Hydantoin derivative formation from oxidation of 7,8-dihydro-8-oxo- 2′-deoxyguanosine (8-oxodG) and incorporation of 14C-labeled 8-oxodG into the DNA of human breast cancer cells

Article abstract of DOI:10.1016/j.bmcl.2005.05.113

One-electron oxidation of 7,8-dihydro-8-oxo-2′-deoxyguanosine (8-oxodG) yielded a guanidinohydantoin derivative (dGh) and a spiroiminodihydantoin derivative (dSp), both putatively mutagenic products that may be formed in vivo. The nucleoside dGh was the major product at room temperature, regardless of pH. The results are contrary to previously published model studies using 2′,3′,5′-triacetoxy-8-oxo-7,8- dihydroguanosine (Luo, W.; Miller, J. G.; Rachlin, E. M.; Burrows, C. J. Org. Lett. 2000, 2, 613; Luo, W.; Miller, J.G.; Rachlin, E.M.; Burrows, C.J. Chem. Res. Toxicol. 2001, 14, 927), who observed a spiroiminodihydantoin derivative as the major product at neutral pH. Clearly, the functional groups attached to the ribose moiety of 8-oxodG influence the oxidation chemistry of the nucleobase derivative. To explore this chemistry in vivo, ¹⁴C-labeled 8-oxodG was synthesized and incubated with growing MCF-7 human breast cancer cells, resulting in the incorporation of the compound into cellular DNA as measured by a novel accelerator mass spectrometry assay.

Full text of DOI:10.1016/j.bmcl.2005.05.113

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Bioorganic & Medicinal Chemistry Letters 15 (2005) 3627–3631
Hydantoin derivative formation from oxidation
of 7,8-dihydro-8-oxo-20-deoxyguanosine (8-oxodG)
and incorporation of C-labeled 8-oxodG into the DNA
14
of human breast cancer cells
Sang Soo Hah, Hyung M. Kim, Rhoda A. Sumbad and Paul T. Henderson*
Biosciences Directorate and Center for Accelerator Mass Spectrometry, Lawrence Livermore National Laboratory,
7000 E. Avenue, L-441, Livermore, CA 94551, USA
Received 8 February 2005; revised 4 May 2005; accepted 10 May 2005
Abstract—One-electron oxidation of 7,8-dihydro-8-oxo-20-deoxyguanosine (8-oxodG) yielded a guanidinohydantoin derivative
(dGh) and a spiroiminodihydantoin derivative (dSp), both putatively mutagenic products that may be formed in vivo. The nucle-
oside dGh was the major product at room temperature, regardless of pH. The results are contrary to previously published model
studies using 20,30,50-triacetoxy-8-oxo-7,8-dihydroguanosine (Luo, W.; Miller, J. G.; Rachlin, E. M.; Burrows, C. J. Org. Lett. 2000,
2, 613; Luo, W.; Miller, J.G.; Rachlin, E.M.; Burrows, C.J. Chem. Res. Toxicol. 2001, 14, 927), who observed a spiroiminodihyd-
antoin derivative as the major product at neutral pH. Clearly, the functional groups attached to the ribose moiety of 8-oxodG influ-
ence the oxidation chemistry of the nucleobase derivative. To explore this chemistry in vivo, 14C-labeled 8-oxodG was synthesized
and incubated with growing MCF-7 human breast cancer cells, resulting in the incorporation of the compound into cellular DNA as
measured by a novel accelerator mass spectrometry assay.
ꢀ 2005 Elsevier Ltd. All rights reserved.
DNA oxidation is a field of interest due to the deleteri-
ous effects that oxidative damage promotes within cells.1
Among the four bases, guanine has the lowest oxidation
potential and constitutes a primary target of oxidants.2
Reactive oxygen species (ROS) are known to damage
genomic and mitochondrial DNA as well as the cellular
nucleotide pool, and are therefore implicated in cancer,
aging, and neurological disorders.1 Among the numer-
ous oxidation products of guanine, 7,8-dihydro-8-oxo-
20-deoxyguanosine (8-oxodG) is the major product,
which is generated by either hydration of the guanine
radical cation formed by direct base oxidation or by hy-
droxy radical addition at the C-8 position of the imidaz-
ole ring of guanine.1a At present, 8-oxodG is recognized
as a highly susceptible site for further oxidation com-
pared to the parent guanine, and may serve as a genomic
Ôhot spotÕ for generation of further oxidative damage
either by direct oxidation or indirectly via hole transfer
from other initially generated radical base cation sites by
one-electron oxidants, singlet oxygen, photochemical
processes, and possibly peroxynitrite, as shown in Fig-
ure 1.3 The redox potential for 8-oxodG is in the range
of 0.58–0.75 V vs a nickel hydrogen electrode (NHE),
compared to 1.29 V vs NHE for 20-deoxyguanosine
(dG).4 It is now fully established that 8-oxodG is a much
better substrate than dG for reaction with a wide set of
oxidants.1a,4 Such a low redox potential makes 8-oxodG
a preferential target for numerous oxidizing agents.
Among the one-electron oxidants, IrIV complexes are
usually exploited for the selective oxidation of 8-oxodG
because the redox potential for IrCl6
2ꢀ
is 0.90 V vs
NHE, just above that of 8-oxodG, but below the redox
potentials of the four normal nucleobases, thus provid-
ing a reasonable mechanism5 by which several 8-oxodG
oxidation products are generated. These nucleoside
derivatives include guanidinohydantoin5a (dGh), spiro-
iminodihydantoin3a (dSp), imidazolone,3f oxazolone,3i
oxaluric acid,6a and urea.6b Recent work on these lesions
in Escherichia coli has demonstrated that they are muta-
genic in vivo.7 For example, both dGh and dSp promote
misincorporation of adenine and guanine opposite the
oxidized bases, respectively, resulting in higher percent-
ages of point mutations compared with 8-oxodG.
*
Corresponding author. Tel.: +1 925 423 2822; fax: +1 925 422
2282; e-mail: henderson48@llnl.gov
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.05.113

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
3628
S. S. Hah et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3627–3631
Figure 1. One-electron oxidation reactions of dG and 8-oxodG.
Importantly, dGh is well bypassed by DNA polymer-
ase(s) and is nearly 100% mutagenic.
HCl. We applied the same oxidation conditions previ-
ously described for OG studies,5a,10 exploiting IrIV as a
one-electron oxidant. Briefly, in a final volume of
300 ll of 75 mM potassium phosphate buffer (pH 4.6,
6.0, 7.2, and 8.4), 7.5 mM 8-oxodG was incubated with
7.5 mM Na2IrCl6 at 22 ꢁC for 1 h to produce dGh as the
major product. The product dSp was generated under
identical conditions, except that reaction required
65 ꢁC in a 10 mM sodium phosphate in 100 mM NaCl
buffer at pH 7.2.11 After filtration, each resulting reac-
tion mixture (10 ll) was directly injected into the HPLC
instrument for analysis or separation.12 The purity of
each isolated product was verified by HPLC for purity
confirmation resulting in a single peak (ꢁ98% purity
for both dGh and dSp, data not shown) and the prod-
ucts were further characterized by mass spectrometry.13
Figure 2a shows an HPLC chromatogram of the 8-
oxodG starting material with a mass spectrogram on
the inset. Figure 2b shows the chromatogram run under
the same conditions after the oxidation reaction was
performed at pH 7.2—the most relevant pH value com-
pared to in vivo conditions. The first two eluting peaks
were separated and characterized by mass spectrometry,
and assigned as dGh and dSp, respectively. Approxi-
mately 99% of the 8-oxodG starting material was con-
sumed during the reaction and the first two peaks
account for the majority of the total peak area for the
Major effort has been recently devoted to the delineation
of mechanistic features of various oxidation reactions
involving 8-oxodG and its analogs either as nucleoside
derivatives or in oligonucleotides containing 8-oxodG
lesion site-specifically.6b,7b For example, studies on oxi-
dation of 20,30,50-triacetoxy-8-oxo-7,8-dihydroguano-
sine (OG) with IrIV yielded two main products as the
hydantoin derivatives, dGh and dSp. Importantly, the
product distributions were highly pH-dependent, sug-
gesting that deprotonation competes with direct rear-
rangement of the 5-hydroxyguanosine intermediate.8
To better understand 8-oxodG reactivity in the nucleo-
tide pool at physiologically relevant concentrations, we
prepared 8-oxodG and 14C-labeled 8-oxodG as shown
in Scheme 1.9 The 14C label allows monitoring of the
incorporation of 8-oxodG and its derivatives into cellu-
lar DNA (vide supra). Briefly, the starting materials of
dG (Chemgenes Corporation) and a mixture of 80%
dG and 20% 14C-labeled dG (specific activity 55 mCi/
mmol, Moravek Biochemicals), respectively, were bro-
minated with saturated bromine solution in water and
the attached bromide was removed by the nucleophilic
attack of benzylate, followed by oxidation with dilute
Scheme 1. Synthetic scheme for 8-oxodG and 14C-labeled 8-oxodG (14C position is asterisked).

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
S. S. Hah et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3627–3631
3629
Figure 3. Relative ratios based on the peak areas corresponding to the
two products of dGh (black bars) and dSp (white bars) formed at (a)
22 ꢁC and pH 4.6, 6.0, 7.2, and 8.4, and at (b) 65 ꢁC and pH 7.2.
Experiments were performed in duplicate and the data shown were
recorded at 220 nm.
major product as dSp. The measured peak area values
gave a conservative estimate of the percent dSp in each
reaction mixture, since the published molar absorptivity
is ꢁ30% greater for dSp than for dGh based on the
known extinction coefficients for these derivatives
formed from OG oxidation at 220–230 nm.5c Percent
peak areas measured at 260 nm, although several fold
lower, resulted in a close correlation with those mea-
sured at 220 nm for dGh at 97.3 0.7, 97.3 0.7,
87.6 1.0, and 86.4 0.5%, respectively. The similar
percentages from measurement at both 220 and
260 nm indicate a minimal contribution from the solvent
front to the Gh peak areas. These data are consistent
with a recent study using the nucleobase 8-oxoG, which
also showed a similar product distribution in the same
pH range.8 However, this report shows that the structur-
ally similar uric acid undergoes deprotonation at N9,
which disfavors spirocyclization. Therefore, this com-
parison must be made tenuously considering that N9
of 8-oxodG is the site of the glycosidic bond to deoxyri-
bose. Clearly, dGh was the major product at physiolog-
ical pH, which was formed with a minor, but significant,
yield of dSp.
Figure 2. HPLC chromatograms showing: (a) 8-oxodG and (b) the 8-
oxodG oxidation products dGh and dSp, from the reaction performed
at 22 ꢁC and pH 7.2. The insets are mass spectra for the purified peaks
assigned 8-oxodG, dGh and dSp, respectively.
entire chromatogram. Although dGh and dSp exist as
diastereomers, we observed only one peak for each
nucleoside derivative, which was likely due to poor res-
olution of the highly polar species that elute from the
HPLC column quickly under reverse-phase conditions
compared to protected nucleosides such as OG and oli-
gonucleotides containing the lesions, in which cases the
diastereomers are resolvable.7b Although dGh eluted at
or near the retention time expected for the solvent front,
the small injection volume of 10 ll produced an essen-
tially negligible salt peak, which was not discernable
from the background on the 8-oxodG control chro-
matogram as shown in Figure 2a. Contamination of
the dGh peak with dSp, or vica versa, was not observed
in the mass spectrometry data. Other minor products
formed were not characterized. Figure 3 shows the per-
cent peak areas of dGh and dSp derived from the one-
electron oxidation reaction of 8-oxodG, and that dGh
is the major product, regardless of pH. The peak areas
of dG and dSp from each chromatogram were added
up and normalized to a value of 100%. The proportion
of dSp and dGh contained in the peaks was plotted as
a percentage of the total peak area of the two products.
The percent peak area from dGh contained in the two
peaks at pH 4.6, 6.0, 7.2, and 8.4 was measured at
220 nm to be 92.8 1.4, 91.3 2.6, 80.4 1.6, and
76.6 2.6% (Fig. 3). The reaction at pH 7.2 and 65 ꢁC
afforded 8.0 0.8% dGh, with the remainder of the
Importantly, though the previously reported one-elec-
tron oxidation reactions of OG were highly pH-depen-
dent and dSp derivative was the major product,5a it
was also reported that the dGh derivative was the major
product in the one-electron oxidation of 8-oxodG in sin-
gle-stranded and duplex DNA under the same condi-
tions employed in this study.3g These differences might
originate from the use of ribose instead of deoxyribose,
or from the use of protecting groups. The free hydroxy
groups in our study may play a role either in the hydro-
lysis step at C-5 or C-4 site by a general base mecha-
nism8 for the production of dGh from 8-oxodG. We
were also able to utilize pH and temperature to synthe-
size 14C-labeled dGh and dSp from 14C-labeled 8-oxodG
for use in experiments with cultured human cells.13

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
3630
S. S. Hah et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3627–3631
Most of the previous model studies on the oxidation of
8-oxodG were performed using protected 8-oxodG
derivatives, such as OG, to prevent oxidation of 50-OH
and to allow longer retention times and better resolution
by HPLC owing to higher hydrophobicity compared to
the unprotected 20-deoxynucleosides. Because of the
possibility of oxidation of 50-OH by IrIV complexes in
this study, the silver mirror experiment for detection
of aldehydes was used, in which the presence of the func-
We are now investigating enzymatic digestion of the
radiolabeled DNA to 20-deoxynucleosides to determine
to what extent the 8-oxodG is oxidized in vivo, either
in the nucleotide pool or in the newly synthesized
DNA. These data indirectly support the formation of
the triphosphate of 8-oxodG or its oxidation products
via the nucleotide salvage pathway which serves as a
substrate for DNA polymerase(s) to enable incorpora-
tion of the modified nucleoside derivative into DNA.
+
tional group reduces Ag(NH3)2 to metallic silver. No
aldehyde was detected in the purified products, dGh
and dSp. In contrast, the oxidation of 8-oxodG with
peroxynitrite, a potent oxidant that is probably formed
in vivo, led to the production of several aldehydes, likely
resulting from the oxidation of 50-OH of the deoxyri-
bose ring (data not shown). Clearly, exploitation of
the proper oxidant for 8-oxodG oxidation can lead to
selective oxidation of the purine ring instead of the
OH groups of the ribose ring. Any aldehyde product(s)
may be of little biological consequence since their 50-OH
group(s) would no longer be available for phosphoryla-
tion by kinases that participate in the nucleoside salvage
pathway.
It is tempting to speculate that dG can be first oxidized
to 8-oxodG via a variety of mechanisms, and can then
be oxidized to dGh and dSp via a one-electron oxidation
mechanism. Alternatively, these nucleoside derivatives
may be formed directly from dG. The nucleosides dGh
and dSp, and their corresponding mono-, di-, and tri-
phosphates, if formed in vivo, are likely to be available
in the nucleotide pool for subsequent promutagenic
incorporation into DNA. The proportion of the dSp
and dGh that is ultimately present in DNA may be gov-
erned by the chemistry of their formation in a substitu-
ent-dependent fashion.
To explore the in vivo oxidation of 8-oxodG, MCF-7
human breast cancer cells were grown in the presence
Acknowledgments
of ꢁ10 pmol of 14C-labeled 8-oxodG for 0, 1, 2, and 4
We acknowledge Kurt Hack for preparation of AMS
samples and John Vogel, Karen Dingley, and Kenneth
Turteltaub for helpful advice and interpretation of the
data. This work was performed at the Research Re-
source for Biomedical Accelerator Mass Spectrometry,
operated at University of California Lawrence Liver-
more National Laboratory under the auspices of the
U.S. DOE Contract W-7405-ENG-48 and partially sup-
ported by NIH/NCRR, Biomedical Technology Pro-
gram Grant P41 RR13461, NIH Grant CA55861, and
by the California Breast Cancer Research Program
Grant 9KB-0179.
14
days, respectively.
Isolated DNA was assayed for
the presence of the radiolabel using accelerator mass
spectrometry, a method for the quantitation of small
amounts of radiocarbon (zeptomol sensitivity) with high
precision.15 Figure 4 shows that radiolabeled nucleoside
incorporation into DNA reached a maximum of
ꢁ10,000 amol/100 lg of DNA after 2 days and then de-
creased ꢁ2-fold by day 4, presumably due to DNA re-
pair competing with radiolabeled 20-deoxynucleoside
incorporation. Clearly, this experiment demonstrated
that modified 20-deoxynucleoside(s) can be incorporated
into cellular DNA at biologically relevant levels. This
experiment, to our knowledge, represents the first report
of direct measurement of 8-oxodG incorporation into
DNA in vivo. The 14C-labeled 8-oxodG may have, in
part, been converted to additional products in vivo.
References and notes
1. (a) Burrows, C. J.; Muller, J. G. Chem. Rev. 1998, 98,
1109; (b) Burrows, C. J.; Muller, J. G.; Kornyushyna, O.;
Luo, W.; Duarte, V.; Leipold, M. D.; David, S. S. Environ.
Health Perspect. 2002, 110, 713; (c) Sugden, K. D.;
Martin, B. D. Environ. Health Perspect. 2002, 110, 725;
(d) Cadet, J.; Douki, T.; Gasparutto, D.; Ravanat, J. L.
Mutat. Res. 2003, 531, 5.
2. Steenken, S.; Jovanovic, S. V. J. Am. Chem. Soc. 1997,
119, 617.
3. (a) Duarte, V.; Muller, J. G.; Burrows, C. J. Nucleic Acids
Res. 1999, 27, 496; (b) Niles, J. C.; Burney, S.; Singh, S. P.;
Wishnok, J. S.; Tannenbaum, S. R. Proc. Natl. Acad. Sci.
USA 1999, 96, 11729; (c) Henderson, P. T.; Jones, D.;
Hampikian, G.; Kan, Y. Z.; Schuster, G. B. Proc. Natl.
Acad. Sci. USA 1999, 96, 8353; (d) Kasai, H.; Yamaizumi,
Z.; Berger, M.; Cadet, J. J. Am. Chem. Soc. 1992, 114,
9692; (e) Ravanat, J. L.; Cadet, J. Chem. Res. Toxicol.
1995, 8, 379; (f) Raoul, S.; Berger, M.; Buchko, G. W.;
Prakash, C.; Joshi, P. C.; Morin, B.; Weinfeld, M.; Cadet,
J. J. Chem. Soc., Perkin Trans. 2 1996, 371; (g) Gasp-
arutto, D.; Ravanat, J. L.; Gerot, O.; Cadet, J. J. Am.
Chem. Soc. 1998, 120, 10283; (h) Muller, J. G.; Duarte, V.;
Hickerson, R. P.; Burrows, C. J. Nucleic Acids Res. 1998,
Figure 4. Incorporation of 14C-labeled 8-oxodG into the DNA of
growing MCF-7 human breast cancer cells. The ratio of 14C to total
carbon contained in the purified genomic DNA was measured with
accelerator mass spectrometry. Experiments were performed in
triplicate.