Inès Le Mao1, Gregory Da Costa1, Jean Martin1, Eric Pedrot1, Soizic Lacampagne1, François Guyon2, Tristan Richard1*
1Université de Bordeaux, Œnologie EA 4577, USC 1366 INRA, INP, ISVV, 210 chemin de Leysotte, 33882 Villenave d’Ornon, France
2Service Commun des Laboratoires, 3 Avenue du Dr. Albert Schweitzer, Pessac Cedex, France
*Corresponding author: tristan.richard@u-bordeaux.fr
Abstract
qNMR Metabolomic applied to wine offers many possibilities. The first application that is increasingly being studied is the authentication of wines through environmental factors such as geographical origin, grape variety or vintage (Gougeon et al., 2019). Another less common approach is from a qualitative point of view by studying the various oenological practices used that are an integral part of the elaboration of a wine. We wondered whether quantitative NMR could be used to dissociate the physical or chemical processes commonly used in oenology. The objective of this work was to provide a better understanding of the interactions between oenological processes and wine by determining the metabolites responsible for differentiation through 1H-NMR fingerprinting and chemometrics.
About 40 molecules were quantified on wine samples that have undergone several physical and chemical processes. Principal Component Analysis (PCA), Partial Least Squares Discriminant Analysis (PLS-DA) and S-score were performed for the analytical discrimination of winemaking processes.
The results obtained show that qNMR combined with chemometrics allows to dissociate not only physical processes such as filtration, but also chemical processes such as maceration temperature, enzymatic treatment and fining. In addition, the metabolites involved in the discrimination of these winemaking processes could also be determined.
Article extracted from the presentation held during Enoforum Web Conference (23-25 February 2021)
Introduction
Nuclear Magnetic Resonance (NMR) is a spectroscopic technique that has been increasingly used for food analysis over the last 20 years. Its non-specific and non-selective properties make it a tool of choice for metabolomics, a recent science that studies all primary and secondary metabolites in a biological sample. NMR has many advantages including the ability to simultaneously analyze different families of molecules in a single analysis without prior preparation and from a small volume of material. Moreover, it is a quantitative technique since the signal area is proportional to the concentration of metabolites detected on the spectrum. This is known as quantitative NMR (qNMR).
Applied to wine, qNMR metabolomics makes it possible to observe the impact of different parameters on the chemical composition of wine such as grape variety, geographical origin or vintage (Valls Fonayet et al., 2021). It is now known that proton NMR is a technique for distinguishing wines from different regions, grape varieties and vintages (López-Rituerto et al., 2012; Godelmann et al., 2013; Gougeon et al., 2018). These results demonstrated that NMR is an efficient tool to guarantee the authenticity of wines. Fraud is a widespread problem in the wine industry. To overcome this problem, the creation of a database was carried out to collect a large number of authentic samples where each analysis is considered as an identity card. The collection leads to the elaboration of a profile for each region, grape variety or vintage. Thus, if there is a suspicion of fraud, the suspect sample is either directly compared to the authentic sample provided by the chateau or compared to a large set of data.
This collection has led to the discovery that NMR spectra distinguish wines not only within the same region but within the same appellation and from châteaux only a few kilometers away. This finding suggested that NMR analysis, rather than being able to distinguish environmental parameters, could be used to discriminate winemaking processes. We have studied several processes commonly used in oenology: winemaking temperature, filtration and the use of enzymes. The aim of this work is to provide a better understanding of the interactions between oenological processes and wine by determining the metabolites responsible for differentiation through 1H-NMR fingerprinting and chemometrics.
Material and methods
Sample preparation and analyses were done as described in our previous article (Gougeon et al., 2019a). Briefly, wine samples were mixed with phosphate buffer and deuterated water solution (ratio 7:2:1). Trimethylsilylpropanoic acid sodium salt (TMSP) and calcium formate were used as frequency and quantification references, respectively. The pH was adjusted to 3.1 using a BTpH unit (Bruker BioSpin, Germany). The 1H-NMR spectra were acquired on a 600 MHz spectrometer (Bruker BioSpin, Germany). Two specific sequences were used for the quantification of wine compounds. The ZGPR pulse program was used for water signal suppression. NOESYGPPS1D sequence was performed for the suppression of water and ethanol signals. Spectra were processed using Topspin 3.2 (Bruker Biospin, Germany) and analyzed using MestReNova 12 (Mestrelab Research, Spain). Spectra were summited to phase and baseline corrections before data extraction. Finally, compounds were semi-automatically quantified using the plugin Simple Mixture Analysis (SMA) of MestReNova software.
Comparison between two wines were performed using an adapted z-score value approach allowing the calculation of a similarity score (s score) for each wine constituent (Gougeon et al., 2019b). A-s-score higher than 2.0 (confidence interval of 95%) suggests significate difference between the two wines. SIMCA 16 software (Umetrics, Sweden) was used for unsupervised principal components analysis (PCA) and partial least squares discriminant analysis (PLS-DA). PCA was carried out in first stage to visualize the data. In a second stage, PLS-DA was completed to refine the discrimination between sets.
Results and discussion
1H-NMR Spectroscopy
A representative spectrum of wine after water and ethanol suppression was presented in Figure 1. Compounds were identified by using pure chemical standard, 2D-NMR experiments and literature data (Fotakis et al., 2013). Today, nearly fifty wine constituents were identified and quantified including organic acids, alcohols, amino acids, phenolic compounds and sugars (Gougeon et al., 2019a).
To perform an accurate quantification, 1H-NMR spectra were exported into MestReNova 12 and treated using the plugin SMA (Gougeon et al., 2018). Peak deconvolution was realized using global spectral deconvolution method (Cobas et al., 2011). The precision of the method was validated by comparison with reference samples and by comparison with official OIV methods (Gougeon et al., 2019b).
Wine authenticity
In order to guarantee the authenticity of a wine, two approaches can be used. The first approach is to directly compare suspect sample with the sample from the authentic bottle. If it is a fraud, the two spectra will be different either in composition (differences detected at the level of the molecules observed) or in content (concentrations of the different constituents of the wine). To allow the comparison of the data and to take into account the wine aging in bottle, a statistical tool called s score has been developed (Gougeon et al., 2019b). This technique was validated by comparison of usual wine analysis performed using OIV official methods and qNMR metabolomics. Compared to classical methods, NMR could reduce the volumes of wine required and the duration of analysis as a first approach of authenticity control of wines. To illustrate the process, the s score comparison of an authentic and a fake wine to the reference wine provided by the winery were presented in the Figure 2. Wines presenting more than four s scores higher than two could be considerate significantly different.
The alternative, when an authentic sample cannot be obtained, is to compare the suspect sample with a database. Indeed, we know that the composition of wine varies according to grape variety, vintage, geographic origin and viticultural practices. Thus, the collection of a large number of samples leads to the elaboration of a profile for each appellation, region or grape variety. When an unknown sample is confronted with the database, we are able to determine its origin through statistical analysis (Gougeon et al., 2019a). To illustrate the ability of qNMR metabolomics to discriminate wines, analytical data of wines from different vintages of three close wineries of Bordeaux were compared by partial least squares discriminant analysis (PLS-DA). As shown in Figure 3, the three sets of wines are clearly discriminate over a near ten-year period
Winemaking processes discrimination
Three processes commonly used in oenology were studied: thermovinification, filtration and enzyme addition. The results are presented in Figure 4.
Thermovinification is a process initially used to eliminate the laccases produced by Botrytis cinerea, responsible for organoleptic defects. In this study two sets of wines processed using thermovinication or conventional temperature were analyzed by qNMR metabolomics. The PLS-DA score plot (Figure 4A) show that thermo-vinified samples are significantly different from samples vinified at conventional temperatures. The loading plot indicates that the molecules involved in the discrimination are: galacturonic, lactic and succinic acids and xylose for thermovinified wines; and fructose, proline, choline and acetic acid for samples vinified at conventional temperature. These results prove that thermovinification does not only have an impact on phenolic compounds and the color of the wine, but also on other compounds such as organic acids, sugars and amino acids.
Filtration is a post-fermentation process that contributes to the clarification and conservation of wines. Filtered and unfiltered wines were analyzed by qNMR metabolomics. The PLS-DA score plot (Figure 4B) shows that NMR analyses discriminate filtered samples from unfiltered samples. The molecules most impacted by filtration are succinic and lactic acids, isobutanol, glycerol and acetoin. These data suggest that filtration process not only impacts large molecules but also low molecular weight ones.
The use of enzyme is a very widely used process for the clarification and stabilization during winemaking. Wines with or without addition of enzymes were analyzed by qNMR metabolomics. Analytical results also show that this process has an impact on the wine metabolome since the untreated samples are significantly different from those that have undergone enzyme addition (Figure 4C). The PLS-DA loadings show that the molecules involved are: glucose, galacturonic acid, methanol, ethyl lactate, mannitol, arabinose and fructose for treated samples and catechin, histidine and isoleucine for untreated samples. This is the first time that molecules impacted by the use of enzymes are highlighted. Indeed, the differences are generally found by sensory analysis but never by classical chemical analysis.
Conclusion
The results obtained illustrate the capacity of qNMR combined with chemometrics to discriminate wines. Applied to winemaking processes, qNMR metabolomics allows discriminating not only physical processes such as filtration, but also chemical processes such as thermovinification, and enzymatic treatment. In addition, the metabolites involved in the discrimination of these winemaking processes are highlighted. To conclude, 1H-NMR metabolomics is a fast technique that could be used as a tool to help professionals decide on technical itineraries.
Acknowledgments
This work was supported by Conseil Interprofessionnel du Vin de Bordeaux (CIVB), Conseil Régional d’Aquitaine, FranceAgriMer program (grant number 2014–0785), Château Mouton Rothschild, Fondation of Bordeaux (donors: Baron Philippe de Rothschild SA, Chateau Cheval Blanc, Château Lafite Rothschild, Le Domaine Clarence Dillon, Château Petrus), and Biolaffort. The experiments were performed thanks to the Bordeaux Metabolome Facility and MetaboHUB program (ANR-11-INBS-0010 project).
Cobas C., Seoane F., Domínguez S., Sykora S., Davies A.N., 2011. A new approach to improving automated analysis of proton NMR spectra through Global Spectral Deconvolution (GSD). Spectrosc. Eur. 23, 26-30.
Fotakis C., Kokkotou K., Zoumpoulakis P., Zervou M., 2013. NMR metabolite fingerprinting in grape derived products: An overview. Food Res. Int. 54, 1184-1194.
Godelmann R., Fang F., Humpfer E., Schütz B., Bansbach M., Schäfer H., Spraul M., 2013. Targeted and Nontargeted Wine Analysis by 1H NMR Spectroscopy Combined with Multivariate Statistical Analysis. Differentiation of Important Parameters: Grape Variety, Geographical Origin, Year of Vintage. J. Agric. Food. Chem. 61, 5610-5619.
Gougeon L., Da Costa G., Le Mao I., Ma W., Teissedre P.L., Guyon F., Richard T., 2018. Wine Analysis and Authenticity Using 1H-NMR Metabolomics Data: Application to Chinese Wines. Food Anal. Meth. 11, 3425-3434.
Gougeon L., da Costa G., Guyon F., Richard T., 2019a. 1H NMR metabolomics applied to Bordeaux red wines. Food Chem. 301, 125257.
Gougeon L., Da Costa G., Richard T., Guyon F., 2019b. Wine Authenticity by Quantitative 1H NMR Versus Multitechnique Analysis: a Case Study. Food Anal. Meth. 12, 956-965.
López-Rituerto E., Savorani F., Avenoza A., Busto J.H., Peregrina J.M., Engelsen S.B., 2012. Investigations of la Rioja terroir for wine production using 1H NMR metabolomics. J. Agric. Food. Chem. 60, 3452-3461.
Valls Fonayet J., Loupit G., Richard T., 2021. MS- and NMR-metabolomic tools for the discrimination of wines: Applications for authenticity, Adv. Bot. Res. 98, 297-357.
Abstract Phenolic compounds are responsible for the most important red..
Abstract This study aims at exploiting an undervalued winemaking by-product,..
Francesca Borghini and Stefano Ferrari ISVEA srl, Via Basilicata 1-3-5..
The seasonal character of harvests limits winemaking considerably, especially oenological..
