No abstract

Proteins are linear polymers built from a repertoire of 20 different amino acids, which are considered building blocks of proteins. The diversity and versatility of these 20 building blocks with regard to their conformations are key to adopting three-dimensional structures that facilitate proteins to undergo important mechanistic biological processes in living systems. The present investigation reports a conformational search of 20 different amino acids, building blocks of proteins, using three different force fields, CHARMM, AMBER, and OPLS-AA, implemented in the gradient gravitational search algorithm. The search technique (ConfGGS) includes the contribution from both bonded and nonbonded terms using Cartesian coordinates. The efficiency of such conformational searches has also been compared with other optimization algorithms: DE/Best, DE/Rand, and PSO algorithms with respect to computational time and accuracy based on the minimum number of iteration steps and computed lowest mean absolute error (MAE) and mean standard deviation (MSD) values for dihedral angles of respective near-optimal structures. Moreover, the ConfGGS technique has also been extended to an ordered protein fragment (PQITL) extracted from HIV-1 protease (PDB ID: 1YTH), an intrinsically disordered protein fragment, i.e., an amyloid-forming segment (AVVTGVTAV), from the NAC domain of Parkinson’s disease protein α-synuclein, residues 69–77 (PDB ID: 4RIK), the experimental NMR atomic-resolution structure of α-synuclein fibrils (PDB ID: 2N0A), and a disulfide bond-containing protein fragment sequence (PCYGWPVCY), residues 59–67 (PDB ID: 6Y4F) toward structure prediction as a close homologue compared with experimental accuracy, using the CHARMM force field. The MolProbity validation results for the protein fragment (PQITL) obtained by ConfGGS/CHARMM are in better agreement with the native protein fragment structure of HIV-1 protease (PDB ID: 1YTH). Furthermore, the computed results have also been compared with the coordinates obtained from the AlphaFold network.

There is growing evidence for the rapid rise of strains that encode mutant proteases resistant to competitive reversible inhibitors of HIV-1 protease, based on enzyme-substrate interactions and with FDA approval. The inhibition potencies of irreversible inhibitors are less sensitive to mutations so as to inactivate the protein completely by stronger covalent interactions. The development of new irreversible protease inhibitors might be interesting to deal with the future handling of HIV. The mechanisms and binding modes of aziridine based inhibitors have been explored in the present investigations using in-silico approaches: (i) ConfGGS towards structure minimization of model aziridine based inhibitors (ii) Molecular Docking towards predicting the best match between model aziridine based inhibitors and HIV-1 PR (iii) Covalent Docking towards exploring the binding affinity for the covalent interaction between model aziridine based inhibitors and HIV-1 PR (iv) MD Simulation of free enzyme HIV-1 PR and complex with the model aziridine based inhibitors to test and check the quality for the description of inhibition process (v) QM/MM computation to understand the inhibition potency and inhibition reaction at molecular level. Furthermore, ConfGGS/CHARMM has also been used to optimize the reactants and products, obtained from QM/MM computations. The correlation coefficient (R2) values for the dihedral angles of the near optimal structures and QM/MM obtained structures, have been computed and compared for the accuracy and efficacy. The computed results may help and provide assistance for experimental optimizations towards design of more potent protease inhibitors.