Exocyclic deoxyadenosine adducts of 1,2,3,4-diepoxybutane: Synthesis, structural elucidation, and mechanistic studies

Article abstract of DOI:10.1021/tx900312e

1,2,3,4-Diepoxybutane (DEB) is considered the ultimate carcinogenic metabolite of 1,3-butadiene, an important industrial chemical and environmental pollutant present in urban air. Although it preferentially modifies guanine within DNA, DEB induces a large number of A → T transversions, suggesting that it forms strongly mispairing lesions at adenine nucleobases. We now report the discovery of three potentially mispairing exocyclic adenine lesions of DEB: N⁶,N⁶-(2,3-dihydroxybutan-1,4-diyl)-2′- deoxyadenosine (compound 2), 1,N⁶-(2-hydroxy-3-hydroxymethylpropan-1, 3-diyl) -2′-deoxyadenosine (compound 3), and 1,N⁶-(1- hydroxymethyl-2-hydroxypropan-1,3-diyl)-2′-deoxyadenosine (compound 4). The structures and stereochemistry of the novel DEB-dA adducts were determined by a combination of UV and NMR spectroscopy, tandem mass spectrometry, and independent synthesis. We found that synthetic N⁶-(2-hydroxy-3,4- epoxybut-1-yl)-2′-deoxyadenosine (compound 1) representing the product of N⁶-adenine alkylation by DEB spontaneously cyclizes to form 3 under aqueous conditions or 2 under anhydrous conditions in the presence of an organic base. Compound 3 can be interconverted with 4 by a reversible unimolecular pericyclic reaction favoring 4 as a more thermodynamically stable product. Both 3 and 4 are present in double stranded DNA treated with DEB in vitro and in liver DNA of laboratory mice exposed to 1,3-butadiene by inhalation. We propose that in DNA under physiological conditions, DEB alkylates the N-1 position of adenine in DNA to form N1-(2-hydroxy-3,4-epoxybut-1-yl)-adenine adducts, which undergo an S_N2-type intramolecular nucleophilic substitution and rearrangement to give 3 (minor) and 4 (major). Formation of exocyclic DEB-adenine lesions following exposure to 1,3-butadiene provides a possible mechanism of mutagenesis at the A:T base pairs.

Full text of DOI:10.1021/tx900312e

118
Chem. Res. Toxicol. 2010, 23, 118–133
Exocyclic Deoxyadenosine Adducts of 1,2,3,4-Diepoxybutane:
Synthesis, Structural Elucidation, and Mechanistic Studies
Uthpala Seneviratne,†,‡ Sergey Antsypovich,†,§ Melissa Goggin,† Danae Quirk Dorr,†,|
Rebecca Guza,† Adam Moser,‡ Carrie Thompson,† Darrin M. York,‡ and
Natalia Tretyakova*,†
Department of Medicinal Chemistry and Masonic Cancer Center, and Department of Chemistry, UniVersity of
Minnesota, Minneapolis, Minnesota 55455
ReceiVed September 1, 2009
1,2,3,4-Diepoxybutane (DEB) is considered the ultimate carcinogenic metabolite of 1,3-butadiene, an
important industrial chemical and environmental pollutant present in urban air. Although it preferentially
modifies guanine within DNA, DEB induces a large number of A f T transversions, suggesting that it
forms strongly mispairing lesions at adenine nucleobases. We now report the discovery of three potentially
mispairing exocyclic adenine lesions of DEB: N6,N6-(2,3-dihydroxybutan-1,4-diyl)-2′-deoxyadenosine
(compound 2), 1,N6-(2-hydroxy-3-hydroxymethylpropan-1,3-diyl)-2′-deoxyadenosine (compound 3), and
1,N6-(1-hydroxymethyl-2-hydroxypropan-1,3-diyl)-2′-deoxyadenosine (compound 4). The structures and
stereochemistry of the novel DEB-dA adducts were determined by a combination of UV and NMR
spectroscopy, tandem mass spectrometry, and independent synthesis. We found that synthetic N6-(2-
hydroxy-3,4-epoxybut-1-yl)-2′-deoxyadenosine (compound 1) representing the product of N6-adenine
alkylation by DEB spontaneously cyclizes to form 3 under aqueous conditions or 2 under anhydrous
conditions in the presence of an organic base. Compound 3 can be interconverted with 4 by a reversible
unimolecular pericyclic reaction favoring 4 as a more thermodynamically stable product. Both 3 and 4
are present in double stranded DNA treated with DEB in vitro and in liver DNA of laboratory mice
exposed to 1,3-butadiene by inhalation. We propose that in DNA under physiological conditions, DEB
alkylates the N-1 position of adenine in DNA to form N1-(2-hydroxy-3,4-epoxybut-1-yl)-adenine adducts,
which undergo an SN2-type intramolecular nucleophilic substitution and rearrangement to give 3 (minor)
and 4 (major). Formation of exocyclic DEB-adenine lesions following exposure to 1,3-butadiene provides
a possible mechanism of mutagenesis at the A:T base pairs.
posed ultimate carcinogenic metabolite of 1,3-butadiene (BD)
(11). 1,3-Butadiene is a known animal and human carcinogen
found in automobile exhaust and in cigarette smoke (12, 13).
All three possible stereoisomers of DEB, S,S, R,R, and meso,
are generated metabolically (14). Although DEB is a relatively
minor metabolite of BD (laboratory mice exposed to 62.5 ppm
BD for 10 days contain ∼247 pmol DEB/g blood (15)),
experimental evidence suggests that it is responsible for many
of the adverse effects of BD. DEB is 50-100-fold more
genotoxic and mutagenic in human cells than its monoepoxide
analogues, 3,4-epoxy-1-butene (EB) and 3,4-epoxy-1,2-butane-
diol (EBD) (16, 17). Efficient metabolism of BD to DEB in
target tissues of laboratory mice is thought to cause the increased
susceptibility of this species to BD carcinogenesis (18).
Introduction
Exocyclic nucleobase adducts are among the most important
types of DNA damage because of their ability to exert significant
biological effects (1-3). These lesions are characterized by
considerable changes of the molecular size/shape and hydrogen
bonding characteristics of the parent nucleobase, leading to
mispairing during DNA synthesis (4-9). For example, 1,N6-
etheno-deoxyadenosine adducts induced by vinyl chloride
preferentially adopt the syn conformation about the glycosidic
bond, forming a Hoogsteen base pair with guanine or cytosine
instead of the normal adenine partner, thymine (4, 10).
One prominent bis-electrophile capable of inducing exocyclic
nucleobase lesions is 1,2,3,4-diepoxybutane (DEB1), the pro-
* Corresponding author. Masonic Cancer Center,
University of Minnesota, MMC 806, 420 Delaware St. S.E., Minneapolis,
MN 55455. Tel: (612) 626-3432. Fax: (612) 626-5135. E-mail:
trety001@umn.edu.
BD, 1,3-butadiene; BzNH2, benzylamine; dA, 2′-deoxyadenosine; DEB,
1,2,3,4-diepoxybutane; DIPEA, N,N-diisopropylethyl amine; DFT, density
functional theory; DMSO, dimethylsulfoxide; dR, 2′-deoxyribose; EtOH,
Ethanol; Fmoc, 9-fluorenylmethoxycarbonyl; gCOSY, 1H-1H gradient
correlation spectroscopy; gHMBC, gradient heteronuclear multiple bond
correlation; gHMQC, gradient heteronuclear multiple quantum correlation;
HEB, 2-hydroxy-3,4-epoxybut-1-yl; HOMO, highest occupied molecular
orbital; HPLC-ESI-MS, high pressure liquid chromatography-electrospray
ionization mass spectrometry; HSQC, heteronuclear single quantum cor-
relation; mCPBA, m-chloroperoxybenzoic acid; MeOH, methanol; N7G-
N1A-BD, 1-(guan-7-yl)-4-(aden-1-yl)-2,3-butanediol; NOESY, nuclear
Overhauser effect spectroscopy; PyBOP, benzotriazol-1-yloxy-tripyrroli-
dinophosphonium hexafluorophosphate; SPE, solid phase extraction; THF,
tetrahydrofuran; TMS-CN, trimethylsilyl cyanide; TOCSY, total correlation
spectroscopy.
† Department of Medicinal Chemistry and Masonic Cancer Center.
‡ Department of Chemistry.
§ Current address: Department of Chemistry, Moscow State University,
Moscow, Russia.
| Current address: Department of Chemistry and Geology, Minnesota
State University, Mankato, MN.
1 Abbreviations: 1, N6-(2-hydroxy-3,4-epoxybut-1-yl)-2′-deoxyadenosine
(N6-HEB-dA); 2, N6,N6-(2,3-dihydroxybutan-1,4-diyl)-2′-deoxyadenosine
(N6,N6-DHB-dA); 3, 1,N6-(2-hydroxy-3-hydroxymethylpropan-1,3-diyl)-2′-
deoxyadenosine (1,N6-γ HMHP-dA); 4, 1,N6-(1-hydroxymethyl-2-hydrox-
ypropan-1,3-diyl)-2′-deoxyadenosine (1,N6-R HMHP-dA); BCNU, 1,3-bis(2-
chloroethyl)-1-nitroso-urea; bis-N7G-BD, 4-bis-(guan-7-yl)-2,3-butanediol;
10.1021/tx900312e CCC: $40.75  2010 American Chemical Society
Published on Web 11/02/2009

Exocyclic Deoxyadenosine Adducts of DEB
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 119
Chart 1.
The adverse biological effects of DEB have been attributed
to its ability to cross-link cellular biomolecules. Initial alkylation
of adenine and guanine bases in DNA by DEB produces
2-hydroxy-3,4-epoxybut-1-yl (HEB) lesions, which contain an
inherently reactive oxirane group and can alkylate neighboring
nucleobases within the DNA duplex to form DNA-DNA cross-
links, namely, 1,4-bis-(guan-7-yl)-2,3-butanediol (bis-N7G-BD)
and 1-(guan-7-yl)-4-(aden-1-yl)-2,3-butanediol (N7G-N1A-BD)
(19, 20). The S,S isomer of DEB produces the highest number
of interstrand DNA-DNA cross-links (21) and is the most
cytotoxic (22, 23) and mutagenic (24). Alternatively, the 3,4-
epoxy group of the HEB adducts can be subject to nucleophilic
attack by another site within the same DNA nucleobase, giving
rise to fused ring structures (25-27).
The documented ability of DEB to induce large numbers of
A f T transversion mutations (28, 29) has led us to hypothesize
that it forms strongly mispairing exocyclic lesions at adenine
nucleobases within DNA. This hypothesis was supported by
our previous work with synthetic DNA oligonucleotides con-
taining site specific N6-(2-hydroxy-3,4-epoxybut-1-yl)deoxyad-
enosine adducts (N6-HEB-dA, 1 in Chart 1) (30). If left in an
aqueous solution at room temperature (pH 7.2), 1 underwent
spontaneous cyclization to form previously unidentified DEB-
dA lesions (see Supporting Information Figure S-1). Another
isomer of the exocyclic DEB-dA species was formed as a side
product during the synthesis of compound 1 (by reacting
6-chloropurine-2′-deoxyriboside with 1-amino-2-hydroxy-3,4-
epoxybutane under basic, anhydrous conditions) (30). In the
present work, we employed a combination of UV and NMR
spectroscopy, tandem mass spectrometry, independent synthesis,
DFT calculations, and kinetic analyses to identify the chemical
structures of these novel DEB-DNA lesions and to establish
the mechanism of their formation.
Materials and Methods
2′-Deoxyinosine was obtained from Berry & Associates, Inc.
(Dexter, MI). 15N5-dA was purchased from Spectra Stable Isotopes,
Div. of Spectra Gases Inc. (Columbia, MD). d,l-DEB was obtained
from the NCI Chemical Carcinogen Repository, and the individual
stereoisomers of DEB were prepared as described previously (21).
All other chemicals and enzymes were purchased from Sigma-
Aldrich Chemical Co. (Milwaukee, WI).
Instrumentation. NMR spectra were acquired with a Varian
Inova 600 MHz, 800 MHz spectrometer (Varian Inc., Palo Alto,

120 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
SeneViratne et al.
Figure 1. UV (left panel) and MS3 spectra (right panel) of synthetic compounds 1 (A), 2 (B), 3 (C), and 4 (D).
CA) and a Bruker Advance 700 MHz instrument (Bruker BioSpin,
Billerica, MA) using DMSO-d6 as the solvent. UV spectrophotom-
etry was performed on a Beckman DU-7400 instrument (Beckman,
Fullerton, CA). Optical rotation mesurements were performed with
a Rudolph Autopol III automatic polarimeter at 589 nm (Rudolph
Research, Flanders, NJ), and HRMS data were obtained using a
Bruker Bio-TOF II instrument.
HPLC-UV Analyses. HPLC was carried out with an Agilent
Technologies model 1100 HPLC system equipped with a photo-
diode array UV detector (Wilmington, DE). Unless specified
otherwise, UV absorbance was monitored at 254 nm. HPLC
columns and solvent elution systems were as follows.
System 1. A semipreparative Zorbax Eclipse XDB-C18 (25 cm
× 9.4 mm, 5 µm) column was eluted with a linear gradient of
acetonitrile (B) in water (A) at a flow rate of 3 mL/min. Solvent
composition was changed from 5% to 10.3% B in 35 min, then
further to 90% B in 2 min. This system was used for the isolation
of exocyclic DEB-dA adducts (compounds 1 and 2) and the
corresponding free bases. To separate 2a from 2b, the same column
with an isocratic elution of 3% acetonitrile in water was used.
System 2. A preparative silica Whatman Particil 10 (50 cm ×
9.4 mm, 10 µm) column was eluted isocratically with 0.5%
methanol in chloroform at a flow rate of 5 mL/min. This system
was used to separate the diastereomers of N-Fmoc-1-amino-2-
hydroxy-3,4-epoxybutanes (compounds 5a, 5b, 5c, and 5d).
System 3. A 100 mm × 3 mm Thermo Hypersil Hypercarb
column (10 cm × 3 mm, 5 µm particle size, Thermo Fisher
Scientific, Inc. Waltham, MA) was eluted with a linear gradient of
acetonitrile (B) in 0.05% acetic acid (A). Solvent composition was
changed from 0 to 20% B in 30 min. The column was eluted at a
flow rate of 0.5 mL/min. This system was used for the separation
of 3 and 4 and the corresponding free bases.
Synthesis of N-Fmoc-1-Amino-2-hydroxy-3,4-epoxybutane,
5a-5d (Scheme 4). Racemic 1-aminobut-3-ene-2-ol was prepared
from acrolein as described previously (31) and separated by
diastereomeric resolution following derivatization with sodium
[(1R)-(endo, anti)]-(+)-3-bromocamphor-8-sulfonic acid ammonium
salt (32, 33), followed by Fmoc protection and epoxidation in the
presence of mCPBA (30). Individual stereoisomers of the amino
epoxide were isolated by preparative HPLC on a silica column
(HPLC system 2). 5a, [R]25 +4.0 (in CHCl3), ee ) 100%; 5b,
D
[R]25 -5.7 (in CHCl3), ee ) 100%; 5c, [R]25 -3.4 (in CHCl3),
D
ee )D92.5% and 5d, [R]25 +3.9 (in CHCl3), ee ) 84.2%.
D
Synthesis of N6,N6-DHB-dA (Scheme 1 and Chart 1). S,S-,
R,R-, and meso-N6,N6-DHB-dA (compounds 2a-c in Chart 1) were
obtained by coupling commercial 6-chloropurine-2′-deoxyriboside
with S,S-, R,R-, or meso-1-amino-2-hydroxy-3,4-epoxybutane,
respectively (Scheme 1) (30). 6-Chloropurine-2′-deoxyriboside (1.45
mg, 5.36 µmol) was dissolved in DIPEA (18.7 µL), and N-Fmoc-
1-amino-2-hydroxy-3,4-epoxybutane (6.95 mg, 21.4 µmol) dis-
solved in DMSO (200 µL) was added. The reaction mixture was
incubated at 37 °C for 2 days under Ar atmosphere. The reaction
products were diluted with water, filtered, and separated using
HPLC system 1. Under these conditions, R,R- and S,S-N6,N6-DHB-

Exocyclic Deoxyadenosine Adducts of DEB
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 121
dA (compounds 2a and 2b) eluted at 15.4 min, while meso-N6,N6-
DHB-dA (compound 2c) was observed as a peak at 16.7 min (2.95
mg, 8.73 µmol, 40.8%). Compounds 2a and 2b were resolved with
a different HPLC system (3% ACN in water): 2a, tR, 55.1 min
(1.59 mg, 22.0% yield),) and 2b, tR, 56.6 min (1.74 mg, 24.0%
yield). UV (water): λmax ) 274 nm (Figure 1, left), ESI+-MS m/z
338.3 [M + H]+; MS/MS m/z 338.3 f 222.2 [M + 2H-dR]+; MS3
m/z 338.3 f m/z 222.2 f m/z 136.0 [Ade + H]+, 147.6 [Ade-CH2
+ H]+, 185.6 [M + 2H-dR-2H2O]+, 204.0 [M + 2H-dR-H2O]+
(Figure 1, right panel).
Chart 2.
1H NMR (DMSO-d6), compound 2a (S,S N6,N6-DHB-dA): 8.32
(s, 1H, H-8); 8.18 (s, 1H, H-2); 6.35 (t, 1H, H-1′, Jvic ) 6.96 Hz);
5.30 (d, 1H, CHOH-3′, Jvic ) 3.66 Hz); 5.22 (t, 1H, CHOH-5′, Jvic
) 5.50 Hz); 5.19 (bs, 1H, CHOHꢀ); 5.17 (bs, 1H, CHOHγ); 4.40
(bs, 1H, H-3′); 4.17 (d, 1H, HR, Jgem ) 12.21 Hz); 4.09 (m, 1H,
Hꢀ); 4.01 (m, 2H, HR, Hγ); 3.87 (d, 1H, H-4′, Jvic ) 2.69 Hz); 3.68
(m, 1H, Hδ); 3.62 (m, 2H, Hδ, H-5′); 3.51 (m, 1H, H-5′); 2.69 (q,
1H, H-2′); 2.26 (dq, 1H, H-2′, Jgem ) 13.18 Hz).
1H NMR (DMSO-d6), compound 2b (R,R N6,N6-DHB-dA): 8.32
(s, 1H, H-8); 8.18 (s, 1H, H-2); 6.35 (t, 1H, H-1′, Jvic ) 6.84 Hz);
5.22 (bs, 4H, OH-groups); 4.40 (q, 1H, H-3′, J ) 2.85 Hz); 4.17
(bd, 1H, HR, J ) 12.6 Hz); 4.10 (bs, 1H, Hꢀ); 4.02 (bs, 1H, HR);
4.00 (bs, 1H, Hγ); 3.87 (q, 1H, H-4′, J ) 4.2 Hz); 3.69 (bd, 1H,
Hδ); 3.63 (m, 1H, Hδ); 3.61 (m, 1H, H-5′); 3.51 (dd, 1H, H-5′, J1
) 11.70 Hz, J2 ) 4.50 Hz); 2.68 (m, 1H, H-2′); 2.25 (dq, 1H,
H-2′, J1 ) 13.20 Hz, J2 ) 3.20 Hz).
temperature). The reaction mixtures were resolved by reversed phase
HPLC (system 1) to give (R,R)-N6-HEB dA, (1a), (R,S)-N6-HEB
dA, (1b), (S,S)-N6-HEB dA, (1c), and (S,R)-N6-HEB dA, (1d)
separately in 95% yield.
Synthesis of Compounds 3 and 4 in Chart 1. Racemic N6-
HEB-dA (compound 1 in Chart 1, 1.5 mg) (30) was dissolved in
250 µL of water and incubated for 15 h at room temperature.
Spontaneous cyclization of N6-HEB-dA produced 3 and 4, which
were isolated with HPLC system 3: 3, tR, 12.5 min and 4, tR, 16.8
min (3, 0.87 mg, 2.55 µmol, 57.3% yield and 4, 0.57 mg, 1.67
1H NMR (DMSO-d6), compound 2c (S,R and R,S N6,N6-DHB-
dA): 8.33 (s, 1H, H-8); 8.19 (s, 1H, H-2); 6.36 (t, 1H, H-1′, Jvic
)
6.9 Hz); 5.31 (d, 1H, CHOH-3′, Jvic ) 4.2 Hz); 5.21 (t, 1H, CHOH-
5′, Jvic ) 5.7 Hz); 5.03 (d, 1H, CHOHꢀ, Jvic ) 3.7 Hz); 4.96 (bs,
1H, CHOHγ); 4.40 (m, 1H, H-3′, J1 ) 5.7 Hz; J2 ) 2.9 Hz); 4.21
(d, 1H, HR, Jgem ) 12.0 Hz); 4.17 (bs, 1H, Hꢀ); 4.13 (bs, 1H, Hγ);
3.91 (d, 1H, HR, Jgem ) 10.2 Hz); 3.88 (q, 1H, H-4′, Jvic ) 4.2
Hz); 3.72 (d, 1H, Hδ Jgem ) 12.0 Hz); 3.61 (m, 2H, H-5′, Hδ, Jgem
) 11.71 Hz; Jvic) 4.4 Hz); 2.70 (m, 1H, H-2′, Jgem ) 13.2 Hz; Jvic
) 6.3 Hz); 2.26 (m, 1H, H-2′, Jgem ) 13.2 Hz; Jvic ) 6.6 Hz).
Computational Methods. Density-functional electronic structure
calculations were performed according to the protocol used in the
QCRNA database (34). The calculations included geometry opti-
mization, vibrational frequency analysis, and solvation energy
corrections. Kohn-Sham density-functional calculations were
performed using the hybrid exchange functional of Becke (35) and
the Lee, Yang, and Parr correlation functional (36) (B3LYP) as
implemented in the Gaussian 03 suite of programs (see http://
www.gaussian.com/citation_g03.htm). Relaxed potential energy
surface scans around the C-N torsion angle were performed at
the B3LYP/6-31+G(d,p) level with the torsion constraint set in
15° intervals, with additional unconstrained (fully optimized) points
at the minima.
µmol, 38.0% yield, ratio of 3:4; 7:3). UV (pH 7) 3 and 4, λmax
)
264 nm. Compound 3: ESI+-MS m/z 338.3 [M + H]+, MS2 m/z
338.3 f m/z 222.2 [M + 2H-dR]+; MS3 m/z 338.3 f m/z 222.2
f m/z 136.0 [Ade + H]+, 148.0 [Ade-CH2 + H]+, 186.0 [M +
2H-dR-2H2O]+. Compound 4: ESI+-MS m/z 338.3 [M + H]+, MS2
m/z 338.3 f m/z 222.2 [M + 2H-dR]+; MS3 m/z 338.3 f m/z
222.2 f m/z 160.6 [Ade-CCH + H]+, 177.5 [Ade-(CH)2OH +
H]+, 203.6 [M + 2H-dR-H2O]+.
Synthesis of 3a-3d and 4a-4d. Compounds 1a-1d were
dissolved in 250 µL of water and incubated for 12 h at room
temperature. The cyclization products were purified with HPLC
system 3. Compounds 3a and 4a were obtained from the cyclization
of 1a; 3b, and 4b were obtained from the cyclization of 1b; 3c,
and 4c were obtained from the cyclization of 1c; 3d, and 4d were
obtained from the cyclization of 1d. Compound 3a, tR, 12.5 min;
4a, tR, 16.8 min; 4b, tR, 19.6 min; and 3b, tR, 21.7 min; 3c, tR, 12.5
min; 4c, tR, 16.8 min; 4d, tR, 19.6 min; and 3d, tR, 21.7 min. The
1H NMR, NOESY, gCOSY, gHSQC, and HMBC data for com-
pounds 3a, 3b, 4a, and 4b are presented in Supporting Information
Figures S-14-S-18.
Molecular orbitals for the axial, equatorial, and pucker transition
state structures of 2a-base were calculated. All structures showed
at least 7 orbitals with significant density above and below the C-N
σ bond. Supporting Information Figures S-12A and S-12B show a
top view and side view of two molecular orbitals (HOMO-2 and
HOMO-20) of compound 2a-base in an equatorial-like conformation.
Independent Synthesis of Compounds 2a, 2b, and 2c
(Scheme 2). 3R,4R-Pyrrolidine-2,3-diol (125 mg, synthesized as
described previously) (37, 38) was coupled with 2′-deoxyinosine
(250 mg) in the presence of DMSO (5 mL), DIPEA (210 µL), and
PyBOP (656 mg) (55 °C for 3 days). The reaction mixture was
resolved by flash chromatography (CH2Cl2/MeOH solvent gradient
using a silica column) and reversed phase HPLC (system 1) to give
optically pure (R,R)-N6,N6-DHB dA, [R]25D +43.0 (in MeOH) (2a)
in 72% yield. The same synthetic route starting with L-tartaric acid
and meso-tartaric acid was followed to obtain the optically pure
15N4-Deoxyinosine. 15N5-Deoxyadenosine (10 mg, 80 µmol) was
incubated in the presence of adenosine deaminase (80 units) in 10
mM Tris-HCl, pH 7.4 (0.75 mL), at 37 °C for 24 h. HPLC analysis
revealed a single peak corresponding to 15N5-deoxyinosine. The
mixture was filtered through a YM-30 Centricon filter to remove
any protein and desalted using Waters C18 SPE cartridges (95%
yield).
Synthesis of 15N4-Labeled Compounds 3 and 4. To a solution
of 15N4-deoxyinosine (0.026 mmol, 6.5 mg) in DMSO (2.5 mL),
PyBOP (0.026 mmol, 13.5 mg), DIPEA (0.11 mmol, 19 µL), and
racemic compound 5 (0.022 mmol, 7 mg) was added (39). The
mixture was stirred for 3 days at room temperature under argon to
give 15N4-N6-HEB-dA. ESI+-MS m/z 342.3 [M + H]+, MS2 m/z
342.3 f m/z 226.2 [M + 2H-dR]+; MS3 m/z 342.3 f m/z 226.2
f m/z 140.0 [15N4-HX + H]+, 152.0 [15N4-HX-CH2 + H]+, 190.0
[M + 2H-dR-2H2O]+, 208.0 [M + 2H-dR-H2O]+. 15N4-N6-HEB-
dA was quantitatively converted to 15N4 -labeled compounds 3 and
4 by incubating in water overnight, followed by HPLC purification
with HPLC system 3. Standard solutions of 15N4-labeled 3 and 4
were prepared in water and quantified by UV spectrophotometry.
(S,S)-N6,N6-DHB dA, [R]25 -42.8 (in MeOH) (2b) and meso-
D
N6,N6-DHB dA, [R]25 -19.4 (in MeOH) (2c), respectively.
D
Synthesis of R,R (1a)-, R,S (1b)-, S,S (1c)-, and S,R
(1d)-N6-HEB-dA. Compounds 5a-5d (15 mg) (37, 38) were
coupled with 2′-deoxyinosine (15 mg) in the presence of DMSO
(1 mL), DIPEA (50 µL), and PyBOP (30 mg) (5 days at room

122 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
SeneViratne et al.
Scheme 1. Formation of 1, 2, 3, and 4 upon the Coupling of 6-chloropurine-2′-deoxyriboside with
N-Fmoc-1-Amino-2-hydroxy-3,4-epoxybutane
Table 1. pH Dependence of the UV Absorbance Spectra of
Exocyclic DEB-dA Adducts
Forced Dimroth Rearrangement of 3a and 4a. Compounds
3a and 4a were separately dissolved in 200 µL of 0.25 M NaOH
solution and kept at room temperature for 1 h, followed by
neutralization with 1 M HCl. The reaction mixtures were then
heated to 60 °C after the addition of 300 µL of 95% EtOH and
kept for 4 more hours. About 8 µL was injected onto a capillary
HPLC column for HPLC-ESI+-MS/MS analysis.
Neutral Thermal Hydrolysis. Nucleoside adducts (0.1 mg each)
were dissolved in 100 µL of water and heated for 1 h at 85 °C.
The resulting reaction mixtures were separated with HPLC systems
1 and 3.
UVmax (nm)
compound
0.1 M HCl
water
0.1 M NaOH
2a
2b
2c
3a
3b
4a
4b
268
268
268
264
264
264
264
274
274
274
264
265
265
265
274
274
274
265
267
266
267
Synthesis of Nucleobase Adducts via Acid Hydrolysis of
the Corresponding Nucleosides. Compounds 2a, 2b, 2c, 3, and 4
were dissolved in 0.1 M HCl (250 µL) and heated for 1 h at 85 °C.
The resulting reaction mixtures were neutralized with 0.1 M NaOH
and separated by HPLC (systems 1 and 3). Compound 2a-base
(Chart 2) was isolated as a peak at 7.7 min (0.62 mg, 2.78 µmol,
94.0% yield). Compound 2b-base was isolated as a peak at 8.1
min (0.62 mg, 2.81 µmol, 95.0% yield). S,R; R,S-N6,N6-(2,3-
dihydroxybutan-1,4-diyl)-adenine (2c-base, Chart 2) was isolated
at 11.7 min (0.65 mg, 2.93 µmol, 99.0% yield). Compounds 2a-
base, 2b-base, and 2c-base had identical UV and mass spectra.
UV (water): λmax 274 nm; ESI+-MS m/z 222.2 [M + H]+, MS2 of
m/z 222.2: m/z 147.9 [Ade-CH2 + H]+, 160.6 [Ade-CHdCH2]+,
175.6 [Ade-CH2-CHdCH2+H]+, 185.6 [M + 2H-dR-2H2O]+,
203.7 [M + 2H-dR-H2O]+.
column, 4 µm particle size (Phenomenex, Torrance, CA) was eluted
with a linear gradient of methanol (B) in 0.05% acetic acid (A).
Solvent composition was changed from 0% to 1% B in 6 min, then
further to 4% B in 11 min, and finally to 30% B in 2 min. The
column was eluted at a flow rate of 12 µL/min. Oxidation products
(dialdehyde derivatives and the corresponding hemiacetals) were
detected by HPLC-ESI+-MS: dialdehyde at m/z 220.2 [M + H]+;
MS2 m/z 220.2 f m/z 202.0 [M+H -H2O]+, and hemiacetal at m/z
238.3 [M +H]+, MS2 m/z 238.3 f m/z 220.2 [M+H -H2O]+
(Supporting Information Figure S-2).
Kinetics of Cyclization of 1a and 1b in Aqueous Medium.
Pure compounds 1a or 1b were dissolved in water and incubated
at 37 °C. The reaction products were monitored by HPLC system
1 (1a, t1/2 ) 58 min; and 1b, t1/2 ) 58 min, Supporting Information
Figure S-19). In a separate experiment, 1a was dissolved in water,
and the formation of 3a and 4a was monitored by HPLC (system
3, Supporting Information Figure S-22).
Compound 3-base was isolated at 8.4 min. λmax 264 nm, ESI+-
MS: m/z 222.2 [M + H]+, MS2 of m/z 222.2: m/z 136.0 [Ade +
H]+, 148.0 [Ade-CH2 + H]+, 186.0 [M + 2H-dR-2H2O]+.
Compound 4-base was isolated at 12.1 min. λmax 264 nm, ESI+-
MS: m/z 222.2 [M + H]+, MS2 of m/z 222.2: m/z 160.6 [Ade-
CCH + H]+, 177.5 [Ade-(CH)2OH + H]+, 203.6 [M + 2H-dR-
H2O]+.
Kinetics of the Conversion of 3a and 4b in Aqueous
Medium. Purified 3a and 4a were incubated in water either at room
temperature or at 37 °C, and the rearrangement products were
detected by HPLC (Supporting Information Figure S-23).
The 1H NMR, NOESY, gCOSY, TOCSY, gHMQC, and HSQC
data for compounds 2a-base, 2b-base, and 2c-base are presented
in Supporting Information Figures S-8 and S-9.
HPLC-ESI-MS/MS Analysis of Exocyclic DEB-dA
Adducts in DEB-Treated DNA. Calf thymus DNA (500 µg
aliquots, in 500 µL of water) was treated with increasing amounts
of R,R-, S,S-, and meso-DEB (0-1 mM) at 37 °C for 24 h. DNA
was precipitated with cold ethanol and resuspended in 500 µL of
water. Following spiking with 15N4-labeled compounds 3 and 4 (500
fmol, internal standards for mass spectrometry), DNA was digested
to deoxynucleosides in the presence of Nuclease P1 (10 U) and
alkaline phosphatase (60 U, 37 °C for 1 h). Samples were filtered
through YM-30 Centricon filters to remove proteins prior to solid
phase extraction (SPE). Carbograph SPE cartridges (3 mL, Alltech)
were prepared by washing with methanol (2 × 3 mL) and water (2
× 3 mL). Samples were loaded in water (1 mL), washed with water
Oxidation of 1,2-Diols with Sodium Periodate and
Analysis of Oxidation Products by HPLC-ESI+-MS/MS. Nu-
cleobase adducts 2a-base, 2b-base, 2c-base, 3-base, and 4-base
(3 nmol aliquots) were dissolved in 15 mM aqueous sodium
periodate (30 µL). Thymidine (0.1 ou260) was added as an internal
standard for quantitation. Aliquots (6 µL) were taken following
incubation for 10 min, 1 h, 2 h, 4 h, and 23 h. The reaction was
terminated by the addition of glycerol (1 µL), and the samples were
frozen until HPLC-ESI+-MS/MS analysis with an Agilent 1100
capillary HPLC-ion trap mass spectrometer system (Agilent Tech-
nologies, Inc.; Wilmington, DE). A 250 mm × 0.5 mm Synergi

Exocyclic Deoxyadenosine Adducts of DEB
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 123
Scheme 2. Stereospecific Synthesis of Compound 2aa
a (a) 1 equiv BzNH2, p-xylene, reflux, 3 h; (b) LiAIH4, THF, 12 h; (c) H2, 10% Pd-C, 55°C, 24 h; (d) 2′-deoxyinosine, DMSO, DIPEA, PyBOP, 55°C,
3 d.
and 10% methanol (3 mL each), and eluted with 30% methanol (3
optimized for maximum response during the infusion of standard
solutions. Both exocyclic DEB-dA lesions were quantified by
isotope dilution with the corresponding 15N4-labeled internal
standards. Quantitative analyses were performed in the selected
reaction monitoring (SRM) mode using HPLC-ESI+-MS/MS peak
areas corresponding to the neutral loss of deoxyribose from
protonated molecules of the adducts (m/z 338.1 [M + H]+ f 222.0
[M + 2H-dR]+ and m/z 342.1 [M + H]+ f 226.0 [M + 2H-dR]+
for analytes and their internal standards, respectively).
mL). The 30% methanol elutions were concentrated under vacuum
and redissolved in 25 µL of water. About 8 µL was injected onto
a capillary HPLC column for HPLC-ESI+-MS/MS analysis. Exo-
cyclic DEB-dA adduct amounts were determined from the relative
response ratios (mass spectrometry peak area of exocyclic DEB-
dA adduct to mass spectrometry peak area of isotopically labeled
internal standard).
HPLC-ESI-MS/MS Analysis of Exocyclic DEB-dA
Adducts in Tissues of B6C3F1 Mice Exposed to BD by
Inhalation. B6C3F1 mice were exposed to 625 ppm butadiene or
air (controls) for 2 weeks, and liver DNA was isolated using
NuceoBond AXG500 anion exchange cartridges (Macherey-Nagel)
(40). DNA purity and amounts were determined by UV spectro-
photometry. The A260/A280 ratios were 1.7-1.8, confirming minimal
protein contamination. DNA (100 µg) was spiked with 15N4-labeled
compounds 3 and 4 (300 fmol each) and subjected to enzymatic
digestion as described above for CT DNA. Samples were filtered
through YM-10 Centricon microfilters, purified by SPE as described
above for CT DNA, dried under vacuum, and dissolved in 25 µL
of water. About 8 µL was injected onto a capillary HPLC column
for HPLC-ESI-MS/MS analysis as described above for CT DNA.
Liquid Chromatography-Mass Spectrometry. The majority
of HPLC-ESI-MS experiments were performed with an Agilent
1100 capillary HPLC-ion trap mass spectrometer (Agilent Tech-
nologies, Inc., Wilmington, DE). The instrument was operated in
the ESI+ mode. Target ion abundance value was set to 30,000, the
maximum accumulation time was 300 ms, and 4 scans were taken
per average. A typical fragmentation amplitude was 0.7 V, with a
scan width of 1.2 m/z. Nitrogen was used as a nebulizing (15 psi)
and drying gas (5 L/min, 200 °C). Electrospray ionization was
achieved at a spray voltage of 3-3.5 kV.
Results
Synthesis and Initial Characterization of Exocyclic
DEB-dA Adducts. Three isomers of exocyclic DEB-dA lesions
were initially prepared. One of them (compound 2 in Scheme
1) was isolated as a side product during the synthesis of
compound 1, which was accomplished by coupling 6-chloro-
purine deoxyriboside with 1-amino-2-hydroxy-3,4-epoxybutane
in the presence of an organic base (Scheme 1) (30). Two
additional DEB-dA exocycles (compounds 3 and 4) were
obtained as major products of spontaneous cyclization of 1 in
an aqueous solution (Scheme 1) (30). All three exocyclic DEB-
dA adducts 2, 3, and 4 had the same molecular weight (337.2
Da) as determined by ESI+ MS (m/z 338, [M + H]+), and their
molecular formula from Q-TOF MS analysis was consistent with
the addition of one molecule of DEB to deoxyadenosine
(C14H19N5O5, calculated, m/z 338.1459; observed, m/z 338.1446).
Unlike compound 1, which has a half-life of ∼100 min under
physiological conditions (30), compounds 2, 3, and 4 were
relatively stable in an aqueous solution at pH 7.2 (with the
exception of the slow interconversion of 3 and 4; see below).
Moreover, no structural changes were observed following
heating for 1 h at 85 °C (not shown). Acid hydrolysis (0.1 N
HCl, 30 min at 85 °C) of nucleosides 2, 3, and 4 cleaved the
deoxyribose, while the alkylated nucleobases retained the DEB
moiety, indicating that both alkylation events took place within
the adenine portion of the molecule.
Three stereoisomers of 2 (2a, 2b, and 2c) were synthesized
(Chart 1). Compounds 2a and 2b originated from R,R + S,S-
1-amino-2-hydroxy-3,4-epoxybutane (5a + 5c) and were re-
solved by reversed phase HPLC. Compound 2c was prepared
from meso-1-amino-2-hydroxy-3,4-epoxybutane (5b + 5d). The
absolute configurations of compounds 2a, 2b, and 2c were
subsequently determined by the independent synthesis described
below. The individual diastereomers of compounds 3 and 4
could not be resolved by conventional HPLC because of their
polar character, which led to poor retention of these nucleosides
on reversed phase HPLC columns. As described in a later
section, stereospecific synthesis was subsequently used to
prepare individual stereoisomers of compounds 3 and 4.
To identify the substitution sites within the adenine hetero-
cycle, electronic spectra of compounds 2, 3, and 4 under
For analyses of nucleoside reaction mixtures, samples were
dissolved in a 1:1 mixture of acetonitrile and 0.1% acetic acid, and
infused at a flow rate of 10-15 µL/min using a syringe pump. The
mass spectrometer was operated in full scan mode over the range
of m/z 50-350.
Enzymatic digests of DEB-treated DNA and liver DNA from
control and butadiene-exposed mice were analyzed with an Agilent
1100 capillary HPLC system (Wilmington, DE) interfaced to a
Thermo-Finnigan TSQ Quantum Ultra mass spectrometer (Thermo
Fisher Scientific Corp., Waltham, MA). A Phenomenex Synergi
(250 × 0.5 mm, 0.4 µm) column was eluted with a gradient of 10
mM ammonium formate, pH 4.2 (A), and methanol (B). The solvent
composition was kept at 100% A for 5 min, changed to 6% B over
1 min, and finally increased linearly to 30% B over 10 min. Under
these conditions, 3 and 4 coeluted at 14.5 min, while 2 eluted at
41.1 min. In order to separate 3 and 4, a capillary Thermo Hypercarb
column (150 × 0.5 mm, 5 µm, Thermo Fisher Scientific, Inc.
Waltham, MA) was eluted with 0.05% acetic acid, pH 4.0 (A),
and 100% acetonitrile (B) at the same gradient. The mass
spectrometer was operated in the positive ion mode, with nitrogen
used as a sheath gas (5 L/min). Electrospray ionization was achieved
at a spray voltage of 4.0 kV and a capillary temperature of 270 °C.
CID was achieved with Ar as a collision gas (1 mTorr) and a
collision energy of 35 V. The mass spectrometer parameters were

124 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
SeneViratne et al.
1
Figure 2. H NMR (A), HMBC (B and D), and phase sensitive HSQC spectra (C) of compound 3 (racemic) (Bruker 700 MHz) (Inset in A:
enlarged area of 3.2-4.3 ppm).
different pH conditions were examined. UV absorbance spectra
of DEB-dA adducts differed from that of compound 1 (Table 1
and Figure 1). While the UV spectra of compound 2 had an
absorbance maximum at neutrality of 274 nm (Figure 1B, Table
1), compounds 3 and 4 had UV absorption spectra that were
nearly identical to each other and were characterized by a
maximum at 264 nm (Figure 1C,D and Table 1). Under low
pH conditions, the UV absorption maximum of compound 2
underwent a 6 nm hypsochromic shift, probably a result of
protonation of the N-1 position of adenine (Table 1 and Figure
1B). This is similar to the pH dependence of the UV spectra
reported for known N6-substituted deoxyadenosines (41). No
such shift was observed for compounds 3 and 4. Instead, the
UV spectra of compounds 3 and 4 were transformed at a high
pH, with the development of a broad shoulder between 280 and
320 nm (Figures 1C and D). This UV absorption behavior is
characteristic for 1,N6-substituted deoxyadenosines (41-44).
Taken together, the UV spectroscopy data were consistent with
N6 substitution of adenine in compound 2 and 1,N6-susbstitution
in compounds 3 and 4.
Further insight into the structures of novel exocyclic DEB-
dA lesions was obtained from tandem mass spectrometry
experiments. The MS/MS fragmentation patterns of compounds
2, 3, and 4 under ESI+-MS/MS conditions were similarly
predominated by the neutral loss of deoxyribose (m/z 338.3 f
222.2 [M + 2H-dR]+), not allowing any insight into the
structures of modified bases (results not shown). However,
characteristic MS3 fragmentation patterns were observed when
the MS/MS product ions at m/z 222.2 were isolated and
subjected to further fragmentation in an ion trap mass spec-
trometer (Figure 1, right panel). Compound 2 exhibited a major
MS3 secondary product ion at m/z 185.9 [M + 2H-dR-2H2O]+
(Figure 1B). In contrast, MS3 fragmentation of compound 3 was
predominated by the fragment ions at m/z 148.0 [Ade-CH2 +

Exocyclic Deoxyadenosine Adducts of DEB
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 125
NMR signal assignment, the deoxyribose group of each nucleo-
side was cleaved in the presence of acid. NMR signal interpreta-
tion for the resulting nucleobase adducts (Chart 2) was facilitated
by comparison with known 1H NMR spectra of monoalkylated
adenines (45-48) and 1,N6-exocyclic dA adducts of BCNU,
vinyl chloride, and R,ꢀ-unsaturated aldehydes (10, 42, 49).
The aromatic region of the 1H NMR spectra recorded for the
compounds 2a-base, 2b-base, and 2c-base was characterized
by signals at 8.15 ppm (H-2 of Ade) and 8.05 ppm (H-8 of
Ade). The signal of N9-H was observed at 12.88-12.89 ppm
(Supporting Information Figures S-6A, S-8, and S-9). These
chemical shifts are comparable to those reported for other N6-
alkyladenines (47, 50). However, unlike singly substituted N6-
dA derivatives, which exhibit a one-proton signal at ∼8 ppm
corresponding to the exocyclic amino group, spectra of com-
pounds 2a-base, 2b-base, and 2c-base lacked the N6-H proton
signal, consistent with bis-alkylation of the N6 position of
adenine. 13C chemical shifts of the aromatic part of the molecule
were similar to those reported for other N6-alkyladenines
(45, 51). NMR spectra of compounds 2a-base and 2b-base were
identical, consistent with their enantiomeric relationship. The
presence of the pyrrolidine-2,3-diol group was supported by
HSQC spectra that exhibited two methylene and two methine
protons within the side chain, the TOCSY spectra that revealed
strong correlations between CH2 and CHꢀ protons, CHꢀ and
R
δ
1
CHγ protons, and CH2 and CHγ protons, H NOESY experi-
ments (Supporting Information Figure S-6), and by comparison
of the NMR spectra of compound 2 with those of other
pyrrolidine derivatives (52). No detectable correlations between
the purine heterocycle and the pyrrolidine moiety of the
molecule were detected in the gHMBC experiments (Supporting
Information Figure S-7), as is often the case for N6-substituted
adenines (20, 31, 53). On the basis of NMR, MS, and periodate
oxidation experiments, compounds 2a and 2b were identified
as the enantiomers of N6,N6-(2,3-dihydroxybutan-1,4-diyl)-2′-
deoxyadenosine (Chart 1). Similar NMR results, with a few
small differences in chemical shifts, were noted for compound
2c-base (Supporting Information Figure S-9). On the basis of
the NMR, UV, and mass spectra, as well as periodate oxidation
results (Supporting Information Figure S-2), compound 2c-base
was identified as meso-N6,N6-(2,3-dihydroxybutan-1,4-diyl)-
adenine (Chart 2).
Figure 3. Molecular models of compounds 3b (A) and 4b (B), and
proposed intramolecular hydrogen bonding in 4b (C). Selected inter-
atomic distances for protons observed in NOESY spectra are indicated.
H]+ (Figure 1C), while the MS3 spectrum of compound 4 was
characterized by intense product ions at m/z 177.5 [Ade-CH-
CH2OH + H]+ (Figure 1D). Our observation of different MS3
fragmentation patterns for the three isomers of DEB-dA adducts
is suggestive of significant structural differences between DEB-
derived deoxyadenosine exocycles 2, 3, and 4.
Periodate Oxidation Experiments to Establish the
Structure of DEB-Derived Butanediol Linker. The structure
of the DEB-derived side chain in compounds 2, 3, and 4 was
explored by periodate oxidation. To simplify the experiment,
periodate oxidation was conducted with free bases of 2a, 2b,
2c, 3, and 4 obtained by acid hydrolysis of the corresponding
nucleosides (Chart 2). These experiments (Supporting Informa-
tion Figure S-2) revealed the presence of a vicinal diol
functionality in compounds 2a, 2b, and 2c, consistent with the
1,4-butandiyl-2,3-diol structure of the DEB-derived exocycle
(Chart 2). In contrast, the lack of periodate reactivity toward
compounds 3 and 4 indicates that they do not contain a 1,2-
diol functionality (Supporting Information Figure S-2). As
described below, these results were subsequently confirmed by
an independent synthesis of 2, 3, and 4.
Final structural assignments of compounds 2a, 2b, and 2c,
including their absolute configurations, were made by comparing
their NMR, UV, optical rotation, and mass spectra to those of
authentic R,R-, S,S-, and meso-N6,N6-DHB-dA prepared inde-
pendently. This stereospecific synthesis was accomplished by
coupling 2′-deoxyinosine with optically pure 3S,4S-, 3R,4R-,
and meso-pyrrolidine-3,4-diols. For example, protected 3R,4R-
pyrrolidine-3,4-diol 8 was synthesized from D-tartaric acid 6
as described previously (Scheme 2) (37, 38). Following deben-
zylation of 8, the resulting 3R,4R pyrrolidine-3,4-diol 9 was
coupled to 2′-deoxyinosine in the presence of DMSO, DIPEA,
and PyBOP (39) to give optically pure compound 2a in 72%
yield. The same synthetic route starting with L-tartaric acid and
meso-tartaric acid was used to obtain optically pure compounds
2b and 2c, respectively. We found that the HPLC retention
times, UV, NMR, and mass spectra of the independently
prepared R,R-, S,S-, and meso-N6,N6-DHB-dA were identical
to those of compounds 2a, 2b, and 2c derived from N6-HEB-
dA (Scheme 1), confirming their chemical structures as S,S-,
R,R-, and meso-N6,N6-(2,3-dihydroxybutan-1,4-diyl)-2′-deoxy-
adenosine, respectively (Chart 1).
Structural Characterization of Compound 2 by NMR
Spectroscopy. Further details of the molecular structure of
compounds 2, 3, and 4 were obtained from one- and two-
dimensional NMR spectroscopy experiments. To simplify the

126 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
Scheme 3. Protonated and Deprotonated Forms of 3 and 4
SeneViratne et al.
Scheme 4. Stereospecific Synthesis of Compounds 3a, 3b, 4a, and 4ba.
a (a) TMS-CN, Zn-I2, reflux; (b) (i) LiAlH4/Et2O; (ii) aq. work up; (c) 1R-endo,anti-(+)-3-bromocamphor-8-sulfonic acid ammonium salt, methanol; (d)
aq., excess Ba(OH)2; (e) Fmoc-Cl, aq. dioxane, aq. Na2CO3, rt; (f) (i) mCPBA, CH2Cl2, rt, 18 h; (ii) HPLC separation of diastereomers; (g) 6-chloropurine-
2′-deoxyriboside, DIPEA, DMSO, 37 °C, 3 d; (h) 2′-deoxyinosine, PyBOP, DMSO, DIPEA, rt, 5 d; (i) H2O, rt, 15 h.
1
As noted above, the NMR spectra of compound 2-base
recorded at an ambient temperature (25 °C) reveal separate
in contrast with the fully symmetrical H and 13C spectra of
synthetic pyrrolidine-3,4-diols prior to their coupling with the
purine heterocycle (Figures S-4 and S-5 in Supporting Informa-
tion). This discrepancy can be explained by a hindered rotation
about the C6-N6 bond in compound 2, resulting in a distinct
stereoelectronic environment for carbons and hydrogens located
δ
signals for the two methylene protons (CH2R and CH2 ) and for
the two methine protons of the pyrrolidine-3,4-diol moiety (CHꢀ
and CHγ). Similarly, 13C NMR signals of carbons ꢀ and γ, as
well as those of carbons R and δ, have slightly different
chemical shifts (Supporting Information Figure S-6B). This is
R
on the opposite sides of the pyrrolidine ring (e.g., CH2 and

Exocyclic Deoxyadenosine Adducts of DEB
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 127
Scheme 5. Dimroth Rearrangement of Compound 3 to Form 4
CHꢀ versus CHγ and CH2 ). This hypothesis is fully supported
γ of the side chain and the H-2 proton of the adenine heterocycle
(Figure 2D). Furthermore, NOESY correlations are observed
between the H-2 proton of the adenine nucleobase and the CHγ
proton of the side chain of compounds 3, 3a, and 3b.
δ
by density functional calculations that reveal that the lone
electron pair of the pyrrolidine nitrogen is shared with the
π-electron system of the purine ring, resulting in a partial double
bond character for the C6-N6 bond. Conformational analyses
(Supporting Information Figures S-10-S-12) predict that the
barrier of rotation about the C6-N6 bond is between 14.5 and
17.5 kcal/mol. As a result, at room temperature, one side of the
The detailed structures of compounds 3a and 3b were
established by considering the results of two-dimensional NMR
experiments. A phase-sensitive HSQC (1H-13C) experiment
(Figure 2C) was instrumental for distinguishing between the
methylene and methine protons within the DEB-derived exo-
cycle. The diastereotopic methylene protons of the CH2δ group
were observed as two multiplets at 3.59 and 3.39 ppm (Figure
2A). According to the HMQC and HSQC data, these two
protons are attached to the same carbon atom (61.9 ppm,
Supporting Information Figure S-12). These two protons couple
with each other in the COSY and NOESY spectra, confirming
their geminal nature. Furthermore, the CH2δ group is connected
to a hydroxyl based on the results of COSY experiments
(Supporting Information Figure S-14). Taken together with the
results of periodate oxidation (Supporting Information Figure
S-2), these data establish the structure of 3 as 1,N6-(2-hydroxy-
3-hydroxymethylpropan-1,3-diyl)-2′-deoxyadenosine (Chart 1).
This structural assignment is supported by comparing the
NMR and UV data of compound 3 with those published for
other 1,N6-dA exocyclic compounds, e.g., 1,N6-(propano)-2′-
deoxyadenosine (53) and 1,N6-(2-hydroxypropano)-2′-deoxy-
adenosine (43). Density functional calculations predict that the
distance between H-2 and CHγ protons in 1,N6-(2-hydroxy-3-
hydroxymethylpropan-1,3-diyl)-2′-deoxyadenosine is 2.21 Å
(Figure 3A), consistent with the NOESY results discussed above.
Despite the single substitution of the exocyclic amino group
in compound 3, no 1H NMR signal corresponding to the
exocyclic NH proton is observed (Figure 2A). Other N1-
alkyladenines are known to be positively charged in aqueous
solutions at a neutral pH (the pKa of 1-methyladenine is 8.25)
(45), with the charge localized mainly at the N1 position. HPLC
behavior of compounds 3 and 4, e.g., their poor retention on
reversed phase HPLC columns in the absence of ion pairing
reagents, but strong retention on cation exchange columns at
pyrrolidine ring (CH2 and CHꢀ) experiences a greater ring
R
current and a more pronounced deshielding than does the other
side (CHγ and CH2 ), leading to a downfield shift in the 1H and
δ
13C NMR spectra.
To confirm the results of this computational study experi-
mentally, temperature-dependent 1H NMR spectra of compounds
2a and 2b were recorded. As the temperature was increased
from 25 °C to 50, 65, and 85 °C, proton signals corresponding
to CHꢀ and CHγ merged, yielding a single two proton signal at
4.04 ppm (Supporting Information Figure S-13). Similarly, the
signals corresponding to CH2 and CH2 coalesced at an
increased temperature, although a higher temperature (∼85 °C)
was required because of a greater separation of their chemical
shifts (4.21 and 4.00, and 3.69 and 3.62 ppm, respectively).
Taken together, our results suggest that the bond rotation along
the C6-N6 bond of compound 2 under physiological temper-
ature is hindered, resulting in an asymmetric structure.
R
δ
Structural Characterization of Compound 3: 1,N6-
(2-Hydroxy-3-hydroxymethylpropan-1,3-diyl)-2′-deoxy-
adenosine (Compound 3). Since compound 3 was obtained
by spontaneous cyclization of compound 1 (in Scheme 1), the
adenine substitution pattern of the resulting exocycle must
include the N6 position of adenine. As discussed above, the UV
spectral data for 3 are indicative of the 1,N6-dialkylation (Figure
1C). Further proof of the 1,N6-substitution of adenine in
compound 3 is obtained from gHMBC and 2-D NOESY spectra
(Figure 2B,D and Supporting Information Figures S-14-S-16
and S-21). In particular, gHMBC signals are found between
R
the C-6 of adenine and the CH2 protons of the side chain
(Figure 2B), and a correlation is observed between the carbon

128 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
SeneViratne et al.
pH 7, suggests that they are charged species under aqueous
conditions, containing a proton at the N6 and a double bond
between the C6 and the N1 (Scheme 3). Further support for
the presence of a positive charge on the N-1 under aqueous
conditions is obtained from the downfield shifts of the NMR
signals of H-2, C-2, and C-6 compared to the corresponding
signals in dA. The proton NMR signal of H-2 in 3 is shifted
downfield by 0.6 ppm when the spectrum is recorded in D2O
relative to that in dry DMSO (Figure S-25 in Supporting
Information). This shift is greater than can be accounted for by
solvent effect and is consistent with the presence of a positive
charge when compound 3a is dissolved in water. This NMR
behavior is analogous to that of structurally similar 1,N6-dA
exocyclic adducts containing a positive charge on the N-1 (43).
We hypothesize that NMR sample drying under high vacuum
conditions leads to deprotonation of the N6-H (and the removal
of acetic acid) and the formation of the N6-C6 double bond
(Scheme 3), explaining the absence of the N6-H proton signal
1
in the H NMR spectrum of compound 3a (Figure 2A).
Structural Characterization of Compound 4: 1,N6-
(1-Hydroxymethyl-2-hydroxypropan-1,3-diyl)-2′-deoxy-
adenosine (Compound 4). The UV spectroscopy results for
compound 4 discussed above (Figure 1 and Table 1) are
consistent with the 1,N6-ring substitution pattern, while periodate
oxidation experiments establish the absence of a vicinal diol
functionality (Supporting Information Figure S-2). Furthermore,
UV absorbance spectra (Figure 1) and the downfield regions of
1
the H NMR spectra of compound 4 (Supporting Information
Figures S-17 and S-18) are identical to those of compound 3,
suggesting that they have the same adenine ring substitution
patterns. However, our observation of distinct MS3 spectra and
significant differences within the upfield regions of the 1H NMR
spectra of compounds 3 and 4 (see Figure 1C and D (right panel)
and Supporting Information Figures S-17 and S-18) suggests
that they have different side chain structures.
Two-dimensional NMR experiments were used to establish
the exact connectivity between the adenine heterocycle and the
DEB-derived side chain in compound 4. In particular, NOESY
γ
correlations between the H-2 adenine proton and both CH2
protons within the propano exocycle were observed (Supporting
Information Figures S-17 and S-18). Furthermore, HMBC
correlations were detected between the two CH2γ protons (3.95
and 3.80 ppm) and the C-2 carbon of adenine (148.6 ppm) of
4a, strongly suggesting that a terminal methylene group of DEB
is connected to the N-1 position of adenine (Supporting
Information Figure S-17). On the other side, the N6 position of
adenine is bound to an internal carbon of DEB as indicated by
the presence of the HMBC correlation between the CHR proton
(3.48 ppm) and the C-6 carbon atom of the nucleobase (142.3
ppm) (Supporting Information Figures S-17 and S-18). Taken
together, these results are consistent with the structure of 1,N6-
(1-hydroxymethyl-2-hydroxypropan-1,3-diyl)-2′-deoxyadenos-
ine for compound 4. According to our density functional
calculations, the distances between the H-2 of adenine and the
two γ methylene protons in 1,N6-(1-hydroxymethyl-2-hydrox-
ypropan-1,3-diyl)-2′-deoxyadenosine are 2.23 and 3.06 Å,
respectively (Figure 3B below), consistent with the NOESY
results for compound 4 (Supporting Information Figures S-17
and S-18).
Figure 4. HPLC-UV analysis of pure standard of 3a (A), pure standard
of 4a (B), and the reaction mixture following forced Dimroth
rearrangement of 3a (C).
2-ol 11 was derived from the addition of trimethylsilyl cyanide
to acrolein to generate a protected cyanohydrin 10, which was
subjected to reduction with lithium aluminum hydride (Scheme
4) (30, 31). The racemic mixture of 1-amino-3-butene-2-ols 11
was separated by diastereomeric resolution following derivati-
zation with sodium [(1R)-(endo, anti)]-(+)-3-bromocamphor-
8-sulfonic acid ammonium salt to give crystals of 12 (Scheme
4) (32, 33). The absolute configuration of the derivatized
1-amino-3-butene-2-ol 12 (S), was established by X-ray crystal-
lography (Supporting Information Figure S-20). Following
crystal isolation, (S)-1-aminobut-3-ene-2-ol 13 was regenerated
by alkaline hydrolysis. Fmoc-protected 1-aminobut-3-ene-2-ol
14 was subjected to epoxidation in the presence of mCPBA to
give diastereomeric 1-amino-3,4-epoxybut-2-ols 5a and 5b (30),
which were resolved by preparative HPLC (Scheme 4). The
other two isomers (5c and 5d in Chart 1) were obtained
analogously using (R)-1-aminobut-3-ene-2-ol isolated from the
mother liquor following the crystallization of 12, followed by
alkaline hydrolysis.
Stereospecific Synthesis of Compounds 3a, 3b, 4a, and
4b and the Assignment of Their Absolute Configuration. In
order to establish the absolute configuration of compounds 3a,
3b, 4a, and 4b, stereospecific synthesis of each isomer was
required. In our synthetic scheme, racemic 1-aminobut-3-ene-

Exocyclic Deoxyadenosine Adducts of DEB
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 129
Scheme 6. Formation of Stereoisomers of 3 and 4 in Animals Treated with 1,3-Butadiene
Two sets of coupling conditions were tested for the coupling
of Fmoc-protected 1-amino-3,4-epoxybut-2-ols with the purine
heterocycle. Coupling of 5a with 6-chloropurine-2′-deoxyribo-
side for 3 days in the presence of DIPEA gave compound 1a
with 71% maximum yield (Scheme 4). An improved yield of
1a (95%) was achieved by coupling of 5a with 2′-deoxyinosine
in the presence of PyBOP for 5 days (39). Optically pure
compounds 3a and 4a were obtained by spontaneous cyclization
of 1a in water overnight at 25 °C, followed by reversed phase
HPLC purification (Scheme 4). Analogous reactions of 5b
produced monoepoxide 1b, which cyclized to compounds 3b
and 4b (7:3 molar ratio by HPLC). Further experiments
indicated that compounds 3a and 3b can be generated selectively
in a 98% yield when the cyclization time is decreased to 3 h.
Interestingly, the stereochemistry of the oxirane ring in 1a and
1b did not affect their cyclization kinetics (t1/2 at 37 °C in water
was ∼58 min for both isomers; see Supporting Information
Figure S-19).
functional calculations (Figure 3B and Supporting Information
Figure S-18).
Compounds 1c and 1d and their cyclization products 3c, 4c,
3d, and 4d (Chart 1) were generated analogously starting with
5c and 5d, respectively.
Solvent-Dependent Regioselectivity for the Cyclization
of Compound 1. As discussed above, compound 1 spontane-
ously cyclizes to form the corresponding 1,N6 exocyclic
compound 3 when allowed to incubate in water under physi-
ological pH (Scheme 1). In contrast, the N6,N6 exocycle 2 is
not observed when 1 is allowed to cyclize in an aqueous
solution. Instead, 2 becomes a significant side product from the
coupling reaction of 6-chloropurine-2′-deoxyriboside with
1-amino-3,4-epoxybut-2-ol when the coupling time is extended
to > 3 days (Scheme 1, Supporting Information Figure S-3).
The same product 2 is generated when purified 1 is incubated
under our standard coupling conditions (DMSO in the presence
of DIPEA) (Scheme 1).
Taken together, these results indicate that the regioselectivity
of the cyclization of compound 1 is determined by solvent
identity. Under aqueous conditions, 1 spontaneously cyclizes
to 3, while 2 is preferentially formed in an aprotic solvent
(DMSO) in the presence of organic base (Scheme 1). These
results can be explained by the altered nucleophilicity of the
N6 and the N1 of 1 under these two sets of conditions, likely a
result of N6H deprotonation in an aprotic solvent in the presence
of N,N-diisopropylethyl amine.
Forced Dimroth Rearrangement of Compound 3 to
Compound 4. While the formation of compound 3 from
compound 1 can be explained by intramolecular nucleophilic
substitution (Scheme 1), no such mechanism can be proposed
for compound 4. Kinetic analyses indicate that 3 originates
directly from 1 but that there is a delay in the formation of 4,
which requires extended incubation times (Supporting Informa-
tion Figure S-22). We hypothesized that compound 4 is formed
from 3 by a base-catalyzed Dimroth rearrangement (Scheme
5). To test this hypothesis, 3a was subjected to forced Dimroth
rearrangement in aqueous NaOH (54). Capillary HPLC analysis
of the reaction mixture revealed the presence of compounds 3a
and 4a in a 3:7 molar ratio (Figure 4C). Interestingly, the same
product ratio was observed when purified 4a was subjected to
forced Dimroth rearrangement (not shown).
An additional nucleoside product 15 was observed while
analyzing the reaction mixture from Dimroth rearrangement by
capillary HPLC. It was tentatively assigned the structure of 2-(5-
amino-1-(2-deoxy-ꢀ-D-erythro-pentofuranosyl)-imidazol-4-yl)-
4-(hydroxymethyl)-tetrahydropyrimidin-5-ol (compound 15 in
Supporting Information Figure S-24) on the basis of its UV
absorption spectra (λmax ) 284 nm) and MS/MS fragmentation
pattern. We hypothesized that this compound was produced by
The absolute configurations of stereogenic centers within the
propano group of compound 3a (ꢀ and γ in Chart 1) were
assigned as follows. The configuration at the ꢀ position of 3a
should be R because it is obtained by the cyclization of 1a,
which has the ꢀR,γR stereochemistry (Scheme 4), and the ꢀ
stereocenter is unchanged in this reaction. In contrast, the optical
configuration of the γ carbon is inverted as a result of SN2 type
nucleophilic substitution at this position, resulting in the S
stereochemistry (Scheme 4). This assignment is experimentally
confirmed by the lack of spacial correlations between the ꢀ and
the γ protons in the NOESY NMR spectrum of 3a (Supporting
Information Figure S-15). Similarly, the chiral centers within
the propane moiety of 3b were assigned as ꢀR,γR because it
was obtained by intramolecular nucleophilic substitution of
ꢀR,γS-1-amino-2-hydroxy-3,4-epoxybutane 1b, which was ex-
perimentally confirmed by the presence of NOESY NMR signals
between the ꢀ and γ protons of 3b (Supporting Information
Figure S-16). Density functional calculations predict that the
distance between ꢀ and γ protons in ꢀR,γR-1,N6-(2-hydroxy-
3-hydroxymethylpropan-1,3-diyl)-2′-deoxyadenosine is between
2.3-2.4 Å, depending on the conformation of the γ-hydroxym-
ethyl group (Figure 3A), corroborating the observed NOESY
correlations and supporting the stereochemical assignment of
3b.
The absolute stereochemistry of compound 4a was assigned
as RS,ꢀR on the basis of the mechanism of its formation via
the isomerization of 3a (see below) and the absence of the NOE
correlations between the ꢀ and R protons (Supporting Informa-
tion Figure S-17). Similarly, compound 4b was assigned the
RR,ꢀR configuration, supported by the observation of NOE
correlations between the ꢀ and R protons of 4b and density

130 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
SeneViratne et al.
a hydroxyl ion attack on compounds 3 or 4, followed by a loss
of CO as proposed by Sugiyama et al. for structurally related
compounds (55).
To determine whether Dimroth rearrangement of compounds
3 and 4 can take place under mild conditions, purified
compounds 3a and 4a were individually dissolved in water and
incubated at room temperature or at 37 °C for up to 5 days.
The rearrangement products were detected by HPLC-ESI+-MS/
MS (Supporting Information Figure S-23). We observed a slow
conversion of 3a to 4a at room temperature (12% in 5 days)
and a faster rearrangement at 37 °C (66% in 5 days). In contrast,
the formation of 3a from 4a was inefficient under both
conditions (only 20% at 37 °C in 5 days). Irrespective of the
temperature, the same molar ratio (70% of 4a and 30% of 3a)
was observed when either 3a or 4a was allowed to equilibrate
in water for a sufficient time (16 days). Kinetic analyses of this
conversion (Supporting Information Figure S-23) reveal that the
half-lives for the conversion of 3a f 4a and 4a f 3a are 71.5
and 365 h, respectively. Taken together, these results indicate
that 4a and 3a can be interconverted under physiological pH,
with the equilibrium shifted toward 4a as a more thermody-
namically stable product. Density functional calculations support
this finding as compound 4a is 1.9 kcal/mol lower in energy
than 3a. We hypothesize that the additional stability of 4a is
derived mainly from the formation of an intramolecular
hydrogen bond between between the N6-H and the hydroxym-
ethyl oxygen of the exocycle (Figure 3C).
Detection of Exocyclic DEB-dA Adducts in
DEB-Treated DNA and in Liver DNA of BD Exposed
Mice. To establish whether exocyclic DEB-dA adducts 2, 3,
and 4 are biologically relevant, aliquots of double stranded DNA
(calf thymus DNA) were separately treated with S,S-, R,R-, or
meso-DEB (Scheme 6). All three isomers of DEB were
employed in order to investigate the potential effects of
diepoxide chirality on exocyclic DEB-dA adduct formation. The
resulting alkylated DNA was enzymatically digested to deoxy-
nucleosides, spiked with 15N4-labeled compounds 3 and 4
(internal standards for mass spectrometry), and purified by solid
phase extraction. HPLC-ESI+-MS/MS analysis of samples
treated with R,R-DEB revealed two major peaks eluting at 12.4
and 17.7 min (Figure 5A), which were identified as compounds
3a and 4a, respectively, on the basis of their MS3 fragmentation
patterns and HPLC retention times. Similar results were obtained
for S,S-DEB (Figure 5B) since compounds 3c and 4c originating
from S,S-DEB have the same HPLC retention times as 3a and
4a, respectively. Samples treated with meso-DEB revealed
HPLC-ESI-MS/MS signals at 19.2 and 21.3 min, corresponding
to compounds 4b/4d and 3b/3d, respectively. In all of the cases,
the amounts of compound 4 were greater than that of compound
3. In contrast, compound 2 was not found in DNA hydrolysates,
although small amounts were detected in a separate experiment
using >200 molar excess of DEB (results not shown).
The amounts of compounds 3 and 4 present in DNA
hydrolysates were calculated from HPLC-ESI-MS/MS peak
areas corresponding to the analytes and the corresponding
internal standards. As shown in Figure 5D, concentrations of
exocyclic 1,N6-HMHP-dA adducts increased linearly as the
concentration of DEB was gradually raised from 50 to 1000
µM. Similar numbers of exocyclic DEB-dA adducts were
observed in DNA treated with S,S-, R,R-, or meso-DEB,
suggesting that the DEB isomers are equally capable of inducing
exocyclic adenine lesions in DNA. This is in contrast with our
earlier results for DNA cross-linking by DEB isomers, which
revealed that the S,S isomer formed the highest numbers of
Figure 5. Formation of exocyclic DEB-dA adducts in DNA in vitro
and in vivo: capillary HPLC-ESI+ MS/MS analysis of compounds 3
and 4 in enzymatic digests of calf thymus DNA treated with R,R DEB
(A); S,S DEB (B); and meso DEB (250 µM DEB, 24 h at 37 °C) (C);
concentration-dependent formation of exocyclic DEB-dA adducts 3 and
4 in calf thymus DNA treated with 0-1 mM DEB (24 h at 37 °C) (D);
and capillary HPLC-ESI+ MS/MS analysis of 3 and 4 in enzymatic
digests of liver DNA from a B6C3F1 mouse exposed to 625 ppm
butadiene for 2 weeks (E). No signals corresponding to 3 and 4 were
observed in DNA digests from control animals.
interstrand guanine-guanine lesions, while meso-DEB induced
equal numbers of intrastrand and interstrand bis-N7G-BD lesions
(21).
Adducts 3 and 4 were also present in genomic DNA isolated
from the liver of laboratory mice exposed to 1,3-butadiene by
inhalation (Figure 5E). Compounds 4a/4c (61%) were the most

Exocyclic Deoxyadenosine Adducts of DEB
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 131
Scheme 7. Proposed Mechanism for the Formation of 3 and 4 in DEB-Treated DNA under Physiological Conditions
Scheme 8. Pericyclic Rearrangement for the Interconversion of 4a and 3a under Physiological Conditions
abundant, followed by 4b/4d (27%) and 3a/3c (12%) (the
individual isomers of 4a/4c and 3a/3c are not resolved under
our HPLC conditions). The presence of stereoisomers of 3 and
4 in vivo is consistent with nonstereoselective metabolic
activation of 1,3-butadiene to S,S-, R,R-, and meso-DEB
(Scheme 6) (14). As previously observed for guanine-guanine
cross-links of DEB (40), exocyclic dA adducts originating from
S,S+R,R-DEB (3a, 4a, 3c, 4c) are more abundant than the
corresponding lesions of meso-DEB (3b, 4b, 3c, 4c, see Scheme
6 and Figure 5E). This may be explained by the greater amounts
of S,S- and R,R-diepoxides produced upon metabolic activation
of BD as compared with that of meso-DEB. Studies using
cDNA-expressed human P450 monooxygenases have shown that
individual P450s can generate meso and racemic DEB in
different molar ratios (56). In particular, P450 2A6 and 2E1
favor meso over racemic DEB, whereas 2C9 forms equal levels
of each (56). Alternatively, mice may detoxify meso-DEB at a
faster rate than racemic DEB. For example, epoxide hydrolase
(EH), an enzyme that detoxifies DEB by hydrolyzing the
epoxide ring to a diol, has been reported to hydrolyze meso-
DEB at a faster rate than S,S- and R,R-diepoxides (56),
potentially leading to smaller amounts of meso-DEB available
for reactions with DNA.
Taken together, our results indicate that the major exocyclic
DEB-dA adducts produced following in vitro or in vivo exposure
of DNA to DEB have the structures of 1,N6-(1-hydroxymethyl-
2-hydroxypropan-1,3-diyl)-2′-deoxyadenosine (compound 4) and
1,N6-(2-hydroxy-3-hydroxymethylpropan-1,3-diyl)-2′-deoxyad-
enosine (compound 3). Although the amounts of adducts 3 and
4 in the liver DNA following exposure to 1,3-butadiene are
significantly lower than the numbers of DEB-induced DNA-
DNA cross-links, 1,4-bis-(guan-7-yl)-2,3-butanediol (bis-N7G-

132 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
SeneViratne et al.
BD), and 1-(guan-7-yl)-4-(aden-1-yl)-2,3-butanediol (N7G-
N1A-BD) (the molar ratio of bis-N7G-BD/N7G-N1A-BD/
adducts 3 and 4 is 1:0.08:0.03), exocyclic DEB-dA lesions are
hydrolytically stable and may persist in vivo. Indeed, our recent
studies indicate that 3 and 4 are present in mouse liver at least
2 weeks following exposure (Goggin, M., and Tretyakova, N.,
manuscript in preparation). Compounds 3 and 4 are not
substrates for base excision repair or nucleotide excision repair,
leading to their potential accumulation in target tissues (Goggin,
M., and Tretyakova, N., manuscript in preparation). On the basis
of these observations, we hypothesize that compounds 3 and 4
play an important role in the mutagenicity and carcinogenicity
of 1,3-butadiene.
Concluding Remarks. Our results presented above are the
first report for the formation of potentially promutagenic
exocyclic DEB-dA lesions, which have the structures of N6,N6-
(2,3-dihydroxybutan-1,4-diyl)-2′-deoxyadenosine (compound 2),
1,N6-(1-hydroxymethyl-2-hydroxypropan-1,3-diyl)-2′-deoxyad-
enosine (compound 4), and 1,N6-(2-hydroxy-3-hydroxymeth-
ylpropan-1,3-diyl)-2′-deoxyadenosine (compound 3). The struc-
tural identities and optical configurations of the novel nucleosides
were confirmed by an independent stereospecific synthesis. We
hypothesize that in DNA under physiological conditions, DEB
alkylates the N-1 position of adenine in DNA to form N1-(2-
hydroxy-3,4-epoxybut-1-yl)-adenine adducts, which undergo
spontaneous intramolecular nucleophilic substitution to yield
compound 4 (Scheme 7A). The minor regioisomer 3 is formed
either by alkylation of the N6 of adenine bases by DEB, followed
by cyclization (Scheme 7B), or by Dimroth-like rearrangement
of 4 (Scheme 5). Compounds 3 and 4 are interconverted under
physiological conditions by a reversible reaction favoring 4 as
a more stable thermodynamic product. While Dimroth rear-
rangement is expected to be slow at physiological pH, an
alternative mechanism based on low energy unimolecular
pericyclic reactions (57) may explain the formation of 4 from
3 (Scheme 8).
Acknowledgment. We thank Kris Murphy for optimizing the
yields of several key synthetic steps, Brock Matter (University
of Minnesota Cancer Center) for his assistance with the HPLC
and mass spectrometry experiments, Victor G. Young, Jr. (X-
ray Crystallographic Laboratory, Department of Chemistry) for
conducting crystallography experiments, and Bob Carlson
(University of Minnesota Cancer Center) for preparing figures
for this manuscript. NMR instrumentation was provided with
funds from the NSF (BIR-961477), the University of Minnesota
Medical School, and the Minnesota Medical Foundation. This
research is supported by grants from the NCI (R01 CA095039)
and the Health Effects Institute (Agreement 05-12). D.Y. is
grateful for financial support provided by the National Institutes
of Health (GM62248). A.M. is grateful for support through a
Department of Energy NDSEG fellowship. Computational
resources were provided by the Minnesota Supercomputing
Institute.
Supporting Information Available: Kinetics data for cy-
clization and rearrangement experiments, NMR data tables,
molecular modeling results, and crystallography data. This
material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) Blair, I. A. (2008) DNA adducts with lipid peroxidation products.
J. Biol. Chem. 283, 15545–15549.
(2) Yang, I. Y., Hashimoto, K., de Wind, N., Blair, I. A., and Moriya, M.
(2009) Two distinct translesion synthesis pathways across a lipid
peroxidation-derived DNA adduct in mammalian cells. J. Biol. Chem.
284, 191–198.
(3) Medeiros, M. H. (2009) Exocyclic DNA adducts as biomarkers of
lipid oxidation and predictors of disease. Challenges in developing
sensitive and specific methods for clinical studies. Chem. Res. Toxicol.
22, 419–425.
(4) Nair, D. T., Johnson, R. E., Prakash, L., Prakash, S., and Aggarwal,
A. K. (2006) Hoogsteen base pair formation promotes synthesis
opposite the 1, N6-ethenodeoxyadenosine lesion by human DNA
polymerase ι. Nat. Struct. Mol. Biol. 13, 619–625.
(5) Basu, A. K., McNulty, J. M., and McGregor, W. G. (1999) Solution
conformation and mutagenic specificity of 1, N6-ethenoadenine. IARC
Sci. Publ. 325–333.
(6) Fink, S. P., Reddy, G. R., and Marnett, L. J. (1997) Mutagenicity in
Escherichia coli of the major DNA adduct derived from the endog-
enous mutagen malondialdehyde. Proc. Natl. Acad. Sci. U.S.A. 94,
8652–8657.
(7) Marnett, L. J., and Burcham, P. C. (1993) Endogenous DNA adducts:
potential and paradox. Chem. Res. Toxicol. 6, 771–785.
(8) VanderVeen, L. A., Hashim, M. F., Nechev, L. V., Harris, T. M.,
Harris, C. M., and Marnett, L. J. (2001) Evaluation of the mutagenic
potential of the principal DNA adduct of acrolein. J. Biol. Chem. 276,
9066–9070.
(9) Basu, A. K., Wood, M. L., Niedernhofer, L. J., Ramos, L. A., and
Essigmann, J. M. (1993) Mutagenic and genotoxic effects of three
vinyl chloride-induced DNA lesions: 1,N6-ethenoadenine, 3,N4-
ethenocytosine, and 4-amino-5-(imidazol-2-yl)imidazole. Biochemistry
32, 12793–12801.
Compounds 4a/4c were the most abundant exocyclic DEB-
dA adducts observed in DEB-treated DNA and also in DNA
extracted from the liver tissues of laboratory mice exposed to
1,3-butadiene by inhalation. Levels of compounds 3 and 4 in
DNA increased with DEB concentration, and little difference
in reactivity was observed for the three isomers of DEB (Figure
5D).
While the biological significance of the novel DEB-dA lesions
discovered in this work remains to be established, they may be
responsible for the ability of 1,3-butadiene and DEB to induce
a large number of A f T mutations (site specific mutagenesis
experiments are in progress). Since Watson-Crick base-pairing
of compounds 3 and 4 with dT is blocked by the (1-
hydroxymethyl-2-hydroxy-1,3-propano) exocycle, the modified
nucleoside is likely to adopt the syn conformation about the
glycosidic bond as previously observed for 1,N6-εdA and N1-
HB-dI adducts (4, 5, 10, 58), preventing accurate replication.
We hypothesize that the modified dA forms a Hoogsteen base
pair with the incoming protonated dATP, leading to A f T
transversions. The synthetic methodologies established in the
present work have been recently used to generate DNA
oligodeoxynucleotides containing site- and stereospecific adducts
2, 3, and 4 (Seneviratne, U., and Tretyakova, N., manuscript in
preparation). Our ongoing studies are employing solution NMR
and site-specific mutagenesis to establish the effects of the novel
DEB-dA lesions on DNA structure and DNA polymerase
fidelity.
(10) de los Santos, C., Kouchakdjian, M., Yarema, K., Basu, A., Essigmann,
J., and Patel, D. J. (1991) NMR studies of the exocyclic 1,
N6-ethenodeoxyadenosine adduct (εdA) opposite deoxyguanosine in
a DNA duplex. ε-dA(syn)·dG(anti) pairing at the lesion site. Biochem-
istry 30, 1828–1835.
(11) Swenberg, J. A., Boysen, G., Georgieva, N., Bird, M. G., and Lewis,
R. J. (2007) Future directions in butadiene risk assessment and the
role of cross-species internal dosimetry. Chem.-Biol. Interact. 166,
78–83.
(12) Morrow, N. L. (1990) The industrial production and use of 1,3-
butadiene. EnViron. Health Perspect. 86, 7–8.
(13) Hecht, S. S. (1999) Tobacco smoke carcinogens and lung cancer.
J. Natl. Cancer Inst. 91, 1194–1210.
(14) Krause, R. J., and Elfarra, A. A. (1997) Oxidation of butadiene
monoxide to meso- and (()-diepoxybutane by cDNA-expressed human
cytochrome P450s and by mouse, rat, and human liver microsomes:
evidence for preferential hydration of meso-diepoxybutane in rat and
human liver microsomes. Arch. Biochem. Biophys. 337, 176–184.

Exocyclic Deoxyadenosine Adducts of DEB
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 133
(15) Thornton-Manning, J. R., Dahl, A. R., Bechtold, W. E., Griffith, W. C.,
Jr., and Henderson, R. F. (1997) Comparison of the disposition of
butadiene epoxides in Sprague-Dawley rats and B6C3F1 mice fol-
lowing a single and repeated exposures to 1,3-butadiene via inhalation.
Toxicology 123, 125–134.
(16) Sasiadek, M., Norppa, H., and Sorsa, M. (1991) 1,3-Butadiene and
its epoxides induce sister-chromatid exchanges in human lymphocytes
in vitro. Mutat. Res. 261, 117–121.
(37) Wong, C. M., Buccini, J., and Te Raa, J. (1968) Synthesis of optically
active deacetyl anisomycin. Can. J. Chem. 46, 3091–3094.
(38) Arakawa, Y., and Yoshifuji, S. (2007) Synthesis of (3S,4S)-3,4-
dihydroxyprolines from L-tartaric acid. Chem. Pharm. Bull. 39, 2219–
2224.
(39) Wan, Z. K., Binnun, E., Wilson, D. P., and Lee, J. (2005) A highly
facile and efficient one-step synthesis of N6-adenosine and N6-2′-
deoxyadenosine derivatives. Org. Lett. 7, 5877–5880.
(40) Goggin, M., Loeber, R., Park, S., Walker, V., Wickliffe, J., and
Tretyakova, N. (2007) HPLC-ESI+-MS/MS analysis of N7-guanine-
N7-guanine DNA cross-links in tissues of mice exposed to 1,3-
butadiene. Chem. Res. Toxicol. 20, 839–847.
(17) Cochrane, J. E., and Skopek, T. R. (1994) Mutagenicity of butadiene
and its epoxide metabolites: I. Mutagenic potential of 1,2-epoxybutene,
1,2,3,4-diepoxybutane and 3,4-epoxy-1,2- butanediol in cultured human
lymphoblasts. Carcinogenesis 15, 713–717.
(41) Singer, B. and Grunberger, D. (1983) Molecular Biology of Mutagens
and Carcinogens, Plenum Press, New York and London.
(42) Maruenda, H., Chenna, A., Liem, L.-K., and Singer, B. (1998)
Synthesis of 1, N6-ethano-2′-deoxyadenosine, a metabolic product of
1,3-bis(2-chloroethyl)nitrosourea, and its incorporation into oligomeric
DNA. J. Org. Chem. 63, 4385–4389.
(43) Pawlowicz, A. J., Munter, T., Klika, K. D., and Kronberg, L. (2006)
Reaction of acrolein with 2′-deoxyadenosine and 9-ethyladenine-
formation of cyclic adducts. Bioorg. Chem. 34, 39–48.
(44) Singh, U. S., Decker-Samuelian, K., and Solomon, J. J. (1996) Reaction
of epichlorohydrin with 2′-deoxynucleosides: characterization of
adducts. Chem.-Biol. Interact. 99, 109–128.
(45) Fujii, T., and Itaya, T. (1999) Systematic tables of mono- and poly-
N-methylated adenines: acid dissociation constants and UV and NMR
spectral data. Heterocycles 51, 2255–2277.
(46) Macon, J. B., and Wolfenden, R. (1968) 1-Methyladenosine. Dimroth
rearrangement and reversible reduction. Biochemistry 7, 3453–3458.
(47) Tretyakova, N., Sangaiah, R., Yen, T. Y., Gold, A., and Swenberg,
J. A. (1997) Adenine adducts with diepoxybutane: isolation and
analysis in exposed calf thymus DNA. Chem. Res. Toxicol. 10, 1171–
1179.
(48) Mikhailov, S. N., Rozenski, J., Efimtseva, E. V., Busson, R., Van
Aerschot, A., and Herdewijn, P. (2002) Chemical incorporation of
1-methyladenosine into oligonucleotides. Nucleic Acids Res. 30, 1124–
1131.
(49) Byrns, M. C., Vu, C., and Peterson, L. A. (2004) The formation of
substituted 1,N6-etheno-2′-deoxyadenosine and 1,N2-etheno-2′-deoxy-
guanosine adducts by cis-2-butene-1,4-dial, a reactive metabolite of
furan. Chem. Res. Toxicol. 17, 1607–1613.
(50) Selzer, R. R., and Elfarra, A. A. (1996) Characterization of N1- and
N6-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. 9, 875–881.
(51) Chang, C. J., Ashworth, D. J., Chern, L. J., Gomes, J. D., Lee, C. G.,
and Narayan, R. (1984) 13C NMR studies of methylnucleosides. Org.
Magn. Reson. 12, 671–675.
(52) Jayaraj, K., Georgieva, N. I., Gold, A., Sangaiah, R., Koc, H., Klapper,
D. G., Ball, L. M., Reddy, A. P., and Swenberg, J. A. (2003) Synthesis
and characterization of peptides containing a cyclic Val adduct of
diepoxybutane, a possible biomarker of human exposure to butadiene.
Chem. Res. Toxicol. 16, 637–643.
(53) Smith, R. A., Williamson, D. S., Cerny, R. L., and Cohen, S. M. (1990)
Detection of 1, N6- propanodeoxyadenosine in acrolein-modified
polydeoxyadenylic acid and DNA by 32P postlabeling. Cancer Res.
50, 3005–3012.
(54) Shallop, A. J., Gaffney, B. L., and Jones, R. A. (2003) Use of 13C as
an indirect tag in 15N specifically labeled nucleosides. Syntheses of
[8-13C-1,7, NH2-15N3]adenosine, -guanosine, and their deoxy analogues.
J. Org. Chem. 68, 8657–8661.
(55) Sugiyama, T., Schweinberger, E., Kazimierczuk, Z., Ramzaeva, N.,
Rosemeyer, H., and Seela, F. (2000) 2-aza-2′-deoxyadenosine: syn-
thesis, base-pairing selectivity, and stacking properties of oligonucle-
otides. Chemistry 6, 369–378.
(56) Kemper, R. A., Krause, R. J., and Elfarra, A. A. (2001) Metabolism
of butadiene monoxide by freshly isolated hepatocytes from mice and
rats: different partitioning between oxidative, hydrolytic, and conjuga-
tion pathways. Drug Metab. Dispos. 29, 830–836.
(57) Silva, L. C., Alvarez, R., Dominguez, M., Nieto, F. O., and de Lera,
A. R. (2009) Complex thermal behavior of 11-cis-retinal, the ligand
of the visual pigments. J. Org. Chem. 74, 1007–1013.
(58) Scholdberg, T. A., Nechev, L. V., Merritt, W. K., Harris, T. M., Harris,
C. M., Lloyd, R. S., and Stone, M. P. (2005) Mispairing of a site
specific major groove (2S,3S)-N6-(2,3,4-trihydroxybutyl)-2′-deoxy-
adenosyl DNA adduct of butadiene diol epoxide with deoxyguanosine:
formation of a dA(anti)·dG(anti) pairing interaction. Chem. Res.
Toxicol. 18, 145–153.
(18) Henderson, R. F., Thornton-Manning, J. R., Bechtold, W. E., and Dahl,
A. R. (1996) Metabolism of 1,3-butadiene: species differences.
Toxicology 113, 17–22.
(19) Park, S., and Tretyakova, N. (2004) Structural characterization of the
major DNA-DNA cross-link of 1,2,3,4-diepoxybutane. Chem. Res.
Toxicol. 17, 129–136.
(20) Park, S., Hodge, J., Anderson, C., and Tretyakova, N. Y. (2004)
Guanine-adenine cross-linking by 1,2,3,4-diepoxybutane: potential
basis for biological activity. Chem. Res. Toxicol. 17, 1638–1651.
(21) Park, S., Anderson, C., Loeber, R., Seetharaman, M., Jones, R., and
Tretyakova, N. (2005) Interstrand and intrastrand DNA-DNA cross-
linking by 1,2,3,4-diepoxybutane: role of stereochemistry. J. Am.
Chem. Soc. 127, 14355–14365.
(22) Verly, W. G., Brakier, L., and Feit, P. W. (1971) Inactivation of the
T7 coliphage by the diepoxybutane stereoisomers. Biochim. Biophys.
Acta 228, 400–406.
(23) Matagne, R. (1969) Induction of chromosomal aberrations and
mutations with isomeric forms of L-threitol-1,4-bismethanesulfonate
in plant materials. Mutat. Res. 7, 241–247.
(24) Kim, M. Y., Tretyakova, N. Y., and Wogan, G. N. (2007) Mutagenesis
of the supF gene by stereoisomers of 1,2,3,4-diepoxybutane. Chem.
Res. Toxicol. 20, 790–797.
(25) Zhang, X. Y., and Elfarra, A. A. (2003) Identification and characteriza-
tion of a series of nucleoside adducts formed by the reaction of 2′-
deoxyguanosine and 1,2,3,4-diepoxybutane under physiological con-
ditions. Chem. Res. Toxicol. 16, 1606–1615.
(26) Zhang, X. Y., and Elfarra, A. A. (2004) Characterization of the reaction
products of 2′-deoxyguanosine and 1,2,3,4-diepoxybutane after acid
hydrolysis: formation of novel guanine and pyrimidine adducts. Chem.
Res. Toxicol. 17, 521–528.
(27) Zhang, X. Y., and Elfarra, A. A. (2005) Reaction of 1,2,3,4-
diepoxybutane with 2′-deoxyguanosine: initial products and their
stabilities and decomposition patterns under physiological conditions.
Chem. Res. Toxicol. 18, 1316–1323.
(28) Recio, L., Steen, A. M., Pluta, L. J., Meyer, K. G., and Saranko, C. J.
(2001) Mutational spectrum of 1,3-butadiene and metabolites 1,2-
epoxybutene and 1,2,3,4-diepoxybutane to assess mutagenic mecha-
nisms. Chem.-Biol. Interact. 135-136, 325–341.
(29) Cochrane, J. E., and Skopek, T. R. (1994) Mutagenicity of butadiene
and its epoxide metabolites: II. Mutational spectra of butadiene, 1,2-
epoxybutene and diepoxybutane at the hprt locus in splenic T cells
from exposed B6C3F1 mice. Carcinogenesis 15, 719–723.
(30) Antsypovich, S., Quirk Dorr, D., Pitts, C., and Tretyakova, N. (2007)
Site specific N6-(2-hydroxy-3,4-epoxybuty-1-yl)adenine oligodeoxy-
nucleotide adducts of 1,2,3,4-diepoxybutane: synthesis and stability
at physiological pH. Chem. Res. Toxicol. 20, 641–649.
(31) Quirk Dorr, D., Murphy, K., and Tretyakova, N. (2007) Synthesis of
DNA oligodeoxynucleotides containing structurally defined N6-(2-
hydroxy-3-buten-1-yl)-adenine adducts of 3,4-epoxy-1-butene. Chem.-
Biol. Interact. 20, 104–111.
(32) Ettlinger, M. G. (1950) Synthesis of the natural antithyroid factor l-5-
vinyl-2-thioo¨xazolidone. J. Am. Chem. Soc. 72, 4792–4796.
(33) Kjaer, A., Christensen, B. W., and Hansen, S. E. (1959) Isothiocy-
anates. XXXIV. The absolute configuration of (-)-5-Vinyl-2-oxazo-
lidinethione (goitrin) and its glucosidic progenitor (progoitrin). Acta
Chem. Scand. 13, 144–150.
(34) Giese, T. J., Gregersen, B. A., Liu, Y., Nam, K., Mayaan, E., Moser,
A., Range, K., Faza, O. N., Lopez, C. S., de Lera, A. R., Schaftenaar,
G., Lopez, X., Lee, T.-S., Karypis, G., and York, D. M. (2006)
QCRNA 1: A database of quantum calculations for RNA catalysis. J.
Mol. Graphics Modell. 25, 423–433.
(35) Becke, A. D. (1988) Density-functional exchange-energy approxima-
tion with correct asymptotic behavior. Phys. ReV. A: At., Mol., Opt.
Phys. 38, 3098–3100.
(36) Lee, C., Yang, W., and Parr, R. G. (1988) Development of the Colle-
Salvetti correlation-energy formula into a functional of the electron
density. Phys. ReV. B: Condens. Matter 37, 785–789.
TX900312E