Intermolecular copigmentation of malvidin-3-O-glucoside with caffeine was studied using a holistic procedure that includes the extension to basic pH values. In moderately basic solutions (7.5 < pH < 9.5) and independently of the copigment presence, there is a pH region where degradation of the quinoidal base and anionic quinoidal bases is faster than hydration and OH- nucleophilic addition, preventing the system from reaching the equilibrium. Intermolecular copigmentation with caffeine reduces significantly the degradation rate of quinoidal bases. In a more basic medium, the equilibrium is reached and degradation occurs from the anionic chalcones. In this case, the addition of caffeine also reduces the degradation rate in the interval 10 < pH < 11.5.
Intermolecular copigmentation of malvidin-3-O-glucoside with caffeine was studied using a holistic procedure that includes the extension to basic pH values. In moderately basic solutions (7.5 < pH < 9.5) and independently of the copigment presence, there is a pH region where degradation of the quinoidal base and anionic quinoidal bases is faster than hydration and OH- nucleophilic addition, preventing the system from reaching the equilibrium. Intermolecular copigmentation with caffeine reduces significantly the degradation rate of quinoidal bases. In a more basic medium, the equilibrium is reached and degradation occurs from the anionic chalcones. In this case, the addition of caffeine also reduces the degradation rate in the interval 10 < pH < 11.5.
Anthocyanins are ubiquitous
molecules that confer color to the
majority of angiosperms.[1] By presenting
an attractive color, they favor plant reproduction, attracting pollinators
and seed dispersers,[2,3] protect plants from several biotic
and abiotic stresses,[4,5] and function as photoprotective
agents by absorbing excess visible and UV light.[6,7] The
consumption of anthocyanin-rich foods has been associated with numerous
beneficial health effects due to their biological properties such
as antiproliferative, anti-inflammatory, and antimicrobial properties.[8−12] However, anthocyanins have a severe drawback. They are not stable
enough for most of the applications that imply relatively long storage
periods.[13−16] Intermolecular copigmentation is used by plants to extend the pH
domain of the red flavylium cation and increase the purple color of
the quinoidal base.[14] The question is as
follows: did intermolecular copigmentation also have some protecting
effect on the anthocyanin’s degradation? From this perspective,
the tools of the physical chemistry, in particular the studies concerning
the pH-dependent thermodynamics and kinetics of anthocyanins, are
indispensable to characterize both reversible and irreversible processes
occurring in these compounds.[14] This knowledge
is critical to design protection strategies to stabilize anthocyanins.
Reversible versus Irreversible Reactions in
Anthocyanins
Reversible Equilibrium
Direct pH Jumps
The reversible
sequence of chemical reactions that follow the addition of the base
to equilibrated solutions of the flavylium cation of malvidin-3-O-glucoside at pH ≤ 1 (direct pH jumps) is illustrated
in Scheme .[14,17,18] The first reaction consists of
flavylium cation deprotonation to give the respective quinoidal bases
with a mole fraction distribution dependent on the final pH. These
deprotonation processes are very fast and occur during the mixing
time of the stopped flow (first step).[19] Nevertheless, it is possible to calculate the respective acid–base
constants (first column of Scheme ) by collecting the absorption spectra using a stopped
flow 10 ms after the mixture of the base and representing the absorption
versus pH plot at representative wavelengths, as reported in Figure a.
Scheme 1
Sequence of Reversible
Reactions upon the Addition of the Base to
Equilibrated Solutions of the Flavylium Cation of Malvidin-3-O-glucoside at pH ≤ 1 (Direct pH Jumps)
This scheme is followed by all
anthocyanins. Adapted with permission from ref (19). Copyright 2022. American
Chemical Society.
Figure 3
(a) Titration curves of malvidin-3-O-glucoside
in the presence of caffeine 0.05 M collected by stopped flow 10 ms
after a direct pH jump from the flavylium cation at pH = 1 at three
representative wavelengths. Fitting was achieved for pKa = 3.5, pKA/A– = 6.7, and pKA–/A2– = 8.3; (b) mole fraction distribution
of the malvidin-3-O-glucoside species, flavylium
cation, and quinoidal bases in the presence of caffeine taken 10 ms
after a direct pH jump; and (c) mole fraction distribution of malvidin-3-O-glucoside in the presence of caffeine at the pseudo-equilibrium
after a series of pH jumps monitored by stopped flow and (d) in the
absence of caffeine for comparison purposes. Figure c,d, Reprinted with permission from ref (21). Copyright 2019. American
Chemical Society.
Sequence of Reversible
Reactions upon the Addition of the Base to
Equilibrated Solutions of the Flavylium Cation of Malvidin-3-O-glucoside at pH ≤ 1 (Direct pH Jumps)
This scheme is followed by all
anthocyanins. Adapted with permission from ref (19). Copyright 2022. American
Chemical Society.The quinoidal bases are
kinetic products that disappear through
the hydration reaction, followed by a faster tautomerization (second
step). In the second step, AH and all quinoidal bases (their mole fraction distribution depending
on the pH) are in fast equilibrium on one side and B and Cc and the respective anionic forms on the other side. One
breakthrough fundamental for the understanding of the multi-equilibria
was reported by Brouillard and Dubois,[17] who showed that quinoidal bases do not hydrate and the systems move
toward the equilibrium due to the hydration of the flavylium cation.
The kinetics of the second step is given by eq , where χAH+ corresponds
to the mole fraction of AH in
its equilibrium with A and A and χB corresponds to the mole fraction of B in its equilibrium with B, Cc, and Cc.In eq , the reaction
channel occurring in the basic medium due to the OH– nucleophilic addition was added.[14] When
the formation of the anionic species was not considered, the experimental
data is accounted for by eq .Inspection of eq shows that the rate of the second step tends to zero
with the increase
in pH and the rate of OH– nucleophilic addition
also tends to zero with the increase in pOH. This is a very relevant
result crucial for the rationalization of anthocyanin degradation,
see below.The system reaches the equilibrium through the much
slower cis–trans
isomerization. In this case, all species except trans-chalcones are in equilibrium, defining the pseudo-equilibrium (identified
by the superscript ^). During isomerization, the
flavylium cation and all quinoidal bases, hemiketals, and cis-chalcones are in faster equilibrium, allowing for the
deduction of the respective observed rate constant, eq .In eq , ki and k–i are the direct
and reverse rate constants for the formation of Ct from Cc, respectively, χCc is the mole fraction
of Cc at the pseudo-equilibrium, and χCt is the mole fractions of Ct in its equilibrium with Ct, mutatis mutandis, for the isomerization of Cc/Ct.The
equilibrium of the first row of Scheme is accounted for by eqs –7This multi-equilibrium can
be simplified considering a single acid–base
system, eqAccording to eq ,
the complex multi-equilibrium given by eqs –7 is equivalent
to the single acid–base equilibrium between the flavylium cation
and its conjugated base CB, equal to the sum of A, B, Cc, and Ct,
given in eq .Extending to the neutral species, eqs –14are obtainedA similar simplification can be done to obtain eq with K″a given by eq and [CB–] = [A–] + [B–] + [Cc–] + [Ct–]In general, in acid to slightly
basic solutions, the complex multi-equilibrium
of anthocyanins can be reduced to a diprotic acid with constants K′a and K″a.Considering that the isomerization reaction is by
far the slowest
kinetic step of the system, a pseudo-equilibrium is attained when
the observation time falls in the time range of seconds to a few minutes
as in the stopped flow measurements.In conclusion, in acid medium up to neutral-slightly basic, the
equilibrium and the pseudo-equilibrium are equivalent to a diprotic
acid and the mole fraction distribution of AH, CB, and CB or the equivalent species, at the pseudo-equilibrium substituting K′a and K″a by K^a and K^̂a, are given by eq .
Reverse pH Jumps
Reverse pH jumps
consist in the addition of the acid, generally back to the flavylium
cation at pH ≤ 1, to equilibrated solutions or pseudo-equilibrated
solutions at higher pH values.[20−22] As shown above in eq , the rate of the hydration reaction
increases with the increasing proton concentration. At sufficiently
acidic solutions (namely, at pH ≤ 1), the hydration reaction
becomes faster than tautomerization, the so-called change of regime.[14,20] The reverse pH jumps rates involving all species except trans-chalcones fall in the domain of seconds, requiring
a stopped flow apparatus, while for the trans-chalcones,
they occur in hours and can be monitored using a common spectrophotometer.
In other words, during stopped flow measurements, the trans-chalcone is “frozen”. In Figure ,the trace of a reverse pH jump of malvidin-3-O-glucoside from pH = 5.5 monitored by stopped flow is shown.
Three amplitudes can be observed (Figure a). The first corresponds to the remaining
flavylium cation and quinoidal bases that are converted in the flavylium
cation during the mixing time of the stopped flow (at pH = 5.5, mainly
A). The amplitude of the faster kinetic process corresponds to the
conversion of hemiketal in the flavylium cation, and the amplitude
of the slower kinetics is the conversion of cis-chalcone
in more flavylium cations via hemiketal.[14,20] Normalization of the amplitudes of Figure a gives the respective mole fractions. Fitting
of the mole fractions of the pseudo-equilibrium species as a function
of pH (see below, Figure c) can be achieved by means of eq .with a0 + b0 + c0 = a1 + b1 + c1 = 1.
Figure 1
(a) Trace of the reverse pH jump of malvidin-3-O-glucoside at pH = 5.5; (b) trace of malvidin-3-O-glucoside obtained using a common spectrophotometer; and (c) mole
fraction distribution of the malvidin-3-O-glucoside
species at the equilibrium. The yellow band corresponds to the region
where equilibrium is not attained. Adapted with permission from ref (21). Copyright 219. American
Chemical Society.
(a) Trace of the reverse pH jump of malvidin-3-O-glucoside at pH = 5.5; (b) trace of malvidin-3-O-glucoside obtained using a common spectrophotometer; and (c) mole
fraction distribution of the malvidin-3-O-glucoside
species at the equilibrium. The yellow band corresponds to the region
where equilibrium is not attained. Adapted with permission from ref (21). Copyright 219. American
Chemical Society.Once the coefficients a, b, and c are obtained,
the equilibrium constants of the respective
species are straightforwardly calculated from the relations of eqs and 23.[21]For the neutral speciesAdditionally, for the anionic speciesThe isomerization rate constants are calculated by means of
a reverse
pH jump from the equilibrium carried out using a common spectrophotometer
(see Figure b). The
amplitude of the flavylium at the initial time corresponds to the
sum of AH + A + B + Cc and the amplitude of the very slow process
(i.e., controlled by the isomerization) permits the calculation of
parameters d0 and d1 using eq and consequently Ki and KCt/Ct– using eq . Having obtained all equilibrium constants, the pH-dependent
mole fraction distribution of the species at the equilibrium is straightforwardly
obtained, as shown in Figure c.[14]
Irreversible Equilibrium
Recently,
we reported on an experimental procedure to account for the calculation
of the degradation rate in anthocyanins.[19] It is based on a sequence of pH jumps as follows: (i) direct pH
jump to an extended pH region including the basic medium; (ii) maintaining
the solution at the final pH of the direct pH jump; (iii) taking one
aliquot of (ii) after 1 day and performing a reverse pH jump back
to pH ≤ 1; and (iv) repeating (iii) for different delay times
and calculating the disappearance of the flavylium cation as a function
of time. The degradation fraction for each pH value, χdecomposition, is calculated using eq .A typical result of the flavylium recovery
after a certain delay time is shown in Figure a. It is possible to calculate the fraction
of the flavylium cation that disappeared for each delay time and obtain
the degradation rate for each pH value, as shown in Figure b. Finally, the degradation
rate as a function of pH, given in Figure c, can be calculated. The comparison between
the reversible processes (hydration and OH– nucleophilic
addition) and the degradation rate allows for the detection of a pH
range where the equilibrium is not attained. In fact, as shown in Figure d, in the pH region
highlighted in red, degradation is faster than any other process toward
the equilibrium. The rationalization of the anthocyanin degradation
implies the determination of this pH interval.
Figure 2
(a). Recovery of the
malvidin-3-O-glucoside flavylium
cation after 1 day of delay time, at pH = 11.6, with the red open
circle representing the initial flavylium cation absorption; (b) decomposition
mole fraction as a function of time, at pH = 11.6, calculated using eq , behaves as a mono-exponential
from which the degradation rate constant can be calculated; (c) degradation
rate of malvidin-3-O-glucoside as a function of pH;
and (d) comparison between the reversible reactions, hydration and
OH– nucleophilic addition, and the degradation rate.
Adapted with permission from ref (19). Copyright 2022. American Chemical Society.
(a). Recovery of the
malvidin-3-O-glucoside flavylium
cation after 1 day of delay time, at pH = 11.6, with the red open
circle representing the initial flavylium cation absorption; (b) decomposition
mole fraction as a function of time, at pH = 11.6, calculated using eq , behaves as a mono-exponential
from which the degradation rate constant can be calculated; (c) degradation
rate of malvidin-3-O-glucoside as a function of pH;
and (d) comparison between the reversible reactions, hydration and
OH– nucleophilic addition, and the degradation rate.
Adapted with permission from ref (19). Copyright 2022. American Chemical Society.The question is how to quantify the effect of intermolecular
copigmentation[23−25] in anthocyanin degradation. In this work, the copigmentation
between
malvidin-3-O-glucoside and caffeine was used to illustrate
this phenomenon. In particular, caffeine was selected to avoid cinnamic
acids, the most common copigments, which have two forms at the working
pH range, neutral and negatively charged, interacting differently
with the several anthocyanin species, as shown in Scheme .[26] These copigments would introduce an additional level of complexity
that would compromise the clarity of the exposition.[27]
Results and Discussion
Reversible Processes
In Figure , the titration of
the quinoidal bases of malvidin-3-O-glucoside in
the presence of caffeine 0.05 M was monitored
by stopped flow 10 ms after a direct pH jump is reported. Fitting
was achieved for pKa = 3.5, pKA/A– =
6.7, and pKA–/A2– = 8.3. This allows for
calculating the mole fraction distribution of the quinoidal species
(first column of Scheme ) in the presence of the copigment, as shown in Figure b. When the acid base constants
in the absence of caffeine, pKa = 3.8, pKA/A– = 6.3, and pKA–/A2– = 8.4, are compared
with those in its presence, there is an evident decrease of pKa in the presence of caffeine. This indicates that the quinoidal
base is more stabilized than the flavylium cation by caffeine and
its pH domain is also extended to higher pH values. This result contrasts
with the interaction of malvidin-3-O-glucoside with
lignosulfonate, where pKa increases to 4.4 due to the stabilization
of the flavylium cation by the negatively charged lignosulfonate.[28](a) Titration curves of malvidin-3-O-glucoside
in the presence of caffeine 0.05 M collected by stopped flow 10 ms
after a direct pH jump from the flavylium cation at pH = 1 at three
representative wavelengths. Fitting was achieved for pKa = 3.5, pKA/A– = 6.7, and pKA–/A2– = 8.3; (b) mole fraction distribution
of the malvidin-3-O-glucoside species, flavylium
cation, and quinoidal bases in the presence of caffeine taken 10 ms
after a direct pH jump; and (c) mole fraction distribution of malvidin-3-O-glucoside in the presence of caffeine at the pseudo-equilibrium
after a series of pH jumps monitored by stopped flow and (d) in the
absence of caffeine for comparison purposes. Figure c,d, Reprinted with permission from ref (21). Copyright 2019. American
Chemical Society.The stabilization of
the quinoidal base by the caffeine is evident
in the comparison of Figure c with Figure d, where the mole fraction distribution of the species at the pseudo-equilibrium
in the absence of caffeine is reported. It is worth noting that in
the reverse pH jump experiments shown in Figure c,d, the parameters are as follows: a1 = 1 and b1 = c1 = 0. This was an unexpected result because it suggests that
near the neutrality, the only observed species is the quinoidal base.
The reason, discussed below in detail, is that in a pH region between
neutral and slightly basic, the irreversible processes involving quinoidal
bases are the rate-determining step of their relatively slow disappearance.The pH-dependent rate constants of the second kinetic step (controlled
by the hydration reaction in the acidic medium and OH– nucleophilic addition in the basic medium) of malvidin-3-O-glucoside in the presence of caffeine 0.05 M as a function
of pH are shown in Figure a.
Figure 4
(a) Hydration reaction of malvidin-3-O-glucoside
in the presence of caffeine 0.05 M. Fitting was achieved by means
of eq for kh = 0.045 s–1, k–h = 150 M–1s–1, and the values of the protonation constants reported in Figure a; (b) relative energy
level of B and A in the absence and presence of caffeine.
(a) Hydration reaction of malvidin-3-O-glucoside
in the presence of caffeine 0.05 M. Fitting was achieved by means
of eq for kh = 0.045 s–1, k–h = 150 M–1s–1, and the values of the protonation constants reported in Figure a; (b) relative energy
level of B and A in the absence and presence of caffeine.Considering that the formation of B– and
Cc– does not take place since they fall in the pH
range
where the degradation reactions of A– and A2– are the rate-determining steps (see below), eq is simplified to eq , and fitting was achieved
for kh = 0.045 s–1, k–h = 150 M–1s–1, kOH = 0.03 M–1s–1, and the values of the protonation constants reported
in Figure a.When the
hydration and OH– nucleophilic addition
rate constants are compared with those obtained in the absence of
caffeine, kh = 0.12 s–1, k–h = 35 M–1s–1, and kOH = 0.09
M–1s–1, and it can be concluded
that the stabilization by intermolecular copigmentation decreases
the rate constant kh and increases k–h, an effect already observed in intermolecular
copigmentation of acylated anthocyanins.[19] The OH– nucleophilic addition is also significantly
reduced by the intermolecular copigmentation. Figure b shows the relative energy levels of the
quinoidal base and hemiketal in the presence and absence of the copigment,
calculated as reported in the literature.[14] The copigmentation stabilizes the quinoidal base and destabilizes
the hemiketal. Clearly, it is a process to increase the color given
by the quinoidal base, as can be visualized when comparing Figure c,d.
Irreversible Processes
The rate constants
of the irreversible processes of malvidin-3-O-glucoside
in the presence of caffeine 0.05 M were calculated and are represented
in Figure a. Fitting
was achieved by means of eq , where χ represents the mole fraction distribution
of the quinoidal bases calculated in Figure b and kirr. the respective rate constants (degradation constants) with kirr.(A) = 2.5 × 10–6 s–1, kirr.(A−) = 5.5
× 10–6 s–1, and kirr.(A2−) = 6 × 10–7 s–1
Figure 5
(a)
Rate constants of the irreversible processes of malvidin-3-O-glucoside in the presence of caffeine 0.05 M. (b) Representation
of the hydration, OH– nucleophilic addition, and
degradation rates of malvidin-3-O-glucoside in the
absence (blue traced lines) and in the presence of caffeine 0.05 M
(black full lines), eq (red band) and eq (higher pH values); the red band corresponds to the pH region where
the equilibrium is not attained. (c) Contribution for the hydration
reaction rate of the term dependent on kh (direct) and the term depending on k–h (reverse) in eq .
(a)
Rate constants of the irreversible processes of malvidin-3-O-glucoside in the presence of caffeine 0.05 M. (b) Representation
of the hydration, OH– nucleophilic addition, and
degradation rates of malvidin-3-O-glucoside in the
absence (blue traced lines) and in the presence of caffeine 0.05 M
(black full lines), eq (red band) and eq (higher pH values); the red band corresponds to the pH region where
the equilibrium is not attained. (c) Contribution for the hydration
reaction rate of the term dependent on kh (direct) and the term depending on k–h (reverse) in eq .In the basic medium, the fitting
was performed considering a similar
process, given by eq , involving most probably Ct2– and Ct3– with pK = 11.3, kirr.(Ct2−) =
7 × 10–7 s–1, and kirr.(Ct3−) = 3 × 10–6 s–1, but there is a huge uncertainty in these
values.In Figure b, the
hydration and OH– nucleophilic addition rates are
represented together with the rates of the irreversible processes
in the absence and presence of caffeine 0.05 M. Some conclusions can
be taken from this figure. The hydration in the presence of the copigment
is shifted to higher pH values. Representing the contributions for
the hydration reaction rate of the term dependent on kh (direct) and the term depending on k–h (reverse) in eq , as shown in Figure c, it can be concluded that there is an inversion.
The hydration rate in the presence of the copigment is shifted to
higher pH values, in comparison with the situation in the absence
of the copigment, only for pH > 6.5. At these pH values, the influence
of the constant k–h is higher.
The shift to higher pH values of the hydration curve in the presence
of the copigment is explained by considering that the k–h increases from 35 to 150 M–1s–1 in the presence of the copigment.The
OH– nucleophilic addition is shifted to higher
pH values, meaning that a higher OH– concentration
is necessary for hydroxylation to occur toward the equilibrium. The
degradation rates of the trans-chalcones are also
significantly decreased in the presence of the copigment caffeine.
Copigmentation at pH = 8.4
In the
pH range where degradation is the controlling step, the rate of the
anionic quinoidal basis disappearance is slow enough to allow for
the determination of the copigmentation constant, as shown in Figure . Caffeine addition
to a malvidin-3-O-glucoside solution 8.4 × 10–6 M at pH = 8.4 gives rise to a red shift and a hyperchromic
effect, reinforcing the blue hue. Representation of the absorbance
at 640 nm as a function of the caffeine concentration can be fitted
for a 1:1 copigmentation with an association constant of 1.5 ×
103 M–1 (estimated 5% error). For the
caffeine concentration of 0.05 M, 99% of the malvidin-3-O-glucoside is involved in the copigmentation process. Despite the
fact that degradation is significantly reduced in these conditions,
it is still not enough for applications requiring relatively long
storage times.
Figure 6
(a) Spectral variations upon the addition of caffeine
(pH = 8.4)
to a malvidin-3-O-glucoside solution at the same
pH and (b) variation of the absorbance at 640 nm vs caffeine concentration.
Fitting was achieved for a 1:1 copigmentation with an association
constant of 1.5 × 103 M–1 (estimated
5% error).
(a) Spectral variations upon the addition of caffeine
(pH = 8.4)
to a malvidin-3-O-glucoside solution at the same
pH and (b) variation of the absorbance at 640 nm vs caffeine concentration.
Fitting was achieved for a 1:1 copigmentation with an association
constant of 1.5 × 103 M–1 (estimated
5% error).
Experimental
Section
Malvidin-3-O-glucoside was purchased
from Extrasynthese,
and caffeine was obtained from Alfa Aesar Company Ltd. The Teorell
and Stenhagen universal buffer[29] was prepared
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 a 1 M NaOH solution
in Millipore water until a volume of 1 L.For the calculation
of pK^a, pK′a, and the
pKs of the ionized forms, direct pH jumps were carried
out by mixing the flavylium stock solution (1 × 10–4 M at pH = 1), NaOH to neutralize the amount of HCl added, buffer
at the desired pH (with and without caffeine), and water (or caffeine
0.1 M in water). Reverse pH jumps were performed by adding enough
HCl (with or without caffeine) to equilibrated or pseudo-equilibrated
solutions to reach pH = 1. pH was recorded on a Radiometer Copenhagen
PHM240 pH/ion meter (Brønshøj, Denmark).UV–vis
spectra were recorded on a Varian-Cary 100 Bio or
5000 spectrophotometer (Palo Alto, CA, USA). The reverse pH jumps
from the pseudo-equilibrium were observed on an SX20 (Applied Photophysics;
Surrey, UK) spectrometer equipped with a PDA.1/UV photodiode array
detector. A filter of 385 nm was used to prevent the photochemical
reactions that could take place from Cc or B.
Conclusions
Intermolecular copigmentation of malvidin-3-O-glucoside
with caffeine affects both reversible and irreversible processes.
It destabilizes hemiketal and stabilizes the colored quinoidal base.
In the presence of the copigment, there is still a pH range where
degradation of quinoidal bases is the faster process. Intermolecular
copigmentation with caffeine is able to decrease the malvidin-3-O-glucoside degradation rate by a few folds. While this
effect is not enough for applications requiring a relatively long
storage time, it constitutes a basic knowledge to find a possible
strategy to stabilize the color at neutral to slightly basic pH values.
In this pH region, the kinetics of both hydration and OH– nucleophilic addition are very slow. Protection of the quinoidal
bases in this pH region could be a way to increase the lifetime of
the blue color given by anthocyanins. Studies using cinnamic acids
as copigments and degradation of acylated anthocyanins are in progress.
Authors: Patrick Trouillas; Juan C Sancho-García; Victor De Freitas; Johannes Gierschner; Michal Otyepka; Olivier Dangles Journal: Chem Rev Date: 2016-03-09 Impact factor: 60.622