496–502
 1999 Oxford University Press
Nucleic Acids Research, 1999, Vol. 27, No. 2
Insertion of dGMP and dAMP during in vitro DNA
synthesis opposite an oxidized form of
7,8-dihydro-8-oxoguanine
Victor Duarte, James G. Muller and Cynthia J. Burrows*
Department of Chemistry, University of Utah, 315 South 1400 East, Salt Lake City, UT 84112-0850, USA
Received October 1, 1998; Revised and Accepted November 18, 1998
ABSTRACT
INTRODUCTION
Oxidative damage to DNA bases commonly results
in the formation of oxidized purines, particularly
7,8-dihydro-8-oxoguanine (8-oxoG) and 7,8-dihydro-8-oxoadenine (8-oxoA), the former being a wellknown mutagenic lesion. Since 8-oxoG is readily
subject to further oxidation compared with normal
bases, the insertion of a base during DNA synthesis
opposite an oxidized form of 8-oxoG was investigated
in vitro. A synthetic template containing a single 8-oxoG
lesion was first treated with different one-electron
oxidants or under singlet oxygen conditions and then
subjected to primer extension catalyzed by Klenow
fragment exo– (Kf exo–), calf thymus DNA polymerase α
(pol α) or human DNA polymerase β (pol β). Consistent
with previous reports, dAMP and dCMP are inserted
selectively opposite 8-oxoG with all three DNA polymerases. Interestingly, oxidation of 8-oxoG was found
to induce dAMP and dGMP insertion opposite the
lesion by Kf exo– with transient inhibition of primer
extension occurring at the site of the modified base.
Furthermore, the lesion constitutes a block during
DNA synthesis by pol α and pol β. Experiments with an
8-oxoA-modified template oligonucleotide show that
both 8-oxoA and an oxidized form of 8-oxoA direct
insertion of dTMP by Kf exo–. Mass spectrometric
analysis of 8-oxoG-containing oligonucleotides before
and after oxidation with IrCl62– are consistent with
oxidation of primarily the 8-oxoG site, resulting in
formation of a guanidinohydantoin moiety as the major
product. No evidence for formation of abasic sites was
obtained. These results demonstrate that an oxidized
form of 8-oxoG, possibly guanidinohydantoin, may
direct misreading and misinsertion of dNTPs during DNA
synthesis. If such a process occurred in vivo, it would
represent a point mutagenic lesion leading to G→T and
G→C transversions. However, the corresponding
oxidized form of 8-oxoA primarily shows correct
insertion of T during DNA synthesis with Kf exo–.
Intracellular DNA damage caused by reactive oxygen species is
known to be mutagenic (1–3) and is possibly involved in the
aging process and many human diseases including cancer (4–6).
The mutagenicity of the damaged DNA arises partly during its
replication by a DNA polymerase (7). Thus, polymerase fidelity
during DNA replication is essential to prevent mutagenesis and
consequently carcinogenesis. Free radical forms of oxygen are
probably the most important source of spontaneous DNA
damage. Oxidative damage to DNA bases commonly results in
the formation of oxidized purines, particularly 7,8-dihydro-8-oxoguanine (8-oxoG) and, to a lesser extent, 7,8-dihydro-8-oxoadenine
(8-oxoA) (8,9). In vitro experiments concerning the mutagenic
potential of 8-oxoG have shown that DNA polymerases can
synthesize past the 8-oxoG lesion and incorporate selectively
dAMP in addition to dCMP opposite the damage (2,10–12).
Furthermore, the ratio between dAMP and dCMP insertion
depends on the polymerase used. Replicative DNA polymerases
insert preferentially dAMP, while the DNA polymerases associated
with repair processes incorporate the correct dCMP (11).
Oxidative DNA damage may be repaired in cells by a variety
of repair enzymes. In both bacteria and mammalian cells, repair
enzymes have been discovered that possess multiple activities
toward oxidative DNA damages. A repair system for the 8-oxoG
lesion involving several enzymes has been elucidated in Escherichia
coli by Michaels et al. (13,14). The MutM protein, also known as
Fpg (15) or 8-oxoguanine glycosylase, removes the 8-oxoG
lesion from the double-stranded DNA when paired to a cytosine.
On the other hand, the 8-oxoG·A mismatch is repaired by a
different DNA glycosylase, MutY. This protein removes the
unmodified base adenine when it is paired with the 8-oxoG
(13,14). Another protein, MutT, is an 8-oxo-dTPase that prevents
incorportation of 8-oxo-dGTP into DNA (16). The mutagenic
process resulting from an 8-oxoG-modified base and the repair
system evolved to prevent these mutations have been extensively
studied. However, the mutagenic potential of an oxidized
8-oxoguanine is unknown.
It has long been recognized that 8-oxoG is readily subjected to
further oxidation since electrochemical detection is used for its
analysis by HPLC (17). A redox potential in the range of
*To whom correspondence should be addressed. Tel: +1 801 585 7290; Fax: +1 801 585 7868; Email: [email protected]
497
Nucleic Acids
Acids Research,
Research,1994,
1999,Vol.
Vol.22,
27,No.
No.12
Nucleic
0.58–0.75 V versus normal hydrogen electrode (NHE) has been
reported for 8-oxo-2′-deoxyguanosine (18–20), significantly
lower than that of 2′-deoxyguanosine nucleoside (1.29 V versus
NHE) (21). In vitro, the products of oxidation of 8-oxoguanosine
have been analyzed under various reaction conditions, although
products in oligonucleotides are unknown. Since the oxidation
product profile of the parent heterocycle guanine differs in
nucleosides versus duplex oligomers, caution must be exercised
in extrapolating nucleoside products to those of the biopolymer(22).
Electrochemical oxidation of 8-oxoguanosine was proposed to form
the corresponding guanidinohydantoin ribonucleoside (Fig. 7),
presumably via the 5-hydroxy-8-oxoG intermediate (23). It is
noteworthy that this pathway is completely analogous to oxidative
degradation of urate, which is converted both chemically and
enzymatically to allantoin, the urea analog of guanidinohydantion,
via 5-hydroxyurate (24). Photo-oxidation studies involving
singlet oxygen generation led to different final products for the
oxidation of 8-oxo-2′-deoxyguanosine (25–29). Raoul and Cadet
(25) found cyanuric acid as the major product of decomposition
of 8-oxoG with singlet oxygen, but interestingly, imidazolone and
oxazolone derivatives were also formed. These products have
been observed in the photochemical one-electron oxidation of
guanosine itself, in competition with formation of 8-oxoG
(22,30,31). Sheu and Foote (29) reported the formation of
cyanuric acid, parabanic acid and a seven-membered ring product
for the oxidation of a silylated 8-oxoguanosine under conditions
of singlet oxygen generation. Because 8-oxoG is easily oxidized,
it is reasonable to assume that once 8-oxoG is formed, the
products of further oxidative degradation of 8-oxoG may play a
role in the mutagenic process associated with DNA damage.
The 8-oxoA-modified base has received less attention in the
literature since the first reports on the mutagenic potential of this
8-oxopurine suggested that during in vitro DNA synthesis, dTMP
was almost exclusively inserted opposite 8-oxoA (32). Wood et al.
indicated that 8-oxoA was at least an order of magnitude less
mutagenic than 8-oxoG in bacteria (33). However, Kamiya et al.
(34) reported the insertion of dGMP, or dAMP plus dGMP, using
DNA polymerases α or β, respectively. Little is known about the
products of further oxidation of 8-oxoA since 8-oxoA is not as
easily oxidized as 8-oxoG (0.92 and 0.58 V versus NHE,
respectively) (18).
We recently reported the selective oxidation of 8-oxoG- and
8-oxoA-containing oligonucleotides by the one-electron oxidant
IrCl62– (35). Moreover, the conversion of these 8-oxopurines into
further oxidized products sensitive to piperidine allows their
detection by gel electrophoresis. IrCl62– is a convenient, watersoluble, one-electron oxidant that does not exhibit any binding
interactions with DNA because of its negative charge. Its redox
potential, 0.90 V versus NHE (35), is convenient for oxidation of
only oxopurines without significant reaction with normal DNA
bases. A number of transition metal complexes also carry out
one-electron oxidation of purines (36), including the carcinogenic
metal nickel (37). Organonickel(II) complexes, including those
of peptide ligands, are redox active in the presence of certain
oxidants leading to oxidation of guanine and 8-oxoguanine in DNA
(38). These species have been studied both as models for nickel
carcinogenesis (39) and as probes of DNA (or RNA) structure due
to their ability to oxidatively modify guanine residues (40).
In the present work, we used IrCl62– in addition to two nickel
catalysts along with HSO5– or SO32–/O2 as one-electron oxidants
as well as singlet oxygen to study the mutagenic potential of
497
oxidized forms of 8-oxoG and 8-oxoA. First, we examined
nucleotide insertion opposite an oxidized form of 8-oxoG during
in vitro DNA synthesis catalyzed by Kf exo–, pol α or pol β, and
the oxidation product of 8-oxoA was studied with Kf exo–. The
present results demonstrate that an oxidized form of 8-oxoG
induces insertion of dAMP and dGMP when DNA synthesis is
performed with Kf exo–, suggesting that it may be a potent
mutagenic lesion leading to G→T and G→C transversions.
However, further oxidation of 8-oxoG represents a block during
replication of the template by pol α and pol β. Electrospray mass
spectrometry was used to identify the major lesion produced from
IrCl62– oxidation as guanidinohydantoin. Secondly, experiments
conducted with 8-oxoA and an oxidized form of 8-oxoA show that
this 8-oxopurine or a further oxidized product of 8-oxoA induces
mainly dTMP insertion during DNA synthesis by Kf exo–.
MATERIALS AND METHODS
Materials
Reagents and enzymes were purchased from the following
sources: Na2IrCl6 from Alfa Aesar, (Ward Hill, MA), Rose
Bengal from Acros (Norcross, GA), potassium monoperoxysulfate
(Oxone) and sodium sulfite were from Sigma and Fluka
(Ronkonkoma, NY), respectively. 8-OxoG and 8-oxoA phosphoramidites were from Glen Research (Sterling, VA), dNTPs from
Pharmacia (Piscataway, NJ), [γ-32P]ATP from Amersham
(Arlington Heights, IL), T4 polynucleotide kinase and Klenow
fragment (3′→5′exo–) were from New England Biolabs
(Beverly, MA). Calf thymus pol α and human pol β were
purchased from Molecular Biology Resources (Milwaukee, WI).
Oligodeoxynucleotides were synthesized with an Applied Biosystems synthesizer (ABI 392 B) using the manufacturer’s
protocols and incorporating 0.25 M into the final manual
deprotection step of oligonucleotides containing 8-oxopurines
(41). Purification was carried out by PAGE using 15 or 20%
polyacrylamide/7 M urea. Oligonucleotides were 5′-end-labeled
using T4 polynucleotide kinase and [γ-32P]ATP. Synthesis of the
complex NiCR (42,43) and the peptide KGH-NH2 (44) were
previously described. The NiKGH-NH2 complex was formed in
situ by the addition of 1 equivalent of Ni(CH3CO2)2 to an aqueous
solution of KGH-NH2 followed by the addition of two equivalents
NaOH (38).
DNA oxidation
Oxidation reactions were carried out with 5 µM oligonucleotides
containing 8-oxoG or 8-oxoA in a 10-µl solution of 10 mM NaPi
buffer (pH 7) with 100 mM NaCl. For the oxidation with IrCl62–,
the template was incubated for 1 h at room temperature in the
presence of 100 µM IrCl62– (35). For the experiments with singlet
oxygen, the reactions were performed with Rose Bengal (50 µM)
for 30 min at 12C under a 300 W tungsten lamp. The template
oxidations with NiCR (42) or NiKGH-NH2 (38) were conducted
as previously described. The DNA was incubated with 5 µM
NiCR or 10 µM NiKGH-NH2 in the presence of 20 µM KHSO5
for 30 min at room temperature or 40 µM Na2SO3 for 1 h at room
temperature, respectively.
498
Nucleic Acids Research, 1999, Vol. 27, No. 2
Figure 1. Sequences of oligodeoxynucleotides.
Primer extension
Reactions catalyzed by Kf exo– were carried out in 10 µl
solutions of 10 mM Tris–HCl (pH 7.5) plus 5 mM MgCl2 and
7.5 mM DTT containing the oligonucleotide template and the
5′-end-labeled primer (Fig. 1) with a template/primer nanomolar
ratio: 1/4 = 50/15; 2/4 and 3/4 = 90/15. The solution was heated
at 90C for 5 min then cooled to room temperature over a period
of 2.5 h. DNA polymerization reactions were conducted with
100 µM of a single dNTP or 100 µM of a mixture of all four
dNTPs and 0.02 U of Kf exo– (for duplex 1/4) or 0.1 U of Kf exo–
(for duplexes 2/4 and 3/4) at 37C for 20 min. Reactions were
stopped by adding 5 µl of termination solution (95% formamide,
0.1% bromophenol blue and 0.1% xylene cyanol). The samples
were then heated at 95C for 3 min, applied to a 15%
polyacrylamide gel in the presence of 7 M urea and then analyzed
by phosphorimagery (Molecular Dynamics Storm 840) using
ImageQuaNT software.
For the experiments with pol α and pol β, the template and the
5′-end-labeled primer (nanomolar ratio 1/4, 2/4, 3/4 = 50/15)
were annealed as described above in 10 µl solutions of 50 mM
Tris–HCl (pH 8) plus 10 mM MgCl2, 2 mM DTT and 0.5 µg/µl
BSA. For DNA synthesis, the reaction mixtures containing the
desired duplex, 100 µM of a single dNTP or 100 µM of a mixture
of the four dNTPs and 0.05 U of pol α or pol β were incubated
at 25C for 1 h. The reactions were stopped and the samples
analyzed as described above.
Mass spectrometric analysis of oxidized
8-oxoguanine-containing oligonucleotides
Oxidations were carried out with 10–50 µl samples containing
12 µM of oligo 5 and 100 µM IrCl62– in 10 mM NaPi buffer
(pH 7) and 100 mM NaCl for 60 min. Reactions were quenched
by the addition of 2 µl of a 50 mM HEPES/250 mM EDTA (pH 7)
solution. The reactions were combined and dialyzed (2 K Dalton
MWCO) against water for 24 h and then against 10 mM
ammonium acetate for an additional 24 h. The existing solution
was lyophilized to dryness and dissolved in 50 µl of water and
50 µl of 10 M ammonium acetate. After 2 h, 150 µl of cold
ethanol–isopropanol (1:1) mixture was added and the DNA was
precipitated for 8 h at –80C. After incubation, the sample was
centrifuged (13 000 r.p.m.) for 30 min at 4C and the supernatant
liquid was carefully removed leaving the DNA as a pellet. After
an additional DNA precipitation using the same procedure, the
DNA pellet was lyophilized for 2 h prior to submission for
electrospray mass spectrometric (ESI-MS) analysis. MS analysis
was carried out on a Micromass Quattro I at the University of
Illinois Mass Spectrometry Facility. A portion of the same sample
Figure 2. Nucleotide incorporation opposite G (template 1), 8-oxoG (template
2) or an oxidized form of 8-oxoG. Primer extension reactions were catalyzed
by Kf exo– using a single dNTP (lanes A, C, G and T) or a mixture of the four
dNTPs (lane Ex). Reaction conditions are described in the text.
used for MS was retained and radiolabeled with a 5′-32P-phosphate,
subjected to hot piperidine treatment (0.2 M, 60 min) and
analyzed by PAGE. The procedure for this has already been
described elsewhere (35).
RESULTS
Insertion of dNTPs opposite an oxidized form of 8-oxoG
To examine the effects of an oxidized form of 8-oxoG during in
vitro DNA synthesis, a 40mer template 2, containing 8-oxoG at
position 10 from the 5′ end (Fig. 1) was subjected to various
oxidants and annealed to a 20mer primer, 4 (Fig. 1). The primer
was then extended (37C, 20 min) using three different DNA
polymerases (0.1 U) in the presence of a single dNTP (100 µM)
or a mixture of all four dNTPs (100 µM). An unmodified template
1, containing a guanine at position 10, and the unoxidized
template 2 were used as controls for nucleotide insertion. In the
case of the unmodified template 1, 0.02 U enzyme concentration
was sufficient.
Results of primer extension reactions catalyzed by Kf exo– are
shown in Figure 2. DNA synthesis on the unmodified template 1
(lanes 1–5) led to the expected insertion of dCMP opposite dG at
position 10 (lane 3). In the presence of the four dNTPs (lane 1),
extension of the oligomer to approximately full length was
observed. When the 8-oxoG-modified oligonucleotide 2 was
used as the template (lanes 6–10), dAMP and dCMP were
preferentially inserted opposite the lesion (lanes 7 and 8), as
expected. In the presence of all four dNTPs, we observed a
transient inhibition opposite the 8-oxoG damage (lane 6) but
essentially full length product was also obtained. The relative
amounts of dAMP and dCMP insertion are dependent on the
concentration of enzyme used in the reaction and on the
polymerization time (data not shown). dAMP insertion opposite
8-oxoG, first reported by Shibutani et al. (11), is representative of
the mutagenic potential of 8-oxoG during DNA replication.
Interestingly, when the 8-oxoG-containing template 2 was
selectively oxidized by IrCl62–, Kf exo– inserts exclusively
dAMP and dGMP opposite the damage (lanes 12 and 14). When
all the dNTPs were present in the reaction, the over-oxidized
8-oxopurine was a stop point for the enzyme and only a small
amount of fully extended primer was observed (lane 11), although
the integrity of this complement was not examined. Similar
499
Nucleic Acids
Acids Research,
Research,1994,
1999,Vol.
Vol.22,
27,No.
No.12
Nucleic
Figure 3. DNA synthesis opposite an oxidized form of 8-oxoG after treatment
of template 2 by Ni complexes in the presence of KHSO5 or Na2SO3/O2. Primer
extension reactions were catalyzed by Kf exo– using a single dNTP (lanes A,
C, G and T) or a mixture of the four dNTPs (lane Ex). Reaction conditions are
described in the text.
results were obtained when template 2 was oxidized in the
presence of singlet oxygen. The oxidized form of 8-oxoG induces
selective insertion of dAMP as well as dGMP (lanes 16 and 18)
opposite the lesion, with strong inhibition during DNA synthesis
when all dNTPs are present (lane 20).
Oxidation of 8-oxoG by nickel complexes that are models for
nickel-mediated DNA damage and toxicity(45) was also studied,
and the resulting template 2 was subjected to primer extension
reaction catalyzed by Kf exo–. The results of nucleotide insertion
by Kf exo– after treatment of 2 by NiCR/HSO5– or NiKGHNH2/SO32–/O2 are presented in Figure 3. In both cases, oxidation
of 8-oxoG by NiCR (lanes 1–5) and NiKGH-NH2 (lanes 6–10)
in the presence of HSO5– and SO32–/O2, respectively, led to
dAMP and dGMP insertion opposite the lesion (lanes 1 and 3, and
6 and 8). With the four dNTPs in the reaction mixture, DNA
polymerization by Kf exo– is strongly inhibited, and no full or
partial extension of the primer was obtained (lanes 5 and 10).
Treatment of 2 by HSO5– or SO32–/O2 without nickel complexes
induced insertion of dAMP and dCMP as expected opposite
8-oxoG (data not shown), indicating that there is little background
reaction of the oxidants in the absence of nickel catalysts.
DNA synthesis catalyzed by pol α and pol β were also
conducted on template 2 after treatment by IrCl62–, and the results
of primer extension are presented in Figure 4. With both enzymes,
extension of the 8-oxoG containing template led to the expected
insertion of dAMP and dCMP (lanes 1 and 2 for pol α, lanes 11
and 12 for pol β). Consistent with previous reports, pol α inserts
predominantly dAMP compared with dCMP (11). The 8-oxoG
lesion retarded the extension but some full length product was
obtained (lanes 5 and 15). Interestingly when 8-oxoG is oxidized
in the presence of IrCl62–, the resulting lesion inhibits the primer
extension by pol α (lanes 6–10) and pol β (lanes 16–20) beyond
the damage. An oxidized form of 8-oxoG may induce a change
in the conformation of the template opposite the damage not
suitable for DNA synthesis by pol α and pol β.
Insertion of dNTPs opposite an oxidized form of 8-oxoA
The 8-oxoA-modified template 3, incubated in the presence of
IrCl62– or NiCR/HSO5–, was used to examine the miscoding
potential of an oxidized form of 8-oxoA. The relative insertion of
dNTPs opposite of 8-oxoA, or a further oxidized form of 8-oxoA,
catalyzed by Kf exo–, is shown in Figure 5. As previously
reported in the literature (32), Kf exo– inserts predominantly
dTMP opposite 8-oxoA (lane 5). In the presence of the four
dNTPs, the 8-oxoA lesion is a stop point for the enzyme but some
extended primer was also observed (lane 1). When 8-oxoA is
499
Figure 4. Nucleotide incorporation opposite 8-oxoG (template 2) or an
oxidized form of 8-oxoG. Primer extension reactions were catalyzed by calf
thymus pol α (lanes 1–10) and by human pol β (lanes 11–20) in the presence
of a single dNTP (lanes A, C, G and T) or a mixture of the four dNTPs (lane Ex).
Reaction conditions are described in the text.
Figure 5. Nucleotide incorporation opposite 8-oxoA (template 3) or an
oxidized form of 8-oxoA. Primer extension reactions were catalyzed by Kf
exo– using a single dNTP (lanes A, C, G and T) or a mixture of the four dNTPs
(lane Ex). Reaction conditions are described in the text.
subjected to further oxidation by IrCl62– (lanes 6–10) or NiCR/
HSO5– (lanes 10–15) prior to primer extension, similar results were
obtained. dTMP is mainly inserted opposite the damage; however,
small amounts of dAMP were also detected (lane 12). Moreover,
essentially fully extended primer was observed with transient
inhibition opposite the lesion (lanes 6 and 11).
Mass spectrometric analysis of oxidized 8-oxoG-containing
oligonucleotides
In order to examine the possible products of 8-oxoG oxidation
that might be responsible for misinsertion of dGMP and dAMP,
we carried out ESI-MS analysis of synthetic oligodeoxynucleotides.
An oligomer consisting of the first 18 residues of template 1 was
initially studied (data not shown), but ∼20% of the products
corresponded to imidazolone and oxazolone residues (although
the major products were as discussed below). Since these are
common one-electron oxidation products of G, and possible
oxidation products of 8-oxoG, it was not clear whether these
minor oxidation products arose from reaction at the single
8-oxoG site or from the five other Gs present in the 18mer. As a
consequence, we chose to study an alternative oligomer with a
single 8-oxoG site and no competing G oxidation sites, namely
oligo 5 (Fig. 1).
Both the starting material, oligo 5, and its IrCl62– oxidized
product were purified and subjected to ESI-MS analysis. As
500
Nucleic Acids Research, 1999, Vol. 27, No. 2
Figure 6. ESI-MS spectra of oligo 5 (dashed line) and of its IrCl62– oxidation product (solid line). Arrows ‘a’ and ‘b’ indicate expected positions for imidazolone and
oxazolone products, respectively.
shown in Figure 6, none of the starting material remained after
treatment with 100 µM IrCl62– for 1 h. The largest peak in the
spectrum represents an overall loss of 10 Da from the mass of the
starting material (5515 versus 5525 Da). This major product is
best explained as the guanidinohydantoin derivative, as shown in
Figure 7. Guanidinohydantoin was proposed to be the twoelectron, electrochemical oxidation product of 8-oxoguanosine
(23), and the pathway is entirely analogous to the well-known
oxidation of uric acid to form allantoin (20,24). The second
largest peak in the product spectrum occurred at a mass of 5541
(M+16) which is consistent with the addition of an oxygen atom to
8-oxoG. This nicely matches with the proposed intermediate,
5-hydroxy-8-oxoguanine, which is thought to be on the pathway to
guanidinohydantoin, just as 5-hydroxyurate is an intermediate in
allantoin formation (20,24). Interestingly, no evidence for
formation of an abasic site was obtained by mass spectral analysis
of the oxidized 8-oxoG containing oligomer, in contrast to a
recent mechanistic proposal concerning γ-irradiated 8-oxoG
lesions (46).
Since the two oligonucleotide products observed in the present
study are likely to give similar responses by ESI-MS, one can
estimate the relative amounts of various oxidation products
formed. Thus, the major products, guanidinohydantoin and
5-hydroxy-8-oxoG, are formed in an ∼70:30 ratio and constitute
>90% of the total products. Interestingly, only trace amounts, if
any, of the oxazolone (or more likely the ring-opened form,
guanidinooxalamide, M-37) and its precursor, imidazolone
(M-55), are observed. When compared with the initial study of a
G-rich 18mer where these products were significant, we now
conclude that they are not generally formed from the one-electron
oxidation of 8-oxoG, but rather from one-electron oxidation of
the parent residue, G.
Alternative interpretations of the mass spectral data are
possible, given the error limits in ESI-MS of approximately
±1 a.m.u. For example, our assignment of guanidinohydantoin
for the M-10 peak in the mass spectrum is based on the
observation of a logical intermediate, 5-hydroxy-8-oxoG
(M+16), its observation in nucleoside oxidation studies,(23) and
the strong analogy to urate oxidation (24). Alternatively, a
seven-membered ring product that has been observed in singlet
oxygen chemistry with 8-oxoG would also lead to an M-10 peak,
but a pathway for its formation by one-electron oxidation is not
evident. Thus, the most consistent explanation of the mass
spectral data is that one-electron oxidation of 8-oxoG leads
initially to 5-hydroxy-8-oxoguanine followed by hydrolysis to
the major product, guanidinohydantoin (Fig. 7, top).
In order to correlate the mass spectrometric results with gel
electrophoretic analysis of 8-oxoG oxidation, a portion of the
sample used in MS analysis was radiolabeled, piperidine treated,
and electrophoresed on a polyacrylamide gel. After 60 min of
piperidine treatment at 90C, the band corresponding to cleavage
at the oxidized 8-oxoG site represented 72% of the total intensity
in the lane, and no other cleavage sites were observed. This nicely
correlates with the ∼70:30 product ratio of guanidinohydantoin:5-hydroxy-8-oxoG and suggests that guanidinohydantoin is
a piperidine-labile lesion while 5-hydroxy-8-oxoG, like 8-oxoG,
is not.
DISCUSSION
In this study, the polymerase insertion of nucleotides opposite
oxidized forms of 8-oxoG and 8-oxoA has been investigated
using modified templates containing these two 8-oxopurine
lesions. Singlet oxygen or one-electron oxidation of 8-oxoG- and
8-oxoA-containing templates, 2 and 3 respectively, were used as
substrates for primer extension catalyzed by Kf exo–, pol α and
pol β. In vitro DNA synthesis opposite the lesion and full
extension of the primer past the damage were studied using a
single dNTP and a mixture of all four dNTPs, respectively.
Replication of the template containing 8-oxoG after oxidation by
IrCl62–, singlet oxygen and NiCR/HSO5– or NiKGHNH2/SO32–/O2 induces insertion of both dAMP and dGMP
501
Nucleic Acids
Acids Research,
Research,1994,
1999,Vol.
Vol.22,
27,No.
No.12
Nucleic
Figure 7. Scheme depicting products of 8-oxoG oxidation and their relative
masses. Electrochemical oxidation (23) and one-electron oxidation (this work)
lead to 5-hydroxy-8-oxoG and guanidinohydantoin. Singlet oxygen leads to a
complex array of products (25–29).
opposite the damage by Kf exo– (Figs 2 and 3). In contrast to the
behavior of 8-oxoG (2,10–12), dCMP is no longer inserted after
further treatment with an oxidant. This result also supports the
proposal that treatment with 100 µM IrCl62– leads to complete
conversion of 8-oxoG to new products, since any unoxidized
8-oxoG would be expected to show at least partial insertion of
dCMP. Complete oxidation is also supported by mass spectrometric results showing the absence of starting material (Fig. 6) as
well as related gel electrophoretic studies with oligonucleotides
(35).
Many reports focused on the oxidation of 8-oxoG suggested
that this modified base is easily oxidized, and an array of products
has already been identified for the degradation of the 8-oxoG
nucleoside under various oxidative conditions (23,25–27,29).
Until now, little was known about the hydrogen bonding, the
base-pairing properties and the potent mutagenicity of these
compounds. Our findings suggest that at least one of the products
of oxidation of 8-oxoG is chemically competent to lead to G→T
and G→C transversions, although further extension of oligomers
containing oxidized 8-oxopurines is much less efficient than
8-oxopurine-containing substrates. Similar nucleotide insertion
was observed when the 8-oxoG-containing template 2 was
oxidized by NiCR and NiKGH-NH2 in the presence of HSO5– or
SO32–/O2, respectively. Nickel complexes such as NiCR and
NiKGH-NH2 mediate oxidative DNA damage, predominantly at
guanine sites (38). Mechanistic studies have shown that the
reaction with guanines does not involve freely diffusible radicals
but a Ni(II) complex with a bound radical, such as a sulfate radical
(38). Furthermore, we have recently observed that the 8-oxoGmodified base present in an oligonucleotide is the preferential site
of reaction with these nickel complexes in the presence of HSO5–
or SO32–/O2 (unpublished results). The lower redox potential of
8-oxoG compared with guanine (0.58 and 1.29 V versus NHE,
respectively) explains the higher reactivity of the 8-oxoG
compared with the unmodified guanine. The reactivity of the
501
nickel-tripeptide complex associated with SO32–/O2 toward
8-oxoG, leading to insertion of purines rather than dCMP
provides some evidence for the toxicity of sulfite mediated by
nickel (45,47).
Primer extension reactions catalyzed by pol α and pol β on
template 2 in the presence of IrCl62– led to different results in
comparison with Kf exo–. Both enzymes were unable to insert a
nucleotide opposite the lesion, leading to a stop at the site of
damage. This result might suggest a change in the template
conformation such that the modified base is no longer a substrate
for pol α and pol β. Alternatively, small amounts of an incorrectly
inserted nucleotide may have been removed by exonuclease
activity of these polymerases, but this remains to be tested.
Nucleotide insertion opposite an oxidized form of 8-oxoG, seems
to be dependent on the DNA polymerase used; in the case of
8-oxoG-containing oligonucleotides, many studies revealed
differences among various polymerases on nucleotide insertion
opposite the damage (11,12,48). Shibutani et al. (11) found that
dAMP was inserted opposite 8-oxoG by pol α and pol δ, while
pol β and Klenow fragment favor insertion of dCMP. Furthermore,
Feig and Loeb (7) pointed out that the mutation spectrum induced
by DNA damage generated by oxygen radicals is highly
dependent on the DNA polymerase.
In conclusion, dAMP and dGMP insertion by Kf exo– provides
a molecular basis for misincorporations opposite an oxidized
form of 8-oxoG during DNA synthesis. Mass spectrometric
analysis is consistent with high-yield conversion of 8-oxoG to
guanidinohydantoin following treatment with IrCl62–, and this is
the first report of this lesion being observed in an oligomer. While
it is tempting to propose that this heterocycle is responsible for the
subsequent incorrect insertions observed, in principle, any or all
of the products in Figure 7 could be active mutagenic lesions.
Based on related work, it is likely that both of the nickel-catalyzed
reactions of 8-oxoG lead to the same guanidinohydantoin
product; however, 1O2 chemistry clearly yields a different
product distribution based on extensive literature surrounding
this reaction. It is therefore quite interesting that the entire group
of oxidation products (Fig. 7) gives the same dAMP + dGMP
incorporation. What these products have in common is that they
are all heterocycles of roughly the same dimensions as a
pyrimidine, and they (or their tautomers) present a variety of
H-bonding opportunities. This may indicate that purine insertion
is favored for Kf exo– on the basis of size and shape as opposed
to highly specific and complementary hydrogen bonding, as
already suggested by Moran et al. for non-hydrogen-bonding
nucleoside analogs (49,50). That very stable base pairs are not
formed may account for the lack of efficient extension beyond the
initial misinsertion. Thus, a tentative conclusion is that oxidative
degradation of purines to one-ring heterocycles generally leads to
purine insertion with Kf exo–.
Indeed, G→C and G→T transversions resulting from dAMP
and dGMP incorporation have been reported during replication of
DNA treated with oxygen radicals. McBride et al. (51) found that
Fe2+/oxygen damage to DNA is a non-random mechanism,
leading to single base substitution with the most frequent being
G→C transversions followed by C→T and G→T. McBride et al.
(52) also found, using singlet oxygen mediated DNA damage,
that among the transversions observed after DNA replication,
G→C transversion was predominant with G→T transversions
occurring at lower frequency. G→C transversions produced by
singlet oxygen have also been detected in a mutation assay in
502
Nucleic Acids Research, 1999, Vol. 27, No. 2
mammalian cells (53). More recently, Valentine et al. (54) studied
the DNA damage generated by peroxyl radicals. The spectrum of
mutations induced in Escherichia coli upon transfection of DNA
exposed to ROO· indicates predominantly G→C and G→T
transversions. The authors reported that the oxidative lesions at
G were sensitive to digestion by MutM, for which 8-oxoG, but not
imidazolone or oxazolone products (22,55), are substrates.
Efforts to detect 8-oxoG by HPLC/electrochemical analysis
suggest that 8-oxoG was not present and that the G→T
transversion was not caused by this 8-oxopurine. Taken together
these results suggest that in oxygen-mediated DNA damage,
8-oxoguanine is not responsible for G→C transversions, and may
not induce all of the G→T transversions observed during DNA
replication. Until now, oxidized products of 8-oxoG have not
been reported to be implicated in DNA mutagenesis. Our findings
suggest that an oxidized form of 8-oxoG may account for some
of the G→C and G→T transversions observed under oxidative
conditions.
Given the high susceptibility of 8-oxoG to further oxidation,
and the findings reported herein that the oxidized 8-oxoG lesion,
possibly the major product guanidinohydantoin, leads to insertion
of dGMP and dAMP during in vitro DNA synthesis, it will now
be important to learn how these lesions might be recognized and
excised by the DNA repair enzymes. Such studies are currently
in progress.
ACKNOWLEDGEMENTS
We thank Prof. T. Widlanski (Indiana University) and Prof. S.
Rokita (University of Maryland) for helpful discussions. Support
of this work by a research grant from the National Science
Foundation (CHE-9521216) and in part by a travel grant from
NSF (INT-9613313) is gratefully acknowledged.
REFERENCES
1 Basu,A.K. and Essigmann,J.M. (1988) Chem. Res. Toxicol., 1, 1–18.
2 Cheng,K.C., Cahill,D.S., Kasai,H., Nishimura,S. and Loeb,L.A. (1992)
J. Biol. Chem., 267, 166–172.
3 Wang,D., Kreutzer,D.A. and Essigmann,J.M. (1998) Mutat. Res., 400,
99–115.
4 Ames,B.N., Shigenaga,M.K. and Hagen,T.M. (1993) Proc. Natl Acad. Sci.
USA, 90, 7915–7922.
5 Ames,B.N. (1983) Science, 221, 1256–1264.
6 Floyd,R.A. (1990) Carcinogenesis, 11, 1447–1450.
7 Feig,D.I. and Loeb,L.A. (1994) J. Mol. Biol., 235, 33–41.
8 Breen,A.P. and Murphy,J.A. (1995) Free Radic. Biol. Med., 18, 1033–1077.
9 von Sonntag,C. (1987) The Chemical Basis of Radiation Biology.
Taylor and Francis, London.
10 Wood,M.L., Dizdaroglu,M., Gajewski,E. and Essigmann,J.M. (1990)
Biochemistry, 29, 7024–7032.
11 Shibutani,S., Takeshita,M. and Grollman,A.P. (1991) Nature, 349, 431–434.
12 Lowe,L.G. and Guengerich,F.P. (1996) Biochemistry, 35, 9840–9849.
13 Michaels,M.L., Tchou,J., Grollman,A.P. and Miller,J.H. (1992)
Biochemistry, 31, 10964–10968.
14 Michaels,M.L., Cruz,C., Grollman,A.P. and Miller,J.H. (1992) Proc. Natl
Acad. Sci. USA, 89, 7022–7025.
15 Boiteux,S., O’Connor,T.R., Lederer,F., Gouyette,A. and Laval,J. (1990)
J. Biol. Chem., 265, 3916–3922.
16 Maki,H. and Sekiguchi,M. (1992) Nature, 355, 273–275.
17 Shigenaga,M.K., Park,J.-W., Cundy,K.C., Gimeno,C.J. and Ames,B.N.
(1990) Methods Enzymol., 186, 521–530.
18 Yanagawa,H., Ogawa,Y. and Ueno,M. (1992) J. Biol. Chem., 267,
13320–13326.
19 Berger,M., Anselmino,C., Mouret,J.-F. and Cadet,J. (1990)
J. Liq. Chromatogr., 13, 929–940.
20 Goyal,R.N., Brajter-Toth,A. and Dryhurst,G. (1982) J. Electroanal. Chem.,
131, 181–202.
21 Steenken,S. and Jovanovic,S. (1997) J. Am. Chem. Soc., 119, 617–618.
22 Angelov,D., Spassky,A., Berger,M. and Cadet,J. (1997) J. Am. Chem. Soc.,
119, 11373–11380.
23 Goyal,R.N., Jain,N. and Garg,D.K. (1997) Bioelectrochem. Bioenerg., 43,
105–114.
24 Kahn,K., Serfozo,P. and Tipton,P.A. (1997) J. Am. Chem. Soc., 119,
5435–5442.
25 Raoul,S. and Cadet,J. (1996) J. Am. Chem. Soc., 118, 1892–1898.
26 Adam,W., Saha-Möller,C.R., Schönberger,A., Berger,M. and Cadet,J.
(1995) Photochem. Photobiol., 62, 231–238.
27 Adam,W., Saha-Möller,C.R. and Schönberger,A. (1996) J. Am. Chem. Soc.,
118, 9233–9238.
28 Sheu,C. and Foote,C.S. (1995) J. Am. Chem. Soc., 117, 6439–6442.
29 Sheu,C. and Foote,C.S. (1995) J. Am. Chem. Soc., 117, 474–477.
30 Cadet,J., Berger,M., Buchko,G.W., Joshi,P.C., Raoul,S. and Ravanat,J.-L.
(1994) J. Am. Chem. Soc., 116, 7403–7404.
31 Kino,K., Saito,I. and Sugiyama,H. (1998) J. Am. Chem. Soc., 120,
7373–7374.
32 Shibutani,S., Bodepudi,V., Johnson,F. and Grollman,A.P. (1993)
Biochemistry, 32, 4615–4621.
33 Wood,M.L., Esteve,A., Morningstar,M.L., Kuziemko,G.M. and
Essigmann,J.M. (1992) Nucleic Acids Res., 20, 6023–6032.
34 Kamiya,H., Miura,H., Murata-Kamiya,N., Ishikawa,H., Sakaguchi,T.,
Inoue,H., Sasaki,T., Masutani,C., Hanaoka,F., Nishimura,S. and Ohtsuka,E.
(1995) Nucleic Acids Res., 23, 2893–2899.
35 Muller,J.G., Duarte,V., Hickerson,R.P. and Burrows,C.J. (1998)
Nucleic Acids Res., 26, 2247–2249.
36 Thorp,H.H. (1995) Adv. Inorg. Chem., 43, 127–175.
37 Oller,A.R., Costa,M. and Oberdorster,G. (1997) Tox. Appl. Pharmacol.,
143, 152–166.
38 Muller,J.G., Hickerson,R.P., Perez,R.J. and Burrows,C.J. (1997)
J. Am. Chem. Soc., 119, 1501–1506.
39 Bal,W., Lukszo,J. and Kasprzak,K.S. (1996) Chem. Res. Toxicol., 9,
535–540.
40 Burrows,C.J. and Rokita,S.E. (1996) In Sigel,A. and Sigel,H. (eds),
Metal Ions in Biological Systems. Dekker, New York, pp. 537–559.
41 Torres,M.C., Rieger,R.A. and Iden,C.R. (1996) Chem. Res. Toxicol, 9,
1313–1318.
42 Muller,J.G., Chen,X., Dadiz,A.C., Rokita,S.E. and Burrows,C.J. (1992)
J. Am. Chem. Soc., 114, 6407–6411.
43 Karn,J.L. and Busch,D.H. (1969) Inorg. Chem., 8, 1149–1153.
44 Perez,R.J. (1997) PhD dissertation, University of Utah.
45 Hickerson,R.P., Duarte,V., Van Horn,J.D., Perez,R.J., Muller,J.G.,
Rokita,S.E. and Burrows,C.J. (1998) In Sarkar,B. (ed.), Metals and
Genetics. Plenum, New York, in press.
46 Doddridge,A.Z., Cullis,P.M., Jones,G.D.D. and Malone,M.E. (1998)
J. Am. Chem. Soc., 120, 10998–10999.
47 Shi,X., Dalal,N. and Kasprzak,K.S. (1994) Environ. Health Perspect., 102
(Suppl. 3), 209–217.
48 Furge,L.L. and Guengerich,F.P. (1997) Biochemistry, 36, 6475–6487.
49 Moran,S., Ren,R.X.-F. and Kool,E.T. (1997) Proc. Natl Acad. Sci. USA,
94, 10506–10511.
50 Moran,S., Ren,R.X.-F., Rumney,S.,IV and Kool,E.T. (1997) J. Am. Chem.
Soc., 119, 2056–2057.
51 McBride,T.J., Preston,B.D. and Loeb,L.A. (1991) Biochemistry, 30,
207–213.
52 McBride,T.J., Schneider,J.E., Floyd,R.A. and Loeb,L.A. (1992) Proc. Natl
Acad. Sci. USA, 89, 6866–6870.
53 Costa de Oliveira,R., Ribeiro,D.T., Nigro,R.G., Mascio,P.D. and
Menck,C.F.M. (1992) Nucleic Acids Res., 20, 4319–4323.
54 Valentine,M.R., Rodriquez,H. and Termini,J. (1998) Biochemistry, 37,
7030–7038.
55 Spassky,A. and Angelov,D. (1997) Biochemistry, 36, 6571–6576.