Mechanism of action of antimicrobial peptides: pore formation and beyond

Antimicrobial peptides (AMPs) are produced by the innate immune system of a variety of organisms, including man, as a first defence line against pathogen infections. They are bactericidal towards a wide range of microbes, usually acting by perturbing the permeability of the bacterial membrane, leading to cell death by collapse of transmembrane electrochemical gradients and loss of important metabolites and cellular components. The absence of receptor-mediated recognition processes makes AMP extremely less sensitive than other antibiotic molecules to the development of bacterial resistance. For this reason, AMPs are excellent candidates as a new class of therapeutics to address the problem of the increasing insurgence of multi-drug resistant superbugs. Several alternative models have been proposed to describe the permeabilization of bacterial membranes by AMPs, and the two most commonly accepted hypotheses are based on the formation of pores into the phospholipid bilayer, by a tansmembrane arrangement of peptide aggregates (barrel-stave model), or by accumulation of peptides on the membrane surface that generates a stress in the bilayer, released through the formation of membrane defects (carpet model). More recently, the idea that the function of antimicrobial peptides is not limited to their bactericidal activity has gained support. Increasing evidences indicate that AMPs play a key role as regulators of the immune response, being involved in processes like chemokine production and regulation, angiogenesis, and wound healing. However, debate is still heated regarding the prevalence of one of the two main functions of AMPs (bactericidal and immunomodulatory) in vivo. The bactericidal behaviour of AMPs is generally investigated trough biophysical studies on model membranes and biological assays on bacterial cells. However, it is still questioned whether bacterial killing requires the same high degree of membrane coverage observed for liposome permeabilization, and whether membrane perturbation is possible under the conditions of bacterial cell density and peptide concentration that take place in vivo. Trying to answer these questions, in this thesis the association of an antimicrobial peptide (PMAP-23) to Escherichia coli cells was quantitatively determined, under the same experimental conditions (temperature, medium and peptide concentration) of the bactericidal assay. These data allowed the estimation of the average number of peptide molecules associated to each E.coli cell at the peptide concentrations causing bacterial death (107 PMAP-23 molecules per cell). These values correspond to complete saturation of bacterial membranes, a finding that validates the carpet model of pore formation hypothesized for PMAP-23 based on studies on model systems. Peptide/membrane association was analysed also on vesicles formed from a lipid extract of E. coli cells. The behaviour of the peptide with liposomes and bacteria was very similar, indicating that, despite their simplicity, liposomes are a good model to mimic the bacterial membrane. The complete membrane coverage by PMAP-23 necessary to exert its antibacterial activity might reasonably induce other effects in the bilayer together with the formation of pores. Possible effects of peptide/membrane association on membrane dynamics, caused by the addition of two peptides that form pores according to the carpet (PMAP-23) and barrel-stave (alamethicin) models were examined, by investigating variations in bilayer dynamics, water penetration and lipid lateral diffusion. Both peptides induced an increase in the rigidity of the bilayer. While for PMAP-23 the membrane stiffening occurred in the same concentration range of pore formation, in the case of alamethicin the effects were significant only at peptide concentrations much higher than those causing membrane leakage. The different behaviour of the two peptides might represent a valid criterion to discriminate between the main mechanisms of membrane permeabilization (carpet and barrelstave). A commonly used test to discriminate between different possible pore formation models is the dependence of the peptide membranepermeabilizing activity on the bilayer thickness or on the intrinsic curvature of the lipids composing the membrane. However, the experimental data supporting this criterion are limited. Here, we tested the effect of lipid properties on the activity of three AMPs whose mechanism of pore formation is well established. Our data indicate that AMP behaviour is more complex than what is commonly assumed, and that therefore a thorough characterization of peptide/membrane interaction is necessary to define the mechanism of membrane perturbation. Overall, the results of this thesis suggest that the effects of AMPs on membranes are not limited to pore formation, and that the bactericidal activity might not be the most important function of these peptides in vivo. In order to reach the high membrane coverage needed for bacterial killing, a peptide concentration at least in the low micromolar range is needed, irrespective of the density of bacterial cells. This concentration might be difficult to attain in vivo, and other effects, such as immunomodulation, might predominate. In addition, when the membrane is covered by peptide molecules, pore formation is not the only effect. The dynamics of the lipids in the bilayer is significantly hindered, and this might lead to inhibition of the function of membrane proteins, contributing to the bactericidal activity of the peptides. These aspects should be considered in the rational design and in the development of new antibiotic molecules based on AMPs, and all the different facets of AMP activity should be assayed when screening for the best drug candidates.