Formation of pigment precursor (+)-1″-methylene-6″-hydroxy2H- furan-5″-one-catechin isomers from (+)-catechin and a degradation product of ascorbic acid in a model wine system

Article abstract of DOI:10.1021/jf902198e

The present study investigates the contribution of ascorbic acid to the formation of pigment precursors in model white wine systems containing (+)-catechin as the oxidizable phenolic substrate. The two main colorless products in these systems were structu

Full text of DOI:10.1021/jf902198e

J. Agric. Food Chem. 2009, 57, 9539–9546 9539
DOI:10.1021/jf902198e
Formation of Pigment Precursor (þ)-100-Methylene-600-hydroxy-
2H-furan-500-one-catechin Isomers from (þ)-Catechin and a
Degradation Product of Ascorbic Acid in a Model Wine System
§
,
CE´ LIA
B
ARRIL,*,† ANDREW C. CLARK,† PAUL D. PRENZLER,† PETER
#
K
ARUSO
AND
G
EOFFREY R. SCOLLARY
†National Wine and Grape Industry Centre, Charles Sturt University, Locked Bag 588, Wagga Wagga,
NSW 2678, Australia, §Department of Chemistry and Biomolecular Sciences, Macquarie University,
Sydney, NSW 2109, Australia, and #School of Chemistry, The University of Melbourne, Parkville,
VIC 3010, Australia
The present study investigates the contribution of ascorbic acid to the formation of pigment precursors in
model white wine systems containing (þ)-catechin as the oxidizable phenolic substrate. The two main
colorless products in these systems were structurally characterized as isomers of (þ)-catechin
substituted at either C6 or C8 on the A ring with a furan-type unit, namely, (þ)-100-methylene-600-
hydroxy-2H-furan-500-one-6-catechin and (þ)-100-methylene-600-hydroxy-2H-furan-500-one-8-catechin.
A known degradation product of ascorbic acid, L-xylosone, was separately prepared and, when reacted
with (þ)-catechin, generated the same (þ)-furanone-catechin isomers as in model white wine systems.
Incubation of these isomers in wine-like conditions yielded yellow xanthylium cation pigments. This study
has shown that undesirable spoilage reactions (yellow coloration) can occur from a breakdown product
of ascorbic acid-L-xylosone.
KEYWORDS: Ascorbic acid; (þ)-catechin; L-xylosone; model white wine; oxidation; xanthylium cation;
biomimetic synthesis
INTRODUCTION
in the wine production process. The levels of ascorbic acid added
may vary considerably, but it is typically added at rates ranging
from 50 to 150 mg/L (9-12).
Flavanols in wine, such as (þ)-catechin (Figure 1), are well-
known to react with aldehydes (1) formed or extracted during the
winemaking process and aging. The most commonly occurring
aldehyde in wine production is acetaldehyde, formed mainly by
yeast during fermentation or by the oxidation of ethanol (2).
Other compounds formed in wine that also contain an aldehyde
group include glyceraldehyde, glyoxylic acid, and furfural. The
first two may be generated under oxidative storage conditions of
wine (3-5), whereas furfural can be extracted during the aging of
wine in oak (6). Of these aldehydes, only glyoxylic acid (7) and
furfural (6) have been shown to participate in xanthylium cation
pigment formation after their reaction with (þ)-catechin in model
systems. However, pigments containing glyoxylic acid have been
reported for red wine (8) but have not yet been observed in white
wine.
Ascorbic acid is added to white wine due to its ability to
scavenge molecular oxygen efficiently, but in the process it is
initially converted to dehydroascorbic acid and hydrogen per-
oxide. The dehydroascorbic acid then undergoes rapid degrada-
tion into a variety of species including numerous carboxylic acid,
ketone, and aldehyde compounds (13-16) (Figure 1). Recent
studies have shown the increased production of phenolic pig-
ments from model wine systems of ascorbic acid and (þ)-catechin
and demonstrated that a degradation product emanating from
ascorbic acid was able to react with (þ)-catechin and form
colored xanthylium cations (17, 18). In general, phenolic pig-
ments formed during the storage of white wine are undesirable as
they are perceived as indicative of wine spoilage.
The behavior of (þ)-catechin solutions in the presence of
relatively high concentrations of ascorbic acid has been the focus
of a recent study by Clark et al. (19). In this study, two colorless
species were observed using HPLC-PDA in model wine samples
containing iron(II), ascorbic acid, and (þ)-catechin. The com-
pounds accumulated while ascorbic acid was still present in
solution and then underwent rapid degradation once the ascorbic
acid was near depletion. Their degradation also coincided with
the production of pigmented compounds in the model wine
system. These colorless compounds thus appear to be key inter-
mediates in the formation of pigments in model white wine
systems. Despite intense effort, these compounds have until
Another wine constituent capable of producing electrophilic
carbonyl compounds, despite its primarily antioxidant proper-
ties, is ascorbic acid (Figure 1). Ascorbic acid is naturally present
in grapes but is usually rapidly consumed after crushing, typically
due to ascorbic acid either scavenging oxygen or reducing
o-quinone compounds formed from the enzymatic oxidation of
phenolic compounds (9, 10). The ascorbic acid present in white
wine is mostly due to exogenous additions, often just before the
bottling of white wine, although it may be used at various stages
*Corresponding author (telephone 612-6933-2158; fax 612-6933-
2107; e-mail cbarril@csu.edu.au).
© 2009 American Chemical Society
Published on Web 10/05/2009
pubs.acs.org/JAFC

9540 J. Agric. Food Chem., Vol. 57, No. 20, 2009
Barril et al.
Figure 1. Molecular structures of (þ)-catechin, ascorbic acid, and some reported degradation products of ascorbic acid.
now eluded identification. Furthermore, direct evidence that
these compounds are indeed intermediates in pigment formation
has not been available. This study addresses these two issues by
identifying the colorless compounds as (þ)-100-methylene-600-
hydroxy-2H-furan-500-one-6-catechin and (þ)-100-methylene-600-
hydroxy-2H-furan-500-one-8-catechin and showing that these
compounds react to give pigments. The study goes further
to directly implicate a breakdown product of ascorbic acid,
0 min; 100%, 1 min; 95%, 1.31 min; 62%, 5.25 min; 56%, 6.27 min; 48%,
6.34 min; 45%, 7.22 min; 0%, 8.85 min; 0%, 9.40 min; 100%, 8.74 min;
and 100%, 9.76 min.
LC-MS studies on the colorless compounds were conducted as
described by Lutter et al. (20). LC-MS studies conducted on the xanthy-
lium cations were conducted on an Agilent 1200 series triple-quadrapole
(6410) HPLC-MS. The column and LC conditions were as described for
the UPLC (above), except for an injection volume of 20 μL. The MS was
operated with drying gas temperature at 350 °C, gas flow of 9 L/min,
nebulizer pressure at 40 psi, and capillary voltage at 4 kV. MS analyses for
the xanthylium cations were carried out in the positive ion mode with the
fragmentor at both 50 and 150 V, the former providing parent ion signals
and the latter inducing fragmentation.
Preparative HPLC separations were conducted on a Perkin-Elmer 250
binary LC pump connected to a Varian 320 Pro Star UV-visible detector
(280 nm) controlled by a Varian Star (v 6.41) chromatography work-
station. The column was a semipreparative 4 μm Phenomenex Synergy
Hydro-RP C18 column (250 ꢀ 10 mm) coupled to a Phenomenex ODS
Octadecyl C18 (10 ꢀ 10 mm) security guard column. Injection volume was
2 mL and flow rate 2 mL/min; the elution gradient consisted of solvent A,
2% formic acid in water, and solvent B, 2% formic acid in 80%
acetonitrile, as follows (expressed in solvent A followed by cumulative
time): 100%, 0 min; 85%, 5 min; 71%, 35 min; 0%, 40 min; 0%, 50 min;
100%, 55 min; 100%, 60 min. Lyophilization was performed on a Christ-
Alpha 2-4D freeze-dryer (Biotech International, Germany).
L
-xylosone, as capable of reacting with (þ)-catechin to generate
the two (þ)-furan-catechin isomers observed in model wine
systems. Finally, the pigments formed by reaction of the
(þ)-furanone-catechin isomers were shown to be xanthylium
cations, which have been reported in other model wine systems.
MATERIALS AND METHODS
General. Water purified through a Milli-Q (Millipore) water system
(ISO 9001) was used for all solution preparations and dilutions. All
glassware was soaked overnight in a 10% nitric acid (AnalR, BDH, Poole,
U.K.) and then rinsed with copious amounts of water.
(99%),
-(þ)-tartaric acid (>99.5%), (þ)-catechin hydrate (98%), potas-
sium bitartrate (99%), d6-DMSO (99.98%), -(-)-xylose (99%), glyoxylic
L-Ascorbic acid
L
L
acid monohydrate (98%), and copper(II) acetate (98%) were purchased
from Sigma-Aldrich (USA). Potassium hydroxide (AR grade, >85%) was
obtained from BDH/Merck (Australia), and sulfuric acid (AR grade,
>95%) and iron(II) sulfate heptahydrate (>98%) were purchased from
Ajax Fine Chemicals (Australia). Ethanol (AR grade, >99.5%, Ajax Fine
Chemicals), methanol (AR grade, >99.9%, Mallinckrodt Chemicals,
USA), glacial acetic acid (AR grade, >99.7%, APS Ajax Fine Chemicals),
hydrochloric acid (31.5%, AnalR, BDH, Australia), formic acid (98%,
Fluka, Germany), and acetonitrile (HPLC grade, >99.9%, Ajax Fine
Chemicals) were used without further purification.
High-resolution MS (HRMS) data were obtained on a 4.7 T FT-ICR
Bruker Apex 3 mass spectrometer, running Bruker X-Mass software
(v 7.0.2). The instrument was calibrated using sodium iodide in methanol.
The samples were dissolved in methanol and infused at a flow rate of
150 μL/min in negative ionization mode (ESI-). The mass range scanned
was 45-1000 amu. Capillary exit voltage was -50 V, and drying gas
temperature was 120 °C. Elemental analysis was run on a VG Opus 3.6
data station running VG elemental program.
Analytical LC analyses were conducted on an Ultra Performance
Liquid Chromatography (UPLC) system consisting of a Waters Acquity
binary solvent manager connected to a sample manager and a PDA
detector all run by Empower2 chromatography manager software. The
column was a Waters Acquity BEH C18 (2.1 ꢀ 50 mm) with 1.7 μm
particle diameter. Injection volume was 7.5 μL, column temperature 35 °C,
and flow rate 0.45 mL/min; the elution gradient consisted of solvent
A, 0.5% acetic acid in water, and solvent B, 0.5% acetic acid in methanol,
as follows (expressed in solvent A followed by cumulative time): 100%,
NMR data were acquired on a Bruker DRX600 NMR spectrometer
operating at a frequency of 600.18 MHz at 25 °C using TopSpin (version 1.3,
Bruker GMBH). Samples were prepared in d6-DMSO in Shigemi tubes
(Sigma-Aldrich) and degassed (argon) before sealing. Spectral widths were set
to allow at least 0.5 ppm either side of observed resonances. 1D NMR spectra
were recorded (6009.62 Hz) with 32K data points zero filled to 64K and
resolution enhanced using a Gaussian multiplication of the raw FID
(Gaussian broadening = 0.2, line broadening = -2 Hz) before Fourier
transformation. All 2D NMR experiments were run with quadrature

Article
J. Agric. Food Chem., Vol. 57, No. 20, 2009 9541
Figure 2. Liquid chromatogram at 280 nm of an ascorbic acid-(þ)-catechin solution after 4 days at 45 °C with spectra of compounds 1 and 2 (inset). The
model wine system consisted of a tartaric acid-buffered 12% ethanol solution (pH 3.2) with 1000 mg/L ascorbic acid, 500 mg/L catechin, and 1.5 mg/L iron(II).
Peaks eluting before (þ)-catechin are ascorbic acid and its degradation products.
1
detection with a H spectral width of 6009 Hz and a recycle delay of 2 s.
Chemical shifts were referenced to the residual d5-DMSO (δH 2.49 ppm; δC
1
39.8 ppm). High-power H π/2 pulses were determined to be 11.2 μs and
low power (for MLEV spin lock), 30 ms. The 13C high-power π/2 pulse was
11.0 μs, and a low-power pulse of 65 μs was used for GARP4 decoupling.
Gradient pulses were delivered along the z-axis using a 100 step sine program.
Homonuclear proton-proton correlation was achieved with the
eCOSY sequence (21) using gradient pulses for selection. The data were
recorded in 4K data points in t1 and in 512K data points in t2. Points were
predicted out to 8K data points in t2 and 1K data points in t1 (maximum
entropy) and zero filled to 8K and 4K data points, respectively. The data
were processed using a π/2 sine-bell shifted apodization inboth dimensions
and carefully phase corrected. TOCSY experiments were recorded over 2K
data points in t2 and 512 in t1 using the X_M16 sequence (22) for the
mixing time (30 ms; P29=50 µs, SP0=8.86 dB). The data were processed
using a π/2 sine-bell shifted apodization in both dimensions. HSQC
spectra were recorded over 2K data points in t2 and 512 data points in
t1 (10-170 ppm) using the sensitivity-enhanced double INEPT transfer
with no trim pulses and adiabatic pulses for 13C decoupling in f2 (23). The
spectra were made phase sensitive using echo-antiecho gradient selection
(80:20:11:-5). Long-range 1H-13C correlation (HMBC) spectra were
recorded with 2K data points in t1 (10-210 ppm) and t2 and processed
with magnitude calculation in f1 to destroy all phase information. Spectra
were obtained with a low-pass J-filter (145 Hz) to suppress one-bond
Figure 3. Formation of compounds 1 and 2 in a tartaric acid-buffered 12%
correlations. No decoupling was used during the acquisition, and gradient
ethanol solution with (þ)-catechin (250 mg/L), ascorbic acid (500 mg/L),
pulses (50:30:40:1) were used for selection. Two experiments were run,
optimized for a 20 and 2 Hz long-range coupling, respectively. Through-
and 1.5 mg/L iron. Error bars indicate the 95% confidence limits (n = 3,
Student’s t test) and were omitted on day 10 for clarity. The peak areas and
95% confidence limits (ꢀ 106) for compounds 1 and 2 at day 10 were 0.3
( 0.9 and 1 ( 9, respectively. The large uncertainty at day 10 was due to
the rapid decrease in concentration of the compounds at this time.
space connectivities were obtained using 2D homonuclear correlation via
dipolar coupling (ROESY) using phase sensitive echo-antiecho gradient
selection with a 250 ms mixing delay (2K ꢀ 512 data points).
Reactions under Various Model Wine Conditions. The composi-
tion of the model wine system was varied to establish the components
essential for the formation of the colorless products. Ascorbic acid
(1000 mg/L, 5.6 mM), (þ)-catechin (500 mg/L, 1.6 mM), and iron(II)
(1.5 mg/L, 0.027 mM) solutions were prepared in a tartaric acid-buffered
model wine system (pH 3.2), with and without ethanol (12% v/v) present,
as per Clark et al. (19). Tartaric acid- and formic acid-buffered model
white wine (pH 3.2), ascorbic acid (500 mg/L, 2.8 mM), and (þ)-catechin
(250 mg/L, 0.8 mM) solutions, with and without ethanol (12% v/v)
present, were prepared as per Barril et al. (17). The concentrations of
the reactants are higher than those generally found in white wine but were
required to generate sufficient amounts of oxidation products. It has been
shown that the high concentration did not affect the species formed but
only their rate of formation (18). All samples were prepared in triplicate,
stored at 45 °C in a covered water bath (i.e., darkness), aerated daily (with
rapid stirring for 2 min), and sampled every 2 or 4 days for chromato-
graphic analysis. The uncertainties quoted are the 95% confidence limits
(n=3, Student’s t test).
Isolation of Reaction Products. A pH 3.2 solution was prepared
by dissolving potassium bitartrate (2.4 g) and tartaric acid (1.2 g) in water
(1 L). Ascorbic acid (1.0 g) and (þ)-catechin (500 mg) were added to the
solution, followed by 1.5 mg/L iron(II) (in the form of iron(II) sulfate
heptahydrate), which was maintained in darkness at 45 °C, with daily
aeration, until the colorless products formed reached a maximum con-
centration (8 days as determined by UPLC-PDA). At this time ascorbic
acid (1.0 g) was added, to slow the otherwise rapid degradation of the
accumulated colorless products, and the 1 L sample was stored at 4 °C.
The 1 L solution was then passed through a solid-phase extraction
cartridge (Strata C18-E, 70 g/150 mL, Phenomenex, USA) to concentrate
and provide a crude purification of the isomers. The cartridge was
first conditioned with 80% aqueous methanol containing 2% formic acid
(200 mL), followed by 2% formic acid (300 mL). The sample was absorbed
and washed with 2% formic acid (300 mL) and eluted with 80% aqueous
methanol with 2% formic acid (200 mL). The eluent was concentrated

9542 J. Agric. Food Chem., Vol. 57, No. 20, 2009
Barril et al.
Figure 4. Heteronuclear multibond correlation spectrum of compound 2 showing correlations between C5OH and C4a and between C7OH and C8 (A); of H6
with both C8 and C4a (B); of (H100)2 and C8 (C); and (H4)2 and C4a (D).
3-fold under a stream of nitrogen and injected directly onto preparative
scale HPLC. Similar fractions were combined and blown down with a
stream of nitrogen, and the aqueous residue was freeze-dried to yield the
first eluting isomer as a pale yellow powder (15.0 mg; 2%): HRMS found,
[M - H]-, 401.088482, C20H17O9 requires 401.087257; λmax (0.5/24.8/
74.7% (v/v) of acetic acid/methanol/water) 284 (300) nm; 1H NMR
(600 MHz, d6-DMSO) δ 6.71 (d, J=2.1 Hz, H20), 6.68 (d, J=8.1 Hz,
H50), 6.58 (dd, J=8.2, 2.1 Hz, H60), 5.87 (s, H8), 4.50 (d, J=7.5 Hz, H2),
4.41 (s, (H400)2), 3.83 (m, H3), 3.78 (s, (H100)2), 2.64 (dd, J=16.0, 5.5 Hz,
H4R), 2.40 (dd, J=16.0, 7.5 Hz, H4β); 13C NMR (150.9 MHz, d6-DMSO)
δ 195.2 (C500), 176.1 (C600), 154.8 (C7), 154.2 (C5), 153.5 (C8a), 144.9
(C40), 144.9 (C30), 133.6 (C200), 130.6 (C10), 118.3 (C60), 115.1 (C50), 114.5
(C20), 101.5 (C6), 99.9 (C4a), 94.5 (C8), 80.7 (C2), 72.3 (C400), 66.3 (C3),
28.3 (C4), 21.6 (C100).
The second (major) isomer to elute was similarly isolated as a pale
yellow powder (25.0 mg; 4%): HRMS found, [M - H]-, 401.086017,
C20H17O9 requires 401.087257; λmax (0.5/24.8/74.7% (v/v) of acetic acid/
methanol/water) 284 (300) nm; 1H NMR (600 MHz, d6-DMSO) δ 9.07
(bs, C5-OH), 9.05 (bs, C7-OH), 8.80 (bs, C40-OH), 8.74 (bs, C30-OH),
7.98 (bs, C600-OH), 6.64 (d, J=8.1 Hz, H50), 6.64 (d, J=2.1 Hz, H20), 6.53
(dd, J=8.2, 2.1 Hz, H60), 5.99 (s, H6), 4.88 (bd, J=5 Hz, C3-OH), 4.60 (d,
J=6.56 Hz, H2), 4.37 (AB, Wh/2=18.4 Hz, (H400)2), 3.82 (m, H3), 3.68
(AB, J=16.05 Hz, (H100)2), 2.54 (dd, J=16.05, 5.14 Hz, H4R), 2.38 (dd, J=
16.05, 7.10 Hz, H4β); 13C NMR (150.9 MHz, d6-DMSO) δ 195.1 (C500),
176.0 (C600), 154.5 (C5), 154.4 (C7), 153.3 (C8a), 144.7 (C40), 144.6 (C30),
133.7 (C200), 130.9 (C10), 117.6 (C60), 115.1 (C50), 114.1 (C20), 98.8 (C4a),
98.7 (C8), 94.7 (C6), 80.7 (C2), 72.2 (C400), 66.3 (C3), 27.0 (C4), 21.3 (C100).
were conducted to ensure that the colorless compounds emanated
specifically from ascorbic acid and (þ)-catechin and were not
dependent on other components of the model wine system. For
instance, an oxidative degradation product of tartaric acid, used
to buffer wine-like media, is known to contribute to the format-
ion of (þ)-catechin oligomers through glyoxylic acid produc-
tion (5, 7).
The model wine systems were prepared at several concentra-
tions of ascorbic acid, (þ)-catechin, and iron(II), with the pre-
sence or absence of ethanol and with different supporting buffer
systems of either tartaric acid or formic acid. Figure 3 provides an
example of the production and decay of compounds 1 and 2 in a
model wine system containing (þ)-catechin and ascorbic acid. In
this case, ascorbic acid was almost depleted by day 10. As
reported previously (19), compounds 1 and 2 in the model wine
samples increased to a maximum concentration and then began
to decrease in concentration once ascorbic acid was near deple-
tion. The formation of compounds 1 and 2 is consistent with the
initial degradation of (þ)-catechin, as the polyphenol decay was
immediate and faster than in the absence of ascorbic acid ((þ)-
catechin alone in Figure 3).
In all combinations of the model wine system, the formation of
compounds 1 and 2 was reliant upon only the presence of both
(þ)-catechin and ascorbic acid. Other components of the model
wine system, particularly iron (0-1.5 mg/L) and/or ethanol
(0-12% (v/v)), had some impact on the rate of formation of
compounds 1 and 2 but were not essential for the formation of
these colorless compounds.
Synthesis of
Catechin. The preparation of
by Salomon et al. (24), following a method originating from Weidenhagen
in 1937 using copper(II) acetate (25). Briefly, -xylose (1 g, 6.7 mmol) was
L
-Xylosone and Subsequent Reaction with (þ)-
L
-xylosone was conducted as described
L
Identification of Compounds 1 and 2. Compounds 1 and 2
yielded identical ESI-MS spectra with peaks at m/z 401 and 403 in
negative and positive ionization mode, respectively. High-resolu-
tion MS data established a molecular formula of C20H18O9 ((3
ppm) for both compounds, indicating 12 degrees of unsaturation.
These facts coupled with identical UV-visible spectra (Figure 2,
inset) suggested that the compounds were isomers and contained
the (þ)-catechin chromophore (λmax = 278 nm), albeit slightly
bathochromically shifted to 284 nm. Loss of 113 amu from the
compounds gave ions with the same mass as catechin-Hþ (m/z
289 (negative ionization mode); C15H13O6), suggesting that the
two compounds were both substituted with a C5H5O3 fragment,
composed, as determined by NMR, of a ketone (δC 195.1) and an
enol (δC 176.0, 133.7), a methylene next to oxygen (δC 72.2), and
an aliphatic methylene (δC 21.3). With 3 degrees of unsaturation
for the substitution unit, this suggested a 5-methylene-4-hydroxy-
2H-furan-3-one structure. Similar structures have been reported
as products of acid treatment of xylitol (26). Two-dimensional
dissolved in aqueous methanol (95%; 60 mL), and the solution was boiled
under reflux until the xylose was fully dissolved. Copper(II) acetate (4 g,
21.6 mmol) was then added in one portion, and the solution was boiled
under reflux for a further 20 min. After cooling, the copper(I) oxide was
removed by filtration, and the concentrated filtrate was percolated
through a Chelex 100 ion-exchange resin (2.5 ꢀ 30 cm, Naþ form; BioRad,
USA) previously converted to the Hþ form.
L-Xylosone was eluted with
methanol to yield a pale yellow syrup (0.4 g, 40%).
-Xylosone (1 g) was dissolved in a tartaric acid-buffered model wine
L
system containing ethanol (50 mL) as prepared per Barril et al. (17).
(þ)-Catechin (500 mg) was added, and the solution was left to react in
darkness at 45 °C. The formation of compounds 1 and 2 was monitored by
UPLC-PDA every 12 h.
RESULTS AND DISCUSSION
Formation of Colorless Products in Various Model White Wine
Systems. In model wine solutions containing ascorbic acid and
(þ)-catechin, colorless compounds giving rise to peaks 1 and 2
(Figure 2) were detected as previously reported (19). Initial studies

Article
J. Agric. Food Chem., Vol. 57, No. 20, 2009 9543
Figure 5. Scheme for the hypothetical formation of compounds 1 and 2.
Table 1. 440 nm Absorbance of Samples during Storage at 45 °C in
NMR evidence supported this structure for the substitution
fragment. In particular, the two methylenes both showed JCH
Darknessa
3
couplings to the enol carbon at 176.0 ppm but no couplings to
each other. The methylene bonded to an oxygen (δC 72.2) was
coupled to the ketone (δC 195.1) in the HMBC spectrum, whereas
the aliphatic methylene (δC 21.3) was coupled to the enol (δC
133.7), indicating that this was the attachment point to the (þ)-
catechin. In the 1H NMR spectrum of compound 2, all of the (þ)-
catechin signals could be identified except one singlet for one
of the A ring protons, suggesting that the fragment was linked to
C6 or C8. This type of substitution is well-known for flava-
nols (5, 27).
sample
day 0
day 14
(þ)-catechin
0.001 ( 0.002
0.005 ( 0.003
0.010 ( 0.004
0.009 ( 0.001
0.011 ( 0.001
0.031 ( 0.002
0.09 ( 0.02
0.154 ( 0.006
0.13 ( 0.02
0.29 ( 0.01
compound 1
(þ)-catechin þ compound 1
compound 2
(þ)-catechin þ compound 2
a All samples were in a formic acid-buffered 12% aqueous ethanol solution. (þ)-
Catechin, 0.25 mM; compound 1 0.02 mM; and compound 2, 0.1 mM. Uncertainties
are expressed as the 95% confidence limits (n = 3, Student’s t-test).
For compound 2, the HSQC spectrum showed three methy-
lenes, one was clearly that of the (þ)-catechin C ring ((H4)2 2.54/
2.38 ppm) as it showed long-range JCH couplings to C5
(δC 154.5), C8a (δC 153.3), C4a (δC 98.8), C3 (δC 66.3), and C2
(δC 80.7) in the HMBC spectrum. C5 was differentiated from C8a
of attachment was confirmed by the fact that in compound 2,
both (H4)2 and (H100)2 were coupled to C8a, whereas in com-
pound 1 both were coupled to C5 (δC 154.2) in the HMBC
spectrum. As in compound 2, (H4)2 was also coupled to C8a and
(H100)2 was coupled to C7. Finally, the attachment point at C6
was confirmed by HMBC correlations between H8 (5.87 ppm)
and C6 (101.5 ppm) and C4a (99.9 ppm) as expected but also C7
(154.8 ppm) and C8a (153.5 ppm).
Compounds 1 and 2 (Figure 5) are the first reportedproducts to
form between a carbon-based degradation product of ascorbic
acid and (þ)-catechin. Their formation explains the accelerated
degradation of (þ)-catechin that has been reported (19) in model
wine solutions containing ascorbic acid. The formation of com-
pounds 1 and 2 also confirms that (þ)-catechin is not being
directly oxidized in the presence of ascorbic acid but is rather
forming an adduct with a degradation compound of ascorbic
acid. These results highlight the multifaceted role that ascorbic
acid may have in wine or other food systems. On the one hand,
ascorbic acid undoubtedly acts as an antioxidant; on the other,
2
on the basis of the former’s JCH coupling to an exchangeable
proton at 9.07 ppm. This exchangeable proton (9.07 ppm) was
also coupled to C4a and C6 (3JCH), indicating that the attachment
point of the -OH was C5. The other exchangeable proton on ring
A (C7OH) was characteristically coupled to C8 (Figure 4). H6
was coupled to C4a and C8, with the latter also coupled to (H100)2
(Figure 4). (H100)2 was also coupled to C8a and C7 in the HMBC
spectrum, unequivocally indicating that the furanone was at-
tached at C8. As expected, H6 showed equally intense ROE
correlation to C5-OH and C7-OH, the latter also showing a
correlation to (H100)2. All other spectral data were consistent with
structure 2 for the later eluting isomer (Figure 5).
Compound 1 was postulated to be the C6-attached isomer as
has been reported for other aldehydes (28). In this case the point

9544 J. Agric. Food Chem., Vol. 57, No. 20, 2009
Barril et al.
Figure 6. Liquid chromatograms at 440 nm of a solution of (þ)-catechin and compound 1 (upper trace) and of (þ)-catechin and glyoxylic acid (lower trace)
after 14 days at 45 °C. The solutions consisted of a formic acid-buffered 12% ethanol solution.
breakdown products of ascorbic acid are capable of fur-
ther reactions. In wines, those reactions may be with flavanols
(e.g., (þ)-catechin), which are the compounds most correlated to
the browning of white wine.
wine-like solutions (19), experiments were conducted to confirm
the ability of compounds 1 and 2 to form colored compounds.
Indeed, when left at room temperature after NMR analyses, the
solution of both isolated isomers developed deep golden-brown
coloration. When 0.02 mM compound 1 and 0.1 mM compound
2 were prepared (n=3) at (þ)-catechin concentrations of 0 and
0.25 mM in the model wine solution (formic buffer), they initially
afforded 440 nm absorbances (a measure of yellow color) of
<0.012 absorbance unit. However, after incubation at 45 °C in
darkness, the samples all increased in 440 nm absorbance to reach
the values indicated in Table 1. It is evident that both compounds
1 and 2 were precursors to colored compounds and that the
presence of 0.25 mM (þ)-catechin enhanced the formation of
colored compounds by compounds 1 and 2.
Analysis by UPLC at day 14 revealed that all samples with
added compound 1 or 2, regardless of catechin addition, had three
major peaks (3-5, Figure 6) in the 440 nm chromatograms. The
UV-visible spectra extracted from the UPLC-PDA demonstrated
that these major peaks (3-5, Figure 6), in all of the samples with
compound 1 or 2 present, showed identical absorbance maxima at
280 and 440 nm with a shoulder at 310 nm. LC-MS data associated
with these peaks all showed m/z signals at 617 in the positive
ionization mode and the major fragmentation ion at m/z 465.
Such MS and UV-visible data were consistent with the
xanthylium cation pigments derived from (þ)-catechin and
glyoxylic acid (Figure 7). The reaction of glyoxylic acid with
(þ)-catechin can generate six isomeric forms of the yellow
xanthylium cations (8), albeit at different rates of formation,
one of which is shown in Figure 7. These yellow pigments have
UV-visible maxima at 280 and 440 nm with a shoulder at 310
nm (7) and a molecular mass of 617 amu (7) as well as the
characteristic retro-Diels-Alder degradation of the dihydropyr-
an moiety to form the m/z 465 fragment (positive ionization
mode) (31).
Compounds 1 and 2 are stable substituted (þ)-catechin mono-
mers, provided some ascorbic acid remains in the model wine
system. Other examples of stable (þ)-catechin-adduct mono-
mers are formaldehyde-substituted (þ)-catechins generated as
minor products in the reaction between glyoxylic acid and (þ)-
catechin (29) and certain oak-derived aldehydes that also form
pyrylium-chromophores on addition to single (þ)-catechin
units (30). This is in contrast to many other (þ)-catechin addi-
tion products that readily undergo further addition and form
bridged-(þ)-catechin dimers. The reliance on ascorbic acidfor the
stability of compounds 1 and 2 suggests that an oxidation step is
required for their conversion to some other species (see below).
Formation and Fate of the Isomers as Model Wine System
Components. Of the many degradation products of ascorbic acid
that are reported in the literature (13-16)), L-xylosone (Figure 1
appeared to be a reasonable candidate as a precursor to com-
pounds 1 and 2. Indeed, a retro-synthetic analysis (Figure 5)
suggested that xylosone could be involved in the formation of 1
and 2. Indeed, no other known degradation product of ascorbic
acid allowed such feasible formation.
L-Xylosone, an aldehyde, is formed from the hydrolysis and
subsequent decarboxylation of dehydroascorbic acid (14), but
rapidly degrades to other aldehydes and ketones (13,15). Its lack
an appropriate chromophore to allow detection above 250 nm,
and its relatively polar character and expected elution near the
injection front under the chromatographic conditions used here
means that its detection by the system utilized in this experiment
was not possible. Consequently, L-xylosone was synthesized as
per the method described by Salomon et al. (24) and reacted with
(þ)-catechin in a model wine system consisting of a tartaric acid-
buffered 12% ethanol solution. Colorless compounds were sub-
sequently generated that had identical retention times, UV-vi-
sible spectra, and MS data as compounds 1 and 2 (as reported
To confirm that compounds 1 and 2 indeed form xanthylium
cations, it was necessary to compare retention times of the
purported xanthylium cations (as reported above) with those of
genuine xanthylium cations synthesized from (þ)-catechin and
glyoxylic acid (7). Thus, a solution of xanthylium cations was
prepared from 0.50 mM (þ)-catechin and 0.25 mM glyoxylic acid
by incubation at 45 °C in a tartaric acid-buffered model wine
solution. The solution was analyzed by UPLC-PDA, and the
above). This confirmed that L-xylosone was the precursor to
compounds 1 and 2 in model wine solutions of (þ)-catechin and
ascorbic acid.
Given that past work showed a link between the degradation of
compounds 1 and 2 with the accelerated production of color in

Article
J. Agric. Food Chem., Vol. 57, No. 20, 2009 9545
Figure 7. Formation of xanthylium cations from (þ)-catechin and ascorbic acid in a tartaric acid-buffered model wine system.
peaks (7-9, Figure 6) corresponding to the xanthylium cations
were confirmed by their UV-visible spectra and HPLC-MS
analysis (data as described earlier). The remaining major peak
in this solution (peak 10, Figure 6) corresponded to the ethyl ester
of the xanthylium cation, which has also been described pre-
viously (29). The retention times of the xanthylium cations
generated from glyoxylic acid with (þ)-catechin (peaks 7-9,
Figure 6) were at retention times identical to those of the major
pigments in all of the samples containing compound 1 or 2 (peaks
3-5, Figure 6). The matching retention times were confirmed
with the use of three different HPLC gradient systems and two
different chromatography columns (data not shown). Conse-
quently, the results demonstrate that compounds 1 and 2 form
the same xanthylium cations as those generated by (þ)-catechin
and glyoxylic acid. Intriguingly, this was the case whether
compound 1 or 2 was incubated with (þ)-catechin or not, but
more xanthylium cations were generated if (þ)-catechin was
present (Table 1).
The actual mechanism for the conversion of compounds 1 and
2 to xanthylium cations is not certain and is the subject of ongoing
investigations. However, the results do establish a mechanistic
link between (þ)-catechin, a degradation product of ascorbic
acid, and the production of xanthylium cations (Figure 7) that has
previously been postulated (17,18). The results also add xylosone
to the list of aldehyde compounds (i.e., glyoxylic acid (7) and
glyoxal (32)) that can react with (þ)-catechin to generate the
xanthylium cation pigments such as that shown in Figure 7. Such
a mechanism will be relevant to all beverages that contain
flavanols and ascorbic acid, such as fruit juices. Furthermore, it
will also be particularly relevant to wines with relatively high
xylosone concentration, such as those produced from grapes
infected with Botrytis cinerea (33).
ascorbic acid and (þ)-catechin. The process is far more complex
than the anticipated and simple transfer of two electrons and two
protons and demonstrates that ascorbic acid is a highly unpre-
dictable molecule capable of inducing a variety of reaction
processes. From a model white wine system, two new “natural”
products derived from (þ)-catechin and ascorbic acid were
isolated. The two compounds were unequivocally identified on
the basis of spectral data, primarily NMR spectroscopy. A retro-
synthetic study indicated that the compounds could be formed
from
and (þ)-catechin. Both isomers were independently synthesized
from -xylosone to establish this compound as a key precursor to
L-xylosone, a known degradation product of ascorbic acid,
L
compounds 1 and 2. The accumulation of the isomers (1 and 2)
depends on the presence of ascorbic acid, and they eventually
react further to give colored xanthylium cation pigments.
ABBREVIATIONS USED
HPLC-DAD, high-performance liquid chromatography with
diode array detection; DMSO, dimethyl sulfoxide; UV, ultravio-
let; HRMS, high-resolution mass spectrometry; FT-ICR-MS,
Fourier transform ion cyclotron resonance mass spectrometry;
ESI, electrospray ionization; NMR, nuclear magnetic resonance;
FID, free induction decay; COSY, correlation spectroscopy;
TOCSY, total correlation spectroscopy; HSQC, heteronu-
clear single quantum coherence; INEPT, insensitive nuclei en-
hanced by polarization transfer; HMBC, heteronuclear multi-
bond correlation; ROESY, rotating frame Overhauser effect
spectroscopy.
ACKNOWLEDGMENT
We are indebted to John Allen from the (Research School of
Chemistry) Australian National University for the high-resolu-
tion mass spectra.
In conclusion, the results of this study show a unique outcome
from the reactions of two “antioxidants” in wine-like conditions,

9546 J. Agric. Food Chem., Vol. 57, No. 20, 2009
Barril et al.
LITERATURE CITED
in a model wine matrix containing (þ)-catechin. J. Agric. Food
Chem. 2003, 51, 4126–4132.
(19) Clark, A. C.; Pedretti, F.; Prenzler, P. D.; Scollary, G. R. Impact of
ascorbic acid on the oxidative colouration and associated reactions
of a model wine solution containing (þ)-catechin, caffeic acid and
iron. Aust. J. Grape Wine Res. 2008, 14, 238–249.
(20) Lutter, M.; Clark, A. C.; Prenzler, P. D.; Scollary, G. R. Oxidation of
caffeic acid in a wine-like medium: production of dihydroxybenzal-
dehyde and its subsequent reactions with (þ)-catechin. Food Chem.
2007, 105, 968–975.
(21) Griesinger, C.; Sorensen, O. W.; Ernst, R. R. Practical aspects of
the E-COSY technique-measurement of scalar spin spin coupling-
constants in peptides. J. Magn. Reson. 1987, 75, 474–492.
(22) Peti, W.; Griesinger, C.; Bermel, W. Adiabatic TOCSY for C, C and
H, H J-transfer. J. Biomol. NMR 2000, 18, 199–205.
(1) Cheynier, V. Flavonoids in wine. In Flavonoids: Chemistry, Bio-
chemistry and Applications; Anderson, O. M., Markham, K. R., Eds.;
Taylor and Francis: London, U.K., 2006; pp 293-298.
(2) Wildenradt, H. L.; Singleton, V. L. The production of aldehydes as a
result of oxidation of polyphenolic compounds and its relation to
wine aging. Am. J. Enol. Vitic. 1974, 25, 119–126.
(3) Laurie, V. F.; Waterhouse, A. L. Oxidation of glycerol in the
presence of hydrogen pereoxide and iron in model solutions and
wine. Potential effects on wine colour. J. Agric. Food Chem. 2006, 54,
4668–4673.
(4) Clark, A. C.; Prenzler, P. D.; Scollary, G. R. Impact of the condition
of storage of tartaric acid solutions on the production and stability of
glyoxylic acid. Food Chem. 2007, 102, 905–916.
(5) Fulcrand, H.; Cheynier, V.; Oszmianski, J.; Moutounet, M. An
oxidized tartaric acid residue as a new bridge potentially competing
with acetaldehyde in flavan-3-ol condensation. Phytochemistry 1997,
46, 223–227.
(6) Es-Safi, N.; Cheynier, V.; Moutounet, M. Study of the reactions
between (þ)-catechin and furfural derivatives in the presence or
absence of anthocyanins and their implication in food color change.
J. Agric. Food Chem. 2000, 48, 5946–5954.
(23) Schleucher, J.; Schwendinger, M.; Sattler, M.; Schmidt, P.; Schedletzky,
O.; Glaser, S. J.; Sorensen, O. W.; Griesinger, C. A general enhance-
ment scheme in heteronuclear multidimensional NMR employing
pulsed-field gradients. J. Biomol. NMR 1994, 4, 301–306.
(24) Salomon, L.; Burns, J.; King, C. Synthesis of L
-ascorbic-1-14C acid
from sorbitol. J. Am. Chem. Soc. 1952, 74, 5161–5162.
(25) Weidenhagen, R. Z. Wirtschaftsgruppe Zuckerind. 1937, 87, 711.
(26) Shono, T.; Matsumura, Y.; Hamaguchi, H.; Naitoh, S. Synthesis of
2-methyl-3-hydroxy-4H-pyran-4-one and 4-hydroxy-5-methyl-2H-
furan-3-one from carbohydrates. J. Org. Chem. 1983, 48, 5126–5128.
(27) Timberlake, C. F.; Bridle, P. Interactions between anthocyanins,
phenolic compounds, and acetaldehyde and their significance in red
wines. Am. J. Enol. Vitic. 1976, 27, 97–105.
(7) Es-Safi, N.; Le Guerneve, C.; Labarbe, B.; Fulcrand, H.; Cheynier,
´
V.; Moutounet, M. Structure of a new xanthylium salt derivative.
Tetrahedron lett. 1999, 40, 5869–5872.
´
(8) Es-Safi, N. E.; Guerneve, C.; Labarbe, B.; Fulcrand, H.; Cheynier,
V.; Moutounet, M. Xanthylium salts formation involved in wine
colour changes. Int. J. Food Sci. Technol. 2000, 35, 63–74.
(9) Boulton, R. B.; Singleton, V. L.; Bisson, L. F.; Kunkee, R. E.
Principles and Practices of Winemaking; Springer: Norwell, MA, 1999;
pp 83-84.
(10) Rankine, B. Making Good Wine; Pan Macmillan Australia: Sydney,
Australia, 2002; pp 196-197.
(28) Saucier, C.; Guerra, C.; Pianet, I.; Laguerre, M.; Glories, Y. (þ)-
Catechin-acetaldehyde condensation products in relation to wine-
ageing. Phytochemistry 1997, 46, 229–234.
(29) Es-Safi, N. E.; Guerneve, C.; Cheynier, V.; Moutounet, M. New
´
phenolic compounds obtained by evolution of (þ)-catechin and
glyoxylic acid in hydroalcoholic medium. Tetrahedron lett. 2000,
41, 1917–1921.
(11) Robinson, J. The Oxford Companion to Wine, 3rd ed.; Oxford
University Press: New York, 2006; pp 35-36.
(30) de Freitas, V.; Sousa, C.; Silva, A. M. S.; Santos-Buelga, C.; Mateus,
N. Synthesis of a new catechin-pyrylium derived pigment. Tetrahe-
dron Lett. 2004, 45, 9349–9352.
(12) Zoecklein, B. W.; Fugelsang, K. C.; Gump, B. H.; Nury, F. S. Wine
Analysis and Production; Champman and Hall: South Melbourne,
Australia, 1995; pp 190-191.
(31) Es-Safi, N.; Le Guerneve, C.; Cheynier, V.; Moutounet, M. 2D
´
(13) Yuan, J. P.; Chen, F. Degradation of ascorbic acid in aqueous
NMR analysis for unambiguous structural elucidation of phenolic
compounds formed through reaction between (þ)-catechin and
glyoxylic acid. Magn. Reson. Chem. 2002, 40, 693–704.
(32) Es-Safi, N.; Cheynier, V.; Moutounet, M. Effect of copper on
oxidation of (þ)-catechin in a model solution system. Int. J. Food
Sci. Technol. 2003, 38, 153–163.
(33) Burroughs, L. F.; Sparks, A. H. Sulphite-binding power of wines and
ciders III. Determination of carbonyl compounds in a wine and
calculation of its sulphite-binding power. J. Sci. Food Agric. 1973, 24,
207–217.
solution. J. Agric. Food Chem. 1998, 46, 5078–5082.
(14) Whiting, G. C.; Coggins, R. A. Formation of L-xylosone from
ascorbic acid. Nature 1960, 185, 843–844.
(15) Shinoda, Y.; Murata, M.; Homma, S.; Komura, H. Browning and
decomposed products of model orange juice. Biosci., Biotechnol.,
Biochem. 2004, 68, 529–536.
(16) Kimoto, E.; Tanaka, H.; Ohmoto, T.; Choami, M. Analysis of the
transformation products of dehydro-L-ascorbic acid by ion-pairing
high-performance liquid chromatography. Anal. Biochem. 1993, 214,
38–44.
(17) Barril, C.; Clark, A. C.; Scollary, G. R. Understanding the con-
tribution of ascorbic acid to the pigment development in model white
wine systems using liquid chromatography with diode array and
mass spectrometry detection techniques. Anal. Chim. Acta 2008, 621,
44–51.
Received June 25, 2009. Revised manuscript received August 27, 2009.
Accepted August 31, 2009. This project was supported by the Australian
Grape and Wine Research and Development Corporation (CSU03/02).
We acknowledge the Australian Research Council and Macquarie
University for funding of the Macquarie University NMR Facility.
(18) Bradshaw, M. P.; Prenzler, P. D.; Scollary, G. R.; Cheynier, V.
Defining the ascorbic acid crossover from anti-oxidant to pro-oxidant