NMR Quantification of Carbohydrates in Complex Mixtures. A Challenge on Honey.

The knowledge of carbohydrate composition is greatly important to determine the properties of natural matrices such as foodstuff and food ingredients. However, because of the structural similarity and the multiple isomeric forms of carbohydrates in solution, their analysis is often a complex task. Here we propose an NMR analytical procedure based on highly selective chemical shift filters followed by TOCSY, which allows us to acquire specific background-free signals for each sugar. The method was tested on raw honey samples dissolved in water with no other pretreatment. In total, 22 sugars typically found in honey were quantified: 4 monosaccharides (glucose, fructose, mannose, rhamnose), 11 disaccharides (sucrose, trehalose, turanose, maltose, maltulose, palatinose, melibiose and melezitose, isomaltose, gentiobiose nigerose, and kojibiose), and 7 trisaccharides (raffinose, isomaltotriose, erlose, melezitose, maltotriose, panose, and 1-kestose). Satisfactory results in terms of limit of quantification (0.03-0.4 g/100g honey), precision (% RSD: 0.99-4.03), trueness (bias % 0.4-4.2), and recovery (97-104%) were obtained. An accurate control of the instrumental temperature and of the sample pH endows an optimal chemical shift reproducibility, making the procedure amenable to automation and suitable to routine analysis. While validated on honey, which is one of the most complex natural matrices in terms of saccharides composition, this innovative approach can be easily transferred to other natural matrices.