Engineering thermoresponsive phase separated vesicles formed via emulsion phase transfer as a content-release platform

Abstract

Giant unilamellar vesicles (GUVs) are a well-established tool for the study of membrane biophysics and are increasingly used as artificial cell models and functional units in biotechnology. This trend is driven by the development of emulsion-based generation methods such as Emulsion Phase Transfer (EPT), which facilitates the encapsulation of almost any water-soluble compounds (including biomolecules) regardless of size or charge, is compatible with droplet microfluidics, and allows GUVs with asymmetric bilayers to be assembled. However, the ability to control the composition of membranes formed via EPT remains an open question; this is key as composition gives rise to an array of biophysical phenomena which can be used to add functionality to membranes. Here, we evaluate the use of GUVs constructed via this method as a platform for phase behaviour studies and take advantage of composition-dependent features to engineer thermally-responsive GUVs. For the first time, we generate ternary GUVs (DOPC/DPPC/cholesterol) using EPT, and by compensating for the lower cholesterol incorporation efficiencies, show that these possess the full range of phase behaviour displayed by electroformed GUVs. As a demonstration of the fine control afforded by this approach, we demonstrate release of dye and peptide cargo when ternary GUVs are heated through the immiscibility transition temperature, and show that release temperature can be tuned by changing vesicle composition. We show that GUVs can be individually addressed and release triggered using a laser beam. Our findings validate EPT as a suitable method for generating phase separated vesicles and provide a valuable proof-of-concept for engineering content release functionality into individually addressable vesicles, which could have a host of applications in the development of smart synthetic biosystems.

Fig. 1. Emulsion phase transfer (EPT) can be used to produce ternary GUVs with different lateral membrane structure. (A) EPT vesicles are produced by driving lipid-monolayer stabilised water-in-oil droplets (with lipid in the oil phase) through an oil–water interface stabilised by a second lipid-monolayer. As the droplet is driven through the interface under gravity, a second monolayer is deposited, forming the final vesicle. (B) Ternary electroformation DOPC/DPPC/cholesterol phase diagram highlighting the three phase behaviours observed in GUVs: (i) a single liquid phase (green), (ii) liquid–liquid domain coexistence (blue) and (iii) gel–liquid coexistence (red). The twelve compositions studied here are overlaid on the diagram, each containing a 1 : 1 molar ratio of DOPC and DPPC, with increasing mol% cholesterol. Phase diagram is based upon the Hamada schematic. The lipid composition shown are those of the initial lipid film, and not necessarily of the GUVs themselves. (C) A comparison of ternary GUV phase behaviour between EPT (left) and electroformed vesicles (right). The observed phase behaviour is shown as a function of the mol% cholesterol in the initial lipid film. Although the same phase behaviours are observed in each case, due to reduced cholesterol incorporation, increased cholesterol is needed to reach key phase changes (∼36 vs. 10%, and ∼80–90 vs. 40% respectively). (D) Selected fluorescence microscopy images of GUVs with increasing mol% cholesterol. Each composition contains 1 mol% rhodamine-PE in order to visualise domains. In all images, the dark regions correspond to the more ordered phase (L_o or gel) and the light regions correspond to the L_d phase. Scale bar = 5 μm, experiments conducted at RT.

Fig. 2. Ternary GUVs can act as a platform for user-defined, thermoresponsive content release. (A) DOPC : DPPC : Chol 1 : 1 : 3 ternary vesicles containing L_o/L_d domains can release their cargo when taken through the vesicle T_mix (∼32 °C), as shown through application of multiple heat cycles and compared to domain-free DOPC GUVs. Error bars = 1 S.D., n = 5. (B) Fluorescence and phase contrast images highlighting calcein-free ternary GUVs after applying a heat cycle. Vesicles that leaked lost contrast during phase contrast imaging (arrows) as their internal and external content equilibrated. White dotted circles represent outlines of vesicles that have leaked scale bar = 30 μm. (C) Graph of content release after 20 thermal cycles, showing ternary GUVs underwent almost complete content release. Error bars = 1 S.D., n = 5. (D) Visualisation of ternary GUVs after 20 cycles with superimposed bright-field and fluorescence channels, showing the majority of vesicles released dye cargo. Scale bar = 50 μm. (E) Image of ternary DOPC : EggSM : Chol 1 : 1 : 3 vesicles successfully formed via phase transfer, showing the same L_o/L_d domain formation. Scale bar = 50 μm. (F) Graph showing that vesicle T_mix defines the release temperature, as shown by the release of content from DOPC : DPPC : Chol (T_mix ∼ 32 °C) when cycled to 40 °C, whilst DOPC : EggSM : Chol vesicles (T_mix ∼ 45 °C) only release their content when cycled to 50 °C. Error bars = 1 S.D., n = 5.

Fig. 3. (A) Image sequence showing the heating process for a DOPC : DPPC : Chol 1 : 1 : 3 phase separated GUV using an optical trap. Once the laser was turned on the sample was locally heated and the ordered and disordered liquid domains mix into a uniform state (above the T_m ∼ 32 °C). After this point the laser was switched off to allow the GUV to cool and the domains to demix and eventually coalesce. Imaging the same vesicles in the calcein channel revealed leakage once the GUV was heated above T_m and allow to cool (∼30 seconds). Scale bars = 10 μm. (B) Graph showing average mean fluorescence of phase separated GUVs (black squares) or single phase DOPC GUVs (red circles) (diameter 7–12 μm) after successive cycling events. Error bars = 1 S.D., n = 10.