Synthesis and thermodynamic studies of oligonucleotides containing the two isomers of thymine glycol

Article abstract of DOI:10.1002/1521-3765(20011015)7:20<4343::AID-CHEM4343>3.0.CO;2-H

Thymine glycol is a major type of base damage, which is formed in DNA by reactive oxygen species. I describe the synthesis of oligonucleotides containing the 5S isomer of thymine glycol, which was not obtained by the oxidation of the oligonucleotides. Bef

Full text of DOI:10.1002/1521-3765(20011015)7:20<4343::AID-CHEM4343>3.0.CO;2-H

S. Iwai
Chem. Eur. J. 2001, 7, No. 20
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0720-4343 $ 17.50+.50/0
4343

FULL PAPER
Synthesis and Thermodynamic Studies of Oligonucleotides Containing
the Two Isomers of Thymine Glycol
Shigenori Iwai*[a, b]
Abstract: Thymine glycol is a major
type of base damage, which is formed
in DNA by reactive oxygen species. I
describe the synthesis of oligonucleoti-
des containing the 5S isomer of thymine
glycol, which was not obtained by the
oxidation of the oligonucleotides. Be-
fore the 5S isomer was synthesized, a
building block without the protection of
the tertiary hydroxy function at the C5
position of thymine glycol was tested by
the use of the previously reported 5R
isomer. In the presence of imidazole,
migration of the silyl group between the
C5 and C6 positions was observed, while
the result of the oligonucleotide syn-
thesis was identical to the case of the
fully protected building block. There-
fore, oligonucleotides containing the
(5S)-thymine glycol were synthesized
with the disilylated building block. In
contrast to the 5R derivative, two
products were detected in the HPLC
analysis of the crude mixture after
deprotection. Analysis by matrix-assist-
ed laser-desorption/ionization time-of-
flight (MALDI-TOF) mass spectrome-
try revealed that the larger peak was the
desired oligonucleotide, and it was
found that the by-product was complete-
ly degraded by a short treatment with
ammonium hydroxide at room temper-
ature. I also report the application of
oligonucleotides containing each isomer
of thymine glycol to thermodynamic
analyses of base-pair formation. The
thermodynamic parameters obtained
for the duplexes containing either the
(5R)- or (5S)-thymine glycol indicated
that the thymine glycol cannot form a
base-pair with any nucleobase, regard-
less of the configuration at the C5
position.
Keywords: chirality ´ DNA damage
´
oligonucleotides
´
solid-phase
synthesis ´ thermodynamics
glycol, namely (5R,6S), (5R,6R), (5S,6R), and (5S,6S)
(Scheme 1). In solution, however, thymine glycol exists as
either the (5R) cis ± trans pair ((5R,6S) and (5R,6R)) or the
(5S) cis ± trans pair ((5S,6R) and (5S,6S)), as a result of
epimerization at the C6 position.[11] It has been reported that
the 5R and 5S isomers are formed in equal amounts in g-
irradiated DNA,[12] whereas oxidation of thymidine or thymi-
dine-containing oligonucleotides preferentially yields (5R)-
thymine glycol.[13]
Introduction
Reactive oxygen species can damage DNA.[1] Thymine glycol
(5,6-dihydro-5,6-dihydroxythymine) is a major type of base
damage that results from the reaction of the base with a
hydroxyl radical generated by ionizing radiation[2] or as a
consequence of aerobic metabolism.[3] While it was reported
that thymine glycol lacks mutagenicity[4] or causes mutations
only at a low frequency,[5] it effectively blocks DNA repli-
cation.[6] This damage is repaired by means of the base-
excision repair pathway initiated by enzymes, such as
endonuclease III[7] and endonuclease VIII,[8] in cells. The
nucleotide-excision repair[9] and the transcription-coupled
repair[10] of this damage have also been shown.
On account of the two chiral carbon atoms at the C5 and C6
positions, there are four possible diastereomers of thymine
[a] Dr. S. Iwai
Scheme 1. Structures of the two isomers of thymine glycol.
Department of Bioorganic Chemistry
Biomolecular Engineering Research Institute
6-2-3 Furuedai, Suita, Osaka 565-0874 (Japan)
E-mail: iwai@chembio.t.u-tokyo.ac.jp
Studies on thymine glycol have been performed on chemi-
cally oxidized oligonucleotides. Although KMnO4 has fre-
quently been used as an oxidizing agent, identification and
purification of the desired product by high-performance
liquid chromatography (HPLC) are extremely difficult,
because KMnO4 oxidation gives various products, even when
[b] Dr. S. Iwai
Present address:
Department of Chemistry and Biotechnology
Graduate School of Engineering, The University of Tokyo
7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656 (Japan)
Fax : (‡81)03-5841-7340
4344
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0720-4344 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 20

Thymine Glycol Containing Oligonucleotides
4343±4351
the oligonucleotide is short and contains only one thymine
base.[5, 13b,c, 14] Alternatively, OsO4 can be used; however, the
yield of thymine glycol is very low.[5] The 5R isomer of
thymine glycol is usually obtained by this postsynthetic
oxidation method.[13b,c]
Results and Discussion
Silyl protection of the hydroxy functions of thymine glycol:
The original synthetic scheme for the (5R,6S)-thymine glycol
building block (5a), reported in a communication,[17] is shown
in Scheme 2. Thymine glycol has a secondary hydroxy
function at the C6 position, which requires protection in the
oligonucleotide synthesis. The protecting group for this
hydroxy function needs to be intact during chain assembly
and the removal of other protecting groups, and to be
removed in the final step without damaging the thymine
glycol moiety. Consequently, the tert-butyldimethylsilyl
(TBDMS) group, which is generally used for the 2'-protection
in oligoribonucleotide synthesis,[19] was chosen for this pur-
pose. The hydroxy function at the C5 position was simulta-
neously protected when an excess amount of the reagent was
used;[17] however, this tertiary hydroxy does not seem to
require protection, analogous to the (6 ± 4) photoproduct, in
which the tertiary hydroxy group was not protected for the
oligonucleotide synthesis.[18] Therefore, a phosphoramidite
building block of thymine glycol without the protection of the
tertiary hydroxy function was prepared, and its usage in the
oligonucleotide synthesis was tested.
The chemical synthesis of oligonucleotides that uses
phosphoramidite building blocks is superior to the postsyn-
thetic methods for the preparation of modified or damaged
DNA because the sequence and the chain length are not
restricted.[15] This direct incorporation method also has
advantages in large-scale preparations, especially for struc-
tural biology. Matray and Greenberg described the synthesis
of oligonucleotides containing (5R)-5,6-dihydro-5-hydroxy-
thymine, the C6-reduced form of thymine glycol.[16] I recently
reported the synthesis of a building block of (5R)-thymine
glycol and its incorporation into oligonucleotides.[17] This
study was undertaken because the chemical structure and the
alkali lability of thymine glycol were very similar to those of
the (6 ± 4) photoproduct in my previous work.[18] In this paper,
I describe the examination of the silyl protection of the
hydroxy functions of thymine glycol and the synthesis of
oligonucleotides containing the 5S isomer of this damaged
base. The oligonucleotides containing each isomer were used
for thermodynamic studies of duplex formation, in order to
gain insight into the in-vivo effect of thymine glycol.
The reaction of 5'- and 3'-protected (5R,6S)-thymidine
glycol ((5R,6S)-5,6-dihydro-5,6-dihydroxythymidine, 2a) with
Scheme 2. Reagents and conditions: i) OsO4, pyridine, room temperature, 2 h; ii) TBDMS-Cl (5 equiv), imidazole, DMF, 378C, 24 h; iii) K2CO3, MeOH,
room temperature, 2 h; iv) [(CH3)2CH]2NP(Cl)OCH2CH2CN, [(CH3)2CH]2NC2H5, THF, room temperature, 30 min.
Chem. Eur. J. 2001, 7, No. 20
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0720-4345 $ 17.50+.50/0
4345

S. Iwai
FULL PAPER
1.2 molar equivalents of tert-butyldimethylchlorosilane gave
two monosilylated compounds in the ratio of 4:1. In the
NOESY spectra, both products gave strong crosspeaks
between H6 and the methyl group at the C5 position and
between H6 and H2', which indicated that the two hydroxy
functions were in the cis orientation and that the base moiety
was in the high anti conformation around the glycosyl bond,
respectively. The difference between the two products was
that only the minor product gave crosspeaks between the
butyl proton of the TBDMS group and the aromatic and
methoxy protons of the 4,4'-dimethoxytrityl (DMT) group
protecting the 5'-hydroxy function. These results indicated
that the major and minor products were the O6- and O5-
silylated derivatives (6a and 6b in Scheme 3), respectively. It
was unusual that the tertiary hydroxy function was protected
with the TBDMS group so easily. I considered the imidazole
used in the silylation reaction as the cause of this phenom-
enon. Isolated 6a and 6b were treated with imidazole
separately, and this treatment caused isomerization of each
compound to a mixture of 6a and 6b in the ratio of 4:1, as
(8), as shown in Scheme 3, and a 13-mer, d(ACGCGATg-
ACGCCA), in which Tg represents (5R)-thymine glycol, was
synthesized. Figure 1 shows a comparison of the HPLC
chromatograms of crude mixtures of the 13-mers synthesized
1
determined by H NMR spectroscopy. The stability of the
thymine glycol against imidazole was confirmed by the same
treatment of 2a. I concluded that the TBDMS group was first
introduced to the secondary hydroxy function at the C6
position, and then migrated to the tertiary hydroxy in the
presence of imidazole. Migration of the TBDMS group has
been reported in ribonucleoside[20] and carbohydrate[21] chem-
istry. The preparation of the fully protected thymidine glycol
(3a) in my previous study[17] must have been enabled by this
migration.
Figure 1. HPLC analysis of the crude mixtures of the 13-mers synthesized
with 5a (a) and 8 (b).
with the disilylated and monosilylated building blocks (5a and
8, respectively). The 13-mer synthesized with 5a has already
been characterized,[17] and the two chromatograms are
identical to each other. This indicates that the protection of
the tertiary hydroxy function is not required for oligonucleo-
tide synthesis by the phosphoramidite chemistry. Considering
the 20% loss by the silyl migration in the preparation of the
building block, however, the disilylated building block (5a) is
recommended for the synthesis
The 3'-hydroxy function of the desired monosilylated
thymidine glycol (6a) was deprotected and phosphitylated
to give the building block for the oligonucleotide synthesis
of oligonucleotides that contain
thymine glycol.
Synthesis of oligonucleotides
containing (5S)-thymine glycol:
As described previously, (5S,6R)-
thymidine glycol (2b) was sepa-
rated on silica gel from the
(5R,6S) isomer (2a) after the
oxidation of the protected thy-
midine with OsO4.[17] Neither of
the isomers gave
a NOESY
crosspeak between H6 and H1',
and a strong NOE was detected
between H6 and H2' only in the
case of 2b. These observations
indicated the (5S,6R) configura-
tion of 2b, as described by Vaish-
nav et al.[13a] Barvian and Green-
berg have described a diastereo-
selective synthesis of (5S,6R)-
thymidine glycol,[22] but an ad-
vantage of the present method is
that both isomers can be ob-
tained at the same time.
Scheme 3. Reagents and conditions: i) TBDMS-Cl (1.2 equiv), imidazole, DMF, room temperature, 20 h;
ii) K2CO3, MeOH, room temperature, 2 h; iii) [(CH3)2CH]2NP(Cl)OCH2CH2CN, [(CH3)2CH]2NC2H5, THF,
room temperature, 30 min.
4346
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0720-4346 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 20

Thymine Glycol Containing Oligonucleotides
4343±4351
For the reason described above, the disilylated building
block (5b) was prepared to synthesize oligonucleotides
containing (5S)-thymine glycol. As shown in Scheme 2, the
procedures were the same as those for the (5R,6S) isomer. The
structure of 3b was analyzed by NOESY again. An NOE
observed between H6 and H1' showed that the conformation
around the glycosyl bond was syn. This is reasonable because
the bulky TBDMS groups must be kept away from the sugar
moiety. As expected, there was a strong crosspeak between
H6 and the methyl group at the C5 position, which indicated
the cis orientation of the two TBDMS-protected hydroxy
functions. The absence of the crosspeaks between H6 and the
methyl proton of the O5-TBDMS group and between the C5
methyl proton and the methyl proton of the O6-TBDMS
group also showed the cis orientation, which indicated that the
configuration of (5S,6R)-thymine glycol was retained. The
fully protected compound (3b) was converted to the phos-
phoramidite building block (5b), as shown in Scheme 2.
From 5b, an 11-mer, d(CGTACTg*CATGC), in which Tg*
represents (5S)-thymine glycol (since the deprotection of
thymine glycol results in the epimerization at the C6
position,[11, 22] I describe the stereochemistry only at the C5
position in the case of deprotected thymine glycol, and (5S)-
thymine glycol means a mixture of the (5S,6R) and (5S,6S)
isomers, which are not separated by HPLC), and a 30-mer,
d(CTCGTCAGCATCTTg*CATCATACAGTCAGTG),
were synthesized. The procedures were the same as those for
the oligonucleotides containing the (5R)-thymine glycol.[17]
For comparison, the Tg 11-mer and the Tg 30-mer, containing
(5R)-thymine glycol, were synthesized. After deprotection
with ammonium hydroxide at room temperature, followed by
treatment with tetrabutylammonium fluoride (TBAF), the
crude mixtures of the Tg and Tg* 11-mers were analyzed by
reversed-phase HPLC. While the Tg 11-mer gave only a single
main peak, two large peaks (peaks 1 and 2 in Figure 2) were
obtained for the Tg* 11-mer. The ratio of the two products
changed, depending on the deprotection methods, as shown in
Figure 2. To designate the desired product, the two products
were isolated by HPLC and analyzed by matrix-assisted laser
desorption/ionization time-of-flight (MALDI-TOF) mass
spectrometry. The results are listed in Table 1. This analysis
demonstrated that peak 1 was the desired 11-mer containing
(5S)-thymine glycol. The smaller m/z value obtained for
peak 2 suggested that this byproduct contained 5-hydroxy-5-
methylhydantoin (HMH) instead of thymine glycol, although
the (5S)-thymine glycol in the oligonucleotide was not
converted to HMH by the procedure described for the
released thymine glycol.[23] The difference in the reactivity of
the (5R)- and (5S)-thymine glycols during deprotection may
be attributed to the difference in the spatial arrangement of
their functional groups. Since an analysis of the products by
mass spectrometry is not always possible, I searched for a
simple method to discriminate between the desired product
containing (5S)-thymine glycol and the by-product. Treatment
with ammonium hydroxide at room temperature for 2 h
resulted in the complete conversion of the by-product to a
new product, whereas most of the thymine glycol-containing
oligonucleotide remained intact after the same treatment
(Figure 3). The new product generated from the by-product
Figure 2. HPLC analysis of the crude mixtures of the Tg 11-mer (a) and
the Tg* 11-mer (b, c, and d). The Tg* 11-mer was deprotected with NH4OH
(b), NH3/MeOH (c), or K2CO3/MeOH (d), and the two products (peaks 1
and 2) were isolated for characterization by MALDI-TOF mass spectrom-
etry and by the ammonia treatment.
Table 1. MALDI-TOF MS analysis of the 11-mers.
Oligonucleotide
Sequence
m/z
observed theoretical[a]
T 11-mer
Tg 11-mer
d(CGTACTCATGC)
d(CGTACTgCATGC)
3287.6
3323.2
3289.6
3323.6
3323.6
3293.6[b]
Tg* 11-mer peak 1 d(CGTACTg*CATGC) 3323.2
peak 2 3295.5
[a] [M H] . [b] An 11-mer containing 5-hydroxy-5-methylhydantoin
Figure 3. Ammonia treatment of the deprotected oligonucleotides. The Tg
11-mer (a) and the two products of the Tg* 11-mer (b, peak 1; c, peak 2)
were treated with 28% ammonia water at room temperature for 2 h and
analyzed by reversed-phase HPLC. The elution profiles of the untreated
oligonucleotides are shown by broken lines.
Chem. Eur. J. 2001, 7, No. 20
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0720-4347 $ 17.50+.50/0
4347

S. Iwai
FULL PAPER
by this treatment was supposed to be an 11-mer in which the
base moiety was decomposed to urea.
A similar result was obtained for the 30-mer (Figure 4).
Since the smaller peak was completely shifted to the other
side of the larger peak by the ammonia treatment, the larger
peak was isolated as the desired product. The presence of
(5S)-thymine glycol in the 30-mer was demonstrated by the
analysis of the nucleoside composition,[14b] as shown in
Figure 5. This analysis also showed that epimerization at the
C5 position of thymine glycol did not occur during the
synthesis and the deprotection of these oligonucleotides.
Thermodynamic analysis of duplex formation: Thermody-
namic analyses are useful to gain insight into the mechanisms
of the damage-induced mutations and the damage recognition
of repair enzymes.[24] The oligonucleotides containing each
isomer of thymine glycol, which were synthesized as described
above, were used for the thermodynamic analyses. Two types
of duplexes were designed, as shown in Table 2. The target
base-pairs are usually placed in the center of a duplex in
thermodynamic studies,[24a] and the first set, d(CGTACX-
CATGC) d(GCATGAGTACG), can be used to yield clues
about the recognition mecha-
Figure 4. HPLC analysis of the crude (a and b) and purified (c) Tg* 30-
mer. Deprotection was performed with NH4OH (a) or K2CO3/MeOH (b).
nisms of the repair enzymes.
The second set, in which the
damaged base is located at the
5' end of the double-stranded
region, is a model for replica-
tion,[25] and the four normal
bases were placed opposite the
undamaged thymine or each
isomer of the thymine glycol,
to study their base-pair forma-
tion. Although NMR studies
Figure 5. Detection of thymidine glycol after enzyme digestion. a) and c) The Tg and Tg* 30-mers were treated
with nuclease P1, phosphodiesterase I, and alkaline phosphatase, and the products were separated by HPLC.
Thymidine glycol in each sample is indicated by an arrow. b) and d) The authentic (5R)- and (5S)-thymidine glycols
were added to the Tg 30-mer and Tg* 30-mer mixtures, respectively, to show the identity. e) Separation of the (5R)-
and (5S)-thymidine glycol was demonstrated by co-injection. The monitoring wavelength was 230 nm in all cases.
have been reported,[13b,c] the
duplex destabilization cannot
be quantified from the struc-
ture, and the latter type of
duplexes cannot be analyzed
by NMR spectroscopy.
Table 2. Thermodynamic parameters of duplex formation.
[a]
Duplex
Tm
DH8
[kcalmol
DS8
[calmol 1 K
DG8(258C)
DDG8[b] (258C)
After hybridization of the
two strands, thermal melting
curves, monitored at l ˆ
260 nm, were measured at total
oligonucleotide concentrations
(Ct) varying between 2.0 and
20.0 mm. Thermodynamic pa-
rameters for each duplex were
determined from the Tm data
by the vanꢁt Hoff method, as-
suming a two-state model for
duplex melting (Figure 6).[26]
The results, as well as the Tm
values at Ct ˆ 2.0 mm, are listed
in Table 2. When thymine gly-
col was placed in the center of
the 11-mer duplex, a large
destabilization was observed
compared to the undamaged
1
1
1
1
[8C]
]
]
[kcalmol
]
[kcalmol
]
5'-CGTACXCATGC-3'
3'-GCATGAGTACG-5'
X ˆ T
44.0
24.7
25.1
79.2 Æ 3.1
67.7 Æ 0.1
68.3 Æ 0.6
221 Æ 10
198 Æ 0
200 Æ 2
13.31 Æ 0.20
8.52 Æ 0.00
8.60 Æ 0.01
X ˆ Tg
4.79 Æ 0.20
4.71 Æ 0.21
X ˆ Tg*
5'-CAXAGCACGAC-3'
3'-
YTCGTGCTG-5'
X ˆ T, Yˆ A
X ˆ T, Yˆ G
X ˆ T, Yˆ C
39.8
38.1
36.0
37.0
37.9
38.7
36.4
37.4
37.7
38.6
36.4
37.1
67.3 Æ 2.4
63.3 Æ 2.3
58.7 Æ 0.7
62.2 Æ 0.4
61.6 Æ 2.2
61.5 Æ 1.5
58.3 Æ 1.9
60.4 Æ 1.7
57.5 Æ 0.9
62.9 Æ 1.5
59.2 Æ 2.1
60.3 Æ 1.2
186 Æ 8
175 Æ 7
161 Æ 2
172 Æ 1
169 Æ 7
169 Æ 5
159 Æ 6
166 Æ 6
156 Æ 3
173 Æ 5
162 Æ 7
166 Æ 4
11.76 Æ 0.13
11.21 Æ 0.13
10.72 Æ 0.04
11.01 Æ 0.02
11.13 Æ 0.10
11.28 Æ 0.07
10.78 Æ 0.10
10.99 Æ 0.08
10.94 Æ 0.04
11.36 Æ 0.08
10.81 Æ 0.10
10.94 Æ 0.05
X ˆ T, Yˆ T
X ˆ Tg, Yˆ A
X ˆ Tg, Yˆ G
X ˆ Tg, Yˆ C
X ˆ Tg, Yˆ T
X ˆ Tg*, Yˆ A
X ˆ Tg*, Yˆ G
X ˆ Tg*, Yˆ C
X ˆ Tg*, Yˆ T
0.63 Æ 0.23
0.07 Æ 0.20
0.06 Æ 0.14
0.02 Æ 0.10
0.82 Æ 0.17
0.15 Æ 0.21
0.09 Æ 0.14
0.07 Æ 0.07
[a] Ct ˆ 2.0 mm. [b] DG8(Tg) DG8(T)
4348
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0720-4348 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 20

Thymine Glycol Containing Oligonucleotides
4343±4351
observed between thymine glycol and any of the four
normal bases when thymine glycol was located at the
end of the duplex. It is possible that this lack of base-
pairing ability, together with the nonplanar structure of
thymine glycol, causes the replication block. It may be
noteworthy that the thermodynamic characteristics, as
well as the reduced thermal stability (DTm), of the
thymine glycol-containing duplexes obtained in this
study are quite similar to those reported for duplexes
containing 3-hydroxy-2-(hydroxymethyl)tetrahydrofur-
an, that is an abasic site analogue.[25b, 27]
Experimental Section
General methods: All solvents and reagents were obtained from
Wako Pure Chemical Industries (Osaka, Japan), except for tert-
butyldimethylchlorosilane (Shin-Etsu Chemical, Tokyo, Japan), (2-
cyanoethyl)-N,N-diisopropylchlorophosphoramidite (Digital Spe-
cialty Chemicals, Ontario, Canada), and TBAF (Tokyo Chemical
Industry, Tokyo, Japan). Reagents for the DNA synthesizer were
purchased from Applied Biosystems Japan (Tokyo, Japan). Nucle-
ase P1 and alkaline phosphatase were purchased from Roche
Diagnostics (Mannheim, Germany), and phosphodiesterase I from
Worthington Biochemical (Lakewood, NJ, USA).
Figure 6. Plots of 1/Tm versus lnCt/4 for the duplexes containing thymine glycol.
~
*
a) d(CGTACXCATGC) d(GCATGAGTACG) where X ˆ T ( ), Tg ( ), and Tg*
TLC analyses were carried out on Merck silica gel 60F254 plates,
which were visualized by UV illumination at l ˆ 254 nm and
spraying of an anisaldehyde/sulfuric acid solution, followed by
heating. For column chromatography, either Wakogel C-200 or
C-300 (Wako Pure Chemical Industries) was used. 1H NMR spectra
were measured on a Bruker DPX300 spectrometer, with tetramethylsilane
as the internal standard. COSY and NOESY spectra were used for the
signal assignment and the configuration determination, respectively. 31P
NMR spectra were measured on the same spectrometer at 121.5 MHz, with
trimethyl phosphate as the internal standard. Mass spectra were obtained
on a JEOL HX-110, VG ZAB-HF, or Hitachi M-4000H spectrometer.
HPLC analyses were carried out on a Gilson gradient-type analytical
system equipped with a Waters 996 photodiode array detector. A mBon-
dasphere C18 5 mm 300 Š column (3.9  150 mm, Waters Corporation,
Milford, MA, USA) was used with a linear gradient of acetonitrile (0 ±
10%, 5 ± 13%, and 7 ± 13% for the 11-, 13-, and 30-mers, respectively) in
0.1m triethylammonium acetate (TEAA, pH 7.0). Anion-exchange HPLC
was performed on a TSK-GEL DEAE-2SW column (4.6 Â 250 mm, Tosoh
Corporation, Tokyo, Japan), with a linear gradient of ammonium formate
(0.3 ± 0.8m and 0.4 ± 1.0m for the 11- and 30-mers, respectively) in 20%
acetonitrile. MALDI-TOF mass analyses of purified oligonucleotides were
carried out on a PerSeptive Biosystems Voyager Elite spectrometer, with a
mixture of 3-hydroxypicolinic acid and picolinic acid (9:1, w/w) as a matrix.
!
(
). b), c), and d) d(CAXAGCACGAC) d(GTCGTGCTY) where X ˆ T (b), Tg
~ !
&
*
(c), and Tg* (d); Yˆ A ( ), G ( ), C ( ), and T ( ).
duplex, regardless of the C5 configuration of the oxidized
base. In the replication model, in which thymine glycol was
located at the end of the double-stranded region, the differ-
ence in the free energy (DG8) for the duplexes with the
correct and mismatched base-pairs was small, as reported
previously,[25] but the correct T± A pair distinctly showed
1
stabilization of the duplex (DDG8 ˆ 0.55 to 1.04 kcalmol at
258C, compared to the mismatches). When thymine was
replaced by thymine glycol, the absolute values of DG8
obtained for the duplexes containing the thymine glycol ±
adenine pair were reduced to the level of the mismatched
pairs, while those for the other duplexes were not altered. No
obvious difference was observed between the 5R and 5S
isomers of thymine glycol.
As shown by the thermodynamic analysis of the first set of
the duplexes, both (5R)- and (5S)-thymine glycols incorpo-
rated into the center of the duplex greatly reduced the
thermodynamic stability. This observation may seem strange
because the functional groups for base-pair formation are
apparently intact after the oxidation of the thymine base.
However, this result supports the structural features of (5R)-
thymine glycol-containing duplexes studied by NMR spectro-
scopy. In the NMR studies by Bolton and co-workers,[13b,c]
thymine glycol induced a highly localized alteration in the
DNA structure, and the oxidized base as well as the opposite
adenine were reported to be extrahelical. The reduced
transition entropy change (DS8) measured in the present
study for the duplexes containing the Tg ± A and Tg* ± A pairs
may reflect the flexibility of the extrahelical bases. As
described previously,[13b,c] this structural feature may be
recognized by repair proteins. Base-pair formation was not
Compounds containing the (5R,6S)-thymine glycol (2a, 3a, 4a, and 5a)
were prepared as described previously.[17]
5'-O-(4,4'-Dimethoxytrityl)-3'-O-benzoyl-(5S,6R)-5,6-dihydro-5,6-dihy-
droxythymidine (2b): Protected thymidine (1) was oxidized with OsO4 in
pyridine, and 2b was separated from 2a on silica gel, as described
1
previously.[17] Rf ˆ 0.62 (CHCl3/MeOH 10/1); H NMR (300 MHz, CDCl3,
258C, TMS): d ˆ 8.0 (d, 3J(H,H) ˆ 7 Hz, 2H; Bz), 7.9 (s, 1H; NH), 7.6
(t, 3J(H,H) ˆ 7 Hz, 1H; Bz), 7.5 ± 7.2 (m, 11H; Bz, DMT), 6.8 (d, 3J(H,H) ˆ
9 Hz, 4H; DMT), 6.5 (t, 3J(H,H) ˆ 7 Hz, 1H; H1'), 5.7 ± 5.6 (m, 1H; H3'),
5.0 (s, 1H; H6), 4.3 ± 4.2 (m, 1H; H4'), 3.8 (s, 6H; OCH3), 3.7 (br, 1H;
3
3
5-OH), 3.5 (dd, J(H,H) ˆ 11, 4 Hz, 1H; H5'), 3.4 (dd, J(H,H) ˆ 11, 3 Hz,
1H; H5'), 3.3 (br, 1H; 6-OH), 2.5 ± 2.4 (m, 2H; H2', H2''), 1.4 (s, 3H; CH3);
HRMS: (FAB): found: 681.2462 [M H] ; calcd: 681.2448.
5'-O-(4,4'-Dimethoxytrityl)-3'-O-benzoyl-(5S,6R)-5,6-dihydro-5,6-di[(tert-
butyl)dimethylsilyloxy]thymidine (3b):
A solution of 5'-O-(4,4'-di-
methoxytrityl)-3'-O-benzoyl-(5S,6R)-5,6-dihydro-5,6-dihydroxythymidine
(2b, 299 mg, 438 mmol), imidazole (298 mg, 4.38 mmol), and tert-butyldi-
methylchlorosilane (330 mg, 2.19 mmol) in N,N-dimethylformamide
(DMF, 2.5 mL) was stirred at 378C for 24 h. This mixture was diluted with
Chem. Eur. J. 2001, 7, No. 20
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0720-4349 $ 17.50+.50/0
4349

S. Iwai
FULL PAPER
chloroform (20 mL) and washed with 0.5m sodium phosphate (pH 5.0). The
organic layer was dried with Na2SO4, and after evaporation and coevapo-
ration with toluene, the residue was chromatographed on silica gel (10 g)
with a step gradient of ethyl acetate in hexane. The product was eluted with
15% ethyl acetate in hexane. Yield: 334 mg (366 mmol, 84%). Rf ˆ 0.71
6a: Yield: 988 mg (1.24 mmol, 42%); Rf ˆ 0.45 (hexane/ethyl acetate 3/2);
3
1H NMR (300 MHz, CDCl3, 258C, TMS): d ˆ 8.0 (d, J(H,H) ˆ 7 Hz, 2H;
3
Bz), 7.8 (s, 1H; NH), 7.6 (t, J(H,H) ˆ 7 Hz, 1H; Bz), 7.5 ± 7.4 (m, 4H; Bz,
DMT), 7.4 ± 7.2 (m, 7H; DMT), 6.8 (d, 3J(H,H) ˆ 9 Hz, 4H; DMT), 6.2 (dd,
3J(H,H) ˆ 9, 6 Hz, 1H; H1'), 5.6 ± 5.5 (m, 1H; H3'), 4.9 (s, 1H; H6), 4.2 (dd,
3J(H,H) ˆ 8, 5 Hz, 1H; H4'), 3.8 (s, 6H; OCH3), 3.4 (dd, 3J(H,H) ˆ 10, 5 Hz,
1H; H5'), 3.3 (dd, 3J(H,H) ˆ 10, 5 Hz, 1H; H5'), 3.3 (s, 1H; -OH), 2.4 ± 2.2
(m, 2H; H2', H2''), 1.4 (s, 3H; CH3), 0.8 (s, 9H; TBDMS), 0.1 (s, 3H;
TBDMS), 0.0 (s, 3H; TBDMS); HRMS: (FAB): found: 795.3311 [M H] ;
calcd: 795.3313.
1
(hexane/ethyl acetate 3/2); H NMR (300 MHz, CDCl3, 258C, TMS): d ˆ
8.0 (d, 3J(H,H) ˆ 7 Hz, 2H; Bz), 7.6 (t, 3J(H,H) ˆ 7 Hz, 1H; Bz), 7.5 ± 7.4 (m,
4H; Bz, DMT), 7.4 ± 7.2 (m, 7H; DMT), 7.1 (s, 1H; NH), 6.8 (d, 3J(H,H) ˆ
3
9 Hz, 4H; DMT), 5.7 (dd, J(H,H) ˆ 8, 5 Hz, 1H; H1'), 5.6 (m, 1H; H3'),
4.9 (s, 1H; H6), 4.3 (dd, 3J(H,H) ˆ 8, 5 Hz, 1H; H4'), 3.8 (s, 6H; OCH3), 3.5
3
(dd, J(H,H) ˆ 10, 5 Hz, 1H; H5'), 3.3 (dd, 3J(H,H) ˆ 10, 5 Hz, 1H; H5'),
6b: Yield: 228 mg (0.29 mmol, 10%); Rf ˆ 0.51 (hexane/ethyl acetate 3/2);
1H NMR (300 MHz, CDCl3, 258C, TMS): d ˆ 8.1 (d, J(H,H) ˆ 8 Hz, 2H;
2.7 ± 2.6 (m, 1H; H2'), 2.5 ± 2.4 (m, 1H; H2''), 1.3 (s, 3H; CH3), 0.9, 0.8 (s,
18H; TBDMS), 0.2 (s, 6H; TBDMS), 0.1, 0.0 (s, 6H; TBDMS); HRMS:
(FAB): found: 909.4149 [M H] ; calcd: 909.4178.
3
3
Bz), 7.7 (s, 1H; NH), 7.6 (t, J(H,H) ˆ 7 Hz, 1H; Bz), 7.5 ± 7.4 (m, 4H; Bz,
DMT), 7.4 ± 7.2 (m, 7H; DMT), 6.8 (d, 3J(H,H) ˆ 9 Hz, 4H; DMT), 6.4 (dd,
3J(H,H) ˆ 10, 5 Hz, 1H; H1'), 5.7 (d, 3J(H,H) ˆ 6 Hz, 1H; H3'), 5.1 (s, 1H;
H6), 4.2 (br, 1H; H4'), 3.8 (s, 6H; OCH3), 3.5 (dd, 3J(H,H) ˆ 10, 3 Hz, 1H;
H5'), 3.4 (dd, 3J(H,H) ˆ 10, 3 Hz, 1H; H5'), 3.1 (s, 1H; -OH), 2.8 ± 2.7 (m,
5'-O-(4,4'-Dimethoxytrityl)-(5S,6R)-5,6-dihydro-5,6-di[(tert-butyl)dimeth-
ylsilyloxy]thymidine
(5S,6R)-5,6-dihydro-5,6-di[(tert-butyl)dimethylsilyloxy]thymidine
(4b):
5'-O-(4,4'-Dimethoxytrityl)-3'-O-benzoyl-
(3b,
3
1H; H2'), 2.5 (dd, J(H,H) ˆ 14, 5 Hz, 1H; H2''), 1.5 (s, 3H; CH3), 0.7 (s,
322 mg, 353 mmol) was dissolved in a 50 mm solution of K2CO3 in methanol
(15 mL). After 2 h, sodium phosphate (0.5m, pH 5.0, 20 mL) was added.
The mixture was extracted with chloroform (30 mL in total). The organic
layer was dried with Na2SO4, and after evaporation, the residue was
chromatographed on silica gel (8 g) with a step gradient of ethyl acetate in
hexane. The product was eluted with 25 ± 30% ethyl acetate in hexane.
Yield: 286 mg (354 mmol, 100%); Rf ˆ 0.37 (hexane/ethyl acetate 3/2);
9H; TBDMS), 0.2 (s, 3H; TBDMS), 0.1 (s, 3H; TBDMS); HRMS: (FAB):
found: 795.3319 [M H] ; calcd: 795.3313.
5'-O-(4,4'-Dimethoxytrityl)-(5R,6S)-5,6-dihydro-5-hydroxy-6-(tert-butyl)-
dimethylsilyloxythymidine (7): 5'-O-(4,4'-Dimethoxytrityl)-3'-O-benzoyl-
(5R,6S)-5,6-dihydro-5-hydroxy-6-(tert-butyl)dimethylsilyloxythymidine
(6a, 988 mg, 1.24 mmol) was dissolved in a 50 mm solution of K2CO3 in
methanol (50 mL). This mixture was stirred for 2 h, and after the solution
was cooled in an ice bath, sodium phosphate (0.5m, pH 5.0, 100 mL) was
added. The mixture was extracted with chloroform (150 mL in total). The
organic layer was dried with Na2SO4, and after evaporation, the residue
was chromatographed on silica gel (20 g) with a step gradient of ethyl
acetate in hexane. The product was eluted with 30 ± 35% ethyl acetate in
hexane and obtained as a powder by precipitation in hexane (60 mL) from
a chloroform solution (3 mL). Yield: 626 mg (904 mmol, 73%); Rf ˆ 0.22
3
1H NMR (300 MHz, CDCl3, 258C, TMS): d ˆ 7.4 (d, J(H,H) ˆ 7 Hz, 2H;
DMT), 7.3 ± 7.2 (m, 7H; DMT), 7.0 (s, 1H; NH), 6.8 (d, 3J(H,H) ˆ 9 Hz, 4H;
DMT), 5.5 (t, 3J(H,H) ˆ 6 Hz, 1H; H1'), 4.8 (s, 1H; H6), 4.5 ± 4.4 (m, 1H;
H3'), 3.9 (dd, 3J(H,H) ˆ 11, 5 Hz, 1H; H4'), 3.8 (s, 6H; OCH3), 3.4 (dd,
3
3J(H,H) ˆ 10, 5 Hz, 1H; H5'), 3.2 (dd, J(H,H) ˆ 10, 6 Hz, 1H; H5'), 2.5 ±
2.3 (m, 2H; H2', H2''), 1.9 (d, 3J(H,H) ˆ 3 Hz, 1H; 3'-OH), 1.3 (s, 3H; CH3),
0.9, 0.8 (s, 18H; TBDMS), 0.2 (s, 6H; TBDMS), 0.1, 0.0 (s, 6H; TBDMS);
HRMS: (FAB): found: 805.3933 [M H] ; calcd: 805.3916.
1
(hexane/ethyl acetate 3/2); H NMR (300 MHz, CDCl3, 258C, TMS): d ˆ
7.4 (d, 3J(H,H) ˆ 7 Hz, 2H; DMT), 7.4 (s, 1H; NH), 7.3 ± 7.2 (m, 7H; DMT),
6.8 (d, 3J(H,H) ˆ 9 Hz, 4H; DMT), 6.1 (dd, 3J(H,H) ˆ 8, 7 Hz, 1H; H1'), 4.8
(s, 1H; H6), 4.4 (br, 1H; H3'), 3.9 (dd, 3J(H,H) ˆ 9, 5 Hz, 1H; H4'), 3.8 (s,
6H; OCH3), 3.4 (dd, 3J(H,H) ˆ 10, 5 Hz, 1H; H5'), 3.3 (s, 1H; -OH), 3.2
(dd, 3J(H,H) ˆ 10, 6 Hz, 1H; H5'), 2.3 ± 2.1 (m, 2H; H2', H2''), 1.9 (br, 1H;
3'-OH), 1.4 (s, 3H; CH3), 0.8 (s, 9H; TBDMS), 0.1 (s, 3H; TBDMS), 0.0 (s,
3H; TBDMS); HRMS: (FAB): found: 691.3063 [M H] ; calcd: 691.3051.
5'-O-(4,4'-Dimethoxytrityl)-(5S,6R)-5,6-dihydro-5,6-di[(tert-butyl)dimeth-
ylsilyloxy]thymidine 3'-(2-cyanoethyl)-N,N-diisopropylphosphoramidite
(5b): N,N-Diisopropylethylamine (242 mL, 1.39 mmol) and (2-cyanoeth-
yl)-N,N-diisopropylchlorophosphoramidite (155 mL, 696 mmol) were added
to a solution of 5'-O-(4,4'-dimethoxytrityl)-(5S,6R)-5,6-dihydro-5,6-di[(tert-
butyl)dimethylsilyloxy]thymidine (4b, 281 mg, 348 mmol) in tetrahydrofur-
an (THF, 3.5 mL). This mixture was stirred for 30 min, diluted with ethyl
acetate, and washed with 2% NaHCO3 and water. The organic layer was
dried with Na2SO4, and after evaporation, the residue was chromato-
graphed on silica gel (7 g) with a step gradient of ethyl acetate in hexane
containing 0.1% pyridine. The product was eluted with 15% ethyl acetate
in hexane, and after evaporation, the pyridine was removed by coevapora-
tion with acetonitrile. Yield: 215 mg (214 mmol, 61%); Rf ˆ 0.66, 0.58
5'-O-(4,4'-Dimethoxytrityl)-(5R,6S)-5,6-dihydro-5-hydroxy-6-(tert-butyl)-
dimethylsilyloxythymidine
3'-(2-cyanoethyl)-N,N-diisopropylphosphor-
amidite (8): N,N-Diisopropylethylamine (105 mL, 600 mmol) and (2-cyano-
ethyl)-N,N-diisopropylchlorophosphoramidite (67 mL, 300 mmol) were
added to a solution of 5'-O-(4,4'-dimethoxytrityl)-(5R,6S)-5,6-dihydro-5-
hydroxy-6-(tert-butyl)dimethylsilyloxythymidine (7, 104 mg, 150 mmol) in
THF (1.5 mL). This mixture was stirred for 30 min, diluted with ethyl
acetate, and washed with 2% NaHCO3 and water. The organic layer was
dried with Na2SO4, and after evaporation, the residue was chromato-
graphed on silica gel (5 g) with a step gradient of ethyl acetate in hexane
containing 0.1% pyridine. The product was eluted with 30% ethyl acetate
in hexane, and after evaporation, the pyridine was removed by precip-
itation of the amidite in pentane (10 mL) from a chloroform solution
(0.5 mL). Yield: 116 mg (130 mmol, 87%); Rf ˆ 0.34 (hexane/ethyl acetate
3/2); 1H NMR (300 MHz, CDCl3, 258C, TMS): d ˆ 7.5 ± 7.2 (m, 10H; DMT,
NH), 6.9 ± 6.8 (m, 4H; DMT), 6.2 ± 6.1 (m, 1H; H1'), 4.9, 4.8 (s, 1H; H6),
4.5 ± 4.4 (m, 1H; H3'), 4.1 ± 4.0 (m, 1H; H4'), 3.8 ± 3.7 (m, 7H; OCH3,
1
(hexane/ethyl acetate 3/2); H NMR (300 MHz, CDCl3, 258C, TMS): d ˆ
7.9 (s, 1H; NH), 7.5 ± 7.4 (m, 2H; DMT), 7.4 ± 7.2 (m, 7H; DMT), 6.9 ± 6.8
(m, 4H; DMT), 5.6 ± 5.5 (m, 1H; H1'), 4.9, 4.8 (s, 1H; H6), 4.6 ± 4.5 (m, 1H;
1
H3'), 4.2 ± 4.1 (m, 1H; H4'), 3.9 ± 3.7 (m, 7H; OCH3, OCH2CH2CN Â ꢂ2),
1
3.7 ± 3.5 (m, 3H; OCH2CH2CN Â ꢂ2, CH(CH3)2), 3.3 ± 3.2 (m, 2H; H5'), 2.6
3
(t, J(H,H) ˆ 6 Hz, 1H; OCH2CH2CN  ꢂ2), 2.4 (t, 3J(H,H) ˆ 6 Hz, 1H;
1
1
OCH2CH2CN Â ꢂ2), 2.6 ± 2.3 (m, 2H; H2', H2''), 1.4, (s, 3H; CH3), 1.2 ± 1.1
(m, 12H; CH(CH3)2), 0.8 (s, 18H; TBDMS), 0.2 (s, 6H; TBDMS), 0.1, 0.0
(s, 6H; TBDMS); 31P NMR (121.5 MHz, CDCl3, 258C, trimethyl phos-
phate): d ˆ 147, 146; HRMS: (SI): found: 1005.4992 [M H] ; calcd:
1005.4990.
1
1
OCH2CH2CN Â ꢂ2), 3.7 ± 3.5 (m, 3H; OCH2CH2CN Â ꢂ2, CH(CH3)2), 3.3 ±
5'-O-(4,4'-Dimethoxytrityl)-3'-O-benzoyl-(5R,6S)-5,6-dihydro-5-hydroxy-
6-(tert-butyl)dimethylsilyloxythymidine (6a) and 5'-O-(4,4'-dimethoxytri-
tyl)-3'-O-benzoyl-(5R,6S)-5,6-dihydro-5-(tert-butyl)dimethylsilyloxy-6-hy-
droxythymidine (6b): A solution of 5'-O-(4,4'-dimethoxytrityl)-3'-O-ben-
zoyl-(5R,6S)-5,6-dihydro-5,6-dihydroxythymidine (2a, 2.03 g, 2.97 mmol),
imidazole (485 mg, 7.12 mmol), and tert-butyldimethylchlorosilane
(537 mg, 3.56 mmol) in DMF (10 mL) was stirred at room temperature
for 20 h. This mixture was diluted with chloroform (100 mL) and washed
with 0.5m sodium phosphate (pH 5.0). The organic layer was dried with
Na2SO4, and after evaporation and coevaporation with toluene, the residue
was chromatographed on silica gel (50 g) with a step gradient of ethyl
acetate in hexane. The two isomers were separately eluted out at 25% ethyl
acetate, and the configurations were determined by the NOESY experi-
ments.
3
1
3.1 (m, 3H; -OH, H5'), 2.6 (t, J(H,H) ˆ 7 Hz, 1H; OCH2CH2CN  ꢂ2), 2.4
3
1
(t, J(H,H) ˆ 7 Hz, 1H; OCH2CH2CN  ꢂ2), 2.3 ± 2.0 (m, 2H; H2', H2''), 1.4
(s, 3H; CH3), 1.2 ± 1.0 (m, 12H; CH(CH3)2), 0.8 (s, 9H; TBDMS), 0.0 (s,
6H; TBDMS); 31P NMR (121.5 MHz, CDCl3, 258C, trimethyl phosphate):
d ˆ 147, 146; HRMS: (FAB): found: 891.4121 [M H] ; calcd: 891.4129.
Oligonucleotide synthesis: Oligonucleotides were synthesized on either a
0.2 or 1.0 mmol scale on an Applied Biosystems Model 394 DNA/RNA
synthesizer. The thymine glycol building blocks (5a, 5b, and 8) were
dissolved in anhydrous acetonitrile at a concentration of 0.1m. Phosphor-
amidites bearing the (4-tert-butylphenoxy)acetyl group for the protection
of the exocyclic amino function were used for dA, dG, and dC. The reaction
time for the coupling of the thymine glycol building blocks was prolonged
to 5 min. After chain assembly and removal of the terminal 5'-DMT group
4350
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0720-4350 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 20

Thymine Glycol Containing Oligonucleotides
4343±4351
on the synthesizer, the solid supports containing the oligonucleotides were
treated with 28% aqueous ammonia (2 mL) at room temperature for 2 h.
The resulting ammoniac solutions were concentrated to dryness on a rotary
evaporator equipped with a vacuum pump. In the case of the (5S)-thymine
glycol, the solid supports containing the oligonucleotides were treated with
50 mm K2CO3 in methanol (1 mL) at room temperature for 4 h.[28] The
supports were removed by filtration, and the solutions were mixed with
sodium phosphate (0.5m, pH 5.0, 1 mL). After concentration, these
mixtures were passed through a NAP10 column (Amersham Pharmacia
Biotech, Uppsala, Sweden), and the eluates were concentrated to dryness.
The residues were dissolved in a 1.0m solution of TBAF in THF (0.1 mL),
and the mixtures were left at room temperature. After 16 h, TEAA (0.1m,
pH 7.0, 0.4 mL) was added, and the solutions were desalted on a NAP10
column. For the deprotection of oligonucleotides synthesized on the
1.0 mmol scale, larger volumes of the reagents and NAP25 columns were
used. The deprotected oligonucleotides were purified by both reversed-
phase and anion-exchange HPLC.
[7] a) B. Demple, S. Linn, Nature 1980, 287, 203 ± 208; b) M. Dizdaroglu, J.
Laval, S. Boiteux, Biochemistry 1993, 32, 12105 ± 12111; c) R. Aspin-
wall, D. G. Rothwell, T. Roldan-Arjona, C. Anselmino, C. J. Ward, J. P.
Cheadle, J. R. Sampson, T. Lindahl, P. C. Harris, I. D. Hickson, Proc.
Natl. Acad. Sci. USA 1997, 94, 109 ± 114; d) T. P. Hilbert, W. Chaung,
R. J. Boorstein, R. P. Cunningham, G. W. Teebor, J. Biol. Chem. 1997,
272, 6733 ± 6740; e) S. Ikeda, T. Biswas, R. Roy, T. Izumi, I. Boldogh,
A. Kurosky, A. H. Sarker, S. Seki, S. Mitra, J. Biol. Chem. 1998, 273,
21585 ± 21593.
[8] R. J. Melamede, Z. Hatahet, Y. W. Kow, H. Ide, S. S. Wallace,
Biochemistry 1994, 33, 1255 ± 1264.
[9] a) J.-J. Lin, A. Sancar, Biochemistry 1989, 28, 7979 ± 7984; b) Y. W.
Kow, S. S. Wallace, B. Van Houten, Mutat. Res. 1990, 235, 147 ± 156;
c) J. T. Reardon, T. Bessho, H. C. Kung, P. H. Bolton, A. Sancar, Proc.
Natl. Acad. Sci. USA 1997, 94, 9463 ± 9468.
[10] S. A. Leadon, S. L. Barbee, A. B. Dunn, Mutat. Res. 1995, 337, 169 ±
178.
[11] M. J. Lustig, J. Cadet, R. J. Boorstein, G. W. Teebor, Nucleic Acids Res.
1992, 20, 4839 ± 4845.
[12] G. Teebor, A. Cummings, K. Frenkel, A. Shaw, L. Voituriez, J. Cadet,
Free Rad. Res. Commun. 1987, 2, 303 ± 309.
Detection of thymidine glycol in oligonucleotides: Aliquots (0.5 A260 units)
of the purified 30-mers were mixed with nuclease P1 (40 mg) in ammonium
acetate (30 mm, pH 5.3, 400 mL) and incubated at 378C. After 20 h, water
(466 mL), Tris-HCl (0.5m, pH 9.0, 100 mL), MgCl2 (1m, 10 mL), phospho-
diesterase I (0.4 units, 4 mL), and alkaline phosphatase (20 units, 20 mL)
were added, and the mixtures were incubated again at 378C for 2 h. After
concentration, the products were subjected to HPLC analysis on an Inertsil
ODS-2 column (4.0 Â 150 mm, GL Sciences, Tokyo, Japan), first by an
isocratic elution with TEAA (0.1m, pH 7.0) for 5 min and then by the use of
an acetonitrile gradient (0 ± 10%) over a period of 20 min. The elution
profiles, monitored at l ˆ 230 nm, are shown in Figure 5. The authentic
samples of (5R)- and (5S)-thymidine glycols were obtained by deprotection
of 2a and 2b, respectively, with K2CO3/MeOH and then with 80% acetic
acid.
[13] a) Y. Vaishnav, E. Holwitt, C. Swenberg, H.-C. Lee, L.-S. Kan, J.
Biomol. Struct. Dyn. 1991, 8, 935 ± 951; b) J. Y. Kao, I. Goljer, T. A.
Phan, P. H. Bolton, J. Biol. Chem. 1993, 268, 17787 ± 17793; c) H. C.
Kung, P. H. Bolton, J. Biol. Chem. 1997, 272, 9227 ± 9236.
[14] a) C. DꢁHam, A. Romieu, M. Jaquinod, D. Gasparutto, J. Cadet,
Biochemistry 1999, 38, 3335 ± 3344; b) K. Asagoshi, H. Odawara, H.
Nakano, T. Miyano, H. Terao, Y. Ohyama, S. Seki, H. Ide, Biochem-
istry 2000, 39, 11389 ± 11398.
[15] a) S. L. Beaucage, R. P. Iyer, Tetrahedron 1993, 49, 6123 ± 6194; b) J.
Butenandt, L. T. Burgdorf, T. Carell, Synthesis 1999, 1085 ± 1105.
[16] T. J. Matray, M. M. Greenberg, J. Am. Chem. Soc. 1994, 116, 6931 ±
6932.
[17] S. Iwai, Angew. Chem. 2000, 112, 4032 ± 4234; Angew. Chem. Int. Ed.
2000, 39, 3874 ± 3876.
[18] S. Iwai, M. Shimizu, H. Kamiya, E. Ohtsuka, J. Am. Chem. Soc. 1996,
118, 7642 ± 7643.
[19] N. Usman, K. K. Ogilvie, M.-Y. Jiang, R. J. Cedergren, J. Am. Chem.
Soc. 1987, 109, 7845 ± 7854.
[20] a) K. K. Ogilvie, S. L. Beaucage, A. L. Schifman, N. Y. Theriault, K. L.
Sadana, Can. J. Chem. 1978, 56, 2768 ± 2780; b) S. S. Jones, C. B. Reese,
J. Chem. Soc. Perkin Trans. 1 1979, 2762 ± 2764.
Thermodynamic studies: Duplexes were formed by cooling the solutions of
the two strands (9 nmol) in water (90 mL) from 708C to 108C, and melting
curves were measured at the total oligonucleotide concentrations (Ct) of
2.00, 3.56, 6.32, 11.2, and 20.0 mm in a buffer (pH 7.0, 350 mL) containing
10 mm sodium phosphate, 100 mm NaCl, and 0.1 mm ethylenediaminetetra-
acetic acid on a Beckman DU-600Tm spectrophotometer. Tm values were
obtained by the two-point average method with the data processing
software supplied by the manufacturer. The reciprocal of Tm was plotted
against lnCt/4, and the thermodynamic parameters were obtained from the
equations, 1/Tm ˆ (R/DH8)lnCt/4 ‡ DS8/DH8 and DG8 ˆ DH8 TDS8, in
which R is the gas constant (ˆ 1.987 calmol 1 K 1) and T is the absolute
temperature.[26] The error analysis was carried out as previously reported.[29]
[21] a) T. Halmos, R. Montserret, J. Filippi, K. Antonakis, Carbohydr. Res.
1987, 170, 57 ± 69; b) D. Icheln, B. Gehrcke, Y. Piprek, P. Mischnick,
W. A. König, M. A. Dessoy, A. F. Morel, Carbohydr. Res. 1996, 280,
237 ± 250.
[22] M. R. Barvian, M. M. Greenberg, J. Org. Chem. 1993, 58, 6151 ± 6154.
[23] S. A. Higgins, K. Frenkel, A. Cummings, G. W. Teebor, Biochemistry
1987, 26, 1683 ± 1688.
[24] a) G. E. Plum, K. J. Breslauer, Ann. N.Y. Acad. Sci. 1994, 726, 45 ± 56;
b) D. S. Pilch, G. E. Plum, K. J. Breslauer, Curr. Opin. Struct. Biol.
1995, 5, 334 ± 342.
[25] a) J. Petruska, M. F. Goodman, M. S. Boosalis, L. C. Sowers, C.
Cheong, I. Tinoco, Proc. Natl. Acad. Sci. USA 1988, 85, 6252 ± 6256;
b) H. Ide, H. Murayama, S. Sakamoto, K. Makino, K. Honda, H.
Nakamuta, M. Sasaki, N. Sugimoto, Nucleic Acids Res. 1995, 23, 123 ±
129; c) Y. Fujiwara, S. Iwai, Biochemistry 1997, 36, 1544 ± 1550.
[26] L. A. Marky, K. J. Breslauer, Biopolymers 1987, 26, 1601 ± 1620.
[27] a) G. Vesnaver, C.-N. Chang, M. Eisenberg, A. P. Grollman, K. J.
Breslauer, Proc. Natl. Acad. Sci. USA 1989, 86, 3614 ± 3618; b) H. Ide,
H. Shimizu, Y. Kimura, S. Sakamoto, K. Makino, M. Glackin, S. S.
Wallace, H. Nakamuta, M. Sasaki, N. Sugimoto, Biochemistry 1995, 34,
6947 ± 6955; c) C. A. Gelfand, G. E. Plum, A. P. Grollman, F. Johnson,
K. J. Breslauer, Biochemistry 1998, 37, 7321 ± 7327.
Acknowledgement
I thank Ms. U. Tagami and Dr. K. Hirayama (Ajinomoto Co. Inc.) and Dr.
H. Urata (Osaka University of Pharmaceutical Sciences) for obtaining the
high-resolution mass data, Dr. K. Nakatani (Kyoto University), Dr. M.
Sekine (Tokyo Institute of Technology), Dr. A. Matsuda (Hokkaido
University), and Dr. H. Ide (Hiroshima University) for helpful discussions,
and Ms. Y. Fujiwara (Biomolecular Engineering Research Institute) for
technical advice regarding the Tm measurement.
[1] H. Wiseman, B. Halliwell, Biochem. J. 1996, 313, 17 ± 29.
[2] a) R. Teoule, C. Bert, A. Bonicel, Radiat. Res. 1977, 72, 190 ± 200; b) K.
Frenkel, M. S. Goldstein, G. W. Teebor, Biochemistry 1981, 20, 7566 ±
7571.
[3] R. Adelman, R. L. Saul, B. N. Ames, Proc. Natl. Acad. Sci. USA 1988,
85, 2706 ± 2708.
[4] R. C. Hayes, L. A. Petrullo, H. Huang, S. S. Wallace, J. E. LeClerc, J.
Mol. Biol. 1988, 201, 239 ± 246.
[28] W. H. A. Kuijpers, J. Huskens, L. H. Koole, C. A. A. van Boeckel,
Nucleic Acids Res. 1990, 18, 5197 ± 5205.
[5] A. K. Basu, E. L. Loechler, S. A. Leadon, J. M. Essigmann, Proc. Natl.
Acad. Sci. USA 1989, 86, 7677 ± 7681.
[6] a) H. Ide, Y. W. Kow, S. S. Wallace, Nucleic Acids Res. 1985, 13, 8035 ±
8052; b) J. M. Clark, G. P. Beardsley, Nucleic Acids Res. 1986, 14, 737 ±
749; c) J. M. McNulty, B. Jerkovic, P. H. Bolton, A. K. Basu, Chem.
Res. Toxicol. 1998, 11, 666 ± 673.
[29] a) S. L. Meyer, Data Analysis for Scientists and Engineers, Wiley, New
York, 1975, pp. 71 ± 75; b) J. SantaLucia, R. Kierzek, D. H. Turner, J.
Am. Chem. Soc. 1991, 113, 4313 ± 4322; c) M. Persmark, F. P.
Guengerich, Biochemistry 1994, 33, 8662 ± 8672.
Received: April 20, 2001 [F3214]
Chem. Eur. J. 2001, 7, No. 20
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0720-4351 $ 17.50+.50/0
4351