Structure prediction for Potential Binding Pockets of Parkin: a Candidate Parkinson disease
Structure prediction:
After sequence retrieval, 3D structure of parkin was predicted through homology modeling, threading and ab initio approaches. Various tools were used for 3D structure prediction. Homology modelling is performed by LOMETS and an offline standalone homology modeling program MODELLER. The query sequence was subjected to BLASTp against PDB to obtain a suitable template for homology modelling. A template with 95% query coverage and 90% of sequence similarity was observed. The extra chains and ligands were removed from template by using UCSF Chimera. 3D structure prediction was performed by using MODELLER for homology modeling approach. Threading approach was also utilized to predict the 3D structure of parking. The online tools including QUARK and CE-threader were also utilized for structure prediction [18, 19]. I-Teaser was also used for 3D structure prediction. HHpred was used to predict the secondary structure of the protein.
Results:
Various structure prediction tools were used to predict the 3D structure of parkin. Various structure evaluation tools were used to evaluate the generated models. The final selected structure was visualized by UCSF Chimera. The selected model has more efficient results as 80.600% quality factor by ERRAT was observed. 92.26% Verify3D value, What-Check 85%, Rampage shows 95.66 favoured region and Procheck predicts that protein has 90% optimizes physiochemical properties. Protein-protein docking was performed by using PatchDock and interacting residues were observed. 39 interacting residues of parkin were observed interacting with 42 active residues of PINK1. Ile-162 was interacting with Gln-63 and Gln-64, Glu-172 showed interaction against Gly-93 and Gly-94. Similarly, Thr-545 showed interaction with Cys-449 and Cys-451. All active residues from both the proteins were observed in the binding domains. The docked complex showed least global binding energy of -45.35 Kcal/mol.