Effect of Helical Kink in Antimicrobial Peptides on Membrane Pore Formation

Abstract

Antimicrobial peptides (AMPs) can kill pathogens via the formation of permeable 14 membrane pores. However, matching peptide properties with their ability to form pores remains 15 elusive. In particular, the proline/glycine kink in helical AMPs was reported to both increase and 16 decrease antimicrobial activity. We used computer simulations and fluorescence leakage 17 experiments to show that a kink in helices affects the formation of membrane pores by stabilizing 18 toroidal pores but disrupting barrel-stave pores. The precise position of the proline/glycine kink in 19 the sequence further controls the formation of specific toroidal pore structure: U-or 20 hourglass-shaped. Moreover, we demonstrate that two helical peptides can form a stable kink-like 21 connection with similar behavior as one long helical peptide with kink. The provided 22 molecular-level insight can be utilized for rational design or modification of antibacterial peptides 23 or toxins to alter their ability to form membrane pores. 24 25 36AMPs possess an amphiphilic character with a sequence composed of both hydrophobic and 37 hydrophilic residues arranged in discrete clusters. (8) Such a distribution is thought to be the key 38 factor for binding to the pathogen's membrane and its disruption, while a common positive net 39 charge is responsible for the increased selectivity towards bacterial cells. (9; 10) AMPs are typically 40 1 of 29 Manuscript submitted to eLife unstructured in solution and frequently adopt an -helical conformation upon interaction with 41 the membrane. Their sequence can contain proline or glycine residues, which cause perturbation 42 in the regular -helical pattern. As a result, a substantial number of the known AMPs has been 43 determined to possess a helix-kink-helix motif. (11) The presence of a helical kink has been shown 44 to be biologically relevant or even responsible for the AMP's activity. (12; 13) However, the effect of 45 the kink on antimicrobial activity has been the subject of controversy over the past few decades. 46 Methodologically diverse studies have produced contradictory results, reporting the helical kink to 47 53 features of various peptides (Magainin II (30), LL-37 (31), Buforin II (32), -lysin (33), Candidalysin (34), 54 and their mutants) with their effect on pore stability. We found that the presence of a kink disrupts 55 barrel-stave, but stabilizes toroidal pores. Moreover, the position of the proline/glycine kink with 56 respect to the hydrophobic patch on AMP determined the peptide arrangement within the toroidal 57 pore. The pore formation of various mutants was further verified using a fluorescence leakage 58 assay on large unilamellar vesicles (LUVs). 59 Results and Discussion60 Contradictory Effect of -helical Kink on Barrel-stave Pore and Toroidal Pore Mech-61 anism 62Firstly, we investigated the effect of general peptide properties on the formation of a barrel-stave 63 pore (BP) and toroidal pore (TP) using MC simulations with phenomenological models (28) (see 64 Figu...

Design guidelines for α-helical peptides that self-assemble into transmembrane barrel pores killing antibiotic-resistant bacteria

“…Peptide-peptide interactions, which are necessary to stabilize TBP, compete with peptidemembrane and peptide-water/ion interactions. Due to such complex environment, the self-assembled pores have been shown to be highly sensitive to peptide amino acid composition, with a single-residue mutation capable of destabilizing the pore structure 7 . There are a few examples of de novo designed helical structures similar to TBPs (e.g., narrow ion channels or double-barrel pores) [8][9][10][11][12] and redesigned naturally-occurring channel-containing proteins [13][14][15] .…”

“…Experimental determination of residue role is challenging because it requires high spatial and temporal resolution of potentially metastable nanostructures 16 . Computer simulations can provide the required atomic resolution and add the details of interaction energies for such dynamic structures 7,17 .…”