Drug Studies Of GCGR And SMO Targets

G protein-coupled receptors（GPCRs）play essential roles in the majority of cell signaling of humans.The abnormal activation or expression of GPCRs is associated with many diseases.The human GPCRs protein families mainly include Class A（rhodopsin‐like）,Class B（secretin receptor family）,Class C（metabotropic glutamate）,and Class F（frizzled/smoothened）.There are already many approved targeted drugs and clinical trial candidates for targets found in these families.In this paper,two proteins of GPCRs were selected for studies of small molecule targeted drugs: one is glucagon receptor（GCGR）associated with type 2 diabetes mellitus（T2DM）that belongs to the Class B family,and the other is Smoothened（SMO）associated with tumors that belongs to the class F family.The research methods used were deep learning,de novo drug design,virtual screening,molecular docking,molecular dynamics simulations,construction and analysis of drug-target interaction network,cell experiments,etc.The specific research works include the following two aspects:1.Mol AICal,as an artificial intelligence drug design software,was used to design small molecule antagonists in the protein binding pocket of GCGR.10 compounds were selected from the clustered and filtered small molecules based on docking scores and MM/GBSA,and each of 10 compounds was subjected to the drug-formation prediction of ADMET properties.According to the network-based approach,the potential targets of 10 compounds were intersected with the disease targets of T2 DM and intersection targets were taken for GO and KEGG functional enrichment.Then the drug-target interaction network was constructed.The correlation between the targets of each compound and T2DM-related pathways were compared and analyzed to predict the best active drug candidate compound among 10 molecules.At the same time,the prediction were further validated by using cell experiments,with the best active compound having the optimal c AMP antagonistic activity.To reveal the mechanism of action of the best active compound targeting GCGR,we performed molecular dynamics simulations on apo-GCGR and antagonist-bound GCGR,respectively.The results showed that the antagonist destabilized the conformation of GCGR,making GCGR unable to bind glucagon and thus blocking signal transduction.These findings not only identified a new small molecule GCGR antagonist with potential for the treatment of T2 DM,but also provided a new low-cost approach to drug discovery.2.The crystal structure of SMO showed residue Asp95 on the extracellular domain had a noncovalent bond with cholesterol,while some experiments pointed out that it should be the covalent bond between Asp95 and cholesterol.To explore these contradictory results,molecular dynamics simulations and Markov state model were employed to study the interaction mechanism of SMO with noncovalent-bound cholesterol and covalent-bound cholesterol.Our results indicated cholesterol cannot keep a stable position around residue Asp95 if there is no covalent bond between Asp95 and cholesterol,which was consistent with the experimental results of covalently modified SMO rather than the reported crystal structure.Based on the above conclusion,virtual screening of small molecules was performed in the protein binding pocket of SMO with covalent-bound cholesterol,and we obtained a number of compounds targeting SMO.Some compounds were selected and their ADMET properties were predicted for drug formation.It was tentatively inferred that the selected compounds could be used for further studies.The results supplied useful information for SMO drug design.More importantly,the results suggested that crystal structures should be carefully analyzed when conducting relevant scientific research,and molecular dynamics simulations could obtain better and more reasonable crystal structures for drug design.