Employing conformational analysis in the molecular modeling of agrochemicals: insights on QSAR parameters of 2,4-D

A common practice to compute ligand conformations of compounds with various degrees of freedom to be used in molecular modeling (QSAR and docking studies) is to perform a conformational distribution based on repeated random sampling, such as Monte-Carlo methods. Further calculations are often required. This short review describes some methods used for conformational analysis and the implications of using selected conformations in QSAR. A case study is developed for 2,4-dichlorophenoxyacetic acid (2,4-D), a widely used herbicide which binds to TIR1 ubiquitin ligase enzyme. The use of such an approach and semi-empirical calculations did not achieve all possible minima for 2,4-D. In addition, the conformations and respective energies obtained by the semi-empirical AM1 method do not match the calculated trends obtained by a high level DFT method. Similar findings were obtained for the carboxylate anion, which is the bioactive form. Finally, the crystal bioactive structure of 2,4-D was not found as a minimum when using Monte-Carlo/AM1 and is similarly populated with another conformer in implicit water solution according to optimization at the B3LYP/aug-cc-pVDZ level. Therefore, quantitative structure-activity relationship (QSAR) methods based on three dimensional chemical structures are not fundamental to provide predictive models for 2,4-D congeners as TIR1 ubiquitin ligase ligands, since they do not necessarily reflect the bioactive conformation of this molecule. This probably extends to other systems