Comparative evaluation of LAB proteolytic activities in soy-based dairy alternatives and bovine milk for selecting plant-based food starter strains

The green transition in food production has intensified the search for novel starter cultures to produce plant-based dairy alternatives (PBDAs) that match dairy products in quality, taste, and texture. Traditional dairy starters have shown clear limitations in these matrices, particularly in nitrogen metabolism and proteolysis. This study aimed to screen and characterize lactic acid bacteria (LAB) with proteolytic traits better suited for plant-based food fermentations. Diverse LAB strains from different genera were screened in a soy-based drink (SBD), identifying 14 different clades of cell envelope proteinases (CEPs). Representative strains were then comparatively profiled in bovine milk and SBD, revealing that Clade I proteases have three distinct phenotypes, with proteases that are active predominantly in milk, predominantly in SBD, or in both matrices. Sequence and three-dimensional structure analyses identified amino acid differences in the Fn2 domain that may modulate substrate specificity. Soy-specific proteases were observed to have a tandem deletion in the pro-peptide and mutation at H763 in the active enzyme. Additionally, proteomics analysis revealed that these proteases target the disordered regions of soy proteins as the primary sources of peptides. The soy proteins dehydrin (K7LEQ5) and a late embryogenesis abundant protein (I1L957), which have highly disordered structures similar to caseins, were found to be susceptible to these proteases. Comparative peptide profiling showed that Lactococcus proteinases generated highly varied peptide sequences with consistent peptide lengths, whereas prtS gene-coded proteinases produced more uniform peptide sets, which highlights key functional difference between these two groups of traditional dairy starters. Overall, our findings indicate that Clade I proteases can be beneficial for hybrid dairy fermentation and plant-based fermentation, while Clade IV are highly conserved and ubiquitous in plant-adapted Leuconostoc spp. Although further investigation of these proteases is needed, these findings help fill a gap in comprehensive studies of the proteolytic capacity across different food matrices and provide detailed criteria for selecting suitable starter strains for plant-based foods.