Investigations of electron-transfer reactions and the redox mechanism of 2′-deoxyguanosine-5′-monophosphate using electrochemical techniques

Article abstract of DOI:10.1039/b415452p

Electron-transfer reactions of 2′-deoxyguanosine-5′- monophosphate (dGMP) have been investigated in phosphate buffers of different pH at a pyrolytic graphite electrode (PGE). In cyclic voltammetry, two well-defined oxidation peaks, 1_a (pH 1.9-1

Full text of DOI:10.1039/b415452p

View Online / Journal Homepage / Table of Contents for this issue
P A P E R
Investigations of electron-transfer reactions and the redox
mechanism of 20-deoxyguanosine-50-monophosphate using
electrochemical techniques
Rajendra N. Goyal,* Sham M. Sondhi and Anand M. Lahoti
Department of Chemistry, Indian Institute of Technology Roorkee, Roorkee, 247 667, India.
E-mail: rngcyfcy@iitr.erent.in
Received (in Montpellier, France) 5th October 2004, Accepted 23rd December 2004
First published as an Advance Article on the web 2nd March 2005
Electron-transfer reactions of 20-deoxyguanosine-50-monophosphate (dGMP) have been investigated in
phosphate buffers of different pH at a pyrolytic graphite electrode (PGE). In cyclic voltammetry, two
well-defined oxidation peaks, Ia (pH 1.9–10.6) and IIa (pH Z 5.8), were noticed. The peak potentials shifted
towards less positive potential when pH was increased. Concentration and sweep rate studies established
the adsorption of dGMP on the electrode surface. UV-vis spectral analysis at pH 2.9 and 7.1 indicated
formation of a UV-absorbing intermediate, which decayed in a pseudo first-order reaction (k = B5.05 ꢀ
10ꢁ4 ꢁ1). Coulometric, voltammetric and UV studies revealed the 4H+,4eꢁ oxidation of dGMP by an
s
EC (electrode reactions followed by chemical reactions) mechanism. The characterization of the oxidation
products was achieved by converting them to their trimethylsilyl derivatives. At pH 7.1, 5-hydroxyhydantoin-
5-carboxamide (9) and a N–O–C8 linked trimer (18) and, at pH 2.9, monohydrated alloxan (12),
deoxyriboside urea (11), two C8O–OC8 and one C8–C8 bridged dimers of dGMP (13, 14 and 15, respectively)
were formed as the major oxidation products. A tentative redox mechanism has been suggested for the
electrooxidation of dGMP.
Compton et al.22 have studied square-wave voltammograms
Introduction
of dGMP at a boron-doped diamond electrode, no attempt
has been made to study the mechanism of the electron-
transfer reactions and to identify the oxidation products of
dGMP.
Purine
nucleotide
20-deoxyguanosine-50-monophosphate
(dGMP) is one of the four essential deoxyribosyl nucleotides
present in deoxyribonucleic acid (DNA) and is apparently a
key compound in biological metabolism.1 dGMP has been
synthesized by oxidative phosphorylation of deoxyguanosine,
catalyzed by deoxyguanosine kinase,2–4 and is a precursor of
the synthesis of 20-deoxynucleoside diphosphate and tripho-
sphate.5,6 Apart from being a vital dietary nucleotide supple-
ment,7 dGMP has been widely used as a DNA model system in
various biological studies. Studies of mutational changes in the
DNA sequence caused by carcinogens present in cigarette
smoke revealed that dGMP prevents double-stranded breaks,
anaphase bridge formation and genomic imbalances.8 Electron
transfer from dGMP to the oxidized radical of riboflavin,
studied using laser flash photolysis,9–11 provided an imperative
pathway to DNA damage both in vivo and in vitro. Fast
electron repair of oxidizing OH adducts of dGMP by various
acid derivatives, studied using pulse radiolysis,12–14 is well-
documented in the literature. Synthetic and spectroscopic
studies of dGMP complexes with Pt(II), Mg(II)15 and with
other metal ions16 have been topics of interest for chemists
during the past decade.
Electrochemical oxidation studies in combination with other
instrumental methods, such as UV spectroscopy and mass
spectrometry, provide invaluable information about biologi-
cally relevant redox reactions of important biomolecules.23,24
Over the past few years, our laboratory has been actively
involved in the investigation of redox properties of purine
bases25,26 and their nucleosides27 and nucleotides,28 using solid
electrodes with the ultimate aim to elucidate the electron-
transfer behavior of nucleic acids, DNA and RNA. In con-
tinuation of our efforts, it was considered worthwhile to study
the electron-transfer reactions of dGMP (1) over a large
pH range, using different electrochemical and analytical
techniques.
Recent reports17 from the American Health Foundation
revealed that cyclic deoxyguanosine adducts formed by reac-
tion of polyunsaturated fatty acids on dGMP under oxidative
conditions are responsible for DNA lesions. Thus, electron-
transfer reactions of dGMP are of significant importance for
understanding the mechanism of oxidative damage in
DNA.18,19 Torun and Morrison20 supported formation of
singlet oxygen during photooxidation studies on dGMP. Still,
a literature survey shows that very limited information
is available on the electrooxidation21 of dGMP. Although
T h i s j o u r n a l i s & T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 5
N e w J . C h e m . , 2 0 0 5 , 2 9 , 5 8 7 – 5 9 5
587

View Online
the absorbance with time was recorded at pre-selected wave-
lengths. The values of the rate constant (k) were calculated
from the linear log(AꢁAN) versus time plots.
Experimental
Materials and general procedures
For identification of the electrooxidation products, 10–15
mg of dGMP were bulk-electrolyzed in an H-cell. The three-
electrode system used was essentially the same as that for the
CPE experiment. The reference arm of the H-cell was filled
with the buffer solution (pH 2.9 or 7.1), whereas the other arm
was filled with a solution of dGMP at the same pH. Two arms
of the H-cell were connected through a sintered disc filled with
a KCl/Agar–Agar mixture. The electrolysis of the solution was
carried out at a potential 100 mV more positive than that of
peak Ia (see Fig. 1) for pH 2.9, or that of peak IIa for pH 7.1.
Nitrogen gas was continuously bubbled through the solution
under magnetic stirring during the course of electrolysis. The
progress of electrolysis was monitored by recording cyclic
voltammograms at different time intervals. When the oxidation
peaks Ia or IIa disappeared, the exhaustively electrolyzed
solution of dGMP was removed from the H-cell and suffi-
ciently acidified (pH B2) by adding a few drops of diluted HCl
in order to convert the sodium salt into the corresponding free
acid. The resulting solution was filtered and lyophilized. The
freeze-dried material obtained was extracted with methanol
(AR grade, 2 ꢀ 10 ml) and the methanolic extract was
The disodium salt of 20-deoxyguanosine-50-monophosphate
(dGMP ꢂ Na2) was purchased from Sigma (USA) and was used
as received. Silylation grade acetonitrile and N,O-bis
(trimethylsilyl)trifluoroacetamide (BSTFA) were from Pierce
Chemical Co. (USA). Phosphate buffers of different pH and
ionic strength were prepared29 from analytical grade chemicals
(NaH2PO4 and Na2HPO4 from Merck). The pH of the buffer
solutions was measured using a Century India Ltd. digital pH-
meter (model CP-901) after due standardization.
The voltammetric studies were carried out by using a
pyrolytic graphite electrode (PGE; surface area: B2.1 mm2)
from Pfizer and was fabricated by using a method reported in
the literature.30 The surface of the PGE was renewed after each
voltammogram recording, by rubbing it with emery paper
(C/P-400), cleaning the surface with water and gently drying
it with tissue paper. As the electrode’s surface area changed
each time due to the cleaning process, the peak current values
showed a variation of ꢃ5%. Hence, voltammetric measure-
ments were performed in triplicate and an average of the
readings was taken as the final value. The working, counter
and reference electrodes were pyrolytic graphite, platinum and
Ag/AgCl electrodes, respectively. All potentials were referred
to Ag/AgCl at ambient temperature, 25 ꢃ 2 1C.
Cyclic sweep voltammetric studies were carried out by using
a Cypress (model CS-1090) computer-controlled electroanaly-
tical system. Coulometric experiments were performed by
using a BAS CV-50W (model MF-9093) potentiostat. Con-
trolled potential electrolysis (CPE) was carried out in a three-
compartment cell using a pyrolytic graphite plate (6 ꢀ 1.5 cm2)
as working electrode, cylindrical platinum gauze (1.5 ꢀ 15 cm2)
as counter electrode and Ag/AgCl as reference electrode.
Spectral changes and kinetic studies associated with the
electron-transfer reactions of UV-absorbing intermediates
were monitored in a 1 cm quartz cell by using a Perkin–Elmer
Lambda 35 UV-vis spectrophotometer. Gas chromatography/
mass spectrometry analysis (GC-MS) of the silylated samples
was performed with a Perkin–Elmer model Clarus-500 mass
spectrometer. High performance liquid chromatography ana-
lysis was performed by using a Shimadzu modular HPLC. The
IR spectrum of compound 9 was recorded in Nujol by using a
Perkin–Elmer 1600 series FTIR spectrophotometer. The 1H
NMR spectrum of compound 9 was recorded in DMSO-d6
with TMS as internal standard, by using a Brucker AC 300 F
instrument.
A stock solution of dGMP (2.0 mM) was prepared using
doubly distilled water. For voltammetric experiments 2.0 ml of
the stock solution was mixed with 2.0 ml of phosphate buffer
(1.0 M) of appropriate pH, so that the effective ionic strength
of the solution would be 0.5 M. The solutions were deoxyge-
nated by bubbling nitrogen gas for 8–10 min before recording
the voltammograms. In the coulometric experiment a potential
100 mV more positive than the oxidation peak potential was
applied to a 0.5 mM solution of dGMP. The number of
electrons (n) involved in the oxidation reaction was determined
by monitoring the exponential decay of the current versus time
curve as reported by Lingane.31 In spectral studies the progress
of the electrolysis of a dGMP solution (0.1 mM) was mon-
itored by withdrawing 2–3 ml of the sample from the working
compartment of the H-cell at different time intervals and
transferring them to a 1 cm quartz cell to record a UV
spectrum in the 200–320 nm region. In another set of experi-
ments, when absorbance at the lmax was reduced to B50%, the
applied potential was turned off and the spectral changes of the
solution were monitored at different time intervals. This helped
to determine the wavelength region in which the UV-absorbing
intermediates were generated during electrolysis. The decay of
Fig. 1 Cyclic voltammograms of 1.0 mM dGMP at PGE: sweep rate
100 mV sꢁ1 and phosphate buffers at pH 2.9 (A), 7.1 (B) and 7.1 with
limited positive sweep potential (C).
588
N e w J . C h e m . , 2 0 0 5 , 2 9 , 5 8 7 – 5 9 5

View Online
concentrated and dried under reduced pressure. The product
mixtures were converted to their thermally stable and volatile
trimethylsilyl derivatives to allow GC-MS analysis.
Silylation was carried out in a 3.0 ml Reacti-vial (Pierce)
glassware. Ca. 100 mg of the sample, acetonitrile (100 ml) and
BSTFA (100 ml) were mixed in the vial, which was then
carefully closed by using a Teflon-coated stopper and heated
in an oil bath to 110 1C for 10–15 min with constant shaking.
The vial was then allowed to cool down to room temperature
and 5 ml of the sample was injected into the GC-MS apparatus;
the spectrum was recorded in EI mode at 70 eV. HPLC was
performed by using a reverse phase 250 ꢀ 4.6 mm 5m
Hypersins BDS C18 column, an SPD-10A UV-vis detector,
and an LC 10 AD pump with methanol–water (1.5 : 8.5) as the
mobile phase.
Fig. 2 Observed variations of peak potentials with pH for 1.0 mM
dGMP at PGE with a sweep rate of 100 mV sꢁ1
Results and discussion
.
Voltammetry
plot of ipvꢁ1/2 versus logv, which can be represented by two
nonlinear curves. Such a behavior suggests a change in me-
chanism at around log v = 1.7.
A typical cyclic voltammogram of a 1.0 mM dGMP solution at
pH 2.9 is depicted in Fig. 1(A). Below pH 5.8, at a sweep rate of
100 mV sꢁ1, dGMP exhibited a first well-defined oxidation
peak Ia in the positive sweep. In the reverse sweep, an ill-
defined cathodic peak IIIc was noticed. In the next cycle, a new
anodic peak IVa was observed in the positive sweep, at a
potential lower than Ia. To check whether peak IIIc corre-
sponds to the reduction of species generated by the oxidation
occurring at peak Ia or is due to an independent reduction of
dGMP, cyclic voltammograms were also recorded by initiating
the sweep in the negative direction. It was observed that, in that
case, no reduction peak IIIc was obtained. Hence, it can be
concluded that IIIc is due to the reduction of a product
resulting from the oxidation of dGMP at the potential of
peak Ia.
At pH Z 5.8, the cyclic voltammogram of dGMP exhibited
two anodic peaks, Ia and IIa, in the initial positive sweep. In the
reverse sweep, reduction peak IIIc was also noticed. In the
following positive sweep, two ill-defined anodic peaks IVa and
Va were recorded [see Fig. 1(B) for a typical example at pH
7.1]. Thus, peaks IIa and Va are only observed at pH Z 5.8,
whereas peaks Ia, IIIc and IVa are observed in the entire pH
range. When the cyclic voltammogram was recorded at pH 7.1
in such a way that the direction of the sweep was reversed after
the first oxidation (peak Ia) but before the second oxidation
[peak IIa not recorded, Fig. 1(C)], peak Va was no longer
observed in the next cycle. This suggests that the species
oxidized at the potential corresponding to peak Va are related
to those formed during the oxidation reactions occurring at the
highest potential (peak IIa) in the first cycle. In other words, a
relation between peak IIa and peak Va is revealed.
The effect of sweep rate on peak potentials was also studied
at pH 7.1 in the sweep range 10–1000 mV sꢁ1. At sweep rates
4700 mV sꢁ1, the peaks became very broad and merged with
the background. The linearity of the plots of the potential of
peaks Ia and IIa versus logv changed at around log v = 2.3, as
shown in Fig. 5. Thus, at sweep rates B200 mV sꢁ1, the
mechanism of electrooxidation of dGMP appeared to change
from diffusion-controlled to a more complicated one due to
adsorption.
In coulometric experiments, the number n of electrons
involved in the electrooxidation of dGMP at different pH
was calculated from the graphical integration of the current
versus time curve.31 The plots of peak current versus time were
exponential for the first 15 min of electrolysis and thereafter
deviation from a straight line was noticed. This deviation
clearly indicates that during the first 15 min, the reactions at
the electrode followed a simple pathway and, thereafter, sub-
sequent chemical reactions played an important role. Similar
observations have been reported by Cauquis and Parker34 on
charge-transfer reactions coupled with follow-up chemical
reactions, that is, EC (electrode reactions followed by chemical
reactions) mechanisms. The average experimental value of n at
different pH was 3.9 ꢃ 0.25.
Spectral study
UV spectra were recorded at different time intervals during the
electrochemical oxidation of 0.1 mM dGMP at pH 2.9 in order
to detect the formation of UV-absorbing intermediates (Fig.
6). The spectrum of dGMP before oxidation exhibits two well-
defined absorption bands at 200 and 255 nm (lmax) and a
shoulder at 280 nm (curve a in Fig. 6). Upon application of a
potential Ep of 1.31 V, that is, 100 mV more positive than
Ep(Ia), absorbance in the ranges 195–210 and 235–295 nm
In all cases, the potentials of the peaks were found to be
dependent on pH and shifted to less positive values with
increasing pH. Peak potential values (Ep) at different pH for
peaks I–V are plotted in Fig. 2. The plots outline a linear
relation between Ep and pH, which is transcribed in Table 1 for
each peak.
The effect of concentration of dGMP, in the range 0.2–3.0
mM, on the peak current (ip) of the well-defined peaks Ia and
IIa was studied. The background current of the phosphate
buffer was subtracted from the observed values of ip. The plots
of peak current versus concentration for peaks Ia and IIa are
shown in Fig. 3. It is clear from these plots that ip increases
linearly with increasing concentration until B2.0 mM, at
which point the slope flattens noticeably. This behavior is
generally indicative of adsorption complications32,33 at the
PGE electrode surface during reactions involving dGMP and
was further confirmed by increasing the sweep rate v. This
resulted in a linear increase of ip, as well as an increase in the
peak current function ipvꢁ1/2 as logv increased. Fig. 4 shows the
Table 1 Peak potential Ep and pH relations calculated from Fig. 2 for
peaks I–V (from CV of 1.0 mM dGMP)
Ep and pH relation
Peak
pH range
Ep (vs. Ag/AgCl)/mV
Correlation coefficient
Ia
1.9–10.6
5.8–10.6
1.9–10.6
1.9–10.6
5.8–07.9
1363.7 ꢁ 53.69 ꢀ pH
1395.3 ꢁ 29.2 ꢀ pH
ꢁ160.8 ꢁ 40.6 ꢀ pH
770.9 ꢁ 53.6 ꢀ pH
816.4 ꢁ 44.7 ꢀ pH
0.976
0.997
0.972
0.974
0.995
IIa
IIIc
IVa
Va
N e w J . C h e m . , 2 0 0 5 , 2 9 , 5 8 7 – 5 9 5
589

View Online
Fig. 5 Variation of peak potentials Ia and IIa with logv for 1.0 mM
dGMP at pH 7.1.
Fig. 3 Dependence of the current intensity (ip) of peaks Ia and IIa on
the concentration of 1.0 mM dGMP at pH 7.1.
subsequent chemical reactions, which corresponds to an EC
mechanism. The kinetics of decay of the UV-absorbing inter-
mediate were monitored by recording the absorbance at se-
lected wavelengths (230, 255 or 280 nm) as a function of time.
The absorbance versus time profiles at different wavelengths
showed an exponential decay (see for example Fig. 7, at 255
nm) and the plots of log(A ꢁ AN) versus time (where AN and A
are the absorbance at infinite time and at the reaction time,
respectively) were linear (see inset of Fig. 7). The pseudo-first-
order rate constant values k for the decay of the UV-absorbing
intermediate generated during the electrooxidation of dGMP
at different pH were calculated from the slopes of the linear
best fit equations corresponding to each log(A ꢁ AN) versus
time plot, and are listed in Table 2.
decreased, while absorbance in the ranges 210–235 and 295–
320 nm increased. In curve i, which was recorded after 120 min
of electrolysis, the band at 250 nm has almost completely
disappeared (Fig. 6). Three isosbestic points are clearly noticed
at 210, 238 and 295 nm.
At pH 7.1, a 0.1 mM solution of dGMP exhibited two well-
defined absorption bands at lmax = 205 and 255 nm and a
shoulder at 275 nm. The spectral changes due to the electro-
chemical oxidation of dGMP at pH 7.1 were essentially the
same as those observed at pH 2.9. However, upon application
of a potential Ep = 1.29 V (100 mV more positive than that of
peak IIa), absorbance in the ranges 210–200 nm and 240–295
nm decreased, while absorbance in the ranges 210–235 nm and
295–320 nm increased. After 120 min of electrolysis, it was
found that the peak at 250 nm had completely disappeared and
a new strong absorption band at 225 nm had appeared. No
clear isosbestic point was observed in this case.
In another set of experiments, after electrolyzing the solution
of dGMP for 45 min (when lmax reaches B50%), the potentio-
stat was open-circuited and UV spectra were recorded at
different time intervals. It was observed that absorbance in
the range 220–290 nm continued to decrease, due to the decay
of the intermediate generated during the electrooxidation.
Thus, a UV-absorbing electroactive intermediate is generated
upon oxidation of dGMP and decays due to involvement in
Product characterization
Attempts to separate the different constituents and determine
their molar masses from the exhaustively electrolyzed (at pH
2.9 and 7.1) samples of dGMP were unsuccessful, probably due
to their nonvolatile nature. Hence, the electrooxidation pro-
ducts of dGMP were first converted to their thermally stable,
volatile trimethylsilyl derivatives. This procedure has been
found to be useful in characterizing the electrooxidation
products of various biomolecules.35
Fig. 6 Spectral changes with time during the electrochemical oxida-
tion of 0.1 mM dGMP at pH 2.9 at Ep = 1.31 V [4 Ep(Ia) ¼ 1.21 V];
spectra recorded after (a) 0, (b) 5, (c) 10, (d) 20, (e) 30, (f) 45, (g) 60, (h)
90, (i) 120 min of electrolysis.
Fig. 4 Variation of the peak current function (ipvꢁ1/2) for peak Ia with
the logarithm of the sweep rate for 1.0 mM dGMP at pH 7.1.
590
N e w J . C h e m . , 2 0 0 5 , 2 9 , 5 8 7 – 5 9 5

View Online
Fig. 7 Plots of absorbance at 255 nm versus time and log(A ꢁ AN) vs.
time (inset) for the decay of the UV-absorbing electroactive intermedi-
ate generated during the electrooxidation of 0.1 mM dGMP at pH 2.9.
Fig. 8 Mass spectra recorded for peaks at Rt ¼ 5.98 (A) and 6.42 (B)
min from the GC-MS analysis of the silylated sample at pH 7.1 after
electrooxidation above Ep(IIa).
The GC-MS analysis of the silylated sample at pH 7.1 was
performed and two major peaks were detected at retention
times Rt = 5.98 and 6.42 min (see corresponding MS spectra in
Fig. 8). In acidic medium (pH 2.9), five major peaks were
recorded at Rt ¼ 4.35, 7.35, 7.42, 9.88 and 11.1 min. The
corresponding molar masses and the relative abundances of the
main peaks for each GC peak are presented in Table 3.
At pH 7.1, the products corresponding to Rt = 5.98 and 6.42
min were identified as being the tetrasilylated derivative of 5-
hydroxyhydantoin-5-carboxamide [9, m/z 447 (M+; 30.5%)]
and the hexasilylated derivative of an N–O–C-linked trimer 18
[m/z 898 (M+; 15.5%)], respectively (see Schemes 1–3). 5-
Hydroxyhydantoin-5-carboxamide is one of the most well-
known metabolites of purines and is reported36,37 to undergo
silylation at four sites. The trimer 18 (calcd 897.338 for
C34H59N13O5Si6) seems to be formed by three monomer units
of dGMP, from which the deoxyribose sugar moiety of all
three monomers has been lost during the reaction at the
electrode surface. This observation is based on a comparison
of the observed molar mass and mass calculations for different
silylated trimer derivatives, with and without deoxyribose
units. One of the possible reasons for losing the sugar units
could be the strong steric hindrance implied by the trimeriza-
tion. Formation of trimers during the electrooxidation of
purine nucleosides has already been described in the litera-
ture38 and obtaining a trimer as the oxidation product of
dGMP at PGE is not unusual.
the C O and NH functions does not permit silylation due to
Q
the bulky nature of the Si(CH3)3 group and compound 9a was
indeed identified by GC-MS at Rt 5.98 min with m/z 447 (calcd
447.186 for C16H37N3O4Si4).
At pH 2.9, the peaks at Rt 4.35 and 11.1 min correspond to
the molar masses 392 (M+; 25.0%) and 448 (M+; 20.5%),
respectively. These molar masses are associated with the
trisilylated derivative of urea deoxyriboside (11) and the tetra-
silylated derivative of 5,5-dihydroxypyrimidine-2,4,6-trione
(12), also know as hydrated alloxan (see Scheme 2).26 Silylation
of urea deoxyriboside has been considered to occur only on one
side of the C O group having the deoxyribose unit. A similar
Q
one-side silylated product of urea riboside has also been
reported to form during the electrochemical oxidation of 9-b-
D-ribofuransyluric acid.39 The peaks at Rt 7.35 and 7.42 min
correspond to the molar masses 952 (M+; 16.9%) and 764
(M+; 22.3%), respectively, and the corresponding products
were identified as the hepta- and hexasilylated derivatives of
the C8O–OC8 bridged dimers 13 and 14, respectively (Scheme
2). The mass calculations imply that the silylated derivative of
compound 13 (calcd 952.397 for C36H72N10O7Si7) possesses
only one deoxyribose sugar moiety, whereas that of 14 (calcd
764.31 for C28H56N10O4Si6) has no sugar unit. It is believed
that the deoxyribose sugar moiety of dGMP undergoes hydro-
lysis readily, due to the steric hindrance of the sugar units
during dimerization. It was also noticed that if a deoxyribose
sugar unit is preserved in the dimer, then, upon silylation, both
reactive OH groups from the five-membered deoxyribose ring
are rapidly silylated. In contrast, both in dimers and trimers,
the NH2 group from the guanine base, upon silylation, changes
to NHSi(CH3)3 rather than N[Si(CH3)3]2. Such observations
Scheme 1 describes the synthesis of the tetrasilylated deriva-
tive 9a from 9 using BSTFA as the silylating agent and
acetonitrile under anhydrous conditions. It is clear from
Scheme 1 that compound 9 possesses five silylation sites and
it can be expected that the reaction would take place at each of
the available sites. However, the OH group situated between
Table 2 Pseudo-first-order rate constant values calculated for the
decay of the UV-absorbing intermediate generated during the electro-
oxidation of dGMP.
pH
7.1
l/nm
k/10ꢁ4 sꢁ1a
280
255
230
280
255
230
4.85 ꢃ 0.02
5.02 ꢃ 0.04
4.84 ꢃ 0.02
5.10 ꢃ 0.04
5.33 ꢃ 0.02
5.14 ꢃ 0.02
2.9
Scheme 1 Formation of the tetrasilylated derivative of 5-hydroxyhy-
dantion-5-carboxamide by reaction with BSTFA in acetonitrile at
pH 7.1.
a
Average of at least three replicate determinations.
N e w J . C h e m . , 2 0 0 5 , 2 9 , 5 8 7 – 5 9 5
591

View Online
Scheme 2 Tentative redox mechanism explaining the formation of compounds 11–15, depending on the pH and redox potential.
about the silylated molecules are also well-documented in the
literature.25,39 The fifth peak at Rt 9.88 min with m/z 920 (M+;
15.6%), corresponds to the heptasilylated derivative of dimer
15 (calcd 920.407 for C36H72N10O5Si7), having a C8–C8 linkage
(Scheme 2). Formation of O–O and C–C-linked dimers during
the oxidation of purine nucleosides and nucleotides is also
reported in the literature.26,39,40
Attempts to separate the products for further characteriza-
tion by standard analytical techniques were unsuccessful, due
to the very similar retention times in HPLC. Despite our
efforts, we found that separation of the dimers was impossible
to achieve. However, when the bulk electrolysis of dGMP was
carried out at pH 7.1 below Ep(IIa), compound 9 could be
isolated as a single oxidation product (TLC Rf B 0.31) and
characterized by FTIR and 1H NMR spectroscopy. 9 exhibited
bands in the IR at vmax ¼ 661, 814, 1281, 1338, 1683 (CONH),
1
1720, 1765 (C O), 3220 and 3401 (NH) cmꢁ1. The H NMR
Q
spectrum of compound 9 showed signals at d ¼ 10.71 (br s,
1 H, O C–NH–C O), 7.95 (s, 1H, NH), and 6.62 (s, 2H,
Q
Q
NH2). However, no clear peak was observed for the tertiary
hydroxyl proton. It is assumed that the missing peak is hidden
under the water peak in DMSO-d6. Upon D2O exchange, all
signals disappeared, except for a peak at 4.68 ppm from HDO.
9 presented all the characteristic features of 4-hydroxy-2,5-
dioxoimidazolidine-4-carboxamide, thus confirming the MS
assignment.
592
N e w J . C h e m . , 2 0 0 5 , 2 9 , 5 8 7 – 5 9 5

View Online
Scheme 3 Proposed mechanistic pathway for the formation of the N–O–C bridged trimer of dGMP at PGE.
process, corresponding to peak Ia in CV. The 4H+,4eꢁ process
may occur in a single step or as the combination of two
2H+,2eꢁ processes. The appearance of peak IVa in the second
positive sweep in CV suggests that oxidation occurs in two
2H+,2eꢁ processes. In Scheme 2, peak IVa is assigned to the
oxidation of the 8-hydroxy dGMP, which can be more easily
oxidized than dGMP. As the coulometric experiments suggest
involvement of coupled chemical reactions, it seems reasonable
Redox mechanism
Tentative redox mechanisms assigning the observed anodic
and cathodic signals I–V to their corresponding reactions with
the probable structure of the products formed during the
electrooxidation of dGMP at pH 7.1 and 2.9 are given in
Schemes 2 and 3. Experimental results revealed that the
electrooxidation of dGMP occurs in an overall 4H+,4eꢁ
Table 3 MS data recorded at each retention time where peaks were detected from the silylated electrolyzed samples of dGMP at pH 7.1 and 2.9
pH
7.1
Rt/min
Product
M+ ion peak
Fragmentation pattern
5.98
6.42
4.35
7.35
7.42
9.88
11.1
9
18
11
13
14
15
12
447
898
392
952
764
920
448
287 (11.4%); 259 (100%); 188 (15.1%); 160 (9.2%).
618 (72.1%); 602 (13.4%); 604 (12.8%); 588 (8.3%); 309 (10.5%); 295 (25.1%); 293 (5.5%).
376 (10.7%); 348 (5.6%); 261 (100%); 131 (7.5%).
2.9
691 (72.2%); 586 (5.5%); 570 (9.4%); 382 (7.8%); 293 (28.1%); 261 (15%).
598 (10.6%); 382 (100%); 366 (16.8%); 253 (8.2%); 129 (9.5%).
863 (5.0%); 659 (15.5%); 554 (85.4%); 366 (9.1%); 261 (50.5%).
331 (13.5%); 270 (11.4%); 246 (100%); 115 (10.3%).
N e w J . C h e m . , 2 0 0 5 , 2 9 , 5 8 7 – 5 9 5
593

View Online
drated at pH 2.9, a hydrated alloxan (12), namely 5,5-dihy-
to conclude that the product of the primary electrode reaction
is unstable and undergoes follow-up chemical reactions.
droxypyrimidine-2,4,6-trione, is the actual product of the
reaction, which was identified as its tetrasilylated derivative
by GC-MS. The opening of the imidazole ring and the forma-
tion of alloxan in acidic conditions has been reported27,48 for
purines. Formation of the dimers (13, 14 and 15) at pH 2.9 is
possible by different combinations of the free radicals 2 and 4,
as shown in Scheme 2. The free radical 4 generated from 3 can
rapidly dimerize to form the C8O–OC8 bridged dimers 13 and
14, depending on whether one or two deoxyribose units are lost
during the O–O bond formation at the C8 position. Dimer 15
could result from the combination of two molecules of radical
2 through the C–C bond at the C8 position. However, it must
be realized that the proposed mechanistic pathway could be
one out of several mechanisms by which the products can be
formed. The most probable structures assigned for the dimers
13, 14 and 15 are shown here.
The first 1H+,1eꢁ oxidation of dGMP (1) gives rise to a free
radical 2 at the electrode surface, which is rapidly oxidized in
the presence of water by a further 1H+,1eꢁ process, yielding
compound 3 (Scheme 2). The initial oxidation of purines at the
C8 position, generating a free radical, is well-documented in the
literature.41,42 The 8-hydroxy dGMP (3) can undergo oxida-
tion by two different routes. In the first case a 1H+,1eꢁ
oxidation would lead to the formation of radical 4, whereas
in the second case a 2H+,2eꢁ oxidation will give diimine 5. It is
expected that the potentials of both oxidation processes would
be very close and, therefore, would explain that no separate
peaks for such oxidations were observed by CV. The diimine
(5) is expected to be highly unstable as previously reported for
several purines43,44 and, hence, it will be readily attacked by a
water molecule to give the imine alcohol 6. Conversion of 5 to 6
is a simple chemical reaction. The imine alcohol can be reduced
at PGE in a 2H+,2eꢁ process will give compound 7. This
reduction process is assigned to peak IIIc. On the other hand,
compound 6 can be further attacked by another molecule of
water to give a diol 8. The values of k observed for the kinetics
of decay are correlated to the conversion of 6 to 8, which
appears to be a slow process (k = B5.05 ꢀ 10ꢁ4
s
ꢁ1). As the
values of k observed were more or less similar at pH 2.9 and
7.1, it is reasonable to conclude that the rate of decay, that is,
the reaction between 6 and a molecule of water, is independent
of pH. At pH 7.1, the diol 8 undergoes opening of the
pyrimidine ring to give 4-hydroxy-2,5-dioxoimidazolidine-4-
carboxamide (5-hydroxyhydantoin-5-carboxamide, 9) as one
of the major electrooxidation products. It is interesting to note
that compound 9 does not possess a ribosyl group. The ribo-
side moiety of diol 8 seems to be hydrolyzed in aqueous
solutions. At neutral pH, the rupture of the ribosyl bond
during the pyrimidine ring-opening of purines is described in
the literature.45,46
Conclusions
The present study clearly demonstrates that the electron-trans-
fer reactions and redox mechanism of dGMP follow a complex
pathway. By extension, this study provides invaluable informa-
tion on the metabolic route of dGMP in the biosystem. It is
likely that reactions, similar those involved in the electrooxida-
tion of dGMP under oxidative conditions, may occur and yield
tumorogenic products in the human organism. Our previous
results on the electrochemical oxidation of guanosine41 already
reported that a dimer with an O–O linkage very similar to
dimer 14 causes nephritis with oedema in albino mice and is
toxic in nature. Identification of the different oxidation pro-
ducts of dGMP at PGE can provide a very important tool for
the biochemist to understand oxidative damage in DNA.
Although complex biological reactions of dGMP occur at the
enzyme–solution interface, the study of its electrochemical
behavior at the electrode–solution interface close to physiolo-
gical conditions can be a perfect model system for in vitro
studies.
The formation of the second major product, trimer 18, at pH
7.1 is mechanistically represented in Scheme 3 and its forma-
tion can also be explained by an EC mechanism as follows. The
1H+,1eꢁ oxidation of 1 generates free radical 2 at the PGE
surface. Radical 2 reacts with another molecule of 1 diffusing
to the surface, forming the secondary radical 2a, which corre-
sponds to a hydrogen atom abstraction reaction. 2a reacts with
a water molecule to give another secondary free radical species
2b, which rapidly combines with a second molecule of 2a to
lead to the formation of a reactive dimer 16, in which both
deoxyribose sugar units of the purine bases have been lost by
hydrolysis. A 2H+,2eꢁ oxidation of 16, followed by attack by a
water molecule, is expected to produce 17. Dimer 17, upon
further 1H+ 1eꢁ oxidation gives rise to compound 17a having
a free radical at the oxygen atom attached to C8. The free
radical species 17a, upon combination with another molecule
of 2a, yields 18 having a N–O–C8 guanine base linkage. The
molar mass of 898 corresponds to the hexasilylated derivative
of compound 18. The appearance of peak IIa in the cyclic
voltammogram of dGMP at pH 7.1 could be due to the overall
3H+,3eꢁ oxidation of dimer 16 into 18 as shown in Scheme 3.
The biodegradation of dGMP is reported47 to produce
xanthine as one of its metabolites, which was not identified
among the final products in the present study. Still, the radical
2b, proposed here as an intermediate to 18 (Scheme 3), is a
reactive moiety of the xanthine nucleotide. The identification
of dimers and trimer formed during oxidation of dGMP
strongly indicates that the radical–radical coupling reactions
play a significant role in comparison to other reactions which
lead to alloxan and ureariboside as the products.
Acknowledgements
One of the authors (AML) is thankful to the Indian Institute of
Technology Roorkee (IITR) for awarding him an Institute
Fellowship, sponsored by the Ministry of Human Resources
Department (MHRD), Govt. of India. Financial assistance for
this work was also provided by MHRD through IITR grant
code no. 106-20-77.
References
Under acidic conditions (pH 2.9), the diol 8 exhibits a
different behavior, the imidazole ring of the diol rupturing,
with the loss of ammonia and H3PO4, to give alloxan 10 and
urea deoxyriboside 11 (Scheme 2). As alloxan is readily hy-
1
G. Hardin and C. Bajema, Biology: Its Principles and Implications,
W. H. Freeman and Co., San Francisco, 3rd edn., 1977, p. 102.
J. Taanman, J. R. Muddle and A. C. Mauntau, Hum. Mol. Genet.,
2003, 12, 1839.
2
594
N e w J . C h e m . , 2 0 0 5 , 2 9 , 5 8 7 – 5 9 5

View Online
28 R. N. Goyal and A. Sangal, J. Electroanal. Chem., 2003, 557,
147.
3
S. Eriksson, B. Munch-Petersen, K. Johansson and H. Eklund,
Cell Mol. Life Sci., 2002, 59, 1327.
4
5
M. Taktakishvili and V. Nair, Tetrahedron Lett., 2000, 41, 7173.
T. S. Stolworthy, E. Krabbenhoft and M. E. Black, Anal. Bio-
chem., 2003, 322, 40.
S. Pal and V. Nair, Biotechnol. Lett., 1998, 20, 1149.
H. Elisabeth and J. Roland, Nutr. Res. (N.Y.), 2004, 24, 197.
L. Z. Luo, K. M. Werner, S. M. Gollin and W. S. Saunders,
Mutat. Res., 2004, 554, 375.
29 G. D. Christian and W. C. Purdy, J. Electroanal. Chem., 1962, 3,
363.
30 J. L. Owens, H. A. Marsh and G. Dryhurst, J. Electroanal. Chem.,
1978, 91, 231.
31 J. J. Lingane, Electroanalytical Chemistry, Wiley, New York, 2nd
edn, 1966, p. 222.
32 R. S. Nicholson and I. Shain, Anal. Chem., 1964, 36, 706.
33 E. C. Brown and R. F. Large, in Physical Methods of Chemistry,
eds. A. Weissberger and B. W. Rossiter, Wiley, New York, 1974,
vol. 1, p. 423.
6
7
8
9
C. Lu, W. Lin, Z. Han, S. Yao and N. Lin, Phys. Chem. Chem.
Phys., 2000, 2, 329.
10 C. Lu, S. Yao and N. Lin, Chem. Phys. Lett., 2000, 330, 389.
11 V. Shafirovich, J. Cadet, D. Gasparutto, A. Dourandin and N. E.
Geacintov, Chem. Res. Toxicol., 2001, 14, 233.
12 Y. Shi, C. Huang, W. Wang, J. Kang, S. Yao, N. Lin and R.
Zheng, Radiat. Phys. Chem., 2000, 58, 253.
34 G. Cauquis and V. D. Parker, in Organic Electrochemistry, ed.
M. M. Baizer, Marcel Dekker, New York, 1973, p. 134.
35 A. A. Rostami and G. Dryhurst, J. Electroanal. Chem., 1987, 223,
143.
13 Y. Jiang, W. Lin, S. Yao, N. Lin and D. Zhu, Radiat. Phys.
Chem., 1999, 54, 349.
36 R. N. Goyal, Indian J. Chem., Sect. A: Inorg., Bio-inorg., Phys.,
Theor. Anal. Chem., 1989, 28, 467.
14 J. Ma, S. Qian, W. Lin and S. Yao, Sci. China, Ser. B, 1998, 41, 556.
15 H. A. Tajmir-Riahi and T. Theophanides, Inorg. Chim. Acta,
1983, 30, 223.
16 H. A. Tajmir-Riahi, Biopolymers, 1991, 31, 1065.
17 J. Pan and F. Chung, Res. Toxicol., 2002, 15, 367.
18 S. Choi, R. B. Cooley, A. S. Hakemian, Y. C. Larrabee, R. C.
Bunt, S. D. Maupas, J. G. Muller and C. J. Burrows, J. Am. Chem.
Soc., 2004, 126, 591.
37 R. N. Goyal, A. K. Srivastava and V. Bansal, J. Chem. Soc.,
Perkin Trans. 2, 1994, 1709.
38 P. Subramanian and G. Dryhurst, J. Electroanal. Chem., 1987,
224, 137.
39 R. N. Goyal, A. Brajter-Toth, J. S. Besca and G. Dryhurst,
J. Electroanal. Chem., 1983, 144, 163.
40 E. F. Strittmatter, P. D. Schnier, J. S. Klassen and E. R. Williams,
J. Am. Soc. Mass Spectrom., 1999, 10, 1095.
41 R. N. Goyal, N. Jain and D. K. Garg, Bioelectrochem. Bioenerg.,
1997, 43, 105.
19 V. Duarte, J. G. Muller and C. J. Burrows, Nucleic Acids Res.,
1999, 27, 496.
20 L. Torun and H. Morrison, Photochem. Photobiol., 2003, 77, 370.
21 A. M. Oliveira-Brett, J. A. P. Piedade, L. A. Silva and V. C.
Diculescu, Anal. Biochem., 2004, 332, 321.
42 L. P. Candeias and S. Steenken, J. Phys. Chem., 1992, 96, 937.
43 R. N. Goyal and G. Dryhurst, J. Electroanal. Chem., 1982, 135,
75.
22 C. Prado, G. Flechsig, P. Gruendler, J. Foord, S. John, F. Marken
and R. G. Compton, Analyst (Cambridge, U.K.), 2002, 127, 329.
23 G. Dryhurst, Chem. Rev., 1990, 90, 795.
44 G. Dryhurst, K. M. Kadish, F. Scheller and R. Renneberg,
Biological Electrochemistry, Academic Press, New York, 1982,
vol. 1, p. 279.
24 T. Tabatabaie, R. N. Goyal, C. L. Blank and G. Dryhurst, J. Med.
Chem., 1993, 36, 229.
45 T. E. C. Peterson and A. Brajter-Toth, J. Electroanal. Chem.,
1988, 239, 161.
25 R. N. Goyal, A. K. Jain and N. Jain, J. Chem. Soc., Perkin Trans.
2, 1996, 1153.
26 R. N. Goyal, P. Thankachan and N. Jain, Bull. Chem. Soc. Jpn.,
2000, 73, 1515.
46 S. Y. Wang, Photochemistry and Photobiology of Nucleic Acids,
Academic Press, New York, 1976, vol. 1, p. 362.
47 H. Kruszewska, A. Misicka and U. Chmielowiec, Il Farmaco,
2004, 59, 13.
27 R. N. Goyal, A. Rastogi and A. Sangal, New J. Chem., 2001, 25,
545.
48 J. L. Owens, H. H. Thomas and G. Dryhurst, Anal. Chim. Acta,
1978, 96, 89.
N e w J . C h e m . , 2 0 0 5 , 2 9 , 5 8 7 – 5 9 5
595