1H NMR Metabolic Profile to Discriminate Pasture Based Alpine Asiago PDO Cheeses

Simple Summary Nowadays, alpine cheese from grazing dairy herds has a premium market value because consumers perceive its higher degree of healthiness and sustainability. The authenticity of pasture-based cheese should be safeguarded from local hay-based milk analogues. The study aimed at assessing the reliability of proton nuclear magnetic resonance (1H NMR) to discriminate pasture-based alpine Asiago PDO cheeses of different ripening time from similar hay-based samples processed in the same dairy plant. Cheeses were produced from raw milk collected from grazing or hay-fed alpine dairy herds and they were ripened for 2 (Pressato), 4 (Allevo_4), and 6 (Allevo_6) months. Samples of the cheeses were submitted to wet chemistry and nuclear magnetic resonance analysis. The outcomes of the 1H NMR spectroscopy were used in a multivariate discriminant procedure. Choline, 2,3-butanediol, lysine, and tyrosine and some residual sugar-like compounds were water-soluble biomarkers of cows’ feeding system. However, the application of 1H NMR based metabolomics was an effective fingerprinting method to correctly identify only cheese samples with the shortest ripening period. The classification of more aged cheese samples according to the cows’ feeding system was less accurate likely due to the chemical and biochemical changes induced by a prolonged maturation process. Abstract The study was carried out in an alpine area of North-Eastern Italy to assess the reliability of proton nuclear magnetic resonance 1H NMR to fingerprint and discriminate Asiago PDO cheeses processed in the same dairy plant from upland pasture-based milk or from upland hay-based milk. Six experimental types of Asiago cheese were made from raw milk considering 2 cows’ feeding systems (pasture- vs. hay-based milk) and 3 ripening times (2 months, Pressato vs. 4 months, Allevo_4 vs. 6 months, Allevo_6). Samples (n = 55) were submitted to chemical analysis and to 1H NMR coupled with multivariate canonical discriminant analysis. Choline, 2,3-butanediol, lysine, tyrosine, and some signals of sugar-like compounds were suggested as the main water-soluble metabolites useful to discriminate cheese according to cows’ feeding system. A wider pool of polar biomarkers explained the variation due to ripening time. The validation procedure based on a predictive set suggested that 1H NMR based metabolomics was an effective fingerprinting tool to identify pasture-based cheese samples with the shortest ripening period (Pressato). The classification to the actual feeding system of more aged cheese samples was less accurate likely due to their chemical and biochemical changes induced by a prolonged maturation process.