Role Of Grape-Extractable Polyphenols in The Generation Of Strecker Aldehydes And in The Instability Of Polyfunctional Mercaptans During Model Wine Oxidation Part 1

Please contact oscar.xiao@wecistanche.com for more information

ABSTRACT: Polyphenolic fractions from Garnacha, Tempranillo, and Moristel grapes were reconstituted to form model wines of identical pH, ethanol, amino acid, metal, and varietal polyfunctional mercaptan (PFM)contents. Models were subjected to a forced oxidation procedure at 35°C and to an equivalent treatment under strict anoxia. Polyphenolic profiles significantly determined oxygen consumption rates(5.6-13.6 mg L-Iday-I), Strecker aldehyde(SA)accumulation (ratios max/min around 2.5), and levels of PFMs remaining (ratio max/min between 1.93 and 4.53). By contrast, acetaldehyde accumulated in small amounts and homogeneously (11-15 mg L-'). Tempranillo samples, with the highest delphinidin and prodelphinidins and smallest catechin, consume O, faster but accumulate less SA and retain the smallest amounts of PFMs under anoxic conditions, Overall. SA accumulation may be related to polyphenols, producing stable quinones. The ability to protect PFMs as disulfides may be negatively related to the increase in tannin activity, while pigmented tannins could be related to a 4-methyl-4-mercaptopentanone decrease.

KEYWORDS: aroma, longevity, premix, shelflife, quinones, disulfides, nucleophiles, phenylacetaldehyde, methional, 3-mercaptoethanol

Please click here to know more

INTRODUCTION

Wine longevity is a complex multifactor phenomenon in which the weight of the different factors is not well known. One of the key factors of wine longevity is related to its resistance to oxidation. This property can be defined as the ability of the wine, under exposure to oxygen, to keep its color, avoid the accumulation of acetaldehyde and Strecker aldehydes(SAs), and keep as long as possible labile varietal aroma compounds, such as polyfunctional mercaptans(PFMs).

The formation of acetaldehyde in the absence of free SO has been widely studied, although some details of the process are not completely understood. The hydrogen peroxide formed in the first two-electron reduction of O, taken from an o-diphenol, reacts with Fe(III) cations to form the powerful hydroxyl radical, OH".Once formed, this radical is a very powerful oxidant, which reacts at diffusion-controlled rates. It is, therefore, proposed that it reacts close to its site of production with the first potential substrate it encounters. This implies that most of it oxidize ethanol to form 1-hydroxyethyl radical(1-HER), and this, in the presence of oxygen, forms 1-hydroxyethyl peroxyl which decomposes into acetaldehyde.,However, the reaction is quite complex. It has been suggested that o-diphenols can quench the 1-HER radical, and it has been demonstrated that cinnamic acids are particularly efficient at trapping it. It has been also suggested that although the reaction of mercaptans with H, O, is kinetically very slow(10-2 or 10-3 M-1 s-1 for cysteine), these compounds can reduce the 1-HER back to ethanol,°which is kinetically much faster(10°M-1s-1).7 A recent report has shown that, quite paradoxically, some antioxidants such as ascorbic acid apparently inhibit the 1-HER radical but do not prevent the accumulation of acetaldehyde, suggesting that in fact, this compound accelerates the oxidation of 1-HER into acetaldehyde. Finally, acetaldehyde could react with the nucleophilic positions of wine polyphenols, particularly in the A ring of the flavonoids, to form different combinations, such as ethylidene-bridged dimers or proanthocyanins."Consequently, the accumulation of acetaldehyde in response to O, consumption is very difficult to predict.

Cistanche can improve the immunity

The SAS, isobutanol, 2-methylbutanal, isovaleraldehyde, methional, and phenylacetaldehyde, are powerful odor molecules, which, along with acetaldehyde, are mainly responsible for wine oxidative aroma. Different studies have demonstrated or suggested the existence of different SA formation routes. One of them is the own fermentation, in which these compounds can be formed via the Ehrlich pathway and remain unnoticed under the form of hydroxyalkylsulfo-nates, the nonvolatile adducts they form with SO. These forms can regenerate free aldehydes during wine oxidation, as SO, is consumed. The second and the most important formation pathway seems to be the Strecker degradation of the corresponding amino acids.1This degradation requires an a-dicarbonyl, which can be a fermentation byproduct, such as methylglyoxal or diacetyl, or the quinones of o-diphenols formed during oxidation, for whose formation metal cations and oxygen are essential. Some authors have demonstrated that at high temperatures (80 and more than 130 ℃C), some polyphenols are more efficient than others for producing phenylacetaldehyde.4,15Under those conditions, single nuclei ortho-diphenols, such as catechol, 4-methyl catechol, and 2,5-dihydroxybenzoic acid, or vicinal triphenols, such as pyrogallol or gallic acid, seem to be more efficient than flavonols, such as catechin or epicatechin (EC), in the accumulation of phenylacetaldehyde. The influence of polyphenols in the ability of a wine to accumulate acetaldehyde and SAs has been indirectly suggested by partial least-squares (PLS)modeling. All models explaining the accumulation rates of aldehydes have in common negative coefficients for anthocyanins, which was therein interpreted as a consequence of their ability to quench aldehydes.°Therefore, the ability of a wine to accumulate SAs is related to the presence of the amino acid precursors, to its tendency to form amino acid reactive quinones, and to its capacity to quench formed aldehydes. Unfortunately, none of these three characteristics have been defined for the different wine polyphenols under wine-like conditions.

Regarding varietal aroma, the most oxygen-sensitive aroma compounds are PFMs, being the most important are 4-methyl-4-mercaptopentanone(4MMP),3-mercaptohexanol(3MH), and its acetate,3-mercaptohexyl acetate(MHA).These compounds are quite reactive. They can form disulfides as demonstrated by Roland et al., but they can also react with wine quinones, as demonstrated by Nikolantonaki et al.8,19 Therefore, their stability will depend again on different compositional factors such as the wine ability to quench the 1-HER radical, the presence of other major mercaptans to form disulfides and the number and reactivity of quinones formed. It follows that such stability will be closely related to the wine polyphenolic composition but, again, the role of the different polyphenols is not known.

The main goal of the present research is to assess, specifically, the role played by the polyphenolic composition on the ability of wine models to accumulate SAs and to retain PFMs and other varietal aroma compounds during oxidation.

MATERIAL AND METHODS

Reagents and Standards. Hydrochloric acid (37%), sodium hydrogencarbonate,and sodium metabisulfite 97% were obtained from Panreac(Barcelona, Spain).L(+)-tartaric acid(99%), glycerol (99,5%), iron(II) chloride tetrahydrate (>99%),manganese(II)chloride tetrahydrate(>99%), copper(I) chloride(99,9%),L-leucine (Leu)(>98%), L-isoleucine(Ile)(>98%), D-valine (Val)(>98%),L-phenylalanine(Phe)(>98%),D-methionine(Met)(>98%),L-cysteine hydrochloride anhydrous (>98%),L-glutathione (GSH) reduced (>98%),hydrogen sulfide(≥99.5%),ethanethiol(97%),2,4-dinitrophenylhydrazine(DNPH)(97%),and acetaldehyde (>99,5%) were obtained from Sigma-Aldrich Madrid, Spain, and malvidin 3-O-glucoside, ovalbumin(≥90%),(-)-EC(purity ≥90%), phloroglucinol, liquid chromatography (LC)-mass spectrometry (MS)grade formic acid used as the mobile phase additive, and all the solvents for the phloroglucinolysis reactions, extraction, isolation, and analysis were purchased from FLUKA Sigma-Aldrich St. Louis, USA.4-Mercapto-4-methyl-2pentanone(4MMP)1% in polyethylene glycol (PG)and 3-MHA were obtained from Oxford Chemicals (Hartlepool, U.K.). 3MH was obtained from Lancaster(Strasbourg, France), as 4-mercapto-4-methyl-2pentanone-d10 (4MMP-d10), 3-MHA-ds(MHA-ds), and 3-mercaptohexanol-ds(3MH-ds).LiChro-lut EN sorbent, 1 mL cartridge and polytetrafluoroethylene frits, dichloromethane, and ethanol were purchased from Merck(Darm-stadt, Germany). Sep Pak-C18 resins, prepacked in 10 g cartridges, were obtained from Waters (Ireland). L-Cysteine hydrochloride anhydrous (99%), sodium citrate trihydrate, and methanol of LC-MS

LiChrosolv grade used for the preparation of mobile phases was obtained from Fluka. Sodium hydroxide 99%, high-performance LC (HPLC)-grade acetonitrile, and o-phosphoric acid were purchased from Scharlab (Sentmenat, Spain).Isobutyraldehyde (Isobut)(99%), 2-methylbutanal (2MB)(95%),3-methylbutanal (3MB)(95%), phenylacetaldehyde (PheAc)(95%) and methional (98%),2-methylpentanal (98%),3-methylpentanal (97%), and O-(2,3,4,5,6 pentafluorobenzyl)hydroxylamine hydrochloride(PFBHA)98% were supplied by Merck USA. Phenylacetaldehyde-d2 (95%)and methional-d2 were purchased from Eptes (Vevey, Switzerland). Water was purified in a Milli-Q system from Millipore (Bedford, UK).Highest purity(>98%)grade(+)-catechin,(-)-EC,(-)-gallocatechin(GC),(-)-epigallocatechin (EGC),(-)-EC gallate (ECG), procyanidin B1, and procyanidin B2 were obtained from TransMIT PlantMetaChem （Gießen， Germany）. The phloroglucinolated derivatives EC 4-phloroglucinol, EC-gallate 4-phloroglucinol, and EGC 4-phloroglucinol were prepared according to Arapitsas et al, 2021.2 Polyphenolic and Aroma Fractions. The 15 polyphenolic aromatic fractions (PAFs)were extracted from 15 lots of grapes from three different Spanish wine-making regions (La Rioja, Ribera del Duero, and Somontano)and three different grape cultivars(7 from Tempranillo, 6 from Garnacha, and 2 from Moristel), as described in Alegre et al.2 Briefly, 10 kg of grapes were collected at technological maturity, kept at 5℃C during the transport to the experimental cellar, destemmed and crushed in the presence of 50 mg/Kg of potassium metabisulfite and ethanol(adjusted to 15% v/v), and left in the dark at 13°C for 7 days in closed recipients with no headspace after pressing to obtain the liquid mistelle (ethanolic must), which after sterile filtration was stored at 5°C in 750 mL wine bottles closed with a natural cork and no headspace. Then, 750 mL aliquots were dealcoholized by rotary evaporation at 23 ℃C(20 bar)to a final volume of 410 mL and then extracted in a 10 g Sep Pak C18 cartridge. Sugars, acids, amino acids, and ions were removed by cleanup with water acidified at pH 3.5. PAFs were eluted with 100 mL of absolute ethanol and kept at -20 °C.

Preparation of Model Wines. This operation was carefully carried out inside a glovebox (complex) containing less than 1 ppm O2. The 100 mL ethanolic extracts were reconstituted with water containing 5 g/L tartaric acid, and pH adjusted at 3.5 and spiked with glycerol(5g/L),FeCl·4 H,O(5 mg/L), MnCl·4 H,O(0,2mg/L), and CuCl(0,2 mg/L)to form 750 mL of model wines 13.3%(v/v)in ethanol. The models were left to stand for 2 weeks within the anoxic chamber and were then spiked with 200 ug/L H, S,25 ug/L ethanethiol,10 mg/L cysteines, and 10 mg/L GSH and left under strict anoxia for 2 additional weeks. After this, the models were spiked with 10 mg/L of Leu， lie， Val， Phe， and Met and with 100μg/L of the three PFMs:4MMP, MHA, and 3 MH. The anoxic controls were prepared by distributing three 60 mL aliquots of each model in three 60 mL screw-capped glass tubes (Wit Deluxe, Denmark), tightly closed and double vacuum bagged, including a layer of powder containing an O2 scavenger (AnaeroGen from Thermo Scientific Waltham, Massachusetts, United States)between both bags.

Forced Oxidation Procedure. The model wines were taken out of the glovebox, saturated with air by vigorous shaking, and then distributed in 60 mL Wit-tubes of internal volume perfectly known and containing Pst3 Nomasense oxygen sensors to measure dissolved oxygen in the liquid sample. Each tube contained the volumes of liquid and headspace required to deliver 50 mg of O, per L of liquid, as described by Marrufo-Curtido et al.22 Tubes were incubated in an orbital shaking thermostatic bath(Grant instruments OLS Aqua Pro)at 35°C for 35 days. Dissolved oxygen was daily controlled.

Chemical Characterization of the PAFs. The detailed analytical conditions are given in the Supporting Information. Anthocyanins were analyzed by ultra-HPLC-MS/MS, as described by Arapitsas et al.2 Flavanols, flavonols, and hydroxycinnamic acids were analyzed, as described by Vrhovsek et al.,24 by UHPLC-MS/MS. The mean degree of polymerization(mDP)was determined by UPLC-MS/MS analysis of the phloroglucinol reaction, as described by Arapitsas et al.20 Tannin activity and total and pigmented tannins were determined by UHPLC with photodiode array detection(280 and

520 nm) at four different temperatures(30,35,40, and 45 °C), as the specific enthalpy of interaction between tannins and a hydrophobic surface(polystyrene divinylbenzene HPLC column), as proposed by Yacco et al.5The concentration of total and pigmented tannins were determined in the chromatogram made at 30°C and they were reported in EC equivalents and area data, respectively.

Chemical Characterization of Oxidized and Unoxidized (Controls) Wine Models. Total acetaldehyde was determined by HPLC with ultraviolet(UV) detection after previous derivatization with DNPH, as described by Han et al.6

Total SAs were analyzed by GC-MS analysis after derivatization with PFBHA. Briefly, samples are introduced within the anoxic chamber and 12 mL aliquots spiked with the internal standards(2-methylpentanal,3-methylpentanal, phenylacetaldehyde-d2, and me thionyl-d2). Samples are taken out and incubated at 50 ℃C for 6h to ensure equilibration. After this， 360μL of a 10 g/L PFBHA solution are added and the reaction is developed at 35℃C for 12 h 10 mL of the sample is then extracted in 1 mL cartridges packed with 30 mg of LiChrolut-EN resins. The cartridge is washed with 10 mL of a solution containing 60% methanol and 1% NaHCO,and then dried and eluted with 1.2 mL of hexane. Three microliters of this extract are injected in the splitless mode in the GC-MS system.

Free PFMs are determined by GC-MS in the negative chemical ionization mode using the procedure described by Mateo-Vivaracho et al.7 Total PFMs are the sum of the free forms and those forming disulfides with themselves or with other mercaptans. For the determination of this total fraction, tris(2-carboxyethyl)phosphine is added to the sample in the anoxia chamber at a concentration of 1 mM prior to the analysis in order to reduce the disulides back to mercaptans.7

Varietal aroma compounds, linalool, geraniol, and 1,1,6-trimethyl-1,2-dihydro naphthalene (TDN), are determined by GC-MS using the procedure described by Lopez et al.9

The color was determined by the measurement of absorbances at 420, 520, and 620 nm as recommended by the OIVand total polyphenol index(TPI)by measurement at 280 nm.

Tannin activity was measured as is described in the Supporting Information.

Redox potential was measured within the anoxic chamber with a commercial platinum electrode versus an Ag-AgCl(s) reference electrode(HI3148 HANNA, instruments, USA)in a potentiometer HI98191 also from HANNA.

Data Analysis. Basic statistical analyses were carried out with an Excel spreadsheet. Analysis of variance(ANOVA) was carried out with XLSTAT version 2015(Addinsoft, XX). PLS modeling was carried out with Unscramble vs(Camo, Norway).

As main data were differences between oxidized samples and controls, their uncertainty was estimated by applying the basic theory of error propagation attending to the formula

RESULTS AND DISCUSSION

The experimental setup is based on the preparation of wine models with standardized composition in metals, amino acids, PFMs, alcoholic degree, and pH, so that the single difference between the wine models in the study are the polyphenolic profles extracted from the grapes. These were from different grape cultivars and different winemaking areas of Spain. The final reconstituted wine models were subjected to an oxidative aging treatment, in which samples were given 50 mg L-I

oxygen and were left for 35 days at 35℃ and to equivalent storage in strict anoxia used as a control.

Overview of Changes Introduced by Oxidation and Effect of Cultivar. The major changes introduced by oxidation, in comparison with the corresponding anoxic controls, are summarized in Table l and in Figure 1 (the complete set of results of the experiment can be found in Supporting Information, Tables S1-S6). Data in Table 1 are the average increments (positive) or decreases (negative)caused by oxidation in the different compositional parameters registered for the individual samples (left part of the table)or averaged by cultivar (right part of the table).

In general, the table reveals that oxidation causes increases of great magnitude in redox potential, tannin activity, and in the levels of SAs and increases of moderate magnitude in total tannins and acetaldehyde. Similarly, oxidation causes decreases of great magnitude in free and total PFMs and of moderate magnitude in TPI, pigmented tannins, and TDN. Most of these changes were expected, although there are very few previous reports about tannin activity, and the decrease of TDN with oxidation has not been previously observed. Average levels of linalool and geraniol did not change significantly with oxidation.

As samples exclusively differ in their polyphenolic composition, differences between samples should be entirely attributed to differences in their specific or varietal polyphenolic profiles. The significance of the effects exerted by these profiles is assessed by means of the p(F) values obtained in the corresponding ANOVAs. Regarding specific sample effects, results in Table 1 reveal that the polyphenolic composition exerted a deep effect on the magnitude and in some cases even on the nature of the effects introduced by oxidation. In fact, changes in all measured chemical parameters, except in the total levels of 4MMP, were significantly related to the polyphenolic profile. Many of the changes were also significantly related to the grape cultivar, as can be seen in the last column of the table. Remarkably, increases in total tannins, acetaldehyde, and tannin activity were not related to the cultivar.

The effects of the varietal polyphenolic profile are most clearly seen in the principal component analysis(PCA)plot given in Figure 1. The figure shows the projection of samples and variables in the plane of the two first principal components obtained from the data matrix containing oxygen consumption rates(OCRs) and the mean(average by replicates)increases or decreases caused by oxidation (vs the anoxic controls)in the 15 different samples. Note that in such a figure, the directions of the variable loadings indicate higher increases for variables increasing with oxidation, but smaller decreases for those decreasing. In any case, the figure reveals the existence of a strong varietal influence because the samples containing polyphenols extracted from Tempranillo are clearly separated from those extracted from Garnacha and Moristel. Those containing polyphenols from Tempranillo consumed oxygen much faster, ended up with less residual oxygen and hence lower redox potential, lost more TPI, more pigmented tannins, and more color, but they lost fewer PFMs due to oxidation and accumulated smaller levels of SAS. Results will be commented on and discussed in more detail later.

OCRs and Redox Potential. OCRs were clearly varietal dependent, as can be seen in Table 1. Samples containing polyphenols from Tempranillo consumed on average 11.0 mg/L O, per day in the first period of oxidation(4 days), while those from Garnacha consumed just 6.6 and those from Moristel 6.1 mg/L per day. The oxidation experiment was finished after 35 days, regardless of whether the O, had been completely consumed or not. This means that samples consuming O more slowly contained higher final residual levels of O, and consequently, higher redox potentials. Samples with PAFs from Moristel were particularly poor at O, consumption so that in the 35 days, they left unconsumed a total of 7.08± 2.2 mg of oxygen per liter of wine(accounting that remaining in the headspace) and their average redox potential was 190 mV. Those samples with PAFs from Garnacha left unconsumed just 2.87± 1.61 mg/L and ended with an average redox potential of 152 mV, while those from Tempranillo left just 1.24±0.25 mg/L and ended with a redox potential of 60.5 mV.

OCRs were positively and significantly correlated to total tannins, to their mDP, to total prodelphinidins, and to the sample content in 3-monoglucoside anthocyanins(delphinidin, petunidin, and cyanidin), as summarized in Table 2. These correlations were expected. Delphinidin and prodelphinidins are easily oxidizable wine polyphenols due to the three vicinal hydroxy groups in the B ring and have been previously found correlated to OCRs. Anthocyanins are more reactive toward superoxide radicals than catechin, and it is known that polymeric tannins are more antioxidant than monomeric forms.33

The negative correlations of OCRs with catechin and to the total content in flavanols, shown in Table 2 may be just statistical artifacts because in the present case, samples with higher levels of catechin and flavanols have also a lower concentration of anthocyanins.

Color and Tannin Activity. Differences in color index introduced by oxygen were not very intense but follow a varietal pattern, as can be seen in Table 1. In the case of samples containing polyphenols from Garnacha and Moristel, the color remained mostly unchanged, while those extracted from Tempranillo lost in average 1.5 units of color, which represents a loss of 10% of the total color of the sample. This is related to their highest OCRs previously seen, confirming that anthocyanins are quickly oxidized.

Tannin activity refers to the specific enthalpy of interaction between tannins and a hydrophobic surface (polystyrene divinylbenzene HPLC column). This parameter has been related to the perception of astringency and dryness in the mouth, and as seen in Table l, it strongly and significantly increases with oxidation in most samples in a nonvarietal related way. Changes were not related to any polyphenolic compositional parameter. However, a significant positive correlation with the redox potential measured in the samples stored in anoxia was observed (leaving out one sample of Tempranillo, r= 0.71, significant at p=0.0027). Although the true meaning of the redox potential in wine and wine-like media is controversial,3in the complete absence of oxygen and in standardized model wine, it can be hypothesized that more negative values of redox potential should be related to higher levels of H, S and of mercaptans, including cysteine and GSH." As the single source of these compounds in our samples is the initial dosage, which was the same for all samples, differences should be most likely related to the specific reactivity of the polyphenolic fractions to mercaptans, as it will later be commented in the PFM section. Therefore, it can be hypothesized that stronger increases in tannin activity during oxidation may be linked to polyphenolic fractions most reactive to mercaptans.

This article is extracted from https://doi.org/10.1021/acs.jafc.1c05880 J. Agric. Food Chem. 2021, 69, 15290−15300