Identification of potent inhibitors against transmembrane serine protease 2 for developing therapeutics against SARS-CoV-2

In viral binding and entry, the Spike(S) protein of SARS-CoV-2 uses transmembrane serine protease 2 (TMPRSS2) for priming to cleavage themselves. In this study, we have screened ‘drug-like’ 7476 ligands and found that over thirty ligands can effectively inhibit the TMPRSS-2 better than the control ligand. Finally, the three best drug agents L1, L2, and L6 were selected according to their average binding affinities and fitting score. These ligands interact with Asp435, Cys437, Ser436, Trp461, and Cys465 amino acid residues. The three best candidates and a reported drug Nafamostat mesylate (NAM) were selected to run 250 ns molecular dynamics (MD) simulations. Various properties of ligand-protein interactions obtained from MD simulation such as bonds, angle, dihedral, planarity, coulomb, and van der Waals (VdW) were used for principal component analysis (PCA) calculation. PCA discloses the evidence of the structural similarities to the corresponding complexes of L1, L2, and L6 with the complex of TMPRSS2(TM) and Nafamostat mesylate (TM-NAM). Moreover, Quantitative structure-activity relationship (QSAR) pattern recognition was generated using PCA for the investigation of structural similarities among the selected ligands. Multiple Linear Regression (MLR) model was built to predict the binding energy compared to the binding energy obtained from molecular docking. The MLR regression model reveals an accuracy of 80% for the prediction of the binding energy of ligands. ADMET analysis demonstrates that these drug agents are appeared to be safer inhibitors. These three ligands can be used as potential inhibitors against the TMPRSS2. Communicated by Ramaswamy H. Sarma

在病毒结合与入侵过程中，严重急性呼吸综合征冠状病毒2（SARS-CoV-2）的刺突（Spike, S）蛋白需借助跨膜丝氨酸蛋白酶2（transmembrane serine protease 2, TMPRSS2）完成剪切激活。本研究筛选了7476个“类药（drug-like）”配体，发现超过30种配体对TMPRSS2的抑制效果优于对照配体。最终依据平均结合亲和力与拟合得分，遴选出3种最优药物候选物L1、L2与L6。这些配体可与Asp435、Cys437、Ser436、Trp461及Cys465氨基酸残基产生相互作用。选取该3种最优候选物以及已报道的药物甲磺酸萘莫司他（Nafamostat mesylate, NAM），开展时长为250 ns的分子动力学（molecular dynamics, MD）模拟。从MD模拟结果中提取配体-蛋白相互作用的多项属性，包括键长、键角、二面角、平面性、库仑力与范德华（van der Waals, VdW）力，用于主成分分析（principal component analysis, PCA）计算。PCA分析结果显示，L1、L2、L6对应的复合物与TMPRSS2-甲磺酸萘莫司他（TM-NAM）复合物存在结构相似性。此外，本研究利用PCA生成定量构效关系（quantitative structure-activity relationship, QSAR）模式识别模型，以探究所选配体间的结构相似性。构建多元线性回归（Multiple Linear Regression, MLR）模型以预测配体结合能，并与分子对接所得结合能进行对比。该MLR回归模型对配体结合能的预测准确率达80%。ADMET分析结果表明，上述候选药物抑制剂具备良好的安全性。该3种配体可作为靶向TMPRSS2的潜在抑制剂。本文由Ramaswamy H. Sarma供稿。