Quantum Mechanical Scoring: Structural and Energetic Insights into Cyclin-Dependent Kinase 2 Inhibition by Pyrazolo[1,5-a]pyrimidines

Abstract

A quantum mechanics (QM)-based scoring function has been applied to complexes of cyclin-dependent kinase 2 (CDK2) and thirty-one pyrazolo[1,5-a]pyrimidine-based inhibitors and their bioisosteres. A hybrid three-layer QM/MM setup (DFT-D/PM6-D3H4X/AMBER in generalized Born solvent) was used here for the first time as an extension of our previous full QM and SQM/MM (SQM means semiempirical QM) approaches. Two approaches to obtain the structures of the CDK2/inhibitor complexes were examined: i) building the modifications from one X-ray structure available coupled with a conformational search and ii) docking the compounds into CDK2. The QM-based scoring entailed a QM/SQM/MM optimization followed by calculations of the binding scores which were subsequently correlated with the experimental binding free energies. The correlation for the building protocol was good (r(2) = 0.64, predictive index = 0.81), whereas the docking approach failed. A decomposition of the interaction energies to ligand fragments enabled us to rationalize the differences in the binding affinities. In conclusion, we have developed and refined a QM-based scoring protocol and successfully applied it to reproduce the binding affinities in congeneric series of CDK2 inhibitors and to rationalize their potency. We thus propose that such a tool can be used in computer-aided rational drug design.

The Semiempirical Quantum Mechanical Scoring Function for In Silico Drug Design

“…[69]). A new consistent notation is in reference [81]. Particular terms describe the gas-phase interaction energy (DE int ), the interaction solvation/desolvation free energy [DDG solv ; Eq.…”

“…The former can be used for a medium-throughput modelling of congeneric series of ligands: based on the experimental structure of the lead compound in the complex with the target, other ligand modifications are built and their most favourable arrangement is found using MD quenching. [81] Docking is a very extensively used tool of CADD that performs a high-throughput generation of binding poses of dissimilar ligands in the protein binding sites. The quality of the pose generation varies across many docking approaches available, but generally experimental binding poses (the RMSD of heavy atoms of a ligand of < 2 ) can be reproduced in about 80 % of cases in redocking studies.…”

“…It can be used in conjunction with SQM methods and is applicable even for large complexes. To speed up the calculations, we have also proposed a SQM/MM procedure [81] in which protein surroundings up to 8 from the ligand were treated by SQM, whereas the rest of the protein was described by MM and surrounded by water on a less accurate and fast GB implicit solvent level. [87] In such a setup, however, caution must be taken regarding amino acid ion pairs, because unphysical proton transfers can occur due to incomplete polarisation of the SQM part from the MM part as well as due to the underestimation of the proton-transfer barrier in SQM.…”

“…[69,81,[92][93][94] In the first study, a series of HIV PR inhibitors were docked using UCSF DOCK and rescored by the PM6-DH2 scoring function. [69] The QM-based rescoring improved the docking results significantly (Figure 4 A,B).…”

“…[81] We implemented not only a doublelayer SQM/MM approach but also its triple-layer QM/SQM/MM extension for an even more accurate treatment of the innermost part on the DFT-D level. We compared ligand building and docking protocols, and only the former one gave a reasonable correlation of R 2 = 0.64 with the experimental inhibition constants.…”