Quantum Chemical Prediction of the Acidities of Sulfonamide Inhibitors of Carbonic Anhydrase

Abstract

This study examined two pK a calculation approaches (direct and proton exchange schemes) that employ high-level quantum chemical methods and implicit solvent models to predict aqueous Brønsted acidities of a large set of sulfonamides. For gas-phase deprotonation energies, the DSD-PBEP86-D3(BJ) double-hybrid functional provided the best agreement with the LNO-CCSD(T)/CBS benchmark with a mean absolute deviation less than 2 kJ mol −1 when the aug-cc-pVTZ or larger basis sets are used. For a large test set of 54 primary and secondary sulfonamides, the use of the DSD-PBEP86-D3(BJ)/aug-cc-pVTZ level of theory in conjunction with SM12 solvation free energies predict their pK a values with a mean accuracy of 0.9 units. In comparison, the SMD and ADF-COSMO-RS models have slightly higher mean errors of 1.4 and 1.1 pK a units provided that the proton exchange scheme was employed to cancel the systematic errors in these models. The performance of these protocols was less ideal when applied to sulfonic acids, sulfamates, and N-substituted sulfonamides, indicating that the degree of error cancellation is sensitive to the chemical environment around the −NH 2 head group. The validated protocols were then used to estimate the pK a values of arylsulfonamide carbonic anhydrase inhibitors, which are used to correct their experimentally measured binding free energies to account for deprotonation of the sulfonamide group upon binding to the enzyme. These corrected values did not have a significant impact on the correlation with MMGBSA binding free energies obtained from classical MD simulations where the ligand is usually considered in the deprotonated form.

MM/GBSA prediction of relative binding affinities of carbonic anhydrase inhibitors: effect of atomic charges and comparison with Autodock4Zn

“…This is because experimental binding constants take into account the deprotonation of the sulfonamide ligand whereas this energetic cost is neglected in classical MD simulations since the ligand is modelled in its deprotonated form. That study concluded that the inclusion of this p K a correction did not result in a noticeable change in the correlation with experimental data presumably because the uncertainty in MM/GBSA predictions are higher than these corrections [ 32 ].…”

High-Level Quantum Chemical Prediction of C–F Bond Dissociation Energies of Perfluoroalkyl Substances

“…Herein, we assess the performance of contemporary DFT functionals, r 2 SCAN-3c minimal basis set composite electronic-structure method, and GFN2-xTB in predicting 550 C–F BDEs of 512 PFASs against the TightPNO-DLPNO-CCSD­(T)/CBS benchmark, hereon referred to as DLPNO-CCSD­(T)/CBS. The DFT functionals in this benchmark study include B3LYP-D3­(BJ), M06-2X, and ωB97M-V that have demonstrated good performance in thermochemical calculations. ,− For a subset of molecules, we have also examined the performance of composite wave function methods such as G4­(MP2)-6X and RO-CBS-QB3 with the view to identify cost-efficient methods for benchmarking. We first determined a suitable basis set that provides the best compromise between cost and accuracy, measured by relative computational speed-up and the deviation of the predicted C–F BDEs against our largest basis set, aug-cc-pVQZ, respectively.…”