Degradation of cyanidin 3-glucoside by caffeic acid o-quinone. Determination of the stoichiometry and characterization of the degradation products

Article abstract of DOI:10.1021/jf981400x

Caffeic acid o-quinone (CQ) was prepared by oxidation of caffeic acid with o-chloranil in organic media. The reaction between the purified CQ and cyanidin 3-glucoside (Cy 3-glc, o-diphenolic anthocyanin) was monitored by HPLC, and quantitative analyses we

Full text of DOI:10.1021/jf981400x

J. Agric. Food Chem. 1999, 47, 4625−4630
4625
Degr a d a tion of Cya n id in 3-Glu cosid e by Ca ffeic Acid o-Qu in on e.
Deter m in a tion of th e Stoich iom etr y a n d Ch a r a cter iza tion of th e
Degr a d a tion P r od u cts
†
†
‡
‡
Farid Kader, Mohammed Irmouli, Nedjma Zitouni, J ean-Pierre Nicolas, and
Maurice Metche*,
†,‡
Laboratoire de Biochimie Appliqu e´ e, Ecole Nationale Sup e´ rieure d’Agronomie et des Industries
Alimentaires, 2 Avenue de la For eˆ t de Haye, B.P. 172, 54505 Vandoeuvre-l e` s-Nancy Cedex, France,
and Laboratoire de Biochimie Nutritionnelle, Unit e´ INSERM 308, Facult e´ de M e´ decine de Nancy,
9
Avenue de la For eˆ t de Haye, B.P. 184, 54500 Vandoeuvre-l e` s-Nancy Cedex, France
Caffeic acid o-quinone (CQ) was prepared by oxidation of caffeic acid with o-chloranil in organic
media. The reaction between the purified CQ and cyanidin 3-glucoside (Cy 3-glc, o-diphenolic
anthocyanin) was monitored by HPLC, and quantitative analyses were performed to establish the
stoichiometry of the reaction. The results indicate that Cy 3-glc is degraded by a coupled oxidation
mechanism with integration of CQ into the degradation products. The ratio of degraded Cy 3-glc to
CQ incorporated into the condensation products was ∼2.0. No brown products could be detected,
only a slight orange color. Moreover, the addition of purified polyphenol oxidase to the slightly colored
media resulted in the disappearance of the caffeic acid formed from the reaction of coupled oxidation
(
Cy 3-glc/CQ) and the formation of brown polymers. The degradation products were isolated by gel
filtration on Sephadex G-25. The UV-vis spectra and chemical analysis (acidic hydrolysis) of the
degradation products suggest that they resulted from the condensation of caffeic acid and Cy 3-glc.
HPLC analysis showed that the partial purified fraction contained a mixture of complex condensation
products.
Keyw or d s: Caffeic acid o-quinone; anthocyanins; stoichiometry; degradation products; partial
characterization
INTRODUCTION
degradation of Cy 3-glc and malvidin 3-glucoside (Mv
-glc) in the presence of caftaric acid and grape PPO.
They have shown that the products of anthocyanin (Cy
-glc and Mv 3-glc) degradation contain both anthocya-
3
Browning reactions that occur after fresh blueberries
(
Vaccinium corymbosum) have been crushed are due
mainly to polyphenol oxidases (PPOs) (Kader et al.,
997b). Chlorogenic acid (CG) is the major hydroxycin-
namic derivative found in blueberries (Kader et al.,
996), and this o-diphenolic compound is a good sub-
3
nin and caftaric acid moieties.
1
The use of model systems has shown that the enzy-
matically generated o-quinones play an essential role
in the process of anthocyanin degradation. However,
this approach does not allow determination of the exact
mechanisms involved in the degradation of both o-
diphenolic and non-o-diphenolic anthocyanins. Pure
solution of quinones should be useful in understanding
the mechanisms of anthocyanin degradation. The prepa-
ration of pure quinones by chemical oxidation should
allow simplification of the model reactions. The reaction
between o-quinones and anthocyanins could facilitate
both the isolation of the reaction products and the study
of their structure to understand the mechanism of
anthocyanin degradation by PPO. However, the syn-
thesis of o-quinones is not easy because of their instabil-
ity in aqueous media.
Sarni-Manchado et al. (1997) studied the reaction of
Mv 3-glc with caftaric acid o-quinone. Liquid chroma-
tography with ion spray mass spectrometry indicates
that the products of degradation are adducts of Mv 3-glc
and caftaric acid. The results show also that the
hemiacetal forms of the pigment are more reactive than
the flavylium form. Recently, we investigated the reac-
tion between anthocyanins and o-chloranil (Kader et al.,
1999b) and demonstrated that the reaction proceeds in
two steps. The first is characterized by the discoloriza-
1
strate for blueberry PPO (Kader et al., 1997a). The first
reaction is the oxidation of chlorogenic acid by PPO to
its o-quinone (chlorogenoquinone, CGQ), followed by the
reaction of the enzymatically generated o-quinone with
the anthocyanins to form brown polymers (Kader et al.,
1
997b). Because of the complexity of the blueberry
anthocyanin extracts, model systems were useful in the
study of the mechanims of anthocyanin degradation by
PPO in the presence of CG. The results showed that the
enzymatically generated CGQ oxidizes Cy 3-glc (o-
diphenolic anthocyanin) by a coupled oxidation mech-
anism with partial regeneration of the CG, which means
that part of the CG is incorporated into the degradation
products of the Cy 3-glc (Kader et al., 1998). Pelargoni-
din 3-glucoside (Pg 3-GLC, non-o-diphenolic anthocya-
nin) might react with the quinones or secondary prod-
ucts of oxidation (formed from the CGQ) to form adducts
(Kader et al., 1999a). Sarni et al. (1995) studied the
*
Author to whom correspondence should be addressed [fax
3
3 (0)3 83 15 34 56; e-mail nicolas@u308.nancy.inserm.fr].
†
Ecole Nationale Sup e´ rieure d’Agronomie et des Industries
Alimentaires.
‡
Facult e´ de M e´ decine de Nancy.
1
0.1021/jf981400x CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/28/1999

4626 J. Agric. Food Chem., Vol. 47, No. 11, 1999
Kader et al.
tion of the anthocyanin solution accompanied by the
formation of three products of combination (anthocya-
nin-chloranil). Whatever the pigment studied (o-diphe-
nolic anthocyanin or non-o-diphenolic anthocyanin), this
step is characterized by a ratio of degraded anthocyanin
to o-chloranil of close to 1. The results showed that the
carbinol-pseudobase form reacts with o-chloranil. Dur-
ing the second step of the reaction (for o-chloranil to
anthocyanins ratios >1) the colorless combinations react
readily with o-chloranil to yield brown condensation
products. Chloranil was chosen because (i) this o-
quinone is stable in aqueous media, (ii) it is available
commercially, (iii) it has a high redox potential, and (iv)
it is totally substituted, which impedes the effect of
water (nucleophilic addition). However, the results
suggested that the reaction mechanism is different from
those described for the PPO-catalyzed oxidation of
anthocyanins in the presence of o-diphenolic substrates
such as chlorogenic and caftaric acids (Kader et al.,
mL of 0.235 mM CQ, and 0.650 mL of McIlvaine buffer (pH
3
.5). To determine the CQ concentration, the reaction mixture
contained 0.250 mL of CQ, 0.050 mL of ascorbic acid or sodium
benzenesulfinate (20 mM in McIlvaine buffer, pH 3.5), and 0.7
mL of McIlvaine buffer (pH 3.5). After 5 min of reaction, 0.1
mL of the reaction mixture was analyzed by HPLC (Merck-
Hitachi L-6200 Intelligent pump equipped with a diode array
detector Merck-Hitachi L-3000 connected to a Chromojet
integrator). The constituents of the medium were separated
on a Lichrosorb 100 RP-18 reversed phase column (250 × 4
mm i.d., 5-µm packing) (Merck, Darmstadt, Germany) pro-
tected by a guard cartridge of the same material. The mobile
phases were the same as described above. The gradient profile
was 0-5 min, 5% B; 5-20 min, 5-40% B; 20-35 min, 40-
8
0% B; 35-45 min, 80-100% B. Elution was performed at a
-
1
flow rate of 1.0 mL min , and 0.1 mL of the reaction mixture
was injected using a Basic Marathon automatic injector
+
(Spark Holland). Caffeic acid and Cy 3-glc were detected at
2
80 nm.
The CQ was determined in equivalents of caffeic acid by
reduction of the quinone by ascorbic acid. Caffeic acid and Cy
3-glc concentrations were determined by HPLC as described
above using a calibration curve ranging from 0.01 to 0.1 mM.
For each assay two analyses were conducted on duplicate
experimentations. Each data point is therefore the mean of
four measurements.
1
1
998, 1999a; Sarni-Manchado et al., 1997; Sarni et al.,
995).
The purpose of this work was to study the reactions
between o-quinones (synthesized from natural phenolic
compounds) and anthocyanins. The CQ was prepared
from caffeic acid by oxidation with o-chloranil and
purified by HPLC on the semipreparative scale. The
reaction between CQ and Cy 3-glc was monitored by
HPLC to determine the stoichiometry of the reaction.
The products of the reaction were separated by gel
filtration and characterized by their UV-vis spectra.
P u r ifica tion of th e Rea ction P r od u cts. The reaction
mixture contained 0.4 mL of McIlvaine buffer (pH 3.5), 0.4
mL of 1 mM Cy 3-glc, and 2.0 mL of 0.2 mM CQ. The reaction
was started by adding the CQ solution. After a reaction time
of 5 min, the solvent was removed under reduced pressure
using a rotary evaporator at 25 °C. The resulting solution was
applied onto a Sephadex G-25 column (15 × 3 cm i.d.),
previously equilibrated with distilled water, at a flow rate of
-
1
1
5 mL h . The reaction products were eluted with distilled
MATERIALS AND METHODS
water, and the absorbances at 280 and 325 nm were recorded
on each 3-mL fraction.
Ch em ica ls. Cy 3-glc (Kuromanin) was of HPLC grade from
Extrasynth e` se (Genay, France). Sephadex G-25 (fine, particule
size ) 20-80 µm) was from Pharmacia (Uppsala, Sweden).
o-Chloranil was obtained from Aldrich (Strasbourg, France).
Caffeic acid, ascorbic acid, trifluoroacetic acid (99.5% of purity),
orthophosphoric acid (minimum 85% purity), and benzene-
sulfinic acid were obtained from Sigma Chemicals (St. Quentin
Fallavier, France). Methanol (HPLC grade), acetic acid (99.5%
purity), chloroform (99.5% purity), ethyl acetate (99.5% purity),
propan-2-ol (analytical grade), and all other chemicals were
of reagent grade from Merck (Darmstadt, Germany). Blueberry
Ch a r a cter iza tion of th e Rea ction P r od u cts. The UV-
vis spectra were recorded from 220 to 600 nm using a
Shimadzu UV-260 spectrophotometer. The reaction products
(0.5 mg) occurring in the fractions C1 were hydrolyzed by
heating at 85 °C in 3.0 mL of 2 M trifluoroacetic acid in
methanol under nitrogen for 45 min. The hydrolyzed solutions
were cooled, and 2.0 mL of water was added. Methanol and
trifluoroacetic acid were removed under vacuum at 35 °C. The
residual aqueous solution was extracted four times with ethyl
acetate (1:1, v/v). The aqueous phase was concentrated in a
rotary evaporator under vacuum at 35 °C. The resulting
solution (0.2 mL) was chromatographed to identify glucose on
a silica gel plate (Merck, ref 5553) with the propan-2-ol/ethyl
acetate/water (50:40:10, v/v/v) system. The glucose was located
on TLC plates by spraying with 2% (w/v) naphthoresorcinol
solution in acetone and 9% orthophosphoric acid (5:1, v/v)
followed by heating in an oven at 105 °C for 10 min.
-
1
PPO (0.194 nkat mL ) was obtained as described by Kader
et al. (1997a). Cy 3-glc (2 mM) was dissolved in McIlvaine
buffer (pH 3.5). This buffer was prepared from 0.1 M citric
acid adjusted to the correct pH by adding 0.2 M dibasic
potassium phosphate.
P r ep a r a tion a n d P u r ifica tion of th e CQ. Caffeic acid
and o-chloranil solutions were prepared in anhydrous metha-
nol. CQ was prepared from caffeic acid by oxidation with
o-chloranil at 25 °C. The reaction mixture contained 0.2 mL
of 50 mM caffeic acid, 0.2 mL of 100 mM o-chloranil, and 1.6
mL of chloroform. The reaction was initiated by adding the
o-chloranil solution. The CQ formed was purified by HPLC.
After 2.5 min of reaction, 1.0 mL of the oxidized solution was
analyzed by HPLC on the semipreparative scale. The HPLC
apparatus was a Spectra Physics system including an SP 8000
ternary HPLC pump, a manual injector (Reodyne, Model 7010
sample injection valve), and a Spectra 100 variable-wavelength
detector set at 400 nm and connected to a Chromojet integra-
tor. The constituents of the reaction mixture were separated
on a Nucleosil C18 (10 µm packing), 250 × 7.5 mm, column
protected by a guard cartridge of the same packing. Elution
conditions: solvent A, 2.5% acetic acid in water (v/v); solvent
HP LC of th e P a r tia lly P u r ified P r od u cts. Fraction C1
was analyzed by HPLC using the same apparatus and column
as described for the degradation of Cy 3-glc by CQ. The elution
conditions were as follows: solvent A, 1% acetic acid in distilled
water (v/v); solvent B, methanol; flow rate, 1.0 mL min-1
;
linear gradients from 0 to 100% B in 50 min; 0.2 mL of the
reaction mixture was injected using a Basic Marathon
+
automatic injector (Spark, Holland). The reaction products
were detected at 280 nm.
RESULTS AND DISCUSSION
Syn th esis a n d Isola tion of th e CQ. CQ was pre-
pared by chemical oxidation of caffeic acid by o-chloranil
in organic media (chloroform). We followed the method
described by Davies (1976) for the synthesis of CQ and
CQ methyl ester. The CQ was isolated by semiprepara-
tive HPLC. The UV-vis spectra of the purified CQ and
caffeic acid were recorded from 240 to 500 nm (Figure
1). The CQ exhibited three absorption maxima at 248,
-
1
B, 80% acetonitrile in solvent A (v/v); flow rate, 1.5 mL min
.
The gradient profile was 0-5 min, 5% B; 5-20 min, 5-20%
B; 20-35 min, 20-40% B; 35-45 min, 40-100% B. The CQ
was collected and used immediately.
Mod el System of th e Degr a d a tion of Cy 3-glc by CQ.
The reaction mixture contained 0.1 mL of 1 mM Cy 3-glc, 0.250

Degradation of Cyanidin 3-Glucoside by Caffeic Acid o-Quinone
J. Agric. Food Chem., Vol. 47, No. 11, 1999 4627
F igu r e 2. HPLC elution profile at 280 nm of a solution of Cy
3
-glc (0.1 mM) incubated for 5 min with the purified CQ (0.06
mM) in McIlvaine buffer (pH 3.5). Peak 1, caffeic acid; peak
, remaining Cy 3-glc.
2
F igu r e 1. UV-vis spectra of caffeic acid (dashed line) and
the purified caffeic acid o-quinone (solid line) collected after
elution from the HPLC column.
CQ as the corresponding phenyl sulfone (Cheynier and
Ricardo da Silva, 1991). The amount of CQ was then
calculated as the difference between the caffeic acid
concentrations in the ascorbic acid and benzenesulfinate
samples.
Degr a d a tion of Cy 3-glc by CQ in McIlva in e
Bu ffer (p H 3.5). The addition of CQ to a Cy 3-glc
solution resulted in the red color disappearing and the
concomitant formation of a slight orange color. The
reaction mixture was analyzed by HPLC at 280 nm
3
00-330, and 400 nm. The UV-vis spectrum was very
close to that published by Cheynier and Moutounet
1992) for the CQ obtained by enzymatic oxidation of
(
caffeic acid at pH 3.6. o-Quinones formed by enzymatic
and chemical oxidation of o-diphenols can also be
characterized through the formation of their stable
phenylsulfonyl derivatives (reaction 1) (Davies and
Pierpoint, 1975). The addition of sodium benzenesulfi-
(
Figure 2). The chromatogram shows that part of the
pigment has disappeared and that a new product (peak
) has formed with an elution time shorter than that of
CQ + benzenesulfinic acid f
1
caffeic acid benzenesulfone (1)
Cy 3-glc (peak 2). This new peak has been identified as
caffeic acid on the basis of its retention time and UV-
vis spectrum (Figure 1). This result shows that Cy 3-glc
is oxidized by CQ, leading to the formation of caffeic
acid according to reaction 3. It is highly probable that
nate to the o-quinone solution (yellow coloration) re-
sulted in an instantaneous discoloration of the solution.
The o-quinone solutions treated with sodium benzene-
sulfinate were analyzed by HPLC on the analytical
scale. The elution profile showed the presence of trace
amounts of caffeic acid along with the expected phenyl-
sulfonyl peak. This result suggests that spontaneous
reduction of the CQ can take place. The amount of
caffeic acid detected in the o-quinone solutions is
relatively constant, ranging from 2 to 3 µM. Cheynier
and Moutounet (1992) observed the same phenomenon
after analyzing a solution of CQ by HPLC. In fact, the
formation of caffeic acid can probably be explained by
dismutation of the CQ in caffeic acid and hydroxycaffeic
acid rather than by a process of spontaneous reduction
of the CQ (Richard-Forget at al., 1992).
CQ + Cy 3-glc f caffeic acid + Cy 3-glc o-Q (3)
Cy 3-glc degradation proceeds by a mechanism of
coupled oxidation as proposed by Kader et al. (1998) and
Sarni et al. (1995). Additional experiments were carried
out to characterize the Cy 3-glc o-Q formed during this
reaction process (reaction 3). Thus, sodium benzene-
sulfinate was added to the discolored reaction mixture
to trap the Cy 3-glc o-Q, which can then be detected as
the corresponding phenyl sulfone by HPLC. However,
no Cy 3-glc phenyl sulfone was detected in the reaction
mixture, which means that Cy 3-glc o-quinone is very
reactive and proceeds readily to condensation products,
as mentioned by Sarni et al. (1995). No brown color was
observed in the reaction mixture, which means that the
degradation products of the Cy 3-glc are not involved
in the formation of brown polymers as observed previ-
ously by Kader et al. (1998). However, the addition of
blueberry PPO induced the formation of brown poly-
mers. The HPLC analysis under the same conditions
showed that the caffeic acid had disappeared from the
reaction mixture. This could indicate that the browning
reaction is mainly due to the condensation products of
caffeic acid oxidation (PPO). On the other hand, the
degradation products of Cy 3-glc are slightly colored.
In the presence of excess reducing agents such as
ascorbic acid, o-quinones are reduced to their original
o-diphenols (reaction 2). The reaction is stoichiometric
CQ + ascorbate f
caffeic acid + dehydroascorbate (2)
with ascorbic acid (Rouet-Mayer et al., 1990), and
consequently this property can be used to determine the
CQ concentration. This was achieved by taking samples
in duplicate and stabilizing one of them by adding
ascorbic acid (1 mM) to reduce the CQ back to the
corresponding hydroquinone (caffeic acid) and adding
sodium benzenesulfinate (1.0 mM) to the other to trap

4628 J. Agric. Food Chem., Vol. 47, No. 11, 1999
Kader et al.
Ta ble 1. Oxid a tion of Cy 3-glc by CQ in McIlva in e Bu ffer (p H 3.5) in Differ en t Rea ction Con d ition sa
initial amount of
initial amount of
amount of caffeic
amount of remaining
theoretical amount of
expt
Cy 3-glc, µmol
CQ, µmol
acid formed, µmol
Cy 3-glc, µmol
Cy 3-glc o-Q formed, µmol
I
II
III
0.1 ( 0.005
0.075 ( 0.004
0.095 ( 0.006
0.047 ( 0.003
0.0525 ( 0.004
0.042 ( 0.003
0.033 ( 0.003
0.039 ( 0.004
0.027 ( 0.002
0.067 ( 0.006
0.036 ( 0.003
0.062 ( 0.005
0.033
0.039
0.033
a
All compounds were determined by HPLC using calibration curves. For each assay two analyses were conducted on duplicate
experimentations. Each data point is therefore the mean of four measurements.
Quantitative analyses were carried out to determine
the stoichiometry of the reaction between Cy 3-glc and
CQ (Table 1). The reaction was monitored by HPLC
(
analytical scale). The CQ concentration was determined
as indicated above in the presence of ascorbic acid and
sodium benzenesulfinate (reactions 1 and 2). Cy 3-glc
and caffeic acid concentrations were determined by
HPLC using calibration curves that were constructed
by injection (100 µL) of 10 standards containing different
concentrations of caffeic acid or Cy 3-glc ranging from
0
.01 to 0.1 mM. In the experiments, the initial amount
of Cy 3-glc was always higher than that of CQ. Under
these experimental conditions, after 5 min of reaction,
the residual CQ content was almost zero, which means
that all of the CQ reacted. The results (Table 1) show
that the amount of caffeic acid formed is equivalent to
the amount of Cy 3-glc which has reacted. Moreover,
the amount of caffeic acid formed was much lower than
that expected from the initial content of CQ. This result
indicates that part of the CQ is integrated into the
degradation products of the pigment. In experiment I
F igu r e 3. Elution profile on Sephadex G-25 of a solution of
Cy 3-glc (0.133 mM) incubated for 5 min with the purified CQ
(0.133 mM) at pH 3.5. Absorbance at 280 nm (0); absorbance
at 335 nm (]).
these conditions, three peaks were obtained: a small
peak (C1) that absorbed at 280 and 330 nm, followed
by two large peaks (C2 and C3) that were shown to be
caffeic acid and Cy 3-glc, respectively, by HPLC analysis
of the corresponding fractions.
Examining the elution profile (Figure 3) reveals that
the Sephadex G-25 gel has separated the Cy 3-glc (MW
(
Table 1), 0.033 µmol of Cy 3-glc reacted with 0.033 µmol
of CQ, leading to 0.033 µmol of caffeic acid and 0.033
µmol of Cy 3-glc o-Q. The latter then reacted readily
with 0.0165 µmol of CQ, leading to the formation of
several combinations with a global stoichiometry cor-
responding to [2 Cy 3-glc o-Q/1 caffeic acid o-Q]. To
explain our results, the following pathway of Cy 3-glc
degradation could be proposed (reactions 4-6). Such a
pathway has already been proposed by Kader et al.
)
484.85) from the caffeic acid (MW ) 180.2). As
expected from the fractional domain (1000-5000 Da),
Cy 3-glc and caffeic acid should be eluted in a single
peak. The chromatographic behavior of Cy 3-glc in the
Sephadex G-25 gel could be explained by the existence
of interactions (hydrogen bonds) between the pigment
and the gel. Consequently, calibrating the gel to esti-
mate the molecular weight of the different fractions
would appear to be difficult.
(1998).
2
Cy 3-glc + 2 CQ f
2
caffeic acid + 2 Cy 3-glc o-Q (4)
2
2
Cy 3-glc o-Q + 1 CQ f [2 Cy 3-glc o-Q/1 CQ] (5)
The fractions corresponding to peak C1 were pooled
and concentrated under vacuum in a rotary evaporator.
The UV-vis spectra were recorded between 220 and 600
nm (Figure 4). The spectrum of C1 showed a maximum
at 266 nm and a shoulder at 335-345 nm. The absorp-
tion at 335-345 nm could suggest the incorporation of
the caffeic moiety into the degradation products of the
pigment. The degradation products occurring in peaks
C1 were hydrolyzed by heating at 85 °C in 2 M
trifluoroacetic acid/methanol (v/v) for 45 min. The
fractions obtained after acidic hydrolysis were analyzed
by TLC (see Materials and Methods). After spraying the
developed chromatogram, we identified glucose, which
is linked at the C-3- position of the cyanidin. The overall
results suggest that the products of Cy 3-glc degradation
contain both caffeic acid and Cy 3-glc moieties. This is
consistent with previous papers on the oxidative deg-
radation of Cy 3-glc and Mv 3-glc in the presence of
caffeoyltartaric acid and grape PPO (Sarni-Manchado
et al., 1997; Sarni et al., 1995), which reported that the
degradation products of Cy 3-glc and Mv 3-glc contained
both caffeoyltartaric acid and anthocyanin moieties.
Moreover, these degradation products were gradually
Cy 3-glc + 3 CQ f
2
caffeic acid + [2 Cy 3-glc o-Q/1 CQ] (6)
The total amount of CQ that reacted was 0.0495 µmol
(
0.033 + 0.0165). This value was close to the initial
amount of CQ (0.047 µmol). The ratio of degraded Cy
-glc to the CQ integrated into the reaction products
3
was close to 2:1. The same ratio value has been reported
by Kader et al. (1998) and Sarni et al. (1995).
P a r tia l P u r ifica tion of th e Rea ction P r od u cts.
HPLC analysis did not detect the reaction products. This
result prompted us to develop a method to isolate the
condensation products arising from the reactions be-
tween anthocyanins and CQ to characterize them. For
this purpose, the reaction products were separated on
a Sephadex G-25 column eluted with distilled water
(
Figure 3). This gel filtration allowed the separation of
the condensation products from (i) the caffeic acid and
the remaining anthocyanin used and (ii) the salts of the
buffer solution. The reaction products obtained after 5
min of reaction between the Cy 3-glc and CQ were
chromatographed on a Sephadex G-25 column. Under

Degradation of Cyanidin 3-Glucoside by Caffeic Acid o-Quinone
J. Agric. Food Chem., Vol. 47, No. 11, 1999 4629
graded by a mechanism of coupled oxidation involving
the CGQ generated by PPO-oxidation of CG. This
reaction is accompanied by a partial regeneration of the
CG, which means that part of the CG is incorporated
into the degradation products of the Cy 3-glc. The
overall results suggest the formation of condensation
products with the following stoichiometry: [2 Cy 3-glc/1
CGQ].
The use of simpler model systems containing purified
anthocyanins and CQ has confirmed most of the results
described above. Indeed, Cy 3-glc reacts with the CQ
according to a mechanism of coupled oxidation leading
to the formation of combinations with the following
stoichiometry: [2 Cy 3-glc/1 CQ]. However, the exact
mechanism remains to be determined by characterizing
the structure of the degradation products. For this
reason, the next step in this work will involve both
isolating and determining the structure of the resolved
condensation products. The combination of gel filtration
and HPLC analysis (on the semipreparative scale) could
provide the condensation products in sufficient amounts
to allow their structural characterization. We are con-
vinced that chemical oxidations will prove to be a
valuable contribution to achieving this aim.
ABBREVIATIONS USED
CG, chlorogenic acid; CGQ, chlorogenoquinone; Cy
3-glc, cyanidin 3-glucoside; Cy 3-glc o-Q, cyanidin 3-glu-
coside o-quinone; Mv 3-glc, malvidin 3-glucoside; Pg
F igu r e 4. UV-vis spectrum of the concentrated fractions
corresponding to peak C1 obtained from the fractionation on
Sephadex G-25 of the degraded Cy 3-glc (C1). Spectra were in
distilled water.
3
-glc, pelargonidin 3-glucoside; CQ, caffeic acid o-
quinone (caffeic acid o-Q); PPO, polyphenol oxidase;
UV-vis, ultraviolet-visible.
LITERATURE CITED
Cheynier, V.; Moutounet, M. Oxidative reactions of caffeic acid
in model systems containing polyphenol oxidase. J . Agric.
Food Chem. 1992, 40, 2038-2044.
Cheynier, V.; Ricardo da Silva, J . M. Oxidation of grape
procyanidins in model solutions containing trans-caffeoyl-
tartaric acid and polyphenol oxidase. J . Agric. Food Chem.
1
991, 39, 1047-1049.
Davies, R. The synthesis and isolation of caffeoquinone and
caffeoquinone methyl ester. Tetrahedron Lett. 1976, 4, 313-
3
14.
Davies, R.; Pierpoint, W. S. Problem of the reactive species
from enzymatic and chemical oxidations of o-diphenols:
anomalies in the trapping of o-quinonoids with benzenesul-
phinic acid. Biochem. Soc. Trans. 1975, 3, 671-674.
Kader, F.; Rovel, B.; Girardin, M.; Metche, M. Fractionation
and identification of the phenolic compounds of Highbush
blueberries (Vaccinium corymbosum, L.). Food Chem. 1996,
5
5, 35-40.
F igu r e 5. HPLC elution profile at 280 nm of the degradation
Kader, F.; Rovel, B.; Girardin, M.; Metche, M. Mechanism of
browning in fresh Highbush blueberry fruit (Vaccinium
corymbosum, L.). Partial purification and characterization
of blueberry polyphenol oxidase. J . Sci. Food Agric. 1997a ,
products eluted in peak C1.
replaced by colorless products as a result of further
oxidative degradation.
7
3, 513-516.
HPLC analysis of fraction C1 (Figure 5) shows that
several products of degradation can be formed from the
reaction between Cy 3-glc and CQ, which supports the
idea of the complex degradation process mentioned by
Kader et al. (1998, 1999a).
A question raised by this work, and one that is to be
examined in future work, is how well chemical oxidation
mimics enzymatic oxidation. Recently, we studied the
mechanism of anthocyanin degradation in model solu-
tions containing CG, blueberry PPO, and purified an-
thocyanins (Cy 3-glc and Pg 3-glc). Kinetic studies have
shown that Cy 3-glc (o-diphenolic anthocyanin) is de-
Kader, F.; Rovel, B.; Girardin, M.; Metche, M. Mechanism of
browning in fresh Highbush blueberry fruit (Vaccinium
corymbosum, L.). Role of blueberry polyphenol oxidase,
chlorogenic acid and anthocyanins. J . Sci. Food Agric.
1
997b, 74, 31-34.
Kader, F.; Haluk, J . P.; Nicolas, J . P.; Metche, M. Degradation
of cyanidin 3-glucoside by blueberry polyphenoloxidase:
Kinetic studies and mechanisms. J . Agric. Food Chem. 1998,
4
6, 3060-3065.
Kader, F.; Nicolas, J . P.; Metche, M. Degradation of pelar-
gonidin 3-glucoside in the presence of chlorogenic acid and
blueberry polyphenol oxidase. J . Sci. Food Agric. 1999a , 79,
517-522.

4630 J. Agric. Food Chem., Vol. 47, No. 11, 1999
Kader et al.
Kader, F.; Irmouli, M.; Zitouni, N.; Nicolas, J . P.; Metche, M.
Degradation of pelargonidin 3-glucoside and cyanidin 3-glu-
coside by o-chloranil in model solutions. J . Agric. Food
Chem. 1999b, submitted for publication.
Sarni, P.; Fulcrand, H.; Souillol, V.; Souquet, J . M.; Cheynier,
V. Mechanisms of anthocyanin degradation in grape must-
like model solutions. J . Sci. Food Agric. 1995, 69, 385-391.
Sarni-Manchado, P.; Cheynier, V.; Moutounet, M. Reactions
of polyphenoloxidase generated caftaric acid o-quinone with
malvidin 3-O-glucoside. Phytochemistry 1997, 45, 1365-
1369.
Richard-Forget, F. C.; Rouet-Mayer, M. A.; Goupy, P. M.;
Philippon, J .; Nicolas, J . J . Oxidation of chlorogenic acid,
catechins, and 4-methylcatechol in model solutions by apple
polyphenol oxidase. J . Agric. Food Chem. 1992, 40, 2114-
2
122.
Received for review December 28, 1998. Revised manuscript
received August 12, 1999. Accepted August 31, 1999.
Rouet-Mayer, M. A.; Ralambosoa, J .; Philippon, J . J . Role of
o-quinones and their polymers in the enzymatic browning
of apples. Phytochemistry 1990, 29, 435-440.
J F981400X