Pathway leading to the formation of anthocyanin-vinylphenol adducts and related pigments in red wines

Article abstract of DOI:10.1021/jf0340963

On the basis of observations from Vitis vinifera cv. Pinotage wines and experiments performed in model wine medium, a new chemical pathway responsible for the formation of anthocyanin-vinylphenol adducts in red wines is described. Until now, these pigments have been considered to be reaction products of anthocyanins and vinylphenols, the latter being generated during fermentation by enzymatic decarboxylation of the respective cinnamic acids. The mechanism of the novel pathway, involving intact hydroxycinnamic acid and anthocyanin, is explained. Only cinnamic acids with electrondonating substituents on the aromatic ring, such as coumaric acid, ferulic acid, caffeic acid, and sinapic acid, undergo this conversion, as they stabilize an intermediately formed carbenium ion. Decarboxylation and oxidation of the pyran moieties are the final steps in the generation of the corresponding 4-vinylphenol, 4-vinylguaiacol, 4-vinylcatechol, and 4-vinylsyringol adducts of anthocyanins in red wine.

Full text of DOI:10.1021/jf0340963

3682 J. Agric. Food Chem. 2003, 51, 3682−3687
Pathway Leading to the Formation of Anthocyanin−Vinylphenol
Adducts and Related Pigments in Red Wines
MICHAEL SCHWARZ, TOBIAS C. WABNITZ,§ AND PETER WINTERHALTER*,†
†
Institute of Food Chemistry, Technical University of Braunschweig, Schleinitzstrasse 20,
38106 Braunschweig, Germany, and University Chemical Laboratory, Lensfield Road,
Cambridge CB2 1EW, United Kingdom
On the basis of observations from Vitis vinifera cv. Pinotage wines and experiments performed in
model wine medium, a new chemical pathway responsible for the formation of anthocyanin-
vinylphenol adducts in red wines is described. Until now, these pigments have been considered to
be reaction products of anthocyanins and vinylphenols, the latter being generated during fermentation
by enzymatic decarboxylation of the respective cinnamic acids. The mechanism of the novel pathway,
involving intact hydroxycinnamic acid and anthocyanin, is explained. Only cinnamic acids with electron-
donating substituents on the aromatic ring, such as coumaric acid, ferulic acid, caffeic acid, and
sinapic acid, undergo this conversion, as they stabilize an intermediately formed carbenium ion.
Decarboxylation and oxidation of the pyran moieties are the final steps in the generation of the
corresponding 4-vinylphenol, 4-vinylguaiacol, 4-vinylcatechol, and 4-vinylsyringol adducts of antho-
cyanins in red wine.
KEYWORDS: Anthocyanins; red wine; Vitis vinifera; malvidin 3-glucoside; vinylphenols; aging products;
pinotin A; hydroxycinnamic acids
INTRODUCTION
4-vinylsyringol derivatives, in red wines from Vitis Vinifera cv.
Shiraz and in grape skin extracts has been postulated on the
basis of data obtained by nano-electrospray tandem mass
spectrometric analysis (9). However, the tiny amounts of
pigments present in wine did not allow isolation and structural
characterization. Most recently, we were able to isolate pinotin
A from red wines made of Vitis Vinifera cv. Pinotage grapes
The most abundant red wine pigments are the 3-O-glucosides
of malvidin and peonidin, as well as their 6′′-acetylated and
coumaroylated derivatives. The 3-O-glucoconjugates of petuni-
din, delphinidin, and cyanidin are also widespread. During wine
aging, these pigments polymerize to a large extent and form a
heterogeneous and not well characterized group of compounds,
which is thought to be of major importance for the color of
aged wines (1). Anthocyanins and colorless flavanols such as
catechin and epicatechin form oligomeric pigments (2-4)s
possibly the starting point in the genesis of polymers. Multiple
other reactions take place during the aging of red wines, and
several colored low molecular reaction products have been
isolated and characterized in the past decade. Some of the aged
pigments bear an additional pyran ring between C-4 and the
hydroxyl group in position 5 of the aglycon moiety. Among
these, the vitisin-type pigments originate from the reaction of
malvidin 3-glucoside and acylated derivatives with pyruvic acid
(
10). The structure was fully characterized by HPLC-ESI-MSn
in combination with one-/two-dimensional NMR measurements
and determined to be the 4-vinylcatechol adduct of malvidin
3
-glucoside. Anthocyanin-vinylphenol-related adducts that have
been isolated and tentatively assigned are summarized in Figure
1
.
The formation of the malvidin 3-glucoside-4-vinylphenol
adduct is considered to be due to a direct reaction of malvidin
3-glucoside with 4-vinylphenol (Figure 2). Obviously, through
an enzymatic side activity of Saccharomyces cereVisiae, forma-
tion of 4-vinylphenol via enzymatic decarboxylation of p-
coumaric acid takes place during fermentation (8). It has been
suggested (9) that the formation of the 4-vinylcatechol, 4-vi-
nylguaiacol, and 4-vinylsyringol derivatives follows the same
mechanism with the respective vinylphenols emerging from
the enzymatic decarboxylation of caffeic, ferulic, and sinapic
acid.
(
4
5, 6). Another pyranoanthocyanin, the malvidin 3-glucoside-
-vinylphenol adduct, was first detected on membranes used
for cross-flow microfiltration of Carignane red wines (7) and
later isolated and structurally identified through comparison with
a synthetic sample (8). The occurrence of similar reaction
products, namely, the 4-vinylcatechol, 4-vinylguaiacol, and
The aim of the present study was to clarify whether the
proposed mechanism is the exclusive pathway for the formation
of anthocyanin-vinylphenol adducts and related pigments in
red wine.
*
Corresponding author (telephone ++49-531-3917200; fax ++49-531-
3
917230; e-mail P.Winterhalter@tu-bs.de).
†
Technical University of Braunschweig.
University Chemical Laboratory.
§
1
0.1021/jf0340963 CCC: $25.00 © 2003 American Chemical Society
Published on Web 05/02/2003

Formation of Anthocyanin−Vinylphenol Adducts
J. Agric. Food Chem., Vol. 51, No. 12, 2003 3683
tartrate in deionized water (adjusted to pH 3.2 with hydrochloric acid).
To 27 mL of the malvidin 3-glucoside solution (concentration ∼ 400
mg/L) was added 3 mL of an ethanolic solution of each of the following
trans-configured cinnamic acid derivatives: cinnamic acid, p-coumaric
acid, ferulic acid, caffeic acid, and sinapic acid (∼400 mg/L) as well
as 4-dimethylaminocinnamic acid, 4-nitrocinnamic acid, and caffeic
acid methyl ester (∼1200 mg/L). The winelike model solutions were
filled into 30 mL amber glass bottles, air in the headspace was replaced
with argon, and the solutions were stored in the dark at 15 °C in a
climatized room. Samples were analyzed in regular intervals by HPLC-
n
Figure 1. Structure of malvidin 3-glucoside adduct with 4-vinylphenol [R
H, R ) H (8)], 4-vinylcatechol [pinotin A; R ) OH, R ) H (10), 6,
cf. Figure 3], 4-vinylguaiacol [R ) OCH , R ) H(9)], and 4-vinylsyringol
) OCH , R ) OCH (9)].
1
DAD and HPLC-ESI-MS for 4 months. Relative reaction rates were
determined by HPLC-DAD analyses comparing the concentrations of
the newly formed pigments (expressed as pinotin A equivalents;
authentic reference standard used for calibration between 0 and 15 mg/
L, detection wavelength ) 510 nm).
)
2
1
2
1
3
2
[R
1
3
2
3
Influence of Oxygen on the Formation of Pinotin A. A winelike
model solution was prepared as described above containing malvidin
3
-glucoside (353 mg/L), caffeic acid (1326 mg/L), and 10% ethanol.
Thirty milliliters of each of the solutions was filled into four glass
bottles. The first bottle was ultrasonicated, and argon was bubbled
through the solution for 10 min to remove dissolved oxygen. Also, the
headspace was filled with argon. The second bottle was not degassed,
but the headspace was filled with argon. The third bottle was not
degassed nor was the air in the headspace replaced. The fourth solution
was saturated with oxygen by bubbling air through it for 10 min. Thus,
four solutions containing different amounts of oxygen were created.
The bottles were stored for 5 weeks and then analyzed by HPLC-DAD.
HPLC with Diode Array Detection (HPLC-DAD). A PU-980
Intelligent HPLC pump equipped with a DG-980-50 3-line degasser,
an LG-980-02 ternary gradient unit, and an MD-1510 multiwavelength
detector were used (Jasco). Samples were injected via a Rheodyne 7175
injection valve (Techlab) equipped with a 20 µL loop, and separations
were carried out on a Synergi MaxRP-11, 4 µm, 250 × 4.6 mm column
(Phenomenex). Solvents were water/formic acid/acetonitrile (87:10:3,
v/v/v, solvent A; 40:10:50, v/v/v, solvent B), and the flow rate was
0.5 mL/min. The linear gradient was from 6 to 20% B at 0-20 min,
from 20 to 40% B at 20-35 min, from 40 to 60% B at 35-40 min,
from 60 to 90% B at 40-45 min, and held at 90% B at 45-50 min.
HPLC with Electrospray Ionization Multiple Mass Spectrometry
Figure 2. Formation of the malvidin 3-glucoside−4-vinylphenol adduct
as proposed by Fulcrand et al. (8).
MATERIALS AND METHODS
Chemicals. All solvents were of HPLC quality and all chemicals
of p.a. grade.
Synthesis and Reactivity of 4-Vinylcatechol. 4-Vinylcatechol was
synthesized according to the method of B u¨ cking (11). 3,4-Dihydroxy-
benzaldehyde (3 g) and malonic acid (4 g) were dissolved in 50 mL of
ethanol. Six drops of piperidine were added, and the solution was
refluxed for 8 h. Five hundred milliliters of deionized water was added
and the solution extracted three times with 300 mL of diethyl ether.
The combined ether fractions were evaporated in vacuo. The residue
was redissolved in 5 mL of ethanol without further purification and
added to 50 mL of a solution of malvidin 3-glucoside (10 mg) in a
winelike model solution. This solution was analyzed by HPLC-DAD
immediately after the addition of 4-vinylcatechol and after stirring
overnight.
n
(HPLC-ESI-MS ). A Bruker Esquire LC-MS system was used (Bruker
Daltonik). The HPLC system consisted of a System 1100 binary pump
G1312A (Agilent) and a Rheodyne 7725i injection valve with a 20 µL
loop (Techlab). MS parameters were as follows: positive ion mode;
Synthesis of Caffeic Acid Methyl Ester. Caffeic acid (3 g) was
dissolved in 50 mL of methanol, a few drops of concentrated sulfuric
acid were added, and the solution was refluxed for 6 h. Three hundred
milliliters of deionized water was added, methanol was evaporated in
vacuo, and the solution was freeze-dried. Caffeic acid methyl ester was
purified by countercurrent chromatography. The solvent system was
n-hexane/ethyl acetate/methanol/water (3:5:3:5, v/v/v/v), and the upper
layer was used as stationary phase, flow rate ) 3.5 mL/min, detection
at 323 nm [chromatogram not shown, CCC apparatus as described
previously (12)]. Organic solvents were evaporated in vacuo from the
fraction containing the product, whereupon caffeic acid methyl ester
started to crystallize. For complete crystallization, the solution was
stored at 5 °C for 12 h. Crystals were removed by filtration, and the
remaining solution was lyophilized. Combined crystals were found to
be of >99% purity by HPLC-DAD and free from residual caffeic acid.
2
dry gas, N , 11 L/min, dry temperature, 325 °C; nebulizer, 60 psi;
capillary, -2500 V; capillary exit offset, 70 V; end plate offset, -500
V; skimmer 1, 20 V; skimmer 2, 10 V; scan range, m/z 50-1000,
chromatographic conditions as above.
n
Alternatively, for MS experiments the sample solution was delivered
directly by a syringe pump 74900 (Cole-Parmer) into the ESI source
at a flow rate of 240 µL/h. MS parameters were as follows: positive
ion mode; dry gas, N , 4.0 L/min; dry temperature, 300 °C; nebulizer,
2
10 psi; capillary, -3500 V; capillary exit offset, 60 V; end plate offset,
-500 V; skimmer 1, 30 V; skimmer 2, 10 V; scan range, m/z 50-
1000.
1H and 13C NMR spectra of caffeic acid methyl ester dissolved
NMR.
in methanol-d were recorded on a Bruker AMX 300 spectrometer
4
(Bruker Biospin) at 300.13 and 75.47 MHz, respectively.
Identity was confirmed by ESI-MS and NMR analyses: ESI-MS
RESULTS AND DISCUSSION
-
(
(
negative ion mode), m/z 193 [M - H] ; daughter ions (%), m/z 178
1
3
41), 161 (100), 134 (81); H NMR δ 3.75 (3H, s, OCH ), 6.25 (1H,
Up to now, the formation of anthocyanin-vinylphenol
adducts was considered to take place through the reaction of
anthocyanins with the respective 4-vinylphenols, which are
formed during fermentation by enzyme-mediated decarboxyla-
tion of the corresponding cinnamic acids (8, 9). In the course
of a study on the anthocyanin profile of red wines from V.
Vinifera cv. Pinotage, we observed that the concentration of
pinotin A (6, cf. Figure 3) was ∼10 times higher in aged wines
(5-6 years old) compared to that in very young (<1 year) wine
samples. This finding suggests that the reaction between
d, J ) 16.0 Hz, H-R), 6.78 (1H, d, J ) 8.1 Hz, H-5), 6.93 (1H, dd, J
8.1, 2.1 Hz, H-6), 7.03 (1H, d, J ) 2.1 Hz, H-2), 7.54 (1H, d, J )
)
1
3
1
1
6.0 Hz, H-â); C NMR δ 51.9 (OCH
3
), 114.9 (C-R), 115.2 (C-2),
16.5 (C-5), 122.9 (C-6), 127.8 (C-1), 146.8 (C-â), 146.9 (C-3), 149.5
(C-4), 169.8 (CdO).
Reactivity of Cinnamic Acid Derivatives. Malvidin 3-glucoside
of ∼90% purity (calculated from the 520 and 280 nm traces of a diode
array chromatogram) was isolated from various red wines by counter-
current chromatography according to methods previously described
(12-14) and dissolved in a 0.02 M solution of potassium hydrogen

3684 J. Agric. Food Chem., Vol. 51, No. 12, 2003
Schwarz et al.
Figure 3. Postulated pathway of pinotin A (6) formation involving caffeic acid (2a) and malvidin 3-glucoside 1. (A concerted cycloaddition is also
conceivable in the generation of 4; however, the stepwise addition process delineated here is preferred as the cycloaddition pathway would require 1
to react in an unfavorable keto form.)
4
-vinylcatechol and malvidin 3-glucoside proceeds rather slowly
of S. cereVisiae to decarboxylate cinnamic acids and found that
certain wine constituents, especially catechin, epicatechin, and
oligomeric procyanidins, but not anthocyanins and polymeric
constituents (i.e., polymeric procyanidins or other highly
condensed macromolecules with a molecular weight >3000),
strongly inhibited the decarboxylation of p-coumaric acid. The
concentration of these inhibitors is much higher in red wines.
Hence, it can been concluded that the cinnamate carboxy-lyase
is largely inactive during red wine fermentation and that the
enzyme-mediated synthesis of 4-vinylphenols in red wine is not
significant with regard to anthocyanin-vinylphenol adduct
formation. The authors also discovered that all of the yeast
strains investigated were able to decarboxylate p-coumaric and
ferulic acid, but none of them was capable of transforming either
caffeic or sinapic or unsubstituted cinnamic acid into 4-vinyl-
catechol, 4-vinylsyringol, or styrene, respectively.
and requires years of storage to complete. To verify the slow
reaction rate, model experiments with 4-vinylcatechol and
malvidin 3-glucoside were carried out. As 4-vinylcatechol is
not commercially available, a synthesis was performed according
to the method of B u¨ cking (11). Synthesized 4-vinylcatechol and
malvidin 3-glucoside were reacted in a model wine medium.
n
Subsequent HPLC-DAD and HPLC-ESI-MS analyses revealed
that malvidin 3-glucoside had been almost quantitatively reacted
to a new pigment, the retention time and mass spectral data of
which were identical to those of an authentic sample of pinotin
A (10). It is important to note that our experiment is in line
with earlier observations of Fulcrand et al. (8) and Sarni-
Manchado et al. (15), who reported that the reaction of malvidin
3-glucoside and 4-vinylphenol went to completion within hours.
This observation prompted us to search for another explana-
tion for the increase of pinotin A during storage as 4-vinylcat-
echol, if present at all, would be consumed rapidly and the
gradual formation of 6 over prolonged periods could not be
rationalized in this way. Our suspicion was strengthened by the
fact that despite a comprehensive literature search, we were
unable to find evidence that either 4-vinylcatechol or 4-vinyl-
syringol had ever been detected in red or white wines, whereas
the presence of 4-vinylphenol and 4-vinylguaiacol is well
documented (16-20). Although red wines contain much more
of the precursors coumaric and ferulic acid, the amount of the
corresponding volatile vinylphenols in white wines was deter-
mined to be orders of magnitude higher and beyond a certain
limit responsible for unpleasant “phenolic” off-flavors. A
possible explanation for this phenomenon could be seen in the
reaction of anthocyanins with vinylphenols (8). However,
Chatonnet et al. (18) investigated the ability of different strains
The decisive hint of a novel pathway of anthocyanin-
vinylphenol adduct formation was finally given by the unusually
high concentrations of caffeic acid in Pinotage wines. The
caffeic acid content in most of the common red wine varieties
does not exceed 10 mg/L (21-23). Only in Pinotage wine were
we able to detect as much as 77 mg/L of caffeic acid with a
mean value of ∼35 mg/L. Hereupon we started to consider the
possibility of a direct reaction between malvidin 3-glucoside
and caffeic acid or anthocyanins and cinnamic acids in general.
Hence, another experiment was performed with malvidin
3-glucoside and caffeic acid in a winelike model solution. This
solution was stored in the dark at 15 °C and analyzed at regular
intervals. Just after 5 days we were able to detect the formation
of a new pigment with the same retention time and mass spectral
properties as pinotin A. The concentration of this peak constantly

Formation of Anthocyanin−Vinylphenol Adducts
J. Agric. Food Chem., Vol. 51, No. 12, 2003 3685
Figure 4. Reactivity of cinnamic acid derivatives with regard to anthocyanin
adduct formation with 1.
Table 1. Mass Spectral Properties of the Generated Malvidin
3-Glucoside Cinnamic Acid Derivatives
molecular
+
malvidin 3-glucoside
adduct
ion [M ]
aglycon
(m/z)
elimination masses
from aglycon (u)
(m/z)
4
4
4
4
-vinylphenol
-vinylcatechol (pinotin A)
-vinylguaiacol
-vinylsyringol
-dimethylaminostyrene
609
625
639
669
636
447
463
477
507
474
16, 32, 44, 61, 89
16, 32, 44, 61, 89
16, 32, 44, 61, 89
16, 32, 44, 58, 61, 89
16, 33, 44, 61, 89
4
increased during the following months as shown by HPLC-DAD
analyses.
Figure 5. Course of the reaction of malvidin 3-glucoside with coumaric
acid (2e, [), ferulic acid (2c, b), sinapic acid (2d, 2), and caffeic acid
These results showed for the first time that pyranoanthocya-
nins can be formed directly in a reaction between intact cinnamic
acid derivatives and anthocyanin. To obtain a mechanistic
rationale for this unprecedented conversion, further experiments
were conducted, including investigations on the reactivity of
other cinnamic acid derivatives and the possible influence of
oxygen on the reaction rate. On the basis of our findings the
following reaction sequence can be suggested (Figure 3).
The initial bond formation between the C-4 position of
malvidin 3-glucoside (1) and the C-2 position of caffeic acid
(
2a, 9).
Figure 6. Mechanism of the acid-catalyzed Prins cyclization.
Negative Hammett reaction parameters F have been obtained,
for example, in the bromate oxidation of cinnamic acids (26-
29). These data confirm that electrophilic attack on cinnamic
acids as shown here can be favored over nucleophilic pathways
such as conjugate addition to the double bond, which can be
accelerated by electron-withdrawing substituents (30).
The intermediate carbenium ion 3 can be trapped intramo-
lecularly by the phenolic hydroxy group of 1 to form the pyran
ring in 4 (cf. Figure 3). Additional support for this cyclization
mechanism is drawn from the striking similarity to the exten-
sively studied acid-catalyzed addition of formaldehyde to
alkenes, the Prins reaction, where the six-membered intermediate
7 has been postulated (31) (Figure 6).
(2a) is in line with the strongly electrophilic nature of the
benzopyrylium unit and the nucleophilicity of the R-carbon atom
of acid 2a. Given the electron-deficient character of the resulting
intermediate 3, it can be expected that electron-donating
substituents on the aromatic ring of the cinnamic acid moiety
facilitate this reaction due to stabilization of the intermediate
carbenium ion 3. This has indeed been observed in a series of
conversion experiments. Whereas p-coumaric acid (2e), ferulic
acid (2c), sinapic acid (2d), and 4-dimethylaminocinnamic acid
(2b) could be successfully reacted, no adduct formation at all
was observed with 4-nitrocinnamic acid (2g) or the parent
compound cinnamic acid (2f) (Figure 4).
Mass spectral properties of the newly formed anthocyanin
derivatives and their aglycons were in line with data published
for the isolated 4-vinylphenol and 4-vinylcatechol derivatives
of malvidin 3-glucoside (8, 10) and the tentatively identified
Subsequently, the final product 6 is formed via oxidation and
decarboxylation of the intermediate 4. Few details are known
about the precise mechanism of pyran oxidation involving
reactive benzylic carbon-hydrogen bonds under the conditions
used here, but these conversions can generally be achieved with
remarkable ease (32, 33). Hydride abstractions (e.g., by unre-
acted 1) (34-36), enzymatic/coenzymatic oxidations (37, 38)
and involvement of other known oxidants such as nitrite (39)
are generally conceivable in authentic wine samples. Radical
pathways (autoxidation) in the presence of oxygen and a
catalytic species (40) are less likely due to the antioxidative
properties of phenolic compounds. Correspondingly, the rate
of our model reaction exhibited no dependence on the presence
of oxygen. The amount of neither malvidin 3-glucoside (mean
( 1.4%) nor pinotin A (mean ( 4.3%) was significantly
different in any of the solutions containing different amounts
of oxygen, thus making hydride abstraction the most likely
pathway under the conditions employed.
4
-vinylguaiacol and 4-vinylsyringol adducts (9). In addition,
upon further fragmentation of the aglycons the same elimination
masses were observed as by Hayasaka and Asenstorfer, who
found this fragmentation pathway to be typical for malvidin-
derived pigments (9). The p-dimethylaminocinnamic acid
product has not been reported before (Table 1).
The kinetics of reaction of the di- and trisubstituted acids
2a, 2c, and 2d with anthocyanins were moderately enhanced
compared to that of 2e (Figures 4 and 5). With the amino moiety
in 2b, however, the rate of formation was accelerated by a factor
>
100 (data not included in Figure 5). These observed reac-
tivities coincide with the expected order of reaction rates for
para-substituent effects when electron-deficient transition states
+
are involved, as given by Brown’s σp constants (24, 25).
It is known that electron-donating substituents can accelerate
reactions involving the olefinic double bond of cinnamates.
There is also ample evidence for oxidative and nonoxidative
decarboxylations in the vicinity of a variety of different oxygen

3686 J. Agric. Food Chem., Vol. 51, No. 12, 2003
Schwarz et al.
substituents, in both biochemical and synthetic processes (37,
Research is currently underway to explore the suitability of
pinotin A as an aging indicator for Pinotage wines.
3
9, 41, 42), and the strongly accelerating effect of â-alkoxy
groups has been explained in terms of the polar nature of the
transition state (43). In fact, a facile thermal decarboxylation
of a phenyl pyranone-based acid has been noted more than 100
years ago (44). The presence of a free acid functionality is
essential for decarboxylation; consequently, only traces of the
LITERATURE CITED
(
1) Somers, T. C. The polymeric nature of wine pigments. Phy-
tochemistry 1971, 10, 2175-2186.
(
2) Francia-Aricha, E. M.; Guerra, M. T.; Rivas-Gonzalo, J. C.;
Santos-Buelga, C. New anthocyanin pigments formed after
condensation with flavanols. J. Agric. Food Chem. 1997, 45,
n
product 6 could be detected upon analysis by HPLC-ESI-MS
when caffeic acid methyl ester (2h) was employed as a substrate
(Figure 4). On the basis of these data, oxidative generation of
2
262-2266.
a stabilized oxonium intermediate such as 5 can be proposed.
Decarboxylation of this intermediate is expected to occur under
mild conditions, and further oxidation of the pyran moieties to
the aromatic heterocycles results in the final product 6.
(
(
(
3) Rivas-Gonzalo, J. C.; Bravo-Haro, S.; Santos-Buelga, C. Detec-
tion of compounds formed through the reaction of malvidin
3-monoglucoside and catechin in the presence of acetaldehyde.
J. Agric. Food Chem. 1995, 43, 1444-1449.
4) Mateus, N.; Silva, A. M. S.; Santos-Buelga, C.; Rivas-Gonzalo,
J. C.; de Freitas, V. Identification of anthocyanin-flavanol
pigments in red wines by NMR and mass spectrometry. J. Agric.
Food Chem. 2002, 50, 2110-2116.
5) Bakker, J.; Bridle, P.; Honda, T.; Kuwano, H.; Saito, N.;
Terahara, N.; Timberlake, C. F. Identification of an anthocyanin
occurring in some red wines. Phytochemistry 1997, 44, 1375-
1382.
The direct reaction between anthocyanins and hydroxycin-
namic acids readily explains the formation of all anthocyanin-
vinylphenol-type adducts in red wines. At the time of writing,
this is the only experimentally verified mechanism leading to
the development of 4-vinylcatechol and 4-vinylsyringol pig-
ments, as the free vinylphenols have neither been detected in
wines nor was it possible to generate these compounds via
enzymatic decarboxylation using yeasts commonly applied to
red wine fermentation (18). Small amounts of 4-vinylphenol
and 4-vinylguaiacol were detected in experimental Shiraz red
wines only if the grape juice was treated with pectic enzymes
possessing cinnamoyl esterase activity prior to fermentation, thus
increasing the amount of free coumaric and ferulic acids (20).
However, even in this case the concentrations of the respective
vinylphenols were in the low microgram range, and they would
have to react quantitatively with malvidin 3-glucoside to
generate detectable amounts of the derived pigments. Given the
high reactivity of vinylphenols toward other constituents of
young red wine, including different anthocyanins, it is extremely
unlikely that high levels of anthocyanin-vinylphenol adducts
can arise in this way.
(6) Fulcrand, H.; Benabdeljalil, C.; Rigaud, J.; Cheynier, V.;
Moutounet, M. A new class of wine pigments generated by
reaction between pyruvic acid and grape anthocyanins. Phy-
tochemistry 1998, 47, 1401-1407.
(
(
(
7) Cameira dos Santos, P.-J.; Brillouet, J.-M.; Cheynier, V.;
Moutounet, M. Detection and partial characterisation of new
anthocyanin-derived pigments in wine. J. Sci. Food Agric. 1996,
7
0, 204-208.
8) Fulcrand, H.; Cameira dos Santos, P.-J.; Sarni-Manchado, P.;
Cheynier, V.; Favre-Bonvin, J. Structure of new anthocyanin-
derived wine pigments. J. Chem. Soc., Perkin Trans. 1 1996,
735-739.
9) Hayasaka, Y.; Asenstorfer, R. E. Screening for potential pigments
derived from anthocyanins in red wine using nanoelectrospray
tandem mass spectrometry. J. Agric. Food Chem. 2002, 50, 756-
7
61.
Therefore, the pathway starting from free hydroxycinnamic
acids is with utmost certainty responsible for the formation of
the vast majority of anthocyanin-vinylphenol pigments in red
wine. The reaction rate in wines is expected to be slower
compared to our model solutions: on the one hand because of
the lower concentration of the reactants, and on the other hand
because of competing reactions of malvidin 3-glucoside with
other wine constituents, leading to the formation of, for example,
vitisin A or anthocyanin-flavanol condensation products. This
makes the vinylphenol-derived pigments potentially attractive
for use as aging indicators for red wines as their concentration
will constantly increase during storage as long as free antho-
cyanins and cinnamic acids are available. During wine storage,
the latter can be constantly replenished through slow hydrolysis
of the corresponding tartaric esters, which are normally present
in higher concentrations than the free acids (45, 46). The fact
that unsubstituted cinnamic acid, also present in red wines, can
neither directly react with anthocyanins nor be decarboxylated
to styrene by wine yeast explains why the occurrence of such
a pigment has not been reported.
(
10) Schwarz, M.; Jerz, G.; Winterhalter, P. Isolation and structure
of Pinotin A, a new anthocyanin derivative from Pinotage wine.
Vitis 2003, in press.
(11) B u¨ cking, W. U¨ ber die Synthese der Kaffees a¨ ure und ihrer
Isomeren, der Dioxybenzalmalons a¨ ure und des Kaffeealdehyds.
Thesis, University of Leipzig, 1928.
(
12) Schwarz, M.; Hillebrand, S.; Habben, S.; Degenhardt, A.;
Winterhalter, P. Application of high-speed countercurrent chro-
matography to the large-scale isolation of anthocyanins. Biochem.
Eng. J. 2003, 14, 179-189.
(
13) Degenhardt, A.; Knapp, H.; Winterhalter, P. Rapid isolation of
malvidin 3-glucoside from red wine by high-speed countercurrent
chromatography (HSCCC). Vitis 2000, 39, 43-44.
(14) Degenhardt, A.; Hofmann, S.; Knapp, H.; Winterhalter, P.
Preparative isolation of anthocyanins by high-speed countercur-
rent chromatography and application of the color activity concept
to red wine. J. Agric. Food Chem. 2000, 48, 5812-5818.
(
15) Sarni-Manchado, P.; Fulcrand, H.; Souquet, J.-M.; Cheynier, V.;
Moutounet, M. Stability and color of unreported wine antho-
cyanin-derived pigments. J. Food Sci. 1996, 61, 938-941.
16) Etievant, P. X. Volatile phenol determination in wine. J. Agric.
Food Chem. 1981, 29, 65-67.
(
Both 4-vinylguaiacol and 4-vinylcatechol adducts of other
anthocyanins have most likely been generated in several studies
on copigmentation in red wines (47, 48) and model solutions
(17) Baumes, R.; Cordonnier, R.; Nitz, S.; Drawert, F. Identification
and determination of volatile constituents in wines from different
vine cultivars. J. Sci. Food Agric. 1986, 37, 927-943.
(
18) Chatonnet, P.; Dubourdieu, D.; Boidron, J.-N.; Lavigne, V.
Synthesis of volatile phenols by Saccharomyces cereVisiae in
wines. J. Sci. Food Agric. 1993, 62, 191-202.
(49). However, the stabilization of color was solely interpreted
by assuming the formation of non-covalent anthocyanin-
copigment complexes.
(
19) Lao, C. L.; L o´ pez-Tamames, E.; Lamuela-Ravent o´ s, R. M.;
Buxaderas, S.; de la Torre-Boronat, M. d. C. Pectic enzyme
treatment effects on quality of white grape musts and wines. J.
Food Sci. 1997, 62, 1142-1144, 1149.
Due to the pure chemical nature of the novel pathway without
the need for enzymatic support, the formation of anthocyanin-
vinylphenol adducts can take place during years of storage.

Formation of Anthocyanin−Vinylphenol Adducts
J. Agric. Food Chem., Vol. 51, No. 12, 2003 3687
(
(
(
(
(
20) Dugelay, I.; Gunata, Z.; Sapis, J.-C.; Baumes, R.; Bayonove, C.
Role of cinnamoyl esterase activities from enzyme preparations
on the formation of volatile phenols during winemaking. J. Agric.
Food Chem. 1993, 41, 2092-2096.
21) Rodr ´ı guez-Delgado, M. A.; Gonz a´ lez-Hern a´ ndez, G.; Conde-
Gonz a´ lez, J. E.; P e´ rez-Trujillo, J. P. Principal component analysis
of the polyphenol content in young red wines. Food Chem. 2002,
(37) Ohshiro, H.; Mitsui, K.; Ando, N.; Ohsawa, Y.; Koinuma, W.;
Takahashi, H.; Kondo, S.; Nabeshima, T.; Yano, Y. Oxidation-
active flavin models: Oxidation of R-hydroxy acids by benzo-
dipteridine bearing metal-binding site in the presence of divalent
metal ion and base in organic solvents. J. Am. Chem. Soc. 2001,
1
23, 2478-2486.
(
38) Fukuzumi, S.; Itoh, S.; Komori, T.; Suenobu, T.; Ishida, A.;
Fujitsuka, M.; Ito, O. Photochemical reactions of coenzyme PQQ
7
8, 523-532.
22) Soleas, G. J.; Dam, J.; Carey, M.; Goldberg, D. M. Toward the
fingerprinting of wines: cultivar-related patterns of polyphenolic
constituents in Ontario wines. J. Agric. Food Chem. 1997, 45,
(
pyrroloquinolinequinone) and analogues with benzyl alcohol
derivatives via photoinduced electron transfer. J. Am. Chem. Soc.
2000, 122, 8435-8443.
3
871-3880.
23) Landrault, N.; Poucheret, P.; Ravel, P.; Gasc, F.; Cros, G.;
Teissedre, P. L. Antioxidant capacities and phenolics levels of
French wines from different varieties and vintages. J. Agric. Food
Chem. 2001, 49, 3341-3348.
(39) Napolitano, A.; d’Ischia, M. New insights into the acid-promoted
reaction of caffeic acid and its esters with nitrite: Decarboxy-
lation drives chain nitrosation pathways toward novel oxime
derivatives and oxidation/fragmentation products thereof. J. Org.
Chem. 2002, 67, 803-810.
24) Okamoto, Y.; Inukai, T.; Brown, H. C. Rates of solvolysis of
phenyldimethylcarbinyl chlorides containing meta directing
substituents. J. Am. Chem. Soc. 1958, 80, 4969-4972.
25) Brown, H. C.; Okamoto, Y. Electrophilic substituent constants.
J. Am. Chem. Soc. 1958, 80, 4979-4987.
(
40) Ishii, Y.; Sakaguchi, S.; Iwahama, T. Innovation of hydrocarbon
oxidation with molecular oxygen and related reactions. AdV.
Synth. Catal. 2001, 343, 393-427.
(
(
26) Reddy, C. S.; Sundaram, E. V. Mechanism of acid bromate
oxidation of trans mono-substituted cinnamic acids: Structure
reactivity correlationsa non-linear Hammett’s plot. Tetrahedron
(41) Cleland, W. W. Mechanisms of enzymatic oxidative decarbox-
ylation. Acc. Chem. Res. 1999, 32, 862-868.
(42) Blay, G.; Fernandez, B.; Formentin, P.; Pedro, J. R.; Rosello,
A. L.; Ruiz, R.; Journaux, Y. Catalytic aerobic oxidative
decarboxylation of R-hydroxy-acids. Methyl mandelate as a
benzoyl anion equivalent. Tetrahedron Lett. 1998, 39, 3327-
1
989, 45, 2109-2126.
(
27) Sabapathy Mohan, R. T.; Gopalakrishnan, M.; Sekar, M. Kinetics
and mechanism of Os(VIII)-catalysed oxidation of some sub-
stituted trans-cinnamic acids by chloramine-T in alkaline mediums
a non-linear Hammett plot. Tetrahedron 1994, 50, 10945-10954.
28) Aruna, K.; Manikyamba, P. Substituent effect in the oxidation
of cinnamic acids by quinolinium dichromate. Indian J. Chem.,
Sect. A 1995, 34, 822-825.
3
330.
(
43) Al-Borno, A.; Bigley, D. B. Studies in decarboxylation. Part 15.
The effect of 3-substitution on the rate of decarboxylation of
âø-unsaturated acids. J. Chem. Soc., Perkin Trans. 2 1982, 15-
(
(
1
7.
29) Lee, D. G.; Brown, K. C. Oxidation of hydrocarbons. 11. Kinetics
and mechanism of the reaction between methyl (E)-cinnamate
and quaternary ammonium permanganates. J. Am. Chem. Soc.
(44) Fichter, F.; Bauer, A. U¨ ber die Phenyl-øδ-pentens a¨ ure. Chem.
Ber. 1898, 31, 2001-2004.
(45) Somers, T. C.; V e´ rette, E.; Pocock, K. F. Hydroxycinnamate
esters of Vitis Vinifera: changes during white vinification, and
effects of exogenous enzymic hydrolysis. J. Sci. Food Agric.
1
982, 104, 5076-5081.
(
30) Ogino, T.; Watanabe, T.; Matsuura, M.; Watanabe, C.; Ozaki,
H. Semiquantitative FMO analysis of substituent effect on the
reaction of permanganate ion with unsymmetrical alkenes. J. Org.
Chem. 1998, 63, 2627-2633.
1
987, 40, 67-78.
(
46) Karagiannis, S.; Economou, A.; Lanaridis, P. Phenolic and
volatile composition of wines made from Vitis Vinifera cv.
Muscat Lefko grapes from the island of Samos. J. Agric. Food
Chem. 2000, 48, 5369-5375.
(
31) Smissman, E.; Schnettler, R. A.; Portoghese, P. S. Mechanism
of the Prins reaction. Stereoaspects of the formation of 1,3-
dioxanes. J. Org. Chem. 1965, 30, 797-801.
(
32) Roehri-Stoeckel, C.; Gonzalez, E.; Fougerousse, A.; Brouillard,
R. Synthetic dyes: Simple and original ways to 4-substituted
flavylium salts and their corresponding vitisin derivatives. Can.
J. Chem. 2001, 79, 1173-1178.
(47) Darias-Mart ´ı n, J.; Carrillo, M.; D ´ı az, E.; Boulton, R. B. Enhance-
ment of red wine colour by pre-fermentation addition of
copigments. Food Chem. 2001, 73, 217-220.
(
33) Ishii, Y.; Nakayama, K.; Takeno, M.; Sakaguchi, S.; Iwahama,
T.; Nishiyama, Y. A novel catalysis of N-hydroxyphthalimide
in the oxidation of organic substrates by molecular oxygen. J.
Org. Chem. 1995, 60, 3934-3935.
(
48) Darias-Mart ´ı n, J.; Mart ´ı n-Luis, B.; Carrillo-L o´ pez, M.; Lamuela-
Ravent o´ s, R.; D ´ı az-Romero, C.; Boulton, R. Effect of caffeic
acid on the color of red wine. J. Agric. Food Chem. 2002, 50,
2
062-2067.
(
34) Fernandez, I.; Pedro, J. R.; Rosello, A. L.; Ruiz, R.; Castro, I.;
Ottenwaelder, X.; Journaux, Y. Alcohol oxidation by dioxygen
and aldehydes catalysed by square-planar cobalt(III) complexes
of disubstituted oxamides and related ligands. Eur. J. Org. Chem.
(
49) Eiro, M. J.; Heinonen, M. Anthocyanin color behavior and
stability during storage: effect of intermolecular copigmentation.
J. Agric. Food Chem. 2002, 50, 7461-7466.
2
001, 1235-1247.
(
35) Sakai, A.; Hendrickson, D. G.; Hendrickson, W. H. Mechanism
of the oxidation of para-substituted 1-phenylethanols with sodium
hypochlorite in acetic acid. Tetrahedron Lett. 2000, 41, 2759-
Received for review January 29, 2003. Revised manuscript received
April 1, 2003. Accepted April 6, 2003. This project was partially
financed by the German Ministry of Economy via the Arbeitskreis
Industrieller Forschungsvereinigungen/Forschungskreis der Ern a1 h-
rungsindustrie (AIF- FV: 12896N).
2
763.
(
36) Nikalje, M. D.; Sudalai, A. Catalytic selective oxidation of alkyl
arenes to aryl tert-butyl peroxides with TBHQ over Ru-
exchanged montmorillonite K10. Tetrahedron 1999, 55, 5903-
5
908.
JF0340963