Open collaboration that uses NMR data to judge the correctness of phospholipid glycerol and head group structures

Patrick F. J. Fuchs, Matti Javanainen, Antti Lamberg , Markus S. Miettinen, Luca Monticelli, Jukka Määttä, O. H. Samuli Ollila, Marius Retegan, and Hubert Santuz

We compare the C-H order parameters measured by Nuclear Magnetic Resonance (NMR) experiments to those predicted by 12 different molecular dynamics (MD) simulation models. We focus on the order parameters of the lipid headgroups and glycerol backbones in phospholipid bilayers.

Only two of the models (CHARMM36 [1] and Maciejewski-Rog [2]) give a reasonable agreement with experiments for a fully hydrated lipid bilayer.

We then compare (for the two best-performing models at full hydration and for the Berger model [3], the most used lipid model in the literature) to NMR experiments the changes in the order parameters as a function of hydration level, NaCl and CaCl2 concentrations, and cholesterol content. The results clearly show that the glycerol and headgroup structures in the Berger model are not realistic, the Na ion partitioning is significantly too strong and cholesterol-induced structural changes are overestimated. The CHARMM36 and Maciejewski-Rog perform better, but the Na partitioning is too strong at least in the latter.

This is an open science project that is progressed at http://nmrlipids.blogspot.com. All the results and discussions are available at that address.

[1] J. B. Klauda, ..., R. W. Pastor. J. Phys. Chem. B 114 7830 (2010).
[2] A. Maciejewski, M. Pasenkiewicz-Gierula, O. Cramariuc, I. Vattulainen, T. Rog. J. Phys. Chem. B 118 4571 (2014).
[3] O. Berger, O. Edholm, F. Jähnig. Biophys. J. 72 2002 (1997).