Microbial Diversity and Interaction Specificity in Kombucha Tea Fermentations

Abstract

Despite the popularity of kombucha tea, the distribution of different microbes across kombucha ferments and how those microbes interact within communities are not well characterized. Using metagenomics, comparative genomics, synthetic community experiments, and metabolomics, we determined the taxonomic, ecological, and functional diversity of 23 distinct kombuchas from across the United States. Shotgun metagenomic sequencing demonstrated that the bacterium Komagataeibacter rhaeticus and the yeast Brettanomyces bruxellensis were the most common microbes in the sampled kombucha communities. To determine the specificity of bacterium-yeast interactions, we experimentally quantified microbial interactions within kombucha biofilms by measuring densities of interacting species and biofilm production. In pairwise combinations of bacteria and yeast, B. bruxellensis and individual strains of Komagataeibacter spp. were sufficient to form kombucha fermentations with robust biofilms, but Zygosaccharomyces bisporus, another yeast found in kombucha, did not stimulate bacteria to produce biofilms. Profiling the spent media of both yeast species using nuclear magnetic resonance spectroscopy suggested that the enhanced ability of B. bruxellensis to ferment and produce key metabolites in sucrose-sweetened tea may explain why it stimulates biofilm formation. Comparative genomics demonstrated that Komagataeibacter spp. with >99% genomic similarity can still have dramatic differences in biofilm production, with strong producers yielding five times more biofilm than the weakest producers. IMPORTANCE Through an integration of metagenomic and experimental approaches, our work reveals the diversity and nature of interactions among key taxa in kombucha microbiomes through the construction of synthetic microbial pairs. Manipulation of these microbes in kombucha has the potential to shape both the fermentation qualities of kombucha and the production of biofilms and is valuable for kombucha beverage producers, biofilm engineers, and synthetic ecologists.

Keywords: Brettanomyces; Komagataeibacter; acetic acid bacteria; fermentation; kombucha; microbiome; yeast.

FIG 1

Taxonomic diversity of 23 kombucha microbiomes. (A) Kombucha fermentation, showing the floating biofilm and liquid tea (50). Samples of kombucha starter from across the United States, containing both biofilm and liquid, were acquired through the website Etsy.com or through donations, and the liquid portions were sequenced using shotgun metagenomic sequencing. Each black dot on the map represents the geographic origin of kombuchas sampled in this study. (B) Order-level taxonomies were assigned to unassembled reads using Kaiju and the NCBI BLAST nr+euk database, which contains bacteria, yeasts, viruses, and microbial eukaryotes. The orders Saccharomycetales (yeast), Rhodospirillales (AAB), and Lactobacillales (LAB) were the most frequent and abundant of classified reads. (C) Genus-level taxonomy assignments were obtained in the same manner. The genus Komagataeibacter was the most abundant AAB, Enterococcus the most abundant LAB, and Brettanomyces the most abundant yeast detected at the genus level. (D) At the species level, Komagataeibacter rhaeticus was the most common and abundant bacterium detected. Brettanomyces bruxellensis was the most common yeast detected. See Tables S1 to S3 for taxon tables.

FIG 2

Genus-level differences in yeast, as well as strain-level differences in bacteria, have consequences for kombucha biofilm formation. (A) Experimental design of synthetic pairings of yeast and bacteria. Different color shades of the same yeast or bacterial species represent distinct strains of the species. (B) Biofilm formation measured as wet weight (grams) in synthetic pairs. White indicates no biofilm formation. Select pairs that formed biofilms are linked to images of those biofilms. For images of all biofilms produced, including all replicates, see Fig. S1. (C and D) Log₁₀ CFU of yeast and bacteria, respectively, in the liquid portion of synthetic pair fermentations. (E) pH of synthetic pairings at the end of 21 days of incubation. In panels B to E, n = 6 replicate fermentations; dots represent the means, and error bars represent the standard deviations of the means. CFU counts from 1 of the 60 experimental treatments—the pair NH2Y1/LC2B1—were not collected due to a plating error during the experimental harvest (indicated by an asterisk [*] in panels C and D).

FIG 3

Growth and metabolite profiles of the kombucha yeasts Z. bisporus and Brettanomyces bruxellensis. (A) OD₆₀₀ of yeast strains used in synthetic pair experiments growing in 10% (wt/vol) sucrose green tea. Lines show mean values, and error bars represent the standard deviations of the means of seven replicates. (B) NMR analysis of filtered spent media of Z. bisporus and B. bruxellensis isolates (a subset of those that were used in synthetic pairs) after incubation in 10% sucrose green tea for 10 and 21 days. The concentration of each compound is scaled from zero to the maximum level observed across all measurements (10 and 21 days). A portion of spent medium was also inoculated with K. rhaeticus strain M2B4, and biofilms were weighed. Bacterial biofilm wet weight (g) is displayed as vertical bars, where error bars represent the standard deviations of the means.

FIG 4

Pangenomic comparison of K. rhaeticus isolates. (A) Pangenome of K. rhaeticus isolates from separate kombucha samples, which were used to make defined cocultures of bacteria and yeast. Each layer represents an isolate genome, and each item radiating from the center of the circle is a gene cluster. The presence of a gene cluster in an isolate is indicated by a darker shade of color. On the top right, a dedrogram illustrates similarity of genomes based on hierarchical clustering of the gene cluster presence/absence matrix of all genes from all the genomes. The isolate that forms insubstantial biofilms is highlighted in light pink. Isolates that formed substantial biofilms are in dark pink. Reference genomes are in dark gray. The grayscale heatmap shows the ANI between isolates. The red/orange heatmap originates from Fig. 2 and shows biofilm formation when these bacteria are paired with yeast. Genes annotated as cellulose synthase are labeled in blue.