Methodological Issues in Free Energy Calculation of Lipid Flip-Flop in Bilayers

Abstract

Lipid flip-flop is an important biological process and may play an important role in trafficking biological relevant molecules such as ions, proteins and drug molecules. Recent molecular dynamics simulations suggest that it is accompanied by water pore formation. However in this work, we find that this molecular observation is sensitive to the simulation parameters. In Berger forcefield, the choice of the van del Waals (vdw) cutoff length makes significant difference. In the literature, various vdw cutoff lengths have been used and it is usually assumed not to matter as long as dispersion correction is applied. However the dispersion correction does not work well in lipid bilayers due to the inhomogeneity in space. By varying the vdw cutoff from 0.9 to 1.4 nm, the area per lipid reduces from 68.0±1.0 to 60.4±0.9 A2. With larger cutoff length the free energy is 20 kJ/mol higher and no water pore is spontaneously forms. Since the area per lipid is only indirectly determined from experiments and has been reported in a wider range in literature, it does not serve as strong evidence to choose the right cutoff. To determine the appropriate cutoff length, we adopt a similar principle as used to develop the Berger forcefield. We determine the cutoff length that can produce the right density and heat vaporization energy of pentadecane but without applying dispersion correction. We recommend this cutoff for future simulations especially for free energy calculations of lipid flip-flops and transfer cell-penetrating and antimicrobial peptides.