Radiation-induced and photosensitized splitting of C5-C5′-linked dihydrothymine dimers: Product and laser flash photolysis studies on the oxidative splitting mechanism

Article abstract of DOI:10.1021/jp991877r

Radiation-induced and photosensitized one-electron oxidation of stereoisomeric C5-C5′-linked dihydrothymine dimers (1a,b[meso], meso compound of (5R,5′S)-bi-5,6-dihydrothymine; 1a,b[rac], racemic compound of (5R,5′R)-and (5S,5′S)-bi-5,6-dihydrothymines), which are the major products yielded by radiolytic reduction of 1-methylthymine (2a) and 1,3-dimethylthymine (2b) in aqueous solution, was studied to compare with the photoreactivating repair mechanism of cyclobutane pyrimidine photodimers. Reacting with sulfate radical anion (SO₄^.-), azide radical (N₃^.), or photoexcited anthraquinone-2-sulfonate (AQS) as oxidants, the C5-C5′-linked dihydrothymine dimers 1a,b split to regenerate the corresponding thymine monomers 2a,b along with 5,6-dihydrothymines (3a,b) in a pH dependent manner. The transient absorption spectra of 5,6-dihydrothymin-5-yl radicals (6a,b) were observed in the nanosecond laser flash photolysis of 1a,b in phosphate buffer under conditions of SO₄^.- generation. Both the product study and the laser flash photolysis study indicated an oxidative splitting mechanism by which one-electron oxidation of the C5-C5′-linked dimers 1a,b produces the radical cation intermediates (4a,b), which undergo facile fragmentation into 5,6-dihydrothymin-5-yl radicals 6a,b and C5-cations (5a,b), followed by deprotonation at C6 of 5a,b to regenerate the monomers 2a,b.

Full text of DOI:10.1021/jp991877r

J. Phys. Chem. A 1999, 103, 8413-8420
8413
Radiation-Induced and Photosensitized Splitting of C5-C5′-Linked Dihydrothymine
Dimers: Product and Laser Flash Photolysis Studies on the Oxidative Splitting Mechanism
Takeo Ito, Hideki Shinohara, Hiroshi Hatta, and Sei-ichi Nishimoto*
Department of Energy and Hydrocarbon Chemistry, Graduate School of Engineering, Kyoto UniVersity,
Sakyo-ku, Kyoto 606-8501, Japan
Shin-ichi Fujita
Research Institute of AdVanced Technology, UniVersity of Osaka Prefecture,Osaka 599-8231, Japan
ReceiVed: June 9, 1999; In Final Form: August 23, 1999
Radiation-induced and photosensitized one-electron oxidation of stereoisomeric C5-C5′-linked dihydrothymine
dimers (1a,b[meso], meso compound of (5R,5′S)-bi-5,6-dihydrothymine; 1a,b[rac], racemic compound of
(5R,5′R)- and (5S,5′S)-bi-5,6-dihydrothymines), which are the major products yielded by radiolytic reduction
of 1-methylthymine (2a) and 1,3-dimethylthymine (2b) in aqueous solution, was studied to compare with the
photoreactivating repair mechanism of cyclobutane pyrimidine photodimers. Reacting with sulfate radical
•
anion (SO4•-), azide radical (N3 ), or photoexcited anthraquinone-2-sulfonate (AQS) as oxidants, the C5-
C5′-linked dihydrothymine dimers 1a,b split to regenerate the corresponding thymine monomers 2a,b along
with 5,6-dihydrothymines (3a,b) in a pH dependent manner. The transient absorption spectra of 5,6-
dihydrothymin-5-yl radicals (6a,b) were observed in the nanosecond laser flash photolysis of 1a,b in phosphate
•-
buffer under conditions of SO4 generation. Both the product study and the laser flash photolysis study
indicated an oxidative splitting mechanism by which one-electron oxidation of the C5-C5′-linked dimers
1a,b produces the radical cation intermediates (4a,b), which undergo facile fragmentation into 5,6-
dihydrothymin-5-yl radicals 6a,b and C5-cations (5a,b), followed by deprotonation at C6 of 5a,b to regenerate
the monomers 2a,b.
Introduction
bonds of the resulting photodimer radical anion and successive
electron transfer back to the flavin reproduce the monomeric
pyrimidines.4a As demonstrated by photochemical and radiation
chemical studies, the splitting reaction of photodimers could
also occur by a one-electron oxidation mechanism involving
appropriate oxidants such as sulfate radical anion (SO4•-), OH
radical (•OH),5 and photoexcited anthraquinone-2-sulfonate
(AQS*).6 The oxidative splitting mechanism is thought to be
distinct from the one-electron reduction mechanism,4,7,8 by
which the photodimer radical cation undergoes facile splitting
at the C6-C6′ bond in the initial step and successively at the
C5-C5′ bond. Such a mechanistic difference in the splitting of
photodimers has been accounted for by recent computational
studies based on the semiempirical AM1 and ab initio HF and
MP2 methods.9
It has been established that DNA is a vital target of genetoxic
agents such as a variety of chemical oxidants, ionizing radiation,
UV light, and certain antibiotics, that cause chemical modifica-
tion in DNA.1 For certain structures of the DNA toxic lesions
a number of repair enzymes in the cells have been identified
and characterized with respect to their structures and biological
functions. Most typically, excision-repair enzymes recognize
and excise damaged bases or oligo sequences containing
damaged base sites of DNA and then replace them with normal
DNA bases by the aid of DNA polymerase.2 There are also a
different class of direct in situ repair enzymes that regenerate
normal bases without excising the damaged bases or the DNA
backbone. For example, DNA photolyase (Mr ) 55000-65000)
absorbs UV-visible light and thereby catalyzes a redox splitting
of the cyclobutane pyrimidine photodimers, which are well-
known to be highly mutagenic and carcinogenic lesions induced
by UV exposure.3,4 This enzyme utilizes a deprotonated reduced
chromophore (FADH-) of flavin adenine dinucleotide and an
antenna chromophore of methyltetrahydrofolate (MTHF) or
8-hydroxy-5-deazaflavin (8-HDF). Thus, in the binding of DNA
photolyase to DNA as a light-independent reaction, the cofactor
MTHF or 8-HDF absorbs visible light and transfers energy to
form the electronically excited state of FADH- (FADH-*),
which in turn transfers an electron to the cyclobutane pyrimidine
photodimer. Concerted splitting at the C5-C5′ and C6-C6′
In the course of our study on the DNA base damaged
structures induced by radiolytic reduction, we have isolated and
identified a novel C5-C5′-linked dihydrothymine dimers, the
meso compound of (5R,5′S)-bi-5,6-dihydrothymine (1a,b[meso])
and the racemic compound of (5R,5′R)- and (5S,5′S)-bi-5,6-
dihydrothymines (1a,b[rac]).10,11 Interestingly, the three-
dimensional structures of these C5-C5′-linked dimers as
characterized by X-ray crystallography are similar to the
cyclobutane thymine photodimers possessing both C5-C5′- and
C6-C6′-bonds. This implied that the C5-C5′-linked dihy-
drothymine dimers may cause some distortion within a DNA
duplex if they were incorporated and could be substrates for
DNA photolyase. Our attention has been focused on the splitting
of the C5-C5′-linked dihydrothymine dimers into thymine
* Corresponding author. Fax: +81(75)753-3351. E-mail: nishimot@scl.
kyoto-u.ac.jp.
10.1021/jp991877r CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/04/1999

8414 J. Phys. Chem. A, Vol. 103, No. 42, 1999
Ito et al.
monomers to get structural insight into a redox mechanism,
compared with the well-characterized fragmentation reactivity
of the C5-C5′ and C6-C6′ bonds of pyrimidine photodimers.
This paper reports radiation-induced and photosensitized one-
electron oxidation of the stereoisomeric C5-C5′-linked dimers
of 1-methylthymine (1a[meso], 1a[rac]) and 1,3-dimethyl-
thymine (1b[meso], 1b[rac]) by several oxidizing species such
(CA 263) was guided with an optical fiber scope to be arranged
in an orientation perpendicular to the exciting laser beam. The
probe beam was monitored with a Hamamatsu R2949 photo-
multiplier tube through a Hamamatsu S3701-512Q MOS linear
image sensor (512 photodiodes). Timing of the exciting pulsed
laser, the probe beam, and the detection system was achieved
through a Tektronix model TDS 320 double channel oscil-
loscope that was interfaced to an NEC PC-9801 computer.
Solutions of 1a[meso] and 1a[rac] (1.0 mM) at pH 4, 7, and
10 in phosphate buffer (5.0 mM) containing K2S2O8 (50 mM)
were deaerated by Ar bubbling prior to the laser flash photolysis
experiments.
•
as sulfate radical anion SO4•-, azide radical (N3 ), and photo-
excited anthraquinone-2-sulfonate AQS*. Along with the prod-
uct study, a laser flash photolysis study was also performed to
observe the transient intermediates involved in the oxidative
splitting of C5-C5′-linked dihydrothymine dimers under condi-
•-
tions of photochemical SO4 generation.
Results and Discussion
Experimental Section
Splitting of C5-C5′-Linked Dihydrothymine Dimers by
Oxidizing Radicals Generated in the Radiolysis. The oxidiz-
Materials. 1-Methylthymine (2a) (Sigma Chemical) was used
as received. Purified 1,3-dimethylthymine (2b) was kindly
supplied by Fujii Memorial Research Institute, Otsuka phar-
maceutical. Anthraquinone-2-sulfonate (AQS), potassium per-
oxodisulfate (K2S2O8), potassium bromide (KBr), sodium azide
(NaN3), and 2-methyl-2-propanol were purchased from Nacalai
Tesque and were used without further purification. Reagents
for high-performance liquid chromatography (HPLC) including
solvents, sodium dihydrogen phosphate (NaH2PO4), and metha-
nol (HPLC grade) were used as received from Wako Pure
Chemical Industries. C5-C5′-linked dihydrothymine dimers
(1a,b) of 1-methylthymine and 1,3-dimethylthymine were
synthesized and purified by following the methods reported
previously.10,11
Radiolytic Oxidation. Aqueous solutions of C5-C5′-linked
dihydrothymine dimers (1a,b: 0.58 mM) containing [1] 2-meth-
yl-2-propanol (50 mM) and K2S2O8 (5.0 mM), [2] KBr (100
mM), or [3] NaN3 (50 mM) were prepared with water ion-
exchanged by Corning Mega-Pure System MP-190 (>16 MΩ
cm). The aqueous solutions were buffered at pH 4, 7, and 10
with phosphate buffer (2 mM) and sodium hydroxide and then
purged with [1] Ar or [2, 3] N2O before γ-irradiation. Steady-
state γ-irradiation was performed in a sealed ampule at room
temperature with a 60Co γ-ray source (dose rate: 0.79-0.70
Gy min-1). The γ-irradiated sample solutions, which were stable
at room temperature in the dark, were subjected to HPLC
analysis as described below.
Photosensitized Oxidation. Typically, solutions of 1a,b (1
mM) in phosphate buffer (5.0 mM) containing AQS (0.4 mM)
were adjusted to pH 4, 7, and 10 and then purged with Ar before
photoirradiation. The solutions in sealed Pyrex glass tubes (λex
> 280 nm) were photoirradiated under magnetic stirring (1000
rpm) at 24 °C with a high-pressure Hg arc (450 W, Eiko-sha
400).
HPLC Analysis. HPLC analyses were performed with
Shimadzu 3A and 10A HPLC systems equipped with Rheodyne
7725 sample injector. Aliquots (10 µL) of irradiated sample
solutions were injected onto a 5 µm C18 reversed-phase column
(Wakosil 5C18, L 4.6 mm × 150 mm, Wako). The phosphate
buffer solutions (10 mM, pH 3.0) containing various concentra-
tions of methanol (15-25 vol %) were delivered as the mobile
phase. The column eluents were monitored by the UV absor-
bance at 210 nm.
ing sulfate radical anion (SO4•-), as generated from the reaction
-
of peroxodisulfate ion (S2O82-) with hydrated electron (eaq
)
in the aqueous radiolysis system, has been conveniently utilized
for investigating the one-electron oxidation reactivity of pyri-
midines and purines.12 This radiation chemical method was
employed in the present study to get mechanistic insight into
one-electron oxidative splitting of C5-C5′-linked dihydrothy-
mine dimers (1a,b[meso] and 1a,b[rac]). Thus, Ar-saturated
solutions of 1a,b[meso] and 1a,b[rac] (0.5 mM) in phosphate
buffer containing 2-methyl-2-propanol (50 mM) and K2S2O8
(5.0 mM) were γ-irradiated up to 800 Gy. In the radiolysis of
aqueous solution, strongly oxidizing hydroxyl radicals (•OH)
and reducing species of eaq- are generated along with a smaller
amount of reducing hydrogen atoms (•H) (reaction 1). The G
-
H2O f •OH + •H + eaq
(1)
•OH + (CH3)3COH f H2O + •CH2(CH3)2COH (2)
eaq- (•H) + S2O82- f SO4•- + SO42- (HSO4 ) (3)
-
-
values13 of such primary water radicals are known as G(eaq
)
) G(•OH) ) 2.8 × 10-7 mol J-1, and G(•H) ) 0.6 × 10-7 mol
J-1 in neutral aqueous solution. Under the present conditions,
OH radicals are scavenged by 2-methyl-2-propanol into sub-
stantially unreactive 2-methyl-2-propanol radicals (•CH2(CH3)2-
COH) as in reaction 2 (k(•OH + (CH3)3COH) ) 6 × 108 dm3
mol-1 s-1).1a,14 On the other hand, hydrated electrons eaq are
-
2-
scavenged by S2O8 to produce oxidant SO4 radical anions
-
(E°(SO4•-/SO42-) ) 2.4 V vs NHE)15 as in reaction 3 (k(eaq
+ S2O82-) ) 1.2 × 1010 dm3 mol-1 s-1, k(•H + S2O82-) ) 2.5
× 107 dm3 mol-1 s-1).1a,14,16
As analyzed by HPLC, 1a,b[meso] and 1a,b[rac] were
oxidized by the radiation chemically generated SO4 radical
anions and their C5-C5′ bonds split to produce the correspond-
ing thymine monomers (2a,b) and 5,6-dihydrothymine deriva-
tives (3a,b) (Chart 1), which were accompanied by isomerization
from meso dimers 1a,b[meso] to racemic dimers 1a,b[rac] or
vice versa. Figure 1 shows radiation dose responses of the dimer
decomposition and the product formation in phosphate buffer
at pH 7, the initial slopes of which gave the respective G values
as listed in Table 1. The fairly high yields of 2a,b are in accord
with a prediction that the C5-C5′-linked dimers 1a,b, as
afforded by one-electron reductive dimerization of thymine
monomers 2a,b,11 may undergo the reverse splitting by a one-
electron oxidation mechanism.
Nanosecond Laser Flash Photolysis. The laser flash pho-
tolysis experiments were carried out with a Unisoku TSP-601
flash spectrometer. A Continuum Surelite-I Nd:YAG (Q-
switched) laser with the fourth harmonic at 266 nm (ca. 50 mJ
per 6 ns pulse) was employed for the flash photoirradiation.
The probe beam from a Hamamatsu 150 W xenon short arc
Table 1 indicates that the G value of dimer decomposition
(G(-1)) exceeds that of SO4 radical anion (G(SO4•-) < 3.2 ×

Splitting of C5-C5′-Linked Dihydrothymine Dimers
J. Phys. Chem. A, Vol. 103, No. 42, 1999 8415
TABLE 1: Oxidative Splitting of 1a,b (0.58 mM) in
Phosphate Buffer Solutions by Radiation Chemically
CHART 1
•-
•
Generated Oxidizing Radicals SO4 or N3
G × 107/mol J-1
substrate pH -1
2
3
isomerization
•-
Sulfate Radical Anion (SO4
)
1a[meso]
1a[rac]
4
7
4.1
3.6
2.1 (49%)a
2.3 (43%)
2.8 (74%)
1.6 (48%)
2.2 (68%)
2.3 (77%)
3.0 (58%)
3.3 (78%)
3.4 (85%)
2.8 (65%)
3.1 (86%)
3.2 (92%)
2.6 (62%)
1.6 (42%)
0.75 (20%)
0.83 (26%)
0.77 (23%)
0.99 (34%)
0.52 (10%)
0.57 (13%)
0.55 (14%)
0.66 (15%)
0.65 (18%)
0.65 (19%)
•
0.08 (2%)
0.31 (9%)
0.31 (8%)
0.21 (6%)
0.34 (10%)
0.23 (8%)
0.33 (6%)
0.50 (12%)
0.38 (10%)
0.27 (6%)
0.31 (9%)
1.3 (40%)
10 3.7
4
7
10 2.9
4
7
10 4.0
4
7
3.2
3.3
1b[meso]
1b[rac]
5.2
4.3
4.4
3.6
10 3.4
Azide Radical (N3 )
1a[meso]
4
7
0.23 0.18 (77%) 0.10 (45%)
0.16 0.17(107%) 0.11 (73%)
10 0.33 0.37(113%) 0.15 (44%)
a Selectivity based on G values for the decomposition of 1a,b.
C5′ bond may generate 5,6-dihydrothymin-5-yl radicals (6a,b),
as supported by the occurrence of isomerization (see Scheme
1).
The SO4•--induced splitting of 1a,b to produce 2a,b and 3a,b
was dependent on the pH of the aqueous solution. A trend is
seen in Table 1 that the restoration of 1a,b to 2a,b becomes
more efficient upon increasing the pH value: the yields of 2a,b
in basic solution at pH 10 are about 1.5 times greater than those
in acidic solution at pH 4. On the other hand, there was
substantially no influence of stereoisomerism on the splitting
reactivity at each pH. Table 1 also suggests that 1,3-dimethyl-
5,6-dihydrothymine dimers 1b[meso] and 1b[rac] are oxidized
by SO4 radical anions somewhat more readily than 1-methyl-
5,6-dihydrothymine dimers 1a[meso] and 1a[rac]. It is likely
that the inductive electron-donating 3-methyl group may
facilitate one-electron oxidation of 1b relative to 1a. As a
different aspect in this context, a methyl substitution results in
a more negative reduction potential of 5,6-dihydropyrimidines.18
Less oxidizing Br2 radical anions (Br2•-; E°(Br2•-/2Br-) )
1.7 V vs NHE)15 were also generated by γ-radiolysis of N2O-
saturated phosphate buffer solution containing an excess amount
of KBr (100 mM) to examine the oxidative splitting reactivity
of the C5-C5′-linked dihydrothymine dimers. Under the present
conditions of radiolysis, more than 98% of hydrated electrons
-
are converted into OH radicals by N2O (reaction 4; k(eaq
+
N2O) ) 9.1 × 109 dm3 mol-1 s-1).14 The OH radicals are in
turn scavenged by Br- ions to produce oxidizing Br2 radical
anions (reactions 5 and 6).1 In contrast to the reactivity toward
Figure 1. Oxidative splitting of (a) 1a[meso] and (b) 1a[rac] (0.58
mM) in the γ-radiolysis of Ar-purged phosphate buffer solution
containing K2S2O8 (5.0 mM) and 2-methyl-2-propanol (50 mM): (O)
decomposition of 1a[meso] or 1a[rac]; (() formation of 2a; (0)
isomerization of 1a.
eaq- + N2O f N2 + •OH + OH-
Br- + •OH f Br• + OH-
(4)
(5)
(6)
•-
Br• + Br- f Br2
10-7 mol J-1 depending on pH) to some extent. This suggests
that some radical intermediates involved in the oxidation of 1a,b
by SO4 radical anions are able to decompose 1a,b. In view of
the previous result that the â-oxoalkyl compounds such as
6-hydroxy-5,6-dihydrothymin-5-yl radicals possess oxidizing
properties,17 one-electron oxidation by such oxidizing 5-yl
radicals may account for the excess G values of dimer
decomposition. Indeed, the oxidative splitting of 1a,b at the C5-
SO4 radical anions, neither 1a[meso] nor 1a[rac] (0.5 mM) was
decomposed by Br2 radical anions thus generated in phosphate
buffer at pH 7. Such an ineffective oxidizing reactivity of Br2
radical anions has been observed for the reactions with cy-
clobutane pyrimidine photodimers.5 Even for electron-rich
thymine bearing a C5-C6 double bond, Br2 radical anions
showed practically negligible reactivity (k(Br2•- + thymine) <
107 dm3 mol-1 s-1).19

8416 J. Phys. Chem. A, Vol. 103, No. 42, 1999
Ito et al.
SCHEME 1
Using azide ions (N3-; 50 mM NaN3) instead of Br- ions in
generated OH radicals.
TH2 + •OH f TH• + H2O
•
N2O-saturated phosphate buffer, oxidizing azide radicals (N3 )
were also generated as in reaction 7 (k ) 1.2 × 1010 dm3 mol-1
(8)
s-1, E°(N3 /N3-) ) 1.3 V vs NHE).14,15 Table 1 shows the results
•
TH• + •OH f TH+ + OH-
TH+ + OH- f T + H2O
(9)
•
N3- + •OH f N3 + OH-
(7)
(10)
Upon γ-radiolysis (600 Gy) of N2O-saturated solution in
phosphate buffer 1a,b (0.5 mM) could not be restored to 2a,b
in contrast to cyclobutane thymine photodimer. This result
clearly rules out a hypothetical one-electron oxidation of the
C5-C5′-linked dihydrothymine dimers 1a,b by OH radicals to
form the dimer radical cations 4a,b. Although C5-C5′-linked
dihydrothymine dimer 6-yl radicals (7a,b) are most likely
intermediates derived from C6-hydrogen abstraction by OH
radicals (path 6 in Scheme 1), fragmentation of 7a,b to yield
thymines 2a,b and 5,6-dihydrothymin-5-yl radicals 6a,b (path
5) should not occur in the OH radical reaction. Concerning a
mechanism of the oxidative splitting of 1a,b, on the other hand,
such a reactivity of 7a,b implies that the dimer radical cations
4a,b as the probable intermediates could not yield thymines
2a,b, if deprotonation at the C6 of 4a,b was predominant (path
4). Thus, it is plausible that the dimer radical cations 4a,b are
liable to undergo C5-C5′-bond splitting into the corresponding
5-yl radical 6a,b and C5-cation (5a,b) (path 2).
Splitting of C5-C5′-Linked Dihydrothymine Dimers by
Photoexcited Oxidizing Sensitizer. As demonstrated previously
by means of CIDNP,6 oxidative splitting of cyclobutane
pyrimidine photodimers (Pyr<>Pyr) into pyrimidine monomers
(Pyr) was effectively induced upon photoexcitation of oxidizing
sensitizers such as anthraquinone-2-sulfonate (AQS), which
possesses a reduction potential of -0.39 V vs NHE.15 The
observation of substrate and product photo-CIDNP signals led
to a one-electron oxidation mechanism of the photodimer
splitting by which electron transfer from photodimer Pyr<>Pyr
to the excited triplet state of AQS (3AQS*) occurs to produce
the photodimer radical cation (Pyr<>Pyr•+) and ASQ radical
•
of N3 -induced oxidation of 1a[meso] (0.5 mM) in phosphate
buffer solutions at pH 4, 7, and 10. The dimer 1a[meso] split
into the monomer 2a in high yield at each pH, despite the
reduction potential of the azide radical being more negative by
about 0.25 V than that of Br2 radical anion. While the G value
•
of the N3 -induced splitting of 1a[meso] was reduced to
approximately 10% the level of the SO4•--induced splitting,
selectivity of the restoration to 2a was significantly enhanced
(Table 1). A similar difference in oxidizing reactivity between
azide radicals and Br2 radical anions was also observed for the
reactions with other nucleic acids. It has been thus suggested
that azide radicals react by an outer-sphere electron-transfer
mechanism depending on the reduction potential of a given
nucleic acid, whereas Br2 radical anions react by an inner-sphere
electron-transfer mechanism.20 Therefore, the splitting of C5-
C5′-linked dihydrothymine dimers 1a,b into 2a,b by SO4 radical
anion or azide radical proceeds via one-electron oxidation (path
1 in Scheme 1) most possibly by an outer-sphere electron-
transfer mechanism.
Previously, strongly oxidizing OH radicals with a high
reduction potential (E°(•OH/OH-) ) 2.7 V vs NHE)15 were
shown to restore cyclobutane thymine photodimer to thymine
with G values of 2.0-3.1.5,21 This reaction might proceed by
C6- or C6′-hydrogen abstraction rather than direct electron
transfer, on analogy of similar •OH-induced restoration of 5,6-
dihydrothymine to thymine involving a hydrogen-abstraction
mechanism as in reactions 8-10.21 For comparison, we further
studied the reactivity of 1a,b toward radiation chemically

Splitting of C5-C5′-Linked Dihydrothymine Dimers
J. Phys. Chem. A, Vol. 103, No. 42, 1999 8417
anion (AQS•-). The radical cation Pyr<>Pyr•+ thus photoge-
nerated may undergo splitting into pyrimidine and its radical
cation, the latter of which is one-electron reduced to pyrimidine
by the radical anion AQS•-.
AQS 9 9 +
hν8 1AQS* isc8 3AQS* Pyr<>Pyr8 AQS•-
Pyr<>Pyr•+
In this light, we also examined similar AQS-sensitized
photooxidation reactivity of the C5-C5′-linked dihydrothymine
dimers 1a,b. Upon photoexcitation of AQS (0.4 mM) in
deoxygenated phosphate buffer (5.0 mM, pH 4, 7 and 10)
containing 1a,b (1.0 mM), efficient splitting of 1a,b into 2a,b
and 3a,b was observed in a similar manner as in the one-electron
oxidative splitting by SO4 radical anions or azide radicals. This
photooxidative splitting of 1a,b is likely to proceed via electron
transfer from 1a,b to the 3AQS*, as in the case of cyclobutane
pyrimidine photodimers, since it was suppressed by the addition
of diacetyl (4.0 mM), a typical triplet quencher (ET ) 236 kJ
mol-1).22 In contrast to the splitting by oxidizing radicals in
the radiolysis systems, the photosensitized splitting by AQS
involved no isomerization of the dimers 1a,b. This suggests
that one-electron reduction of the 5-yl radical intermediates 6a,b
by the radical anion AQS•- may occur more efficiently than
radical coupling and disproportionation (paths 7 and 8 in Scheme
1), followed by protonation to yield 3a,b. Therefore, in the AQS-
sensitized photoreaction system, the electron transfer from
3AQS* and the back electron transfer from AQS•- would lead
to a net splitting of the C5-C5′-linked dihydrothymine dimers
1a,b into the corresponding thymines 2a,b and 5,6-dihy-
drothymines 3a,b in a 1:1 ratio.
Figure 2 shows the time course of the AQS-sensitized
photooxidative splitting of 1a,b in phosphate buffer at pH 7. It
is evident that 2a,b and 3a,b were produced in almost equivalent
yields in the initial stage of photoirradiation, while the yield of
2a,b became considerably higher than that of 3a,b during the
prolonged photoirradiation. This behavior suggests concurrent
oxidation of 3a,b to produce 2a,b as a secondary reaction in
the AQS-sensitized photooxidation system. In a separate experi-
ment, we confirmed that photoexcitation of AQS (0.2 mM)
induced oxidation of 3a (1.0 mM) to yield 2a almost quanti-
tatively in Ar-purged phosphate buffer at pH 7 (Figure 3).
Therefore, direct one-electron oxidation of dihydrothymines 3a,b
by 3AQS* accounts for the secondary conversion to the
corresponding monomers 2a,b. According to a semiquinone-
semiquinolate equilibrium with pKa ) 8.2 (reaction 11) as
Figure 2. Oxidative splitting of (a) 1a[meso] and (b) 1a[rac] (1.0 mM)
in Ar-purged phosphate buffer upon photoexcitation (λex > 280 nm)
of AQS (0.4 mM): (0) decomposition of 1a[meso]; formation of (b)
2a and (O) 3a.
AQS•- + H+ f AQSH•
(11)
Figure 3. Oxidation of 3a (1.0 mM) in Ar-purged phosphate buffer
upon photoexcitation (λex > 280 nm) of AQS (0.2 mM): (O)
decomposition of 3a; (b) formation of 2a.
characterized by the previous flash photolysis study,23 on the
other hand, the semiquinone radicals (AQSH•) are derived from
protonation of the semiquinolate ions AQS•- at an almost
diffusion-controlled rate under neutral conditions. An alternative
reaction pathway is also possible where the primary photoprod-
uct 3a,b may undergo hydrogen abstraction by the semiquinone
radicals AQSH• to yield 2a,b. Similar hydrogen abstraction is
splittings, a trend was also seen in the AQS-sensitized photo-
reaction that 1b[meso] and 1b[rac] favor a one-electron oxida-
tive splitting more than 1a[meso] and 1a[rac] (Tables 1 and
2).
Laser Flash Photolysis. Laser flash photolyses at 355 nm
of aqueous solutions of C5-C5′-linked dihydrothymine dimers
1a,b (1.0 mM) and AQS (50 µM) as an oxidizing sensitizer
were first attempted to detect transient absorptions originating
from oxidative splitting of 1a,b but were unsuccessful because
of overlapping with intense absorption due to AQS•- in the
wavelength region of 300-530 nm (ꢀ(AQS•-)500 ) 6.8 × 103
dm3 mol-1 cm-1; ꢀ(AQS•-)385 ) 5.37 × 103 dm3 mol-1 cm-1).24
In light of the SO4•--induced oxidation in the radiolysis
experiment as above, oxidizing SO4 radical anions were
•
also involved in the OH-induced restoration of 3a,b to 2a,b
(reactions 8-10).21
Table 2 shows the influence of pH on the photosensitized
splitting of 1a,b into 2a,b and 3a,b as observed in the initial
stage of photoirradiation for 8 min. Although the pH dependence
was not clear, the decomposition of 1a,b seemed to become
more facilitated with increasing pH of aqueous solution, as
expected from the semiquinone-semiquinolate equilibrium, as
in reaction 11. As observed in the SO4•-- and N3 -induced
•

8418 J. Phys. Chem. A, Vol. 103, No. 42, 1999
Ito et al.
TABLE 2: Oxidative Splitting of 1a,b (1.0 mM) in
Ar-Purged Phosphate Buffer upon Photoexcitation (λex
>
280 nm) of AQS (0.4 mM) for 8 min with a High Pressure
Hg Lamp
yield/%
conversion
substrate
pH
of 1a,b/%
2a,b
3a,b
1a[meso]
4
7
10
4
7
10
4
7
10
4
7
10
17
16
27
17
17
24
20
18
15
29
30
30
65
63
48
47
53
58
75
106
73
55
60
43
42
44
52
35
41
71
65
106
100
34
1a[rac]
1b[meso]
1b[rac]
Figure 5. Time courses of the absorbencies at (a) 400 nm and (b) 500
67
60
2-
nm in the 266 nm laser photolysis of S2O8 (50 mM) in phosphate
buffer solution containing 1a[meso] (1.0 mM). (c) Buildup of the
absorbance at 400 nm derived from the decay profile at 400 nm by
subtracting a decay component following the kinetics as in the
absorbance at 500 nm.
mM) and K2S2O8 (50 mM) at pH 10.2, the SO4 radical anions
decayed much more rapidly due to a reaction with OH- to
generate OH radicals (k ) 6.5 × 107 dm3 mol-1 s-1, reaction
13),27 which occurs in competition with the one-electron
oxidation of 1a[meso].
SO4•- + OH- f SO42- + •OH
(13)
In the present experiments, we could not identify the dimer
radical cations 4a that would possess a characteristic absorption
at wavelengths shorter than 350 nm. It follows that the dimer
radical cations 4a would be liable to split into C5-cations 5a
and radicals 6a on the nanosecond time scale.
Figure 4. Transient absorption spectra of the intermediates as observed
(O, pH 3.3) 0.4 µs and (b, pH 3.3; 4, pH 10.2) 2.0 µs after the 266
2-
nm laser photolysis of S2O8 (50 mM) in phosphate buffer solution
containing 1a[meso] (1.0 mM).
Parts a and b of Figure 5 compare the decays of absorbancies
due to 1-methyl-5,6-dihydrothymin-5-yl radical 6a and SO4
generated to react with 1a,b by the laser flash photolysis of
•-
S2O8 in aqueous solution, as in reaction 12.25
2-
at 400 and 500 nm, respectively. In the initial stage up to 20 µs
after the laser flash, the transient absorbance at 400 nm is likely
to involve the overlapped absorption tail of SO4•-. The
contribution of SO4 radical anions to the absorbance at 400 nm
was therefore estimated from the corresponding decay behavior
at 500 nm, taking into account the relative molar extinction
coefficients (ꢀ400(SO4•-)/ꢀ500(SO4•-) ) 1.1). Thus, the buildup
of 1-methyl-5,6-dihydrothymin-5-yl radicals 6a (λmax ) 400 nm)
could be obtained by subtracting the contribution of SO4 radical
anions from the apparent absorbance at 400 nm, as shown in
Figure 5c. The decay of 6a was of the second-order kinetics
(see Figure 6), indicating the occurrence of disproportionation
(formation of 2a and 3a) and/or recombination (isomerization),
as in reaction 14. The second-order rate constants were
•-
S2O82- + hν f 2SO4
(12)
In a control experiment, upon laser flash excitation at 266
nm of peroxodisulfate ion S2O8 (50 mM) in deoxygenated
2-
aqueous solution, a broad transient absorption band assigned
to the SO4 radical anion at wavelengths of 300-550 nm (λmax
) 460 nm) was observed as reported previously.25 Similar 266
nm laser flash photolyses of Ar-purged acidic solution (pH 3.3)
of 1a[meso] (1.0 mM) in the presence of K2S2O8 (50 mM)
resulted in the transient absorption spectra shown in Figure 4.
Substantially the same behavior was observed for a solution of
1a[rac] (1.0 mM) under similar conditions. The spectrum
recorded within 1 µs after the laser flash contained an absorption
maximum at λmax ) 460 nm (ꢀ(SO4
)
) 4.6 × 102 dm3
•-
455
TH• + TH• f TH2 + T or TH-TH
(14)
mol-1 cm-1) characteristic of SO4 radical anions.25 The SO4
radical anions decayed rapidly and thereafter a common
absorption maximum at λmax ) 400 nm emerged from both
aqueous solutions of 1a[meso] and 1a[rac], as was evidently
observed 2 µs after the laser flash. This transient may be
assigned to 1-methyl-5,6-dihydrothymin-5-yl radical 6a by
reference to the previous pulse radiolysis study, in which 5,6-
dihydrothymin-5-yl radicals generated by one-electron reduc-
tions of 5-bromo-5,6-dihydrothymine and 5-bromo-5,6-dihy-
approximately evaluated as 2k ) 2.7 × 109 dm3 mol-1 s-1 in
the splitting of 1a[meso] and 2.2 × 109 dm3 mol-1 s-1 in the
splitting of 1a[rac] from the slopes of the reciprocal of
absorbance at 400 nm vs time plots (Figure 6), assuming that
the molar extinction coefficient of 6a is equal to the reported
value of the 5,6-dihydrothymin-5-yl radical at 380 nm (ꢀ380
)
1150 dm3 mol-1 cm-1).26 The rate constants thus evaluated are
in good agreement with the literature value (2k ) 2.35 × 109
dm3 mol-1 s-1 at pH 6.8) for the 5,6-dihydrothymin-5-yl radical
as derived from the pulse radiolysis study on the reaction of
5,6-dihydrothymine with OH radicals.28
-
drouracil derivatives by hydrated electrons eaq in aqueous
solution showed broad absorptions at around 400 nm (λmax
)
380 nm, ꢀ380 ) 1150 dm3 mol-1 cm-1).26 When the laser flash
photolysis was performed in a basic solution of 1a[meso] (1.0

Splitting of C5-C5′-Linked Dihydrothymine Dimers
J. Phys. Chem. A, Vol. 103, No. 42, 1999 8419
drothymin-5-yl radicals 6a,b, involving C5-C5′-bond splitting
followed by deprotonation (paths 2 and 3) or vice versa (paths
4 and 5), can be assumed, as shown in Scheme 1. The latter
route is, however, ruled out from the experimental evidence
that OH radicals could not produce 2a,b. Thus, the C5-C5′-
linked dihydrothymine dimer 6-yl radicals 7a,b, that are
potentially derived from C6-deprotonation of 4a,b (path 4) or
C6-hydrogen abstraction from 1a,b by OH radicals (path 6) are
unlikely to undergo the C5-C5′-bond splitting into 2a,b and
6a,b (path 5). Alternatively, 4a,b may favor the C5-C5′-bond
splitting into the 5-yl radicals 6a,b and C5′-cations 5a,b (path
2). In view of the oxidizing property of 5,6-dihydropyrimidin-
5-yl radicals,17 the C5-cations 5a,b must be so unstable as to
be deprotonated at C6 into thymines 2a,b. In relation to the
C5-C5′-bond splitting reactivity of 4a,b, it is interesting to note
that the cyclobutane photodimer radical cation undergoes a
stepwise fragmentation involving sequential cleavage of the
C6-C6′ and the C5-C5′ bonds, as supported by the successful
trapping of a single-bond-cleaved intermediate in a model
reaction system.30 Such a preferential C6-C6′-bond splitting
of the photodimer radical cations resulting in intramolecular
formation of a pair of a dimer C6 radical and a C6′ cation seems
to be rationalized by the reducing property of 5,6-dihy-
drothymin-6-yl radicals.1a Accordingly, a radical and carbocation
pair formed by the C6-C6′-bond splitting of the photodimer
radical cation would be energetically more stable than that by
the C5-C5′-bond splitting of 4a,b.
Figure 6. Plot of 1/OD vs time for the buildup component at 400 nm
in the 266 nm laser photolysis of S2O82- (50 mM) in phosphate buffer
solution containing 1a[meso] (1.0 mM). Inset: pseudo-first-order
kinetics for the buildup component with absorbance at 400 nm. ODinf
corresponds to a plateau value of the absorbance at 400 nm due to the
buildup component.
The SO4 radical anions decayed following the pseudo-first-
order kinetics in the presence of 1a[meso] or 1a[rac], while
being of the second-order kinetics in the control photolysis of
peroxodisulfate ion S2O82- alone. The pseudo-first-order decays
gave the corresponding rate constants for the oxidation of dimers
by the SO4 radical anion as k ) 2.4 × 108 dm3 mol-1 s-1
(1a[meso] ) 1.0 mM) and k ) 1.5 × 108 dm3 mol-1 s-1
(1a[rac] ) 1.0 mM). In accord with the decay behavior of SO4
radical anions, the buildup of 5-yl radicals 6a to attain its plateau
level was represented in terms of a pseudo-first-order kinetics
(the inset in Figure 6), leading to similar rate constants of k )
1.8 × 108 dm3 mol-1 s-1 (1a[meso]) and k ) 1.0 × 108 dm3
mol-1 s-1 (1a[rac]). On the other hand, the SO4•--induced
oxidation of 5,6-dihydrothymine (DHT) at varying concentra-
tions of 0.5-5.0 mM was determined to show a rate constant
of k(DHT + SO4•-) ) 8.5 × 107 dm3 mol-1 s-1. These values
were at least 1 order of magnitude smaller than those of 5,6-
unsaturated pyrimidines such as 1-methylthymine (k(2a +
SO4•-) ) 5.0 × 109 dm3 mol-1 s-1) and 1,3-dimethylthymine
(k(2b + SO4•-) ) 4.6 × 109 dm3 mol-1 s-1).29 This is probably
attributable to the lower nucleophilicity of 5,6-dihydrothymine
compounds including 1a[meso] and 1a[rac] toward SO4 radical
anions.
The C5-C5′-linked dimers 1a[meso] and 1a[rac] are of the
N3-protonated forms in the pH range 4-10, since their pKa
values in aqueous solution are 11.7, as measured by the UV-
absorption spectral changes (data not shown). The pH depen-
•
dence of the SO4•-- and N3 -induced splitting of 1a,b that the
restoration to 2a,b becomes more efficient with increasing the
value (Table 1) can be explained by the more facilitated C6-
deprotonation of 5a,b (path 3). On the other hand, base-enhanced
C6-deprotonation of 4a,b to form 7a,b is unlikely, because it
will reduce the restoration efficiency if occurred.
In accord with the decay of the second-order kinetics shown
in Figure 5, two types of bimolecular reactions of 5,6-
dihydrothymin-5-yl radicals 6a,b may proceed competitively
in the reaction system. Recombination and disproportionation
of 6a,b accounts for the apparent isomerization of the dimers
1a,b (path 7) and formation of 2a,b and 3a,b (path 8),
respectively. As a concomitant minor reaction, oxidizing radical
6a,b could one-electron oxidize the C5-C5′-linked dimers 1a,b,
thereby resulting in the enhanced G value for the decomposition
of 1a,b (see Table 1; G(-1a,b) > G(SO4•-) ) 3.3 × 10-7 mol
J-1).17
Oxidative Splitting Mechanism of the C5-C5′-linked
dihydrothymine Dimers. Scheme 1 illustrates a proposal for
a mechanism by which the C5-C5′-linked dihydrothymine
dimers 1a,b undergo splitting into thymines 2a,b and 5,6-
dihydrothymines 3a,b by oxidizing radicals such as SO4•- and
•
N3 , but not Br2•-, and by photoexcited oxidizing sensitizers
3
•-
such as AQS*. The failure of the oxidant Br2 in reacting
Conclusion
•-
toward 1a,b is not surprising in view of the fact that Br2 is
also unreactive toward the cyclobutane thymine dimers.5 The
laser flash photolysis study has provided spectroscopic evidence
in support of 5,6-dihydrothymin-5-yl radical intermediates 6a,b
involved in the SO4•--induced oxidative splitting of 1a,b. The
C5-C5′-linked dihydrothymine dimer radical cations 4a,b
would be the most likely precursor leading to the observed 5-yl
radicals 6a,b, although spectroscopic detection of the transients
4a,b was unsuccessful in the laser flash photolysis experiments.
A similar mechanistic aspect has been proposed for the oxidative
splitting of cyclobutane pyrimidine photodimers in that the
radical cations are generated by electron transfer from the
photodimers to a photoexcited oxidizing sensitizer.4a
We have demonstrated oxidative splitting of the C5-C5′-
linked dihydrothymine dimers 1a,b into thymine monomers 2a,b
along with byproducts of 5,6-dihydrothymines 3a,b by oxidizing
•-
•
radicals (SO4 and N3 ) generated in the radiolysis and by
photoexcited oxidizing sensitizer (3AQS*). As characterized by
the laser flash photolysis, 5,6-dihydrothymin-5-yl radicals 6a,b
are intermediates involved in the oxidative splitting. A mech-
anism involving generation of the C5-C5′-linked dihydrothy-
mine dimer radical cations 4a,b by electron transfer from 1a,b
to an oxidant has been proposed, by reference to a similar
oxidative splitting mechanism of cyclobutane pyrimidine pho-
todimers. While the pyrimidine photodimer radical cation prefers
the C6-C6′-bond splitting to the counterpart C5-C5′-bond
splitting, the C5-C5′-linked dihydrothymine dimer radical
Concerning the fate of dihydrothymine dimer radical cations
4a,b, two possible routes to thymines 2a,b and 5,6-dihy-

8420 J. Phys. Chem. A, Vol. 103, No. 42, 1999
Ito et al.
(12) (a) Angelov, D.; Berger, M.; Cadet, J.; Getoff, N.; Keskinova, E.;
Solar, S. Radiat. Phys. Chem. 1991, 37, 717-727. (b) Bansal, K. M.;
Fessenden, R. W. Radiat. Res. 1978, 75, 497-507.
(13) The number of molecules produced or changed per 1 J of radiation
energy absorbed by the reaction system.
cations 4a,b may undergo the C5-C5′-bond splitting into the
5,6-dihydrothymin-5-yl radicals 6a,b and the C5-cations 5a,b.
Because of the oxidizing property, the intermediate radicals 6a,b
can also be one-electron oxidants toward 1a,b.
(14) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. J.
Phys. Chem. Ref. Data 1988, 17, 513-886.
Acknowledgment. This work was supported by the Grant-
in-Aids for JSPS Fellows (No. 00086200) and for Scientific
Research (C) (No. 09640632) from the Ministry of Education,
Science, Sports, and Culture of Japan.
(15) Wardman, P. J. Phys. Chem. Ref. Data 1989, 18, 1637-1755.
-
(16) Direct reaction of eaq with the substrate was negligible at 5 mM
of K2S2O8 employed in the present γ-radiolysis study, as confirmed by a
separate comparative experiment: the yields of the radiolysis products were
virtually invariant, when γ-irradiated in the presence of 10-fold and 40-
fold excess amounts (5 and 20 mM) of K2S2O8 (see also ref 29).
(17) (a) Steenken, S.; Neta, P. J. Phys. Chem. 1979, 83, 1134-1137.
(b) Fujita, S.; Steenken, S. J. Am. Chem. Soc. 1981, 103, 2540-2545.
(18) (a) Yeh, S.-R.; Falvey, D. E. J. Am. Chem. Soc. 1992, 114, 7313-
7314. (b) Scannell, M. P.; Prakash, G.; Falvey, D. E. J. Phys. Chem. A
1997, 101, 4332-4337. (c) Scannell, M. P.; Fenick, D. J.; Yeh, S.-R.;
Falvey, D. E. J. Am. Chem. Soc. 1997, 119, 1971-1977.
(19) Neta, P.; Huie, R.; Ross, A. B. J. Phys. Chem. Ref. Data 1988, 17,
1027-1284.
(20) (a) Steenken, S. In Topics in Current Chemistry; Mattay, J., Ed.;
Springer-Verlag: Berlin, 1996; Vol. 177, pp 125-145. (b) Faraggi, M.;
Broitman, F.; Trent, J. B.; Klapper, M. H. J. Phys. Chem. 1996, 100, 14751-
14761. (c) Jovanovic, S. V.; Simic, M. G. J. Phys. Chem. 1986, 90, 974-
978.
(21) Grossweiner, L. I.; Kepka, A. G.; Santus, R.; Vigil, J. A. Int. J.
Radiat. Biol. 1974, 25, 521-523.
(22) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochem-
istry, 2nd ed., Revised and Expanded; Marcel Dekker: New York, 1993.
(23) Roy, A.; Aditya, S. J. Photochem. 1983, 22, 361-367.
(24) Phillips, G. O.; Worthington, N. W.; McKellar, J. F.; Sharpe, R.
R. J. Chem. Soc. A 1969, 767-773.
(25) (a) Dogliotti, L.; Hayon, E. J. Phys. Chem. 1967, 71, 2511-2516.
(b) Neda, O.; Yamauchi, K.; Masuda, T. Bull. Chem. Soc. Jpn. 1985, 58,
227-229.
(26) Deeble, D. J.; Das, S.; von Sonntag, C. J. Phys. Chem. 1985, 89,
5784-5788.
References and Notes
(1) (a) von Sonntag, C. The Chemical Basis of Radiation Biology;
Taylor and Francis: London, 1987. (b) Fuciarelli, A. F.; Zimbrick, J. D.
Radiation Damage in DNA: Structure/Function Relationship at Early Times;
Battelle Press: Columbus, 1995. (c) Bensasson, R. V.; Land, E. J.; Truscott,
T. G. Excited States and Free Radicals in Biology and Medicine; Oxford
University Press: Oxford, U.K., 1993.
(2) (a) Friedberg, E. C.; Walker, G. C.; Siede, W. DNA Repair and
Mutagenesis; ASM Press: Washington, D.C., 1995. (b) David, S. S.;
Williams, S. D. Chem. ReV. 1998, 98, 1221-1261. (c) Wallace, S. S. Radiat.
Res. 1998, 150, Suppl., S60-S79.
(3) (a) Cadet, J.; Vigny, P. In Bioorganic Photochemistry: Photo-
chemistry and the Nucleic Acids; Morrison, H., Ed.; John Wiley & Sons:
New York, 1990; pp 1-272. (b) Taylor, J.-S. Acc. Chem. Res. 1994, 27,
76-82. (c) Go¨rner, H. J. Photochem. Photobiol. B: Biol. 1994, 26, 117-
139. (d) Pfeifer, G. P. Photochem. Photobiol. 1997, 65, 270-283.
(4) (a) Begley, T. P. Acc. Chem. Res. 1994, 27, 394-401. (b) Kim,
S.-T.; Sancar, A. Photochem. Photobiol. 1993, 57, 895-904. (c) Carell,
T.; Epple, R. Eur. J. Org. Chem.. 1998, 1245-1258.
(5) Heelis, P. F.; Deeble, D. J.; Kim, S.-T.; Sancar, A. Int. J. Radiat.
Biol. 1992, 62, 137-143.
(6) (a) Young, T.; Nieman, R.; Rose, S. D. Photochem. Photobiol. 1990,
52, 661-668. (b) Pouwels, P. J. W.; Hartman, R. F.; Rose, S. D.; Kaptein,
R. Photochem. Photobiol. 1995, 61, 563-574.
(7) Pouwels, P. J. W.; Hartman, R. F.; Rose, S. D.; Kaptein, R.
Photochem. Photobiol. 1995, 61, 575-583.
(8) Heelis, P. F.; Hartman, R. F.; Rose, S. D. J. Photochem. Photobiol.
A: Chem. 1996, 95, 89-98.
(27) Hayon, E.; Treinin, A.; Wilf, J. J. Am. Chem. Soc. 1972, 94, 47-
57.
(9) (a) Voityuk, A. A.; Michel-Beyerle, M.-E.; Ro¨sch, N. J. Am. Chem.
Soc. 1996, 118, 9750-9758. (b) Voityuk, A. A.; Ro¨sch, N. J. Phys. Chem.
A 1997, 101, 8335-8338.
(28) Myers, L. S., Jr.; Theard, L. M. J. Am. Chem. Soc. 1970, 92, 2868-
2870.
(29) Deeble, D. J.; Schuchmann, M. N.; Steenken, S.; von Sonntag, C.
J. Phys. Chem. 1990, 94, 8186-8192.
(10) Nishimoto, S.; Ide, H.; Nakamichi, K.; Kagiya, T. J. Am. Chem.
Soc. 1983, 105, 6740-6741.
(11) Ito, T.; Shinohara, H.; Hatta, H.; Nishimoto, S. J. Org. Chem. 1999,
64, 5100-5108.
(30) Burdi, D.; Begley, T. P. J. Am. Chem. Soc. 1991, 113, 7768-7770.