Monosaccharide-H2O2 reactions as a source of glycolate and their stimulation by hydroxyl radicals

Article abstract of DOI:10.1016/j.carres.2006.06.023

An analysis of the H₂O₂-induced breakdown and transformation of different keto-monosaccharides at physiological concentrations reveals that glycolate and other short-chained carbohydrates and organic acids are produced. Depletion of monosaccharides and glycolate synthesis occurs at increased rates as the length of the carbohydrate chain is decreased, and is significantly increased in the presence of trace amounts of Fe²⁺ ions (10 μM). Rates of monosaccharide depletion (initial concentration of 3 mM) observed were up to 1.55 mmol h^-1 in the case of fructose, and 2.59 mmol h^-1 in the case of dihydroxyacetone, depending upon pH, H₂O₂ concentration, temperature and the presence or absence of catalytic amounts of Fe²⁺. Glycolate was produced by dihydroxyacetone cleavage at rates up to 0.45 mmol h^-1 in the absence, and up to 1.88 mmol h^-1 in the presence of Fe²⁺ ions (pH 8). Besides glycolate, other sugars (ribose, glyceraldehyde, glucose), glucitol (sorbitol) and organic acids (formic and 2-oxogluconic acid) were produced in such H₂O₂-induced reactions with fructose or dihydroxyacetone. EPR measurements demonstrated the participation of the ^{{radical dot}}OH radical, especially at higher pH. Presence of metal ions at higher pH values, resulting in increased glycolate synthesis, was accompanied by enhanced hydroxyl radical generation. Observed changes in intensity of DEPMPO-OH signals recorded from dihydroxyacetone and fructose reactions demonstrate a strong correlation with changes in glycolate yield, suggesting that ^{{radical dot}}OH radical formation enhances glycolate synthesis. The results presented suggest that different mechanisms are responsible for the cleavage or other reactions (isomerisation, auto- or free-radical-mediated oxidation) of keto-monosaccharides depending of experimental conditions.

Full text of DOI:10.1016/j.carres.2006.06.023

Carbohydrate Research 341 (2006) 2360–2369
Monosaccharide–H2O2 reactions as a source of glycolate
and their stimulation by hydroxyl radicals
a,
b
a
*
ˇ
´
ˇ
´
ˇ
´
Vuk Maksimovic, Milos Mojovic and Zeljko Vucinic
aLaboratory for Biophysics, Center for Multidisciplinary Studies, University of Belgrade, Kneza Visˇeslava 1a,
YU-11000 Belgrade, Serbia and Montenegro
bFaculty of Physical Chemistry, University of Belgrade, Studentski trg 3, YU-11000 Belgrade, Serbia and Montenegro
Received 4 April 2006; received in revised form 26 June 2006; accepted 27 June 2006
Available online 25 July 2006
Abstract—An analysis of the H2O2-induced breakdown and transformation of different keto-monosaccharides at physiological con-
centrations reveals that glycolate and other short-chained carbohydrates and organic acids are produced. Depletion of monosaccha-
rides and glycolate synthesis occurs at increased rates as the length of the carbohydrate chain is decreased, and is significantly
increased in the presence of trace amounts of Fe2+ ions (10 lM). Rates of monosaccharide depletion (initial concentration of
3 mM) observed were up to 1.55 mmol hÀ1 in the case of fructose, and 2.59 mmol hÀ1 in the case of dihydroxyacetone, depending
upon pH, H2O2 concentration, temperature and the presence or absence of catalytic amounts of Fe2+. Glycolate was produced by
dihydroxyacetone cleavage at rates up to 0.45 mmol hÀ1 in the absence, and up to 1.88 mmol hÀ1 in the presence of Fe2+ ions
(pH 8). Besides glycolate, other sugars (ribose, glyceraldehyde, glucose), glucitol (sorbitol) and organic acids (formic and 2-oxoglu-
conic acid) were produced in such H2O2-induced reactions with fructose or dihydroxyacetone. EPR measurements demonstrated the
participation of the ꢀOH radical, especially at higher pH. Presence of metal ions at higher pH values, resulting in increased glycolate
synthesis, was accompanied by enhanced hydroxyl radical generation. Observed changes in intensity of DEPMPO-OH signals
recorded from dihydroxyacetone and fructose reactions demonstrate a strong correlation with changes in glycolate yield, suggesting
that ꢀOH radical formation enhances glycolate synthesis. The results presented suggest that different mechanisms are responsible for
the cleavage or other reactions (isomerisation, auto- or free-radical-mediated oxidation) of keto-monosaccharides depending of
experimental conditions.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Keywords: Fenton reaction; Dihydroxyacetone; Glycolate; Hydrogen peroxide; Hydroxyl radical; Monosaccharides
1. Introduction
been shown using high concentrations of hydrogen
peroxide and 24-hour-long incubations.5,6 In our previ-
ous reports, we tested the ability of different mono-
saccharides to produce glycolate in the reaction with
hydrogen peroxide under conditions close to those of
the physiological state and have shown that keto forms
can produce glycolate, as opposed to the corresponding
aldo forms, especially at higher pH values (pH P8).7
Presence of metal ions in trace amounts stimulated this
reaction, indicating the involvement of free radicals from
a Fenton-type reaction. However, those results were ob-
tained using the classical Calkins colorimetric assay and
liquid ion-exchange chromatography,8 a procedure with
a number of drawbacks and lack of specificity.
Basic organic chemistry stipulates that strongly alkaline
conditions make sugars more susceptible to oxidation,
hydrogen peroxide interactions and carbon chain frag-
mentations.1 Simple monosaccharides have been shown
to be capable of glycolate production under nonphysio-
logical conditions, such as extreme peroxide concentra-
tions, pH, temperatures and the presence of complex
catalysts.2–4 Also, synthesis of carboxylic acids (includ-
ing glycolate) from fructose and dihydroxyacetone has
*
Corresponding author. Tel.: +381 11 2078 456; fax: +381 11 3055
289; e-mail: maxivuk@ibiss.bg.ac.yu
0008-6215/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2006.06.023

V. Maksimovic´ et al. / Carbohydrate Research 341 (2006) 2360–2369
2361
The aim of this work was to test the ability of hydro-
gen peroxide to cleave keto-monosaccharides of
different chain length under conditions close to physio-
logical and analyse the products of such a reaction
(especially glycolate). In order to bypass various non-
specific photometric derivatisation methods prone to
formation of artifacts, quantification of products was
carried out by anion-exchange HPLC with electrochem-
ical and UV detection, its sensitivity being such that no
derivatisation is required. Also, we wanted to determine
if and what oxygen-centred free-radical species partici-
pate in such a reaction, in the presence and absence of
trace metals and chelators, using a selective electron spin
resonance trapping technique.
published results, as well as older data obtained using
Calkins colorimetric assay.11,12
2.1. HPLC analysis of reaction
Dihydroxyacetone and fructose were used for a detailed
analysis of the reaction with H2O2 at two pH values (4
and 8) and in the absence or presence of Fe2+, the rela-
tive share of the products after 30 min of incubation at
40 ꢁC being presented in Tables 1 and 2. In the dihy-
droxyacetone–H2O2 reaction, in the case of the absence
of Fe2+ ions, dihydroxyacetone decomposed to a greater
degree at lower pH 4. Addition of Fe2+ ions greatly en-
hanced the cleavage of the monosaccharide, and the pH
effect was reversed, that is, dihydroxyacetone decom-
posed to a greater degree at higher pH 8. The main
product of H2O2-induced decomposition in the presence
of Fe2+ ions, as well as in the absence of iron at low pH,
was glycolate with formate as the main byproduct. Gly-
ceraldehyde became the predominant product in the
absence of metal ions at pH 8.
In the fructose–H2O2 reaction, the main products
were 2-oxogluconic acid, glucitol (sorbitol) and a small
amount of glycolate in the reaction without metal ions.
Addition of 10 lM FeSO4 did not alter the product dis-
tribution, with the exception of ribose, which appeared
as the fragmentation product (Table 2).
The observed differences in glycolate yield between
dihydroxyacetone and fructose could be explained by
different levels of enolisation of those two ketoses.13
The susceptibility of a monosaccharide to enolisation
is inversely proportional to the extent of formation of
2. Results and discussion
Preliminary skim tests demonstrated that keto forms of
simple monosaccharides in reaction with hydrogen per-
oxide produce glycolate, as opposed to aldo isomers.
Those findings are supported by earlier reports in which
a stepwise degradation mechanism of aldoses is pro-
posed, yielding formic acid as final product.9 Tested
ketoses (dihydroxyacetone, erythrulose and fructose)
produced glycolate as a characteristic product, as de-
scribed in previous investigations employing higher
alkaline conditions and concentrations of reactants.10
Glycolate yield increased with decreasing carbon chain
length (0.11 mmol hÀ1 at 0.75 mM H2O2 for dihydroxy-
acetone in comparison to 0.012 mmol hÀ1 from fruc-
tose), these results being in line with our recently
Table 1. Relative share (in %) in overall chromatogram area of dihydroxyacetone/H2O2 reaction products, in the absence or presence of 10 lM
a
FeSO4
Reactant/product
Structure
DHA/H2O2 reaction
DHA/H2O2 + FeSO4
pH 4
pH 8
pH 4
pH 8
HO
Dihydroxyacetone
70.05
18.93
77.39
49.92
23.96
31.06
O
HO
HO
O
Glycolic acid
6.58
39.74
OH
Formic acid
Glyceraldehyde
Unidentified
2.68
2.49
ꢀ6
1.04
10.11
ꢀ5
15.14
1.23
18.81
2.76
HO
O
OH
O
HO
—
ꢀ9.5
ꢀ7.5
a Reaction mixtures contained 3 mM dihydroxyacetone, 0.75 mM H2O2 and 5 mM phosphate buffer at pH 8, the reaction being carried out for
30 min at 40 ꢁC.

2362
V. Maksimovic´ et al. / Carbohydrate Research 341 (2006) 2360–2369
a
Table 2. Relative share (in %) in overall chromatogram area of fructose/H2O2 reaction products in the absence or presence of 10 lM FeSO4
Reactant/product
Fructose
Structure
Fructose/H2O2 reaction
Fructose/H2O2 + FeSO4
pH 4
pH 8
pH 4
pH 8
O
OH HO
87.32
95.52
68.20
73.50
HO
HO
OH
O
Glycolic acid
3.45
6.62
0.27
0.14
7.92
0.85
3.04
HO
OH
O
OH HO
OH HO
OH HO
2-Oxogluconic acid
10.47
HO
O
HO
OH
HO
Glucose
Sorbitol
n.d.
2.65
n.d.
n.d.
O
HO
OH
HO
0.35
n.d.
1.33
6.54
0.13
HO
HO
OH
HO
O
Ribose
n.d.
n.d.
n.d.
n.d.
12.47
n.d.
HO
HO
OH
OH
Formic acid
n.d.
O
OH
Glyceraldehyde
Unidentified
n.d.
n.d.
n.d.
—
n.d.
O
HO
—
ꢀ2.5
ꢀ1.5
ꢀ16
a The mixtures contained 3 mM fructose, 0.75 mM H2O2 at pH 4 or 8 adjusted with 5 mM phosphate buffer, the reaction time being 30 min at 40 ꢁC.
furanose or pyranose rings (i.e., C-chain length); thus
the presence of an overt carbonyl group in dihydroxy-
acetone (that exists in the open-chain form) could be the
main reason for its higher reactivity towards H2O2.14
Additionally, the enediol intermediate was shown to
possess different properties, depending on the pH.15
Lower pH stimulates the conversion of the enediol to
an a-oxoaldehyde intermediate,16 which through further
oxidation and decarboxylation could produce glycolate,
as we have indeed observed in metal-free reactions of
dihydroxyacetone and fructose at pH 4. Also, unlike
aldoses, in alkaline solutions ketoses could be degraded
to glycolic acid via nucleophilic attack of the perhydr-
oxyl anion to the carbonyl group, the mechanism being
similar to that described that involves an enediolate
intermediate.10,17
2.2. Concentration dependence
The depletion of monosaccharides and quantity of gly-
colate produced as a function of hydrogen peroxide con-
centration is shown in Figure 1A and C. At all H2O2

V. Maksimovic´ et al. / Carbohydrate Research 341 (2006) 2360–2369
2363
Figure 1. Monosaccharide decomposition and glycolate formation during 30 min of incubation (at 40 ꢁC) of 3 mM dihydroxyacetone or fructose
with different concentrations of H2O2 in the absence (A–C) and presence (B–D) of 10 lM FeSO4, at pH 4 or pH 8.
concentrations and both pH values, dihydroxyacetone
reacted more rapidly than fructose. There was not much
difference in the depletion of dihydroxyacetone at pH 4
(1.22 mmol hÀ1 at 0.75 mM H2O2) compared to pH 8
(1.04 mmol hÀ1 at 0.75 mM H2O2). In the case of fruc-
tose, acidic pH significantly increased the rate of mono-
saccharide depletion, compared to alkaline pH. The fact
that similar rates of dihydroxyacetone breakdown were
observed at either pH value of the reaction medium,
while glycolate yield differed (being lower in alkaline
conditions), is an indication of the existence of reaction
routes other than glycolate synthesis at higher pH. The
most probable mechanisms would be auto-oxidation
and enediol polymerisation, these having been shown
to be stimulated under similar reaction conditions.18
Auto-oxidation is also initiated by the formation of an
enediol intermediate, which by oxidation with O2 or
H2O2 subsequently forms various other intermediates,19
glycolate not necessarily being among them. Thus, alka-
line conditions could favour the transition of a newly
formed enediol intermediate in the presence of H2O2
into two directions: formation of glycolate by hydroxyl
radicals or by auto-oxidation.
times as rapid at pH 8 compared to pH 4 at 0.75 mM
H2O2, the stimulation induced by the trace metal pres-
ence decreasing with increasing hydrogen peroxide
concentrations. A fourfold increase in glycolate yield
was observed in the reaction of hydrogen peroxide
with dihydroxyacetone compared to fructose (0.29 vs
0.08 mmol hÀ1, respectively) at pH 4 and 0.75 mM
H2O2, almost negligible quantities of glycolate being
released from fructose at higher pH (0.012 mmol hÀ1).
At low pH (4), the rate of glycolate synthesis from
dihydroxyacetone was slightly enhanced (from 0.29 to
0.34 mmol hÀ1 at 0.75 mM H2O2), with similar tendency
throughout the whole range of hydrogen peroxide con-
centrations used. A striking difference was observed at
weakly alkaline pH (pH 7.5–8.5), the rate of glycolate
production increasing from 0.11 to 0.89 mmol hÀ1 at
0.75 mM H2O2 (pH 8) in the presence of metal ions.
Increased reactivity of dihydroxyacetone can again be
explained by its open-chain form and greater susceptibil-
ꢀ
ity to free-radical attack, probably by OH radical-med-
iated decarboxylation.20,23 Also, it is known that the
enediol oxidation mechanism that leads to monosaccha-
ride fragmentation at higher pH is enhanced by the Fen-
ton reaction.17,24 In the case of fructose, at low pH a
small increase of glycolate synthesis (from 0.07 to
0.2 mmol hÀ1 at 0.75 mM H2O2) could be observed.
However, alkaline pH (8) did not exhibit such an effect
as observed in the case of dihydroxyacetone, the low
rate production of glycolate being monotonous
(0.04 mmol hÀ1at 0.75 mM H2O2). In reactions proceed-
ing with relevant rates, first-order reaction kinetics are
not supported by the log-derived plots of presented
data.
In order to study the stimulating effect of free radicals
on the monosaccharide–H2O2 reaction,20,21 we em-
ployed the Fenton reaction as a simple method for
ꢀ
obtaining OH radicals.22 Catalytic amounts of metal
ions added to the reaction medium exhibited a differen-
tial effect on the monosaccharide reaction at low and
high pH (Fig. 1B and D). Higher pH stimulated the rate
of breakdown of dihydroxyacetone, whereas lower pH
stimulated the rate of fructose decomposition. Thus,
the depletion of dihydroxyacetone was almost three

2364
V. Maksimovic´ et al. / Carbohydrate Research 341 (2006) 2360–2369
2.3. pH and temperature dependence
strated that an increase in temperature resulted in a
significant increase of glycolate synthesis in the case of
dihydroxyacetone, both in the presence or absence of
metal ions. Arrhenius plots and similar activation ener-
gies calculated for Fe2+-catalysed reactions supports
In the case of fructose, acidification of the medium in-
creased the yield of glycolate, higher pH values slowly
decreasing the yield. Dihydroxyacetone exhibited a
more complex dependence with two distinct regions of
stimulation, one strong in acidified solution in the region
of pH < 5 (0.51 mmol hÀ1 at 0.75 mM H2O2) and an-
other weaker at pH 8 (0.21 mmol hÀ1 at 0.75 mM
H2O2). In the presence of metal ions, a pronounced in-
crease in glycolate yield occurred in the case dihydroxy-
acetone at higher pH values, a maximum being observed
at pH 8 (Fig. 2).
ꢀ
hypothesis that activation of the OH radical is the
rate-limiting step for both keto-monosaccharides. The
stimulation of glycolate synthesis observed at higher
temperatures supports the proposed free-radical mecha-
ꢀ
nism, since it has been shown that the reactivity of OH
radicals rapidly declines at temperatures below 30 ꢁC
and significantly increases above 40 ꢁC (Fig. 3).26
Differences in dihydroxyacetone depletion to glycolate
yield ratio at various pH values argue in favour of a
change in the reaction mechanism—either the formation
ꢀ
of intermediates capable of quenching the OH radicals
produced by the Fenton reaction, or the formation of
ꢀOH radicals participating in dihydroxyacetone cleavage
and glycolate synthesis. The decrease in glycolate yield
from dihydroxyacetone at pH 6 7, despite the expected
stimulation of the Fenton reaction, can be explained
by the formation of a hydroxyalkyl radical intermediate
ꢀ
capable of quenching OH radicals.25 It should be men-
tioned that hydroxyalkyl radicals formed in such a reac-
tion could, through decarboxylation, produce glycolate,
but at much slower rate. Our HPLC analysis confirmed
the presence of smaller quantities of glycolate among the
products at lower pH, accompanied by a slight stimula-
tion of the reaction by FeSO4.
Temperature dependence of glycolate production in
the monosaccharide–H2O2 reaction (at pH 8 for dihy-
droxyacetone and pH 4 for fructose) again demon-
Figure 3. Arrhenius plots of dihydroxyacetone and fructose reactions
after 30 min of incubation of 3 mM dihydroxyacetone or fructose with
0.75 mM H2O2 in the absence and presence of 10 lM FeSO4.
Figure 2. Effect of pH on monosaccharide decomposition and glycolate production during 30 min of incubation (at 40 ꢁC) of 3 mM
dihydroxyacetone or fructose with 0.75 mM H2O2 in the absence (A–B) and presence (C–D) of 10 lM FeSO4.

V. Maksimovic´ et al. / Carbohydrate Research 341 (2006) 2360–2369
2365
2.4. Effect of EDTA and DETAPAC on glycolate
synthesis
between different oxygen-centred free-radical species
present. EPR spectra of DEPMPO shown in Figure 4,
obtained in the reaction of dihydroxyacetone with
H2O2 at pH 8, in the absence or presence of metal ions,
showed the existence of a DEPMPO-OH adduct, no
other oxygen centred radical species being observed. In
the spectra shown in Figure 4, no difference in the mea-
sured signals could be observed in the case of N2-purged
samples from those kept in air, even after repeated sub-
stitution of the two. We have recently demonstrated that
it is possible to discern between free-radical species
dependent on external molecular oxygen, from endoge-
nously generated, oxygen-centred free radicals, by use
of gas-permeable Teflon tubes to contain measured sam-
ples.33 Control measurements using DEPMPO in buf-
fered water solution were preformed showing that the
system is free of possible artifacts. Thus, the DEPMPO
adducts detected in our measurements were due to
endogenously generated, oxygen-centred free-radical
species. This is in line with our O2 polarographic mea-
surements, where no change in oxygen concentration
could be observed in any of the samples (data not
shown), arguing that the cleavage or transformation of
monosaccharides is an anaerobic reaction.
To test the stimulatory effect of metal ions on glycolate
synthesis through the presumed formation of reactive
oxygen species, changes in glycolate yield were moni-
tored upon addition of chelators EDTA and DETA-
PAC. DETAPAC is known to be a very potent
inhibitor of the formation of highly reactive species
through metal–hydrogen peroxide interactions,27 while
EDTA was shown to be inefficient in qꢀuenching hydr-
oxyl radicals,28 even being capable of OH formation
by EDTA–iron complexes.29,30 Various FeSO4–chelator
concentration ratios were used in order to avoid
problems described elsewhere, regarding the influence
of different chelator concentrations on free-radical pro-
duction and possible competition for the spin trap.31
A
strong suppression of glycolate synthesis (P93%) was
present at all DETAPAC/FeSO4 ratios with minor dif-
ferences between experimental variations (Table 3),
while the presence of EDTA did not inhibit to such a
degree the glycolate production, no inhibition being
observed for dihydroxyacetone at pH 8.
2.5. EPR spectroscopy
Indeed, our EPR measurements (Figs. 4 and 5) show
that in the presence of EDTA a DEPMPO-OH adduct
signal can be observed, accompanied by an incomplete
inhibition of glycolate production in the dihydroxyace-
DEPMPO iꢀs a spin trap capable of forming different ad-
ꢀ
ducts with OH and O2 radicals,32 thus differentiating
Table 3. Inhibition of glycolate synthesis after addition of DETAPAC or EDTA to dihydroxyacetone or fructose at different pH valuesa,b
DHA Fructose
pH 4
pH 8
pH 4
pH 8
10 lM Fe2+ 50 lM DETAPAC
30 lM Fe2+ 150 lM DETAPAC
10 lM Fe2+ 50 lM EDTA
98.95 0.79
97.47 1.16
14.49 1.67
9.42 1.23
96.82 1.85
93.91 3.59
11.46 1.20
0.86 0.38
96.35 0.45
97.47 0.16
26.12 4.43
19.50 3.64
99.02 0.25
99.21 0.18
23.25 8.67
14.76 4.16
30 lM Fe2+ 150 lM EDTA
a Expressed as % of control in the same reaction without chelator added.
b Reaction mixtures contained 3 mM monosaccharide, 0.75 mM H2O2 at pH 4 or 8 adjusted with 5 mM phosphate buffer, the reactions being carried
out for 30 min at 40 ꢁC.
Figure 4. EPR measurements of the DEPMPO spin trap signal after 15 min of incubation in N2 or O2 atmosphere of the reaction of 3 mM
dihydroxyacetone with 1 mM H2O2, in the absence or presence of 10 lmol FeSO4, 0.5 mM EDTA or 0.5 mM DETAPAC (at 40 ꢁC and pH 8,
adjusted with 5 mM phosphate buffer).

2366
V. Maksimovic´ et al. / Carbohydrate Research 341 (2006) 2360–2369
Figure 5. EPR measurements of the DEPMPO spin trap signal after 15 min of incubation of 3 mM dihydroxyacetone or fructose with 1 mM H2O2,
in the absence or presence of 10 lmol FeSO4, 0.5 mM EDTA or 0.5 mM DETAPAC (at 40 ꢁC and pH 8, adjusted with 5 mM phosphate buffer).
tone–H2O2 reaction, suggesting that ꢀOH radical forma-
tion was not prevented. Nearly complete inhibition of
glycolate production, following the addition of DETA-
PAC, as well as the disappearance of the DEPMPO-
OH adduct signal, additionally proves that ꢀOH radicals
are required for efficient glycolate synthesis. The absence
of a DEPMPO-hydroxyalkyl radical adduct signal can
be explained by the competition between OH radicals
produced by the Fenton reaction and a low concentra-
tion of unstable hydroxyalkyl radicals.34
Figure 5 shows the EPR spectra of DEPMPO adducts
in the reaction of dihydroxyacetone and fructose with
hydrogen peroxide in the presence and absence of chela-
tors and at two different proton concentrations of the
reaction medium.
DEPMPO-OH signal supports our previous hypothesis
that in the absence of Fe2+, the free-radical mechanism
of glycolate synthesis is predominant at higher pH.
Thus, EPR measurements clearly demonstrate that
hydroxyl radicals are involved in the reaction of H2O2-
induced cleavage of monosaccharides, especially at high
pH and with the shorter chain sugar, dihydroxyacetone.
Furthermore, the observed formation of a DEPMPO-
OH adduct signal at higher pH exhibited a positive cor-
relation with enhanced glycolate formation from the
dihydroxyacetone–H2O2 reaction, resulting in maximal
amplitudes of the DEPMPO-OH adduct signal re-
corded. Observed anaerobic properties and absence of
any adducts except the DEPMPO-OH signal argues in
ꢀ
ꢀ
favour of a Fenton-like generation of OH radicals as
Presence of Fe2+ ions in the reaction markedly in-
creased the signal, showing only the presence of the
DEPMPO-OH adduct, confirmed by comparison of
recorded signals with computed simulation of the
DEPMPO-OH signal. An important difference could
be observed between dihydroxyacetone and fructose at
pH 8 (without metal present), where dihydroxyacetone
exhibited a DEPMPO-OH signal ꢀand fructose did not.
No indications of the presence of OH radicals were ob-
served at lower pH in the case of dihydroxyacetone or
any experiments with fructose. The absence of a DEP-
MPO-OH adduct signal at pH 4 in the case of dihy-
being the sole source of radicals, excluding the super-
oxide forming Haber–Weiss reaction.
2.6. Physiological relevance
A number of simple sugars capable of tautomerising to
enediols, such as those present in various biochemical
pathways, are possible targets for H2O2 and free-radical
action.36,37 The results presented above demonstrate
that different products of such action on monosac
charides (auto-oxidation, cleavage, isomerisation, free-
radical formation) can be expected under physiological
conditions, depending upon the microenvironment
(pH, presence and concentration of H2O2 and metal
ions, or hydroxyl radicals produced by other reactions),
which will affect the direction of the reaction. Also, it is
obvious that at the concentrations of monosaccharides
and reactive oxygen species that can be expected within
the cell or its organelles, especially under stress, a signi-
ficant quantity of biochemical intermediates can be
diverted to other biochemical cycles.
ꢀ
droxyacetone is a possible outcome of OH quenching
by the present enediol in its more stable unionised form,
leading to the formation of hydroxyalkyl radicals by the
reaction of addition. At higher pH dihydroxyacetone
enolisation is more rapid, leading to the formation of
an unstable enediolate anion that rapidly reduces hydro-
ꢀ
gen peroxide, thus forming small amounts of OH radi-
cals, even in the absence of metal ions.35 Correlation of
low glycolate yield from dihydroxyacetone with a weak

V. Maksimovic´ et al. / Carbohydrate Research 341 (2006) 2360–2369
2367
Glycolate, first isolated from plant tissue, is usually
3.3. HPLC determination of reaction products
identified with the process of photorespiration38 in pho-
tosynthetic cells, but it has been found in other tissues,
for example, cerebrospinal fluid and kidney in ani-
mals,39,40 often associated with urolithiasis or liver dis-
orders.41,42 In the case of photorespiration, enzymatic
synthesis of glycolate via the oxidation of the transketo-
lase reaction intermediate with H2O2,43 or oxygenase
action of Rubisco are considered to be the two
mechanisms responsible for the phenomenon.38 How-
ever, they are not able to explain the stoichiometry
and outburst of glycolate synthesis with the rise in tem-
perature, high light intensities and (drought) stress.44
Direct oxygenation of monosaccharides with H2O2,
studied previously under physiological conditions did
not show rates efficient for the estimated glycolate pro-
duction.45,46 In those reports, investigators did not
monitor the whole pH range of the reaction, with con-
comitant problems concerning quantification of reaction
products as well as sugar depletion with the methods
employed.
Separations were performed on a Waters Breeze chro-
matographic system (Waters, Milford, MA) containing
binary pumps system, a thermostated column compart-
ment and a model 2465 Waters electrochemical detector.
Injections were made manually with a Rheodyne 7725i
valve (Rheodyne, Rohnert Park, CA), with injection-
data start signal triggered through a 20 lL sample loop.
The 2465 EC detector was equipped with flow cell body
made of moulded polyurethane and a 3-mm gold work-
ing electrode with a hydrogen reference electrode. The
2465 detector was connected to the system and data
acquisition devices by a BUS-SATIN module, as an
external channel. Data acquisition and evaluation were
carried out by Waters Breeze software.
3.3.1. HPLC determination of sugar utilisation. Sepa-
ration of sugars was performed on CarboPac PA1 (Dio-
nex, Sunnyvale, CA) 250 · 4 mm column equipped with
corresponding CarboPac PA1 guard column. Sugars
were eluted for 20 min at a flow rate of 1.0 mL minÀ1
at a constant temperature of 30 ꢁC. Signals were de-
tected at in the pulse mode with the waveform:
E1 = +0.1 V for 280 ms; E2 = +0.75 V for 150 ms;
E3 = À0.85 V for 150 ms and within 80 ms of integra-
tion time. The filter timescale was of 0.2 s, and the range
was 200–500 nA for the full mV scale.
3. Experimental
3.1. Chemicals and reagents
All sugar standards, DETAPAC and EDTA were ob-
tained from Sigma–Aldrich (Sigma Co. St. Louis,
MO). We tested all of the sugars and standards for
traces of all possible reaction intermediates and prod-
ucts, especially glycolate and formate. Standard and
control solutions were prepared daily by dissolving in
18 MX redistilled and deionised water (Millipore, Bed-
ford, MA) or in the mobile phase, which were addi-
tionally filtered and degassed through a 0.22-lm filter
(Pall Gelman Lab. MI). For the preparation of
200 mM NaOH solution, 10.5 mL of sodium hydroxide
solution (50% w/w, low carbonate, J.T. Baker, Deven-
ter, Holland) was diluted in 1 L of previously vacuum-
degassed deionised water. For organic acid analysis,
5 mM sulfuric acid was prepared daily by dissolving
of 5 mL of filtered 1 M acid in 1 L of 18 MX deionised
water.
3.3.2. HPLC analysis of organic acids. Separation of
organic acids was performed on 250 · 4 mm Aminex
HPX-87H (Bio-Rad Laboratories, CA) anion-exchange
column with 5 mM H2SO4 as the mobile phase. Elution
used was isocratic with a flow of 0.6 mL minÀ1 at 30 ꢁC
with two different HPLC systems: (a) a model 2465
Waters electrochemical detector adjusted for the pulse
mode with the waveform: E1 = +0.3 V for 280 ms;
E2 = +1.4 V for 150 ms; E3 = À0.4 V for 280 ms and
40 ms integration time; (b) a Hewlett–Packard HP1100
series chromatograph (Palo, Alto, CA) composed of
an inline degasser, an autosampler, a thermostated com-
partment and an HP 1100 photo diode array detector
adjusted at 191 and 210 nm, with a reference signal at
600 nm. Additional peak confirmations were made with
peak spectral evaluation via HP Chemstation chromato-
graphic software, also used for data acquisition and
method/run control.
3.2. Reaction parameters
If it not otherwise specified, reaction conditions in all
experiments were carried out for 30 min at pH 4 or 8
(adjusted with 5 mM phosphate buffer), 40 ꢁC, with
the concentration of dihydroxyacetone and fructose
being 3 mM and hydrogen peroxide 0.1–1.5 mM. The
concentration of metals (FeSO4), when added, was
10 lM and chelators were at 500 lM. Reactions were
stopped by freezing in liquid nitrogen and stored at
À20 ꢁC prior to HPLC analysis.
3.4. EPR measurements
Spin-trapping EPR measurements were performed
using a Varian EPR spectrometer, model E 104A,
operating at X-band frequency (9.5 GHz). Measuring
conditions (H = 3410 G, SW = 200 G, MA = 2 G,
P = 10 mW) were the same in all EPR measurements,

2368
V. Maksimovic´ et al. / Carbohydrate Research 341 (2006) 2360–2369
using the spin-trap DEPMPO (5-(diethoxyphosphoryl)-
5-methyl-1-pyrroline-N-oxide) (Alexis Biochemicals,
Switzerland) in either oxygen-permeable Teflon tubes
(Zeus industries, Raritan, NJ) or in glass capillaries.
We tested all of the reaction mixtures and buffers for
occurrence of free radicals, and we did not detect any
except the ones described in our results. Computer
simulations of EPR spectra were performed using the
computer program, WINEPR SimFonia (Bruker
Analytische Messtechnik GmbH). Distilled and deion-
ised 18 MX water (Millipore) was used in all
experiments.
References
1. Moody, G. J. Adv. Carbohydr. Chem. 1964, 19, 149–179.
2. Friedmann, T. E. J. Biol. Chem. 1927, 73, 331–335.
3. Varela, O. Adv. Carbohydr. Chem. Biochem. 2003, 58, 307–
369.
4. Arts, S.; Mombarg, E.; van Bekkum, H.; Sheldon, R.
Synthesis 1997, 6, 597–613.
5. Dubois, M. F.; Rigal, L.; Gaset, A. Can. J. Chem. 1995,
73, 1582–1590.
6. Rodio, R. T.; Pereira, E. M.; Tavares, M. F. M.; Ferreira,
A. M. D. Carbohydr. Res. 1999, 315, 319–329.
ˇ
´
´
´
ˇ
7. Maksimovic, V.; Sladic, D.; Vucinic, Z. Yugoslav. Physiol.
Pharmacol. Acta 1998, 34, 135–144.
8. Calkins, V. P. Anal. Chem. 1943, 15, 762–763.
9. Isbell, H. S.; Frush, H. L.; Martin, E. T. Carbohydr. Res.
1973, 26, 287–295.
4. Conclusions
10. Isbell, H. S.; Frush, H. L. Carbohydr. Res. 1973, 28, 295–
301.
ˇ
´
´
´
ˇ
11. Maksimovic, V.; Mojovic, M.; Neumann, G.; Vucinic, Z.
Our results for the first time demonstrate that
keto-monosaccharides could under appropriate micro-
environmental physiological conditions and oxidative
stress be transformed nonenzymatically into a number
of different intermediates including glycolate, but also
2-oxogluconic acid, glucose, ribose, glucitol (sorbitol),
glyceraldehyde, formate and others, depending upon
the monosaccharide. The results presented suggest that
different mechanisms are responsible for the cleavage
or other reactions (isomerisation, auto- or free-radical-
mediated oxidation) of keto-monosaccharides depend-
ing on the keto-monosaccharide carbon chain length
and presence of metal ions. Dihydroxyacetone deriva-
tives could especially be a source of glycolate in a
cleavage reaction with H2O2, especially at higher
temperatures and pH value of the solution. At pH 8,
the presence of metal ions stimulates the reaction by a
Fenton-type of mechanism, with the participation of
only ꢀOH radicals. Addition of DETAPAC to the
reaction medium suppressed glycolate formation and
prevented ꢀOH radical formation, supporting our
hypothesis that a Fenton-type reaction is necessary
for efficient glycolate production from keto-
monosaccharides. These results could explain some of
the discrepancies in the photorespiration-related literature.
Ann. N.Y. Acad. Sci. 2005, 1048, 461–465.
ˇ
´
´
ˇ
12. Maksimovic, V.; Vucinic, Z. In Photosynthesis: Mecha-
nisms and Effects; Garab, G., Ed.; Kluwer Academic:
Dordecht–Boston–London, 1998; Vol. 5, pp 3719–3722.
13. De Wit, G.; KIeboom, A. P. G.; Van Bekkum, H.
Carbohydr. Res. 1979, 74, 157–175.
14. Isbell, H. Adv. Chem. Ser. 1973, 117, 70–87.
15. Crabbe, M. J. C.; Goode, D. Prog. Ret. E. Res. 1998, 17,
313–383.
16. Stanek, J.; Gerny, M.; Kocourek, J.; Pacak, J. In The
Monosaccharides; Academic Press: New York, 1963; pp
139–164.
17. Shalaby, M. A.; Isbell, H. S.; Elkhadem, H. S. J.
Carbohydr. Chem. 1995, 14, 429–437.
18. Thornalley, P. J. Environ. Health Persp. 1985, 64, 297–
307.
19. Mashino, T.; Fridovich, I. Arch. Biochem. Biophys. 1987,
254, 547–551.
20. Isbell, H. S.; Czubarow, P. Carbohydr. Res. 1990, 203,
287–289.
21. Isbell, H. S.; Frush, H. L.; Parks, E. W. Carbohydr. Res.
1976, 51, C5–C9.
22. Fenton, H. J. H. J. Chem. Soc. 1894, 65, 899.
23. Isbell, H. S.; Frush, H. L. Carbohydr. Res. 1977, 59, C25–
C31.
24. Wolff, S. P.; Dean, R. T. Bioelectrochem. Bioenerg. 1987,
18, 283–293.
25. Davis, G.; Thornalley, P. J. Biochim. Biophys. Acta—
Bioenergetics 1983, 724, 456–464.
26. Willson, R. L. Chem. Ind. 1977, 1, 183–192.
27. Buettner, G. R.; Doherty, T. P.; Patterson, L. K. FEBS
Lett. 1983, 158, 143–146.
28. Mc Cord, J. M.; Day, E. D., Jr. FEBS Lett. 1978, 86, 139–
142.
Acknowledgement
29. Elstner, E. F.; Osswald, W.; Konze, J. R. FEBS Lett. 1980,
121, 219–221.
30. Halliwell, B. FEBS Lett. 1978, 92, 321–326.
31. Burkitt, M. J. Free Radical Res. Commun. 1993, 18, 43–
57.
This work was supported by grant 143016B from the
Ministry of Science and Environmental Protection of
Republic of Serbia.
32. Frejaville, C.; Karoui, H.; Tuccio, B.; Le Moigne, F.;
Culcasi, M.; Pietri, S. L.; Tordo, P. J. Med. Chem. 1995,
38, 258–265.
Supplementary data
ˇ
´
´
´
ˇ ´
ˇ
33. Mojovic, M.; Vuletic, M.; Bacic, G.; Vucinic, Z. J. Exp.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.carres.
2006.06.023.
Bot. 2004, 55, 2523–2531.
34. Welch, K. D.; Davis, T. Z.; Aust, S. D. Arch. Biochem.
Biophys. 2002, 397, 360–369.

V. Maksimovic´ et al. / Carbohydrate Research 341 (2006) 2360–2369
2369
35. Wolff, S. P.; Crabbe, M. J.; Thornalley, P. J. Experientia
1984, 40, 244–246.
40. Lederer, F.; Belmouden, A.; Gondry, M. Biochem. Soc.
Trans. 1996, 24, 77–83.
36. Weiss, M. F.; Erhard, P.; Kader-Attia, F. A.; Wu, Y. C.;
Deoreo, P. B.; Araki, A.; Glomb, M. A.; Monnier, V. M.
Kidney Int. 2000, 57, 2571–2585.
37. Exner, M.; Hermann, M.; Hofbauer, R.; Kapiotis, S.;
Speiser, W.; Held, I.; Seelos, C.; Gmeiner, B. M. K. FEBS
Lett. 2000, 470, 47–50.
38. Andrews, T. J.; Lorimer, G. H.; Tolbert, N. E. Biochem-
istry 1971, 10, 4777–4782.
39. Hoffmann, G. F.; Meieraugenstein, W.; Stockler, S.;
Surtees, R.; Rating, D.; Nyhan, W. L. J. Inher. Metab.
Dis. 1993, 16, 648–669.
41. Baker, P. W.; Rofe, A. M.; Bais, R. Crit. Rev. Clin. Lab.
Sci. 1996, 33, 39–82.
42. Yanagawa, M.; Maedanakai, E.; Yamakawa, K.; Yama-
moto, I.; Kawamura, J.; Tada, S.; Ichiyama, A. Biochim.
Biophys. Acta 1990, 1036, 24–33.
43. Wilson, T. A.; Calvin, M. J. J. Am. Chem. Soc. 1955, 77,
5948–5953.
ˇ
´
44. Vucˇinic, Z.; Grodzinski, B. Fiziol. Biokhim. Kulturnih
Rastenii 1986, 18, 1–23 (in Russian).
45. Parnik, T.; Keerberg, O. J. Exp. Bot. 1995, 46, 1439–1447.
46. Oliver, J. D. Plant Sci. Lett. 1979, 15, 35–40.