Biochimica et Biophysica Acta 1569 (2002) 35^44

www.bba-direct.com

Kinetic analysis and mechanistic aspects of autoxidation of catechins

Manabu Mochizuki, Shin-ichi Yamazaki, Kenji Kano *, Tokuji Ikeda
Division of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, Sakyo, Kyoto 606-8502, Japan
Received 20 June 2001; received in revised form 2 October 2001; accepted 4 October 2001

Abstract
A peroxidase-based bioelectrochemical sensor of hydrogen peroxide (H2 O2 ) and a Clark-type oxygen electrode were applied to
continuous monitoring and kinetic analysis of the autoxidation of catechins. Four major catechins in green tea, (3)-epicatechin, (3)epicatechin gallate, (3)-epigallocatechin, and (3)-epigallocatechin gallate, were used as model compounds. It was found that dioxygen (O2 )
is quantitatively reduced to H2 O2 . The initial rate of autoxidation is suppressed by superoxide dismutase and H , but is independent of
buffer capacity. Based on these results, a mechanism of autoxidation is proposed; the initial step is the one-electron oxidation of the B ring
of catechins by O2 to generate a superoxide anion (O3
2 ) and a semiquinone radical, as supported in part by electron spin resonance
measurements. O3
2 works as a stronger one-electron oxidant than O2 against catechins and is reduced to H2 O2 . The semiquinone radical is
more susceptible to oxidation with O2 than fully reduced catechins. The autoxidation rate increases with pH. This behavior can be
interpreted in terms of the increase in the stability of O3
2 and the semiquinone radical with increasing pH, rather than the acid dissociation
of phenolic groups. Cupric ion enhances autoxidation; most probably it functions as a catalyst of the initial oxidation step of catechins. The
product cuprous ion can trigger a Fenton reaction to generate hydroxyl radical. On the other hand, borate ion suppresses autoxidation
drastically, due to the strong complex formation with catechins. The biological significance of autoxidation and its effectors are also
discussed. 2002 Elsevier Science B.V. All rights reserved.
Keywords : Catechin ; Autoxidation ; Superoxide dismutase; Cupric ion; Borate; H2 O2 sensor

1. Introduction
It has been pointed out that oxygen free radicals such as
superoxide anion radical (O3
2 ) and hydroxyl radical
(OH ) cleave DNA and peroxidize low-density lipoproteins and lipids to cause several kinds of disease [1]. Ascorbic acid and K-tocopherol are known to function as
antioxidants to scavenge those toxic free radicals and to
prevent the related diseases. Recently, increasing attention
has been paid to polyphenols as water- and fat-soluble
radical scavengers [2]. Strong reactivity of polyphenols toward O3
2 has been evidenced with electron spin resonance
(ESR) measurements [3]. Haliwell et al. have revealed high
reactivity of catechins against OH with the deoxyribose
assay method [4]. It was also reported that catechins suppress lipid peroxidation induced by water-soluble radical
generators,
2,2P-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) [5], azo-bis(2-amidinopropane)hydrochloride
Abbreviations : SOD, superoxide dismutase ; SHE, standard hydrogen
electrode
* Corresponding authors. Fax: +81-75-753-6456.
E-mail address : kkano@kais.kyoto-u.ac.jp (K. Kano).

[6], or peroxynitrite [7]. These benecial radical-scavenging

activities of catechins are proposed to correlate with the
oxidation potential [8,9].
On the other hand, catechins are easily oxidized with
dioxygen (O2 ) to generate active oxygen species such as
and hydrogen peroxide (H2 O2 ) under certain condiO3
2
tions [10,11]. It has also been reported that catechins cause
DNA cleavage and fatty acid peroxidation through the
generation of such active oxygen species, especially in
the presence of cupric ion (Cu2 ) [12,13]. Plant tissues
contain a large amount of polyphenols, which are oxidized
with O2 enzymatically [14,15] or non-enzymatically [10,11].
The product quinones have cytotoxicity such as inactivation of L-glucan synthase [15]. Therefore, native plants
should have some defense systems against the autoxidation of polyphenols. Borate is one of the candidates of
such systems and is proposed to inhibit the polyphenol
oxidase reaction [14], while it has also been reported
that borate promotes the formation of polyphenol [16].
These advantages and disadvantages of catechins and
also the roles of several eectors have been the subject
of much controversy. Most of these biological actions of
catechins seem to be closely related to their redox prop-

0304-4165 / 02 / $ ^ see front matter 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 3 0 4 - 4 1 6 5 ( 0 1 ) 0 0 2 3 0 - 6

BBAGEN 25265 23-1-02

36

M. Mochizuki et al. / Biochimica et Biophysica Acta 1569 (2002) 35^44

erty. In spite of increased interest in these biological activities, quantitative and mechanistic aspects of the autoxidation of catechins remain to be elucidated. Although consumption of O2 to be measured by a Clark-type oxygen
electrode is one of the useful methods for continuous monitoring of autoxidation [17], the autoxidation of catechins
has been monitored so far by the absorbance change of
catechins or by the determination of H2 O2 . The interpretation of spectral change is, however, complicated because
of the generation of several oxidized products including
polymerized ones. H2 O2 can be detected by the spectrophotometric or chemiluminescent method using peroxidase. However, these homogeneous enzymatic methods
cannot be applied to continuous monitoring of H2 O2 during the reaction. Furthermore, catechins must be separated
from H2 O2 -containing samples before the measurement
[10,11], since catechins function as electron donors in the
peroxidase reaction and interfere with the H2 O2 measurement. All these situations make it dicult to study the
autoxidation of catechins from kinetic viewpoints.
In the present study, we attempted to monitor the generation of H2 O2 and the consumption of O2 with a peroxidase-based amperometric sensor [18^21] and a Clarktype oxygen electrode, respectively, during the autoxidation of catechins in order to get insight into the mechanism. Four types of catechins as major polyphenol components of green tea were used as model compounds.
Special attention was paid to epigallocatechin gallate,
which is very susceptible to autoxidation. The eects of
superoxide dismutase (SOD), buer capacity, pH, Cu2 ,
and borate on autoxidation kinetics were investigated. The
result shows that O3
2 and the semiquinone intermediate of
catechins play signicant roles in autoxidation. The structural inuence on autoxidation is also discussed. Furthermore, this work reveals that Cu2 accelerates the autoxidation rate, while borate ion can suppress it drastically.
The biological signicance of these eectors is also discussed.
2. Materials and methods
2.1. Materials
(3)-Epicatechin, (3)-epicatechin gallate, (3)-epigallocatechin, and (3)-epigallocatechin gallate (see Scheme 1 for
their structures) were purchased from Sigma. H2 O2 , cuprous chloride (CuCl), cupric chloride (CuCl2 ) and glutaraldehyde were purchased form Wako. Horseradish peroxidase, SOD and bovine serum albumin were purchased
from Toyobo, Wako, and Nacalai Tesque, respectively.
Graphite powder was kindly donated by Nippon Kokuen.
Partially saponied (22 mol%) polyvinyl acetate lm
(UMR-150L, 20 Wm thickness) was kindly donated by
Unitika. All other chemicals were of reagent grade. Water
was puried on Nanopure II (Barnstead). The stock solu-

tion of CuCl was prepared in acetonitrile in order to avoid

disproportionation of Cu (into Cu2 and Cu) [22]. The
basal solutions used were 0.1 M acetate buer (3.0 6
pH 6 7.0), 0.1 M phosphate buer (pH 6 3.0, 7.0 6
pH 6 8.0, pH s 11.5), 0.1 M Tris buer (8.0 6 pH 6 9.0),
and 0.1 M carbonate buer (9.0 6 pH 6 11.5).
2.2. Amperometry of H2 O2
The principle of amperometric detection of H2 O2 used
here is a peroxidase-catalyzed and ferrocene-mediated
electrochemical reduction of H2 O2 [18,19], where ferrocene
functions as the electron donor for the peroxidase reaction
near the electrode surface and ferricenium ion is re-reduced at the electrode. In the sensor, peroxidase and mediator are, respectively, entrapped on and embedded in a
carbon paste electrode. Since catechins can also work as
electron donors of peroxidase and compete with ferrocene
even in this detection system, the contact between catechins in the test solution and peroxidase on the electrode
must be completely avoided. Covering the peroxidase/mediator-immobilized (or entrapped) electrode surface with
H2 O2 -permeable membranes is very eective for this purpose [18,19,23] as well as for stabilizing the current signal
[24].
In this study, the peroxidase-entrapped and ferroceneembedded carbon paste electrode was fabricated according
to the literature [18] with some modications. Ferrocene,
paran oil and graphite powder were thoroughly mixed in
a weight ratio of 1:3:5. The mixture was tightly packed
into the well of carbon paste electrodes (Bioanalytical Systems (BAS), inner diameter = 3.0 mm). The surface of the
electrode was smoothed on weigh paper with light pressure. On the electrode surface, 4 Wl of peroxidase solution
(1.0 mg ml31 in water) containing 0.5 mg ml31 of albumin
and 1 Wl of 0.1% glutaraldehyde in ethanol were mixed.
After the solvent was evaporated, the immobilized layer
was covered with the partially saponied polyvinyl acetate
lm. The lm was xed at the side of the electrode with an
O ring.
Constant potential amperometry was performed on a
BAS CV-50W electrochemical analyzer at +100 mV vs.
Ag/AgCl/KCl(sat.) in a three-electrode system with the
peroxidase-entrapped and ferrocene-embedded carbon
paste sensor as the working electrode, a platinum disk as
the counter electrode, and Ag/AgCl/KCl(sat.) as the reference electrode. All experiments were carried out under
moderate stirring with a magnetic stirrer at 28C. The
observed reduction current increased linearly with the concentration of H2 O2 up to about 40 WM. The detection
limit was 0.5 WM. The current intensity was independent
of pH in the range from pH 2 to pH 10, where the current
is governed by the rate of permeation of H2 O2 through the
outer membrane. The leakage of peroxidase from the immobilized layer was checked by addition of 1.0 M
K4 Fe(CN)6 to the 20 WM H2 O2 -containing test solution;

BBAGEN 25265 23-1-02

M. Mochizuki et al. / Biochimica et Biophysica Acta 1569 (2002) 35^44

when peroxidase leaking occurs from the sensor, H2 O2 in

the test solution is catalytically reduced by Fe(CN)43
6 , and
then the current decreases. This test is essential to avoid
the interference from catechins in H2 O2 measurements.
2.3. Other electrochemical measurements
The O2 concentration was measured at 28C using a
Clark-type oxygen electrode and a closed cell of 1.3 ml
volume in a two-electrode system. The applied potential
was set at 3400 mV vs. Ag/AgCl/KCl(sat.). Cyclic voltammetry was performed with a glassy carbon electrode as the
working electrode in a three-electrode system. Other details in electrochemical measurements were identical with
those described in Section 2.2.
2.4. Spectroscopic measurements
Absorption spectroscopic measurements were performed on a Shimadzu Multi-Spec 1500 at 28C. ESR
spectra were measured on a Nikkiso ES-10 spectrophotometer using a glass capillary (i.d. = 0.7 mm) as an ESR
cell, of which the bottom was capped with rubber after
introduction of samples.
3. Results and discussion
3.1. Continuous monitoring of H2 O2 and stoichiometric
consideration
Fig. 1, curve A shows the time course of the concentration of H2 O2 monitored with a peroxidase-based H2 O2
sensor after the addition of epigallocatechin gallate under
aerobic conditions (pH 9.0). The H2 O2 concentration increased just after addition of the catechin. Since no change
was observed under anaerobic conditions, H2 O2 is the
reductive product of O2 , as reported in the literature

37

[10,11]. This is the rst report on the continuous monitoring of H2 O2 generation during the autoxidation of catechins, and this method allows the kinetic analysis of the
autoxidation. In separate experiments, the reaction between catechins and H2 O2 was also examined under anaerobic conditions. However, the H2 O2 concentration (20
WM, pH 9.0) did not decrease after the addition of 20 WM
of any catechin used here. This indicates that the scavenging activity of catechins toward H2 O2 , if any, is very weak
and that the direct consumption of H2 O2 by catechins can
be ignored under the experimental time windows and at
the experimental concentrations.
We also measured the concentration prole of O2 with
an oxygen electrode during the autoxidation of catechins
(Fig. 1, curve B for epigallocatechin gallate). As compared
to curve A in Fig. 1, the generation of H2 O2 and the
consumption of O2 are almost identical within the experimental error. A similar behavior was observed for other
catechins. This result indicates that O2 is quantitatively
reduced to H2 O2 in the autoxidation. It is expected that
the ortho-di- or trihydroxyl groups of catechins are oxidized with O2 to generate the corresponding quinones
(two-electron transfer), although the quinones are not so
stable and undergo subsequent polymerization. The concentration of the generated H2 O2 (or consumed O2 ) exceeds the initial concentration of catechins (see curves A
and B in Fig. 1 for epigallocatechin gallate). This means
that two-electron oxidation of catechins is followed by
subsequent oxidation, although the subsequent oxidation
is not so fast compared with the rst two-electron oxidation. This is in accord with electrochemical studies of polyphenols, in which more than two-electron oxidation,
often giving two or more oxidation waves, is reported
[9,25].
3.2. Participation of O3
2 in autoxidation
The rst step of the autoxidation of quinols is the gen-

Scheme 1. Proposed mechanism of autoxidation of catechins.

BBAGEN 25265 23-1-02

38

M. Mochizuki et al. / Biochimica et Biophysica Acta 1569 (2002) 35^44

concentration range from 0.02 M to 0.5 M. Therefore,

does not work as a base but as
we concluded that O3
2
an oxidant in aqueous solution, as suggested by several
other authors [10,11,30]. This is in marked contrast with
what was observed in aprotic solvents, where O3
2 works
as an eective base to induce autoxidation or oxygenation
[29,31].
Considering these results, we propose a mechanism of
the autoxidation of catechins (Scheme 1). The initiation
step would be the oxidation of catechins by O2 itself, as
written in Eq. 1 [10,26]:
Fig. 1. Time course of the autoxidation of 40 WM epigallocatechin gallate in the absence (A,B) and in the presence (C) of SOD. (A,C) Amperometric monitoring of H2 O2 generation using a peroxidase-based
H2 O2 sensor. (B) Amperometric monitoring of O2 consumption using a
Clark-type oxygen electrode. (C) SOD was injected at the time indicated
by the arrow at a concentration of 80 units ml31 . All measurements
were performed in air-saturated 0.1 M Tris buer (pH 9.0) at 28C.
The current value was converted into the corresponding concentration
using the working curve prepared separately.

eration of O3
2 [26], which is disproportioned rapidly into
O2 and H2 O2 at least under neutral or slightly acidic conditions [27]. The generation of O3
2 has also been reported
for the autoxidation of catechins [10,11]. O3
2 can work as
a strong oxidant (the redox potential at pH 7 being 0.89 V
vs. the standard hydrogen electrode (SHE) [27] and/or a
weak base (pKa of HO2 = 4.68 [27]). Due to these properties, O3
2 itself can participate in the autoxidation of catechins. In order to clarify this point, the eect of SOD on
the autoxidation rate was examined with epigallocatechin
gallate as the model compound. As shown in Fig. 1, curve
C, the generation of H2 O2 (that is, the rate of autoxidation) was suppressed in part by the addition of SOD.
Similar SOD eects have been reported in the literature
[10,11].
Apo-SOD in the enzyme solution has chelating activity
toward metal ions [28]. However, in this experiment, the
addition of EDTA did not aect the autoxidation rate and
the SOD eect. This means that metal ion contaminants
do not exist in the system and that the chelating activity of
apo-SOD, if any, is not responsible for the suppression
eect. Therefore, it can be considered that the O3
2 intermediate plays an important role in the autoxidation. The
suppression eect increased with SOD concentration, but
it was saturated at SOD over 80 units ml31 . Even under
saturated conditions, the autoxidation was not completely
suppressed by SOD. Therefore, O3
is not essential for
2
autoxidation ; rather it works as a catalyst.
If O3
works as a base to enhance autoxidation, as
2
suggested by Nanni et al. [29], it is expected that the suppression eect by SOD decreases with an increase in buer
capacity, since the conjugate acid component of the buer
enhances the protonation of O3
2 and the subsequent dis3
proportionation of O2 into O2 and H2 O2 [27]. However,
no signicant dierence was observed within the buer

1
The rate of this reaction would be slow due to the low
redox potential of the O2 /O3
2 couple (30.16 V vs. SHE at
pH 7 [27]) and spin restriction between the reactants. The
product O3
2 would also oxidize catechins, as evidenced by

Fig. 2. (A) ESR spectra of epigallocatechin gallate semiquinone generated

by autoxidation at pH 9.5. The width of 100 kHz magnetic eld modulation is 0.39 mT (or 1.57 mT for the inset). (B) The simulated Lorentzian
spectrum using the hyperne coupling constants given in the text with a
line width of 0.020 mT.

BBAGEN 25265 23-1-02

M. Mochizuki et al. / Biochimica et Biophysica Acta 1569 (2002) 35^44

the SOD addition experiments mentioned above.

2
O3
is a stronger oxidant than O2 , as judged from the
2
redox potential in the case of one-electron transfer (see
above). Therefore, O3
would oxidize catechins faster
2
than O2 , and the autoxidation rate is enhanced in the
presence of O3
2 . In this context, catechins function as
good scavengers against O3
2 . However, H2 O2 is generated
in Eq. 2. Unfortunately, the reaction between catechins
and H2 O2 is very slow, if any, as stated in Section 3.1.
Therefore, one should be careful in using the expression
that catechins are powerful antioxidants.
SOD and/or H can suppress the autoxidation of catechins, since they scavenge O3
2 by accelerating the disproportionation reaction [27].

2O3
2 2H ! O2 H2 O2

This reaction competes with Eq. 2 in terms of the reduction of O3

2 . This would be the reason why the autoxidation proceeds in a steady sate (at least in the beginning of
the reaction (see Fig. 1)).
3.3. Semiquinone radical intermediate in autoxidation
The semiquinone intermediate generated in Eqs. 1 and 2
can work as a better electron donor to O2 than the fully
reduced form (catechins), as evidenced electrochemically
for quinones and avins [26].
4
Since the rate of the reaction in Eq. 4 is very large [26],
the rate determining step of catechin autoxidation would
be the one-electron oxidation of the fully reduced state
(Eqs. 1 and 2). Eq. 4 also generates O3
2 , which oxidizes
3
the fully reduced state (Eq. 2). Thus, O2 and semiquinone
work as catalysts, and the autoxidation of catechins proceeds via radical chain processes, as shown in Scheme 1.
The existence of the semiquinone was clearly evidenced
by ESR spectroscopy, as shown in Fig. 2A. The ESR
signal was fairly stable within the capillary cell, most probably due to the limited supply of O2 to the ESR cell and
the following redox equilibrium.
5
The major part of the ESR spectrum was simulated by
considering four non-equivalent 1 H nuclei with hyperne
coupling constants of 0.371, 0.097, 0.093, and 0.036 mT,
as shown in Fig. 2B, although the generation of minor
radical(s) is also suggested, as judged from some disagree-

39

ment between the observed and simulated spectra in the

middle portion of the spectra. The major radical signal can
be assigned to the B ring-derived semiquinone anion radial
of epigallocatechin gallate (one hydroxyl 1 H and 2P- and
6P-1 Hs on the B ring, and 2-1 H on the C ring). This supports that the unpaired electron is localized on the B ring,
and that the B ring does not strongly conjugate with the C
ring. The ESR intensity became strong with increasing pH.
This means that the thermodynamic stabilization of the
semiquinone generated in Eq. 5 increases with pH, as in
the case of other typical quinones [32]. This is due to the
facts that the redox reaction is coupled with the proton
transfer and that semiquinone is a stronger acid than the
fully reduced form [26,32].
3.4. pH dependence of the autoxidation
Fig. 3 shows the pH dependence of the initial rate of the
autoxidation of catechins. The H2 O2 monitoring employed
here is more sensitive than the O2 monitoring, because of
the relatively small change in the O2 signal (small consumption of O2 ). However, the application of the peroxidase-based H2 O2 sensor is limited to the pH range between 2 and 10. Therefore, the initial rate was also
evaluated from the initial slope of O2 consumption (for
example, see Fig. 1, curve B).
The rate of autoxidation increased with pH for all catechins used (Fig. 3). Under weak basic conditions (pH 7^
9.5), the autoxidation rate of epicatechin is almost the
same as that of its gallate derivative (epicatechin gallate).
A similar relationship was observed for epigallocatechin
and its gallate derivative. The gallate substituent (pyrogallol moiety) on the C ring scarcely aects the initial rate.
The result strongly suggests that the initial step of the
oxidation occurs at the B ring under such conditions.
This is in accord with the assignment of the ESR spectrum
described in Section 3.3. In addition, the electron-donating
property of the C ring pyrogallol moiety would be weakened by the adjacent electron-withdrawing carbonyl group
compared with that of the B ring.
The rate of epicatechins (epicatechin and its gallate derivative) with dihydroxyl groups (catechol moiety) on the
B ring is lower than that of epigallocatechins with trihydroxyl groups (pyrogallol moiety) under weak basic conditions. This means that the pyrogallol moiety on the B
ring is more susceptible to autoxidation than the catechol
moiety. This would be due to the electron-donating property of the additional phenolic substituent of the pyrogallol moiety on the B ring.
In contrast, in strong basic solutions (pH s 10), the reaction rate of epicatechin gallate increases drastically with
pH, and becomes comparable with that of epigallocatechin. A similar drastic increase is observed for epicatechin
gallate. Under such strong basic conditions, the gallate
substituent on the C ring would also be oxidized.
It might be expected that the pH dependence of the

BBAGEN 25265 23-1-02

40

M. Mochizuki et al. / Biochimica et Biophysica Acta 1569 (2002) 35^44

autoxidation rate is closely related to the acid dissociation

of the phenolic groups of the reactants, since the dissociated phenolic group has a stronger electron-donating
property than the undissociated one. In this context, we
attempted to evaluate the acid dissociation constants (Ka )
of epicatechin gallate by means of spectroscopic titration
under completely anaerobic conditions. Since no isosbestic
points were observed during the pH change, non-linear
regression analysis was adapted to reproduce the absorbance change at 273 and 322 nm. As shown in the inset of
Fig. 3, the data were well reproduced by assuming four
pKa values (5.30, 7.96, 8.88 and 10.96). However, the pH
dependence of the autoxidation of the catechin is not simply interpreted from the pKa values alone. This suggests
that the acid dissociation of the phenolic groups on the A
and/or C rings might not be so important due to the low
extent of Z-conjugation of the B ring with these rings, and/
or that the autoxidation is not a simple one-step oxidation
but a more complicated reaction. As described in Section
3.3, the semiquinone radical becomes stable with pH. In
addition, the stability of O3
2 increases with an increase in
pH due to the decreased kinetics of Eq. 3 [27]. This stabilization of the semiquinone and O3
at increased pH
2
seems to be responsible for the pH dependence of autoxidation.
Under physiological conditions (at pH 7.4), the autoxidation of catechins would be very slow, as judged from
Fig. 3 and the rather low concentrations of dissolved O2 in
tissues (67 WM). Even under such conditions, the autoxidation cannot be ignored in longer time windows. This
property would be responsible for the benecial and harmful activity of catechins. Expression of these contrary ef-

Fig. 3. Initial autoxidation rate of epicatechin (F), epigallocatechin (b),

epicatechin gallate (E) and epigallocatechin gallate (a) as a function of
pH. The measurements were performed at 28C with a peroxidase-based
H2 O2 sensor or a Clark-type oxygen electrode. The concentration of
catechins was xed at 20 WM. The electrolysis basal solutions used
were: 0.1 M acetate buer (pH 6 7.0), 0.1 M phosphate buer
(7.0 6 pH 6 8.0, pH s 11.5), 0.1 M Tris buer (8.0 6 pH 6 9.0), and 0.1
M carbonate buer (9.0 6 pH 6 11.5). The inset shows the absorbance
change of epigallocatechin gallate (20 WM) at 273 nm (b) and 322 nm
(a) under completely anaerobic conditions at 28C. The solid lines represent the regression curves on a model with four-step acid dissociation.
The evaluated values of pKa are given in the inset.

Fig. 4. Amperometric monitoring of the autoxidation of epigallocatechin

gallate in the presence of (A) 0, (B) 2.0, (C) 5.0, (D) 10, (E) 20, and (F)
50 WM CuCl2 . The measurements were performed in 0.1 M Tris buer
(pH 9.0) with a Clark-type oxygen electrode at 28C. The epigallocatechin gallate concentration was xed at 50 WM. The catechin stock solution was injected into the test solution at t = 0. The inset shows the (initial) steady-state autoxidation rate as a function of Cu2 concentration.

fects might be determined by the relative rates of the autoxidation of catechins and active oxygen species
scavenging catalyzed by SOD and catalase. The local concentration of O2 would also be an important factor.
3.5. Enhancement of autoxidation by Cu2+
It has been reported that addition of Cu2 accelerates
the autoxidation of ascorbate [33] and quinols [17,34].
Therefore, we investigated the eects of Cu2 on the autoxidation rate of epigallocatechin gallate with an oxygen
electrode. The result is given in Fig. 4. The reaction
reached a steady state after a short time lag, and the O2
consumption tended to reach a saturated value. Thereafter
slow O2 consumption followed (data not shown). The (initial) steady-state rate of the autoxidation increased with an
increase in the CuCl2 concentration. The CuCl2 concentration dependence showed curved characteristics, as depicted in the inset of Fig. 4.
In order to clarify the mechanism, we compared the
spectral change of the catechin on the addition of CuCl2
under anaerobic conditions with that observed during autoxidation (Fig. 5). The spectral change after the addition
of Cu2 was very similar to that observed on autoxidation.
Therefore, we can conclude that Cu2 works as a oneelectron oxidant in place of O2 .

6
A similar mechanism has been proposed for the reaction
between p-hydroquinones and Cu2 [17]. The generated
semiquinone is quickly oxidized by O2 [26], as described
above (Eq. 4). Therefore, the one-electron oxidation of the
fully reduced state (catechins) by Cu2 enhances the over-

BBAGEN 25265 23-1-02

M. Mochizuki et al. / Biochimica et Biophysica Acta 1569 (2002) 35^44

all autoxidation rate. Transition metals exist in several

spin states, and this property has an important role in
the relief of spin restriction for the autoxidation of catechins [35].
The saturated value in Fig. 4 is close to that of the
initial concentration of the catechin, and seems to be independent of the CuCl2 concentration. This suggests that
Cu is reoxidized by O2 eectively as written in Eq. 7, and
that Cu2 works as a catalyst for the two-electron oxidation of catechins.
Cu O2 ! Cu2 O3
2

Therefore, we attempted to conrm Eq. 7 using an oxygen

electrode. The O2 concentration decreased rapidly on the
addition of a CuCl acetonitrile solution. The rough evaluation of the bimolecular rate constant between Cu and
O2 yielded 1.6U102 M31 s31 at pH 9.0, although the
eects of the disproportionation of both Cu and O3
2
were ignored in the kinetic analysis.
Since the CuCl2 concentration dependence of the initial
rate constant shows curved characteristics (Fig. 4, inset),
the reaction between the catechin and Cu2 would not be
a simple bimolecular reaction. Some complex formation
might precede the oxidation. Such a complex formation
was proposed for the reaction between ascorbate and
Cu2 [33]. However, we could not get clear spectral evidence of complex formation between Cu2 and the catechin, due to the fast (succeeding) oxidation.
Hayakawa et al. reported that catechins accelerate DNA
cleavage and fatty acid peroxidation in the presence of
Cu2 [12,13]. They suggested that OH generated from
H2 O2 plays an important role in these reactions, although
the eects of Cu2 on H2 O2 generation were not examined. In the present study, we have revealed that Cu2
promotes H2 O2 generation in the autoxidation of catechins. This result can support the proposal of Hayakawa
et al.
Our mechanism suggests that Cu is generated as an
intermediate during the acceleration of autoxidation. It
is considered that Cu is responsible for the generation
of OH by a Fenton reaction:
H2 O2 Cu ! OH OH3 Cu2

Eq. 8 was monitored with the peroxidase-based H2 O2 sensor. On the addition of an aliquot of a CuCl acetonitrile
solution to the H2 O2 solution (20 WM, at pH 9.0), the
H2 O2 concentration decreased immediately. A rough evaluation of the bimolecular rate constant was approx.
2.8U102 M31 s31 . This result also supports the proposal
of Hayakawa et al. In conclusion, Cu2 accelerates the
autoxidation of catechins and the generated H2 O2 and
Cu are responsible for the generation of OH . The rate
constant of Eq. 8 is larger than that of the Fe2 -derived
Fenton reaction (68 M31 s31 [36]), but in the same order
as that of Eq. 7. Therefore, the catechins and the Cu2 derived generation of OH would become very serious

41

under relatively low concentrations of O2 , since Eq. 8

would become predominant over Eq. 7 under such conditions.
3.6. Inhibition of autoxidation by borate
The one-electron oxidation of the fully reduced state is
the important initiation step in the autoxidation of catechins, as described in Section 3.3. Therefore, if Eq. 1 can
be suppressed, the autoxidation rate will decrease. It is
well known that borate ion forms complexes with diol
compounds [37], such as catechins [38], polyhydric alcohol
(saccharides) [37] and o-dihydroxyl benzene derivatives
[39]. Therefore, we focus our attention on the eects of
borate ion on autoxidation.
The inset of Fig. 6 shows the absorption spectral change
of epigallocatechin gallate in the presence of various concentrations of borate ion under anaerobic conditions.
With increasing the borate concentration, the absorption
peak at 322 nm decreases and a new peak goes up at 301
nm, giving an isosbestic point at 309 nm. The data clearly
support the complex formation between borate and the
catechin.

9
The dissociation constant of the complex was evaluated
as 4.5U1038 M2 from the absorbance change at 322 nm
(Fig. 6, ) by assuming 2:1 complex formation (Eq. 9)
using non-liner regression analysis. The calculated values
can well reproduce the experimental data, as shown by the
solid line in Fig. 6.
As expected, the rate of autoxidation was drastically
suppressed in the presence of borate. The initial rate of
autoxidation decreased with an increase in borate concentration, as indicated by the open circles in Fig. 6. The
concentration dependence is closely related with the anaerobic absorbance change (Fig. 6, ) reecting the complex
formation. Almost all epigallocatechin gallate in solution
is complexed with borate at 20 mM of borate, where the
autoxidation is almost completely suppressed. Similar suppression eects were observed for the other catechins used.
Therefore, we can conclude that the borate complex of
catechins has strong resistance against autoxidation.
The eect of borate ion complex formation on the redox
property of catechins was also examined by cyclic voltammetry. As shown in Fig. 7, epigallocatechin gallate gave
complicated irreversible oxidation waves (curve A). In the
presence of borate, the oxidation wave shifted to the direction of the positive potential (curve B). Although thermodynamic analysis could not be performed due to the
irreversible characteristics, this result qualitatively supports the idea that the complex formation of the catechin

BBAGEN 25265 23-1-02

42

M. Mochizuki et al. / Biochimica et Biophysica Acta 1569 (2002) 35^44

Fig. 5. (A) Spectrum change of 20 WM epigallocatechin gallate under

aerobic conditions. (B) Spectrum change of 8 WM epigallocatechin gallate under anaerobic conditions in the presence of 20 WM CuCl2 . The
curves in panel A are the spectra recorded (a) 0, (b) 2700, and (c) 4900
s after the addition of the catechin, and the curves in panel B are the
spectra recorded (a) 0, (b) 900 and (c) 2700 s after the addition of the
catechin to the test solution containing Cu2 , while curve (d) shows the
spectrum of 20 WM CuCl2 in the absence of the catechin.

with borate causes the positive shift of the oxidation potential (that is, the decrease in reduction power) due to the
stabilization of the reduced form. This would be the reason why the autoxidation of the catechin is suppressed in
the presence of borate.
Although the autoxidation of the catechin was completely suppressed in the presence of appropriate amounts
of borate, it was found that the addition of CuCl2 causes
generation of H2 O2 even in the presence of borate (data
not shown). This also supports that Cu2 can work as an
initiator (or catalyst) of autoxidation. However, this may
suggest that Cu2 can coordinate catechins more strongly
than borate before oxidation, although the possibility that
Cu2 can oxidize the catechin^borate complex directly is
not ruled out.
It is considered that borate plays an important role in
plants, for example as a promoter of sugar transport

Fig. 6. The absorbance at 322 nm (b, left axis) and the autoxidation
rate (a, right axis) of epigallocatechin gallate as a function of borate
concentration. The spectroscopic measurements were performed in 0.1
M Tris buer solution (pH 9.0) under anaerobic conditions. The autoxidation rate was measured using an oxygen electrode or a peroxidasebased H2 O2 sensor at 28C. The inset shows the spectra of the catechin
in the presence of (a) 0, (b) 1.0, (c) 2.0, (d) 4.0, (e) 10, (f) 14, (g) 26, (h)
38, and (i) 100 mM borate.

Fig. 7. Cyclic voltammograms of 200 WM epigallocatechin gallate at pH

9.0 in 0.1 M Tris buer (A), and in 0.1 M borate buer (B) under anaerobic conditions. Scan rate was 500 mV s31 . The measurements were
performed at 28C.

[40,41], as a component of the cell wall [42,43], and as

an inhibitor of phenol metabolism [44]. These activities
are strongly related with the complex formation with ortho-diol compounds [41,43,44]. It has also been pointed
out that borate may restrict polyphenol oxidation in
plants [45]. The present work supports in part the proposed function of borate as an inhibitor of the autoxidation of polyphenols.
4. Conclusion
We have successfully performed the kinetic analysis of
the autoxidation of catechins using a peroxidase-based
H2 O2 sensor as well as an oxygen electrode. The kinetic
analysis revealed that the rst step of autoxidation is the
one-electron oxidation of catechins to generate a semiquinone intermediate and O3
(Eq. 1), both of which accel2
erate autoxidation. SOD and H are eective quenchers of
these radical intermediates and suppress autoxidation. The
proposed mechanism is illustrated in Scheme 1. Catechins
can work as good scavengers of O2 and O3
2 , but during
the reaction they produce H2 O2 , which is one of the active
oxygen species. In order to bring out the benecial activity
of catechins (antioxidant action), it would be important to
increase the H2 O2 scavenging activity, most probably of
catalase or peroxidase, because the reactivity of catechins
to H2 O2 is, if any, very low. At an increased rate of autoxidation, some harmful eects would become predominant. In this sense, the SOD activity would be very important to scavenge O3
2 and control the oxidation rate of
catechins.
Cu2 causes the one-electron oxidation of catechins
(Fig. 5) and enhances the overall autoxidation rate. The
product Cu can trigger a Fenton reaction to yield very
harmful OH , especially under low concentrations of O2 .
These properties would be responsible for DNA strand
cleavage induced by catechins and Cu2 [12,13]. On the
other hand, it has been found that borate inhibits the

BBAGEN 25265 23-1-02

M. Mochizuki et al. / Biochimica et Biophysica Acta 1569 (2002) 35^44

autoxidation of catechins by complex formation, leading

to a decrease in the reduction power of catechins. This
property might be in part responsible for the defense system of plants against autoxidation of polyphenols.
With regard to beverages, the autoxidation of polyphenol and related components is one of the causes of quality
deterioration. In addition, H2 O2 generated from traces of
dissolved O2 can be reduced to OH with Fe2 or other
metal ion contaminants. Oxidation by OH causes further
staling [46]. Therefore, it is very important to prevent autoxidation of the components for quality control. Fortunately, autoxidation would be negligible in slightly acidic
pH, as in the case of most of the catechin-containing beverages including green tea, beer and wine, within relatively
short periods. However, in order to maintain the quality,
including avor, it is important to avoid O2 ingress during
the production and storage process.
Acknowledgements
The authors wish to thank Dr. Hasegawa, Ichimaru
Pharcos Co., Ltd., for valuable discussion, and Nippon
Kokuen Co. (Japan) and Unitika Chemical Co. (Japan)
for their kind donation of graphite powder and partially
saponied polyvinyl acetate lm, respectively. This work
has been in part supported by Grand-in-Aids for Scientic
Research from the Ministry of Education, Science, and
Culture of Japan.

References
[1] B. Halliwell, J.M. Guttenridge, Role of free radicals and catalytic
metal ions in human disease : an overview, Methods Enzymol. 186
(1987) 1^85.
[2] C.A. Rice-Evans, N.J. Miller, G. Paganga, Structure-antioxidant activity relationships of avonoids and phenolic acids, Free Radic. Biol.
Med. 20 (1996) 933^956.
[3] Y. Tsujimoto, H. Hashizume, M. Yamazaki, Superoxide radical scavenging activity of phenolic compounds, Int. J. Biochem. 25 (1993)
491^494.
[4] B. Halliwell, J.M.C. Gutteridge, O.I. Aruoma, The deoxyribose
method : a simple `test-tube' assay for determination of rate constants
for reactions of hydroxyl radicals, Anal. Biochem. 165 (1987) 215^
219.
[5] N. Salah, N.J. Miller, G. Paganga, L. Tijburg, G.P. Bolwell, C. RiceEvans, Polyphenolic avenols as scavengers of aqueous phase radicals and as chain-breaking antioxidants, Arch. Biochem. Biophys.
322 (1995) 339^346.
[6] J. Terao, M. Piskula, Q. Yao, Protective eect of epicatechin, epicatechin gallate, and quercetin on lipid peroxidation in phospholipid
bilayers, Arch. Biochem. Biophys. 308 (1994) 278^284.
[7] A.S. Pannala, C.A. Rice-Evans, B. Halliwell, S. Singh, Inhibition of
peroxynitrite-mediated tyrosine nitration by catechin polyphenols,
Biochem. Biophys. Res. Commun. 232 (1997) 164^168.
[8] S.A.B.E. van Acker, D.J. van den Berg, M.N.J.L. Tromp, D.H.
Grioen, W.P. van Bennekom, W.J.F. van der Vijgh, A. Bast, Structural aspects of antioxidant activity of avonoids, Free Radic. Biol.
Med. 20 (1996) 331^342.

43

[9] B. Yang, K. Arai, F. Kusu, Oxidation potentials of avonoids determined by ow-through column electrolysis, Electrochemistry 69
(2001) 519^525.
[10] T. Nakayama, Y. Enoki, K. Hashimoto, Hydrogen peroxide formation during catechin oxidation is inhibited by superoxide dismutase,
Food Sci. Technol. Int. 1 (1995) 65^69.
[11] Y.H. Miura, I. Tomita, T. Watanabe, T. Hirayama, S. Fukui, Active
oxygen generation by avonoids, Biol. Pharm. Bull. 21 (1998) 93^96.
[12] F. Hayakawa, T. Kimura, T. Maeda, M. Fujita, H. Sohmiya, M.
Fujii, T. Ando, DNA cleavage reaction and linoleic acid peroxidation
induced by tea catechins in the presence of cupric ion, Biochim. Biophys. Acta 1336 (1997) 123^131.
[13] F. Hayakawa, T. Kimura, N. Hoshino, T. Ando, DNA cleavage
activities of (3)-epigallocatechin, (3)-epicatechin, (+)-catechin, and
(3)-epigallocatechin gallate with various kind of metal ions, Biosci.
Biotechnol. Biochem. 63 (1999) 1654^1656.
[14] I. Cakmak, V. Romheld, Boron deciency-induced impairments of
cellular functions in plants, Plant Soil 193 (1997) 71^83.
[15] T.L. Mason, B.P. Wasserman, Inactivation of red beet L-glucan synthase by native and oxidized phenolic compounds, Phytochemistry 26
(1987) 2197^2202.
[16] J.A. Rajaratnam, J.B. Lowry, P.N. Avadhani, R.H.V. Corley, Boron: possible role in plant metabolism, Science 172 (1971) 1142^1143.
[17] Y. Li, M.A. Trush, Oxidation of hydroquinone by copper: chemical
mechanism and biological eects, Arch. Biochem. Biophys. 300
(1993) 346^355.
[18] H. Kinoshita, M. Torimura, K. Kano, T. Ikeda, Peroxidase-based
amperometric sensor of hydrogen peroxide generated in oxidase reaction: application to creatinine and creatine assay, Electroanalysis 9
(1997) 1234^1238.
[19] H. Kinoshita, M. Torimura, K. Kano, T. Ikeda, Amperometric determination of high-density lipoprotein cholesterol using polyethylene
glycol-modied enzymes and a peroxidase-entrapped electrode, Ann.
Clin. Biochem. 35 (1998) 739^744.
[20] T. Ikeda, T. Shiraichi, M. Senda, An ecient method for entrapping
ionic mediators in the enzyme layer of mediated amperometric biosensors, Agric. Biol. Chem. 52 (1988) 3187^3188.
[21] L. Yang, E. Janle, T. Huang, J. Gitzen, P.T. Kissinger, M. Vreeke, A.
Heller, Applications of `wired' peroxidase electrodes for peroxide
determination in liquid chromatography coupled to oxidase immobilized enzyme reactors, Anal. Chem. 67 (1995) 1326^1331.
[22] A.J. Parker, Hydrometallurgy of copper and silver in solvent mixtures, Search 4 (1973) 426^432.
[23] K. Yamamoto, T. Ohgaru, M. Torimura, H. Kinoshita, K. Kano, T.
Ikeda, Highly-sensitive ow injection determination of hydrogen peroxide with a peroxidase-immobilized electrode and its application to
clinical chemistry, Anal. Chim. Acta 406 (2000) 201^207.
[24] K. Kano, T. Ikeda, Fundamentals and practices of mediated bioelectrocatalysis, Anal. Sci. 16 (2000) 1013^1021.
[25] H. Hotta, H. Sakamoto, S. Nagano, T. Osakai, Y. Tsujino, Unusually large numbers of electrons for the oxidation of polyphenolic
antioxidants, Biochim. Biophys. Acta 1526 (2001) 159^167.
[26] H. Tatsumi, H. Nakase, K. Kano, T. Ikeda, Mechanistic study of the
autoxidation of reduced avin and quinone compounds, J. Electroanal. Chem. 443 (1998) 236^242.
[27] D.T. Sawyer, J.S. Valentine, How super is superoxide ?, Acc. Chem.
Res. 14 (1981) 393^400.
[28] T.D. Rae, A.S. Torres, R.A. Pufahl, T.V. O'Halloran, Mechanism of
Cu,Zn-superoxide dismutase activation by the human metallochaperone hCCS, J. Biol. Chem. 276 (2001) 5166^5176.
[29] E.J. Nanni Jr., M.D. Stallings, D.T. Sawyer, Does superoxide ion
oxidize catechol, K-tocopherol, and ascorbic acid by direct electron
transfer?, J. Am. Chem. Soc. 102 (1980) 4481^4485.
[30] J.M. Ricardo da Silva, N. Darmon, Y. Fernandez, S. Mitjavila, Oxygen free radical scavenger capacity in aqueous models of dierent
procyanidins from grape seeds, J. Agric. Food Chem. 39 (1991)
1549^1552.

BBAGEN 25265 23-1-02

44

M. Mochizuki et al. / Biochimica et Biophysica Acta 1569 (2002) 35^44

[31] K. Kano, T. Mabuchi, B. Uno, Y. Esaka, T. Tanaka, M. Iinuma,

Superoxide anion radical-induced dioxygenolysis of quercetin as a
mimic of quercetinase, J. Chem. Soc. Chem. Commun. (1994) 593^
594.
[32] J.Q. Chambers, Electrochemistry of quinones, in: S. Patai, Z. Rappoport (Eds.), The Chemistry of the Quinonoid Compounds, Vol. 2,
Part 1, John Wiley, New York, 1988, pp. 719^757.
[33] R.F. Jameson, N.J. Blackburn, Role of copper dimers and the participation of copper(III) in the copper-catalysed autooxidation of
ascorbic acid. Part II. Kinetics and mechanism in 0.100 mol dm33
potassium nitrate, J. Chem. Soc. Dalton (1976) 534^541.
[34] L. Zhang, B. Bandy, A.J. Davison, Eects of metals, ligands and
antioxidants on the reaction of oxygen with 1,2,4-benzenetriol, Free
Radic. Biol. Med. 20 (1996) 495^505.
[35] D.M. Miller, G.R. Buettner, S.D. Aust, Transition metals as catalysts
of `autoxidation' reactions, Free Radic. Biol. Med. 8 (1990) 95^108.
[36] G.B. Shul'pin, Y.N. Kozlov, G.V. Nizova, G. Suss-Fink, S. Stanislas,
A. Kitaygorodskiy, V.S. Kulikova, Oxidations by the reagent `O2 H2 O2 -vanadium derivative-pyrazine-2-carboxylic acid' Part 12. Main
features, kinetics and mechanism of alkane hydroperoxidation,
J. Chem. Soc. Perkin Trans. 2 (2001) 1351^1371.
[37] J. Boeseken, The use of boric acid for the determination of the conguration of carbohydrates, Adv. Carbohydr. Chem. 4 (1949) 189^
210.
[38] H. Tsuchiya, M. Sato, H. Kato, T. Okubo, L.R. Juneja, M. Kim,

[39]

[40]

[41]
[42]
[43]

[44]
[45]
[46]

Simultaneous determination of catechins in human saliva by highperformance liquid chromatography, J. Chromatogr. B Biomed.
Sci. Appl. 703 (1997) 253^258.
E. Shoji, M.S. Freund, Potentiometric sensors based on the inductive
eect on the pKa of poly(aniline): a nonenzymatic glucose sensor,
J. Am. Chem. Soc. 123 (2001) 3383^3384.
W.M. Dugger, Boron in plant metabolism, in: Encyclopedia of Plant
Physiology, New Series, Vol. 15, Springer-Verlag, Heidelberg 1983,
pp. 626^650.
K.W. Lee, C.M. Whittle, H.J. Dyer, Boron deciency and translocation proles in sunower, Physiol. Plant. 19 (1966) 919^924.
T. Matoh, Boron in plant cell walls, Plant Soil 193 (1997) 59^69.
M.A. O'Neill, D. Warrenfeltz, K. Kates, P. Pellerin, T. Doco, A.G.
Darvill, P. Albersheim, Rhamnogalacturonan-II, a pectic polysaccharide in the walls of growing plant cells, forms a dimer that is covalently cross-linked by a borate ester, J. Biol. Chem. 271 (1996) 22923^
22930.
S. Lee, Boron in plants: a biochemical role, Science 158 (1967) 798^
799.
D.H. Lewis, Boron, lignication and the origin of vascular plants ^ a
unied hypothesis, New Phytol. 84 (1980) 209^229.
M. Uchida, M. Ono, Determination of hydrogen peroxide in beer
and its role in beer oxidation, J. Am. Soc. Brew. Chem. 57 (1999)
145^150.

BBAGEN 25265 23-1-02