LC-HRMS data analysis of natural polymer homologue series Application on wine neutral oligosaccharides

Abstract

Although oligosaccharides have much impact both on health (prevention of diabetes, cardiovascular disease), and on the perception of wine (sweetness, astringency, acidity or bitterness), information on their composition in wine is still limited. In a previous work, neutral oligosaccharide fractions isolated from wine were analyzed. The results present a composition of different monosaccharide units (hexose, pentose, uronic acid and deoxyhexose) and show the presence of several structures of oligo-rhamnogalacturonan type I substituted through the rhamnose moieties by arabinan and/or galactan chains.
The aim of this work is to explore new approaches for processing LC-HRMS data to identify these compounds containing repeating units (homologous series) such as arabinans or galactans.  The presented approach allows visualization of these series in the form of a Kendrick mass defect (KMD) plot to facilitate their characterization.
The chromatographic profiles obtained by LC-HRMS analysis of these fractions showed a poorly resolved bump, and the mass spectra were very complex consisting of mono, di, and tricharged ions peaks over a mass range between 500 and 2500. They allowed however to visualize numerous series formed by separated monocharged peaks of 132 m/z, or dicharged peaks of 66 m/z, i.e. a pentose unit.
The construction of the KMD plot is done with a change of scale for which the mass taken into account of (C5H8O4) is 132.0000 (nominal mass) instead of 132.0423 Da (exact mass). All masses of the spectrum are thus recalculated and called Kendrick masses (KM). The mass defect (KMD) for each peak of the spectrum is the difference between its Kendrick mass and its nominal mass. All compounds of the form R-(C5H8O4)n, R being a common radical, will have the same mass defect. The graphical representation, thus makes it possible to visualize the set of compounds that differ only by their number of pentose units on the same line.
In our case, the accuracy of the measurement at masses above m/z 1000 does not allow this calculation. This limitation was resolved by processing the data with the Compound DiscovererTM software (ThermoScientific) to obtain a list of monocharged masses, for which several crude formulas were proposed. A sorting of these crude formulas was carried out considering the possible ratios between number of carbons, oxygens and hydrogens. The exact masses of the 2045 remaining formulas were then calculated and allowed to draw the Kendrick mass defect plot.
Finally, the Kendrick diagram approach allows visualization of the homologous series of arabinoses. Identification hypotheses were proposed for 555 compounds attributed to oligo-rhamnogalacturonan type I, and its arabinans/galactans side chains degradation products. This study demonstrated the relevance of this analytical approach for the determination of the structure of wine oligosaccharides.