Assessing the mechanism of facilitated proton transport across GUVs trapped in a microfluidic device

Abstract

Proton transport across lipid membranes is one of the most fundamental reactions that make up living organisms. In vitro, however, the study of proton transport reactions can be very challenging due to limitations imposed by proton concentrations, compartment size, and unstirred layers as well as buffer exchange and buffer capacity. In this study, we have developed a proton permeation assay based on the microfluidic trapping of giant vesicles enclosing the pH-sensitive dye pyranine to address some of these challenges. Time-resolved fluorescence imaging upon a rapid pH shift enabled us to investigate the facilitated H⁺ permeation mediated by either a channel or a carrier. Specifically, we compared the proton transport rates as a function of different proton gradients of the channel gramicidin D and the proton carrier carbonyl cyanide-m-chlorophenyl hydrazone. Our results demonstrate the efficacy of the assay in monitoring proton transport reactions and distinguishing between a channel-like and a carrier-like mechanism. This groundbreaking result enabled us to elucidate the enigmatic mode of the proton permeation mechanism of the recently discovered natural fibupeptide lugdunin.

Figure 1

Schematic illustration of the proton translocation assay. GUVs are obtained by electroformation on ITO-covered glass slides with or without the channel peptide gramicidin. Gramicidin-doped GUVs were directly exposed to a pH gradient, and proton influx $\left({\text{pH}}_{\text{out}}<{\text{pH}}_{\text{in}}\right)$ led to pyranine protonation and fluorescence quenching (left). GUVs for the carrier-based assay were first incubated with CCCP and valinomycin and then exposed to the pH gradient (middle). For lugdunin, GUVs were incubated with the cyclopeptide and then exposed to the pH gradient (right).

Figure 2

Setup of the microfluidic device for the trapping of GUVs. (A) Schematic design of the microfluidic channels comprising 12 lines with 17 traps each. (B) Confocal micrographs of pyranine-filled GUVs in the traps (1) and with (2) side posts. Lipid composition: POPC/TxR (99.6/0.4, $n/n$).

Figure 3

Normalized fluorescence intensities I of encapsulated pyranine at different pH values. Fluorescence intensities were normalized to that measured at pH 7.4. Eq. (1) was fit to the data and used to calculate changes in luminal pH values. Each data point is the mean $\pm$ standard deviation for $n\ge 30$ GUVs. Two exemplary epifluorescence micrographs of individual traps at pH 5.5 and 9.6 are shown below.

Figure 4

Proton transport properties of gramicidin and CCCP. (A) Fluorescence micrographs obtained at different time points of the proton permeation assay ($\Delta c\left({\mathrm{H}}^{+}\right)=$ 0.079 μmol L^–1) (A1) without and (A2) with gramicidin (peptide/lipid ratio 1:750). Lipid composition: POPC/TxR (99.6/0.4, $n/n$). (B) Time-dependent change in pH ($n=6$; mean: black solid line; gray area: standard deviation) mediated by gramicidin (peptide/lipid ratio 1:750, $n/n$) at $\Delta c\left({\mathrm{H}}^{+}\right)=$ 0.079 μmol L^–1. In the absence of a proton transporter, no significant pH change was observed (dashed line, $\Delta c\left({\mathrm{H}}^{+}\right)=$ 0.584 μmol L^–1). From the slope (green solid line), the rate of pH change is extracted. Perfusing pH was 6.9 and 6.2, respectively. (C) Rate of pH change dependent on the proton concentration gradient across the membrane for the channel gramicidin (black, linear fit) and the proton carrier CCCP (green, Eq. (3)). Each data point represents the mean $\pm$ standard deviation of $n\ge 3$ experiments. Each experimental condition encompasses $90\le n\le 416$ individual GUVs (Table S2).

Figure 5

Membrane channel properties of lugdunin. (A) Structure of the fibupeptide lugdunin. (B) Schematic representation of the possible ion channel activity of lugdunin by the assembly into a membrane-spanning nanotube. (C) Rate of pH change dependent on the proton concentration gradient across the vesicular membrane showing a distinct channel-like behavior of lugdunin (linear fit). Each data point represents the mean $\pm$ standard deviation for $n\ge 3$ experiments. Each experimental condition encompasses $120\le n\le 289$ individual GUVs (Table S2).