Thermodynamic and Kinetic Properties of a Red Wine Pigment

Oct 6, 2010 - Formation of pyranoanthocyanins in red wines: a new and diverse class of anthocyanin derivatives. Victor Freitas , Nuno Mateus. Analytic...
1 downloads 0 Views 796KB Size
J. Phys. Chem. B 2010, 114, 13487–13496

13487

Thermodynamic and Kinetic Properties of a Red Wine Pigment: Catechin-(4,8)-malvidin-3-O-glucoside Frederico Nave,† Vesselin Petrov,‡ Fernando Pina,*,‡ Nate´rcia Teixeira,† Nuno Mateus,† and Victor de Freitas*,† Centro de InVestigac¸a˜o em Quı´mica, Departamento de Quı´mica, Faculdade de Cieˆncias, UniVersidade do Porto, Rua do Campo Alegre 687, 4169-007 Porto, Portugal, and REQUIMTE, Departamento de Quı´mica, Faculdade de Cieˆncias e Tecnologia, UniVersidade NoVa de Lisboa, 2829-516 Caparica, Portugal ReceiVed: May 24, 2010; ReVised Manuscript ReceiVed: July 30, 2010

Catechin-(4,8)-malvidin-3-glucoside, a red pigment adduct (at acid pH) found in red wine and resulting from the reaction between malvidin-3-glucoside and flavan-3-ols during wine aging, was synthesized. The thermodynamic and kinetic constants of the network of chemical reactions were fully determined by stopped flow: (i) Direct pH jumps, from thermal equilibrated solutions at pH ) 1.0 (flavylium cation, AH+), show three kinetic processes. The first one occurs within the mixing time of the stopped flow and leads to the formation of quinoidal bases A and/or A- depending on the final pH; the second one takes place with a rate constant equal to 0.075 + 33[H+] and was attributed to the hydration reaction that forms the pseudobases (hemiketals), B/B-. The third process is much slower, 2 × 10-4 s-1, and is due to the cis-trans isomerization giving rise to a small fraction of trans-chalcones, Ct/Ct-. (ii) Reverse pH jumps from the thermally equilibrated solution at moderate to neutral pH values back to a sufficiently acidic medium clearly distinguish three kinetic processes: the first one takes place within the dead time and is due to the protonation of the bases; the second process occurs with the same rate constant of the hydration reaction monitored by direct pH jumps and is attributed to the formation of flavylium cation from the B; the last process occurs with a rate constant of 1.8 s-1, and results from the formation of AH+ from Ct through B, reflecting the rate of the ring closure (tautomerization). The separation of the hydration from the tautomerization upon a reverse pH jump is only possible because at pH < 1 the former reaction is faster than the last. An identical situation was observed for malvidin-3-glucoside (oenin) for pH < 2. Introduction +

Flavanol-(4,8)-anthocyanin adducts (F-A ) are one of the numerous products of direct condensation between flavan-3ols and anthocyanins that are able to occur in red wine.1-3 In the past decade, these compounds have also been identified in strawberries,4 beans, grapes,5 and corn.5,6 Flavanol-(4,8)-anthocyanin adducts are described to be a flavan-3-ol linked through a B type interflavanic bond to an anthocyanin. The bond occurs between a C4 carbon of the first and the C8 carbon of the latter, Scheme 1. These compounds present multifarious chemical characteristics of both flavan-3-ols (nucleophilicity at the C6 and C8 carbons of the upper unit) and anthocyanin (electrophilicity at the C2 and C4 carbons of the lower unit).7-9 When compared to their anthocyanin analogues, the flavanol-(4,8)-anthocyanin adducts present different color properties and higher resistance to thiolysis3 but equal vulnerability to sulphite bleaching and hydration.8 Despite these slightly different properties, the flavanol-(4,8)anthocyanin adducts could also endure, like anthocyanins, as dyes,10 solar energy converters,11 or food colorants.12 In the latter case, besides the chromatic features, these compounds can also present higher antioxidant properties than anthocyanins due to

the polymeric flavan-3-ol moiety. To study such applications it is important to chart their chemical behavior in aqueous solutions.12 Similarly to their anthocyanic precursors, in aqueous solutions, the anthocyanic lower unit of the F-A+ adducts suffers a series of pH dependent structural transformations which affect the visible spectrum displayed by the molecule. These transformations have been studied by several research groups in the past decades using both natural and synthetic anthocyanins. Good examples are, among others, the seminal works from Brouillard et al.,13-18 McClelland et al.,19-23 and Asenstorfer et al.,24,25 and more recently the contributions from some of us in particular which regard the photochemistry of these molecules.26-31 The general equilibria taking place in aqueous medium are shown in Scheme 2. In acidic media the most stable species is SCHEME 1

* Corresponding authors. E-mail: [email protected] (V.d.F.); fjp@ dq.fct.unl.pt (F.P.). † Universidade do Porto. ‡ Universidade Nova de Lisboa.

10.1021/jp104749f  2010 American Chemical Society Published on Web 10/06/2010

13488

J. Phys. Chem. B, Vol. 114, No. 42, 2010

Nave et al.

SCHEME 2

the flavylium cation (AH+). As the pH is raised, the base (A), the hemiketal (B), and the cis- (Cc) and trans- (Ct) chalcones are formed. Further rise of pH (toward basic medium) leads to formation of the respective ionized forms from deprotonation of the hydroxyl substituents. Prior to reaching thermodynamic equilibrium, kinetic processes lead to the formation of transient species. For example, proton transfer is faster than the other processes, and consequently, if a solution of flavylium cation (stable in acidic medium) becomes less acidic, the base A is immediately formed. However, in the case of anthocyanins this is not the most stable species at equilibrium, and thus, A tends to disappear to form the most stable B. This species is in fast equilibrium with some Cc (minor species). Finally, the slowest process is the cis-trans isomerization to lead to a small mole fraction of Ct at equilibrium. One useful way to follow the kinetic process occurring in the network of flavylium compounds is the use of relaxation techniques, in which the equilibrium of the system is shifted by means of an external influence, applied very rapidly, forcing the metastable system thus formed to shift to the new state of equilibrium. The most common perturbations used are instantaneous variations of temperature, pressure or pH, the latter being more advantageous since it does not produce secondary physical effects.13 The application of these relaxation methods allows the determination of both equilibrium constants and the reaction rate constants.31 It is expected that the flavanol-(4,8)-anthocyanin adducts present similar behavior in aqueous solutions. It was thus interesting to evaluate the changes in the aqueous equilibria of oenin induced by the flavan-3-ol substitution at the carbon C8 of the flavanic skeleton.

Experimental Section Materials. Acetonitrile and ethanol used were HPLC grade. All other solvents used were analytical grade. All aqueous solutions were prepared with ultrapure water, obtained by passing distilled water through a Millipore system (Simplicity 185). This system allows us to obtain type 1 ultrapure water. (+)-Taxifolin (purity 98%) was purchased from Extrasynthe`se, and (-)-epicatechin (purity 98%) and sodium borohydride (purity 98%) were purchased from Sigma-Aldrich. Malvidin-3-O-glucoside (Mv3glc) was extracted and purified in the laboratory from a young red wine (varietal caste Touriga Nacional) by semipreparative HPLC with C18 reversed phase column, as described elsewhere.7,32 Procyanidin B4 dimer was hemisynthesized, as previously described.7,33 A universal buffer of Theorell and Stenhagen34 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 water (until 1 L completion). Hemisynthesis of the F-A+ Adducts (Cat-(4,8)-Mv3glc). The hemisynthesis of the flavanol-(4,8)-anthocyanin adducts, in this case Cat-(4,8)-Mv3glc, was performed as previously described.7 Briefly, in a 50 mL Schlenk flask, 60 mg of Mv3glc was placed with 4 mg of procyanidin B4 in 45 mL of a biphasic system of 10% formic acid aqueous solution and ethyl ether (1:2), under argon atmosphere. The Schlenk flask was heated at 85 °C for 3 h, under magnetic stirring. At that point, the organic phase was discarded, and 4 mg of procyanidin and 30 mL of ethyl ether were added prior to a new heating cycle. The successive additions of procyanidin and ether were repeated 4 times.

Catechin-(4,8)-malvidin-3-O-glucoside The resulting aqueous phase was evaporated in a rotary evaporator (T ∼ 35 °C) and fractionated in 3 steps. The first fractionation consisted of the application of the reaction mixture into 50 g of Toyopearl HW40(s) placed in a medium porosity sintered glass funnel and elution with 20%, 80%, and 100% (v/v) of acidified methanol. The fraction of 80% acidified methanolic solution was refractionated by column chromatography (160 mm × 16 mm i.d.), using 30 g of Toyopearl HW40(s) as stationary phase and 40%, 75%, and 100% (v/v) of acidified methanol as mobile phase. The flow rate was set at 0.8 mL min-1, and the fraction collection was made upon visual detection of the colored bands. The identification of the compounds was made by HPLC-DAD. The fraction corresponding to Cat-Mv3glc was further fractionated by preparative RP-C18 HPLC. The purity of Cat-Mv (>95%) was assessed by ESI-MS and NMR. HPLC. Two HPLC systems were used throughout the present work: an analytical system and a preparative system. The analytical HPLC system was composed by an Elite LaChrom chromatograph fitted with an L-2130 quaternary pump, an L-2200 autosampler, an L-2300 column oven, and an L-2455 diode array detector (DAD). The injection volume used was 20 µL. The stationary phase was composed by a Merck RPC18 column (250 mm × 4 mm i.d.., 5 µm pore size) and the mobile phase by a 2 solvent system: solvent A, a 10% (v/v) aqueous formic acid solution, and solvent B, a formic acid/ acetonitrile/water (1:3:6) solution. The flow rate was set at 1 mL min-1 with the following separation program: a linear gradient from 20% of B to 66% of B in 50 min followed by a new linear gradient to 100% of B for 10 min succeeded by an isocratic elution of 100% of B for 10 min. Column washing was carried out by an isocratic elution with 100% of solvent C for a 7 min period followed by a 15 min re-equilibration stage with the initial conditions. The absorbance spectra were registered from 250 to 750 nm, and the separations were followed at 280 and 520 nm. The preparative HPLC system was composed by a Knauer K-1001 quaternary pump, a rheodyne 7125 manual injector with a 200 µL loop, and a Merck Hitachi L-2420 UV-vis detector. The stationary phase was composed by a Thermofinnigan Hypersil Gold column (250 mm × 10 mm i.d., 5 µm pore size) and the mobile phase composed by solvent A, 10% (v/v) of formic acid, and solvent B, an aqueous solution of 10% (v/v) formic acid and 30% (v/v) methanol. The flow rate was 4 mL min-1, and the gradient used was the same as the analytical one. The separation was followed at 520 nm. 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 7.00 and pH 4.00 purchased from Crison. Absorption Spectroscopy. The UV-vis absorption spectra were recorded in a Varian-Cary 100 Bio spectrophotometer, and the determination of the isomerization equilibrium was made with a Varian-Cary 5000i, both using a quartz cell cuvette of 800 µL capacity and 1 cm of optical length. Stopped Flow Equipment. The stopped flow experiments were conducted in a spectrophotometric SFM-300 mixing chamber, controlled by an MPS-60 unit (Bio-Logic). The data were collected by a TIDAS diode array (J&M), with wavelength range between 300 and 800 nm. The standard cuvette has an observation path length of 1 cm. For these experiments, the dead time of each shot was calculated to be ∼6 ms. Equilibrium and Kinetic Constants. All thermodynamic and kinetic constants were determined by spectrophotometric meth-

J. Phys. Chem. B, Vol. 114, No. 42, 2010 13489 ods, although with different techniques. The mother solution of the F-A+ adduct (1.18 × 10-4 M) was prepared using a solution of HCl 0.1 M, which was protected from the sunlight at room temperature. All solutions analyzed were prepared from this mother solution. The low pH value of the mother solution ensured that most of the F-A+ adducts were in the flavylium form and thus prevent pigment degradation. The determination of the apparent acidity constant (thermodynamic pseudoequilibrium) was accomplished by measuring, in the common UV-vis spectrophotometer, solutions composed of 1/3 of CatMv3glc mother solution, 1/3 of solution of NaOH (0.1 M), and 1 /3 of universal buffer solution (with 30% ethanol) at different pH values. Results and Discussion Anthocyanins, like many other synthetic flavylium compounds, follow the same basic network of chemical reactions.12,18-22,27,28,31,35,36 In this work, evidence for the formation of ionized forms (anions) was achieved at moderately acidic to neutral solutions, Scheme 2. The network was studied by performing in thermally equilibrated solutions pH jumps from acidic (direct pH jumps) to basic and vice versa (reverse pH jumps). The terms “direct” and “reverse” are used in order to distinguish the two types of pH jumps. Proton transfer is much faster than any other reaction of the network, and consequently, it is the first species to appear after a direct pH jump if the final pH is high enough to form a base, a requirement that depends solely on the pKa. Direct pH jumps carried out in flavylium compounds lacking hydroxyl substituents in the pyranic ring of the flavonoid structure lead to the formation of the pseudobases B2 (hydration in carbon 2) as well as B4 (hydration in carbon 4) as the primary products of the pH jump.19 In the case of the present compound (Cat-Mv3glc), B4 was not observed. Pseudobases can also be the first species to appear even in a flavylium derivative possessing hydroxy substituents, as long as the pKa of the compound is substantially higher than the pKh. However, the hydration reaction to give B2 is much slower than the proton transfer (subseconds to seconds compared with microseconds, respectively).31,37 Therefore, it is possible to measure the absorption spectra of the quinoidal bases, using a stopped flow apparatus, after a few milliseconds (mixing time) before formation of the pseudobases. As further discussed in this work, the formation of Cc from B takes place on a time scale of subseconds, and consequently, during a direct pH jump these two species could be considered to be in fast equilibrium because at moderately acidic medium hydration is much slower than tautomerization. The thermodynamic equilibrium is reached by the appearance of transchalcone (Ct) via isomerization of cis-chalcone (Cc) as was proposed by some authors in synthetic and natural flavylium derivatives.12,18-23,27,28,31,35,36 The identification in oenin of the structures of A, B, and Cc and Ct in equilibrium was firmly established by Houbier et al (1998) through 1H NMR.12 Preston and Timberlake (1981) have separated both Cc and Ct of oenin by HPLC.36 They have stated that, in the case of oenin and malvin, the molar fraction of Ct at the equilibrium is relatively small at room temperature, opposite of some synthetic flavylium cations where Ct, rather than B, is the dominant species at moderately acidic pH values.29 As previously reported in acidic media, the following equations account for the thermodynamic and kinetic processes:14,28,38

13490

J. Phys. Chem. B, Vol. 114, No. 42, 2010

AH+ + H2O h A + H3O+

Ka

Nave et al.

B- h Cc-

proton transfer

Kt-

tautomerization

(13)

(1) +

+

AH + 2H2O h B + H3O B h Cc Cc h Ct

Kt

Kh

(2)

hydration

tautomerization

(3)

isomerization

(4)

Ki

Equations 1-4 can be substituted by a single acid-base equilibrium with constant K′a. CB stands for conjugated base.

AH+ + H2O h CB + H3O+

K'a )

Kt- )

[CB][H3O+] [AH ]

χAH+ )

[CB] ) [A] + [B] + [Cc] + [Ct] (6)

However, it is possible to detect, by careful analysis of the absorption spectrum in a moderately acidic medium, the formation of ionized base (A-). Considering the possibility of deprotonation of the other species, eqs 7-10 can be written as follows: -

+

A + H2O h A + H3O

KA-

[H+]2 [H+]2 + K′a[H+] + K′aK′′a Ka[H+]

[H+]2 + K′a[H+] + K′aK′′a

(5) K'a ) Ka + Kh + KhKt + KhKtKi

χA- )

KA-Ka [H ] + K′a[H+] + K′aK′′a + 2

Kh[H+]

χB )

[H+]2 + K′a[H+] + K′aK′′a

χB- )

proton transfer

KB-Kh [H+]2 + K′a[H+] + K′aK′′a

(7) -

KhKt[H+]

χCc )

+

B + H2O h B + H3O

KB-

(14)

Thermodynamic Equilibrium. The mole fraction distribution of the several species appearing in solution can now be determined by eqs 15-23.

χA )

+

KCcKB-

[H+]2 + K′a[H+] + K′aK′′a

proton transfer

(15)

(16)

(17)

(18)

(19)

(20)

(8) -

+

Cc + H2O h Cc + H3O

KCc-

χCc- )

proton transfer

KCc-KhKt [H ] + K′a[H+] + K′aK′′a + 2

(21)

(9) -

+

Ct + H2O h Ct + H3O

KCt-

χCt)

proton transfer

(10) It is easy to demonstrate10,28 that eqs 7-10 can be simplified as a second acid-base equilibrium, as in eq 11:

-

+

CB + H2O h CB + H3O

K′′a )

[CB-][H3O+] K''a ) [CB] (11)

KA-Ka + KB-Kh + KCc-KtKh + KCt-KhKtKi K′a [CB-] ) [A-] + [B-] + [Cc-] + [Ct-] (12)

Accordingly, the network of the flavylium system including the ionized species CB- can be accounted for by considering that AH+ is a diprotic acid yielding CB upon releasing the first proton and CB- upon releasing the second one. There are other alternatives to writing eqs 7-10, but not all are linearly independent. This example follows:

KhKtKi[H+] [H+]2 + K′a[H+] + K′aK′′a

χCt- )

KCt-KhKtKi [H ] + K′a[H+] + K′aK′′a + 2

(22)

(23)

Proton Transfer To Form Quinoidal Bases. The rate of the proton transfer to give the quinoidal bases is too fast to be studied by stopped flow.13,14 However, through this technique, the absorption spectra of the flavylium cation and the respective quinoidal bases A and A- can be monitored after 10 ms upon the direct pH jump before formation of significant amounts of the pseudobase B2 (Figure 1). The spectra suggest the presence of two successive isosbestic points (580 nm, 590 nm), and the data can be fitted considering two acid-base consecutive processes with pKa ) 3.8 and pKA- ) 5.9 (see inset in Figure 1I). Taking the values of Ka and KA- obtained through the fitting, the absorption spectra of the species AH+, A, and A- can be calculated by mathematical decomposition, see Figure 1II. In particular, it is useful to retain for further discussion the ratio of the absorption of these three species at 535 nm, 1:0.73:0.41, respectively, for AH+, A, and A-.

Catechin-(4,8)-malvidin-3-O-glucoside

J. Phys. Chem. B, Vol. 114, No. 42, 2010 13491

Figure 1. (I) Absorption spectra of the Cat-Mv3glc obtained after a pH jump from acidic equilibrated solutions at pH ) 1.0 (AH+) to higher pH values (3.00, 3.59, 4.50, 5.70, 5.96, 6.25, and 7.15, respectively), obtained by stopped flow 10 ms after mixing. The data can be fitted assuming two consecutive acid-base equilibria with pKa1 ) 3.8 and pKa2 ) 5.9. (II) Spectra of the components obtained by mathematical decomposition.

Figure 2. (I) Spectral changes of the Cat-Mv3glc as a function of time between 10 ms and 9 s, due to a direct pH jump from pH ) 1 to pH ) 2.45, monitored by stopped flow. (II) The same from pH ) 1 to pH ) 4.5.

Hydration and Isomerization. The spectral variations taking place from 10 ms after the pH jump up to reach the equilibrium are reported in Figures 2 and 3I. Two processes have been clearly identified, the faster being very pH dependent. Immediately after the pH jump to pH ) 2.45 only a small amount of the base A is formed (pKa ) 3.8), and thus, Figure 2I regards essentially the disappearance of AH+ (note that the stopped flow apparatus used was limited to the visible region). At higher pH values, as shown in Figure 2II, the bases (essentially A at this pH value) are formed immediately after the pH jump to pH ) 4.5. However, as shown by Dubois and Brouillard, the evolution of the system to the equilibrium takes place through the hydration of AH+, and not from A. Consequently, the formation rate of B2 from AH+ becomes slow at higher pH values where the quinonoidal species (A and A-) are dominant. The AH+ available to hydrate depends on its mole fraction, [H+]2/[H+]2 + Ka[H+] + KaKA-. On the other hand, the hydration also depends on the proton concentration through the term k-h[H+] and could thus become faster than tautometization in very acidic solutions. However, direct pH jumps from AH+ to very acidic

pH values ([H+]2 . Ka[H+] + KaKA-) are useless for observing the variation of the rate constant of hydration because at higher proton concentrations AH+ is the stable species at the starting and final pH values and no spectral modifications can take place. The situation is overcome, and kinetic processes exhibiting the hydration faster than tautomerization can be achieved through the so-called reverse pH jumps, where at sufficiently acidic final pH formation of AH+ from B2 is observed, as further discussed. Direct pH jumps followed in a longer scale of time (using a common spectrophotometer) put in evidence the existence of a much slower process that was assigned to the Cc-Ct isomerization (Figures 3 and 4). Four pH jumps were performed and followed during several hours in order to study the isomerization process. An example of one of those experiments is depicted in Figure 3. Taking the absorbance values of the consecutive spectra at λ ) 535 nm and plotting them against time, one can obtain the graph from Figure 3II (b). The fitting of the biexponential curve allowed the determination of the observed rate constants (kobs) of the hydration (fastest process, see inset) and the isomerization. The

13492

J. Phys. Chem. B, Vol. 114, No. 42, 2010

Nave et al.

Figure 3. (I) Spectral variations followed during 5 h of the Cat-Mv3glc that followed a direct pH jump from 1 to 5.9. (II) Traces of the absorption versus time evidencing the existence of biexponential decay. Inset is the magnification of the fastest process.

kobs ) χAH+kh + χBk-h[H+]

Figure 4. pH dependence of the observed rate constants of the three kinetic processes occurring in the network of the Cat-Mv3glc. Circles correspond to data obtained from reversed pH jumps, and squares and triangles correspond to data obtained from pH jumps: (9) and (b) hydration reaction, (O) tautomerization reaction, (2) cis-trans isomerization.

values of the isomerization observed rate constants are represented in Figure 4 as black triangles as a function of final pH of the pH jump performed. It is possible to conclude that the isomerization occurs with a mean lifetime of 1.4 h [1/kobs in secondary axis (right vertical axis) of Figure 4]. The initial part of the fitting of the biexponential curve (inset in Figure 3II) allows the calculation of the observed rate constant of the hydration process for the same pH values, black squares of Figure 4. Summarizing, Figure 4 displays the rate constants of the three processes, hydration, tautomerization, and isomerization, represented as a function of pH obtained by pH jump techniques: black squares for hydration and black triangles for Cc-Ct isomerization obtained by direct pH jumps. The other values of hydration (b) and tautomerization (O) were obtained by reversed pH jumps as described below. The rate constant of the hydration is expected to exhibit a variation according to eq 24.

(24)

Equation 24 takes into account that the hydration reaction is preceded by a faster reaction (AH+ h A + H+) and followed by another one (B h Cc) which is also faster. At this stage, formation of Ct can be neglected since it occurs in a much higher scale of time. On this basis, the hydration reaction is equivalent to a reversible first order process whose rate constant is the sum of the forward and backward reactions.39 However, different from a single reversible first order reaction, in the case of the hydration, both AH+ and B2 are involved in other faster equilibria (comparatively to the hydration) that reduce their concentration, and by consequence, each rate constant should be multiplied by the respective mole fraction. Fitting of the data using eq 24 was achieved for kh ) 0.075 s-1 and χBk-h ) 33 s-1 M-1. Thermal Equilibrium. The absorption variations of the compound after thermal equilibration are displayed in Figure 5. This figure shows the disappearance of the characteristic flavylium cation absorption (533 nm) as pH is raised. At moderately acidic-neutral pH values the absorption band suggests the presence of B2 (280 nm) as a major species in equilibrium with Cc (see below for evidence for B2 and Cc) as well as a small absorption, around 600 nm, that can be attributed to a minor fraction of quinoidal bases. It is interesting to note the formation of the ionized quinoidal base (A-) at higher pH values (Figure 5I, inset), by comparison with the position and shape of the data obtained by stopped flow as reported in Figure 1. The data represented in Figure 5 allow us to calculate pK′a ) 2.5, which is comparable with the values determined (2.3) for oenin, and pK′′a) 6.2. From eqs 6 and 12 the following relations are obtained.

Ka′ ) Ka + Kh + KhKt + KhKtKi ) 10-2.5

(6a)

K″a ) KA-Ka + KB-Kh + KCc-KtKh + KCt-KhKtKi ) 10-8.7

(12a)

The fitting at 535 nm (Figure 5II) was achieved considering the mole fraction distribution of AH+, CB, and CB- defined

Catechin-(4,8)-malvidin-3-O-glucoside

J. Phys. Chem. B, Vol. 114, No. 42, 2010 13493

Figure 5. (I) Absorption spectra of the Cat-Mv3glc at different pH values upon equilibration. (II) Fitting of the absorbance at λ ) 535 nm was achieved for pK′a ) 2.5 and pK′′a ) 6.2.

through the two acid-base equilibria through pK′a and pK′′a according to eq 25:

A(535nm) ) 1 × 0.53 × χAH+ + 0.08 × 0.73 × 0.53 × χCB + 0.16 × 0.41 × 0.53χCB-

TABLE 1: Thermodynamic Constants Determined for the Transformations of the F-A+ Adduct pK′a pK′′a pKa Kh Kt Ki pKApKBpKCcpKCt-

(25)

In eq 25, 0.53 is an experimental parameter corresponding to the absorbance of AH+ taken at lower pH values. The values 1, 0.73, and 0.41 are the relative absorbances of AH+/A/A- at 535 nm calculated from the data displayed in Figure 1II. The values 0.08 and 0.16 are the fraction of A and A- at the equilibrium that fit the curve. In other words, only the pKa’s and the fraction of the bases at the equilibrium needed to be fitted. Reverse pH Jumps. More insights on the system were obtained by carrying out the so-called reverse pH jumps (Figure 6). This experiment consists of a pH jump back to pH ) 1 from solutions previously equilibrated at neutral to moderately acidic pH values, in the present case at pH ) 6.9, following the appearance of the flavylium cation by stopped flow. This procedure is very useful because during the mixing time of the stopped flow all the proton transfer reactions take place. In other words, all the ionized species

a

10-2.5(0.1 M-1 10-6.2(0.2 M-1 10-3.8(0.1 M-1 2.3 × 10-3 a M-1 0.13 ( 0.01 0.15 ( 0.03 5.9 ( 0.1 M-1 6.2 ( 0.1 M-1 6.9 ( 0.2 M-1 6.7 ( 0.2 M-1

Estimated error 20%.

(A-, B2-, Cc-, and Ct-, if present) are frozen in the form of the neutral analogs (A, B2, Cc, and Ct). Analysis of Figure 6 shows the existence of three distinct processes that lead to the appearance of flavylium cation. The first absorption appears during the mixing time of the stopped flow and is due to the protonation of the quinoidal bases (A and/or A-) at the equilibrium (at pH ) 4.0 or 6.9). From this point, the new equilibrium is reached by two consecutive first order reactions, one due to the formation of flavylium cation from B2 and the other from Cc through B2. This interpretation follows the one already reported by McClelland and

Figure 6. Reverse pH jumps of Cat-Mv3glc from previously equilibrated solutions at 4.0 (I) and 6.9 (II).

13494

J. Phys. Chem. B, Vol. 114, No. 42, 2010

Nave et al.

TABLE 2: Kinetic Constants Determined for the Transformations of F-A+ Adduct

isomerization is observed (rate determining step of the process). On the other hand, reverse pH jumps are very useful when applied to the anthocyanin equilibrium. Using reverse pH jumps, it is possible to study the kinetics of the system at very low pH values, which is not possible by direct pH jump, as discussed previously. Whenever pH < pKa, the ratio [H+]2/([H+]2 + Ka[H+] + KaKA-) is approximately equal to 1 and the hydration becomes proportional to the proton concentration, eq 24. Moreover, at sufficiently low pH values hydration is expected to become faster than tautomerization, as it was proved below, and the two processes can be separated. In other words, while in direct pH jumps the tautomerization reaction could not be distinguished, because as soon as B2 is formed from AH+ it immediately equilibrates with Cc, reverse pH jumps clearly allow the observation of the tautomerization reaction in a stopped flow apparatus at very low pH values. The rate constants of the hydration and tautomerization reactions at lowest pH determined by reversed pH jumps are also represented in Figure 4, respectively as black circles and open circle. Moreover, the rates of the hydration are determined by fitting the data obtained as described in Figure 6 into eq 24, also used in direct pH jump measurements (inset of Figure 4). Figure 4 shows that the hydration becomes slower than tautomerization for ca. pH > 1.5, and consequently, the pH window to observe by reverse pH jumps both processes separately is limited to pH < 1. For example, the trace from a reversed pH jump to 1.8 is fitted with a single exponential, with a rate constant that is expected for the hydration reaction. In fact, at this pH value, hydration is already the rate determining step.

kh (s-1) k-h (M-1 s-1) kt (s-1) k-t (s-1) ki (s-1) k-i (s-1) kOH (M-1 s-1) kw (s-1) a

0.075 ( 0.01 33 ( 1 M-1 0.23 ( 0.05 1.8 ( 0.3 2.6 × 10-5 a 1.7 × 10-4 a 8 ( 0.5 × 106 0.07

Estimated error 20%.

Figure 7. Mole fraction distribution of the Cat-Mv3glc calculated by means of eqs 15-23 and the data reported in Table 1.

co-workers19,20 in the case of synthetic flavylium compounds bearing a cis-trans isomerization barrier. These authors used the pseudoequilibrium involving the species AH+, A, B2, and Cc that is formed a few minutes before a direct pH jump, prior to the formation of significant amounts of Ct. However, reverse pH jumps are useless when applied to the McClelland synthetic flavylium compounds at the equilibrium because Ct is the major species, and consequetly, only the cis-trans

Another aspect to take into account in the case of the reverse pH jumps is the fact that the rate constant of the tautomerization reaction approximately 1.8 s-1 (at pH 1) accounts for k-t since no reversibility is expected to occur: once B2 formed is transformed into AH+ faster than go back to Cc. As shown in Figure 4, the rate constants kh ) 0.075 s-1, k-h ) 33 s-1, and k-t ) 1.8 s-1 have been directly obtained by the respective fittings. The others were obtained by the calculation kt ) Ktk-t ) 0.23 s-1 and from the relations ki + k-i ) 1.98 × 10-4 s-1 and Ki ) 0.15 with the values of ki ) 2.6 × 10-5 s-1

Figure 8. (I) Reverse pH jumps for oenin from equilibrated solutions at pH ) 5 to pH ) 1.1 at the absorption wavelength of 518 nm. (II) Representation of the two rate constants for a series of reverse pH jumps (solutions equilibrated at pH ) 5).

Catechin-(4,8)-malvidin-3-O-glucoside

J. Phys. Chem. B, Vol. 114, No. 42, 2010 13495

TABLE 3: Thermodynamic Constants Associated with the Transformations of Oenin (in s-1)

K′a (M-1) Ka (M-1) Kh (M-1) Kt Ki a

described values in s-1 (pK)

determined values (pK)a

2.74 × 10-3 (2.54)13 5.7 × 10-5 (4.3);14 1.99 × 10-4 (3.7)31 2.5 × 10-3 (2.6);14 1.6 × 10-3 (2.8)12 1.2 × 10-1 (0.92);14 2.6 × 10-1 (0.59)12 6.0 × 10-1 (0.22)12

5.35 × 10-3 (2.3 ( 0.1) 1.6 × 10-4 (3.8 ( 0.1) 4.3 × 10-3 (2.4) 0.24 ( 0.01

10% ethanol.

TABLE 4: Kinetic Constants Associated with the Transformations of Oenin (in s-1) kobs ka (s-1) k-a kh2 (s-1) k-h2 (s-1) kt k-t a

equilibrium constants and their associated rate constants may be now compared to the anthocyanic equivalent precursor oenin. In that way, it is possible to determine whether the substitution at the C8 carbon, with a flavan-3-ol, stabilizes the flavylium form, making the anthocyanin slightly more resistant to hydration and thus discoloration. The comparisons can be made from Tables 1 and 3 (thermodynamic data) and Tables 2 and 4 (kinetic data). No significant modifications regarding the kinetic and thermodynamic behaviors were observed, indicating that the network of chemical reactions involving catechin-(4,8)-malvidin-3-glucoside is controlled by its oenin moiety.

described values in s-1 4.7 × 104 (14); 5.0 × 106 31 6.7 × 108 s-1 M-1;14 2.51 × 1010 s-1 M-1 31 8.5 × 10-2 14,40 34 s-1 M-1 14

determined valuesa

0.08 ( 0.01 19 ( 1 M-1 0.084 ( 0.05 0.33 ( 0.05

10% ethanol.

and k-i ) 1.7 × 10-4 s-1. Tables 1 and 2 report, respectively, the equilibrium and the rate constants of the network. Finally, the mole fraction distribution at the equilibrium can be calculated from the data reported in Table 1 by means of eqs 15-23, see Figure 7. Malvidin-3-glucoside (oenin) Revisited through Reverse pH Jumps. The usefulness of the reverse pH jumps to study the anthocyanin network of chemical reactions was also exemplified for the first time for oenin (the anthocyanin moiety of the adduct Cat-Mv3glc studied). In this compound, the situation is more advantageous than the one for the F-A+ adduct because it was observed that the inversion of the rate constants between hydration and tautomerization occurs for pH e 2, leading to a larger pH window to observe both processes well separated (Figure 8II). The values of kh ) 0.08 s-1 and k-h ) 19 M-1 s-1, and k-t ) 0.33 s-1, were obtained. Following a similar approach used to characterize the thermodynamic and equilibrium constants of the adduct, its anthocyanin component, oenin was revisited, taking into account the stopped flow data, in particular those from the reverse pH jumps (Tables 2 and 3). The values of the

Conclusion Reverse pH jumps monitored by stopped flow have shown to be a powerful tool to study the network of chemical reactions involving anthocyanins. In particular, it was possible to measure with accuracy the rates of the ring-opening closure of oenin and catechin-(4,8)-malvidin-3-glucoside. Finally, an energy level diagram of the several species at the equilibrium can be constructed from the equilibrium constants reported in Table 1, by means of the thermodynamic relation ∆G° ) -RT ln Keq, together with the rate constant of the interconversion of the species taken from Table 2 (Figure 9). As it can be seen, at very acidic conditions, the most stable species is AH+. If the pH is raised up to ∼2.7, the AH+ will immediately equilibrate with B. If the pH is further raised to pH ) 3.9, A is instantaneously formed, but as B is in a lower energy level, A will be slowly converted into B via AH+. B is also in equilibrium with Cc, but as it is thermodynamically less stable, it will only be formed in small quantities. The comparison of the values of the thermodynamic and kinetic equilibrium constants between the F-A+ adduct and its anthocyanin moiety (oenin) showed that the flavanolic unit (catechin) shifts the absorption maximum of the flavylium cation from 518 to 535 nm but has no significant modification in the kinetic and thermodynamic properties of the parent compound. Acknowledgment. This research was supported by 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. The authors acknowledge the financial support of FCT (Fundac¸a˜o para a Cieˆncia e Tecnologia). V.P. gratefully acknowledges Grant SFRH/BPD/18214/ 2004. F.N. gratefully acknowledges a Ph.D. Scholarship from FCT (PhD SFRH/BD/30294/2006).

Figure 9. Energy level diagram for Cat-Mv3glc (with the values/expressions for the kinetic processes that occur before equilibration).

13496

J. Phys. Chem. B, Vol. 114, No. 42, 2010

References and Notes (1) Duenas, M.; Fulcrand, H.; Cheynier, V. Anal. Chim. Acta 2005, 563, 15–25. (2) Remy, S.; Fulcrand, H.; Labarbe, B.; Cheynier, V.; Moutounet, M. J. Sci. Food Agric. 2000, 80, 745–751. (3) Salas, E.; Atanasova, V.; Poncet-Legrand, C.; Meudec, E.; Mazauric, J. P.; Cheynier, V. Anal. Chim. Acta 2004, 513, 325–332. (4) Fossen, T.; Rayyan, S.; Andersen, O. M. Phytochemistry 2004, 65, 1421–1428. (5) Gonzalez-Paramas, A. M.; da Silva, F. L.; Martin-Lopez, P.; MaczPop, G.; Gonzalez-Manzano, S.; Alcalde-Eon, C.; Perez-Alonso, J. J.; Escribano-Bailon, M. T.; Rivas-Gonzalo, J. C.; Santos-Buelga, C. Food Chem. 2006, 94, 428–436. (6) Gonzalez-Manzano, S.; Perez-Alonso, J. J.; Salinas-Moreno, Y.; Mateus, N.; Silva, A. M. S.; de Freitas, V.; Santos-Buelga, C. J. Food Compos. Anal. 2008, 21, 521–526. (7) Nave, F.; Teixeira, N.; Mateus, N.; de Freitas, V. Food Chem. 2010, 121, 1129–1138. (8) Salas, E.; Le Guerneve, C.; Fulcrand, H.; Poncet-Legrand, C.; Cheynier, W. Tetrahedron Lett. 2004, 45, 8725–8729. (9) Nave, F.; Teixeira, N.; Mateus, N.; de Freitas, V. Rapid Commun. Mass Spectrom. 2010, 24, 1964–1970. (10) Melo, M. J.; Pina, F.; Andary, C. In Handbook of Natural Colorants; Bechtold, T., Mussak, R., Eds.; John Wiley and Sons: 2009, pp 135-150. (11) Wongcharee, K.; Meeyoo, V.; Chavadej, S. Sol. Energy Mater. Sol. Cells 2007, 91, 566–571. (12) Houbiers, C.; Lima, J. C.; Macanita, A. L.; Santos, H. J. Phys. Chem. B 1998, 102, 3578–3585. (13) Brouillard, R.; Delaporte, B.; Dubois, J. E. J. Am. Chem. Soc. 1978, 100, 6202–6205. (14) Brouillard, R.; Delaporte, B. J. Am. Chem. Soc. 1977, 99, 8461– 8468. (15) Alcais, P.; Brouilla., R. J. Chem. Soc., Perkin Trans. 2 1972, 1214– 1219. (16) Brouillard, R.; Delaporte, B.; Elhagechahine, J. M.; Dubois, J. E. J. Chim. Phys. Phys.-Chim. Biol. 1979, 76, 273–276. (17) Brouillard, R.; Dubois, J. E. J. Am. Chem. Soc. 1977, 99, 1359– 1364. (18) Brouillard, R.; Lang, J. Can. J. Chem. 1990, 68, 755–761. (19) McClelland, R. A.; Gedge, S. J. Am. Chem. Soc. 1980, 102, 5838– 5848.

Nave et al. (20) McClelland, R. A.; McGall, G. H. J. Org. Chem. 1982, 47, 3730– 3736. (21) Devine, D. B.; McClelland, R. A. J. Org. Chem. 1985, 50, 5656– 5660. (22) McClelland, R. A.; Devine, D. B.; Sorensen, P. E. J. Am. Chem. Soc. 1985, 107, 5459–5463. (23) McClelland, R. A.; Sorensen, P. E. Acta Chem. Scand. 1990, 44, 1082–1084. (24) Asenstorfer, R. E.; Lee, D. F.; Jones, G. P. Anal. Chim. Acta 2006, 563, 10–14. (25) Asenstorfer, R. E.; Jones, G. P. Tetrahedron 2007, 63, 4788–4792. (26) Lima, J. C.; Danesh, P.; Figueiredo, P.; Pina, F. S.; Macanita, A. Photochem. Photobiol. 1994, 59, 412–418. (27) Pina, F.; Benedito, L.; Melo, J. M.; Parola, A. J.; Lima, J. C.; Macanita, A. L. An. Quim. 1997, 93, 111–118. (28) Pina, F. J. Chem. Soc., Faraday Trans. 1998, 94, 2109–2116. (29) Moncada, M. C.; Pina, F.; Roque, A.; Parola, A. J.; Maestri, M.; Balzani, V. Eur. J. Org. Chem. 2004, 304–312. (30) Petrov, V.; Gomes, R.; Parola, A. J.; Pina, F. Dyes Pigm. 2009, 80, 149–155. (31) 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. (32) Pissarra, J.; Mateus, N.; Rivas-Gonzalo, J.; Buelga, C. S.; de Freitas, V. J. Food Sci. 2003, 68, 476–481. (33) Pissarra, J.; Lourenco, S.; Gonzalez-Paramas, A. M.; Mateus, N.; Santos-Buelga, C.; De Freitas, V. Anal. Chim. Acta 2004, 513, 215–221. (34) Ku¨ster, W. F.; Thiel, A. Tabelle per le Analisi Chimiche e ChimicoFisiche, 12 ed.; Hoepli: Milano, 1982. (35) Melo, M. J.; Sousa, M.; Parola, A. J.; de Melo, J. S. S.; Catarino, F.; Marcalo, J.; Pina, F. Chem.sEur. J. 2007, 13, 1417–1422. (36) Preston, N. W.; Timberlake, C. F. J. Chromatogr. 1981, 214, 222– 228. (37) Quina, F. H.; Freitas, A. A.; Macanita, A. L.; Silva, P. F. d.; Lima, J. C. ChemInform 2009, 40, 18–23. (38) Photochromic Systems Based on Synthetic FlaVylium Compounds and Their Potential Use as Molecular-LeVel Memory Elements; Pina, F., Maestri, M., Balzani, V., Eds.; American Chemical Society: Washington, DC, 2003; Vol. 3. (39) Atkins, P. Physical Chemistry, 6th ed.; Oxford University Press: Oxford, U.K., 1998. (40) Lima, J. C.; Vautier-Giongo, C.; Lopes, A.; Melo, E.; Quina, F. H.; Macanita, A. L. J. Phys. Chem. A 2002, 106, 5851–5859.

JP104749F