Detection and characterization of two major ethylated deoxyguanosine adducts by high performance liquid chromatography, electrospray mass spectrometry, and 32P-postlabeling. Development of an approach for detection of phosphotriesters

Article abstract of DOI:10.1021/tx960135b

Postlabeling can be one of the most sensitive methods for the measurement of DNA adducts. However, for the determination of alkylated adducts, essential requirements are standards which must be fully chemically characterized. In order to develop a postlabeling assay for monitoring exposure to genotoxic ethylating agents, the reaction of diethyl sulfate with 2'-deoxynucleoside 3'- and 5'-monophosphates was examined. The adducts generated were fully characterized using HPLC, electrospray tandem mass spectrometry, UV, and postlabeling. The major product was the phosphodiester derived from alkylation of the phosphate, and alkylation of the base occurred to a lesser extent. The phosphodiester standard, 2'-deoxyguanosine 3'-(mono- O-ethyl phosphate) (3'Et-pdG), was used to develop a postlabeling assay for the detection of this adduct in DNA samples. Since alkylated phosphodiesters in DNA are not susceptible to the actions of micrococcal nuclease and calf spleen phosphodiesterase, they can be obtained as alkylated phosphodiester dinucleosides from DNA. Nuclease P₁ was used as an enhancement step which allowed the separation of these adducted phosphotriesters from the unmodified nucleotides by HPLC. Subsequent hydrolysis of the phosphotriester dinucleosides in alkali yielded phosphodiesters, including 3'Et-pdG, which was efficiently postlabeled. This approach was shown to be capable of detecting this adduct in liver DNA from mice treated intraperitoneally with N-nitrosodiethylamine.

Full text of DOI:10.1021/tx960135b

70
Chem. Res. Toxicol. 1997, 10, 70-77
Detection a n d Ch a r a cter iza tion of Tw o Ma jor Eth yla ted
Deoxygu a n osin e Ad d u cts by High P er for m a n ce Liqu id
Ch r om a togr a p h y, Electr osp r a y Ma ss Sp ectr om etr y, a n d
32P -P ostla belin g. Develop m en t of a n Ap p r oa ch for
Detection of P h osp h otr iester s
Rajinder Singh,*,† Gavain M. A. Sweetman,† Peter B. Farmer,*,†
David E. G. Shuker,† and Kim J . Rich‡
MRC Toxicology Unit, Hodgkin Building, University of Leicester, P.O. Box 138, Lancaster Road,
Leicester LE1 9HN, U.K., and Proctor & Gamble (HABC) Ltd., Lovett House, Lovett Road,
Staines, Middlesex TW18 3AZ, U.K.
Received J uly 31, 1996X
Postlabeling can be one of the most sensitive methods for the measurement of DNA adducts.
However, for the determination of alkylated adducts, essential requirements are standards
which must be fully chemically characterized. In order to develop a postlabeling assay for
monitoring exposure to genotoxic ethylating agents, the reaction of diethyl sulfate with 2′-
deoxynucleoside 3′- and 5′-monophosphates was examined. The adducts generated were fully
characterized using HPLC, electrospray tandem mass spectrometry, UV, and postlabeling. The
major product was the phosphodiester derived from alkylation of the phosphate, and alkylation
of the base occurred to a lesser extent. The phosphodiester standard, 2′-deoxyguanosine 3′-
(mono-O-ethyl phosphate) (3′Et-pdG), was used to develop a postlabeling assay for the detection
of this adduct in DNA samples. Since alkylated phosphodiesters in DNA are not susceptible
to the actions of micrococcal nuclease and calf spleen phosphodiesterase, they can be obtained
as alkylated phosphodiester dinucleosides from DNA. Nuclease P1 was used as an enhancement
step which allowed the separation of these adducted phosphotriesters from the unmodified
nucleotides by HPLC. Subsequent hydrolysis of the phosphotriester dinucleosides in alkali
yielded phosphodiesters, including 3′Et-pdG, which was efficiently postlabeled. This approach
was shown to be capable of detecting this adduct in liver DNA from mice treated intraperi-
toneally with N-nitrosodiethylamine.
as an index of exposure. Further sites for DNA modifica-
tion following exposures to simple alkylating agents
include the exocyclic oxygens of guanine and thymine (O6-
alkylguanine and O4-alkylthymine, respectively). Le-
sions at these sites are believed to be promutagenic.
Although in many cases repair enzymes remove the alkyl
groups relatively rapidly from these bases, long term
exposure may result in steady state levels of some of
these adducts (13).
Alkylating agents also react with the phosphate oxy-
gens in DNA, yielding phosphotriesters (14). For simple
alkylating agents, these adducts tend to be stable and
not rapidly repaired, and hence are long-lived (15, 16).
Mutation resulting from phosphotriester formation has
In tr od u ction
The formation of DNA adducts is believed to be a
critical event in the initiation of carcinogenesis (1, 2). The
measurement of DNA adducts has consequently become
an increasingly popular approach for studying chemical
mutagenesis and carcinogenesis and for indicating ex-
posure to genotoxic compounds such as alkylating agents
(3, 4). Exposure of man to these compounds occurs
principally in the environment (5, 6) but may also occur
following chemotherapy for treatment of tumors (7). In
spite of the numerous reports on the detection of alky-
lated adducts in both nontarget and target organs (8),
few studies have been designed to assess their value in
predicting interindividual susceptibility to tumor forma-
tion. At the moment, DNA adducts are being used
primarily as highly specific and sensitive indices of
exposure in epidemiological studies (9, 10), the advantage
of this approach being that they account for interindi-
vidual variations in carcinogen absorption, distribution,
metabolism, and excretion.
1
Abbreviations: N-7 EtG, N-7-ethylguanine; N-7 Et 3′pdG, N-7-
ethyl-2′-deoxyguanosine 3′-monophosphate; 3′pdG, 2′-deoxyguanosine
3′-monophosphate; 3′pdN, 2′-deoxynucleoside 3′-monophosphate; 3′Et-
pdG, 2′-deoxyguanosine 3′-mono(O-ethyl phosphate); 3′Et-pdN, 2′-
deoxynucleotide 3′-mono(O-ethyl phosphate); 3′Et-pdGp, 2′-deoxygua-
nosine; 3′-(O-ethyl phosphate) 5′-phosphate; 3′Et-pdNp, 2′-deoxy-
nucleoside 3′-(O-ethyl phosphate) 5′-phosphate; N-7 Et 5′pdG, N-7-
ethyl-2′-deoxyguanosine 5′-monophosphate; 5′pdG, 2′-deoxyguanosine
5′-monophosphate; 5′pdN, 2′-deoxynucleoside 5′-monophosphate; 5′Et-
pdG, 2′-deoxyguanosine 5′-mono(O-ethyl phosphate); 5′Et-pdN, 2′-
deoxynucleoside 5′-mono(O-ethyl phosphate); pdGp, 2′-deoxyguanosine
3′,5′-bisphosphate; pdNp, 2′-deoxynucleoside 3′,5′-bisphosphate; 2′dN,
2′-deoxynucleoside; dNpdN, 2′-deoxynucleoside-(3′,5′)-2′-deoxynucleo-
side; dNp(Et)dN, 2′-deoxynucleoside-(3,′5′)-2′-deoxynucleoside mono-
(O-ethyl phosphate); DES, diethyl sulfate; ES-MS, electrospray mass
spectrometry; ES-MS/MS, electrospray tandem mass spectrometry;
MN, micrococcal nuclease; CSPD, calf spleen phosphodiesterase; PNK,
polynucleotide T4 kinase, NDEA, N-nitrosodiethylamine; TEAF, tri-
ethylammonium formate; TEA, triethylammonium acetate; AF, am-
monium formate; Pi, inorganic phosphate.
One of the principal lesions produced in vivo following
exposure to simple monoalkylating agents occurs at the
nucleophilic N-7 position of guanine (11). This adduct
is not considered to be promutagenic; nevertheless, it is
relatively slowly repaired (12) and thus is widely used
* To whom correspondence should be addressed.
† University of Leicester.
‡ Proctor & Gamble.
X Abstract published in Advance ACS Abstracts, December 15, 1996.
S0893-228x(96)00135-X CCC: $14.00 © 1997 American Chemical Society

32P-Postlabeling of Ethyl Phosphotriesters in DNA
Chem. Res. Toxicol., Vol. 10, No. 1, 1997 71
not been fully addressed, but such adducts may influence
cellular function by altering the binding of proteins to
DNA through neutralization of the negative charge along
the sugar-phosphate backbone. This may affect chro-
matin structure, leading to DNA strand breaks. Phos-
photriester adducts hold considerable potential as valid
surrogates for monitoring exposure to carcinogens and
perhaps for risk assessment (16, 17).
N-Nitrosodiethylamine (NDEA)1 is a simple alkylating
agent that requires metabolic activation to generate its
ethylating electrophile (18, 19), which is capable of
interacting with DNA at many different sites (20). This
compound is currently being used as a model carcinogen
to investigate the predictive value of adducts for carci-
nogenic risk assessment in an animal model. In order
to achieve this objective, we required sensitive assays for
ethylated DNA, and in particular, for two of the products
formed with 2′-deoxyguanosine 3′-monophosphate (3′pdG),
N-7-ethyl-2′-deoxyguanosine 3′-monophosphate (N-7 Et
3′pdG), and 2′-deoxyguanosine 3′-mono(O-ethyl phos-
phate) (3′Et-pdG).
Numerous assays have been used to detect alkyl DNA
adducts, such as high performance liquid chromatogra-
phy (HPLC)/immunoassay (21), HPLC/fluorescence (22),
HPLC/electrochemical detection (23, 24), or immunoslot
blot or ELISA detection (25-27). Recent advances
include the development of fluorescent postlabeling tech-
niques where the adduct of interest is derivatized with
highly fluorescent compounds for detection by HPLC (28).
Mass spectrometry, though more useful as a tool for the
structural characterization of adducts, has been employed
to determine levels of DNA adducts (29). Although these
methods have been successful, they may require rela-
tively large quantities of DNA for analysis or rely upon
specific antibodies and are often limited in their sensitiv-
ity. The technique which gives potentially the highest
degree of sensitivity for detection of alkylated DNA
adducts and which has the additional advantage of only
requiring a few micrograms of DNA is 32P-postlabeling
(30). Coupling of 32P-postlabeling with HPLC separation
of adducts results in a highly specific detection method
(31). However 32P-postlabeling of N-7 alkyl 3′pdG is not
so successful because of its unstable nature, due to
spontaneous depurination (32-34), and consequent low
labeling efficiency (35). Attempts have been made to
circumvent this problem by labeling the ring-opened
derivatives (36). A further consideration for postlabeling
of a low molecular weight adduct such as ethyl is that
the introduction of an enhancement step (e.g., HPLC or
immunoaffinity purification) to separate the adduct from
similar unmodified 2′-deoxynucleotides may result in
some loss of adduct.
describes the characterization of two major adducts
derived from reaction of 3′pdG with the model ethylating
agent diethyl sulfate (DES), N-7 Et 3′pdG, and 3′Et-pdG.
Conditions were optimized for their detection by HPLC
and 32P-postlabeling, and labeling efficiencies were de-
termined. The possibility that 3′Et-pdG would be liber-
ated following enzymic digestion of DNA in the postla-
beling procedure was also investigated. A new approach
for the detection of alkyl phosphotriesters in DNA was
developed, and the possibility for the formation of ethyl
phosphotriesters in DNA in vivo was investigated in mice
treated intraperitoneally with NDEA.
Ma ter ia ls a n d Meth od s
Ch em ica ls. 2′-Deoxynucleoside 3′- or 5′-monophosphate
(3′pdN or 5′pdN), ATP, DES, dithiothreitol, nuclease P1, sper-
midine, micrococcal nuclease (MN), calf thymus DNA (Type 1),
and potato apyrase, grade VI, were purchased from the Sigma
Chemical Co. Ltd. (Poole, Dorset, U.K.). Ca u tion : DES is a
mutagen and carcinogen. Protective clothing should be worn
and appropriate safety procedures followed when working with
this compound. Polynucleotide T4 kinase (3′-phosphatase free)
(PNK) and calf spleen phosphodiesterase (CSPD) were pur-
chased from Boehringer Mannheim (Lewes, East Sussex, U.K.).
[γ-32P]ATP, >5000 Ci/mmol, was purchased from Amersham
International (Amersham, Buckinghamshire, U.K.). Ca u tion :
Appropriate procedures should be used to minimize exposure to
γ - radiation when using [γ-32P] ATP. All other reagents and
solvents of analytical grade were purchased from either Merck
Ltd. (Poole, Dorset, U.K.) or Fisher Scientific Ltd. (Loughbor-
ough, Leicestershire, U.K.). Authentic N-7 ethylguanine (N-7
EtG) was kindly provided by Professor M. J arman (Institute
of Cancer Research, Sutton, Surrey, U.K.).
Ethylated nucleotide standards were generated by reacting
the 2′-deoxynucleotide (2 mg) with 100 mM DES in 0.5 M
sodium phosphate buffer (pH 6.0) at room temperature for 8 h
as reported previously (35) and were then separated by HPLC.
Calf thymus DNA (ca. 50 mg) was ethylated by treatment with
DES (1-100 mM) in 0.5 M sodium phosphate buffer (pH 6.0)
at room temperature for 8 h and precipitated by the addition of
3 M sodium acetate (1:10, v/v) and 2-propanol (0.8 volume).
HP LC Sep a r a t ion P r oced u r es. HPLC was routinely
performed using LKB (Pharmacia Biotech Ltd., Milton Keynes,
U.K.) or Gilson (Gilson Medical Electronics Inc., Middleton, WI)
pumps connected to a fixed wavelength detector or diode array
UV detector. Absorbance of eluate was monitored at 254 nm.
All analyses were performed at room temperature, using a
reverse-phase C18 column (Apex, 5 µm; 4.6 × 250 mm; J ones
Chromatography) fitted with a C18 guard column (Techsphere,
5 µm; 3 × 10 mm; HPLC Technology). The flow rate was 1 mL/
min. Three different buffered mobile phase systems were
utilized for gradient elution as described below:
(1) Triethylammonium formate (TEAF). Solvent A: methanol/
0.1 M TEAF (1:99 v/v) (pH 6.0); solvent B: methanol/0.1 M
TEAF (20:80 v/v) (pH 6.0); gradient: 0 min: 0% B, 20 min: 45%
B, 28 min: 100% B, 35 min: 100% B, 40 min: 0% B. (2)
Triethylammonium acetate (TEA). Solvent A: 0.1 M TEA (pH
7.0); solvent B: acetonitrile; gradient: 0 min: 1% B, 20 min:
1% B, 45 min: 2% B, 55 min: 95% B, 60 min: 1% B. (3)
Ammonium formate (AF). Solvent A: 0.05 M AF, (pH 5.4);
solvent B: methanol; gradient: 0 min: 0% B, 15 min: 20% B,
20 min: 0% B.
HPLC fractions were collected, and either evaporated to
dryness using a DNA centrifugal evaporator (DNA 110, Savant,
Farmingdale, NY) or freeze-dried (IEC Hyoprep 3000, Interna-
tional Equipment Co., Dunstable, U.K.) for 32P-postlabeling or
mass spectrometry analysis.
In contrast to the situation for N-7-alkylguanine, there
have only been limited attempts made to analyze alkyl
phosphotriesters in DNA (37), although the potential for
32P-postlabeling of such adducts has, however, recently
been demonstrated by Saris et al. (38). Phosphate
alkylation in DNA will yield a phosphotriester, which is
not a substrate for postlabeling. However, alkaline
hydrolysis of alkyl phosphotriesters yields 3′-phosphodi-
esters. One of the products is the 2′-deoxynucleoside 3′-
mono(O-alkyl phosphate), which may be postlabeled and
could be used as a marker of the extent of phosphate
alkylation.
An a lysis of N-7-Eth ylgu a n in e by HP LC-Electr och em i-
ca l Detection . The DES-treated calf thymus DNA samples
were hydrolyzed in 0.1 M formic acid (pH 2.3) at 70 °C for 1 h
to release adducted bases. An enhancement step was performed
For development of postlabeling assays, it is necessary
to obtain standards of the adduct of interest, which
should be fully structurally characterized. This paper

72 Chem. Res. Toxicol., Vol. 10, No. 1, 1997
Singh et al.
Ta ble 1. Com p a r ison of HP LC P h a se System s for th e
Sep a r a tion of Eth yla ted Gu a n in e P r od u cts Exp r essed a s
Ca p a city Ra tiosa
on the sample using reverse-phase HPLC (AF mobile phase) to
separate N-7 EtG from guanine and adenine, which were also
generated by the hydrolysis conditions utilized. The HPLC
fraction containing N-7 EtG was lyophilized, redissolved in
water, and analyzed by electrochemical detection (Antec EC
detector, Antec Analytical Technology, Leiden, The Nether-
lands). The optimum oxidation potential of 1.10 V for the
detector cell was determined by chromatographing a standard
sample of N-7 EtG over a range of settings. N-7 EtG was
isocratically separated on a reverse-phase C18 column (Hypersil,
5 µm; 4.6 × 250 mm; Shandon HPLC) using 25 mM potassium
phosphate (pH 6.0)/methanol (90:10 v/v). Determination of the
levels of N-7 EtG following the treatment of calf thymus DNA
with various concentrations of DES demonstrated a linear dose-
dependent increase in the level of adduct produced, and
subsequently these samples were used in the development of a
postlabeling method for 3′Et-pdG. The limit of detection of the
method was 6 N-7 EtG adducts per 106 nucleotides (2.0 pmol)
per 100 µg of DNA.
capacity ratio (k′)
mobile phase
nucleotide/base
TEAF
TEA
AF
N-7 Et 3′pdG
3′Et-pdG
N-7 Et 5′pdG
5′Et-pdG
6.24
11.48
4.44
10.20
8.16
11.48
18.96
5.64
19.00
12.96
4.00
5.80
3.88
4.96
6.19
N-7 EtG
a
(tR - t0)/t0, where tR is the retention time of the ethylated
product and t0 is the elution time of an unretained component.
Ma ss Sp ectr om etr y. Mass spectrometry was carried out
using a Quattro BQ tandem quadrupole instrument (Micromass,
Manchester, U.K.) fitted with an electrospray source operating
in the negative ion mode. The carrier solvent used in all the
experiments was acetonitrile and water (1:1 v/v) maintained at
a constant flow rate of 50 µL/min. Dried HPLC fractions were
resuspended in 20 µL of water prior to MS analysis. Samples
were introduced into the solvent flow using a Rheodyne 7125
injector, fitted with a 10 µL loop, and the instrument (MS1) was
scanned over a mass range of 300-450. Product ion spectra
were obtained with a collision energy of 110 eV and argon
collision gas at a pressure of 2.0 × 10-4 mbar and scanning MS2
over a mass range of 50-400.
P ostla belin g of N-7 Et 3′p d G a n d 3′Et-p d G w ith Non -
r a d ioa ctive ATP . Labeling of N-7 Et 3′pdG and 3′Et-pdG with
cold ATP was carried out by incubation with labeling buffer (12
µL, 200 mM Tris-HCl, 100 mM MgCl2, 100 mM dithiothreitol,
10 mM spermidine, pH 7.6), 2 mM ATP (15 µL), HPLC grade
water (6 µL), and PNK (2 µL, 10 U/µL) at 37 °C for 1 h. The
total mixture (55 µL) was separated by using the TEAF mobile
phase. The HPLC fraction containing the ethylated 2′-deoxy-
nucleoside bisphosphate product of 3′Et-pdG was evaporated to
dryness, redissolved in 15 µL of HPLC grade water, and
incubated with nuclease P1 (5 µL, 2 U/µL) at 37 °C for 30 min
prior to HPLC.
F igu r e 1. Typical reverse-phase HPLC chromatograms ob-
tained for the reaction of (A) 3′pdG (5.76 µmol) and (B) 5′pdG
(5.76 µmol) with 100 mM DES in 0.5 M sodium phosphate
buffer, pH 6.0 (1.0 mL). A 50 µL aliquot of the reaction mixture
was separated using the TEA mobile phase. (I, N-7 Et 5′pdG;
II, 5′pdG; III, N-7 EtG; IV, 5′Et-pdG).
32P -P ostla belin g. La belin g Efficien cies of N-7 Et 3′p d G
a n d 3′Et-p d G. The labeling efficiency of each adduct, dissolved
in HPLC grade water (5 µL), was investigated by incubation
with 500 fmol of 3′pdG (as the internal standard, 1 µL), labeling
buffer (4 µL, 200 mM Tris-HCl, 100 mM MgCl2, 100mM
dithiothreitol, 10 mM spermidine, pH 7.6), 10 µCi of [γ-32P]ATP
(specific activity >5000 Ci/mmol), and PNK (1 µL, 10 U/µL) for
1 h at 37 °C. The postlabeling mixture was then applied to
prewashed (with water) poly(ethylenimine) (PEI) cellulose TLC
plates (20 × 20 cm, Macherey-Nagel, Camlab, Cambridge, U.K.).
One-dimensional chromatography was performed using 1 M
ammonium formate (pH 3.5). Radiolabeled spots were visual-
ized by storage phosphor image analysis (ImageQuant software
(v. 3.3) and Phosphoimager, Molecular Dynamics, Sunnyvale,
CA) and adduct levels quantified as (fmol of 3′pdG) × (screen
volume postlabeled N-7 Et 3′pdG or 3′Et-pdG/screen volume
postlabeled 3′pdG).
mM dithiothreitol, 10 mM spermidine, pH 7.6), [γ-32P]ATP (10
µCi, specific activity >5000 Ci/mmol), and PNK (1 µL, 10 U/µL)
were added, and the mixture was incubated at 37 °C for 1 h.
Following a further incubation with apyrase (4 µL, 40 mU/µL),
the postlabeled mixture was spotted onto cellulose-coated glass
plates (20 × 20 cm, 0.1 mm, UV 254 nm, Macherey-Nagel,
Camlab, Cambridge) at a position 2 cm from each edge in the
lower left-hand corner. These plates were prespotted with an
unlabeled UV marker 5′pdG which runs in the same position
as the postlabeled 3′Et-pdG adduct. The plates were developed
in the first dimension (D1: isobutyric acid/water/concentrated
NH4OH, 66:20:1 v/v), rotated through 90°, and developed
similarly in the second dimension (D2: saturated (NH4)2SO4/
2-propanol/1 M sodium acetate, 80:2:18 v/v) (39). Each chro-
matographic run took 7-8 h for the solvent front to reach the
top of the plate. The UV marker was visualized under UV light
(254 nm) and the 32P-postlabeled 3′Et-pdG adduct detected by
storage phosphor image analysis.
Deter m in a tion of 3′Et-p d G in DNA Sa m p les. DNA (50
µg) extracted from animal tissue or calf thymus DNA was
incubated with MN (5 µL, 0.4 U/µL), CSPD, (35 µL, 1 mU/µL),
and digestion buffer (10 µL, 100 mM sodium succinate, 50 mM
CaC12, pH 6.0) at 37 °C overnight. The mixture was further
incubated with nuclease P1 (30 µL, 2 U/µL) at 37 °C for 6 h and
then with 25% ammonium hydroxide solution (70 µL) for 24 h
at 70°C. The fraction corresponding to 3′Et-pdG was collected,
following HPLC separation of the reaction mixture using 50 mM
ammonium formate/methanol gradient and evaporation to dry-
ness, and 32P-postlabeled after resuspension in 5 µL of water.
Labeling buffer (5 µL, 200 mM Tris-HCl, 100 mM MgCl2, 100
Resu lts
Syn th esis an d Ch ar acter ization of Eth ylated Gu a-
n in e Deoxyn u cleotid es. The ultimate requirement to
carry out 32P-postlabeling necessitated the generation of
ethylated standards from 3′pdG. However, for reasons
of greater accessibility of the 2′-deoxynucleotide, initial
optimization experiments of the HPLC and ES-MS
methods used standards generated from 5′pdG. The

32P-Postlabeling of Ethyl Phosphotriesters in DNA
Chem. Res. Toxicol., Vol. 10, No. 1, 1997 73
F igu r e 3. Typical HPLC chromatograms obtained for the
postlabeling of 3′Et-pdG (tR: 31.2 min) using nonradioactive
ATP (s), and after nuclease P1 treatment (- - -). The incubation
mixture was separated using the TEAF mobile phase. (X denotes
the peak observed for enzyme/labeling buffer blank.)
ES-MS analysis of peaks I and IV gave ions of m/ z
374 (M - H)- for both products, consistent with the mass
of ethylated derivatives of 5′pdG. The site of ethylation
for each of these products was further investigated by
ES-MS/MS.
Peak I gave a parent ion of m/ z 374 (M - H)- and
product ions of m/ z 195 ([sugar + phosphate] - H)-, 178
([ethylated guanine] - H)-, 97 (H2PO4)-, and 79 (PO3)-
consistent with an ethylated guanine moiety, probably
at the N-7 position (Figure 2A). ES-MS/MS does not
allow an assignation site for the ethyl group on the
guanine base. To confirm that the ethyl moiety was at
the N-7 position, the product was depurinated by heating
to 70 °C. The heat-treated fraction was subsequently
shown to coelute with an authentic N-7 EtG standard,
and further analysis of this HPLC fraction by ES-MS
confirmed its identity as that of an ethylated guanine
species (data not shown). Treatment of the product in
peak I with 30% ammonium hydroxide in water for 24 h
at room temperature produced the imidazole ring-opened
N-7 Et 5′pdG, as confirmed by ES-MS analysis, m/ z 392
(M - H)- (Figure 2C). However, this treatment was not
sufficient to allow complete conversion to the ring-opened
form, as indicated by the presence of an ion of m/ z 374
in the ES-MS spectrum. The identity of peak I was
further confirmed by performing UV absorption spec-
troscopy at neutral, acidic, and alkaline pH, resulting in
the predictable change of λmax from 254 to 266 nm at
alkaline pH (40).
F igu r e 2. (A) Typical ES-MS/MS spectrum obtained for peak
I, N-7 Et 5′pdG; fragments m/ z 178 and 97 (M - H)- represent
the product ions of ethylated guanine and the phosphate group,
respectively, formed from the molecular ion m/ z 374 of N-7 Et
5′pdG. (B) Typical ES-MS/MS spectrum obtained for peak IV,
5′Et-pdG; fragments m/ z 150 and 125 (M - H)- represent the
product ions of guanine and the ethylated phosphate group,
respectively, formed from the molecular ion m/ z 374 of 5′Et-
pdG. (C) ES-MS spectrum of peak I following imidazole ring-
opening with ammonium hydroxide/water (30:70 v/v) for 24 h;
ions m/ z 392 and 374 (M - H)- represent the molecular ions of
imidazole ring-opened N-7 Et 5′pdG (2′-deoxy-N5-ethyl-N5-
formyl-2,5,6-triamino-4-oxopyrimidine 5′-monophosphate) and
unreacted N-7 Et 5′pdG, respectively.
retention times of analogous 5′pdG or 3′pdG products
were different using the three different mobile phase
systems (Table 1).
Two major and one minor product resulted from the
reaction of DES with 3′pdG, as shown in the HPLC
chromatogram (Figure 1A). Analogous products were
produced from 5′pdG (Figure 1B). Peak II from the latter
reaction mixture was confirmed as unreacted starting
material following ES-MS analysis (m/ z 346) and by
coelution with an authentic standard. Peak I was
identified as N-7-ethyl 2′-deoxyguanosine 5′-monophos-
phate (N-7 Et 5′pdG); peak III as the spontaneous
depurination product of peak I, N-7 EtG; peak IV as 2′-
deoxyguanosine 5′-mono(O-ethyl phosphate) (5′Et-pdG).
The variation in retention time of N-7 EtG observed
between these two experiments (Figures 1A and 1B) was
within the standard deviation for repeated analyses
carried out using this system. In each case, the structure
of N-7 EtG was confirmed by ES-MS and cochromatog-
raphy with the authentic standard.
The ES-MS/MS analysis of peak IV from the reaction
of 5′pdG with DES gave a parent ion of m/ z 374 and
product ions of m/ z 150 ([guanine] - H)- and m/ z 125
(C2H6PO4)- consistent with ethylation of the phosphate
group of the nucleotide (Figure 2B).
32P -P ostla belin g of N-7 Et 3′p d G a n d 3′Et-p d G.
Analogous analytical approaches were applied to the
products from 3′pdG. Initial attempts to postlabel (with
nonradioactive ATP) and perform ES-MS on N-7 Et
3′pdG were unsuccessful. Although ATP was dephos-
phorylated to ADP by PNK as evidenced by HPLC of the
reaction mixture, there was no peak corresponding to
either N-7 ethyl-2′-deoxyguanosine 3′,5′-bisphosphate or

74 Chem. Res. Toxicol., Vol. 10, No. 1, 1997
Singh et al.
F igu r e 6. Typical HPLC chromatogram of the products formed
when DNA is digested with MN and CSPD, followed by a further
incubation with nuclease P1 and hydrolyzed in 25% ammonium
hydroxide (as shown in Scheme 1). Products were separated
using the AF mobile phase: 0.0 min: 0% B, 15.0 min: 20% B,
20.0 min: 0% B, 25.0 min: 100% B, 30.0 min: 0% B. The arrow
indicates the retention time of 3′Et-pdG (X denotes an unidenti-
fied peak).
F igu r e 4. Typical storage phosphor image following the one-
dimensional TLC separation of 32P-postlabeled 3′Et-pdG, with
3′pdG as the internal standard. The amounts of 3′Et-pdG labeled
are shown at the bottom of the image. Adducts were separated
using 1 M ammonium formate (pH 3.5); 15 min exposure.
Sch em e 1
F igu r e 5. Standard curve for determination of 3′Et-pdG by
32P-postlabeling using 500 fmol of 3′pdG as the internal stan-
dard.
quent treatment of this 2′-deoxyguanosine bisphosphate
with nuclease P1, which acts as a 3′-phosphatase, resulted
in the formation of the unethylated 2′-deoxynucleoside
5′-monophosphate. The structural identity of this was
confirmed by ES-MS, which gave an ion at m/ z 346 (M
- H)- representing the molecular ion of 5′pdG, thus
unequivocally demonstrating original ethylation of the
3′-phosphate group (Figure 3).
Incubations of DES with the 2′-deoxynucleotides of the
other three DNA bases yielded similar results to those
obtained with 3′ or 5′pdG, in that the major product of
reaction was the ethylated phosphodiester (data not
shown).
Once characterized, the two 3′pdG adduct standards
were investigated for their postlabeling efficiencies using
3′pdG as an internal standard. The concentration of N-7
Et 3′pdG was determined by HPLC-electrochemical
detection following neutral depurination at 70 °C for 1
h. The concentration of the 3′Et-pdG standard was
determined using a standard curve of 3′pdG, where
of the parent compound. The parent compound had
depurinated during evaporation in the DNA centrifugal
evaporator, probably as a result of the increase in
temperature generated during centrifugation and of the
apparent greater chemical instability of the 3′ relative
to the 5′ derivative. Freeze-drying of samples thereafter
rectified this problem. Postlabeling of N-7 Et 3′pdG with
nonradioactive ATP gave one major product which was
separated by HPLC from unreacted ATP and ADP
produced as a result of the transfer of the terminal
phosphate group from ATP to the nucleotide. This had
a parent ion of m/ z 454 (M - H)- and product ions at
m/ z 97 and 79 when further analyzed by ES-MS/MS
consistent with an ethylated 2′-deoxyguanosine bispho-
sphate product with ethylation of the guanine moiety.
Postlabeling of 3′Et-pdG with cold ATP gave one major
product with an HPLC retention time of 26.0 min and a
molecular ion of m/ z 454 upon ES-MS analysis. Subse-

32P-Postlabeling of Ethyl Phosphotriesters in DNA
Chem. Res. Toxicol., Vol. 10, No. 1, 1997 75
F igu r e 7. Typical storage phosphor images following the two-dimensional TLC separation of 32P-postlabeled: (A) HPLC fraction
corresponding to authentic 3′Et-pdG; 20 min exposure. (B) HPLC fraction corresponding to 3′Et-pdG (17.0-19.0 min) derived from
untreated calf thymus DNA. The dotted circle represents the position of the UV marker (unlabeled 5′pdG) which comigrates with
postlabeled 3′Et-pdG; 20 min exposure. (C) HPLC fraction (19.0-21.0 min) derived from untreated calf thymus DNA; 30 min exposure.
(D) HPLC fraction corresponding to 3′Et-pdG (17.0-19.0 min, Figure 6) derived from calf thymus DNA treated with 100 mM DES;
20 min exposure; x, y, and z denote unidentified spots. (E) HPLC fraction following the 3′Et-pdG fraction (19.0-21.0 min, Figure 6)
derived from the same DNA sample; 20 min exposure; t, u, w, and x denote unidentified spots.
absorbance was measured at 253 nm. The postlabeled
mixture was applied to prewashed PEI cellulose TLC
plates and developed one-dimensionally in 1 M am-
monium formate, pH 3.5 (3′Et-pdG plate depicted in
Figure 4). Radiolabeled spots were visualized and quan-
titated by storage phosphor image analysis. The labeling
efficiency for 3′ Et-pdG was determined to be 71.3 ( 2.3%
compared to 3′pdG, the standard curve being shown in
Figure 5. The labeling efficiency for N-7 Et 3′pdG could
not be determined because the postlabeled adduct was
undetectable using N-7 Et 3′pdG in the range 170-679
fmol.
from compounds of unknown identity appear (spots w,
u, and t); however, one spot appears to be unchanged
from the previous fraction (spot x). None of these spots
was present in postlabeling maps from corresponding
fractions from untreated DNA (Figures 7B and 7C). The
method has initially been applied to tissues at various
time points generated from an animal study in which
mice were dosed with NDEA, with results showing the
presence of 3′Et-pdG (Figures 8A and 8B). This adduct,
which has so far not been quantified, is detectable up to
56 days following dosing in Balb/c and SWR mice.2
Detection of 3′Et-p d G in DNA. Attempts to detect
3′Et-pdG in calf thymus DNA treated with 100 mM DES
following digestion with MN and CSPD, by using a two-
step HPLC purification procedure to remove unmodified
nucleotides, followed by postlabeling, were unsuccessful.
This was expected, since the 3′Et-pdG adduct is not
generated following digestion of DNA with MN and
CSPD alone, which generates ethylated 2′-deoxynucleo-
side phosphotriesters. The approach therefore followed
for analysis of 3′Et-pdG involved digestion of DNA with
MN and CSPD to 3′pdNs, incubation with nuclease P1,
and alkaline hydrolysis according to Scheme 1. Figure
6 shows a typical HPLC elution profile of the products
from such an approach, and Figure 7D shows the typical
TLC images that are obtained for alkylated DNA follow-
ing postlabeling of the HPLC fraction corresponding to
3′Et-pdG marked with an arrow in Figure 6. Postlabel-
ing of a fraction collected following the 3′Et-pdG fraction
(Figure 7E) reveals a different pattern. The spot for 3′Et-
pdG was not present, confirming that this was not an
artifact from the postlabeling process. Three new spots
Discu ssion
The major products formed when DES reacts with
3′pdN have been shown to be the phosphodiesters, e.g.,
3′Et-pdG, which has been chemically characterized using,
in particular, ES-MS/MS. For reactions with 3′pdG, the
ring substituted product N-7 Et 3′pdG was similarly
characterized, and postlabeling conditions were estab-
lished using routine procedures.
Alkylated phosphotriesters are not substrates for PNK,
and consequently, they have to be converted into com-
pounds that are substrates, namely, phosphodiester
dinucleosides or 2′-deoxynucleoside 3′-mono(O-alkyl phos-
phates). Saris et al. (38) have attempted to determine
the whole range of adducted phosphotriesters present in
DNA following exposure to alkylating agents. Their
method relies on the conversion of the adducted phos-
phodiester dinucleosides to normal dinucleosides and this
may be disadvantageous, since contamination from nor-
2
Singh et al., unpublished results.

76 Chem. Res. Toxicol., Vol. 10, No. 1, 1997
Singh et al.
In addition to 3′Et-pdG, one would theoretically expect
to find 16 dinucleoside products as well as the ethylated
phosphodiester nucleosides for thymine, adenine, and
cytosine being generated in our analytical procedure.
Since the elution times for these products are not known,
a 100% methanol wash step is included at the end of the
HPLC gradient, ensuring no carry-over of these products
into the next run. In the TLC images that were obtained
for alkylated DNA following postlabeling of the HPLC
fraction corresponding to 3′Et-pdG (Figure 7D), and the
subsequent fraction (Figure 7E), there were unidentified
spots which could represent the reaction products de-
scribed above. This suggestion is supported by the
observation that these spots are only present after
treatment of DNA with the ethylating agent (Figures
7B-7E). Further work clearly is required to determine
the yield of 3′Et-pdG following each of the steps in the
method, for instance, by using a radiolabeled standard
of 3′Et-pdG. The possibility of using an internal standard
also needs to be investigated, as this would allow more
accurate quantitation of the adduct and define the
sensitivity of the method. We have demonstrated that
3′Et-pdG has a labeling efficiency that is very much
greater than that for N-7 Et 3′pdG. The low labeling
efficiency of N-7 Et 3′pdG may be explained by chemical
instability of the adduct and loss by depurination during
the postlabeling procedure, or by it being a poor substrate
for PNK. Similar results were obtained by Koivisto and
Hemminki (35) who determined a labeling efficiency of
1.5% for N-7 Et 3′pdG. The adduct 3′Et-pdG shows both
chemical stability and persistence in vivo, its presence
being determined in DNA extracted from the livers of
mice treated with NDEA. Although the biological con-
sequences of phosphotriester formation remain unknown,
the determination of this lesion may provide valuable
information on earlier exposure to genotoxic compounds.
F igu r e 8. Typical storage phosphor images following the two-
dimensional TLC separation of 32P-postlabeled: (A) liver DNA
from Balb/c mice, 5 h after ip injection with saline; 30 min
exposure; (B) liver DNA from Balb/c mice, 24 h after ip injection
with NDEA (90 mg/kg); 30 min exposure.
mal dinucleosides may occur if these are not adequately
resolved from the adducted phosphodiesters during the
workup procedures of the method. We have used nu-
clease P1 treatment as an enhancement step for alkylated
phosphodiesters. Unmodified nucleotides and base alky-
lated nucleotide adducts are converted to their corre-
sponding 2′-deoxynucleosides following treatment, ren-
dering them no longer substrates for phosphorylation by
PNK. This eliminates contamination of the adduct
fraction by unmodified nucleotides that occurs as a result
of peak tailing, which is normally a problem encountered
when HPLC is used as an enhancement step for the
separation of alkylated adducts from unmodified nucle-
otides, due to their similar chromatographic properties.
As 3′Et-pdG will not be generated from ethylated DNA
by digestion with MN and CSPD, the overall extent of
ethylation of DNA may be seriously underestimated by
simply measuring by postlabeling the extent of base
alkylation, especially since many alkylating agents such
as nitrosamines tend to ethylate the phosphate backbone
of DNA to a greater extent than the base of the nucleotide
(11).
A further advantage of studying 3′Et-pdG as a marker
for ethylation damage of DNA is that it is a very stable
adduct and has been shown to withstand heating to 100
°C (16), in contrast to adducts formed at the N-7 position
of guanine, which are very labile and can easily be lost
during the lengthy workup procedures involved in post-
labeling. In comparison to the O6-ethylguanine adduct,
a well-characterized repair mechanism does not exist for
3′Et-pdG; hence measurement of the latter provides a
longer-term marker of exposure.
Ack n ow led gm en t. The authors acknowledge finan-
cial support from the Medical Research Council and Dow
Chemical Co.
Refer en ces
(1) Phillips, D. H. (1985) Chemical Carcinogenesis. In The Molecular
Basis of Cancer (Farmer, P. B., and Walker, J ., Eds.) pp 133-
179, Croom Helm, London.
(2) Cooper, D. P., O’Connor, P. J ., Povey, A. C., and Rafferty, J . A.
(1995) Cell and Molecular Mechanisms in Chemical Carcinogen-
esis. In Oxford Textbook of Oncology (Peckham, M., Pinedo, H.,
and Veronesi, U., Eds.) Vol. 1, pp 135-147, Oxford University
Press, Oxford.
(3) Poirier, M. C., and Beland, F. A. (1992) DNA adduct measure-
ments and tumor incidence during chronic carcinogen exposure
in animal models: implications for DNA adduct-based human
cancer risk assessment. Chem. Res. Toxicol. 5, 749-755.
(4) Saffhill, R., Margison, G. P., and O’Connor, P. J . (1985) Mecha-
nism of carcinogenesis induced by alkylating agents. Biochim.
Biophys. Acta 823, 111-145.
(5) Bartsch, H., and Montesano, R. (1984) Commentary: Relevance
of nitrosamines to human cancer. Carcinogenesis 5, 1381-1393.
(6) Greim, H., Csanady, G., Filser, J . G., Kreuzer, P., Schwarz, L.,
Wolff, T., and Werner, S. (1995) Biomarkers as tools in human
health risk assessment. Clin. Chem. 41, 1804-1808.
(7) Farmer, P. B. (1994) Metabolism and reaction of alkylating
agents. In Anticancer Drugs: Reactive Metabolism and Drug
Interactions (Powis, G., Ed.) pp 1-77, Pergamon Press, Oxford.
(8) Swenberg, J . A., Hoel, D. G., and Magee, P. N. (1991) Mechanistic
and statistical insight into the large carcinogenesis bioassays on
N-nitrosodiethylamine and N-nitrosodimethylamine. Cancer Res.
51, 6409-6414.
(9) Wogan, G. N., and Gorelick, N. J . (1985) Chemical and biochemi-
cal dosimetry of exposure to genotoxic chemicals. Environ. Health
Perspect. 62, 5-18.

32P-Postlabeling of Ethyl Phosphotriesters in DNA
Chem. Res. Toxicol., Vol. 10, No. 1, 1997 77
(10) Wogan, G. N. (1992) Molecular epidemiology in cancer risk
assessment and prevention: Recent progress and avenues for
future research. Environ. Health Perspect. 98, 167-178.
(11) Beranek, D. T. (1990) Distribution of methyl and ethyl adducts
following alkylation with monofunctional alkylating agents. Mu-
tat. Res. 231, 11-30.
(12) Margison, G. P., and Pegg, A. E. (1981) Enzymatic release of
7-methylguanine from methylated DNA by rodent liver extracts.
Proc. Natl. Acad. Sci U.S.A. 78, 861-865.
(13) Boucheron, J . A., Richardson, F. C., Morgan, P. H., and Swenberg,
J . A. (1987) Molecular dosimetry of O4-ethyldeoxythymidine in
rats continuously exposed to diethylnitrosamine. Cancer Res. 47,
1577-1581.
(14) Shooter, K. V. (1978) DNA phosphotriesters as indicators of
cumulative carcinogen-induced damage. Nature 274, 612-614.
(15) Weinfield, M., Drake, A. F., Saunders, J . K., and Paterson, M. C.
(1985) Stereospecific removal of methyl phosphotriesters from
DNA by an Escherichia coli ada+ extract. Nucleic Acids Res. 13,
7067-7077.
(16) Bannon, P., and Verly, W. (1972) Alkylation of phosphates and
stability of phosphate triesters in DNA. Eur. J . Biochem. 31, 103-
111.
(17) Tates, A. D., Neuteboom, I., De Vogel, N., and Den Engelse, L.
(1983) The induction of chromosomal damage in rat hepatocytes
and lymphocytes. Mutat. Res. 107, 131-151.
(18) Shuker, D. E. G., and Bartsch, H. (1994) DNA adducts of
nitrosamines. In DNA Adducts: Identification and Biological
Significance (Hemminki, K., Dipple, A., Shuker, D. E. G., Kad-
lubar, F. F., Segerback, D., and Bartsch, H., Eds.) IARC Scientific
Publication No. 125, pp 73-89, International Agency for Research
on Cancer, Lyon.
urea. An immuno-slot-blot study. Carcinogenesis 9, 147-151.
(26) Degan, P., Montesano, R., and Wild, C. P. (1988) Antibodies
against 7-methyldeoxyguanosine: Its detection in rat peripheral
blood lymphocyte DNA and potential applications to molecular
epidemiology. Cancer Res. 48, 5065-5070.
(27) Van Delft, J . H. M., Van Winden, M. J . M., Van Den Ende, A. M.
C., and Baan, R. A. (1993) Determining N7-alkylguanine adducts
by immunochemical methods and HPLC with electrochemical de-
tection: applications in animal studies and in monitoring human
exposure to alkylating agents. Environ. Health Perspect. 99, 25-
32.
(28) Shuker, D. E. G., Durand, M.-J ., and Molko, D. (1993) Fluorescent
postlabeling of modified DNA bases. In Postlabelling Methods for
Detection of DNA Adducts (Phillips, D. H., Castegnaro, M., and
Bartsch, H., Eds.) IARC Scientific Publication No. 124, pp 227-
232, International Agency for Research on Cancer, Lyon.
(29) Farmer, P. B., and Sweetman, G. M. A. (1995) Mass spectrometric
detection of carcinogen adducts. J . Mass Spectrom. 30, 1369-
1379.
(30) Randerath, K., and Randerath, E. (1994) 32P-Postlabeling methods
for DNA adduct detection: overview and critical evaluation. Drug
Metab. Rev. 26, 67-85.
(31) Gorelick, N. J . (1993) Application of HPLC in the 32P-postlabeling
assay. Mutat. Res. 288, 5-18.
(32) Muller, N., and Eisenbrand, G. (1985) The influence of N7
substituents on the stability of N7-alkylated guanosines. Chem.-
Biol. Interact. 53, 173-181.
(33) Kato, K., Petruzzelli, S., Bowman, E. D., Turteltaub, K. W.,
Blomeke, B., Weston, A., and Shields, P. G. (1993) 7-Alkyldeox-
yguanosine adduct detection by two-step HPLC and the 32P-
postlabeling assay. Carcinogenesis 14, 545-550.
(19) Camus, A.-M., Geneste, O., Honkakoski, P., Bereziat, J .-C.,
Henderson, C. J ., Wolf, C. R., Bartsch, H., and Lang, M. A. (1993)
High variability of nitrosamine metabolism among individuals:
role of cytochromes P450 2A6 and 2E1 in the dealkylation of
N-nitrosodimethylamine and N-nitrosodiethylamine in mice and
humans. Mol. Carcinog. 7, 268-275.
(20) Preussmann, R., and Wiessler, M. (1987) The enigma of the organ-
specificity of carcinogenic nitrosamines. Trends Pharmacol. Sci.
8, 185-189.
(21) Wild, C. P., Smart, G., Saffhill, R., and Boyle, J . M. (1983)
Radioimmunoassay of O6-methyldeoxyguanosine in DNA of cells
alkylated in vitro and in vivo. Carcinogenesis 4, 1605-1609.
(22) Herron, D. C., and Shank, R. C. (1980) Methylated purines in
human liver DNA after probable dimethylnitrosamine poisoning.
Cancer Res. 40, 3116-3117.
(23) De Groot, A. J . L., J ansen, J . G., Van Valkenburg, C. F. M., and
Van Zeeland, A. A. (1994) Molecular dosimetry of 7-alkyl- and
O6-alkylguanine in DNA by electrochemical detection. Mutat. Res.
307, 61-66.
(34) Haque, K., Cooper, D. P., and Povey, A. C. (1994) Optimization
of 32P-postlabeling for the quantification of O6-methyl and N7-
methyldeoxyguanosine-3′-monophosphates in human DNA. Car-
cinogenesis 15, 2485-2490.
(35) Koivisto, P., and Hemminki, K. (1990) 32P-postlabeling of 2-hy-
droxyethylated, ethylated and methylated adducts of 2′-deoxy-
guanosine 3′-monophosphate. Carcinogenesis 11, 1389-1392.
(36) Hemminki, K. (1989) Ring-opened 7-methylguanine nucleotides
are resistant to nuclease P1 digestion and good substrates to
polynucleotide kinase. Carcinogenesis 10, 1761-1763.
(37) Shooter, K. V., and Merrifield, R. K. (1980) Analysis of mam-
malian DNA for the presence of carcinogen-induced phospho-
triesters: application of the technique of difference sedimentation.
Anal. Biochem. 103, 110-117.
(38) Saris, C. P., Damman, S. J ., Van den Ende, A. M. C., Westra, J .
G., and Den Engelse, L. (1995) A 32P-postlabeling assay for the
detection of alkylphosphotriesters in DNA. Carcinogenesis 16,
1543-1548.
(39) Wilson, V. L., Smith, R. A., Autrup, H., Krokan, H., Musci, D. E.,
Le, N., Longoria, J ., Ziska, D., and Harris, C. C. (1986) Genomic
5-methylcytosine determination by 32P-postlabeling analysis.
Anal. Biochem. 152, 275-284.
(24) Bianchini, F., Montesano, R., Shuker, D. E. G., Cuzick, J ., and
Wild, C. P. (1993) Quantification of 7-methyldeoxyguanosine using
immunoaffinity purification and HPLC with electrochemical
detection. Carcinogenesis 14, 1677-1682.
(25) Ludeke, B. I., and Kleihues, P. (1988) Formation and persistence
of O6-(2-hydroxyethyl)-2′-deoxyguanosine in DNA of various rat
tissues following a single dose of N-nitroso-N-(2-hydroxyethyl)-
(40) Singer, B. (1972) Reaction of guanosine with ethylating agents.
Biochemistry 11, 3939-3947.
TX960135B