ABSTRACT
The kinetics of excision of damaged purine bases from oxidatively damaged DNA by
Escherichia coli
Fpg protein were investigated. DNA substrates, prepared by treatment with H
2
O
2
/Fe(III)-EDTA or by
[gamma]
-irradiation under N
2
O or air, were incubated with Fpg protein, followed by precipitation of DNA.
Precipitated DNA and supernatant fractions were analyzed by gas
chromatography/isotope-dilution mass spectrometry. Kinetic studies revealed efficient excision of
8-hydroxyguanine (8-OH-Gua), 2,6-diamino-4-hydroxy-5-formamidopyrimidine (FapyGua)
and 4,6-diamino-5-formamidopyrimidine (FapyAde). Thirteen other modified bases
in the oxidized DNA substrates, including 5-hydroxycytosine and 5-hydroxyuracil, were not excised. Excision was measured as a function
of enzyme concentration, substrate concentration, time and temperature. The
rate of release of modified purine bases from the three damaged DNA substrates
varied significantly even though each DNA substrate contained similar levels of
oxidative damage. Specificity constants (
k
cat/
K
M) for the excision reaction indicated similar preferences of Fpg protein for
excision of 8-OH-Gua, FapyGua and FapyAde from each DNA substrate. These findings
suggest that, in addition to 8-OH-Gua, FapyGua and FapyAde may be primary substrates for this enzyme
in cells.
Oxidative DNA damage produced by endogenously- and exogenously-generated reactive oxygen species has been implicated in mutagenesis
and carcinogenesis and may play an important role in the pathogenesis of aging
(reviewed in
1
). Among oxygen- derived species, the hydroxyl radical is highly reactive, producing a
variety of lesions in DNA (reviewed in
2
,
3
). Most of these lesions are substrates for enzymes engaged in DNA repair in
bacteria and mammalian cells (reviewed in
4
-
7
).
Fpg protein of
Escherichia coli
(
7
-
10
) has been purified to apparent homogeneity and shown to possess N-glycosylase and [beta]-lyase activities (
11
-
13
). Several products of oxidative DNA damage, including 8-hydroxyguanine (8-OH-Gua) (
14
,
15
), 2,6-diamino-4-hydroxy-5-formamidopyrimidine (FapyGua) (
15
) and 4,6-diamino-5-formamidopyrimidine (FapyAde) (
9
,
15
), are substrates for this enzyme. Fpg protein appears to act weakly on 8-hydroxyadenine (8-OH-Ade) (
15
), but several laboratories, using different analytical techniques, did not
detect this activity (
16
,
17
). Recently, 5-hydroxyuracil (5-OH-Ura) and 5-hydroxycytosine (5-OH-Cyt) were reported to be efficiently excised
from duplex oligonucleotides by Fpg protein (
18
). The specificity of Fpg protein has been examined in terms of its binding
affinity and kinetic parameters for cleavage of defined DNA substrates (
17
). Duplex oligonucleotides containing a single lesion were used for this purpose
and a catalytic mechanism was proposed to explain the facile excision of
formamidopyrimidines and 8-OH-Gua by the enzyme (
17
).
Simultaneous measurement of the kinetics of excision of 8-OH-Gua, FapyGua and FapyAde by
E.coli
Fpg protein from oxidatively damaged DNA has not been reported. It is not known
whether Fpg protein excises these purine lesions with similar specificity from
oxidatively damaged DNA, and whether the kinetics of excision depends on the
nature of DNA substrate. Oxidatively damaged DNA contains a variety of
pyrimidine and purine lesions (
2
,
3
). The distribution and concentration of the lesions are affected by many
factors including the radical environment of DNA, and the presence or absence
of oxygen. In this context, the kinetics of excision may depend on the nature
of DNA substrates, which may differ from one another in terms of types,
distribution and quantities of lesions.
The objective of the present study was to simultaneously measure the kinetics of
excision of 8-OH-Gua, FapyGua and FapyAde by
E.coli
Fpg protein from oxidatively damaged DNA, and thus to provide a quantitative
comparison of the specificity of the enzyme for these lesions under similar conditions. Furthermore, we wished to determine if the rate of excision was affected by the nature of
the DNA substrate, using several oxidatively damaged DNA substrates. Calf
thymus DNA was used as a model for these experiments. We utilized the technique
of gas chromatography/isotope-dilution mass spectrometry (GC/IDMS) to measure the excision rates of
purine lesions. This technique permits precise identification and
quantification of numerous base lesions in a given DNA sample (
3
,
19
), and is well suited for the determination of the substrate specificity of DNA
repair enzymes (
15
,
20
-
23
).
Modified DNA bases, their stable isotope-labeled analogues, and materials for gas chromatography/isotope-dilution mass spectrometry (GC/IDMS) were obtained as described
previously (
19
). Calf thymus DNA and hydrogen peroxide solution were purchased from Sigma.
Calf thymus DNA was dissolved in 10 mM phosphate buffer (pH 7.4) (0.3 mg/ml), and then dialyzed against 10 mM phosphate buffer using
dialysis membranes with a molecular weight cutoff of 6000-8000. Aliquots of the DNA solution were bubbled with N
2
O or air for 30 min and subsequently irradiated with [gamma]-rays in a
60
Co [gamma]-source at a dose of 80 Gy (dose rate 56 Gy/min). Another aliquot of
the DNA solution was treated with 3 mM H
2
O
2
in the presence of 25 [mu]M FeCl
3
and 100 [mu]M EDTA at 4oC for 30 min. FeCl
3
and EDTA were mixed before addition to the DNA solution. Subsequently, all DNA
samples were dialyzed against 10 mM phosphate buffer (pH 7.4) for 18 h at 4oC.
The purification of homogenous Fpg protein from an overproducing strain of
E.coli
has been described elsewhere (
24
). The specific activity of enzyme preparations used for our studies was >2 * 10
8
u/mg.
Aliquots of 100 [mu]g of irradiated or H
2
O
2
/Fe(III)-EDTA-treated DNA samples were dried in a SpeedVac under vacuum. Samples
were then dissolved in a 1 ml Eppendorf tube in the incubation mixture
containing 50 mM phosphate buffer (pH 7.4), 100 mM KCl, 1 mM EDTA, 0.1 mM
dithiothreitol and bovine serum albumin (0.1 mg/ml). Where indicated, 1, 2 or 5
[mu]g of Fpg protein were added to each mixture. Some samples contained no Fpg
protein, but the equivalent amount of buffer (25 mM HEPES, 200 mM NaCl, 1 mM
EDTA and 50% glycerol). The total volume of the mixture was 240 [mu]l. The enzyme was inactivated by heating at 100oC for 15 min. Three replicates of each mixture were incubated at 37oC in a water bath. Incubation time varied depending on the
experiment.
For determination of excision as a function of the product concentration, 12.5, 20, 35, 50 and 75 [mu]g of irradiated or H
2
O
2
/Fe(III)-EDTA-treated DNA samples were supplemented with 87.5, 80, 65, 50 and 25 [mu]g of control DNA samples, respectively. An additional sample containing 100 [mu]g of irradiated or H
2
O
2
/Fe(III)-EDTA-treated DNA samples was also used. Three replicates of these samples were incubated with 1 or 2 [mu]g Fpg protein, or without Fpg protein at 37oC for 15 min as described above. For determination of
excision as a function of incubation temperature, three replicates of 100 [mu]g aliquots of irradiated or H
2
O
2
/Fe(III)-EDTA-treated DNA samples were incubated with 1 [mu]g Fpg protein or without Fpg protein at 5, 15, 20, 25, 30 and 37oC for 15 min.
Following incubation, 540 [mu]l of cold ethanol (-20oC) were added to each sample. Samples were kept at -20oC for 2 h and then centrifuged at 4oC for 30 min at 10000 r.p.m. DNA pellets and
supernatant fractions were separated. The pellets were washed with 100 [mu]l of a cold mixture (-20oC) of ethanol and water (80/20; v/v). DNA pellets were dried in
a SpeedVac under vacuum, 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). The recovery of DNA by precipitation with ethanol was ~100%.
Aliquots of stable isotope-labeled analogues of modified DNA bases were added as internal standards
to pellets with known amounts of DNA and to the supernatant fractions (
19
). Pellets 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 18 h.
Supernatant fractions were freed of ethanol in a SpeedVac under vacuum for 30
min, and then lyophilized for 18 h. Supernatant fractions were not hydrolyzed. The derivatization of the lyophilized samples and the subsequent analysis by GC/IDMS with selected-ion monitoring were performed as described elsewhere (
25
).
Three different DNA substrates were used in this work. Sixteen and 12 modified
bases were identified and quantified by GC/IDMS in DNA samples damaged under
anoxic (irradiation under N
2
O) and oxic conditions [H
2
O
2
/Fe(III)-EDTA-treatment or irradiation under air], respectively. These modified
bases were FapyGua, 8-OH-Gua, FapyAde, 8-OH-Ade, 2-hydroxyadenine, 5-hydroxy-5-methylhydantoin, 5-hydroxyhydantoin, 5-OH-Ura, 5-OH-Cyt, 5-(hydroxymethyl)uracil, thymine glycol, 5,6-dihydroxyuracil, 5,6-dihydrothymine, 5,6-dihydrouracil, 5-hydroxy-6-
hydrothymine and 5-hydroxy-6-hydrouracil. Oxygen inhibits the formation of the latter four
compounds (
26
,
27
). The levels of the modified bases in H
2
O
2
/Fe(III)-EDTA-treated DNA, in DNA irradiated under N
2
O and in DNA irradiated under air were 4.9+-0.2 (S.D.), 7.0+-0.3 and 7.8+-0.3 lesions per 10
3
DNA bases, respectively.
8-OH-Gua, FapyGua and FapyAde were excised from all three DNA substrates,
in agreement with a previous report (
15
). Figure
1
illustrates the excision of these products from H
2
O
2
/Fe(III)-EDTA-treated DNA. The amounts recovered in the pellets of DNA samples
incubated with the heat-inactivated enzyme were similar to those found after incubation without
the enzyme (data not shown). These experiments show that the amounts of
products in the supernatant fractions of DNA samples incubated with the active
enzyme were similar to the amounts removed from the pellets of the same
samples. None of the other lesions in the DNA substrates were excised
significantly.
We have studied the kinetics of excision of purine lesions by
E.coli
Fpg protein from three differently prepared DNA substrates. Each DNA substrate
contained a variety of pyrimidine- and purine-derived lesions. In this respect, this work differs from structure-function studies that use a single chemically-defined lesion embedded in an oligonucleotide substrate
at a defined position (
14
,
17
,
29
). Previously, the excision of 8-OH-Gua, FapyGua and FapyAde by
E.coli
Fpg protein from oxidatively damaged DNA has been demonstrated using GC/MS;
however, the kinetics of excision of these lesions has not been reported (
15
). The present paper is the first report on the simultaneous measurement of the
kinetics of excision of 8-OH-Gua, FapyGua and FapyAde by
E.coli
Fpg protein from oxidatively damaged DNA. The salient feature of this work is
the evidence that these purine lesions were excised by Fpg protein with similar
specificity from various oxidatively damaged DNA substrates and there was a
significant influence of the nature of DNA substrate on the efficiency of
excision. The results suggest that, in addition to 8-OH-Gua which is thought to be the main physiological substrate of Fpg
protein (
14
,
17
,
29
), FapyGua and FapyAde may be important physiological substrates of this enzyme.
It should be pointed out that calf thymus DNA was used as a model for the
present study. The experimental conditions used may not be necessarily similar
to those found
in vivo
.
Values for
k
cat
(maximum velocity/[enzyme]) and
K
M
were generally higher for excision of 8-OH-Gua, FapyGua and FapyAde from H
2
O
2
/Fe(III)-EDTA-treated DNA than from other substrates. Kinetic constants varied
significantly. For example,
k
cat
for excision of FapyGua from H
2
O
2
/Fe(III)-EDTA-treated DNA was twice and seven times as high as
k
cat
for excision from DNA substrates irradiated under N
2
O and under air, respectively. Values for
k
cat
/
K
M
were greater for excision from DNA irradiated under N
2
O than from other substrates, indicating the preferred excision of lesions from
this DNA substrate. The preference for FapyGua excision was pronounced as
k
cat
/
K
M
was almost twice as high as the comparable value for H
2
O
2
/Fe(III)-EDTA-treated DNA and three to five times higher than for DNA irradiated
under air. The
k
cat
/
K
M
ratios for excision from H
2
O
2
/Fe(III)-EDTA-treated DNA were greater than those for excision from DNA irradiated
under air. Taken together, these results indicate a strong dependence of
excision of purine lesions by Fpg protein on the nature of the DNA substrate.
A comparison of kinetic constants for the same DNA substrate reveals that
k
cat
/
K
M
ratios for excision of 8-OH-Gua, FapyGua and FapyAde from H
2
O
2
/Fe(III)-EDTA-treated DNA were similar, indicating comparable preference of the
enzyme for these lesions. The
k
cat
/
K
M
ratio for excision of FapyGua from DNA irradiated under N
2
O was significantly greater than that for excision of 8-OH-Gua and similar to that for excision of FapyAde. FapyGua and FapyAde
showed similar
k
cat
/
K
M
ratios for DNA irradiated under air. In this case,
k
cat
/
K
M
for excision of 8-OH-Gua was significantly greater than for excision of FapyGua and
FapyAde. When the concentration of the enzyme was doubled, no significant
differences between the
k
cat
/
K
M
ratios for excision of 8-OH-Gua, FapyGua and FapyAde were observed for any of the DNA
substrates. These results suggest a comparable preference of the enzyme for all
three lesions.
The measurement of base excision as a function of temperature revealed a
profound dependence of this reaction on temperature. 8-OH-Gua had the highest activation energy (E
a
) in the case of H
2
O
2
/Fe(III)-EDTA-treated DNA and DNA irradiated under air. In both cases, E
a
of FapyGua was greater than that of FapyAde. With DNA irradiated under N
2
O, E
a
of 8-OH-Gua was similar to that of FapyGua, but greater than that of
FapyAde. FapyAde had the lowest E
a
in all cases. Significant differences between DNA substrates were noted. Values
of E
a
for excision from DNA irradiated under N
2
O were higher than those for excision from other substrates. Values of E
a
were lowest for DNA irradiated under air. Taken together, the results indicate
a significant dependence of activation energies on the nature of the DNA
substrate and a significant difference between activation energies for each DNA
substrate tested.
Differences in the mechanisms by which H
2
O
2
/Fe(III)-EDTA and ionizing radiation cause damage to DNA may affect the excision
rates of modified purines. Oxygen profoundly modifies radiation damage in DNA (
30
), and alters the distribution of lesions (
3
,
26
,
27
). Ionizing radiation may generate multiply damaged sites consisting of base
damages and strand breaks within less than 20 base pairs (
31
). Closely spaced lesions on opposite strands of DNA may present difficulties
for DNA repair enzymes (
32
,
33
). There may be differences between end groups due to strand breaks in DNA. The
ratio of base damage to strand breaks among DNA substrates and the nucleotide
sequence context of damage may vary. In fact, there is evidence that the
excision of 2,6-diamino-4-hydroxy-5-
N
-methylformamidopyrimidine by Fpg protein from methylated, alkali-treated DNA shows some sequence specificity (
34
). Some or all of these factors may significantly influence the kinetics of
excision of modified bases from different DNA substrates. The methodology used
in this work is not conducive to a precise molecular explanation of the effect
of different DNA substrates on excision rates.
Excision of synthetically prepared 5-OH-Ura and 5-OH-Cyt by Fpg protein from defined oligonucleotides
containing a single lesion has been observed (
18
). In our study, these oxidized pyrimidines were not excised from any of three
DNA substrates tested, all of which contained substantial amounts of these
products. Generally, enzymatic excision of single oxidized bases from
oligodeoxynucleotide substrates parallels their excision from oxidized DNA. The
reason for the apparent discrepancy is unclear.
The action of Fpg protein on defined oligodeoxynuclotides containing 8-OH-Gua and structurally related lesions has been systematically
investigated; binding parameters and specificity constants were related to the
three dimensional structure of duplex DNA containing either 8-OH-Gua:Cyt or 8-OH-Gua:Ade (
17
,
35
,
36
). The presence of the 8-oxo function in the major groove of DNA correlates with the apparent
binding affinity of Fpg protein for its several substrates. Both 8-oxo and 6-oxo functions are present in the major groove in bases that were
efficiently excised from duplex DNA. A catalytic mechanism involving attack at
C1' by a nucleophilic residue in Fpg protein, accompanied by O-protonation of the 6-oxo group has been postulated to explain these observations (
17
,
37
). Similar mechanisms have been proposed for Fpg protein and other DNA
N
-glycosylases by Lloyd
et al
. (
38
,
39
). The kinetic data presented here for excision of 8-OH-Gua and FapyGua support these mechanisms and account for the lack of
excision that was noted when Fpg protein acts on damaged DNA containing 8-OH-Ade, 5-OH-Ura or 5-OH-Cyt. However, they fail to explain
satisfactorily the efficient excision of FapyAde. The latter may exist in two rotameric forms
as was shown for another formamidopyrimidine (
40
), and the conformational flexibility inherent in ring-opened structures may be important if modified bases are flipped out of
the duplex prior to excision by the cognate DNA glycosylase (
41
). Comparative studies of the chemical and biological properties of 8-OH-Ade and FapyAde will be required to elucidate the interaction of
these lesions with Fpg protein.
In conclusion, we have shown that Fpg protein excised 8-OH-Gua, FapyGua and FapyAde efficiently from three heterogenous DNA
substrates. The rates of excision of these lesions among DNA substrates varied,
even though each substrate contained similar levels of oxidative damage.
Differences in the mechanisms of DNA damage may account for these observations. The simultaneous measurement of the kinetic constants revealed that the enzyme
had similar specificity for 8-OH-Gua, FapyGua and FapyAde. These data suggest that, in addition to 8-OH-Gua, FapyGua and FapyAde may be important physiological
substrates for this enzyme. The results also emphasize the utility of the
GC/IDMS technique as an assay of high specificity and sensitivity to measure
the kinetics of excision of oxidatively damaged bases from DNA by DNA repair
enzymes.
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.
Mr Dmitry Zharkov of State University of New York at Stony Brook kindly provided
the Fpg protein used in this study. We are grateful to Dr Thomas J. Buckley of
National Institute of Standards and Technology for his help in the linear least
squares and statistical analyses of the data. A.P.G. acknowledges support from
NIH grant CA17395.
REFERENCES
Return