6]DMSO, 400 MHz): d=
HPLC. The synthesized ODN 1ACHTNUTRGNENU(G AQ2) was identified by ESI-TOF mass
spectrometry ([MÀ4H]4À; calcd: 1392.4; found: 1392.5).
Photosensitized oxidation of 5-methyl-2’-deoxycytidine by anthraquinone
2-sulfonate: A solution of dmC (200 mm) in 2 mm sodium cacodylate
buffer (pH 7.0) containing 20 mm NaCl was added to an aqueous solution
À
8.19 (d, 2H, J=7.1 Hz; Bz), 8.06 (s, 1H; C6 H), 7.64–7.47 (3H; Bz), 6.16
Chem. Eur. J. 2011, 17, 2225 – 2235
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2233

H. Yamada, S.-I Nishimoto et al.
of AQS (200 mm). The solution (100 mL) was exposed to 312 nm UV light
with a Lourmat TFX-20m transilluminator (Vilber Lourmat, France) at
08C. Analytical HPLC was performed with a HITACHI Lachrom Elite
HPLC system. Sample solutions (10 mL) were injected onto a reversed-
phase column (Inertsil ODS-3, GL Sciences Inc., 4.6ꢁ250 mm). A sol-
vent mixture of TEAA buffer solution (0.1m, pH 7.0) containing 5 vol%
acetonitrile was delivered as the mobile phase. The column eluents were
monitored by UV absorbance at 260 nm. In the deuterium isotope experi-
ment, aqueous solutions of dmC (200 mm) or [D3]dmC (200 mm) containing
AQS (200 mm) were photoirradiated under the conditions described
above. After 312 nm photoirradiation, the solution was immediately sub-
jected to HPLC analysis. Each deuterium isotope effect (kH/kD) of degra-
dation and formation was estimated from the initial rates of dmC and
[D3]dmC degradation and those of dfC and deuterated dfC formation.
lich, Oncogene 2002, 21, 5400–5413; d) A. Bird, Genes Dev. 2002,
16, 6–21.
[2] a) P. A. Jones, D. Takai, Science 2001, 293, 1068–1070; b) R. A. Mar-
tienssen, V. Colot, Science 2001, 293, 1070–1074; c) K. D. Robert-
son, Nat. Rev. Genet. 2005, 6, 597–610; d) P. A. Jones, S. B. Baylin,
Cell 2007, 128, 683–692.
[3] a) H. Ohmori, J. Tomizawa, A. M. Maxam, Nucleic Acids Res. 1978,
5, 1479–1485; b) G. M. Church, W. Gilbert, Proc. Natl. Acad. Sci.
USA 1984, 81, 1991–1995; c) G. P. Pfeifer, S. D. Steigerwald, P. R.
Mueller, B. Wold, A. D. Riggs, Science 1989, 246, 810–813.
[4] a) H. Hayatsu, Y. Wataya, K. Kai, S. Iida, Biochemistry 1970, 9,
2858–2866; b) R. Shapiro, R. E. Servis, M. Welcher, J. Am. Chem.
Soc. 1970, 92, 422–424.
[5] a) M. Frommer, L. E. McDonald, D. S. Millar, C. M. Collis, F. Watt,
G. W. Grigg, P. L. Molloy, C. L. Paul, Proc. Natl. Acad. Sci. USA
1992, 89, 1827–1831; b) J. G. Herman, J. R. Graff, S. Myçhꢃnen,
B. D. Nelkin, S. B. Baylin, Proc. Natl. Acad. Sci. USA 1996, 93,
9821–9826; c) M. L. Gonzalgo, P. A. Jones, Nucleic Acids Res. 1997,
25, 2529–2531; d) Z. Xiong, P. W. Laird, Nucleic Acids Res. 1997, 25,
2532–2534.
Nanosecond laser flash photolysis: The laser flash photolysis experiments
were carried out with a Unisoku TSP-601 flash spectrometer. A Continu-
um Surelite-I Nd:YAG (Q-switched) laser with the third harmonic at
355 nm (approximately 50 mJ per 6 ns pulse) was employed for the flash
photoirradiation. Further details of the laser flash photolysis system have
been described previously.[45] Aqueous acetonitrile solutions (acetonitrile/
water, 9:1) of dmC (500 mm) containing AQS (50 mm) were deaerated by
passing argon through the solution prior to the laser flash photolysis ex-
periments.
Photooxidative cleavage reaction and PAGE analysis: ODNs were 5’-32P-
labeled by phosphorylation with [g-32P]ATP (4 mL) and T4 polynucleo-
tide kinase (4 mL). The reaction mixtures were purified by using a QIA-
quick Nucleotide Removal Kit (QIAGEN) to remove excess unincorpo-
rated nucleotide. The 32P-5’-end labeled ODNs (<0.4 mm strand concen-
tration) were hybridized by their complementary ODNs possessing the
AQ1 or NQ1 chromophore (0.5 mm) in 2 mm sodium cacodylate buffer
(pH 7.0) containing 20 mm NaCl. Hybridization was achieved by heating
the sample at 908C for 5 min and slowly cooling to room temperature.
The solutions of 32P-5’-end labeled duplex were irradiated at 312 nm UV
light at 208C on exposure to air. After irradiation, all reaction mixtures
were subjected to precipitation by ethanol. The precipitated DNA was re-
solved in 10% piperidine (50 mL), heated at 908C for 20 min, and con-
centrated. The resulting samples were then analyzed by PAGE, for which
the experimental details were described previously.[12]
[6] For a recent review, see: A. Okamoto, Org. Biomol. Chem. 2009, 7,
21–26.
[7] a) S. Bareyt, T. Carell, Angew. Chem. 2008, 120, 187–190; Angew.
Chem. Int. Ed. 2008, 47, 181–184; b) K. Tanaka, K. Tainaka, T.
Kamei, A. Okamoto, J. Am. Chem. Soc. 2007, 129, 5612–5620; c) K.
Tanaka, K. Tainaka, T. Umemoto, A. Nomura, A. Okamoto, J. Am.
Chem. Soc. 2007, 129, 14511–14517.
[8] a) Long-Range Charge Transfer in DNA
I (Topics in Current
Chemistry Vol. 236) (Ed.: G. B. Schuster), Springer-Verlag, New
York, 2004; b) Long-Range Charge Transfer in DNA II (Topics in
Current Chemistry Vol. 237) (Ed.: G. B. Schuster), Springer-Verlag,
New York, 2004; c) Charge Transfer in DNA (Ed.: H.-A. Wagen-
knecht), Wiley-VCH, Weinheim, 2005; d) H.-A. Wagenknecht, Nat.
Prod. Rep. 2006, 23, 973–1006.
[9] For examples, see: a) M. E. NfflÇez, J. K. Barton, Curr. Opin. Chem.
Biol. 2000, 4, 199–206; b) G. B. Schuster, Acc. Chem. Res. 2000, 33,
253–260; c) B. Giese, Acc. Chem. Res. 2000, 33, 631–636; d) F. D.
Lewis, R. L. Letsinger, M. R. Wasielewski, Acc. Chem. Res. 2001, 34,
159–170; e) K. Nakatani, C. Dohno, I. Saito, J. Am. Chem. Soc.
2000, 122, 5893–5894; f) K. Kawai, T. Takada, T. Nagai, X. C. Cai,
A. Sugimoto, M. Fujitsuka, T. Majima, J. Am. Chem. Soc. 2003, 125,
16198–16199; g) F. Bergeron, D. Houde, D. J. Hunting, J. R.
Wagner, Nucleic Acids Res. 2004, 32, 6154–6163.
[10] a) H. Sugiyama, I. Saito, J. Am. Chem. Soc. 1996, 118, 7063–7068;
b) S. Steenken, S. V. Jovanovic, J. Am. Chem. Soc. 1997, 119, 617–
618; c) I. Saito, T. Nakamura, K. Nakatani, Y. Yoshioka, K. Yama-
guchi, H. Sugiyama, J. Am. Chem. Soc. 1998, 120, 12686–12687;
d) V. Shafirovich, A. Dourandin, N. E. Geacintov, J. Phys. Chem. B
2001, 105, 8431–8435; e) V. Shafirovich, J. Cadet, D. Gasparutto, A.
Dourandin, W. D. Huang, N. E. Geacintov, J. Phys. Chem. B 2001,
105, 586–592.
Determination of quantum yields of photosensitized oxidation and strand
cleavage reaction: Quantum yield measurements were carried out by
using a phenylglyoxylic acid actinometer.[38] The light flux was estimated
to be 1.66ꢁ1015 photonsÀ1. The relative quantum yields of strand cleav-
age at the mC site were calculated from the gel electrophoresis study. In
determining the quantum yields, molar extinction coefficients (e) of the
respective photosensitizers at their absorption wavelengths were estimat-
ed: AQ 2-carboxylic acid: e312 =3605mÀ1 cmÀ1; NQ: e312 =1793mÀ1 cmÀ1
.
Melting temperature (Tm) of hybridized ODNs: 1 mm appropriate ODNs
were dissolved in 10 mm sodium cacodylate buffer (pH 7.0) containing
100 mm NaCl. UV melting curves were recorded on a JASCO-V630 spec-
trophotometer equipped with a multicell block and a Peltier temperature
controller. Melting curves were obtained by monitoring the UV absorb-
ance at 260 nm with elevating temperature from 4 to 808C at a rate of
[11] a) A. Joy, A. K. Ghosh, G. B. Schuster, J. Am. Chem. Soc. 2006, 128,
5346–5347; b) A. Ghosh, A. Joy, G. B. Schuster, T. Douki, J. Cadet,
Org. Biomol. Chem. 2008, 6, 916–928; c) J. Joseph, G. B. Schuster, J.
Am. Chem. Soc. 2009, 131, 13904–13905.
18CminÀ1
.
[12] a) H. Yamada, K. Tanabe, S. Nishimoto, Bioorg. Med. Chem. Lett.
2005, 15, 665–668; b) K. Tanabe, H. Yamada, S. Nishimoto, J. Am.
Chem. Soc. 2007, 129, 8034–8040.
[13] H. Yamada, K. Tanabe, S. Nishimoto, Bioconjugate Chem. 2008, 19,
20–23.
[14] a) B. Armitage, Chem. Rev. 1998, 98, 1171–1200; b) R. P. Fahlman,
R. D. Sharma, D. Sen, J. Am. Chem. Soc. 2002, 124, 12477–12485;
c) T. T. Williams, C. Dohno, E. D. A. Stemp, J. K. Barton, J. Am.
Chem. Soc. 2004, 126, 8148–8158; d) S. Kanvah, G. B. Schuster, J.
Am. Chem. Soc. 2004, 126, 7341–7344; e) F. Shao, K. Augustyn,
J. K. Barton, J. Am. Chem. Soc. 2005, 127, 17445–17452; f) S.
Kanvah, J. Joseph, G. B. Schuster, R. N. Barnett, C. L. Cleveland, U.
Landman, Acc. Chem. Res. 2010, 43, 280–287.
Acknowledgements
This research was financially supported by a Ministry of Education, Sci-
ence, Sports, and Culture Grant-in-Aid for Young Scientists (B). We
thank Prof. Teruyuki Kondo and Dr. Yu Kimura (Advanced Biomedical
Engineering Research Unit, Kyoto University) for LC–ESI-TOF mass
measurements.
[1] a) P. A. Jones, P. W. Laird, Nat. Genet. 1999, 21, 163–167; b) K. D.
Robertson, A. P. Wolffe, Nat. Rev. Genet. 2000, 1, 11–19; c) M. Ehr-
2234
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2225 – 2235

Piperidine-Induced DNA Strand Cleavage at 5-Methylcytosine
FULL PAPER
[15] C. Bienvenu, J. R. Wagner, J. Cadet, J. Am. Chem. Soc. 1996, 118,
11406–11411.
[16] H. Yamada, K. Tanabe, T. Ito, S. Nishimoto, Chem. Eur. J. 2008, 14,
10453–10461.
labile lesions.[31b] We could not identify the formation of such dimer-
ic photoproducts, so it is plausible that the present photoirradiation
conditions did not produce the alkali-labile dimeric photoproducts,
in accordance with the suppression of piperidine-induced strand
cleavage at mC bases in the control duplex (ODN 1(G)/
ODN 2(mC)). See: a) Photochemistry and Photobiology of Nucleic
Acids: Chemistry Vol. I (Ed. S. Y. Wang), Academic Press, New
York, 1976; b) T. Douki, J. Cadet, Biochemistry 1994, 33, 11942–
11950.
[17] A small amount of dfC degradation was observed in the photosensi-
tized oxidation of dmC by the AQS chromophore, which suggests
that prolonged AQS photosensization of dmC may result in the fur-
ther photoreaction of dfC. This behavior was consistent with the pre-
vious NQ-sensitized photooxidation of dmC.[15,16] In contrast, such
degradation was substantially suppressed in a similar photoreaction
of dfC without AQS.
[32] a) M. J. Lustig, J. Cadet, R. J. Boorstein, W. Teebor, Nucleic Acids
Res. 1992, 20, 4839–4845; b) T. Berthod, Y. PØtillot, A. Guy, J.
Cadet, E. Forest, D. Molkot, Nucleosides Nucleotides 1996, 15,
1287–1305; c) H. Ding, M. M. Greenberg, J. Am. Chem. Soc. 2007,
129, 772–773.
[18] I. Loeff, S. Goldstein, A. Treinin, H. Linschitz, J. Phys. Chem. 1991,
95, 4423–4430.
[19] The oxidation potential (Eox) of 5-methylcytosine was calculated
from the ionization potential (8.39 eV)[19a] by using the following
[33] The free energy changes of charge separation (ÀDGCS) for photoox-
equation:[19b]
E
ox =0.827ꢁ[ionization potential]À(5.20 V vs. NHE).
idation of C and mC by AQ* were estimated to be 0.02 and 0.42 eV,
3
See: a) D. M. Close, J. Phys. Chem. B 2003, 107, 864–867; b) C. J.
Burrows, J. G. Muller, Chem. Rev. 1998, 98, 1109–1151.
respectively, by using the Rehm–Weller equation,[33a] DGCS
=
À
(ET+Erdn)+Eox, in which ET is the AQ triplet energy[33b] (2.73 eV),
ACHTUNGTRENNUNG
[20] a) I. Loeff, A. Treinin, H. Linschltz, J. Phys. Chem. 1983, 87, 2536–
2544; b) J. Ma, W. Lin, W. Wang, Z. Han, S. Yao, N. Lin, Radiat.
Phys. Chem. 1999, 54, 491–497.
[21] I. Loeff, A. Treinin, H. Linschitz, J. Phys. Chem. 1984, 88, 4931–
4937.
Erdn is its reduction potential[33b] (À0.58 V vs. NHE), and Eox is the
oxidation potential of the nucleobase (C: 2.13 V; mC: 1.73 V vs.
NHE).[19] See: a) D. Rehm, A. Weller, Isr. J. Chem. 1970, 8, 259–
271; b) B. Armitage, C. Yu, C. Devadoss, G. B. Schuster, J. Am.
Chem. Soc. 1994, 116, 9847–9859.
[22] The 5-methyl-2’-deoxycytidine derivative labeled with deuterium
atoms at the C5 and C6 positions has been synthesized from dU by
deuterium exchange; see: A. Burdzy, K. T. Noyes, V. Valinluck,
L. C. Sowers, Nucleic Acids Res. 2002, 30, 4068–4074.
[23] a) J. Cadet, T. Douki, D. Gasparutto, J.-L. Ravanat, J. R. Wagner in
Radical and Radical Ion Reactivity in Nucleic Acid Chemistry (Ed.:
M. M. Greenberg), John Wiley & Sons, New Jersey, 2009, pp. 69–
97; b) J. R. Wagner, J. Cadet, Acc. Chem. Res. 2010, 43, 564–571.
[24] G. A. Russell, J. Am. Chem. Soc. 1957, 79, 3871–3877.
[25] F. M. Prado, M. C. Oliveira, S. Miyamoto, G. R. Martinez, M. H.
Medeiros, G. E. Ronsein, P. Di Mascio, Free Radical Biol. Med.
2009, 47, 401–409.
[34] Wagner, Cadet, and co-workers have shown that excited type I pho-
tosensitizers such as 2-methyl-1,4-naphthoquinone can one-electron
oxidize 2’-deoxycytidine to final oxidation products.[23] Another pos-
sibility that cannot be ruled out is that alkali-stable photooxidation
products are generated through the present AQ-photosensitized re-
action of the cytosine base.
m
+
C
[35] Although the rate of the deprotonation of
C
into a methyl-cen-
m
+
C
tered radical is still unknown, it is at least presumable that the
C
species may undergo deprotonation on a time-scale comparable to
that of positive-charge transfer through DNA bases.
[36] D. Ly, L. Sanii, G. B. Schuster, J. Am. Chem. Soc. 1999, 121, 9400–
9410.
[26] a) S. Miyamoto, G. R. Martinez, M. H. G. Medeiros, P. Di Mascio, J.
Am. Chem. Soc. 2003, 125, 6172–6179; b) S. Miyamoto, G. R. Marti-
nez, D. Rettori, O. Augusto, M. H. G. Medeiros, P. Di Mascio, Proc.
Natl. Acad. Sci. USA 2006, 103, 293–298.
[37] D. T. Breslin, G. B. Schuster, J. Am. Chem. Soc. 1996, 118, 2311–
2319.
[38] A. Defoin, R. Defoin-Straatmann, K. Hildenbrand, E. Bittersmann,
D. Kreft, H. J. Kuhn, J. Photochem. 1986, 33, 237–255.
[39] F. Wilkinson, J. Phys. Chem. 1962, 66, 2569–2574.
[40] G. J. Fisher, E. J. Land, Photochem. Photobiol. 1983, 37, 27–32.
[41] F. D. Lewis, A. K. Thazhathveetil, T. A. Zeidan, J. Vura-Weis, M. R.
Wasielewski, J. Am. Chem. Soc. 2010, 132, 444–445.
[42] N. Murata-Kamiya, H. Kamiya, N. Karino, Y. Ueno, H. Kaji, A.
Matsuda, H. Kasai, Nucleic Acids Res. 1999, 27, 4385–4390.
[43] T. Itahara, T. Yoshitake, S. Koga, A. Nishino, Bull. Chem. Soc. Jpn.
1994, 67, 2257–2264.
[27] A. Ono, A. Dan, A. Matsuda, Bioconjugate Chem. 1993, 4, 499–508.
[28] AQ-tethered ODNs were synthesized by reference to a previous
report; see: F. Bergeron, V. K. Nair, J. R. Wagner, J. Am. Chem.
Soc. 2006, 128, 14798–14799.
[29] mC in codon 282 of exon 8 in the human p53 gene is marked as a
methylation hot spot; see: S. Tornaletti, G. P. Pfeifer, Oncogene
1995, 10, 1493–1499.
[30] The AQ chromophore has a larger molar extinction coefficient at
312 nm (e312 =3605 mÀ1 cmÀ1) than at 365 nm (e365 =1116 mÀ1 cmÀ1),
so photoirradiation was carried out at 312 nm to enhance the appar-
ent efficiencies of strand cleavage at the mC residue.
[31] UV-B (290–320 nm) radiations can induce the formation of dimeric
photoproducts between adjacent pyrimidine bases in DNA.[31a]
Among the dimeric pyrimidine photooxidation products, pyrimidine
(6–4) photodimers and their Dewar valence isomers may be alkali-
[44] K. Nakatani, T. Yoshida, I. Saito, J. Am. Chem. Soc. 2002, 124,
2118–2119.
[45] T. Ito, A. Kondo, S. Terada, S. Nishimoto, J. Am. Chem. Soc. 2006,
128, 10934–10942.
Received: July 5, 2010
Published online: January 14, 2011
Chem. Eur. J. 2011, 17, 2225 – 2235
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2235