Measurement of oxidative damage at pyrimidine bases in γ-irradiated DNA

Article abstract of DOI:10.1021/tx960095b

Oxidized nucleobases represent one of the main classes of damage induced in DNA by ionizing radiation. Emphasis was placed in this work on the measurement of four oxidized pyrimidine bases, including 5- (hydroxymethyl)uracil (5-HMUra), 5-formyluracil (5-ForUra), 5-hydroxy- cytosine (5-OHCyt), and 5-hydroxyuracil (5-OHUra), in isolated DNA upon exposure to γ radiation in aerated aqueous solution. For this purpose, both high performance liquid chromatography associated with electrochemical detection (HPLC-EC) and gas chromatography coupled to mass spectrometry (GC- MS) were used. Conditions of hydrolysis of the N-glycosidic bond were carefully checked in order to achieve a quantitative release of the lesions. We showed that 60% formic acid treatment leads to the decomposition of the four lesions studied. On the other hand, hydrolysis based on the use of either 88% formic acid or 70% hydrogen fluoride in pyridine (HF/Pyr) allowed the quantitative release of the modified bases, with the exception of 5- HMUra when the latter reagent was utilized. A dose course study of the radiation-induced formation of 5-HMUra and 5-ForUra in DNA by using the GC- MS assay showed that the latter lesion was produced in a 2.1-fold higher yield than the former one. HF/Pyr and 88% formic acid hydrolysis provided similar results for 5-ForUra, indicating the reliability of both techniques for the measurement of this lesion. For 5-OHUra and 5-OHCyt, the level of modification determined by GC-MS analysis was higher after 88% formic acid treatment than upon HF/Pyr hydrolysis. When DNA was enzymatically digested and analyzed by HPLC-EC for 5-OHdCyd and 5-OHdUrd, the results were very close to those obtained by GC-MS following HF/Pyr treatment. It was concluded that additional amounts of both 5-OHUra and 5-OHCyt are produced during the 88% formic acid treatment from radiation-induced 5,6-saturated pyrimidine precursors. It is likely that cytosine and uracil diols are involved in this reaction. The radiochemical yields of formation (in μmol · J^-1) for the products studied are in the following decreasing order: 5-ForUra (0.0083) > 5-OHCyt (0.0046) > 5-HMUra (0.0039) > 5-OHUra (0.0035).

Full text of DOI:10.1021/tx960095b

Chem. Res. Toxicol. 1996, 9, 1145-1151
1145
Mea su r em en t of Oxid a tive Da m a ge a t P yr im id in e Ba ses
in γ-Ir r a d ia ted DNA
Thierry Douki, Thierry Delatour, Fre´de´rique Paganon, and J ean Cadet*
CEA/ De´partement de Recherche Fondamentale sur la Matie`re Condense´e, SCIB/ LAN,
F-38054 Grenoble Cedex 9, France
Received J une 13, 1996X
Oxidized nucleobases represent one of the main classes of damage induced in DNA by ionizing
radiation. Emphasis was placed in this work on the measurement of four oxidized pyrimidine
bases, including 5-(hydroxymethyl)uracil (5-HMUra), 5-formyluracil (5-ForUra), 5-hydroxy-
cytosine (5-OHCyt), and 5-hydroxyuracil (5-OHUra), in isolated DNA upon exposure to γ
radiation in aerated aqueous solution. For this purpose, both high performance liquid
chromatography associated with electrochemical detection (HPLC-EC) and gas chromatography
coupled to mass spectrometry (GC-MS) were used. Conditions of hydrolysis of the N-glycosidic
bond were carefully checked in order to achieve a quantitative release of the lesions. We showed
that 60% formic acid treatment leads to the decomposition of the four lesions studied. On the
other hand, hydrolysis based on the use of either 88% formic acid or 70% hydrogen fluoride in
pyridine (HF/Pyr) allowed the quantitative release of the modified bases, with the exception
of 5-HMUra when the latter reagent was utilized. A dose course study of the radiation-induced
formation of 5-HMUra and 5-ForUra in DNA by using the GC-MS assay showed that the latter
lesion was produced in a 2.1-fold higher yield than the former one. HF/Pyr and 88% formic
acid hydrolysis provided similar results for 5-ForUra, indicating the reliability of both
techniques for the measurement of this lesion. For 5-OHUra and 5-OHCyt, the level of
modification determined by GC-MS analysis was higher after 88% formic acid treatment than
upon HF/Pyr hydrolysis. When DNA was enzymatically digested and analyzed by HPLC-EC
for 5-OHdCyd and 5-OHdUrd, the results were very close to those obtained by GC-MS following
HF/Pyr treatment. It was concluded that additional amounts of both 5-OHUra and 5-OHCyt
are produced during the 88% formic acid treatment from radiation-induced 5,6-saturated
pyrimidine precursors. It is likely that cytosine and uracil diols are involved in this reaction.
The radiochemical yields of formation (in µmol‚J -1) for the products studied are in the following
decreasing order: 5-ForUra (0.0083) > 5-OHCyt (0.0046) > 5-HMUra (0.0039) > 5-OHUra
(0.0035).
formation of 8-oxo-7,8-dihydro-2′-deoxyadenosine (2), 5-hy-
droxy-2′-deoxycytidine (5-OHdCyd), and 5-hydroxy-2′-
deoxyuridine (5-OHdUrd) (3) in DNA exposed to γ
radiation. Other approaches, based either on HPLC
analysis with UV detection of enzymatically digested
DNA (4, 5) or on detection of bases released by repair
enzymes (6, 7) have been proposed to measure oxidized
bases in irradiated DNA. However, they are either not
sensitive enough or not quantitative. Gas chromatogra-
phy analysis coupled to mass spectrometry is another
powerful tool for the study of the formation of oxidized
DNA bases. It has been applied to the measurement of
radiation-induced base damage to DNA in both in vitro
(8-10) and in vivo experiments (11). However, recent
results have shown that artifactual oxidation of normal
DNA bases takes place during the silylation step aimed
at preparing volatile derivatives for GC analysis (12-
14). It should be added that most of the early GC-MS
measurements of radiation-induced oxidized DNA bases
were carried out without using isotopically labeled in-
ternal standards, which are required for an accurate GC-
MS analysis. Altogether, this prompts us to reconsider
the data previously obtained by GC-MS analyses of OH•
mediated oxidation of pyrimidine bases. The measure-
ment of four characteristic oxidized bases was acheived
by using an improved version of the GC-MS assay.
Emphasis was placed on the optimization of the condi-
tions of DNA hydrolysis in order to obtain quantitative
In tr od u ction
The indirect effect of ionizing radiation which is mostly
mediated by OH• radicals gives rise to strand breaks and
base lesions as the main classes of DNA damage. The
bulk of the oxidative damage to purine and pyrimidine
bases has been characterized either in model compounds
or in isolated DNA. However, the quantitative aspect of
the formation of most of these lesions remains to be
established mostly because of a lack of sensitive and
reliable methods of measurement. High performance
liquid chromatography (HPLC)1 associated with electro-
chemical detection (EC) allowed the measurement of
8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodGuo) in iso-
lated DNA exposed to ionizing radiation (1, 2). Applica-
tion of HPLC-EC has been extended to the study of the
* Corresponding author: J ean Cadet, CEA, De´partement de Re-
cherche Fondamentale sur la Matie`re Condense´e, SCIB/LAN, 17
avenue des Martyrs, F-38054 Grenoble Cedex 9, France; tel (33) 76
88 49 87; fax (33) 76 88 50 90; e-mail cadet@drfmc.ceng.cea.fr.
X Abstract published in Advance ACS Abstracts, September 1, 1996.
1
Abbreviations: 5-FordUrd: 5-formyl-2′-deoxyuridine; 5-ForUra:
5-formyluracil; 5-HMdUrd: 5-(hydroxymethyl)-2′-deoxyuridine; 5-HM-
Ura: 5-(hydroxymethyl)uracil; 5-OHCyt: 5-hydroxycytosine; 5-OH-
dCyd: 5-hydroxy-2′-deoxycytosine; 5-OHdUrd: 5-hydroxy-2′-deoxyuri-
dine; 5-OHUra: 5-hydroxyuracil; 8-oxoAde: 8-oxo-7,8-dihydroadenine;
8-oxodGuo: 8-oxo-7,8-dihydro-2′-deoxyguanosine; 8-oxoGua: 8-oxo-7,8-
dihydroguanine; GC-MS: gas chromatography coupled to mass spec-
trometry; HF/Pyr: 70% w/w solution of hydrogen fluoride in pyridine;
HPLC-EC: high performance liquid chromatography with electro-
chemical detection.
S0893-228x(96)00095-1 CCC: $12.00 © 1996 American Chemical Society

1146 Chem. Res. Toxicol., Vol. 9, No. 7, 1996
Douki et al.
addition of 50 µL of chloroform. The samples were centrifuged
and the aqueous layers collected. Another aliquot fraction (200
µL) of the irradiated DNA solution was freeze-dried and the
resulting residue suspended in 1 mL of 88% v/v formic acid to
which isotopically labeled internal standards (300 pmol) were
added. The resulting solutions were placed in sealed glass vials
and held at 140 °C for 45 min in an aluminum heating block.
They were subsequently dried under vacuum. An additional
aliquot fraction (300 µL) of the DNA solution was dried under
vacuum in polypropylene vial. HF/Pyr (100 µL) was added and
the sample was held at 37 °C for 45 min. Then, the hydrolysis
mixture was poured down in a suspension of 300 mg of calcium
carbonate in 1 mL of water. The mixture was shaken until the
pH reached 7. The resulting suspension was centrifuged and
the aqueous phase collected. The solid fraction was washed with
1 mL of a [1:1] mixture of water and ethanol. The liquid phase
obtained was combined with the first one, and the resulting
solution was freeze-dried. The whole irradiation and hydrolysis
procedure was done in triplicate for each dose. Time course
studies of the hydrolysis were performed using DNA samples
which were exposed to 75 Gy. Hydrolysis experiments using
60% formic acid were carried out in a similar way as with 88%
formic acid.
F igu r e 1. Chemical structure of the oxidative lesions studied.
results. The modified bases studied include 5-(hydroxy-
methyl)uracil (5-HMUra) and 5-formyluracil (5-ForUra)
which result from the OH• mediated oxidation of the
methyl group of thymine (5). In addition, 5-hydroxycyto-
sine (5-OHCyt) and 5-hydroxyuracil (5-OHUra), which
arise from the addition of OH• to the C5-C6 double bond
of cytosine, were also measured (Figure 1). The detection
of the two latter lesions was also made by HPLC-EC
following enzymatic digestion. The role of cytosine and
uracil diols in the detection of 5-hydroxylated cytosine
derivatives was investigated through a study of the
stability of 5,6-dihydroxy-5,6-dihydro-2′-deoxyuridine
(dUrd diols) under various hydrolysis conditions.
The samples were then injected on
a HPLC apparatus
(Gilson, Middletown, WI) consisting of two pumps (Model 306)
controlled by a third one (Model 305). Samples dissolved in 1
mL of water were injected using a 231 XL autoinjector, and
fractions of interest were collected on a FC204 fraction collector.
The elution was monitored by a 111B UV spectrometer set at
280 nm connected to a D2500 integrator (Hitachi, Tokyo, J apan).
The system was equipped with a home-packed semipreparative
(250 × 7 mm i.d., particle size: 10 µm) Nucleosil 100-10 C18
octadecylsilyl silica gel column (Macherey Nagel, Du¨ren, Ger-
many). The elution started with a 25 mM solution of am-
monium formate in water for 10 min. Methanol was linearly
added over 15 min until its proportion reaches 20%. The latter
composition was then kept for 5 min. The flow rate was 3
mL‚min-1. The two fractions eluting between 4 and 5 min on
one hand, and 5 and 8 min on the other hand, were collected
and dried under vaccum. Water was added (350 µL) and the
resulting solution transferred into silylation vials. They were
then freeze-dried overnight. Samples were subsequently de-
rivatized for 20 min in 100 µL of a [50:50] v/v mixture of
acetonitrile (silylation grade, Pierce, Rockford, IL) and N-(tert-
butyldimethylsilyl)-N-methyltrifluoroacetamide (Fluka, Buchs,
Switzerland) at 110 °C. GC-MS analyses were performed on a
HP 5890 Serie II gas chromatograph (Hewlett-Packard, Les
Ulis, France) equipped with a capillary column (0.25 mm, 30
m) coated with a 0.25 µm film of methylsiloxane substituted by
5% phenylsiloxane (HP5-MS, Hewlett-Packard). The constant
Ma ter ia ls a n d Meth od s
Ch em ica ls. 5-HMUra, 5-(hydroxymethyl)-2′-deoxyuridine
(5-HMdUrd) and calf thymus DNA were purchased from Sigma
(St Louis, MO). 5-ForUra and the 70% solution of hydrogen
fluoride in pyridine (HF/Pyr) were provided by Aldrich (Stein-
heim, Germany). 5-OHUra (isobarbituric acid) was obtained
from Fluka (Buchs, Switzerland). 5-Formyl-2′-deoxyuridine (5-
FordUrd) was synthesized by menadione photosensitization of
thymidine (15). 5-OHCyt was prepared by using the method
reported by Moshel and Behrman (16). 5-OHdCyd and 5-
OHdUrd were prepared by collidine treatment of 5-bromo-2′-
deoxycytidine and 5-bromo-2′-deoxyuridine, respectively. The
latter nucleosides were obtained by addition of bromine to a
solution of 2′-deoxycytidine and 2′-deoxyuridine, respectively.
Condensation of [1,3-15N]uracil with [2H2]paraformaldehyde
(MSD Isotopes, Montreal, Canada) provided [1,3-15N,5-2H2]-5-
HMUra (15). An aliquot fraction of the latter product was
subsequently oxidized into [1,3-15N,5-2H]-5-ForUra (17). [2-
13C,1,3-15N]-5-OHCyt was prepared as described for 5-OHCyt,
using [2-13C,1,3-15N]Cyt obtained from [13C,15N2]urea (Eurisotop,
Saint Aubin, France) according to Bendich et al. (18). [1,3-15N,5-
18O]-5-OHUra was synthesized from [1,3-15N]uracil by bromi-
nation followed by collidine treatment in [18O]H2O (Eurisotop,
St. Aubin, France). The two cis diastereoisomers of dUrd diols
were prepared by permanganate oxidation of an aqueous
solution of 2′-deoxyuridine. The diols were purified by reverse
phase HPLC and characterized by 1H NMR and GC-MS
analyses.
Ir r a d ia tion a n d Hyd r olysis of Isola ted DNA. A 0.7
mg‚mL-1 solution (3 mL) of calf thymus DNA was exposed under
constant air bubbling to the γ rays of a 60Co source placed in a
pool. The dose rate was 45 Gy‚min-1 as determined by poly-
(methyl methacrylate) dosimetry. An aliquot fraction (100 µL)
was incubated with 10 µL of nuclease P1 buffer (300 mM sodium
acetate, 1 mM ZnSO4, pH 5.3) and 10 U of nuclease P1
(Boerhinger, Mannheim, Germany). The samples were held at
37 °C for 2 h. Then, dephosphorylation of the resulting
nucleotides was achieved by addition of 12 µL of alkaline
phosphatase buffer (500 mM Tris-HCl, 1 mM EDTA, pH 8) and
1 U of alkaline phosphatase (Boerhinger, Mannheim, Germany).
After incubation for 1 h at 37 °C, proteins were precipitated by
flow rate was 1.6 mL‚min-1
. The injection (injection volume:
1 µL) was performed in the splitless mode with the temperature
of the injection port set at 250 °C. The temperature of the GC
oven was maintained at 70 °C for 1 min and then raised to 300
°C at a rate of 20 °C‚min-1, and finally left at the latter
temperature for 3.5 min. Detection of positive ions was provided
by a HP 5972 mass detector using electron impact ionization
(Hewlett-Packard, Les Ulis, France). Three (tert-butyldimeth-
ylsilyl)-N-methyl- (tBDMS) groups were added to 5-OHUra,
5-OHCyt, and 5-HMUra while 5-ForUra was derivatized into
its bisilylated derivative. The ion at m/ z ) M - 57 (M - butyl)
was the base peak in all mass spectra and was thus used for
the quantitative analysis, performed in the ion collection mode.
The retention times were 9.9, 10.9, 11.1, and 11.3 min for the
silylation products of 5-ForUra, 5-OHUra, 5-HMUra, and
5-OHCyt, respectively.
Hyd r olysis of th e Nu cleosid es. A solution containing
5-HMdUrd, 5-FordUrd, 5-OHdCyd, and 5-OHdUrd (1 nmol of
each) was freeze-dried in polypropylene tubes. HF/Pyr (100 µL)
was added and the sample held at 37 °C for increasing periods
of time. The solution was poured down into a suspension of
300 µg of calcium carbonate in 1 mL of water. Mixture of the
isotopically labeled bases (300 pmol of 5-OHUra, 5-OHCyt,

Radiation-Induced Pyrimidine Lesions in DNA
Chem. Res. Toxicol., Vol. 9, No. 7, 1996 1147
5-HMUra, and 5-ForUra) was added. The mixture was then
shaken until pH reached 7. The suspension was centrifuged
and the liquid phase collected. The solid fraction was washed
with 500 µL of a [1:1] mixture of ethanol and water. The two
liquid phases were mixed and freeze-dried. The resulting
samples were then prepurified by HPLC and analyzed by GC-
MS as reported above for DNA. Other aliquot fractions of the
solution of the four nucleosides were freeze-dried in glass vials,
and 500 µL of either 60% or 88% formic acid was added. The
vials were sealed and held at 140 °C for increasing periods of
time. Formic acid was removed under vacuum, and 300 pmol
of the isotopically labeled bases dissolved in 1 mL of water were
added. The samples were then prepurified by HPLC and
analyzed by GC-MS as reported above for DNA. HF/Pyr and
88% formic acid hydrolysis of the dUrd diols (250 µg) was
performed for 45 min in a similar way, without adding isoto-
pically labeled internal standard. A third sample of dUrd diols
was incubated for 2 h in the nuclease P1 buffer and then 1 h in
the alkaline phosphatase buffer used in the enzymatic digestion
(vide supra). The hydrolyzed dUrd diol samples were analyzed
by reverse phase HPLC using a Interchrom HC18-25F octade-
cylsilyl silica gel column (250 × 4.6 mm i.d.) (Interchim,
Montluc¸on, France) and a 25 mM ammonium formate aqueous
solution as the isocratic eluent. The detection was provided by
a Waters 990 diode array UV spectrometer. Under these
conditions, the retention time of the two diastereoisomers of
dUrd diols was 4.12 and 4.68 min. Their UV absorption spectra
only exhibited a maximum at 210 nm. Pure 5-OHUra (tR: 4.3
min) and 5-OHdUrd (tR: 13.1 min) were injected on the HPLC
system in order to determine their retention times. Their UV
spectra exhibited an additional absorption maximum at 280 nm.
In HF/Pyr hydrolyzed samples, a compound exhibiting a reten-
tion time of 3.4 min was identified as cis-5,6-dihydroxy-5,6-
dihydrouracil (Ura diol). Hydrolyzed dUrd diols samples were
also analyzed by GC-MS in the SCAN mode following trimeth-
ylsilylation. The respective retention times of the two hexa-
silylated diastereoisomers of dUrd diols and trisilylated 5-O-
HUra were 10.99, 11.40, and 7.97 min. The mass spectra of
the two silylated diastereoisomers of dUrd diols exhibited
pseudomolecular ions at m/ z ) 622. The tetrasilylated Ura diol
exhibited a retention time of 7.97 min. A pseudomolecular ion
at m/ z ) 434 was observed in its EI mass spectrum. In all
mass spectra the ion at m/ z ) M - 15 (M - methyl) was
observed together with characteristic fragments.
F igu r e 2. Time course study of acidic hydrolysis of oxidized
nucleosides: (a) 5-OHdCyd and (b) 5-HMdUrd. Hydrolysis
conditions: 60% fa ) 60% formic acid, 140 °C; 88% fa ) 88%
formic acid, 140 °C; HF/Pyr ) 70% hydrogen fluoride in pyridine,
37 °C.
standards required by the GC-MS analysis were added
at the end of the hydrolysis in order to estimate the
stability of the compounds. For all analyses of DNA,
internal standards were added prior to the hydrolysis in
order to compensate for any loss of product which could
be a major drawback when working at the trace level. It
should also be added that hydrolyzed samples were
purified by HPLC prior to GC-MS analysis in order to
remove remaining nucleosides from the solution. Indeed,
the N-glycosidic bond of most of the pyrimidine nucleo-
sides is cleaved during the silylation, leading to the
release of the corresponding base moiety. In addition,
the salts produced during the neutralization step of the
HF/Pyr hydrolysis have to be removed in order to avoid
interferences during the GC-MS analysis. Similar ob-
servations were made regarding the stability of 5-OHUra,
5-ForUra, and 5-OHCyt (Figure 2a for the latter lesion).
The latter products underwent a fast decomposition in
60% formic acid but were stable upon 88% formic acid
and HF/Pyr treatments. With the two latter hydrolyzing
reagents, the total amount of nucleoside (1 nmol) was
recovered as free base, which is indicative of a complete
hydrolysis. 5-HMUra (Figure 2b) was stable and quan-
titatively released upon treatment with 88% formic acid
but not with HF/Pyr. The amount of 5-HMUra detected
in samples of 5-HMdUrd hydrolyzed in 60% formic was
only half of the total amount of nucleoside treated. This
indicates that, in spite of the plateau observed for the
release of 5-HMUra during the time course study, the
compound is also unstable under the latter hydrolysis
conditions. It is worth mentioning that the four pyrimi-
dine oxidation products studied are also unstable at the
base level in 60% formic acid (data not shown). Similar
conclusions were drawn when samples of DNA exposed
to 75 Gy of γ radiation in aqueous aerated solution were
hydrolyzed for increasing periods of time in the presence
of internal standards. In 60% formic acid, the amount
of the four oxidized bases detected in the hydrolysate
decreased with hydrolysis time. On the other hand, the
HP LC-EC An a lysis. The HPLC-EC system consisted of a
model 2150 LKB pump (Pharmacia LKB Biotechnology, Upp-
sala, Sweden) connected to a SIL-9A autosampler (Shimadzu,
Kyoto, J apan) equipped with a Interchrom HC18-25F octa-
decylsilyl silica gel column (250 × 4.6 mm i.d.) (Interchim,
Montluc¸on, France). The isocratic eluent was a mixture of 50
mM sodium citrate (pH 5) containing 1.5% of methanol. The
coulometric detection was provided by a Coulochem II detector
(Esa, Chelmsford, MA). The detection potentials were set at
+200 and +450 mV for electrodes 1 and 2, respectively. The
retention times of 5-OHdCyd and 5-OHdUrd were 10.5 and 12.9
min, respectively. Unmodified nucleosides were monitored by
a Waters 484 UV detector set at 280 nm. Both signals were
collected on a D7500 Hitachi (Tokyo, J apan) integrator. The
amount of DNA analyzed was inferred from the area of the peak
of 2′-deoxyguanosine (tR: 45 min) after appropriate calibration.
Resu lts
Stu d y of th e Hyd r olysis Step . A preliminary study
was made to assess the quantitative aspects of the acidic
hydrolysis step prior to GC-MS analysis for the four
lesions studied. A first experiment was carried out with
1 nmol of each of the corresponding nucleosides (5-
HMdUrd, 5-FordUrd, 5-OHdCyd, and 5-OHdUrd) that
were hydrolyzed for increasing periods of time. Three
acidic treatments were investigated, namely, 88% formic
acid at 140 °C, 60% formic acid at 140 °C, and HF/Pyr at
37 °C. In these series of experiments, the internal

1148 Chem. Res. Toxicol., Vol. 9, No. 7, 1996
Douki et al.
F igu r e 3. Effect of the conditions of enzymatic digestion on
the modification rate of 5-OHdCyd measured in DNA by HPLC-
EC. (a) 10 units of nuclease P1 for increasing period of time; (b)
increasing amount of nuclease P1 for 2 h.
release of the four oxidized bases reached a plateau after
20 min in 88% formic acid. A similar observation was
made for the release of oxidized bases induced by HF/
Pyr, with the exception of 5-HMUra which was not stable
under these conditions. The quantitative aspect of the
DNA enzymatic hydrolysis by nuclease P1 was also
checked (Figure 3). A time course study (from 0.5 to 18
h) with 2 U of the enzyme was performed. Optimization
of the enzymatic hydrolysis conditions was also investi-
gated by incubating DNA samples with increasing
amounts of enzyme (from 0.3 to 17 U) for 2 h. Samples
were then analyzed for their content in 5-OHdCyd by
HPLC-EC. It was found that both the amount of
hydrolyzed DNA and the value of the yield of lesion
reached a plateau either after 1 h when 10 U of nuclease
P1 was used or when DNA sample was incubated with 2
U of the latter enzyme for 2 h. Therefore, it was
concluded that treatment of the samples by 10 U of
nuclease P1 for 2 h is most likely to allow a complete
digestion. Study of the stability of dUrd diols (Figure
4A) showed that they are completely stable under the
enzymatic digestion conditions used. Indeed, the HPLC
elution profiles were identical after and before incubation
in the buffer used for nuclease P1 and alkaline phos-
phatase. In contrast, 5-OHUra was the only detectable
product in dUrd diol samples hydrolyzed by 88% formic
acid. 5-HMura and the faster eluted dUrd diol exhibit
identical HPLC retention time but are easily distinguish-
able on the basis on their UV spectrum. In addition to
5-OHUra, a second major compound, which UV spectra
exhibited only one maximum centered around 210 nm
like dUrd diols, was detected in dUrd diol samples after
treatment by HF/Pyr. The yield of the latter hydrolysis
product was approximately half of that of 5-OHUra.
Additional structural information was inferred from a
GC-MS analysis showing the presence of a compound
whose mass spectrum corresponds to that of the tetra-
silylated Ura diol (Figure 4B). It is worth mentioning
that dUrd diols are partly converted into 5-OHUra during
the silylation step of the GC-MS analysis. Therefore, it
F igu r e 4. Stability of dUrd diols under different hydrolysis
conditions. (A) HPLC analysis of samples treated under condi-
tions of (1) enzymatic digestion, (2) 88% formic acid, (3) HF/
Pyr. (B) Electron impact mass spectrum of the Ura diol obtained
by GC-MS after trimethylsilylation of dUrd diol samples
hydrolyzed by HF/Pyr.
is not possible to directly estimate the yield of each of
the hydrolysis products on the basis of the area of the
peaks of the GC-MS elution profile. Altogether, it was
concluded that Ura diol is a major product of the HF/
Pyr mediated hydrolysis of dUrd diols.
Mea su r em en ts of Oxid ized P yr im id in e Ba ses in
γ-Ir r a d ia ted DNA Sa m p les. A first series of results
was obtained for DNA samples exposed to γ rays in
aerated aqueous solution and then hydrolyzed by either
88% formic acid or HF/Pyr. 5-HMUra, 5-ForUra, 5-
OHUra, and 5-OHCyt were measured by GC-MS. The
modification rate was found to be proportional to the dose
for the four radiation-induced lesions over the dose range
(0-225 Gy) investigated. In addition, the results dealing
with the formation of 5-ForUra upon either HF/Pyr or
88% formic acid mediated release were similar (Figure
5). Such a comparison could not be made for 5-HMUra
since the latter lesion is not stable in HF/Pyr. The
homogeneity of the results obtained for 5-ForUra follow-
ing either formic acid or HF/Pyr was not observed for
5-OHUra and 5-OHCyt (Figure 6a). When 88% formic
acid was used, 5-OHUra was found to be predominant
over 5-OHCyt (ratio 5-OHUra/5-OHCyt ) 2.39). In
contrast, when DNA was hydrolyzed by HF/Pyr, 5-OHCyt
was detected in a higher yield than 5-OHUra (5-OHUra/
5-OHCyt ) 0.78). It should be added that the yield of
5-OHCyt and 5-OHUra was higher upon formic acid
hydrolysis than following HF/Pyr treatment. The effect
was stronger for 5-OHUra since the yield increases by a

Radiation-Induced Pyrimidine Lesions in DNA
Chem. Res. Toxicol., Vol. 9, No. 7, 1996 1149
and cellular DNA (19). It was observed that the level of
both 5-OHUra (Figure 6b) and 5-OHCyt was higher upon
formic acid hydrolysis than after enzymatic digestion
(ratio: 4.0 and 1.6, respectively). In addition, 5-OHCyt
was the predominant lesion upon enzymatic digestion
(ratio 5-OHUra/5-OHCyt ) 0.75) whereas 5-OHUra was
detected in the highest yield after formic acid hydrolysis
(5-OHUra/5-OHCyt ) 1.87).
Discu ssion
GC-MS Meth od ology. GC-MS analysis is a now
widely used technique for the detection of oxidative base
lesions within DNA after acidic hydrolysis. However,
until recently, yields of 8-oxoGua measured by using this
approach were much higher than those obtained by
HPLC-EC (20). Observation of an artifactual oxidation
of guanine moieties during the silylation step of the GC-
MS analysis is likely to explain these discrepancies (12,
13). This observation has then been extended to four
other oxidized nucleobases, including 5-HMUra, 5-
OHCyt, 5-ForUra, and 8-oxo-7,8-dihydroadenine (14). In
order to overcome this problem, GC-MS measurements
of the radiation-induced oxidized DNA bases released
upon acidic treatment of DNA were carried out after
HPLC prepurification of the samples. This step was
aimed at removing the normal bases from their oxidized
derivatives prior to silylation. Under these conditions,
the amount of oxidized bases detected by GC-MS is only
accounted for by their release from DNA. Hydrolysis at
high temperature could be another source of artifactual
oxidation of the DNA samples. However, this would lead
to a continuous increase of the amount of oxidized bases
measured in the same sample with increasing periods of
acidic treatment. In contrast, the time course study of
the release of the four lesions studied in 88% formic acid
reached a plateau. This is strongly indicative that no
oxidation occurs during the hot acidic hydrolysis. Similar
observations were made for HF/Pyr and enzymatic treat-
ments. Another aspect of the present GC-MS assay is
the use of isotopically labeled internal standard. This is
necessary to avoid any bias due to the possible loss of
sample in the experimental workup and to compensate
for variation in the silylation rate or injection efficiency.
The accuracy of the improved GC-MS assay allowed the
precise study of the effect of the hydrolysis conditions. It
was observed that 60% formic acid, a reagent used by
some authors (21, 22), leads to an extensive degradation
of the four lesions studied. In contrast, when the
concentration of formic acid was raised to 88% (23, 24),
the four molecules were stable over a period of time
allowing their quantitative release from DNA. These
results can be explained by the molecular ratio between
water and formic acid which is 1.4 in 60% formic acid
and 0.5 in 88% formic acid. This means that dissociation
of the carboxylic acid group will be higher in the less
concentrated solution. Similar conclusions on the deg-
radative effects of diluted formic acid on oxidized DNA
bases were reported in a recent work (25). In addition
to the obvious loss of sensitivity, the use of a hydrolyzing
reagent able to decompose the molecules of interest
induces a drastic decrease in the reliability of the
measurement. It can be stated that since both the lesion
and its internal standard are degraded at a similar rate,
the ratio between the amount of these two compounds
will remain constant over the hydrolysis period. How-
ever, the normal oxidized base is slowly released from
F igu r e 5. Formation of (a) 5-ForUra and (b) 5-HMUra in DNA
exposed to γ rays in aqueous aerated solution determined by
GC-MS. Hydrolysis conditions: 88% fa ) 88% formic acid, 140
°C, 45 min; HF/pyr ) 70% hydrogen fluoride in pyridine, 37 °C,
45 min. The error bars correspond to the average of 3 indepen-
dent determinations.
F igu r e 6. Formation of (a) 5-OHCyt and (b) 5-OHUra in DNA
exposed to γ radiation in aqueous aerated solution. Hydrolysis
conditions: 88% fa ) 88% formic acid, 140 °C, 45 min; HF/Pyr
) 70% hydrogen fluoride in pyridine, 37 °C, 45 min; P1/AP )
enzymatic digestion by nuclease P1 and alkaline phosphatase.
The two former series of samples were analyzed by GC-MS. The
enzymatically digested samples were analyzed by HPLC-EC.
The error bars correspond to the average of 3 independent
determinations.
factor of 4.4 whereas it was only 1.6-fold higher for
5-OHCyt. To further assess the role of the hydrolysis
technique, another series of γ-irradiated DNA samples
were analyzed by both GC-MS following 88% formic acid
hydrolysis and HPLC-EC after enzymatic digestion. The
latter technique has been successfully applied to the
detection of both 5-OHdCyd and 5-OHdUrd in isolated

1150 Chem. Res. Toxicol., Vol. 9, No. 7, 1996
Douki et al.
Ta ble 1. Ra d ia tion -In d u ced F or m a tion of 5-HMUr a ,
5-F or Ur a , 5-OHUr a , a n d 5-OHCyt follow in g Differ en t
Hyd r olysis P r oced u r esa
0.0069 and 0.0154, respectively. In the case of the
enzymatic treatment, HPLC-EC detection was applied
instead of GC-MS because internal isotopically labeled
standards for the nucleosides were not available. How-
ever, the difference observed in the results cannot be
explained by the detection technique used since we
previously showed that, for the same hydrolysis tech-
nique, the use of either HPLC-EC or improved GC-MS
led to the same level of 5-OHCyt in calf thymus DNA
(14). Moreover, similar G values for 5-OHCyt were
obtained by GC-MS following HF/Pyr hydrolysis and
HPLC-EC following enzymatic digestion (G ) 0.0043 and
0.0048, respectively). This observation also applied to
5-OHUra (G ) 0.0035 and 0.0035, respectively). This
further supports the fact that the apparent discrepancies
in the results cannot be explained by the detection
methods used. These differences which are dependent
on the hydrolysis conditions can be rationalized in terms
of formation of 5-hydroxylated derivatives from unstable
5,6-saturated products of cytosine during the strong
acidic treatment. Likely candidates for these precursors
are the cytosine diols. As far as enzymatic digestion and
formic acid hydrolysis are concerned, this is strongly
supported by studies involving dUrd diols. The use of
formic acid leads to the quantitative release of 5-OHUra
whereas no decomposition of dUrd diols was observed
under the conditions of enzymatic hydrolysis. Results
are more ambiguous for HF/Pyr hydrolysis. When used
for the hydrolysis of dUrd diols, HF/Pyr leads to the
partial conversion of dUrd diols as 5-OHUra. However,
the latter hydrolysis conditions also allow the release of
uracil diol, as shown by HPLC and GC-MS. Therefore,
it can be concluded that HF/Pyr is a milder hydrolyzing
agent than hot 88% formic acid. The difference in the
results between nucleoside and DNA may be accounted
for by the role of the phosphate groups. The presence of
the latter electron withdrawing group has been shown
to decrease the hydrolysis rate in modified dinucleoside
monophosphates (26). The involvement of cytosine and
uracil diols is also supported by other experimental data.
In acidic media, cytosine diols can undergo two main
transformations. One reaction is a fast deamination
which leads to the corresponding uracil derivative. In
addition, both cytosine and uracil diols can dehydrate,
producing the related 5-hydroxylated derivatives. This
is in agreement with the observation of a much higher
increase in the amount of 5-OHUra by comparison to
5-OHCyt in strong acidic conditions. This is likely due
to the fact that deamination is faster than dehydration.
Altogether, it can be proposed that 5-OHCyt and 5-
OHUra detected using mild hydrolysis conditions are
produced by a direct mechanism, as proposed for cytosine
by von Sonntag (30), whereas the higher amount detected
after 88% formic acid hydrolysis accounts for both directly
produced 5-hydroxylated derivatives as well as cytosine
and uracil diols. Based on the values obtained following
either enzymatic digestion or HF/Pyr hydrolysis, the G
values of 5-OHCyt and 5-OHUra were determined as
0.0046 and 0.0035 µmol‚J -1, respectively. The differences
between the yields of 5-OHUra and 5-OHCyt determined
using enzymatic and strong hydrolysis conditions provide
an estimation of the overall yield of cytosine and uracil
diols produced by γ irradiation (G ) 0.0131). These
observations are in agreement with semiquantitative
results based on the analysis of the supernatant fraction
of γ-irradiated DNA incubated with endonuclease III
compd
HF/Pyr
88% formic
P1/AP digestion
5-HMUra
5-ForUra
5-OHCyt
5-OHUra
unstable
0.0077
0.0043
0.0035
0.0042
0.0083
0.0069
0.0154
0.0035
ndb
0.0048
0.0035
a
b
Yields expressed in µmol‚J -1
.
nd, not determined.
DNA by the hydrolysis reagent. Therefore, the overall
stability of the normal oxidized pyrimidine in the hy-
drolysis mixture partly depends on its stability when still
linked to DNA. This will not be compensated by the
addition of internal standard which is added as a free
base from the very beginning of the hydrolysis. A third
hydrolysis technique, based on the use of hydrogen
fluoride in solution in pyridine, was investigated. This
reagent has been applied in our laboratory to various
modified DNA bases, including (6-4) photoproducts (26)
and adducts of R,â-unsaturated aldehydes to 2′-deoxy-
guanosine (27) and 8-oxodGuo (2). HF/Pyr was also
found to lead to the quantitative hydrolysis of the
N-glycosidic bond of modified nucleosides such as thy-
midine diols (28) and 4-hydroxy-8-oxo-4,8-dihydro-2′-
deoxyguanosine (29). HF/Pyr was found to be a mild
reagent allowing the hydrolysis of lesions which were
unstable under other conditions. In the present study,
HF/Pyr was successfully applied to the quantitative
release of 5-OHCyt, 5-OHUra, and 5-ForUra. Altogether,
this work shows that optimization of the whole analytical
procedure has to be achieved when a new lesion is
studied. GC-MS, like any other detection technique,
cannot be used as a universal method of measurement.
R ela t ive Yield of F or m a t ion of 5-F or Ur a a n d
5-HMUr a . In the following discussion, the formation of
oxidized bases is discussed in terms of radiochemical
yield of formation (G) (Table 1). The formation rate of
the two oxidized thymine derivatives studied, namely,
5-ForUra and 5-HMUra, was inferred from GC-MS
analysis. 5-ForUra was measured by GC-MS using two
different hydrolysis techniques involving 88% formic acid
and HF/Pyr, respectively. The data obtained by using
these two approaches are similar (G: 0.0083 and 0.0077
µmol‚J -1, respectively). This is strongly indicative that
we obtained reliable results with respect to the hydrolysis
method used. For 5-HMUra, only 88% formic acid
hydrolysis was performed. The G value obtained (G )
0.0042) is in good agreement with the value obtained
previously by GC-MS analysis following enzymatic diges-
tion (G ) 0.0035) (14). Altogether, these results show
that 5-ForUra is generated in a 2.1-fold higher yield than
5-HMUra. This is in agreement with menadione medi-
ated photosensitization experiments at the nucleoside
level. 5-FordUrd was isolated in a 3.7-fold higher yield
than 5-HMUra (15). This is expected since both oxidizing
conditions lead to the formation of the same methyl
centered thymine radical which further gives rise to
5-HMUra and 5-ForUra under aerobic conditions.
F or m a tion of 5-OHUr a , 5-OHCyt, a n d Cytosin e
a n d Ur a cil Diols. The yields of 5-OHCyt and 5-OHUra
in γ-irradiated DNA are highly dependent on the hy-
drolysis technique used. When DNA was hydrolyzed by
88% formic acid, the yield of both lesions was much
higher than when milder conditions (HF/Pyr or enzy-
matic digestion) were used. Following formic acid hy-
drolysis, the G values for 5-OHCyt and 5-OHUra were

Radiation-Induced Pyrimidine Lesions in DNA
Chem. Res. Toxicol., Vol. 9, No. 7, 1996 1151
gas chromatography-mass spectrometry and HPLC-electro-
chemical detection: overestimation of the background level of the
oxidized base by the gas chromatography-mass spectrometry
assay. Chem. Res. Toxicol. 8, 1039-1045.
where uracil diols were detected in higher amount than
5-OHCyt and 5-OHUra (3).
(14) Douki, T., Delatour, T., Bianchini, F., and Cadet, J . (1996)
Observation and prevention of an artefactual formation of oxi-
dized DNA bases and nucleosides in the GC-EIMS method.
Carcinogenesis 17, 347-353.
(15) Decarroz, C., Wagner, J . R., van Lier, J . E., Murali, K. C., Riesz,
P., and Cadet, J . (1986) Sensitized photo-oxidation of thymidine
by 2-methyl-1,4-naphthoquinone. Characterization of stable prod-
ucts. Int. J . Radiat. Biol. 50, 491-505.
Con clu sion
The yields of formation of 5-ForUra, 5-HMUra, 5-
OHUra, and 5-OHCyt obtained in this study range
between those of 8-oxoGua and 8-oxoAde, which were
determined by HPLC-EC to be 0.0178 and 0.0009
µmol‚J -1, respectively (2). This suggests that the oxidized
pyrimidine derivatives studied in the present work are
among the main lesions induced by ionizing radiation in
DNA. Work is in progress to extend the assays used in
the present study to the measurement of other radiation-
induced DNA base damage in order to obtain reliable
values for their yields of formation.
(16) Moshel, R. C., and Behrman, E. J . (1974) Oxidation of nucleic
acid bases by potassium peroxodisulfate in alkaline aqueous
solution. J . Org. Chem. 39, 1983-1989.
(17) Brossmer, R., and Ziegler, D. (1966) Zur darstellung heterocy-
clisher aldehyde. Tetrahedron Lett. 43, 5253-5256.
(18) Bendich, A., Gelter, H., and Brown, G. B. (1949) Synthesis of
isotopic cytosine and a study of its metabolism in rat. J . Biol.
Chem. 177, 565-570.
(19) Wagner, J . R., Hu, C.-C., and Ames, B. N. (1992) Endogenous
oxidative damage of deoxycytidine in DNA. Proc. Natl. Acad. Sci.
U.S.A. 89, 3380-3384.
(20) Halliwell, B., and Dizdaroglu, M. (1992) The measurement of
oxidative damage to DNA by HPLC and GC/MS techniques. Free
Radical Res. Commun. 16, 15-28.
Refer en ces
(1) Fischer-Nielsen, A., Bøgh, I., and Loft, S. (1994) Radiation-induced
formation of 8-hydroxy-2′-deoxyguanosine and its prevention by
scavengers. Carcinogenesis 15, 1609-1612.
(2) Douki, T., Berger, M., Raoul, S., Ravanat, J .-L., and Cadet, J .
(1995) Determination of 8-oxopurines in DNA by HPLC using
amperometric detection. In Analysis of Free Radicals in Biological
Systems Favier, A. E., Cadet, J ., Kalyanaraman, B., Fontecave,
M., and Pierre, J .-L., Eds.) pp 213-224, Birkhau¨ser Verlag, Basel.
(3) Wagner, J . R., Blount, B. C., and Weinfeld, M. (1996) Excision of
oxidative cytosine modifications from γ-irradiated DNA by Es-
cherichia coli endonuclease III and human whole-cell extracts.
Anal. Biochem. 233, 76-86.
(4) Kasai, H., Tanooka, H., and Nishimura, S. (1984) Formation of
8-hydroxyguanine residues in DNA by X-irradiation. Gann 75,
1037-1039.
(5) Kasai, H., Iida, A., Yamaizumi, Z., Nishimura, S., and Tanooka,
H. (1990) 5-Formyldeoxyuridine: a new type of DNA damage
induced by ionizing radiation and its mutagenicity to Salmonella
strain TA102. Mutat. Res. 243, 249-253.
(6) Breimer, L. H., and Lindahl, T. (1985) Thymine lesions produced
by ionizing radiation in double-stranded DNA. Biochemistry 24,
4018-4022.
(7) Zuo, S., Boorstein, R., and Teebor, G. W. (1995) Oxidative damage
to 5-methylcytosine. Nucleic Acids Res. 23, 3239-3243.
(8) Dizdaroglu, M. (1985) Formation of an 8-hydroxyguanine moiety
in deoxyribonucleic acid on γ-irradiation in aqueous solution.
Biochemistry 24, 4476-4481.
(9) Dizdaroglu, M., and Bergtold, D. S. (1986) Characterization of
free radical-induced base damage in DNA at biologically relevant
levels. Anal. Biochem. 156, 182-188.
(10) Fuciarelli, A. F., Wegher, B. J ., Blakely, W., and Dizdaroglu, M.
(1990) Yields of radiation-induced base products in DNA: effects
of DNA conformation and gassing conditions. Int. J . Radiat. Biol.
58, 397-415.
(21) Dizdaroglu, M., Karakaya, A., J agura, P., Slupphaug, G., and
Krokan, H. E. (1996) Novel activities of human uracil DNA
N-glycosylase for cytosine-derived products of oxidative DNA
damage. Nucleic Acids Res. 24, 418-422.
(22) Spencer, J . P. E., J enner, A., Aruoma, O. I., Evans, P. J ., Kaur,
H., Dexter, D. T., J enner, P., Lees, A. J ., Marsden, D. C., and
Halliwell, B. (1994) Intense oxidative DNA damage promoted by
L-DOPA and its metabolites. Implication for neurodegenerative
disease. FEBS Lett. 353, 246-250.
(23) Aruoma, O. I., Halliwell, B., and Dizdaroglu, M. (1989) Iron ion-
dependent modification of bases in DNA by the superoxide radical-
generating system hypoxanthine/xanthine oxidase. J . Biol. Chem.
264, 13024-13028.
(24) Fuciarelli, A. F., Sisk, E. C., Thomas, R. M., and Miller, D. L.
(1995) Induction of base damage in DNA solutions by ultrasonic
cavitation. Free Radical Biol. Med. 18, 231-238.
(25) Swarts, S. G., Smith, G. S., Miao, L., and Wheeler, K. T. (1996)
Effects of formic acid hydrolysis on the quantitative analysis of
radiation-induced DNA base damage products assayed by gas
chromatography. Radiat. Environ. Biophys. 35, 41-53.
(26) Douki, T., Voituriez, L., and Cadet, J . (1995) Measurement of
pyrimidine (6-4) photoproducts in DNA by a mild acidic hydroly-
sis-HPLC fluorescence detection assay. Chem. Res. Toxicol. 8,
244-253.
(27) Douki, T., and Ames, B. N. (1994) An HPLC-EC assay for 1,N2
-
propano adducts of 2′-deoxyguanosine with 4-hydroxynonenal and
other R,â-unsaturated aldehydes. Chem. Res. Toxicol. 7, 511-518.
(28) Polverelli, M., Berger, M., Mouret, J .-F., Odin, F., and Cadet, J .
(1990) Acidic hydrolysis of the N-glycosidic bond of deoxyribo-
nucleic acid by hydrogen fluoride stabilized in pyridine. Nucleo-
sides Nucleotides 9, 451-452.
(29) Cadet, J ., Ravanat, J .-L., Buchko, G. W., Yeo, H. C., and Ames,
B. N. (1994) Singlet oxygen DNA damage: chromatographic and
mass spectrometry of damage products. Methods Enzymol. 234,
79-88.
(11) Mori, T., and Dizdaroglu, M. (1994) Ionizing radiation causes
greater DNA base damage in radiation-sensitive mutant M10 cells
than in parent mouse lymphoma L5178Y cells. Radiat. Res. 140,
85-90.
(12) Hamberg, M., and Zhang, L.-Y. (1995) Quantitative determination
of 8-hydroxyguanine and guanine by isotope dilution mass
spectrometry. Anal. Biochem. 229, 336-344.
(30) von Sonntag, C. (1987) The chemical bases of radiation biology,
Taylor & Francis, London.
(13) Ravanat, J .-L., Turesky, R. J ., Gremaud, E., Trudel, L. J ., and
Stader, R. H. (1995) Determination of 8-oxoguanine in DNA by
TX960095B