Dose-Dependent Softening of Bacterial Model Membranes by Structurally Distinct Antimicrobial Peptides: A Coarse-Grained Molecular Dynamics Study

Antimicrobial peptides (AMPs) act at the membrane interface, where they remodel lipid packing defects and redistribute lateral stresses, yet a quantitative, dose-dependent understanding of how they alter membrane mechanical properties remains incomplete. We use coarse-grained MARTINI 3 molecular dynamics simulations to systematically characterize the mechanical and microstructural response of a 70:30 POPE:POPG bilayer to three AMPs spanning distinct structural classes: aedesin (alpha-helical, 2MMM), arenicin-1 (beta-hairpin, 2JSB), and indolicidin (disordered, 1G89). For each peptide we vary the surface loading from one to four peptides per leaflet and extract the bending modulus K _c , the area compressibility modulus K _A , peptide localization depth, bilayer thickness, peptide-lipid and peptide-peptide spatial organization, and leaflet-resolved lipid packing defect distributions. All three peptides soften K _c monotonically with loading, but at per-peptide rates that span a threefold range and order systematically by structural class: − 1.39 ± 0.09, − 0.66 ± 0.04, and − 0.44 ± 0.01 k _B T per peptide for aedesin, arenicin-1, and indolicidin, respectively. The tilt and twist moduli remain invariant across all conditions, indicating that the perturbation operates selectively on long-wavelength collective deformation modes. K _A softens for the two structured peptides but is statistically indistinguishable from the control for indolicidin, a dissociation we trace to a supraphosphate adsorption versus interfacial insertion dichotomy: structured peptides sit above the phosphate plane and act as supraphosphate wedges, while the disordered peptide threads into the interface without coherently displacing lipids. Independent geometric, spatial-organization, and microstructural observables corroborate this framework, with the deep versus shallow defect remodeling asymmetry providing a clean microstructural counterpart to the K _c – K _A dichotomy. Acyl chain order parameters resolve the per-lipid splay from the bilayer-averaged response and show that the per-lipid perturbation tracks conformational state rather than peptide length: the two structured peptides impose comparable per-lipid chain disordering despite differing in length, while the disordered peptide imposes far less. These findings establish a quantitative connection between peptide-induced defect remodeling and the elastic response of the bilayer, and suggest a design principle in which conformational restriction maximizes the per-peptide membrane perturbation, motivating experimental tests on stapled-peptide AMP analogs.