What Vesicle Leakage Reveals about Antimicrobial Activity (and What It Doesn't)

Sara Hovakeemian,Runhui Liu,Samuel H. Gellman,Heiko Heerklotz
DOI: https://doi.org/10.1016/j.bpj.2014.11.3021
IF: 3.4
2015-01-01
Biophysical Journal
Abstract:The mode of action of antimicrobial peptides and their mimics is often assessed by vesicle leakage experiments, and most of these compounds are believed to act by membrane permeabilization. This work aims at improving the interpretation of vesicle leakage data in general, and at understanding and optimizing the fungicidal activity of nylon-3 polymers. We have studied the membrane permeabilizing properties of the cationic homopolymer poly-NM (Liu, R et al. JACS, 2013, 135, 5270), which displays significant antifungal activity, and two related cationic/hydrophobic binary copolymers using the lifetime-based leakage assay of calcein-loaded vesicles. We compared the results with biological activities against Candida albicans. Poly-NM induces all-or-none leakage of vesicles that are made from yeast polar lipid extract (YPLE), at the polymer's MIC against C. albicans (3 μg/mL). At this and higher concentrations, leakage requires a lag time but then proceeds to 100%. Concerted activity tests imply that the activity of the polymer does not involve detergent-like effects. Both vesicle leakage and antimicrobial activity against C. albicans spheroplasts are independent of the presence of a detergent, octyl glucoside (OG). Negligible activity is found against zwitterionic vesicles or red blood cells. All these characteristics provide a consistent, detailed picture of membrane leakage induced by electrostatic lipid clustering. The cationic/hydrophobic binary copolymer 40:60 MM:CO shows a fundamentally different pattern. Vesicle leakage is transient (limited to <100%) and graded, unspecific between zwitterionic and YPLE vesicles, additive with detergent action, and correlates poorly with biological activity. This activity profile suggests action by membrane asymmetry stress. We conclude that comprehensive leakage experiments can identify the mode of action for model membrane disruption, and we hypothesize that the correlation between vesicle leakage and antimicrobial activity is good for some types of membrane leakage but not for others.
What problem does this paper attempt to address?