[Analysis of differences between unifloral honeys from different botanical origins based on non-targeted metabolomics by ultra-high performance liquid chromatography-quadrupole time-of-flight mass spectrometry].

Different nectar plants contain various secondary metabolites. Herein, the differences in the contents of endogenous metabolites in honeys from eight botanical origins (i. e., acacia, jujube, vitex, linden, buckwheat, manuka, wolfberry, and motherwort honeys) were investigated by a non-targeted metabolomics-based method. This method involved solid-phase extraction pretreatment and ultra-high performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UPLC-Q-TOF-MS). An oasis HLB cartridge was used for the removal of many saccharides. Chromatographic experiments were performed on an HSS T3 column (100 mm×2.1 mm, 1.8 μm) using a mobile phase that consisted of 0.1% (v/v) formic acid in acetonitrile and water. Mass spectrometry was conducted in the positive and negative modes by electrospray ionization (ESI). Metabolic information about the honeys from different botanical origins was acquired using a multivariate statistical analysis model. Principal component analysis (PCA) and partial least square discriminant analysis (PLS-DA) were conducted for pattern recognition and difference analysis. PCA was performed for 10557 and 2706 data variables in the positive and negative ion modes, respectively. The distribution of honeys from different botanical origins was investigated in 88 honey samples. The three principal components exhibited 48.05% and 57.88% of the total variance in positive and negative ion modes, respectively. The samples studied were divided into six different groups on the basis of their botanical origins and metabolic compounds: linden, vitex, buckwheat, manuka, jujube, and acacia honeys. A permutation test (=200) was conducted to verify the fit of the model. The differential metabolites were screened on the basis of variable importance in project (VIP; >1), analysis of variance (ANOVA; <0.05), and maximum fold change (>1.5) by using the PLS-DA model. The compounds were identified based on the data retrieved from the Chemspider and HMDB databases according to the quality information of precursor ions and fragment ions. Thirty-two differential metabolites were screened and primarily identified according to the characteristic fragmentation rules of specific structure types and data retrieval, including 18 flavonoids, 7 phenolic acids, 6 phenyl and terpenoid glycosides, and 1 steroid. Various flavonoids in buckwheat and manuka honeys, such as quercetin, sakuranetin, 7-hydroxy-2-(4-hydroxy-3,5-dimethoxyphenyl)-4-chromen-4-one, 5,7-dihydroxy-2-(3-methoxyphenyl)-4-chromen-4-one, luteolin-7-methyl ether, and pollenitin, were found. In buckwheat honey, the contents of 3-methoxy-2-(4-methylbenzoyl)-4-chromen-4-one, 2-hydroxy-3,4-diphenylpentanedioic acid, 3'-methoxydihydroformononetin, phenylpyruvic acid, 2--coumaroyltartronic acid, 2-(3-hydroxy-4,5-dimethoxyphenyl)-4-chromen-4-one, 7-hydroxy-6-methoxy-3-(4-methoxyphenyl)-4-chromen-4-one, 4-[(2)-3-(4-hydroxyphenyl)prop-2-en-1-yl]-3-methoxyphenol, and 7-hydroxy-5-methoxyflavan were the highest; these compounds are the characteristic metabolites of buckwheat honey. In addition, manuka honey possessed the highest contents of gnaphaliin and galangin 3-methyl ether. Moreover, linden honey contained the characteristic phenyl glycosides of ()-multifidol 2-[apiosyl-(1➝6)-glucoside], 2-phenylethyl--D-glucopyranoside, benzyl -[arabinofuranosyl-(1➝6)-glucoside], crosatoside B, and terpenoid glycosides of isopentyl gentiobioside and 6--oleuropeoylsucrose. Vitex honey was found to be rich in quinic acid derivatives such as caffeoyl-3--feruloyl-quinic acid/1-feruloyl-5-caffeoylquinic acid, 3--caffeoyl-4--methyl-quinic acid/3-feruloylquinic acid, and 3--caffeoyl-1--methyl-quinic acid, in addition to the flavonoids of vitexin, namely, 6″-(3-hydroxy-3-methylglutarate) and apigenin-7-[galactosyl-(1➝4)-mannoside]. Moreover, ponasteroside A was a characteristic marker of jujube honey, and the contents of 6--fucosylluteolin and kaempferol 3-(2″-rhamnosylrutinoside) were the highest in acacia honey. In conclusion, the method based on non-targeted metabolomics involving UPLC-Q-TOF-MS for different unifloral honeys was found to be fast, effective, specific, and accurate. The differences in metabolite contents and the characteristic compounds in various unifloral honeys were preliminarily illustrated. This study provides an effective analytical strategy for honey traceability and quality analysis of unifloral honey.