Free energy calculation of single molecular interaction using Jarzynski’s identity method: the case of HIV-1 protease inhibitor system

• Published in: Acta mechanica Sinica Vol. 28; no. 3; pp. 891 - 903
• Main Authors: Li, De-Chang, Ji, Bao-Hua
• Format: Journal Article
• Language: English
• Published: Heidelberg The Chinese Society of Theoretical and Applied Mechanics; Institute of Mechanics, Chinese Academy of Sciences 01.06.2012
• Subjects: Classical and Continuum Physics; Computational Intelligence; Computer simulation; Engineering; Engineering Fluid Dynamics; Free energy; Human immunodeficiency virus 1; Inhibitors; Landscapes; Mathematical analysis; Mathematical models; Protease; Research Paper; Simulation; Theoretical and Applied Mechanics; Molecular biomechanics; Single molecular interaction; Molecular dynamics simulation; Free energy calculation; Steered molecular dynamics
• ISSN: 0567-7718; 1614-3116
• DOI: 10.1007/s10409-012-0112-9

Jarzynski’ identity (JI) method was suggested a promising tool for reconstructing free energy landscape of biomolecular interactions in numerical simulations and experiments. However, JI method has not yet been well tested in complex systems such as ligand-receptor molecular pairs. In this paper, we applied a huge number of steered molecular dynamics (SMD) simulations to dissociate the protease of human immunodeficiency type I virus (HIV-1 protease) and its inhibitors. We showed that because of intrinsic complexity of the ligand-receptor system, the energy barrier predicted by JI method at high pulling rates is much higher than experimental results. However, with a slower pulling rate and fewer switch times of simulations, the predictions of JI method can approach to the experiments. These results suggested that the JI method is more appropriate for reconstructing free energy landscape using the data taken from experiments, since the pulling rates used in experiments are often much slower than those in SMD simulations. Furthermore, we showed that a higher loading stiffness can produce higher precision of calculation of energy landscape because it yields a lower mean value and narrower bandwidth of work distribution in SMD simulations.