Food Chemistry 168 (2015) 464–470
Contents lists available at ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Experimental and theoretical binding affinity between polyvinylpolypyrrolidone and selected phenolic compounds from food matrices Esteban F. Durán-Lara a,b, Xaviera A. López-Cortés a,b, Ricardo I. Castro a,b, Fabián Avila-Salas a,b, Fernando D. González-Nilo c, V. Felipe Laurie a,d, Leonardo S. Santos a,b,⇑ a
Nanobiotechnology Division, Fraunhofer Chile Research Foundation – Center for Systems Biotechnology (FCR-CSB), Talca, Maule, Chile Laboratory of Asymmetric Synthesis, Institute of Chemistry and Natural Resources, Universidad de Talca, Talca, Chile c Universidad Andres Bello, Center for Bioinformatics and Integrative Biology (CBIB), Faculty of Biological Sciences, Santiago, Chile d School of Agricultural Sciences, Universidad de Talca, Talca, Chile b
a r t i c l e
i n f o
Article history: Received 17 March 2014 Received in revised form 1 July 2014 Accepted 7 July 2014 Available online 14 July 2014 Keywords: PVPP Phenolic compounds Interaction energy Semi-empirical methods PM7
a b s t r a c t Polyvinylpolypyrrolidone (PVPP) is a fining agent, widely used in winemaking and brewing, whose mode of action in removing phenolic compounds has not been fully characterised. The aim of this study was to evaluate the experimental and theoretical binding affinity of PVPP towards six phenolic compounds representing different types of phenolic species. The interaction between PVPP and phenolics was evaluated in model solutions, where hydroxyl groups, hydrophobic bonding and steric hindrance were characterised. The results of the study indicated that PVPP exhibits high affinity for quercetin and catechin, moderate affinity for epicatechin, gallic acid and lower affinity for 4-methylcatechol and caffeic acid. The affinity has a direct correlation with the hydroxylation degree of each compound. The results show that the affinity of PVPP towards phenols is related with frontier orbitals. This work demonstrates a direct correlation between the experimental affinity and the interaction energy calculations obtained through computational chemistry methods. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Phenolic compounds are important secondary metabolites that are ubiquitous to several fresh and processed food products (Balasundram, Sundram, & Samman, 2006; Cheynier, 2012). In fermented alcoholic beverages, such as wines, they contribute with sensorial characteristics that are critical to the quality of the finished product, such as colour, astringency and bitterness (Baiano et al., 2014; Marquez, Serratosa, & Merida, 2014; Sun, Liang, Bin, Li, & Duan, 2007). Moreover, phenolic compounds have been proven to be effective radical scavengers that upon regular consumption have been linked with health benefits such as anti-inflammatory, antimutagenic, and anticancer effects (Dong et al., 2011; Fernandez, Oliva, Barba, & Camara, 2005; Gollucke et al., 2013; Nunes et al., 2013; Sergent, Piront, Meurice, Toussaint, & Schneider, 2010). In wines, the presence of large amounts of these ⇑ Corresponding author at: Laboratory of Asymmetric Synthesis, Institute of Chemistry and Natural Resources, Universidad de Talca, Talca, Chile. Tel.: +56 (71) 2 201575; fax: +56 (71) 2 200448. E-mail address: [email protected] (L.S. Santos). http://dx.doi.org/10.1016/j.foodchem.2014.07.048 0308-8146/Ó 2014 Elsevier Ltd. All rights reserved.
phenolic compounds may relate with sensorial or cosmetic problems (Lorrain et al., 2013; Villamor & Ross, 2013) which, in some instances, are solved by reducing their concentration using fining agents such as polyvinylpolypyrrolidone, PVPP (Caceres-Mella et al., 2013; Mcmurrough, Madigan, & Smyth, 1995; Sen et al., 2012). The PVPP polymer (Fig. 1) was introduced commercially as an adsorbent for beer phenolics in 1961 (Caceres-Mella et al., 2013; Mcmurrough et al., 1995). Since then, it has been widely used as an agent for prolonging the stability of beers against haze formation (McMurrough, Kelly, Byrne, & O’Brien, 1992), as well as for the modulation of the concentration of phenolics in wines (Magalhães et al., 2010; Mcmurrough et al., 1995). Although few studies have applied computational chemistry methodologies to estimate the molecular interactions between PVPP and target molecules of industrial interest (Laborde et al., 2006; Le Bourvellec & Renard, 2012; Panarin, Kalninsh, & Pestov, 2001), none of them have correlated their results with experimental data, particularly regarding to phenolic compounds. Therefore, the aim of this study was to evaluate the experimental affinity between phenolic compounds such as quercetin, catechin, epicatechin, 4-methylcatechol, gallic acid and caffeic acid towards PVPP
E.F. Durán-Lara et al. / Food Chemistry 168 (2015) 464–470
465
The adsorption efficiency of each phenolic compound by PVPP was evaluated by determining the percentage decrease in the absorbance at each specific maximum absorbance wavelength using the following Eq. (1):
Adsorption ð%Þ ¼
Fig. 1. (A) Chemical structure of PVPP – (C6H9NO)n = (111.1)n. (B) Monomeric unit used to calculate the interaction energy .
polymer (adsorption test in model solutions evaluated by HPLC–DAD), and correlate these results with information obtained through the structure-based molecular characterisation using semi-empirical and quantum mechanical studies (Gontijo et al., 2012; Louli, Ragoussis, & Magoulas, 2004; Stewart, 2007; Yilmaz & Toledo, 2006). 2. Materials and methods 2.1. Reagents Quercetin hydrate P 95% (Sigma), ()-epicatechin (Sigma), 4-methylcatechol P 95% (Aldrich), gallic acid monohydrate P 98% (Aldrich), caffeic acid (Sigma), (+)-catechin hydrate P 98% (Sigma), ethanol HPLC grade P 99.8% (Merck), acetonitrile HPLC grade P 99.5% (Sigma) and polyvinylpyrrolidone powder, average Mw 40,000 (Sigma). All solutions were prepared using MilliQ water. 2.2. Adsorption of phenolic compounds by PVPP in model solutions Model solutions containing a mixture of phenolic compounds were prepared in (a) 12% v/v ethanol solution at pH 3.5 (a mixture commonly used to model wine samples), and (b) water at pH 6.5 adjusted with formic acid. The concentration of phenolic compound was adjusted to 0.16 mM for each of the following phenolic species (gallic acid, catechin, caffeic acid, epicatechin, 4-methylcatecol and quercetin) affording a 2.25 mg mL1 of phenolics. Considering that quercetin was insoluble, or slightly soluble in pure water, this phenol was only evaluated in the acidified 12% ethanol solution. An initial trial was performed to evaluate the time of interaction required between PVPP and phenolics in order to produce the maximum adsorption. The adsorption kinetics was evaluated at nine different times (0, 5, 10, 15, 20, 25, 30, 45 and 60 min after the addition of PVPP) at room temperature by measuring the solution’s absorption at 280 nm (i.e., a simple and commonly used procedure to quantify the concentration of phenolics in solution). After identifying the time of maximum absorption, an HPLC–DAD method was tested (as detailed below) in order to assess the concentration of each individual phenol in solution. To evaluate the affinity of PVPP towards each selected phenolic compounds, 1.8 mg of PVPP (0.45 mg mL1) was added to a 4 mL solution containing all phenols (2.25 mg mL1 of total phenols). The ratio by weight of the total phenolic compounds to PVPP was 5 (total phenolic compounds:PVPP, 5:1 w/w). The samples were incubated with stirring for 15 min at constant room temperature (25 °C) and then centrifuged at 10,000 rpm for 10 min. The concentration of phenolic compounds in the supernatant was analysed by HPLC–DAD (as further explained below). The quantification was conducted at 280 nm for catechin, 4-methylcatechol and epicatechin, for gallic acid at 271 nm, for caffeic acid at 323 nm, and for quercetin at 265 nm. Three independent measurements were made for each sample, and the results are presented as mean values with standard deviations.
A0 A 100 A0
ð1Þ
where A0 is the initial absorbance at specific wavelength and A is the final absorbance at the same wavelength (Table 1). 2.3. Chromatographic method and instrument employed The HPLC system (Agilent ChemStation, 1200, USA) consisted of a low-pressure quaternary pump (model Agilent 1200), autosampler (model Agilent 1260 Infinity Autosampler), an in-line DG-model G1322A degasser and a model G1329B, and a photodiode array detector (model MD-1510 UV/visible multiwavelength detector). Separations were achieved on a LiChrospherRP-18 column of 250 mm (5 lm) particle size. The chromatographic conditions were the following: Mobile phases were prepared with formic acid 4.5% and milliQ water (solution A) (filtered through a 0.45 lm nylon filters), and acetonitrile, HPLC grade (solution B). The gradient was programmed as specified in Table S1 (Supplemental material) running 35 min. The injection volumes used were of 100 lL, and the wavelengths selected for evaluation were at 265, 271, 280 and 323 nm. Column room temperature and flow rate were optimised for phenols separation, initial 1.2 mL min1 for 10 min, 1.3 mL min1 between 10 and 15 min, and 1.5 mL min1 between 15 and 35 min. Peak purity was checked to exclude any contribution from interfering peaks. The identification and quantification of phenolic compounds were performed by comparing their retention time against high purity standard. For this purpose, calibration curves were performed for each phenolic compound (catechin, epicatechin, gallic acid, caffeic acid, quercetin and 4-methylcatechol) in the range of 5–50 ppm. 2.4. Computational methods To carry out the computational calculations a nanoinformatics strategy (Avila-Salas et al., 2012; Metropolis, Rosenbluth, Rosenbluth, Teller, & Teller, 1953) was used to calculate the interaction energy between pairs of molecules (Fan, Olafson, & Blanco, 1992; Gonzalez-Nilo, Urzua, Leiva, Gargallo, & Radic, 2003) using 1-methyl-2-pyrrolidone (monomeric unit of PVPP) with each phenol, as a way to estimate the molecular properties of PVPP. To obtain an accurate prediction of the interaction energies of these systems, it is important to use a representative sampling of the available conformations for each complex (molecule 1–molecule 2). For this reason, the algorithm used starts with a random sampling through Euler Angles between the pair of molecules under study (Fan et al., 1992; Metropolis et al., 1953). That strategy generates thousands of pairs of 1-methyl-2-pyrrolydone–phenol, allowing an exhaustive conformational sampling. Then, for each new pair conformation, a single energy point was calculated using the latter semi-empirical quantum mechanical methods, in this case, the Parameterization Method 7 (PM7) MOPAC, 2012. The advantage of using this new version of PM7 method is that allows the study of intermolecular interactions with a good accuracy and great speed. PM7 also includes empirical corrections for dispersion and hydrogen-bond interactions (Hostas, Rezac, & Hobza, 2013; Korth, 2010). Semi-empirical calculations consider the valence shell electrons, and use empirical data (experimental information) to account for the energies of the inner shell electrons (Louli et al., 2004). The use of empirical information allows simplifying the quantum mechanical calculations. Although they are not as accurate as
2.4.1. Building molecular structures The molecular structures of the six phenols (catechin, quercetin, epicatechin, gallic acid, caffeic acid and 4-methylcatechol) (see Table 2) and 1-methyl-2-pyrrolidone were built using the GaussView program (Dennington et al., 2003). The molecular structures of these molecules were optimised with Density Functional theory (DFT) Andzelm & Wimmer, 1992, using the B3LYP method (Becke, 1997; Lee, Yang, & Parr, 1988) with 6-31G⁄ as basis set, which has been implemented in Gaussian 03 package program (Pople & Beveridge, 1970). The structures were built considering an environment at low pH and another at neutral pH. In the latter, both gallic and caffeic acids have their carboxylic group deprotonated (see Table 2).
Only the carboxyl groups of gallic and caffeic acids were deprotonated when they went from low pH to a neutral pH.
C15H10O7 Empirical formula
Molecular structure
ab initio methods, they are fast, and results are often good for large molecular structures that are well characterised by experimental data, such as the polymer-molecule or monomer-molecule systems (Feng, Wang, Zhuang, Wu, & Han, 2004). In addition, the molecular structure and electronic parameters of phenols and the polymer of PVPP (size comprising 30 monomers) were calculated through quantum mechanics calculations, these are the higher occupied molecular orbital (HOMO), the lower unoccupied molecular orbital (LUMO) and the energy gap (DE = ELUMO EHOMO). The energy gap between each phenol with the PVPP monomer was compared with experimental data, which afforded the identification of the main substituent groups in the phenols involved in the affinity for PVPP.
**
C7H8O3 C9H8O4 C7H6O5 C15H14O6 C15H14O6
4-Methylcatechol Caffeic acid** Epicatechin Chemical name
Quercetin
Catechin
Gallic acid**
E.F. Durán-Lara et al. / Food Chemistry 168 (2015) 464–470
Table 1 Characteristic data for the eight phenols at pH 3.5 and pH 6.5.
466
2.4.2. In silico calculation of interaction energies A computational chemistry strategy complemented with conformational sampling, using Metropolis Monte Carlo simulations (Avila-Salas et al., 2012; Metropolis et al., 1953), was used to calculate the interaction energy between 1-methyl-2-pyrrolidone–phenol pairs (molecule 1–molecule 2). These simulations use a random sampling algorithm to generate thousands of different molecular orientations of one specific pair, calculates the energy using semi-empirical quantum mechanical methods, and distributes the calculations on a High Performance Computing System (HPC). Six pairs combining a monomer of PVPP and phenols were generated (monomer–catechin, monomer–quercetin, monomer–epicatechin, monomer–gallic acid, monomer–caffeic acid and monomer–4-methylcatechol). They were used to calculate the intermolecular interaction energies (Eq. (2)) through the Metropolis Monte Carlo algorithm. The above algorithm involves a methodology for generating a conformational sampling described by Fan et al. (1992) and calculations of single-point energy (1SCF) using a semi-empirical quantum chemistry method (PM7) implemented in MOPAC2012™ packaged program, version 12.310L (LINUX) MOPAC, 2012. The values of heat of formation (Hf) are extracted of 1SCF calculations from both of the pair monomer-phenol as its isolated parts (monomer and phenols).
DE ¼ Hfðmolecule1—molecule2Þ ðHfðmolecule1Þ þ Hfðmolecule2Þ Þ
ð2Þ
2.4.3. HOMO and LUMO calculations The set of phenols and the polymer of PVPP (size comprising 30 monomers) were built using GaussView program (Dennington et al., 2003). For these calculations we work at pH 3.5 in order to simulate the wine conditions. Density functional theory (DFT) methods Andzelm & Wimmer, 1992 were used. All structures were optimised with the B3LYP method (Becke, 1997; Lee et al., 1988) and the STO-3G⁄ basis set (Pople & Beveridge, 1970), then, frequency analyses were carried out in order to verify the nature of minimum state of the resulting stationary points. HOMO and LUMO energy of the phenols and PVPP were obtained and energy
467
E.F. Durán-Lara et al. / Food Chemistry 168 (2015) 464–470 Table 2 Gap Energy between the LUMO of the polymer PVPP and the HOMO of the phenols.
Quercetin Catechin Epicatechin Gallic acid Caffeic acid 4-Methylcatechol PVPP
HOMO kcal mol1
LUMO kcal mol1
Gap energy kcal mol1
Adsorption capacity %
58.566 75.238 71.084 73.174 79.380 72.967 69.189
17.319 56.230 57.127 62.920 57.204 63.573 69.810
86.509 125.419 126.316 132.110 126.393 132.762
100 67 57 45 26 25 r2 = 0.73
gap parameters were calculated. All calculations were performed with the software Gaussian03 (Frisch et al., 2004). 2.5. Statistical analysis All the experiments were carried out in triplicate and the student t-test was used to calculate the statistical differences between the experimental compound and the negative control. A p-value of catechin > epicatechin > gallic acid > caffeic acid > 4-methylcatechol) and neutral pH (catechin > epicatechin > caffeic acid > gallic acid > 4-methylcatechol) (Fig. 3A). Therefore, the use of semi-empirical methods for this type of calculations may be in the first instance, a good option to
-1
Interaction Energy kcal mol
carry out large sampling conformations to determine with accuracy and speed the intermolecular interactions in similar polymeric systems to those studied in this article. To observe the behaviour of phenols versus PVPP monomer, the 100 lowest interaction energy conformations of 1-methyl-2-pyrrolidone–phenol complexes were selected (both for the series to low and neutral pH). The spatial distributions of these structures showed an interaction capacity similar to that obtained experimentally for PVPP–phenols (see Fig. 4). 3.3. HOMO and LUMO calculations Quantum mechanics calculations of phenols and PVPP polymer (length 30) were performed in order to obtain HOMO and LUMO energy, and to describe their molecular properties. The energy gap between the polymer and every phenol was calculated
-1
Interaction Energy kcal mol
Fig. 3. Correlation between experimental percentages of absorption capacity of PVPP by six phenols versus the average of the interaction energies. The r2 values in a low and neutral pH environment were higher than 0.95. The average of the interaction energies showed the same order of percentage of absorption capacity observed experimentally for PVPP–phenols at low pH (A): quercetin > catechin > epicatechin > gallic acid > caffeic acid > 4-methylcatechol and at neutral pH (B): catechin > epicatechin > caffeic acid > gallic acid > 4-methylcatechol.
Fig. 4. Spatial distributions for the 100 lowest energy conformations of 1-ethyl-2-pyrrolidone-phenol complexes. The spatial distributions showed the same order of percentage of absorption capacity observed experimentally for PVPP–phenols at low pH (A) and at neutral pH (B).
E.F. Durán-Lara et al. / Food Chemistry 168 (2015) 464–470
according to DE = ELUMO EHOMO, and this energy gap was compared with the experimental values of HPLC (Table 2). Two energy gaps were calculated, the first corresponding to the energy between the LUMO of the phenol and the HOMO of the polymer, and the second representing the LUMO of the polymer and the HOMO of the phenol. From these results, it was possible to see that the favored interaction was between the LUMO of the phenol and the HOMO of the polymer (size comprising 30 monomers). The results of energy gap between the HOMO of the polymer and the LUMO of each phenol were in accordance with the experimental data giving a correlation coefficient (r2) of 0.73 (Table 2). Moreover, the number of substituent groups involved in the phenolic compounds was used trying to explain the preference of PVPP towards quercetin, catechin and epicatechin. Table S4 (Supplemental Material) shows the different substituent groups present in each phenol. According to the data, it was possible to see that hydroxyls were not the only groups involved in the interaction with PVPP (Lorrain et al., 2013), and that the interaction was governed by more than only hydrogen bonds. In fact, it was possible to see that the compounds with higher affinity had more aromatic rings and hydroxyl groups (such as in quercetin, catechin and epicatechin), as opposed to the carbonated chains and carboxyl groups present in the less affine phenols. The former might be due to the need of the phenol to extend the LUMO zone to have a better possibility to interact with PVPP. Therefore, the interaction between these phenols and PVPP is at molecular orbital level, between the HOMO of the polymer and the LUMO of the phenols. This is due to the groups involved in the LUMO of the phenols, which increase their affinity for the PVPP polymer. When the energy gaps are compared, it can be seen that the lowest energy gap is obtained for quercetin-PVPP. This implies that quercetin and PVPP have a strong chemical enhancement, which favours the electronic transference from the PVPP to quercetin according to gallic acid, caffeic acid and 4-methylcatechol that have the higher gap energies. 4. Conclusions This work studied the adsorption, or trapping capacity, of six phenolic compounds by PVPP. The results indicated that their interaction has a direct correlation with the number of hydroxyl and aromatic rings presented in the phenolic molecules. The LUMO of phenolic compounds proves that their substituents play a key role in the increase of polymer affinity toward phenols. The interaction of phenols and PVPP occurs at molecular orbital level, between the frontier orbitals HOMO-PVPP and LUMO-phenols, respectively. That might be the reason why a higher electronic transfer between the phenols and PVPP was favored in the phenols that had a higher number of hydroxyl groups and aromatic rings. For this reason, the PVPP polymer has the highest affinity for quercetin, catechin and epicatechin. Interaction energy averages obtained through a computational chemistry strategy proved to be highly predictive for estimating the affinity of specific phenols by PVPP. Acknowledgements E.F.D.L. thanks Fondecyt (Postdoctoral Grant 3120178), Fraunhofer Chile Research, and Innova Chile CORFO (Code FCR-CSB 09CEII-6991). X.A.L.C. and F.A.S. thank CONICYT-PCHA/Doctorado Nacional/2013-21130553 and 2013-21130308 for doctoral scholarship. L.S.S. and F.D.G.N. thank project Anillo ACT 1107 (Integracion de la Biologia Estructural al desarrollo de la Bionanotecnologia).
469
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.foodchem.2014. 07.048.
References Andzelm, J., & Wimmer, E. (1992). Density functional gaussian-type-orbital approach to molecular geometries, vibrations, and reaction energies. Journal of Chemical Physics, 96, 1280–1303. Avila-Salas, F., Sandoval, C., Caballero, J., Guinez-Molinos, S., Santos, L. S., Cachau, R. E., et al. (2012). Study of interaction energies between PAMAM dendrimer and non-steroidal anti-inflammatory drug using a distributed computational strategy and experimental analysis by ESI–MS/MS. Journal of Physical Chemistry B, 116, 2031–2039. Baiano, A., Varva, G., De Gianni, A., Viggiani, I., Terracone, C., & Del Nobile, M. A. (2014). Influence of type of amphora on physico-chemical properties and antioxidant capacity of ‘Falanghina’ white wines. Food Chemistry, 146, 226–233. Balasundram, N., Sundram, K., & Samman, S. (2006). Phenolic compounds in plants and agri-industrial by-products: Antioxidant activity, occurrence, and potential uses. Food Chemistry, 99, 191–203. Becke, A. D. (1997). Density-functional thermochemistry 0.5. Systematic optimization of exchange-correlation functionals. Journal of Chemical Physics, 107, 8554–8560. Caceres-Mella, A., Peña-Neira, A., Parraguez, J., Lopez-Solís, R., Laurie, F., & Canals, J. M. (2013). Effect of inert gas and prefermentative treatment with polyvinylpolypyrrolidone on the phenolic composition of Chilean Sauvignon blanc wines. Journal of the Science Food and Agriculture, 93, 1928–1934. Cheynier, V. (2012). Phenolic compounds: from plants to foods. Phytochemistry Reviews, 11, 153–177. Dennington, R., II, Keith, T., Milliam, J., Eppinnett, K., Hovell, W. L., Gilliland, R. (2003). GaussView, Version 3.09, Semichem Inc., Shawnee Mission. Dong, Z. B., Liang, Y. R., Fan, F. Y., Ye, J. H., Zheng, X. Q., & Lu, J. L. (2011). Adsorption behavior of the catechins and caffeine onto polyvinylpolypyrrolidone. Journal of Agricultural and Food Chemistry, 59, 4238–4247. Fan, C. F., Olafson, B. D., & Blanco, M. (1992). Application of molecular simulation to derive phase diagrams of binary mixtures. Macromolecules, 25, 3667–3676. Feng, D. D., Wang, S. F., Zhuang, Q. X., Wu, P. P., & Han, Z. W. (2004). Semi-empirical calculation and spectroscopic study of protonated poly(p-phenylene benzobisoxazole) models. Polymer, 45, 8871–8879. Fernandez, M. J., Oliva, J., Barba, A., & Camara, M. A. (2005). Effects of clarification and filtration processes on the removal of fungicide residues in red wines (var. Monastrell). Journal of Agricultural and Food Chemistry, 53, 6156–6161. Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A., Cheeseman, J. R., et al. (2004). Gaussian 03, Revision C. 02, Gaussian Inc., Wallingford, CT. Gollucke, A. P., Aguiar, O., Jr., Barbisan, L. F., & Ribeiro, D. A. (2013). Use of grape polyphenols against carcinogenesis: Putative molecular mechanisms of action using in vitro and in vivo test systems. Journal of Medicinal Food, 16, 199–205. Gonzalez-Nilo, F. D., Urzua, M., Leiva, A., Gargallo, L., & Radic, D. (2003). Interaction energy analysis of polymer blends containing N-1-alkylitaconamic acid moiety. Journal of Macromolecular Science, Part B: Physics, 42, 1281–1291. Hostas, J., Rezac, J., & Hobza, P. (2013). On the performance of the semiempirical quantum mechanical PM6 and PM7 methods for noncovalent interactions. Chemical Physics Letters, 568, 161–166. Korth, M. (2010). Third-generation hydrogen-bonding corrections for semiempirical QM methods and force fields. Chemical Theory and Computation, 6, 3808–3816. Laborde, B., Moine Ledoux, V., Richard, T., Saucier, C., Dubourdieu, D., & Monti, J. P. (2006). PVPP-polyphenol complexes: A molecular approach. Journal of Agricultural and Food Chemistry, 54, 4383–4389. Le Bourvellec, C., & Renard, C. M. G. C. (2012). Interactions between polyphenols and macromolecules: Quantification methods and mechanisms. Critical Reviews Food Science Nutrition, 52, 213–248. Lee, C. T., Yang, W. T., & Parr, R. G. (1988). Development of the Colle-Salvetti correlation-energy formula into a functional of the electron-density. Physical Review B, 37, 785–789. Lorrain, B., Tempere, S., Iturmendi, N., Moine, V., de Revel, G., & Teissedre, P. L. (2013). Influence of phenolic compounds on the sensorial perception and volatility of red wine esters in model solution: An insight at the molecular level. Food Chemistry, 140, 76–82. Louli, V., Ragoussis, N., & Magoulas, K. (2004). Recovery of phenolic antioxidants from wine industry by-products. Bioresource Technology, 92, 201–208. Magalhães, P. J., Vieira, J. S., Goncalves, L. M., Pacheco, J. G., Guido, L. F., Barros, A. A., et al. (2010). Isolation of phenolic compounds from hop extracts using polyvinylpolypyrrolidone: Characterization by high-performance liquid chromatography–diode array detection–electrospray tandem mass spectrometry. Journal of Chromatography A, 1217, 3258–3268. Marquez, A., Serratosa, M. P., & Merida, J. (2014). Influence of bottle storage time on colour, phenolic composition and sensory properties of sweet red wines. Food Chemistry, 146, 507–514. McMurrough, I., Kelly, R. J., Byrne, J. R., & O’Brien, M. (1992). Effect of the removal of sensitive proteins and proanthocyanidins on the colloidal stability of lager beer. Journal American Society Brewing Chemistry, 50, 67–76.
470
E.F. Durán-Lara et al. / Food Chemistry 168 (2015) 464–470
Mcmurrough, I., Madigan, D., & Smyth, M. R. (1995). Adsorption by polyvinylpolypyrrolidone of catechins and proanthocyanidins from beer. Journal of Agricultural and Food Chemistry, 43, 2687–2691. Metropolis, N., Rosenbluth, A. W., Rosenbluth, M. N., Teller, A. H., & Teller, E. (1953). Equation of state calculations by fast computing machines. Journal Chemical Physics, 21, 1087–1109. MOPAC2012, Stewart, James J. P. (2012). Stewart Computational Chemistry, Colorado Springs, CO, USA. Nunes, C., Elisabete Ferreira, E., Freitas, V., Almeida, L., Barbosa, R. M., & Laranjinha, J. (2013). Intestinal anti-inflammatory activity of red wine extract: Unveiling the mechanisms in colonic epithelial cells. Food and Functional, 4, 373–383. Panarin, E. F., Kalninsh, K. K., & Pestov, D. V. (2001). Complexation of hydrogen peroxide with polyvinylpyrrolidone: Ab initio calculations. European Polymer Journal, 37, 375–379. Pople, J. A., & Beveridge, D. L. (1970). Approximate molecular orbital theory (1st ed.). New York: McGraw-Hill. Sen, K., Cabaroglub, T., & Yilmaz, H. (2012). The influence of fining agents on the removal of some pesticides from white wine of Vitis vinifera L. cv. Emir. Food and Chemical Toxicology, 50, 3990–3995.
Sergent, T., Piront, N., Meurice, J., Toussaint, O., & Schneider, Y. (2010). Antiinflammatory effects of dietary phenolic compounds in an in vitro model of inflamed human intestinal epithelium. Chemico-Biological Interactions, 188, 659–667. Silva Gontijo, V., Judice, W. A., Codonho, B., Oliveira Pereirac, I., Assis, D. M., Pereira Januário, J., et al. (2012). Leishmanicidal antiproteolytic and antioxidant evaluation of natural biflavonoids isolated from Garcinia brasiliensis and their semisynthetic derivatives. European Journal of Medicinal Chemistry, 58, 613–623. Stewart, J. J., & Stewart, J. J. P. (2007). Optimization of parameters for semiempirical methods V: Modification of NDDO approximations and application to 70 elements. Journal of Molecular Modeling, 13, 1173–1213. Sun, J., Liang, F., Bin, Y., Li, P., & Duan, C. Q. (2007). Screening non-colored phenolics in red wines using liquid chromatography/ultraviolet and mass spectrometry/ mass spectrometry libraries. Molecules, 12, 679–693. Villamor, R. R., & Ross, C. F. (2013). Wine matrix compounds affect perception of wine aromas. Annual Review of Food Science and Technology, 4, 1–20. Yilmaz, Y., & Toledo, R. T. (2006). Oxygen radical absorbance capacities of grape/ wine industry byproducts and effect of solvent type on extraction of grape seed polyphenols. Journal of Food Composition and Analysis, 19, 41–48.
Contents lists available at ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Experimental and theoretical binding affinity between polyvinylpolypyrrolidone and selected phenolic compounds from food matrices Esteban F. Durán-Lara a,b, Xaviera A. López-Cortés a,b, Ricardo I. Castro a,b, Fabián Avila-Salas a,b, Fernando D. González-Nilo c, V. Felipe Laurie a,d, Leonardo S. Santos a,b,⇑ a
Nanobiotechnology Division, Fraunhofer Chile Research Foundation – Center for Systems Biotechnology (FCR-CSB), Talca, Maule, Chile Laboratory of Asymmetric Synthesis, Institute of Chemistry and Natural Resources, Universidad de Talca, Talca, Chile c Universidad Andres Bello, Center for Bioinformatics and Integrative Biology (CBIB), Faculty of Biological Sciences, Santiago, Chile d School of Agricultural Sciences, Universidad de Talca, Talca, Chile b
a r t i c l e
i n f o
Article history: Received 17 March 2014 Received in revised form 1 July 2014 Accepted 7 July 2014 Available online 14 July 2014 Keywords: PVPP Phenolic compounds Interaction energy Semi-empirical methods PM7
a b s t r a c t Polyvinylpolypyrrolidone (PVPP) is a fining agent, widely used in winemaking and brewing, whose mode of action in removing phenolic compounds has not been fully characterised. The aim of this study was to evaluate the experimental and theoretical binding affinity of PVPP towards six phenolic compounds representing different types of phenolic species. The interaction between PVPP and phenolics was evaluated in model solutions, where hydroxyl groups, hydrophobic bonding and steric hindrance were characterised. The results of the study indicated that PVPP exhibits high affinity for quercetin and catechin, moderate affinity for epicatechin, gallic acid and lower affinity for 4-methylcatechol and caffeic acid. The affinity has a direct correlation with the hydroxylation degree of each compound. The results show that the affinity of PVPP towards phenols is related with frontier orbitals. This work demonstrates a direct correlation between the experimental affinity and the interaction energy calculations obtained through computational chemistry methods. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Phenolic compounds are important secondary metabolites that are ubiquitous to several fresh and processed food products (Balasundram, Sundram, & Samman, 2006; Cheynier, 2012). In fermented alcoholic beverages, such as wines, they contribute with sensorial characteristics that are critical to the quality of the finished product, such as colour, astringency and bitterness (Baiano et al., 2014; Marquez, Serratosa, & Merida, 2014; Sun, Liang, Bin, Li, & Duan, 2007). Moreover, phenolic compounds have been proven to be effective radical scavengers that upon regular consumption have been linked with health benefits such as anti-inflammatory, antimutagenic, and anticancer effects (Dong et al., 2011; Fernandez, Oliva, Barba, & Camara, 2005; Gollucke et al., 2013; Nunes et al., 2013; Sergent, Piront, Meurice, Toussaint, & Schneider, 2010). In wines, the presence of large amounts of these ⇑ Corresponding author at: Laboratory of Asymmetric Synthesis, Institute of Chemistry and Natural Resources, Universidad de Talca, Talca, Chile. Tel.: +56 (71) 2 201575; fax: +56 (71) 2 200448. E-mail address: [email protected] (L.S. Santos). http://dx.doi.org/10.1016/j.foodchem.2014.07.048 0308-8146/Ó 2014 Elsevier Ltd. All rights reserved.
phenolic compounds may relate with sensorial or cosmetic problems (Lorrain et al., 2013; Villamor & Ross, 2013) which, in some instances, are solved by reducing their concentration using fining agents such as polyvinylpolypyrrolidone, PVPP (Caceres-Mella et al., 2013; Mcmurrough, Madigan, & Smyth, 1995; Sen et al., 2012). The PVPP polymer (Fig. 1) was introduced commercially as an adsorbent for beer phenolics in 1961 (Caceres-Mella et al., 2013; Mcmurrough et al., 1995). Since then, it has been widely used as an agent for prolonging the stability of beers against haze formation (McMurrough, Kelly, Byrne, & O’Brien, 1992), as well as for the modulation of the concentration of phenolics in wines (Magalhães et al., 2010; Mcmurrough et al., 1995). Although few studies have applied computational chemistry methodologies to estimate the molecular interactions between PVPP and target molecules of industrial interest (Laborde et al., 2006; Le Bourvellec & Renard, 2012; Panarin, Kalninsh, & Pestov, 2001), none of them have correlated their results with experimental data, particularly regarding to phenolic compounds. Therefore, the aim of this study was to evaluate the experimental affinity between phenolic compounds such as quercetin, catechin, epicatechin, 4-methylcatechol, gallic acid and caffeic acid towards PVPP
E.F. Durán-Lara et al. / Food Chemistry 168 (2015) 464–470
465
The adsorption efficiency of each phenolic compound by PVPP was evaluated by determining the percentage decrease in the absorbance at each specific maximum absorbance wavelength using the following Eq. (1):
Adsorption ð%Þ ¼
Fig. 1. (A) Chemical structure of PVPP – (C6H9NO)n = (111.1)n. (B) Monomeric unit used to calculate the interaction energy .
polymer (adsorption test in model solutions evaluated by HPLC–DAD), and correlate these results with information obtained through the structure-based molecular characterisation using semi-empirical and quantum mechanical studies (Gontijo et al., 2012; Louli, Ragoussis, & Magoulas, 2004; Stewart, 2007; Yilmaz & Toledo, 2006). 2. Materials and methods 2.1. Reagents Quercetin hydrate P 95% (Sigma), ()-epicatechin (Sigma), 4-methylcatechol P 95% (Aldrich), gallic acid monohydrate P 98% (Aldrich), caffeic acid (Sigma), (+)-catechin hydrate P 98% (Sigma), ethanol HPLC grade P 99.8% (Merck), acetonitrile HPLC grade P 99.5% (Sigma) and polyvinylpyrrolidone powder, average Mw 40,000 (Sigma). All solutions were prepared using MilliQ water. 2.2. Adsorption of phenolic compounds by PVPP in model solutions Model solutions containing a mixture of phenolic compounds were prepared in (a) 12% v/v ethanol solution at pH 3.5 (a mixture commonly used to model wine samples), and (b) water at pH 6.5 adjusted with formic acid. The concentration of phenolic compound was adjusted to 0.16 mM for each of the following phenolic species (gallic acid, catechin, caffeic acid, epicatechin, 4-methylcatecol and quercetin) affording a 2.25 mg mL1 of phenolics. Considering that quercetin was insoluble, or slightly soluble in pure water, this phenol was only evaluated in the acidified 12% ethanol solution. An initial trial was performed to evaluate the time of interaction required between PVPP and phenolics in order to produce the maximum adsorption. The adsorption kinetics was evaluated at nine different times (0, 5, 10, 15, 20, 25, 30, 45 and 60 min after the addition of PVPP) at room temperature by measuring the solution’s absorption at 280 nm (i.e., a simple and commonly used procedure to quantify the concentration of phenolics in solution). After identifying the time of maximum absorption, an HPLC–DAD method was tested (as detailed below) in order to assess the concentration of each individual phenol in solution. To evaluate the affinity of PVPP towards each selected phenolic compounds, 1.8 mg of PVPP (0.45 mg mL1) was added to a 4 mL solution containing all phenols (2.25 mg mL1 of total phenols). The ratio by weight of the total phenolic compounds to PVPP was 5 (total phenolic compounds:PVPP, 5:1 w/w). The samples were incubated with stirring for 15 min at constant room temperature (25 °C) and then centrifuged at 10,000 rpm for 10 min. The concentration of phenolic compounds in the supernatant was analysed by HPLC–DAD (as further explained below). The quantification was conducted at 280 nm for catechin, 4-methylcatechol and epicatechin, for gallic acid at 271 nm, for caffeic acid at 323 nm, and for quercetin at 265 nm. Three independent measurements were made for each sample, and the results are presented as mean values with standard deviations.
A0 A 100 A0
ð1Þ
where A0 is the initial absorbance at specific wavelength and A is the final absorbance at the same wavelength (Table 1). 2.3. Chromatographic method and instrument employed The HPLC system (Agilent ChemStation, 1200, USA) consisted of a low-pressure quaternary pump (model Agilent 1200), autosampler (model Agilent 1260 Infinity Autosampler), an in-line DG-model G1322A degasser and a model G1329B, and a photodiode array detector (model MD-1510 UV/visible multiwavelength detector). Separations were achieved on a LiChrospherRP-18 column of 250 mm (5 lm) particle size. The chromatographic conditions were the following: Mobile phases were prepared with formic acid 4.5% and milliQ water (solution A) (filtered through a 0.45 lm nylon filters), and acetonitrile, HPLC grade (solution B). The gradient was programmed as specified in Table S1 (Supplemental material) running 35 min. The injection volumes used were of 100 lL, and the wavelengths selected for evaluation were at 265, 271, 280 and 323 nm. Column room temperature and flow rate were optimised for phenols separation, initial 1.2 mL min1 for 10 min, 1.3 mL min1 between 10 and 15 min, and 1.5 mL min1 between 15 and 35 min. Peak purity was checked to exclude any contribution from interfering peaks. The identification and quantification of phenolic compounds were performed by comparing their retention time against high purity standard. For this purpose, calibration curves were performed for each phenolic compound (catechin, epicatechin, gallic acid, caffeic acid, quercetin and 4-methylcatechol) in the range of 5–50 ppm. 2.4. Computational methods To carry out the computational calculations a nanoinformatics strategy (Avila-Salas et al., 2012; Metropolis, Rosenbluth, Rosenbluth, Teller, & Teller, 1953) was used to calculate the interaction energy between pairs of molecules (Fan, Olafson, & Blanco, 1992; Gonzalez-Nilo, Urzua, Leiva, Gargallo, & Radic, 2003) using 1-methyl-2-pyrrolidone (monomeric unit of PVPP) with each phenol, as a way to estimate the molecular properties of PVPP. To obtain an accurate prediction of the interaction energies of these systems, it is important to use a representative sampling of the available conformations for each complex (molecule 1–molecule 2). For this reason, the algorithm used starts with a random sampling through Euler Angles between the pair of molecules under study (Fan et al., 1992; Metropolis et al., 1953). That strategy generates thousands of pairs of 1-methyl-2-pyrrolydone–phenol, allowing an exhaustive conformational sampling. Then, for each new pair conformation, a single energy point was calculated using the latter semi-empirical quantum mechanical methods, in this case, the Parameterization Method 7 (PM7) MOPAC, 2012. The advantage of using this new version of PM7 method is that allows the study of intermolecular interactions with a good accuracy and great speed. PM7 also includes empirical corrections for dispersion and hydrogen-bond interactions (Hostas, Rezac, & Hobza, 2013; Korth, 2010). Semi-empirical calculations consider the valence shell electrons, and use empirical data (experimental information) to account for the energies of the inner shell electrons (Louli et al., 2004). The use of empirical information allows simplifying the quantum mechanical calculations. Although they are not as accurate as
2.4.1. Building molecular structures The molecular structures of the six phenols (catechin, quercetin, epicatechin, gallic acid, caffeic acid and 4-methylcatechol) (see Table 2) and 1-methyl-2-pyrrolidone were built using the GaussView program (Dennington et al., 2003). The molecular structures of these molecules were optimised with Density Functional theory (DFT) Andzelm & Wimmer, 1992, using the B3LYP method (Becke, 1997; Lee, Yang, & Parr, 1988) with 6-31G⁄ as basis set, which has been implemented in Gaussian 03 package program (Pople & Beveridge, 1970). The structures were built considering an environment at low pH and another at neutral pH. In the latter, both gallic and caffeic acids have their carboxylic group deprotonated (see Table 2).
Only the carboxyl groups of gallic and caffeic acids were deprotonated when they went from low pH to a neutral pH.
C15H10O7 Empirical formula
Molecular structure
ab initio methods, they are fast, and results are often good for large molecular structures that are well characterised by experimental data, such as the polymer-molecule or monomer-molecule systems (Feng, Wang, Zhuang, Wu, & Han, 2004). In addition, the molecular structure and electronic parameters of phenols and the polymer of PVPP (size comprising 30 monomers) were calculated through quantum mechanics calculations, these are the higher occupied molecular orbital (HOMO), the lower unoccupied molecular orbital (LUMO) and the energy gap (DE = ELUMO EHOMO). The energy gap between each phenol with the PVPP monomer was compared with experimental data, which afforded the identification of the main substituent groups in the phenols involved in the affinity for PVPP.
**
C7H8O3 C9H8O4 C7H6O5 C15H14O6 C15H14O6
4-Methylcatechol Caffeic acid** Epicatechin Chemical name
Quercetin
Catechin
Gallic acid**
E.F. Durán-Lara et al. / Food Chemistry 168 (2015) 464–470
Table 1 Characteristic data for the eight phenols at pH 3.5 and pH 6.5.
466
2.4.2. In silico calculation of interaction energies A computational chemistry strategy complemented with conformational sampling, using Metropolis Monte Carlo simulations (Avila-Salas et al., 2012; Metropolis et al., 1953), was used to calculate the interaction energy between 1-methyl-2-pyrrolidone–phenol pairs (molecule 1–molecule 2). These simulations use a random sampling algorithm to generate thousands of different molecular orientations of one specific pair, calculates the energy using semi-empirical quantum mechanical methods, and distributes the calculations on a High Performance Computing System (HPC). Six pairs combining a monomer of PVPP and phenols were generated (monomer–catechin, monomer–quercetin, monomer–epicatechin, monomer–gallic acid, monomer–caffeic acid and monomer–4-methylcatechol). They were used to calculate the intermolecular interaction energies (Eq. (2)) through the Metropolis Monte Carlo algorithm. The above algorithm involves a methodology for generating a conformational sampling described by Fan et al. (1992) and calculations of single-point energy (1SCF) using a semi-empirical quantum chemistry method (PM7) implemented in MOPAC2012™ packaged program, version 12.310L (LINUX) MOPAC, 2012. The values of heat of formation (Hf) are extracted of 1SCF calculations from both of the pair monomer-phenol as its isolated parts (monomer and phenols).
DE ¼ Hfðmolecule1—molecule2Þ ðHfðmolecule1Þ þ Hfðmolecule2Þ Þ
ð2Þ
2.4.3. HOMO and LUMO calculations The set of phenols and the polymer of PVPP (size comprising 30 monomers) were built using GaussView program (Dennington et al., 2003). For these calculations we work at pH 3.5 in order to simulate the wine conditions. Density functional theory (DFT) methods Andzelm & Wimmer, 1992 were used. All structures were optimised with the B3LYP method (Becke, 1997; Lee et al., 1988) and the STO-3G⁄ basis set (Pople & Beveridge, 1970), then, frequency analyses were carried out in order to verify the nature of minimum state of the resulting stationary points. HOMO and LUMO energy of the phenols and PVPP were obtained and energy
467
E.F. Durán-Lara et al. / Food Chemistry 168 (2015) 464–470 Table 2 Gap Energy between the LUMO of the polymer PVPP and the HOMO of the phenols.
Quercetin Catechin Epicatechin Gallic acid Caffeic acid 4-Methylcatechol PVPP
HOMO kcal mol1
LUMO kcal mol1
Gap energy kcal mol1
Adsorption capacity %
58.566 75.238 71.084 73.174 79.380 72.967 69.189
17.319 56.230 57.127 62.920 57.204 63.573 69.810
86.509 125.419 126.316 132.110 126.393 132.762
100 67 57 45 26 25 r2 = 0.73
gap parameters were calculated. All calculations were performed with the software Gaussian03 (Frisch et al., 2004). 2.5. Statistical analysis All the experiments were carried out in triplicate and the student t-test was used to calculate the statistical differences between the experimental compound and the negative control. A p-value of catechin > epicatechin > gallic acid > caffeic acid > 4-methylcatechol) and neutral pH (catechin > epicatechin > caffeic acid > gallic acid > 4-methylcatechol) (Fig. 3A). Therefore, the use of semi-empirical methods for this type of calculations may be in the first instance, a good option to
-1
Interaction Energy kcal mol
carry out large sampling conformations to determine with accuracy and speed the intermolecular interactions in similar polymeric systems to those studied in this article. To observe the behaviour of phenols versus PVPP monomer, the 100 lowest interaction energy conformations of 1-methyl-2-pyrrolidone–phenol complexes were selected (both for the series to low and neutral pH). The spatial distributions of these structures showed an interaction capacity similar to that obtained experimentally for PVPP–phenols (see Fig. 4). 3.3. HOMO and LUMO calculations Quantum mechanics calculations of phenols and PVPP polymer (length 30) were performed in order to obtain HOMO and LUMO energy, and to describe their molecular properties. The energy gap between the polymer and every phenol was calculated
-1
Interaction Energy kcal mol
Fig. 3. Correlation between experimental percentages of absorption capacity of PVPP by six phenols versus the average of the interaction energies. The r2 values in a low and neutral pH environment were higher than 0.95. The average of the interaction energies showed the same order of percentage of absorption capacity observed experimentally for PVPP–phenols at low pH (A): quercetin > catechin > epicatechin > gallic acid > caffeic acid > 4-methylcatechol and at neutral pH (B): catechin > epicatechin > caffeic acid > gallic acid > 4-methylcatechol.
Fig. 4. Spatial distributions for the 100 lowest energy conformations of 1-ethyl-2-pyrrolidone-phenol complexes. The spatial distributions showed the same order of percentage of absorption capacity observed experimentally for PVPP–phenols at low pH (A) and at neutral pH (B).
E.F. Durán-Lara et al. / Food Chemistry 168 (2015) 464–470
according to DE = ELUMO EHOMO, and this energy gap was compared with the experimental values of HPLC (Table 2). Two energy gaps were calculated, the first corresponding to the energy between the LUMO of the phenol and the HOMO of the polymer, and the second representing the LUMO of the polymer and the HOMO of the phenol. From these results, it was possible to see that the favored interaction was between the LUMO of the phenol and the HOMO of the polymer (size comprising 30 monomers). The results of energy gap between the HOMO of the polymer and the LUMO of each phenol were in accordance with the experimental data giving a correlation coefficient (r2) of 0.73 (Table 2). Moreover, the number of substituent groups involved in the phenolic compounds was used trying to explain the preference of PVPP towards quercetin, catechin and epicatechin. Table S4 (Supplemental Material) shows the different substituent groups present in each phenol. According to the data, it was possible to see that hydroxyls were not the only groups involved in the interaction with PVPP (Lorrain et al., 2013), and that the interaction was governed by more than only hydrogen bonds. In fact, it was possible to see that the compounds with higher affinity had more aromatic rings and hydroxyl groups (such as in quercetin, catechin and epicatechin), as opposed to the carbonated chains and carboxyl groups present in the less affine phenols. The former might be due to the need of the phenol to extend the LUMO zone to have a better possibility to interact with PVPP. Therefore, the interaction between these phenols and PVPP is at molecular orbital level, between the HOMO of the polymer and the LUMO of the phenols. This is due to the groups involved in the LUMO of the phenols, which increase their affinity for the PVPP polymer. When the energy gaps are compared, it can be seen that the lowest energy gap is obtained for quercetin-PVPP. This implies that quercetin and PVPP have a strong chemical enhancement, which favours the electronic transference from the PVPP to quercetin according to gallic acid, caffeic acid and 4-methylcatechol that have the higher gap energies. 4. Conclusions This work studied the adsorption, or trapping capacity, of six phenolic compounds by PVPP. The results indicated that their interaction has a direct correlation with the number of hydroxyl and aromatic rings presented in the phenolic molecules. The LUMO of phenolic compounds proves that their substituents play a key role in the increase of polymer affinity toward phenols. The interaction of phenols and PVPP occurs at molecular orbital level, between the frontier orbitals HOMO-PVPP and LUMO-phenols, respectively. That might be the reason why a higher electronic transfer between the phenols and PVPP was favored in the phenols that had a higher number of hydroxyl groups and aromatic rings. For this reason, the PVPP polymer has the highest affinity for quercetin, catechin and epicatechin. Interaction energy averages obtained through a computational chemistry strategy proved to be highly predictive for estimating the affinity of specific phenols by PVPP. Acknowledgements E.F.D.L. thanks Fondecyt (Postdoctoral Grant 3120178), Fraunhofer Chile Research, and Innova Chile CORFO (Code FCR-CSB 09CEII-6991). X.A.L.C. and F.A.S. thank CONICYT-PCHA/Doctorado Nacional/2013-21130553 and 2013-21130308 for doctoral scholarship. L.S.S. and F.D.G.N. thank project Anillo ACT 1107 (Integracion de la Biologia Estructural al desarrollo de la Bionanotecnologia).
469
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.foodchem.2014. 07.048.
References Andzelm, J., & Wimmer, E. (1992). Density functional gaussian-type-orbital approach to molecular geometries, vibrations, and reaction energies. Journal of Chemical Physics, 96, 1280–1303. Avila-Salas, F., Sandoval, C., Caballero, J., Guinez-Molinos, S., Santos, L. S., Cachau, R. E., et al. (2012). Study of interaction energies between PAMAM dendrimer and non-steroidal anti-inflammatory drug using a distributed computational strategy and experimental analysis by ESI–MS/MS. Journal of Physical Chemistry B, 116, 2031–2039. Baiano, A., Varva, G., De Gianni, A., Viggiani, I., Terracone, C., & Del Nobile, M. A. (2014). Influence of type of amphora on physico-chemical properties and antioxidant capacity of ‘Falanghina’ white wines. Food Chemistry, 146, 226–233. Balasundram, N., Sundram, K., & Samman, S. (2006). Phenolic compounds in plants and agri-industrial by-products: Antioxidant activity, occurrence, and potential uses. Food Chemistry, 99, 191–203. Becke, A. D. (1997). Density-functional thermochemistry 0.5. Systematic optimization of exchange-correlation functionals. Journal of Chemical Physics, 107, 8554–8560. Caceres-Mella, A., Peña-Neira, A., Parraguez, J., Lopez-Solís, R., Laurie, F., & Canals, J. M. (2013). Effect of inert gas and prefermentative treatment with polyvinylpolypyrrolidone on the phenolic composition of Chilean Sauvignon blanc wines. Journal of the Science Food and Agriculture, 93, 1928–1934. Cheynier, V. (2012). Phenolic compounds: from plants to foods. Phytochemistry Reviews, 11, 153–177. Dennington, R., II, Keith, T., Milliam, J., Eppinnett, K., Hovell, W. L., Gilliland, R. (2003). GaussView, Version 3.09, Semichem Inc., Shawnee Mission. Dong, Z. B., Liang, Y. R., Fan, F. Y., Ye, J. H., Zheng, X. Q., & Lu, J. L. (2011). Adsorption behavior of the catechins and caffeine onto polyvinylpolypyrrolidone. Journal of Agricultural and Food Chemistry, 59, 4238–4247. Fan, C. F., Olafson, B. D., & Blanco, M. (1992). Application of molecular simulation to derive phase diagrams of binary mixtures. Macromolecules, 25, 3667–3676. Feng, D. D., Wang, S. F., Zhuang, Q. X., Wu, P. P., & Han, Z. W. (2004). Semi-empirical calculation and spectroscopic study of protonated poly(p-phenylene benzobisoxazole) models. Polymer, 45, 8871–8879. Fernandez, M. J., Oliva, J., Barba, A., & Camara, M. A. (2005). Effects of clarification and filtration processes on the removal of fungicide residues in red wines (var. Monastrell). Journal of Agricultural and Food Chemistry, 53, 6156–6161. Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A., Cheeseman, J. R., et al. (2004). Gaussian 03, Revision C. 02, Gaussian Inc., Wallingford, CT. Gollucke, A. P., Aguiar, O., Jr., Barbisan, L. F., & Ribeiro, D. A. (2013). Use of grape polyphenols against carcinogenesis: Putative molecular mechanisms of action using in vitro and in vivo test systems. Journal of Medicinal Food, 16, 199–205. Gonzalez-Nilo, F. D., Urzua, M., Leiva, A., Gargallo, L., & Radic, D. (2003). Interaction energy analysis of polymer blends containing N-1-alkylitaconamic acid moiety. Journal of Macromolecular Science, Part B: Physics, 42, 1281–1291. Hostas, J., Rezac, J., & Hobza, P. (2013). On the performance of the semiempirical quantum mechanical PM6 and PM7 methods for noncovalent interactions. Chemical Physics Letters, 568, 161–166. Korth, M. (2010). Third-generation hydrogen-bonding corrections for semiempirical QM methods and force fields. Chemical Theory and Computation, 6, 3808–3816. Laborde, B., Moine Ledoux, V., Richard, T., Saucier, C., Dubourdieu, D., & Monti, J. P. (2006). PVPP-polyphenol complexes: A molecular approach. Journal of Agricultural and Food Chemistry, 54, 4383–4389. Le Bourvellec, C., & Renard, C. M. G. C. (2012). Interactions between polyphenols and macromolecules: Quantification methods and mechanisms. Critical Reviews Food Science Nutrition, 52, 213–248. Lee, C. T., Yang, W. T., & Parr, R. G. (1988). Development of the Colle-Salvetti correlation-energy formula into a functional of the electron-density. Physical Review B, 37, 785–789. Lorrain, B., Tempere, S., Iturmendi, N., Moine, V., de Revel, G., & Teissedre, P. L. (2013). Influence of phenolic compounds on the sensorial perception and volatility of red wine esters in model solution: An insight at the molecular level. Food Chemistry, 140, 76–82. Louli, V., Ragoussis, N., & Magoulas, K. (2004). Recovery of phenolic antioxidants from wine industry by-products. Bioresource Technology, 92, 201–208. Magalhães, P. J., Vieira, J. S., Goncalves, L. M., Pacheco, J. G., Guido, L. F., Barros, A. A., et al. (2010). Isolation of phenolic compounds from hop extracts using polyvinylpolypyrrolidone: Characterization by high-performance liquid chromatography–diode array detection–electrospray tandem mass spectrometry. Journal of Chromatography A, 1217, 3258–3268. Marquez, A., Serratosa, M. P., & Merida, J. (2014). Influence of bottle storage time on colour, phenolic composition and sensory properties of sweet red wines. Food Chemistry, 146, 507–514. McMurrough, I., Kelly, R. J., Byrne, J. R., & O’Brien, M. (1992). Effect of the removal of sensitive proteins and proanthocyanidins on the colloidal stability of lager beer. Journal American Society Brewing Chemistry, 50, 67–76.
470
E.F. Durán-Lara et al. / Food Chemistry 168 (2015) 464–470
Mcmurrough, I., Madigan, D., & Smyth, M. R. (1995). Adsorption by polyvinylpolypyrrolidone of catechins and proanthocyanidins from beer. Journal of Agricultural and Food Chemistry, 43, 2687–2691. Metropolis, N., Rosenbluth, A. W., Rosenbluth, M. N., Teller, A. H., & Teller, E. (1953). Equation of state calculations by fast computing machines. Journal Chemical Physics, 21, 1087–1109. MOPAC2012, Stewart, James J. P. (2012). Stewart Computational Chemistry, Colorado Springs, CO, USA. Nunes, C., Elisabete Ferreira, E., Freitas, V., Almeida, L., Barbosa, R. M., & Laranjinha, J. (2013). Intestinal anti-inflammatory activity of red wine extract: Unveiling the mechanisms in colonic epithelial cells. Food and Functional, 4, 373–383. Panarin, E. F., Kalninsh, K. K., & Pestov, D. V. (2001). Complexation of hydrogen peroxide with polyvinylpyrrolidone: Ab initio calculations. European Polymer Journal, 37, 375–379. Pople, J. A., & Beveridge, D. L. (1970). Approximate molecular orbital theory (1st ed.). New York: McGraw-Hill. Sen, K., Cabaroglub, T., & Yilmaz, H. (2012). The influence of fining agents on the removal of some pesticides from white wine of Vitis vinifera L. cv. Emir. Food and Chemical Toxicology, 50, 3990–3995.
Sergent, T., Piront, N., Meurice, J., Toussaint, O., & Schneider, Y. (2010). Antiinflammatory effects of dietary phenolic compounds in an in vitro model of inflamed human intestinal epithelium. Chemico-Biological Interactions, 188, 659–667. Silva Gontijo, V., Judice, W. A., Codonho, B., Oliveira Pereirac, I., Assis, D. M., Pereira Januário, J., et al. (2012). Leishmanicidal antiproteolytic and antioxidant evaluation of natural biflavonoids isolated from Garcinia brasiliensis and their semisynthetic derivatives. European Journal of Medicinal Chemistry, 58, 613–623. Stewart, J. J., & Stewart, J. J. P. (2007). Optimization of parameters for semiempirical methods V: Modification of NDDO approximations and application to 70 elements. Journal of Molecular Modeling, 13, 1173–1213. Sun, J., Liang, F., Bin, Y., Li, P., & Duan, C. Q. (2007). Screening non-colored phenolics in red wines using liquid chromatography/ultraviolet and mass spectrometry/ mass spectrometry libraries. Molecules, 12, 679–693. Villamor, R. R., & Ross, C. F. (2013). Wine matrix compounds affect perception of wine aromas. Annual Review of Food Science and Technology, 4, 1–20. Yilmaz, Y., & Toledo, R. T. (2006). Oxygen radical absorbance capacities of grape/ wine industry byproducts and effect of solvent type on extraction of grape seed polyphenols. Journal of Food Composition and Analysis, 19, 41–48.
