Molecular dynamics methods to predict peptide locations in membranes: LAH4 as a stringent test case

Determining the structure of membrane-active peptides inside lipid bilayers is essential to understand their mechanism of action. Molecular dynamics simulations can easily provide atomistic details, but need experimental validation. We assessed the reliability of self-assembling (or "minimum-bias") and potential of mean force (PMF) approaches, using all-atom (AA) and coarse-grained (CG) force-fields. The LAH4 peptide was selected as a stringent test case, since it is known to attain different orientations depending on the protonation state of its four histidine residues. In all simulations the histidine side-chains inserted in the membrane when neutral, while they interacted with phospholipid headgroups in their charged state. This led to transmembrane orientations for neutral-His LAH4 in all minimum-bias AA simulations and in most CG trajectories. By contrast, the charged-His peptide stabilized membrane defects in AA simulations, whereas it was located at the membrane surface in some CG trajectories, and interacted with both lipid leaflets in others. This behavior is consistent with the higher antimicrobial activity and membrane-permeabilizing behavior of the charged-His LAH4. In addition, good agreement with solid-state NMR orientational data was observed in AA simulations. PMF calculations correctly predicted a higher membrane affinity for the neutral-His peptide. Interestingly, the structures and relative populations of PMF local free-energy minima corresponded to those determined in the less computationally demanding minimum-bias simulations. These data provide an indication about the possible membrane-perturbation mechanism of the charged-His LAH4 peptide: by interacting with lipid headgroups of both leaflets through its cationic side-chains, it could favor membrane defects and facilitate translocation across the bilayer.