Giant Lipid Vesicle Formation Using Vapor-Deposited Charged

Jul 2, 2018 - Joseph PazziMelissa XuAnand Bala SubramaniamJoseph Pazzi, Melissa Xu, and Anand Bala Subramaniam. Langmuir 2018 Article ASAP...
3 downloads 0 Views 5MB Size
Biological and Environmental Phenomena at the Interface

Subscriber access provided by BOSTON COLLEGE

Giant Lipid Vesicle Formation Using VaporDeposited Charged Porous Polymers Nareh Movsesian, Matthew Tittensor, Golnaz Dianat, Malancha Gupta, and Noah Malmstadt Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.8b00736 • Publication Date (Web): 02 Jul 2018 Downloaded from http://pubs.acs.org on July 10, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

1

Giant Lipid Vesicle Formation Using Vapor-Deposited Charged Porous Polymers Nareh Movsesian,1 Matthew Tittensor,1 Golnaz Dianat,1 Malancha Gupta,*1,2 and Noah Malmstadt*1,2,3 Departments of Chemical Engineering and Materials science1, Chemistry2, and Biomedical Engineering3, University of Southern California, Los Angeles, California 90089, United States

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 29

2

ABSTRACT In this study, we prepare giant lipid vesicles using vapor-deposited charged microporous poly(methacrylic acid-co-ethylene glycol diacrylate) polymer membranes with different morphologies and thicknesses. Our results suggest that vesicle formation is favored by thinner, more structured porous hydrogel substrates. Electrostatic interactions between the polymer and the lipid head groups affect vesicle yield and size distribution. Repulsive electrostatic interactions between the hydrogel and the lipid head groups promote vesicle formation; attractive electrostatic interactions suppress vesicle formation. Ionic strength and sugar concentration are also major parameters affecting the yield and size of giant vesicles. The presence of both ions and sugars in the hydration buffer results in increased vesicle yields. These results indicate that lipid-polymer interactions and osmotic effects in addition to the substrate morphology and surface charge are key factors affecting vesicle formation. Our data suggest that surface chemistry should be designed to tune electrostatic interactions with the lipid mixture of interest to promote vesicle formation. This vapor-deposited hydrogel fabrication technique offers tunability over physicochemical properties of the hydrogel substrate for the production of giant vesicles with different sizes and compositions.

KEYWORDS GUV, hydrogel swelling, poly(methacrylic acid), iCVD, lipid bilayer

ACS Paragon Plus Environment

Page 3 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

3

INTRODUCTION Giant unilamellar vesicles (GUVs) are commonly employed as practical cell models to examine various aspects of biomembranes1,2 such as interactions with nanoparticles3,4, membrane dynamics5, incorporation of membrane proteins and their activities6-8, lipid bilayer permeability9, and transport of ions through the membrane10. They have also been explored as potential drug delivery vehicles.11 Emerging techniques for GUV fabrication are opening new research vistas by facilitating the construction of systems that better mimic biology. Unlike the traditional electroformation12 and gentle hydration13 techniques for giant vesicle formation, rehydration of lipids using hydrophilic hydrogels has proven efficient to form vesicles over a short timescale under physiological conditions at high yields.14 Even though electroformation using longer hydration times and high field frequencies allows vesicle formation at high buffer ionic strength, the yield is low and undesired lipid peroxidation and hydrolysis reactions occur.15,16 Microfluidic techniques, on the other hand, produce monodisperse vesicles of various compositions at high yields; yet they often require specialized devices and leave oil impurities in the lipid membranes.17-19 Hydrogel-assisted rehydration easily forms vesicles made of charged lipid mixtures under physiological conditions, overcoming one of the main disadvantages of electroformation. Recent studies on the dynamics of GUV formation on agarose hydrogels have given insight into the formation mechanism.20 Small vesicles form upon rehydration and coalesce on the surface of agarose to form larger vesicles. Higher buffer ionic strength results in Debye screening of vesicle surface charges, leading to increased rate of coalescence and larger average vesicle size regardless of lipid type. In addition, higher agarose density increases the rate of hydrogel rehydration and GUV coalescence due to water retention in the hydrogel which allows for faster preorientation and self-assembly of the lipid lamellae. Following the introduction of agarose as the first effective hydrogel substrate for the rehydration method, submillimeter glass beads21,22, cellulose paper23,24, and synthetic hydrogel materials such as crosslinked polyacrylamide12, poly(vinyl alcohol)25, and crosslinked dextran(ethylene glycol)26 have been introduced. Altered vesicle mechanical properties due to polymer incorporation in the bilayer system arising from contamination of vesicles by agarose and poly(vinyl alcohol) chains have been reported.27,28 Cellulose and synthetic crosslinked hydrogels, however, are insoluble and therefore do not result in vesicle contamination from the polymer substrate. Lopez Mora and coworkers recently introduced chemically crosslinked dextran (poly ethylene glycol) (DexPEG) to control vesicle size and production by controlling crosslinking density.26 They further showed that chemical functionality and crosslinking conditions affecting surface roughness of the hydrogel significantly impacts the yield of GUV production. Moreover, weaker interactions between the polymer and lipid were shown to favor the formation of vesicles.29 Clearly, physical and chemical properties of the hydrogels are major control parameters for vesicle formation. Here we introduce a method of GUV formation based on swelling from a covalently crossed-linked, vapor-deposited, porous, hydrophilic polymer with well controlled surface chemistry and morphology. In this work, porous hydrophilic polymers with distinct morphologies and thicknesses are fabricated by a modified initiated chemical vapor deposition (iCVD) technique and are used as hydrogel substrates for GUV production. The iCVD process is a solventless polymerization method that is typically used to deposit dense polymer films via a free radical polymerization mechanism.30 In this report, our modified iCVD technique was used to form porous polymer membranes by introducing the gaseous monomer into a vacuum chamber where the monomer

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 29

4 partial pressure and substrate temperature are kept below the triple point pressure and freezing temperature of the monomer. Hence, the vapor undergoes a gas to solid phase change on the surface of the substrate and forms solid microstructures.31-34 Polymerization subsequently occurs under a heated filament array in the presence of an initiator and a crosslinking agent is used to render the polymer insoluble in aqueous media.35 Following polymerization, unreacted solid monomer is removed via sublimation resulting in membranes with dual-scale porosity. The morphology and large-scale porosity (tens to hundreds of microns) of the membranes are dependent on the shape of the deposited solid monomer which is controlled by substrate temperature and monomer deposition time while smaller scale pores (hundreds of nanometers to a few microns) are formed during sublimation of the unreacted monomer. We previously demonstrated that monomer deposition at low temperatures (-20°C) results in 3D pillar-like microstructures while deposition at higher temperatures (0 °C) results in 2D web-like structures.36 Here we report on the fabrication of crosslinked poly(methacrylic acid-co-ethylene glycol diacrylate) (xPMAA) porous membranes with distinct morphologies and thicknesses to form GUVs from zwitterionic and charged lipid mixtures. PMAA is a pH-responsive polymer (pKa = 5.7) 37 and is negatively charged at physiological pH. We show that the yield and the size distribution of the vesicles are controlled by the morphology and charge of the porous hydrogels. Furthermore, we study the effect of buffer ionic strength and charge of the lipid on vesicle formation. Our results show that the morphology and charge of the polymer both play significant roles in the yield of vesicles, specifically for charged lipid mixtures.

EXPERIMENTAL SECTION Materials. 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), 1-palmitoyl-2oleoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (POPG), 1,2-dioleoyl-3-trimethylammoniumpropane (DOTAP), and 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(Liss Rhod PE) were purchased from Avanti polar lipids (USA) and used without purification. Liss Rhod PE was used as a fluorescent lipid probe. Bovine serum albumin (BSA), chloroform (CHCl3), phosphatebuffered saline (PBS), sucrose, glucose, calcein, α-hemolysin from staphylococcus aureus, methacrylic acid (MAA; 99%), and tert-butyl peroxide (TBPO; 98%) were purchased from Sigma Aldrich and used as received. Ethylene glycol diacrylate (EGDA) was purchased from Polyscience was used as a crosslinking agent. Deuterium oxide (D2O, 99.9%) was purchased from Cambridge Isotope Laboratories, Inc. 18.2 MΩ-cm milli-Q water (EMD Millipore, USA), Sykes-Moore chambers (Bellco, USA), and standard 25 mm no. 1 glass coverslips (ChemGlass, USA) were used in the GUV formation experiments. Fabrication of Porous Membranes. The membranes were made in a custom-designed pancake-shaped iCVD vacuum chamber of 48 mm in height and 250 mm in diameter (GVD Corporation) containing a nichrome filament array (Omega Engineering, 80%/20% Ni/Cr). A thermoelectric cooler (TEC) was incorporated onto the chamber stage for temperature control. Vacuum was achieved using a rotary vane pump (Edwards E2M40) and maintained with a throttling valve (MKS 153D) and capacitance manometer (MKS Baratron 622A01TDE). MAA and EGDA feed jars were maintained at 30 °C and 35 °C, respectively. Needle valves were used to meter MAA and EGDA flows, and a mass flow controller (MKS 1479 A) was used

ACS Paragon Plus Environment

Page 5 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

5 to meter TBPO flow. Porous polymer membranes were deposited onto 1.5 cm x 1 cm silicon wafer substrates (Wafer World) that were located on top of the TEC. To synthesize xPMAA porous membranes, TBPO was first introduced into the reactor at 0.7 sccm to achieve a total pressure of 650 mTorr. Then MAA monomer was introduced at 3.5 sccm while keeping the reactor pressure at 650 mTorr. The deposition time was varied between 10 s to 2 min and the TEC temperature was varied between and -20 °C and 0 °C during MAA deposition. MAA flow was then halted and the reactor pressure was decreased to 200 mTorr with TBPO flowing at 1.0 sccm. The TEC temperature was increased to 0 °C and EGDA was introduced into the reactor at 0.2 sccm. Subsequently, polymerization started by heating the filament to 230 °C for 30 minutes for all depositions. Finally, the TBPO and EGDA flows were stopped and the samples were pumped down to allow the unreacted excess MAA to sublimate which was confirmed by the system returning to base pressure (Figure 1). Sublimation time across the samples was variable and lasted from 10 to 40 minutes. In order to ensure the membranes were not soluble in aqueous media, solvent extraction in deuterium oxide was performed on multiple samples and the 1H-NMR spectra were collected using a Varian Mercury 400 instrument. Samples were imaged using scanning electron microscopy (SEM; Topcon Aquila) at 20 kV accelerating voltage. ImageJ was used to process the SEM images for estimation of the thickness and pore sizes of the membranes. Gold was sputter-coated on polymer samples for 30 s prior to imaging in order to avoid charging. Giant Vesicle Formation. Glass coverslips were sonicated in MeOH at 38 °C for 30 minutes, dried at 45°C, and plasma treated in a PDC-32G benchtop plasma cleaner (Harrick Plasma, USA) for 10 minutes. The wafers with the porous polymer membranes were glued on coverslips using epoxy and held in Sykes-Moore chambers. For GUV formation, lipid solution was applied by syringe on the porous polymer substrate, dried with a gentle stream of nitrogen, and subsequently hydrated with buffer for 1 hour. 2 mg/ml of the lipid solutions containing 0.4 mole % Liss Rhod PE (POPC, 20 POPG: 80 POPC (mol%) and 20 DOTAP: 80 POPC (mol%)) were prepared in CHCl3. For each vesicle formation experiment, 10 µl of the lipid solution was added dropwise in 2 µl drops on the polymer membranes and quickly dried with a gentle flow of nitrogen gas to evaporate the solvent and form a uniform lipid film on the substrate. Following the lipid film formation process on the substrate, 600 µl hydration buffer was added to each sample and the system was allowed to hydrate for 1 hour. To determine the yield and size distribution of the vesicles, GUVs formed in sucrose-containing PBS buffers (in 200 mM, 400 mM, and 440 mM sucrose concentrations) were transferred into an isosmotic PBS buffer containing glucose. Osmolarities of the buffers were measured using an Osmomat 3000 basic freezing point osmometer. GUVs were harvested and transferred to 700 µl glucose solution and were allowed to settle for 40 minutes. From the bottom of the harvested GUV solution, 400 µl was gently withdrawn and transferred to an observation chamber for viewing. Observation chambers were pre-treated by incubation with 1 mg/ml BSA solution in milliQ water to passivate their surfaces and avoid vesicle rupture. After 40 minutes of further settling down in the observation chambers, samples were imaged using fluorescence microscopy and vesicle count and size were determined using ImageJ. Lamellarity Measurement. Unilamellarity of the vesicles were tested using α-hemolysin as the pore forming protein and calcein as the fluorescent dye. Vesicles were prepared using -20

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 29

6 ℃_36 µm membrane in 185 mM PBS buffer containing 200 mM sucrose as described previously and were harvested in185 mM PBS buffer containing 200 mM glucose and 0.1 mM Calcein. Final alpha-hemolysin concentration was 14 µg/ml and images were taken 30 minutes after adding the protein. Microscopy. Fluorescent micrographs were acquired on an Axio Observer Z1 (Zeiss, Germany) inverted microscope using an EC Plan-Neofluar 40× objective (numerical aperture of 0.75) with 1.6× optovar magnification and equipped with a Hamamatsu CMOS camera (Hamamatsu, Japan). Each image taken contains 2048×2048 pixels (pixel size of 0.102 µm). Illumination was provided by a Colibri 2 LED Illumination System with a 120 V LED illuminator (Zeiss, Germany). Liss Rhod PE was illuminated using a green filter (555/25 nm bandpass, double band emission filter 575/98 nm). For the lamellarity experiments, imaging was performed using a Nikon TI-E inverted microscope with a spinning-disk CSUX confocal head (Yokogawa, Japan), a S Plan-Fluor 40× objective (numerical aperture of 0.6) and a 16-bit Cascade II 512 electronmultiplied charge- coupled device camera (Photometrics, US). 50 mW solid-state lasers (Coherent, US) at 491 nm (emission filter centered at 525 nm) and 561 nm (emission filter centered at 595 nm) were used as the illumination sources for calcein and rhodamine, respectively. Image Processing and Data Analysis. Image processing was conducted in ImageJ by utilizing a batch macro to process and collect vesicle count and size data from samples. Particle size measurements were performed by making the images binary using the triangle thresholding enabling the built-in particle analyzer function to operate. Processed vesicle size data was then transferred to JMP for statistical analysis. Particulate structures with diameters less than 3 micron were not easily distinguishable as vesicles and hence were eliminated from data analysis. These structures were not frequently observed.

RESULTS AND DISCUSSION Understanding the role that polymer structure and chemistry plays in GUV swelling from hydrogel substrates is an important step in optimizing this increasingly important technique. Here, we use a vapor-phase polymerization technique to create polymer substrates with controlled structure and chemistry. By varying these parameters, we are able to study their impacts on GUV formation. To form controlled polymer hydrogel substrates, we used a modified iCVD technique. Porous polymer membranes of xPMAA were formed in a two-step process. First, MAA monomer vapor was deposited as solid microstructures on a cooled silicon substrate by keeping the substrate temperature and monomer partial pressure below the triple point temperature and pressure of the monomer, respectively. This solid MAA was then copolymerized with a crosslinking agent (EGDA) from the vapor phase in the presence of initiator free radicals that were formed under the heated filament array (Figure 1). Excess unreacted monomer was subsequently sublimated after polymerization by decreasing the reactor pressure. Decoupling the monomer deposition and the polymerization steps allows for changes in process parameters (e.g., temperature and pressure) that allowed for crosslinking and parametric studies of the relationships between hydrogel properties and GUVs formed from these hydrogels. Additionally, solvent extraction and 1H-NMR confirm that the membranes were not soluble in aqueous media. No signal pertaining to the polymer was detected after solvent extraction on several membranes (Figure S6).

ACS Paragon Plus Environment

Page 7 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

7 The morphology and thickness of the xPMAA hydrogels are the physical properties that were varied by changing the substrate temperature and the monomer deposition time. The micrographs in Figure 2a correspond to membranes with average thicknesses of 36 ± 11µm, 82 ± 8 µm, and 124 ± 19 µm fabricated by deposition at substrate temperatures of -20, -10, and 0 °C. Membranes of all thicknesses were prepared by varying the deposition time at each of the deposition temperatures (Figure 2b). SEM images indicate that lower temperatures result in both closer spacing of nucleation sites and shorter deposition periods for a given thickness. Higher temperatures require longer deposition times to achieve the same thickness. Moreover, the microstructures increase in width from several microns to tens of microns when the total membrane thickness is increased at -20 °C. Increasing the deposition temperature to -10 °C and further to 0 °C forms structures with considerably fewer void spaces. These data are consistent with previously demonstrated monomer depositions at temperatures of -20 °C and 0 °C resulting in 3D pillar-like and 2D web-like microstructures, respectively. Effects of the physicochemical properties of xPMAA hydrogels on giant vesicle formation were studied in terms of vesicle yield and size distribution. Vesicle formation experiments were performed using the hydrogel assisted method using xPMAA porous membranes in Figure 2a. Figure 3 shows the count for zwitterionic and net-charged lipid GUVs formed upon hydration of the porous membranes. Vesicle count declines at higher deposition temperatures and thicker membranes regardless of the lipid type; recall that both of these parameters are correlated with longer deposition times. The increase in POPC vesicle count at lower deposition temperatures is likely due to increased structures in low temperature depositions which provide more nucleation sites for the lipid film to grow into vesicles (Figure 3a). Increasing thickness leads to a decrease in vesicle yield due to structural densification and a corresponding decrease in nucleation sites. Vesicle yield for membranes with higher average thicknesses of 82 ± 8 µm and 124 ± 19 µm at deposition temperatures of -20 and -10 °C are on the same order of magnitude while membranes formed at 0 °C result in a very poor vesicle yield. The surface roughness of crosslinked DexPEG hydrogels was previously shown to alter the yield of vesicles in a similar manner.29 The lamellarity of several POPC vesicles was tested using α-hemolysin as a membrane pore forming protein and calcein as a fluorescent dye that does not passively cross the lipid bilayer.38,39 Lumen intensity of some vesicles increased after incubation with α-hemolysin (due to pore formation and calcein transport into the vesicles). However, lumen intensity of other vesicles did not increase in the presence of α-hemolysin (Figure S5). This indicates that both unilamellar and multilamellar vesicles are formed in this method. We also observed a mixture of unilamellar and multilamellar vesicles, as determined by relative fluorescent intensity, under various conditions (Figure S4). Net-negatively charged and net-positively charged lipid compositions (20% POPG or 20% DOTAP with the remainder POPC) show similar trends in vesicle yield. For both lipids, decreased yield is associated with higher deposition temperatures and thicker membranes as shown in Figure 3b. However, DOTAP yields were overall much lower. This indicates that electrostatic interactions between the negatively charged polymer and the charged lipid head groups are crucial to vesicle formation under physiological conditions (185 mM PBS with 200mM Sucrose at pH=7.4). Electrostatic repulsion between the anionic POPG lipid mixture and the negatively charged polymer favors formation of vesicles while electrostatic attraction between the cationic DOTAP lipid mixture and the polymer has an adverse impact on vesicle yield. For the case of positively charged DOTAP lipid mixture, increasing ionic strength of the hydration buffer 4-fold from 185 mM to 740 mM results in a drastic increase in vesicle yield, as

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 29

8 shown in Figure 3b. Shielding the negative charges of the polymer by higher ionic strength weakens lipid-polymer interactions and allows the charged lipid film to swell and grow into vesicles (Figure 3c). Figure 4a shows the log-scale size distribution of vesicles formed upon hydration of POPC-coated porous membranes. For a given thickness, increasing deposition temperature leads to smaller vesicles. However, there is not a strong dependence of vesicle size on membrane thickness except in the case of the 124 µm membranes, which produce significantly smaller vesicles (Figure 4a, Table 1a, Table S2). The size and yield results do not correlate with the measured equilibrium mass swelling data (Table S6). Higher swelling in the case of thicker and denser membranes does not result in larger vesicles. We surmise that polymer substrates resulting in high vesicle yields (such as the 36 µm thick membrane deposited at -20 °C) also result in larger vesicles; this is likely due to coalescence when many vesicles are present in a given area. Similar to vesicle bulge merging in electroformation40, vesicle coalescence via merging through the surface of polymer has been shown for both agarose and dextran PEG hydrogels.20,29 In the case of negatively charged POPG, there is no trend relating yield and size. This is consistent with charge-charge repulsion inhibiting coalescence of these vesicles (Figure 4b, Table 1b). Given the small number of DOTAP vesicles formed at low ionic strength, no statistically significant correlation between growth conditions and size can be identified. However, higher buffer ionic strength increases the average DOTAP vesicle size significantly and produces multilamellar vesicles indicated by higher fluorescence intensities (Figure 4c, Table 1c, Figure S4). This is likely due to screening of charges on the positively charged lipid lamella and the negatively charged polymer surfaces facilitating vesicle growth and coalescence (Figure 4f). We further studied the effect of hydration buffer ionic strength on POPC GUV formation. The goal of these studies was to decouple ionic effects from osmotic effects. 36 µm thick membranes deposited at -20 ℃ were hydrated with PBS buffer containing 200 mM sucrose at varying ionic strengths of PBS (1X PBS with a total ionic strength of 185 mM contains 137 mM NaCl, 10 mM PO4-3 and 2.7 mM KCl). We then repeated these experiments with the concentration of sucrose varied such that buffer osmolarity was constant across varying ionic strengths. At constant concentrations of sucrose, GUV count is significantly greater in the presence of salt. However, increasing ionic strength above 18.5 mM does not significantly increase yield (Figure 5a). Average vesicle size increases in the presence of PBS (Figure 5b, Table 2) in a trend that mirrors that observed for vesicle count: there is a significant increase upon the addition of salt but adding salt above 18.5 mM results in no additional increases. In the isosmotic case, more vesicles are formed in the higher sugar concentrations, and a combination of both sugar and salt in the buffer results in the highest observed yield. Together with the high vesicle yields observed in the presence of sucrose, high sucrose concentrations also result in the formation of tubules and GUVs that are accumulated inside each other (Figure S3). A large difference in osmotic pressure and enhanced repulsion between lipid lamellae through gentle hydration of sugar-containing dried lipid films have previously shown to promote GUV formation.41 Moreover, Tubule formation has been reported in systems with high sucrose concentrations and charged lipid mixtures due to excess membrane area.42-44 High concentrations of sugar have shown to cause instabilities by affecting interactions and lateral expansion of the membranes.45 These results suggest that high osmolarity or, possibly, direct

ACS Paragon Plus Environment

Page 9 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

9 interactions between the lipids and sugars, is more important in controlling vesicle characteristics than ionic strength. Giant vesicle yields in this method are comparable to the vesicle yields produced using the agarose rehydration method. Other phenomena such as ionic, osmotic, surface interactions and polymer morphological effects are also consistent with previous reports. The high surface area provided by the porous hydrogel network similar to the large areas in the submillimeter sized glass beads and the fibrous cellulose paper lead to larger number of vesicles produced in this method. 20-24,26

CONCLUSION In comparison to previously introduced hydrogel supports for giant vesicle preparation, this vapor phase fabrication method offers control over different process parameters such as temperature, pressure, and time that can be independently adjusted to modify the characteristics of the resulting polymer for vesicle production. These characteristics include crosslinking density, thickness, morphology, and surface chemistry of the polymer. In this report we show that largest vesicle yields are achieved by highly porous thin hydrogel films, lipid films that have weak interactions with the polymer substrate, and the presence of ions and sugars in the hydration buffer. Electrostatic interactions between the lipid and the polymer also contribute to vesicle yield. Our results indicate that minimized electrostatic attractions between the polymer and the lipid head groups favor vesicle formation. Electrostatic interactions between the lipid and the underlying hydrogel and its impact on vesicle size and yield are further borne out by the effects of hydration buffer ionic strength. Both higher ionic strength and sugar concentration lead to increased vesicle yields. In order to produce vesicles of certain size and composition, surface chemistry can be modified to minimize attractive interactions with the lipid mixture of interest. Manipulation of surface chemistry can be achieved by the application of a polymer top coat on the porous scaffolds or copolymerization with various functionalities. As with the previous hydrogel assisted methods, porous xPMAA forms polydisperse vesicles. Further studies regarding the effect of surface chemistry on vesicle growth could give insights in designing hydrogel conditions for producing low polydispersity vesicles at higher yields. These findings set important guidelines toward the design of novel hydrogels in practical applications that involve engineering surface interactions.

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 29

10

Figure 1. Schematic (not to scale) of the vapor phase fabrication of porous membranes in an iCVD chamber proceeded by hydrogel-supported vesicle formation experiments. Steps a-c depict the process of (a) monomer deposition at variable substrate temperatures, followed by (b) polymerization in the presence of a crosslinking agent under the heated filament array and ultimately, (c) sublimation of the unreacted monomer in the reactor. Steps d-f represent the hydrogel-supported formation of vesicles: (d) applying the lipid film on the polymer membrane and subsequently (e) hydrating it with sucrose buffer and (f) harvesting the vesicles in glucose buffer for observation with fluorescence microscopy.

ACS Paragon Plus Environment

Page 11 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

11

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 29

12

Figure 2. (a) SEM micrographs of xPMAA porous membranes deposited at temperatures -20 ℃, -10 °C, and 0 ℃ with three different thicknesses. Note that deposition time is varied with temperature to maintain a fixed thickness. Insets show the angled-view of the corresponding membranes. The scale bars for the top-down and angled micrographs are 150 µm and 20 µm, respectively. (b) Deposition time versus thickness of the porous membranes measured at the three temperatures: circle, diamond and square correspond to deposition temperatures -20 ℃, -10 °C, and 0 ℃ and red, blue and black refer to average thicknesses of 36 ± 11µm, 82 ± 8 µm, and 124 ± 19 µm, respectively. Error bars for each of the depositions are calculated from thickness measured at three different locations in each sample. The average thicknesses are then calculated by taking the average of the measured mean thicknesses for the three temperatures.

ACS Paragon Plus Environment

Page 13 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

13

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 29

14

ACS Paragon Plus Environment

Page 15 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

15

Figure 3. (a) Yield of vesicles formed from POPC lipid-coated porous membranes (membrane labels correspond to the designations in Figure 2a) upon hydration with 185 mM PBS (137 mM NaCl, 10 mM PO4-3, and 2.7 mM KCl) buffer containing 200mM sucrose at pH 7.4. (b) Yield of vesicles formed from charged lipid mixture (20%POPG:80% POPC or 20%DOTAP:80% POPC). The first four pairs of bars describe hydration with 185 mM PBS buffer containing 200mM sucrose at pH 7.4. The final bar shows the results for 20%DOTAP:80% POPC hydrated with 740 mM PBS containing 200 mM Sucrose at pH=7.4. (c) Schematic for the proposed charge screening mechanism in oppositely charged polymer and lipid mixtures. Increasing hydration buffer ionic strength decreases electrostatic interactions between the polymer and the lipid and promotes vesicle growth. The Debye length in 185 mM PBS buffer is 0.7 nm. Increasing the ionic strength by 4 fold should decrease the Debye length by half.

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 29

16

ACS Paragon Plus Environment

Page 17 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

17

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 29

18

ACS Paragon Plus Environment

Page 19 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

19

Figure 4. (a) Box plots for natural log-scale size distribution of the vesicles formed from POPC- coated porous membranes (figure 2a) upon hydration with 185 mM PBS buffer containing 200mM sucrose at pH 7.4. (b) and (c) Box plots for log-scale size distribution of the vesicles formed from 20%POPG and 20% DOTAP coated porous membranes upon hydration with 185 mM PBS buffer containing 200mM sucrose at pH 7.4 (except if otherwise mentioned). Note that the x axes refer to the polymer iCVD deposition temperature and nominal thickness of the membranes. The central box spans the interquartile range with the confidence diamond containing the mean (horizontal line passing in the middle) and the upper and lower 95% of the mean. The line segment inside each box shows the median for each set. Whiskers (lines connected to the box) extend from each end of the box to the outermost data point that falls within the range computed as follows: 1st quartile - 1.5*(interquartile range) and 3rd quartile + 1.5*(interquartile range). All data points outside the range are considered outliers. (d) and (e) Fluorescent micrographs of the vesicles formed from pure POPC and 20% POPG lipid-coated 36 µm porous membranes (deposition temperatures of -20, -10, and 0 ℃) upon hydration with 185 mM PBS containing 200 mM Sucrose at pH=7.4. (f) Fluorescent micrographs of the vesicles formed from 20% DOTAP lipid-coated 36 µm thick porous membrane at deposition temperature -20 ℃ upon hydration with 185 mM and 740 mM PBS containing 200 mM Sucrose at pH=7.4. False color was added and the contrast in all the images was enhanced for better visualization. All scale bars are 10 µm.

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 29

20 Table 1. Non-parametric size comparisons for sample pairs of the same thickness and deposition temperature using Wilcoxon rank-sum test for (a) pure POPC (b) 20% POPG (c) 20% DOTAP vesicles formed using different membranes upon hydration with 185 mM PBS buffer containing 200 mM sucrose except the case for 20% DOTAP vesicles. This method was used to determine whether two independent data sets come from the same distribution. Unlike the t-test, this test does not require the assumption of normal distribution. The null hypothesis is that the data for each pair comes from the same distribution. p values smaller than the calculated Z reject the hypothesis. Stars represent significance in an ascending order: p