Faba Bean (Vicia faba L. minor) Bitterness: An Untargeted Metabolomic Approach to Highlight the Impact of the Non-Volatile Fraction

Abstract

In the context of climate change, faba beans are an interesting alternative to animal proteins but are characterised by off-notes and bitterness that decrease consumer acceptability. However, research on pulse bitterness is often limited to soybeans and peas. This study aimed to highlight potential bitter non-volatile compounds in faba beans. First, the bitterness of flours and air-classified fractions (starch and protein) of three faba bean cultivars was evaluated by a trained panel. The fractions from the high-alkaloid cultivars and the protein fractions exhibited higher bitter intensity. Second, an untargeted metabolomic approach using ultra-high-performance liquid chromatography-diode array detector-tandem-high resolution mass spectrometry (UHPLC-DAD-HRMS) was correlated with the bitter perception of the fractions. Third, 42 tentatively identified non-volatile compounds were associated with faba bean bitterness by correlated sensory and metabolomic data. These compounds mainly belonged to different chemical classes such as alkaloids, amino acids, phenolic compounds, organic acids, and terpenoids. This research provided a better understanding of the molecules responsible for bitterness in faba beans and the impact of cultivar and air-classification on the bitter content. The bitter character of these highlighted compounds needs to be confirmed by sensory and/or cellular analyses to identify removal or masking strategies.

Keywords: UHPLC-HRMS; bitterness; faba beans; metabolomic approach; off-flavours; pulses; sensory analysis.

Figure 1

Bitter intensity (over 10) of the 9 fractions evaluated by a trained panel. Significant differences are indicated by different letters (Tukey’s HSD test, α = 5.0%). S: starch fraction; F: flour; P: protein fraction—the number after the fraction corresponds to the cultivar (1, 2, or 3).

Figure 2

Biplot representation of the PCA (centred reduced variables, Pearson correlation, α = 5.0%) of the detected compound areas in the negative (A) or positive (B) modes and the perceived bitterness (as a supplementary variable in light green) of the 9 faba bean fractions. The compounds positively correlated with bitterness are related to variables in blue for the linear model, in dark green for the logarithmic model and in pink for both the linear and logarithmic models, whereas the black highlighted compounds correspond to the mis-dereplicated data of the positively correlated compounds. S: starch fraction; F: flour; P: protein fraction—the number after the fraction corresponds to the cultivar (1, 2, or 3).