An ammonium bicarbonate-enhanced stable isotope dilution UHPLC-MS/MS method for sensitive and accurate quantification of acrolein-DNA adducts in human leukocytes

Article abstract of DOI:10.1021/ac3034695

Acrolein (Acr), a ubiquitous environmental pollutant, can react directly with genomic DNA to form mutagenic adducts without undergoing metabolic activation. To sensitively and accurately quantify Acr-DNA adducts (including structural isomers and stereoiso

Full text of DOI:10.1021/ac3034695

Article
pubs.acs.org/ac
An Ammonium Bicarbonate-Enhanced Stable Isotope Dilution
UHPLC-MS/MS Method for Sensitive and Accurate Quantification of
Acrolein−DNA Adducts in Human Leukocytes
Ruichuan Yin,† Shengquan Liu,† Chao Zhao,† Meiling Lu,†,‡ Moon-shong Tang,§ and Hailin Wang*,†
†State Key Laboratory of Environmental Chemistry and Ecotoxicology, Research Center for Eco-Environmental Sciences, Chinese
Academy of Sciences, Beijing, 100085, China
‡Chemical Analysis Group, Agilent Technologies, Beijing, 100102, China
§Department of Environmental Medicine, Pathology and Medicine, New York University School of Medicine, Tuxedo Park, New
York 10987, United States
ABSTRACT: Acrolein (Acr), a ubiquitous environmental
pollutant, can react directly with genomic DNA to form
mutagenic adducts without undergoing metabolic activation. To
sensitively and accurately quantify Acr−DNA adducts (includ-
ing structural isomers and stereoisomers) in human leukocytes,
we developed an enhanced stable isotope dilution ultrahigh
performance liquid chromatography (UHPLC)−tandem mass
spectrometry (MS/MS) method using ammonium bicarbonate
(NH4HCO3), which is thermally unstable and degrades readily
to carbon dioxide and ammonia in heated gas phase.
Interestingly, ammonium bicarbonate (as an additive to the
mobile phase) not only improves the protonation of AcrdG adducts but also suppresses the formation of MS signal-deteriorating
metal−AcrdG complexes during electrospray ionization, leading to the enhancement of their MS detection by 2.3−8.7 times. In
contrast, routinely used ammonium salts (ammonium acetate and ammonium formate) and formic acid do not show similar
enhancement. The developed method is potentially useful for enhancing ESI-MS detection of other modified 2′-
deoxyribonucleosides that have difficulty in protonation and may form excess metal complexes during electrospray ionization.
The limits of detection (LODs, S/N = 3) are estimated to be about 40−80 amol. By the use of the developed method, we found
that the Acr adducts of three nucleotides (dG, dA, and dC) can be detected in human leukocytes. In addition to the known γ-
AcrdG, α-AcrdA is also identified as an Acr-adduct of high abundance (2.5−20 adducts per108 nts).
crolein (Acr) is one of the most reactive and harmful α,β-
AcrdG).15 Meanwhile, α-AcrdG has two mutually convertible
stereoisomers. Both α- and γ-AcrdG are mutagenic in
bacteria16,17 and mammalian cells.18,19 In human cells, AcrdG
mainly induces mutations of G:C to T:A or A:T.18,19 In an early
report, γ-AcrdG could be detected in human tissues without
Acr treatment, but α-AcrdG was not detectable in most
samples.20,21 The levels of γ-AcrdG in smokers are significantly
higher than that in nonsmokers.22 Recently, both α-AcrdG and
γ-AcrdG could be detected in human placenta, leukocytes, and
lung tissues by LC-MS/MS analysis.23,24 Inconsistently, the
levels of α-AcrdG in human lung were found to be even higher
than that of γ-AcrdG.24 What factors contributed to such
inconsistency is not clear. Nonetheless, an advanced analytical
method for accurate and sensitive quantitation of Acr−DNA
adducts is highly desirable, particularly for the minor but highly
mutagenic α-AcrdG.
A
unsaturated aldehydes and is found in cigarette smoke,1,2
the volatiles generated during cooking,3 and automobile
exhaust.4 Acr has been proposed as the major carcinogenic
agent responsible for cigarette-related lung cancer, replacing the
well-known carcinogen benzo(α)pyrene.5 Moreover, Acr can
be endogenously produced during peroxidation of polyunsatu-
rated fatty acids6,7 and degradation of amino acids8 and
polyamines.9 Elevated levels of Acr were found in vulnerable
brain regions of subjects with wild cognitive impairment10 and
preclinical11 and late-stage Alzheimer’s disease (AD).12,13
Endogenously produced Acr may play an important role in
the pathogenesis of AD10−13 and cardiovascular diseases.14
Currently, it is not clear how Acr causes these pathogenic
effects. Because Acr can react directly with the bases of genomic
DNA due to its strong electrophilicity, the formation of stable
Acr−DNA adducts may play essential roles in the pathogenesis
of Acr.
The reaction of Acr with guanine in genomic DNA can
produce two structural isomers of exocyclic adducts (Figure 1):
6-hydroxy-1,N2-propano-2′-deoxyguanosine (α-AcrdG) and 8-
hydroxy-1,N2-propano-2′-deoxyguanosine adducts (γ-
Received: November 29, 2012
Accepted: February 22, 2013
Published: February 22, 2013
© 2013 American Chemical Society
3190
dx.doi.org/10.1021/ac3034695 | Anal. Chem. 2013, 85, 3190−3197

Analytical Chemistry
Article
ammonium bicarbonate were obtained from Sigma-Aldrich Co.
(St. Louis, MO). Ammonium formate and ammonium acetate
were purchased from Agilent (Palo Alto, CA) and Signopharm
Chemical Reagent Co. (Beijing, China), respectively. Deoxy-
ribonuclease I (DNase I) and calf intestinal alkaline
phosphatase (CIP) were obtained from New England Biolabs
(Ipswich, MA).
Synthesis of Stable Isotopic Standards of Acr−DNA
Adducts. Each Acr−DNA standard was prepared by the direct
reaction of Acr with the corresponding 2′-deoxyribonucleoside
(with or without stable isotopic label). In brief, dG, dA, or dC
was separately incubated with an equal mole equivalent of Acr
in 50 mM phosphate buffer, pH 7.0 at 37 °C for 48h, and the
products were purified by reversed-phase HPLC. The collected
fractions were dried by lyophilization and then were weighed
with a Mettler Toledo XS 105 Analytical Balance (Greifensee,
Switzerland). All the Acr−DNA standards were characterized
by ESI-MS/MS and quantitated by UV at their maximum
absorption wavelength (257 nm for AcrdG; 261 nm for α-
AcrdA; 281 nm for α-AcrdC).
UHPLC-ESI-MS/MS Analysis. The UHPLC-MS/MS anal-
ysis was performed on Agilent 1290 UHPLC system coupled
with a G6410B triple quadrupole mass spectrometer (Agilent
Technologies, Palo Alto, CA). Zorbax Eclipse Plus C18 column
(100 mm × 2.1 mm i.d., 1.8 μm particle size, Agilent
Technologies, Palo Alto, CA) was used for separation of
modified 2′-deoxyribonucleoside. The optimized mobile phase
consisted of two solvents: 10 mM NH4HCO3, pH 9.0 (solvent
A), and pure methanol (solvent B). An optimized gradient
elution was used for UHPLC separation: 0−6 min, 5% B; 6−10
min, 15% B; 10−15 min, 5% B. The flow-rate was 0.25 mL/
min, and the column temperature was set at 30 °C. We also
tested formic acid (0.1%, 0.01%, and 0.001%) and three volatile
ammonium buffers (0.5−10 mM) as an additive to the mobile
phase.
The mass spectrometer was operated in the positive ion
mode. A multiple reaction monitoring (MRM) mode was
adopted for selective detection of the Acr−DNA adducts: m/z
324 → 208 for AcrdG (collision energy, 5 eV); m/z 308 → 192
for α-AcrdA (15 eV), and m/z 284 → 168 for α-AcrdC (5 eV);
m/z 329 → 213 for [15N5] AcrdG (5 eV), m/z 313 → 179 for
[15N5] α-AcrdA (35 eV), and m/z 287 → 171 for [15N3] α-
AcrdC (5 eV). Nitrogen was used for nebulization and
desolvation. The nebulization gas was set at 40 psi, the flow-
rate of desolvation gas was 9 L/min, and source temperature
was set at 300 °C. Capillary voltage was set at 3500 V. High
purity nitrogen (99.999%) was used as the collision gas. Each
sample was at least analyzed for three times with an injection
volume of 15 μL. The adduct frequency was calibrated
internally by the stable isotopic standards.
Figure 1. Chemical structures of Acr−DNA adducts studied in this
work.
In addition to AcrdG, the adducts of 2′-deoxyadenosine,25−29
2′-deoxycytidine,30,31 and thymidine32 were also identified,
including 9-hydroxy-1,N6-propano-2′-deoxyadenosine (α-
AcrdA)26−28 and 9-hydroxy-3, N4-propano-2′-deoxycytidine
(α-AcrdC).31 Although these non-dG adducts have been
identified in naked DNA or cultured cells treated by Acr, it is
unclear whether they are present in human tissues and their fate
and genotoxicity are also unknown.
Several approaches for the detection of AcrdG have been
developed, including 32P-postlabeling,20−22 immunoassays,33,34
and LC-MS/MS.13,23,24 Due to its advantages in specificity,
structural characterization, separation efficiency, and safety, LC-
MS/MS is of choice. By taking advantage of nano-LC and
nanospray-enhanced MS detection, a highly sensitive method
for simultaneous detection of AcrdG adducts was developed.23
However, one stereoisomer of α-AcrdG coeluted with γ-AcrdG.
The same problem also faced in the capillary LC-MS/MS
method.13 On the other hand, without the application of
nanoflow and nanospray technologies that are not widely
applicable, highly sensitive detection has not been achieved yet.
Moreover, no assay was reported to detect AcrdA and dC
adducts in human tissues.
The objectives of this study are (1) to improve the ESI-MS
detection of Acr−DNA adducts by enhancing protonation, (2)
to improve the separation of Acr adducts and their structural
isomers and stereoisomers, and (3) to accurately quantify α-
AcrdG, γ-AcrdG, α-AcrdA, and α-AcrdC in human tissues by
the use of stable isotope as internal standard. Therefore, we
synthesized the stable isotope standards of Acr−DNA adducts,
investigated the ESI-MS detection and UHPLC separation of
DNA adducts, and tested the abundance of Acr−dG and non-
dG adducts in human leukocytes.
The separation resolution (Rs) was calculated following
Chinese National Standards (GB4946-85) as below.
Rs = 2 × (tR,2 − tR,1)/(tw,1 + tw,2
)
(1)
where tR,1 and tR,2 are the respective retention time of peaks 1
and 2, and tw,1 and tw,2 are the respective peak widths of peaks 1
and 2.
Calibration Curves and Recovery. Calibration curves
were constructed using the prepared standards of Acr adducts.
The stable isotopic standards of a constant amount were mixed
with nonisotope Acr adducts of varying amounts, and then the
mixtures were analyzed by UHPLC-MS/MS. Calibration curves
were obtained by linearly plotting the peak area ratios of
EXPERIMENTAL SECTION
Caution: Acr is highly mutagenic and should be handled with
caution.
■
Chemicals and Materials. Acr was ordered from Aladdin
(Shanghai, China). 2′-Deoxyguanosine monohydrate
(dG·H2O), 2′-deoxyadenosine monohydrate (dA·H2O), 2′-
deoxycytidine (dC), snake venom phosphodiesterase I, and
3191
dx.doi.org/10.1021/ac3034695 | Anal. Chem. 2013, 85, 3190−3197

Analytical Chemistry
Article
nonisotope Acr adducts to the stable isotopic standards versus
the amount of the added nonisotopic Acr adducts.
The recovery was tested by measuring the content of adducts
in A549 genomic DNA spiked with the known amounts of Acr
standards (1.5, 4.5, and 15 fmol) and calculated according to
the formula shown below.
recovery% = (the measured level − the background level)
(2)
/added amount × 100%
A549 Cell Culture and DNA Extract. Human lung
epithelial carcinoma A549 cells (5 × 105 cells/dish) were
seeded in 10 cm plastic dishes (Corning, NY) and cultured in
10 mL (per dish) RPMI 1640 growth medium supplemented
with 10% fetal bovine serum (Invitrogen, Carlsbad, CA),
containing 100 U/mL penicillin G and 100 U/mL
streptomycin sulfate. The dishes were put in a humidified 37
°C incubator supplied with 5% CO2. After the A549 cells were
cultured for 48 h, the genomic DNA was extracted with the
Promega Wizard genomic DNA purification kit according to
the manufacturer’s instructions. The isolated DNA was air-
dried and stored at −20 °C prior to enzymatic digestion.
Genomic DNA Isolation from Human Leukocytes.
Human blood samples were kindly provided by five volunteers.
Freshly drawn human blood was stored at 4 °C in the presence
of 10% (v/v) citrate−dextrose solution as anticoagulant. Red
blood cells were lysed by addition of cell lysis solution, and the
leucocyte pellets were obtained by centrifuging the mixture at
12000 rpm for 1 min at room temperature.
Genomic DNAs were extracted from the isolated leucocytes
with the Promega Wizard genomic DNA purification kit. The
extracted genomic DNA was redissolved in 50 μL of water. The
amount of genomic DNA was determined at 260 nm using
Nanodrop 2000 by assuming that 1.0 OD equals 50 μg/mL.
The DNA samples (20 μg each) were mixed with a known
amount of [15N] Acr−DNA and digested with 2.0 units of
DNase I, 4.0 units of CIP, and 0.008 units of snake venom
phosphodiesterase I at 37 °C for 12 h. The total volume was
100 μL. The enzymatic digests were filtered by ultrafiltration
tubes (MW cutoff: 3 KDa, Pall, Port Washington, NY) to
remove the enzymes and then analyzed by UHPLC-ESI-MS/
MS. Note, the DNA samples were stored at −20 °C prior to
enzymatic digestion.
Figure 2. The effects of mobile-phase additives on the UHPLC
separation of and MS detection sensitivity to Acr−dG isomers. (a)
The UHPLC separation of α-AcrdG and γ-AcrdG using HCOOH as
an additive to the mobile phase. Peaks 1 and 2 represent the two
stereoisomers of α-AcrdG, and peak 3 is γ-AcrdG. (b) The separation
of AcrdG isomers was obtained by using one additive (NH4HCO3,
HCOONH4, CH3COONH4, or HCOOH) in the mobile phase as
indicated in the figure. (c) The plots of the peak area of AcrdG
standards (10 nM for each isomer) against the concentration of the
additive ammonium salt (0.5−10 mM) or HCOOH (0.001%−0.10%)
as indicated. The peak areas were estimated as the sum of peaks 1−3.
The mobile phases consisted of two solvents A and B (pure
methanol). Solvent A: 10 mM ammonium salt or 0.001−0.1% (v/v)
HCOOH in water. Gradient elution was used: 0−6 min, 5% B; 6−10
min, 15% B; 10−15 min, 5% B. The other chromatographic conditions
are described in Experimental Section.
RESULTS AND DISCUSSION
■
Ammonium Bicarbonate Improves the Separation of
the Structural Isomers and Stereoisomers of AcrdG. We
first optimized the separation of Acr−dG adducts and their
isomers by a reversed-phase UHPLC. By adding 0.1% formic
acid into the mobile phase, we observed that the peaks of
AcrdG isomers are broadened and cannot be well resolved
(Figure 2a). The two stereoisomers of α-AcrdG are coeluted
with γ-AcrdG. When the concentration of formic acid is
reduced to 0.001%, the two stereoisomers of α-AcrdG are
completely separated, but one stereoisomer of α-AcrdG (with
longer retention) could not be well resolved from γ-AcrdG (R3,2
< 0.8). For all the tested concentrations of formic acid, the peak
of γ-AcrdG is broadened (Figure 2a). The separation of α-
AcrdG and γ-AcrdG can be markedly improved with any tested
ammonium salt (formate, acetate, or bicarbonate salt) (Figure
2b). The optimum separation of α-AcrdG and γ-AcrdG can be
obtained with ammonium bicarbonate (R3,2 = 1.7). However,
ammonium acetate (R3,2 = 1.4) and ammonium formate (R3,2
=
1.2) can provide satisfactory separation.
Ammonium Bicarbonate Enhances the MS Detection
of AcrdG Isomers. By including ammonium bicarbonate as an
additive of the mobile phase, we found that ammonium
bicarbonate can greatly enhance the MS detection of AcrdG
isomers no matter whether the MS signal is evaluated by peak
height or peak area (Figure 2b). Even involved with high
concentrations of the additive (5−10 mM), ammonium
bicarbonate can enhance the MS detection of AcrdG isomers
(Figure 2c). Compared with ammonium formate and
ammonium acetate in the same concentration range,
ammonium bicarbonate can increase the MS signals of all
AcrdG isomers significantly by 2.3−8.7 times (Figure 2b and
2c). In contrast to formic acid, ammonium bicarbonate also
3192
dx.doi.org/10.1021/ac3034695 | Anal. Chem. 2013, 85, 3190−3197

Analytical Chemistry
Article
increases the MS detection of AcrdG isomers by 5.2 fold (10
mM NH4HCO3 vs 0.1% HCOOH).
To explore the mechanism of the enhancement of MS
detection by ammonium bicarbonate, the UHPLC fractions of
AcrdG are directly scanned from m/z 100 to 380 in the positive
ion mode by ESI-triple quadrupole mass spectrometer. Figure 3
form carbonic acid and thermally degrade to carbon dioxide
and water. This process is expected to favor the protonation of
AcrdG isomers. This mechanism should not be limited to
AcrdG, and we found that a number of modified 2′-
deoxyribonucleosides can benefit from this enhancement
mechanism (unpublished data). Therefore, we provide a
general method for enhancing ESI-MS detection of modified
2′-deoxyribonucleosides that have difficulty in protonation due
to the formation of excess metal complexes during the
ionization process.
Separation and Detection of 8-oxodG and α-AcrdC.
Unexpectedly, oxidatively generated damage 8-oxo-7,8-dihydro-
2′-deoxyguanosine (8-oxodG), which is unavoidably present in
genomic DNA, may have the same chromatographic retention
as α-AcrdC when ammonium bicarbonate is included in the
mobile phase (Figure 4a). Moreover, 8-oxodG and α-AcrdC
have the same molecular weight (283 Da) and same MRM
transition (m/z 284 → 168) for sensitive MS/MS detection.
Therefore, the overlapped peaks of 8-oxodG and α-AcrdC
cannot be indirectly resolved by MS detection without
separation. In light of this problem, we optimized the pH of
the mobile phase without sacrificing the MS detection
sensitivity. This was achieved by titration of the mobile phase
with ammonia. We found that, by increasing the pH to 9.0, the
peak of α-AcrdC will have stronger retention than the two
peaks of 8-oxodG (Figure 4b), rendering satisfactory
separation.
Consistent with previous work,36 two peaks (peaks 1 and 2)
were observed as assigned to 8-oxodG in Figure 4a and 4b. The
first peak of 8-oxodG has the same chromatographic retention
pattern as that of dG regardless of the change in the stationary
and mobile phases of UHPLC (data not shown), assuming it is
the product of the spontaneous oxidation of dG during the ESI
process. The injection of authentic dG also generated this peak,
further supporting the peak assignment. The second peak of 8-
oxodG has the same chromatographic retention as that of
standard 8-oxodG and is assigned to 8-oxodG present in
genomic DNA (data not shown).
Calibration Curves, Limits of Detection (LOD), and
Recovery. The obtained calibration curves are shown in Table
1. Excellent linearity for each Acr−DNA adduct (correlation
coefficient: R2 ≥ 0.998) is achieved with a dynamic range of
0.15−15 fmol. The LODs of these adducts (S/N = 3) are
estimated to be about 40 amol except for α-AcrdG because α-
AcrdG can be split into two peaks and thus has a higher LOD
(80 amol, Table 1).
The recovery is measured by spiking A549 DNA with the
known amounts of standard Acr adducts. By deducting the
corresponding control level, the estimated recovery for the
added adducts ranges from 83 to 92% for α-AcrdA, 84 to 99%
for α-AcrdC, 85 to 95% for α-AcrdG, and 81 to 97% for γ-
AcrdG (Table 2).
Figure 3. The typical ESI-MS spectra of AcrdG isomers were obtained
when different additives to the mobile phase were used. The
concentrations of NH4HCO3, HCOONH4, and CH3COONH4 in
mobile phase A were kept at 10 mM, and HCOOH was kept at 0.1%
(v/v).
shows the formation of [AcrdG + Na]+ and [AcrdG + K]+
complexes by using HCOOH, HCOONH4, or CH3COONH4
as an additive to the mobile phase. By assuming that the
abundance of [AcrdG] is 100%, the metal−AcrdG complexes
account for 22.8−72%. Regarding the instability of these metal
complexes during ESI-MS detection, the observation of metal
complexes at an un-neglected ratio may indicate the formation
of paramount metal complexes during the ESI process, which
will significantly deteriorate the MS detection.35 Interestingly,
the metal−AcrdG complexes significantly reduce down to 3.6%
by using NH4HCO3 as an additive. These data suggest that
ammonium bicarbonate can suppress the formation of MS
signal-deteriorating metal−AcrdG complexes during the ESI
process. Probably, ammonium group of NH4HCO3 may form
ionic complexes with AcrdG, which can be further deammon-
ized to form protonated AcrdGs during the ionization process
presumably in a heated gas phase. Meanwhile, the HCO3− may
Acr−DNA Adducts in A549 Cells. We further applied the
developed UHPLC-MS/MS assay to the quantification of Acr−
DNA adducts in cultured A549 cells. Both α-AcrdG and γ-
AcrdG are detected in A549 cells (Figure 5a), which is not
treated with any Acr. The mean frequency of γ-AcrdG (15
3
adducts per 108 nts) is 4 times higher than that of α-AcrdG (3.5
0.8 adducts in 108 nts) (Table 3). The observed peaks of
AcrdG in A549 DNA (m/z 324 → 208) are further validated by
spiking with the nonisotope standards of α-AcrdG and γ-AcrdG
(Figure 5b). Although it is a minor adduct, α-AcrdG is >10
times more mutagenic than γ-AcrdG.37 Surprisingly, the signal
3193
dx.doi.org/10.1021/ac3034695 | Anal. Chem. 2013, 85, 3190−3197

Analytical Chemistry
Article
Figure 4. The effects of pH of mobile phase A on the separation of α-AcrdC and 8-oxodG. Mobile phase A consisted of 10 mM NH4HCO3, pH 8.0
(a), 10 mM NH4HCO3 adjusted to pH 9.0 by NH3·H2O (b). A 5.0 μL volume of the mixture of 10.0 nM α-AcrdC, 10.0 nM 8-oxodG, and 100 nM
dG were each injected. Gradient elution was used: 0−3 min, 5% B (methanol); 3−8 min, 15% B; 8−15 min, 5% B. The flow rate is 0.25 mL/min,
and the column temperature is set at 30 °C. Peak 1 is 8-oxodG from the spontaneous oxidation of dG in the ESI source, and peak 2 represents 8-
oxodG generated in genomic DNA.
implicating the spontaneous generation of non-dG adducts in
cultured human cells, which are probably attributed to
metabolic processes.
Table 1. Dynamic Range, Linearity, and LOD of Acr−DNA
Adducts by UHPLC-MS/MS Analysis
dynamic range
(fmol)
LOD
analyte
α-
AcrdA
α-
AcrdC
linearity
R2
(amol)
Quantification of Acr−DNA Adducts in Human
Leukocytes. It is known that AcrdG isomers were found in
various human tissues, including lung, liver, brain, placenta, and
leukocytes. However, it is not clear whether the non-dG
adducts discussed above are present in human tissues, although
they are detected in cultured human A549 cells as described
above. So we further characterized the Acr adducts in human
leukocytes by employing the developed UHPLC-MS/MS
method.
The major adduct γ-AcrdG is detectable in all the genomic
DNA samples of extracted human leukocytes (n = 5, Table 3).
The adduct level ranges from 7.5 to 11 adducts per 108 nts for
γ-AcrdG, and the mean level is about 9.0 1.3 adducts per 108
nts. α-AcrdG is detected only in two subjects with a much
lower level than that of γ-AcrdG (ND−3.3 adducts in 108 nts).
These results are consistent with previous reports.23
0.15−15
0.15−15
0.15−15
0.15−15
y = 0.0575x − 0.0066
0.999
40
40
80
40
y = 0.0570x − 0.0117
y = 0.0869x − 0.0217
y = 0.0885x − 0.0156
0.999
0.998
0.999
α-
AcrdG
γ-AcrdG
Table 2. Recovery of the UHPLC-MS/MS Method for
Determination of Acr−DNA Adducts in A549 DNA
α-AcrdA (%) α-AcrdC (%) α-AcrdG (%) γ-AcrdG (%)
1.5 fmol
4.5 fmol
15 fmol
83
85
92
4
2
4
99
84
87
4
2
1
94
95
85
6
7
5
97
86
81
9
2
1
α-AcrdA is detected in all five leukocyte samples with a
frequency of 2.5−20 adducts per108 nts, and the mean adduct
level is 8.7 6.6 adducts per 108 nts. Obviously, the level of α-
AcrdA in leukocytes is comparable to that of γ-AcrdG and is
much higher than that in A549 cells (1.4 0.3 adducts per 108
nts). Another non-dG adduct AcrdC is detected in three
leukocyte samples with a frequency of about ND−16 adducts
per 108 nts.
In this study, we demonstrate that ammonium bicarbonate
can enhance the protonation of AcrdG and suppress the
formation of metal−AcrdG complexes, thereby improving the
MS detection. Due to the enhancement of MS detection, even
without the use of nano-LC and nano-ESI, the LODs (S/N =
3) for detection of Acr adducts by our approach are estimated
about 40−80 amol, which are comparable to that of nano-LC-
MS/MS.23 This is the first report showing the enhancement of
MS detection of modified 2′-deoxyribonucleoside by ammo-
of the second stereoisomer (10.0 min) of α-AcrdG is much
smaller than that of the first one (9.3 min) (bottom trace in
Figure 5a). This is also true for the stereoisomers of isotopic
and nonisotopic standards of the α-AcrdG when spiked with
the digested genomic DNA (top trace in Figure 5a and top and
bottom traces in Figure 5b). The exact reason is not clear, but
we speculated that it is probably due to the ionization
suppression by other unseen fractions that are coeluted with
the second stereoisomer during the ESI process.
Both AcrdC and AcrdA adducts are detected in untreated
A549 cells (Figure 5c−f). The level of α-AcrdC is 3.6 0.3
adducts per 108 nts, while α-AcrdA is the least adduct (1.4
0.3 adducts per 108 nts, Table 3). The two non-dG adducts in
A549 cells (m/z 284 → 168 and 308 → 192) are also validated
by spiking corresponding standard (Figure 5d and 5f). It is clear
that both α-AcrdC and α-AcrdA can be detected in genomic
DNA from cultured human A549 cells without Acr treatment,
3194
dx.doi.org/10.1021/ac3034695 | Anal. Chem. 2013, 85, 3190−3197

Analytical Chemistry
Article
Figure 5. UHPLC-MS/MS chromatograms obtained from analysis of A549 DNA for detection of Acr−DNA adducts. The chromatograms a, c, and
e obtained from the unspiked genomic DNA, and b, d, and f from the genomic DNA spiked with 1.5 fmol standard Acr adducts (as indicated) to
confirm the identity of each Acr−DNA adduct in A549 DNA. The optimized chromatographic conditions were used, and the details are described in
Experimental Section.
3195
dx.doi.org/10.1021/ac3034695 | Anal. Chem. 2013, 85, 3190−3197

Analytical Chemistry
Article
Table 3. Quantitation of Acr−DNA Adducts in Genomic DNA from A549 and Human Leukocytes
Acr−DNA level (adducts per 108 bases)
sample
gender
amount of DNA (μg)
α-AcrdA
α-AcrdC
α-AcrdG
γ-AcrdG
A549
−
male
20
18
13
12
9
1.4 0.3
8.0 0.4
2.5 0.3
8.7 0.3
4.8 0.2
20 0.3
8.7 6.6
3.6 0.3
6.2 0.3
3.5 0.8
1.8 0.7
ND
15 3.4
8.4 0.5
9.3 0.8
11 0.8
8.8 0.6
7.5 0.2
9.0 1.3
1
2
3
4
5
a
female
male
ND
6.8 0.2
ND
ND
female
male
ND
9
16 0.8
3.3 0.7
b
mean
a
b
ND indicates “undetectable”. Calculated from samples 1 to 5, not including the A549 DNA sample. Each sample was analyzed at least three times.
(7) Uchida, K.; Kanematsu, M.; Morimitsu, Y.; Osawa, T.; Noguchi,
N.; Niki, E. J. Biol. Chem. 1998, 273, 16058−16066.
(8) Anderson, M. M.; Hazen, S. L.; Hsu, F. F.; Heinecke, J. W. J. Clin.
Invest. 1997, 99, 424−432.
nium bicarbonate. In contrast, the other ammonium salts do
not have this enhancement effect. These improvements allow
us to accurately determine the levels of each isomer of AcrdG.
By employing the developed sensitive method, α-AcrdG is
found in cultured human cells even without Acr treatment.
Although its level is just comparable to 30% of γ-AcrdG, α-
AcrdG should be paid more attention regarding its higher
mutation potency. Besides AcrdG isomers, non-dG adducts
(AcrdA and AcrdC) are also found in untreated A549 cells and
leukocytes. Some non-dG adducts (α-AcrdA) may be present at
levels even comparable to that of the predominant γ-AcrdG.
These results may suggest that the Acr−DNA adducts can be
spontaneously generated through normal cellular metabolism.
(9) O’Brien, P. J.; Siraki, A. G.; Shangari, N. Crit. Rev. Toxicol. 2005,
35, 609−612.
(10) Williams, T. I.; Lynn, B. C.; Markesbery, W. R.; Lovell, M. A.
Neurobiol. Aging 2006, 27, 1094−1099.
(11) Bradley, M. A.; Markesbery, W. R.; Lovell, M. A. Free Radic. Biol.
Med. 2010, 48, 1570−1576.
(12) Lovell, M. A.; Xie, C.; Markesbery, W. R. Neurobiol. Aging 2001,
22, 187−194.
(13) Liu, X.; Lovell, M. A.; Lynn, B. C. Anal. Chem. 2005, 77, 5982−
5989.
(14) Uchida, K. Free Radic. Biol. Med. 2000, 28, 1685.
(15) Chung, F. L.; Young, R.; Hecht, S. S. Cancer Res. 1984, 44, 990−
995.
(16) Parent, R. A.; Caravello, H. E.; San, R. H. C. J. Appl. Toxicol.
1996, 16, 103−108.
CONCLUSIONS
■
We developed a sensitive and accurate method for simulta-
neous quantification of α-AcrdG (two stereoisomers), γ-AcrdG,
α-AcrdC, and α-AcrdA in human cells using stable isotope
dilution UHPLC-MS/MS, which only requires 3 μg of genomic
DNA for each analysis. We also demonstrate that ammonium
bicarbonate not only improves the separation of the Acr−DNA
adducts and their isomers but also enhances the MS detection
of modified 2′-deoxyribonucleosides with difficulty in proto-
nation during the ESI process.
(17) Marnett, L. J.; Hurd, H. K.; Hollstein, M. C.; Levin, D. E.;
Esterbauer, H.; Ames, B. N. Mutat. Res. 1985, 148, 25−34.
(18) Kawanishi, M.; Matsuda, T.; Nakayama, A.; Takebe, H.; Matsui,
S.; Yagi, T. Mutat. Res. 1998, 417, 65−73.
(19) Wang, H. T.; Zhang, S.; Hu, Y.; Tang, M. S. Chem. Res. Toxicol.
2009, 22, 511−517.
(20) Nath, R. G.; Chung, F. L. Proc. Natl. Acad. Sci. U.S.A. 1994, 91,
7491−7495.
(21) Nath, R. G.; Ocando, J. E.; Chung, F. L. Cancer Res. 1996, 56,
452−456.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: hlwang@rcees.ac.cn.
Notes
(22) Nath, R. G.; Ocando, J. E.; Guttenplan, J. B.; Chung, F. L.
Cancer Res. 1998, 58, 581−584.
(23) Chen, H. J.; Lin, W. P. Anal. Chem. 2009, 81, 9812−9818.
(24) Zhang, S.; Villalta, P. W.; Wang, M.; Hecht, S. S. Chem. Res.
Toxicol. 2007, 20, 565−571.
The authors declare no competing financial interest.
(25) Sodum, R.; Shapiro, R. Bioorg. Chem. 1988, 16, 272−282.
(26) Smith, R. A.; Williamson, D. S.; Cerny, R. L.; Cohen, S. M.
Cancer Res. 1990, 50, 3005−3012.
ACKNOWLEDGMENTS
■
This work was supported by grants from the National Basic
Research Program of China (2011CB936001, 2009CB421065,
and 2011YQ060084) and the National Natural Science
Foundation of China (21077129, 20890112, and 21125523).
(27) Pawiowicz, A. J.; Munter, T.; Klika, K. D.; Kronberg, L. Bioorg.
Chem. 2006, 34, 39−48.
(28) Pawłowicz, A. J.; Munter, T.; Zhao, Y.; Kronberg, L. Chem. Res.
Toxicol. 2006, 19, 571−576.
(29) Kawai, Y.; Furuhata, A.; Toyokuni, S.; Aratani, Y.; Uchida, K. J.
Biol. Chem. 2003, 278, 50346−50354.
REFERENCES
■
(30) Chenna, A.; Iden, C. R. Chem. Res. Toxicol. 1993, 6, 261−268.
(31) Pawłowicz, A. J.; Klika, K. D.; Kronberg, L. Eur. J. Org. Chem.
2007, 1429−1437.
(1) Stevens, J. F.; Maier, C. S. Mol. Nutr. Food Res. 2008, 52, 7−25.
(2) Hecht, S. S. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 15725−15726.
(3) Ho, S. S.; Yu, J. Z.; Chu, K. W.; Yeung, L. L. J. Air Waste Manage.
Assoc. 2006, 56, 1091−1098.
(4) A comprehensive analysis of biodiesel impacts on exhaust
emissions, Draft Technical Report, United States Environmental
Protection Agency, 2002.
(32) Pawłowicz, A. J.; Kronberg, L. Chem. Biodivers. 2008, 5, 177−
188.
(33) Foiles, P. G.; Akerkar, S. A.; Chung, F. L. Carcinogenesis 1989,
10, 87−90.
(34) McDiarmid, M. A.; Iype, T.; Kolodner, K.; Jacobson-Kram, D.;
Strickland, P. T. Mutat. Res. 1991, 248, 93−99.
(35) Nordhoff, E.; Kirpekar, F.; Roepstorff, P. Mass Spectrom. Rev.
1996, 15, 67−138.
(5) Feng, Z.; Hu, W.; Hu, Y.; Tang, M. S. Proc. Natl. Acad. Sci. U.S.A.
2006, 103, 15404−15409.
(6) Chung, F. L.; Chen, H. J.; Nath, R. G. Carcinogenesis 1996, 17,
2105−2111.
3196
dx.doi.org/10.1021/ac3034695 | Anal. Chem. 2013, 85, 3190−3197

Analytical Chemistry
Article
(36) Frelon, S.; Douki, T.; Ravanat, J. L.; Pouget, J. P.; Tornabene,
C.; Cadet, J. Chem. Res. Toxicol. 2000, 13, 1002−1010.
(37) Yang, I. Y.; Chan, G.; Miller, H.; Huang, Y.; Torres, M. C.;
Johnson, F.; Moriya, M. Biochemistry 2002, 41, 13826−13832.
3197
dx.doi.org/10.1021/ac3034695 | Anal. Chem. 2013, 85, 3190−3197