Abstract
The transition toward sustainable, plant-based diets requires innovations in fermentation to enhance the nutritional, functional, and sensory quality of pulse-based matrices. This PhD research investigated the functional potential and metabolic adaptability of lactic acid bacteria (LAB) and yeasts in pulse-based systems, focusing on faba bean and yellow pea as sustainable protein sources for healthy food fermentations. A co-designed citizen science (CS) initiative was first launched within the HealthFerm project to collect sourdough fermentation data and samples across Europe. Over 1000 participants from 33 countries registered, and 671 samples were collected, enabling large-scale analysis of fermentation practices, motivations, and sourdough characteristics. This effort also led to the establishment of a microbial biobank from the collected sourdough samples. Building on these findings, optimized binary LAB–yeast consortia were developed for sourdough fermentation of various pulse-based matrices. From 288 type II sourdoughs screened, seventeen best performers were further propagated via backslopping to type IV sourdoughs exhibiting greater maturity and functionality. Six of the most promising pulse-based type IV sourdoughs were used as tailored inocula for breadmaking, leading to pulse-based sourdough breads with significantly improved protein content, amino acid profiles, and phenolic compounds enrichment, like catechin, rutin, epicatechin, sinapic acid, and quercetin. Co-culturing Leuc. mesenteroides and Saccharomyces cerevisiae in yellow pea notably reduced anti-nutrient levels. Despite some textural limitations, most sourdough breads exhibited more complex volatile organic compound profiles compared to wholewheat control breads leavened with baker´s yeasts. In a parallel study, novel fermented cereal-pulse-based beverages were developed using tailored binary, ternary, and quaternary LAB consortia. Most formulations achieved dairy-like consistency, enhanced antioxidant activity, and distinct nutritional advantages. Among the consortia, the binary composed of Lacticaseibacillus paracasei and Leuconostoc citreum or of Lc. paracasei and Levilactobacillus brevis effectively reduced anti-nutrients and improved protein digestibility. Starter–substrate interactions strongly influenced both microbial metabolite production and consumer preference. Complementing the screening approach and deepening the understanding of microbial adaptation during pulse-based fermentation, the high performing LAB (Leuconostoc mesenteroides GSL1) was characterized following three sequential 2 propagation cycle in faba bean- and yellow pea-based model systems, with de Man Rogosa and Sharpe (MRS) medium used as the control. Substrate-dependent shifts were observed in carbon source utilization patterns and the expression of genes linked to carbohydrate metabolism. Although the strain GSL1 consistently metabolized sucrose, maltose, glucose, and fructose, phenotype microarray (OmniLog) analysis indicated an expanded carbon utilization profile in pulse-propagated cultures, including N-acetyl D-glucosamine, galactose, and mannose. Genomic and gene expression analyses supported this phenotypic adaptation, identifying mutations and upregulation of genes associated with carbohydrate transport and metabolism. Future integration of other omics approaches, such as metabolomics, will facilitate predictive modelling of metabolic adaptation and guide the rational design of tailored pulse based fermentations. Collectively, this PhD thesis demonstrates the feasibility and potential of using tailored microbial strains and consortia, guided by multi-omics insights, for the design of nutritious, sustainable, and functional food products from pulse-based matrices.