基于结构和药效团特征的人类腺苷受体拮抗剂选择性比较

腺苷受体是重要的治疗靶标,选择性腺苷受体拮抗剂具有广泛的临床应用前景.本文通过同源模建构建了腺苷A₁、A_2B和A₃受体的结构,采用LigandScout 3.12软件分别构建了腺苷受体四种亚型的拮抗剂药效团模型.然后利用Schrödinger程序中的Induced Fit Docking模块完成受体-拮抗剂结合模式的预测,并与药效团结果进行比对.结果发现,由于结合口袋部位的残基在家族间高度保守,模建得到的各个亚型受体的初始结构活性口袋部位极为相似,无法用于亚型选择性拮抗剂的识别.而腺苷受体四种亚型拮抗剂药效团的药效特征与空间排布都不同,并与以前突变实验信息相吻合.研究结果说明,结合口袋部位的优化是模建中的关键步骤,基于配体的药效团模型所包含的一系列药效特征元素如氢键受体、氢键供体、疏水基团、芳环中心,可以很好地表征受体结合部位氢键、疏水空腔的位置及其方向.本文研究结果可以为进一步的优化同源模建结果,寻找新型的人类腺苷受体选择性拮抗剂提供理论依据.

Comparison of the Selectivity of Human Adenosine Receptor Antagonists Based on Structure and Pharmacophore Features

Adenosine receptors (ARs) are crucial therapeutic targets, and selective adenosine receptor antagonists are promising for numerous therapeutic applications. In this study, three dimensional models of human adenosine A₁, A_2B, and A₃ receptors (A₁AR, A_2BAR, A₃AR, respectively) were generated by homology modeling. In addition, pharmacophore models of the antagonists of four human adenosine receptor subtypes were developed using the LigandScout 3.12 program. Furthermore, Induced Fit Docking module of Schrödinger program was implemented to investigate receptor-ligand interactions. The results show that because of the subfamily-wide conservation of the core pocket residues, the ligand binding pockets of the three raw AR homology models are extremely similar, which poses challenges for subtype selective ligand recognition. However, the pharmacophore models of the four AR subtypes differ in pharmacophore features and spatial configuration, which are also consistent with previous site-directed mutagenesis studies. This indicates that binding site optimization is a crucial step in model generation, and the distributions for a set of pharmacophore features in ligand-based pharmacophore, including hydrogen bond acceptors, hydrogen bond donors, hydrophobic centroids, and aromatic rings, can reflect the position and direction characterization of hydrogen bonds and hydrophobic cavities, which aid identification and characterization of binding sites. This study may provide a significant theoretical foundation for further raw model optimization in homology modeling and discovery of novel selective human adenosine receptor antagonists.