Comparative profiling of volatile and non-volatile compounds in three representative flue-cured tobacco samples using HS-GC-IMS and UHPLC-Q-Exactive/MS

Flue-cured tobacco samples assigned to fresh-flavor (FF), middle-flavor (MF), and strong-flavor (SF) categories show differences in sensory characteristics and chemical composition. However, the metabolite patterns associated with these representative sample groups remain insufficiently characterized. In this exploratory study, we compared the volatile and non-volatile metabolite profiles of three region–cultivar samples representing different flavor-style categories. HS-GC-IMS identified 70 volatile compounds, including ketones, aldehydes, and alcohols, from the three sample groups. OPLS-DA identified 29 volatile compounds that differed in relative abundance among the groups. n-Pentanol, benzaldehyde, and benzene acetaldehyde showed higher relative abundances in the FF-representing sample group, whereas 2-butanone, octanal, and diethyl disulfide showed higher relative abundances in the MF-representing group. UHPLC-Q-Exactive/MS identified 134 non-volatile metabolites belonging to 11 chemical categories, including 41 differential metabolites. Organic acids such as quinic acid, malic acid, and malonic acid showed higher levels in the FF-representing group, whereas shikimic acid, pimelic acid, decanedioic acid, and succinic acid showed higher levels in the SF-representing group. Nicotine and cotinine were relatively more abundant in the MF-representing group. KEGG analysis indicated that amino acid biosynthesis, alkaloid biosynthesis, taurine and hypotaurine metabolism, tryptophan metabolism, and phenylpropanoid biosynthesis were associated with the differences observed among the three sample groups. The combined HS-GC-IMS and UHPLC-Q-Exactive/MS approach revealed distinct volatile and non-volatile metabolite profiles among three region–cultivar samples representing different flue-cured tobacco flavor-style categories. These metabolites and metabolic pathways may serve as candidate chemical features associated with the observed differences. However, because flavor-style category was confounded with geographical origin and cultivar in the present design, the findings should be considered exploratory and should not be interpreted as universal flavor-style-specific markers. These findings provide candidate metabolites and metabolic pathways for further investigation of quality differences among flue-cured tobacco samples representing different flavor categories.