The compartmentalization of cell-free gene expression systems in liposomes provides an attractive route to the formation of protocells, but these models do not capture the physical (crowded) environment found in living systems. Here, we present a microfluidics-based route to produce monodisperse liposomes that can shrink almost 3 orders of magnitude without compromising their stability. We demonstrate that our strategy is compatible with cell-free gene expression and show increased protein production rates in crowded liposome protocells.
The compartmentalization of cell-free gene expression systems in liposomes provides an attractive route to the formation of protocells, but these models do not capture the physical (crowded) environment found in living systems. Here, we present a microfluidics-based route to produce monodisperse liposomes that can shrink almost 3 orders of magnitude without compromising their stability. We demonstrate that our strategy is compatible with cell-free gene expression and show increased protein production rates in crowded liposome protocells.
In recent years, there has been
a significant research effort to produce synthetic cell-like compartments
including liposomes,[1] fatty acid vesicles,[2] polymersomes,[3] proteinosomes,[4] water-in-oil droplets[5] and coacervate droplets[6] as protocells
to study key aspects of living systems such as compartmentalization,[5,7] replication,[8] or metabolism.[6b,9] Liposomes have been used to produce protocells for studying in vitro transcription and translation (IVTT),[1a,1b] cell division[8b,10] and spatially confined catalysis.[11] However, little progress has been made in generating
protocells with adaptive performances: tunable volumes and surface
areas in response to environmental changes.Cells regulate their
surface area and volume through membrane folding,
invagination, or vesicle fusion and fission.[12] Recent studies have demonstrated that fusion of micelles into protocell
membranes,[13] or in situ generation of lipids inside membranes,[14] can lead to slow membrane growth, and lipid bilayers can generate
lipid tubes or vesicles to balance hypertonic osmotic pressure.[12b,15] However, there is no method to robustly swell and shrink liposome
protocells without compromising their stability. Here, we propose
a new route to regulate protocell volumes via incorporation of artificial
oil-based organelles–lipid droplets (LDs). LDs are present
in some cells as oil droplets with a lipid core surrounded by a phospholipid
monolayer, and store lipids for energy metabolism and membrane synthesis.[16] Here. we designed an oil-based lipid reservoir
attached to liposomes to regulate the volume of protocells. The artificial
LD allows collection or supply of lipids from/to the bilayer membrane
as osmotic pressure fluctuates, resulting in shrinking or swelling
of the protocells, opening up opportunities to study dynamic cellular
features in vitro.We used microfluidic emulsification
to produce monodispersewater-in-oil-in-water
(W/O/W) double emulsion droplets as templates (Figures a, S1 in Supporting Information (SI)), which subsequently undergo partial dewetting (Figure a). The dewetting process is
determined by the spreading coefficient defined as S = γ – (γ + γ), where γ is
the interfacial tension between fluids i and j.[17] Only if Sw1 < 0, So < 0 and Sw2 < 0, will the emulsion templates form
a stable partial dewetting configuration (see SI Experimental Section for details of the typically used
inner, middle and outer fluids).[17a] The
success of partial dewetting relies on careful control of the concentration
of surfactant F-68 which tunes the interfacial energies. Without addition
of F-68, the emulsion templates will keep the core–shell structure
due to a positive So (Figures a,b, first panel).[1b] When F-68 is added into W2, γw1w2 and γow2 will decrease and So becomes negative, triggering the dewetting process. A lipid
bilayer is formed via combining two lipid monolayers at the two water–oil
interfaces (Figures a and S1). Higher concentrations of F-68
promote the formation of more bilayer until an intact liposome is
generated (C(F-68) > 0.2 wt
%),[1b] as F-68 in W2 reduces γw1w2 and γow2, which make the adhesion energy
ΔF (ΔF = γow2 + γw1w2 – γow1)[18] smaller, resulting in less contact
area between oil droplet and
liposome. Other combinations of liquids and surfactants reported in
the literature enable tuning of the structures of double emulsion
droplets[17,20] and these also appear promising for making
vesicles by the partial dewetting method.
Figure 1
(a) Schematics and (b)
confocal images of diverse configurations
of protocells prepared from partial dewetting of W/O/W emulsion templates
with 0.00–0.20 wt % F-68 in the W2 phase. (c) Reconstructed
confocal image showing the 3D structure of protocells. (d) As-formed
model protocells with different-sized oil organelles (in green). Scale
bars, 100 μm.
(a) Schematics and (b)
confocal images of diverse configurations
of protocells prepared from partial dewetting of W/O/W emulsion templates
with 0.00–0.20 wt % F-68 in the W2 phase. (c) Reconstructed
confocal image showing the 3D structure of protocells. (d) As-formed
model protocells with different-sized oil organelles (in green). Scale
bars, 100 μm.Our approach yields excellent
control over the liposomal structures
by adjusting concentrations of F-68 or the flow rates (Figure b,c). As the concentrations
of F-68 increase from 0%, 0.01%, 0.02%, 0.075% to 0.20%, the bilayer
area gradually increases from 0%, 41%, 49%, 72% to 84% of the surface
area of inner droplets (D = 69 μm), respectively.
Moreover, the sizes of attached oil droplets can be easily tuned (Figure d), and complex structures
with multiple compartments (so-called multisomes[19]) were also easily prepared in our method (Figure S2). Importantly, the as-formed liposomes are stable
(Figures S3b,c), with no obvious loss in
numbers after storage for 4 days (Figure S4).Next, we studied the swelling and shrinking of the liposomes
(Figure ). As Figure a–c shows,
when the
protocells consisting of 0.05 wt % PEG were dispersed in a hypertonic
solution of 2 M sucrose, they rapidly lost water and shrunk to balance
the osmotic difference (Movies S1, S2).
We did not observe any lipid spots, vesicles or tubule formation in
the bilayers during the shrinking process, indicating the extra lipids
from the bilayers were collected into the attached oil droplets. Remarkably,
the membrane surface and volume are only 1/77 and 1/670, respectively,
of their initial states (Figure S5, Movie S2). In contrast, liposomes without LDs collapsed and burst immediately
under the same conditions (Figure S6).
Inversely, LDs also allow the supply of lipids to the bilayers to
induce membrane growth (Figure d,e), because a negative pressure outside the liposomes reverses
the water flux, leading to the growth of surface area and volume.
The shrinking process is reversible, as demonstrated in Figure S7. Liposomes with an inner phase of 1.7
wt % PEG and 170 mM sucrose were shrunk from 72 to 47 μm in
750 mM sucrose solution with 0.05 wt % F-68 added. Subsequently, an
aqueous solution of 0.05 wt % F-68 was carefully added and the liposomes
swelled to 60 μm.
Figure 2
(a–c) Schematics, confocal and optical
images of the shrinking
process of protocells in response to hypertonic shock. (d,e) Schematics
and confocal images of the swelling process of protocells when in
hypotonic solution. Scale bars, 100 μm.
(a–c) Schematics, confocal and optical
images of the shrinking
process of protocells in response to hypertonic shock. (d,e) Schematics
and confocal images of the swelling process of protocells when in
hypotonic solution. Scale bars, 100 μm.To verify the lipid exchange between the bilayers and the
LDs,
we performed fluorescence recovery after photobleaching (FRAP) experiments
via labeling the bilayers and the inner water droplets with two different
fluorophores (Figure a, see SI for details). As Figure b,c shows, the fluorescence
of the whole bilayers recovered gradually in about 2 min after photobleaching
(Movie S3), but the fluorescence in the
interior did not recover due to shortage of dye supply (Figure S9). This experiment directly demonstrates
the successful lipid exchange between the bilayers and the attached
LDs.
Figure 3
(a) Schematics of the FRAP experiments. (b) Confocal images of
the recovery of fluorescence of bilayers after photobleaching. Insets
in panel b showing the fluorescence recovery in the inner water droplet.
(c) Fluorescence intensity–distance profiles along the line
in panel b. Scale bars, 50 μm.
(a) Schematics of the FRAP experiments. (b) Confocal images of
the recovery of fluorescence of bilayers after photobleaching. Insets
in panel b showing the fluorescence recovery in the inner water droplet.
(c) Fluorescence intensity–distance profiles along the line
in panel b. Scale bars, 50 μm.Cells are densely packed with macromolecules (total macromolecule
concentrations in excess of 300 g·L–1 in E. coli),[21] which influences
biochemical kinetics.[22] However, no method
enables the production of liposomes with levels of crowding found
in cells, because high concentrations of macromolecules are too viscous
to encapsulate. To solve this issue and to reconstitute a realistic
cell-like internal environment, we encapsulated cell lysate (60 g·L–1), nucleoids into E. colilipid liposomes of 88 μm in diameter (see SI for details). We then shrunk them to 54 μm in diameter
to form protocells with a concentration of macromolecules at about
260 g·L–1 (Figure a,b). To illustrate the dense interior, FRAP
experiments were performed to probe the diffusion of enhanced green
fluorescent protein (eGFP) encapsulated. As Figure c shows, the fluorescence recovers within
1 s in liposomes before shrinking, while it takes more than 40 s to
recover after shrinking, which demonstrates the crowded interior of
the protocells.
Figure 4
(a,b) Illustration and images of macromolecularly crowded
protocells.
Inset in panel b1 is the sample before shrinking. (c) Fluorescence
recovery of eGFP in liposomes before (upper sequence) and after (lower
sequence) shrinking after photobleaching. (d) Expression profiles
of mRFP in normal and shrunk liposomes and confocal images of liposomes
after expression for 6 h. Scale bars: 50 μm in panels b, c;
100 μm in panel d.
(a,b) Illustration and images of macromolecularly crowded
protocells.
Inset in panel b1 is the sample before shrinking. (c) Fluorescence
recovery of eGFP in liposomes before (upper sequence) and after (lower
sequence) shrinking after photobleaching. (d) Expression profiles
of mRFP in normal and shrunk liposomes and confocal images of liposomes
after expression for 6 h. Scale bars: 50 μm in panels b, c;
100 μm in panel d.We then performed IVTT in both crowded and noncrowded protocells
to investigate the influence of crowded interiors on gene expression.
We encapsulated a mix of cell lysate, feeding buffers and plasmids
coding for monomeric red fluorescent protein (mRFP) (total concentration
is about 40 g L–1) into l-α-phosphatidylcholine
(eggPC) liposomes (see SI Experimental
Section for details), then collected them into two containers (one
with hypertonic solution, the other without) to form crowded protocells
(diameter: 48 μm) and noncrowded protocells (diameter: 91 μm)
(Figure d). In crowded
protocells, the concentration of the interior solution increases approximately
6.8 times, yielding a concentration of IVTT mix of about 272 g L–1. The expression of mRFP in shrunk liposomes is notably
enhanced compared to expression in normal liposomes which is very
slow and barely detectable after 6 h (Movie S4). We postulate that the rate enhancement in gene expression is not
only due to increased concentration of key components such as DNA
or ribosomes in the IVTT mixture but also because of the molecularly
crowded interior.[5,6c] The slightly increased ratio
of the surface to volume of the bilayer membrane is not expected to
alter gene expression significantly.[23]To extend the technological scope for constructing protocells with
more synthetic complexity, we induced complex coacervation of the
cell lysates to create subcompartments in protocells. Complex coacervation
is a form of liquid–liquid phase separation of oppositely charged
polyelectrolytes, and provides powerful means of membrane-free compartmentalization.
Coacervation has been explored extensively in protocell models for
the construction of artificial cells or organelles.[6,7] The
phase separation of cell lysates to form crowded coacervates has been
accomplished in water-in-oil droplets recently,[6c] but it has not been demonstrated in biological vesicles
because of the use of concentrated salt solution (as high as 6 M)
and fragile nature of vesicles. To address this problem, we shrunk
the liposomes containing cell lysate, feeding buffers and 8 g·L–1 PEG via multistep osmotic shocks (Figure a, see SI Experimental Section for details). Meanwhile, plasmids
coding for eGFP were also encapsulated into the protocells to perform
IVTT. As Figures b,c
and S10 show, shrinking the volume of liposomes
induced coacervate droplet formation with cell lysate and PEG in the
liposomes (Movie S5), due to the phase
transition of salt and PEG as well as partitioning of cell lysate
into the PEG phase. Notably, the expressed eGFP also prefers to partition
into the innermost coacervate droplet (bright core in Figure b).[6c]
Figure 5
(a,b)
Coacervate formation in liposomes induced by decrease of
volume. (c) Optical images of liquid–liquid phase separation
process in protocells. Scale bars, 50 μm.
(a,b)
Coacervate formation in liposomes induced by decrease of
volume. (c) Optical images of liquid–liquid phase separation
process in protocells. Scale bars, 50 μm.In summary, we have presented a novel design to regulate
membrane
area and volume of microfluidically prepared liposomes by exploiting
the oil droplet as a reservoir which collects or supplies lipids from/to
the bilayer membrane during shrinking or swelling. We demonstrated
cell-free gene expression in cell-like conditions inside shrunk liposomes.
Control over liposome volume may find use in research as diverse as
maintaining artificial intracellular conditions for homeostasis,[24] tuning biochemical reaction rates on the basis
of membrane curvature,[25] preparing monodisperse
sub-micrometer-sized vesicles (Figure S11), and performing protein crystallization and growth.[26]
Authors: Chiara Martino; Shin-Hyun Kim; Louise Horsfall; Alireza Abbaspourrad; Susan J Rosser; Jonathan Cooper; David A Weitz Journal: Angew Chem Int Ed Engl Date: 2012-05-29 Impact factor: 15.336
Authors: Michael D Hardy; Jun Yang; Jangir Selimkhanov; Christian M Cole; Lev S Tsimring; Neal K Devaraj Journal: Proc Natl Acad Sci U S A Date: 2015-06-22 Impact factor: 11.205
Authors: Alexander F Mason; N Amy Yewdall; Pascal L W Welzen; Jingxin Shao; Marleen van Stevendaal; Jan C M van Hest; David S Williams; Loai K E A Abdelmohsen Journal: ACS Cent Sci Date: 2019-07-03 Impact factor: 14.553
Authors: Xuejing Wang; Liangfei Tian; Hang Du; Mei Li; Wei Mu; Bruce W Drinkwater; Xiaojun Han; Stephen Mann Journal: Chem Sci Date: 2019-09-18 Impact factor: 9.825
Authors: Luis P B Guerzoni; André V C de Goes; Milara Kalacheva; Jakub Haduła; Matthias Mork; Laura De Laporte; Arnold J Boersma Journal: Adv Sci (Weinh) Date: 2022-07-29 Impact factor: 17.521
Authors: Mahesh A Vibhute; Mark H Schaap; Roel J M Maas; Frank H T Nelissen; Evan Spruijt; Hans A Heus; Maike M K Hansen; Wilhelm T S Huck Journal: ACS Synth Biol Date: 2020-10-05 Impact factor: 5.110