Phosphatidylcholine reverse micelles on the wrong track in molecular dynamics simulations of phospholipids in an organic solvent.

Here, we examine a well-characterized model system of phospholipids in cyclohexane via molecular dynamics simulations using a force field known for reproducing both phospholipid behavior in water and cyclohexane bulk properties to a high accuracy, CHARMM36, with the aim of evaluating the transferability of a force field parametrization from an aqueous environment to an organic solvent. We compare the resulting reverse micelles with their expected experimental shape and size, and find the model struggles with reproducing basic, experimentally known reverse micellar structural characteristics for common phosphadidylcholine lipids such as 1,2-dipalmitoyl-sn-glycero-3-phosphatidylcholine (DPPC), 1,2-dioleyl-sn-glycero-3-phosphatidylcholine (DOPC), and 1,2-dilinoleyl-sn-glycero-3-phosphatidylcholine (DLPC) in cyclohexane solvent. We find evidence that the deviation from the experimental behavior originates from an underestimation of the lipid tail-cyclohexane interaction in the model. We compensate for this, obtain reverse micellar structures within the experimentally expected range, and characterize these structurally in molecular detail. Our findings indicate extra caution and verification of model applicability is warranted in simulational studies employing standard biomolecular models outside the usual aqueous environment.