LC–MS untargeted metabolomics reveals early biomarkers to predict browning of fresh-cut lettuce

Abstract

A complete untargeted metabolomics study was developed to identify biomarkers related to the browning of fresh-cut lettuce which is the main cause of quality loss. For this purpose, UPLC-MS-QTOF analysis was optimized to explore the metabolome of 30 selected cultivars of romaine lettuce with different browning susceptibility harvested at three different harvest dates. Different multivariate analyses and statistics software, such as Agilent Mass Profiler Professional (MPP), SIMCA and The Unscrambler, were used for the selection of entities correlated with browning induced after cutting and storage. A group of metabolites that were identified through the analysis of different databases and comparison with authentic standards when available, highly correlated with browning measured by image analysis measuring Hue angle difference between day 0 and day 5 of storage at 7 °C. A Multiple Linear Regression (MLR) model combined entities matrix and browning. At day 0 the metabolites that correlated positively (P ≤ 0.01) with browning development at day 5 were caffeoylquinic acid and 3-hydroxy-tetradecadienoic acid while ferulic acid methyl ester and 2-O-p-hydroxyphenyl-6-O-galloyl glucose correlated negatively (P ≤ 0.01). This study also confirmed the involvement of different types of metabolites (phenolic compounds, lipids and, terpenes) in the development of browning. A ratio ferulic acid methyl ester/caffeoylquinic acid at time 0 was able to predict browning after 5 days of storage in 70% of the cases.