Activity modulation of the Escherichia coli F1FO ATP synthase by a designed antimicrobial peptide via cardiolipin sequestering

Abstract

Most antimicrobial peptides (AMPs) exert their microbicidal activity through membrane permeabilization. The designed AMP EcDBS1R4 has a cryptic mechanism of action involving the membrane hyperpolarization of Escherichia coli, suggesting that EcDBS1R4 may hinder processes involved in membrane potential dissipation. We show that EcDBS1R4 can sequester cardiolipin, a phospholipid that interacts with several respiratory complexes of E. coli. Among these, F₁F_O ATP synthase uses membrane potential to fuel ATP synthesis. We found that EcDBS1R4 can modulate the activity of ATP synthase upon partition to membranes containing cardiolipin. Molecular dynamics simulations suggest that EcDBS1R4 alters the membrane environment of the transmembrane F_O motor, impairing cardiolipin interactions with the cytoplasmic face of the peripheral stalk that binds the catalytic F₁ domain to the F_O domain. The proposed mechanism of action, targeting membrane protein function through lipid reorganization may open new venues of research on the mode of action and design of other AMPs.

Keywords: Applied microbiology; Microbiology.

Graphical abstract

Figure 1

Bilayer remodeling activity by EcDBS1R4 (A) Diffusion coefficient measurements by FRAP of a POPC (circles) and a POPC:CL (4:1) (squares) lipid supported bilayer (lipid concentration of 0.8 mM) in the absence (white) and presence (red) of 20 μM of EcDBS1R4, using 0.5% of TopFluor-CL as the fluorescent probe. ∗, p < 0.05; ∗∗, p < 0.01 by 1-way ANOVA. (B) Representative confocal image before and after 45 min of addition of EcDBS1R4 on a POPC:CL (4:1) and a POPC lipid bilayer. The scalebar of the bottom confocal micrograph applies to the confocal micrographs at the top (50 μm). (C) AFM corrected height images of the time evolution of an incubation with 20 μM of EcDBS1R4 on a PC:CL (4:1) lipid bilayer. The scalebar of the rightmost AFM micrograph applies to the rest of AFM micrographs (2 μm). (D) Left: the height profile along the white line signaled in C, at three time points; Right: height distribution histogram of the same four time points. The dashed lines in the height profile and height histogram indicate the approximate position of a lipid monolayer (fine dashes) or a lipid bilayer (coarse dashes) below and above the baseline bilayer.

Figure 2

Molecular dynamics simulations of EcDBS1R4 interacting with bicelles containing CL (A) Snapshots of the top view (top) and side view (bottom) of a POPC:CL (4:1) bicelle on its initial configuration and after 20 μs of simulation. Peptide was added to only one side of the bicelle. For clarity, only the phosphate groups of lipids (yellow: POPC, red: CL) and peptide backbone beads (green) are represented. The green arrows point to the corral borders imposed by a flat-bottomed potential, within which the peptide is forced to stay. (B) Ratio of lipid distribution inside vs. outside the “peptide corral” in a POPC:CL (4:1) and a PE:PG:CL (65:30:5) bilayer (left and right, respectively). Ratios above or below 1 mean enrichment or depletion of a lipid inside the peptide corral, indicating lipid sequestration or exclusion by the peptide. Lipid ratios at the top (PEP+) and bottom (PEP-) leaflets.

Figure 3

EcDBS1R4 preferential position across the bilayer plane assessed by electron paramagnetic spectroscopy (A) Chemical structures of the probes used, namely 5-doxyl-PC as the interfacial sensitive probe (reporting close to the bilayer-water interface) and 16-doxyl-PC, sensitive to alterations at the core of the lipid bilayer. (B) Example of EPR first-derivative absorption spectra of 2.5 mM PC:CL (4:1) + 0.75% spin label (top: 5-doxyl-PC, bottom: 16-doxyl-PC). The spectral parameters used to obtain τ_c are indicated in the 16-doxyl spectrum: h₀, is the mid-field line height; h_-1, high-field line height; ΔH₀, mid-field line width; and 2A_max maximum hyperfine splitting. (C) Plots of the hyperfine splitting (2Amax) of the 5-doxyl-PC (which is proportional to τ_c), and the rotational times (τ_c) of the 16-doxyl-PC. The τ_c plots show that only 5-doxyl-PC is significantly perturbed by EcDBS1R4, indicating a more interfacial localization of the peptide. Spectra were recorded at 298 K. Error bars indicate standard deviation.

Figure 4

ATPase activity of inner membrane vesicles (IMVs) of E. coli and F₁F₀ reconstituted vesicles with different lipid compositions, as a function of EcDBS1R4 concentration ATPase activity was measured by the colorimetric malachite assay. Lipid concentrations of IMVs (green) and F₁F_O-reconsituted vesicles was 1 mM. The lipid compositions of the reconstituted proteoliposomes were POPC (yellow), POPC:POPG:CL (65:30:5; blue), and POPE:POPG:CL (65:30:5; orange). Each data points represents the average of at least three experiments. Error bars indicate the coefficient of variation.

Figure 5

Effects of EcDBS1R4 on the lipid interactions with the F_O motor of E. coli (A) Overview of the structure of the F_O motor of E. coli’s ATP synthase (PDB ID: 6VWK). Three views are presented, from left to right: the side view focusing on the c-ring; the top view from the cytoplasmic side; and the side view focusing on the a-subunit. The main subunits that made up the rotor are color-coded as follows: gray for the c-ring; purple for the a-subunit; yellow and orange for helices 1 and 2 that conform the b-subunit of the peripheral stalk; in green, the two helices perpendicular to the bilayer plane that channel the protons to the c-ring. (B) Occupancy maps of the phosphate residues of each lipid species (red: CL, blue: POPE, and black: POPG) reveal defined lipid territories in the surroundings of the F_O motor. (C) Occupancy maps of the CL phosphates in the absence of peptide (ice blue), unrestrained peptide (salmon red), and restrained peptide (PEP + (PR); orange). Red and lime dashed highlights represent loss and gain of CL in a territory that is number coded from 1 to 5 for clarity. (D) Radial distribution functions of lipids in the cytoplasmic and periplasmic leaflets as a function of the distance to the surface of the protein. The color code is shared with C.

Figure 6

Cardiolipin interactions with the subunits of the F_O motor (A) Protein-centric view of the most frequent interactions of the F_O motor with CL. The spheres indicate residues that spent at least 30% of the time bound to a CL phosphate bead. The gradient color goes from blue to red, maximum binding is 60% (red residue at b-subunit). F_O subunits are color-coded as in Figure 4A. (B) Normalized total time of CL phosphate-binding to each subunit. Per-subunit break-down of the amino acid residues that participate the most in the binding to CL: c-ring: THR 48. a-subunit: Ser58, Lys62, Leu96, Met164; b-subunit: Tyr 23 (helix 1), Tyr71 (helix 2). (C) Protein-centric viewpoint of the total time of binding of CL to the F_O motor in the absence and presence of peptide. From top to bottom: normalized bound time of the residues of the F_O motor with the phosphate residues of CL in the absence (ice blue) and presence of unrestricted (salmon red) and restricted peptide (orange); center: difference of normalized binding times between the peptide free simulation and the simulation with unrestrained peptide. Bottom: difference between peptide free and the simulation with restrained peptide. Points above 0 represent residues that interact more with CL in the peptide-free simulation than in the simulation with peptide and vice-versa. Dashed lines represent the boundaries between subunits. The background indicates the different F_O subunits: c-ring: gray; b-subunit: checkered orange-yellow; a-subunit: purple.

Figure 7

Molecular dynamics studies of the effect of EcDBS1R4 on dynamic and spatial properties of a model bioenergetic membrane A POPE:POPG:CL (65:30:5) composition was used in all simulated systems: F_O+ systems: bilayers carrying the transmembrane F_O motor of the ATP synthase of E. coli; and, F_O- systems: naked bilayers lacking the F_O motor. (A) Diffusion coefficients of the different lipids. (B) Average z-coordinate position of the phosphate beads of lipids as a function of the distance to the protein center. (C) Thicknesses of the bilayers measured by the inter-leaflet distance between phosphate beads (P–P distance). On the right, the thickness as a function of the distance to the F_O motor. All simulations were performed in triplicate. Error bars represent the 95% confidence interval.