Quantitative Structure-Activity Relationship Analysis of Angiotensin-Converting Enzyme Inhibitory Pentapeptides Based on Amino Acid Descriptors

This study aimed to investigate the structure-function relationship of angiotensin-converting enzyme (ACE) inhibitory peptides and to elucidate the action mechanism of food-derived ACE inhibitory peptides. Based on the amino acid sequences of recently reported ACE inhibitory pentapeptides and their half-maximal inhibitory concentration (IC₅₀) values, a library of ACE inhibitory pentapeptides was generated and the structures of the ACE inhibitory pentapeptides were characterized using three amino acid descriptors, Z-scales, VHSE and SVHEHS. A partial least square (PLS) model for describing the quantitative structure-activity relationship (QSAR) of the ACE inhibitory peptides with the hydrophobic properties, steric properties, and electrical properties of amino acids as the independent variables and the lg IC₅₀ of the ACE inhibitory pentapeptides as the dependent variable was established using Matlab software. The results showed that the R² and Q² of the QSAR model based on Z-scales descriptor were 0.6411 and 0.5369, respectively, and Gln-Arg-Pro-Asn-Met showed higher ACE inhibitory activity as predicted by this model. The predicted and measured IC₅₀ were 0.0517 and (0.0400 ± 0.0083) μmol/L, respectively, and the error between them was 0.0117 μmol/L. The R² and Q² of the QSAR model based on VHSE descriptor were 0.7636 and 0.5081, respectively, and Leu-Arg-Ala-Phe-Gln exhibited better ACE inhibitory activity as predicted by this model. The predicted and measured IC₅₀ were 0.0438 and (0.0273 ± 0.0053) μmol/L, respectively, and the error between them was 0.0165 μmol/L. The R² and Q² of the QSAR model based on SVHEHS descriptor were 0.8405 and 0.4005, respectively, and Leu-Arg-Ala-Phe-Gln displayed better ACE inhibitory activity as predicted by this model. The predicted and measured IC₅₀ were 0.0055 and (0.0312 ± 0.0042) μmol/L, and the error between them was 0.0257 μmol/L. Among the three QSAR models, the one based on SVHEHS descriptor had the strongest fitting capability but weak predictive capacity, while the models based on Z-scales and VHSE descriptors could allow good QSAR analysis of the pentapeptides. Our modeling analysis showed that the activity of the ACE inhibitory peptides was negatively correlated with the hydrophobic characteristics of amino acids and positively correlated with the steric characteristics of amino acids. Molecular docking of three ACE inhibitory peptides to ACE protein (2X8J) showed that all the ACE inhibitory peptides could bind to ACE protein. This study provides a new tool for developing ACE inhibitory peptides and a theoretical basis for the development and application of food-derived ACE inhibitory peptides.