Scarpino Andrea, Ferenczy György G, Keserű György M
Medicinal Chemistry Research Group, Research Centre for Natural Sciences, Budapest, Hungary.
Methods Mol Biol. 2021;2266:73-88. doi: 10.1007/978-1-0716-1209-5_4.
The mechanism of action of covalent drugs involves the formation of a bond between their electrophilic warhead group and a nucleophilic residue of the protein target. The recent advances in covalent drug discovery have accelerated the development of computational tools for the design and characterization of covalent binders. Covalent docking algorithms can predict the binding mode of covalent ligands by modeling the bonds and interactions formed at the reaction site. Their scoring functions can estimate the relative binding affinity of ligands towards the target of interest, thus allowing virtual screening of compound libraries. However, most of the scoring schemes have no specific terms for the bond formation, and therefore it prevents the direct comparison of warheads with different intrinsic reactivity. Herein, we describe a protocol for the binding mode prediction of covalent ligands, a typical virtual screening of compound sets with a single warhead chemistry, and an alternative approach to screen libraries that include various warhead types, as applied in recently validated studies.
共价药物的作用机制涉及其亲电弹头基团与蛋白质靶点的亲核残基之间形成化学键。共价药物发现的最新进展加速了用于共价结合物设计和表征的计算工具的开发。共价对接算法可以通过对反应位点形成的键和相互作用进行建模来预测共价配体的结合模式。它们的评分函数可以估计配体对目标靶点的相对结合亲和力，从而允许对化合物库进行虚拟筛选。然而，大多数评分方案没有针对键形成的特定项，因此无法直接比较具有不同固有反应性的弹头。在此，我们描述了一种用于共价配体结合模式预测的方案、一种典型的对具有单一弹头化学性质的化合物集进行虚拟筛选的方法，以及一种筛选包含各种弹头类型的文库的替代方法，这些方法已应用于最近经过验证的研究中。
