Engineering plant membranes using droplet interface bilayers

Abstract

Droplet interface bilayers (DIBs) have become widely recognised as a robust platform for constructing model membranes and are emerging as a key technology for the bottom-up assembly of synthetic cell-like and tissue-like structures. DIBs are formed when lipid-monolayer coated water droplets are brought together inside a well of oil, which is excluded from the interface as the DIB forms. The unique features of the system, compared to traditional approaches (e.g., supported lipid bilayers, black lipid membranes, and liposomes), is the ability to engineer multi-layered bilayer networks by connecting multiple droplets together in 3D, and the capability to impart bilayer asymmetry freely within these droplet architectures by supplying droplets with different lipids. Yet despite these achievements, one potential limitation of the technology is that DIBs formed from biologically relevant components have not been well studied. This could limit the reach of the platform to biological systems where bilayer composition and asymmetry are understood to play a key role. Herein, we address this issue by reporting the assembly of asymmetric DIBs designed to replicate the plasma membrane compositions of three different plant species; Arabidopsis thaliana, tobacco, and oats, by engineering vesicles with different amounts of plant phospholipids, sterols and cerebrosides for the first time. We show that vesicles made from our plant lipid formulations are stable and can be used to assemble asymmetric plant DIBs. We verify this using a bilayer permeation assay, from which we extract values for absolute effective bilayer permeation and bilayer stability. Our results confirm that stable DIBs can be assembled from our plant membrane mimics and could lead to new approaches for assembling model systems to study membrane translocation and to screen new agrochemicals in plants.

FIG. 1.

Schematic diagram showing the assembly of droplet interface bilayers in different operating modes. (a) In the lipid-out approach, lipids are added directly to the oil, whereas in (b) the lipid-in approach, lipid vesicles are supplied to the water droplets. (c) Asymmetric DIBs can be engineered by supplying droplets with different vesicles as illustrated. In each case, the lipids supplied to the system assemble to form a monolayer around each droplet and a DIB is formed when the droplets are manipulated into contact. The number of lipid species (indicated by the different colors) in part (c) is to reflect the complex compositions of our plant lipid preparations.

FIG. 2.

Calcein leakage assay of vesicles engineered from (a) SPLE, (b) PSC-1 (c) PSC-2, and (d) PSC-3 loaded with 50 mM calcein and monitored for 3 hours on a 96 well plate using a fluorescence spectrophotometer. The data shows that no calcein leakage was observed until detergent was added to the system (denoted by the asterisk), indicating that our plant lipid vesicles were both formed successfully and were stable for the duration of our experiments.

FIG. 3.

Brightfield micrographs of symmetric and asymmetric DIBs formed from plant lipids. (a)–(d) Symmetric DIBs formed between droplets containing vesicles of (a) SPLE (b) PSC-1, (c) PSC-2, and (d) PSC-3 with interfacial areas of 86, 65, 87, and 53 μm² respectively. (e)–(h) Asymmetric DIBs formed between a droplet containing DOPC vesicles (loaded with calcein) and a second droplet containing (e) SPLE, (f) PSC-1, (g) PSC-2, and (h) PSC-3 vesicles with interfacial areas of 68, 87, 88, and 68 μm². (i) DIB network assembled from one DOPC droplet (left) and three droplets containing SPLE vesicles with interfacial areas of 34, 48, and 48 μm², respectively. All images were obtained 5 min after the droplets were placed into contact. Scale bars = 200 μm.

FIG. 4.

Fluorescence assay investigating bilayer permeability following the formation of asymmetric DIBs from plant lipids. DIBs were formed between a source droplet containing DOPC vesicles and resorufin and a sink droplet containing vesicles made with (a) DOPC (n = 17), (b) SPLE (n = 8), (c) PSC-1 (n = 5), (d) PSC-2 (n = 16), (e) PSC-3 (n = 9), and (f) control (n = 6). The normalized fluorescence intensity is shown in the source (black squares) and sink (red circles), along with their standard deviations. The data show resorufin diffusing out of the source droplet into the sink.

FIG. 5.

(a) Measurements of effective bilayer permeability values of resorufin across symmetric and asymmetric membranes composed of DOPC and the indicated lipid formulation. The addition of cerebrosides appears to decrease the effective bilayer permeability. (b) Bilayer survival rates of symmetric and asymmetric DIBs composed of DOPC and the indicated lipid composition.