Identification of a Fused-Ring 2′-Deoxyadenosine Adduct Formed in Human Cells Incubated with 1-Chloro-3-buten-2-one, a Potential Reactive Metabolite of 1,3-Butadiene

Article abstract of DOI:10.1021/acs.chemrestox.6b00095

1-Chloro-3-buten-2-one (CBO) is an in vitro metabolite of 1,3-butadiene (BD), a carcinogenic air pollutant. CBO exhibited potent cytotoxicity and genotoxicity that have been attributed in part to its reactivity toward DNA. Previously, we have characterized the CBO adducts with 2′-deoxycytidine and 2′-deoxyguanosine. In the present study, we report on the reaction of CBO with 2′-deoxyadenosine (dA) under in vitro physiological conditions (pH 7.4, 37 °C). We used the synthesized standards and their decomposition and acid-hydrolysis products to characterize the CBO-DNA adducts formed in human cells. The fused-ring dA adducts (dA-1 and dA-2) were readily synthesized and were structurally characterized as 1,N⁶-(1-hydroxy-1-hydroxymethylpropan-1,3-diyl)-2′-deoxyadenosine and 1,N⁶-(1-hydroxy-1-chloromethylpropan-1,3-diyl)-2′-deoxyadenosine, respectively. dA-1 exhibited a half-life of 16.0 ± 0.7 h and decomposed to dA at pH 7.4 and 37 °C. At similar conditions, dA-2 decomposed to dA-1 and dA, and had a half-life of 0.9 ± 0.1 h. These results provide strong evidence for dA-1 being a degradation product of dA-2. dA-1 is formed by replacement of the chlorine atom of dA-2 by a hydroxyl group. The slow decomposition of dA-1 to dA, along with the detection of hydroxymethyl vinyl ketone (HMVK) as another degradation product, suggested equilibrium between dA-1 and a ring-opened carbonyl-containing intermediate that undergoes a retro-Michael reaction to yield dA and HMVK. Acid hydrolysis of dA-1 and dA-2 yielded the corresponding deribosylated products A-1D and A-2D, respectively. In the acid-hydrolyzed reaction mixture of CBO with calf thymus DNA, both A-1D and A-2D could be detected; however, the amount of A-2D was significantly larger than that of A-1D. Interestingly, only A-2D could be detected by LC-MS analysis of acid-hydrolyzed DNA from cells incubated with CBO, suggesting that dA-2 was stable in DNA and thus may play an important role in the genotoxicity and carcinogenicity of BD. In addition, A-2D could be developed as a biomarker of CBO formation in human cells.

Full text of DOI:10.1021/acs.chemrestox.6b00095

Article
pubs.acs.org/crt
Identification of a Fused-Ring 2′-Deoxyadenosine Adduct Formed in
Human Cells Incubated with 1‑Chloro-3-buten-2-one, a Potential
Reactive Metabolite of 1,3-Butadiene
Fang-Mao Zeng,†,∥ Ling-Yan Liu,†,∥ Jin Zheng,† Cong Kong,‡ Jing An,† Ying-Xin Yu,† Xin-Yu Zhang,*,†
and Adnan A. Elfarra*,§
†Institute of Environmental Pollution and Health, School of Environmental and Chemical Engineering, Shanghai University, Shanghai
200444, China
‡East China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Shanghai 200090, China
§Department of Comparative Biosciences and the Molecular and Environmental Toxicology Center, University of
WisconsinMadison, Madison, Wisconsin 53706, United States
S
* Supporting Information
ABSTRACT: 1-Chloro-3-buten-2-one (CBO) is an in vitro
metabolite of 1,3-butadiene (BD), a carcinogenic air pollutant.
CBO exhibited potent cytotoxicity and genotoxicity that have been
attributed in part to its reactivity toward DNA. Previously, we have
characterized the CBO adducts with 2′-deoxycytidine and 2′-
deoxyguanosine. In the present study, we report on the reaction of
CBO with 2′-deoxyadenosine (dA) under in vitro physiological
conditions (pH 7.4, 37 °C). We used the synthesized standards and
their decomposition and acid-hydrolysis products to characterize the
CBO−DNA adducts formed in human cells. The fused-ring dA
adducts (dA-1 and dA-2) were readily synthesized and were
structurally characterized as 1,N6-(1-hydroxy-1-hydroxymethylpro-
pan-1,3-diyl)-2′-deoxyadenosine and 1,N6-(1-hydroxy-1-chloromethylpropan-1,3-diyl)-2′-deoxyadenosine, respectively. dA-1
exhibited a half-life of 16.0 0.7 h and decomposed to dA at pH 7.4 and 37 °C. At similar conditions, dA-2 decomposed
to dA-1 and dA, and had a half-life of 0.9 0.1 h. These results provide strong evidence for dA-1 being a degradation product of
dA-2. dA-1 is formed by replacement of the chlorine atom of dA-2 by a hydroxyl group. The slow decomposition of dA-1 to dA,
along with the detection of hydroxymethyl vinyl ketone (HMVK) as another degradation product, suggested equilibrium
between dA-1 and a ring-opened carbonyl-containing intermediate that undergoes a retro-Michael reaction to yield dA and
HMVK. Acid hydrolysis of dA-1 and dA-2 yielded the corresponding deribosylated products A-1D and A-2D, respectively. In the
acid-hydrolyzed reaction mixture of CBO with calf thymus DNA, both A-1D and A-2D could be detected; however, the amount
of A-2D was significantly larger than that of A-1D. Interestingly, only A-2D could be detected by LC-MS analysis of acid-
hydrolyzed DNA from cells incubated with CBO, suggesting that dA-2 was stable in DNA and thus may play an important role in
the genotoxicity and carcinogenicity of BD. In addition, A-2D could be developed as a biomarker of CBO formation in human
cells.
The major environmental sources include automobile exhaust,
cigarette smoke, and exhaust from biomass burning.1
Carcinogenicity of BD originates from its reactive metabo-
lites, including 3,4-epoxy-1-butene, 3,4-epoxy-1,2-butanediol,
1,2,3,4-diepoxybutane, and likely hydroxymethyl vinyl ketone
(HMVK).9 The metabolites can form DNA adducts and cause
DNA damage and mutation, which is probably one of the
underlying mechanisms of carcinogenicity of these metabo-
lites.10
INTRODUCTION
■
1-Chloro-3-buten-2-one (CBO), an α,β-unsaturated ketone, is a
potential metabolite of 1,3-butadiene (BD), a known human
carcinogen.1 It is formed from BD via a two-step process
(Scheme 1). The first step is a myeloperoxidase (MPO)-
mediated reaction of BD with hydrogen peroxide in the
presence of high concentrations of chloride ion (>50 mM) to
form 1-chloro-2-hydroxy-3-butene (CHB).2 The second step is
an alcohol dehydrogenase (ADH)-mediated oxidation of CHB
to produce CBO.3
Evidence for an alternative metabolic pathway, in which two
different metabolites, i.e., CHB and CBO, are formed, has been
BD is an air pollutant with high cancer risks.4−7 It is
considered the most carcinogenic agent in cigarette smoke.8 BD
is ubiquitous in the environment because it is formed as a
product of incomplete combustion of fossil fuels and biomass.
Received: March 19, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.chemrestox.6b00095
Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Chemical Research in Toxicology
Article
All phosphate buffers (100 mM) contained 100 mM KCl. The HepG2
cell line was purchased from the Center of Cell Resources of the
Shanghai Institutes for Biological Sciences, Chinese Academy of
Sciences (Shanghai, China). Dulbecco’s modified Eagle’s medium
(DMEM) and fetal bovine serum (FBS) were purchased from Life
Technologies (Grand Island, NY, U.S.). Biomiga EZgene Tissue
gDNA Miniprep Kit was obtained from Biomiga, Inc. (San Diego, CA,
U.S.).
Scheme 1. Formation of CHB and CBO from BD
Instruments and Methods. Mass spectra were obtained on a
Bruker microTOF II mass spectrometer with electrospray ionization
(ESI). NMR spectra were recorded on a Bruker Instruments Avance
AV500 spectrometer (500 MHz). Analysis of samples and isolation of
products were performed on a Waters Alliance 2695 HPLC system
(Milford, MA, U.S.) equipped with a Waters 2996 diode array
detector. An Agilent ZORBAX 80 Å StableBond SB-C18 analytical
column (4.6 × 250 mm, 5 μm) was used for analysis of samples.
Different gradient programs and isocratic elution conditions were used
in the analysis of samples. For analysis of CBO-dA reaction mixtures,
the acid hydrolysis products of dA-1 and dA-2, and the decomposition
products of dA-1 and dA-2 under in vitro physiological conditions, a
linear gradient program was used, starting at 1 min from 0% pump B
to 22% pump B over 11 min [pump A, 1% (v/v) methanol with the
pH being adjusted to 2.5 with TFA; pump B, 10% (v/v) methanol at
pH 2.5] at a flow rate of 1 mL/min and then at 12 min from 22% to
100% pump B over 2 min, at 22 min from 100% to 0% pump B over 1
min, and stopped at 23 min. The gradient for analyses of the CBO−
DNA reaction mixtures started at 1 min from 0.1% pump B to 10%
pump B over 19 min (pump A, water with the pH being adjusted to
2.5 with TFA; pump B, acetonitrile containing 0.1% TFA) at a flow
rate of 1 mL/min and then at 20 min from 10% to 15.3% pump B over
9 min, at 29 min from 15.3% to 0.1% pump B over 1 min, and stopped
at 30 min. The column temperature was set at 30 °C. For preparative
isolation of the products, a Waters SunFire C18 preparative column
(10 × 150 mm, 5 μm) was employed with the same mobile phases.
dA-1 and its decomposition product dA were isolated through
isocratic elution with 50% pump B at a flow rate of 3 mL/min. A-1D
was purified through isocratic elution with 100% pump A at a flow rate
of 2 mL/min, and A-2D was similarly isolated with 20% pump B at a
flow rate of 3 mL/min. Fractions were collected manually and were
lyophilized on a CHRIST ALPHA 1−4 LD-2 freeze-dryer (Osterode
am Harz, Germany). For analysis of the HMVK standard, and HMVK-
GSH and dA-1-GSH reaction mixtures, the linear gradient program
started at 1 min from 0% pump B to 10% pump B over 19 min (pump
A, water with the pH being adjusted to 2.5 using TFA; pump B,
acetonitrile containing 0.019% TFA) at a flow rate of 1 mL/min and
then at 20 min from 10% to 0% pump B over 1 min, and stopped at 21
min. The column temperature was set at 30 °C except the preparative
isolation of the products. LC-MS analysis was performed on an Agilent
1290 Infinity HPLC system connected to an Agilent 6224 TOF mass
spectrometer with ESI. An Agilent Poroshell 120 PFP column (3.0 ×
100 mm, 2.7 μm) was used with a linear gradient program, starting at 1
min from 0.1% pump B to 53.2% pump B over 10 min [pump A, water
containing 0.1% (v/v) formic acid; pump B, acetonitrile containing
0.1% formic acid] at a flow rate of 0.3 mL/min, at 11 min from 53.2%
to 0.1% pump B over 1 min, and stopped at 12 min.
obtained.2,3 CHB is expected to be generated in immune and
bone marrow cells due to the requirement for MPO,2 and BD
has been known to cause lymphohematopoietic cancers in
occupationally exposed workers,1 implying an association
between the carcinogenicity of BD in humans and the
alternative MPO-mediated metabolic pathway. However, the
toxicological significance of CBO has not been well investigated
in part because of the high reactivity of CBO which prevents
the development of direct methods to assess its formation in
vivo. Recently, we reported both CHB and CBO to be
cytotoxic and genotoxic to human liver cells in culture with
CBO being approximately 100-fold more potent than CHB.11
Results obtained with the Ames test indicated that CHB was a
mutagen; however, the high toxicity of CBO prevented a
reliable assessment of its mutagenicity.11
CBO is a Michael acceptor that is expected to react with
nucleobases to form DNA adducts. Indeed, we have found that
CBO reacted readily with 2′-deoxycytidine (dC) and 2′-
deoxyguanosine (dG) to produce multiple adducts under in
vitro physiological conditions (pH 7.4, 37 °C).12,13 The
reactivity of CBO toward nucleobases is probably the basis of
CBO genotoxicity and implies that CBO may be a mutagen. In
the present study, we investigated the reaction of CBO with 2′-
deoxyadenosine (dA) under in vitro physiological conditions
and characterized the formed adducts, their stabilities, and the
decomposition and acid-hydrolysis products to obtain stand-
ards that were used to characterize the major CBO-dA adducts
formed in the reactions with DNA and in human cells in vitro.
Synthesis of HMVK and the HMVK-Glutathione (GSH)
Conjugate. HMVK was synthesized according to the procedure in
the literature with some modifications.14 Briefly, 2-butyn-1,4-diol (1.72
g, 20 mmol) was dissolved in 15 mL of water. Mercury(II) sulfate
(0.09 g, 0.30 mmol) and 0.11 mL of concentrated sulfuric acid were
added, and the suspension was stirred at ambient temperature for 20 h.
Barium carbonate (1.3 g, 6.6 mmol) was added to neutralize the
mixture to pH 6−7. After filtration, a pale yellow solution was
obtained, whose HPLC chromatogram exhibited only one peak at 7.4
min with the λmax at 212 nm. The solution was directly used to prepare
the HMVK-GSH conjugate as described previously.15 The HMVK-
GSH conjugate was eluted at 10.1 min.
EXPERIMENTAL PROCEDURES
■
Materials. BD (99%) was purchased from Dalian Da-Te Gas Ltd.
(Dalian, China). Calcium hypochlorite and dA monohydrate were
obtained from Alfa Aesar (Ward Hill, MA, U.S.). Chromium oxide was
obtained from Adamas Reagent Company (Shanghai, China). L-
Buthionine-sulfoximine (BSO), calf thymus DNA (CT DNA, Cat. No.
D1501), and dimethyl sulfoxide-d6 (DMSO-d6) was obtained from
Sigma-Aldrich (St. Louis, MO, U.S.). Trifluoroacetic acid (TFA, purity
≥99.0%), tris(hydroxymethyl)aminomethane (Tris), and other
reagents (analytical reagent grade) were purchased from Sinopharm
Chemical Reagent Company (Shanghai, China). CBO was synthesized
as reported previously and was usually used without further
purification (purity ∼90%).3 Pure CBO (>98%) was obtained through
purification on a silica gel column with dichloromethane as the eluent.
Reactions of CBO with dA at in Vitro Physiological
Conditions and at Other pH Values. Time-Course Experiments.
dA (5.0 mg, 0.019 mmol) was dissolved in 2 mL of phosphate buffer
B
DOI: 10.1021/acs.chemrestox.6b00095
Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Chemical Research in Toxicology
Article
with the specified pH value, and CBO (6.3 μL, purity 95%, 0.065
mmol, CBO/dA = 3.5:1) was added. The mixture was incubated at 37
°C for 7 h. Aliquots (50 μL) were withdrawn every hour and were
subjected to HPLC analysis.
Isolation of Products for Structural Characterization. To
prepare the products for structural characterization, the reactions were
performed on a larger scale. To prepare dA-1, dA (9.0 mg, 0.033
mmol) was dissolved in 3.6 mL of pH 7.4 phosphate buffer, and crude
CBO (12 μL, purity 90%, 0.12 mmol, CBO/dA = 3.5:1) was added.
The solution was incubated at 37 °C for 6 h, extracted with diethyl
ether three times, and subjected to isolation of products on preparative
HPLC.
dA-2 was quite labile with a short half-life (0.9 h at pH 7.4 and 37
°C). Attempts to isolate pure dA-2 for structural characterization
failed, even when the beakers to collect fractions were kept on ice. The
problem was partly avoided by directly lyophilizing the reaction
mixture at pH 4.0 without isolation of dA-2, although the dA-2
obtained was not highly pure (purity ∼80%). Nonetheless, structural
characterization of dA-2 by NMR spectroscopy was achieved in this
way. To prepare dA-2 for structural characterization, the reaction of
CBO with dA was carried out in TFA-acidified water instead of
phosphate buffer as in the preparation of other products. Specifically,
dA (11.0 mg, 0.041 mmol) was dissolved in 4.4 mL of water with the
pH being adjusted to 4.0 using TFA, and pure CBO (75 μL, 0.82
mmol, CBO/dA = 20:1) was added. The mixture was incubated at 37
°C for 6 h, extracted with diethyl ether three times, and lyophilized.
The resulting solid, which predominantly contained dA-2 (accounted
for ∼82% of all compounds except CBO) as estimated by HPLC, was
directly dissolved in DMSO-d6 for NMR analysis or in acetonitrile−
water (1:1, v/v) for ESI-MS analysis.
in 50 mM, pH 7.2 Tris buffer containing 1 mM MgCl2. CBO (2 μL,
purity 80%, 17 μmol) was dissolved in 40 μL of isopropanol. To 0.5
mL of DNA solution, 2 μL of CBO solution in isopropanol was added.
The reaction mixture was then incubated at 37 °C for 9 h. After
hydrolysis in 0.1 M HCl at 75 °C for 20 min, the solution was cooled
to room temperature, neutralized to pH ∼5, filtered with a 0.22 μm
membrane, and then subjected to HPLC analysis.
Detection of A-2D in Cells Incubated with CBO. HepG2 cells
were cultured as monolayers in DMEM supplemented with 10% FBS
and antibiotics (100 U/mL penicillin and 100 μg/mL streptomycin) at
37 °C in a humidified atmosphere with 5% CO2 and routinely
passaged by trypsinization when nearly confluent. Cells at ∼60%
confluence were incubated with 100 μM BSO in FBS-containing
DMEM at 37 °C for 24 h, and then with 1 μM CBO in D-Hank’s
balanced salt solution at 37 °C for 1 h. Cellular DNA was extracted
with Biomiga EZgene Tissue gDNA Miniprep Kit. The DNA extracted
was immediately hydrolyzed in 0.1 M HCl at 75 °C for 20 min. After
cooling to room temperature, adjusting pH to ∼3, and filtering with a
0.22 μm membrane, the hydrolyzed solution was subjected to LC-MS
analysis.
RESULTS
■
Reaction of CBO with dA under in Vitro Physiological
Conditions. The reaction of CBO with dA under in vitro
physiological conditions proceeded readily. Two major
products, which were designated as dA-1 and dA-2 in the
order of the retention times (which were 10.6 and 20.6 min,
respectively), were formed (Figure 1). Their UV/vis absorption
To prepare dA formed from decomposition of dA-1, 15.0 mg of dA
(0.056 mmol) was dissolved in 6.0 mL of pH 4.0 phosphate buffer, and
22 μL of crude CBO (purity 90%, 0.22 mmol, CBO:dA = 3.9:1) was
added. Incubation was carried out at 37 °C for 2.75 h, and HPLC
analysis indicated that the solution contained almost pure dA-2 (98%
as estimated by peak areas at 260 nm). After being extracted with
diethyl ether 5 times, the aqueous solution was divided into two equal
parts (3.0 mL per part). One part was used to prepare the
decomposition product of dA-1 (i.e., dA) and the other to prepare
A-2D. To investigate the decomposition of dA-1, the pH of the
solution was adjusted to 7.4 with KOH, and dA was isolated on
preparative HPLC after incubation at 37 °C for 8 h. The other part of
the solution was acidified by adding 61 μL of 5 M HCl and was left in
a boiling water bath for 1 h. After cooling down, the pH was adjusted
to ∼5, and then A-2D was isolated using preparative HPLC.
A-1D was obtained by dissolving purified dA-1 in 0.1 M HCl and
heating the acidic solution in a boiling water bath for 1 h. After cooling
down and adjusting the pH to ∼6, A-1D was isolated on preparative
HPLC.
Figure 1. Typical chromatogram of the reaction mixture of CBO and
dA under in vitro physiological conditions (monitored at 220 nm).
The CBO/dA molar ratio was 3.5:1, and the reaction time was 1 h.
Stabilities of Products, Determination of the Half-Lives, and
Decomposition of Products under in Vitro Physiological
Conditions. General Procedure. A small quantity of purified product
(1−2 mg) was dissolved in pH 7.4 phosphate buffer and incubated at
37 °C for different time points. The solution was then analyzed using
HPLC. For the determination of the half-lives, aliquots were
withdrawn every hour and were subjected to HPLC analysis. The
peak areas of the compounds to be examined were determined and
were regressed against time using exponential decay (y = a × e‑bx). The
half-lives were calculated using the data obtained after regression. The
experiments were carried out in triplicate, and the reported values were
the means and standard deviations of three independent experiments.
Detection of HMVK Generated during the Decomposition of
dA-1 under in Vitro Physiological Conditions. To detect HMVK
in the incubation mixture, GSH was used as the trapping agent. A pH
7.4 phosphate buffer (305 μL) containing ∼0.024 μmol of purified dA-
1 (estimated by the peak area) and 0.14 μmol of GSH was incubated
at 37 °C for 29 h. The solution was then analyzed for the formation of
the HMVK-GSH conjugate by HPLC as described above.
spectra were similar to that of dA with the absorption maxima
being 212 and 263 nm for dA-1, and 212 and 264 nm for dA-2
(λmax for dA was 257 nm) (Figure S1). Thus, dA-1 and dA-2
were expected to be adducts of dA.
Time- and pH-Dependence of the Formation of
Products. The formation of dA-2 was rapid; in fact, it could
be observed immediately after CBO and dA were mixed at
room temperature. However, the formation of dA-1 could be
observed only after starting the incubation of the reaction
mixture at 37 °C. The two products showed different patterns
of formation (Figure 2).
The reaction of CBO with dA was examined at pH 4, 7.4,
and 10 to determine the optimal conditions for preparation of
the products. Under all three pH conditions, the amounts of
dA-2 increased over time at the early stages of the reactions
(within 1−2 h) and subsequently decreased, whereas those of
dA-1 only showed increases up to 7 h (Figure 2), suggesting
that dA-2 underwent an initial formation and subsequent
Reaction of CBO with CT DNA under in Vitro Physiological
Conditions. CT DNA (2 mg/mL) was rehydrated overnight at 4 °C
C
DOI: 10.1021/acs.chemrestox.6b00095
Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Chemical Research in Toxicology
Article
overlapped the signals of three protons at 3.52 and 3.59 ppm,
rendering the integrals of these signals less accurate (Figure
S3). The proton number suggested that dA-1 was a quaternary
ammonium salt. In the spectrum, five broad single peaks
appeared at 10.80, 6.75, 5.49, 5.35, and 5.07 ppm, and the
corresponding signals were absent from the heteronuclear
multiple-quantum correlation (HMQC) spectrum (Figure S4),
indicating that these peaks were the signals of protons at
hydroxyl or amino groups. The peak at 10.80 ppm apparently
could only be assigned as the proton at the exocyclic amino
group. The fact that it contained only one proton and appeared
at low field (the signal of exocyclic amino group in dA in
DMSO-d6 appears at 7.40 ppm) clearly indicated that both N1-
and N6-positions of dA were alkylated. Alkylation at the N1-
position led to the formation of a positive charge at this
nitrogen atom. The positive charge can move to the nitrogen
atom of the exocyclic amino group via tautomerization,
explaining why the proton at the exocyclic amino group in
dA-1 appeared at such a low field (the feature seems a reliable
tool to identify adenine adducts containing a quaternary
nitrogen at the N1 position of adenine, as similar proton shifts
were also observed with dA-2 and A-2D; see Table 1). dA-1
contains four hydroxyl groups, consistent with cyclization by
the addition of a nitrogen atom to the carbonyl carbon of CBO.
The two protons on the purine ring (H2 and H8) exhibited
very similar chemical shifts (8.69 and 8.74 ppm). The
assignment of the proton and carbon signals at positions 2
and 8 could be established through the heteronuclear multiple-
bond correlation (HMBC) spectrum (see below). The signals
of the deoxyribose protons of dA-1 shifted little in comparison
to the deoxyribose protons of dA, thus being easily identified
(Table 1). After these protons were assigned, the remaining six
protons could only be assigned to those on three methylene
groups in the CBO moiety. The HMQC spectrum shows
corresponding 13C signals at 65.3, 43.5, and 27.5 ppm (Table
1), which are assigned to 1″-, 4″-, and 3″-CH2 (the numbering
of the carbon atoms in dA-1 is shown in Scheme 2),
respectively, based on the most downfield signal belonging to
the carbon bearing a hydroxyl substituent (1″-CH2) and the
highest field signal belonging to the carbon attached only to H
and C (3″-CH2).
Figure 2. Time-dependent formation of dA-1 and dA-2 at different
pH values. (A) pH 7.4; (B) pH 4.0; and (C) pH 10.0. The CBO/dA
molar ratios were all 3.5:1, and the detection wavelength was 260 nm.
decomposition. The incubation pH significantly affected the
yields and ratios of the products. At pH 7.4, the peak area ratio
of dA-2 to dA-1 changed over time, from 1.7 at 1 h to 0.11 at 7
h. At pH 10, the yields and ratios of products were similar to
those obtained at pH 7.4. However, different results were
observed at pH 4. At this pH, the yield of dA-2 greatly
increased and that of dA-1 significantly decreased, leading to a
dA-2:dA-1 ratio of 135:1 at 1 h and 12:1 at 7 h. This indicated
that the formation of dA-2 was greatly favored at pH 4.
Meanwhile, decomposition of dA-2 was considerably decreased
at this pH (Figure 2A).
In the HMBC spectrum (Figure S5), H1′ was observed to
couple to two carbon atoms at 142.4 and 146.3 ppm, which
were thus assigned as C8 and C4, respectively, because the
carbon atom at 142.4 ppm carries a hydrogen atom as indicated
by the HMQC spectrum. The assignment was further
supported by a very weak coupling of H8 at 8.74 ppm with
C1′. Therefore, the other hydrogen-carrying carbon atom on
the purine ring, i.e., C2, which had a 13C chemical shift at 146.9
ppm, was identified. The determination of the H2 and C2
chemical shifts was critical to establish the structure of dA-1.
The HMBC spectrum showed that H2 coupled with C4″, and
in turn, H-4″ coupled with C2, indicating that C4″ was
connected to N1. Thus, it was C2″ (i.e., the carbon atom at the
original carbonyl group in the CBO moiety) that was
connected to the exocyclic amino group. Other coupling
information was consistent with such a structure. Therefore,
dA-1 was characterized as 1,N6-(1-hydroxy-1-hydroxymethyl-
propan-1,3-diyl)-2′-deoxyadenosine (Scheme 2).
Structural Characterization of dA-1. The high-resolution
ESI-mass spectrum of dA-1 (Figure S2A) showed the base peak
at m/z 222.0978 and a strong signal at m/z 338.1453, which did
not contain a chlorine atom based on the intensity of the
isotopic peak at m/z 340.1503. The data matched a structure
with one molecule of CBO being added to dA with the chlorine
atom being replaced by a hydroxyl group (the calculated
formula weight for such a structure with a formula of
C14H20N5O5 is 338.1464). It was also consistent with the
observed base peak, which was an ion formed after losing the
deoxyribose moiety (the calculated mass for C9H12N5O2 was
222.0991).
Structural Characterization of dA-2. The first clue for
the structure of dA-2 came from the fact that incubation of dA-
2 at pH 7.4 led to the formation of dA-1 as the predominant
product (see below).
1
The H NMR spectrum of dA-1 in DMSO-d6 showed the
presence of 20 protons, although the water peak at 3.42 ppm
D
DOI: 10.1021/acs.chemrestox.6b00095
Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Chemical Research in Toxicology
Article
a
Table 1. NMR Data of the Products Characterized in the Present Study
dA-1
dA-2
A-1D
A-2D
position
1H
8.69 (s, 1H)
13C
1H
13C
1H
13C
1H
13C
2
146.9 8.77 (s, 1H)
146.9 8.60 (s, 1H)
146.0 8.66 (s, 1H)
146.3
148.0
117.6
146.4
143.6
4
146.3
146.5
e
5
118.5
118.3
e
6
146.8
146.6
e
8
8.74 (s, 1H)
142.4 8.80 (s, 1H)
142.7 8.50 (s, 1H)
144.0 8.56 (s, 1H)
1′
2′
6.39 (t, 1H, J = 6.5 Hz)
2.66 (m, 1H)
83.8
39.6
6.40 (t, 1H, J = 6.5 Hz)
2.67 (m, 1H)
83.9
39.5
2.38 (m, 1H)
2.40 (m, 1H)
3′
4′
5′
4.43 (s, 1H)
70.2
88.1
61.1
4.43 (m, 1H)
70.1
88.0
61.1
3.90 (m, 1H)
3.91 (m, 1H)
3.61 (m, 1H)
3.59 (m, 1H)
3.52 (m, 1H)
3.52 (m, 1H)
1″
3.75 (d, 1H, J = 10.5 Hz) 65.3
4.07 (d, 1H, J = 11.5 Hz)
3.94 (dd, 1H, J = 11.5, 2 Hz)
47.9
3.75 (d, 1H, J = 12 Hz) 65.4
4.06 (d, 1H, J = 11 Hz) 48.0
3.94 (d, 1H, J = 11 Hz)
78.3
b
d
3.58 (1H)
3.58 (1H)
2″
3″
79.4
78.4
27.4
e
2.24 (m, 1H)
2.09 (m, 1H)
4.58 (m, 1H)
4.34 (m, 1H)
10.80 (br s, 1H)
6.75 (br s, 1H)
5.49 (br s, 1H)
5.35 (br s, 1H)
5.07 (br s, 1H)
27.5
2.33 (m, 1H)
2.28 (m, 1H)
4.63 (m, 1H)
4.38 (m, 1H)
11.22 (br s, 1H)
7.35 (br s, 1H)
5.47 (br s, 1H)
5.41 (br s, 1H)
2.23 (m, 1H)
2.09 (m, 1H)
4.58 (m, 1H)
4.34 (m, 1H)
c
27.4
2.33 (m, 1H)
2.25 (m, 1H)
4.60 (m, 1H)
4.33 (m, 1H)
11.09 (br s, 1H)
c
27.5
4″
43.5
43.3
43.3
43.2
NH
OH
OH
OH
OH
c
c
c
a
The 13C chemical shifts of the carbon atoms carrying hydrogen were obtained from the HMQC spectra, and those without hydrogen (i.e., C4, C5,
b
C6, and C2″) were from the HMBC spectra. The signal should be a double peak. However, it overlapped with the signal of H5′ at 3.61 ppm;
therefore, its multiplicity could not be identified. The signals were not observed. The signal was expected to be a double peak. However, its
multiplicity was unable to be determined due to overlapping with the peaks of impurities. The data were not available because the HMBC spectrum
c
d
e
of A-1D was not collected.
To obtain the NMR data, a sufficient amount of dA-2 (at
milligram level) needed to be isolated with HPLC. However, it
was found that dA-2 was so labile that most of dA-2
decomposed during isolation; keeping the collection beakers
on ice seemed not to decrease the decomposition.
accounted for 10%, and unreacted dA for 1%). Therefore, CBO
was reacted with dA under these conditions, the reaction
mixture was lyophilized after extraction with diethyl ether, and
the formed residue was dissolved in DMSO-d6 and was
subjected to NMR data collection. Using this approach, the
NMR and ESI-mass spectra were obtained.
Considering that dA-2 was the overwhelmingly predominant
product at pH 4, especially at the early stage of the reaction
(2−3 h) (Figure 2B), the possibility to obtain the spectroscopic
data directly using the reaction mixture was explored. Indeed, as
estimated by peak areas, dA-2 accounted for ∼94% of all
compounds in the reaction mixture at 2 h (the monitoring
wavelength was 260 nm), which included dA-1, dA-2, and
unreacted dA (unreacted CBO was not considered because it
can easily be removed by extraction and lyophilization). Thus,
in our first attempt, the reaction was performed in pH 4
phosphate buffer at 37 °C for 3 h, and the excess of CBO was
removed by extraction with diethyl ether. The obtained
solution was lyophilized, DMSO-d6 was directly added to the
residual solid, and the 1H NMR spectrum was recorded.
However, the spectrum exhibited broad peaks and a poor
baseline, probably due to the presence of inorganic salts, such
as phosphate and KCl. To avoid this problem, we examined the
reaction of CBO with dA in water whose pH was adjusted to 4
with TFA. However, in this situation it was found that a higher
CBO/dA molar ratio was needed to obtain a high yield of dA-2.
Finally, the reaction conditions were optimized to a CBO/dA
molar ratio of 20:1 and incubation time of 6 h at 37 °C. Under
these conditions, 99% of dA was consumed/ and the peak area
of dA-2 accounted for 89% of the total peak areas (dA-1
The high-resolution ESI-mass spectrum of dA-2 exhibited
the base peak at m/z 240.0651 and a strong peak at m/z
356.1125 (Figure S2B) with a peak at m/z 358.1095 having
approximately one-third the intensity, consistent with the
presence of a chlorine atom. The formula weight was consistent
with a structure with a CBO molecule being added to dA (the
calculated formula weight for C14H19N5O4Cl is 356.1126).
Obviously, the base peak was the signal of the daughter ion of
the molecular ion after losing the deoxyribose moiety (the
calculated formula weight for C9H11N5OCl is 240.0652).
1
Although the H NMR spectrum exhibited the presence of
significant amounts of impurities (Figure S6), the proton
signals of dA-2 could be identified with assistance from the
HMQC and correlation spectroscopy (COSY) spectra (Figures
1
S7 and S8) and also the H NMR data of dA-1. Overall, the
chemical shifts of protons in dA-2 were similar to those in dA-1
(Table 1). Similar to dA-1, the HMBC spectrum of dA-2
(Figure S9) exhibited the coupling between H2 and C4″, and
between H4″ and C2, confirming that the vinyl terminus of
CBO had added to N1. H1″, H3″, and H4″ were all observed
to couple to a carbon atom at 78.4 ppm, which was assigned as
C2″. The chemical shift precludes the possibility that C2″ is a
carbonyl carbon and is consistent with addition of the exocyclic
E
DOI: 10.1021/acs.chemrestox.6b00095
Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Chemical Research in Toxicology
Article
Scheme 2. Structures of the Reaction Products of CBO with
dA under in Vitro Physiological Conditions and of the
Decomposition/Acid Hydrolysis Products of the Reaction
a
Products
Figure 3. Decomposition of dA-1 and dA-2 over time at pH 7.4 and
37 °C as monitored by HPLC at 260 nm.
a
The numbering of the positions in the compounds, which was used in
the text and Table 1, was shown in dA-1 and A-1D. On the basis of the
numbering rules, for the CBO moieties in A-1D and A-2D, 1′, 2′, 3′,
and 4′ (rather than 1″, 2″, 3″, and 4″) should be used because the
deoxyribose moieties are already lost. However, the numbering with
double primes (i.e., 1″, 2″, 3″, and 4″) were still used in A-1D and A-
2D to facilitate comparison of data.
Figure 4. Decomposition of dA-1 and dA-2. (A) The HPLC
chromatogram of dA-1 monitored at 260 nm after incubation at pH
7.4 and 37 °C for 9 h. (B) The HPLC chromatogram of dA-2
monitored at 260 nm after incubation at pH 7.4 and 37 °C for 1 h.
amino group. Therefore, CBO reacted with N1 and N6 of dA to
form a fused ring, and dA-2 was characterized as 1,N6-(1-
hydroxy-1-chloromethylpropan-1,3-diyl)-2′-deoxyadenosine
(Scheme 2).
peak at m/z 136.0620, which was consistent with the data of
1
dA. The H NMR spectrum of dA-1B in DMSO-d6 matched
Stabilities of dA-1 and dA-2 under in Vitro Physio-
logical Conditions. At pH 7.4 and 37 °C, dA-1 and dA-2
were both observed to undergo decomposition with dA-2 being
much more labile than dA-1 (Figure 3). The half-lives of dA-1
and dA-2 were determined to be 16.0 0.7 and 0.9 0.1 h,
respectively.
Characterization of the Decomposition Products of
dA-1 and dA-2. Incubation of dA-1 at pH 7.4 and 37 °C led
to the formation of a product with the retention time at 16.9
min and λmax at 257 nm, which was designated as dA-1B
(Figure 4A). dA-1B had the same retention time and λmax as dA
(see Figure 1 for comparison). The product was isolated and
subjected to ESI-MS and NMR analysis.
that of dA in DMSO-d6 well. In addition, dA-1B coeluted with
the dA standard. Furthermore, acid hydrolysis of dA-1B yielded
a product with the retention time at 6.3 min, which coeluted
with the acid hydrolysis product of dA. The HPLC chromato-
gram of the reaction mixture of CBO with dA-1B matched that
of CBO with dA well. Collectively, these results provided clear
evidence that dA-1B was dA.
Incubation of dA-2 at pH 7.4 and 37 °C yielded two
products, which were identified as dA-1 and dA-1B (dA) based
on the coelution experiments with the purified dA-1 and dA
standard (Figure 4B). The decrease in the peak area of dA-2
(monitored at 260 nm) was equal to the combined increases in
the peak areas of dA-1 and dA within the experimental errors
(data not shown), indicating that decomposed dA-2 was
completely converted to dA-1 and dA.
The high-resolution ESI-mass spectrum of dA-1B exhibited
the protonated molecular ion peak at m/z 252.1097 and a base
F
DOI: 10.1021/acs.chemrestox.6b00095
Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Chemical Research in Toxicology
Article
Scheme 3. Proposed Mechanism of Decomposition of dA-1 to dA-1B (i.e., dA) and HMVK under in Vitro Physiological
Conditions
Further Confirmation of the Decomposition Mecha-
nism of dA-1. To explain the decomposition of dA-1 to dA, a
mechanism was proposed (Scheme 3). On the basis of the
mechanism, dA-1 decomposed to dA and HMVK. To provide
evidence for this mechanism, we used GSH to trap HMVK as
the HMVK-GSH conjugate because HMVK is unstable at pH
7.4 and 37 °C.15 dA-1 was incubated with an excess of GSH
(the molar ratio of GSH/dA-1 was about 5:1) at pH 7.4 and 37
°C for 29 h. As expected, a new peak at 10.1 min was observed
in the HPLC chromatogram of the reaction mixture (Figure 5),
which was determined to be the HMVK-GSH conjugate by
matching its elution time and UV absorption spectrum with
those of the HMVK-GSH standard.
the absence of the deoxyribose moiety. The 1H NMR spectrum
showed that purified A-1D still contained impurities (Figure
S10); however, the impurities did not contain aromatic rings
because only two peaks, which appeared at 8.50 and 8.60 ppm
(Table 1), were observed at low field (>5.5 ppm). Apparently,
the two peaks were the signals of the two protons (H2 and H8)
on the purine ring of A-1D. The signals of H3″, which appeared
at 2.09 and 2.23 ppm, were easily identified because their
chemical shifts changed little in comparison with dA-1 (Table
1), and no other signals from 2.0 to 3.0 ppm could be assigned
to H3″. However, it was somewhat difficult to identify the
signals of H1″ and H4″ due to overlapping with the peaks of
impurities (Figure S10). To overcome the problem, the
HMQC spectrum of A-1D was collected (Figure S11). The
signals of H1″ and H4″ were assigned on the basis of the 13C
chemical shifts of the corresponding carbon atoms, which
changed little in comparison with dA-1 (for example, the 13C
chemical shift of C4″ in A-1D was 43.3 ppm, and this value in
dA-1 was 43.5 ppm, indicating that the loss of the deoxyribose
moiety virtually had no effect on the 13C chemical shifts of the
carbon atoms in the CBO moiety; Table 1). In fact, the final
results showed that, compared to its precursor dA-1, the signals
of the protons in the CBO moiety in A-1D (i.e., H1″, H3″, and
H4″) did not change in terms of chemical shifts and peak
splitting. Thus, A-1D was characterized as 1,N6-(1-hydroxy-1-
hydroxymethylpropan-1,3-diyl)adenine (Scheme 2).
Hydrolysis of dA-2 in 0.1 M HCl at 100 °C for 1 h gave one
major product with a retention time of 15.9 min and λmax at 267
nm (see Figure S1 for the UV absorption spectrum), which was
designated as A-2D. In the high-resolution ESI-mass spectrum
(Figure S2D), its protonated molecular ion appeared at m/z
240.0653 and contained a chlorine atom, which was consistent
with the expected structure of a deribosylated product (the
calculated formula weight of protonated molecular ion was
240.0652). The 1H NMR spectrum indicated the absence of the
deoxyribose moiety (Figure S12). Obviously, like A-1D, it was
the hydrolysis product after dA-2 lost the deoxyribose moiety.
The signals of the two protons on the purine ring (H2 and H8)
appeared at 8.56 and 8.66 ppm (Table 1, Figure S12). The
HMBC spectrum (Figure S13) indicated that the proton at
8.66 ppm coupled to carbon atoms at 43.2, 117.6, and 146.4
ppm, whereas the proton at 8.56 ppm coupled to carbon atoms
at 117.6 and 148.0 ppm. Therefore, apparently the proton at
8.66 ppm was H2 (the 13C chemical shift of C2 was 146.3 ppm
based on the HMQC spectrum) because the carbon atom at
43.2 ppm was one at the CBO moiety (see below). The
HMQC spectrum (Figure S14) indicated that the signals at
4.60 and 4.33 ppm, 4.06 and 3.94 ppm, and 2.33 and 2.25 ppm
Figure 5. HPLC chromatogram of the reaction mixture of dA-1 with
GSH after incubation at pH 7.4 and 37 °C for 29 h to confirm the
formation of HMVK during the incubation of dA-1 (monitoring
wavelength: 210 nm). The retention times of GSH and oxidized GSH
(GSSG) were 6.5 and 10.5/10.7 (a doublet peak), respectively.
Characterization of the Acid Hydrolysis Products of
dA-1 and dA-2. To be used as potential biomarkers, the acid
hydrolysis products of dA-1 and dA-2 were prepared and
characterized. Hydrolysis of dA-1 in 0.1 M HCl at 100 °C for 1
h yielded a predominant product (which was designated as A-
1D) with a retention time of 4.5 min and λmax at 265 nm (see
Figure S1 for the UV absorption spectrum).
The formula weight of A-1D was 222.0992 (contained no
chlorine atom) as determined by its high-resolution ESI-mass
spectrum (Figure S2C), which was consistent with the expected
formula C9H12N5O2 (the calculated formula weight is
222.0991). Thus, A-1D was the hydrolysis product of dA-1
following the loss of the deoxyribose moiety. This was further
confirmed by the 1H NMR spectrum of A-1D, which indicated
G
DOI: 10.1021/acs.chemrestox.6b00095
Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Chemical Research in Toxicology
Article
were those of three methylene groups with the corresponding
13C chemical shifts at 43.2, 48.0, and 27.5 ppm, respectively.
The signals at 4.60 and 4.33 ppm were identified as those of
H4″ because the HMBC spectrum indicated that the two
signals coupled to a carbon atom at 146.3 ppm, i.e., C2. The
COSY spectrum (Figure S15) indicated that the signals at 4.60
and 4.33 ppm coupled with those at 2.33 and 2.25 ppm, which
were thus assigned as the signals of H3″. Thus, the peaks at
4.06 and 3.94 ppm were the signals of H1″ (Table 1). In
addition, A-2D was converted to A-1D when incubated at pH
7.4 and 37 °C. Therefore, A-2D was characterized as 1,N6-(1-
hydroxy-1-chloromethylpropan-1,3-diyl)adenine (Scheme 2).
Detection of A-2D from the Reaction of CBO with CT
DNA under in Vitro Physiological Conditions. To examine
if CBO could react with dA residues in DNA to generate
products corresponding to dA-1 and dA-2, CT DNA (2 mg/
mL) was incubated with CBO (1.7 mM) at pH 7.4 and 37 °C
for 9 h and then was hydrolyzed in 0.1 M HCl at 75 °C for 20
min to release purines. The resulting solution was analyzed by
HPLC. On the basis of identical retention times (confirmed by
coelution experiments with the standards) and similar UV
absorption spectra to those of the standards, A-1D and A-2D
were identified (Figure 6). It should be noted that the amount
Figure 7. LC-MS chromatogram of acid-hydrolyzed cellular DNA
under selected ion monitoring mode (the mass range was from
240.0649 to 240.0655). Cellular DNA was extracted from HepG2 cells
incubated with 100 μM BSO at 37 °C for 24 h and then 1 μM CBO
for 1 h.
cells to form dA adducts, which predominantly consisted of dA-
2, i.e., dA-2 was stable in DNA and retained its chlorine atom.
DISCUSSION
■
CBO is a potential reactive metabolite of BD. We recently
demonstrated that it was genotoxic in cultured human liver
cells.11 CBO induced the formation of alkali-labile sites on
DNA and was capable of directly generating DNA breaks as
examined by the comet assay.11 Because CBO is expected to be
reactive toward DNA, we characterized the reaction products of
CBO with dC and dG under in vitro physiological
conditions.12,13 In the present study, we report on the reactivity
of CBO with dA and the chemical characterization of products,
and their degradation and acid-hydrolyzed products. We also
provide evidence for the formation of a fused-ring dA adduct in
CT DNA and human cells incubated with CBO.
The reaction of CBO with dA (Scheme 2) is similar to that
of CBO with dC described previously. Both reactions
proceeded through initial Michael addition of CBO to
nucleoside ring nitrogen atoms and subsequent cyclization of
the resulting intermediates via reaction of the exocyclic amino
groups with the carbonyl groups of the CBO moieties, forming
fused-ring adducts. These adducts underwent similar hydrolysis
of the chloromethyl groups to yield the corresponding
hydroxylated products (Scheme 2). A major difference between
the two reaction profiles was, however, that there existed a
minor product in the reaction of CBO with dC, which was
formed through an initial Michael addition of CBO with the
exocyclic amino group of dC. The corresponding product was
not detected in the reaction of CBO with dA. This may be
attributed to multiple factors, including the relative nucleophilic
strength of the exocyclic amino groups in comparison with the
ring nitrogen atoms, “hardness/softness” of the electrophiles
and nucleophiles,16 and kinetic or thermodynamic control of
the reactions.17,18
It was interesting to observe the decomposition of dA-1 to
produce dA under the used incubation conditions. To explain
this decomposition, a mechanism (Scheme 3), in which dA-1 is
present in equilibrium with 1, a carbonyl-containing inter-
mediate, and 1 undergoes a retro-Michael reaction to produce
dA and HMVK, is proposed. The formation of HMVK was
confirmed through its trapping with GSH, providing evidence
for this mechanism. Similarly, the observed rapid equilibrium of
the two diastereomers of the CBO-dC adduct with 3,N4-fused
Figure 6. HPLC chromatogram of the reaction mixture of CBO (1.7
mM) with CT DNA after incubation at pH 7.4 and 37 °C for 9 h, and
acid-hydrolysis (monitoring wavelength: 260 nm). The peaks of A-1D
and A-2D were indicated by arrows. Inset: the magnified chromato-
gram to clearly show the peaks of A-1D and A-2D.
of A-2D was significantly larger than that of A-1D (the peak
area of A-2D was 3-fold as large as that of A-1D). In addition,
no guanine adducts were detected. The results provide strong
evidence that CBO reacted with the dA residues in DNA to
form products corresponding to dA-1 and dA-2.
Detection of A-2D in Cells Incubated with CBO. To
examine if CBO could cause the formation of dA adducts in
living cells, GSH-depleted HepG2 cells were incubated with 1
μM CBO at 37 °C for 1 h. DNA was extracted, acid-
hydrolyzed, and then subjected to LC-MS analysis. Indeed, as
shown in Figure 7, a peak with the elution time at 2.26 min was
observed under selected ion monitoring mode (the mass range
was set at 240.0649−240.0655). The compound that the peak
represents exhibited a signal at m/z 240.0650 with an isotope
one at m/z 242.0609 (∼1/3 intensity) and had the same
retention time as the A-2D standard. Apparently, this
compound was A-2D. However, A-1D was not detected.
Moreover, no guanine adducts were detected. Therefore, the
results demonstrated that CBO could react with DNA in living
H
DOI: 10.1021/acs.chemrestox.6b00095
Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Chemical Research in Toxicology
Article
rings and hydroxyl groups in place of the chlorine atoms
suggested the formation of a carbonyl-containing intermedi-
ate.12 While dA-2 could also undergo retro-Michael reaction to
regenerate dA and CBO, we were unable to obtain evidence for
this hypothesis, possibly because of the rapid decomposition of
dA-2 to yield dA-1.
*(A.A.E.) University of WisconsinMadison, Department of
Comparative Biosciences, 2015 Linden Dr., Madison, WI
53706, United States. Tel: + 1-608-262-6518. Fax: + 1-608-262-
3926. E-mail: adnan.elfarra@wisc.edu.
Author Contributions
∥F.-M.Z. and L.-Y.L. contributed to the work equally.
CBO can form fused-ring adducts with dA, dC, and dG.
Similarly, other α,β-unsaturated aldehydes/ketones, such as
acrolein,19 4-hydroxy-2-nonenal and crotonaldehyde,20 4-oxo-2-
nonenal,21 and lipid peroxidation products,22 also form fused-
ring adducts with the N1- and N6-positions of dA. Bifunctional
alkylating agents, such as 1,2,3,4-diepoxybutane, can also lead to
the formation of dA adducts with fused-rings.23 These adducts
could be mutagenic because they have potential to significantly
disrupt the Watson−Crick base pairing in DNA.24,25 Levine et
al. demonstrated that 1,N6-ethenodeoxyadenosine, the simplest
dA adduct with a fused-ring, was substantially mutagenic in
human cells.26 The formation of 1,N6-ethenodeoxyadenosine
and 3,N4-ethenodeoxycytidine during stage-specific tumor
development in mouse skin has been demonstrated.27 1,N6-
Ethenodeoxyadenosine was detectable in human urine and its
levels were associated with lipid peroxidation-derived DNA
damage in vivo, and these adducts have been implicated in the
development of hepatocellular carcinoma.28,29 Therefore, the
CBO-dA adducts could also be mutagenic and play an
important role in the genotoxicity of CHB and BD. In addition,
A-2D may be used as a biomarker of CBO formation in human
cells.
Funding
This work was supported by the National Nature Science
Foundation of China (Nos. 21077070 and 21377080), the
Shanghai Municipal Education Commission (No. 11ZZ90), the
Shanghai Leading Academic Discipline Project (No. S30109),
and the National Institutes of Health of the U.S.A. (R01
DK044295 and R01 ES06841).
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The NMR spectra of all samples were recorded on the Bruker
Avance AV500 spectrometer in the Instrumental Analysis &
Research Center of Shanghai University. We are grateful to Dr.
Hong-Mei Deng of the Nuclear Magnetic Resonance
Laboratory in this center for her assistance with the collection
of the NMR spectra. We also thank Dr. Ying Wen and
Professor Tao Yi of the Chemistry Department in Fudan
University for their assistance in submitting samples for ESI-MS
and LC-MS analysis, and Mr. Gang-Feng Tang in this
department for the ESI-MS and LC-MS analysis of samples.
It was somewhat surprising that the amount of A-2D
obtained in the CT DNA experiment was significantly greater
than that of A-1D. The result suggested that, unlike the adduct
at the free nucleoside level, hydrolysis of dA-2 at the DNA level
was largely inhibited, probably due to the stacking of
nucleobases in DNA. In addition, the fact that no guanine
adducts were detected suggested that CBO predominantly
reacted with the dA residues in DNA.
The result obtained with HepG2 cells was qualitatively
consistent with that obtained with CT DNA, supporting the
observation that dA-2 was stable in DNA. A-1D was detected
in CT DNA but not in HepG2 cells, probably due to the large
difference in the CBO concentrations used in these experi-
ments (1.7 mM in the DNA experiment vs 1 μM in the cell
one). In addition, no guanine adducts were detected in both
experiments, suggesting that, in terms of purines, CBO
predominantly reacted with adenine in DNA. The basis for
the selectivity of CBO toward dA residues in DNA is presently
unclear. More studies with multiple time points and different
CBO concentrations are being conducted to clarify the basis for
the observed CBO selectivity toward dA residues in DNA.
ABBREVIATIONS
■
ADH, alcohol dehydrogenase; BD, 1,3-butadiene; BSO, L-
buthionine-sulfoximine; CBO, 1-chloro-3-buten-2-one; CHB,
1-chloro-2-hydroxy-3-butene; COSY, correlation spectroscopy;
CT DNA, calf thymus DNA; dA, 2′-deoxyadenosine; dC, 2′-
deoxycytidine; DMEM, Dulbecco’s modified Eagle’s medium;
DMSO, dimethyl sulfoxide; ESI, electrospray ionization; FBS,
fetal bovine serum; GSH, glutathione; HMBC, heteronuclear
multiple-bond correlation; HMQC, heteronuclear multiple-
quantum correlation; HMVK, hydroxymethyl vinyl ketone;
MPO, myeloperoxidase; TFA, trifluoroacetic acid; Tris, tris-
(hydroxymethyl)aminomethane
REFERENCES
■
(1) Koppikar, A. M., Wu, C., Leavens, T., Valcovic, L., Kimmel, C.,
Faust, R., and Jinot, J. (2002) Health Assessment of 1,3-Butadiene,
EPA/600/P-98/001F, U.S. Environmental Protection Agency, Office
of Research and Development, National Center for Environmental
Assessment-Washington Office, U.S. Government Printing Office,
Washington, DC, http://cfpub.epa.gov/ncea/cfm/recordisplay.
cfm?deid=54499 (accessed Mar 19, 2016).
(2) Duescher, R. J., and Elfarra, A. A. (1992) 1,3-Butadiene oxidation
by human myeloperoxidase. J. Biol. Chem. 267, 19859−19865.
(3) Elfarra, A. A., and Zhang, X.-Y. (2012) Alcohol dehydrogenase-
and rat liver cytosol-dependent bioactivation of 1-chloro-2-hydroxy-3-
butene to 1-chloro-3-buten-2-one, a bifunctional alkylating agent.
Chem. Res. Toxicol. 25, 2600−2607.
(4) Loh, M. M., Levy, J. I., Spengler, J. D., Houseman, E. A., and
Bennett, D. H. (2007) Ranking cancer risks of organic hazardous air
pollutants in the United States. Environ. Health Perspect. 115, 1160−
1168.
(5) McCarthy, M. C., O’Brien, T. E., Charrier, J. G., and Hafner, H.
R. (2009) Characterization of the chronic risk and hazard of hazardous
air pollutants in the United States using ambient monitoring data.
Environ. Health Perspect. 117, 790−796.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.chemres-
tox.6b00095.
UV absorption, NMR, and high-resolution ESI-mass
spectra of dA-1, dA-2, A-1D, and A-2D (PDF)
AUTHOR INFORMATION
Corresponding Authors
■
*(X.-Y.Z.) Tel: + 86-21-6613-7736. Fax: + 86-21-6613-6928. E-
mail: xyzhang999@shu.edu.cn.
I
DOI: 10.1021/acs.chemrestox.6b00095
Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Chemical Research in Toxicology
Article
(6) Zhou, J., You, Y., Bai, Z., Hu, Y., Zhang, J., and Zhang, N. (2011)
Health risk assessment of personal inhalation exposure to volatile
organic compounds in Tianjin, China. Sci. Total Environ. 409, 452−
459.
(7) Logue, J. M., McKone, T. E., Sherman, M. H., and Singer, B. C.
(2011) Hazard assessment of chemical air contaminants measured in
residences. Indoor Air 21, 92−109.
(8) Fowles, J., and Dybing, E. (2003) Application of toxicological risk
assessment principles to the chemical constituents of cigarette smoke.
Tob. Control 12, 424−430.
(9) Kirman, C. R., Albertini, R. J., Sweeney, L. M., and Gargas, M. L.
(2010) 1,3-Butadiene: I. Review of metabolism and the implications to
human health risk assessment. Crit. Rev. Toxicol. 40 (Suppl. 1), 1−11.
(10) Albertini, R. J., Carson, M. L., Kirman, C. R., and Gargas, M. L.
(2010) 1,3-Butadiene: II. Genotoxicity profile. Crit. Rev. Toxicol. 40
(Suppl. 1), 12−73.
(11) Liu, X.-J., Zeng, F.-M., An, J., Yu, Y.-X., Zhang, X.-Y., and Elfarra,
A. A. (2013) Cytotoxicity, genotoxicity, and mutagenicity of 1-chloro-
2-hydroxy-3-butene and 1-chloro-3-buten-2-one, two alternative
metabolites of 1,3-butadiene. Toxicol. Appl. Pharmacol. 271, 13−19.
(12) Sun, L., Pelah, A., Zhang, D.-P., Zhong, Y.-F., An, J., Yu, Y.-X.,
Zhang, X.-Y., and Elfarra, A. A. (2013) Formation of fused-ring 2′-
deoxycytidine adducts from 1-chloro-3-buten-2-one, an in vitro 1,3-
butadiene metabolite, under in vitro physiological conditions. Chem.
Res. Toxicol. 26, 1545−1553.
(13) Zheng, J., Li, Y., Yu, Y.-X., An, J., Zhang, X.-Y., and Elfarra, A. A.
(2015) Novel adducts from the reaction of 1-chloro-3-buten-2-one
with 2′-deoxyguanosine. Structural characterization and potential as
tools to investigate 1,3-butadiene carcinogenicity. Chem.-Biol. Interact.
226, 40−48.
(2015) Polymerase bypass of N6-deoxyadenosine adducts derived from
epoxide metabolites of 1,3-butadiene. Chem. Res. Toxicol. 28, 1496−
1507.
(25) Kotapati, S., Maddukuri, L., Wickramaratne, S., Seneviratne, U.,
Goggin, M., Pence, M. G., Villalta, P., Guengerich, F. P., Marnett, L.,
and Tretyakova, N. (2012) Translesion synthesis across 1,N6-(2-
hydroxy-3-hydroxymethylpropan-1,3-diyl)-2′-deoxyadenosine (1,N6-γ-
HMHP-dA) adducts by human and archebacterial DNA polymerases.
J. Biol. Chem. 287, 38800−38811.
(26) Levine, R. L., Yang, I.-Y., Hossain, M., Pandya, G. A., Grollman,
A. P., and Moriya, M. (2000) Mutagenesis induced by a single 1,N6-
ethenodeoxyadenosine adduct in human cells. Cancer Res. 60, 4098−
4104.
(27) Nair, J., Furstenberger, G., Burger, F., Marks, F., and Bartsch, H.
(2000) Promutagenic etheno-DNA adducts in multistage mouse skin
carcinogenesis: correlation with lipoxygenase-catalyzed arachidonic
acid metabolism. Chem. Res. Toxicol. 13, 703−709.
(28) Nair, J., Srivatanakul, P., Haas, C., Jedpiyawongse, A.,
Khuhaprema, T., Seitz, H. K., and Bartsch, H. (2010) High urinary
excretion of lipid peroxidation-derived DNA damage in patients with
cancer-prone liver diseases. Mutat. Res., Fundam. Mol. Mech. Mutagen.
683, 23−28.
(29) Bartsch, H., and Nair, J. (2005) Accumulation of lipid
peroxidation-derived DNA lesions: potential lead markers for
chemoprevention of inflammation-driven malignancies. Mutat. Res.,
Fundam. Mol. Mech. Mutagen. 591, 34−44.
(14) Reppe, W., and Mitarbeiter (1955) Ethynylation. IV. Reactions
of α-alkynols and γ-alkynediols. Justus Liebigs Ann. Chem. 596, 38−79.
(15) Krause, R. J., Kemper, R. A., and Elfarra, A. A. (2001)
Hydroxymethylvinyl ketone: a reactive Michael acceptor formed by
the oxidation of 3-butene-1,2-diol by cDNA-expressed human
cytochrome P450s and mouse, rat, and human liver microsomes.
Chem. Res. Toxicol. 14, 1590−1595.
(16) LoPachin, R. M., Gavin, T., Petersen, D. R., and Barber, D. S.
(2009) Molecular mechanisms of 4-hydroxy-2-nonenal and acrolein
toxicity: nucleophilic targets and adduct formation. Chem. Res. Toxicol.
22, 1499−1508.
(17) Freccero, M., Gandolfi, R., and Sarzi-Amade, M. (2003)
Selectivity of purine alkylation by a quinone methide. Kinetic or
thermodynamic control? J. Org. Chem. 68, 6411−6423.
(18) Veldhuyzen, W. F., Shallop, A. J., Jones, R. A., and Rokita, S. E.
(2001) Thermodynamic versus kinetic products of DNA alkylation as
modeled by reaction of deoxyadenosine. J. Am. Chem. Soc. 123,
11126−11132.
(19) Pawlowicz, A. J., Munter, T., Zhao, Y., and Kronberg, L. (2006)
Formation of acrolein adducts with 2′-deoxyadenosine in calf thymus
DNA. Chem. Res. Toxicol. 19, 571−576.
(20) Chen, H.-J. C., and Chung, F.-L. (1994) Formation of etheno
adducts in reactions of enals via autoxidation. Chem. Res. Toxicol. 7,
857−860.
(21) Rindgen, D., Lee, S. H., Nakajima, M., and Blair, I. A. (2000)
Formation of a substituted 1,N6-etheno-2′-deoxyadenosine adduct by
lipid hydroperoxide-mediated generation of 4-oxo-2-nonenal. Chem.
Res. Toxicol. 13, 846−852.
(22) el Ghissassi, F., Barbin, A., Nair, J., and Bartsch, H. (1995)
Formation of 1,N6-ethenoadenine and 3,N4-ethenocytosine by lipid
peroxidation products and nucleic acid bases. Chem. Res. Toxicol. 8,
278−283.
(23) Seneviratne, U., Antsypovich, S., Goggin, M., Dorr, D. Q., Guza,
R., Moser, A., Thompson, C., York, D. M., and Tretyakova, N. (2010)
Exocyclic deoxyadenosine adducts of 1,2,3,4-diepoxybutane: synthesis,
structural elucidation, and mechanistic studies. Chem. Res. Toxicol. 23,
118−133.
(24) Kotapati, S., Wickramaratne, S., Esades, A., Boldry, E. J., Quirk
Dorr, D., Pence, M. G., Guengerich, F. P., and Tretyakova, N. Y.
J
DOI: 10.1021/acs.chemrestox.6b00095
Chem. Res. Toxicol. XXXX, XXX, XXX−XXX