Sensoproteomic Characterization of Lactobacillus Johnsonii-Fermented Pea Protein-Based Beverage: A Promising Strategy for Enhancing Umami and Kokumi Sensations while Mitigating Bitterness

This study investigated the mechanism underlying the flavor improvement observed during fermentation of a pea protein-based beverage using Lactobacillus johnsonii NCC533. A combination of sensomics and sensoproteomics approach revealed that the fermentation process enriched or generated well-known basic taste ingredients, such as amino acids, nucleotides, organic acids, and dipeptides, besides six new taste-active peptide sequences that enhance kokumi and umami notes. The six new umami and kokumi enhancing peptides, with human recognition thresholds ranging from 0.046 to 0.555 mM, are produced through the degradation of Pisum sativum’s storage protein. Our findings suggest that compounds derived from fermentation enhance umami and kokumi sensations and reduce bitterness, thus improving the overall flavor perception of pea proteins. In addition, the analysis of intraspecific variations in the proteolytic activity of L. johnsonii and the genome–peptidome correlation analysis performed in this study point at cell-wall-bound proteinases such as PrtP and PrtM as the key genes necessary to initiate the flavor improving proteolytic cascade. This study provides valuable insights into the molecular mechanisms underlying the flavor improvement of pea protein during fermentation and identifies potential future research directions. The results highlight the importance of combining fermentation and senso(proteo)mics techniques in developing tastier and more palatable plant-based protein products.


Additional Tables:
Table S1.Compositions of the two raw materials employed to produce the pea beverages.

Figure S1 :
Figure S1: Tandem mass spectrometry (MS/MS) fragmentation spectrum of the peptide with sequence DKEEEQEETSKQVQ.The spectrum displays the relative abundance of detected ions versus the mass-to-charge ratio (m/z).Annotation of b-and y-type fragment ions is indicated, corresponding to the cleavage of peptide bonds.The precursor ion is indicated with a m/z of 612.61 [M+3H]3+, suggesting a triply charged state.

Figure S2 :
Figure S2: Tandem mass spectrometry (MS/MS) fragmentation spectrum of the peptide with sequence AGEENDNVIS.The spectrum displays the relative abundance of detected ions versus the mass-to-charge ratio (m/z).Annotation of b-, z-and y-type fragment ions is indicated, corresponding to the cleavage of peptide bonds.The precursor ion is indicated with a m/z of 612.61 [M+2H]2+, suggesting a double charged state.

Figure S3 :
Figure S3: Tandem mass spectrometry (MS/MS) fragmentation spectrum of the peptide with sequence EENVIVKV.The spectrum displays the relative abundance of detected ions versus the mass-to-charge ratio (m/z).Annotation of b-and y-type fragment ions is indicated, corresponding to the cleavage of peptide bonds.The precursor ion is indicated with a m/z of 929.53 [M+1H]1+, suggesting a single charged state.

Figure S4 :
Figure S4: Tandem mass spectrometry (MS/MS) fragmentation spectrum of the peptide with sequence GQIEEL.The spectrum displays the relative abundance of detected ions versus the mass-to-charge ratio (m/z).Annotation of b-and y-type fragment ions is indicated, corresponding to the cleavage of peptide bonds.The precursor ion is indicated with a m/z of 688.35 [M+1H]1+, suggesting a single charged state.

Figure S5 :
Figure S5: Tandem mass spectrometry (MS/MS) fragmentation spectrum of the peptide with sequence GSSHEVD.The spectrum displays the relative abundance of detected ions versus the mass-to-charge ratio (m/z).Annotation of b-and y-type fragment ions is indicated, corresponding to the cleavage of peptide bonds.The precursor ion is indicated with a m/z of 730.30 [M+1H]1+, suggesting a single charged state.

Figure S6 :
Figure S6: Tandem mass spectrometry (MS/MS) fragmentation spectrum of the peptide with sequence GSAQEVD.The spectrum displays the relative abundance of detected ions versus the mass-to-charge ratio (m/z).Annotation of b-and y-type fragment ions is indicated, corresponding to the cleavage of peptide bonds.The precursor ion is indicated with a m/z of 705.30 [M+1H]1+, suggesting a single charged state.

Figure S7 :
Figure S7: Tandem mass spectrometry (MS/MS) fragmentation spectrum of the peptide with sequence SREQIEEL.The spectrum displays the relative abundance of detected ions versus the mass-to-charge ratio (m/z).Annotation of b-and y-type fragment ions is indicated, corresponding to the cleavage of peptide bonds.The precursor ion is indicated with a m/z of 705.30 [M+2H]2+, suggesting a single charged state.

Table S2 .
Summary of the strains employed, the materials, the P1 and P2 preculture time, the preculture OD, the inoculum (%), the growth conditions, the fermentation time and the fermentation volume.

Table S3 .
Summary of extraction and solid phase extraction (SPE) Yields from pea beverage materials, including solvent usage, material recovery, and desired concentrations for subsequent application in pea beverage formulations

Table S4 .
SWATH method table: MS experiments conducted at each cycle time for each swath window.The upper part of the table reports the SWATH windows for the reverse phase method, whereas the lower part reports the windows for the HILIC method.The HILIC method is acquired up from 50 to 1000 Da, whereas the reversedphase method is acquired up to 1500 Da.

Table S6 .
Quantitation results of the upregulated metabolites in the extracted pea beverage fermented at 0 h (UEPB), fermented for 24 h (FEPB 24 hours ) and fermented for 48 h (FEPB 48 hours ).The table reports each analyte's concentration at 0h, 24h, and 48h.Samples were run in triplicates, and the mean concentration was reported in μmol/L with Relative Standard Deviations (RSD [%]).

Table S7
Comparative Analysis of Gene Presence in L. johnsonii Strains as Identified by BLAST Analysis.This table lists genes compared across different strains of L. johnsonii, including their UniProt IDs, GI numbers, gene names, and protein categories.The data is based on research by Liu et al. (2010), employing a stringent e-value threshold of 0.001 and a minimum coverage requirement of 80% for BLAST hits.It includes various protein categories such as ABC transporter systems, aminopeptidases, cell-wall bound proteinases, di/tripeptides uptake proteins, dipeptidases, endopeptidases, oligopeptides uptake components, proline peptidases, and tripeptidases.

Table S8 .
Evaluation of Peptide Sequences from Pea Protein Hydrolysates for Sensory Attributes Using iUmami and iBitter Sequence-Composition Models (SCM).7,8Thistable includes peptide length, fraction assignment, umami scores, predicted umami attributes, bitterness scores, and predicted bitterness attributes for sequences identified exclusively in the 48-hour fermented taste-active SPE fractions F1 and F2, except for the sequence ANAQPLQRE, which was unique to the SPE unfermented fraction U2

Table S9 .
BLAST scores, raw data.The provided table outlines the results obtained from the results obteained from the process where protein sequences were search through BLAST (Basic Local Alignment Search Tool) to compare against genomes of L. johnsonii strains, focusing on identifying the presence of specific genes within these strains.The criteria for selecting significant BLAST hits included e-value threshold of 0.001 and a minimum coverage of 80% relative to the query sequence.These data points, were then used to construct heatmaps, providing a visual representation of the genetic similarity and diversity among the L. johnsonii strains tested.