Oxidation of adenosine and inosine: The chemistry of 8-Oxo-7,8- dihydropurines, purine iminoquinones, and purine quinones as observed by ultrafast spectroscopy

Article abstract of DOI:10.1021/ja3068148

Oxidative damage to purine nucleic acid bases proceeds through quinoidal intermediates derived from their corresponding 8-oxo-7,8-dihydropurine bases. Oxidation studies of 8-oxo-7,8-dihyroadenosine and 8-oxo-7,8-dihydroinosine indicate that these quinoidal species can produce stable cross-links with a wide variety of nucleophiles in the 2-positions of the purines. An azide precursor for the adenosine iminoquinone has been synthesized and applied in ultrafast transient absorption spectroscopic studies. Thus, the adenosine iminoquinone can be observed directly, and its susceptibility to nucleophilic attack with various nucleophiles as well as the stability of the resulting cross-linked species have been evaluated. Finally, these observations indicate that this azide might be a very useful photoaffinity labeling agent, because the reactive intermediate, adenosine iminoquinone, is such a good mimic for the universal purine base adenosine.

Full text of DOI:10.1021/ja3068148

Article
pubs.acs.org/JACS
Oxidation of Adenosine and Inosine: The Chemistry of 8‑Oxo-7,8-
dihydropurines, Purine Iminoquinones, and Purine Quinones as
Observed by Ultrafast Spectroscopy
Denis I. Nilov, Dmitry Y. Komarov, Maxim S. Panov, Kanykey E. Karabaeva, Andrey S. Mereshchenko,
Alexander N. Tarnovsky,* and R. Marshall Wilson*
Department of Chemistry and Center for Photochemical Sciences, Bowling Green State University, Bowling Green, Ohio 43403,
United States
S
* Supporting Information
ABSTRACT: Oxidative damage to purine nucleic acid bases proceeds through quinoidal intermediates derived from their
corresponding 8-oxo-7,8-dihydropurine bases. Oxidation studies of 8-oxo-7,8-dihyroadenosine and 8-oxo-7,8-dihydroinosine
indicate that these quinoidal species can produce stable cross-links with a wide variety of nucleophiles in the 2-positions of the
purines. An azide precursor for the adenosine iminoquinone has been synthesized and applied in ultrafast transient absorption
spectroscopic studies. Thus, the adenosine iminoquinone can be observed directly, and its susceptibility to nucleophilic attack
with various nucleophiles as well as the stability of the resulting cross-linked species have been evaluated. Finally, these
observations indicate that this azide might be a very useful photoaffinity labeling agent, because the reactive intermediate,
adenosine iminoquinone, is such a good mimic for the universal purine base adenosine.
not formed in such high yields, are more difficult to detect, and
because 8-oxodA (2) is less mutagenic than 8-oxodG (1).12
These oxidized bases might lead to miscoding in DNA
replication and transcription processes,13,14 and may be
associated with aging.15 In addition, it has been noted that 8-
oxo derivatives are formed more readily in RNA than in DNA,
possibly because single-stranded RNA is more susceptible to
oxidation than is duplex DNA or DNA protected by
chromatin.16,17
Repair enzymes for replacing 8-oxodG (1) with undamaged
guanosine are well-known.1,18 The comparable repair enzymes
for replacing 8-oxodA (2) with undamaged adenosine seem to
be much less well studied possibly because 8-oxodA (2) lesions
are much less mutagenic than 8-oxodG lesions, and are only
INTRODUCTION
■
Oxidatively damaged nucleic acids are now widely accepted as
one of the principal sources of genetic damage including
possible sources of carcinogenesis, mutation, aging, and cell
death.1−8 Purines are more susceptible to single-electron
oxidative damage than are pyrimidines. The oxidation
potentials of unmodified bases are G (1.29 V), A (1.42 V), C
(1.6 V), and T (1.7 V) versus NHE.9 Oxidation of purine bases
initially leads to 8-oxo-7,8-dihydroguanosine (8-oxoG or 8-
oxodG (1)), 8-oxo-7,8-dihydroadenosine (8-oxoA or 8-oxodA
(2)), and 8-oxo-7,8-dihydroinosine (8-oxoI, or 8-oxodI (3)),
which can occur both within existing RNA/DNA strands, as
well as in nucleotide pools. As indicated by the aforementioned
reduction potentials, guanosine is the most easily oxidized base
and has been studied in this regard extensively.10,11 In contrast,
the oxidation of adenosine and inosine has been studied much
less extensively, possibly because their oxidation products are
Received: August 7, 2012
Published: January 22, 2013
© 2013 American Chemical Society
3423
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
might form relatively stable “cross-linked” adducts at the 2-
position related to 8, in contrast to the much more extensively
studied 8-oxoG (1), which forms more labile “cross-linked”
adducts at the 5-position as indicated in Scheme 1.27−31
RESULTS AND DISCUSSION
■
about 1/3 as common.19−22 In fact, there seems to be less
effective repair of 8-oxodA:T lesions in certain cases.23,24 In a
similar vein, the course of further oxidation of 8-oxodG (1) has
been well studied and found to afford 4 and 5 (Nuc2 = OH or
O in water), Scheme 1.25 Ultimately, the modified bases 4 and
5 may hydrolyze to form the urea derivative, 6. These oxidation
processes are usually attributed to attack on the purine nucleus
by hydroxyl radicals and other reactive oxygen species (ROS)
to form the 8-oxo species 1 or 2 and their further oxidation by a
variety of mechanisms.1−8,12−14,16 However, in a separate line
of research, photolysis of 8-azidoadenosine has been found to
produce a diiminoquinone purine intermediate related to 7,
O replaced by HN.26 This intermediate is readily attacked
by a variety of nucleophiles at the 2-position to form products
related to 8, O replaced by HN. Nucleophiles examined
by transient absorption spectroscopy in this work include
cysteine, diethylamine, sodium hydroxide, as well as amino acid
esters and nucleic acid bases. A similar quinone intermediate,
9NH2, is now thought to play a pivotal role in the oxidation of
8-oxoG (1).27−31 These observations suggest that oxidation of
8-oxopurines might follow similar oxidative pathways passing
through quinoidal purines rather than via direct attack of
hydroxyl radical on the purine nucleus. Thus, 8-oxoA (2) might
also be susceptible to oxidative substitution to form 8 via 7,
which in turn might link the oxidation of adenosine with that of
guanosine, as outlined in Scheme 1, and arrive at the same final
products as ROS mechanisms. Perhaps more significantly,
attack on 7 by nucleophiles that do not appreciably enhance the
susceptibility of 8 to further oxidation might lead to stable
adducts giving rise to covalent cross-links between nucleic acids
and proteins or other biomolecules. For these reasons, we have
undertaken a study of adenosine oxidation using the
mechanism outlined in Scheme 1 as our working hypothesis.
In this study, we have concentrated on the oxidation of the 8-
oxopurines 8-oxoA (2) and 8-oxoI (3), because these molecules
Chemical Oxidation Studies. Using the model for purine
oxidation outlined in Scheme 1, we have investigated the
oxidation of 8-oxoA (2) with a variety of oxidizing agents. The
full description of the following oxidation reactions along with
the spectroscopic characterization of the products is provided
in the Supporting Information. Thus, treatment of 8-oxoA (2)
with NBS in methanol alone at room temperature leads to
formation of the methanol adduct 8OMe, R = 1′-ribosyl, Nuc1
= OCH3, or in a methanolic imidazole solution, the imidazole
adduct 8IM, R = 1′-ribosyl, Nuc1 = imidazolyl as the major
products (Scheme 2). From these results, it is clear that
oxidation of 8-oxoA (2) produces an intermediate that is
susceptible to nucleophilic attack at the purine 2-position, and
that virtually any nucleophile, even weak ones like methanol
and water, vide infra, will add to the 2-position of this
intermediate. We attempted to observe this reactive inter-
mediate by conducting the oxidation reactions in the cold, but
were unsuccessful. These transient studies are described in full
detail in the Supporting Information. Apparently, these putative
quinoidal intermediates are too reactive to survive into the
second-time domain. Sigma adducts related to 10OMe have
been observed previously and found to aromatize in the
microsecond- and millisecond-time domains.32
As might be expected, oxidation of 8-oxoA (2) often
produces nucleophile adducts that are very susceptible to
further oxidation as outlined in Scheme 1. This subsequent
oxidation is indicated by the formation of N 1′-ribosyl urea (6)
observed in the NBS oxidation of 2, Scheme 2. In many cases of
DNA oxidation,14 2′-deoxy-1′-ribosyl urea has been observed,
and N 1′-ribosyl urea is occasionally observed in small amounts
in the oxidations conducted in this work. It was found that the
initial adducts in the 2-position are often more easily oxidized
than the starting 8-oxoA (2). Thus, addition of aliphatic amines,
such as diethyl amine and primary amine related to lysine,
quench the intermediate 7 quite rapidly even in the cold.
Scheme 1
3424
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
Scheme 2
Scheme 3
Unfortunately, it has not been possible to isolate these amine
adducts, possibly because they are highly susceptible to further
oxidation. Related amine adducts formed in the photochemistry
of 8-azidoadenosine are oxidized upon standing in air.33 Thus,
we suspect that these susceptible adducts are oxidized and
destroyed more rapidly than they are formed or during the
extensive HPLC procedures necessary for their isolation.
An amine that does not significantly activate the purine
nucleus to further oxidation is imidazole. Thus, when the NBS
oxidation of 8-oxoA (2) is conducted in a methanolic solution
of imidazole, the imidazole adduct 8IM (R = 1′-ribosyl, Nuc =
imidazolyl) (Scheme 2) can be isolated. In many ways,
imidazole is a unique amine nucleophile. The n-lone pair on
nitrogen serves as a good nucleophile, but once attached to the
purine ring, the imidazole electron pair on the nucleophilic
nitrogen becomes part of the imidazole aromatic sextet, and
thus is not effectively donated to the purine ring where such
donation would activate the purine ring to further oxidation.
An alternative strategy for reducing overoxidation associated
with the addition of electron-rich nucleophiles to the purine
ring is to begin with a less electron-rich purine. The nucleic acid
8-oxoI (3) would seem to be an ideal candidate for testing this
approach, because the electron-donating 6-amino group in
adenosine has been replaced by the electron-withdrawing
carbonyl group in inosine. The potential required to oxidize 8-
oxoI (3) has been measured and found to be 1.2 V/NHE as
compared to about 0.92 V/NHE for 8-oxoA (2),34 and 0.74 V/
NHE for 8-oxoG (1).35 The relationship between these
potentials is supported by the observation that treatment of
8-oxoI (3) with methanolic NBS, the same conditions that were
successful for the oxidation of 8-oxoA (2), does not lead to the
formation of adducts. However, treatment of 3 with the more
powerful oxidizing agent sodium hypochlorite in aqueous
media in the presence of imidazole leads to the formation of the
imidazole adduct in the 2-position, 12IM, Scheme 3.
transformation to 8IM and 12IM, respectively. These experi-
ments produced only trace amounts of products. However,
when Na2S2O8 was added to the reactions photosensitized with
riboflavin and visible light, the imidazole adducts 8IM and
12IM (11%) were formed, albeit in low yields. Thus, the
•−
riboflavin sensitized generation of SO4 is apparently a viable
method for the oxidation of 2 and 3. Of particular note is the
observation that direct irradiation of 3 at 254 nm in the
presence of Na2S2O8 and imidazole significantly enhances the
yield of 12IM; see Table 1. In Burrows’s studies of the
Table 1. Yields of Purine Oxidation Products
purine
oxidized
reaction conditions
NBS, CH3OH
products and % yield
2
8OMe triacetate, 6-acetamido,
a
6%
b
2
2
2
NBS, CH3OH, imidazole
NBS, H2O, imidazole
8IM tri-TBDMS, 19%
8IM, 23%
8IM, 21%
Na2IrCl6·6H2O, H2O,
imidazole
3
3
aq NaOCl, imidazole
12IM, 40.5%
12IM, 35%
Na2S2O8, hν (254 nm),
imidazole
3
Na2IrCl6·6H2O, H2O,
imidazole
12IM, 69%
a
Product isolated following treatment of crude reaction mixture with
b
acetic anhydride. Product isolated following treatment of crude
reaction mixture with TBDMSCl.
oxidation of 8-oxoG (1), Na2IrCl6 has been found to be the
oxidizing agent of choice.42−44 This reagent proved to be one of
the more effective in the oxidation of 8-oxoA (2) and 8-oxoI
(3) as well; see Table 1.
While the aforementioned observations provide good
circumstantial evidence for the intermediacy of the quinoidal
intermediates 7 and 11, alternative mechanistic explanations for
these reactions are conceivable. Furthermore, simple oxidative
methods for formation of 7 and 11 suffer from the fact that
many of products that might be formed through the addition of
various nucleophiles are themselves more easily oxidized than
are the starting 8-oxopurines, and thus cannot be isolated or
readily observed. Therefore, an alternative strategy for the
Oxidative damage to nucleic acids is thought to be associated
with “reactive” oxygen species (ROS) such as hydroxyl radicals
and singlet oxygen.36−41 Therefore, the visible light irradiation
of both 8-oxoA (2) and 8-oxoI (3) in the presence of oxygen,
imidazole, and rose bengal or riboflavin was examined in an
effort to determine if singlet oxygen might affect their
3425
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
formation of iminoquinone 7 has been devised that provides
much more information about the reactivity of 7 and at the
same time offers the possibility of isolation of products that
would be susceptible to further oxidation. Thus, the nitrene
derived from 6-azido-8-oxo-7,8-dihydropurine (13) should
undergo rapid, possibly solvent mediated, proton reallocation
as outlined in Scheme 4 to form iminoquinone 7, and the
formation and reactivity of 7 should be much more easily
studied when generated under these spectroscopically favorable
photochemical conditions.
and in the presence of imidazole (0.1 M). The analysis of these
reaction mixtures indicated the formation of reduction product
16, water addition product 17OH, over oxidation product 18,
dimerization products of undetermined structure, and in the
presence of imidazole, imidazole adduct 17IM, Scheme 7 and
Table 2.
The imidazole adducts and their acetylated and benzoylated
derivatives have been obtained by several chemical and
photochemical methods. In addition to the imidazole adducts
8IM and 17IM, the 8-oxoI imidazole adduct 12IM, Scheme 3,
has been observed in the chemical oxidation of 8-oxoI (3).
These observations, as well as similar observations for 8-
oxoG,25,27−31 are consistent with the general mechanism
outlined in Scheme 1 in which the pivotal intermediates are
quinoidal purines. The 8-oxoA and 8-oxoI adducts in the 2-
position, 8 and 12, respectively, are distinguished from the 8-
oxoG adduct in the 5-position, 19 in Scheme 1, in that the
purine nuclei remain intact following nucleophile attack, and,
consequently, might form stable cross-links with many other
biological molecules. It should be emphasized that in these
studies of quinoidal purines, 7, 11, and 21, adducts in the 5-
position were not identified. In addition, the further oxidation
of adducts in the 2-position, such as 8IM and 12IM, has yet to
be studied in detail. The implication of the above observations
with imidazole is that histidine adducts might provide relatively
stable protein cross-links, but this correlation remains to be
tested with the amino acid itself. While these and many related
studies will be necessary to evaluate the working hypothesis
outlined in Scheme 1, the availability of azide 13 has provided
an ideal opportunity to probe the intermediacy of iminoqui-
none 7 in the early stages of this reaction scheme.
Scheme 4
Toward this end, the triacetate and tribenzoate of azide 13,
13Ac3, and 13Bz3 have been synthesized45−47 as described in
detail in the Supporting Information. The preparation and
irradiation of the acetylated and benzoylated derivatives greatly
facilitates HPLC product analysis, but does not significantly
affect the product distribution.
Irradiation of 13Ac3 in methanolic solutions of imidazole
affords both reduction, 2Ac3, and the desired imidazole
substitution product, 8IMAc3, as outlined in Scheme 5. In
addition, the iminoquinone hydrolysis product 3Ac3 is formed
as well. The imidazole adduct 8IM isolated from this azide
photochemistry was compared and found to be identical to the
imidazole adduct isolated from the chemical oxidation of 8-
oxoA (2), Scheme 2.
A variety of azidopurines have been used in photoaffinity
labeling studies,26 but 6-azido-8-oxo-7,8-dihydropurines (13)
have never been applied in such studies. The azide 13 would
seem to be the ideal tool for probing the fate/role of
iminoquinone 7 in biological systems leading to gene
malfunction. One might argue that the 6-azido group of 13
would alter the usual Watson−Crick hydrogen bonding or
modify the normal syn/anti base conformer distribution. In
addition, it is well-known that azidopyrimidines undergo
azide−tetrazine tautomerization such as that shown for azide
13 in Scheme 8.45−47 However, azide 13 displays an intense
absorption at 2150 cm−1. So its azide tautomer 13 is the major,
if not the only, form present. Even so, both tautomers might be
expected to form the nitrene precursor of 7 upon irradiation at
254 nm.48 Thus, 13 rapidly forms the iminoquinone 7 upon
irradiation (<100 ps, Scheme 4), and 7 reacts with nucleophiles
so slowly (μs to ms) that there is adequate time for it to probe
multiple binding sites and react with nucleophiles in the most
suitable of those sites. Thus, while one might question the
Scheme 5 and Table 2 summarize the products formed in the
photolyses of azides 13Ac3 and 13Bz3 in methanolic solutions.
Because methanol by itself is a poor nucleophile, the
photodecomposition of azide 13Bz3 in pure methanol leads
to little, if any, formation of adduct 8OMeAc3. However,
photolysis of methanolic solutions of 13Ac3 and 13Bz3 in the
presence of the powerful nucleophile imidazole leads to the
formation of appreciable yields of imidazole adduct 8IMAc3
and 8IMBz3. The formation of 3Ac3 and 3Bz3 is apparently due
to the hydrolysis of the imino group in the iminoquinone 7Ac3
and 7Bz3, respectively, followed by reduction of the quinone 11
as outlined in Scheme 6.
The photochemical behavior of 8-azidoadenosine, 14, is also
of considerable interest, because it is one of the most widely
applied photoaffinity labeling agents used to cross-link many
biological molecules. Its preparative photochemistry has been
studied in detail26 and found to proceed through the
diiminoquinone 15. In this work, preparative irradiation of
azide 14 was carried out at 356 nm in water in the absence of
Scheme 5
3426
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
Table 2. Irradiation of Azides 13Ac3 and 13Bz3 in MeOH and 14 in Water
a
product distribution (%)
b
azide
solvent
MeOH
conditions, hν = 1 h/mmol
nuc
2
3
8IM
unidentified
c
13Bz3
254 nm
MeOH
30
Bz3
10
Ac3
0
30
8OMe
40
10
20
20
5 mmol/L
254 nm
Bz3
20
13Ac3
13Bz3
13Ac3
14
MeOH
MeOH
MeOH
imidazole
imidazole
imidazole
H2O
50
5 mmol/L
356 nm
Ac3
30
Ac3
50
5 mmol/L
356 nm
Bz3
Bz3
60
20
5 mmol/L
356 nm
Ac3
16
Ac3
H2O
17OH
∼10
17OH
18
0.5% HOAc
H2O
5 mmol/L
356 nm
∼7
16
∼20
18
∼20
∼20
14
H2O
17IM
0.5% HOAc
5 mmol/L
imidazole
∼20
∼10
∼10
∼20
a
The sugar esters tend to hydrolyze partially during the course of the reaction due to the basic alcoholic conditions of the imidazole reactions.
These products seem to constitute a complex mixture of over oxidation or coupling products. Adduct observed in only trace amounts.
b
c
Scheme 6
Scheme 7
Scheme 8
Scheme 9
viability of 13 as a mimic for the affinities of 8-oxoA (2), the
iminoquinone 7 should be an ideal mimic for 2.
Many 8-oxoA (2) and 8-oxoI (3) adducts are susceptible to
further oxidation, which ultimately leads to the destruction of
the purine nucleus, and the release of any cross-linked
molecules. While such further oxidation can be minimized by
the photochemical generation of the iminoquinone 7 from
azide 13, even this route suffers from adduct overoxidation
possibly as outlined in Scheme 9. This problem of oxidation of
cross-linked molecules requires further study and evaluation.
Further oxidation of amine adducts of 8-azidoadenosine (14)
such as the triamine 17NR2 is a major problem, becauase these
adducts are extremely susceptible to oxidation even by
atmospheric oxygen, which leads to ephemeral cross-links and
complex reaction mixtures. Therefore, further studies compar-
ing the stabilities of the very electron-rich amine adducts such
as 17NR2 with the much less electron-rich amine adducts of 13,
see Scheme 10, would seem to be most desirable. However, at
this time, it would seem safe to say that 8NR2 should be less
susceptible to further oxidation than 17NR2.
Finally, in this section, the chemical oxidation of 8-oxo-7,8-
dihydropurines has been investigated and shown to lead to
Scheme 10
3427
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
Scheme 11
Figure 1. Ultrafast transient absorption spectra of azide 14 in water solution at pH = 2.6 (A), 3.7 (B), and 9.2 (C). Time delays between pump and
probe pulses are expressed in picoseconds and given in legends.
nucleophile attack and substitution at the purine 2-position.
The photochemistry of 6- and 8-azidopurines leads to similar
products so long as the 2-position is unsubstituted in the
starting azide. Because this azide photochemistry provides an
excellent tool for the spectroscopic mechanistic study of the
formation of quinoidal purines and their chemistry, we have
undertaken such a study and report the results in the following
section.
Transient Absorption Spectroscopic Studies. Because
the azide 14 provides an ideal precursor for the diiminoquinone
15 (Scheme 11),26 the azide 13 would seem to provide an ideal
precursor to the iminoquinone 7, and the azide 20 to the
iminoquinone 21. We have studied the photochemistry of these
azides in aqueous media using ultrafast transient absorption
spectroscopy. The iminoquinone 7, derived from the oxidation
of 8-oxo-7,8-dihydroadenosine, 2, and the photochemistry of
the azide 13, is probably the most interesting of these, becauase
adenosine is the universal base found in DNA, RNA, and many
coenzymes. However, 8-azidoadenosine, 14, is a widely used
photoaffinity labeling agent, which has been shown to be a
precursor of the diiminoquinone 15 involved in the cross-
linking of this reagent to many biomolecules,49 and is included
in this study to contrast its ease of formation and the reactivity
with that of iminoquinones 7 and 21.
Photodecomposition of azide 14 following 270 nm excitation
has been studied previously by ultrafast transient absorption
spectroscopy with a 400 fs time resolution.26 In the current
work, transient absorption, ΔA, spectra of 13, 14, and 20 were
measured and studied following 300 nm excitation with a time
resolution of 200−300 fs (limited by solvent artifact, and, as a
result, solvent dependent, see Supporting Information). The
transient spectra of 14 were found to be strongly pH dependent
in the present work. At pH = 2.6 or lower, a band centered at
λmax = 390 nm develops within ca. 500 fs, spectrally narrows
indicative of vibrational relaxation of the responsible product
species (exponential time constants (τ) are wavelength
dependent ranging between ca. 0.5−10 ps), and remains
constant for at least 1 ns (Figure 1A). In contrast, at pH = 9.2
or higher, the formation of the 390 nm band was not observed
in aqueous solution of azide 14. Instead, a new band centered at
λmax = 465 nm develops within 300 fs, undergoes several-
picosecond spectral sharpening, and then survives for at least 1
ns (Figure 1C). Transient absorption spectra of azide 14,
recorded at intermediate pH values such as pH = 3.7 (Figure
1B), have a broad band centered at λmax = 440 nm. This band is
broader as compared to the bands at pH = 2.6 and 9.2 and
seems to be a combination of the 390 and 465 nm bands.
3428
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
Scheme 12
Table 3. Calculated Vertical Excitation Energies and Oscillator Strengths (f) for Intermediates 15Me, 23HMe, and 22Me in
a
Methanol
a
All calculations were performed in the gas phase.
Clearly two quite different intermediates are being generated
from azide 14 at different pH’s. Under acidic conditions, the 6-
amino group of 14 will be protonated, 14H in Scheme 12. The
nitrene 22H will not be stabilized by the powerful electron-
donating effect of the unprotonated amino group. Under these
conditions, rapid opening of the five-membered ring to the
iminonitrile 23H might occur.26,50−53 In contrast, under basic
conditions, the nitrene will be stabilized by the powerful
electron-donating effect of this amino group and becomes the
powerful base, 22, which will lead to proton reallocation to
form 15. The formation of 15 is complete within ca. 300 fs. The
resulting vibrationally hot diiminoquinone 15 then undergoes
vibrational cooling manifested by spectral narrowing over
several picoseconds, Figure 1C. Spectral assignments of
intermediates and products observed throughout the current
work were assisted by calculating vertical excitation transitions
(VETs) of these species at the TDDFT (B3LYP functional, 6-
31G(d,p) basis set) and CASPT2/ANO-RCC levels of
theory.54 The calculated absorption spectra (Scheme 12,
Table 3) support the aforementioned assignments. Thus, the
ring-cleaved iminonitrile 23H is calculated to have a major
electronic transition at 389 nm, Scheme 12, and this ring-
opening might easily occur within 500 fs, Figure 1A. In
contrast, the diiminoquinone 15 is calculated to display
absorption in the 430−450 nm region.26 The fast sub-300 fs
proton reallocation necessary to transform 22 in to 15 in
Scheme 12 apparently takes place via some type of proton relay
mechanism, which might be a coordinated process related to
that shown in Scheme 12. Calculations of energetics involved in
the transformation of 22 to 15 via such a proton relay process
(a relaxed reaction coordinated scan involving two water
molecules) indicated rather high energy barriers on the order of
18 kcal/mol for the proton-relay process shown in Scheme 12.
Therefore, if this type of proton reallocation is to take place
3429
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
Scheme 13
Figure 2. Ultrafast transient absorption spectra of acylated derivatives of azide 14 in nonprotic solvents. Column A: 14Bz5 (1 mM) in
dichloromethane. Column B: 14Ac5 (1 mM) in methylcyclohexane. Time delays in picoseconds between pump and probe pluses are given in the
legends. The 395 and 375 nm absorption bands develop within the first 300 fs after the excitation pulse, and they decay either with formation of two
absorption bands at 370 and 416 nm (τ = 2.5 ps, A) or very slowly on a several nanosecond time scale (B).
within the <300 fs window indicated, either proton tunneling
will have to play a very significant role, or the energetics of the
system must change appreciably upon solvation with more than
two water molecules in the formation of the species absorbing
at 465 nm.
group, and leads to five-membered ring fragmentation in both
14Bz5 and 14Ac5, Figure 2. Thus, excitation of 14Bz5 in a
nonprotic solvent leads to transient absorption bands at λmax
=
395 nm within ca. 300 fs. This band relaxes into two bands at
370 and 416 nm with τ = 2.5 ps (Figure 2A). In a similar
fashion, irradiation of 14Ac5 in a nonprotic solvent affords a
band centered λmax = 375 nm with τ = 1.5 ps that undergoes
little or no further change (Figure 2B). These bands centered at
λmax = 365−375 nm are assigned to corresponding nitrenes and
the mature bands in the 375−416 region to the corresponding
ring-opened iminonitriles. Apparently even acyl groups on the
6-amino group are sufficiently electron withdrawing to favor the
ring-opening pathway, if alternative protonation pathways are
not available.
Because there are apparently two alternative nitrene decay
pathways, one via five-membered ring cleavage and the other
via diiminoquinone formation, there must be two convergent
product forming pathways, because the same products are
apparently formed from both types of intermediates. Two such
alternative pathways for nucleophile attack are outlined in
Scheme 13. This same dichotomy has been discussed in detail
in previous papers,26,50−53 and little new can be added to that
discussion except to note that these two pathways are regulated
by the pH of the solution and other factors that alter the
availability of electron density on the nitrene nitrogen.
One such factor is the acylation of the 6-amino group, which
has an effect similar to that of the protonation of this amino
In previous work with electron-rich nitrenes, we observed
that the rate of nitrene protonation/relaxation was related to
the structure of the proton source with the process in methanol
being ca. 4 times faster than that in 2-propanol.32 This same
3430
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
diminioquinone 15 is only clearly visible in Figure 4B, the pH =
6.5 spectrum, λmax = 450 nm at 32 μs. It is formed in the
picosecond time domain and slowly reacts with nucleophiles
and rearomatizes within the microsecond−millisecond time
domain. It is this rearomatization process that is visible in
Figure 4A and C and Figure 5A and B. It is interesting to
compare the behavior of 8-azidoadenosine, 14, with that of 8-
azidoinosine, 20. The nitrene 25 generated from 20 is not
conjugated to a powerful electron-donating group, and
therefore might not undergo proton reallocation to form the
iminoquinone 21, Scheme 15. Thus, irradiation of 20 in water,
pH = 1.7−7.5, alcohols, or aprotic solvents such as dichloro-
methane all lead to the rapid development of a 380 nm band
characteristic of opening of the five-membered ring and
formation of the iminonitrile 26, Scheme 15. Because no
proton reallocation patterns could be observed under any
conditions examined, it would seem that irradiation of 20 leads
to opening of the five-membered ring to form 26 exclusively,
Scheme 15. It is worthy of note that during theoretical
structural optimization of nitrene 25 using either DFT or MP2
calculations, the C8−N9 bond always broke to form the
iminonitrile isomer 26. The most informative of these transient
experiments was that conducted in water at pH 7.5 (Figure
6A). The very broad absorption at 450−600 nm is assigned to
the nitrene 25 that decays with the same time constant as the
growth of the iminonitrile 26 band at 380 nm, τ = 7 ps. In
contrast, opening of the five-membered ring in methanol
(Figure 6B) and dichloromethane (Figure 6C) is more rapid
and apparently occurs in near-concert with nitrene formation,
because there is only a slight narrowing of the 380 nm peak,
which can be attributed to vibrational cooling of the open
iminonitrile 26. Furthermore, the iminonitrile 26Ac3 displays
spectroscopic behavior similar to that of the diiminoquinone 15
in the microsecond−millisecond time domains (Figures 7 and
4, respectively). Thus, both 26Ac3 and 23H react with water or
methanol within 15−20 ms. However, it seems that the
iminonitriles are better electrophiles than are the iminoqui-
nones: iminonitrile 26Ac3 reacts much more rapidly with
general trend is observed with azide 14, Figure 3. Thus, in
methanol, nitrene 22 (λmax = 460 nm) undergoes proton
reallocation/relaxation, which is characterized by an exponen-
tial rise-time constant of τ = 1.5 ps. While in 2-propanol, a
significantly longer reallocation/relaxation time constant is
observed, τ = 2.5 ps. Thus, proton transfer, Scheme 12,
proceeds with about the same very rapid rate in water and
methanol, but with a slightly slower rate in 2-propanol. In all of
these systems, this proton transfer ranks among the most rapid
yet observed.55−58
The reactions of diiminoquinone 15 with various nucleo-
philes were studied by nanosecond transient absorption
spectroscopy (Table 4). Thus, azide 14 was irradiated in
water at pH = 3.1, 6.5, and 11.5 (Figure 4), and in 0.1 M
aqueous solutions of imidazole (Figure 5A), a model
compound for histidine, and in 0.1 M sodium phenoxide
(Figure 5B), a model compound for tyrosine. According to
these spectra, quenching of diiminoquinone 15 under neutral
conditions, pH = 6.5, Figure 4B, is a relatively slow process
(Schemes 7, 13, and 14), with quenching of diiminoquinone 15
occurring in τ1 = ca. 100 μs, and subsequent aromatization of
the corresponding C2-adduct proceeding with τ2 > 35 ms.
However, these processes are accelerated under both acidic and
basic conditions to τ2 = ca. 20 ms at pH = 3.1 and τ2 = 4.2 ms at
pH 11.5. The reaction of 15 with 0.1 M aqueous sodium
phenoxide takes place with τ2 = 3 ms (Figure 5B) and more
rapidly with imidazole, τ2 = 900 μs (Figure 5A). These
nucleophilic substitution reactions all proceed through
intermediates similar to 24Nuc in Scheme 13. The initial
Figure 3. Ultrafast transient absorption spectra of azide 14. Column A: 14, 1 mM solution in methanol. Column B: 14, 0.75 mM in 2-propanol.
Time delays in picoseconds between pump and probe pluses are given in the legends. The major time constants describing the ΔA evolution shown
are τ = 1.5 ps (A) and τ = 2.5 ps (B).
3431
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
Table 4. Time Constants τ1 and τ2 for the Reactions of Purine Intermediates 15, 23H, and 26 under Various Conditions
Figure 4. Nanosecond−millisecond transient absorption spectra of azide 14 in water at different pH values.
Figure 5. Nanosecond−millisecond transient absorption spectra of azide 14 in water in the presence of (A) imidazole (0.1 M) and (B) sodium
phenoxide (0.1 M).
Scheme 14
imidazole, τ2 = 75 μs, than does the diiminoquinone 15, τ2 =
900 μs (Figures 7B and 5A, respectively). It is worthy of note
that the absorption associated with the reaction of imidazole
with the iminonitrile, 26Ac3, Figure 7B, derived from 20Ac3 is
unusually broad as compared to other transient absorptions
involving reactions of imidazole with iminoquinones 15 and 7
(Figure 5A or 10B, respectively). This extremely broad band
may be due to intermediates involved in the ring closure.
Azide 13 is perhaps the most interesting azide studied in this
work, because it is a precursor for the natural iminoquinone 7,
3432
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
Scheme 15
Figure 7. Microsecond−millisecond transient absorption spectra of
azide 20Ac3 in methanol (A) and in 0.1 M methanolic solution of
imidazole (B).
Scheme 16. Ring-opening does not seem to be a viable reaction
pathway for nitrene 27. Instead, nitrene 27 is stabilized via
proton reallocation. There are at least two possible mechanisms
for proton reallocation as outlined in Scheme 16. The first of
these is a direct 1,4-proton shift, and the second is a proton
reallocation mediated by a water or an alcohol molecule. The
nitrene 27 is expected to be of a similar basicity as nitrene 25,
and therefore not be particularly susceptible to protonation.
Consequently, the opening of the six-membered ring of nitrene
27 must be quite slow, if it occurs at all, for the slow proton
reallocation to compete effectively. In fact, protonation seems
to be the only nitrene stabilization process active in this system.
When azide 13Bz3 was photolyzed in alcohol solvents, Figure 8,
bands at λmax = 375 and 550625 nm are initially formed and
slowly shift to a new band centered at λmax = 405 nm. In all
cases, the broad band in the 550−625 nm region decays with
the same time constant as the growth of the 405 nm band.
Therefore, the 375 and 550−625 nm bands have been assigned
to the nitrene 27Bz3 and the 405 nm band assigned to the
iminoquinone 7Bz3 (Scheme 16 and Table 5). In methanol, the
proton reallocation time constant is τ = 90 ps and in 2-propanol
τ = 300 ps. These observations indicate that the solvent plays
an active role in proton reallocation, and therefore the direct
1,4-proton shift mechanism probably is not the only route for
the formation of the iminoquinone 7. This more rapid alcohol-
mediated route is further supported by femtosecond/pico-
second flash photolysis data for azide 13Bz3 in aprotic solvents,
Figure 9, in which the strong iminoquinone band of 7Bz3 at
λmax = 405 nm is not observed. In both methylcyclohexane and
dichloromethane, 13Bz3 affords bands at λmax = 345−350 nm
and a broad band in the 550−625 nm region. In dichloro-
methane, the short wavelength band tends to slowly shift to the
red to form a new band at ca. 400 nm with a time constant of τ
= 600 ps (Figure 9B), while the long wavelength band at 500−
650 nm remains stable. These observations are consistent with
the calculated absorption spectrum for 27 in Table 5. The short
wavelength band centered at λmax = 345−350 nm in
methylcyclohexane is assigned to an initially formed singlet
nitrene 27s, which is stable for at least 1 ns, Figure 9A. The
same short-time band is observed in dichloromethane, but a
new band slowly grows with λmax = 410 nm, τ = 600 ps, Figure
9B.
The broad absorption in the region 550−625 nm changes
little, indicating that these changes might be associated with
ISC from the singlet nitrene to the triplet nitrene 27t (λmax
=
546 nm). This spectral evolution also might be interpreted as
due to formation of both iminoquinone 7 (λmax = 405 nm)
produced by the unimolecular proton shift pathway in Scheme
16 and the triplet nitrene 27t via ISC. Triplet nitrenes such as
27t would be expected to react via abstraction of hydrogen
Figure 6. Ultrafast transient absorption spectra of azide 20 (2 mM) solutions in water at pH = 7.5, τ = 7 ps (nitrene 25 decay at 450−600 nm and
iminonitrile 26 growth at 380 nm) (A), methanol, τ < 300 fs (nitrene 25 formation and opening to iminonitrile 26) (B), and 20Ac3 in
dichloromethane, τ < 300 fs (nitrene 25Ac3 formation and opening to iminonitrile 26Ac3) (C). Time delays in picoseconds between pump and
probe pluses are given in the legends.
3433
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
Scheme 16
Figure 8. Ultrafast transient absorption spectra of azide 13Bz3 in MeOH, τ = 90 ps (singlet nitrene 27s decay at 375 nm and 550−625 nm, and
iminoquinone 7Bz3 formation at 405 nm) (A) and in 2-PrOH, τ = 300 ps (singlet nitrene 27s decay at 375 nm and 550−625 nm, and iminoquinone
7Bz3 formation at 405 nm) (B). Time delays between pump and probe pluses are given in picoseconds in the legends.
atoms forming the radical 29, which in turn would ultimately
undergo conversion to amine 2 that was isolated from
preparative reaction of azide 13Bz3, Schemes 5 and 17.
could have been among the unidentified products; see Table 2.
In fact, it is to be expected that nucleophile attack at the 2-
position would be the preferred mode of reaction for
unsubstituted purines, because this chemistry parallels that
previously observed for azide 14.26 In addition, the protonation
of the nitrenes generated from these purine azides closely
parallels the behavior of phenyl nitrenes that are activated by
electron-donating substituents.32 Apparently oxidative addition
of nucleophiles to the purine ring only occurs at the 5-position
when the 2-position is blocked by prior substitution as in the
case of guanosine. Therefore, once the 2-positions of 8-oxoA
(2) and 8-oxoI (3) become substituted, further oxidation
should parallel the well-studied 8-oxoG (1) pathway.25 We did
not observe guanidinohydantioin, 4, or spiroiminodihydantion,
5, Schemes 1 and 18, but the terminal oxidation products, N 1′-
ribosyl urea, 6, from the oxidation of 2, Scheme 2, and the 1-
ribosyl guanidine, 18, from the photoreaction of 14, Scheme 7,
are often observed. Verification of the mechanisms for these
advanced stages of purine oxidation will require the develop-
ment of photochemical methodologies for accessing the 8-
oxoG quinone system 9NH2. The development of these
photochemical tools is currently being investigated.
The microsecond−millisecond transient absorption spectra
for azide 13Bz3 are shown in methanol (Figure 10A) and in 0.1
M solution of imidazole in methanol (Figure 10B).
Iminoquinone 7Bz3 is quenched by methanol with a time
constant of τ > 30 ms (Figure 10A). This value is slower than
that observed for iminonitrile 26Ac3, τ = 15 ms (Figure 7A).
On the other hand, 7Bz3 is quenched by imidazole with a time
constant of τ = 20 μs (Figure 10B), which is the fastest of the
iminoquinone reactions studied in this work. This adduct
formation is associated with a broad adduct band with λmax
=
390 nm (Figure 10B and Scheme 17, 28NUC), but which
displays less broadening than the imidazole adduct bands of
other systems, Figures 5A and 7B.
Finally, in light of the tendency of guanine to undergo
oxidative attack by nucleophiles at the 5-purine ring position,
the exclusive attack of adenine-related systems at the 2-position
may be surprising. In no case was attack at the 5-position
observed in this work. However, in some cases the reaction
mixtures were quite complex, and adducts in the 5-position
3434
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
Table 5. Calculated B3LYP/6-31G(d) Vertical Excitation Energies and Oscillator Strengths (f) for Intermediates 27Me, 27t,
a
7Me, 28OMe, and 28IM
a
All calculations were performed in the gas phase.
Figure 9. Ultrafast transient absorption spectra of azide 13Bz3 in methylcyclohexane (A) and in dichloromethane (B). Time delays in picoseconds
between pump and probe pluses are given in the legends. In (B), following the spectral narrowing and reshaping process on a time scale of several
picoseconds, the spectral evolution displays a τ = 600 ps rise component.
systems.1,2 Burrows and Adam have invoked the quinoidal
CONCLUSIONS
A working hypothesis for the oxidation of purines was outlined
in Scheme 1 and is refined in Scheme 18. The initial oxidation
steps to form the 8-oxo-7,8-dihydropurines 1, 2, and 3 are well
established and have been observed in many biological
■
guanosine 9NH2 to explain the further oxidation of 1,27−31 and
the intermediacy of the related diiminoquinonone 15 has been
established in the photochemistry of purine azide 14.26 In this
work, we have focused upon the possible role of the purine
3435
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
Scheme 17
Figure 10. Microsecond−millisecond transient absorption spectra of azide 13Bz3 in pure methanol, 28OMe (A), and in 1 M solution of imidazole in
methanol, 28IM (B).
Scheme 18
iminoquinone 7 and quinone 11 in the oxidation of adenosine
and inosine, respectively. The observation of oxidative
nucleophilic addition to the 2-position of 2 and 3 in the
presence of imidazole supports the intermediacy of 7 and 11,
and the general hypothesis outlined in Scheme 18.
Furthermore, the intermediacy of 7 and its conversion to 8
has been examined in detail using femtosecond and nano-
second transient absorption spectroscopy. These observations
strongly support the involvement of the iminoquinone 7 in the
further oxidation of 2. Because 2 is derived from the oxidation
3436
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
of adenosine (Eox = 1.42 V/NHE), and has been widely
observed in biological systems, one can assume that the further
oxidation of 2 to 7 (Eox = 0.92 V/NHE) can proceed in living
systems as well, and might lead to a wide variety of cross-linking
processes in biological systems.
The literature has focused upon the further oxidation of 8-
oxo-7,8-dihydroguanosine (1) because it is more easily formed
than 8-oxo-7,8-dihydroadenosine (2). In fact, it has been
estimated that the formation of about 105 oxidative lesions
related to 1 and 2 are formed and repaired each day in the
average rat cell.59 Furthermore, because the ratio of 1/2 is
about 3/1,21 the oxidation of adenosine might possibly be of
equal or greater significance than that of guanosine. The relative
significance of these two types of oxidative lesions might well
be biased by the fact that the adenosine lesions tend to form
much more stable, less readily repaired, cross-links than do
guanosine lesions.
The observation that imidazole forms especially stable
adducts with oxidized adenosine deserves comment. The
amino acid histidine occurs in the active sites of numerous
enzymes, many of which interact with DNA.60,61 Two
particularly interesting cases are His-365 found in the active
site of Escherichia coli DNA topoisomerase I62 and the seryl-
histidine unit found in the active sites of many hydrolytic
enzymes such as the intein−extein junctions of homing
endonucleases such as DNAase I.63 The interaction of these
enzymes that function in the routine maintenance of DNA with
oxidatively damaged DNA would seem to be a most interesting
area for future investigation.
and photoproducts. This material is available free of charge via
the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
atarnov@bgsu.edu; rmw@bgsu.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Dr. Larry Sallans of the Mass Spectrometry Facility at
the University of Cincinnati for providing us with detailed mass
spectrometric analyses. Financial support from the Ohio Super
Computer Center and the NSF (ANT, Career Award, CHE-
0847707) and the R. Marshall and Antonia G. Wilson
Chemistry Fund is acknowledged. We express our appreciation
to Prof. Felix N. Castellano and Valentina A. Prusakova for
their help in acquisition of nanosecond spectroscopy data. We
would like to thank Prof. Massimo Olivucci and Elena V.
Laricheva for their help in theoretical calculations and Prof.
Ksenija Glusac, Dr. Pavel Kucheryavy, and Ekaterina
Mirzakulova for their help in electrochemistry experiments.
REFERENCES
■
(1) Wang, D.; Kreutzer, D. A.; Essigmann, J. M. Mutat. Res. 1998,
400, 99−115.
(2) Cadet, J.; Douki, T.; Gasparutto, D.; Ravanat, J.-L. Mutat. Res.
2003, 531, 5−23.
(3) Barnes, D. E.; Lindahl, T. Annu. Rev. Genet. 2004, 38, 445−476.
(4) Nakabeppu, Y.; Sakumi, K.; Sakamoto, K.; Tsuchimoto, D.;
Tsuzuki, T.; Nakatsu, Y. Biol. Chem. 2006, 387, 373−379.
(5) Wilson, D. M.; Bohr, V. A. DNA Repair 2007, 6, 544−559.
(6) Finkel, T.; Holbrook, N. J. Nature 2000, 408, 239−247.
(7) Feig, D. I.; Loeb, L. A. J. Mol. Biol. 1994, 235, 33−41.
(8) Wiseman, H.; Halliwell, B. Biochem. J. 1996, 313, 17−29.
(9) Steenken, S.; Jovanovic, S. V. J. Am. Chem. Soc. 1997, 119, 617−
618.
While the work described in this Article significantly extends
our understanding of purine oxidation pathways by establishing
the viability of quinoidal purine intermediates, it does not
address the questions of bridges between the various purine
families and the latter stages of purine oxidation to urea 6 and
parabanic acid64,65 shown as dashed arrows in Scheme 18.
Further work will be required to elucidate the possible oxidative
pathways between 11 and 1, 8 and 9, and the ultimate N 1′-
ribosyl urea 6.
Finally, the azide 14 is one of the most widely used
photoaffinity labeling agents. The observations reported here
further support the intermediacy of the diiminoquinone 15 as
the pivotal intermediate in this chemistry.26 Furthermore,
comparison of the transient absorption spectroscopy of 14 with
that of 13 is consistent with previously studied nitrene
chemistry where it was observed that singlet nitrenes in
electronic communication with electron-donating groups
become powerful bases and protonate rapidly.32 Thus, in this
work, the activated basic nitrene 22 can be protonated within
the first 300 fs, while the deactivated nitrene 27 requires 90−
300 ps (Figure 8) for protonation. On the other hand, once
protonated, iminoquinone 7 is a more powerful electrophile
than the diiminoquinone 15, and thus might well be a more
effective cross-linking agent and PAL agent in many biological
systems. Work is continuing to further characterize the
chemical reactivity of iminoquinone 7 in duplex nucleic acid
environments.
(10) Solivio, M. J.; Joy, T. J.; Sallans, L.; Merino, E. J. J. Inorg.
Biochem. 2010, 104, 1000−1005.
(11) Solivio, M. J.; Nemera, D. B.; Sallans, L.; Merino, E. J. Chem. Res.
Technol. 2012, 25, 326−336.
(12) Tan, X.; Grollman, A. P.; Shibutani, S. Carcinogenesis 1999, 20,
2287−2292.
(13) von Sonntag, C. Free-Radical-Induced DNA Damage and Its
Repair: A Chemical Perspective; Springer-Verlag: New York, 2006; pp
211−447.
(14) Bjelland, S.; Seeberg, E. Mutat. Res. 2003, 531, 37−80.
(15) Dirks, A. J.; Hofer, T.; Marzetti, E.; Pahor, M.; Leeuwenburgh,
C. Aging Res. Rev. 2006, 5, 179−195.
(16) Hofer, T.; Seo, A. Y.; Prudencio, M.; Leeuwenburgh, C. Biol.
Chem. 2006, 387, 103−111.
(17) Hofer, T.; Badouard, C.; Bajak, E.; Ravanat, J.-L.; Mattsson, A.;
Cotgreave, I. A. Biol. Chem. 2005, 386, 333−337.
(18) Dizdaroglu, M. Mutat. Res. 2003, 531, 109−126.
(19) Michaels, M. L.; Tchou, J.; Grollman, A. P.; Miller, J. H.
Biochemistry 1992, 31, 10964−10968.
(20) Jaruga, P.; Dizdaroglu, M. Nucleic Acids Res. 1996, 24, 1389−
1394.
(21) Llano, J.; Eriksson, L. A. Phys. Chem. Chem. Phys. 2004, 6,
4707−4713.
ASSOCIATED CONTENT
* Supporting Information
Catalog of chemical structures with their structure numbers,
and a description of the instrumentation used to obtain ultrafast
femto- and picosecond data. Syntheses of starting materials and
characterization of synthetic intermediates, starting materials,
■
(22) Ito, T.; Kuno, S.; Uchida, T.; Fujita, J.-i.; Nishimoto, S.-i. J. Phys.
Chem. B 2009, 113, 389−394.
S
(23) Jensen, A.; Calvayrac, G.; Karahalil, B.; Bohr, V. A.; Stevnsner,
T. J. Biol. Chem. 2003, 278, 19541−19548.
(24) Grin, I. R.; Dianov, G. L.; Zharkov, D. O. FEBS Lett. 2010, 584,
1553−1557.
3437
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438

Journal of the American Chemical Society
Article
(25) Perrier, S.; Hau, J.; Gasparutto, D.; Cadet, J.; Favier, A.; Ravanat,
J.-L. J. Am. Chem. Soc. 2006, 128, 5703−5710.
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
revision A.1; Gaussian, Inc.: Wallingford, CT, 2009. (b) MOLCAS
̈
(26) Polshakov, D.; Rai, S.; Wilson, R. M.; Mack, E. T.; Vogel, M.;
Krause, J. A.; Burdzinski, G.; Platz, M. S. Biochemistry 2005, 44,
11241−11253.
7.4: Aquilante, F.; De Vico, L.; Ferre,
Neogrady, P.; Pedersen, T. B.; Pitonak
Serrano-Andres
́
N.; Ghigo, G.; Malmqvist, P.;
, M.; Reiher, M.; Roos, B. O.;
, L.; Urban, M.; Veryazov, V.; Lindh, R. J. Comput.
(27) Xu, X.; Fleming, A. M.; Muller, J. G.; Burrows, C. J. J. Am. Chem.
Soc. 2008, 130, 10080−10081.
́
̌
́
́
(28) Johansen, M. E.; Muller, J. G.; Xu, X.; Burrows, C. J.
Biochemistry 2005, 44, 5660−5671.
Chem. 2010, 31, 224−247.
(55) Peon, J.; Polshakov, D.; Kohler, B. J. Am. Chem. Soc. 2002, 124,
6428−6438.
(29) Hosford, M. E.; Muller, J. G.; Burrows, C. J. J. Am. Chem. Soc.
2004, 126, 9540−9541.
(56) Dix, E. J.; Goodman, J. L. J. Phys. Chem. 1994, 98, 12609−
(30) Ye, Y.; Muller, J. G.; Luo, W.; Mayne, C. L.; Shallop, A. J.; Jones,
R. A.; Burrows, C. J. J. Am. Chem. Soc. 2003, 125, 13926−13927.
12612.
(57) Pines, E.; Pines, D.; Barak, T.; Magnes, B. Z.; Tolbert, L. M.;
Haubrich, J. E. Ber. Bunsen-Ges. 1998, 102, 511−517.
(58) Lima, J. C.; Abreu, I.; Brouillard, R.; Macanita, A. L. Chem. Phys.
Lett. 1998, 298, 189−195.
(31) Adam, W.; Arnold, M. A.; Nau, W. M.; Pischel, U.; Saha-Moller,
̈
C. R. J. Am. Chem. Soc. 2002, 124, 3893−3904.
(32) Voskresenska, V.; Wilson, R. M.; Panov, M.; Tarnovsky, A. N.;
Krause, J. A.; Vyas, S.; Winter, A. H.; Hadad, C. M. J. Am. Chem. Soc.
2009, 131, 11535−11547.
(59) Park, E.-M.; Shigenaga, M. K.; Degan, P.; Korn, T. S.; Kitzler, J.
W.; Wehr, C. M.; Kolachana, P.; Ames, B. N. Proc. Natl. Acad. Sci.
U.S.A. 1992, 89, 3375−3379.
(33) Mack, E. T. University of Cincinnati, unpublished results.
(34) Yanagawa, H.; Ogawa, Y.; Ueno, M. J. Biol. Chem. 1992, 267,
13320−13326.
(60) Luscombe, N. M.; Laskowski, R. A.; Thornton, J. M. Nucleic
Acids Res. 2001, 29, 2860−2874.
(35) Steenken, S.; Jovanovic, S. V.; Bietti, M.; Bernhard, K. J. Am.
Chem. Soc. 2000, 122, 2373−2374.
(61) Leavens, F. M. V.; Churchill, C. D. M.; Wang, S.; Wetmore, S.
D. J. Phys. Chem. B 2011, 115, 10990−11003.
(36) Adam, W.; Saha-Moller, C. R.; Schonberger, A.; Berger, M.;
̈
̈
(62) Perry, K.; Mondragon
13245.
́
, A. J. Biol. Chem. 2002, 277, 13237−
Cadet, J. J. Photochem. Photobiol. 1995, 62, 231−238.
(37) Sheu, C.; Foote, C. S. J. Am. Chem. Soc. 1995, 117, 474−477.
(38) Sheu, C.; Foote, C. S. J. Am. Chem. Soc. 1995, 117, 6439−6442.
(39) Raoul, S.; Cadet, J. J. Am. Chem. Soc. 1996, 118, 1892−1898.
(63) Li, Y.; Hatfield, S.; Li, J.; McMills, M.; Zhao, Y.; Chen, X. Bioorg.
Med. Chem. 2002, 10, 667−673.
(64) Vialas, C.; Claparols, C.; Pratviel, G.; Meunier, B. J. Am. Chem.
Soc. 2000, 122, 2157−2167.
(40) Adam, W.; Saha-Moller, C. R.; Schonberger, A. J. Am. Chem. Soc.
̈
̈
1996, 118, 9233−9238.
(65) Jena, N. R.; Mishra, P. C. Free Radical Biol. Med. 2012, 53, 81−
94.
(41) Duarte, V.; Gasparutto, D.; Yamaguchi, L. F.; Ravanat, J.-L.;
Martinez, G. R.; Medeiros, M. H. G.; Di Mascio, P.; Cadet, J. J. Am.
Chem. Soc. 2000, 122, 12622−12628.
(42) Luo, W.; Muller, J. G.; Rachlin, E. M.; Burrows, C. J. Chem. Res.
Toxicol. 2001, 14, 927−938.
(43) Duarte, V.; Muller, J. G.; Burrows, C. J. Nucleic Acids Res. 1999,
27, 496−502.
(44) Luo, W.; Muller, J. G.; Rachlin, E. M.; Burrows, C. J. Org. Lett.
2000, 2, 613−616.
(45) Temple, C., Jr.; Montgomery, J. A. J. Org. Chem. 1965, 30, 826−
829.
(46) Temple, C., Jr.; Coburn, W. C., Jr.; Thorpe, M. C.;
Montgomery, J. A. J. Org. Chem. 1965, 30, 2395−2398.
(47) Temple, C., Jr.; Thorpe, M. C.; Corburn, W. C., Jr.;
Montgomery, J. A. J. Org. Chem. 1966, 31, 935−938.
(48) Addicott, C.; Wong, M. W.; Wentrup, C. J. Org. Chem. 2002, 67,
8538−8546.
(49) Marburg, S.; Jorn, D.; Tolman, R. L. J. Heterocycl. Chem. 1982,
19, 671−672.
(50) Gadosy, T. A.; McClelland, R. A. J. Am. Chem. Soc. 1999, 121,
1459−1465.
(51) Dehaen, W.; Becher, J. Acta Chem. Scand. 1993, 47, 244−254.
(52) Funicello, M.; Spagnolo, P.; Zanirato, P. Acta Chem. Scand.
1993, 47, 231−243.
(53) Spinelli, D.; Zanirato, P. J. Chem. Soc., Perkin Trans. 2 1993,
1129−1133.
(54) Assignment of spectral bands was supported by theoretical
calculations using Gaussian 09 and MOLCAS 7.4 software: (a) Frisch,
M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.;
Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson,
G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov,
A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.;
Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda,
Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J.
E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.;
Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
3438
dx.doi.org/10.1021/ja3068148 | J. Am. Chem. Soc. 2013, 135, 3423−3438