Structure-activity relationships for the degradation reaction of 1-beta-O-acyl glucuronides. Part 2: Electronic and steric descriptors predicting the reactivity of 1-beta-O-acyl glucuronides derived from benzoic acids.

Glucuronidation of carboxylic acid drugs has been found to be a metabolic activation pathway, possibly leading to covalent binding of the resultant 1-beta-O-acyl glucuronides (beta GAs) to proteins. Previous studies on the structure-activity relationships (SARs) of the degradation rate constants (k values) of beta GAs have revealed that the electrophilicity of and steric hindrance around the 1-beta-O-acyl linkages cause the diversity and complexity of the observed k values. To evaluate these effects and ultimately predict the k values of structurally diverse beta GAs, we derived further SARs for k values of 18 1-beta-O-benzoyl glucuronides with o-, m-, and p-substituents (BAGAs). In single regression analyses of 10 m- and p-substituted BAGAs, the log k values were well-predicted using an electronic parameter of Hammett's sigma constant, pK(a), (1)H NMR chemical shift (delta(COOH)), computed delta(COOH), or computed partial atomic charge (H(PAC) or O(PAC)) of the parent benzoic acids. The log k values of eight o-substituted BAGAs, although showing a correlation with the (13)C NMR chemical shift of the parent benzoic acids [delta(C=O)OH)], were well-predicted using multiple regression analyses; some combinations of electronic (delta(COOH), H(PAC), or calculated pK(a)) and steric [delta(C=O)OH) or Es] descriptors predicted the 18 observed k values with a high degree of certainty. The standard partial regression coefficients indicate that steric effects affected the k values as strongly as electronic effects, indicating that the k values increase as the acidity of the parent acids increases and as the steric bulkiness around the 1-beta-O-acyl linkages decreases. These single and multiple regression equations, using different electronic and/or steric descriptors of the parent benzoic acids, are expected to be useful for predicting the k values of BAGAs. The applicability domain and mechanistic interpretation of the derived SAR models are also discussed together with the relevant toxicology of beta GAs.