Gram-Positive Bacterial Membrane-Based Biosensor for Multimodal Investigation of Membrane-Antibiotic Interactions

Abstract

As membrane-mediated antibiotic resistance continues to evolve in Gram-positive bacteria, the development of new approaches to elucidate the membrane properties involved in antibiotic resistance has become critical. Membrane vesicles (MVs) secreted by the cytoplasmic membrane of Gram-positive bacteria contain native components, preserving lipid and protein diversity, nucleic acids, and sometimes virulence factors. Thus, MV-derived membrane platforms present a great model for Gram-positive bacterial membranes. In this work, we report the development of a planar bacterial cytoplasmic membrane-based biosensor using MVs isolated from the Bacillus subtilis WT strain that can be coated on multiple surface types such as glass, quartz crystals, and polymeric electrodes, fostering the multimodal assessment of drug-membrane interactions. Retention of native membrane components such as lipoteichoic acids, lipids, and proteins is verified. This biosensor replicates known interaction patterns of the antimicrobial compound, daptomycin, with the Gram-positive bacterial membrane, establishing the applicability of this platform for carrying out biophysical characterization of the interactions of membrane-acting antibiotic compounds with the bacterial cytoplasmic membrane. We report changes in membrane viscoelasticity and permeability that correspond to partial membrane disruption when calcium ions are present with daptomycin but not when these ions are chelated. This biomembrane biosensing platform enables an assessment of membrane biophysical characteristics during exposure to antibiotic drug candidates to aid in identifying compounds that target membrane disruption as a mechanism of action.

Keywords: Gram-positive bacteria; antibiotic sensing; daptomycin; membrane vesicles; microelectrode array; organic electronic; permeability; supported lipid bilayer.

Figure 1

(a) TEM image of negatively stained MVs isolated from B. subtilis shows intact vesicles indicated by black arrows. Scale bar represents 500 nm. (b) Dynamic light scattering (DLS) results for vesicle size distribution. Error bars represent standard deviation.

Figure 2

Changes in Δf and ΔD with time during bilayer formation using QCM-D. (a,b) SLB using B. subtilis MVs was formed by (1) flowing in vesicles for adsorption, (2) rinsing excess vesicles with buffer, (3) adding rupture vesicles, and (4) washing away excess material with buffer upon stabilization of signal. (c,d) The POPC–PEG SLB was formed by (1) flowing in vesicles and (2) washing away excess vesicles with buffer after completion of rupture indicated by signal stabilization. Different colors represent different overtones: orange (3rd = 15 MHz), grey (5th = 25 MHz), yellow (7th = 35 MHz), light blue (9th = 45 MHz), green (11th = 55 MHz), and dark blue (13th = 65 MHz).

Figure 2

Changes in Δf and ΔD with time during bilayer formation using QCM-D. (a,b) SLB using B. subtilis MVs was formed by (1) flowing in vesicles for adsorption, (2) rinsing excess vesicles with buffer, (3) adding rupture vesicles, and (4) washing away excess material with buffer upon stabilization of signal. (c,d) The POPC–PEG SLB was formed by (1) flowing in vesicles and (2) washing away excess vesicles with buffer after completion of rupture indicated by signal stabilization. Different colors represent different overtones: orange (3rd = 15 MHz), grey (5th = 25 MHz), yellow (7th = 35 MHz), light blue (9th = 45 MHz), green (11th = 55 MHz), and dark blue (13th = 65 MHz).

Figure 3

Membrane protein and LTA retention in Gram-positive SLBs. (a) TIRFM images for Alexa Fluor 594 succinimidyl ester binding to POPC–PEG and Gram-positive bilayers. Higher fluorescence is indicative of the presence of primary amines found in proteins. (b) Mean fluorescence intensity from TIRFM images for primary amine binding. (c) TIRFM images for anti-LTA antibody binding to POPC–PEG and Gram-positive bilayers. (d) Mean fluorescence intensity from TIRFM images for antibody binding. For quantification, images (n ≥ 8) obtained using TIRFM were analyzed via ImageJ. Error bars represent standard deviation. Scale bars represent 20 μm.

Figure 4

Graphical representations of the (a) one-layer Kelvin–Voigt model for the POPC–PEG bilayer (grey) and (b) two-layer Voigt–Voinova model for the Gram-positive bilayer. Note that the LTA (green) is not drawn to scale with the bilayer (purple) thickness.

Figure 5

EIS monitoring of SLB interactions with daptomycin. (a) Schematic representation of SLBs formed on PEDOT:PSS electrodes, with an equivalent circuit used for fitting impedance data. (b–d) Representative Bode plots showing impedance responses of SLBs upon addition of daptomycin (DAP) for (b) Gram-positive SLB in the presence of Ca²⁺, (c) Gram-positive SLB in the absence of Ca²⁺, and (d) POPC SLB in the presence of Ca²⁺.

Figure 6

Daptomycin interaction with Gram-positive SLB using QCM-D in the presence of Ca²⁺. The SLB is washed with buffer supplemented with Ca²⁺ before daptomycin addition to isolate the impact of the antibiotic. Representative plots monitor (a) Δf and (b) ΔD, with time for the process. The time of daptomycin addition is marked with green arrows. Upon daptomycin addition, a drop in Δf indicating insertion of the antibiotic into the bilayer and a rise in ΔD indicating destabilization of the bilayer, were recorded. Values are normalized post-SLB formation. Different colors represent different overtones: orange (3rd = 15 MHz), grey (5th = 25 MHz), yellow (7th = 35 MHz), light blue (9th = 45 MHz), green (11th = 55 MHz), and dark blue (13th = 65 MHz).

Figure 7

Impact of daptomycin interaction on membrane viscoelastic properties in the presence of calcium ions. Grey bars correspond to the POPC–PEG bilayer, purple bars correspond to the Gram-positive lipid bilayer, and green bars correspond to the polymeric chain component of LTA extending beyond the bilayer. Changes in modeled parameters, e.g., thickness, viscosity, and shear modulus, after antibiotic addition, are presented as percentage changes normalized with respect to the values for each parameter before the antibiotic was added. Error bars represent standard error from 3 independent experiments.