Constructing protein models for ligand-receptor binding thermodynamic simulations: an application to a set of peptidometic renin inhibitors.

Structure-based design is the application of ligand-receptor modeling to predict the activity of a series of molecules that bind to a common receptor for which the molecular geometry is available. Successful structure-based design requires an accurate receptor model which can be economically employed in the design calculations. One goal of the work reported here has been to reduce the size of a model structure of a macromolecular receptor to allow multiple ligand-receptor molecular dynamic (MD) simulations to be computationally economical yet still provide meaningful binding thermodynamic data. A scaled-down 10 A receptor model of the enzyme renin, when subjected to an alternate atomic mass constraint, maintains the structural integrity of the composite parent crystal structure. A second goal of the work has been to develop schemes to explore and characterize the protonation states of receptors and ligand-receptor systems. Application of the charge state characterization schemes to the hydroxyethylene and statine transition state inhibitors of renin in the training set suggests a monoprotonation state of the two active-site aspartate residues, where the lone proton resides on the outer carboxylate oxygen of Asp226 is most likely. For the reduced amide transition state inhibitors an active site consisting of both aspartates in the totally ionized state, and the ligand carrying a net +1.0 charge, is most stable and consistent with experimental data.