Microbial communities of a variety of 75 homemade fermented vegetables

Abstract

Fermentation is an ancient practice of food preservation. Fermented vegetables are popular in Eastern European and Asian countries. They have received a growing interest in Western countries, where they are mainly manufactured at domestic and artisanal scales and poorly characterized. Our aim was to investigate the microbial communities and the safety of French homemade fermented vegetables, in the frame of a citizen science project. Fermented vegetables and the data associated with their manufacture were collected from citizens and characterized for pH, NaCl concentration, and microbiology by culturomics and 16S DNA metabarcoding analysis. Lactic acid bacteria (LAB) and yeast isolates were identified by 16S rRNA gene sequencing and D1/D2 domains of the large subunit of the rRNA gene, respectively. The 75 collected samples contained 23 types of vegetables, mainly cabbage, followed by carrots and beets, and many mixtures of vegetables. They were 2 weeks to 4 years old, and their median pH was 3.56, except for two samples with a pH over 4.5. LAB represented the dominant viable bacteria. LAB concentrations ranged from non-detectable values to 8.7 log colony-forming units (CFU)/g and only depended on the age of the samples, with the highest most frequently observed in the youngest samples (<100 days). The 93 LAB isolates identified belonged to 23 species, the two mains being Lactiplantibacillus pentosus/plantarum and Levilactobacillus brevis. The other microbial groups enumerated (total aerobic bacteria, halotolerant bacteria, Gram-negative bacteria, and acetic acid bacteria) generally showed lower concentrations compared to LAB concentrations. No pathogenic bacteria were detected. Viable yeasts were observed in nearly half the samples, at concentrations reaching up to 8.0 log CFU/g. The 33 yeast clones identified belonged to 16 species. Bacterial metabarcoding showed two main orders, namely, Lactobacillales (i.e., LAB, 79% of abundance, 177 of the 398 total ASVs) and Enterobacterales (19% of abundance, 191 ASVs). Fifteen LAB genera were identified, with Lactiplantibacillus and Levilactobacillus as the most abundant, with 41 and 12% of total reads, respectively. Enterobacterales members were mainly represented by Enterobacteriaceae and Yersiniaceae. This study is the first wide description of the microbiota of a large variety of homemade fermented vegetables and documents their safety.

Keywords: citizen science; fermented food; lactic acid bacteria; sauerkraut; spontaneous fermentation; yeast.

Figure 1

Vegetables used to manufacture the homemade fermented vegetables characterized in this study. (A) Word cloud scheme built from the list of all vegetables used in the 75 recipes associated with the samples received. (B) Pictures illustrating the different sizes of slicing, either thinly cut (grated, shredded, two upper lines) or roughly cut (into slices, dices, or simply cut in two lengthwise, bottom line).

Figure 2

Boxplots of pH values, NaCl concentrations expressed in g/L, lactic acid bacteria (LAB) and yeast concentrations, expressed in log CFU/g, for 75 homemade fermented vegetables, according to (upper figures) the type of vegetable, used either alone: beetroot (n = 7), cabbage (different types, n = 10), carrot (n = 8), others (n = 21), or used in mixture: Mix_veg (n = 29); (bottom figures) the category of vegetable: bulb/root (n = 26), leaf/stem (n = 13), or fruit (n = 13), or mixed (n = 23).

Figure 3

Principal component analysis performed for the 75 homemade or artisanal fermented vegetable samples collected. (A) Factorial map: individuals are colored according to a hierarchical clustering made on PCA data. (B) Graph of variables: active variables are microbial plate counts enumerated on different media (see Supplementary Table S2), shown in green, and in black for enumeration after a 70°C and 15 min treatment, and pH and measured NaCl concentration, shown in red. Four variables that describe samples are projected as supplementary variables, shown in blue italics and blue dotted lines: the age of the samples, their expected NaCl content calculated from declaration in the corresponding recipes, the number of ingredients in the recipes, and the “diversity score” corresponding to the number of practices and factors that may influence the initial microbial community (see text).

Figure 4

Proportions of the 93 lactic acid bacteria isolates collected on the MRS plates from 59 fermented vegetable samples and identified by 16S rRNA gene sequencing.

Figure 5

Relative abundances of LAB and some bacterial orders in the 47 homemade fermented vegetables successfully analyzed by metabarcoding targeting the 16S RNA gene. Samples are sorted by order of (i) decreasing proportion of Lactobacillales order, shown in different shades of blue, green, and purple according to the genus, and (ii) increasing proportion of Enterobacterales order.

Figure 6

Multiple factor analysis (MFA) performed for 45 samples with metabarcoding results. The variables included as active variables were compositional variables (pH and NaCl content, depicted in red) and the six main taxa from metabarcoding analysis (depicted in green); quantitative variables were used as supplementary variable: descriptors of samples (age, number of ingredients, “diversity score,” depicted in pink) and plate counting results on six media, depicted in blue. Samples are colored according to a hierarchical clustering analysis made from MFA data. (A) Dim 1-Dim 2 plots; (B) Dim 3-Dim 4 plots.