The bacterial succession and its role in flavor compounds formation during the fermentation of cigar tobacco leaves

Abstract

Fermentation is the key process required for developing the characteristic properties of cigar tobacco leaves, complex microorganisms are involved in this process. However, the microbial fermentation mechanisms during the fermentation process have not been well-characterized. This study investigated the dynamic changes in conventional chemical composition, flavor compounds, and bacterial community during the fermentation of cigar tobacco leaves from Hainan and Sichuan provinces in China, as well as the potential roles of bacteria. Fermentation resulted in a reduction of conventional chemical components in tobacco leaves, with the exception of a noteworthy increase in insoluble protein content. Furthermore, the levels of 10 organic acids and 19 amino acids showed a significant decrease, whereas the concentration of 30 aromatic substances exhibited a unimodal trend. Before fermentation, the bacterial community structures and dominant bacteria in Hainan and Sichuan tobacco leaves differed significantly. As fermentation progressed, the community structures in the two regions became relatively similar, with Delftia, Ochrobactrum, Rhodococcus, and Stenotrophomonas being dominant. Furthermore, a total of 12 functional bacterial genera were identified in Hainan and Sichuan tobacco leaves using bidirectional orthogonal partial least squares (O2PLS) analysis. Delftia, Ochrobactrum, and Rhodococcus demonstrated a significant negative correlation with oleic acid and linoleic acid, while Stenotrophomonas and Delftia showed a significant negative correlation with undesirable amino acids, such as Ala and Glu. In addition, Bacillus showed a positive correlation with benzaldehyde, while Kocuria displayed a positive correlation with 2-acetylfuran, isophorone, 2, 6-nonadienal, and β-damascenone. The co-occurrence network analysis of microorganisms revealed a prevalence of positive correlations within the bacterial network, with non-abundant bacteria potentially contributing to the stabilization of the bacterial community. These findings can improve the overall tobacco quality and provide a novel perspective on the utilization of microorganisms in the fermentation of cigar tobacco leaves.

Keywords: Bacterial community; Cigar tobacco leaves; Co-occurrence network; Fermentation; Flavor substances.

Fig. 1

Comparison of the dynamic conventional chemical compounds during the fermentation. A Score scatter plot for PCA model; B HCA plots based on PCA modeling

Fig. 2

Heatmap analysis of 10 organic acids (A), 19 amino acids (B), and 30 aromatic compounds (C) during the fermentation. Legends show the log-transformed scores of their content. A color-coded scale grading from blue to red corresponds to the content of flavor substances from low to high

Fig. 3

Analysis of bacterial community during the fermentation process. A Bar plot analysis of the bacterial community at the phylum level. B Bar plot analysis of the bacterial community at the genus level. C Venn diagram depicting shared and unique genera between Hainan and Sichuan samples. D Pieplot analysis of shared genera between Hainan and Sichuan samples. LEfSe analysis of Hainan E and Sichuan F bacterial communities that differed between pre- and post-fermentation stages based on LDA scores

Fig. 4

Analysis of the differences in bacterial succession during the fermentation process. A Non-metric multidimensional scaling analysis of samples at different fermentation stages. B Analysis of the distance between inter- and intra-groups at the genus level in different samples

Fig. 5

Correlation network of Hainan A and Sichuan B functional bacteria and flavor compounds during the fermentation. The green circles represent the functional bacteria, and the yellow squares represent the flavor compounds. The red lines indicate positive correlations, while the blue lines represent negative correlations

Fig. 6

Co-occurrence network analysis for bacterial community during the fermentation process. The network of bacterial community for Hainan (A) and Sichuan (B) samples. Within or among module connectivity of each node for Hainan (C) and Sichuan (D) samples. The co-occurrence networks are colored by module. The size of each node is proportion to the number of connections (that is, degree). Red and green lines indicate positive and negative correlations, respectively