Generation of 2′-deoxyadenosine N6-aminyl radicals from the photolysis of phenylhydrazone derivatives

Article abstract of DOI:10.1021/tx900268r

Nitrogen-centered radicals are major species generated by the addition of hydroxyl radicals and the one-electron oxidation of adenine derivatives. Aminyl radicals are also generated in the decomposition of adenine chloramines upon reaction of hypochlorite. Here, we report the photochemistry of modified 2′-deoxyadenosine (dAdo) containing photoactive hydrazone substituents as a model to investigate the chemistry of dAdo N⁶-aminyl radicals. Derivatives of dAdo containing a phenylhydrazone moiety at N6 displayed UV absorption between 300 and 400 nm. Upon UV photolysis in the presence of a H-donor, that is, glutathione, two major products were formed, dAdo and benzaldehyde, indicating efficient homolytic cleavage to dAdo N ⁶-aminyl radicals and benzylidene iminyl radicals. dAdo N ⁶-phenylhydrazone was photolyzed in the presence of a molar excess of nonmodified dAdo to mimic the reactions taking place in DNA, and the major photoproducts were identified by high-performance liquid chromatography, mass spectrometry, and nuclear magnetic resonance. The formation of 2-(benzylideneamino)-2′-deoxyadenosine as well as a more extensive oxidation product may be explained by the recombination of initial dAdo N ⁶-aminyl and benzylidene iminyl radicals. The formation of 2′-deoxyinosine may be explained by hydrolytic deamination of dAdo N ⁶-aminyl radicals. Interestingly, a dimeric product containing two dAdo moieties was identified in the photolysis mixture. The present studies demonstrate the ability of dAdo N⁶-aminyl radicals to undergo H-abstraction to give dAdo, deamination to give 2′-deoxyinosine, and addition to the adenine moiety to give dimers.

Full text of DOI:10.1021/tx900268r

48
Chem. Res. Toxicol. 2010, 23, 48–54
Generation of 2′-Deoxyadenosine N6-Aminyl Radicals from the
Photolysis of Phenylhydrazone Derivatives
Vandana Kuttappan-Nair, Francois Samson-Thibault, and J. Richard Wagner*
De´partement de Me´decine Nucle´aire et Radiobiologie, Faculte´ de Me´decine, 3001 12e AVenue Nord,
UniVersite´ de Sherbrooke, Sherbrooke, Que´bec, Canada J1H 5N4
ReceiVed August 3, 2009
Nitrogen-centered radicals are major species generated by the addition of hydroxyl radicals and the
one-electron oxidation of adenine derivatives. Aminyl radicals are also generated in the decomposition
of adenine chloramines upon reaction of hypochlorite. Here, we report the photochemistry of modified
2′-deoxyadenosine (dAdo) containing photoactive hydrazone substituents as a model to investigate the
chemistry of dAdo N6-aminyl radicals. Derivatives of dAdo containing a phenylhydrazone moiety at N6
displayed UV absorption between 300 and 400 nm. Upon UV photolysis in the presence of a H-donor,
that is, glutathione, two major products were formed, dAdo and benzaldehyde, indicating efficient homolytic
cleavage to dAdo N6-aminyl radicals and benzylidene iminyl radicals. dAdo N6-phenylhydrazone was
photolyzed in the presence of a molar excess of nonmodified dAdo to mimic the reactions taking place
in DNA, and the major photoproducts were identified by high-performance liquid chromatography, mass
spectrometry, and nuclear magnetic resonance. The formation of 2-(benzylideneamino)-2′-deoxyadenosine
as well as a more extensive oxidation product may be explained by the recombination of initial dAdo
N6-aminyl and benzylidene iminyl radicals. The formation of 2′-deoxyinosine may be explained by
hydrolytic deamination of dAdo N6-aminyl radicals. Interestingly, a dimeric product containing two dAdo
moieties was identified in the photolysis mixture. The present studies demonstrate the ability of dAdo
N6-aminyl radicals to undergo H-abstraction to give dAdo, deamination to give 2′-deoxyinosine, and
addition to the adenine moiety to give dimers.
chemistry of adenine N6-aminyl radicals is very similar to that
Introduction
of cytosine N4-aminyl radicals, which undergo addition to the
The oxidation of adenine by ionizing radiation involving
5,6-double bond of nonmodified cytosine to give dimers (9, 10).
either hydroxyl radicals or base radical cations remains poorly
understood (1-3). The initial reaction of hydroxyl radicals with
adenine derivatives was investigated by pulse radiolysis and
product analysis (4, 5). The main reaction pathway involves
addition to C4 of the adenine moiety followed by dehydration
to oxidizing adenine N6-aminyl radicals. Other reactions include
addition to C8, addition to C5, and H-atom abstraction from
the 2-deoxyribose moiety. Adenine N6-aminyl radicals are also
the main species arising from one-electron oxidation and
deprotonation of adenine radical cations. Thus, the reaction of
hydroxyl radicals with adenine as well as the one-electron
oxidation of adenine generate adenine N6-aminyl radicals as the
predominant radical species. Interestingly, the yield of damage
arising from the oxidation of adenine derivatives either in
solution or in DNA exposed to radiation is much lower than
damage to other nucleobases (4, 6, 7). The reason for such a
resistance of adenine toward oxidative damage is not clear. It
has been suggested that adenine N6-aminyl radicals efficiently
undergo chemical repair by reaction with reducing radicals (4);
however, this reaction is unlikely in DNA in which radicals
become immobilized within the polymer. Alternatively, it is
reasonable to propose that adenine N6-aminyl radicals react with
other DNA base moieties, leading to intra- and interstrand cross-
links. The photosensitized oxidation of polyA by 2-methyl-1,4-
naphthoquinone induced an increase of Rayleigh scattering,
pointing to the formation of cross-links in polyA (8). The
Recent studies suggest that adenine may be a key player in the
formation of interstrand cross-links by initial 5-methyl-(uracilyl)
radicals (11, 12).
Hypochlorite (HOCl) is a strong oxidant, which is generated
during inflammation by myeloperoxidase enzyme in the presence
of H2O2 and chloride ions. The reaction of adenosine mono-
phosphate (AMP) with HOCl was shown to generate several
nitrogen-centered radicals at N1, N3, and N6 of adenine through
the decomposition of intermediate chloramines (13). The
propensity of radical formation from the chloramines of
nucleosides and polynucleotide followed C > A > G > T, as
inferred by electron paramagnetic resonance spin trapping
(14, 15). The major radical adducts from a mixture of nucleo-
sides appeared to be nitrogen-centered radicals at the exocyclic
amino positions of C and A, together with carbon-centered
radicals, likely arising from the addition of aminyl radicals to
nonmodified nucleosides. HOCl has also been shown to induce
single and double strand breaks in plasmid DNA (16).
Several studies have demonstrated the ability of free alkyl
and aryl aminyl radicals to induce oxidative DNA damage. The
formation of 8-oxo-7,8-dihydroguanine and piperidine-labile
breaks in double-stranded DNA induced by Cu2+/H2O2 was
greatly enhanced in the presence of acrylonitrile, which
undergoes oxidation to iminyl radicals in the reaction mixture.
Interestingly, the profile of damage shifted from a random
distribution without acrylonitrile to one localized on GG and
GGG residues in the presence of acrylonitrile (17). Alternatively,
the oxidation of formamide to dimethylaminyl radicals by Cu2+/
* To whom correspondence should be addressed. Tel: 1-819-820-6868
ext. 12717. E-mail: richard.wagner@usherbrooke.ca.
10.1021/tx900268r  2010 American Chemical Society
Published on Web 12/10/2009

Generation and Fate of dAdo N6-Aminyl Radicals
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 49
Scheme 1. Synthesis of dAdo N6-Phenylhydrazones (3)
performed with an automated syringe at a flow of 0.45 µL/min
and 2.2 kV electrospray voltage. The cone voltage was set to 45
V, and the instrument was calibrated by infusing a standard NaI
solution (0.2 mg/mL) in 50% aqueous isopropanol. All of the
experiments were conducted in the positive ionization mode.
N6-[Benzaldehyde hydrazone]-2′-deoxyadenosine (3a) [Scheme
1, Method 1 (21)]. Compound 1 (30 mg, 0.11 mmol) was dissolved
in dimethyl sulfoxide (DMSO; 0.32 mL), and DIPEA (0.5 mL)
was added slowly, followed by benzaldehyde hydrazone (2a, 40
mg, 0.33 mmol). The reaction mixture was stirred at 60 °C for
46 h and the progress of the reaction was monitored by HPLC-
UV. The final yield of 3a was 39% after purification by HPLC
using standard conditions (note: other unidentified products formed
in the reaction medium). 1H NMR (300 MHz, acetone-d6) δ 10.55
(s, 1H, NH), 8.41 (s, 1H, NdC2H), 8.33-8.35 (m, 2H, N)CH +
NdC8H), 7.84-7.87 (m, 2H, ArH), 7.40-7.48 (m, 3H, ArH), 6.48
(m, 1H, H1′), 5.34 (m, 1H, OH3′), 4.64 (m, 1H, OH5′), 4.44 (m, 1H,
H3′), 4.08 (m, 1H, H4′), 3.70-3.83 (m, 2H, H5′ and H5′′), 2.87-2.93
(m, 1H, H2”), 2.33-2.38 (m, 1H, H2′). 13C NMR (100 MHz, DMSO-
d6) δ 152.2 (C2), 151.7 (NdC(hydrazone)), 150.3 (C4), δ 144.3
(C6), 140.8 (C8), δ 134.9 (C1-benzylidene), 129.3 (C4-benzylidene),
128.7 (C2,C6-benzylidene), 126.7 (C3,C5-benzylidene), 118.9 (C5),
88.0 (C4′), 83.8 (C1′), 70.8 (C3′), 61.8 (C5′), 39.5 (C2′). ESI-MS/MS
(50% acetonitrile, 1 mM NaCl, positive mode): m/z 377.0 (M +
Na+), m/z 355.0 (M+ H+), m/z 239.0 (M-deoxyribose + Na+).
N6-[4-Methoxybenzaldehyde hydrazone]-2′-deoxyadenosine
(3b) [Scheme 1, Method 1 (21)]. Compound 1 (18 mg, 0.065
mmol) was dissolved in DMSO (0.32 mL) and DIPEA (0.33 mL)
was added slowly followed by 4-methoxybenzaldehyde hydrazone
(2b, 32 mg, 0.2 mmol). The reaction mixture was stirred at 60 °C
for 46 h and the progress of the reaction was monitored by HPLC-
UV. The final yield of 3b was 18% after purification by HPLC
using standard conditions (Note that several other unidentified
products were formed in the reaction medium). 1H NMR (300 MHz,
acetone-d6) δ 10.5 (s, 1H, NH), 8.32- 8.35 (m, 3H, N)CH +
NdC2H + NdC8H), 7.7- 7.8 (m, 2H, ArH), 7.0- 7.02 (m, 2H, ArH),
6.48 (m, 1H, H1′), δ 5.4 (br, 1H, OH3′), 4.64 (m, 1H, H3′), 4.4 (br,
1H, OH5′), 4.08 (m, 1H, H4′), 3.83- 3.86 (m, 5H, H5′+H5′′ + OCH3,
2.90 (m, 1H, H2”), 2.37 (m,1H, H2′). 13C NMR (100 MHz, DMSO-
d6) δ 160.4 (C4-benzylidene), 152.2 (C2), 151.7 (C6), 150.2 (C4),
144.4 (NdC(hydrazone)), 140.6 (C8), 128.3 (C2,C6-benzylidene),
127.5 (C1-benzylidene), 118.8 (C5), 114.3 (C3,C5-benzylidene),
88.0 (C4′), δ 83.8 (C1′), 70.9 (C3′), 61.8 (C5′), 55.3 (OCH3-
benzylidene), 39.5 (C2′). ESI- MS/ MS (50% acetonitrile, 1 mM
NaCl, positive mode): m/z 407 (M+ Na+), m/z 291 (M-deoxyribose
+ Na+), m/z 157 (M-adenine+Na+).
H2O2 induced DNA damage favorably at cytosine and thymine
residues as revealed by piperidine-labile breaks in double-
stranded DNA (18). The photolysis of p-cyanobenzoyl an-
thraquinone oxime ester was reported to induce single and
double strand breaks in supercoiled plasmid DNA under
anaerobic conditions, presumably in part by way of the
generation of iminyl radicals of anthraquinone (19). Alkylaminyl
radicals also appear to contribute to the formation of breaks in
plasmid DNA by near-UV-induced photoionization of naphazo-
line (20). Thus, freely diffusing aminyl and iminyl radicals are
able to oxidize DNA bases to give oxidation products, such as
8-oxo-7,8-dihydroguanine, and to induce strand breaks probably
by H-atom abstraction from the sugar moiety.
In the present work, we have prepared three novel modifica-
tions of 2′-deoxyadenosine (dAdo) containing photoactive
hydrazone substituents at N6 as a means to specifically generate
N6-aminyl (and tautomeric N1-iminyl) radicals of dAdo. Thereby,
we report the chemistry of dAdo N6-aminyl radicals in the
presence of a good H-donor, that is, glutathione (GSH), and a
high concentration of nucleoside, dAdo, to mimic the possible
reactions under biological conditions.
Experimental Procedures
All chemicals were purchased from Sigma-Aldrich and were of
reagent grade unless stated otherwise. 6-Chloro-2′-deoxypurine (1)
was purchased from Berry & Associates, Inc. (Dexter, MI). High-
performance liquid chromatography with ultraviolet detection
(HPLC-UV) was carried out using Alliance systems (Waters 2795
or 2690) connected to dual wavelength UV detectors (Waters 2487)
and a Millenium workstation (Waters version 4). The standard
conditions for the analysis of hydrazones by HPLC-UV with a
reversed phase column (5 µm ODS A 250 mm × 6 mm; YMC)
included a gradient starting at 100% of mobile phase A [triethy-
lammonium acetate (25 mM) at pH 7 containing 2.5% acetonitrile]
and going to 60% A and 40% mobile phase B (95% acetonitrile)
in 30 min at a flow rate of 1 mL/min. 1H NMR measurements were
obtained using Bruker 300 MHz and Varian 600 MHz spectrom-
eters. Electrospray ionization-tandem mass spectrometry (ESI-MS/
MS) was carried out using a Q-tof-2 instrument, and matrix-assisted
laser desorption/ionization-time-of-flight (MALDI-TOF) was car-
ried out on a Tof Spec 2E instrument, both from Micromass. For
ESI-MS/MS analysis, samples were dissolved in either 50%
CH3CN/H2O or 50% acetonitrile/D2O solutions containing 1 mM
NaCl. Continuous injection was done with an automated syringe
at a flow of 0.6 µL /min. ESI-MS and ESI-MS/MS analyses of the
crude reaction mixture and purified compounds were carried out
using a Q-Tof-2 instrument (Waters). The samples were diluted in
H2O/25% CH3CN/0.1% HCOOH. Continuous injections were
N6-[4-Nitrobenzaldehyde Hydrazone]-2′-deoxyadenosine (3c)
[Scheme 1, Method 1 (21)]. Compound 1 (30 mg, 0.11 mmol)
was dissolved in DMSO (0.32 mL) and DIPEA (0.5 mL) was added
slowly followed by 4-nitrobenzaldehyde hydrazone (2c, 182 mg,
1.1 mmol). The reaction mixture was stirred at 60 °C for 4 days.

50 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
Kuttappan-Nair et al.
The progress of the reaction and subsequent purification of product
was carried out by HPLC using standard conditions. The final yield
of 3c after purification was 38%. 1H NMR (300 MHz, acetone-d6)
δ 10.9 (s, 1H, NH), 8.1- 8.6 (m, 7H, N)CH + NdC (2) H +
ArH), 6.51 (m, 1H, H1′), 5.19 (br, 1H, OH3′), 4.66 (m, 1H, H3′),
4.45 (m, 1H, H3′), 4.1(m, 1H, H4′), 3.71- 3.77 (m, 2H, H5′ + H5′′),
2.94 (m, 1H, H2”), 2.42 (m, 1H, H2′). 13C NMR (100 MHz, DMSO-
d6) δ 152.1 (C2), 151.5 (C6), 150.7 (C4), 147.3 (C4-benzylidene),
141.6 (CdN(hydrazone)), 141.4 (C1-benzylidene), 141.4 (C8), 127.4
(C2,C6-benzylidene), 124.1 (C3,C5- benzylidene), 119.2 (C5), 88.0
(C4′), 83.7 (C1′), 70.8 (C3′), 61.7 (C5′), 39.5 (C2′). ESI- MS/ MS
(50% acetonitrile, 1 mM NaCl, positive mode): m/z 422 (M+ Na+),
m/z 306 (M-deoxyribose + Na+).
Aldehyde Hydrazones 2a-c (22). Benzaldehyde hydrazone
(2a). Two hundred milligrams (1 equiv) of benzaldehdye in DMSO
(1.9 mL) was added dropwise to 94 mg (1.88 mmol) of hydrazine
monohydrate, and the solution was stirred for 20 min at room
temperature.
4-Methoxybenzaldehyde hydrazone (2b). Two hundred mil-
ligrams (1 equiv) of benzaldehdye in DMSO (1.9 mL) was added
dropwise to 73.1 mg (1.46 mmol; 1 equiv) of hydrazine monohy-
drate, and the reaction was stirred for 5 h at room temperature.
4-Nitrobenzaldehyde hydrazone (2c). The synthesis of 2c was
the same as for 2a except that it required 24 h to complete the
reaction. The formation of aldehyde hydrazones was monitored by
HPLC using standard conditions.
Figure 1. UV spectra of phenylhydrazone derivatives of dAdo (0.066
mM) in 10 mM phosphate buffer (pH 7.0) at room temperature. (a)
Benzaldehydephenylhydrazone (3a), (b) p-methoxy-benzaldehyde hy-
drazone (3b), and (c) nitrobenzaldehydehydrazone (3c).
continuous bubbling during photolysis. The three substituted dAdo
hydrazones displayed strong UV absorption between 300 and 400
nm in contrast to dAdo, whose absorption cuts off at below 280
nm. Quantum yields were estimated from the loss of parent
hydrazone and/or the formation of dAdo as measured by quantitative
HPLC divided by the number of incident photons on the solution
as determined by a photometer (Industrial Fiber Optics, Inc., United
States). The average number of incident photons was 0.85 × 10 16
photons s-1. The initial concentration of hydrazone was adjusted
to correspond to an optical density of 1.0 at the wavelength of
photolysis (313 nm), for example, the concentration of 3a was 0.05
mM.
N6-Hydrazino-2′-deoxyadenosine (4) [Scheme 1, Method 2
(23)]. Compound 1 (200 mg, 0.74 mmol) was dissolved in methanol
(4 mL), and hydrazine hydrate (1 mL) was added dropwise at room
temperature. After it was stirred for 5 min, the reaction mixture
was concentrated in vacuo to provide a crude residue, which was
suspended in methanol and stirred at room temperature for 30 min.
The solid product from the suspension was filtered, washed with
methanol, and dried in vacuo to provide 4 (187 mg, 95% yield).
This product was used in the subsequent step without further
purification. For MS and NMR analyses, it was recrystallized from
methanol.
Results and Discussion
N6-[Benzaldehyde hydrazone]-2′-deoxyadenosine (3a) [Scheme
1, Method 2 (23)]. A suspension of 4 (18 mg, 0.07 mmol) and
benzaldehyde (0.012 mL, 0.102 mmol) in 2 mL of methanol was
stirred at room temperature for 20 min, and the reaction mixture
was concentrated in vacuo to provide a crude residue. The mixture
was purified by reversed phase HPLC (column: 5 µm ODS-AQ
250 mm × 10 mm; YMC) using a gradient starting at 100% A and
going to 70% A and 30% B in 40 min, and then 5% A and 95% B
over the next 40 min, at a flow rate of 2.5 mL/min, where solvent
A is composed of triethylammonium acetate, pH 7, and solvent B
is composed of 95% acetonitrile and 5% water.
Two methods were employed to synthesize dAdo N6-
phenylhydrazones (3a-c). Method 1 involved the reaction of
1 with the hydrazone (2) resulting from the addition of hydrazine
and the benzaldehyde derivative, whereas method 2 involved
the initial conjugation of hydrazine with 1 followed by
subsequent addition to the aldehyde (Scheme 1). The presence
of phenylhydrazone moieties at N6 of dAdo shifted the
absorption band into the near-UV (325-370 nm; Figure 1). In
comparison to nonsubstituted 3a (λmax ) 325 nm), the absorption
shifted to higher wavelengths for the 4-methyoxy-substituted
derivative (3b; λmax ) 330 nm), and this shift was more dramatic
for the 4-nitro-substituted derivative (3c; λmax ) 370 nm).
Similar absorption properties have been reported for other aryl
and alkyl hydrazones (19). The photochemistry of dAdo N6-
phenylhydrazones (3a-c) was studied by steady state photolysis
and product analysis (Table 1). The quantum yield of photo-
decomposition was in the range from 1.2 to 2.0 × 10-4 for 3a,b
but was 100-fold lower for 3c. The relatively low quantum yields
suggest that the excited state of 3a-c largely deactivates by
nonproductive radiative and/or nonradiative pathways. The main
pathway of photodecomposition of 3a-c is believed to involve
homolytic cleavage of the N-N bond leading to dAdo N6-
aminyl radicals (5) and benzylidene iminyl radicals (6) as
reported for various phenylhydrazones (Scheme 2) (19, 24).
The photolysis of 3a (0.05 mM) in N2-bubbled aqueous
solution and neutral pH led to its photodecomposition together
with the formation of dAdo and benzaldehyde (8; Table 1). The
yield and profile of photoproducts did not significantly change
in O2-bubbled solutions. The formation of dAdo likely involves
H-atom abstraction by dAdo N6-aminyl radicals, whereas the
formation of benzaldehyde involves H-atom abstraction by
2-(Benzylideneamino)-2′-deoxyadenosine (9). Compound 3a
(0.10 mM) and dAdo (30 mM) were irradiated for 24 h with Blak-
Ray B-100 lamps until the reduction of 3a to approximately 80%.
The photolysis mixture was concentrated and subjected to repeated
1
purification by HPLC-UV (yield ) 10 mg). H NMR, 600 MHz
(D2O): δ 8.18 (s, 1H, C8H), 7.41-7.74 (m, 5H, ArH), 6.34 (m,
1H, H1′), 4.46 (m, 1H, H3′), 3.99 (m, 1H, H4′), 3.63- 3.65 (m, 2H,
H5′ and H5′′), 2.68-2.70 (m, 1H, H2′′), 2.32-2.40 (m, 1H, H2′). The
ESI-MS/MS spectrum of product 1 has two major peaks at m/z
393 and 277.
2-Benzamido-2′-deoxyadenosine (10). This product was pre-
pared as described for 9. 1H NMR, 600 MHz (D2O): δ 8.96 (s, 1H,
NdCH), 8.35 (s, 1H, C8), 7.66-7.45 (m, 5H, ArH), 6.38 (m, 1H,
H1′), 4.60 (s, 2H, D2O), 4.50 (m, 1H, H3′), 3.98 (m, 1H, H4′),
3.63-3.65 (m, 2H, H5′ and H5′′), 2.72-2.75 (m, 1H, H2′′), 2.44-2.47
(m, 1H, H2′).
Near-UV Photolysis and Quantum Yield Calculations. Pho-
tolysis was carried out using a 1000 W Hg-Xe arc lamp (spectra
Physics, Oriel Instruments, Stratford, CT) fitted with an infrared
filter and diffracting monochromator (spectral energy Corp., Chester,
NY) set at 315 ( 2 nm. The beam of light was focused on a quartz
cuvette (1 cm path length). The samples were bubbled with either
oxygen or nitrogen for 10 min before photolysis followed by

Generation and Fate of dAdo N6-Aminyl Radicals
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 51
Table 1. Yields of Products from the Photolysis of dAdo
N6-Phenylhydrazones (3) under Various Conditions
pmol/minb
hydrazone
(0.05 mM)
GSH
(50 mM)
a
O2
-hydrazone +dAdo +aldehyde (8)
3a
3a
3a
3b
3c
-
+
-
-
-
-
-
+
+
+
-4.09
-4.02
-2.58
-3.93
-0.01
0.45
0.15
1.88
2.23
0.01
0.30
0.15
1.27
2.71
0.01
a Solutions were buffered with phosphate buffer at pH 7 and bubbled
10 min before and continuously during photolysis with either O2 (+) or
N2 (-). b The quantum yield for the decomposition of 3a in the absence
of oxygen and GSH is approximately 2 × 10-4
.
benzylidene iminyl radicals followed by hydrolysis of the
resulting imine in aqueous solution (Scheme 2). Thus, the low
yields of dAdo (11% of total) and benzaldehyde (7% of total),
in the absence of additives, suggest that dAdo N6-aminyl radicals
and benzylidene iminyl radicals, generated by initial photo-
cleavage, readily react with 3a leading to its decomposition.
The nature of decomposition products was not pursued because
of their number and relatively low yield. Previously, the
photolysis of benzaldehyde phenylhydrazone was shown to
induce homolytic cleavage followed by H-atom transfer to give
benzonitrile and aniline (24). However, we can rule out H-atom
transfer in the photochemistry of dAdo N6-phenylhydrazines in
view of the low yield of dAdo in the absence of additives and
lack of formation of benzonitrile.
Photolysis in the Presence of GSH. The initial yields of
dAdo N6-aminyl radicals and benylidene iminyl radicals were
estimated by the photolysis of 3a in the presence of a good
H-donor, that is, GSH. The loss of 3a and formation of dAdo
were monitored as a function of time of photolysis (Figure 2
and Table 1). The reactions of dAdo N6-aminyl radicals and
benzylidene iminyl radicals in the presence of GSH are shown
in Scheme 2. The photodecomposition of 3a decreases in the
presence of GSH because it scavenges dAdo N6-aminyl radicals
(5) and benzylidene iminyl radicals (6), which react with 3a in
the absence of other substrates. In contrast, the formation of
both dAdo and benzaldehyde increases because GSH readily
donates H-atoms to precursor radicals of these products. The
yield of dAdo reaches 73% of the total decomposition of 3a at
a concentration of GSH of 50 mM (1.88/2.58; Table 1). Thus,
the generation of dAdo N6-aminyl radicals and benzylidene
iminyl radicals is the major pathway of UVA-induced decom-
position of dAdo N6-phenylhydrazones.
Figure 2. Decomposition of 3a and formation of dAdo as a function
of photolysis time. Decomposition of 3a (solid squares). Formation of
dAdo (solid circles). Compound 3a (0.05 mM, 1 OD) was irradiated
in the presence of GSH (50 mM) at pH 7 with N2 bubbling. Aliquots
of 50 µL were injected for analyses by HPLC-UV.
photodecomposition of the methoxy-substituted phenylhy-
drazone (3b) was 2-fold more efficient, whereas the nitro-
phenylhydrazone derivative (3c) did not significantly de-
compose under identical conditions. Similarly, the most
efficient arylhydrazone derivatives that induce strand breaks
in DNA were previously shown to be those with a methoxy
group in the para-position, whereas those with a nitro group
at the same position were much less efficient (19). The
methoxy group probably enhances homolytic cleavage of the
hydrazone because it pushes electrons into the phenyl ring,
thereby stabilizing the formation of electron-deficient iminyl
radicals.
Photolysis in the Presence of an Excess of dAdo. The
photolysis of 3a was examined in the presence of an excess of
dAdo to mimic the reactions of nitrogen-centered radicals in
DNA. The mixture of photoproducts was separated by HPLC,
and the major products were identified by chemical analysis
(Figure 3). On the basis of these analyses, we propose four
pathways of decomposition of dAdo N6-aminyl radicals (5):
recombination with benzylidene iminyl radicals (6) to give 9
and 10, H-atom abstraction to dAdo, deamination to 2′-
deoxyinosine (11), and addition with dAdo to dAdo-dAdo
dimers (Scheme 3). The contribution of each reaction was
estimated from the formation of products divided by the
decomposition of 3a as a function of irradiation time as
determined by HPLC-UV. The formation of the above products
is blocked in the presence of 50 mM GSH, indicating that they
arise from dAdo N6-aminyl radicals. Because 10 and dAdo-dAdo
elute together under standard HPLC-UV conditions, the indi-
Effect of Substitutents on Phenylhydrazone Photochem-
istry. The photochemistry of three derivatives of dAdo N6-
phenylhydrazones (3a-c) was examined (Table 1). In
comparison to nonsubstituted phenylhydrazone (3a), the
Scheme 2. Photolysis of dAdo N6-Phenylhydrazones (3) in the Presence of GSH

52 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
Kuttappan-Nair et al.
signal at low field in the region of C2 and C8 protons for an
intact adenine ring. The low field signal was assigned to the
C8 proton in view of the presence of cross-peaks between C8
and the anomeric proton in the nuclear Overhauser effect
spectroscopy (NOESY) spectrum. Thus, the structure of 9
consists of benzylidene amine attached to the C2 position of
dAdo.
The mechanism of formation of 9 in the presence of an excess
of dAdo may involve the recombination of dAdo N6-aminyl and
benzylidene iminyl radicals (Scheme 4) and/or the addition of
benzylidene radicals to nonmodified dAdo. Two lines of
evidence point to the recombination of radicals as the main
pathway. First, there was no significant change in the yield of
9 when 3a was irradiated in the presence of either 10 or 30
mM dAdo. Second, the photolysis of 3a in the presence of an
excess of the ribonucleoside derivative, adenosine in place of
dAdo, led predominantly to the formation of 9 and not to the
corresponding ribose derivative. Although the ribose derivative
of 9 was identified in the photolysis mixture [retention time
∼25 min, m/z 371 (MH+) and m/z 239 (MH+ minus ribose)],
its yield was 15-fold less than the corresponding 2-deoxyribose
derivative. These results indicate that the addition of benzylidene
iminyl radicals to dAdo is a minor pathway in the formation of
9. As a probable mechanism of formation, we propose that dAdo
N6-aminyl and benzylidene iminyl radicals recombine at the C2
position of dAdo followed by rearrangement of the resulting
biradical (Scheme 4). The recombination reaction appears to
involve diffusible radicals in the bulk solution because the
formation of 9 is blocked by the addition of GSH. However,
we cannot rule out the possibility that the reaction takes place
within a radical cage in which GSH cannot scavenge the initial
radicals but interferes with the chemistry of subsequent radicals
in the formation of 9, such as 12 and 13.
The structure of 10 was deduced by MS and NMR analyses.
The mass of 10 [m/z 371 (MH+)] was 16 mass units higher
than that of 9 [m/z 355 (MH+)]. The MS spectra of 9 and 10
showed similar features, including the loss of 2-deoxyribose
from the molecular ion and the presence of five exchangeable
protons. Of the exchangeable protons, three were present on
the exocyclic amino group and two on the sugar moiety. The
1HNMR spectra of 9 and 10 were also very similar except for
one major difference: The imine proton of 9 was not present in
the spectrum of 10. These results are consistent with a structure
in which the imine group of 9 undergoes oxidation to an amide
group as in product 10 (Scheme 4). The proposed structure is
consistent with other NMR features. The phenyl group displayed
five protons, excluding the possibility that the aromatic ring was
hydroxylated. Similar to the analysis of 9, the 1H NMR,
correlation spectroscopy, and NOESY spectra of 10 indicated
the presence of a proton at C8 but not at C2. The formation of
10 may be explained by photooxidation of product 9, which,
Figure 3. HPLC-UV analysis of the mixture from the photolysis of 3a
(0.05 mM) with dAdo (30 mM) in phosphate buffer (10 mM, pH 7).
Photolysis time: 0 (black line), 30 (blue line), and 60 min (red line).
Scheme 3. Reactions of dAdo N6-Aminyl Radicals in the
Presence of dAdo
vidual yield of these products was estimated from the corre-
sponding MS ion signals of the purified HPLC peak. The ion
abundance of dAdo-dAdo dimer is 5-fold less than that of 10,
suggesting that it is less than 1% of the total decomposition of
3a. The rate of formation of dAdo was determined in parallel
experiments using an excess of the ribonucleoside (adenosine)
in place of dAdo. Together, products 9-11, dAdo, and
dAdo-dAdo account for about 90% of the total decomposition
of 3a.
Reactions of Benzylidene Iminyl Radicals. Two products
were identified by MS and NMR analyses, and they have a
similar structure: 2-(benzylideneamino)-2′-deoxyadenosine 9 and
2-benzamido-2′-deoxyadenosine (10; Scheme 4). The ESI-MS
spectrum of 9 displayed the molecular ion and corresponding
Na+ adduct at m/z 355 and m/z 377, respectively. Fragmentation
of m/z 377 gave m/z 261 (loss of 2-deoxyribose), and subsequent
fragmentation of m/z 239 gave m/z 136 (loss of adenine). Indeed,
the MS features of 9 were nearly identical to those of 3a.
Analysis of 9 equilibrated in D2O, however, revealed that this
compound contained four exchangeable protons, while 3a
showed only three exchangeable protons. The position of
attachment between dAdo and benzylidene amine was confirmed
1
by NMR analyses. The H NMR spectrum depicted only one
Scheme 4. Proposed Mechanism of Formation of 9 and Its Oxidation Product 10

Generation and Fate of dAdo N6-Aminyl Radicals
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 53
Figure 4. ESI-MS/MS spectrum of a dAdo-dAdo dimer.
similar to dAdo hydrazones, absorbs UV light at 313 nm.
Alternatively, 9 may undergo autoxidation to product 10 because
it was observed as a minor product in the NMR spectrum of
purified 9.
dAdo. However, we were not able to confirm the structure of
the cross-link by NMR analysis due to its relatively low yield.
Finally, dAdo is a major product of dAdo N6-aminyl radicals
(32%) (Scheme 3). The formation of dAdo likely involves
H-atom abstraction of dAdo N6-aminyl radicals from GSH or
other possible sources such as dAdo in the absence of GSH
and presence of an excess of dAdo. These results agree with
the tendency of primary aminyl radicals to undergo H-atom
abstraction rather than addition to a double bond when compared
to alkylamino or iminyl radicals (29-31). Nevertheless, the
situation may be complicated by the contribution of tautomeric
dAdo N1-iminyl radicals (4, 32). According to natural popula-
tion analysis, the distribution of spin densities of neutral dAdo
N6-aminyl radical is greatest at the N6 position (-0.59) as
compared to other positions: N1 (-0.16), N3 (-0.22), C5
(-0.22), and C8 (-0.14) (33). The photosensitized oxidation
of several dinucleoside monophosphates by 2-methyl-1,4-
naphthoquinone, including d(ApA), d(CpA), and d(ApC), led
to the formation of N6-formyladenine and N6-acetyladenine
modifications (34, 35). When 2-methyl-1,4-naphthoquinone
tethered oligonucleotides were exposed to UVA light, the
formation of interstrand cross-links occurred between the
adenine and the quinone moieties, which may be explained from
the condensation of dAdo N6-aminyl radical and semiquinone
radicals (36).
Reactions of dAdo N6-Aminyl Radicals. The reactivity of
dAdo N6-aminyl radicals was examined by the photolysis of
3a (0.05 mM) in the presence of a molar excess of dAdo (30
mM). From the photolysis mixture, two photoproducts were
identified, which likely arise from the reaction of dAdo N6-
aminyl radicals. The peak at 14 min was identified as 11 on the
basis of HPLC-UV and MS analyses. The formation of 11 may
be explained by the addition of water to dAdo N6-aminyl radicals
followed by deamination. Similarly, dAdo N6-aminyl radicals
have been implicated in the formation of 11 by the radiolysis
of deaerated aqueous solutions of dAdo and the UVA photolysis
of dry films of dAdo and 7-methylpyrido[3,4-c]psoralen (25, 26).
The formation of 11 is very similar to the formation of
2′-deoxyuridine from the hydrolysis of cytosine N4-aminyl
radicals generated in the UVA photosensitized oxidation of 2′-
deoxycytidine by 2-methyl-1,4-naphthoquinone (27). The con-
version of cytosine N4-aminyl radicals to 2′-deoxyuridine was
the subject of a theoretical study supporting the addition of H2O
to the electrophilic C4 followed by loss of the exocyclic amino
group (28). For cytosine and adenine nitrogen-centered radicals,
the deamination reaction may take place via the iminyl tautomer,
which, similar to ground state imines, should be more susceptible
to hydrolysis.
Conclusions
A novel product consisting of two dAdo moieties was
tentatively identified by MS. This product eluted at 22 min on
HPLC-UV with the same retention as 10 (Figure 3). The MS
spectrum of the product displayed a molecular ion peak at m/z
523, which corresponded to the molecular ion of two dAdo
moieties (m/z 502) plus Na+ (m/z 23) minus 2 mass units. The
MS/MS spectrum of m/z 523 showed ion fragments at m/z 407
and 291, indicating the consecutive loss of two 2-deoxyribose
moieties (minus m/z 116) (Figure 4). The yield of dAdo-dAdo
dimer increased upon an increase in the concentration of
nonmodified dAdo in the photolysis solution from 10 to 30 mM.
Furthermore, the formation of dimer was not detected when the
photolysis was carried out in the presence of an excess of
nonmodified adenosine instead of dAdo. In contrast, an analo-
gous product was observed in the photolysis mixture with the
mass of adenosine and dAdo and the loss of ribose (m/z minus
132). These results suggest that dAdo-dAdo dimer is formed
by the addition of dAdo N6-aminyl radicals to nonmodified
dAdo N6-aminyl radicals are the main species generated by
the reaction of hydroxyl radicals with adenine and the depro-
tonation of adenine radical cations. Furthermore, the deproto-
nation of adenine radical cations at N6 is favored in double-
stranded DNA due to the lack of H-bonding and contact with
H2O in the major groove (33). The present study suggests that
dAdo N6-aminyl radicals will undergo three reactions: H-atom
abstraction, hydrolytic deamination, and addition to nonmodified
adenine. H-atom abstraction in DNA may lead to strand breaks
as has been reported for various alkyl and aryl aminyl radicals.
Deamination of adenine produces inosine, which should not
significantly affect DNA structure or its coding ability. In
contrast, the addition of dAdo N6-aminyl radicals to another
adenine moiety, or another DNA base, may cause intra- or
interstrand cross-links in DNA, which may have severe con-
sequences in cell death and mutagenesis. Future studies are in
progress to assess the fate of dAdo N6-aminyl radicals in duplex
DNA.

54 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
Kuttappan-Nair et al.
(18) Midorikawa, K., Murata, M., Oikawa, S., Tada-Oikawa, S., and
Kawanishi, S. (2000) DNA damage by dimethylformamide: Role of
hydrogen peroxide generated during degradation. Chem. Res. Toxicol.
13, 309–315.
(19) Hwu, J. R., Lin, C. C., Chuang, S. H., King, K. Y., Sue, T. R., and
Tsay, S. C. (2004) Aminyl and iminyl radicals from arylhydrazones
in the photo-induced DNA cleavage. Bioorg. Med. Chem. 12, 2509–
2515.
Acknowledgment. We thank Natural Science and Engineer-
ing Research Council of Canada (NSERC) for funding. We also
thank Prof. Klaus Klarskov for mass spectrometry analyses and
Profs. Pierre Lavinge and Luc Tremblay for NMR analyses.
V.K.-N. thanks Mahatma Gandhi University, Kottayam, for
sanctioning leave for this project.
(20) Sortino, S., Giuffrida, S., and Scaiano, J. C. (1999) Phototoxicity of
naphazoline. Evidence that hydrated electrons, nitrogen-centered
radicals, and OH radicals trigger DNA damage: A combined photo-
cleavage and laser flash photolysis study. Chem. Res. Toxicol. 12, 971–
978.
(21) Wang, H., Marnett, L. J., Harris, T. M., and Rizzo, C. J. (2004) A
novel synthesis of malondialdehyde adducts of deoxyguanosine,
deoxyadenosine, and deoxycytidine. Chem. Res. Toxicol. 17, 144–
149.
(22) Nenajdenko, V. G., Reznichenko, A. L., Lenkova, O. N., Shastin,
A. V., and Balenkova, E. S. (2005) A new synthetic approach to alpha-
chlorocinnamaldehydes. Synthesis-Stuttgart 605–609.
(23) Prakash, J., Casaba, S., and Andrew, L. S. (2005) International
Application Published under the Patent Cooporation Treaty, 117910
A117912.
References
(1) von Sonntag, C. (2006) Free-Radical Induced DNA Damage and Its
Repair. A Chemical PerspectiVe, Springer, Heidelberg.
(2) Cadet, J., Berger, M., Douki, T., and Ravanat, J. L. (1997) Oxidative
damage to DNA: Formation, measurement, and biological significance.
ReV. Physiol. Biochem. Pharmacol. 131, 1–87.
(3) Cadet, J., Delatour, T., Douki, T., Gasparutto, D., Pouget, J. P.,
Ravanat, J. L., and Sauvaigo, S. (1999) Hydroxyl radicals and DNA
base damage. Mutat. Res., Fundam. Mol. Mech. Mutagen. 424, 9–21.
(4) Vieira, A., and Steenken, S. (1990) Pattern of OH radical reaction
with adenine and its nucleosides and nucleotidessCharacterization
of 2 types of isomeric oh adduct and their unimolecular transformation
reactions. J. Am. Chem. Soc. 112, 6986–6994.
(24) Binkley, R. W. (1970) Photochemistry of unsaturated nitrogen-
containing compounds. VII. Photolysis of phenylhydrazones. J. Org.
Chem. 35, 2796–2801.
(25) Raoul, S., Bardet, M., and Cadet, J. (1995) Gamma-irradiation of 2′-
deoxyadenosine in oxygen-free aqueous-solutionssIdentification and
conformational features of formamidopyrimidine nucleoside deriva-
tives. Chem. Res. Toxicol. 8, 924–933.
(5) Cadet, J., and Berger, M. (1985) Radiation-induced decomposition of
purine bases within DNA and related model compounds. Int. J. Radiat.
Biol. 47, 127–143.
(6) Cadet, J., Douki, T., Frelon, S., Sauvaigo, S., Pouget, J. P., and
Ravanat, J. L. (2002) Assessment of oxidative base damage to isolated
and cellular DNA by HPLC-MS/MS measurement. Free Radical Biol.
Med. 33, 441–449.
(7) Frelon, S., Douki, T., and Cadet, J. (2002) Radical oxidation of the
adenine moiety of nucleoside and DNA: 2-Hydroxy-2 ′-deoxyadenos-
ine is a minor decomposition product. Free Radical Res. 36, 499–
508.
(8) Melvin, T., Bothe, E., and Schultefrohlinde, D. (1996) The reaction
of triplet 2-methyl-1,4-naphthoquinone (menadione) with DNA and
polynucleotides. Photochem. Photobiol. 64, 769–776.
(26) Voituriez, L., and Cadet, J. (1999) Isolation and characterization of
two furan-side photoadducts of 7-methylpyrido[3,4-c] psoralen to the
sugar moiety of 2′-deoxyadenosine. Photochem. Photobiol. 70, 152–
158.
(27) Wagner, J. R., Decarroz, C., Berger, M., and Cadet, J. (1999) Hydroxyl
radical-induced decomposition of 2′-deoxycytidine in aerated aqueous
solutions. J. Am. Chem. Soc. 121, 4101–4110.
(9) Liu, Z., Gao, Y., and Wang, Y. (2003) Identification and characteriza-
tion of a novel cross-link lesion in d(CpC) upon 365-nm irradiation
in the presence of 2-methyl-1,4-naphthoquinone. Nucleic Acids Res.
31, 5413–5424.
(10) Liu, Z. J., Gao, Y., Zeng, Y., Fang, F., Chi, D., and Wang, Y. S.
(2004) Isolation and characterization of a novel cross-link lesion in
d(CpC) induced by one-electron photooxidation. Photochem. Photo-
biol. 80, 209–215.
(11) Ding, H., Majumdar, A., Tolman, J. R., and Greenberg, M. M. (2008)
Multinuclear NMR and kinetic analysis of DNA interstrand cross-
link formation. J. Am. Chem. Soc. 130, 17981–17987.
(12) Hong, I. S., and Greenberg, M. M. (2005) Efficient DNA interstrand
cross-link formation from a nucleotide radical. J. Am. Chem. Soc. 127,
3692–3693.
(13) Bernofsky, C., Bandara, B. M. R., Hinojosa, O., and Strauss, S. L.
(1990) Hypochlorite-modified adenine-nucleotidessStructure, spin-
trapping, and formation by activated guinea-pig polymorphonuclear
leukocytes. Free Radical Res. Commun. 303–315.
(14) Hawkins, C. L., and Davies, M. J. (2001) Hypochlorite-induced damage
to nucleosides: Formation of chloramines and nitrogen-centered
radicals. Chem. Res. Toxicol. 14, 1071–1081.
(15) Hawkins, C. L., and Davies, M. J. (2002) Hypochlorite-induced damage
to DNA, RNA, and polynucleotides: Formation of chloramines and
nitrogen-centered radicals. Chem. Res. Toxicol. 15, 83–92.
(16) Hawkins, C. L., Pattison, D. I., and Davies, M. J. (2002) Reaction of
protein chloramines with DNA and nucleosides: Evidence for the
formation of radicals, protein-DNA cross-links and DNA fragmenta-
tion. Biochem. J. 365, 605–615.
(17) Murata, M., Ohnishi, S., and Kawanishi, S. (2001) Acrylonitrile
enhances H2O2-mediated DNA damage via nitrogen-centered radical
formation. Chem. Res. Toxicol. 14, 1421–1427.
(28) Labet, V., Grant, A., Cadet, J., and Erriksson, L. A. (2008) Deamination
of the radical cation of the base moiety of 2′-deoxycytidine: A
theoretical study. ChemPhysChem 9, 1195–1203.
(29) Lalevee, J., Gigmes, D., Bertin, D., Graff, B., Allonas, X., and
Fouassier, J. P. (2007) Comparative reactivity of aminyl and ami-
noalkyl radicals. Chem. Phys. Lett. 438, 346–350.
(30) Le Tadic Biadatti, M.-H., Callier-Dublanchet, A.-C., Horner, J. H.,
Quiclet-Sire, B., Zard, S. Z., and Newcomb, M. (1997) Absolute rate
constants for iminyl radical reactions. J. Org. Chem. 62, 559–563.
(31) Zard, S. Z. (2008) Recent progress in the generation and use of
nitrogen-centred radicals. Chem. Soc. ReV. 37, 1603–1618.
(32) Steenken, S. (1989) Purine-bases, nucleosides, and nucleotidessAqueous-
solution redox chemistry and transformation reactions of their radical
cations and e- and oh adducts. Chem. ReV. 89, 503–520.
(33) Reynisson, J., and Steenken, S. (2002) DFT calculations on the
electrophilic reaction with water of the guanine and adenine radical
cations. A model for the situation in DNA. Phys. Chem. Chem. Phys.
4, 527–532.
(34) Wang, Y. S., and Liu, Z. J. (2002) Mechanisms for the formation of
major oxidation products of adenine upon 365-nm irradiation with
2-methyl-1,4-naphthoquinone as a sensitizer. J. Org. Chem. 67, 8507–
8512.
(35) Wang, Y. S., Liu, Z. J., and Dixon, C. (2002) Major adenine products
from 2-methyl-1,4-naphthoquinone-sensitized photoirradiation at 365
nm. Biochem. Biophys. Res. Commun. 291, 1252–1257.
(36) Bergeron, F., Klarskov, K., Hunting, D. J., and Wagner, J. R. (2007)
Near-UV photolysis of 2-methyl-1,4-naphthoquinone-DNA duplexes:
Characterization of reversible and stable interstrand cross-links between
quinone and adenine moieties. Chem. Res. Toxicol. 20, 745–756.
TX900268R