Towards the quantification of residual milk allergens in caseinate-fined white wines using HPLC coupled with single-stage Orbitrap mass spectrometry.

A method based on LC-ESI-high-resolution (HR)-MS analysis, using a single-stage Orbitrap mass spectrometer, was developed for the quantification of casein allergens potentially present in white wines as a result of fining by caseinate. The method consists of (1) extraction from the matrix by ultrafiltration, (2) digestion with trypsin and (3) detection/quantification of residual caseins, obtained by monitoring the LC-MS response of representative tryptic peptides (peak areas in extracted-ion chromatograms). Method linearity was assessed first on caseinate solutions prepared either in water or in wine matrix (the ultrafiltration residue of a protein-free white wine). Limits of detection (LOD) ranged from 0.1 to 0.3 mg ml1 (S/N ¼ 3) in water, and between 0.15 and 0.7 mg ml1 in wine matrix, depending on the selected peptide. Method repeatability and reproducibility, measured as response variability (standard deviation) due to LC-MS analysis alone and to both enzymatic digestion and LC-MS analysis, were assessed on caseinate standard solutions in water and ranged from 5 to 12% and from 8 to 20%, respectively. A higher variability was usually observed for the peptide marker response in the case of matrixmatched samples, the only exception being peptide GPFPIIV from -casein, the marker also providing the highest sensitivity. The method was finally applied to a casein-free white wine ('Greco di Tufo') fined with caseinate at different concentrations, after discarding the precipitate due to casein-wine components aggregation. Minimum detectable added caseinate concentrations (i.e. those corresponding to responses with S/N ¼ 3) were estimated between 39 and 51 mg ml1 , according to the peptide marker chosen. These limits are compatible with caseinate concentrations typically adopted for wine-fining purposes. Moreover, a cross-check with the calibration performed in wine matrix led to an estimation of the concentration of dissolved caseinate to be in the low ng ml1 range.