Computational Modeling of the Interactions of Drugs with Human Serum Albumin (HSA)

Abstract

Human serum albumin (HSA) is the most abundant protein in the circulatory system that shows a remarkable capacity to bind a wide range of drugs impacting their therapeutic effect. Therefore, the binding to HSA represents a fundamental factor to consider when designing and developing new drugs. Although biophysical techniques (e.g. spectroscopy) are commonly employed to measure the extent to which drugs bind to HSA, these methods are time consuming and usually extremely expensive. Hence, there is an urgent need to incorporate more efficient methods in an attempt to streamline the development of new drugs. Here we present the implementation of a robust and cost-effective computational method to the prediction of the binding affinity of drugs towards HSA. Our method incorporates the program AutoDock Vina to perform in silico molecular docking of a highly diverse set of drugs against the 3D crystal structure of HSA. The 3D structure of HSA was retrieved from the Protein Data Bank and prepared to be used as receptor in our docking simulations. 3D structures of drugs were generated and optimized using Open Babel. Our protocol using AutoDock Vina as the docking engine was capable of reproducing the binding mode of indoxyl sulfate within the X-ray crystal structure of HSA (RMSD < 2.0 Å). In addition, our protocol correlated accurately predicted affinity values with experimentally determined association constants (r2=0.61). Our computational-based molecular docking approach incorporating AutoDock Vina may prove useful to the prediction of the binding affinities of drugs towards human serum albumin, and thus, could help alleviate a major bottleneck of the drug discovery process.