Reaction of glycidamide with 2′-deoxyadenosine and 2′-deoxyguanosine - Mechanism for the amide hydrolysis

Article abstract of DOI:10.1080/15257770601112697

2′-Deoxyadenosine (dA) and 2′-deoxyguanosine (dG) were reacted with the mutagenic epoxide glycidamide (GA, Scheme 1). The reactions yielded three GA-dA adducts (N1-GA-dA, N⁶-GA-dA and N1-GA-dI) and two GA-dG adducts (N1-GA-dG I and N1-GA-dG II) (). The structures of the adducts were characterized by spectroscopic and spectrometric methods (¹H-, ¹³C, and 2D NMR, MS, UV). The mechanism of the amide hydrolysis taking place during formation of the adducts N1-GA-dA and N1-GA-dG I was studied. We propose a mechanism where a transamidation is the key step in the hydrolysis of the amide function of GA. Copyright Taylor & Francis Group, LLC.

Full text of DOI:10.1080/15257770601112697

Nucleosides, Nucleotides, and Nucleic Acids, 26:129–148, 2007
Copyright ꢀC Taylor & Francis Group, LLC
ISSN: 1525-7770 print / 1532-2335 online
DOI: 10.1080/15257770601112697
REACTION OF GLYCIDAMIDE WITH 2ꢀ-DEOXYADENOSINE AND
2ꢀ-DEOXYGUANOSINE—MECHANISM FOR THE AMIDE HYDROLYSIS
Josefin Backman and Leif Kronberg
Abo Akademi University, Turku/Abo, Finland
Laboratory of Organic Chemistry,
2
˚
˚
2ꢁ-Deoxyadenosine (dA) and 2ꢁ-deoxyguanosine (dG) were reacted with the mutagenic epoxide
2
glycidamide (GA, Scheme 1). The reactions yielded three GA-dA adducts (N1-GA-dA, N6-GA-dA
and N1-GA-dI) and two GA-dG adducts (N1-GA-dG I and N1-GA-dG II) (Scheme 2). The
structures of the adducts were characterized by spectroscopic and spectrometric methods (1H-, 13C,
and 2D NMR, MS, UV). The mechanism of the amide hydrolysis taking place during formation
of the adducts N1-GA-dA and N1-GA-dG I was studied. We propose a mechanism where a
transamidation is the key step in the hydrolysis of the amide function of GA.
Keywords Glycidamide; nucleosides; adducts; amide hydrolysis
INTRODUCTION
Acrylamide (AM) (Scheme 1) is a high production volume chemical
with a wide variety of industrial applications. AM is neurotoxic, clastogenic,
and carcinogenic in animal experiments, and it is classified as a probable
human carcinogen.[1] Recently, it was discovered that AM is present in
a wide range of different starchy foodstuffs.[2] This has led to extensive
investigations worldwide to determine the genotoxic mechanisms of AM.
Previously, it has been shown that AM reacts with the bases in DNA, but
the reactivity seems to be quite low.[3] It now has been concluded that
the mutagenicity of acrylamide in human and mouse cells is due to the
AM epoxide metabolite glycidamide (GA) (Scheme 2) and its enhanced
reactivity with DNA.[4–6]
The first GA adduct that was structurally characterized, was the N7-GA-
Gua adduct.[7] This adduct also is formed in DNA of mice and rats treated
Received 15 May 2006; accepted 27 September 2006.
This work was a part of the EU-project “Heat-generated food-toxicants, identification, character-
isation and risk minimisation” and was financed by the European Commission under Contract No.
FOOD-CT-2003–506820 (STREP). This work was also financed by the Foundation of Magnus Ehrnrooth
˚
and the Foundation of the Research Institute at Abo Akademi University.
˚
Address correspondence to Leif Kronberg, Laboratory of Organic Chemistry, Abo Akademi
˚
University, Biskopsgatan 8, FIN-20500 Turku/Abo, Finlan. E-mail: leif.kronberg@abo.fi
129

130
J. Backman and L. Kronberg
SCHEME 1 Structures of AM and GA.
with AM and GA.[8] More recently, Gamboa da Costa et al. reported on two
adducts, N3-GA-Ade and N1-GA-dA, formed in the reaction of GA with 2ꢁ-
deoxyadenosine (dA).[8] In addition, it was proposed that the N6-GA-adduct
along with a cyclic unidentified adduct were formed in the reactions of dA
and GA. The N3-GA-Ade and N1-GA-dA also were found in salmon testis
DNA modified with GA.[8]
We previously have reported that GA is reactive toward cytidine and
thymidine.[9]
Thymidine is considered to be the least reactive of the nucleosides, but
we found that in the reaction with GA a surprisingly high yield (12.3 mol%
isolated yield) of a thymidine adduct could be obtained.
In this work, we further explored the reactions of GA with 2ꢁ-
deoxyadenosine and 2ꢁ-deoxyguanosine (dG). The work resulted in iden-
tification of one dA adduct (N1-GA-dI) (Scheme 2) and two dG adducts
(N1-GA-dG I, N1-GA-dG II) (Scheme 2) not previously described in
SCHEME 2 Structures of the GA-derived nucleoside adducts identified in this study.

Mechanism of Formation of Glycidamide Adducts
131
the literature. In addition, the N6-dA adduct is for the first time fully
characterized. Furthermore, we wanted to explore the mechanism for the
facile hydrolysis of the amide function in the GA tail.
MATERIALS AND METHODS
(Caution: GA has been found to be mutagenic on Salmonella typhimurium.
Caution should therefore be exercised in the handling of the compound.)
Chemicals and Enzymes
GA was purchased from Toronto Research Chemicals (North York, ON,
Canada). 2ꢁ-deoxyadenosine and 2ꢁ-deoxyguanosine were obtained from
Sigma Chemical Co. (St. Louis, MO, USA). Acetonitrile (p.a.) was from
Labscan Ltd. (Dublin, Ireland). The water used was distilled water purified
with a Millipore system (Simplicity 185, Billerica, MA, USA).
Chromatographic Methods
Liquid chromatography-diode array detector (LC-DAD) analyses were
performed on an Agilent 1100 Series liquid chromatographic system (Agi-
lent Technologies, Esbo, Finland) consisting of a quaternary pump, a vac-
uum degasser, an autosampler, a thermostated column compartment, and a
diode array detector (DAD). The reaction mixtures were chromatographed
on a 5 µm, 4 mm × 125 mm reversed phase C18 analytical column (Hypersil
BDS-C18, Agilent Technologies, Esbo, Finland). The column was eluted
isocratically for 3 minutes with 2% acetonitrile in ammonium acetate buffer
(0.01 M, pH 7) and then with a gradient from 2 to 30% acetonitrile over the
course of 17 minutes at a flow rate of 1 mL/minute. The products were
purified on a semipreparative 5 µm, 10 mm × 250 mm reversed phase
C18 column (BDS Hypersil C18, Thermo Hypersil-Keystone, Krotek Oy,
Tammerfors, Finland). When the 2ꢁ-deoxyadenosine adducts were purified
the semipreparative column was coupled to a Varian 5000 Liquid Chro-
matograph (Varian Aerograph, Walnut Creek, CA, USA) and a variable-
wavelength Shimadzu SPD-6 A UV spectrophotometric detector (Shimadzu
Europe, Duisburg, Germany). When the 2ꢁ-deoxyguanosine adducts were
purified the semipreparative column was coupled to the above mentioned
Agilent 1100 Series liquid chromatographic system, with a fraction collector
(Agilent 1100) added to the system.
Spectroscopic and Spectrometric Methods
The 1H and 13C NMR spectra of the adducts were recorded at 25◦C on
a Bruker Avance 600 NMR spectrometer at 600 and 150 MHz, respectively
(Bruker, Rheinstetten, Germany). The samples were dissolved in Me2SO-d6,
except for one of the compounds (N6-GA-dA), which was dissolved in D2O.

132
J. Backman and L. Kronberg
1
The H NMR signal assignments were based on chemical shifts and H-H
and C-H correlation data. The assignment of carbon signals was based on
chemical shifts and C-H correlations.
The LC-ESI-MS/MS analyses were performed on an Agilent 1100 Series
LC/MSD SL Trap instrument (Agilent Technologies, Esbo, Finland) consist-
ing of a binary pump, a vacuum degasser, an autosampler, a thermostated
column compartment, and a UV detector. The ion trap mass spectrometer
was equipped with an electrospray source and operated in the positive ion
mode. Operation parameters of the ESI ion source were as follows: drying
gas temperature was 350◦C, drying gas flow 12 L/minute, nebulizer gas
pressure 40 psi, end plate voltage –3500 V, end plate offset –500 V, capillary
exit 115 V and skim 1 was set at 40 V. Ion trap parameters were as follows:
accumulation time was 20 ms, averages 5, rolling averaging on and ion
charge control on. Nitrogen gas was used as drying and nebulizing gas.
Collision induced dissociation (CID) experiments coupled with multiple
tandem mass spectrometry (MSn) employed helium as collision gas. The
fragmentation amplitude was varied between 0.7 and 1.0 V. The samples
from the reaction mixtures were chromatographed on a 5 µm, 4 mm ×
125 mm reversed phase C18 analytical column (Hypersil BDS-C18, Agilent
Technologies, Esbo, Finland). The column was eluted with a gradient
consisting of 0.01 M aqueous ammonium acetate and acetonitrile. First, the
column was eluted isocratically for 5 minutes with 1% acetonitrile and then
with a gradient from 1 to 30% acetonitrile over the course of 14 minutes at
a flow rate 0.5 mL/minute.
The pure compounds also were analyzed by direct inlet infusion to
the source by a syringe pump at a flow rate of 5 µL/minute and at a
concentration of about 1 µg/mL. The compounds were dissolved in a
1:1 v/v mixture of 0.01 M aqueous ammonium acetate/acetonitrile. The
operation parameters were as follows: drying gas temperature was 325◦C,
drying gas flow 5 L/minute, and nebulizer gas pressure 15 psi.
The UV spectra of the isolated compounds were recorded with the diode
array detector as the peaks eluted from the HPLC column.
Preparative-Scale Reaction of GA with 2ꢁ-Deoxyadenosine at pH 7.
Preparation of N1-(2-Carboxy-2-hydroxyethyl)-2ꢁ-Deoxyadenosine
(N1-GA-dA), N6-(2-Carboxy-2-hydroxyethyl)-2ꢁ-Deoxyadenosine (N6-GA-dA)
and N1-(2-Carboxy-2-hydroxyethyl)-2ꢁ-Deoxyinosine (N1-GA-dI)
GA (690 mg, 7.96 mmol) was reacted with 2ꢁ-deoxyadenosine (100 mg,
0.39 mmol) in 20 mL of 0.5 M phosphate buffer solution (pH 7.0).
The reaction was performed at 37◦C. The progress of the reaction was
followed by LC-DAD and LC-ESI-MS/MS analyses on the C18 analytical
column. The reaction was stopped at optimum yield (4 days). The reaction
mixture was concentrated by rotary evaporation to about 4 mL, and the
precipitated phosphate salts were removed by filtration. The adducts were

Mechanism of Formation of Glycidamide Adducts
133
isolated from the mixture by chromatography on the semi-preparative
column. The column was first eluted isocratically with 2% acetonitrile in
ammonium acetate buffer (0.01 M, pH 7) in 3 minutes, and then with
a gradient from 2 to 30% over the course of 17 minutes. The flow rate
was 3 mL/minute. One fraction contained the adducts N1-GA-dA and
N1-GA-dI, and another fraction contained the adduct N6-GA-dA. N1-GA-dA
and N1-GA-dI were separated from each other by re-injecting the fraction
using the same conditions as during the initial fractionation, but the sample
amount injected was lower, thus improving the separation. The solutions
containing the pure compounds were then combined and rotary evaporated
to dryness and further dried under vacuum. The residues were subjected
to spectroscopic and spectrometric studies. The isolated amounts of the
compounds were as follows: N1-GA-dA, 8.02 mg; N6-GA-dA, 6.8 mg; and
N1-GA-dI, 2.25 mg. The yields were 9.5, 12.5, and 8.8 mol%.
N1-GA-dA had the following spectral characteristics: UV spectrum,
UVmax 260, 212 nm, UVmin 234 nm (HPLC eluent: approximately 2% ACN
in ammonium acetate buffer; pH 7). In the positive ion electrospray mass
spectrum, the following ions were observed (m/z, relative abundance, for-
mation): MS, 340 (100, MH+), 362 (5, MNa+), MS2 of 340, 224 (100, MH+-
deoxyribosyl + H), MS3 of 340→224, 206 (21, MH+-deoxyribosyl-H2O +
H), 178 (100, MH+-deoxyribosyl-CH2O2 + H), 136 (37, MH+-deoxyribosyl-
C3H5O3), MS4 of 340→224→178, 160 (100, MH+-deoxyribosyl-CH3O3
1
+ H). The H and 13C NMR spectroscopic data of the compound are
presented in Table 1.
TABLE 1 1H and 13C chemical shifts (δ), spin-spin coupling constants, J H,H (Hz) of protons, and
long-range C-H correlations (HMBC) in the 2ꢁ-deoxyadenosine adduct N1-GA-dA (in DMSO-d6)
Proton
δ (ppm)
Multiplicity J H,H
Carbon
δ (ppm)
HMBC
H-8 (1H)a
H-2 (1H)a
H-10a (1H)
H-10b (1H)a
H-11 (1H)
8.27; 8.28
8.08; 8.09
4.59
3.79; 3.81
3.97
s
s
C-8a
C-2
139.0; 139.1 H-1ꢁ
149.2
H-10a; H-10b
br dd
dd
br d
13.5; 2.2
13.5; 7.8
7.7
C-10a
52.48; 52.51 H-2; H-11
C-11
C-12
C-4a
C-5
68.4
H-10b
174.2
H-11; H-10a; H-10b
142.6; 142.7 H-8; H-1ꢁ; H-2
121.9
153.3
84.0
H-8
C-6ꢁ
H-2; H-10a; H-10b
H-1ꢁꢁ (1H)
H-2 (1H)
H-2ꢁꢁ (1H)
H-3ꢁ (1H)
H-4ꢁꢁ (1H)
H-5 (1H)a
H-5ꢁꢁ (1H)a
6.25
2.62
2.29
4.39
3.86
t
6.8
C-1
H-8; H-2ꢁ
m
m
dt
dt
C-2ꢁb
H-4ꢁ
6.5; 3.3
4.5
11.7; 4.3
11.8; 4.6
C-3ꢁꢁa
C-4
70.82; 70.84 H-1ꢁ
88.2
H-1ꢁ
3.595; 3.588 dd
3.513; 3.506 dd
C-5ꢁa
61.80; 61.81
aSeparate shifts due to mixture of diastereomers.
bSignal overlapped by solvent.

134
J. Backman and L. Kronberg
TABLE 2 1H and 13C chemical shifts (δ) and spin-spin coupling constants, J H,H (Hz) of protons, and
long-range C-H correlations (HMBC) in the 2ꢁ-deoxyadenosine adduct N6-GA-dA (in D2O)
Proton
δ (ppm)
Multiplicity
J H,H
Carbon
δ (ppm)
HMBC
H-8 (1H)
H-2 (1H)
H-11 (1H)
H-10a (1H)
H-10b (1H)
8.15
8.07
4.26
3.83
3.68
s
s
C-8
139.6
152.0
C-2
br s
br s
br s
C-11a
C-10
44.5
C-12
C-4
C-5
178.9
146.9
119.0
154.3
84.5
H-2; H-8
H-8
C-6
H-2
H-1ꢁꢁ (1H)
H-2 (1H)
H-2ꢁꢁ (1H)
H-3ꢁ (1H)
H-4ꢁꢁ (1H)
H-5 (1H)
H-5ꢁꢁ (1H)
6.31
2.70
2.48
4.56
4.11
3.78
3.71
t
6.7
C-1ꢁ
C-2ꢁ
H-2ꢁ
ddd
ddd
dt
dt
dd
dd
13.9; 6.7
13.9; 6.1; 3.3
6.1; 2.9
3.4; 3.2
12.7; 3.1
12.7; 4.2
39.1
C-3ꢁ
C-4ꢁ
C-5ꢁ
71.2
87.3
61.6
H-4ꢁ; H-5ꢁ
aSignal overlapped by C-3ꢁ.
N6-GA-dA had the following spectral characteristics: UV spectrum,
UVmax 270, 210 nm, UVmin 230 nm (HPLC eluent: approximately 8%
ACN in ammonium acetate buffer, pH 7). In the positive ion electrospray
mass spectrum, the following ions were observed (m/z, relative abundance,
formation): MS, 340 (100, MH+), MS2 of 340, 224 (100, MH+-deoxyribosyl
+ H), MS3 of 340→224, 136 (36, MH+-deoxyribosyl-C3H5O3), 178 (100,
MH+-deoxyribosyl-CH2O2 + H), 207 (9, MH+-deoxyribosyl-H2O + H), MS4
of 340→224→207, 178 (100, MH+-deoxyribosyl-CH3O2 + H). The 1H and
13C NMR spectroscopic data of the compound are presented in Table 2.
N1-GA-dI had the following spectral characteristics: UV spectrum,
UVmax 250, 206, UVmin 226, with a shoulder between 266 and 294 nm (HPLC
eluent: approximately 3% ACN in ammonium acetate buffer, pH 7). In the
positive ion electrospray mass spectrum, the following ions were observed
(m/z, relative abundance, formation): MS, 341 (20, MH+), 225 (100, MH+-
deoxyribosyl + H), MS2 of 225, 137 (100, MH+-deoxyribosyl-C3H5O3), 179
1
(65, MH+-deoxyribosyl-CH2O2). The H and 13C NMR spectroscopic data
of the compound are presented in Table 3.
Preparative-Scale Reaction of GA with 2ꢁ-Deoxyguanosine at pH 9.
Preparation of N1-(2- Carboxy -2-hydroxyethyl)-2ꢁ-deoxyguanosine (N1-GA-dG
I) and N1-(2-Carbamoyl-2-hydroxyethyl)-2ꢁ-deoxyguanosine (N1-GA-dG II)
GA (652 mg, 7.49 mmol) was reacted with 2ꢁ-deoxyguanosine (100 mg,
0.37 mmol) in 30 mL of 0.5 M phosphate buffer solution (pH 9.0). The
reaction was performed at 37◦C. The progress of the reaction was followed

Mechanism of Formation of Glycidamide Adducts
135
TABLE 3 1H and 13C chemical shifts (δ) and spin-spin coupling constants, J H,H (Hz) of protons, and
long-range C-H correlations (HMBC) in the 2ꢁ-deoxyadenosine adduct N1-GA-dI (in DMSO-d6)
Proton
δ (ppm)
Multiplicity
J H,H
Carbon
δ (ppm)
HMBC
H-8 (1H)a
H-2 (1H)
8.291; 8.294
8.23
4.54; 4.55
s
s
dd
C-8a
C-2a
C-10a
139.0; 139.1 H-1ꢁ
149.5; 149.6 H-10a; H-10b
H-10a (1H)a
13.3; 3.6
9.3
50.5; 50.6
H-2
H-10b (1H)b app. 3.58
H-11 (1H)
3.74
br d
C-11
C-12
C-4a
69.0
H-10a
173.6
147.27;
147.31
123.81;
123.87
156.1
83.8
H-10b
H-2; H-8; H-1ꢁ
C-5a
H-8
C-6ꢁ
C-1
H-2; H-8
H-1ꢁꢁ (1H)
H-2 (1H)a
H-2ꢁꢁ (1H)
H-3ꢁ (1H)a
6.29
2.62; 2.64
2.29
dd
7.1; 6.6
H-2ꢁ; H-4ꢁ
ddd
ddd
dt
13.3; 7.3; 5.8 C-2ꢁc
13.3; 6.4; 3.4
4.38; 4.39
2.8
C-3ꢁa
C-4ꢁ
C-5ꢁ
70.83; 70.85 H-4ꢁ; H-1ꢁ;
H-2ꢁ
H-4ꢁ (1H)a
3.854; 3.858
dt
4.5; 2.8
88.2
H-5ꢁ; H-5ꢁꢁ;
H-2ꢁꢁ
H-5ꢁ (1H)a
H-5ꢁꢁ (1H)a
3.585; 3.593
3.50; 3.51
dd
dd
11.7; 4.5
11.7; 4.5
61.8
aSeparate shifts due to mixture of diastereomers.
bSignal overlapped by H-5ꢁ.
cSignal overlapped by solvent.
by LC-DAD and LC-ESI-MS/MS analyses on the C18 analytical column.
The reaction was stopped at optimum yield (6 days). The reaction mixture
was concentrated by rotary evaporation to about 5 mL. The adducts were
isolated from the mixture by chromatography on the semi-preparative
column. The column was first eluted isocratically with 2% acetonitrile in
ammonium acetate buffer (0.01 M, pH 7) in 3 minutes, and then with a
gradient from 2 to 30% over the course of 17 minutes. The flow rate was
3 mL/minute. Fractions containing the adducts N1-GA-dG I and N1-GA-
dG II were collected. The solutions containing the pure compounds were
then combined and rotary evaporated to dryness and further dried under
vacuum. The residues were subjected to spectroscopic and spectrometric
studies. The isolated amounts of N1-GA-dG I and of N1-GA-dG II were 3.78
mg and 0.55 mg, respectively, and the molar yields were 4.86 and 0.68 mol%,
respectively.
N1-GA-dG I had the following spectral characteristics: UV spectrum,
UVmax 254, 204 nm, UVmin 224 nm with a shoulder between 270 and 286 nm
(HPLC eluent: approximately 12% ACN in ammonium acetate buffer, pH
7). In the positive ion electrospray mass spectrum the following ions were
observed (m/z, relative abundance, formation): MS, 356 (32, MH+), MS2
of 356, 240 (100, MH+ -deoxyribosyl + H), MS3 of 356→240, 194 (100,

136
J. Backman and L. Kronberg
TABLE 4 1H and 13C chemical shifts (δ) and spin-spin coupling constants, J H,H (Hz) of protons, and
long-range C-H correlations (HMBC) in the 2ꢁ-deoxyguanosine adduct N1-GA-dG I (in DMSO-d6)
Proton
δ (ppm)
Multiplicity J H,H
Carbon δ (ppm)
HMBC
H-1ꢁ
H-8 (1H)
H-10a (1H) 4.29
H-10b (1H) 3.99
7.93
s
C-8
C-10
135.5
47.5
broad
broad
broad
H-11 (1H)
3.74
C-11b
C-12
C-2
C-4
C-5
72.1
171.5
156.6
148.9
116.0
155.1
82.3
H-8
H-8; H-1ꢁ
H-8
C-6ꢁ
H-1ꢁꢁ (1H)
H-2 (1H)a
H-2ꢁꢁ (1H)
H-3ꢁ (1H)
6.12
dd
7.8; 6.1
C-1
H-2ꢁ;H-3ꢁ; H-4ꢁ
C-2ꢁ a
2.19
4.34
ddd
13.0; 6.0; 3.2
C-3ꢁ
C-4ꢁ
C-5ꢁ
71.2
87.5
61.6
H-2ꢁ; H-4ꢁ; H-5ꢁ; H-5ꢁꢁ
H-4ꢁꢁ (1H)c 3.799; 3.804
4.7
H-5 (1H)
H-5ꢁꢁ (1H)
3.55
3.49
dd
dd
11.7; 4.8
11.7; 4.6
aSignal overlapped by solvent.
bCarbon signal from the HSQC spectrum.
cSeparate shifts due to mixture of diastereomers.
1
MH+-deoxyribosyl-CH2O2), 152 (100, MH+-deoxyribosyl-C3H5O3).The H
and 13C NMR spectroscopic data of the compound are presented in
Table 4.
N1-GA-dG II had the following spectral characteristics: UV spectrum,
UVmax 254, 204 nm, UVmin 222 nm with a shoulder between 270 and 286 nm
(HPLC eluent: approximately 16% ACN in ammonium acetate buffer, pH
7). In the positive ion electrospray mass spectrum the following ions were
observed (m/z, relative abundance, formation): MS, 355 (10, MH+), 239
(13, MH+-deoxyribosyl + H), MS2 of 355, 239 (100, MH+-deoxyribosyl +
H), MS3 of 355→239, 222 (100, MH+-deoxyribosyl-NH3 + H), 194 (100,
MH+-deoxyribosyl-CH3NO), 152 (100, MH+-deoxyribosyl-C3H5NO2). The
1H and 13C NMR spectroscopic data of the compound are presented in
Table 5.
Small-Scale Reactions of GA with 2ꢁ-deoxyadenosine
and 2ꢁ-deoxyguanosine
Small-scale reactions were performed to find the reaction conditions
giving the highest possible yield of the adducts. These conditions were then
applied in the large scale synthesis. GA (69.6 mg, 0.80 mmol was reacted with
2ꢁ-deoxyadenosine (10 mg, 0.04 mmol) in 3 mL of 0.5 M phosphate buffer
solutions at pH 4.6 and 7.0. The reactions were performed at 37◦C. The

Mechanism of Formation of Glycidamide Adducts
137
TABLE 5 1H and 13C chemical shifts (δ) and spin-spin coupling constants, J H,H (Hz) of protons, and
H-H correlations (COSY) in the 2-deoxyguanosine adduct N1-GA-dG II (in DMSO-d6)
Proton
δ (ppm)
Multiplicity J H,H
Carbona δ (ppm)
COSY
H-8 (1H)
7.90
s
m
m
dd
s
s
C-8
C-10
136.0
44.6
H-10a (1H) 3.63
H-10b (1H) 3.33
H-10b; H-11; OH-11
H-10a; H-11; OH-11
H-10a; H-10b
H-11b
4.02; 4.03
7.34
7.29
7.4; 4.2
C-11
69.9
NHa (1H)
NHb (1H)
OH-11 (1H) 7.15
br s
H-10a; H-10b
C-12
C-2
C-4
C-5
C-6
C-1ꢁ
H-1ꢁ (1H)
H-2ꢁ (1H)
H-2ꢁꢁ (1H)
H-3ꢁ (1H)
H-4ꢁꢁ (1H)b
H-5 (1H)b
6.15
2.60
2.21
4.36
t
m
ddd
dt
dt
dd
dd
br s
6.9
82.7
H-2ꢁ; H-2ꢁꢁ
H-1ꢁ; H-3ꢁ
H-1ꢁ; H-3ꢁ
H-4ꢁ
13.2, 6.7; 3.2 C-2ꢁ
39.4
71.0
87.9
62.0
2.9
C-3ꢁ
C-4ꢁ
C-5ꢁ
3.808; 3.811
3.55; 3.56
4.8; 2.8
11.6; 5.03
11.6; 4.6
H-5ꢁ; H-5ꢁꢁ
H-5ꢁꢁ (1H)b 3.49; 3.48
OH (1H)
6.67
aAll carbon chemical shifts are based on C-H correlations, due to insufficient sample amount for
recording a 13C spectrum.
bSeparate shifts due to mixture of diastereomers.
progress of the reaction was followed daily by LC-DAD and LC-ESI-MS/MS
analyses of aliquots of the reaction mixture using the C18 analytical column.
GA (65 mg, 0.75 mmol) was reacted with 2ꢁ-deoxyguanosine (10 mg,
0.037 mmol) in 3 mL of 0.5 M phosphate buffer solutions at pH 7.0 and 9.0.
The reactions were performed at 37◦C. The progress of the reactions was
followed daily by LC-DAD and LC-ESI-MS/MS analyses of aliquots of the
reaction mixtures using the C18 analytical column.
Rearrangement of N1-GA-dA to N6-GA-dA and to N1-GA-dI
A sample of N1-GA-dA (130 µg) in 250 µL of 0.5 M phosphate buffer
(pH 7.0 and 9.0) was incubated at 37◦C, and aliquots were analyzed at
various times by LC-DAD and LC-ESI-MS/MS.
Determination of Nucleoside Adduct Yields
Quantitative 1H NMR analysis, using chloroform as an internal standard
was performed on the purified adducts. Standard solutions were prepared
for HPLC by taking an exact volume of NMR sample and diluting with an
appropriate volume of water. The quantitative determination of adducts in

138
J. Backman and L. Kronberg
the preparative scale reaction mixtures was made by comparing the peak
area of the adducts in HPLC standard solutions with the peak area of the
adducts in the reaction mixtures. The molar yields were calculated from the
original amount of deoxynucleoside in the reaction mixture.
RESULTS AND DISCUSSION
Reaction of GA with 2ꢁ-deoxyadenosine
A small-scale reaction was performed at various pH conditions to find
out the conditions that give the optimal yields of the adducts. The highest
yields of the compounds were obtained in the reaction carried out for 4 days
at pH 7.0. At these conditions three major products were formed. In the
LC-DAD chromatogram shown in Figure 1, the product peaks are marked
N1-GA-dA, N1-GA-dI, and N6-GA-dA. When the reaction was performed at
pH 4.6, only the adduct N1-GA-dA was formed.
For the determination of the structure of the compounds, a large-scale
reaction was performed at pH 7.0. After 4 days of reaction, the compounds
were isolated from the reaction mixture by semipreparative C18 column
chromatography. The isolated yields of N1-GA-dA, N1-GA-dI, and N6-GA-
dA were 8.0, 6.8, and 2.3 mol%, respectively. On the basis of data from
NMR and UV spectroscopy and mass spectrometry, the structures of the
adducts were assigned as N1-GA-dA, N1-GA-dI, and N6-GA-dA, respectively
(Scheme 2).
The mass spectrometric and UV and 1H NMR spectroscopic data
recorded for N1-GA-dA were in all essential features identical to those
presented by Gamboa da Costa et al.[8] In the work of Gamboa da Costa,
FIGURE 1 C18 analytical column HPLC chromatogram of the reaction mixture of glycidamide and
2ꢁ-deoxyadenosine held at 37◦C and pH 7.0 for 4 days. The chromatogram was recorded at 254 nm by
the UV diode-array detector.

Mechanism of Formation of Glycidamide Adducts
139
1
1
the H NMR signals were not fully assigned, but in Table 1 all H and 13C
NMR signals are assigned.
In the paper of Gamboa da Costa et al. it was found that the N1-GA-dA
adduct was unstable when stored at pH = 13 and yielded a chromophore
which was proposed to be N6-GA-dA (8). However, no further proof was
provided for the structural assignment of the adduct. In our hands, the N6-
dA adduct was found to be formed in the reaction of GA and dA (Figure 1).
The compound was isolated from the reaction mixture and was structurally
characterized by UV and NMR spectroscopy and mass spectrometry.
The UV spectrum of N6-GA-dA exhibited absorption maxima at 270 and
210 nm and an absorption minimum at 230 nm (Figure 2). These values
FIGURE 2 C18 analytical column HPLC chromatogram of the reaction mixture of glycidamide and
2ꢁ-deoxyguanosine held at 37◦C and pH 9.0 for 6 days. The chromatogram was recorded at 254 nm by
the UV diode-array detector.

140
J. Backman and L. Kronberg
are similar to the absorption maxima and minima for other reported N6-
deoxyadenosine adducts.[10].
In the positive ion electrospray mass spectrum of N6-GA-dA, the pro-
tonated molecular ion peak was observed at m/z = 340 and was the most
abundant ion. In the MS2 spectrum, the ion peak recorded at m/z = 224
corresponds to the cleavage of the deoxyribosyl moiety from the protonated
molecular ion (followed by the attachment of a proton to N-9). Upon
isolation and fragmentation of m/z = 224 (MS3), ion peaks were observed at
m/z = 206, m/z = 178, and m/z = 136. These correspond to the loss of water,
the loss of formic acid and to protonated adenine, respectively.
The NMR spectra of N6-GA-dA were recorded in D2O, since the reso-
1
nance peaks did not resolve as well in Me2SO-d6 as in D2O. The H NMR
spectrum of N6-GA-dA displayed, besides the signals from the protons of
the deoxyribosyl moiety, two one-proton singlets at δ = 8.15 ppm and 8.07
ppm (Table 2). The protons were assigned to H-8 and H-2, respectively.
The one-bond C-H correlation spectrum (heteronuclear multiple quantum
coherence, HMQC) showed the carbon signals at δ = 139.6 and 152 ppm
to be due to C-8 and C-2, respectively. Moreover, the spectrum showed
one-proton broad singlets at δ = 4.26 ppm, 3.83 ppm and 3.68 ppm. The
signals were assigned to H-11, H-10a, and H-10b, respectively. In the COSY
spectrum, correlations between these signals could be observed. The HMQC
spectrum showed the carbon signal at δ = 44.5 ppm to be due to C-10. A
correlation between H-11 and C-11 also could be observed, but the carbon
signal was overlapped by C-3ꢁ. When the N6-GA-dA spectrum was recorded in
DMSO-d6, the signals for the H-8 and H-2 protons appeared at δ = 8.31 ppm
and 8.20 ppm, respectively. These chemical shifts are in accordance with
those reported for the H-2 and H-8 protons in N6-adducts recorded in
Me2SO-d6.[11,12]
The UV spectrum of N1-GA-dI exhibited absorption maxima at 250 and
206 nm and an absorption minimum at 226 nm, with a shoulder between
266 and 290 nm (Figure 2). The UV spectrum of N1-GA-dI is consistent
with the spectral shape maxima of other N1-inosine adducts.[13]
In the positive ion electrospray mass spectrum of N1-GA-dI, the proto-
nated molecular ion peak was observed at m/z = 341. The product ion peak
of m/z = 341 was observed at m/z = 225 and was due to the loss of the
deoxyribosyl unit. This ion peak produced fragment peaks (MS3) at m/z =
179 and m/z = 137, which correspond to the loss of formic acid and to
protonated hypoxanthine, respectively (Scheme 3).
In the NMR spectra of N1-GA-dI (Table 3) some of the proton and
carbon atoms were represented by two signals with only slight differences in
chemical shifts, which show that two diastereomers of the compound were
present in the sample. In the 1H NMR spectrum, the resonance signals were
observed at chemical shifts close to those of N1-GA-dA. The only exception
was the one-proton signal for H-2 which was observed at δ = 8.23 ppm,

Mechanism of Formation of Glycidamide Adducts
141
SCHEME 3 The ions observed in the positive electrospray mass spectrum and proposed fragment
structures.
that is, 0.15 ppm more deshielded than in N1-GA-dA. This low field shift
can be explained by the more deshielding effect of the carbonyl group in
N1-GA-dI, compared with the exocyclic imino group in N1-GA-dA. The MS
fragmentation pattern was identical to that of N1-GA-dA, but the fragment
masses were one unit higher for the inosine adduct (Scheme 3).
Gamboa da Costa et al. reported that the N3-adenine adduct is formed
in the reaction of deoxyadenosine and GA.[8] The LC-MS/MS analyses of
our reaction mixtures showed the formation of a compound with the mass
of the N3-adenine adduct (m/z = 223), but in our hands the compound was
formed in yields too low for preparative isolation and subsequent structural
characterization. However, when DNA is reacted with GA, N3-GA-Ade is
one of the major adducts found in the reaction.[8] The reason for the
difference in the reactivity of N3 of the adenine base between the site in
the nucleoside and in DNA can be that the N3-position of adenine is not
involved in hydrogen bonding in DNA and is sterically exposed in DNA.
As a free nucleoside however, the N3-position may be blocked by the sugar
group or by intra- or intermolecular bonding.[14]
It is well known that N1-substituted adenosines rearrange to N6-
substituted adenosines through the Dimroth rearrangement. Simple N1-
alkyladenosines require an alkaline pH to undergo the rearrangement,[15]
while various N1-(2ꢁ-hydroxy)alkyladenines may undergo Dimroth rear-
rangement at neutral pH.[16] This is consistent with our findings; the
reaction of GA with 2ꢁ-deoxyadenosine was performed at pH 7 and the
N6-GA-dA adduct was detected already after 1 day of reaction. As further
proof for correct identification of the N6-adduct, we incubated the N1-GA-
dA adduct in a buffer solution at pH = 9.0, and the adduct was found to

142
J. Backman and L. Kronberg
be completely transformed to the N6-adduct within 24 hours. This finding
is consistent with other reports for N1-dA adducts of simple epoxides[17,18]
and of the findings of Gamboa da Costa et al.[8]
The deamination of N1-GA-dA yielding the N1-inosine adduct seems
to occur readily since the inosine adduct could be detected already after
1 day of reaction. To further study the conversion of N1-GA-dA to N1-GA-dI,
N1-GA-dA was incubated in neutral buffer solution (pH 7.0). After 2 days
of incubation, both the N1-GA-dI and N6-GA-dA could be observed. The
deamination of N1-GA-dA could proceed through a similar 5-membered
intermediate as was proposed by Barlow et al. when studying the deamina-
tion of the styrene oxide adduct 1-(2-hydroxy-1-phenylethyl)adenosine.[19]
But in the case of N1-GA-dA the deamination also could proceed through
a 6-membered lactone ring intermediate formed through displacement of
the amino group by the carboxyl group and subsequently the lactone ring
intermediate is opened by attack of water at the 6-position in the purine
ring. Which of the mechanisms are responsible for the formation of the
inosine adduct was not further explored.
Reaction of GA with 2ꢁ-deoxyguanosine
A small-scale reaction was performed at various pH conditions to find
out the conditions giving the optimal yields of the adducts. LC-DAD and
LC-ESI-MS/MS analyses of the reaction carried out at pH 7.0 showed the
formation of a single adduct which on the basis of the mass spectral data was
identified as the previously known N7-GA-Gua adduct.[7,8] Analyses of the
reaction performed at pH 9.0 revealed the formation of two major adducts,
N1-GA-dG I and N1-GA-dG II (Figure 3). In this reaction the N7 adduct
could not be detected.
For the purpose of determining the structure of the compounds, a
large-scale reaction was performed at pH 9.0. After 6 days of reaction, the
compounds were isolated from the reaction mixture by semipreparative
C18 column chromatography. On the basis of data from NMR and UV
spectroscopy and mass spectrometry, the structures of the adducts were
assigned as N1-GA-dG I and N1-GA-dG II, respectively (Scheme 2). The
isolated yields of the compounds were 3.8 and 0.6 mol%, respectively.
Identical UV spectra were recorded for the two adducts (Figure 4).
Absorption maxima at 254 and 204 nm with a shoulder between 270 nm
and 286 nm are features consistent with other N1-substituted guanosine
adducts.[11]
In the positive ion electrospray mass spectrum of N1-GA-dG I, the
protonated molecular ion peak was observed at m/z = 356. The product
ion peak of m/z = 356 was observed at m/z = 240 and was due to the loss
of the deoxyribosyl unit. This ion peak produced fragment peaks (MS3) at

Mechanism of Formation of Glycidamide Adducts
143
FIGURE 3 UV absorbance spectra of N1-GA-dA, N6-GA-dA, and N1-GA-dI. The UV spectra were
recorded with the diode-array detector as the compounds eluted from the column (N1-GA-dA: app.
2% ACN; N6-GA-dA: app. 3% ACN and N1-GA-dI: app. 8% ACN).
m/z = 194 and m/z = 152, which correspond to the loss of formic acid and
to protonated guanine, respectively (Scheme 3).
1
The H NMR spectrum of N1-GA-dG I displayed, besides the signals
from the deoxyribosyl moiety, a one-proton singlet at δ = 7.93 ppm (Table
4). This signal was assigned to the H-8 of the guanosine unit. A strong
correlation between this signal and a signal at δ = 135.5 ppm was observed
in the HMQC spectrum. Consequently, this signal was assigned to C-8.
Moreover, the spectrum showed broad singlets at δ = 4.29 ppm, 3.99 ppm,
and 3.74 ppm. These signals were assigned as H-10a, H-10b, and H-11,
respectively. Strong correlations between all these signals were observed in
the COSY spectrum. On the basis of C-H correlations, the carbon signals at
δ = 47.5 ppm and 72.1 ppm were assigned to C-10 and C-11. In the carbon
spectrum a weak signal was observed at δ = 171.5 ppm. This was assigned
to C-12. This low field chemical shift was similar to the chemical shifts of
C-12 observed in the spectra of N1-GA-dA and N1-GA-dI. On the basis of
the spectral data, the compound was assigned to N1-GA-dG I where the “GA
tail” had been hydrolyzed to the corresponding carboxylic acid.
In the positive ion electrospray mass spectrum of N1-GA-dG II, the
protonated molecular ion peak was observed at m/z = 355. In the MS2
spectrum, the ion peak recorded at m/z = 239 corresponds to the cleavage of
the deoxyribosyl moiety from the protonated molecular ion. Upon isolation
and fragmentation of m/z = 239 (MS3), ion peaks were observed at m/z =
222, m/z = 194 and m/z = 152. These correspond to the loss of ammonia,
the loss of formamide and to protonated guanine.

144
J. Backman and L. Kronberg
FIGURE 4 UV absorbance spectra of N1-GA-dG I and N1-GA-dG II. The UV spectra were recorded with
the diode-array detector as the compounds eluted from the column (N1-GA-dG I: app. 12% CAN and
N1-GA-dG II: app.16% ACN).
Since some of the protons gave rise to two resonance peaks close
1
to each other in the H NMR spectrum, N1-GA-dG II was isolated as a
mixture of two diastereomers (Table 5). The HPLC profile shows two close
chromatographic peaks of similar size for the adduct, also an indication of
the presence of two diastereomers. The 1H NMR spectrum of N1-GA-dG II
displayed, besides the signals from the deoxyribosyl moiety, a one-proton
singlet at δ = 7.90 ppm. This signal was assigned to H-8 of the guanosine
unit. Moreover, the spectrum showed two multiplets at δ = 3.63 and
3.33 ppm, and a doublet of doublets at δ = 4.02 ppm. All these signals
correlated strongly in the COSY spectrum. The HSQC spectrum revealed
that the multiplets were attached to the same carbon, and the carbon signal
was at δ = 44.6 ppm. Consequently, the multiplet signals were assigned
to H-10a and H-10b, respectively, and the doublet of doublets signal was
assigned to H-11. A C-H correlation in the HSQC spectrum showed that the
resonance signal at δ = 69.9 ppm was due to C-11. The proton signals at δ =
7.29 and 7.34 ppm were assigned to the amide group.

Mechanism of Formation of Glycidamide Adducts
145
The mass spectral data and the observation of the signals due to the
1
amide protons in the H NMR proved that the original amide function in
the “GA tail” was present and the structure of the compound was N1-GA-dG
II (Scheme 2).
Selzer and Elfarra have previously reported on an N1 adduct formed
in the reaction of butadiene monoxide and guanosine.[20] In this reaction,
the nucleophilic N1 reacted with the carbon adjacent to the double bond
in butadiene monoxide. This gives an adduct with a branched butadiene
monoxide tail on N1 of guanosine, instead of a straight butadiene monoxide
tail. We wanted to rule out this possible branched structure for our N1-
GA-dG adducts. Since no correlations could be observed between the H-10
protons and C-2 in the HMBC spectra for the two N1-GA-dG adducts (this
correlation could be observed for N1-GA-dA and N1-GA-dI), we instead
focused on the chemical shifts for C-10 and C-11. By comparing the
chemical shifts for C-10 (δ = 47.5 ppm for N1-GA-dG I and δ = 44.6 ppm
for N1-GA-dG II) and C-11 (δ = 72.1 ppm for N1-GA-dG I and δ = 69.9 ppm
for N1-GA-dG II) with chemical shifts found in the literature for similar
structures.[11] it was clear that our N1-GA-dG adducts had straight “GA-tails”
and not branched “GA-tails.” In a branched structure, C-10 adjacent to a
hydroxyl group could not have such a low field chemical shift as δ = 47.5
ppm or δ = 44.6 ppm. Instead, this C-10 must be attached to N1 of dG, thus
giving a straight “GA-tail.”
The Hydrolysis of the Amide Function
We previously have reported on a cytidine adduct of GA where the amide
function had been transformed to a carboxyl function.[9] In the current
work we found that the major adducts formed, that is, the N1-GA-dA, the
N6-GA-dA and the N1-GA-dG I adducts have the carboxyl function instead of
the amide function. The formation of the adducts with a carboxyl function
is surprising, since the amide group should not easily undergo hydrolysis.
In the reaction with dG, we found both the amide and the carboxyl
adduct. Any conversion of the amide adduct to the carboxyl adduct could
not be observed upon storage of the pure amide adduct for several days at
pH 7 and pH 9. This observation shows that the hydrolysis of the amide
function takes place before the epoxide ring of GA is attacked by N1 of
guanosine.
In the work of Golding et al.[21] concerning the reaction of glycidalde-
hyde and guanosine, it was found that at pH 10 the initial reaction is an
attack of the exocyclic amino group of guanosine on the carbonyl function
in glycidaldehyde followed by a ring closure by attack of N1 on the epoxide
ring. A similar sequence of reactions may also be valid for the GA reaction
(Scheme 4), but in this case we end up with the N1 carboxyl adduct. The
amide function is attacked by the exocyclic nitrogen and the resulting

146
J. Backman and L. Kronberg
SCHEME 4 The proposed reaction mechanisms for the formation of the carboxyl function in adducts
N1-GA-dG I and N1-GA-dA.
intermediate A loses ammonia and B is obtained. The N1 anion in the
intermediate B (the pKa of guanosine is approximately 9.4) attacks the less
substituted carbon in the epoxide ring forming the cyclic intermediate C.
This ring intermediate is hydrolysed and the ring is opened at N2, and the
carboxyl adduct N1-GA-dG is formed. The cyclization is favored by basic
conditions and seems not to take place at neutral conditions (only the N7
adduct was obtained at pH 7).
Upon LC-MS analyses of the dG reaction mixture, we could not find
any evidence for the presence of the intermediates A, B or C. However,
in the reaction mixture of dA and GA, a compound was present with
[M+H]+ = 322 mass units (D in Scheme 4). An intermediate with this mass
has previously been detected by Gamboa da Costa et al.[8]and Solomon[17]
and has been tentatively assigned as a cyclic structure (E). We isolated the
compound by collecting the peak when it eluted from the UV-detector and
found it to be unstable in water (pH = 7, 25 and 37◦C) yielding dA and
N1-GA-dA. The formation of dA shows that this intermediate cannot possess
the cyclic structure depicted as E, since the bond between N1 and the
exocyclic carbon should be stable. This argument is based on the fact that
no dA was formed upon storage of N1-GA-dA in solution (pH 7) for several
days, only Dimroth rearrangement yielding N6-GA-dA was observed. On the
other hand, structure D for the intermediate explains our observations.
D may be hydrolysed to dA or it can form the intermediate E, which in
turn is hydrolysed to N1-GA-dA. The conclusion of this discussion is that
the hydrolysis of the amide seems to take place through a transamidation
mechanism originally proposed by us in the paper dealing with the identifi-
cation of cytidine and thymidine adducts of GA.[9] Further support for the

Mechanism of Formation of Glycidamide Adducts
147
transamidation was the observation that the 1H NMR spectrum recorded of
a D2O solution of GA and n-propylamine showed that the amino function
attacked the carbonyl group of GA and not the epoxide ring (data not
shown). Although it could be expected that the exocyclic aminogroup in dA
should be less reactive than the amino group in n-propylamine, it is obvious
that the dA amino group attacks GA and that the attack is on the amide
function.
The amide adducts, N1-GA-dG II, N7-GA-Gua,[7] and N3-GA-Ade[8] are
most likely formed by direct attack of the nucleophilic endocyclic nitrogens
on the oxirane ring of GA.
CONCLUSIONS
In this study, we fully characterized the major adducts formed in the
reaction of GA with 2ꢁ-deoxyadenosine at physiological conditions, and the
adducts formed in the reaction of GA with 2ꢁ-deoxyguanosine at pH 9.0 and
37◦C.
We also propose a plausible mechanism for the hydrolysis of the amide
function. The mechanism is supported by the identification of the degrada-
tion products of the intermediate in the dA reaction and by NMR studies
of the reaction product of n-propylamine and GA. The adducts N7-GA-Gua,
N3-GA-Ade and N1-GA-dG II are most likely formed through a direct attack
of the endocyclic N1 in the purine on the β-carbon in the oxirane ring,
in the case of N1-GA-dA and N1-GA-dG I the exocyclic aminogroups of
dA and dG attack the carbonyl carbon in GA and the attack is followed by
deamination and ring closure through reaction of the oxirane ring with the
nucleophilic ring nitrogen of dA and/or dG. N6-GA-dA and N1-GA-dI are
formed from N1-GA-dA through Dimroth rearrangement and a succeeding
deamination, respectively.
The adducts characterized in this work may be used as reference
compounds in studies aiming at the identification of adducts in laboratory
animals exposed to GA.
REFERENCES
1. IARC, IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans (60, IARC,
Lyon, France, 1994) 389–433.
2. Tareke, E.; Rydberg, P.; Karlsson, P.; Eriksson, S.; To¨rnqvist, M. Analysis of acrylamide, a carcinogen
formed in heated foodstuffs. J. Agric. Food. Chem. 2002, 50, 4998–5006.
3. Solomon, J.J.; Fedyk, J.; Mukai, F.; Segal, A. Direct alkylation of 2ꢁ-deoxynucleosides and DNA
following in vitro reaction with acrylamide. Cancer Res. 1985, 45, 3465–3470.
4. Besaratinia, A.; Pfeifer, G.P. Genotoxicity of acrylamide and glycidamide. J. Natl. Cancer Inst. 2004,
96, 1023–1029.
5. Paulsson, B.; Kotova, N.; Grawe´, J.; Henderson, A.; Granath, F.; Golding, B.; To¨rnqvist, M. Induction
of micronuclei in mouse and rat by glycidamide, genotoxic metabolite of acrylamide. Mutation
Research 2003, 535, 15–24.

148
J. Backman and L. Kronberg
6. Ghanayem, B.I.; Witt, K.L.; Kissling, G.E.; Tice, R.R.; Recio, L. Absence of acrylamide-induced
genotoxicity in CYP2E1-null mice: Evidence consistent with a glycidamide-mediated effect. Mutation
Research 2005, 578, 284–297.
7. Segerba¨ck, D.; Calleman, C.J.; Schroeder, J.L.; Costa, L.G.; Faustman, E.M. Formation of N-7-(2-
carbamoyl-2-hydroxyethyl)guanine in DNA of the mouse and the rat following intraperitoneal
administration of [14C]acrylamide. Carcinogenesis 1995, 16, 1161–1165.
8. Gamboa da Costa, G.; Churchwell, M.I.; Hamilton, L.P.; Von Tungeln, L.S.; Beland, F.A.; Marques,
M.M.; Doerge, D.R. DNA Adduct formation from acrylamide via conversion to glycidamide in adult
and neonatal mice. Chem. Res. Toxicol. 2003, 16, 1328–1337.
9. Backman, J.; Sjo¨holm, R.; Kronberg, L. Characterization of the adducts formed in the reactions of
glycidamide with thymidine and cytidine. Chem. Res. Toxicol. 2004, 17, 1652–1658.
10. Florea-Wang, D.; Haapala, E.; Mattinen, J.; Hakala, K.; Vilpo, J.; Hovinen, J. Reactions of N,N-bis(2-
chloroethyl)-p-aminophenylbutyric acid (chlorambucil) with 2ꢁ-deoxyadenosine. Chem. Res. Toxicol.
2003, 16(3), 403–408.
11. Munter, T.; Cottrell, L.; Hill, S.; Kronberg, L.; Watson, W.P.; Golding, B.T. Identification of adducts
derived from reactions of (1-chloroethenyl)oxirane with nucleosides and calf thymus DNA. Chem.
Res. Toxicol. 2002, 15, 1549–1560.
12. Olsen, R.; Molander, P.; Ovrebo, S.; Ellingsen, D.G.; Thorud, S.; Thomassen, Y.; Lundanes, E.;
Greibrokk, T.; Backman, J.; Sjoeholm, R.; Kronberg, L. Reaction of glyoxal with 2ꢁ-deoxyguanosine,
2ꢁ-deoxyadenosine, 2ꢁ-deoxycytidine, cytidine, thymidine, and calf thymus DNA: Identification of
DNA adducts. Chem. Res. Toxicol. 2005, 18(4), 730–739.
13. Singer, B.; Grunberger, D. Molecular Biology of Mutagens & Carcinogens (Plenum Press, New York,
1983).
14. Carrell, H.L.; Glusker, J.P.; Moschel, R.C.; Hudgins, W.R.; Dipple, A. Crystal structure of a
carcinogen: nucleoside adduct. Cancer Res. 1981, 41, 2230–2234.
15. Macon, J.B.; Wolfenden, R. 1-Methyladenosine. Dimroth rearrangement and reversible reduction.
Biochemistry 1968, 7, 3453–3458.
16. Fujii, T.; Saito, T.; Terahara, N. Purines. XXVII. Hydrolytic deamination versus Dimroth rearrange-
ment in the 9-substituted adenine ring: Effect of an omega-hydroxyalkyl group at the 1-position.
Chem. Pharm. Bull. 1986, 34, 1094–1107.
17. Solomon, J.J. Cyclic adducts and intermediates induced by simple epoxides. In Exocyclic DNA Adducts
in Mutagenesis and Carcinogenesis, Eds. Singer, B., Bartsch, H. (IARC Scientific Publication No. 150, p
International Agency for Research on Cancer, Lyon, France, 1999) 123–135.
18. Selzer, R.R.; Elfarra, A.A. Characterization of N1- and N 6-adenosine adducts and N1-inosine adducts
formed by the reaction of butadiene monoxide with adenosine: evidence for the N1-adenosine
adducts as major initial products. Chem. Res. Toxicol. 1996, 9, 875–881.
19. Barlow, T.; Ding, J.; Vouros, P.; Dipple, A. Investigation of hydrolytic deamination of 1-(2-hydroxy-1-
phenylethyl)adenosine. Chem. Res. Toxicol. 1997, 10, 1247–1249.
20. Selzer, R.R.; Elfarra, A.A. Synthesis and biochemical characterization of N1-, N2-, and N7-guanosine
adducts of butadiene monoxide. Chem. Res. Toxicol. 1996, 9, 126–132.
21. Golding, B.T.; Slaich, P.T.; Kennedy, G.; Bleasdale, C.; Watson, P.T. Mechanisms of formation of
adducts from reactions of glycidaldehyde with 2ꢁ-deoxyguanosine and/or guanosine. Chem. Res.
Toxicol. 1996, 9, 147–157.