Giant Vesicles Produced with Phosphatidylcholines (PCs) and Phosphatidylethanolamines (PEs) by Water-in-Oil Inverted Emulsions

(1) Background: giant vesicles (GVs) are widely employed as models for studying physicochemical properties of bio-membranes and artificial cell construction due to their similarities to natural cell membranes. Considering the critical roles of GVs, various methods have been developed to prepare them. Notably, the water-in-oil (w/o) inverted emulsion-transfer method is reported to be the most promising, owning to the relatively higher productivity and better encapsulation efficiency of biomolecules. Previously, we successfully established an improved approach to acquire detailed information of 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC)-derived GVs with imaging flow cytometry (IFC); (2) Methods: we prepared GVs with different lipid compositions, including phosphatidylcholines (PCs), phosphatidylethanolamines (PEs), and PC/PE mixtures by w/o inverted emulsion methods. We comprehensively compared the yield, purity, size, and encapsulation efficiency of the resulting vesicles; (3) Results: the relatively higher productivities of GVs could be obtained from POPC, 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2-dilauroyl-sn-glycero-3-phosphoethanolamine (DLPE), DOPC: DLPE (7:3), and POPC: DLPE (6:4) pools. Furthermore, we also demonstrate that these GVs are stable during long term preservation in 4 °C. (4) Conclusions: our results will be useful for the analytical study of GVs and GV-based applications.


Introduction
To understand the organization and dynamics of lipid bilayers, research on giant vesicles (GVs) as an artificial cellular membrane model has gained considerable interest in recent decades [1][2][3][4][5][6]. GVs consisting of phospholipid membranes are similar to modern cellular membranes in terms of physical and chemical properties. Additionally, GVs have large dimensions of their typical cell sizes, making them suitable artificial compartments for the construction of synthetic cells [7,8] through the bottom-up approach [9], is very attractive in the synthetic biology field. Thus, the development of reliable protocols for the preparation and analysis of GVs in a designed environment is crucial for promoting the synthetic cell project to the next stage.
The importance of GVs has initiated a massive effort in developing an easy and efficient preparation method to produce vesicles from known lipid compositions [6]. The challenge lies in establishing a reproducible way to produce GVs with a considerable encapsulation efficacy for biomolecules in a desired yield and quality. To date, various techniques for GV preparations have been established, such as lipid hydration including simple film hydration [10] and gel- [11] or paper-assisted hydration [12], electro-formation [13,14], and emulsion phase transfer method [15][16][17]. Lipid hydration, as one of the most commonly Table 1. Chemical information on all lipids used to produce giant vesicles (GVs).

Solubilized Lipids, Inner and Bottom Aqueous Solutions Preparation
All

Solubilized Lipids, Inner and Bottom Aqueous Solutions Preparation
All  (Table 2). Finally, the chloroform was evaporated by heating the lipid-oil solutions in an open tube at 80 • C for 30 min. Two aqueous solutions, inner and bottom aqueous solutions, were made using ddH 2 O water. The inner aqueous solution contained 50 mM HEPES-KOH (pH 7.6), 500 mM sucrose, 0.2 mM calcein. The bottom aqueous solution was also prepared with 50 mM HEPES-KOH (pH 7.6) and 500 mM glucose to maintain the isosmotic conditions across the lipid bilayer membranes.

Preparation of Calcein-Stained Vesicles by the Inverted Emulsion Method
The w/o emulsion transfer method was chosen to produce GVs as previously described [15] with some minor modifications for some lipids. Calcein was added to the inner aqueous solution enabling quantitative analysis of GVs by flow cytometry. After the lipids were dissolved and the chloroform was entirely evaporated, the w/o emulsion was made by sonication using an ultrasonic processor (SONICS & MATERIALS, INC, VCX750, Newton, Connecticut, USA) on ice for 1 min (Ampl. 40%, pulse on/off 2 s/1 s) and equilibrated on ice for 10 min. Considering that lipid species used in this study have different structures and transition temperatures, the sonication conditions for the emulsion produced with those lipids were optimized, as listed in Table 2. The bottom aqueous solution was placed in a new microcentrifuge tube, and the w/o emulsion obtained in the previous step was gently poured on top of the bottom aqueous solution. After standing for 5−10 min to equilibrate the interface, the emulsion was centrifuged to form a pellet of GVs. The supernatant oil was removed and a fresh bottom aqueous solution was added to resuspend the pellet. Oil traces were removed by another centrifugation round. Finally, 100 µL of this GV suspension was prepared and immediately analyzed by IFC or stored at 4 • C for further stability analysis.

GV Analysis by Imaging Flow Cytometry (IFC)
Like Ishiodori et al. [15], we used the ImageStream Mark II system and INSPIRE acquisition software (Amnis/Millipore, Seattle, Washington, USA) to measure GVs based on the fluorescence intensities from encapsulated calcein. The emission from calcein (488 nm) was detected in Channel 2 with a 505−560 nm filter; bright field and side scatter (SSC) data were collected in Channel 4 and Channel 6 (785 nm), respectively. All samples were acquired with 60× magnification with a setup of low flow rate/high sensitivity. 50,000 events were recorded for each GV composition and the measurement data was analyzed by IDEAS analysis software (Amnis/Millipore, Seattle, Washington, USA). In this type of flow cytometry, internal size standard beads are run concurrently and used for daily calibration as well as real-time velocity detection and autofocus.

Confocal Microscopy Observation
The produced GVs were dropped (8 µL) on a microscope slide with a silicon imaging spacer (1 well, diam. × thickness 9 mm × 0.12 mm, Sigma-Aldrich, St. Louis, MO, USA), which were ultra-thin adhesive spacers peeled and stuck to slides to confine specimens without compression. The sample was then covered with a coverslip and left to stand for 10 min to allow the sediment of vesicles to travel to the bottom of the chamber by gravity. GV samples were observed with a Ti2-E inverted microscope (Nikon Ti2-E, Yokohama, Japan), and images were acquired using a confocal microscope (Nikon C2plus, Yokohama, Japan). The green fluorescence of calcein inside the GVs was excited by using a 488 nm laser with emission collected at 498-535 nm. Image analysis was performed using NIS-ELEMENTS C-ER software (Nikon, Yokohama, Japan).

Stability Analysis
GVs produced with POPC, DOPC, DLPE, DOPC: DLPE (7:3) mixture, and POPC: DLPE (6:4) mixture were stored at 4 • C for up to one month. We investigated the concentration and size distribution of the resulting GVs via IFC at indicated time points, 0 h, 6 h, 12 h, 24 h, 3 days, 1 week, 2 weeks, and 1 month. We also directly observed these GV samples under the confocal microscope (Nikon C2plus, Yokohama, Japan).

Using IFC to Analyze GVs Produced by the w/o Emulsion Transfer Method
The experimental procedures are summarized in Figure 1a. An advantage of using IFC to characterize GVs was that the size distribution of all GVs could be clearly displayed and recorded. We employed IFC and microscopy to analyze the properties of GVs produced from different lipid components. As vesicles produced by the inverted w/o emulsion transfer method are of high encapsulation efficiency and mainly in the uni-lamellar form [15,16], we chose this method to prepare GVs with PCs and PEs with or without calcein encapsulations [15]. The obtained GVs were then subjected to IFC analysis as described in the "Materials and Methods" Section.   Following the procedure developed previously [15], we further made some fine-tuning for each kind of GV sample. Based on a sequence diagram for various SSC features versus brightfield, we could separate target GVs from the internal calibration beads and other byproducts such as oil droplets. Finally, we could also obtain characteristics of target GVs such as purity, calcein encapsulation efficiency, and vesicle size from the IFC analysis results as listed in Table 3 and Table S1. All the data were summarized from at least three independent replications, showing low standard deviations in each sample and outstanding reproducibility via IFC-based analysis.

Comparing GVs Prepared with Different PC, PE, and PC:DLPE Mixture
To investigate the effects of PC and PE species on vesicle formation, we firstly prepared GVs with seven kinds of PCs (POPC, DMPC, DOPC, DPPC, DSPC, DPHPC, and DLPC) or six kinds of PEs (DLPE, DOPE, DPPE, DSPE, POPE, and DMPE). Before IFC and confocal observation, generated GVs were viewed under a fluorescence microscope as soon as they were prepared. As shown in Figure 1b (upper panel), the fluorescence microscope showed that most PCs successfully produced GVs, except for DHPC and DLPC which yielded inferior productions ( Figure S1). For GVs prepared with PEs, DLPE obtained a relatively higher yield of GVs as compared with DOPE, POPE, and DMPE species (Figure 1b, lower  panel). However, the fluorescence microscope could barely detect GV products with DSPE (data not shown). The size of DLPE-derived vesicles seems smaller than DMPC-derived ones as judged from the fluorescence microscopy.
Then we investigated all the GV samples via IFC, with which we calculated the concentration of produced GVs. As plotted in Figure 1c, POPC, DMPC, and DOPC had higher yields of GVs than DPPC, DSPC, DPHPC, and DLPC. Overall, DOPC could achieve the highest vesicle production (6.54 ± 0.17 × 10 9 vesicles/mL) and calcein encapsulation efficiency (94.12 ± 0.30%, Table 3). Out of 13 kinds of lipids tested in this study, the concentration of GVs from DLPE reaches the highest at 7.02 ± 0.90 × 10 9 vesicles/mL ( Figure 1d and Table 3). These results acquired from IFC were mainly consistent with the microscopy images. Moreover, with IFC we could compare detailed analysis among samples, and we conclude that POPC, DMPC, DOPC, and DLPE allow for efficient GV production.
According to the above concentration results, we tried to prepare GVs with mixed POPC, DOPC, and DMPC with DLPE in ratios of 6:4, 7:3, 8:2, and 9:1, because it was reported that the PC/PE ratio is a critical modulator of membrane integrity [30,41]. For GVs with POPC: DLPE mixture, the yields of vesicles suggested that the extra DLPE could enhance the vesicle formation at a ratio of around 6:4 and 7:3, which is comparable to the concentration with pure DLPE (Figure 2a). We obtained the highest yield vesicle product of 6.68 ± 0.78 × 10 9 vesicles/mL in the ratio of 6:4 (Table 3). When DOPC was mixed with DLPE in a ratio of 7:3, the mixture reached the highest yields of 6.94 ± 0.60 × 10 9 vesicles/mL, which was a little higher than the mix in the ratio of 8:2 (Figure 2a). Once DMPC mixed with DLPE in the proportion 6:4, the concentration of produced GV (5.68 ± 0.14 × 10 9 vesicles/mL) was obviously higher than GVs in the ratio of 7:3, 8:2, and 9:1 (Figure 2a). The fluorescence microscope showed an apparent

Size Distribution-based Stability Analysis of GVs via IFC
Since IFC can measure the size of single particles, we compared size distributions among GV composition with higher yields: POPC, DOPC, DLPE, DOPC: DLPE (7:3), and POPC:DLPE (6:4). For every sample, 50,000 particles were registered regardless of the concentrations. Purity and encapsulation efficiency were defined as the ratio of GVs excluding beads and other byproducts such as oil droplets in all obtained objects and the ratio of GVs containing calcein over all GVs, respectively. As listed in Table 3, the mean sizes of POPC and DOPC vesicles were almost identical (~4.1 μm), which are slightly smaller than DMPC (~4.7 μm). Comparatively, DLPE-derived GVs are the smallest (~3.7 μm). When mixed with DLPE, PCs/DLPE also generate smaller GVs, ranging from 3.4 to 3.8 μm. As for the encapsulation efficiency, although internal calcein can leak from inside

Size Distribution-Based Stability Analysis of GVs via IFC
Since IFC can measure the size of single particles, we compared size distributions among GV composition with higher yields: POPC, DOPC, DLPE, DOPC: DLPE (7:3), and POPC:DLPE (6:4). For every sample, 50,000 particles were registered regardless of the concentrations. Purity and encapsulation efficiency were defined as the ratio of GVs excluding beads and other byproducts such as oil droplets in all obtained objects and the ratio of GVs containing calcein over all GVs, respectively. As listed in Table 3, the mean sizes of POPC and DOPC vesicles were almost identical (~4.1 µm), which are slightly smaller than DMPC (~4.7 µm). Comparatively, DLPE-derived GVs are the smallest (~3.7 µm). When mixed with DLPE, PCs/DLPE also generate smaller GVs, ranging from 3.4 to 3.8 µm.
As for the encapsulation efficiency, although internal calcein can leak from inside of the GVs to outside solutions, all GVs tested have~90% or more vesicles encapsulated indicating that the w/o emulsion method is suitable for encapsulating materials inside of GVs.
We then investigated the dynamics of the concentration (Figure 3a) and size distribution (Figure 3b) of the above GVs (with or without calcein encapsulation) stored at 4 • C for Life 2021, 11, 223 9 of 13 indicated time periods up to one month. IFC analysis results showed that the yield of DLPE vesicles reduced almost by half at 6 h, and the concentration was the lowest at 3 d, to~10%, indicating the instability of DLPE-derived GVs. The overall reduction of GV concentrations with DOPC or POPC fluctuated at a range of 20% to 40%. For GVs of DOPC:DLPE (7:3), the trend of the concentration changes at different time points is similar to DOPC vesicles. For GVs of POPC:DLPE (6:4), it is interesting to observe that the yields dropped to almost 20% at 12 h. Afterward, it increased to 1.8-fold at 1 w and then started to decline again ( Figure 3a). As shown in Figures 3b and 4, there was an apparent increase in the population of smaller size POPC:DLPE (6:4)-derived vesicles starting from 1 week, which could be the consequence of the long-term storage at 4 • C. For the whole size distribution of GVs, most GVs produced with DLPE, DOPC, POPC, and DOPC:DLPE were almost the same during the 1-month storage (Figure 3b). Collectively, except for DLPE, GVs produced with DOPC, POPC, DOPC: DLPE (7:3), and POPC: DLPE (6:4) were stable. To show a comparative analysis between IFC and microscopy, we further provide a series of observations under the confocal microscope, as shown in Figure 4. Basically, the fluorescence images could not give quantitive results, but rather trends for quality control. As experienced in our study, the results from microscopy images are greatly influenced by the resolution of the machine and the spot or area observed. For instance, fluorescent  at 0 h (Figure 4), probably due to the heterogeneity of GVs when we prepared the glass sample. Another example is that GVs of POPC:DLPE (6:4) became smaller at 2 weeks and 1 month, but the concentrations showed almost no significant difference to the control (Figures 3a and 4).

Discussion
Recently, there has been a lively interest in using GVs to build either protocells or artificial cells that mimic the biological functions of cells [4,8,42]. The broad application prospects of GVs incentivizes the development of GV preparation methods and state-ofthe-art analytical approaches [6]. The w/o inverted emulsion-based process, compared with others such as gentle hydration, electro-formation or microfluidic, has benefits of high encapsulation efficiency and yield of uni-lamellar vesicles with a wide range of lipid compositions [15,27,28]. In this work, using the w/o inverted emulsion-based method, we prepared GVs with different kinds of phosphatidylcholines (PCs) and phosphatidylethanolamines (PEs), which are two major phospholipids of the plasma membrane. Once we prepared the GVs, we firstly used a fluorescence microscope to observe them. Based on microscopy results, we can only detect whether or not the GVs formed and estimate the rough concentrations.
However, following the analysis of the imaging flow cytometer, which combines high-resolution microscopy technology with traditional flow cytometry [15], we can perform extensive qualitative and quantitative imaging analysis of particles in a straightforward manner. The vesicle and non-vesicle particles, such as oil droplets or vesicle-oil mixture, and vesicles with or without encapsulations, can be divided using, via imaging, all vesicles at a single particle level. Moreover, vesicle shape changes such as budding or To show a comparative analysis between IFC and microscopy, we further provide a series of observations under the confocal microscope, as shown in Figure 4. Basically, the fluorescence images could not give quantitive results, but rather trends for quality control. As experienced in our study, the results from microscopy images are greatly influenced by the resolution of the machine and the spot or area observed. For instance, fluorescent images showed that the yield of GVs of DOPC:DLPE (7:3) at 1 month seems higher than at 0 h ( Figure 4), probably due to the heterogeneity of GVs when we prepared the glass sample. Another example is that GVs of POPC:DLPE (6:4) became smaller at 2 weeks and 1 month, but the concentrations showed almost no significant difference to the control (Figures 3a and 4).

Discussion
Recently, there has been a lively interest in using GVs to build either protocells or artificial cells that mimic the biological functions of cells [4,8,42]. The broad application prospects of GVs incentivizes the development of GV preparation methods and state-of-theart analytical approaches [6]. The w/o inverted emulsion-based process, compared with others such as gentle hydration, electro-formation or microfluidic, has benefits of high encapsulation efficiency and yield of uni-lamellar vesicles with a wide range of lipid compositions [15,27,28]. In this work, using the w/o inverted emulsion-based method, we prepared GVs with different kinds of phosphatidylcholines (PCs) and phosphatidylethanolamines (PEs), which are two major phospholipids of the plasma membrane. Once we prepared the GVs, we firstly used a fluorescence microscope to observe them. Based on microscopy results, we can only detect whether or not the GVs formed and estimate the rough concentrations.
However, following the analysis of the imaging flow cytometer, which combines highresolution microscopy technology with traditional flow cytometry [15], we can perform extensive qualitative and quantitative imaging analysis of particles in a straightforward manner. The vesicle and non-vesicle particles, such as oil droplets or vesicle-oil mixture, and vesicles with or without encapsulations, can be divided using, via imaging, all vesicles at a single particle level. Moreover, vesicle shape changes such as budding or bursting, vesicle fusion, or fission can also be detected and analyzed accordingly, depending on the experimental design. More importantly, the concentration, the size or size distribution, and the encapsulation efficiency of GVs can be measured accurately, which increase reproducibility significantly.
It has been reported in part that PC and PE, mainly egg PC [31,[35][36][37], DOPC [33,36,37,43], POPC [17,44,45], DMPC [34,46], and DPPE [46] can be used to produce GVs by gentle hydration, electro-formation, or the w/o emulsion transfer method. By the gentle hydration method, DOPC/DOPE-PEG2000 [47], POPC/DPPE-PEG2000 [44] can form GVs; e.g., PC/POPE can prepare GVs by the electro-formation method [35]. So far, there is a lack of comprehensive study aiming at producing GVs with various kinds of PC and PE lipids. Here, we have prepared GVs with various kinds of single PC, PE, and PC:PE mixtures. We demonstrate that PCs including POPC, DOPC, DMPC, PEs containing DLPE, DMPE, and POPC:DLPE mixture, DOPC:DLPE mixture, and DMPC:DLPE mixture can yield a high production of GVs. The concentration and the size distribution of these GVs are relatively stable during lengthy time storage. Considering that cylindrical PCs and cone-shaped PEs promote vesicle-vesicle hemi-fusion [48], our findings will aid in vesicle preparation research, vesicle-vesicle fusion studies, and further developments of GVs in artificial membrane and artificial cell research.
Supplementary Materials: The following are available online at https://www.mdpi.com/2075-172 9/11/3/223/s1, Figure S1: The fluorescence microscope images of GVs prepared with all PCs and PEs used in this study, Figure S2: The fluorescence microscope images of GVs prepared with PCs: DLPE, Table S1: Characterization of vesicles produced with other lipids resulting lower yields.