Mechanism of action of beta-stranded antimicrobial (KL)n model peptides.

Abstract

Here, we investigate model amphipathic β-stranded peptides, consisting of alternating lysine and leucine residues. These so called KLn peptides ([KL]n/2-NH2) show high antimicrobial activity against various bacterial strains. At the same time, they tend to aggregate into β-pleated amyloid-like fibrils. We systematically investigated their membrane affinity and their structural behaviour in the presence of lipid membranes as a function of peptide length, charge and hydrophobicity, using peptides with 6-26 amino acids. Different biophysical techniques (circular dichroism, solid-state NMR and fluorescence assays) were used and the results correlated with results of biological assays (antimicrobial and hemolytic assays), to identify possible mechanisms of action of KLn peptides. We find that the very short KL peptides are inactive, but longer peptides are highly hemolytic and aggregates strongly, especially in the presence of phosphate [1]. There is an optimal length of around 10 amino acids where the KL peptides have a high antimicrobial effect and low hemolysis. At this length, peptides are long enough to form stable β-stranded structure but flexible enough to not aggregate very fast in solution. A similar length behaviour has been observed for other similar peptides made from alternating hydrophobic and charged residues [2]. Vesicle leakage is length-dependent and peptides need at least 12 amino acids to induce high leakage. This threshold length does not depend on membrane thickness. This indicates that pores are probably not formed by these peptides, and this was further confirmed by solid-state 15N- and 19F-NMR experiments, whose results were incompatible with pore models [3]. Thus, it seems that the peptides are bound to the membrane surface and destabilize the membrane by some kind of carpet mechanism.