摘要

Rho-associated protein kinase (ROCK) has been recognized as an attractive therapeutic target to promote neurogenesis, neuroregeneration, and neurorecovery after cerebral injury. Here, a high-throughput screening protocol was described to discover novel ROCK inhibitors from a large chemical library containing 6.1 million structurally diverse, lead-like compounds. The protocol employed empirical rules such as ADMET evaluation and chemical similarity analysis to exclude those of drug-unlike candidates, and then molecular docking and binding affinity predictions were performed to suggest few promising candidates with high scores. Consequently, five compounds were successfully identified to have satisfactory activity profile with values at nanomolar level. In order to elucidate the molecular mechanism of inhibitor binding to target, the complex structures of ROCK kinase domain with the five identified compounds were modeled and examined in detail. A number of polar chemical forces such as hydrogen bonds and cation- interactions as well as nonpolar contacts such as - stacking and hydrophobic forces were revealed at the complex interface, conferring high affinity and strong specificity to inhibitor binding. In addition, several key residues around the kinase active site, including Val90, Lys105, Asn203, and Phe368, were found to play an important role in binding.