Membrane tension-mediated growth of liposomes

(1)

Membrane

tension-mediated

growth

of

liposomes

A

step

closer

to

synthetic

cells

Sai

Sreekar

Wunnava

Venkata

Degree project inbiology, Master ofscience (2years), 2018 Examensarbete ibiologi 30 hp tillmasterexamen, 2018

Biology Education Centre, Uppsala University, and Department ofBionanoscience, Kavli Institute of Nanoscience Delft, Delft University of Technology, Building 58, Van der Maasweg 9,2629 HZ Delft, The Netherlands

(2)

ABSTRACT

Living cells are highly complex, making it an extremely challenging task to understand how they function. A possible solution is the bottom-up assembly of non-living components and building up life-like features from scratch, i.e., using synthetic cells as a tool to understand the basic characteristics of life. One such chassis for synthetic cells are liposomes, which, like the cell membrane of living cells, are made of phospholipids. As living cells grow, lipids are incorporated into their membrane in order to cope up with the volume increase of the cell. In a similar fashion, a variety of ways are currently being investigated to achieve growth of synthetic cells. Few examples include incorporation of fatty acids from the surrounding environment, reconstituting the enzymes for fatty acid or lipid biosynthesis in the liposome, or by carrying out the synthesis of artificial membrane components through the external addition of precursor molecules. Here, we demonstrate the membrane-tension mediated growth of giant unilamellar vesicles (GUVs) by fusing sub-micrometre-sized feeder vesicles to them. We use a recently developed microfluidic technique, octanol-assisted liposome assembly (OLA), to produce cell-sized (~10 µm) GUVs on-chip. Following the density-based separation of the liposomes from the waste product (1-octanol droplets), we supply small unilamellar vesicles (SUVs, ~30 nm in diameter) which act as a lipid reserve for growth by fusing with the GUVs. The lipids molecules, being very stable in bilayer conformation, require energy to reorient themselves and undergo membrane fusion. We show that increased membrane tension of GUVs can act as a sole driver to carry out multiple fusion events and cause significant growth. By placing a mass population (>1000) of GUVs in a sufficiently hypotonic solution (delta c 3−5 mM), we build up the membrane tension (~10 mN/m) driving multiple SUV-GUV fusion events, eventually doubling the volume of a part of the population. We probe a variety of lipid compositions, including hybrid (composed of lipids and fatty acids) GUVs and find the growth to be dependent on the lipid composition. Maximum growth is obtained when using a hybrid system, as compared to pure lipids. Our results show the possibility to use a protein-free minimal system to induce growth in a minimalistic manner and the demonstrated high-throughput microfluidic approach may have useful implications towards realizing an autonomous entity capable of undergoing a continuous growth-division cycle.

Keywords: synthetic cells, liposomes, growth, membrane fusion, bottom-up biology,

(4)

LIST OF ABBREVIATIONS

RNA ribonucleic acid

DNA deoxyribonucleic acid

DPPC dipalmitoyl phosphatidylcholine

DOPC dioleoyl phosphatidylcholine

SNAREs Soluble N-ethylmaleimide-sensitive factorAttachment Protein receptor

PEG Polyethylene glycol

SUV small unilamellar vesicles

GUV giant unilamellar vesicles

OLA Octanol-assisted liposome assembly

OA Outer aqueous

IA Inner aqueous

LO lipid-carrying organic

PDMS polydimethylsiloxane

PVA polyvinyl alcohol

DOPE dioleoyl phosphatidyl ethanolamine

LPC lyso phosphatidyl choline (oleoyl-hydroxy phosphatidyl choline)

PE-CF dioleoyl phosphoethanolamine-N-(carboxyfluorescein)

(5)

1 INTRODUCTION

1.1 Bottom-up assembly of synthetic cells

Life, as we know, is highly complex. Even one of the smallest and the simplest organisms,

Mycoplasma genitalium, which has 525 known genes(Razin, 1997), is complex enough to not

fully understand the functioning of these genes and the emergent properties arising from their interactions. To understand this complexity of life, one approach is to strip down the genome to just leave only the genes essential for survival. Using this top-down approach, the

Mycoplasma genome has been reconstructed to make a new cell, JCVI-syn3.0, with 438

essential protein-coding genes and 35 RNA coding genes in a 531 kbp genome(Hutchison et

al., 2016). Even with such minimal genome, for about 32% of the “essential” genes, the

function remains unknown.

An alternative way to go is a bottom-up reconstitution of a protein-based self-replicating system, from biological components, such as a reduced genome encoding for DNA, RNA and protein synthesis and encapsulating it in a membrane to form a synthetic minimal cell(Forster and Church, 2006; Jewett and Forster, 2010). Both these ways towards synthesizing life have major disadvantages of either the lack of understanding of the functions or the practical hindrances to integrating the biological subsystems which require radically different conditions for their activity. To circumvent this, one can either start assembling life completely bottom-up from simple chemical entities(Luisi, Ferri and Stano, 2006) or understand the origins of life and try to retrace the path nature took in evolving these systems to the current form(Schrum, Zhu and Szostak, 2010). However, except for the top-down approach, the bottom-up approaches would require a container/ compartment which separates the system from the environment. To this aid, most utilized container so far is a liposome, due to its resemblance with natural living cells.

1.2 Liposome as the synthetic cell container

(6)

This self-assembly is dependent on the packing parameter of the lipid molecule. Packing parameter is defined as the ratio of the volume occupied by the hydrophobic tails per molecule (v) to the product of the head group area (a) and the lipid tail length (l), i.e., 𝑣

𝑎𝑙. Commonly used

phospholipids like dipalmitoyl phosphatidylcholine (DPPC) or dioleoyl phosphatidylcholine (DOPC) have an approximately cylindrical shape with a packing parameter close to 1. They spontaneously assemble into bilayers which, to minimize the surface of the aggregate, eventually form spherical vesicles, i.e., liposomes(Salim et al., 2014). Figure 1 gives an overview of different lipid structures based on their packing parameters.

Liposomes have been used as models for cell membrane, to study the biophysical aspects of lipid bilayers as well as for studying membrane proteins. Apart from this, they have been extensively probed as drug delivery vehicles(Colley and Ryman, 1976). With their similarity to cell membranes, they have been extensively used in the bottom-up assembly of synthetic cells. In recent years, liposomes have been used as synthetic cells to demonstrate DNA replication, RNA polymerization, protein synthesis and rudimentary metabolism (Noireaux and Libchaber, 2004; Gardner, Winzer and Davis, 2009; Xu, Hu and Chen, 2016; van Nies et

al., 2018). Such synthetic cells expressing a genetic pathway for quorum sensing molecules,

have even shown to carry out two-way communication with their living counterparts and are said to have passed a cellular version of a Turing test (Lentini et al., 2017). Thus, by

Figure 1: Different self-assembled structures of lipids based on the critical packing parameter and the molecular shape. < ½ ≈ ½ − 1 > 1 Micelles Bilayer Inverted micelles a l v

(7)

encapsulating cell-free expression systems and building up complexity, liposomes have served well to mimic biological cells.

1.3 Growing synthetic cells

Though the creation of synthetic cells inches forward towards mimicking a biological system, one factor which it falls short is the ability to replicate. One of the basic characteristics of life is reproduction. For a living cell, the basic unit of life, this process of reproduction is reflected in the division process that gives rise to two or more daughter cells. However, before each division cycle, the mother cell undergoes a growth phase where it increases in mass by sequestering nutrients, increasing protein synthesis and replicating the genetic material which will be eventually distributed to the daughter cells. To keep up with the increasing mass of the cell and to have a consistent average cell size, the mother cell needs to increase in volume and correspondingly, in surface area. This is achieved by synthesizing lipids to make new plasma membrane and also the membrane for organelles, in the case of eukaryotic cells (Guertin and Sabatini, 2015). For prokaryotes, which lack membrane-bound organelles, membrane assembly takes place at the cell boundary itself, where the enzymes needed for lipid synthesis are located and the synthesised lipids are directly incorporated into the plasma membrane. In case of eukaryotic cells, lipid synthesis happens at the membrane of a specialized organelle, the endoplasmic reticulum, and the newly synthesized lipids are then transported to the plasma membrane (Lodish and Rothman, 1979).

In recent years, various research groups have been working towards growth and division of synthetic cells. Taking cues from the chemical origins of life, fatty acid vesicles, as model protocells, have been also employed in this regard.

(8)

Attempts are also being made to incorporate the lipid synthesis machinery within liposomes to achieve controlled growth of synthetic cells. Purified enzymes for lipid synthesis have been used to synthesise phospholipids in liposomes, however, without success, as no growth was observed (Schmidli, Schurtenberger and Luisi, 1991; Wick and Luisi, 1996). Also, using a cell-free expression system, the genetic pathway for specific lipids has been reconstituted in liposomes (Scott et al., 2016). However, due to the low yield of phospholipid production, no significant growth was observed. Synthetic amphiphiles and phospholipid analogues have also been incorporated in liposomes which resulted in either birthing of smaller vesicles or growth followed by shape deformations (Takakura, Toyota and Sugawara, 2003; Kurihara et al., 2011). Another way to induce membrane growth would be to add the lipids to the growing vesicles using smaller vesicles. This can be achieved by the fusion of multiple small unilamellar vesicles to a growing mother liposome.

1.4 Membrane Fusion

Membrane fusion is a ubiquitous biological process in which the membrane of two independent lipid bilayers fuses to form one continuous structure. For vesicles, which are closed assemblies, it also involves the mixing of the encapsulated contents. The process of membrane fusions is multifaceted and can be broadly divided into adhesion of the involved membranes, stalk formation, the formation of hemifusion diaphragm, and subsequent rupture of the hemifusion diaphragm to form the fusion pore (Chizmadzhev, 2012). Figure 2 gives a schematic representation of the different stages of fusion.

(9)

Membrane fusion plays a crucial role in biological systems. As mentioned earlier, the cell boundary consists of a semi-permeable plasma membrane. It allows free passage of only small, uncharged molecules like water, glycerol and urea but is effectively impermeable to charged molecules, molecules with a higher molecular mass, and other supramolecular assemblies (Milo and Phillips, 2015). Small ions and some organic molecules can be transported across the membrane with the help of pore-forming proteins or specific transporters. For larger molecules like peptide hormones, enzymes etc. and other supramolecular structures like viral particles, to move in and out of the cell, or in between different cell compartments, a sophisticated vesicle-mediated transport is evolved (Cooper and Hausman, 2000).

Events such as exocytosis, endocytosis, and vesicular transport between the Golgi apparatus and the endoplasmic reticulum are important for proper functioning of diverse cell types such as neurons and immune cells like phagocytes. All of these processes require membrane fusion events to finally deliver the cargo to the delivery site. Viruses too employ fusion machinery for successfully infecting a cell.

In various biological systems, membrane fusion is achieved by highly specialized proteins and protein complexes such as SNAREs (for vesicle trafficking), and hemagglutinin (for viral invasions) (Jahn, Lang and Südhof, 2003). Apart from these, different artificial methods have been developed in the biotechnology field to bring about membrane fusion. Polyethylene glycol (PEG)-mediated fusion is mostly used to fuse cells, plant protoplasts, and also liposomes (Lentz and Lee, 1999; Yang and Shen, 2006).

Other methods used for membrane fusion include electrostatic interactions between bilayers, divalent or multivalent cations to bridge membranes, and even DNA complementarity (Pantazatos and MacDonald, 1999; Tanaka and Yamazaki, 2004; Chan, van Lengerich and Boxer, 2008).

(10)

membrane to rearrange and fuse. Both of these methods rely on disrupting the bilayer structures of the interacting membranes by lowering the energy requirement for the stalk formation (inducing membrane curvature and/or increasing hydrophobic interactions).

1.5 Membrane tension as the driving force for fusion

When two solutions with different concentrations of solutes are separated by a semi-permeable membrane, the solvent has an overall movement towards the higher solute concentration. This is known as osmosis. Osmotic pressure is external pressure that needs to be applied so that there is no net movement of the solvent across the semi-permeable membrane. When a liposome is placed in a hypotonic solution, the water moves into the liposome making it tense and this turgor pressure is equivalent to the osmotic pressure difference (ΔP) is given by,

ΔP 𝑖∆𝑐𝑅𝑇

Here, c is the concentration difference (Cin-Cout in Molarity) across the membrane, R is the

gas constant ( 𝐽

𝑚𝑜𝑙.𝐾) and T is the temperature(K) and i is the van’t Hoff factor. This influx of

water is countered by the Laplace pressure of the liposome which tries to maintain the integrity of the membrane. For a curved surface, Laplace pressure is defined as the difference between the pressures on either side of the surface which is due to the surface tension of the interface. Laplace pressure is defined as following.

𝑃_𝑙 𝛾 (1 𝑟₁ +

1 𝑟₂)

Here,  is the surface tension at the interface, r1 and r2 are the two radii of curvatures. For a

liposome, r1  r2 = r. As a liposome has two interfaces, the Laplace pressure across one surface

can be calculated as 𝑃_𝑙 2

𝑟 .

Therefore, the Laplace pressure across the liposome is 4𝛾

𝑟. At equilibrium, the osmotic pressure

(11)

𝛾 ∆𝑐𝑅𝑇𝑟 4

Thus, the concentration difference across the membrane, at equilibrium, gets translated to the surface tension of the lipid bilayer of the liposome. This surface tension can provide the energy to drive the fusion process. Theoretically speaking, under isotonic conditions there is negligible surface tension, only due to thermal fluctuations.(Koslov and Markin, 1984; Alam Shibly et

al., 2016).

It needs to be noted that, depending on the composition, a lipid bilayer can withstand a maximum membrane tension 3-30 mN/m (Olbrich et al., 2000; Reddy, Warshaviak and Chachisvilis, 2012; Oglęcka et al., 2014), after which the interactions between the lipid molecules is disrupted and the membrane ruptures.

Probing the potential of osmotic stress as a fusogen has been conducted using various systems. When a vesicle that is adhered to a planar bilayer is swollen osmotically, the vesicle fuses with the bilayer (Finkelstein, Zimmerberg and Cohen, 1986). The swelling can be achieved by either changing the osmolarity across the planar membrane or by having a membrane permeable osmoticum (Cohen, Akabas and Finkelstein, 1982; Finkelstein, Zimmerberg and Cohen, 1986). Also, osmotic stress has been shown to enhance the rate of fusions with fusogens like Ca2+

(Zimmerberg, Cohen and Finkelstein, 1980). Molecular dynamics simulations suggest increased fusion events when osmotic stress is involved, with the membrane tension driving the fusion pore opening (Knecht and Marrink, 2007).

(12)

1.6 Aim

In order to achieve a completely synthetic cell cycle, one needs repeated growth and division cycles. In the growth phase, the volume of the synthetic cell doubles making it competent for the division phase of the cycle where it is split into two daughter cells. The aim of this project was to establish the growth phase of the synthetic cell cycle in a simplistic manner. This can be then coupled to the mechanical division of the liposomes to finally complete a simplistic cell cycle (Deshpande et al., 2018).

The specific aims of my project can be divided as follows :-

• Theoretical modelling of experimental conditions to decide optimal parameters (e.g. concentration of osmolytes) to induce maximum growth.

• Testing different osmolytes for density-based separation. • Testing different lipid compositions to facilitate growth

• Ultimately, doubling of the volume of liposomes, which are under osmotic stress by fusing SUVs (Figure 3)

(13)

2 METHODS AND MATERIALS

2.1 Octanol-assisted Liposome Assembly (OLA)

All the experiments described in this work are done using a microfluidic technique known as octanol-assisted liposome assembly (OLA) (Deshpande et al., 2016). Figure 4 from Deshpande

et al., 2016 gives an overview of the process. It employs a microfluidic chip with a six-way

junction including one inner aqueous channel (IA), two lipid-carrying organic phase channels (LO), two outer aqueous phase channels (OA) and one post junction channel. With this architecture, the LO phase forms a film at the junction which is pushed by the IA phase into the post junction channel. Here, due to the hydrodynamic focusing by the OA, it gets pinched off to form a double emulsion of water/alcohol/water. The formed film then pinched off by an outer aqueous (OA) solution forming a double emulsion of water/alcohol/water. The cohesive interaction between the octanol molecules is stronger than the adhesive interactions between octanol and water. Due to this, the octanol soon gets pushed to one side forming a prominent pocket while the IA phase remains surrounded by a lipid bilayer (Deshpande and Dekker, 2018). Due to interfacial tensions between the three phases, the octanol pocket soon buds-off leaving a unilamellar liposome. The formed liposomes have the IA phase as the inner solution and OA forms the surrounding solution.

It is critical to modulate the flow of the three phases for the formation of double emulsions. This is done using pressure pumps (Fluigent MFCS-EZ) which is controlled by a computer with the help of the software MAESFLO. To begin, the OA is flown first past the post-junction channel. This is followed by the LO and IA phases. By carefully modulating the pressures, monodisperse double emulsions can be obtained which then mature further in the post junction channel to form liposomes and 1-octanol droplets.

(14)

The octanol droplet that is formed is an unwanted by-product of the process and may interfere with further experiments. To avoid this, it is separated from the liposomes on the basis of the density differences. A modified version of the previously reported density-based separation was used (Deshpande, Birnie and Dekker, 2017). Instead of a small (~330µm) separation hole, a 4 mm exit well is made. To further aid the settling of liposomes in the exit well, it was loaded with sucrose or dextran.

(15)

This method not only helps in removing the contaminating octanol but also provides a large number of liposomes at the bottom of the well to be used for experiments. Figure 5 gives a schematic representation of this density-based separation process and Figure 6 gives a region of the field of view showing settled liposomes.

Figure 5: Schematic representation of the separation process. 1-octanol pockets having a lower density than the surrounding solution float up while the liposomes, which are made denser, sink to the bottom of the exit well and are used for further experiments.

(16)

2.2 Casting of the microfluidic device

The following casting and surface treatment procedures are described in detail by Deshpande and Dekker, 2018. The microfluidic device used for the experiments were cast using a pre-fabricated master which has the pattern for the device etched onto it. For the experiments, two different masters with different channel heights (6.8 µm and 9.4 µm) were used. Figure 7 shows a design for the device used in the experiments.

The silicon master wafer was thoroughly cleaned with acetone and blow dried with a nitrogen stream before use. 40 g of polydimethylsiloxane base (PDMS) (Sylgard 184 Silicone Elastomer base) was thoroughly mixed with 4 g of the corresponding curing agent (Sylgard 184 silicone elastomer curing agent) and kept under vacuum to remove trapped air bubbles. A wall of aluminium foil was made around the master wafer and ~30 g of the PDMS-curing agent mix was poured onto it. It was desiccated again to get rid of any air bubbles that might have formed while pouring. To another silanized wafer, without any pattern, a thin layer (~5 g) of the PDMS-curing agent mix was poured. Glass coverslips were pressed onto this to have a thin layer of PDMS under them. Both the wafers were kept at 80℃ for 4 hours for the PDMS to polymerise after which the devices were cut from the PDMS block and the coverslips were peeled off.

Holes were punched in the devices for the inlets of OA, IA, and LO using a 0.75 mm punch and 4 mm punch for the exit hole. The exit hole should be ideally punched ~4-5 mm from the junction.

(17)

The punched devices and the coverslips were cleaned using isopropanol and dried with the nitrogen. Using a plasma preen, the coverslips and the devices were plasma treated for 10 s using oxygen plasma following which the patterned side of the device was bonded to the PDMS side of the coverslip and baked in the oven for ~20 minutes at 80℃ and kept at room temperature for at least 4 hours before surface treatment.

2.3 Surface treatment with polyvinyl alcohol

PDMS is hydrophobic in nature and the 1-octanol used in the process would stick to it preventing the formation of double emulsions in the first place. To avoid this, the devices were partially treated with polyvinyl alcohol (PVA) to make the post junction channel hydrophilic. This was done by carefully injecting 5% w/v PVA (30-70k MW and 87-90% hydrolysed) through the OA channel and letting it flow past the junction into the exit which was also filled with PVA later. To avoid PVA from going into the LO and IA channels, the air was pumped through them. The PVA was allowed to flow for 5 minutes after which it was sucked out from the exit hole and the OA inlet using a vacuum. The PVA treated devices were kept in the oven at 120℃ for ~20 minutes to heat immobilize the PVA onto PDMS. As the plasma-treated surface is highly hydrophilic, PVA treatment is usually done between 4-20 hours after bonding.

2.4 Lipids used in different compositions

• 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), MW 786.113 g/mol • 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE), MW 744.034 g/mol • 1-oleoyl-2-hydroxy-sn-glycero-3-phosphocholine (LPC), MW 521.667 g/mol • Oleic acid MW, 282.46 g/mol

• 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(carboxyfluorescein) ammonium salt (PE-CF), MW 1136.395 g/mol

• 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl) ammonium salt (Rh-PE), MW 1301.715 g/mol

(18)

2.5 Preparation of small unilamellar vesicles (SUVs)

Small unilamellar vesicles, used as feed for the growing GUVs, were prepared by extrusion. First, a lipid solution, containing 5 mg of total lipids (species depending on the composition) was dispensed in a round-bottom flask. Chloroform was evaporated under a stream of nitrogen while spreading the lipids at the bottom of the flask to obtain a thin film. Traces of chloroform were evaporated by keeping the flask under vacuum for at least 2 hours. The thin film so obtained was hydrated with 250 µL of desired hydration solution and kept shaking at 30 ℃ for 30 minutes to evenly hydrate the lipid film. The hydrated lipid film was then sonicated for 30 minutes in a bath ultra-sonicator to break larger aggregates and suspend them in the solution. The suspension was then extruded, first through 100 nm and then through 30 nm polycarbonate filter in a mini-extruder(Zhu, Budin and Szostak, 2013). The suspension was passed through each filter 21 times to finally obtain a suspension of SUVs with 30 nm diameter. These were then stored at 4℃ for further use.

2.6 Experimental Setup

OLA requires three solutions for the formation of double emulsions as mentioned before. Also, the exit well must be filled with a solution which acts as a hypotonic solution and also as a solution containing SUVs. Broadly the compositions of the 4 solutions are described below, though changes were made for different experiments.

• IA phase: osmolyte (sucrose or dextran of varying concentrations), 15% glycerol and 0.04 mg/mL dextran-Alexa Fluor 647

• LO phase: 0.2% Lipids in 1-Octanol. The lipids are stored in 10% stock in ethanol and further diluted in 1-octanol.

• OA phase: osmolyte (glucose to balance the osmolarity of the IA phase), 15% glycerol, 5% Poloxamer 188.

(19)

(20)

3 RESULTS

3.1 Back-of-the-envelope calculations to figure out experimental parameters

Prior to experiments, the system was modelled mathematically to optimise the parameters. With every fusion, a finite amount of SUV membrane is incorporated into the original GUV membrane, and the inner content of the SUV gets added to the original GUV volume. As the SUVs have a relatively larger surface area to volume ratio (S/V = 1/R), a lot more surface area is added to the GUV as compared to the volume. However, as the GUV is still under hypotonic stress, the excess membrane re-stretches with a slight influx of water. This influx of water, in turn, dilutes the inner contents, slightly decreasing the osmolyte concentration. Thus, at every fusion, the membrane tension is relieved by addition of excess area and dilution of the solutes inside. This gives a limit to the maximum number of fusions after which the GUV become isotonic. It needs to be noted that this assumes the probability of fusion to be constant for any concentration difference. In reality, this is not the case as the membrane tension would be less than that required to promote fusion even before the solution becomes completely isotonic. Thus, the number of possible fusions calculated this way is an overestimation and the actual number will be much lower than the calculated maximum.

For a GUV of 10 µm and SUVs of 30 nm in diameter, the number of fusions and the corresponding volume increase was calculated using the logic described above. Taking the solute concentration inside (Cin) the GUV and the SUVs to be 100 mM and the surrounding

solute concentration (Cout) of 95 mM (Δc = 5 mM), we can see in the Figure 8

,

the total number

(21)

This is similar to the dilution effect given by the dilution equation C1V1 = C2V2. Where plugging in 100mM as C1 and 95 mM as C2 gives us 100 𝑉₁ 95 𝑉₂ 𝑉₂ 𝑉₁ 100 95 1.052

This exactly corresponds to a 5.2% increase in volume as calculated before. Thus, for maintaining the osmotic stress long enough and prevent the fast dilution of the inner contents, we must consider both the concentration difference (Δc = Cin-Cout) as well as the ratio of

concentrations, Cin/Cout.

For a pure DOPC membrane, the critical membrane tension is reported to be around 10 mN/m (Olbrich et al., 2000), which corresponds to a maximum concentration difference of ~3.23 mM. Considering the importance of the ratio and the maximum concentration difference, we devised the experiments so as to keep the concentration of solutes inside less than 10 mM and the difference of less than 5 mM.

Figure 8 Prediction of the maximum number of fusions and corresponding volume and area increase for

(22)

Considering Cin = 5 mM and Cout = 2 mM gives us the following predictions further proving

the hypothesis (Figure 9).

A

B

Figure 9: Prediction of the maximum number of fusions and corresponding volume and area increase for a

Δc = 3 mM (Cin = 5 mM and Cout =2 mM). A) The total number of predicted fusions is ~ 𝟗. 𝟑 𝟏𝟎𝟒. B)

(23)

Another practical consideration needed to be taken into account. Since the SUVs were made by the extrusion, the surrounding buffer would have the same osmolarity as the SUV contents. In order to maintain the required hypotonicity, the SUV suspension thus needs to be diluted before adding it into the exit well. As a simple solution, the SUVs were made with a higher lipid concentration of 20 mg/mL and diluted to a final concentration of 1–2 mg/mL in the exit well.

3.2 Growing pure lipid vesicles

3.2.1 Using cylindrical (DOPC) and inverted cone-shaped (DOPE) lipids

Since DOPC is a cylindrically shaped lipid, liposomes formed by DOPC alone are less fusogenic due to low spontaneous curvatures. DOPE, on the other hand, has an inverted cone shape, which increases the membrane curvature and hence is more fusogenic(Chernomordik et

al., 1985). That is why a combination of these two lipids (DOPC-DOPE in 7:3 molar ratio) was

chosen. However, the production of liposomes with OLA with this lipid composition was quite troublesome. Not all octanol pockets separated to produce mature liposomes. Liposomes with octanol pockets could not be used for the analysis as the pockets could potentially act as a lipid reserve to induce growth. Nonetheless, a sufficient number of pure liposomes were collected over time at the bottom of the well and SUVs were added later. They were then incubated overnight in the SUV bath. The frequency distribution of the GUV radii before and after overnight incubation with SUVs is given in Figure 11. In total, 5054 liposomes were analysed before and after, demonstrating the high-throughput nature of the method. As can be observed, both the initial and the final distributions look very similar with the initial radius being 4.8±0.8 µm, compared to the final radius of 4.7±0.8 µm, suggesting no growth occurred.

(24)

Figure 11: Population distribution of liposome radii before and after the incubation with SUVs. The osmotic stress applied was of Δc = 2.7 mM. Red and blue curves indicate the size distribution of the population before and after incubation respectively. Initial mean size is 4.83 ± 0.83 µm and the final mean size is 4.68 ± 0.84 µm.

(25)

3.2.2 Using cylindrical (DOPC), inverted cone-shaped (DOPE), and cone-shaped (LPC) lipids

When using the DOPC-DOPE composition, separation of octanol pockets was a major issue. This was attributed to the negative curvature of DOPE lipids. Prior observations in the lab suggested that having lipids with positive curvature promote pocket separation. To aid the pocket separation, it was decided to incorporate LPC into the mixture. This resulted in a very effective pocket separation. The value of Δc was kept around 3 mM in order to have a high enough membrane tension but prevent bursting of the liposomes at the same time. The initial images of liposomes were obtained in the channel because the SUV suspension was already present in exit well. This change was made in order to simplify the experimentation and also avoid any flow-induced movement of liposomes due to the addition of SUV solution to start the experiment. As the liposomes could be potentially squeezed in the channel depending on their size (channel height 6.8 µm or 9.4 µm depending on the device), the actual sizes were obtained through a correction factor as described in the APPENDIX.

For a Δc of 2.75 mM, the population distribution of the GUV sizes before and after overnight incubation with SUVs is given in Figure 12. As can be seen from figure 12A, we observed a slight right shift in the distribution. The size distribution before the incubation was quite skewed with a left tail and with a sharp cut-off on the right edge of the peak. This sharp right cut-off is attributed to the fact that the production of liposomes using OLA is quite mono-disperse. The left tail is due to the unwanted bursting and resealing of liposomes in the post-junction channel.

This can happen due to less-than-ideal surface treatment, rough channel walls, and presence of physical obstacles (e.g. dirt particles) in the channel. After the incubation with SUVs, the left tail could still be seen, but more interestingly, the distribution became broader with a shift in the mean value, from 𝑟𝑖𝑛𝑖𝑡𝑖𝑎𝑙 5.3 ± 1.2 𝜇𝑚 ( n = 461) to 𝑟𝑓𝑖𝑛𝑎𝑙 5.7 ± 1.2 𝜇𝑚 ( n = 1319)

suggesting a growth of 7.1% in radius and a corresponding 22.9% increase in the volume.

(26)

It must be noted that the maximum size before incubation was 7.3 µm and that after incubation was 8.1 µm. Assuming that these data points are given by the same liposomes, it gives us a minimum estimate of 10.8% increase in the radius, which corresponds to a 36% increase in volume. It is definitely likely that the actual growth was higher than this estimate.

(27)

We wondered whether the orientation of SUVs with respect to the bilayer curvature of the GUVs would make a difference. This was tested by encapsulating the SUVs inside the GUVs. The SUVs were thus added in the inner aqueous solution and equivalent osmolarity was maintained in the exit solution. For a similar Δc of 2.75 mM, encapsulating SUVs fared better than having SUVs outside. As seen from the Figure 13, prior to the induction of osmotic stress, the size distribution was highly monodisperse, with a mean radius 𝑟_{𝑖𝑛𝑖𝑡𝑖𝑎𝑙} 5.1 ± 0.5 𝜇𝑚 ( n = 351). Again, a weak left tail was seen, corresponding to the unwanted bursting-resealing of liposomes. However, after the induction of osmotic stress, the distribution clearly became wider on the right side, suggesting that a significant fraction of liposomes had grown. The mean of the distribution changed to 𝑟_{𝑓𝑖𝑛𝑎𝑙} 5.4 ± 0.8 𝜇𝑚 ( n = 1234) which gives an average increase of 4.5% in radius and a corresponding 15.8% increase in volume.

The difference in the shapes of the two distributions suggests a heterogeneous growth pattern. Most of the burst-and-resealed liposomes, representing the left tail, might not have grown as losing their inner contents led to the excess surface area, reducing the membrane tension. On the other hand, bigger liposomes have a higher probability of achieving fusion with SUVs due to higher membrane stress.

(28)

Thus, the mean is not very a good measure to capture growth, as it could substantially underestimate its value. The extent of growth can be aided by looking at the maximum liposome size observed before and after growth. The max liposome size initially was found to be 5.7 µm whereas the largest liposome observed after incubation was 7.8 µm. This gives us an underestimation of maximum increase in radius being 36% and a corresponding volume increase to be 151.6%. Thus, a small fraction of the population actually more than doubled in volume, which is exactly what we were aiming for.

A better representation is a cumulative percentage frequency distribution shown in Figure 14. It was seen that for a radius < 5 µm, the trend looked similar but for larger sizes the difference between the distributions was prominent. Initially, 99% of liposomes were under the size of ~5.75 µm whereas after incubation only 60% of the liposomes were smaller than that, meaning 40% of the liposome population grew, although to a variable extent.

(29)

3.3 Growing hybrid vesicles (DOPC and oleic acid)

As mentioned earlier, fatty acid vesicles have a higher chance of growth due to the fact that fatty acids tend to be dynamic in the bilayer due to their single aliphatic chain. To use this mobility of fatty acids to aid growth, a hybrid vesicle system, comprising of DOPC and oleic acid (1:1 molar ratio) was used. The production of hybrid GUVs was found to be feasible with OLA, with a very similar operating procedure. Similar to DOPC-DOPE-LPC system, the initial images were obtained in the channel. Figure 15 shows the population distribution of the liposome sizes before and after incubation with the SUVs.

As can be seen in the figure, the change in distributions is similar to the DOPC-DOPE-LPC system. Prior to incubation, the initial distribution is monodisperse, with a slight left shoulder due to the bursting of liposomes. For Δc = 2.75 mM, the mean radius of the liposomes before incubation, 𝑟𝑖𝑛𝑖𝑡𝑖𝑎𝑙 4.5 ± 0.6 𝜇𝑚 (n = 631) with a maximum value of 5.8 µm. After

incubation, the distribution becomes broader with the mean radius 𝑟_{𝑓𝑖𝑛𝑎𝑙} 5.1 ± 1.0 𝜇𝑚 (n = 1314) and a maximum radius of 8.15 μm. Taking the means gives us an increase in the radius of 12.7% and a corresponding the volume increase of 43.14%.

However, considering the maxima of the distributions gives an increase of 40.5% in radius which corresponds to a 175% increase in volume. These two analyses suggest that a significant population of liposomes achieved at least a 100% increase in volume, i.e., doubling their inner content.

(30)

The observed broadening of the distribution, preferentially on the right side, suggests inherent heterogeneity in growth. The cumulative percentage frequency distribution given in Figure 16, shows that while no changes are observed for the smaller radii, the difference is much more pronounced for higher sizes. Initially, about 99% of the liposomes were under 5.2 µm, whereas after incubation only ~62% of liposomes were under 5.2 µm suggesting that at least ~35% of liposomes grew.

Similar results were obtained with a device of higher height (9.4 µm as compared to 6.8 µm, to take into account the possible effect of squeezing on the analysis). The osmotic pressure difference was kept the same (2.75 mM, Fig.15). Figure 17 shows the size distribution and the corresponding cumulative percentage frequency distribution of the experiments. The mean radius increased by 10.17%, i.e., a volume increase of 33.7%. However, the maximum radius of the distribution increased from 6 µm to 8.9 µm which is a staggering 49.2% increase in radius corresponding to a huge 237.5% increase in volume.

(31)

As can be seen from the cumulative percentage distribution of liposome sizes in Figure 18

,

about 90% of the liposomes initially were under a size of 5.5 µm.

The number dropped to 55% after incubation. Also, about 30% of the liposomes grew beyond the initial maximum value of ~5.8 µm.

Figure 17: Population distribution of liposome sizes before and after incubation with SUVs. The osmotic stress was by Δc of 2.75 mM. Red and blue curves indicate the size distribution of the population before and after incubation respectively. Initial mean size is 4.72±0.59 µm (n = 313) and final mean size is 5.2±0.74 µm (n = 2361), i.e 10.17% increase in radius.

(32)

A similar experiment was performed, but with a higher osmotic stress of Δc = 3.5 mM. However, it had a drastic effect. The mean initial radius 𝑟_{𝑖𝑛𝑖𝑡𝑖𝑎𝑙} 5.3 ± 0.5 𝜇𝑚 ( n= 496) changed to 𝑟𝑓𝑖𝑛𝑎𝑙 5.6 ± 0.9 𝜇𝑚 (n = 5008) suggesting a growth of a mere 5% increase in

radius corresponding to 16.1% increase in volume which seems significantly lower than the previous results. However, looking at the distribution in Figure 19

,

suggests that the initial production was not as mono-disperse as previous cases which might explain the means being so close.

Taking into account the maximum sizes in the initial and final distribution, suggests an increase of 53.2% in the radius from a maximum of 6.18 µm to 9.5 µm which corresponds to an enormous amount of 259.7% growth in the volume. Also, note a slight hump on the left of the peak in the distribution after incubation, this could be due bursting of liposomes due to higher osmotic pressure.

This trend of growth is much better evident from the cumulative percentage distribution in Figure 20. It shows that almost 99% of the liposomes were under a size of 6 µm initially whereas, in the final distribution, only about 70% of the liposomes were under the size of 6 µm.

(33)

3.4 No growth is observed in absence of membrane fusion

To further corroborate that the observed growth was driven via membrane fusion with SUVs, we carried out experiments in absence of SUVs but in a hypotonic bath to still have the osmotic stress. This control was also important to ascertain that the correction factor used to find the initial sizes, arising from liposomes being slightly squeezed inside the channels, was justified and not prone to errors. Furthermore, hybrid vesicles were chosen for the negative control as fatty acid vesicles are more prone to growth at the expense of smaller ones.

As can be observed from the size distributions in Figure 21, the initial and the final distributions were quite identical with the initial radius. 𝑟𝑖𝑛𝑖𝑡𝑖𝑎𝑙 5.7 ± 0.7 𝜇𝑚 (n = 176) and the final

radius, 𝑟𝑓𝑖𝑛𝑎𝑙 5.4 ± 0.9 𝜇𝑚 (n= 299). Even though it might seem that the liposomes have

shrunk, a look at the distribution shows more small liposomes which greatly affect the mean value. The presence of small liposomes in the well after overnight incubation is much more prominent from the cumulative percentage distribution. From Figure 22

,

it can be seen that apart from the excess number of smaller liposomes, both the distributions are quite similar.

(34)

In both the cases, ~99% of liposomes are under 6.6 µm. Comparing the maximum radii of the distribution gives an initial maximum of 7 µm and a final maximum of 8.15 µm which suggests an increase of 16% in the radius.

Figure 21: Population distribution of liposome sizes before and after incubation without SUVs. The osmotic stress was by Δc of ~2.7 mM. Red and blue curves indicate the size distribution of the population before and after incubation respectively. Initial mean size is 5.66±0.7 µm (n = 176) and final mean size is 5.36±0.94 µm (n = 299).

(35)

4 DISCUSSION

In this project, we successfully demonstrated the growth of osmotically stressed cell-sized liposomes by fusing multiple small unilamellar vesicles with them. Under hypotonic conditions, the increased membrane tension increased the energy of the bilayer, making the fusion of SUVs to the GUV an energetically favourable process by releasing the membrane tension. For SUVs, the membrane is highly curved due to the small size, putting a strain on the lipid molecules and making the SUVs fusogenic

With DOPC-DOPE liposomes, no significant growth was observed. In this experiment, the SUVs were added after the liposomes have settled in the well. Addition of the SUV suspension increases the solute concentration in the well. As it was added later, the GUVs settled in the well experience a slightly higher osmotic pressure than with the addition of SUVs. Under these circumstances, the membrane tension is higher than the lysis tension and the giant vesicles might open up transient pores and release solutes lowering the osmotic pressure difference(Chabanon et al., 2017). This happens continuously till the tension is relieved. During this time, the vesicle seems to pulsate as the pores open and close. Over this time, the osmotic pressure might have dropped and thus no growth was observed. Also, we can see that the initial distribution suggested that the liposomes in the well were not monodisperse. The initial broad distribution can be attributed to the presence of octanol pockets and bursting of liposomes, both of which reduces the osmolarity by dilution or by leakage.

(36)

SUVs were fusing from the outside. This is since, as SUVs fuse from outside, a small amount of solute gets added to the GUV contents, whereas, if the SUVs fuse from inside the inner contents of SUVs are released outside.

Fusion of bilayers requires an intricate system with different stages in the process governed by different aspects of the bilayer lipids (Markvoort and Marrink, 2011). Even though the initial adhesion stalk formation and the formation of hemifusion diaphragm are highly favourable with the PE lipids. The final opening of the fusion pore is repressed by it. (Grafmüller, Shillcock and Lipowsky, 2009; Markvoort and Marrink, 2011). Once the hemifusion diaphragm is formed, the lipids with high positive curvature like LPC and the membrane tension promote the opening and expansion of the fusion pore(Nikolaus et al., 2011). This was also observed in our experiments with DOPC-DOPE-LPC.

Another composition tested was an equimolar mixture of DOPC and oleic acid to make hybrid (lipid-fatty acid) vesicles. As fatty acids are much more dynamic and make the bilayer unstable, they can be assumed to be more fusogenic. This was found to be true since hybrid vesicles showed the highest amount of growth, exemplified by the mean increase of around 40% in volume as compared to around 16% with lipid vesicle (DOPC-DOPE-LPC).

To draw a conclusion whether the SUVs are fusing with the GUVs, a fusion assay with Eu3+_{/SDIP was performed. Eu}3+ _{was encapsulated within the SUVs while SDIP was}

encapsulated inside the GUVs. Both the components are not fluorescent on their own but upon fusion, Eu3+_{forms a fluorescent complex with SDIP. Surprisingly, we could not observe a}

significant fluorescence signal in the GUVs. This could be due to the low amounts Eu3+_which

get even further diluted in the GUVs, yielding negligible fluorescence. Increasing the concentration of Eu3+_{was, however, not feasible due to the fact that it would act as an osmolyte}

(37)

with the conducted overnight experiments as the liposomes were free to diffuse in the exit well. Even though the liposomes were initially scattered across the well, overnight, they tend to drift towards the edge. We speculated that this could be due to the evaporation of water through the PDMS. Nevertheless, I would still expect them to behave quite heterogeneously due to inherent and minutes differences between the GUVs.

Due to the difference in the shapes of the distributions and the heterogeneity, the population mean is not a robust measure for quantifying growth. However, for the scope of this project, the mean along with the maximum sizes was taken to be a good enough measure for quantifying growth. As the distribution is skewed to the left in the initial observations, percentiles would be a better measure. The cumulative percentage frequency distribution gives a range of sizes where most liposomes lie under a certain value. It is evident from them that this size range for the majority of liposomes increases after incubation. One may expect the liposomes with still intact 1-octanol pockets to grow as the pocket acts as a lipid reserve. These liposomes are visibly distinct, as the octanol pocket would be brightly fluorescent. Such contaminating liposomes were excluded manually or within the MATLAB script based on this increased fluorescence intensity .

With the obtained results, it can be stated that the stressed liposomes have the potential to grow substantially (> 100% increase in volume). However, quantifying such growth and giving a general growth rate is not an easy task. A conclusive experiment would be having the liposomes trapped at the bottom of the well to restrict their movement or to employ tracking software for the microscope to move the stage accordingly. This would give us both a size distribution trend and individual growth events. Fusion assay also needs to be improved, by encapsulating a higher concentration of Eu3+_{which may be a challenging task as the multivalent cations also}

interact with the membrane and destabilize it. Previous experiments in the group were done with trapped liposomes. Due to the heterogeneity of growth and a low number of trapped liposomes, the trend of growth was quite inconsistent.

Although the experiments done in this project are better at gauging the population trends, for a conclusion of growth, it would be better to have a large number of liposomes with their individual growths tracked and analysed.

Once grown to double the volume, these liposomes can then be cycled into the mechanical division system developed in the lab to have a completely synthetic cell cycle (Deshpande et

(38)

5

REFERENCES

• Alam Shibly, S. U., Ghatak, C., Sayem Karal, M. A., Moniruzzaman, M. and Yamazaki, M. (2016) ‘Experimental Estimation of Membrane Tension Induced by Osmotic Pressure.’, Biophysical journal. The Biophysical Society, 111(10), pp. 2190–2201. doi: 10.1016/j.bpj.2016.09.043.

• Chabanon, M., Ho, J. C. S., Liedberg, B., Parikh, A. N. and Rangamani, P. (2017) ‘Pulsatile Lipid Vesicles under Osmotic Stress’, Biophysical Journal. Biophysical Society, 112(8), pp. 1682–1691. doi: 10.1016/j.bpj.2017.03.018.

• Chan, Y.-H. M., van Lengerich, B. and Boxer, S. G. (2008) ‘Lipid-anchored DNA mediates vesicle fusion as observed by lipid and content mixing’, Biointerphases, 3(2), pp. FA17-FA21. doi: 10.1116/1.2889062.

• Chen, I. A. and Szostak, J. W. (2004) ‘A kinetic study of the growth of fatty acid vesicles.’, Biophysical journal. The Biophysical Society, 87(2), pp. 988–98. doi: 10.1529/biophysj.104.039875.

• Chernomordik, L. V., Kozlov, M. M., Melikyan, G. B., Abidor, I. G., Markin, V. S. and Chizmadzhev, Y. A. (1985) ‘The shape of lipid molecules and monolayer membrane fusion’, Biochimica et Biophysica Acta (BBA) - Biomembranes. Elsevier, 812(3), pp. 643–655. doi: 10.1016/0005-2736(85)90257-3.

• Chizmadzhev, Y. A. (2012) ‘Membrane fusion’, Biochemistry (Moscow) Supplement

Series A: Membrane and Cell Biology. SP MAIK Nauka/Interperiodica, 6(2), pp. 152–

158. doi: 10.1134/S1990747812010047.

• Cohen, F. S., Akabas, M. H. and Finkelstein, A. (1982) ‘Osmotic swelling of phospholipid vesicles causes them to fuse with a planar phospholipid bilayer membrane.’, Science (New York, N.Y.). American Association for the Advancement of Science, 217(4558), pp. 458–60. doi: 10.1126/SCIENCE.6283637.

• Colley, C. M. and Ryman, B. E. (1976) The liposome: from membrane model to

therapeutic agent. Available at: https://www.cell.com/trends/biochemical-sciences/pdf/S0968-0004(76)80030-8.pdf (Accessed: 14 August 2018).

• Cooper, G. M. and Hausman, R. E. (2000) The cell: a molecular approach, Boston

University, Sunderland. ASM Press. Available at: https://books.google.nl/books?id=c9fEQgAACAAJ.

(39)

phospholipid vesicle in a bath of fatty acids’, European Physical Journal Plus, 132(6), pp. 1–5. doi: 10.1140/epjp/i2017-11554-1.

• Deshpande, S., Birnie, A. and Dekker, C. (2017) ‘On-chip density-based purification of liposomes’, Biomicrofluidics, 11(3), p. 034106. doi: 10.1063/1.4983174.

• Deshpande, S., Caspi, Y., Meijering, A. E. C. and Dekker, C. (2016) ‘Octanol-assisted liposome assembly on-chip’, Nature Communications, 7, p. 10447. doi: 10.1038/ncomms10447.

• Deshpande, S. and Dekker, C. (2018) ‘On-chip microfluidic production of cell-sized liposomes’, Nature Protocols, 13(5), pp. 856–874. doi: 10.1038/nprot.2017.160. • Deshpande, S., Spoelstra, W. K., van Doorn, M., Kerssemakers, J. and Dekker, C.

(2018) ‘Mechanical Division of Cell-Sized Liposomes’, ACS Nano. American Chemical Society, 12(3), pp. 2560–2568. doi: 10.1021/acsnano.7b08411.

• Finkelstein, A., Zimmerberg, J. and Cohen, F. S. (1986) ‘OSMOTIC SWELLING OF VESICLES: Its Role in the Fusion of Vesicles with Planar Phospholipid Bilayer Membranes and its Possible Role in Exocytosis 1’, Ann. Rev. Physiol, 48, pp. 163–74.

Available at:

https://www.annualreviews.org/doi/pdf/10.1146/annurev.ph.48.030186.001115 (Accessed: 27 February 2018).

• Forster, A. C. and Church, G. M. (2006) ‘Towards synthesis of a minimal cell.’,

Molecular systems biology. European Molecular Biology Organization, 2, p. 45. doi:

10.1038/msb4100090.

• Gardner, P. M., Winzer, K. and Davis, B. G. (2009) ‘Sugar synthesis in a protocellular model leads to a cell signalling response in bacteria’, Nature Chemistry. Nature Publishing Group, 1(5), pp. 377–383. doi: 10.1038/nchem.296.

• Grafmüller, A., Shillcock, J. and Lipowsky, R. (2009) ‘The fusion of membranes and vesicles: pathway and energy barriers from dissipative particle dynamics.’, Biophysical

journal. The Biophysical Society, 96(7), pp. 2658–75. doi: 10.1016/j.bpj.2008.11.073.

• Guertin, D. A. and Sabatini, D. M. (2015) ‘Cell Growth’, in John Mendelsohn, Joe W. Gray, Craig B. Thompson,Peter M. Howley, M. A. I. (ed.) The Molecular Basis of

Cancer. 4th edn, p. 179–190.e1. doi: 10.1016/B978-1-4557-4066-6.00012-3.

(40)

Smith, H. O., Glass, J. I., Merryman, C., Gibson, D. G. and Venter, J. C. (2016) ‘Design and synthesis of a minimal bacterial genome’, Science, 351(6280), pp. aad6253-aad6253. doi: 10.1126/science.aad6253-aad6253.

• Jahn, R., Lang, T. and Südhof, T. C. (2003) ‘Membrane fusion.’, Cell. Elsevier, 112(4), pp. 519–33. doi: 10.1016/S0092-8674(03)00112-0.

• Jewett, M. C. and Forster, A. C. (2010) ‘Update on designing and building minimal cells.’, Current opinion in biotechnology. NIH Public Access, 21(5), pp. 697–703. doi: 10.1016/j.copbio.2010.06.008.

• Knecht, V. and Marrink, S.-J. (2007) ‘Molecular Dynamics Simulations of Lipid Vesicle Fusion in Atomic Detail’, Biophysical Journal, 92(12), pp. 4254–4261. doi: 10.1529/biophysj.106.103572.

• Koslov, M. M. and Markin, V. S. (1984) ‘A theory of osmotic lysis of lipid vesicles’,

Journal of Theoretical Biology. Academic Press, 109(1), pp. 17–39. doi:

10.1016/S0022-5193(84)80108-3.

• Kurihara, K., Tamura, M., Shohda, K., Toyota, T., Suzuki, K. and Sugawara, T. (2011) ‘Self-reproduction of supramolecular giant vesicles combined with the amplification of encapsulated DNA’, Nature Chemistry, 3(10), pp. 775–781. doi: 10.1038/nchem.1127. • Lentini, R., Martín, N. Y., Forlin, M., Belmonte, L., Fontana, J., Cornella, M., Martini, L., Tamburini, S., Bentley, W. E., Jousson, O. and Mansy, S. S. (2017) ‘Two-Way Chemical Communication between Artificial and Natural Cells’, ACS Central Science. American Chemical Society, 3(2), pp. 117–123. doi: 10.1021/acscentsci.6b00330. • Lentz, B. R. and Lee, J. K. (1999) ‘Poly(ethylene glycol) (PEG)-mediated fusion

between pure lipid bilayers: a mechanism in common with viral fusion and secretory vesicle release?’, Molecular membrane biology, 16(4), pp. 279–96. Available at: http://www.ncbi.nlm.nih.gov/pubmed/10766128 (Accessed: 15 August 2018).

• Lodish, H. F. and Rothman, J. E. (1979) ‘The Assembly of Cell Membranes’, Scientific

American. Scientific American, a division of Nature America, Inc., pp. 48–63. doi:

10.2307/24965067.

• Luisi, P. L., Ferri, F. and Stano, P. (2006) ‘Approaches to semi-synthetic minimal cells: a review’, Naturwissenschaften. Springer-Verlag, 93(1), pp. 1–13. doi: 10.1007/s00114-005-0056-z.

(41)

10.1016/B978-0-12-385891-7.00011-8.

• Milo, R. and Phillips, R. (2015) ‘Cell Biology By the Numbers: What Is the Density of Cells?’ Garland Science, p. 1.

• Murtas, G. (2010) ‘Internal lipid synthesis and vesicle growth as a step toward self-reproduction of the minimal cell.’, Systems and synthetic biology. Springer, 4(2), pp. 85–93. doi: 10.1007/s11693-009-9048-1.

• van Nies, P., Westerlaken, I., Blanken, D., Salas, M., Mencía, M. and Danelon, C. (2018) ‘Self-replication of DNA by its encoded proteins in liposome-based synthetic cells’, Nature Communications. Nature Publishing Group, 9(1), p. 1583. doi: 10.1038/s41467-018-03926-1.

• Nikolaus, J., Warner, J. M., O’Shaughnessy, B. and Herrmann, A. (2011) ‘The Pathway to Membrane Fusion through Hemifusion’, Current Topics in Membranes. Academic Press, 68, pp. 1–32. doi: 10.1016/B978-0-12-385891-7.00001-5.

• Noireaux, V. and Libchaber, A. (2004) ‘A vesicle bioreactor as a step toward an artificial cell assembly.’, Proceedings of the National Academy of Sciences of the

United States of America. National Academy of Sciences, 101(51), pp. 17669–74. doi:

10.1073/pnas.0408236101.

• Oglęcka, K., Rangamani, P., Liedberg, B., Kraut, R. S. and Parikh, A. N. (2014) ‘Oscillatory phase separation in giant lipid vesicles induced by transmembrane osmotic differentials’, eLife. eLife Sciences Publications Limited, 3, p. e03695. doi: 10.7554/eLife.03695.

• Olbrich, K., Rawicz, W., Needham, D. and Evans, E. (2000) ‘Water Permeability and Mechanical Strength of Polyunsaturated Lipid Bilayers’, Biophysical Journal, 79(1), pp. 321–327. doi: 10.1016/S0006-3495(00)76294-1.

• Pantazatos, D. P. and MacDonald, R. C. (1999) ‘Directly observed membrane fusion between oppositely charged phospholipid bilayers.’, The Journal of membrane biology, 170(1), pp. 27–38. Available at: http://www.ncbi.nlm.nih.gov/pubmed/10398758 (Accessed: 15 August 2018).

• Razin, S. (1997) ‘The minimal cellular genome of mycoplasma.’, Indian journal of

biochemistry & biophysics, 34(1–2), pp. 124–30. Available at: http://www.ncbi.nlm.nih.gov/pubmed/9343940 (Accessed: 13 August 2018).

(42)

Biophysica Acta (BBA) - Biomembranes, 1818(9), pp. 2271–2281. doi:

10.1016/j.bbamem.2012.05.006.

• Salim, M., Minamikawa, H., Sugimura, A. and Hashim, R. (2014) ‘Amphiphilic designer nano-carriers for controlled release: from drug delivery to diagnostics’, Med.

Chem. Commun. The Royal Society of Chemistry, 5(11), pp. 1602–1618. doi:

10.1039/C4MD00085D.

• Schmidli, P. K., Schurtenberger, P. and Luisi, P. L. (1991) ‘Liposome-Mediated Enzymatic Synthesis of Phosphatidylcholine as an Approach to Self-Replicating Liposomes’, Journal of the American Chemical Society, 113(21), pp. 8127–8130. doi: 10.1021/ja00021a043.

• Schrum, J. P., Zhu, T. F. and Szostak, J. W. (2010) ‘The origins of cellular life.’, Cold

Spring Harbor perspectives in biology. Cold Spring Harbor Laboratory Press, 2(9), p.

a002212. doi: 10.1101/cshperspect.a002212.

• Scott, A., Noga, M. J., de Graaf, P., Westerlaken, I., Yildirim, E. and Danelon, C. (2016) ‘Cell-Free Phospholipid Biosynthesis by Gene-Encoded Enzymes Reconstituted in Liposomes’, PLOS ONE. Edited by H. W. van Veen, 11(10), p. e0163058. doi: 10.1371/journal.pone.0163058.

• Takakura, K., Toyota, T. and Sugawara, T. (2003) ‘A Novel System of Self-Reproducing Giant Vesicles’, Journal of the American Chemical Society, 125(27), pp. 8134–40. doi: 10.1021/ja029379a.

• Tanaka, T. and Yamazaki, M. (2004) ‘Membrane fusion of giant unilamellar vesicles of neutral phospholipid membranes induced by La3+.’, Langmuir : the ACS journal of

surfaces and colloids, 20(13), pp. 5160–4. Available at: http://www.ncbi.nlm.nih.gov/pubmed/15986643 (Accessed: 15 August 2018).

• Weisstein, E. W. (2018) Pappus’s Centroid Theorem, MathWords-Wolfram Web

Resource. Wolfram Research, Inc. Available at: http://mathworld.wolfram.com/PappussCentroidTheorem.html (Accessed: 18 August 2018).

• Wick, R. and Luisi, P. L. (1996) ‘Enzyme-containing liposomes can endogenously produce membrane-constituting lipids’, Chemistry and Biology, 3(4), pp. 277–285. doi: 10.1016/S1074-5521(96)90107-6.

• Xu, C., Hu, S. and Chen, X. (2016) ‘Artificial cells: from basic science to applications.’,

(43)

10.1016/j.mattod.2016.02.020.

• Yang, J. and Shen, M. H. (2006) ‘Polyethylene Glycol-Mediated Cell Fusion’, in

Nuclear Reprogramming. New Jersey: Humana Press, pp. 59–66. doi:

10.1385/1-59745-005-7:59.

• Zhu, T. F., Budin, I. and Szostak, J. W. (2013) ‘Vesicle Extrusion Through Polycarbonate Track-etched Membranes using a Hand-held Mini-extruder’, in Lorsch, J. (ed.) Methods in Enzymology, pp. 275–282. doi: 10.1016/B978-0-12-420067-8.00021-0.

• Zimmerberg, J., Cohen, F. and Finkelstein, A. (1980) ‘Micromolar Ca2+ stimulates fusion of lipid vesicles with planar bilayers containing a calcium-binding protein’,

Science. American Association for the Advancement of Science, 210(4472), pp. 906–

(44)

6

APPENDIX

A liposome in normal circumstances would assume a spherical shape in order to minimize the surface area. However, when in the channel with a height smaller than the observed liposome diameter (D = 2R), the liposome would acquire a pancake-like geometry

.

Figure 23 shows a representation of the top and a side view of a liposome inside a channel with a channel height less than the diameter of the observed liposome.

The part of the liposome not in contact with the channels would assume a curved shape to minimise the surface area. To have a smooth edge, the curved part needs to be a semi-circle in the cross-section. The radius of the semi-circular part would be equal to half the channel height, i.e., h/2. Thus, the cross-section can be represented as a rectangle with dimensions (R-h/2) sandwiched between two semi-circular caps of radius h/2.

As the liposome exits the well, it is no longer constrained and thus would assume a spherical shape. As no lipids can be added or removed, the area of the liposome in the well should be

Figure 23: A schematic representation of a squeezed liposome inside the channel of height less than the observed diameter. The top part represents the top view at the equator, whereas the bottom part represents the side view of the middle cross-section. The shape of the liposome can be represented as the rotation of this cross-section with a rectangular centre and two semi-circular caps on both ends.

Top-view

(equatorial)

(45)

equal to the area of the liposome in the channel. This is used as the correction for liposome sizes in the channel.

The shape of the liposome is obtained by the rotation of the side view, cross-section shape which is in effect, a rectangle with two semi-circular caps on both ends.

As the shape is symmetrical about the central axis, one 360o_{rotation of either half would give}

us the solid. So, the area of the given shape can be represented as the sum of the areas obtained by the rotation of the semi-circular arc around the axis and the area of the flat faces on top and bottom. 𝑇𝑜𝑡𝑎𝑙 𝑠𝑢𝑟𝑓𝑎𝑐𝑒 𝑎𝑟𝑒𝑎 𝐴𝑟𝑒𝑎 𝑜𝑓 𝑓𝑙𝑎𝑡 𝑓𝑎𝑐𝑒 + 𝐴𝑟𝑒𝑎 𝑜𝑓 𝑡ℎ𝑒 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑠𝑒𝑚𝑖𝑐𝑖𝑟𝑐𝑙𝑒 2 𝜋 (𝑅 −ℎ 2) 2 + 𝐴𝑟𝑒𝑎 𝑜𝑓 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑠𝑒𝑚𝑖𝑐𝑖𝑟𝑐𝑙𝑒

To find the surface area of rotation, Pappus’s first centroid theorem is employed. The theorem states that when a curve is rotated about an external axis, the area of rotation is equal to the product of the arc length and the distance travelled by the centroid (Weisstein, 2018).

The centroid of a semi-circular arc lies at a perpendicular distance of 2𝑟/𝜋 where r is the radius of the semi-circle. In our case, the centroid would be ℎ/𝜋 from the centre of the semi-circle. Upon rotation about the central axis, the centroid would trace a circular path with a radius equalling 𝑅 −ℎ

2+ ℎ

𝜋. Thus, the distance travelled by the centroid would be 2𝜋(𝑅 − ℎ 2+

ℎ 𝜋).

Also, the length of the semi-circular arc would be half the circumference, i.e., 𝜋ℎ/2. Therefore, according to Pappus’s centroid theorem, the area of surface of rotation of the semi-circle is

2𝜋(𝑅 −ℎ 2+ ℎ 𝜋) 𝜋 ℎ 2