An exploratory metabolomic analysis of non-alcoholic wines via 1H NMR, ICP-MS and chemometrics

Abstract

Non-alcoholic wines contain Ôëñ 0.5% alcohol by volume, which is comparable to the alcoholic content of grape juices. These wines are consumed to avoid intoxication while being nutritionally beneficial (i.e. low in sugar, low in calories and high in antioxidants). Although non-alcoholic wines have been growing in popularity in the past 20 years, global inconsistencies in their classifications remain. Many dealcoholised wines, premium grape juices, and a combination thereof, are being sold as non-alcoholic wines with limited understanding of their metabolomic compositions. It has been well established that wines and grape juices are complex mixtures containing numerous metabolites (i.e. sugars, amino acids, organic acids, alcohols, polyphenols, minerals, and so on). The complexity of metabolomic profiling is increased as the chemical composition of wines are further influenced by the external factors of fermentation, oenological practices, geographical origin, ageing and blending. In the case of many alcoholic and nonalcoholic wines, the process of dealcoholisation adds to this metabolomic complexity. The analytical techniques of 1H nuclear magnetic resonance (NMR) and inductively coupled plasma mass spectroscopy (ICP-MS) can provide comprehensive metabolomic profiles of the non-volatile wine matrix (i.e. organic and elemental metabolites). Since the large amount of chemical information obtained via these techniques, observing trends in the resulting data matrices is difficult without the aid of chemometric analysis. The application of unsupervised learning methods, such as principal component analysis (PCA), hierarchical clustering analysis (HCA) and K-Means clustering (K-Means), can reduce data dimensionality and reveal natural clusters within the sample. A better understanding of the metabolomic differences between dealcoholised wines and premium grape juices can be gained through the methods described above. Comprehensive analysis of this kind can better inform formal classification of nonalcoholic wines as well as explicate their nutritional claims. In this research, 78 commercial non-alcoholic wines and their non-volatile metabolites were analysed via 1D low field and high field 1H NMR and ICP-MS. Characterisation of organic metabolites was further aided by 1H NMR Diffusion Ordered Spectroscopy (DOSY). Unsupervised learning was used to visualise the natural clusters and sample variance in spectral 1H NMR data and quantitative ICP-MS data. Previous literature has reported similar metabolomic profiling of alcoholic wines, non-alcoholic beers and fruit juices, however, to the best of the researcher's knowledge, these profiling strategies have not previously been conducted on commercial non-alcoholic wines nor to this scale.

Date of Award	2021
Original language	English