Identification of novel aromatic precursors in winemaking grapes using an optimized fractionation and UHPLC-MS analysis

Abstract

Winemaking grapes contain a diverse array of non-volatile precursors that become noticeable only after hydrolysis reactions or molecular rearrangements, during which aroma compounds are generated and released [1]. Among these, glycosidic precursors are the most abundant and play a key role in the development of wine aroma [2]. Although direct determination of these molecules is possible using UHPLC-MS [3], it remains challenging due to the lack of commercial standards, the wide variety of precursors associated with the same aroma, and their typically low concentrations. A deeper understanding of the nature and behaviour of these molecules during the different stages of the winemaking process would allow for the optimisation of certain stages, such as maceration time or enzyme addition, among others. This study aims to identify the largest possible number of aroma precursors through an optimized fractionation approach [4], involving a sequence of semi-preparative chromatographies, first by size exclusion and then by normal phase fractionation. This method is designed to reduce the excessive complexity of grape samples in UHPLC-MS analyses, thereby facilitating precursor identification.

For this purpose, a phenolic-aromatic fraction from Garnacha grapes was subjected to silica gel fractionation, improving a previously established protocol by increasing the sample size and adjusting the polarity of the mobile phases. A total of 96 fractions were obtained, 92 of which released aroma upon hydrolysis. These fractions were analysed using SPME-GC-MS to identify where the precursors of key varietal aromas -such as terpenes, norisoprenoids, phenols, vanillins, and cinnamates- were located. The hydrolysates from 55 of these fractions contained aroma molecules of interest and were subsequently analysed by UHPLC-MS to detect potential precursors. The identification process was based on three criteria: a) the expected molecular mass of the precursor, b) the presence of fragments in the MS/MS spectrum that were consistent with the precursor’s structure, and c) the correlation between the precursor signal in UHPLC-MS and the corresponding aroma signal in GC-MS for the same fractions.

As a result, 175 precursors were identified, including 67 novel compounds not previously reported in the literature. Among these, 54 precursors were confirmed based on all the three identification criteria. The majority of the identified precursors were disaccharides (79) and monosaccharides (63), with a smaller proportion of trisaccharides (33). Noteworthy among them were 54 terpene precursors, 24 phenol precursors, 22 vanillin derivatives, and 14 cinnamate derivatives. This represents both a quantitative and qualitative advancement in the evaluation of the potential aromatic quality of winemaking grapes using UHPLC-MS. Furthermore, this approach would be useful for proposing new winemaking techniques or for developing new grape-derived products. It will also be essential for optimising grape quality and mitigating some effects of climate change.

References

[1] Ferreira, V., and Lopez, R. (2019). Biomolecules. 9(12), 818.

[2] Liu, J., Zhu, X.-L., Ullah, N., and Tao, Y.-S. (2017). Journal of Food Science. 82(2), 248-259.

[3] Caffrey, A., Lerno, L., Zweigenbaum, J., and Ebeler, S. E. (2020). Journal of Agricultural and Food Chemistry. 68(12), 3817-3833.

[4] Sánchez-Acevedo, E. (2024). Doctoral Thesis. University of Zaragoza, Zaragoza, Spain.