Establishment of the chemical equilibria of different types of pyranoanthocyanins in aqueous solutions: Evidence for the formation of aggregation in pyranomalvidin-3- O -coumaroylglucoside-(+)-catechin

Article abstract of DOI:10.1021/jp1045673

The chemical equilibria of the pyranomalvidin-3-glucosides linked to (+)-catechin, (-)-epicatechin, and catechol moieties (and the respective coumaroylglucoside compounds) were established by means of UV-vis spectroscopy. The conjugated double bonds among pyranic rings C and D provide a higher electronic delocalization that prevents the nucleophilic attack of water at position 2. Consequently, besides flavylium cation (AH⁺), the bases A, A^-, and A^2- have been identified by increasing pH, and the respective acidity constants were determined by spectrophotometry. The formation of dimers at higher concentration was observed for pyranomalvidin-3-O-coumaroylglucoside-(+)-catechin, and the respective data treated by the exciton model suggests the formation of a dimer where the monomers form J-type aggregates with the dipolar moments in opposite directions and rotated by 174° at a distance of 5.2 A? (from the center).

Full text of DOI:10.1021/jp1045673

13232
J. Phys. Chem. B 2010, 114, 13232–13240
Establishment of the Chemical Equilibria of Different Types of Pyranoanthocyanins in
Aqueous Solutions: Evidence for the Formation of Aggregation in
Pyranomalvidin-3-O-coumaroylglucoside-(+)-catechin
Lu´ıs Cruz,† Vesselin Petrov,‡ Nate´rcia Teixeira,† Nuno Mateus,† Fernando Pina,‡ and
Victor de Freitas*,†
Centro de InVestigac¸a˜o em Qu´ımica, Departamento de Qu´ımica, Faculdade de Cieˆncias, UniVersidade do
Porto, Rua do Campo Alegre, 687, 4169-007 Porto, Portugal and REQUIMTE, Departamento de Qu´ımica,
Faculdade de Cieˆncias e Tecnologia, UniVersidade NoVa de Lisboa, 2829-516 Caparica, Portugal
ReceiVed: May 19, 2010; ReVised Manuscript ReceiVed: September 9, 2010
The chemical equilibria of the pyranomalvidin-3-glucosides linked to (+)-catechin, (-)-epicatechin, and
catechol moieties (and the respective coumaroylglucoside compounds) were established by means of UV-vis
spectroscopy. The conjugated double bonds among pyranic rings C and D provide a higher electronic
delocalization that prevents the nucleophilic attack of water at position 2. Consequently, besides flavylium
cation (AH+), the bases A, A-, and A2- have been identified by increasing pH, and the respective acidity
constants were determined by spectrophotometry. The formation of dimers at higher concentration was observed
for pyranomalvidin-3-O-coumaroylglucoside-(+)-catechin, and the respective data treated by the exciton model
suggests the formation of a dimer where the monomers form J-type aggregates with the dipolar moments in
opposite directions and rotated by 174° at a distance of 5.2 Å (from the center).
Introduction
wine color changes from purple to orange-brown during storage
and aging.1 However, their equilibrium forms present in red wine
are poorly reported in the literature. For all these reasons and
because of the large number of compounds detected in wines
so far, pyranoanthocyanin-type compounds have gained great
interest. It is well known that the color of anthocyanins is greatly
affected by the pH of the solution. The sequence of chemical
reactions was correctly established by Brouillard et al.21-23 using
temperature, pressure, and pH jump experiments: at very acidic
pH, the red flavylium cation (AH+) is the predominant species;
when the pH is raised the flavylium cation immediately
undergoes a proton transfer reaction, giving rise the purple
quinonoidal base (A), and simultaneously but more slowly the
flavylium cation leads formation of colorless hemiketal (B) in
the hydration reaction. The hemiketal compound further un-
dergoes a tautomerization reaction to give the pale yellow cis-
chalcone (Cc), which isomerizes to trans-chalcone (Ct). How-
ever, these authors gave less importance to formation of Ct
(minor species in common anthocyanins at room temperature).
After the works of McClelland (1980)24 and Preston and
Timberlake (1981)25 in which trans-chalcone was identified and
isolated, complete elucidation of all pH-dependent species of
malvidin-3,5-diglucoside was carried out by Santos et al.
(1993).26 Recently, Asenstorfer et al. (2003)27 determined by
electrophoresis studies that the first proton transfer reaction of
malvidin 3-glucoside occurs at a quite lower pH (pKa 1.76) than
the ones previously reported (3.7 and 4.3),22,28 which is
controversial. Besides, similar results published with vitisin A
and B are also somehow divergent. For example, Asenstorfer
et al. (2007)29 determined two pKh values for vitisin A by
absorption spectroscopy, while Oliveira et al. (2009)30 by means
of NMR and visible spectroscopy reported that the hydration
reaction does not occur in vitisin B. In this work, six different
types of pyranoanthocyanins were studied by UV-vis spec-
troscopy in order to comprehend what kind of chemical
Pyranoanthocyanins belong to an important group of antho-
cyanin-derived pigments that occur essentially in processed
foodstuffs like fruits and vegetable juices and mainly in red
wines.1 Several different classes of these pigments have been
identified in the past decade such as vitisins,2-4 hydroxyphenyl-
pyranoanthocyanins,5,6 methylpyranoanthocyanins,7,8 vinylfla-
vanol-pyranoanthocyanins,9-11 portisins,12-14 and more recently
a new family of pyranoanthocyanin dimers.15 Formation of this
kind of compounds in food matrixes results from a cycloadition
between C-4/5-OH of the anthocyanin and a double bond from
another molecule such as pyruvic acid,4 vinylcatechin,16 and
vinylcatechol,17,18 resulting in an additional pyran ring D (see
Figure 1). Pyranoanthocyanins differ from their genuine precur-
sors (anthocyanins) in many aspects, especially in color. In
general, they have a λmax hypsochromically shifted (between
478 and 510 nm) conferring orange colors, except for portisins,
which show a bathochromic shift in their λmax values to more
bluish hues around 580 nm. Due to their new pyranic ring, the
color of these compounds is much more stable toward pH
variations and bleaching by SO2 in comparison to the genuine
anthocyanins.19 In fact, while anthocyanins lose about 80% of
their color between pH 1 and 5, as a result of the formation of
colorless hemiketal forms (pKh 2-3), pyranoanthocyanins
practically do not change the color intensity.20 Therefore, they
also have the possibility of acting as pseudonatural food
colorants in the substitution of synthetic ones. Elucidation of
structural transformations that anthocyanin derivatives undergo
in aqueous solutions is important in order to understand the color
expression of red wines. Formation of pyranoanthocyanins has
been appointed by many authors as the main process for which
* To whom correspondence should be addressed. Phone: +351.226082862.
Fax: +351.226082959. E-mail: vfreitas@fc.up.pt.
† Centro de Investigac¸a˜o em Qu´ımica.
‡ REQUIMTE.
10.1021/jp1045673  2010 American Chemical Society
Published on Web 09/29/2010

Pyranoanthocyanins in Aqueous Solutions
J. Phys. Chem. B, Vol. 114, No. 41, 2010 13233
Figure 1. Structures of the pyranoanthocyanin pigments (1-6) studied.
equilibrium could be established and provide the respective
equilibrium thermodynamic constants. This knowledge will
allow predicting which species of this type of pyranoanthocya-
nins could occur at wine pH and how they may contribute to
the red wine color. (Usually in a monomer-dimer equilibrium
the mathematical definition of the mole fraction of the dimer is
2[D]/([M] + 2[D]) while the mole part, [D]/([M] + [D]), and
mutatis mutandis for the monomer. In common equilibria these
definitions are coincident and used without distinction.)
Pigment Purification. The ethanol of each solution was
eliminated by evaporation, and liquid-liquid extraction with
ethyl acetate was performed in order to remove the catechins
and other organic impurities. The aqueous phases containing
the pyranoanthocyanin derivatives were applied onto a TSK
Toyopearl gel HW-40 (S) column (250 mm × 16 mm i.d.),
and the pigments were eluted with 40-50% aqueous methanol
acidified with 2% HCl at a flow rate of 0.8 mL/min. The
fractions collections were made upon visual detection of the
colored bands. Methanol was removed under vacuum, protected
from light and at room temperature. The pyranoanthocyanin
pigments were freeze dried and then stored at -18 °C until
use. When necessary, preparative HPLC was used in order to
obtain pure pigments. The compounds were further characterized
by LC-DAD/ESI-MS, and their structures and purity were
assessed by NMR analyses.11,17,34
HPLC. The samples were analyzed by HPLC (Merck-Hitachi
L-7100) on a 150 mm × 4.6 mm i.d. reversed-phase C18 column
(Merck, Darmstadt, Germany) thermostatted at 25 °C; detection
was carried out at 500 nm using a diode array detector (Merck-
Hitachi L-7450A). Solvents were (A) water/formic acid 9:1
(v:v) and (B) acetonitrile with the following gradient: 10-35%
B over 50 min at a flow rate of 0.5 mL/min. The sample
injection volume was 20 µL. The chromatographic column was
washed with 100% B for 10 min and then stabilized with the
initial conditions for another 10 min.
Materials and Methods
Reagents. TSK Toyopearl gel HW-40(S) was purchased from
Tosoh (Tokyo, Japan). Malvidin-3-O-glucoside (mv3glc) and
malvidin-3-O-coumaroylglucoside (mv3coumglc) were isolated
from a young red table wine (Vitis Vinifera L. cv. Touriga
Nacional) by semipreparative HPLC using a reversed-phase C18
column (250 mm × 4.6 mm i.d.) as reported elsewhere.31
8-Vinyl-(+)-catechin, 8-vinyl-(-)-epicatechin, and 4-vinylcat-
echol were obtained through organic hemisynthesis from (+)-
catechin, (-)-epicatechin, and catechol, respectively, following
a strategy already reported.32 The structures of these compounds
were assessed by NMR by comparison with the data previously
published.17,32 Acetonitrile (HPLC grade) and ethanol (absolute)
were obtained from Carlo Erba (Val de Reuil, France); formic
acid (98%, PA-ACS) and methanol (99.5%) were purchased
from Panreac Qu´ımica Sau (Barcelona, Spain); hydrochloridric
acid (37% v/v), sodium hydroxide, and ethyl acetate were
purchase from Jose´ M. Vaz Pereira, S.A. (Lisboa, Portugal). A
universal buffer of Theorell and Stenhagen33 was made by
dissolving 2.25 mL of phosphoric acid (85% w/w), 7.00 g of
monohydrated citric acid, 3.54 g of boric acid, and 343 mL of
1 M NaOH solution in Millipore water to 1 L. Solutions were
made using Millipore water. All other chemicals used were
HPLC grade.
Hemisynthesis of Pyranoanthocyanin Pigments. The hemi-
syntheses of six pyranoanthocyanin-type pigments were pro-
ceeded through the procedures described elsewhere.16,17 Briefly,
solutions containing mv3glc or mv3coumglc (2.5 mM) were
incubated with 8-vinyl-(+)-catechin, 8-vinyl-(-)-epicatechin,
or 4-vinylcatechol (2-3 equiv) in 5-10% ethanol/water at pH
3.5 (adjusted with dilute HCl or NaOH) at 30 °C. Formation of
the respective pigments was monitored by high-performance
liquid chromatography (HPLC) with diode array detection
(DAD) at 500 nm. When their formation reached the maximum,
the reactions were stopped and the respective purifications were
performed.
The preparative HPLC system (Elite LaChrom) was com-
posed of a L-2130 quaternary pump, a manual injector with a
loop of 250 µL, and a L-2420 UV-vis detector. The stationary
phase was composed of a reversed-phase C18 column (Merck,
Darmstadt, Germany) (150 mm × 4.6 mm i.d., 5 µm pore size)
thermostatted at 25 °C, and the eluents were (A) water/formic
acid 9:1 (v:v) and (B) methanol with the following gradient:
20-100% B over 60 min at a flow rate of 1 mL/min. The
detection wavelength was set to 500 nm.
LC-DAD/ESI-MS Analysis. A Finnigan Surveyor series
liquid chromatograph equipped with a Thermo Finnigan (Hy-
persil Gold) reversed-phase column (150 mm × 4.6 mm, 5 µm,
C18) thermostatted at 25 °C was used. The samples were
analyzed using the same solvents, gradients, injection volume,
and flow rate referred to above for HPLC analysis. Double-
online detection was done by a photodiode spectrophotometer
and mass spectrometry. The mass detector was a Finnigan LCQ
DECA XP MAX (Finnigan Corp., San Jose, CA) quadrupole
ion trap equipped with an atmospheric pressure ionization (API)
source, using an electrospray ionization (ESI) interface. The

13234 J. Phys. Chem. B, Vol. 114, No. 41, 2010
Cruz et al.
TABLE 1: Mass and Spectroscopic Data of the Synthesized
Pyranoanthocyanins
vaporizer and capillary voltages were 5 kV and 4 V, respec-
tively. The capillary temperature was set at 325 °C. Nitrogen
was used as both sheath and auxiliary gas at flow rates of 90
and 25, respectively (in arbitrary units). Spectra were recorded
in positive-ion mode between m/z 250 and 1500.
λmax
[M]+,
m/z
MS2,
m/z
MS3,
m/z
pigments
(nm)a
pymv3glc-(+)-catechin, 1
pymv3coumglc-(+)-catechin, 2
pymv3glc-(-)-epicatechin, 3
pymv3coumglc-(-)-epicatechin, 4
pymv3glc-catechol, 5
505
502
505
508
502
505
805
951
805
951
625
771
643
643
643
643
463
463
491
491
491
491
447
447
NMR Analysis. 1H NMR (500.13 MHz) and 13C NMR
(125.77 MHz) spectra were recorded in CD3OD/TFA (98:2) on
a Bruker-Avance 500 spectrometer at 303 K and with TMS as
an internal standard (chemical shifts (δ) in parts per million,
coupling constants (J) in Hertz). Multiplicities are recorded as
singlets (s), doublets (d), triplets (t), doublets of doublets (dd),
pymv3coumglc-catechol, 6
a Recorded from the LC-DAD.
1
multiplets (m), and unresolved (*). H chemical shifts were
assigned using 2D NMR (COSY) experiments, while 13C
resonances were assigned using 2D NMR techniques (gHMBC
and gHSQC).35,36 The delay for the long-range C/H coupling
constant was optimized to 7 Hz.
sized in model solution through the reactions of the respective
anthocyanin (malvidin-3-O-glucoside or malvidin-3-O-couma-
roylglucoside) with 8-vinyl-(+)-catechin, 8-vinyl-(-)-epicat-
echin, and 4-vinylcatechol (Figure 1), according to previous
studies.16,17,32
After the purification steps, the pure pigments were individu-
ally analyzed by LC-DAD/ESI-MS in positive-ion mode (Table
1). Full 1H and 13C NMR characterization confirmed the
structures of pigments 1-6, which are in agreement with the
data already described in the literature.11,17,34
UV-vis spectra of the synthesized pigments recorded from
the LC-DAD revealed an absorbance maximum λmax between
502 and 508 nm, which is typical of pyranoanthocyanin
compounds.37 Additionally, spectra of pigments 2, 4, and 6 show
a shoulder around 310 nm, which is characteristic of acylation
of p-coumaric acid.38
Absorption Spectroscopy. UV-vis absorption spectra were
recorded on a Varian-Cary 100 Bio spectrophotometer. Spectro-
scopic absorbance curves were recorded at 25 °C for all of the
solutions from 200 to 800 nm with a 1 nm sampling interval using
a quartz cell cuvette of 3.5 mL capacity and 1 cm optical path.
Determination of the Thermodynamic Pseudoequilibrium:
Apparent Acidity Constant (K′a). Stock solutions of pyra-
noanthocyanin pigments (0.1 mM) were prepared in 10%
ethanol/aqueous solution of 0.1 M HCl, which was kept,
protected against the sunlight, at 4 °C. The resulting low pH
value of the stock solutions (pH ≈ 1) ensured that most of the
pyranoanthocyanins is in its flavylium form, thus preventing
pigment degradation.
Thermodynamic Equilibrium. In the case of natural an-
thocyanin compounds, it is well known that the flavylium cation,
AH+, is thermodynamically stable at very acidic pH values (pH
) 1), being the dominant species of the equilibrium (Figure 2).
When the pH becomes higher, the flavylium cation could be
involved in two parallel reactions: (i) acid-base reaction to form
the neutral quinonoidal base, A, and (ii) hydration reaction to
give the colorless hemiketal, B. Proton transfer reactions take
place immediately after the pH jump on a microseconds time
scale, while the hydration reaction is much slower (subseconds
to seconds). On the other hand, at higher pH values (moderately
acidic) A is not the most stable species at the equilibrium but
B, which means that A is formed during the first instants of the
process and disappears partially later to give the equilibrium
distribution at these pH values, including also minor fractions
of Cc and Ct (Figure 2).
In order to study what kind of reactions pyranoanthocyanins
undergo, pH jumps from 1 to 3, 5, 7, 9, and 11 were carried
out. The respective UV-vis absorption spectra were recorded
and the kinetic processes followed over time. The shape and
position of the absorption bands do not change significantly after
70 h, which is indicative of the stability of the structures and
means that no chemical transformations/degradations may occur
during this period (see Figure 9 in the Supporting Information).
However, during this large period of time a continuous
decrease of the absorbance was observed in a wide range of
wavelengths for each pH. This behavior was initially thought
to be due to a very slow hydration process that could occur
preferentially in the positively charged carbon 2 since the other
active positions (carbons 4 and 10) are blocked. Later, it was
concluded that the observed behavior is due essentially to
precipitation of the compounds. In fact, after some days, some
solid particles were observed at the bottom of the flasks.
However, despite a greater decrease in the absorbance intensity
observed at pH 3 (see Figure 9 in the Supporting Information),
there were more solid particles mainly at pH 5 and 7, where
Determination of the apparent acidity constant (thermodynamic
pseudoequilibrium) was accomplished by measuring, on the com-
mon UV-vis spectrophotometer, solutions composed of one-third
of a pyranoanthocyanin stock solution, one-third of a solution of
NaOH (0.1 M), and one-third of a universal buffer solution (with
20% ethanol) at pH’s of 3, 5, 7, 9, and 11.
Titrations. The acidic thermodynamic constants were deter-
mined by spectophotometric titrations. Subsequently added to
the cell cuvette were 1 mL of NaOH solution (0.1 M), 1 mL of
universal buffer solution (with 20% ethanol) at pH 3, and 1
mL of a stock solution of the pyranoanthocyanins, where the
final concentration of each pigment was 0.03 mM. The first
UV-vis spectrum was immediately recorded. Increasing amounts
of NaOH solutions 10 or 1 M were successively added to the
cell cuvette, achieving a pH range between 3 and 12. After each
addition, the mixture was rapidly shacked, a UV-vis spectrum
was recorded, and the pH was measured. Final volumes were
appropriately corrected.
pH Measurements. All pH measurements were made in a
Crison micro pH meter 2002 fitted with a Crison 5209 electrode.
The calibration was made with standard buffers pH 7.0 and 4.0
purchased from Crison.
Aggregation of Pigment 2. A highly concentrated stock
solution of pigment 2 (∼1 mg in 500 µL) was prepared in a 10%
ethanol/aqueous solution of 0.1 M HCl (pH ) 1.3), and a UV-vis
spectrum was recorded using a quartz cell cuvette with a 0.2 cm
optical path. Increasing volumes of solvent were added to the stock
solution, and after each dilution a UV-vis spectrum was recorded.
When the absorbance at λmax reaches 0.45, the following UV-vis
spectra were recorded using a quartz cell cuvette with a 1 cm optical
path until a 40 times dilution had been reached.
Results and Discussion
Hemisynthesis of Pyranoanthocyanin Pigments. Six dif-
ferent types of pyranoanthocyanin pigments 1-6 were synthe-

Pyranoanthocyanins in Aqueous Solutions
J. Phys. Chem. B, Vol. 114, No. 41, 2010 13235
Figure 2. Equilibrium forms of natural anthocyanins in aqueous solutions.
Figure 3. UV-vis spectra of the solutions of pyranomv3coumglc-catechol pigment 6 at different pH values between 3.8 and 11.8. (Inset) UV-vis
spectra of the same pigment from 450 to 650 nm at pH values between 3.8 and 5.7.
the concentration of neutral quinonoidal base A is higher. The
conjugated double bonds among the pyranic rings C and D seem
to provide a higher electronic delocalization that should avoid
nucleophilic attack of water at position 2.
Figure 3 show the evolution of the UV-vis spectra of
solutions of pyranomv3coumglc-catechol pigment 6 at different
pH measured immediately after pH change. As pH increases
an isosbestic point was found around 558 nm, which is indicative
of the neutral quinonoidal base appearance (see inset of Figure
3).
Determination of pKa Values. Titrations of all pigments were
performed in order to determine the respective pKa values. The
titration was fast enough to avoid significant precipitation, which
is a relatively slow process. The UV-vis spectrum of each
pigment was obtained for the range of pH between 3 and 12
(for example, for pigment 6, Figure 3).
Overall, by increasing the pH the following behavior was
observed for all pigments: a decrease in the absorbance
accompanied by a blue shift of λmax followed by an absorbance
increase with a red shift of λmax
.
Once again, no evidence of the hydration process was
observed immediately after the pH increase from slightly acid
to neutral. Indeed, some reverse pH jump experiments were
performed in the stopped flow equipment in order to determine
amounts of hemiketal. A solution of the compound (initially at
pH ) 1.0) was left at pH ) 6.0 for approximately 30 min, and
a pH jump back to acid was followed by stopped flow (reverse
pH jumps). Observation of a small increase of the flavylium
absorption within the lifetime of seconds does not exclude
formation of residual amounts of hemiketal. This process was
neglected due to its small amplitude.
Furthermore, no evidence of formation of B2 hemiketal
(appearance of a band with an absorption at 280 nm, see
Figure 9 in the Supporting Information) is given in the recorded
UV-vis spectra. A similar behavior was observed for all
compounds studied. This means that only acid-base reactions
should take place as recently reported for vitisin B-type
compounds by UV-vis and NMR studies.30
For pigment 6, the first isosbestic point was achieved at
approximately 558 nm and can be attributed to the equilibrium
between AH+ and A. The second isosbestic point was found
around 468 nm and is due to the equilibrium between A and
A- species, while the third isosbestic point (423 nm) is related
to the equilibrium between A- and A2-
.
Treatment of the experimental data reported in Figure 3 using
a simple mathematical decomposition allowed achieving the
UV-vis absorption spectra of the four species, AH+, A, A-,
and A2- (see the cases of pigments 1 and 2, Figure 4).
For pigment 1, the first species formed is the neutral quinonoidal
base (A) which displays a λmax hypsochromatically shifted (∼477
nm) in comparison with AH+. The second deprotonation lead to
the anionic quinonoidal base (A-), bathochromatically shifted (λmax
≈ 557 nm). The third species formed is the dianionic quinonoidal
base, A2-, which is blue (λmax ≈ 574 nm).
In Figure 4II, it is interesting to note that the A2- species
showed a band with λmax around 360 nm. This behavior was

13236 J. Phys. Chem. B, Vol. 114, No. 41, 2010
Cruz et al.
Figure 4. UV-vis spectra of the individual species of the pigments pyranomv3glc-(+)-catechin 1 (I) and pyranomv3coumglc-(+)-catechin 2 (II).
only observed for pigments 2, 4, and 6 which have a p-
coumaroyl residue attached. In fact, when the A2- species is
formed (pH > 9.5), the ionized form of the p-coumaroyl residue
is already present in solution (pKa2 p-coumaric acid ) 8.98),39
which might contribute to the appearance of a band at 360 nm.
On this basis, the flavylium cation can be considered as a
three protic acid and the global process can be accounted for
by eqs 1-3
[A]
C0
ꢀA )
Ka1[H3O+]2
[H3O+]3 + Ka1[H3O+]2 + Ka1Ka2[H3O+] + Ka1Ka2Ka3
(7)
)
[A-]
C0
ꢀA
-
)
AH+ + H2O T A + H3O+ Ka1, first proton transfer
Ka1Ka2[H3O+]
[H3O+]3 + Ka1[H3O+]2 + Ka1Ka2[H3O+] + Ka1Ka2Ka3
(8)
(1)
)
A + H2O T A- + H3O+ Ka2, second proton transfer
(2)
[A2-
C0
]
ꢀA2-
)
)
A- + H2O T A2- + H3O+ Ka3, third proton transfer
Ka1Ka2Ka3
(3)
[H3O+]3 + Ka1[H3O+]2 + Ka1Ka2[H3O+] + Ka1Ka2Ka3
(9)
[A][H3O+]
Ka1
)
[AH+]
[A-][H3O+]
[A]
The absorbance at wavelength λ shows a pH dependence which
is given by the following equation
Ka2
Ka3
)
)
[A2-][H3O+]
Acalcd ) εAH+ꢀAH + ε ꢀ + ε ꢀ + εA2-ꢀA2-
+
-
-
A
A
A
A
(4)
[A-]
(10)
+
-
2-
where εAH , εA, εA , and εA are the mole absorption coefficients
of the individual species or components at the considered
wavelength.
The total concentration of individual species, C0, is given by
eq 5
The pKa values were determined by fitting the calculated
absorbance values by eq 10 to the ones obtained experimentally.
The solver function was then used to minimize the sum of the
difference between the theoretical and the experimental values
squared.
In Figure 5, the molar fractions of each species and the
experimental absorbance values at two chosen wavelengths are
plotted for pigment 1, with the respective fittings, as a function
of pH.
The pKa values for pigments 1-6 are given in Table 2. It is
reported in the literature that the pKa for 7-hydroxyflavylium is
3.55 while in 4′-hydroxyflavylium is 4.5. Therefore, it is easier
to remove a proton from position 7 than from position 4′. When
C0 ) [AH+] + [A] + [A-] + [A2-]
(5)
and the mole fraction distribution of the different species, ꢀi, is
calculated as follows
[AH+]
C0
ꢀAH
)
)
+
[H3O+]3
[H3O+]3 + Ka1[H3O+]2 + Ka1Ka2[H3O+] + Ka1Ka2Ka3
(6)

Pyranoanthocyanins in Aqueous Solutions
J. Phys. Chem. B, Vol. 114, No. 41, 2010 13237
Figure 5. Determination of respective pKa values of pigment 1 at λ ) 506 and 572 nm. The cuvette cells show the colors of each individual
species.
TABLE 2: pKa Values for Pigments 1-6 Obtained from UV-Vis Spectroscopya
AH+
A
A-
A2-
pigment
λmax (nm)
pKa1
λmax (nm)
pKa2
λmax (nm)
pKa3
λmax (nm)
1
2
3
4
5
6
507
512
508
509
511
513
5.05 ((0.05)
5.35 ((0.08)
4.80 ((0.09)
5.24 ((0.11)
4.20 ((0.06)
4.31 ((0.07)
477
480
481
476
487
471
7.90 ((0.07)
8.06 ((0.09)
7.82 ((0.09)
8.02 ((0.10)
7.84 ((0.05)
8.34 ((0.06)
557
562
567
559
577
562
9.77 ((0.06)
9.76 ((0.07)
9.49 ((0.10)
9.60 ((0.08)
10.28 ((0.07)
10.20 ((0.06)
574
577
576
577
583
583
a Values in parentheses are the standard deviations of the curve-fitting procedure.
two hydroxyl groups are present, release of the first proton
influences the pKa of the second one and successively. For
example, in the case of 7,4′-dihydroxyflavylium, the first (7-
OH) pKa is 4 while the second (4′-OH) is 8,40-43 which are in
agreement with the ones obtained for all pigments studied (pKa1
4.20-5.35 and pKa2 7.82-8.34). This data also agrees with the
pKa values obtained for the vitisin B pigment which has the
same pyranoflavylium core in its structure.30 The third pKa
obtained for the pigments bearing a (+)-catechin or (-)-
epicatechin moiety (1-4) is observed between 9.49 and 9.77,
and the two pKa values for catechin and epicatechin described
in the literature are 8.16 and 9.20.44 The third pKa for the
pigments with catechol moieties (5 and 6) are higher (around
10.20), which is consistent with the catechol pKa values reported
in aqueous solutions (pKa1 9.45 and pKa2 12.8).45 Overall, the
pigments that bear a coumaroyl residue (2, 4, and 6) revealed
slightly higher pKa values than the respective glucosides (1, 3,
and 5), which may be explained by the higher steric hindrance
brought by the coumaroyl group to the overall structure, making
it more difficult to remove an additional proton.
the appearance of only one band (λmax 509 nm). This observation
is indicative of a monomer-dimer equilibrium.
In order to characterize the monomer-dimer equilibrium,
several UV-vis spectra of pigment 2 were recorded as a
function of increasing concentrations from 5.7 × 10-5 to 2.6 ×
10-4 M (Figure 7A). The absorbance data that would be obtained
for the initial concentration C0 and for an optical path of 1 cm
were appropriately corrected (Figure 7B).49
Two sets of isosbestic points were clearly identified, sug-
gesting not only formation of a dimer but also high-order
aggregates, as can be confirmed from Figure 10 of the
Supporting Information.
Calculation of the monomer-dimer equilibrium was achieved
taking into account only the points where the absorption can
be considered linearly dependent on the logarithm of the
concentration, and the high-order aggregates could be neglected.
The monomer (M)-dimer (D) equilibrium is described by
eq 11
The general equilibrium forms of the pyranoanthocyanins
studied in aqueous solutions for the pH range between 3 and
12 are depicted in Figure 6.
Aggregation of Pigment 2. Aggregation of the ionic dyes
in solution is one of the properties which affects their spectral
and physicochemical properties.46,47 It is well known that these
compounds form dimers even at very low concentrations from
10-5 to 10-6 M.48
Formation of aggregates was only observed for pigment 2
by recording the UV-vis spectrum of the stock solution (0.1
mM), which showed two bands with a λmax bathochromatically
shifted from 509 to 526 nm. The UV-vis spectrum of this
diluted solution (5 × 10-5 M) was then recorded, resulting in
M + M T D
(11)
(12)
and the dimeric constant (KD) is defined by eq 12
[D]eq
([M]eq)2
ꢀDeq
KD )
)
2C0i (ꢀM)2
eq
Considering the mass balance of the system is given by eq 13

13238 J. Phys. Chem. B, Vol. 114, No. 41, 2010
Cruz et al.
Figure 6. Equilibrium forms of pyranomv3glc-(+)-catechin pigment 1 in aqueous solutions for the pH range between 3 and 12.
Figure 7. (A) Absorption spectra of pigment 2 as a function of the concentration from 5.7 × 10-5 to 2.6 × 10-4 M; (B) UV-vis spectra of
pigment 2 solutions with different concentrations keeping C0l constant.
ꢀMeq + ꢀDeq ) 1
2[D]eq
Ci0
ꢀDeq
)
(15)
(16)
[M]eq + 2[D]eq ) C0i
(13)
A ) AMꢀMeq + ADꢀDeq
where the molar fractions (ꢀ) are defined according eqs 14 and
15
where AM and AD are the absorptions that the pure species would
have at an initial concentration C0i and at 526 nm.
The fitting was achieved using eqs 13-16 for log Kd ) 3.5
( 0.1, and ꢀM and ꢀD are equal to 12 600 and 53 000 M-1 cm-1,
respectively (AD ) (εDC0i )/2 and AM ) εMCi0) (Figure 8A and
see Figure 11 in the Supporting Information).
[M]eq
Ci0
-1 + 1 + 8C0K
√
i
d
ꢀMeq
)
)
(14)
4KdC0i

Pyranoanthocyanins in Aqueous Solutions
J. Phys. Chem. B, Vol. 114, No. 41, 2010 13239
Figure 8. (A) Fitting of the absorption of compound 2 at 526 nm using eqs 14-16. (B) Calculated monomer (dotted line) and dimer (full line)
individual absorption spectra of pigment 2. Decomposition bands used for calculation of the dimer parameters are also represented.
TABLE 3: Results for Parameters of the Dimer Formed by
Conclusions
Pigment 2
The reactions of four pyranoanthocyanin-flavan-3-ols and two
pyranoanthocyanin-catechol pigments in hydroalcoholic aque-
ous solutions at different pH values were studied. These
compounds were shown to not undergo hydration reactions but
only acid-base reactions as recently described for vitisin B
pigment. The acid-base equilibrium gives rise to three quinon-
oidal bases (neutral A, anionic A-, and dianionic A2-) between
pH 3 and 12. The pKa values obtained for all pigments were
pKa1 between 4.20 and 5.35, pKa2 between 7.82 and 8.34, and
pKa3 between 9.49 and 10.28. The first two pKa values obtained
agree with the ones available in the literature for natural and
synthetic 7,4′-dihydroxyflavylium compounds. The third pKa
value (due to a proton transfer from (+)-catechin, (-)-
epicatechin, or catechol unit) is also close to the respective
values already reported. At wine pH (3.2-4.0) the pyranoan-
thocyanins studied are in their flavylium form AH+ (red-orange
color), and therefore, their contribution to the orange hues
observed in aged wines could be important. Pyranomalvidin-
3-O-coumaroylglucoside-(+)-catechin have shown the ability
to form dimer and high-order aggregates at higher concentrations.
pigment 2
∆ν (cm-1
R (deg)
R (Å)
)
1232
174
5.2
log KD
3.5
The set of absorption spectra regarding the monomer-dimer
equilibrium were proceeded by the FiNAl algorithm50 as
described in ref 49. This algorithm recovers the individual
spectrum of the monomer and dimer and gives their contribution
(mole fraction) for each concentration. Moreover, the overall
spectrum of the monomer and dimer is resolved in their
individual bands (Gaussian shape) (Figure 8B), where monomer
and the H and J bands are represented (see below).
Figure 8B shows that the monomer absorption splits into two
new bands in the dimer, one red shifted (λmax 526 nm) of major
intensity and the other blue shifted (λmax 509 nm).
The monomer-dimer absorption spectrum can be interpreted
through the exciton model.51,52 According to this model the angle
(R) and distance (R in Å) between the transition moment of the
monomers in the dimer are defined by eqs 17 and 18,
respectively
Acknowledgment. This research was supported by the
research project grants (PTDC/QUI/67681/2006) and CONC-
REEQ/275/2001 funding from FCT (Fundac¸a˜o para a Cieˆncia
e a Tecnologia) from Portugal. L.C. gratefully acknowledges
the Ph.D. grant from FCT (SFRH/BD/30915/2006), and V.P.
acknowledges the Postdoctoral Grant SFRH/BPD/18214/2004.
νHfJ
R ) 2 arctan
(17)
(18)
ν f
ꢀ
J H
Supporting Information Available: Three figures that help
to explain better the study of pyranoanthocyanin reactions in
aqueous solutions and the aggregation phenomena of pigment
2. This material is available free of charge via the Internet at
http://pubs.acs.org.
2.14 × 1010 cos RfM
∆ννM
3
R )
ꢀ
References and Notes
where νx is the position of band X in cm-1 and fX is the oscillator
(1) Rentzsch, W.; Schwarz, M.; Winterhalter, P. Trends Food Sci.
Technol. 2007, 18, 526–534.
(2) Bakker, J.; Bridle, P.; Honda, T.; Kuwano, H.; Saito, N.; Terahara,
N.; Timberlake, C. F. Phytochemistry 1997, 44, 1375–1382.
(3) Bakker, J.; Timberlake, C. F. J. Agric. Food Chem. 1997, 45, 35–
43.
(4) Fulcrand, H.; Benabdeljalil, C.; Rigaud, J.; Cheynier, V.; Moutounet,
M. Phytochemistry 1998, 47, 1401–1407.
(5) Cameira dos Santos, P. J.; Brillouet, J. M.; Cheynier, V.; Moutounet,
M. J. Sci. Food Agric. 1996, 70, 204–208.
∞
straight defined as fX ) 1.3 × 10-8Θ(n)∫-∞ε(ν)dν, Θ(n) ) 9n/
((n2 + 2)2) the factor of the environment, n the refraction index,
and ∆ν˜ the difference of the positions of the split bands (J and
H) in cm-1
.
The results obtained for KD, R, and R for pigment 2 (Table
3) suggest that the interaction is a J-type aggregate as described
in the literature.51

13240 J. Phys. Chem. B, Vol. 114, No. 41, 2010
Cruz et al.
(6) Fulcrand, H.; Cameira dos Santos, P.; Sarni-Manchado, P.; Chey-
nier, V.; Favre-Bonvin, J. J. Chem. Soc., Perkin Trans 1 1996, 735–739.
(7) He, J.; Santos-Buelga, C.; Silva, A. M. S.; Mateus, N.; De Freitas,
V. J. Agric. Food Chem. 2006, 54, 9598–9603.
(8) Lu, Y. R.; Sun, Y.; Foo, L. Y. Tetrahedron Lett. 2000, 41, 5975–
5978.
(30) Oliveira, J.; Mateus, N.; Silva, A. M. S.; De Freitas, V. J. Phys.
Chem. B 2009, 113, 11352–11358.
(31) Pissarra, J.; Mateus, N.; Rivas-Gonzalo, J. C.; Santos-Buelga, C.;
De Freitas, V. J. Food Sci. 2003, 68, 476–481.
(32) Cruz, L.; Borges, E.; Silva, A. M. S.; Mateus, N.; De Freitas, V.
Lett. Org. Chem. 2008, 5, 530–536.
(33) Ku¨ster, W. F.; Thiel, A. Tabelle per le analisi chimiche e chimico-
(9) Asenstorfer, R. E.; Hayasaka, Y.; Jones, G. P. J. Agric. Food Chem.
2001, 49, 5957–5963.
fisiche, 12th ed.; Hoepli: Milano, 1982.
(10) Francia-Aricha, E. M.; Guerra, M. T.; Rivas-Gonzalo, J. C.; Santos-
Buelga, C. J. Agric. Food Chem. 1997, 45, 2262–2265.
(11) Mateus, N.; Silva, A. M. S.; Santos-Buelga, C.; Rivas-Gonzalo,
J. C.; De Freitas, V. J. Agric. Food Chem. 2002, 50, 2110–2116.
(12) Mateus, N.; Oliveira, J.; Haettich-Motta, M.; De Freitas, V.
J. Biomed. Biotechnol. 2004, 299–305.
(13) Mateus, N.; Silva, A. M. S.; Rivas-Gonzalo, J. C.; Santos-Buelga,
C.; De Freitas, V. J. Agric. Food Chem. 2003, 51, 1919–1923.
(14) Oliveira, J.; De Freitas, V.; Silva, A. M. S.; Mateus, N. J. Agric.
Food Chem. 2007, 55, 6349–6356.
(15) Oliveira, J.; Azevedo, J.; Silva, A. M. S.; Teixeira, N. r.; Cruz, L.;
Mateus, N.; De Freitas, V. J. Agric. Food Chem. 2010, 58, 5154–5159.
(16) Cruz, L.; Teixeira, N.; Silva, A. M. S.; Mateus, N.; Borges, J.; De
Freitas, V. J. Agric. Food Chem. 2008, 56, 10980–10987.
(17) Hakansson, A. E.; Pardon, K.; Hayasaka, Y.; de Sa, M.; Herderich,
M. Tetrahedron Lett. 2003, 44, 4887–4891.
(18) Schwarz, M.; Jerz, G.; Winterhalter, P. Vitis 2003, 42, 105–106.
(19) SarniManchado, P.; Fulcrand, H.; Souquet, J. M.; Cheynier, V.;
Moutounet, M. J. Food Sci. 1996, 61, 938–941.
(20) Mazza, G.; Minati, E. Anthocyanins in fruits, Vegetables and grains;
CRC Press: Boca Raton, FL, 1993.
(21) Brouillard, R.; Dubois, J. E. J. Am. Chem. Soc. 1977, 99, 1359–
1364.
(22) Brouillard, R.; Delaporte, B. J. Am. Chem. Soc. 1977, 99, 8461–
8468.
(23) Brouillard, R.; Delaporte, B.; Dubois, J. E. J. Am. Chem. Soc. 1978,
100, 6202–6205.
(24) McClelland, R. A.; Gedge, S. J. Am. Chem. Soc. 1980, 102, 5838–
5848.
(34) Mateus, N.; Carvalho, E.; Carvalho, A. R. F.; Melo, A.; Gonzalez-
Paramas, A. M.; Santos-Buelga, C.; Silva, A. M. S.; De Freitas, V. J. Agric.
Food Chem. 2003, 51, 277–282.
(35) Bax, A.; Subramanian, S. J. Magn. Reson. 1986, 67, 565–569.
(36) Bax, A.; Summers, M. F. J. Am. Chem. Soc. 1986, 108, 2093–
2094.
(37) Mateus, N.; Silva, A. M. S.; Vercauteren, J.; De Freitas, V. J. Agric.
Food Chem. 2001, 49, 4836–4840.
(38) Harborne, J. B.; Grayer, R. J. In The flaVonoids, adVances in
research since 1980; Harborne, J. B., ed.; Chapman & Hall: London, 1988;
pp 1-20.
¨
(39) Beltra´n, J. L.; Sanli, N.; Fonrodona, G.; Barro´n, D.; Ozkan, G.;
Barbosa, J. Anal. Chim. Acta 2003, 484, 253–264.
(40) McClelland, R. A.; McGall, G. H. J. Org. Chem. 1982, 47, 3730–
3736.
(41) Pina, F.; Melo, M. J.; Ballardini, R.; Flamigni, L.; Maestri, M. New
J. Chem. 1997, 21, 969–976.
(42) Pina, F.; Melo, M. J.; Parola, A. J.; Maestri, M.; Balzani, V.
Chem.sEur. J. 1998, 4, 2001–2007.
(43) Pina, F.; Roque, A.; Melo, M. J.; Maestri, I.; Belladelli, L.; Balzani,
V. Chem.sEur. J. 1998, 4, 1184–1191.
(44) Slabbert, N. P. Tetrahedron 1977, 33, 821–824.
(45) Timberlake, C. F. J. Chem. Soc. 1957, 4987–4993.
(46) Bergmann, K.; O’Konski, C. T. J. Phys. Chem. 1963, 67, 2169–
2177.
(47) Monahan, A. R.; Blossey, D. F. J. Phys. Chem. 1970, 74, 4014–
4021.
(48) Georges, J. Spectrochim. Acta, A: M 1995, 51, 985–994.
(49) Antonov, L.; Gergov, G.; Petrov, V.; Kubista, M.; Nygren, J.
Talanta 1999, 49, 99–106.
(25) Preston, N. W.; Timberlake, C. F. J. Chromatogr. 1981, 214, 222–
228.
(26) Santos, H.; Turner, D. L.; Lima, J. C.; Figueiredo, P.; Pina, F. S.;
Macanita, A. L. Phytochemistry 1993, 33, 1227–1232.
(27) Asenstorfer, R. E.; Iland, P. G.; Tate, M. E.; Jones, G. P. Anal.
Biochem. 2003, 318, 291–299.
(28) Macanita, A. L.; Moreira, P. F.; Lima, J. C.; Quina, F. H.; Yihwa,
C.; Vautier-Giongo, C. J. Phys. Chem. A 2002, 106, 1248–1255.
(29) Asenstorfer, R. E.; Jones, G. P. Tetrahedron 2007, 63, 4788–4792.
(50) Antonov, L.; Petrov, V. Anal. Bioanal. Chem. 2002, 374, 1312–
1317.
(51) Kasha, M.; Rawls, H. R.; Ashraf El-Bayoumi, M. Pure Appl. Chem.
1965, 11, 371–392.
(52) Mason, S. F. J. Soc. Dyers Colour 1968, 84, 604–612.
JP1045673