Hierarchical Effects of Lactic Fermentation and Grain Germination on the Microbial and Metabolomic Profile of Rye Doughs

Abstract

A multi-omics approach was adopted to investigate the impact of lactic acid fermentation and seed germination on the composition and physicochemical properties of rye doughs. Doughs were prepared with either native or germinated rye flour and fermented with Saccharomyces cerevisiae, combined or not with a sourdough starter including Limosilactobacillus fermentum, Weissella confusa and Weissella cibaria. LAB fermentation significantly increased total titrable acidity and dough rise regardless of the flour used. Targeted metagenomics revealed a strong impact of germination on the bacterial community profile of sprouted rye flour. Doughs made with germinated rye displayed higher levels of Latilactobacillus curvatus, while native rye doughs were associated with higher proportions of Lactoplantibacillus plantarum. The oligosaccharide profile of rye doughs indicated a lower carbohydrate content in native doughs as compared to the sprouted counterparts. Mixed fermentation promoted a consistent decrease in both monosaccharides and low-polymerization degree (PD)-oligosaccharides, but not in high-PD carbohydrates. Untargeted metabolomic analysis showed that native and germinated rye doughs differed in the relative abundance of phenolic compounds, terpenoids, and phospholipids. Sourdough fermentation promoted the accumulation of terpenoids, phenolic compounds and proteinogenic and non-proteinogenic amino acids. Present findings offer an integrated perspective on rye dough as a multi-constituent system and on cereal-sourced bioactive compounds potentially affecting the functional properties of derived food products.

Keywords: LAB; dough; fermentation; germination; metabolomics; metagenomics; rye; yeast.

Figure 1

Beta-diversity analysis of rye dough and flour samples. PCoA plot based on unweighted UniFrac distances of microbial communities among all samples. Solid symbols represent doughs made with native rye flour (squares) and sprouted rye flour (circles); empty symbols represent native rye flour (squares) and sprouted rye flour (circles). Dough samples were colored according to the starter used for fermentation: blue = LAB + SC, green = SC, pink = no inoculated starter.

Figure 2

Relative abundance of bacterial species found in dough and flour samples. The bar plot represents the top 15 most abundant taxa among all samples identified to the species level. “Others” refers to merged species that individually showed a relative abundance below 0.25%. “Unclassified” are bacterial taxa not identified at the phylum level. The results are expressed as average of replicates for each type of dough. In the x-axis are reported the name of the samples as follow: sD, sprouted rye dough; nD, native rye dough; s_F, sprouted rye flour; n_F, native rye flour, CTR, dough control; LAB + SC, dough fermented with LAB + SC; SC, dough fermented with SC.

Figure 3

Differential abundance analysis of bacterial species among rye doughs. Box and whiskers plot indicate the proportion of differentially abundant taxa between starters used for fermentation in native rye doughs (A) and in sprouted rye doughs (B), and between native and sprouted rye doughs (C). Only significant species (FDR < 0.05) detected by DESeq2 are reported in the figure. Boxes represent the minimum and maximum of abundance values of replicates for each condition; the line in the box is the median.

Figure 4

HPAEC-PAD carbohydrate content of rye sourdoughs. Carbohydrate content of native rye doughs (A). Carbohydrate content of sprouted rye doughs (B). Results are expressed as carbohydrate content (mg g-1) referred as xylose, arabinotriose, and arabinoctaose equivalents for monosaccharides, low-polymerization degree oligosaccharides (low-PD), and high-polymerization degree oligosaccharides (high-PD) contents, respectively (B). Vertical bars indicate standard deviation (n = 6). Different letters indicate statistically significant differences (p < 0.05) among treatments.

Figure 5

Unsupervised hierarchical cluster analysis on the untargeted metabolic profiling of rye dough. The fold change values, represented by a color range, for each compound were calculated with respect to the median of all samples, and further used to obtain a fold change-based heatmap, according to Ward’s algorithm (Euclidean distance). The factors involved in clustering were NATIVE and SPROUTED, for native and sprouted-derived rye sourdough, respectively; and control, SC, and LAB + SC, for unfermented, yeast-fermented and yeast and LAB-combined fermented rye sourdough, respectively.

Figure 6

Orthogonal projection to latent structures discriminant analysis and variable importance in projection analysis of rye doughs. The OPLS models were combined with the proportion of VIP markers on the discrimination of metabolomic effects caused by germination conditions (A,B), fermentation starters on native rye doughs (C,D), fermentation starters on sprouted rye doughs (E,F).