2.2 Virtual Screening and Docking
The crystal structures of TP53, CASP1 and HSP90AB1 proteins were obtained from RCSB Protein Data Bank20 (PDB ID: 2Z2121,1RWX22 ,5FWK23, respectively). The sequences of four protein were obtained from Uniprot Knowledgebase24 (Identifier: NLRP3-Q96P20, TP53-P04637, CASP1-P29466 and HSP90AB1-P08238). With I-TASSER242526 (https://zhanglab.ccmb.med.umich.edu/I-TASSER/)25-27, the crystal structures of NLRP3 was constructed complete protein structure. The complete protein structure contained ligands for predicting binding sites and the highest confidence model was chosen. Ramachandran plot validation of our modeling structure was shown in Figure 4 and the 3D-profilevalidation of modeling structure was shown in Figure 5. It can help us to find whether the structure is reliable. We used ‘Prepared protein’ module in Accelrys Discovery Studio (DS) software to prepare the proteins in order to make the result more credible. 18776 TCM compounds from TCM Database @ Taiwan (http://tcm.cmu.edu.tw) 8 which were filtered by Linpinski’s Rule of Five, were applied to dock into NLRP3, TP53, 1RWX and 5FWK protein structures by using LigandFit module28 in DS. To minimized docking poses between ligands and proteins, Chemistry at HARvard Molecular Mechanics (CHARMm) force field 29,30 were used. During the docking process, the DREIDING force field containing Gasteiger charges was chosen to calculate the interaction energy for each ligand. The ADMET Descriptors module in DS was used to calculate absorption, distribution, metabolism, excretion, toxicity (ADMET) properties such as solubility, blood-brain barrier (BBB) penetration, hepatotoxicity, cytochrome P450 2D6 (CYP2D6) inhibition, plasma protein binding (PPB) level and absorption. Results of the docking studied were ranked according to Dock Score which was calculated mainly based on the interaction energy between ligand and protein.