Application of untargeted screening technology based on high-resolution mass spectrometry for the authentication of honey adulteration
基于高分辨质谱的非靶向筛查技术在鉴别蜂蜜掺假中的应用研究

摘要

目的 蜂蜜掺假是当前食品欺诈的突出问题，严重损害消费者权益与市场公平。本研究旨在建立一种基于液相色谱-四极杆飞行时间质谱（LC-Q-TOF-MS）的非靶向代谢组学方法，用于高效、准确地鉴别纯正蜂蜜与掺假蜂蜜。方法 收集来自不同蜜源（槐花、荆条、枣花）的纯正蜂蜜样本，并以常见掺假物（大米糖浆、甜菜糖浆）制备不同掺假比例（10%，20%，30%，50%，w/w）的模拟样本。采用LC-Q-TOF-MS技术获取所有样本的代谢谱图，通过主成分分析（PCA）和正交偏最小二乘-判别分析（OPLS-DA）进行模式识别，筛选差异性标志物，并构建鉴定模型。结果 PCA和OPLS-DA模型能清晰区分纯正蜂蜜与掺假蜂蜜。通过变量重要性投影（VIP>1.5）和t检验（p<0.05），共筛选出15个显著差异性化合物作为潜在化学标志物，包括寡糖、氨基酸、酚酸和黄酮类化合物。其中，大米糖浆标志物（如麦芽糖寡糖特定异构体）和甜菜糖浆标志物（如海藻糖、棉子糖）在掺假样本中显著上调，而某些蜂蜜特征性花粉源植物化合物在掺假样本中显著下调。基于这些标志物建立的判别模型对验证集的鉴别准确率达到98.5%。结论 本研究建立的非靶向筛查方法能够克服传统靶向方法的局限性，无需预先设定目标物即可有效鉴别蜂蜜中是否存在外源糖浆掺假，并提示可能的掺假源，为蜂蜜真实性鉴别提供了强有力的技术支撑。

Abstract

Objective Honey adulteration is a prominent issue of food fraud, severely undermining consumer rights and market fairness. This study aimed to establish an untargeted metabolomics approach based on liquid chromatography-quadrupole time-of-flight mass spectrometry (LC-Q-TOF-MS) for the efficient and accurate discrimination of pure and adulterated honey.
Methods Pure honey samples from different botanical origins (acacia, vitex, jujube) were collected, and simulated adulterated samples were prepared by adding common adulterants (rice syrup, beet syrup) at different ratios (10%, 20%, 30%, 50%, w/w). The metabolic profiles of all samples were acquired using LC-Q-TOF-MS. Pattern recognition was performed using principal component analysis (PCA) and orthogonal partial least squares-discriminant analysis (OPLS-DA) to screen for differential markers and construct identification models.
Results PCA and OPLS-DA models effectively distinguished pure honey from adulterated honey. Fifteen significant differential compounds were screened out as potential chemical markers through variable importance in projection (VIP > 1.5) and t-test (p < 0.05), including oligosaccharides, amino acids, phenolic acids, and flavonoids. Among them, markers for rice syrup (e.g., specific isomers of maltooligosaccharides) and beet syrup (e.g., trehalose, raffinose) were significantly up-regulated in adulterated samples, while some characteristic plant-derived compounds of honey were significantly down-regulated. The discriminant model established based on these markers achieved an accuracy of 98.5% for the validation set.
Conclusion   The untargeted screening method established in this study can overcome the limitations of traditional targeted methods, effectively identifying the presence of exogenous syrup adulteration in honey without pre-defined targets, and suggesting possible adulteration sources. It provides a powerful technical support for honey authenticity identification.