Molecular simulation of transport proteins in interaction with physiological and pharmacological ligands

Abstract

Molecular complexes of transport proteins with small compounds have been studied by using docking techniques and molecular dynamics simulations. The macromolecules considered are β- lactoglobulin and albumin, i.e. the most abundant proteins in bovine milk and human blood serum, respectively. The ligands are long-chain fatty acids of different length and ibuprofen, a molecule of pharmaceutical interest. Simulations of β-lactoglobulin with fatty acids, ranging from caprylic to stearic acid, revealed the key protein residues that contribute to the binding process. In particular, a rationale was found for the high binding affinity of both stearic and palmitic acid compared to shorter lipids. Moreover, the location of two low-affinity external binding sites was predicted for palmitic acid, by comparing docking results with those obtained for vitamin D3, for which an external site has already been identified in crystallography. For human serum albumin, docking results suggest different candidate binding locations for both charged and neutral ibuprofen. An alchemical free energy approach has been used to estimate the binding affinity for each pose. The results show that charged ibuprofen has a greater affinity for albumin compared to the ligand in the neutral form, suggesting that the former corresponds to the physiological binding state. The simulation findings were compared to experimental results and show an overall good agreement, predicting details of the protein-ligand interaction that include binding geometries and contacts with specific amino acid residues. The overall findings reveal significant features of the binding of well-known ligands to two extensively investigated transport proteins, and show how computational tools can be used to support experimental techniques in a variety of cases.