Ligand Binding Rate Constants in Heme Proteins Using Markov State Models and Molecular Dynamics Simulations

Abstract

Computer simulation studies of the molecular basis for ligand migration in proteins allow the description and quantification of the key events implicated in this process as, such as the transition between docking sites, displacements of existing ligands and solvent molecules, and open/closure of specific 'gates', among other factors. In heme proteins, especially in globins, these phenomena are related to the regulation of protein function, since ligand migration from the solvent to the active site preludes ligand binding to the iron in the distal cavity, which in turn triggers the different globin functions. In this work, a combination of molecular dynamics simulations with a Markov-state model of ligand migration is used to the study the migration of O2 and ·NO in two truncated hemoglobins of Mycobacterium tuberculosis (truncated hemoglobin N -Mt-TrHbN- and O -Mt-TrHbO). The results indicate that the proposed model provides trends in kinetic association constants in agreement with experimental data. In particular, for Mt-TrHbN, we show that the difference in the association constant in the oxy and deoxy states relies mainly in the displacement of water molecules anchored in the distal cavity by O2 in the deoxy form, whereas the conformational transition of PheE15 between open and closed states plays a minor role. On the other hand, the results also show the relevant effect played by easily diffusive tunnels, as the ones present in Mt-TrHbN, compared to the more impeded passage in Mt-TrHbO, which contributes to justify the different .NO dioxygenation rates in these proteins. Altogether, the results in this work provide a valuable approach to study ligand migration in globins using molecular dynamics simulations and Markov-state model analysis.