Direct Alkylation of Deoxyguanosine by Azaserine Leads to O6-Carboxymethyldeoxyguanosine

Article abstract of DOI:10.1021/acs.chemrestox.0c00471

The O6-alkylguanosine adduct O6-carboxymethyldeoxyguanosine (O6-CMdG) has been detected at elevated levels in blood and tissue samples from colorectal cancer patients and from healthy volunteers after consuming red meat. The diazo compound l-azaserine leads to the formation of O6-CMdG as well as the corresponding methyl adduct O6-methyldeoxyguanosine (O6-MedG) in cells and is therefore in wide use as a chemical probe in cellular studies concerning DNA damage and mutation. However, there remain knowledge gaps concerning the chemical basis of DNA adduct formation by l-azaserine. To characterize O6-CMdG formation by l-azaserine, we carried out a combination of chemical and enzymatic stability and reactivity studies supported by liquid chromatography tandem mass spectrometry for the simultaneous quantification of O6-CMdG and O6-MedG. We found that l-azaserine is stable under physiological and alkaline conditions as well as in active biological matrices but undergoes acid-catalyzed hydrolysis. We show, for the first time, that l-azaserine reacts directly with guanosine (dG) and oligonucleotides to form an O6-serine-CMdG (O6-Ser-CMdG) adduct. Moreover, by characterizing the reaction of dG with l-azaserine, we demonstrate that O6-Ser-CMdG forms as an intermediate that spontaneously decomposes to form O6-CMdG. Finally, we quantified levels of O6-CMdG and O6-MedG in a human cell line exposed to l-azaserine and found maximal adduct levels after 48 h. The findings of this work elucidate the chemical basis of how l-azaserine reacts with deoxyguanosine and support its use as a chemical probe for N-nitroso compound exposure in carcinogenesis research, particularly concerning the identification of pathways and factors that promote adduct formation.

Full text of DOI:10.1021/acs.chemrestox.0c00471

pubs.acs.org/crt
Article
Direct Alkylation of Deoxyguanosine by Azaserine Leads to
O6‑Carboxymethyldeoxyguanosine
Susanne M. Geisen, Claudia M. N. Aloisi, Sabrina M. Huber, Emma S. Sandell, Nora A. Escher,
and Shana J. Sturla*
Cite This: Chem. Res. Toxicol. 2021, 34, 1518−1529
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: The O6-alkylguanosine adduct O6-carboxymethyl-
deoxyguanosine (O6-CMdG) has been detected at elevated levels
in blood and tissue samples from colorectal cancer patients and
from healthy volunteers after consuming red meat. The diazo
compound L-azaserine leads to the formation of O6-CMdG as well
as the corresponding methyl adduct O6-methyldeoxyguanosine
(O6-MedG) in cells and is therefore in wide use as a chemical
probe in cellular studies concerning DNA damage and mutation.
However, there remain knowledge gaps concerning the chemical
basis of DNA adduct formation by L-azaserine. To characterize O6-
CMdG formation by L-azaserine, we carried out a combination of
chemical and enzymatic stability and reactivity studies supported
by liquid chromatography tandem mass spectrometry for the
simultaneous quantification of O6-CMdG and O6-MedG. We found that L-azaserine is stable under physiological and alkaline
conditions as well as in active biological matrices but undergoes acid-catalyzed hydrolysis. We show, for the first time, that L-
azaserine reacts directly with guanosine (dG) and oligonucleotides to form an O6-serine-CMdG (O6-Ser-CMdG) adduct. Moreover,
by characterizing the reaction of dG with L-azaserine, we demonstrate that O6-Ser-CMdG forms as an intermediate that
spontaneously decomposes to form O6-CMdG. Finally, we quantified levels of O6-CMdG and O6-MedG in a human cell line exposed
to L-azaserine and found maximal adduct levels after 48 h. The findings of this work elucidate the chemical basis of how L-azaserine
reacts with deoxyguanosine and support its use as a chemical probe for N-nitroso compound exposure in carcinogenesis research,
particularly concerning the identification of pathways and factors that promote adduct formation.
or can be derived from nitrogenous dietary precursors via
nitrite-mediated nitrosation of secondary amines that occurs
largely endogenously.19 If unrepaired, the high miscoding
potential of O6-MedG to pair with thymine results in GC →
AT transition mutations upon replication.4,5 For O6-CMdG,
the misincorporation of thymine, and to a smaller extent of
adenine, could be demonstrated,20,21 therefore providing a
mechanistic basis for the GC→ AT transition and GC → TA
transversion mutations found in the p53 gene after treatment
with a carboxymethylating agent and replication in yeast.22
Interestingly, these are also the two most common mutations
found in colorectal cancer.22,23
INTRODUCTION
■
The structural integrity of nucleobases in DNA is essential for
the correct functioning of cellular processes. The reaction of
DNA with electrophilic exogenous and endogenous chemicals
can result in DNA adducts that are mutagenic and initiate
carcinogenesis. The DNA adducts O6-carboxymethyldeoxygua-
nosine (O6-CMdG) and O6-methyldeoxyguanosine (O6-
MedG) are pro-mutagenic lesions of current interest in the
context of colon carcinogenesis associated with meat
consumption.1−5 Both adducts were elevated in blood and
tissue samples from colorectal cancer patients.6−9 Further, O6-
CMdG was detected in colonic exfoliated cells from human
feces10 and in human blood samples from individuals on high-
red-meat diets.11 The association with meat, particularly red,
was confirmed in other studies where O6-CMdG was detected
at higher levels in tissues from volunteers with a high-red-meat
diet compared to a vegetarian one12 and in studies where
digestion of red and white meat was modeled in vitro.13−15
The formation of O6-CMdG and O6-MedG was also observed
after exposure of DNA to N-nitroso compounds
(NOCs),11,16,17 which are present in red and cured meat18
Carboxymethlyating agents, such as compounds derived
from N-carboxmethyl-NOCs and other chemical probes,24
Received: October 30, 2020
Published: June 1, 2021
© 2021 The Authors. Published by
American Chemical Society
https://doi.org/10.1021/acs.chemrestox.0c00471
Chem. Res. Toxicol. 2021, 34, 1518−1529
1518

Chemical Research in Toxicology
pubs.acs.org/crt
Article
Figure 1. Proposed mechanisms for L-azaserine-induced O6-CMdG formation. (A) Nucleophilic attack of dG on L-azaserine results in an O6-Ser-
CMdG intermediate followed by hydrolysis of L-serine to yield O6-CMdG. (B) L-Azaserine decomposes spontaneously to L-serine and diazoacetate,
which subsequently carboxymethylates dG. (C) Enzymatic β-elimination of L-azaserine results in pyruvate and diazoacetate.
promote the formation of O6-CMdG (Figure 1) and to a lesser
extent O6-MedG.11,17,25 For example, L-azaserine and potas-
sium diazoacetate are commonly used to generate O6-CMdG
and O6-MedG.8,10,17,26,27 In previous studies, potassium
diazoacetate induced higher absolute levels of O6-CMdG and
O6-MedG than L-azaserine, but the carboxymethyl:methyl ratio
was lower for potassium diazoacetate.16,17 The average level of
O6-CMdG and O6-MedG adducts after exposure to potassium
diazoacetate was similar in cells and naked DNA, suggesting
that no enzymatic activation is required for potassium
diazoacetate. However, the high activity of potassium
diazoacetate limits its use as a carboxymethylating agent, and
no reproducible results were obtained in cells.10 Moreover, it is
difficult to determine potassium diazoacetate concentration,
due to its high instability and fast conversion during
characterization. Hence, a nominal concentration is normally
assumed based on quantitative production from alkaline
hydrolysis of ethyldiazoacetate,10,28 but a true yield to our
knowledge has never been determined. Based on technical
limitations in the use of potassium diazoacetate, L-azaserine
seems to be the preferred chemical for inducing O6-CMdG and
O6-MedG, and exposure of cells to L-azaserine resulted in
reproducible adduct levels.27
L-Azaserine was discovered in 1954 as the active component
of a crude filtrate from a culture of Streptomyces bacteria with
antibiotic activity and potent antitumor activity in rodents.29 It
was positive in the Ames test and induced adenocarcinoma in
rats30 via the formation of DNA damage.31 The formation of
O6-CMdG by L-azaserine was demonstrated by reaction with
calf thymus DNA (ctDNA) and exposure in cells.16,17,26,27
Despite the long-standing knowledge that L-azaserine gives rise
to O6-CMdG, details concerning the mechanism of this
reaction remain limited.
diazoacetate and pyruvate (Figure 1C) or spontaneous
hydrolysis to yield diazoacetate and L-serine (Figure 1B).
Evidence for enzymatic β-elimination of L-azaserine was
established using enzymatic extracts from mouse liver, and a
stoichiometric relationship could be demonstrated for L-
azaserine disappearance and glycolic acid, ammonia, and
pyruvate formation, with the last two being the hydrolysis
products of the β-elimination intermediate 2-aminoacrylate.32
The responsible enzyme was presumed to be a dehydrogen-
ase.33 Decomposition of L-azaserine to pyruvate and ammonia
was also demonstrated in bicarbonate buffer, but only in the
presence of pyridoxal ion as a cofactor. A pyridoxal ion-
catalyzed β-elimination of L-azaserine via a Schiff base was
suggested, resulting in diazoacetate and aminoacrylate that
subsequent hydrolyses to pyruvate and ammonia.34 However,
the observation that O6-CMdG and O6-MedG form from L-
azaserine and ctDNA in the absence of enzymes suggested an
additional, nonenzymatic mechanism of L-azaserine breakdown
that could yield DNA adducts. This mechanism has been
proposed to involve spontaneous hydrolysis to yield serine and
diazoacetate (Figure 1B).17 A further possibility is that L-
azaserine may react directly with DNA and that subsequent
hydrolysis yields O6-CMdG (Figure 1A).
The objective of this study was to elucidate the chemical
mechanism of O6-CMdG formation by L-azaserine and
establish a chemical basis for its use to address the
corresponding DNA damage in cells. Thus, we evaluated the
stability of L-azaserine under varying pH conditions and in the
presence of active biological matrices. We characterized in
detail the basis of O6-CMdG formation from the direct
reaction of L-azaserine with guanosine and oligonucleotides.
Finally, we quantified the formation of O6-CMdG and O6-
MedG by L-azaserine in human colon epithelial cells (HCEC)
on the basis of dose and time by establishing a mass
spectrometry approach for the simultaneous and sensitive
quantification of O6-CMdG and O6-MedG. The findings
suggest a new direct reaction mechanism between L-azaserine
and dG that gives rise to a O6-Ser-CMdG intermediate, which
It has been hypothesized that L-azaserine gives rise to
carboxymethyl adducts by conversion to diazoacetate, which
can carboxymethylate DNA (Figure 1). Two potential
mechanisms for the conversion of L-azaserine in diazoacetate
have been suggested: enzymatic β-elimination of L-azaserine to
1519
https://doi.org/10.1021/acs.chemrestox.0c00471
Chem. Res. Toxicol. 2021, 34, 1518−1529

Chemical Research in Toxicology
pubs.acs.org/crt
Article
undergoes spontaneous hydrolysis to form O6-CMdG. Further,
the characterization of adduct formation in cells as a function
of L-azaserine dose and time of exposure supports the use of L-
azaserine as a chemical probe in future studies addressing
biological consequences of O6-CMdG.
30 min on a shaker. Cell lysate was collected using a cell scraper and
transferred into a 1.5 mL Eppendorf tube for immediate use. The
mixture was incubated at 37 °C between 30 min and 24 h. After
incubation, the mixture was loaded onto a prewashed 10 kDa MWCO
filter unit (VWR, prewashed with H2O (300 μL) by centrifugation at
12,000g for 5 min) and centrifuged at 12,000g for 10 min. The flow
through was loaded onto a Phenomenex Strata-X 33 μm Polymeric
Reversed Phase 30 mg/1 mL SPE column that was prewashed with 2
× 1 mL MeOH followed by 2 × 1 mL H2O. The L-azaserine-
containing fraction was eluted using 400 μL of H2O, concentrated to
dryness by vacuum centrifugation and redissolved in 50 μL H2O for
HPLC analysis using the identical method as described in the pH
Stability of l-Azaserine section. Results are shown in Figure S1A. As a
positive control, fluorescein diacetate (FDA) was also tested as a
substrate. Stock solutions (20 mg/mL in acetone) were stored at −20
°C, and immediately before use, an aliquot of FDA stock solution was
diluted in acetone to a final concentration of 2 mg/mL. FDA (2 μL of
2 mg/mL acetone solution) was added to fresh cell lysate (400 μL) or
to CelLytic M (400 μL) to obtain a final FDA concentration of 24
μM. The resulting mixtures were allowed to react at room
temperature for 2 h and applied to prewashed 10 kDaA MWCO
filter units and centrifuged (12,000g for 10 min). The 50 μL aliquots
of the flow throughs were loaded on a black 96 half-well plate for
analysis. The experiment was performed three times.
To quantify the amount of FDA converted to fluorescein, a stock
solution was prepared fresh at a concentration of 2 mg/mL in
acetone. This stock solution was used to generate a fluorescein
calibration line in CelLytic M with fluorescein concentrations of 75
μM, 50 μM, 25 μM, 10 μM, 5 μM, 1 μM, and 0.5 μM. The 50 μL
aliquots (triplicates) of each concentration were loaded on a black 96
half-well plate for analysis. Fluorescence measurements were acquired
on an Infinite M200 Pro plate reader from Tecan at 18 °C using
excitation wavelength of 490 nm and an emission wavelength of 525
nm. No fluorescence readout was obtained for the FDA in CelLytic
control, confirming that there is no spontaneous hydrolysis of FDA
occurring.
Incubation of L-Azaserine with Porcine Liver Esterase.
Lyophilized powder of porcine liver esterase (Sigma, E3019) was
reconstituted in 10 mM borate buffer pH 8.0 at a concentration of 40
mU/μL. L-Azaserine (100 μM) in 10 mM borate buffer pH 8.0 was
incubated at 37 °C with varying amounts of porcine liver esterase
(0.04−4 U) between 30 min and 24 h. After incubation, the esterase
was removed by filtration over a 10 kDa MWCO filter unit at 12,000g
for 5 min. The flow through was directly used for HPLC
quantification of L-azaserine using the identical method as described
in the pH Stability of l-Azaserine section. Results are shown in Figure
S1B. FDA was tested as positive control as described in Activity of
Whole Cell Lysate toward l-Azaserine section.
Reaction of an Oligonucleotide with L-Azaserine. A 9-mer
oligonucleotide (TTTTGTTTT, 5 mM) was allowed to react with L-
azaserine (50 mM) in 1 × sodium phosphate buffer (20 μL) at 37 °C
for 24 h. The resulting mixture (20 μL) was directly analyzed by
reversed-phase HPLC with a diode array detector set to monitor
absorbance at λ = 260 nm on a Clarity oligo-RP column, 250 × 4.6
mm, particle size 5 μm, pore size 110 Å (Phenomenex, Torrance,
California) using TEAA 50 mM as mobile phase A and acetonitrile
(ACN) as mobile phase B. Chromatographic separation was achieved
by a gradient from 92% A to 85% A in 30 min at a flow rate of 1 mL/
min. The unreacted oligonucleotide eluted with a retention time of 21
min, followed by three small peaks with retention times 21.5, 22.0,
and 22.7 min (Figure S2A). Fractions of these peaks were collected
and further analyzed by electrospray ionization mass spectrometry on
a Velos Linear Ion Trap (Thermo Scientific, Waltham, Massachu-
setts). The electrospray mass spectrum for the material that eluted at
21.96 min in the UV-chromatogram corresponded to the anticipated
m/z value of the electrospray ionization mass for the O6-Ser-CMG-
oligonucleotide (2845.46, Figure S2C). To verify the structure and
sequence, this material was further fragmented to obtain its MS2 mass
spectrum. The collision energy was set at 30 with parent m/z 948,
which predominantly generated a-B and w type fragment ions (Figure
EXPERIMENTAL PROCEDURES
■
Materials. Potassium diazoacetate was synthesized via alkaline
hydrolysis of ethyldiazoacetate as previously described, and neutral
solutions for further usage were obtained by neutralizing with 0.1 M
HCl.10,17,22 L-Azaserine was purchased from Abcam (Cambridge, UK)
1
and purity confirmed by H NMR. Ethyldiazoacetate was obtained
from Pfaltz & Bauer Inc. (Waterbury, Connecticut), deoxygunaosine
(dG) from Fluorochem (Hadfield, UK), and labeled 15N5-dG from
Cambridge Isotope Laboratories Inc., (Tewskbury, Massachusetts)
and were used without further purification. N2-MedG standard was
purchased from Carbosynth (Compton, UK). Oligonucleotide was
ordered high-performance liquid chromatography (HPLC)-pure from
Eurogentec. If not stated otherwise, all chemicals were purchased
from Sigma-Aldrich (Buchs, Switzerland) and used without further
purification. Milli-Q water was from a Millipore Synergy UV device.
Mobile phases for liquid chromatography (LC) were obtained from
Sigma-Aldrich (Buchs, Switzerland), and mobile phases for LC
coupled with mass spectrometry (LC-MS) were purchased from
Biosolve Chimie (Dieuze, France). All media and buffers for cell
culture were purchased from Invitrogen (Carlsbad, California).
HCEC-1CT cell lines were kindly provided by Prof J.W. Shay
(University of Texas, USA).35 NMR analysis was performed on a
Bruker Biospin 400 MHz NMR instrument (Bruker, Billerica,
Massachusetts), and chemical shifts are reported in parts per million
(ppm, δ) referenced to the chemical shift of the respective solvent.
NMR solvents were purchased from Sigma-Aldrich (Buchs, Switzer-
land). NMR files were evaluated by Mnova version 11 (Mestrelab
Research, Santiago de Compostela, Spain). Reversed-phase HPLC
analysis was performed on either an Agilent 1100 or 1200 Series
HPLC (Agilent Technologies, Santa Clara, California).
Data were analyzed and plotted using GraphPad Prism, Version 8
(GraphPad Software Inc., San Diego, California). Applied statistical
tests are specified for each data set in the corresponding figure
caption.
pH Stability of L-Azaserine. L-Azaserine (25 mM) was dissolved
in 500−980 μL of 0.1 M buffer solutions at a pH ranging from pH 2
to 11: pH 2, 2.4, and 3.4 glycine-HCl buffer, pH 5.8 sodium
phosphate, pH 7.2 sodium phosphate buffer, pH 8.5 carbonate buffer,
and pH 10.5 tris-buffer. Solutions were analyzed by reversed-phase
HPLC with a diode array detector set to monitor absorbance at λ =
254 nm (L-azaserine) and 210 nm (serine, pyruvate) on a Luna C18
column, 4.6 × 250 mm, particle size 5 μm, and pore size 100 Å
(Phenomenex, Torrance, California) using H2O as mobile phase A
and ACN as mobile phase B. Chromatographic separation was
achieved by an initial isocratic hold at 100% A for 5 min, followed by
a gradient to 40% B in 10 min at a flow rate of 1 mL/min. Injection
volume was 10 μL. The starting concentration of L-azaserine was
chosen to allow for detection of 0.5% and quantification of 2%
hydrolysis to L-serine (LOQ 0.5 mM L-serine). Each sample (pH 2−
11) was analyzed every 3 h for a total of 120 h. The analytical method
was shortened for samples at pH 2−5.8 to an isocratic hold at 100% A
for 6.0 min. L-Serine eluted at retention time 2.6 min, pyruvate at 2.9
min, and L-azaserine at 3.8 min. A calibration curve for L-azaserine and
L-serine from 0.5, 1, 6, 12, 25, and 50 mM was analyzed.
Activity of Whole Cell Lysate toward L-Azaserine. L-Azaserine
(5 μL of 2 mM stock solution) was added to freshly prepared whole
cell lysate (100 μL) derived from human lung epithelial cells (BEAS-
2B cells, ATCC CRL-9609) cultured in bronchial epithelial basal
medium (BEBM, Lonza) at 37 °C in 5% CO2 atmosphere. To obtain
the lysate, growth medium was removed when the cells reached a
confluency of ∼80%, and 1 × DPBS (10 mL) was added to wash the
cells. Per 100 mm cell culture dish, 1 mL of CelLytic M (Sigma,
C2978) was added followed by incubation at room temperature for
1520
https://doi.org/10.1021/acs.chemrestox.0c00471
Chem. Res. Toxicol. 2021, 34, 1518−1529

Chemical Research in Toxicology
pubs.acs.org/crt
Article
S2B). The CID mass fragments corresponded to the calculated mass
fragments of the O6-Ser-CMG-oligonucleotide, indicating the
anticipated sequence context and the serine modification to be on
guanine. HRMS was calculated for TTTT-G-TTTT: [M − H]− m/z
2699.813, found 674.12, 898.994, 1348.94, Magtran deconvolution
2700.26, TTTT- O6-CMG-TTTT: [M − H]− m/z 2759.7, found
688.32, 918.46, 1377.94, Magtran deconvolution 2758.48, TTTT- O6-
Ser-CMG-TTTT: [M − H]− m/z 2846.69, found 568.044, 710.483,
947.45, 1421.45, Magtran deconvolution 2845.46
Target inclusion list was expanded to include O6-Ser-CMdG m/z
413.1421 [M + H]+, O6-Ser-CMG m/z 297.0947 [M + H]+, L-
azaserine m/z 174.0515 [M + H]+, serine m/z 106.0504 [M + H]+,
and dG m/z 268.1046 [M + H]+. The mass tolerance was set to 50
ppm. Fragmentation of a O6-Ser-CMdG was targeted in MS2 XICs
with neutral loss of dR, O6-Ser-CMG m/z 297.0947 [M + H − dR]+,
and serine fragment,38−40 with m/z 251.0893 [M + H − (dR + H2O +
CO)]+. Mass tolerance for all XICs was 10 ppm.
Synthesis of O6-Ser-CMdG. Four mM dG was dissolved in 1200
μL of 0.1 M sodium phosphate buffer at pH 7.2. Twenty-five mM L-
azaserine was added, and the mixture was allowed to react at 22 °C for
120 h while being continuously analyzed by reversed-phase HPLC
with a diode array detector s set to monitor absorbance at λ = 254 nm
(L-azaserine, dG, G, O6-CMdG, O6-Ser-CMdG), 210 nm (L-serine),
and 280 nm (O6-CMdG, O6-Ser-CMdG) on a Luna C18 column, 4.6
× 250 mm, particle size 5 μm, pore size 100 Å (Phenomenex,
Torrance, California) using H2O as mobile phase A and ACN as
mobile phase B. Chromatographic separation was achieved by an
initial isocratic hold at 100% A for 5 min, followed by a gradient to
45% B in 45 min at a flow rate of 1 mL/min. Injection volume was 10
μL. All eluting peaks were identified by m/z on a Velos Linear Ion
Trap (Thermo Scientific, Waltham, Massachusetts) and in case of L-
azaserine, dG, guanine, and L-serine with coelution of respective
standards and corresponding DAD spectrum. L-Azaserine eluted at
retention time 3.9 min (m/z 174), dG at 16.8 min (m/z 268), L-serine
at 2.8 min (m/z 106), and guanine at 11.4 min (m/z 152) A peak at
retention time 7.9 min had m/z 297 matching the calculated mass for
Ser-G, while the peak of interest eluted at retention time of 18.7 min
with m/z 413 as calculated for O6-Ser-CMdG (Figure S5A). This peak
was collected over 200 runs and characterized. HRMS was calculated
for C15H20N6O8: [M + H]+ m/z 413.1421, found 413.1411 (Figure
S5B), and MS2 was calculated for C10H12N6O5: [M + H − Gua]+ m/z
Preparation of Standards for Mass Spectrometry. O6-CMdG
was prepared by the copper carbene-based method described by
Geigle et al.36 In short, 2.5 mg (9.35 μmol) of dG was allowed to react
with 10.7 mg (93.5 μmol) of ethyldiazoacetate in the presence of 300
μg (1.87 μmol) CuSO4. O6-CMdG was obtained by alkaline
hydrolysis followed by reversed-phase HPLC purification with a
diode array detector set to monitor absorbance at λ = 280 and 250 nm
on a Luna C18 column, 4.6 × 250 mm, particle size 5 μm, pore size
100 Å (Phenomenex, Torrance, California) using 0.1% HOAc as
mobile phase A and ACN as mobile phase B. Chromatographic
separation was achieved by an initial isocratic hold at 100% A for 10
min, followed by a gradient to 16% B in 45 min at a flow rate of 1
mL/min. Injection volume was 20 μL. The product (RT 41.3 min, λ =
280, 250 nm) was obtained in 40% yield as a white powder. HRMS
(ESI) calculated for C12H15N5O6: [M + H]+ m/z 326.1101, found
326.1095, MS2 calculated for C7H7N5O3: [M + H − Gua]+ m/z
1
210.0628, found 210.0617. H NMR (Figure S3) matched published
data.16 15N5-O6-CMdG was prepared by the same protocol using 2.5
mg (9.35 μmol) of dG 15N5-dG as starting material, yielding 1 mg
(40% yield) of a white powder of 15N5-O6-CMdG. HRMS (ESI)
calculated for C12H1515N5O6: [M + H]+ m/z 331.0952, found
331.0944, MS2 calculated for C7H715N5O3: [M + H − Gua]+ m/z
215.0479, found 215.0469 [M + H − Gua]+.
O6-MedG was synthesized by the method reported by Reza et al.37
In short, 0.57 g (2 mmol) dG was allowed to react in 10 mL of
pyridine with 4.3 g (79.6 mmol) of sodium methoxide in 300 mL of
MeOH. The product was purified by reversed-phase HPLC with a
diode array detector set to monitor absorbance at λ = 260 nm on a
Luna C18 column 10 × 250 mm, particle size 5 μm, pore size 100 Å
(Phenomenenx, Torrance, California) using 5% ACN in H2O as
mobile phase A and ACN as mobile phase B. Chromatographic
separation was achieved by an initial isocratic hold at 100% A for 15
min, followed by gradient to 35% B in 10 min at a flow rate of 2 mL/
min. Injection volume was 200 μL. Product (RT 20.7 min, λ = 260
nm) was obtained in a 60% yield as a light yellow powder. HRMS
(ESI) calculated for C11H15N5O4: [M + H]+ m/z 282.1202, found
282.1197, MS2 calculated for C6H7N5O: [M + H − Gua]+ m/z
166.0729, found 166.0715. 1H NMR (Figure S4) matched the
published data.37 D3-O6-MedG was synthesized accordingly with 0.19
g (0.67 mmol) of dG and 1.51 g (26.5 mmol) of D3-sodium
methoxide in 100 mL MeOD (both Armar chemicals, Switzerland).
Product was purified by reversed-phase LC as described for unlabeled
O6-MedG and obtained as a light-yellow powder in 60% yield. HRMS
(ESI) calculated for C11H12D3N5O4: [M + H]+ m/z 285.1391, found
285.1383, MS2 calculated for C6H5D3N5O: [M + H − Gua]+ m/z
169.0918, found 169.0904. We did not observe any D/H exchange by
analyzing a 100 nM D3-O6-MedG standard solution by nanoLC-ESI-
HRMS2 and following the intensity in the extracted ion chromato-
grams (XICs) for D3-O6-MedG (m/z 285.1383 [M + H]+) and O6-
MedG (m/z 282.1197 [M + H]+, not detectable (nd)) over time of
the analysis.
297.0947, found 297.0938 H NMR in deuterium oxide confirmed
1
presence of O6-Ser-CMdG (Figure S5C). 1H NMR (400 MHz,
deuterium oxide) δ 8.14 (s, 1H, Ar−H), 6.38 (t, J = 8.2 Hz, 1H, 1′-
H), 5.07 (s, 1H, NH2−CH), 4.83−4.81 (m, 2H, O−CH2−COO),
4.67−4.62 (m, 1H, 4′-H), 4.18−4.13 (m, 1H, 3′-H), 3.93−3.84 (m,
2H, O−CH2−CNH2), 3.83−3.74 (m, 2H, 5′-H), 2.84 (dt, J = 14.7,
7.4 Hz, 1H, 2′-H), 2.57−2.46 (m, 1H, 2′-H).
Hydrolysis of O6-Ser-CMdG. O6-Ser-CMdG was dissolved in 0.1
M phosphate buffer at pH 7.2, and the solution was analyzed by
reversed-phase HPLC every hour for a total of 30 h and peaks
identified as described above but using HOAc as mobile phase A and
a gradient consisting of an initial isocratic hold for 10 min at 100% A,
followed by a gradient to 16% B in 45 min. A product eluting at 42.5
min (Figure S6A) was isolated and confirmed to be O6-CMdG on the
basis of 1H NMR (Figure S6C) and HRMS (ESI) calculated for
C12H15N5O6: [M + H]+ 326.1101, found 326.1094, matching the
authentic standard.
Cell Culture of HCEC Cells and L-Azaserine Exposure. Cells
were maintained as monolayers in 10 cm dishes in a humidified, 5%
CO2 atmosphere at 37 °C. Media consisted of 80% DMEM and 20%
M199 Earle’s salt medium, supplemented with 2% Hyclone fetal
bovine serum (Hyclone Laboratories Inc., San Angelo, Texas), 25 ng/
L epidermal growth factor, 1 μg/L hydrocortisone, 10 μg/L insulin, 2
μg/L transferrin, 50 μg/L gentamycin, and 0.9 ng/L sodium selenite.
Cells were regularly confirmed to be mycoplasma free using the
MycoAlert Kit (Lonza, Basel, Switzerland).41 L-Azaserine stock
solutions were prepared in Milli-Q water. O6-Benzylguanine (O6-
BG) stock solutions were prepared in dimethyl sulfoxide (DMSO),
and the final concentration of DMSO was 0.1%. To assess cell
viability, cells were seeded in 96-well plates at a density of 1 × 104 and
exposed to increasing L-azaserine concentrations (0, 1, 10, 50, 250,
750, 1000, 2500 μM) for 120 h. Cells were exposed to TritonX
(Sigma-Aldrich, Buchs, Switzerland) as a positive control for
cytotoxicity. Cell survival was measured using the CellTiterGlo
assay (Promega, Madison, Wisconsin) following the manufacturer’s
instructions.42
Reaction of dG with L-Azaserine. dG (5 μM in 500 μL Milli-Q
or tris-buffer (10 mM) containing 2 mM MgCl2, at pH 7, 8, 9, and
10.6) was combined with L-azaserine or potassium diazoacetate (5
mM), and the reaction mixture was stirred for 0.5, 1, 2, 3, and 24 h at
37 °C. After the indicated time, samples were processed for
quantification of O6-CMdG and O6-MedG and analyzed by
nanoLC-ESI- HRMS2, as described in the detailed method validated
for analysis of cell samples elaborated below, with the following
changes: Trapping time was set to 0.1 min. Resolution was set to 60 K
for the orbitrap mass analyzer in MS1, 68−444 m/z, and 1 microscan.
For DNA adduct analysis, cells were seeded in 10 cm dishes at a
density of 2 × 106. Cells were exposed to 0, 125, 250, 500, and 1000
1521
https://doi.org/10.1021/acs.chemrestox.0c00471
Chem. Res. Toxicol. 2021, 34, 1518−1529

Chemical Research in Toxicology
pubs.acs.org/crt
Article
μM L-azaserine for 4 h to evaluate thhe dose−response relationship
and with 500 μM L-azaserine for 0.5, 4, 10, 24, 48, 72, 96, and 120 h
to determine the time-dependent O6-CMdG and O6-MedG
formation. Cell viability, dose−response, and time course studies
were each performed three independent times.
25, 50, and 100 nM of O6-CMdG and O6-MedG standards
undergoing sample preparation served for quantification (Figure S7).
Specificity was guaranteed by processing a blank sample for each
analysis and selectivity by coelution with internal standards and
accurate parent and fragment mass. Pure standards were used to
determine the limit of detection (LOD) and limit of quantification
(LOQ) defined as the amount of analyte generating a signal-to-noise
ratio (S/N) of 3 or 10 in the XICs. LOD (S/N > 3) was 5 pM (10
amol) and 1 pM (2 amol) for O6-CMdG and O6-MedG, respectively,
and LOQ (S/N > 10) was 10 pM (20 amol) for both adducts. We
further assessed LOQ in matrix, intra- and interday accuracy
(expressed as % recovery), and precision (expressed as the relative
standard deviation, RSD %) using quality control samples consisting
of calf thymus DNA (ctDNA, 15 μg) or HCEC genomic DNA (15
μg) spiked with increasing amounts of O6-CMdG and O6-MedG
(0.01, 0.05, 0.1, 1.0, 10, 50 100 and 200 nM). LOQ in matrix was
determined to be 50 pM (100 amol) for O6-CMdG and O6-MedG.
The precision and accuracy for the measurement of O6-CMdG and
O6-MedG in ctDNA and HCEC genomic DNA as matrix (Table S1)
had a relative standard deviation (RSD) ranging from 1.7 to 25.6%
(O6-CMdG) or 0.8 to 23.5% (O6-MedG) and recovery 83.7 to
113.4% (O6-CMdG) or 92.1 to 122.6% (O6-MedG) of the actual
adduct level.
dG Quantification by Liquid Chromatography. Quantitation
of dG was carried out by reversed-phase HPLC with a diode array
detector set to monitor absorbance at λ = 254 nm on a C18 Kinetex
column, 2.1 × 150 mm, particle size 2.6 μm, pore size 100 Å column
(Phenomenex, Torrance, California) using 3% ACN in H2O as
mobile phase A and ACN as mobile phase B. Chromatographic
separation was achieved by a gradient from 100% A to 88% A in 6 min
at a flow rate of 0.2 mL/min. The injection volume was 20 μL. The
retention times for nucleosides were the following: dC 3.5 min, dG
8.4 min, dT 9.2 min, and dA 10.6 min. A calibration curve for dG
(0.1, 1, 10, 25, 50, and 100 μM) was analzyed. The amount of dG was
then used to calculate the total number of nucleotides based on the
assumption that the GC content is 40.9% in human DNA.43
DNA Isolation and Sample Preparation for Adduct Analysis
by Mass Spectrometry. DNA was isolated from cell pellets using
the QIAamp DNA Mini Kit (©QIAGEN, Switzerland). Extracted and
dried DNA (20 μg) was dissolved in buffer (10 μL, 10 mM Tris pH 7,
2 mM MgCl2). To this sample was added 50 μL of a master mix
solution comprised of 2.5 U/μg DNA benzonase, 3 mU/μg DNA
phosphodiesterase I, 2 U/μg DNA alkaline phosphatase, 2 nM D3-O6-
MedG and 4 nM 15N5-O6-CMdG in the aforementioned buffer. A
solution of ctDNA (20 μg in 10 μL buffer) and a blank sample (10 μL
buffer) were combined with 50 μL of the master mix as controls for
DNA digestion and background signal, respectively. Following the
digestion, MS grade H2O (350 μL) was added to the sample (total
volume 410 μL), and enzymes were removed by molecular weight
filtration with a cutoff of 10 kDa (VWR, Radnor, Pennsylvania). An
aliquot (60 μL) of the resulting solution was removed for dG
quantification by HPLC (described below). Samples were concen-
trated to dryness in a Thermo Scientific Savant Universal SpeedVac
Vacuum system, resuspended in 200 μL MS grade H2O + 0.1% formic
acid (FA), and further processed by solid-phase extraction on a C18
cartridge, Sep-Pak Vac 1 cm3 (50 mg) (Waters, Milford,
Massachusetts) using MS grade CH3OH, CH3OH + FA, and H2O.
Cartridges were washed with 2 × 1 mL CH3OH and 1 mL CH3OH +
0.1% FA followed by 2 × 1 mL H2O + 0.1% FA. The samples were
loaded on the cartridge, washed with 2 × 1 mL H2O + 0.1% FA and 1
mL 3% CH3OH, and eluted with 2 × 500 μL of 80% CH3OH in H2O.
The eluted fractions were dried under vacuum and dissolved in 10 μL
H2O for subsequent MS analysis.
High-Resolution Mass Spectrometry Method for Simulta-
neous Quantification of O6-CMdG and O6-MedG. Simultaneous
quantification of O6-CMdG and O6-MedG was achieved on an
Orbitrap Fusion Lumos (Thermo Scientific, Waltham, Massachusetts)
equipped with a nano electrospray ionization source and a
nanoAcquity UPLC M-class system (Waters, Milford, Massachusetts).
Liquid chromatography was performed with an Acquity UPLC M-
Class Symmetry C18 Trap column (100 Å, 5 μm, 180 μm x 20 mm)
and an Acquity UPLC M-Class HSS T3 column (100 Å, 1.8 μm, 75
μm × 250 mm) at a column temperature of 40 °C. The sample loop
volume was 5 μL, with an injection volume of 2 μL. Mobile phases A
and B consisted of mass spectrometry grade H2O with 0.1% FA and
ACN with 0.1% FA, respectively (Biosolve Chimie, Dieuze, France).
The chromatographic method included trapping for 0.5 min at 99.5%
A at a flow rate of 15 μL/min. The chromatographic separation was
achieved by a 1 min hold at 3% B, followed by a linear gradient from
3% to 45% B over 19 min, at a flow rate of 0.3 μL/min. Nano
electrospray ionization was used in positive mode, 2.2 kV at 270 °C.
Mass spectrometry-based detection was performed in parallel reaction
monitoring (PRM) with higher-energy collisional dissociation in
orbitrap technology (HCD-OT) detection mode with the following
settings: The RF lens was set to 35%. Resolution of the orbitrap was
set to 120 K in MS1, 160−400 m/z, and 1 microscan. AGC target was
set to 4 × 105 with a maximum injection time of 100 ms. Targeted
inclusion list was comprised of O6-CMdG with m/z 326.1096 [M +
H]+, heavy 15N5-O6-CMdG with m/z 331.0946 [M + H]+, O6-MedG
with m/z 282.1995 [M + H]+, and heavy D3-O6-MedG with m/z
285.1384 [M + H]+. The mass tolerance was set to 25 ppm.
Isolation for MS2 was performed in the quadrupole with an isolation
window of 1.6 m/z and the first mass set to 140 m/z. HCD was used
for fragmentation with a collision energy of 25%. Fragments were
detected in the orbitrap with a resolution of 60 K. The AGC target
was set to 5 × 104 with a maximum injection time of 118 ms. For
quantification, neutral loss of deoxyribose was targeted in XICs using
O6-CMG m/z 210.0622 [M + H − Gua]+, heavy 15N5-O6-CMG m/z
215.0472 [M + H − Gua]+, O6-MeG m/z 166.0722 [M + H − Gua]+,
and heavy D3-O6-MeG m/z 169.0911 [M + H − Gua]+ with a mass
tolerance of 10 ppm. Calibration standards containing 0.5, 1, 5, 10,
RESULTS
■
Stability of L-Azaserine at Varying pH and Biological
Matrices. The hydrolysis of L-azaserine to yield L-serine and
diazoacetate has been previously suggested as a basis for the
formation of O6-CMdG (Figure 1B).17 However, this
mechanism raises concern about the wide use of L-azaserine
in cell-based studies and whether the compound is stable
under the experimental conditions of in vitro toxicological
assays. Therefore, we sought to characterize the stability of L-
azaserine as a function of pH and in the presence of biological
matrices. First, levels of L-azaserine and its hydrolysis product
L-serine were measured in solutions ranging in pH from pH 2
to 11 at 25 °C over the course of 5 days. No decomposition
was observed between pH 5.8 and 10.5, while at pH 2, L-
azaserine was completely hydrolyzed to L-serine after 2 h
(Figure S8). The acid-catalyzed hydrolysis was modeled as a
pseudo-first-order reaction (Figure 2A), and the half-life (t1/2
)
was calculated to be 0.4 h at pH 2, 0.7 h at pH 2.4, and 4.7 h at
pH 3.4. Next, we assessed the stability of L-azaserine in the
presence of cell culture medium, active whole cell lysates, and
excess commercial porcine liver esterase. None of these
biological matrices were active in catalyzing any conversion
of L-azaserine at 37 °C from 10 min to 24 h (Figure S1). On
the other hand, fluorescein diacetate, tested as a positive
control at the same concentration, was completely hydrolyzed
within a few minutes. Together, these data emphasize that L-
azaserine resists hydrolysis under neutral physiological
conditions in the presence of biological matrices and suggest
1522
https://doi.org/10.1021/acs.chemrestox.0c00471
Chem. Res. Toxicol. 2021, 34, 1518−1529

Chemical Research in Toxicology
pubs.acs.org/crt
Article
carboxymethylated dG with retention times close to the
potential Ser-CMdG isomers, and one of them coeluted with
the standard 15N5-O6-CMdG (18.35 min), thereby assigned as
O6-CMdG. Similarly, two peaks were detected with m/z
corresponding to isomers of MedG; one peak coeluted with an
N2-MedG standard (18.01 min), while the second peak
coeluted with the standard for O6-MedG (18.75 min). As a
negative control, we performed the same reaction with
potassium diazoacetate, and no peaks were detected
corresponding to Ser-CMdG or fragments thereof, but the
same two peaks for each carboxymethylation and methylation
of dG as for the L-azaserine treated samples were observed
(Figure 3B). None of the targeted m/z values were detected in
control samples with any of the single reagents. These results
suggest that L-azaserine reacts directly with dG, forming
adducts that retain the serine moiety at the O6- and N2-
positions.
O6-Ser-CMdG Decomposes to O6-CMdG. Having gained
evidence for the formation of O6-Ser-CMdG intermediate in
the reaction of dG with L-azaserine by mass spectrometry, we
further characterized this novel structure and confirmed that it
can be converted to O6-CMdG. Thus, by HPLC analysis, Ser-
CMdG was observed in the reaction of dG with L-azaserine
already after 30 min and remained detectable up to 120 h (18.7
min, m/z of 413, Figure S5A). The compound was purified by
Figure 2. Acid-promoted hydrolysis of L-azaserine. (A) Rate law for
hydrolysis of L-azaserine and relationship with observed values. (B)
Proposed mechanism of acid-catalyzed hydrolysis of L-azaserine to L-
serine.
1
HPLC, and its H NMR spectrum was consistent with Ser-
CMdG due to the presence of three characteristic peaks at 5.07
(s, 1H, NH2−CH), 4.83−4.81 (m, 2H, O−CH2−COO), and
3.93−3.84 ppm (m, 2H, O−CH2), accounting for five
additional Hs in addition to those present in dG (Figure
S5C). The HRMS of the isolated peak revealed m/z 413.1411
(RT 17.56 min), corresponding to Ser-CMdG (Figure 3C).
After isolation of Ser-CMdG by HPLC, analysis of the sample
by mass spectrometry always revealed the presence of O6-
CMdG, but not N2-CMdG, (Figure 3C), suggesting that the
Ser-CMdG adduct eluting at retention time 17.56 is O6-Ser-
CMdG.
against it being the initial step in the reaction of L-azaserine
with DNA to give rise to O6-CMdG.
Evidence for O6-Ser-CMdG as the Intermediate in the
Formation of O6-CMdG by L-Azaserine. Having established
that L-azaserine does not undergo hydrolysis under physiolog-
ically relevant conditions, we considered whether under those
conditions L-azaserine may directly react with dG to give rise
to the direct adduct O6-Ser-CMdG, which then yields O6-
CMdG upon spontaneous hydrolysis (Figure 1A). To gain
support for this mechanism, we first explored whether O6-Ser-
CMdG could be formed by reacting a 9-mer nucleotide with an
excess L-azaserine in sodium phosphate buffer at 37 °C for 24
h. The crude reaction mixture was then analyzed by HPLC for
the presence of newly formed oligonucleotide peaks (Figure
S2), which were collected and further analyzed by electrospray
ionization mass spectrometry. The electrospray mass spectrum
and the fragmentation for the oligonucleotide that eluted at
22.0 min in the UV-chromatogram corresponded to the
anticipated m/z values of the electrospray ionization mass for
the O6-Ser-CMG-oligonucleotide (Figure S2), providing the
first support for the direct reaction of L-azaserine with DNA.
To allow robust and unambiguous characterization of the
direct adduct, we decided to reduce the complexity of the
reaction and allow dG to react with an excess of L-azaserine in
tris-buffer at pH in the range of 7−11 for 0.5−24 h. The
resulting reaction mixtures were analyzed for the presence of
O6-Ser-CMdG, O6-CMdG, and O6-MedG by nanoLC-ESI-
HRMS2 in PRM mode. We monitored mass transitions
accounting for neutral loss of deoxyribose followed by possible
fragmentation of serine (Table 1).38−40 XICs of all reaction
mixtures revealed the appearance of two peaks with the mass
corresponding to O6-Ser-CMdG at a retention time of 17.9 and
18.1 min (Figure 3A). Their mass-fragments were the same,
consistent with being two positional isomers of Ser-CMdG.
Further, two peaks were detected with m/z corresponding to
Having established that O6-Ser-CMdG forms as a direct
product of the reaction between dG and L-azaserine, we
characterized further its aforementioned apparent hydrolysis to
yield O6-CMdG. At pH 7.2, the disappearance of O6-Ser-
CMdG (HPLC retention time 23 min) coincided with the
appearance of O6-CMdG (HPLC retention time 43 min;
Figure S6, identity confirmed by coelution with the authentic
standard and consistent HRMS data (Figure 4)). We
concluded that O6-Ser-CMdG is an intermediate in L-
azaserine-induced O6-CMdG formation, thereby suggesting a
possible mechanism for the formation of O6-CMdG in DNA
under physiological conditions in which L-azaserine is stable.
Exposure of Cells to L-Azaserine and Quantification
of Resulting Adducts. Having established that under neutral
conditions direct alkylation of dG by L-azaserine occurs to
form O6-Ser-CMdG that spontaneously hydrolyzes to yield O6-
CMdG, we were interested in characterizing this process in
cells and understanding the kinetic basis of adduct formation
upon exposure to L-azaserine. Immortalized but normal diploid
human colonic epithelial HCEC cells were employed as a
model for adduct formation in healthy colon tissue.35 HCEC
cells express epithelial and stem cell markers and do not show
mutations in genes known to be involved in colon cancer
progression (APC, KRAS, TP53).44 HCEC cells were exposed
to L-azaserine for 0.5 to 120 h and at concentrations ranging
from 125 to 1000 μM. Following DNA isolation and
1523
https://doi.org/10.1021/acs.chemrestox.0c00471
Chem. Res. Toxicol. 2021, 34, 1518−1529

Chemical Research in Toxicology
pubs.acs.org/crt
Article
Table 1. Targeted m/z with Proposed Structures for O6-Ser-CMdG and Transitions
a
O6-substitution was confirmed by isotopically labeled O6-CMdG and O6-MedG standards. N2-coordination is confirmed by N2-MedG standard.
b
Structures and m/z of O6-CMdG and O6-MedG as hydrolysis products are included. Detected structures are indicated for diazoacetate, L-
azaserine, and control samples from reaction mixture with dG and for the isolated compound by reversed-phase liquid chromatography (LC RT
18.7 min). nd = not detectable.
hydrolysis, products O6-CMdG and O6-MedG were detected,
but no O6-Ser-CMdG intermediate in cells, either because
levels were too low or because it is not formed in cells. While
inconclusive regarding whether the chemical mechanism
established herein is the same as what occurs in cells, the
fact that we could not detect it was not surprising, given our
observation of the high instability of the serine intermediate on
the nucleotide level, and the procedure required to digest
genomic DNA for analysis of modified nucleosides from cells.
Despite not observing the transient intermediate in samples
obtained from cells, in order to characterize overall rates of
DNA damage induced by L-azaserine in cells, we validated the
nanoLC-ESI-HRMS2 method for the concomitant quantifica-
tion of O6-CMdG and O6-MedG in cells with a stable isotope
dilution method that accounts for losses during sample
preparation and ionization and allows for accurate quantifica-
tion of the analyte based on the ratio of its signal area to that of
labeled internal standard. PRM was applied for the mass
transitions corresponding to a neutral loss of deoxyribose (dR)
from the protonated analytes: O6-CMdG and O6-MedG, and
their isotopically labeled internal standards 15N5-O6-CMdG
and D3-O6-MedG.
O6-CMdG and O6-MedG Levels in HCEC Cells. The
robust nanoLC-ESI-HRMS2 approach for simultaneously
quantifying O6-CMdG and O6-MedG enabled the measure-
ment of L-azaserine-induced adducts in HCEC cells in a dose-
and time-dependent manner. The level of both adducts
increased linearly with L-azaserine concentration (0−1000
μM, 4 h exposure), with maximum adduct levels of 32.8 7.8
O6-CMdG and 4.3
0.9 O6-MedG lesions/107 nucleotides
1524
https://doi.org/10.1021/acs.chemrestox.0c00471
Chem. Res. Toxicol. 2021, 34, 1518−1529

Chemical Research in Toxicology
pubs.acs.org/crt
Article
Figure 3. Representative chromatograms for the reaction of dG with L-azaserine or potassium diazoacetate analyzed by nanoLC-ESI-HRMS2. (A,
B) 5 μM dG was reacted with 5 mM L-azaserine (A) or potassium diazoacetate (B) as control in 10 mM tris-buffer at pH 7, followed by SPE and
nanoLC-ESI-HRMS2. (C) 4 mM dG was reacted with 25 mM L-azaserine in 0.1 M sodium phosphate buffer and peak with m/z 413 isolated by LC,
concentrated, and analyzed by nanoLC-ESI-HRMS2. XICs represent parent mass for O6-Ser-CMdG and two fragmentations, followed by O6-
CMdG, 15N5-O6-CMdG as internal standard, O6-MedG and D3-O6-MedG as internal standard. The sample represented in (C) was analyzed on a
new column, explaining the shift in absolute retention time, while the Δ between O6-Ser-CMdG and O6-CMdG was equal.
Figure 5. Dose-dependent formation of O6-CMdG and O6-MedG in
HCEC cells. Cells were exposed for 4 h to 0, 125, 250, 500, or 1000
μM L-azaserine, and extracted DNA was analyzed for O6-CMdG and
O6-MedG by nanoLC-ESI-HRMS2. Adduct levels are expressed as
lesions/107 nucleotides (mean SD, n = 3). P-values were calculated
with one-way ANOVA and Tukey’s multiple comparison’s test, **p <
0.005, ****p < 0.0001.
CMdG was previously reported in human colorectal cells
(HCT-166) and in XPA-deficient human fibroblast cells, and
O6-MedG is very well-known to be formed endogenously.27,45
Figure 4. NanoLC-ESI-HRMS2 chromatogram for the decomposition
Having established the dose-dependent formation of O6-
product of O6-Ser-CMdG. Decomposition product was isolated by
CMdG and O6-MedG after 4 h exposure to L-azaserine, we
reversed-phase LC and identified as O6-CMdG by nanoLC-ESI-
HRMS2. No methylated dG could be detected. The m/z values are
determined experimentally.
were interested in the temporal profiles of the DNA adducts.
Therefore, we exposed HCEC cells to 500 μM L-azaserine
(EC40 after 120 h, Figure S11) for various durations (0.5, 4, 10,
24, 48, 72, 96, 120 h). After 0.5 h, adduct levels were below the
LOQ for O6-CMdG and O6-MedG. O6-CMdG levels increased
during the first 48 h to a maximum adduct level of 116 21
O6-CMdG lesions/107 nucleotides (Figure 6) and plateaued
after, displaying no changes in adduct level for the remaining 3
days. Also, O6-MedG levels reached a maximum value after 48
h; however, we could not establish a time-dependent
(Figure 5, for representative MS2 spectra, see Figures S9 and
S10). Levels of O6-CMdG were low compared to O6-MedG, as
reported previously.17,27 We also detected O6-CMdG in
unexposed control samples but at levels below LOQ, while
0.7
quantified in unexposed control samples. Endogenous O6-
1.3 O6-MedG adducts/107 nucleotides could be
1525
https://doi.org/10.1021/acs.chemrestox.0c00471
Chem. Res. Toxicol. 2021, 34, 1518−1529

Chemical Research in Toxicology
pubs.acs.org/crt
Article
while similar diazo compounds such as diazomethane have a
pKa of 10, such a high value would be inconsistent with the
pH-stability profile of L-azaserine.48 The observation of acid-
promoted decomposition of L-azaserine is consistent with a
previous report of loss of antibiotic activity and N2 formation
at pH 2 for L-azaserine.49 Indeed, α-diazo compounds are acid
labile, resulting in formation of reactive alkenium ions and loss
of N2.50 Acid-promoted decomposition was also demonstrated
for diazopeptides and N-nitrosopeptides, while under physio-
logical conditions they are stable.51,52 Finally, under neutral
conditions, diazoacetate was demonstrated to be an
intermediate of N-nitroso-glycine-induced O6-CMdG and O6-
MedG formation, potentially via formation of carboxymethyl-
diazonium ion and diazomethane, whereas, to our knowledge,
the decomposition of NOCs with more stabilized α-
diazocarboxy-groups,53 for example, diazopeptides, has not
been characterized under physiological conditions.11
Figure 6. Time-dependent formation of O6-CMdG and O6-MedG in
HCEC cells. Cells were exposed to 500 μM L-azaserine, and DNA was
extracted after 0.5, 4, 10, 24, 48, 72, 96, and 120 h and further
analyzed for O6-CMdG and O6-MedG levels by nanoLC-ESI-HRMS2.
Adduct levels are expressed as lesions/107 nucleotides (mean SD, n
= 3).
Based on the lack of evidence for the decomposition of L-
azaserine at neutral pH in the presence of active whole cell
lysates, we hypothesized that L-azaserine reacts directly with
DNA under physiological conditions. The formation of a direct
reaction product Ser-CMdG from L-azaserine has been
previously proposed, but to our knowledge was never
investigated.17,54 Data from reactions with oligonucleotides
and with dG revealed the existence of the intermediate O6-Ser-
CMdG (Figure 1). We did not observe any correlation of levels
of the direct adduct with time or pH; however, the amount
that could be obtained from oligonucleotides was extremely
small and transient on the nucleotide level, making any clear
kinetic conclusions elusive. The other regioisomer observed
was assigned as N2-SerCMdG. On the basis of peak area, and
approximating the same recovery and ionization behavior, it
appears to be on average about 3 times less abundant than O6-
SerCMdG. However, for CMdG and MedG adducts, the
N2:O6 ratio was 1.3 and 1.1, respectively. Interestingly, for the
reaction with diazoacetate, the N2:O6 ratio was 2 for O6-CMdG
and 0.7 for O6-MedG, suggesting the adducts were formed via
a different chemical mechanism. Diazo compounds with
stabilizing electron withdrawing groups are a useful tool in
the modification and detection of biomolecules,53 as
demonstrated by the use of ethyldiazoacetate in carboxylation
of dG.36 Next to the O6- position, N7, N3, N2, and N1 represent
nucleophilic sites in dG that are susceptible to modification by
electrophilic reagents.55 N7-CMG and N7-MeG have been
detected after in vitro NOCs exposure, however, these readily
depurinate.25,54 Further, two methylated-dG adducts with
modification other than at the O6-position were detected
previously after exposure of ctDNA to potassium diazoacetate,
and one carboxymethylated-dG adduct other than O6 was
present in colon tumor biopsy samples analyzed by a HRMS2
approach for establishing a diet-related DNA adduct database.8
While evidence for the direct serine adduct could not be
gained from samples of genomic DNA from L-azaserine-
exposed cells, we characterized the formation and persistence
of O6-CMdG and O6-MedG as a function of L-azaserine dose
and time of exposure with a robust and sensitive nanoLC-ESI-
HRMS2 PRM method validated for analysis of cell samples. To
our knowledge, this is the first HRMS approach for the
absolute quantification of O6-MedG and O6-CMdG in
biological samples using stable isotope-labeled internal stand-
ards for accurate quantification of both adducts. A few
methods involving triple quadrupole or ion traps as mass
analyzer exist for the detection and quantification of O6-
relationship due to the overall low level of O6-MedG adducts
formed in HCEC exposed to 500 μM L-azaserine.
DISCUSSION
■
NOCs are well-known carcinogens linked to diet and lifestyle.
The formation of the pro-mutagenic DNA adducts O6-CMdG
and O6-MedG by NOCs is well established;8,10,16,17,27
however, understanding of the chemical mechanism of
NOC-induced O6-CMdG formation and factors that influence
their biological consequences is limited. It has been
hypothesized that the release of diazoacetate from NOCs
and analogs is responsible for the carboxymethylation and
methylation of DNA. However, if the reactive intermediate is
the same for various NOCs, such as N-(N-acetyl-L-prolyl)-N-
nitrosoglycine (APNG) and N-nitrosoglycocholic acid
(NOGC), and for the reactive intermediates L-azaserine and
diazoacetate, it remains unclear why they all result in different
ratios of O6-CMdG to O6-MedG.17,46 Further, the carbox-
ymethylating decomposition products of N-nitroso bile acid
conjugates have not been recovered quantitatively.47 The lack
of characterization of the metabolism of NOCs and of reactive
intermediates raises the question of what are the mechanisms
of carboxymethylation and methylation of DNA by NOCs,
including the possibility of several pathways and intermediates.
Hydrolysis of L-azaserine to L-serine and diazoacetate has
been suggested earlier by Harrison et al.,17 and decomposition
to diazoacetate as the intermediate in O6-CMdG formation has
been demonstrated for N-nitroso-glycine under physiological
conditions.11 However, in this study, we demonstrated that the
carboxymethylating agent L-azaserine undergoes hydrolysis to
L-serine only under acidic conditions and that under
physiologically relevant conditions in the presence of biological
matrices, L-azaserine is stable. The half-life of L-azaserine was
0.4 h at pH 2, and its stability increased with increasing pH,
consistent with an acid-catalyzed hydrolysis mechanism.
Protonation of the diazocarbonyl moiety, possibly at the
carbon bound to both the diazo and carbonyl groups, is
expected to promote hydrolysis of the ester and release L-serine
(Figure 2B). On the basis of the pH dependence of the
protonation/hydrolysis rates and having derived kobs for each
pH value (Table S2), we estimated the ratio khydrolysis/Ka to be
0.90 × 10−4 M−1 s−1 (Figure 2A, eq 5). The pKa of the
protonated L-azaserine (Figure 2A) could not be extracted,
1526
https://doi.org/10.1021/acs.chemrestox.0c00471
Chem. Res. Toxicol. 2021, 34, 1518−1529

Chemical Research in Toxicology
pubs.acs.org/crt
Article
CM(d)G in biological samples,56 also in combination with O6-
Me(d)G.10,27 A HRMS methodology targeting O6-MeG and
O6-CMG and other targeted and untargeted DNA adducts was
successfully applied to colon biopsy samples to establish a diet-
related DNA adducts database and to compare the DNA
adduct profile from white vs red meat in gastrointestinal
digestion samples.8,15 O6-MedG was further one of the targeted
adducts in a reported adductomic approach for nitrosamine-
induced DNA adducts.57 Validation of our nanoLC-ESI-
HRMS2 approach revealed LOQs in matrix were in the low
amol range and comparable to LOQs previously reported for
pure standards on nanoLC-ESI-MS3,27 and improved
compared to a HRMS method targeting four adducts in
biological matrix.8 Applying our methodology to HCEC cell
exposed to L-azaserine, we found a linear response between L-
azaserine dose and O6-CMdG and O6-MedG adduct levels in
HCEC cells (Figure 5). Both adducts were present in
unexposed control samples, though not quantifiable for O6-
CMdG and below one O6-MedG adduct/107 nucleotides. We
measured maximal O6-CMdG adduct levels in cells after 48 h
of exposure, followed by a steady level up to 120 h (Figure 6).
The O6-MedG profile showed a similar trend, but the adduct
levels were too low for statistical analysis. The low O6-MedG
levels observed are consistent with the more potent O6-CMdG
formation by L-azaserine.16,17 Additionally, low levels of O6-
MedG measured in cells might be due to efficient repair of O6-
MedG by O6-methylguanine DNA methyl transferase.58 In a
previous study, a maximum N7-MeG adduct level was achieved
also after 48 h exposure to NDMA, however, in a process
requiring metabolic activation by CYP enzymes.59 Adduct
levels were on the same order of magnitude as previously
reported for colorectal cancer cells exposed to L-azaserine,27
which was more potent in inducing O6-CMdG.27 Endogenous
O6-CMdG and O6-MedG formation has been reported
previously and is likely due to the endogenous formation of
s-adenosylmethionine and NOCs,27,60−62 which account for
45−75% of the total human NOC exposure.19,63
Figure S1: Stability of L-azaserine in the presence of
biological matrices. Figure S2: Formation of O6-Ser-
CMG in an oligonucleotide. Figures S3−S6: Compound
characterizations. Figure S7: Analytical calibration
curves. Figure S8: Acid-catalyzed hydrolysis of L-
azaserine. Figure S9: Cell viability data. Figures S10
and S11: Representative MS2 data from analysis of
genomic DNA samples. Table S1: L-Azaserine decom-
position rates. Table S2: Precision and accuracy of
adduct analysis method (PDF)
AUTHOR INFORMATION
Corresponding Author
■
Shana J. Sturla − Department of Health Science and
Technology, ETH Zurich, 8092 Zurich, Switzerland;
orcid.org/0000-0001-6808-5950; Email: sturlas@
ethz.ch
Authors
Susanne M. Geisen − Department of Health Science and
Technology, ETH Zurich, 8092 Zurich, Switzerland;
orcid.org/0000-0002-9394-0274
Claudia M. N. Aloisi − Department of Health Science and
Technology, ETH Zurich, 8092 Zurich, Switzerland;
orcid.org/0000-0002-2279-7757
Sabrina M. Huber − Department of Health Science and
Technology, ETH Zurich, 8092 Zurich, Switzerland;
orcid.org/0000-0001-9204-5537
Emma S. Sandell − Department of Health Science and
Technology, ETH Zurich, 8092 Zurich, Switzerland;
orcid.org/0000-0002-8066-5293
Nora A. Escher − Department of Health Science and
Technology, ETH Zurich, 8092 Zurich, Switzerland
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.chemrestox.0c00471
Funding
CONCLUSION
This work was supported by the Swiss National Science
Foundation (185020, 186332) and the Swiss Cancer Research
Foundation (KFS-4443-02-2018-R).
■
In this study, we characterized the chemical basis of L-
azaserine-induced O6-CMdG formation. We identified two
mechanisms, occurring under acidic and physiological
conditions, respectively, by which L-azaserine converts into
reactive intermediates and induces O6-CMdG. Our findings
have implications for endogenously formed NOCs that do not
breakdown to diazoacetate under physiological conditions but
can form DNA adducts via nucleophilic attack. We established
a highly sensitive nanoLC-ESI-HRMS2 methodology and
measured a linear dose-response relationship for O6-CMdG
lesions in cells after exposure to L-azaserine. The knowledge of
L-azaserine-induced DNA adduct formation has important
implications to understand O6-CMdG formation and accumu-
lation in cells and to study its biological consequences. Overall,
this work provides fundamental chemical insight concerning
the basis of how L-azaserine may give rise to DNA damage and
supports its use as a chemical probe in NOC-induced
carcinogenesis research, particularly concerning the identifica-
tion of factors that mitigate adverse effects of O6-CMdG.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors wish to thank the staff of the Functional
Genomics Center Zurich (FGCZ) for providing nanoLC-
ESI-HRMS2 instrumentation and technical support, namely
Dr. Bernd Roschitzki, Dr. Peter Gehrig, Dr. Serena di Palma,
Dr. Endre Laczko, and Laura Kunz. We also thank Dr. Hailey
Gahlon for scientific advice and Jonne van Dijk for her
technical support in sample preparation.
REFERENCES
■
(1) Bouvard, V., Loomis, D., Guyton, K. Z., Grosse, Y., Ghissassi, F.
E., Benbrahim-Tallaa, L., Guha, N., Mattock, H., and Straif, K. (2015)
Carcinogenicity of consumption of red and processed meat. Lancet
Oncol. 16 (16), 1599−1600.
(2) Aloisi, C. M. N., Sandell, E. S., and Sturla, S. J. (2021) A
chemical link between meat consumption and colorectal cancer
development? Chem. Res. Toxicol. 34, 12−23.
(3) Turesky, R. J. (2018) Mechanistic Evidence for Red Meat and
Processed Meat Intake and Cancer Risk: A Follow-up on the
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.chemrestox.0c00471.
■
sı
1527
https://doi.org/10.1021/acs.chemrestox.0c00471
Chem. Res. Toxicol. 2021, 34, 1518−1529

Chemical Research in Toxicology
pubs.acs.org/crt
Article
International Agency for Research on Cancer Evaluation of 2015.
Chimia 72 (10), 718−724.
(20) Wu, J., Wang, P., Li, L., Williams, N. L., Ji, D., Zahurancik, W.
J., You, C., Wang, J., Suo, Z., and Wang, Y. (2017) Replication studies
of carboxymethylated DNA lesions in human cells. Nucleic Acids Res.
45 (12), 7276−7284.
(21) Räz, M. H., Dexter, H. R., Millington, C. L., van Loon, B.,
Williams, D. M., and Sturla, S. J. (2016) Bypass of Mutagenic O6-
Carboxymethylguanine DNA Adducts by Human Y- and B-Family
Polymerases. Chem. Res. Toxicol. 29 (9), 1493−503.
(22) Gottschalg, E., Scott, G. B., Burns, P. A., and Shuker, D. E.
(2006) Potassium diazoacetate-induced p53 mutations in vitro in
relation to formation of O6-carboxymethyl- and O6-methyl-2′-
deoxyguanosine DNA adducts: relevance for gastrointestinal cancer.
Carcinogenesis 28 (2), 356−62.
(23) Andreyev, H. J., Norman, A. R., Cunningham, D., Oates, J., Dix,
B. R., Iacopetta, B. J., Young, J., Walsh, T., Ward, R., Hawkins, N.,
Beranek, M., Jandik, P., Benamouzig, R., Jullian, E., Laurent-Puig, P.,
Olschwang, S., Muller, O., Hoffmann, I., Rabes, H. M., Zietz, C.,
Troungos, C., Valavanis, C., Yuen, S. T., Ho, J. W., Croke, C. T.,
O’Donoghue, D. P., Giaretti, W., Rapallo, A., Russo, A., Bazan, V.,
Tanaka, M., Omura, K., Azuma, T., Ohkusa, T., Fujimori, T., Ono, Y.,
Pauly, M., Faber, C., Glaesener, R., de Goeij, A. F., Arends, J. W.,
Andersen, S. N., Lövig, T., Breivik, J., Gaudernack, G., Clausen, O. P.,
De Angelis, P. D., Meling, G. I., Rognum, T. O., Smith, R., Goh, H. S.,
Font, A., Rosell, R., Sun, X. F., Zhang, H., Benhattar, J., Losi, L., Lee, J.
Q., Wang, S. T., Clarke, P. A., Bell, S., Quirke, P., Bubb, V. J., Piris, J.,
Cruickshank, N. R., Morton, D., Fox, J. C., Al-Mulla, F., Lees, N., Hall,
C. N., Snary, D., Wilkinson, K., Dillon, D., Costa, J., Pricolo, V. E.,
Finkelstein, S. D., Thebo, J. S., Senagore, A. J., Halter, S. A., Wadler,
S., Malik, S., Krtolica, K., and Urosevic, N. (2001) Kirsten ras
mutations in patients with colorectal cancer: the ’RASCAL II’ study.
Br. J. Cancer 85 (5), 692−696.
(24) Wang, J., and Wang, Y. (2011) Carboxymethylation of DNA
Induced by N-Nitroso Compounds and Its Biological Implications.
Advances in Molecular Toxicology 5, 219−243.
(25) Shuker, D. E. G., and Margison, G. P. (1997) Nitrosated
Glycine Derivatives as a Potential Source of O6-Methylguanine in
DNA. Cancer Res. 57 (3), 366−369.
(26) O’Driscoll, M., Macpherson, P., Xu, Y.-Z., and Karran, P.
(1999) The cytotoxicity of DNA carboxymethylation and methylation
by the model carboxymethylating agent azaserine in human cells.
Carcinogenesis 20 (9), 1855−1862.
(27) Yu, Y., Wang, J., Wang, P., and Wang, Y. (2016) Quantification
of Azaserine-Induced Carboxymethylated and Methylated DNA
Lesions in Cells by Nanoflow Liquid Chromatography-Nano-
electrospray Ionization Tandem Mass Spectrometry Coupled with
the Stable Isotope-Dilution Method. Anal. Chem. 88 (16), 8036−
8042.
(28) Wang, J., and Wang, Y. (2009) Chemical synthesis of
oligodeoxyribonucleotides containing N3- and O4-carboxymethylth-
ymidine and their formation in DNA. Nucleic Acids Res. 37 (2), 336−
45.
(29) Stock, C. C., Reilly, H. C., Buckley, S. M., Clarke, D. A., and
Rhodas, C. P. (1954) Azaserine, A New Tumour-Inhibitory
Substance: Studies with Crocker Mouse Sarcoma 180. Nature 173
(4393), 71−72.
(30) Longnecker, D. S., and Curphey, T. J. (1975) Adenocarcinoma
of the Pancreas in Azaserine-treated Rats. Cancer Res. 35 (8), 2249−
2257.
(4) Loechler, E. L., Green, C. L., and Essigmann, J. M. (1984) In
vivo mutagenesis by O6-methylguanine built into a unique site in a
viral genome. Proc. Natl. Acad. Sci. U. S. A. 81 (20), 6271−6275.
(5) Pauly, G. T., and Moschel, R. C. (2001) Mutagenesis by O6-
Methyl-, O6-Ethyl-, and O6-Benzylguanine and O4-Methylthymine in
Human Cells: Effects of O6-Alkylguanine-DNA Alkyltransferase and
Mismatch Repair. Chem. Res. Toxicol. 14 (7), 894−900.
(6) Hall, C. N., Badawi, A. F., O’Connor, P. J., and Saffhill, R.
(1991) The detection of alkylation damage in the DNA of human
gastrointestinal tissues. Br. J. Cancer 64 (1), 59−63.
(7) O6-methylguanine in blood leucocyte DNA: an association with
the geographic prevalence of gastric cancer and with low levels of
serum pepsinogen A, a marker of severe chronic atrophic gastritis.
Carcinogenesis 1994, 15 (9), 1815−20.
(8) Hemeryck, L. Y., Decloedt, A. I., Vanden Bussche, J., Geboes, K.
P., and Vanhaecke, L. (2015) High resolution mass spectrometry
based profiling of diet-related deoxyribonucleic acid adducts. Anal.
Chim. Acta 892, 123−131.
(9) Povey, A. C., Hall, C. N., Badawi, A. F., Cooper, D. P., and
O’Connor, P. J. (2000) Elevated levels of the pro-carcinogenic adduct,
O6-methylguanine, in normal DNA from the cancer prone regions of
the large bowel. Gut 47 (3), 362−365.
(10) Vanden Bussche, J., Moore, S. A., Pasmans, F., Kuhnle, G. G.
C., and Vanhaecke, L. (2012) An approach based on ultra-high
pressure liquid chromatography−tandem mass spectrometry to
quantify O6-methyl and O6-carboxymethylguanine DNA adducts in
intestinal cell lines. Journal of Chromatography A 1257, 25−33.
(11) Cupid, B. C., Zeng, Z., Singh, R., and Shuker, D. E. G. (2004)
Detection of O6-Carboxymethyl-2′-deoxyguanosine in DNA Follow-
ing Reaction of Nitric Oxide with Glycine and in Human Blood DNA
Using a Quantitative Immunoslot Blot Assay. Chem. Res. Toxicol. 17
(3), 294−300.
(12) Lewin, M. H., Bailey, N., Bandaletova, T., Bowman, R., Cross,
A. J., Pollock, J., Shuker, D. E. G., and Bingham, S. A. (2006) Red
Meat Enhances the Colonic Formation of the DNA Adduct O6-
Carboxymethyl Guanine: Implications for Colorectal Cancer Risk.
Cancer Res. 66 (3), 1859−1865.
(13) Bussche, J. V., Hemeryck, L. Y., Van Hecke, T., Kuhnle, G. G.
C., Pasmans, F., Moore, S. A., Van de Wiele, T., De Smet, S., and
Vanhaecke, L. (2014) O6-carboxymethylguanine DNA adduct
formation and lipid peroxidation upon in vitro gastrointestinal
digestion of haem-rich meat. Mol. Nutr. Food Res. 58 (9), 1883−1896.
(14) Hemeryck, L. Y., Rombouts, C., Hecke, T. V., Van Meulebroek,
L., Bussche, J. V., De Smet, S., and Vanhaecke, L. (2016) In vitro
DNA adduct profiling to mechanistically link red meat consumption
to colon cancer promotion. Toxicol. Res. (Cambridge, U. K.) 5 (5),
1346−1358.
(15) Hemeryck, L. Y., Rombouts, C., De Paepe, E., and Vanhaecke,
L. (2018) DNA adduct profiling of in vitro colonic meat digests to
map red vs. white meat genotoxicity. Food Chem. Toxicol. 115, 73−87.
(16) Harrison, K. L., Fairhurst, N., Challis, B. C., and Shuker, D. E.
G. (1997) Synthesis, Characterization, and Immunochemical
Detection of O6-(Carboxymethyl)-2′-deoxyguanosine: A DNA
Adduct Formed by Nitrosated Glycine Derivatives. Chem. Res.
Toxicol. 10 (6), 652−659.
(17) Harrison, K. L., Jukes, R., Cooper, D. P., and Shuker, D. E. G.
(1999) Detection of Concomitant Formation of O6-Carboxymethyl-
and O6-Methyl-2′-deoxyguanosine in DNA Exposed to Nitrosated
Glycine Derivatives Using a Combined Immunoaffinity/HPLC
Method. Chem. Res. Toxicol. 12 (1), 106−111.
(18) Stuff, J. E., Goh, E. T., Barrera, S. L., Bondy, M. L., and Forman,
M. R. (2009) Construction of an N-nitroso database for assessing
dietary intake. J. Food Compos. Anal. 22, S42−S47.
(31) Lilja, H. S., Hyde, E., Longnecker, D. S., and Yager, J. D. (1977)
DNA Damage and Repair in Rat Tissues following Administration of
Azaserine. Cancer Res. 37 (11), 3925−3931.
(32) Rosenkrantz, H., Sprague, R., Schaeppi, U., Pittillo, R. T.,
Cooney, D. A., and Davis, R. D. (1972) The stoichiometric fate of
azaserine metabolized in vitro by tissues from azaserine-treated dogs
and mice. Toxicol. Appl. Pharmacol. 22 (4), 607−620.
(19) Mirvish, S. S. (1995) Role of N-nitroso compounds (NOC)
and N-nitrosation in etiology of gastric, esophageal, nasopharyngeal
and bladder cancer and contribution to cancer of known exposures to
NOC. Cancer Lett. 93 (1), 17−48.
(33) Jacquez, J. A., and Sherman, J. H. (1962) Enzymatic
Degradation of Azaserine. Cancer Res. 22 (1), 56−−61.
1528
https://doi.org/10.1021/acs.chemrestox.0c00471
Chem. Res. Toxicol. 2021, 34, 1518−1529

Chemical Research in Toxicology
pubs.acs.org/crt
Article
(34) Longenecker, J. B., and Snell, E. E. (1957) Pyridoxal and metal
ion catalysis of alpha, beta elimination reactions of serine-3-phosphate
and related compounds. J. Biol. Chem. 225 (1), 409−418.
(35) Zhang, L., Kim, S., Jia, G., Buhmeida, A., Dallol, A., Wright, W.
E., Fornace, A. J., Al-Qahtani, M., and Shay, J. W. (2015) Exome
Sequencing of Normal and Isogenic Transformed Human Colonic
Epithelial Cells (HCECs) Reveals Novel Genes Potentially Involved
in the Early Stages of Colorectal Tumorigenesis. BMC Genomics 16
(1), S8.
(36) Geigle, S. N., Wyss, L. A., Sturla, S. J., and Gillingham, D. G.
(2017) Copper carbenes alkylate guanine chemoselectively through a
substrate directed reaction. Chemical Science 8 (1), 499−506.
(37) Reza, F., Bhaswati, G., Pei-Pei, K., Gaffney, B. L., and Jones, R.
A. (1990) Synthesis of 6-substituted 2′-deoxyguanosine derivatives
using trifluoroacetic anhydride in pyridine. Tetrahedron Lett. 31 (3),
319−321.
(53) Mix, K. A., Aronoff, M. R., and Raines, R. T. (2016) Diazo
Compounds: Versatile Tools for Chemical Biology. ACS Chem. Biol.
11 (12), 3233−3244.
(54) Zurlo, J., Curphey, T. J., Hiley, R., and Longnecker, D. S.
(1982) Identification of 7-Carboxymethylguanine in DNA from
Pancreatic Acinar Cells Exposed to Azaserine. Cancer Res. 42 (4),
1286−1288.
(55) Liu, S., and Wang, Y. (2015) Mass spectrometry for the
assessment of the occurrence and biological consequences of DNA
adducts. Chem. Soc. Rev. 44 (21), 7829−7854.
(56) Da Pieve, C., Sahgal, N., Moore, S. A., and Velasco-Garcia, M.
N. (2013) Development of a liquid chromatography/tandem mass
spectrometry method to investigate the presence of biomarkers of
DNA damage in urine related to red meat consumption and risk of
colorectal cancer. Rapid Commun. Mass Spectrom. 27 (21), 2493−503.
(57) Balbo, S., Hecht, S. S., Upadhyaya, P., and Villalta, P. W.
(2014) Application of a High-Resolution Mass-Spectrometry-Based
DNA Adductomics Approach for Identification of DNA Adducts in
Complex Mixtures. Anal. Chem. 86 (3), 1744−1752.
(38) Zhang, P., Chan, W., Ang, I. L., Wei, R., Lam, M. M. T., Lei, K.
M. K., and Poon, T. C. W. (2019) Revisiting Fragmentation Reactions
of Protonated α-Amino Acids by High-Resolution Electrospray
Ionization Tandem Mass Spectrometry with Collision-Induced
Dissociation. Sci. Rep. 9 (1), 6453.
(58) Pegg, A. E. (2000) Repair of O(6)-alkylguanine by
alkyltransferases. Mutat. Res., Rev. Mutat. Res. 462 (2−3), 83−100.
(59) Lin, H.-l., and Hollenberg, P. F. (2001) N-Nitrosodimethyl-
amine-Mediated Formation of Oxidized and Methylated DNA Bases
in a Cytochrome P450 2E1 Expressing Cell Line. Chem. Res. Toxicol.
14 (5), 562−566.
(60) Sharma, V., Collins, L. B., Clement, J. M., Zhang, Z., Nakamura,
J., and Swenberg, J. A. (2014) Molecular Dosimetry of Endogenous
and Exogenous O6-Methyl-dG and N7-Methyl-G Adducts Following
Low Dose [D3]-Methylnitrosourea Exposures in Cultured Human
Cells. Chem. Res. Toxicol. 27 (4), 480−482.
(61) Kang, H., Konishi, C., Kuroki, T., and Huh, N. (1995)
Detection of O6-methylguanine, O4-methylthymine and O4-ethyl-
thymine in human liver and peripheral blood leukocyte DNA.
Carcinogenesis 16 (6), 1277−1280.
(62) De Bont, R., and van Larebeke, N. (2004) Endogenous DNA
damage in humans: a review of quantitative data. Mutagenesis 19 (3),
169−185.
(63) Tricker, A. R. (1997) N-nitroso compounds and man: sources
of exposure, endogenous formation and occurrence in body fluids.
Eur. J. Cancer Prev. 6 (3), 226−268.
(39) Rogalewicz, F., Hoppilliard, Y., and Ohanessian, G. (2000)
Fragmentation mechanisms of α-amino acids protonated under
electrospray ionization: a collisional activation and ab initio
theoretical study. Int. J. Mass Spectrom. 195−196, 565−590.
(40) Piraud, M., Vianey-Saban, C., Petritis, K., Elfakir, C., Steghens,
J.-P., Morla, A., and Bouchu, D. (2003) ESI-MS/MS analysis of
underivatised amino acids: a new tool for the diagnosis of inherited
disorders of amino acid metabolism. Fragmentation study of 79
molecules of biological interest in positive and negative ionisation
mode. Rapid Commun. Mass Spectrom. 17 (12), 1297−1311.
(41) MycoAlert Assay Control Set, Lonza online protocol, https://
bioscience.lonza.com/lonza_bs/CH/en/download/product/asset/
27672 (accessed 2021-5-25).
(42) CellTiter-Glo Luminescent Cell Viability Assay, Promega.https://
ch.promega.com/-/media/files/resources/protocols/technical-
bulletins/0/celltiter-glo-luminescent-cell-viability-assay-protocol.pdf
(accessed 2021-5-25).
(43) Piovesan, A., Pelleri, M. C., Antonaros, F., Strippoli, P.,
Caracausi, M., and Vitale, L. (2019) On the length, weight and GC
content of the human genome. BMC Res. Notes 12 (1), 106.
(44) Roig, A. I., Eskiocak, U., Hight, S. K., Kim, S. B., Delgado, O.,
Souza, R. F., Spechler, S. J., Wright, W. E., and Shay, J. W. (2010)
Immortalized Epithelial Cells Derived From Human Colon Biopsies
Express Stem Cell Markers and Differentiate In Vitro. Gastroenterology
138 (3), 1012−1021.e5.
(45) Nakamura, J., Mutlu, E., Sharma, V., Collins, L., Bodnar, W.,
Yu, R., Lai, Y., Moeller, B., Lu, K., and Swenberg, J. (2014) The
endogenous exposome. DNA Repair 19, 3−13.
(46) Shuker, D. E. G., and Margison, G. P. (1997) Nitrosated
Glycine Derivatives as a Potential Source of O6-Methylguanine in
DNA. Cancer Res. 57 (3), 366−−369.
(47) Shuker, D. E. G., Tannenbaum, S. R., and Wishnok, J. S. (1981)
N-Nitroso bile acid conjugates. 1. Synthesis, chemical reactivity and
mutagenic activity. J. Org. Chem. 46 (10), 2092−2096.
(48) McGarrity, J. F., and Smyth, T. (1980) Hydrolysis of
diazomethane-kinetics and mechanism. J. Am. Chem. Soc. 102 (24),
7303−7308.
(49) Bartz, Q. R., Elder, C. C., Frohardt, R. P., Fusari, S. A., Haskell,
T. H., Johannessen, D. W., and Ryder, A. (1954) Isolation and
characterization of azaserine. Nature 173 (4393), 72−73.
(50) Reglitz, M. (2012) Diazo Compounds: Porperties and Synthesis,
Academic Press, Orlando, FL.
(51) Challis, B. C. (1989) Chemistry and biology of nitrosated
peptides. Cancer Surv. 8 (2), 363−384.
(52) Challis, B. C., Hopkins, A. R., Milligan, J. R., Mitchell, R. C.,
and Massey, R. C. (1984) Nitrosation of peptides. IARC Sci. Publ. 57,
61−70.
1529
https://doi.org/10.1021/acs.chemrestox.0c00471
Chem. Res. Toxicol. 2021, 34, 1518−1529