Molecular dynamics simulations of the interaction of phospholipid bilayers with polycaprolactone

ABSTRACT

The molecular interaction between common polymer chains and the cell membrane is unknown. Molecular dynamics simulations offer an emerging tool to characterise the nature of the interaction between common degradable polymer chains used in biomedical applications, such as polycaprolactone, and model cell membranes. Herein we characterise with all-atomistic and coarse-grained molecular dynamics simulations the interaction between single polycaprolactone chains of varying chain lengths with a phospholipid membrane. We find that the length of the polymer chain greatly affects the nature of interaction with the membrane, as well as the membrane properties. Furthermore, we next utilise advanced sampling techniques in molecular dynamics to characterise the two-dimensional free energy surface for the interaction of varying polymer chain lengths (short, intermediate, and long) with model cell membranes. We find that the free energy minimum shifts from the membrane-water interface to the hydrophobic core of the phospholipid membrane as a function of chain length. Finally, we perform coarse-grained molecular dynamics simulations of slightly larger membranes with polymers of the same length and characterise the results as compared with all-atomistic molecular dynamics simulations. These results can be used to design polymer chain lengths and chemistries to optimise their interaction with cell membranes at the molecular level.

KEYWORDS:

• Phospholipid membrane
• polycaprolactone
• hydrophobic polymer
• molecular dynamics

Acknowledgements

MD would like to thank Dr Alex Tzanov from CUNY High Performance Computing Center and Lei Huang from the Texas Advanced Computing Center.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This research was supported, in part, by the NSF through TeraGrid resources under grant number TG-CHE130099 and a grant of computer time from the City University of New York High Performance Computing Center under NSF Grants CNS-0855217, CNS-0958379 and ACI-1126113. S. M. L. acknowledges start-up funding received from College of Staten Island and City University of New York. S. M. L. would also like to acknowledge PRF grant 54235-DNI6, NSF Grant DMR-1506937 and NSF Grant DMR-1750694. This work was supported by the American Chemical Society Petroleum Research Fund [grant number 54235-DNI6]; Directorate for Mathematical and Physical Sciences [grant number DMR-1506937 and DMR-1750694].