QSAR of Flavylium Salts as Inhibitors of Xanthine Oxidase Dragan Amic ´,* Dus ˇanka Davidovic ´-Amic ´, and Drago Bes ˇlo Faculty of Agriculture, The Josip Juraj Strossmayer University, P.O. Box 117, HR-31001 Osijek, The Republic of Croatia Bono Luc ˇic ´ and Nenad Trinajstic ´ The Rugjer Bos ˇkovic ´ Institute, P.O. Box 1016, HR-10001 Zagreb, The Republic of Croatia Received January 5, 1998 A simple QSAR model of xanthine oxidase (XO) inhibitory flavylium salts, which enables prediction of the inhibitory potency of anthocyanidins as a function of their molecular properties, has been developed. The results obtained in the present work help to understand which of the several tautomeric anhydrobase species present in nearly neutral solution are mainly responsible for the inhibition of XO. INTRODUCTION Flavylium salts (anthocyanins, anthocyanidins, and related compounds) are a part of the very large and widespread group of plant pigments known collectively as flavonoids. 1 Many of the red and blue fruits, vegetables, and flowers owe their attractive coloration to the anthocyanins dissolved in the cell sap. Apparently harmless to health, anthocyanins have a considerable potential in the food industry as safe and effective food colorants. 2 There has also been current medicinal interest in anthocyanins as biologically active substances. Several reviews on the pharmacological and medicinal properties of flavylium salts have been published. 3-6 For example, anthocyanin extracts are effective in decreasing capillary permeability and fragility; they possess antiinflam- matory, antioedematic, antioxidative, and antiulcer activities, show inhibitory activity of larval growth in insects, and exert pharmacological activities toward several enzymes. Very recently, inhibitory activity of some flavylium salts on xanthine oxidase (XO) has been published. 7 XO is an unusual enzyme (containing iron and molybdenum) which oxidizes xanthine to uric acid. 8 Since accumulation of excess uric acid in the body results in the painful disease gout (caused by crystallization of uric acid in the joints), there has been considerable interest in designing XO inhibitors. Species of Flavylium Salts Present at Physiological pH. Using QSAR for compounds having a wide variety of structures appears to be an extremely complicated problem. Tied to this is the question of the form of the compound which is the active species. At physiological pH, depending on the hydroxylation pattern, flavylium salts can exist in many different chemical entities, such as various neutral anhydrobases and their anionic forms in prototropic tauto- meric equilibrium. 9 All the tautomers shown in Figure 1 can be present in neutral aqueous solution. In addition, carbinol pseudobases and chalkones resulting from the hydration reactions also exist. 9-11 Recently, Rastelli et al. 12,13 have proposed a method for calculating the percentage composition of the various forms of anthocyanidins present in solution. These tautomers are likely to play different roles in pharmacological activity. Owing to the specificity of the enzyme-inhibitor interac- tions, one has to take into account that unfavored tautomeric forms could even become the preferred enzyme-bound forms. Thus, to establish a quantitative relationship between the chemical structure and biological activity we had to predict which of the species present in solution is/are mainly responsible for the exerted activity. RESULTS AND DISCUSSION To compare the inhibitory potency of the flavylium salts considered and to establish the relationships between struc- ture and activity, the dissociation constant of the enzyme- inhibitor complex, K EI , was used as experimental biological activity. The structures of the flavylium salts used in the analysis are shown in Figure 2. The variety in molecular descriptors is considered to clarify the effect of electronic, hydrophobic, and steric properties on K EI . We started with a descriptor pool containing the same descriptors as described in our recent paper. 14 In addition, Hansch’s hydrophobocity parameter π, molar refractivity MR, and Hammett’s electronic constants of substituents σ p were used. At physiological pH, both the neutral and the anionic anhydrobase forms of flavylium salts are present and could be involved in the interaction with the enzyme. Therefore, descriptors were calculated for all possible tautomeric species. To determine the contribution of each tautomeric form to explaining the enzyme inhibitory activity, models were calculated for all combinations of one to six tautomeric forms. For example, compound 1 possesses six tautomeric forms (A 5 ,A 7 ,A 4′ ,A 54′ - ,A 74′ - , and A 57 - ), compound 2 has three tautomeric forms (A 7 ,A 4′ , and A 74′ - ), while compound 8 possesses only one tautomeric form (A 4′ ). As usual, classical QSARs were calculated using stepwise multiple linear regression to fit the biological activity to molecular properties. This served to elucidate the relative significance of each independent variable in explaining the XO inhibitory activity as determined by regression analysis. Different models for predicting the inhibition of XO by a particular tautomeric form of anthocyanidins can be achieved by considering various combinations of molecular descrip- 815 J. Chem. Inf. Comput. Sci. 1998, 38, 815-818 S0095-2338(98)00002-X CCC: $15.00 © 1998 American Chemical Society Published on Web 08/20/1998