Exploring the Mechanism of Liquid Smoke and Human Taste Perception Based on the Synergy of the Electronic Tongue, Molecular Docking, and Multiple Linear Regression

Abstract

The flavor and peculiarity of modern smoked meat products must be adjusted and controlled with the use of liquid smoke. In this study, to explore the interaction between human taste receptors and flavor components, an electronic tongue, molecular docking, and statistical methods were used to build a structure-activity relationship model of smoked flavor components based on multiple linear regression (MLR). Eighteen molecular descriptors of 44 flavor components in liquid smoke (including phenols, furans, aldehydes, alcohols, pyrazines, ketones, and esters) were collected and analyzed via systematic cluster analysis, and finally classified into five categories. The liquid smoke taste analysis indicates that bitterness contributes significantly (P < 0.01) to the flavor of liquid smoke. The data from the molecular docking of the human bitter taste receptor TA2R1 and 4-ethyl-2-methoxyphenol (EMP) reveals that the perceptual process may depend on the main flavor component of liquid smoke as a ligand, as determined by AutoDock Vina; the affinity of the highest conformation was found to be -5.2 kcal/mol, EMP was stable at the active pocket and formed hydrogen bonds with residue Arg55, and EMP formed hydrophobic interactions with the residues Leu48, Leu51, Ala96, Leu99, Gly100, and Leu277. The structure-activity relationship model of smoked flavor components based on MLR reveals a good linear relationship between the molecular descriptors and the affinities between the flavor components in liquid smoke and the human taste receptor (R2 = 0.872). This model can be used to predict the affinities between smoked flavor components as ligands and the human bitter taste receptor.