Importance of hydrogen bonding and aromaticity indices in QSAR modeling of the antioxidative capacity of selected (poly)phenolic antioxidants.

The quantitative structure-activity relationship (QSAR) models for predicting antioxidative capacity of 21 structurally similar natural and synthetic phenolic antioxidants was considered. The one-, two- and three-descriptor QSAR models were developed. For this purpose the literature data on the vitamin C equivalent antioxidative capacity (VCEAC) values were used as experimental descriptor of antioxidative capacity. Some thermodynamic and aromaticity properties, as well as the natural bond analysis (NBO) based quantities aimed at measuring the strength of intramolecular hydrogen bonds, were used as independent variables. It was examined whether a combination of these variables can yield a mathematical function that is in good correlation with the VCEAC values. It was shown that a combination of a certain thermodynamic descriptor (related to the single proton loss electron transfer mechanism) with the NBO-based quantities results in several two-descriptor models with the correlation coefficient greater than 0.950. Thus, a significant influence of internal hydrogen bonds on the antioxidative capacity of the studied molecules was confirmed. The best correlation with the VCEAC values was achieved within a three-descriptor QSAR model. This model was obtained by including a magnetic aromaticity index. It was found that aromaticity has only secondary effects on the antioxidative capacity.