ABSTRACT
Oxidative DNA damage is the most frequent type of damage encountered by aerobic
cells and may play an important role in biological processes such as mutagenesis, carcinogenesis and aging in humans. Oxidative damage generates a
myriad of modifications in DNA. We investigated the cellular repair of DNA base
damage products in DNA of cultured human lymphoblast cells, which were exposed
to oxidative stress by H
2
O
2
. This DNA-damaging agent is known to cause base modifications in genomic DNA of
mammalian cells [Dizdaroglu, M., Nackerdien, Z., Chao, B.-C., Gajewski, E. and Rao, G. (1991)
Arch. Biochem. Biophys
. 285, 388-390]. Following treatment with H
2
O
2
, the culture medium was freed from H
2
O
2
and cells were incubated for time periods ranging from 10 min to 6 h. DNA was
isolated from control cells, hydrogen peroxide-treated cells and cells incubated after H
2
O
2
exposure. DNA samples were analyzed by gas chromatography/isotope-dilution mass spectrometry. Eleven modified bases were identified and
quantified. The results showed a significant formation of these DNA base
products upon H
2
O
2
-treatment of cells. Subsequent incubation of cells caused a time-dependent excision of these products from cellular DNA. The cell viability did not change significantly by various
treatments. There were distinct differences between the kinetics of excision of
individual products. The observed excisions were attributed to DNA repair in
cells. The rate of repair of purine lesions was slower than that of pyrimidine
lesions. Most of the identified products are known to possess various
premutagenic properties. The results of this work may contribute to the
understanding of the cellular repair of oxidative DNA damage in human and other
mammalian cells.
Oxygen-derived species including free radicals are formed in living cells by
normal metabolism and by exogenous sources (reviewed in
1
). Of free radicals, the hydroxyl radical (
@
OH) is the most reactive toward biological molecules and generates a multitude of modifications in DNA such as base damage, sugar damage and DNA-protein crosslinks (reviewed in
2
,
3
). This type of DNA damage, also called oxidative DNA damage, has been
implicated in biological processes such as mutagenesis, carcinogenesis and
aging (
1
).
Oxidative DNA damage may be repaired in cells by a variety of repair enzymes. In
both bacteria and mammalian cells, a multitude of repair enzymes have been
discovered, which possess multiple activities toward products of oxidative DNA
damage (reviewed in
4
-
6
). DNA base products are repaired by both base-excision and nucleotide-excision repair, but predominantly by the former. There are 20 or so
major products resulting from reactions of free radicals with four heterocyclic
bases in DNA (
2
,
3
). In the past,
in vitro
studies have determined the specificities of the repair enzymes for most of
these DNA base products. In particular, two well known
Escherichia coli
enzymes Nth and Fpg proteins (endonuclease III and formamidopyrimidine-DNA glycosylase, respectively) account for the excision of most modified
bases from DNA (
4
,
5
). Eukaryotic counterparts of these enzymes exist in other organisms. Nucleotide-excision repair systems also act on oxidative DNA damage (
4
,
6
).
Little is known about the repair of individual products of oxidative DNA base
damage and their repair kinetics in mammalian cells. We present here a study of
the repair of oxidative DNA base damage in human cells. The objective was to
investigate the cellular repair of individual DNA base modifications that are
formed in human cells upon exposure to oxidative stress. Hydrogen peroxide was
chosen as the agent that causes oxidative stress because it can cross the
cellular membranes, reach the nucleus and cause damage to nuclear DNA by
generating
@
OH in close proximity to DNA (
1
). Hydrogen peroxide is also relevant in terms of endogenous oxidative stress
because it is continuously produced in aerobic cells (reviewed in
7
). A previous work has shown the formation of typical
@
OH-induced products from all four DNA bases in mammalian cells upon treatment
with H
2
O
2
(
8
). In the present work, the formation and subsequent time-dependent removal of modified DNA bases in cells were determined by using
the technique of gas chromatography/isotope-dilution mass spectrometry (GC/IDMS). This technique permits precise identification and quantification of modified DNA bases in cells (
3
,
9
).
RPMI-1640 medium, Hanks' balanced salt solution, fetal bovine serum (heat
inactivated), L-glutamine, penicillin-streptomycin solution, sodium bicarbonate solution (7.5%) were
purchased from Sigma Chemical Company. DNA NOW reagent was obtained from PGC Scientific.
Modified DNA bases, their stable isotope-labeled analogues, and materials for GC/IDMS were obtained as described
previously (
9
).
Human lymphoblast cells were used. A sample of these cells (GM03798A) was
purchased from NIGMS Human Genetic Mutant Cells Repository, Coriel Institute
for Medical Research. Cells were grown in suspension in a culture medium
consisting of RPMI-1640 modified medium supplemented with 15% fetal bovine serum, sodium
bicarbonate (0.2%), L-glutamine (2 mM), penicillin (10 U/ml) and streptomycin (1 [mu]g/ml) at 37oC under an atmosphere of 5% CO
2
mixed with room air. When cells reached a density of ~5 * 10
7
cells/ml, they were divided into three 25 ml aliquots and transferred into 50
ml tissue culture flasks. This was done for each data point. Cell viability was
determined by the trypan blue exclusion test.
The cells were centrifuged at 800
g
for 5 min, washed once with 25 ml of Hanks' balanced salt solution and then
centrifuged. The solution was removed and the cells were resuspended in 25 ml
of Hanks' balanced salt solution. H
2
O
2
was added to the flasks at concentrations of 1-5 mM. Flasks were placed under an atmosphere of 5% CO
2
mixed with room air and kept at 37oC for 60 min. Control cells were treated in the same manner except for H
2
O
2
treatment. For repair studies, cells were centrifuged at 800
g
for 5 min, washed once with RPMI-1640 medium with fetal bovine serum and resuspended in fresh RPMI-1640 medium with fetal bovine serum. Aliquots of cells were
incubated at 37oC for time periods of 10 min-6 h.
Aliquots (1 ml) of cell suspensions containing ~5 * 10
7
cells were centrifuged. An aliquot (1 ml) of DNA NOW reagent was added to each
cell pellet. The pellet was homogenized by repetitive pipetting. Subsequently,
0.2 ml chloroform (kept at -20oC) was added and samples were shaken by hand for 20 s, kept on ice
for 5 min and centrifuged. The aqueous phase was transferred to a clean tube
and 2 vol of cold isopropanol were added. Samples were kept on ice for 1 h. DNA
precipitate was removed with a glass rod and washed twice with cold ethanol
(70%) and air dried.
DNA samples were dissolved in 150 [mu]l of 10 mM phosphate buffer (pH 7.4), and the concentration of DNA was
determined by the absorbance at 260 nm (absorbance of 1 = 50 [mu]g of DNA/ml). Aliquots of stable isotope-labeled analogues of modified DNA bases were added as internal
standards to 50 [mu]g aliquots of DNA samples (
9
). Samples were dried in a SpeedVac under vacuum and then hydrolyzed with 0.5 ml of 60% formic acid in evacuated and
sealed tubes at 140oC for 30 min. The hydrolyzates were lyophilized in vials. For
derivatization, a mixture (0.1 ml) of nitrogen-bubbled bis(trimethylsilyl)trifluoroacetamide (containing 1% trimethylchlorosilane) and acetonitrile (4:1, v/v) was added to the vials. They were sealed under nitrogen with Teflon-coated septa and heated at 120oC for 30 min. Analyses of derivatized samples were performed by
GC/IDMS with selected-ion monitoring (SIM) (
9
,
10
). For this purpose, 2 [mu]l of derivatized samples were injected without further treatment into the
injection port of the gas chromatograph. The split mode of injection with a
split ratio of 20:1 was used.
In order to study the cellular repair of DNA base products, their levels in
cells must be elevated significantly over the background levels upon treatment
of cells with the damaging agent. Furthermore, the cell viability after the treatment must remain at the same level as that of
untreated control cells. For these reasons, we investigated first the formation
of DNA base products in cells as a function of the concentration of H
2
O
2
. At the same time, the cell viability was determined. At H
2
O
2
concentrations of 1-5 mM in the culture medium, a significant extent of modification of all
four DNA bases was observed. The cell viability did not change significantly
upon treatment of cells with <= 5 mM H
2
O
2
(data not shown). Further experiments were undertaken using 5 mM H
2
O
2
.
Using the GC/IDMS-SIM, the following DNA base products were identified and quantified in cellular DNA: 2,6-diamino- 4-hydroxy-5-formamidopyrimidine (FapyGua), 8-hydroxyguanine (8-OH-Gua), xanthine (Xan), 4,6-diamino-5-formamidopyrimidine (FapyAde), 8-hydroxyadenine (8-OH-Ade), 2-hydroxyadenine (2-OH-Ade), isodialuric acid, 5-hydroxyuracil (5-OH-Ura), 5-hydroxycytosine (5-OH-Cyt), 5-(hydroxymethyl)uracil (5-OHMeUra) and 5-hydroxy-5-methylhydantoin (5-OH-5-MeHyd). Of these compounds, the uracil derivatives are products of cytosine modification in DNA, except for 5-OHMeUra, which is a product of thymine modification (
2
,
3
). Isodialuric acid was detected as 5,6-dihydroxyuracil (5,6-diOH-Ura) because it enolizes during derivatization (
11
). Figure
1
illustrates the structures of these products. Upon treatment of cells with 5 mM
H
2
O
2
, the product levels increased above control levels by 3-5-fold with no significant loss of cell viability (Fig.
2
).
Having observed significant formation of DNA base products, we studied their
excision from cellular DNA upon incubation of cells for different time
intervals. Following H
2
O
2
treatment, cells were washed with the medium without H
2
O
2
, suspended in fresh medium and then incubated at 37oC for 10 min-6 h. DNA was isolated from cells and analyzed. The results are
illustrated in Figure
2
.
There were three guanine-derived products, namely FapyGua, 8-OH-Gua and Xan, among those measured in cells. FapyGua and 8-OH-Gua were formed significantly in cellular DNA upon
treatment of cells with H
2
O
2
(Fig.
2
A). Their levels increased by ~5- and 3-fold over background levels, respectively. An ~50% reduction in their levels was observed after 30-40 min of incubation. Afterwards, the rates of
removal became slower. The level of FapyGua reached the background level at 2 h
of incubation and remained constant thereafter up to 6 h (the last data point
not shown in Fig.
2
A). The excision of 8-OH-Gua required >2 h. The kinetic results in Figure
3
were analyzed to see which order of reaction applies to the excisions of
FapyGua and 8-OH-Gua from cellular DNA. The logarithm of the ratio of the initial
amount (a
0
) to the amount (a) at a given incubation time was plotted against the
incubation time (
12
). The plots yielded linear relationships up to 45 min of incubation (Fig.
3
). This revealed that the excisions of FapyGua and 8-OH-Gua followed first-order kinetics within this time period. First-order rate constants and half-lives were calculated using the initial amounts
and the amounts at incubation times from 10 to 45 min. The means (+- standard deviation) of these kinetic constants are given in Table
1
. The treatment of cells with H
2
O
2
caused an ~3-fold increase in the level of Xan over its background level (Fig.
2
B). Upon incubation, this product was excised from cellular DNA in 45 min with
its level almost reduced to the background level (see also Table
1
).
Typical
@
OH-induced DNA base products were formed in human cells upon oxidative stress
by H
2
O
2
with no significant loss of cell viability. No loss of cell viability upon
oxidative stress was a prerequisite for the subsequent study of the repair of
DNA damage. The salient feature of this work is the evidence that these
modified bases were efficiently excised from cellular DNA, but with distinct
differences between the kinetics of excision for individual products. Since the
cell viability did not change upon various treatments, this observed excision
of modified bases was attributed to actions of DNA repair systems in cells. In
general, the rate of repair of purine-derived lesions was slower than that of pyrimidine-derived lesions. The kinetics of repair of guanine-derived lesions FapyGua, 8-OH-Gua and Xan were different. FapyGua was removed
from cellular DNA within 2 h, whereas the repair of 8-OH-Gua lasted almost 4 h. The first-order rate constant for FapyGua was ~60% higher than that for 8-OH-Gua in the first 45 min of repair. The slower
repair kinetics of 8-OH-Gua is in agreement with recent results on the repair of 8-OH-Gua in a different human cell line (
13
) and in liver DNA of mice (
14
). Xan was removed from cellular DNA within 45 min with a first-order rate constant similar to that for the excision of FapyGua.
FapyGua and 8-OH-Gua in DNA are substrates for the DNA repair enzyme Fpg protein of
E.coli
(
4
,
15
,
16
). Mammalian cells also possess activities that remove formamidopyrimidine
lesions from DNA (reviewed in
17
). There is evidence for the existence in human and other mammalian cells of DNA
glycosylase and endonuclease activities for removal of 8-OH-Gua (
18
-
20
). The observed excision of FapyGua and 8-OH-Gua from DNA in human cells in this work may be due to the activity
of such DNA repair enzymes. No repair enzyme specific for Xan has been
described. Considering the wide substrate range of human excinuclease (
21
), these and other products may also be repaired by the nucleotide-excision repair system.
Of the adenine-derived products, the repair of FapyAde was slow, lasting >2 h, whereas
the repair of 8-OH-Ade was complete within 20 min. The rate of excision of FapyAde
within the first 45 min of repair was similar to that of 8-OH-Gua, but ~60 and 300% slower than those of FapyGua and 8-OH-Ade, respectively. FapyAde is efficiently excised
from DNA by
E.coli
Fpg protein (
16
,
22
), and also by
E.coli
T4 endonuclease V (
23
). In contrast with 8-OH-Gua, 8-OH-Ade is a poor substrate for Fpg protein (
16
). The rate of repair of 2-OH-Ade was rather slow. After 2 h of repair, the level of this product
was reduced by 50% only, with repair completed after 4 h. No enzymatic activity
has thus far been described for 2-OH-Ade excision.
Of three cytosine-derived lesions detected, the rate of repair of 5-OH-Ura was approximately twice as fast as that of 5-OH-Cyt and isodialuric acid, reaching completion
within 10 min. The repair of 5-OH-Cyt and isodialuric acid were complete in 20-30 min. The first-order rate constants for excision of 5-OH-Cyt and isodialuric acid were similar. The thymine-derived lesion, 5-OHMeUra, was also excised
quickly from cellular DNA. On the other hand, 5-OH-5-MeHyd was removed at a slower rate. 5-OH-Ura, 5-OH-Cyt and 5-OH-5MeHyd are substrates for
E.coli
Nth protein (
24
-
26
), whereas
E.coli
and human uracil DNA
N
-glycosylases possess activities for isodialuric acid and 5-OH-Ura (
11
,
26
,
27
). A DNA
N
-glycosylase that excises 5-OHMeUra from DNA was detected in mammalian cells (
4
). Possible eukaryotic counterparts to
E.coli
Nth protein exist in other organisms (reviewed in
17
). Human DNA
N
-glycosylases and human excinuclease repair system may be involved in the
repair of cytosine- and thymine-derived lesions in human cells observed in this work.
Some of the DNA base lesions, of which repair was studied in this work, have
been shown to possess premutagenic properties. In this respect, 8-OH-Gua is the most investigated lesion and has been shown to cause GC -> TA transversions (
28
-
31
). FapyGua may lead to GC -> CG transversions (
32
). The repair of these two guanine-derived lesions in human cells required up to 4 h. Considering also the
extent of their formation, these two products may contribute significantly to
the mutagenic effects of oxidative DNA damage. In support of this view, H
2
O
2
has been shown to induce GC -> TA and GC -> CG transversions in the supF gene of
E.coli
(
33
) and in the same system after passage through a mammalian host (
34
). A recent work has shown that 8-OH-Ade also possesses premutagenic properties (
35
). Likewise, 2-OH-Ade may be potentially premutagenic in cells because it pairs with
adenine and guanine (
36
). Our data indicate that cells may be able to repair 8-OH-Ade with a faster rate than its guanine-derived analog, whereas 2-OH-Ade repair may last as long as that of 8-OH-Gua. There is no information on the
base-pairing characteristics of the other prominent adenine-derived lesion FapyAde (
5
) and of the guanine-derived lesion Xan.
Of the pyrimidine-derived lesions, 5-OH-Cyt and 5-OH-Ura have been shown to be potentially premutagenic
lesions leading to GC -> AT transitions and GC -> CG transversions (
37
). 5-OH-Cyt appears to be more mutagenic than any other product of oxidative
DNA damage (
38
). In mammalian cells, H
2
O
2
predominantly produces GC -> AT transitions followed by GC -> CG and GC -> TA transversions (
34
). The first two types of mutations may indicate the role of cytosine-derived lesions in mutagenesis induced by oxidative DNA damage. The
present data show that 5-OH-Cyt and 5-OH-Ura may be repaired rapidly in human cells. The other
prominent cytosine-derived lesion isodialuric acid has not been investigated for its biological effects. 5-OHMeUra codes as thymine and its observed mutagenicity has been
attributed to incorporative mutagenesis (
39
,
40
). There is no information on possible mutagenic consequences of 5-OH-5-MeHyd (
5
). The contribution of these products to H
2
O
2
-induced mutagenesis is not known.
In conclusion, the results show the formation of a myriad of modified bases in
cellular DNA and their subsequent cellular repair as a function of time. This
is the first systematic study on the repair of oxidative damage-induced products of all four DNA bases in a human cell line. These
products substantially differ from one another in terms of their kinetics of
excision from cellular DNA. The results of this work may contribute to the
understanding of cellular repair of oxidative DNA damage in terms of individual
DNA base products. The approach used may be applicable to studies of repair of
oxidative DNA base damage in other mammalian cells and of possible differences
in repair capacity between cell lines for oxidative DNA damage.
Certain commercial equipment or materials are identified in this paper in order
to specify adequately the experimental procedure. Such identification does not
imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that
the materials or equipment identified are necessarily the best available for
the purpose.
REFERENCES
Return