We will perform equilibrium Molecular Dynamics (MD) calculations, using the Amber suite, with a mix of co-solvent molecules at different concentrations (co-solvent MD simulations).In Co-solvent MD simulations water and co-solvents compete for pockets on the protein surface and the simulation results provide insight into desolvation contributions and also provide information regarding interaction between co-solvents and the interaction surfaces of LRRK2's WD40 domain. The mix of co-solvents for these simulations will be chosen to reflect functional groups found commonly in approved drugs. Our co-solvent approach will facilitate the exploration of potential cryptic binding sites as well as define areas where specific chemistry is preferred. An important feature of this approach is that the whole domain will be explored for potential binding sites without a priori selection of sites of interest. From the simulations, we will obtain density maps for each co-solvent which we will use to create pharmacophore models to screen databases built using ZINC compound libraries. The potential hits identified by this approach will be refined through a round of molecular docking in order to select /design ligands for this target. We plan to initially build a set of pharmacophore databases using ZINC compound libraries of interest using Schrodinger phase program. we will customize the libraries by adding additional compounds from ZINC database depending on the pharmacophore models for each site of interest identified in the simulations. we will use Schrodinger phase program to perform pharmacophore screening of large compound libraries to filter potential hits. we will use Schrodinger Glide to dock these compounds into their respective pockets and use the refined poses to understand the protein-ligand interactions and make compound selections. we refer to this approach as ' Co-solvent MD informed pharmacophore screening ' and we have found it to be a very powerful method to rapidly filter through large libraries to identify compounds of interest and to be able to find druggable pockets on challenging protein targets. We intend to use of VR technology in every step of this process to assess the co-solvent density generated through the simulations and identify potential binding sites to be considered. It will also help us analyze our pharmacophore hypotheses as well as review of binding poses for the final set of selected ligands.
