DFT-based QSAR modelling of selectivity and inhibitory activity of coumarins and sulfocoumarins against tumor-associated carbonic anhydrase isoform IX.

In this study, we present four multilinear regression quantitative structure-activity relationship (QSAR) models, the first of which tackles the experimental CA IX isozyme inhibitory activities of a mixed set of 59 compounds (coumarins and sulfocoumarins), while the second and the third ones are local models for coumarins (42 compounds) and sulfocoumarins (17 compounds) for the same endpoint respectively. The fourth model deals with the selective inhibitory activity of 29 compounds on CA II and CA IX isozymes which was defined as differences of inhibition constants between CA II and CA IX (ΔpK=pK_CAII- pK_CAIX) in logarithmic scale. All presented models are statistically significant, robust and internally and externally validated. Most of the descriptors involved in these models are DFT-based physically-interpretable ones. Existence of the eigenvalues of chemical reactivity-related orbitals such as HOMO-1 Energy and LUMO Energy in the models allow us to speculate that the hydrolyzing reactions of the coumarins and sulfocoumarins in the active pocket of CA isozymes play a critical role in (or modulate) the binding free energy of the ligand-isozyme complexes. We believe that presented models may be used to design new virtual analogues of existing compounds with desired activity and selectivity towards CA IX or CA II.