The impact of non-ideality of lipid mixing on peptide induced lipid clustering

Abstract

The influence of several antimicrobial trivalent cyclic hexapeptides on the mixing behavior of bilayer lipid membranes containing phosphatidylglycerol (PG) and phosphatidylethanolamine (PE) with varying composition was studied using DSC and ITC. The peptides contained three arginines and three aromatic amino acids and had different sequences. All of them induce clustering of PG-rich clusters with bound peptides after binding. In a previous publication we could show that a correlation between clustering efficacy and the antimicrobial activity of the peptides exists (S. Finger et al., Biochim. Biophys. Acta 1848 (2015) 2998-3006). In the current study we investigated whether the non-ideality of the lipid mixture had any effect on the clustering efficacy and the critical peptide/lipid clustering ratio. We could show that for PG/PE membranes containing 1:1 M ratios and lipids with equal or unequal chain lengths, the amount of clustered PG depended only slightly on the absolute chain length and on the chain length difference between PG and PE. Much larger differences were observed when the PG/ PE mixing ratio was changed. In mixtures of DPPG/DPPE with high PG content, the amount of clustered PG per added peptide was much higher than in PE-rich mixtures. The ITC experiments showed that the critical peptide/lipid ratio for cluster formation is also strongly dependent on the PG/PE ratio in the mixture. In the PG/PE 3:1 mixture, the formation of clusters with bound peptide is much more likely than for mixtures with less PG. For 1:1 and 1:3 lipid mixtures, the critical peptide/lipid ratio for demixing is between 0.002 and 0.004. Therefore, even in these mixtures clustering occurs way below charge saturation of the PG in the mixture and the PG-rich clusters are not charge compensated either. The peptide concentration necessary for inducing clustering amounts to ~8 μM, a value well within the range of minimal inhibitory concentration values observed for the cyclic peptides studied here. Our results show that not only the structure of the cyclic peptide influences the clustering efficacy but also the mixing behavior of the lipids in the bilayers has an influence on the amount of clustering induced by binding of cyclic peptides.

Keywords: Antimicrobial peptides; Differential scanning calorimetry; Isothermal titration calorimetry; Lipid clustering; Lipid-peptide interactions; Model membranes.