Integrated metabolomics and metagenomics reveal plant-microbe interactions driving aroma differentiation in flue-cured tobacco leaves

Abstract

Current research on tobacco aroma predominantly focuses on single-omics approaches. In this study, we conducted a comprehensive investigation of the relationships between tobacco metabolite profiles, microbial communities, and aroma characteristics. Untargeted metabolomics and metagenomic analyses were performed on flue-cured upper tobacco leaves to compare light aromatic tobacco (LAT) and strong aromatic tobacco (SAT). The results showed that sugar metabolite levels in LAT were significantly higher than those in SAT, whereas levels of specific acids and amino acid metabolites in SAT exceeded those in LAT. Redundancy analysis (RDA) and metabolomic correlation analyses indicated that the genera Methylorubrum and Pseudomonas may promote sugar metabolite accumulation, while Pseudokineococcus potentially regulates both sugar and acid metabolites. In contrast, Methylobacterium and Sphingomonas were associated with acid and amino acid metabolism, with Methylobacterium additionally exhibiting inhibitory effects on sugar metabolism. Metagenomic analysis revealed that Methylorubrum, Pseudomonas, and Pseudokineococcus were abundant in LAT, whereas Methylobacterium and Sphingomonas dominated in SAT. Notably, the bidirectional regulation of aromatic metabolites by microbial genera such as Pseudokineococcus highlights the universality of plant-microbe interactions in shaping metabolic networks-a mechanism potentially applicable to other crop systems. These findings reveal conserved microbial functional traits (e.g., metabolic pathway modulation) that may drive plant phenotypic differentiation beyond tobacco, offering insights into microbiome-mediated crop quality improvement. The results provide theoretical guidance for tobacco aging and aroma regulation and underscore the broader significance of microbial community engineering in agriculture for manipulating plant metabolic outputs.

Keywords: aroma; flue-cured tobacco; metagenomics; plant-microbe interaction; untargeted metabolomics.

Figure 1

Untargeted metabolomics data of tobacco from GC-MS. (A) Metabolic molecule (Bar plot) and their proportion (Pie plot). (B) The relative contents of metabolites in different aromatic flue-cured tobacco. *P<0.05, ns, no significance. SAT, strong aromatic tobacco; LAT, light aromatic tobacco.

Figure 2

Multivariate analysis of the aromatic flue-cured tobacco leaf metabolites. (A) PLS-DA score plot, ellipses represent 95% confidence intervals for each group. (B) Dot plot of metabolites with variable importance in projection (VIP) scores >1.0 in the PLS-DA model. (C) Metabolite composition, bar plot showing the number of the metabolite molecules, pie plot displaying the proportion of metabolites. (D) Heat map of differential metabolites relative content in tobacco leaves with different flavors (Euclidean distance). SAT, strong aromatic tobacco; LAT, light aromatic tobacco.

Figure 3

Metabolome analysis of differential metabolites in light and strong aromatic tobacco leaves. (A) Overview of metabolic pathways. (B) Metabolic pathway map of the differential metabolites. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. In (A), the bubbles represent the metabolic pathways, and the color of the bubbles (from yellow to red) indicates the significant level of the metabolites in the data.

Figure 4

Metabolites and microorganisms of the flue-cured tobacco leaves. (A) RDA analysis, the angle between two variables less than 90° indicates a positive correlation, otherwise a negative correlation. (B) Correlation analysis between microorganisms at the genus level and metabolites. *P<0.05, **P<0.001.