Antimicrobial peptides bind more strongly to membrane pores

Abstract

Antimicrobial peptides (AMPs) are small, usually cationic peptides, which permeabilize bacterial membranes. Understanding their mechanism of action might help design better antibiotics. Using an implicit membrane model, modified to include pores of different shapes, we show that four AMPs (alamethicin, melittin, a magainin analogue, MG-H2, and piscidin 1) bind more strongly to membrane pores, consistent with the idea that they stabilize them. The effective energy of alamethicin in cylindrical pores is similar to that in toroidal pores, whereas the effective energy of the other three peptides is lower in toroidal pores. Only alamethicin intercalates into the membrane core; MG-H2, melittin and piscidin are located exclusively at the hydrophobic/hydrophilic interface. In toroidal pores, the latter three peptides often bind at the edge of the pore, and are in an oblique orientation. The calculated binding energies of the peptides are correlated with their hemolytic activities. We hypothesize that one distinguishing feature of AMPs may be the fact that they are imperfectly amphipathic which allows them to bind more strongly to toroidal pores. An initial test on a melittin-based mutant seems to support this hypothesis.

Figure 1

Thermodynamic cycle for binding of antimicrobial peptides (AMPs) to pores.

Figure 2

Geometry of a cylindrical (A) and a parabolic (B) pore. The solid lines represent the hydrophobic/hydrophilic interface.

Figure 3

The relative binding energies of the peptides in the pores of radius R_o (Å) and the curvature determined by k (k=0 corresponds to a cylindrical pore). Error bars are the standard deviation. A two-tailed Student’s t-test was used to determine significant differences between ΔΔW in each toroidal pore and ΔΔW in the cylindrical pore of the same R_o; p value < 0.05 is denoted by asterisk.

Figure 4

Contributions from the electrostatic (ΔΔW_elec) and solvation energy (ΔΔW_solv) to the relative binding energy of peptides in different pores, as well as contributions from polar (ΔΔW_polar) and aliphatic and aromatic groups (ΔΔW_hfob) to ΔΔW_solv. Error bars are the standard deviation. The radius of the pores, R_o, is in Å. The curvature of the pores is determined by k; the smaller the k, the smaller the curvature; k=0 corresponds to a cylindrical pore.

Figure 5

The optimal orientation of alamethicin, melittin, MG-H2 and piscidin on the flat membrane (intf) and in the pores of R_o=15 Å. The energy-minimized average structure is calculated from a MD simulation between 0.5 ns and 1 ns. The hydrophilic/hydrophobic interface is denoted by lines; only one side of the pore is shown.

Figure 6

Tilt angle of alamethicin, melittin, MG-H2 and piscidin vs. time, in the selected pores. The four lines in the same color represent the tilt angle calculated from 4 MD trajectories generated using different seeds.

Figure 7

The relative binding energies of melittin confined to the transmembrane orientation in the pores of radius R_o (Å) and the curvature determined by k (k=0 corresponds to a cylindrical pore). Error bars are the standard deviation.

Figure 8

The relative binding energy of a perfectly amphipathic melittin-based mutant in the pores of radius Ro (Å) and the curvature determined by k (k=0 corresponds to a cylindrical pore). Error bars are the standard deviation.