Computer-Aided Drug Design Of Selective α_1A-Adrenoceptor Antagonist Molecules And Cell Screening Model Studies

The α_1adrenergic receptors （α₁-ARs） play an important role in humannormal physiological regulation. They are also effective treatment targets of benignprostatic hyperplasia. The α₁-ARs can be divided into α_1a、α_1band α_1dsubtype atthe molecular level. Because different receptors in the same family are homologous,they will lead to similar effects when interacting with the ligands in vivo. For lowselectiveness antagonist, multiple activities would be emerged when interacting withreceptors, which will lead to reduce the efficacy and improve the side-effects.Therefore, the selectiveness improvement of the α₁-AR subtype has become the keyfactor in designing safe and efficient α₁-AR antagonists. The introduction ofcomputer-aided drug design （CADD） and high-throughput drug screening （HTS）method in drug design can greatly reduce the time to find lead compounds, improvethe drug development efficiency, reduce drug development costs and promote theoverall level of research. The purpose of the paper is trying to rich the theoretical indesigning highly selective and specific α_1A-AR ligand using CADD and HTSmethods. The paper contains the following:1. High-throughput screening in cell level: Dual-luciferase reportergene method was used to high-throughput compound screening. The reportergene vector was built by inserting the pGL3vector, which can express luciferase,into the response element of CRE. The reporter gene vectors and eukaryoticexpression of human α_1A-AR subtype plasmids （ADRA1A） were transfectedinto PC-12cells. The established drug-screening model, which based on the reportergene vectors and GPCR pathway of α_1Areceptor subtypes, can be used toinitial screening the small ligands with an effect on α_1A-AR subtype by measuringthe fluorescence intensity. To identify and optimize the assay condition, influencingfactors has been examined by using phenylephrine and prazosin. The result showedthat screening can successful done without use forskolin, which was a CRE activator.The optimum screening concentration of system is10^-5mol/L. The agonist incubationtime was from8hour and the final concentration of DMSO （methanol） was less than1%. The Z-factor value of0.7165shows the established screening model can be usedto high-throughput screening for α₁-AR antagonist.2. Pharmacophore model: pharmacophore model was established based on theligands of aryl piperazine α_1A-AR antagonist using GALAHAD module of SYBYL8.1software. The best pharmacophore model contains a hydrogen bond donor,two H-bond acceptors, a positively charged center, and two hydrophobic centers.Most of the pharmacophore parameters consistent with those reported publics. ThePharmacophore model also can match the better activity of the compounds with thetraining set.The established pharmacophore model has the ability to predict activity ofcompounds, the predicted values is closer to the experimental data. A small α_1A-ARantagonist data from ZINC database was used to validate the pharmacophore models.Active compounds from the α_1A-AR antagonist database can be screening, with a highenrichment factor of8.9.3．3D-QSAR：Classic3D-QSAR models （CoMFA and CoMSIA） of α_1A-ARantagonists were generated based on pharmacophore alignment by GALAHAD. In CoMFA study, the field filtration and grid step were examined. In CoMSIA study, thefield filtration factor, grid step, attenuation factors and different combinations of fieldwere checked. Both the CoMFA and CoMSIA models have a good cross-validation coefficient and linear regression coefficient, which showed that they havegood stability and strong predictive power. The CoMFA and CoMSIA contour mapsof the PLS regression coefficients at each region grid point provide a graphicalvisualization of the various field contributions, which can explain the differences inthe biological activities of each compound provides further theory for structureoptimization. The activities of existing ligands in laboratory were predicted using theestablished3D-QSAR model.4．Homology modeling and molecular docking: Three-dimensional structure ofα_1Aα_lBand α_1D-AR were constructed using homology modeling and moleculardynamics methods, with a high sequence similarity adrenergic receptor2-AR crystalstructure as template. Three classic receptor subtype antagonists （silodosin, L-765314and BMY-7378） were docking with the optimized three-dimensional structureusing flexible docking method. The ligands were deleted from the receptor-antagonistcomplexes after optimized using molecular mechanics and molecular dynamicsmethods. The antagonistic states of the models were reasonable and have high qualityusing program verification. Compounds with known antagonistic activity weredocking with each model, respectively. Regression analysis was done with dockingtotal score and the respective experimental values （pKi）.It was found that dockingresults could predict activity in some extent. The paper used the establish models andthe results of structure-activity relationship equation to predict the existing lignads inlaboratory in order to find highly selective α₁-AR antagonist agent.5. Virtual Screening: Based on previously established pharmacophore model andhomology modeling results, virtual screening of of compound libraries of ourlaboratories and zinc database were completed. In the screening progress, the methodsof Unity3D search, molecular similarity search and molecular docking combinedwith graphics inspection were used. After ADME/T prediction,15aryl piperazine α₁-AR antagonists of our laboratory libraries and16compounds of zinc database hadbeen successfully screened out. They were expected to become candidate compoundsof highly selective and high affinity of α₁-AR antagonists.6. New compounds’ design: Based on the pharmacophore distribution,3D-QSARand homology modeling above, using ligand fishing and fragment connection method,new compounds have been designed. The head group, the linker and the tail group ofthe aryl-piperazine compound have been transformed using MedChem Studio. Thefragments were connected with classic combinatorial chemistry design strategies. Avirtual screening library of α_1A-AR antagonist containing17920compounds was built.The theoretically high subtype-selective activity antagonists were screening usingmolecular docking. The further optimization of these compounds also be completedcombined with the3D-QSAR model and the feasibility synthetic route of the highestactivity compouds was designed.