Studies of the mechanisms of adduction of 2'-deoxyadenosine with styrene oxide and polycyclic aromatic hydrocarbon dihydrodiol epoxides

Article abstract of DOI:10.1021/tx000054m

The mechanism of adduction of 2'-deoxyadenosine by styrene oxide and polycyclic aromatic hydrocarbon dihydrodiol epoxides has been explored using ¹⁵N⁶-labeled adenine nucleosides. The extent of reaction at N1 versus N⁶ was

Full text of DOI:10.1021/tx000054m

Chem. Res. Toxicol. 2000, 13, 625-637
625
Stu d ies of th e Mech a n ism s of Ad d u ction of
2′-Deoxya d en osin e w ith Styr en e Oxid e a n d P olycyclic
Ar om a tic Hyd r oca r bon Dih yd r od iol Ep oxid es
Hye-Young H. Kim, J ari I. Finneman,† Constance M. Harris,* and
Thomas M. Harris*
Chemistry Department and Center in Molecular Toxicology, Vanderbilt University,
Nashville, Tennessee 37235
Received March 6, 2000
The mechanism of adduction of 2′-deoxyadenosine by styrene oxide and polycyclic aromatic
hydrocarbon dihydrodiol epoxides has been explored using 15N6-labeled adenine nucleosides.
1
The extent of reaction at N1 versus N6 was evaluated by H NMR of the N6 adducts after
allowing Dimroth rearrangement to occur. Products arising from attack at N1 followed by
1
Dimroth rearrangement exhibited a small two-bond H-15N coupling constant (N1-H2 J ∼
1
13 Hz); products from direct attack exhibited a much larger one-bond H-15N coupling constant
(J ∼ 90 Hz). In the case of styrene oxide, all of the N6 â adduct arose by initial attack at N1,
whereas the majority (70-80%) of the N6 R adducts came from direct attack. The styrene oxide
reaction was also studied with a self-complementary oligodeoxynucleotide (24-mer) containing
nine 15N6-labeled adenine residues. NMR examination of the N6 R- and â-styrene oxide adducts
isolated after enzymatic degradation of the 24-mer gave very similar results, indicating that
N1 attack can occur readily even with a duplexed oligonucleotide. With the PAH dihydrodiol
epoxides, only naphthalene dihydrodiol epoxide exhibited significant initial reaction at N1 (50%).
No detectable rearranged product was seen in reactions with benzo[a]pyrene dihydrodiol epoxide
or non-bay or bay region benz[a]anthracene dihydrodiol epoxide; interestingly, a small amount
of N1 attack (5-7%) was seen in the case of benzo[c]phenanthrene dihydrodiol epoxide. It
appears that initial attack at N1 is only a significant reaction pathway for epoxides attached
to a single aromatic ring.
Dimroth rearrangement is a base-catalyzed cleavage of
the six-membered ring followed by a recyclization that
interchanges the locations of the endocyclic and exocyclic
nitrogens (Scheme 1). With the dihydrodiol epoxides of
polycyclic aromatic hydrocarbons and other highly con-
jugated epoxides, N1 adducts have not been observed.
Although it has generally been presumed that this is due
to reaction occurring directly at N6, the possibility that
a very rapid Dimroth rearrangement with bulky N1
derivatives makes the N1 adducts difficult to detect
cannot be excluded.
Both the simple epoxides and those of the PAHs cause
mutations at adenine sites. For example, butadiene
monoepoxide produces a significant fraction of mutations
at A‚T base pairs but so do the bay region dihydrodiol
epoxides of PAHs.1 The mutagenicity of N6 adducts of the
PAHs, styrene oxide, and to a lesser extent the simpler
epoxides has been studied using synthetic oligonucle-
otides containing adducts of unambiguous structure.
However, these studies do not preclude the possibility
that mutations observed in vivo may be arising from N1
lesions rather than from N6.
In tr od u ction
The genotoxicity of xenobiotic compounds involves
reactions of the compounds or more frequently their
metabolic products with DNA. The resulting adducts or
lesions can degrade replication fidelity, leading to muta-
tions. In most of the adduction reactions, the xenobiotic
species acts as an electrophile. There are well-docu-
mented examples of attack on many different sites in the
DNA, but the identity of the adduct giving rise to any
specific mutation is frequently not known. As a conse-
quence, the reactivity of electrophiles with nucleic acid
constituents and the biochemical impact of the resulting
adducts continue to be the subject of intense study with
the hope of better understanding the structural basis of
mutagenesis and carcinogenesis.
A particularly interesting situation exists with ad-
enine. The reactivity of its nucleosides with alkylating
agents, including both the alkyl and aralkyl halides and
the related epoxides, has been studied extensively, and
adducts have been observed at N1, N3, N6, and N7 with
the distribution of products depending upon the structure
of the electrophile and the reaction conditions. With
simple epoxides such as propylene oxide (1), butadiene
monoepoxide (2), and glycidyl ethers (3), both N1 and N6
products have been observed. In these cases, the adduc-
tion reaction may always involve initial attack at N1 with
the N6 adduct arising by a Dimroth rearrangement. The
1
Abbreviations: HRMS-FAB+, high-resolution mass spectrometry-
fast atom bombardment positive mode; EI-MS, electron impact mass
spectrometry; COSY, correlation spectroscopy; HMQC, heteronuclear
correlation through multiple-bond quantum coherence; INEPT, insen-
sitive nuclei enhanced by polarization transfer; PAH, polycyclic
aromatic hydrocarbon; SO, styrene oxide; NA, naphthalene; nBA, non-
bay region benz[a]anthracene; BA, bay region benz[a]anthracene; BP,
benzo[a]pyrene; BC, benzo[c]phenanthrene; DE, dihydrodiol epoxide.
† Currently at Pfizer, Inc., Groton, CT.
10.1021/tx000054m CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/20/2000

626 Chem. Res. Toxicol., Vol. 13, No. 7, 2000
Kim et al.
NCI Chemical Carcinogen Repository (Midwest Research In-
stitute). (R)-Styrene oxide was purchased from Aldrich. 15N-
labeled ammonium chloride (>98 at. %) was purchased from
Aldrich and 15NH4OH (3.3 M, >98 at. %) from Cambridge
Isotopes. 1H, one-dimensional (1D) COSY, two-dimensional (2D)
COSY, and HMQC NMR spectra were obtained at either 400
or 500 MHz. The spectra were recorded in DMSO-d6 or mixtures
of DMSO-d6 with D2O. The 1H spectra were internally refer-
enced to the residual 1H signal in DMSO-d6 (δ ) 2.49 ppm);
the 15N spectra were referenced externally to 2.9 M 15NH4Cl in
1% HCl /DMSO (δ ) 0 ppm). Mass spectra were obtained in a
positive ion FAB mode (glycerol/DMSO/TFA). Reactions were
monitored by TLC on silica gel plates (EM Science, Kieselgel
60, F254). TLC plates were visualized by UV and/or anisalde-
hyde stain. Column chromatography was conducted on silica
gel 60 (70-230 mesh) from EM Science. HPLC analyses and
isolations of nucleosides were carried out on a gradient HPLC
system (Beckman System Gold) equipped with a diode array
detector. Analytical runs were performed on a 4.6 mm × 250
Sch em e 1. Dim r oth Rea r r a n gem en t of N1 Alk yl
Ad en in e Nu cleosid es
The situation with regard to styrene oxide adducts is
particularly interesting. Styrene oxide appears to have
properties intermediate between the extremes repre-
sented by propylene oxide and PAH dihydrodiol epoxides
(4-6). Qian and Dipple (6) found that styrene oxide reacts
with adenosine at N1 and N6 to give both R- and
â-products. They proposed that the N6 â-product arises
entirely by Dimroth rearrangement, whereas the N6
R-product is formed through a combination of Dimroth
rearrangement from N1 and direct attack at N6 (Scheme
2). The N6 R adducts of styrene oxide have been reported
to show negligible mutagenicity in a bacterial system (7),
but styrene oxide has been reported to be mutagenic in
human t-lymphocytes with a significant fraction of these
mutations involving A‚T base pairs (8). These results
raise the question of whether transient N1 adducts may
be the source of some or all of the AT mutations observed
in vivo.
mm YMC-ODS-AQ column at
a flow rate of 1.0 mL/min;
preparative-scale isolations were carried out on a 10 mm × 250
mm YMC-ODS-AQ column at a flow rate of 3.0 mL/min.
Syn th esis of 15N6-La beled Nu cleosid es. (1) [15N6]-2′-
Deoxya d en osin e. 6-Chloropurinyl 2′-deoxyriboside (0.30 g, 1.1
mmol) was dissolved in 100 µL of DMSO and 1 mL of dioxane
in a heavy-walled reaction vial. 15NH4OH, generated by the
reaction of 15NH4Cl (240 mg, 4.4 mmol) and NaOH (4.4 M, 1
mL), was added. The vial was sealed, and the reaction mixture
was heated at 50 °C for 62 h, evaporated to dryness, and
redissolved in methanol. Precipitated NaCl was filtered off, and
the filtrate was passed through an anion exchange column (Bio-
Rad; AG1-8X, Cl- form, 20 mm × 140 mm; eluted with 4:6
MeOH/H2O). The resulting solution was evaporated to dryness
and then redissolved in MeOH for purification by silica gel
column chromatography (8:2 CHCl3/EtOH) to give [15N6]-2′-
deoxyadenosine in 60-80% yield: 1H NMR (400 MHz, DMSO-
d6) δ 8.32 (s, 1H, H8), 8.11 (s, 1H, H2), 7.29 (d, 2H, 15NH2, 1J NH
) 90.1 Hz, can be exchanged in D2O), 6.33 (dd, 1H, H1′, J 1 ) J 2
) 7.7 Hz), 5.30 (d, 1H, OH, J ) 3.6 Hz, can be exchanged in
D2O), 5.22 (t, 1H, OH, J ) 1.3 Hz, can be exchanged in D2O),
4.40 (br, 1H, H3′), 3.87 (br, 1H, H4′), 3.59-3.50 (m, 2H, H5′,
H5′′), 2.76 (m, 1H, H2′′), 2.22 (m, 2H, H2′); 15N NMR (DMSO-
d6) δ 60.03 (J ) 90.1 Hz); HRMS-FAB+ m/z 253.1062, calcd for
C10H1414N415N1O3 (MH+) 253.0989.
We wanted to probe further the formation and reactiv-
ity of adducts of styrene oxide and PAH dihydrodiol
epoxides on the N1 position of deoxyadenosine. The
relatively rapid rearrangement of the N1 styrene oxide
adducts, and possibly even faster rearrangement of as
yet unobserved N1 adducts of PAH dihydrodiol epoxides,
complicate the investigation of this issue. It occurred to
us that by using 15N6-labeled deoxyadenosine in such
alkylation reactions we would be able to detect rear-
rangement even if we were unable to isolate the N1
adducts. The inherently low sensitivity of 15N NMR
spectroscopy even with highly enriched samples is an
experimental complication for such a study in that direct
observation of 15N would require large samples and/or
long acquisition times. However, greatly enhanced sen-
(2) [15N6]Aden osin e. 6-Chloropurinyl riboside (Sigma Chemi-
cal Co., 100 mg, 0.35 mmol) was dissolved in 3.3 M 15NH4OH
(1.4 mL, Cambridge Isotopes). The mixture was heated (50 °C)
with stirring for 24 h in a sealed vial. The reaction mixture was
evaporated to dryness and redissolved in water. Silica gel (1.4
g) was added, and the mixture was lyophilized. Purification of
the labeled nucleoside was accomplished by column chroma-
tography on silica gel (7:3 CHCl3/EtOH) to give the pure
nucleoside in 40% yield (37.5 mg, 0.18 mmol): Rf ) 0.3 (7:3
CHCl3/EtOH); 1H NMR (400 MHz, DMSO-d6) δ 8.34 (s, 1H,
1
1
sitivity can be obtained by using H NMR to detect H-
15N one- and two-bond coupling so that microgram
samples of 15N-enriched nucleosides would be adequate
for assigning the location of 15N labels.
In this paper, we report the results of a study of the
reactions of 15N6-labeled deoxyadenosine with a series of
aralkyl epoxides, ranging from styrene oxide to the
dihydrodiol epoxides of the highly carcinogenic PAHs
benzo[c]phenanthrene and benzo[a]pyrene. The goal of
these experiments was to determine the extent to which
the N6 deoxyadenosine derivatives of these compounds
arise by initial attack at N1. In addition, the reaction of
styrene oxide with a 24-mer oligonucleotide containing
15N6-labeled adenine was examined to determine the
effect of duplex structure on the relative reactivity of the
N1 and N6 sites.
1
adenine H8), 8.12 (s, 1H, adenine H2), 7.35 (d, 1H, 15NH2, J NH
) 90.2 Hz, can be exchanged in D2O), 5.86 (d, 1H, H1′, J ) 6.2
Hz), 5.42 (m, 2H, OH, can be exchanged in D2O), 5.16 (br, 1H,
OH, can be exchanged in D2O), 4.59 (m, 1H, H2′), 4.13 (br, 1H,
H3′), 3.95 (dd, 1H, H4′, J ) 6.6 Hz), 3.65-3.52 (m, 2H, H5′,
H5′′); 15N NMR (INEPT, DMSO-d6) δ 60.48; HRMS-FAB+ m/z
269.1016, calcd for C10H1414N415N1O4 (MH+) 269.1018.
R ea ct ion of (R)-St yr en e Oxid e w it h Un la b eled 2′-
Deoxya d en osin e. (R)-Styrene oxide (40 µL, 0.17 mmol) and
DMSO (40 µL) were added to a solution of deoxyadenosine (20
mg, 0.08 mmol) dissolved in 50 mM Tris-HCl (2 mL, pH 7.2).
The mixture was sonicated for 30 min, and then stirred for 72
h at 42 °C. The sonication was repeated every 24 h. The mixture
was loaded onto C-18 Sep-Pak cartridges which were eluted with
water (40 mL) and finally with 100% MeOH (2 × 2 mL) to elute
the more lipophilic styrene oxide adducts. The resulting MeOH
fractions were evaporated to dryness and redissolved in MeOH/
H2O (50:50) for HPLC purification. UV absorbance was moni-
Exp er im en ta l P r oced u r es
Ca u tion : The dihydrodiol epoxides of polycyclic aromatic
hydrocarbons are potent mutagens and/ or carcinogens and
should be handled with care, as outlined in National Cancer
Institute guidelines.
Gen er a l. (()-3â,4R-Dihydroxy-1â,2â-epoxy-1,2,3,4-tetrahy-
drobenzo[c]phenanthrene and (()-3â,4R-dihydroxy-1â,2â-epoxy-
1,2,3,4-tetrahydrobenz[a]anthracene were purchased from the

1H NMR Study of [15N6]dAdo Alkylation by Epoxides
Chem. Res. Toxicol., Vol. 13, No. 7, 2000 627
Sch em e 2. Rea ction of Styr en e Oxid e w ith Ad en osin e
tored at 260 and 280 nm. Peaks from multiple chromatographic
runs were pooled and concentrated to dryness (combined yield
of ∼25%). The purity of samples was established by HPLC under
the same conditions.
labeled products, using 2′-deoxyadenosine labeled with 15N at
the N6 position (yield of 28%).
N6r(S)SOd 15Ad o. The isolated N6R(S)SOd15Ado was a mix-
ture of rearranged and un-rearranged products: 1H NMR (400
MHz, DMSO-d6) δ 8.36 (s, 1H, adenine H8), 8.13 (s, 1H, adenine
N6-[2-Hydr oxy-1-(S)-ph en yleth yl]-2′-deoxyaden osin e [N6r-
(S)SOd Ad o]: 1H NMR (400 MHz, DMSO-d6) δ 8.35 (s, 1H,
adenine H8), 8.13 (s, 1H, adenine H2), 8.01 (d, 1H, NH, J ) 8.6
Hz, can be exchanged in D2O), 7.41 (d, 2H, Ar-H, ortho, J ) 7.3
Hz), 7.27 (dd, 2H, Ar-H, meta, J 1 ) 7.3 Hz, J 2 ) 7.7 Hz), 7.19
(t, 1H, Ar-H, para, J ) 7.7 Hz), 6.33 (dd, 1H, H1′, J 1 ) J 2 ) 6.9
Hz), 5.37 (br, 1H, R-CH), 5.27 (d, 1H, 3′-OH, J ) 4.0 Hz, can be
exchanged in D2O), 5.16 (t, 1H, 5′-OH, J ) 5.1 Hz, can be
exchanged in D2O), 4.92 (t, 1H, â-OH, J ) 5.8 Hz, can be
exchanged in D2O), 4.39 (br, 1H, H3′), 3.86-3.85 (m, 1H, H4′),
3.74-3.70 (m, 2H, â-CH2), 3.58-3.49 (m, 2H, H5′, H5′′), 2.75-
2.65 (m, 2H, H2′′), 2.30-2.20 (m, 1H, H2′); HRMS-FAB+ m/z
372.1676, calcd for C18H22N5O4 (MH+) 372.1672.
N6-[2-H yd r oxy-1-(R)-p h en ylet h yl]-2′-d eoxya d en osin e
[N6r(R)SOd Ad o]: 1H NMR (400 MHz, DMSO-d6) δ 8.35 (s, 1H,
adenine H8), 8.13 (s, 1H, adenine H2), 8.01 (d, 1H, NH, J ) 8.6
Hz, can be exchanged in D2O), 7.41 (d, 2H, Ar-H, ortho, J ) 7.3
Hz), 7.27 (dd, 2H, Ar-H, meta, J 1 ) 7.3 Hz, J 2 ) 7.6 Hz), 7.19
(t, 1H, Ar-H, para, J ) 7.6 Hz), 6.33 (dd, 1H, H1′, J 1 ) J 2 ) 6.9
Hz), 5.37 (br, 1H, R-CH), 5.27 (d, 1H, 3′-OH, J ) 4.0 Hz, can be
exchanged in D2O), 5.16 (t, 1H, 5′-OH, J ) 5.1 Hz, can be
exchanged in D2O), 4.92 (t, 1H, â-OH, J ) 5.8 Hz, can be
exchanged in D2O), 4.39 (br, 1H, H3′), 3.86-3.85 (m, 1H, H4′),
3.74-3.70 (m, 2H, â-CH2), 3.58-3.49 (m, 2H, H5′, H5′′), 2.75-
2.65 (m, 2H, H2′′), 2.30-2.20 (m, 1H, H2′); HRMS-FAB+ m/z
372.1675, calcd for C18H22O4N5 (MH+) 372.1672.
N6-[1-H yd r oxy-2-(R)-p h en ylet h yl]-2′-d eoxya d en osin e
[N6â(R)SOd Ad o]: 1H NMR (400 MHz, DMSO-d6) δ 8.33 (br,
1H, adenine H8), 8.21 (br, 1H, adenine H2), 7.57 (br, 1H, NH,
can be exchanged in D2O), 7.36 (br, 2H, Ar-H, ortho), 7.31 (t,
2H, Ar-H, meta, J ) 7.4 Hz), 7.23 (t, 1H, Ar-H, para, J ) 5.2
Hz), 6.34 (t, 1H, H1′, J ) 7.6 Hz), 5.57 (br, 1H, R-OH, can be
exchanged in D2O), 5.29 (br, 1H, 3′-OH, can be exchanged in
D2O), 5.19 (t, 1H, 5′-OH, J ) 5.8 Hz, can be exchanged in D2O),
4.87 (m, 1H, R-CH), 4.39 (m, 1H, H3′), 3.87-3.86 (m, 1H, H4′),
3.80-3.70 (m, 1H, â-CH2), 3.63-3.50 (m, 3H, â-CH2, H5′, H5′′),
2.71-2.65 (m, 1H, H2′′), 2.32-2.22 (m, 1H, H2′); HRMS-FAB+
m/z 372.1676, calcd for C18H22N5O4 (MH+) 372.1672.
The structure and stereochemistry of the styrene oxide
adducts were confirmed by independent syntheses from 6-chlo-
ropurine 2′-deoxyriboside and (R)- and (S)-2-amino-2-phenyl-
ethanol (9) for the R adducts and (R)- and (S)-2-amino-1-
phenylethanol (10) for the â adducts and also by comparison
with literature values (6, 11).
1
H2), 8.01 (dd, 1H, 15NH, J NH ) 92.0 Hz, J HH ) 8.5 Hz, can be
exchanged in D2O), 7.41 (d, 2H, Ar-H, ortho, J ) 7.5 Hz), 7.27
(dd, 2H, Ar-H, meta, J ) 7.5 and 7.3 Hz), 7.18 (t, 1H, Ar-H,
para, J ) 7.3 Hz), 6.33 (dd, 1H, H1′, J 1 ) 7.4 Hz, J 2 ) 6.5 Hz),
5.38 (br, 1H, R-CH), 5.28 (d, 1H, OH, J ) 4.0 Hz, can be
exchanged in D2O), 5.16 (t, 1H, OH, J 1 ) J 2 ) 5.7 Hz, can be
exchanged in D2O), 4.93 (t, 1H, OH, J ) 5.6 Hz, can be
exchanged in D2O), 4.39 (br, 1H, OH, J 1 ) J 2 ) 5.7 Hz, can be
exchanged in D2O), 3.86-3.85 (m, 1H, H4′), 3.74-3.70 (m, 2H,
â-CH2), 3.58-3.49 (m, 2H, H5′, H5′′), 2.75-2.65 (m, 1H, H2′′),
2.30-2.20 (m, 1H, H2′). The NMR spectra established that the
material was 20% rearranged and 80% un-rearranged: HRMS-
FAB+ m/z 373.1646, calcd for C18H2214N415N1O4 (MH+) 373.1644.
1
N6r(R)SOd 15Ad o: H NMR (400 MHz, DMSO-d6) δ 8.36 (s,
1H, adenine H8), 8.13 (s, 1H, adenine H2), 8.01 (dd, 1H, 15NH,
1J NH ) 92.0 Hz, J HH ) 8.5 Hz, can be exchanged in D2O), 7.41
(dd, 2H, Ar-H, ortho, J ) 7.5 Hz), 7.27 (dd, 2H, Ar-H, meta, J 1
) 7.5 Hz, J 2 ) 7.3 Hz), 7.18 (t, 1H, Ar-H, para, J ) 7.3 Hz),
6.33 (dd, 1H, H1′, J 1 ) 7.4 Hz, J 2 ) 6.5 Hz), 5.38 (br, 1H, R-CH),
5.28 (d, 1H, OH, J ) 4.0 Hz, can be exchanged in D2O), 5.16 (t,
1H, OH, J 1 ) J 2 ) 5.7 Hz, can be exchanged in D2O), 4.93 (t,
1H, OH, J ) 5.6 Hz, can be exchanged in D2O), 4.39 (br, 1H,
OH, can be exchanged in D2O), 3.86-3.85 (m, 1H, H4′), 3.74-
3.70 (m, 2H, â-CH2), 3.58-3.49 (m, 2H, H5′, H5′′), 2.75-2.65
(m, 1H, H2′′), 2.30-2.20 (m, 1H, H2′). The NMR spectra
established that the material was essentially un-rearranged:
HRMS-FAB+ m/z 373.1646, calcd for C18H2214N415N1O4 373.1644
(MH+).
N6âSOd 15Ad o: 1H NMR (400 MHz, DMSO-d6) δ 8.35 (br, 1H,
2
adenine H8), 8.24 (d, 1H, adenine H2, J NH ) 16.4 Hz), 7.55
(br, 1H, NH, can be exchanged in D2O), 7.36 (br, 2H, Ar-H,
ortho), 7.31 (t, 2H, Ar-H, meta, J ) 7.5 Hz), 7.23 (t, 1H, Ar-H,
para, J ) 7.2 Hz), 6.34 (dd, 1H, H1′, J 1 ) J 2 ) 6.3 Hz), 5.57 (br,
1H, R-OH, can be exchanged in D2O), 5.29 (br, 1H, OH, can be
exchanged in D2O), 5.19 (t, 1H, OH, J 1 ) 5.2 Hz, J 2 ) 5.8 Hz,
can be exchanged in D2O), 4.87 (m, 1H, R-CH), 4.39 (m, 1H,
H3′), 3.87-3.86 (m, 1H, H4′), 3.8-3.7 (m, 1H, â-CH2), 3.63-
3.50 (m, 3H, â-CH2, H5′, H5′′), 2.71-2.65 (m, 1H, H2′′), 2.32-
2.22 (m, 1H, H2′). The NMR spectra established that the
material was completely rearranged from N1 to N6: HRMS-
FAB+ m/z 373.1646, calcd for C18H2214N415N1O4 (MH+) 373.1644.
R ea ct ion of (R)-St yr en e Oxid e w it h [15N6]Ad en osin e.
[
15N6]Adenosine (20 mg, 0.07 mmol) was stirred in 2 mL of H2O
Rea ction of (R)-Styr en e Oxid e w ith [15N6]-2′-Deoxya d -
en osin e. The reaction conditions described above for the
nonlabeled styrene oxide adducts were used to prepare the
at 42 °C to form a homogeneous solution. (R)-Styrene oxide (40
µL, 0.17 mmol) was added. The reaction mixture was stirred at
42 °C for 72 h. The reaction mixture was extracted with EtOAc

628 Chem. Res. Toxicol., Vol. 13, No. 7, 2000
Kim et al.
(3 × 1 mL), followed by Et2O (2 × 1 mL), to remove epoxide
hydrolysis products. The aqueous solution was then loaded onto
C-18 cartridges (Alltech), and most of the unreacted nucleoside
was removed by washing the cartridges with water (20 mL).
The more lipophilic styrene oxide-nucleoside adducts were
subsequently eluted with 100% MeOH (2 × 2 mL). The resulting
MeOH fraction was evaporated to dryness and redissolved in
MeOH/H2O (50:50, 1 mL). Aliquots (50 µL) were injected onto
a 10 mm × 250 mm C-18 column (YMC-ODS-AQ) for purifica-
tion. The peaks of interest were collected over a 20 min linear
gradient from 40 to 99% MeOH in H2O with retention times of
11.7 min (R adduct) and 12.9 min (â adduct) at a flow rate of
3.0 mL/min (combined yield of ∼2%). UV absorbance was
monitored at 258 and 278 nm.
N6r(S)SO15Ad o: 1H NMR (400 MHz, DMSO-d6) δ 8.38 (s,
1H, adenine H8), 8.14 (s, 1H, adenine H2), 8.14 (d, adenine H2,
2J NH ) 16 Hz), 8.08 (dd, 1H, 15NH, 1J NH ) 92 Hz, J HH ) 8.2 Hz,
can be exchanged in D2O), 7.41 (d, 2H, Ar-H, ortho, J ) 7.5 Hz),
7.28 (dd, 2H, Ar-H, meta, J 1 ) 7.3 Hz, J 2 ) 7.5 Hz), 7.19 (t, 1H,
Ar-H, para, J ) 7.3 Hz), 5.86 (d, 1H, H1′, J ) 6.2 Hz), 5.42 (d,
1H, OH, J ) 6.0 Hz, can be exchanged in D2O), 5.36 (m, 1H,
R-CH), 5.16 (d, 1H, OH, J ) 4.6 Hz, can be exchanged in D2O),
4.94 (t, 1H, OH, J ) 5.7 Hz, can be exchanged in D2O), 4.61-
4.57 (m, 1H, H2′), 4.14-4.11 (m, 1H, H3′), 3.94 (dd, 1H, H4′, J 1
) 6.5 Hz, J 2 ) 3.2 Hz), 3.78-3.50 (m, 4H, â-CH2, H5′, H5′′).
The NMR spectra established that the material was 30%
rearranged and 70% un-rearranged: HRMS-FAB+ m/z 389.1588,
calcd for C18H2214N415N1O5 (MH+) 389.1591. No N6R(R) was
detected.
by 1H NMR as before.
Rea ction s of P AH Dih yd r od iol Ep oxid es w ith [15N6]-2′-
Deoxya d en osin e. (1) (+)-3â,4r-Dih yd r oxy-1â,2â-ep oxy-
1,2,3,4-tetr a h yd r on a p h th a len e (NADE). Racemic anti-NADE
was synthesized from 1,2-dihydronaphthalene (16, 17). [15N6]-
2′-Deoxyadenosine (40 mg, 0.16 mmol) was dissolved in 4 mL
of 50 mM Tris-HCl (pH 7.2) to give a homogeneous solution at
40 °C. Racemic NADE (20 mg/50 µL of DMSO, 0.12 mmol) was
added. The mixture was stirred for 5 days at 40 °C. The reaction
mixture was extracted with water-saturated EtOAc (3 × 2 mL),
followed by ethyl ether (2 × 2 mL) to remove diol epoxide
hydrolysis products. The aqueous solution was lyophilized to
dryness and redissolved in 1 mL of MeOH/H2O (50:50) for HPLC
purification. Aliquots (50 µL) were injected onto a 10 mm × 250
mm C-18 column (Kromasil, Higgins Analytical, Inc.). The (1R)-
and (1S)-anti-trans diastereomers (retention times of 9.0 and
10.7 min, respectively) were collected (combined yield of ∼2%)
using a 35 min linear gradient from 3 to 80% MeOH in H2O.
UV absorbance was monitored at 258 and 268 nm.
(1r,2â,3â,4r)-2′-Deoxy-[15N6](1,2,3,4-tetr a h yd r o-2,3,4-tr i-
h yd r oxyn a p h t h a len -1-yl)a d en osin e [(+)- a n d (-)-a n ti-
tr a n s]. The 1H NMR spectra of the diastereomers were indis-
tinguishable: 1H NMR (500 MHz, DMSO-d6) δ 8.34 (s, 1H, H8),
8.23 (s, br, 1H, adenine H2), 7.72 (d, 1H, J HH ) 9.0 Hz, 14NH,
3
1
3
dd, 1H, J NH ) 93 Hz, J HH ) 9.0 Hz, 15NH, can be exchanged
in D2O), 7.34 (d, 1H, H8, J ) 7.5 Hz), 7.21 (dd, 1H, H7, J 1 ) 8.0
Hz, J 2 ) 7.5 Hz), 7.14 (dd, 1H, H6, J 1 ) 6.5 Hz, J 2 ) 8.0 Hz),
7.09 (br, 1H, H5), 6.36 (t, 1H, H1′, J ) 7.1 Hz), 5.64 (br, 1H,
H4), 4.50 (t, 1H, H1), 4.41 (br, 1H, H3′), 4.17 (br, 1H, H3), 3.96
(br, 1H, H2), 3.88 (br, 1H, H4′), 3.62-3.57 (m, 2H, H5′, H5′′),
2.75 (m, 1H, H2′′), 2.27 (m, 1H, H2′). The NMR spectrum estab-
lished that the material was 50% rearranged and 50% un-re-
arranged: HRMS-FAB+ m/z 431.1702, calcd for C20H2414N415N1O6
431.1697 (MH+). The NMR spectra matched those reported by
Kim et al. (18) for adducts prepared by an independent pathway.
(2) (+)-8(R),9(S)-Dih yd r oxy-10(S),11(R)-ep oxy-8,9,10,11-
tetr a h yd r oben z[a ]a n th r a cen e (n BADE). Non-bay region
enantiopure (+)-anti-nBADE was prepared by J acobsen oxida-
tion (19, 20).2 [15N6]-2′-Deoxyadenosine (10 mg, 0.04 mmol) was
dissolved in 1 mL of 50 mM Tris-HCl (pH 7.2) to give a
homogeneous solution at 42 °C. Enantiopure epoxide (20 µL of
a 10 mg/mL solution of epoxide in DMSO) was added to the
solution, which was stirred at 42 °C for 72 h. The reaction
mixture was extracted with EtOAc (3 × 1 mL), followed by Et2O
(2 × 1 mL) to remove hydrolysis products. The aqueous solution
was then loaded onto C-18 Sep-Pak cartridges (Waters), and
most of the unreacted nucleoside was removed by washing with
H2O (20 mL). The more lipophilic nBADE-nucleoside adducts
were subsequently eluted with 100% MeOH (2 × 2 mL). The
resulting methanol fraction was evaporated to dryness and
redissolved in MeOH/H2O (50:50). Aliquots (50 µL) were frac-
tionated on a 10 mm × 250 mm C-18 column (YMC-ODS-AQ).
Using a 30 min linear gradient from 40 to 90% MeOH in H2O,
the anti-trans adduct (retention time of 21.9 min) was collected
(one-isomer yield of ∼1%). The following peak at 23.2 min was
most likely the cis adduct, but there was not enough material
to identify it by NMR. UV absorbance was monitored at 258
and 300 nm.
1
N6âSO15Ad o: H NMR (400 MHz, DMSO-d6) δ 8.35 (s, 1H,
adenine H8), 8.22 (d, 1H, adenine H2, J ) 16.5 Hz), 7.63 (s, br,
1H, 14NH, can be exchanged in D2O), 7.36-7.21 (m, 5H, Ar-H),
5.87 (d, 1H, H1′, J ) 6.1 Hz), 5.58 (br, 1H, OH, can be exchanged
in D2O), 5.44 (d, 1H, OH, can be exchanged in D2O, J ) 6.2
Hz), 5.37 (dd, 1H, OH, can be exchanged in D2O, J 1 ) 7.0 Hz,
J 2 ) 2.4 Hz), 5.18 (d, 1H, OH, J ) 4.7 Hz, can be exchanged in
D2O), 4.88 (br, 1H, R-CH), 4.62-4.59 (m, 1H, H2′), 4.15-4.08
(m, 1H, H3′), 3.95 (dd, 1H, H4′, J 1 ) 6.6 Hz, J 2 ) 3.2 Hz), 3.69-
3.53 (m, 4H, â-CH2, H5′, H5′′). The NMR spectrum established
that the material was 100% rearranged: HRMS-FAB+ calcd for
C18H2214N415N1O5, 389.1591 (MH+), found 389.1598.
Syn th esis of th e [15N6]d Ad o 24-m er d (CGATTAATAT-
AGCTATATTAATCG-3′). [15N6]-2′-Deoxyadenosine was con-
verted to the 5′-O-(dimethoxytrityl)-3′-O-(N,N-diisopropylamino)-
(2-cyanoethyl)phosphinyl derivative by literature procedures
(12-15). The phosphoramidite was used to prepare a self-
complementary 24-mer with all adenines labeled at the N6
position. The terminal DMT group was left on, and after removal
of the protecting groups by treatment with concentrated NH4-
OH (8 h, 60 °C), the oligomer was purified using Nensorb
(DuPont) cartridges according to the manufacturer’s directions.
A portion of the Nensorb-purified material was subjected to
additional PAGE purification. Similar results were obtained
from both samples.
To the 15N6-labeled 24-mer (50 A260 units) dissolved in 1 mL
of 0.05 M potassium phosphate buffer (pH 7.00) was added (R)-
styrene oxide (3.5 µL) at 24 h intervals for 30 days. After
removal of most of the unreacted SO by ether extraction (5 × 1
mL), the remainder was removed by HPLC (YMC-ODS-AQ C-18
column; gradient, (A) 0.1 M ammonium formate and (B) CH3-
CN, from 10 to 90% B over the course of 10 min). All peaks were
collected and combined except the last one which was unreacted
epoxide. The solvents were removed by centrifugal evaporation
in vacuo followed by lyophilization. The dry oligonucleotide
mixture was dissolved in 50 mM Na2CO3 (500 µL) and heated
for 10 h at 55 °C to ensure complete Dimroth rearrangement of
N1 adducts. The pH of the reaction was adjusted to pH 8, and
the solution was lyophilized. The dry material was enzymati-
cally hydrolyzed [500 µL of 10 mM MgCl2, 1 unit of snake venom
phosphodiesterase (Sigma, P-7027), and 10 units of alkaline
phosphatase (Sigma, P-4252), 12 h, 37 °C]. The N6 dAdo adducts
[(S)-N6-R, (R)-N6-R, and (R)-N6-â] were collected and analyzed
(+)-(8R,9S,10R,11S)-2′-Deoxy-[15N6](8,9,10,11-tetr a h yd r o-
8,9,10-tr ih yd r oxyben z[a ]a n th r a cen -11-yl)a d en osin e [Non -
Ba y Region (+)-a n ti-tr a n s]: 1H NMR (400 MHz, DMSO-d6)
δ 8.53 (br, 1H, H8), 8.37 (s, 1H, adenine H2), 8.33 (br, 2H, H12,
H1), 8.00 (s, 1H, H7), 7.96 (dd, 1H, 15NH, 1J NH ) 92 Hz, 3J HH
)
8.9 Hz, can be exchanged in D2O), 7.93-7.91 (m, 1H, H2), 7.82
(dd, 2H, H5, H6, J 1 ) J 2 ) 8.9 Hz), 7.76 (d, 1H, H7, J ) 8.9
Hz), 7.58-7.55 (m, 2H, H3, H4), 6.40 (t, 1H, H1′, J 1 ) 6.9 Hz,
J 2 ) 6.7 Hz), 5.88 (br, 1H, H11), 5.32 (d, 1H, OH, J ) 4.0 Hz,
can be exchanged in D2O), 5.23 (t, 1H, OH, J ) 5.7 Hz, can be
exchanged in D2O), 5.06 (d, 1H, OH, J ) 4.0 Hz, can be
2
Synthesized by L. N. Nechev and S. Han via chiral epoxidations
employing a J acobsen catalyst.

1H NMR Study of [15N6]dAdo Alkylation by Epoxides
Chem. Res. Toxicol., Vol. 13, No. 7, 2000 629
comparison with literature data (18, 23, 24).
exchanged in D2O). Because of the very limited amount of
sample, the upfield region of the spectrum was obscured by
solvent peaks. However, the regions of interest in the spectrum
showed that no rearrangement had occurred: HRMS-FAB+ m/z
531.2018, calcd for C28H2814N415N1O6 (MH+) 531.2010. The 1H
NMR spectrum was in agreement with that of the adducted
nucleoside synthesized independently by McNees et al. (21). The
non-biomimetic synthesis had been performed with racemic
aminotriol, and the stereochemistry of the adducts had been
assigned from CD spectra by analogy to other PAH adducts;
the CD spectrum of the adduct prepared in this work from chiral
nBADE confirmed the earlier assignments.
(7R,8S,9R,10S)- an d (7R,8S,9R,10R)-2′-Deoxy-[15N6](7,8,9,-
10-t et r a h yd r o-7,8,9-t r ih yd r oxyb en zo[a ]p yr en -10-yl)a d e-
n osin e [(+)-a n ti-tr a n s a n d (+)-a n ti-cis]. (10S)-a n ti-tr a n s:
1H NMR (400 MHz, DMSO-d6) δ 8.51 (d, 1H, H11, J ) 8.7 Hz),
8.50 (s, 1H, adenine H8), 8.29 (s, 1H, adenine H2), 8.25 (d, 1H,
H1, J ) 7.5 Hz), 8.19 (d, 1H, H4, J ) 3.2 Hz), 8.17 (s, 1H, H6),
1
8.14-8.08 (m, 3H, H5, H2, H3), 8.10 (dd, 1H, 15NH, J NH ) 95
3
Hz, J HH ) 7.5 Hz, can be exchanged in D2O), 8.01 (t, 1H, H2,
J ) 7.6 Hz), 6.42-6.36 (m, 2H, H10, H1′), 5.25 (br, 1H, H7),
4.97 (d, 1H, H8), 4.40 (br, 1H, H3′), 4.24 (m, 2H, H8, H9), 4.02
(m, H4′), 3.54 (m, 2H, H5′, H5′′), 2.87 (m, 2Hs, H2′, H2′′). The
spectrum indicated no rearrangement: HRMS-FAB+ m/z
555.2009, calcd for C30H2814N415N1O6 (MH+) 555.2010.
(3) (()-3â,4r-Dih yd r oxy-1â,2â-ep oxy-1,2,3,4-tetr a h yd r o-
ben z[a ]a n th r a cen e (BADE). [15N6]-2′-Deoxyadenosine (20 mg,
0.08 mmol) was dissolved in 2 mL of 50 mM Tris-HCl (pH 7.2)
to give a homogeneous solution at 42 °C. Racemic BADE (40
µL of 10 mg of epoxide/mL of DMSO) was added to the solution,
which was stirred at 42 °C for 72 h. The reaction mixture was
extracted three times with EtOAc (1 mL), and then twice with
Et2O (1 mL) to remove hydrolysis products. The aqueous
solution was loaded onto C-18 Sep-Pak cartridges, and most of
the unreacted nucleoside was removed by washing with H2O
(20 mL). The more lipophilic BA-nucleoside adducts were
subsequently eluted with 100% MeOH (2 × 2 mL×). The
resulting methanol fraction was evaporated to dryness and
redissolved in MeOH/H2O (50:50). Aliquots (50 µL) were injected
onto a 10 mm × 250 mm C-18 column (YMC-ODS-AQ) for
purification. The peaks of interest [retention times of 22.0 and
22.7 min corresponding to (1S)- and (1R)-anti-trans isomers,
respectively] were collected using a 30 min linear gradient from
40 to 90% MeOH in H2O (combined yield of ∼2%). UV absor-
(10R)-a n ti-cis: 1H NMR (400 MHz, DMSO-d6) δ 8.81 (d, 1H,
H11, J ) 9.4 Hz), 8.53 (s, 1H, adenine H8), 8.31 (s, 1H, adenine
H2), 8.25 (d, 1H, H1, J ) 7.5 Hz), 8.21 (s, 1H, H6), 8.19 (d, 1H,
H3, J ) 4.5 Hz), 8.15 (d, 2H, H4, H5, J ) 5.4 Hz), 8.11 (d, 1H,
H12, J ) 9.4 Hz), 8.02 (t, 1H, H2, J ) 7.6 Hz), 7.45 (dd, 1H,
15NH, J NH ) 92 Hz, J HH ) 10 Hz, can be exchanged in D2O),
1
3
6.69-6.65 (m, 1H, H10), 6.34 (dd, 1H, H1′, J 1 ) 7.4 Hz, J 2
)
6.4 Hz), 5.90 (d, 1H, OH, can be exchanged in D2O), 5.72 (d,
1H, OH, J ) 5.2 Hz, can be exchanged in D2O), 5.29 (d, 1H,
OH, J ) 3.8 Hz, can be exchanged in D2O), 5.18 (br, 1H, H7),
5.10 (dd, 1H, H9, J 1 ) 7.7 Hz, J 2 ) 4.0 Hz), 4.55 (m, 1H, H3′),
4.40 (br, 1H, H8), 3.88 (m, 1H, H4′), 3.61-3.49 (m, 2H, H5′,
H5′′), 2.74 (m, 1H, H2′′), 2.26 (m, 1H, H2′). The spectrum
indicated that no rearrangement had taken place: HRMS-FAB+
m/z 555.1998, calcd for C30H2814N415N1O6 (MH+) 555.2010.
(5) (+)-3â,4r-Dih yd r oxy-1â,2â-ep oxy-1,2,3,4-tetr a h yd r o-
ben zo[c]p h en a n th r en e (BCDE). 15N6-labeled 2′-deoxyadeno-
sine (10 mg, 0.04 mmol) was dissolved completely in 1 mL of 50
mM Tris-HCl (pH 7.2) at 42 °C. An acetone solution of racemic
anti-BCDE (100 µL of a 1 mg/mL solution) was added to the
2′-deoxyadenosine solution. The reaction mixture was stirred
at 42 °C for 72 h. The reaction mixture was extracted twice
with EtOAc and Et2O to remove hydrolysis products. The
aqueous solution was loaded onto a C-18 cartridge and treated
with the same method described previously. The resulting
methanol fractions were evaporated to dryness and redissolved
in MeOH/H2O (50:50). Aliquots (50 µL) were injected onto a
YMC-ODS-AQ HPLC column (10 mm × 250 mm). The peaks of
interest were collected over a 15 min linear gradient from 50 to
100% MeOH in H2O (combined yield of ∼2%). UV absorbance
was monitored at 250 and 303 nm. The two adducts had
retention times of 11.2 and 11.6 min. Peaks were pooled from
multiple HPLC runs. These adducts were subjected to NMR,
CD, and MS analyses. The structural and stereochemical
assignments were confirmed by comparison with literature data
(18, 25, 26).
1
bance was monitored at 258 and 300 nm. The H NMR spectra
were in accord with literature values (21, 22).
(()-(1r,2â,3â,4r)-2′-Deoxy-[15N6](1,2,3,4-tetr a h yd r o-2,3,4-
tr ih yd r oxyben z[a ]a n th r a cen -1-yl)a d en osin e [Ba y Region
1
(+)- a n d (-)-a n ti-tr a n s]. The H NMR spectra of the diaster-
eomers were indistinguishable: 1H NMR (400 MHz, DMSO-d6)
δ 8.58 (s, 1H, adenine H8), 8.53 (s, 1H, H12), 8.43 (s, 1H, H7),
8.28 (s, 1H, adenine H2), 8.03 (dd, 2H, H6, H8, J 1 ) 8.6 Hz, J 2
) 8.0 Hz), 7.70 (d, 1H, H5, J ) 8.6 Hz), 7.65 (dd, 1H, 15NH,
3
1J NH ) 93 Hz, J HH ) 8.3 Hz, can be exchanged in D2O), 7.65-
7.63 (m, 2H, H11), 7.46-7.36 (m, 2H, H9, H10), 6.37 (br, 1H, J
) 5.3 Hz, H1′), 6.29 (m, 1H, H1), 5.31 (d, 1H, OH, J ) 4.0 Hz,
can be exchanged in D2O), 5.26 (br, 1H, H4), 5.20 (d, 1H, H2, J
) 8.3 Hz), 5.14 (t, br, 1H, H3′), 4.96 (d, 1H, OH, J ) 5.2 Hz, can
be exchanged in D2O), 4.67 (t, 1H, OH, J ) 8.3 Hz, can be
exchanged in D2O), 4.40 (br, 1H, H3), 4.14 (m, 1H, H4′), 3.62-
3.50 (m, 2H, H5′, H5′′), 2.75 (m, 1H, H2′′), 2.40 (m, 1H, H2′);
HRMS-FAB+ m/z 531.2002, calcd for C28H2814N415N1O6 (MH+)
531.2001.
(()-(1r,2â,3â,4r)-2′-Deoxy-[15N6](1,2,3,4-tetr a h yd r o-2,3,4-
tr ih yd r oxyben zo[c]p h en a n th r en -1-yl)a d en osin e [(+)- a n d
(-)-a n ti-tr a n s]. (1S)-a n ti-tr a n s: 1H NMR (400 MHz, DMSO-
d6 with D2O) δ 8.61 (d, 1H, H12, J ) 8.8 Hz), 8.53 (s, 1H, adenine
H8), 8.35 (s, 1H, adenine H2), 8.01 (d, 1H, H5, J ) 8.3 Hz),
7.96 (dd, 1H, 15NH, 1J NH ∼ 93 Hz, 3J HH ) 9 Hz, can be exchanged
in D2O; calcd from 2D COSY data), 7.94 (d, 1H, H6, J ) 8.3
Hz), 7.86 (d, 1H, H9, J ) 8.3 Hz), 7.81 (t, 2H, H7, H8, J ) 5.0
Hz), 7.49 (t, 1H, H10, J 1 ) J 2 ) 8.0 Hz), 7.12 (t, 1H, H11, J )
8.8 Hz), 6.39 (dd, 1H, H1′, J 1 ) 9.2 Hz, J 2 ) 8.5 Hz), 6.26 (d,
1H, H1, J ) 4.3 Hz), 4.60 (d, 1H, H4, J ) 7.5 Hz), 4.42 (m, 1H,
H3′), 4.35 (m, 1H, H2), 4.15 (dd, br, 1H, H3, J ) 2.5 Hz), 3.89-
3.88 (m, 1H, H4′), 3.70-3.55 (m, 2H, H5′, H5′′), 2.76 (m, 1H,
H2′′), 2.30 (m, 1H, H2′). The NMR spectra established that
the material was ∼7% rearranged from N1 to N6 (calcd
from 2D COSY data): HRMS-FAB+ m/z 531.2005, calcd for
C28H2814N415N1O6 (MH+) 531.2010.
(4) (+)-7(R),8(S)-Dih yd r oxy-9(S),10(R)-ep oxy-7,8,9,10-
tetr a h yd r oben z[a ]p yr en e (BP DE). [15N6]-2′-Deoxyadenosine
(10 mg, 0.04 mmol) was dissolved in 1 mL of 50 mM Tris-HCl
(pH 7.2) at 42 °C to give a homogeneous solution. To this was
added enantiopure (+)-anti-BPDE (20 µL of a solution of 1 mg
of epoxide/100 µL of DMSO), itself prepared by J acobsen
oxidation (19, 20), and the mixture was stirred at 42 °C for 72
h. The reaction mixture was extracted three times with ethyl
acetate (1 mL), and then twice with Et2O (1 mL) to remove
hydrolysis products. The aqueous solutions were then loaded
onto C-18 Sep-Pak cartridges, and most of the unreacted
nucleoside was removed by washing with H2O (20 mL). The
more lipophilic BPDE-nucleoside adducts were subsequently
eluted with 100% methanol (2 × 2 mL). The resulting MeOH
fraction was evaporated to dryness and redissolved in 1 mL of
MeOH/H2O (50:50). Aliquots (50 µL) were injected onto a 10 mm
× 250 mm C-18 column (YMC-ODS-AQ) for purification. The
peaks of interest, with retention times of 24.2 min (anti-trans)
and 25.5 min (anti-cis), were collected over a 35 min linear
gradient from 40 to 100% MeOH in H2O (combined yield of
∼3%). UV absorbance was monitored at 258 and 333 nm.
Structure (cis or trans) and stereochemistry were confirmed by
(1R)-a n ti-tr a n s: 1H NMR (400 MHz, DMSO-d6 with D2O) δ
8.61 (d, 1H, H12, J ) 8.8 Hz), 8.53 (s, 1H, adenine H8), 8.35 (s,
1H, adenine H2), 8.01 (d, 1H, H5, J ) 8.3 Hz), 7.96 (dd, 1H,
1
3
15NH, J NH ∼ 93 Hz, J HH ) 9 Hz, can be exchanged in D2O;
calcd from 2D COSY data), 7.94 (d, 1H, H6, J ) 8.3 Hz), 7.86

630 Chem. Res. Toxicol., Vol. 13, No. 7, 2000
Kim et al.
Sch em e 3. Syn th esis of [15N6]Ad o a n d d Ad o
Sch em e 4. Rea ction of [15N6]d Ad o w ith (R)-Styr en e
Oxid e, w ith P r ed icted 15N-1H Cou p lin g Con sta n ts
1
(d, 1H, H9, J ) 8.3 Hz), 7.81 (dd, 2H, H7, H8, J 1 ) J 2 ) 5.0
Hz), 7.49 (dd, 1H, H10, J 1 ) J 2 ) 8.0 Hz), 7.12 (t, 1H, H11, J )
8.8 Hz), 6.39 (dd, 1H, H1′, J 1 ) 9.2 Hz, J 2 ) 8.5 Hz), 6.26 (d,
1H, H1, J ) 4.3 Hz), 4.60 (d, 1H, H4, J ) 7.5 Hz), 4.42 (m,
1H, H3′), 4.35 (m, 1H, H2), 4.15 (dd, br, 1H, H3, J ) 2.5 Hz),
3.89-3.88 (m, 1H, H4′), 3.70-3.55 (m, 2H, H5′, H5′′), 2.76 (m,
1H, H2′′), 2.30 (m, 1H, H2′). The NMR spectra established
that the material was ∼7% rearranged from N1 to N6 (calcd
from 2D COSY data): HRMS-FAB+ m/z 531.2010, calcd for
F igu r e 1. 400 MHz H NMR spectra (DMSO-d6, 25 °C) of (a)
N6R(S)-SOd15Ado, (b) N6R(S)-SOd14Ado, and (c) N6R(R)-SOd15
-
Ado adducts.
gave three major components, the structures of which
were established by comparison with unlabeled stan-
dards. Two of them were identified as the N6R (R) and
(S) adducts. CD spectra were used to confirm the con-
figuration of the R adducts (11). The third product was
an N6â adduct. The (R) and (S) â diastereomers had
identical HPLC retention times; however, it was assumed
that the product had the (R) configuration since opening
of the epoxide at the â position does not affect the
configuration of the R position. The order of elution from
a reversed-phase HPLC column was R(S), R(R), and â;
the products were found in a ratio of 7:1:8. The reaction
conditions were chosen such that all N1 products would
be converted to N6, and as a consequence, no N1 products
were detected. The N1 adducts can also undergo deami-
nation (11, 29); however, little (<5%) deaminated product
was seen.
C
28H2814N415N1O6 (MH+) 531.2010.
Resu lts
The reactions of deoxyadenosine with epoxides were
initially carried out with unlabeled nucleoside to estab-
lish chromatography conditions and to obtain standards
for comparison with the labeled products. The structures
of these standards were established by NMR and fast
atom bombardment mass spectrometry (FAB-MS). Stan-
dards were also prepared independently by reaction of
6-chloro- or 6-fluoropurine 2′-deoxyribonucleoside or ri-
bonucleoside (in the case of the styrene oxide reaction)
with the appropriate amino alcohols (18, 27).
The H NMR spectrum of the N6R(S) adduct (Figure
1
[15N6]Deoxya d en osin e a n d [15N6]Ad en osin e. The
[15N6]deoxyadenosine and [15N6]adenosine required for
these studies were prepared by treatment of the corre-
sponding 6-chloronucleoside with 15NH4OH (Scheme 3).
The NMR data for the products were consistent with
those reported by Gao and J ones (28); the 15N6-H
coupling constant was 91 Hz. Integration of the 1H NMR
spectrum showed less than 1% contamination with 14N6;
15N NMR showed that the label was exclusively at N6.
Rea ction of Styr en e Oxid e w ith [15N6]Deoxy-
a d en osin e a n d [15N6]Ad en osin e. The initial reaction
of [15N6]deoxyadenosine with (R)-styrene oxide was car-
ried out in 0.05 M Tris-HCl at pH 7.2 (Scheme 4). HPLC
1a) formed from the 15N6-labeled deoxyadenosine showed
that the product arose from both direct reaction at N6
and reaction at N1 followed by rearrangement. The
doublet (8.5 Hz, vicinal coupling to the benzylic proton)
for the N6 proton signal seen in the unlabeled sample
(Figure 1b) became a doublet of doublets in the isotopi-
cally labeled sample with a large coupling constant (∼92
Hz), due to coupling between the proton and 15N, super-
imposed on the smaller doublet. There was still a
significant 14NH doublet centered under the doublet of
doublets. This result implies that the product was formed
by two pathways. The majority of the material arose by
direct reaction at N6 and the remainder by reaction at

1H NMR Study of [15N6]dAdo Alkylation by Epoxides
Chem. Res. Toxicol., Vol. 13, No. 7, 2000 631
partial racemization. The mixture of products was very
similar to that obtained in the presence of chloride ion.
We conclude that 0.05 M chloride ion is insufficient for
the chlorohydrin mechanism to play a significant role in
the reaction of styrene oxide with deoxyadenosine.
A third reaction was carried out in water without the
addition of any buffers or salts. For this reaction, [15N6]-
adenosine was used rather than the labeled 2′-deoxyad-
enosine. The reaction at 40 °C showed no loss of configu-
rational integrity and gave only N6R(S) and N6â adducts.
The N6R(S) adduct was formed in a 70:30 ratio by direct
reaction at N6 and rearrangement from N1; the N6â
adduct was formed entirely by rearrangement from N1.
A small amount, less than 5%, of an R adduct on the N1
position of 2′-deoxyinosine was also found. The (R)-
styrene oxide adduction reactions of 15N6-labeled adenos-
ine and deoxyadenosine are summarized in Table 1.
It is likely that the only noteworthy difference between
deoxyadenosine and adenosine is the better solubility of
the latter in water. However, there are two significant
differences between the adenosine experiment and the
previous ones. First, the lack of salts in the adenosine
reaction mixture decreased the ability of the medium to
support cations involved in SN1 type mechanisms. More
subtly, the absence of buffer also means the pH is
uncontrolled. Consequently, the reaction mechanism may
be shifting toward SN2 type processes.
1
F igu r e 2. 400 MHz H NMR spectra (DMSO-d6, 25 °C) of (a)
N6âSOd15Ado and (b) N6âSOd14Ado adducts.
N1 followed by Dimroth rearrangement. The ratio of the
un-rearranged to rearranged products was 4:1. The NMR
spectrum of the N6R(R) adduct showed that it arose
almost entirely by direct reaction at N6 (Figure 1c); the
small amount of this isomer that was obtained did not
permit an accurate determination of the 15N content at
N1.
Rea ction of (R)-Styr en e Oxid e w ith th e [15N6]-
d Ad o 24-m er d (CGATTAATATAGCTATATTAATCG-
3′). A self-complementary 24-mer containing nine 15N6-
labeled adenines was used to examine the reaction of
styrene oxide with deoxyadenosine in duplexed DNA. The
oligonucleotide was reacted with (R)-styrene oxide at pH
7.0 and 25 °C (well below the TM of this sequence which
was 44 °C) for 30 days, treated with base at 55 °C to effect
Dimroth rearrangement of N1 products, and enzymati-
cally hydrolyzed to nucleosides. The deoxyadenosine
adducts of styrene oxide were isolated and analyzed by
NMR as before. The results, shown in Table 1, are very
similar to those found for the nucleosides in terms of
relative yields of adducts and the extent of rearrange-
ment, indicating that the existence of duplex structure
in which the N1 and N6 of adenine are involved in
hydrogen bonding does not significantly change the
course of the reaction or the relative reactivity of the two
nitrogens. These results support those of Barlow et al.
(11), who studied the reaction of styrene oxide with DNA
The NMR spectrum of the N6â adduct revealed that it
arose entirely by rearrangement (Figure 2a). The proton
at C2 of adenine, which is a singlet in the adduct of
unlabeled deoxyadenosine (Figure 2b), became a doublet
2
in the labeled material with a J NH coupling constant of
16.4 Hz, indicating that because of rearrangement, the
labeled nitrogen had moved to the N1 position. In the
NMR spectrum, the peaks were broad at room temper-
ature but the lines could be sharpened by raising the
temperature to 52 °C such that the doublet could be
clearly resolved. The line broadening at lower tempera-
tures is presumed to be the result of slow equilibration
among conformers.
Thus, under the conditions of this initial experiment,
the N6â product arose exclusively by reaction at N1
followed by rearrangement whereas the N6R adduct was
formed by a combination of N1 and N6 reaction. More
than 80% of the R product had arisen by direct reaction
at N6. Loss of configurational integrity occurred during
formation of the R product. The R product had mainly
the inverted configuration, i.e., (S); however, 12% of it
had been formed with retention.
One possible explanation for partial racemization
involves reaction of the epoxide with chloride ion. Forma-
tion of an R-chlorohydrin by chloride ion attack on the
epoxide such as that reported by Meehan et al. (30) with
the diol epoxide of benzo[a]pyrene would occur with
inversion of configuration. A second inversion during a
subsequent reaction with the nucleoside would give a net
retention of configuration (30). As a test of this possibility,
the adduction reaction was carried out in the absence of
chloride ion using 0.05 M potassium phosphate at pH 7.0
and 25 °C. The reaction still gave net inversion with
3
containing H-labeled deoxyadenosine. They found the
level of adducts in native DNA decreased 5-fold relative
to that in denatured DNA but the ratio of R to â adducts
was relatively unchanged. They observed somewhat
increased levels of the deaminated â N1-styrene oxide
product in the native DNA sample.
Rea ction of [15N6]Deoxya d en osin e w ith P AH Di-
h yd r od iol Ep oxid es. The PAH dihydrodiol epoxides
with which mechanisms were explored are shown in
Scheme 5. Reactions of the epoxides with deoxyadenosine
were carried out under conditions previously employed
by Canella et al. (24), except that 2′-deoxyadenosine was
used instead of the 3′-phosphate (Scheme 6). The buffer
was 50 mM Tris-HCl (pH 7.2). The nucleoside was stirred
in the buffer until the solution became homogeneous
before adding the epoxide in a small amount of DMSO
solution. After the reaction was complete, the products
were separated by HPLC, and NMR spectra were used
to identify them and to probe the extent of rearrange-

632 Chem. Res. Toxicol., Vol. 13, No. 7, 2000
Ta ble 1. Rea ction of [15N6]Ad en in e Nu cleosid es a n d Oligon u cleotid e w ith (R)-Styr en e Oxid e
Kim et al.
isolated 15N6 adduct
% rearranged
(relative yield)
nucleoside
reaction conditions
H2O, 40 °C, 72 h
R-(S)
R-(R)
â
Ado
2′-dAdo
30 (40-45)
20 (44)
∼0
100 (55-60)
100 (50)
0.05 M Tris-HCl,
not determined (6)
pH 7.2, 40 °C, 72 h
2′-dAdo
0.05 M potassium phosphate,
pH 7.0, 25 °C, 96 h
0.05 M potassium phosphate,
pH 7.0, 25 °C, 30 days
15 (49)
15 (44)
10 (5)
100 (46)
100 (47)
oligonucleotide
d(CGATTAATATAGCTATATTAATCG)
<10 (9)
Sch em e 5. P AH Diol Ep oxid es Used in Rea ction s
w ith [15N6]d Ad o
F igu r e 3. 500 MHz NMR spectrum (DMSO-d6, 23 °C) of anti-
trans-NADE-d15Ado adduct.
Sch em e 6. Rea ction of [15N6]d Ad o w ith P AH
Dih yd r od iol Ep oxid es, w ith P r ed icted
Cou p lin g Con sta n ts
The two major products, eluting from a reverse-phase
HPLC column at 9.5 and 9.8 min, were the diastereo-
meric N6(S) and -(R) trans adducts previously synthe-
sized by Kim et al. (18). Minor components at 8.7 and
8.9 min were obtained in quantities insufficient to permit
identification, but were probably cis adducts.
The diastereomers were collected, and the NMR spec-
tra were analyzed in the manner described for the
styrene oxide adducts (Figure 3). The N6 proton signal
was used to analyze the reaction. The narrowly spaced
doublet for rearranged material was centered within the
wide doublet of doublets (93.0 and 9.0 Hz) for material
that had arisen by direct reaction at N6. The proton at
C2 of adenine should also be useful for analyzing the
extent of rearrangement and should appear as a doublet
(∼13 Hz) for rearranged and a singlet for un-rearranged
material; however, the lines in this region of the spectrum
were too broad to permit analysis. The analysis showed
that 50% of the product arose by reaction at N1 followed
by rearrangement and the remainder by direct reaction
at N6.
Non -Ba y Region (11R)-a n ti-Ben z[a ]a n th r a cen e
Dih yd r od iol Ep oxid e. The 11R enantiomer of anti-
nBADE was used.2 A mixture of cis and trans adducts
was formed in which the trans adduct predominated. The
trans adduct, i.e., 8R,9S,10R,11S, eluted from the HPLC
column at 22 min; the cis adduct, i.e., 8R,9S,10R,11R,
eluted at 23.3 min. The identity of the trans adduct was
confirmed by comparison with independently synthesized
material (21). The cis adduct was obtained in quantities
insufficient for NMR study. The cis and trans tetrols were
obtained as byproducts; they eluted at 18 and 19 min,
respectively.
ment. As in the studies with styrene oxide, the 15N-1H
coupling constants were used to distinguish direct reac-
tion at N6 from Dimroth rearrangement of N1 adducts
(Scheme 6). In the case of the PAH dihydrodiol epoxides,
the major products arose in all cases from trans opening
at the benzylic position of the epoxide; only in the case
of BPDE were cis products isolated in sufficient quantity
for NMR analysis. The structure and stereochemistry of
products were confirmed by comparison with literature
values and by comparison with samples prepared by
independent synthesis.
(()-a n ti-Na p h th a len e Dih yd r od iol Ep oxid e. The
reaction of racemic anti-NADE with [15N6]deoxyadenosine
was carried out under the conditions described above. The
reaction temperature was maintained at 40 °C for 72 h.
Analysis of the NMR spectrum (Figure 4) of the trans
adduct showed that essentially the entire reaction had

1H NMR Study of [15N6]dAdo Alkylation by Epoxides
Chem. Res. Toxicol., Vol. 13, No. 7, 2000 633
1
F igu r e 5. 400 MHz H NMR spectra (DMSO-d6, 25 °C) of (a)
anti-trans-BADE d15Ado adduct (1R) with decoupling at 6.3 ppm
(H1) and (b) the same adduct without decoupling at 6.3 ppm.
spectra with those of products prepared by Kim et al. (18)
using an independent method and with literature values
(22) permitted the first component to be assigned as 1S
and the second as 1R.
F igu r e 4. 400 MHz 1H NMR spectra (DMSO-d6, 25 °C) of anti-
trans-nBADE dAdo adduct (11S) with (a) 15N6 and (b) 14N6.
In summary, neither the bay region nor the non-bay
region benz[a]anthracene dihydrodiol epoxide reacted
significantly at the N1 position of deoxyadenosine.
(10R)-a n ti-Ben zo[a ]p yr en e Dih yd r od iol Ep oxid e.
The reaction of racemic anti-BPDE with 2′-deoxyadeno-
sine gave four products, resulting from the trans and cis
opening of the enantiomeric anti-epoxides at C10. These
diastereomers eluted close to one another, which made
them difficult to purify by HPLC. The availability of
enantiopure (10R)-anti-BPDE provided a solution to the
problem (20). Optically active BPDE was reacted with
[15N6]-2′-deoxyadenosine to give two products, the trans
(10S) and cis (10R) forms, in a ratio of 1:1.3. The peak
that eluted first on HPLC was the anti-trans adduct
(10S), and the second was the anti-cis adduct (10R). The
cis product was formed to a significantly greater extent
than had been seen with the three previous dihydrodiol
epoxides. The reaction was carried out in Tris-HCl buffer,
and this result is in accordance with Meehan’s previous
observation (30) that chloride ion can react with BPDE
to give a chlorohydrin. The reaction competes with direct
attack on the nucleoside. The chlorohydrin can in turn
alkylate the nucleoside. If both steps involve inversion,
the resulting adduct will be cis instead of trans. As
discussed previously, the chloride ion effect was not very
significant in the case of the styrene oxide reaction. The
assignments were made by comparison with materials
synthesized by Kim et al. by reaction of the (()-anti-
trans-BP aminotriols with 6-fluoropurine deoxyriboside
(18, 27) and with other literature values (23, 24). The
relative HPLC elution order is consistent with those
reported by Harvey and co-workers (31).
occurred directly at N6; i.e., there was no evidence of any
rearrangement. In the NMR spectrum of the unlabeled
adduct (Figure 4b), the NH signal is seen as a doublet at
7.97 ppm with a vicinal coupling constant of ∼8.5 Hz.
The adduct from the reaction with [15N6]-2′-deoxyadenos-
ine (Figure 4a) gave a doublet of doublets with a large
coupling constant (1J NH ) 92 Hz) accompanied by a
vicinal coupling constant (8.5 Hz). Any material arising
via Dimroth rearrangement was below the level of
detection (<5%). By the addition of two more benzene
rings to the naphthalene dihydrodiol epoxide, the electron
delocalization in the transition state with this softer
electrophile favored direct attack at the N6 position.
Ba y Region (()-a n ti-Ben z[a ]a n th r a cen e Dih y-
d r od iol Ep oxid e. Racemic material was used for the
study of the BADE reaction. Therefore, diastereomeric
products were expected. Two major adducts were detected
by HPLC between 21 and 22 min. The second-eluting
product overlapped with a following peak which was
probably epoxide. Nevertheless, the adduct could be
purified by multiple HPLC fractionations. Analysis of the
1H NMR spectra of the two adducts showed the two
products to be diastereomeric trans adducts (Figure 5).
Moreover, the spectra showed they arose by direct
adduction at N6. The NH signal for the labeled sample
is partially obscured by another signal (Figure 5b). A
doublet at 7.89 ppm is half of the 15NH signal. The other
half is under the signal for H11 of the hydrocarbon at
7.67 ppm. When the R-proton of the hydrocarbon (H1)
was decoupled (Figure 5a), the doublet at 7.89 ppm
became a singlet and the shape of the peak containing
H11 and the other arm of the 15N doublet of doublets
The trans and cis isomers exhibited distinctly different
NMR spectra that are shown in Figures 6 and 7. The
NMR spectra of the 15N-labeled trans isomer (Figure 6b)
showed that reaction had occurred directly at N6. The
NH signal for the unlabeled trans adduct (Figure 6c)
1
changed. The coupling constants were as follows: J NH
3
) 93.2 Hz and J HH ) 8.3 Hz. Comparison of the CD

634 Chem. Res. Toxicol., Vol. 13, No. 7, 2000
Kim et al.
The NH signal of the cis adduct was sharper than that
of the trans adduct and was 0.5 ppm further upfield
(Figure 7). The doublet of the NH in the unlabeled sample
(Figure 7c) became a doublet of doublets (91.6 and 10.0
Hz) in the labeled material (Figure 7b) which collapsed
to a doublet when the peak at 6.68 ppm (H10) was
irradiated.
As seen with the benz[a]anthracene adducts, additional
benzene rings give more stabilization of charge in the
transition state to yield N6 adducts directly.
(()-a n ti-Ben zo[c]p h en a n th r en e Dih yd r od iol Ep -
oxid e. Commercially available racemic BCDE was used
for this experiment. Two major products, the diastereo-
meric anti-trans adducts, were obtained with retention
times of 11.4 and 11.8 min using a gradient of 50 to 100%
methanol in water over the course of 15 min. Using a
different gradient, 36 to 90% methanol over the course
of 30 min, two minor peaks eluting earlier than the trans
adducts were also observed. These were provisionally
assigned as the cis adducts on the basis of comparison of
UV spectra with those reported by Dipple (32). The
structure of the trans adducts was confirmed by com-
parison with literature values (18, 25), and the stereo-
chemical assignments were confirmed by CD spectros-
copy; the 1S isomer eluted before the 1R.
An interesting observation was made with the BCDE
reaction. The NMR spectra revealed that a small amount
of rearrangement (5-7%) had occurred. To quantify this
observation, many different NMR and MS techniques
1
F igu r e 6. 400 MHz H NMR spectra (DMSO-d6, 25 °C) of the
anti-trans-BPDE dAdo adduct (10S): (a) from the [15N6]dAdo
spectrum acquired with decoupling at 6.4 ppm (H10), (b) the
same adduct without decoupling at 6.4 ppm, and (c) the adduct
derived from [14N6]dAdo.
1
were examined. In the one-dimensional H NMR spec-
trum, the 15NH peaks fell on top of the benzo[c]phenan-
threne aromatic protons. Therefore, they could not be
used to quantitate the NH peaks. For 15N NMR, a
prohibitively costly amount of epoxide would have been
required to prepare sufficient material to achieve spectra
with adequate signal-to-noise ratios.
Proton-carbon HMQC spectra were explored as an
indirect route for observation of 15N spectra. These
spectra are more readily acquired than proton-15N
spectra. Using inverse detection, the natural abundance
of 13C is easier to observe in this way than the 100%
15
1
13C-15
N
isotopically enriched N. It was hoped that the J
coupling (∼20 Hz) would be visible in the 1H-13C spectra.
Spectra with sufficient signal-to-noise levels for seeing
all the key 13C signals were easily obtained. However, it
was not possible to achieve sufficient resolution in the
13C domain to resolve the carbon-nitrogen coupling.
Several mass spectrometric methods were tried with-
out success. Meehan had reported that a permethylated
derivative of the BP Ado adduct showed fragmentation
between C6 and N6 of adenine with electron impact (EI)
ionization (33). Attempts to use this method to observe
the corresponding fragmentation with the unmethylated
and permethylated benzo[a]pyrene and benzo[c]phenan-
threne adducts were unsuccessful using both EI and
electrospray methodologies. FAB+ and FAB- were also
investigated. It is possible that the FAB methods gave
appropriate fragment ions but the background noise in
the spectra was too high for the method to be quantita-
tively useful.
1
F igu r e 7. 400 MHz H NMR spectra (DMSO-d6, 25 °C) of the
anti-cis-BPDE dAdo adduct (10R): (a) adduct from [15N6]dAdo,
where the spectrum was acquired with decoupling at 6.7 ppm
(H10), (b) the same adduct without decoupling at 6.7 ppm, and
(c) the adduct derived from [14N6]dAdo.
Finally, this analytical problem was solved with COSY
spectra (Figure 8). Cross-peaks were observed for vicinal
coupling between the NH and the R-proton of the benzo-
[c]phenanthrene moiety when the spectra were acquired
in anhydrous DMSO-d6. In the isotopically labeled sample,
the cross-peak was split (∼93 Hz) by 15N when the
appeared as a doublet lying at 7.98 ppm. The NH signal
of the product from the reaction with 15N-labeled 2′-
deoxyadenosine appeared as a doublet of doublets (95.4
and 7.5 Hz) which collapsed to a doublet upon irradiation
at 6.4 ppm (H10). No N1 label was detected.

1H NMR Study of [15N6]dAdo Alkylation by Epoxides
Chem. Res. Toxicol., Vol. 13, No. 7, 2000 635
Ta ble 2. Exten t of Dim r oth Rea r r a n gem en t in th e
Rea ction of Cer ta in P AH Dih yd r od iol Ep oxid es w ith
[
15N6]Deoxya d en osin e a s Eva lu a ted by 1H NMRa
anti-dihydrodiol
[
15N6]dAdo
adduct
Dimroth
rearrangement (%)
epoxide
(()-naphthalene
(-)-anti-trans
(+)-anti-trans
(+)-anti-trans
50
∼0
∼0
∼0
5-7
non-bay region
(+)-benz[a]anthracene
bay region
(()-benz[a]anthracene
(+)-benzo[a]pyrene
(()-anti-trans
(+)-anti-trans
(+)-anti-cis
(()-benzo[c]phenanthrene (()-anti-trans
(()-anti-cis or â
a
The reactions were carried out in 0.05 M Tris-HCl buffer at
pH 7.2 and 42 °C for 60-72 h.
N6 adducts did not contain 15N at the N6 position. Thus,
since 15N labeling of the deoxyadenosine was at a level
of ∼99%, it can be concluded that 5-7% of the product
had arisen by Dimroth rearrangement. Both diastereo-
mers were analyzed and gave similar results.
F igu r e 8. 400 MHz NMR spectra (DMSO-d6, 25 °C) of the anti-
trans-BCDE d15Ado adduct (1R): 2D COSY and (inset) cross
section of the 15N6-H region (∼8 ppm).
The data for the reaction of 2′-deoxyadenosine with
PAH dihydrodiol epoxides are summarized in Table 2.
Discu ssion
The adduction reaction between (R)-styrene oxide and
2′-deoxyadenosine when carried out in Tris-HCl (pH 7.2)
yielded N6R(S), N6R(R), and N6â adducts in a ratio of 7:1:
8. The reaction between the riboside and (R)-styrene
oxide carried out in the absence of Tris-HCl gave only
N6R(S) and N6â adducts. The product ratio was 1:1.5. In
both cases, overall combined yields were 20-30%. Deam-
inated N1 adducts were seen in minor amounts (<5%).
Adduction at N1 followed by Dimroth rearrangement
accounted for 20-30% of the N6R adducts. The â adducts
arose entirely by this rearrangement. These reactions
reflect the relative reactivities of N1 and N6 in nucleo-
sides. The studies with adenine incorporated into duplex
DNA by Barlow et al. (11) and by us indicate that,
although the extent of reaction may be decreased in
native DNA, the mechanism of reaction and relative
reactivity of the nitrogens is not significantly affected by
duplex structure. One must conclude that breathing
motions in duplexed DNA lead to disruption of Watson-
Crick A‚T base pairs a significant portion of the time so
that reactions are able to occur at the N1 position.
The reaction of the simplest dihydrodiol epoxide, i.e.,
that of naphthalene, with deoxyadenosine yielded a
significant amount (50%) of N6 adduct via rearrange-
ment. However, in this case, all reaction occurred at the
benzylic position of the epoxide; no products from â attack
were observed. The overall yield was low (3%, one
diastereomer); only trace amounts (<1% of the total
product mixture) of deaminated products were detected.
The polycyclic aromatic hydrocarbon dihydrodiol epoxides
reacted almost exclusively at N6. In the reaction of benzo-
[c]phenanthrene dihydrodiol epoxide, a small amount of
N1 to N6 rearrangement in the formation of the N6
adducts was detected. The other PAH dihydrodiol ep-
oxides may also have given small quantities of N1
adducts which rearranged to the N6 position, but the
amount of adduct formed by this pathway was below the
level of detection.
F igu r e 9. 500 MHz NMR spectrum (DMSO-d6, 28 °C) of the
1
anti-trans-BCDE d15Ado adduct (1R): (a) H 1D spectrum, (b)
1D COSY spectrum, and (c) integration spectrum of the proton
attached to N6.
isotopic label lay at N6 but not when it was at N1. The
spectra exhibited intense signals for the 15N-split peak
but also a weak signal, centered within the wide doublet
due to rearranged material (plus any unlabeled material
that might have been present). It was not clear that the
spectra could be used to obtain good quantitation due to
a number of factors involved in the acquisition and
processing of the 2D spectra. In particular, differences
in relaxation could lead to biasing of results during
spectral apodization. Therefore, a specific 1D COSY
spectrum (Figure 9) for the 15NH region was obtained.
By this experiment, it was determined that 6-8% of the
The interpretation of these results is not totally
straightforward. If electron delocalization in the elec-

636 Chem. Res. Toxicol., Vol. 13, No. 7, 2000
Kim et al.
(6) Qian, C., and Dipple, A. (1995) Different mechanisms of aralky-
lation of adenosine at the 1- and N6-positions. Chem. Res. Toxicol.
8, 389-395.
(7) Latham, G. J ., Zhou, L., Harris, C. M., Harris, T. M., and Lloyd,
R. S. (1993) The replication fate of R- and S-styrene oxide adducts
on adenine N6 is dependent on both the chirality of the lesion
and the local sequence context. J . Biol. Chem. 268, 23427-23434.
(8) Bastlova, T., and Podlutsky, A. (1996) Molecular analysis of
styrene oxide-induced hprt mutation in human T-lymphocytes.
Mutagenesis 11, 581-591.
(9) Harris, C. M., Zhou, L., Strand, E. A., and Harris, T. M. (1991) A
new strategy for synthesis of oligonucleotides bearing adducts at
exocyclic amino sites of purine nucleosides. J . Am. Chem. Soc.
113, 4328-4329.
(10) Kanerva, L. T., Rahiala, K., and Va¨nttinen, E. (1992) Lipase
catalysis in the optical resolution of 2-amino-1-phenylethanol. J .
Chem. Soc., Perkin Trans. 1, 1759-1762.
(11) Barlow, T., Takeshita, J ., and Dipple, A. (1998) Deamination and
Dimroth rearrangement of deoxyadenosine-styrene oxide adducts
in DNA. Chem. Res. Toxicol. 11, 838-845.
(12) Ti, G. S., Gaffney, B. L., and J ones, R. A. (1982) Transient
protection: efficient one-flask syntheses of protected deoxynucleo-
sides. J . Am. Chem. Soc. 104, 1316-1319.
(13) J ones, R. (1984) in Oligonucleotide Synthesis, A Practical Ap-
proach (Gait, M. J ., Ed.) pp 22-34, IRL Press, Washington, DC.
(14) Barone, A. D., Tang, J .-Y., and Caruthers, M. H. (1984) In situ
trophile were the sole factor, one would expect the
naphthalene dihydrodiol epoxide to attack N6 at least as
much as styrene oxide. However, NADE showed an
approximately 1:1 ratio of initial attack at N1 versus N6,
whereas for the SO R adducts, the ratio is closer to 1:3.
The difference may lie in the greater importance that
steric hindrance has for the rigidly constrained electro-
philic site in NADE. Thus, attack at the sterically more
accessible N1 position may be more favored by the
naphthalene epoxide than it is for styrene oxide, the
reaction of which is less affected by steric considerations.
Similarly, the steric hindrance caused by the fjord region
of benzo[c]phenanthrene dihydrodiol epoxide may pro-
duce a higher preference for reaction at N1 than is seen
with the less hindered bay region epoxides or the non-
bay region epoxide. It should be recognized that benz[a]-
anthracene and benzo[c]phenanthrene are isomeric to
each other, and charge in the transition states of the
adduction reactions will be stabilized by the same
number of aromatic rings. However, as Szeliga and
Dipple (34) have suggested, the nonplanar benzo[c]-
phenanthrene fjord region epoxide has less tendency to
ionize, as evidenced by its much slower rate of hydrolysis,
and hence, it has more opportunity to react at the
nucleophilic sites in DNA.
activation of bis-dialkylaminophosphines:
a new method for
synthesizing deoxyoligonucleotides on polymer supports. Nucleic
Acids Res. 12, 4051-4061.
(15) Bannwarth, W., and Trzeciak, A. (1987) A simple and effective
chemical phosphorylation procedure for biomolecules. Helv. Chim.
Acta 70, 175-186.
The conclusion to be drawn from the studies reported
herein is that, whereas aliphatic epoxides react prefer-
entially at the N1 of adenine, the dihydrodiol epoxides
of PAHs react almost entirely by direct reaction with N6
and the epoxides attached to only one aromatic ring give
a mixture of initial attack at N1 and at N6. For aliphatic
and aralkyl epoxides, it is clear that both N1 and N6
adducts (and their deaminated products) need to be
examined carefully for their biological and structural
properties.
(16) Becker, A., J anusz, J . M., and Bruice, T. C. (1979) Solution
chemistry of the syn- and anti-tetrahydrodiol epoxides, the syn-
and anti-tetrahydrodimethoxy epoxides, and the 1,2- and 1,4-
tetrahydro epoxides of naphthalene. J . Am. Chem. Soc. 101,
5679-5687.
(17) Yagi, H., Thakker, D. R., Hernandez, O., Koreeda, M., and J erina,
D. M. (1977) Synthesis and reactions of the highly mutagenic 7,8-
diol 9,10-epoxides of the carcinogen benzo[a]pyrene. J . Am. Chem.
Soc. 99, 1604-1611.
(18) Kim, S. J ., J ajoo, H. K., Kim, H.-Y., Zhou, L., Horton, P., Harris,
C. M., and Harris, T. M. (1995) An efficient route to N6 deoxy-
adenosine adducts of diol epoxides of carcinogenic polycyclic
aromatic hydrocarbons. Bioorg. Med. Chem. 3, 811-822.
(19) Harvey, R. G., and Tang, X.-Q. (1995) Enantioselective synthesis
of the tumorigenic anti-diol epoxide metabolites of benzo[a]pyrene.
Tetrahedron Lett. 36, 2737-2740.
(20) Huang, X., and Harris, T. M. (1995) Enantioselective synthesis
of the potent carcinogen (7R,8S,9S,10R)-benzo[a]pyrene-7,8-
dihydrodiol-9,10-epoxide. J . Chem. Soc., Chem. Commun., 1699-
1700.
(21) McNees, A. G., O’Donnell, M., Horton, P. H., Kim, H.-Y., Kim, S.
J ., Harris, C. M., Harris, T. M., and Lloyd, R. S. (1997) Lack of
correlation between in vitro and in vivo replication of precisely
defined benz[a]anthracene adducted DNAs. J . Biol. Chem. 272,
33211-33219.
(22) Cheh, A. M., Chadha, A., Sayer, J . M., Yeh, H. J . C., Yagi, H.,
Pannell, L. K., and J erina, D. M. (1993) Structures of covalent
nucleoside adducts formed from adenine, guanine, and cytosine
bases of DNA and the optically active bay-region 3,4-diol 1,2-
epoxides of benz[a]anthracene. J . Org. Chem. 58, 4013-4022.
(23) Cheng, S. C., Hilton, B. D., Roman, J . M., and Dipple, A. (1989)
DNA adducts from carcinogenic and noncarcinogenic enantiomers
of benzo[a]pyrene dihydrodiol epoxide. Chem. Res. Toxicol. 2,
334-340.
Ack n ow led gm en t. We thank Dr. Lubomir Nechev
for the synthesis of (+)-8(R),9(S)-dihydroxy-10(S),11(R)-
epoxy-8,9,10,11-tetrahydrobenz[a]anthracene (nBADE)
and Dr. Shin Han for the synthesis of (+)-7(R),8(S)-
dihydroxy-9(S),10(R)-epoxy-7,8,9,10-tetrahydrobenz[a]py-
rene (BPDE). We also thank Professor Liisa Kanerva
(University of Turku, Turku, Finland) for the generous
gift of the chiral 2-amino-1-phenylethanols used in the
synthesis of the N6âSOdAdo nucleoside standards and
Markus Voehler (Vanderbilt University) for his invalu-
able assistance with the HMQC and 1D COSY NMR
experiments. This research was generously supported by
Grants ES 05355 and ES 00267 from the United States
Public Health Service.
Refer en ces
(24) Canella, K., Peltonen, K., and Dipple, A. (1991) Identification of
(+) and (-) anti-benzo[a]pyrene dihydrodiol epoxide-nucleic acid
adducts by the 32P-postlabeling assay. Carcinogenesis 12, 1109-
1114.
(25) Agarwal, S. K., Sayer, J . M., Yeh, H. J . C., Pannell, L. K., Hilton,
B. D., Pigott, M. A., Dipple, A., Yagi, H., and J erina, D. M. (1987)
Chemical characterization of DNA adducts derived from the
configurationally isomeric benzo[c]phenanthrene-3,4-diol 1,2-
epoxides. J . Am. Chem. Soc. 109, 2497-2504.
(26) Lakshman, M. K., Yeh, H. J . C., Yagi, H., and J erina, D. M. (1992)
Synthesis of deoxyadenosine adducts from the highly carcinogenic
fjord region diol epoxide of benzo[c]phenanthrene. Tetrahedron
Lett. 33, 7121-7124.
(27) Kim, S. J ., Harris, C. M., J ung, K.-Y., Koreeda, M., and Harris,
T. M. (1991) Non-biomimetic route to deoxyadenosine adducts of
carcinogenic polycyclic aromatic hydrocarbons. Tetrahedron Lett.
32, 6073-6076.
(1) Solomon, J . J ., Mukai, F., Fedyk, J ., and Segal, A. (1988) Reactions
of propylene oxide with 2′-deoxynucleosides and in vitro with calf
thymus DNA. Chem.-Biol. Interact. 67, 275-294.
(2) 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.
(3) Plna, K., Segerback, D., and Schweda, E. K. (1996) DNA adduct
formation by allyl glycidyl ether. Carcinogenesis 17, 1465-1471.
(4) Vodicka, P., and Hemminki, K. (1988) Identification of alkylation
products of styrene oxide in single- and double-stranded DNA.
Carcinogenesis 9, 1657-1660.
(5) Hemminki, K., and Hesso, A. (1984) Reaction products of styrene
oxide with guanosine in aqueous media. Carcinogenesis 5, 601-
607.

1H NMR Study of [15N6]dAdo Alkylation by Epoxides
Chem. Res. Toxicol., Vol. 13, No. 7, 2000 637
(28) Gao, X., and J ones, R. A. (1987) Nitrogen-15-labeled deoxynucleo-
sides. Synthesis of [6-15N]- and [1-15N]deoxyadenosines from
deoxyadenosine. J . Am. Chem. Soc. 109, 1275-1278.
(29) Barlow, T., Ding, J ., Vouros, P., and Dipple, A. (1997) Investiga-
tion of hydrolytic deamination of 1-(2-hydroxy-1-phenylethyl)-
adenosine. Chem. Res. Toxicol. 10, 1247-1249.
(30) Wolfe, A. R., and Meehan, T. (1994) Halide-catalyzed cis product
formation in the hydrolysis of anti-benzo[a]pyrene diol epoxide
and its alkylation of poly(A). Chem. Res. Toxicol. 7, 110-119.
(31) J effrey, A. M., Grzeskowiak, K., Weinstein, I. B., Nakanishi, K.,
Roller, P., and Harvey, R. G. (1979) Benzo[a]pyrene-7,8-dihy-
drodiol 9,10-oxide adenosine and deoxyadenosine adducts: struc-
ture and stereochemistry. Science 206, 1309-1311.
(32) Dipple, A., Pigott, M. A., Agarwal, S. K., Yagi, H., Sayer, J . M.,
and J erina, D. M. (1987) Optically active benzo[c]phenanthrene
diol epoxides bind extensively to adenine in DNA. Nature 327,
535-536.
(33) Yamamoto, J ., Subramaniam, R., Wolfe, A. R., and Meehan, T.
(1990) The formation of covalent adducts between benzo[a]-
pyrenediol epoxide and RNA: structural analysis by mass
spectrometry. Biochemistry 29, 3966-3972.
(34) Szeliga, J ., and Dipple, A. (1998) DNA adduct formation by
polycyclic aromatic hydrocarbon dihydrodiol epoxides. Chem. Res.
Toxicol. 11, 1-11.
TX000054M