云计算技术下的靶向胰高血糖素受体的药物大数据的精准筛选及抗2型糖尿病药物活性的评价

The human glucagon (GCGR), which belongs to one member of class B G protein-coupled receptors, is an important drug target of type 2 diabetes mellitus. It is an important significance to design the drugs targeting to type 2 diabetes mellitus. There exist the problems of expensive screening cost and the evaluation of anti-diabetes drug activity on high throughput drugs targeting into GCGR. To solve these problems, this project will propose to build the low consumption and high efficiency screening platform based on the technologies of cloud computing and virtual screening. And the 35 million compounds are stored and screened on basis of the built platform. The accelerated molecular dynamics and free energy calculations will be further employed to study the dynamical interaction between GCGR and drug molecules at the atomic level. The screened drugs are further modified to improve the accuracy of targeting GCGR according to calculated information. Meanwhile, the project will construct the assay system to test binding ability between GCGR and drugs through the experiments of fluorescence label and calcium flux at molecular cell level. This project not only builds the screening and assaying platforms of drug big data, but also possibly screens the potential drugs which are effective in reducing blood glucose. This project has the profound significance on accelerated drugs development of anti-type 2 diabetes mellitus and cost reduction of medication in diabetic patients.

胰高血糖素受体（GCGR）是一个治疗2型糖尿病的重要药物靶点，它属于B类G蛋白偶联受体家族的成员之一，设计靶向GCGR晶体结构的拮抗剂对治疗2型糖尿病有着重要的意义。高通量筛选靶向GCGR的药物面临着昂贵的筛选成本和抗糖尿病药物的活性评价等问题，针对上述问题本项目采用云计算和虚拟筛选的技术构建低耗、高效的药物筛选平台，在此基础上完成3500万药物分子的存储和筛选，并通过加速分子动力学模拟和自由能计算从原子水平上研究配体与受体之间的动态相互作用，根据计算获得的信息改造筛选出的药物分子，进一步提高药物靶向GCGR的精准性；同时，本项目建立药物分子与GCGR结合效果的检测系统，通过荧光标记和钙流实验在分子细胞水平上检测筛选改造的药物与GCGR的作用活性。本项目既可以构建药物大数据筛选与测试的平台又可能筛选出有效降血糖的药物，这对加速研发抗2型糖尿病的药物和降低糖尿病患者的用药成本有着深远的意义。

人的胰高血糖素受体（GCGR）是一个治疗2型糖尿病的重要药物靶点，设计GCGR的拮抗剂可以用于治疗2型糖尿病。本项目成功设计了基于云计算的分布式软件MolGridCal，搭建了药物虚拟筛选平台，在此基础上筛选了~3500万药物分子，通过优化分子结构和K-means算法选取了30个药物分子，这种筛选方法节约了药物发现的成本和周期；通过加速分子动力学模拟和自由能计算从原子水平上研究了配体与受体之间的动态相互作用机制，结果显示10条平行加速分子动力学模拟轨迹区分了结合拮抗剂、结合激动剂和未结合任何配体的GCGR胞内区构象变化的机制。同时，为了验证挑选的30个药物分子是否有效，通过cAMP实验激活荧光标记的方法在HEK293细胞上检测30个药物分子的活性，发现了有活性的药物分子，为进一步降血糖药物的深度研发奠定了基础。