EcDBS1R4, an Antimicrobial Peptide Effective against Escherichia coli with In Vitro Fusogenic Ability

Abstract

Discovering antibiotic molecules able to hold the growing spread of antimicrobial resistance is one of the most urgent endeavors that public health must tackle. The case of Gram-negative bacterial pathogens is of special concern, as they are intrinsically resistant to many antibiotics, due to an outer membrane that constitutes an effective permeability barrier. Antimicrobial peptides (AMPs) have been pointed out as potential alternatives to conventional antibiotics, as their main mechanism of action is membrane disruption, arguably less prone to elicit resistance in pathogens. Here, we investigate the in vitro activity and selectivity of EcDBS1R4, a bioinspired AMP. To this purpose, we have used bacterial cells and model membrane systems mimicking both the inner and the outer membranes of Escherichia coli, and a variety of optical spectroscopic methodologies. EcDBS1R4 is effective against the Gram-negative E. coli, ineffective against the Gram-positive Staphylococcus aureus and noncytotoxic for human cells. EcDBS1R4 does not form stable pores in E. coli, as the peptide does not dissipate its membrane potential, suggesting an unusual mechanism of action. Interestingly, EcDBS1R4 promotes a hemi-fusion of vesicles mimicking the inner membrane of E. coli. This fusogenic ability of EcDBS1R4 requires the presence of phospholipids with a negative curvature and a negative charge. This finding suggests that EcDBS1R4 promotes a large lipid spatial reorganization able to reshape membrane curvature, with interesting biological implications herein discussed.

Keywords: EcDBS1R4; Escherichia coli; Gram-negative bacteria; antimicrobial peptide; cardiolipin; hemifusion; hyperpolarization.

Figure 1

E. coli cell viability kinetics after EcDBS1R4 treatment. (A) Bacteria cell viability followed for 1 h by flow cytometry as a function of peptide concentration. (B) Time kinetics of cell viability after peptide treatment. Each point represents the time necessary to produce 50% of death for each concentration tested. Dashed lines indicate 95% confidence intervals.

Figure 2

In silico and experimental structural studies of EcDBS1R4 structure. (A) Amino acid sequence of EcDBS1R4 and theoretical three-dimensional structure predicted by homology and constructed using Modeller v. 9.12. (B) α-helical wheel (constructed using the server http://rzlab.ucr.edu/scripts/wheel/wheel.cgi; hydrophobic residues are represented as green diamonds, hydrophilic, non-charged residues are coded as orange circles, and positively charged residues are blue-grey pentagons) and adaptive Poisson-Boltzmann solver electrostatic potential of EcDBS1R4, ranging from −5.0 k_BT/e (red) to +5.0 k_BT/e (blue). (C,D) Experimental structural studies of EcDBS1R4 in the presence of different LUV compositions. (C) Circular dichroism spectra of EcDBS1R4 in solution (dotted black) at 16 µM, and in the presence of 500 µM of lipid vesicles of POPC (black), POPC:Chol (70:30) (blue), POPC:POPG (70:30) (red), inner (green) and outer (orange) membrane of E. coli mimetic systems. (D) Comparative plot of the θ signal at 222 nm (a local minimum for α-helices) at different lipid concentrations, with the same lipid composition described above. Solid lines represent fits to the experimental data using Equation (1). Each experiment was conducted in triplicate and represented as mean ± standard deviation (SD). Calculated parameters are presented in Table 2.

Figure 3

Experimental and in silico studies of the membrane selectivity of EcDBS1R4. (A) Partition curves of 6 µM EcDBS1R4 to LUVs of different compositions. Solid lines represent fits obtained using Equations (2) and (3). The LUV compositions used were POPC (black), POPC:Chol 70:30 (blue), POPC:POPG 70:30 (red), IML (POPE:POPG:CL 63:33:4) (green) and OML (POPE:POPG:CL:LPS 80:16:1:3) (orange). (B) Fluorescence quenching by acrylamide of 6 µM EcDBS1R4 in the absence (open circles) and presence of LUVs (color code as in A). Solid lines represent fits obtained using Equations (4) and (5). (C) Changes in membrane dipole potential as a function of EcDBS1R4 concentration, measured with di-8-ANEPPS (4-(2- [6-(dioctylamino)-2-naphthalenyl]ethenyl)-1-(3-sulfopropyl)pyridinium inner salt) in LUVs (200 µM lipid concentration; color code as in A) and E. coli cells (1 × 10⁴ cells/mL), represented by grey filled circles. The plot represents the di-8-ANEPPS excitation ratio R (I_455nm/I_525nm) for each peptide concentration, normalized divideing by R₀ (the R value in the absence of peptide). Solid lines represent fits obtained using Equation (6). The parameters obtained from the fittings are summarized in Table 2. All experiments were conducted in triplicate. (D–F) Three-dimensional theoretical representation of the peptide interacting with membranes composed of POPC (D), POPC:Chol (70:30) (E) and POPC:POPG (70:30) (F), indicating the amino acid residues (in green) and the phospholipid molecules (in white) possibly involved in the interactions (residues involved and distances of interactions occurring for each membrane are detailed in Supporting Information Table 1).

Figure 4

Alterations in membrane properties induced by EcDBS1R4. (A) Effect of EcDBS1R4 on the transmembrane potential of E. coli cells, calculated from the DiOC₂(3) green shift. (B) Generalized polarization of LUVs (3 mM lipid) and E. coli cells (1 × 10⁴ cells/mL) labelled with Laurdan, as a function of peptide concentration. (C) Zeta-potential of LUVs (200 µM lipid) and E. coli cells as a function of peptide concentration. (D) Hydrodynamic diameter of LUVs (200 µM lipid), measured by dynamic light scattering as a function of EcDBS1R4 concentration. (E) Fusion/hemifusion efficiency of IML (PE:PG:CL (63:33:4)) LUVs (200 µM lipid) as a function of peptide concentration, calculated using Equation (8). The LUV compositions used were POPC (black), POPC:Chol 70:30 (blue), POPC:POPG 70:30 (red), IML (POPE:POPG:CL 63:33:4) (green) and OML (POPE:POPG:CL:LPS 80:16:1:3) (orange). E. coli cells are represented by grey open circles. All experiments were conducted in triplicate. Data are presented as mean ± SD.