Abstract
Antimicrobial peptides (AMPs) are naturally-occurring peptide antibiotics. AMPs are typically cationic and utilize their electrostatic interactions with the bacterial membrane to selectively attack bacteria. The way they work has inspired a vigorous search for optimized peptides for fighting resistant bacteria. Here, we present a physical model of membrane selectivity of AMPs. The challenge for theoretical modeling of membrane-peptide systems arises from the simultaneous presence of several competing effects, including lipid demixing and peptide-peptide interactions on the membrane surface. We first examine critically a number of models of peptide-membrane interactions and map out one, which incorporates adequately these competing effects as well as the geometry of various regions in membranes, occupied by bound peptides, anionic lipids within the interaction range of each peptide, and those outside this range. This effort leads to a systematically-improved model for peptide selectivity. Using the model, we relate peptide's intrinsic (Ccell-independent) selectivity to an apparent, Ccell-dependent one, and clarify the relative roles of peptide parameters and cell densities in determining their selectivity. This relationship suggests that the selectivity is more sensitive to peptide parameters at low cell densities; as a result, the optimal peptide charge, at which the selectivity is maximized, increases with the cell density in such a manner that this notion becomes less meaningful at high cell densities.