Incorporating Target Protein Structure Flexibility and Dynamics in Computational Drug Discovery Using Ensemble-Based Docking Analysis

Our research focuses on applying computational techniques to design more effective drugs, aiming to accelerate drug discovery and ultimately improve treatment outcomes and patient's quality of life.

Current computer-aided drug design often overlook target protein flexibility. The discussed protocol addresses this gap by incorporating multiple protein conformation derived from molecular dynamic simulation. Ensemble-based drug design improve accuracy by considering protein flexibility. We aim to integrate artificial intelligence for faster, personalized, and more effective drug discovery in the future.

[Presenter] To begin, launch the Avogadro software on a computer system. For cluster analysis, type the command given on-screen. When prompted, type 1 for the protein group to calculate least squares fit and root-mean-square deviation, or RMSD. Then type 1 again for system output. Open the cluster-size.xvg file. If the number of clusters is low, increase the RMSD cutoff value. Alternatively, if the number is high, decrease the cutoff. Perform grace analysis with the commands shown on screen using different RMSD cutoff values as needed. Now open the Chimera software and search for cluster.pdb. Click on Presents and Publication 1, silhouette, rounded ribbon. for visual representation. Sequentially click Select, Chain, No ID, followed by cluster.pdb, #10, then select and invert selected models. Go to Actions, then press Atoms/Bonds and click on Delete to isolate the chain. Now select File. Click on Save PDB. Name the file cluster1.pdb, and press Save to save the file. For ensemble-based docking, launch the Autodock tool software to open it. Place files cluster1.pdb and ligand.pdb into a new folder. Now click on File, Preferences, and Set. In the popup, paste the folder address into the startup directory field, and click Set. Click the blue folder icon, choose cluster1.pdb, and click Open. Go to Edit, then press Charges, Add Kollman Charges, and click OK. Click on Grid, Macromolecules, Choose. Select cluster 1 in the Choose Macromolecules box, then press Select Molecules. Click OK to generate a Modified AutoDock4 Macromolecule file. Save it as cluster1.pdbqt. Next, empty the workspace by clicking Edit, then press Delete, and delete all molecules before clicking Continue, then press Ligand, Input, Open. When a Ligand file for Autodock4 folder appears, select ligand.pdb and click Open, and then OK. Now choose Ligand, Torsion Tree, and Detect Root to define the torsional flexibility of the ligand. Go to Ligand, Output, Save as PDBQT, and save the Formatted Autotors Molecules folder as ligand.pdbqt. After emptying the workspace, open the cluster1.pdbqt file by clicking on Grid, Macromolecules, and Open, then press Yes and OK. Navigate to Grid again, and press the Set Map Types, and choose Open Ligand. Select and open ligand.pdbqt. Now navigate to the Grid Box option under Grid. In the Grid Options box, set number of points in the X, Y, and Z dimensions to 120, and spacing to 0.375 angstrom. Leave the center settings as default. Then click File and Close Saving Current. Go to Grid, Output, and press Save GPF. When the Grid Parameter Output file appears, enter grid.gpf as the file name, and click Save. Next, click on Run and Run AutoGrid. At Parameter Filename tab, click Browse. Open the grid.gpf file. Now browse through the program pathname. Search for autogrid4.exe and click Open and Launch. Sequentially, click on Docking, followed by Macromolecules and Set Rigid Filenames. When a PDBQT Macromolecules file appears, select cluster1.pdbqt and click Open. Choose the ligand from the Docking menu. When the Choose Ligands box appears, select Ligand and click on Select Ligand, then press Accept. Now navigate to Genetic Algorithm from Docking. When the Genetic Algorithm Parameters box appears, set GA Runs to 100 and click Accept. Click on Docking, Output, Lamarckian GA 4.2. When an Autodock4.2 GALS Docking Parameter Output file appears, name it as docking.dpf and click Save. Now press Run and Run Autodoc. A Run Autodoc box will appear. At Parameter File Name, click Browse. When an autodock4 Parameter file appears, select docking.dpf and click on Open. At Program Pathname, click Browse. An autodock4 file will appear. Search for autodock4.exe and click Open, followed by Launch. Delete all molecules as demonstrated previously, and repeat the process for all cluster files. The chemical structure and the 3D structural representation of flavokawain B and lysozyme at the initial state before molecular dynamic simulation was obtained. The total energy of the protein structure was stable during the simulation, and root-mean-square deviation stabilized after 20 nanoseconds. Root-mean-square fluctuation revealed high flexibility in regions between residues 40 to 50, 60 to 80, and 100 to the end. A total of 15 structural clusters were obtained from root-mean-square deviation-based clustering of 10,001 trajectory frames, with the largest cluster containing 5,818 members. Superimposed conformations of all clusters showed visible structural variations among the trajectories. Molecular docking of flavokawain B with the representative structures of the top four clusters showed consistent binding at the same site across all conformations, with cluster 2 showing the lowest binding energy of -29.37 kJ/mol. Electrostatic surface mapping confirmed identical binding sites in all cluster conformations with flavokawain B nested in the same pocket region. Detailed interaction analysis showed flavokawain B binding was stabilized by several surrounding residues, including alanine 31, glutamine 35, leucine 56, gamma-carboxyglutamic acid 57, isoleucine 58, alanine 95, isoleucine 98, tryptophan 108, valine 109, alanine 110, tryptophan 111, and arginine 114.