Souza Paulo C T, Limongelli Vittorio, Wu Sangwook, Marrink Siewert J, Monticelli Luca
Groningen Biomolecular Sciences and Biotechnology Institute and Zernike Institute for Advanced Materials, University of Groningen, Groningen, Netherlands.
PharmCADD, Busan, South Korea.
Front Mol Biosci. 2021 Mar 29;8:657222. doi: 10.3389/fmolb.2021.657222. eCollection 2021.
Molecular docking is central to rational drug design. Current docking techniques suffer, however, from limitations in protein flexibility and solvation models and by the use of simplified scoring functions. All-atom molecular dynamics simulations, on the other hand, feature a realistic representation of protein flexibility and solvent, but require knowledge of the binding site. Recently we showed that coarse-grained molecular dynamics simulations, based on the most recent version of the Martini force field, can be used to predict protein/ligand binding sites and pathways, without requiring any information, and offer a level of accuracy approaching all-atom simulations. Given the excellent computational efficiency of Martini, this opens the way to high-throughput drug screening based on dynamic docking pipelines. In this opinion article, we sketch the roadmap to achieve this goal.
分子对接是合理药物设计的核心。然而，当前的对接技术存在蛋白质灵活性和溶剂化模型方面的局限性，并且使用的是简化的评分函数。另一方面，全原子分子动力学模拟能够真实地呈现蛋白质的灵活性和溶剂情况，但需要知道结合位点。最近我们表明，基于最新版马蒂尼力场的粗粒度分子动力学模拟可用于预测蛋白质/配体结合位点和途径，无需任何信息，且准确性接近全原子模拟。鉴于马蒂尼出色的计算效率，这为基于动态对接流程的高通量药物筛选开辟了道路。在这篇观点文章中，我们勾勒了实现这一目标的路线图。
