Development and single‐particle analysis of hybrid extracellular vesicles fused with liposomes using viral fusogenic proteins

Extracellular vesicles (EVs) have potential biomedical applications, particularly as a means of transport for therapeutic agents. There is a need for rapid and efficient EV‐liposome membrane fusion that maintains the integrity of hybrid EVs. We recently described Sf9 insect cell‐derived EVs on which functional membrane proteins were presented using a baculovirus‐expression system. Here, we developed hybrid EVs by membrane fusion of small liposomes and EVs equipped with baculoviral fusogenic proteins. Single‐particle analysis of EV‐liposome complexes revealed controlled introduction of liposome components into EVs. Our findings and methodology will support further applications of EV engineering in biomedicine.

Extracellular vesicles (EVs) have potential biomedical applications, particularly as a means of transport for therapeutic agents. There is a need for rapid and efficient EV-liposome membrane fusion that maintains the integrity of hybrid EVs. We recently described Sf9 insect cell-derived EVs on which functional membrane proteins were presented using a baculovirusexpression system. Here, we developed hybrid EVs by membrane fusion of small liposomes and EVs equipped with baculoviral fusogenic proteins. Single-particle analysis of EV-liposome complexes revealed controlled introduction of liposome components into EVs. Our findings and methodology will support further applications of EV engineering in biomedicine.
Genetic engineering techniques are often used to control EV functions. By gene transfection modifications of parent cells, target proteins or peptides can be presented on released EVs, whereas bioactive cargos (e.g. nucleic acids or proteins) are loaded inside the EVs [6]. These engineered EVs have been used in applications including drug delivery [7], vaccines [8], and biosensors [9]. In addition to genetic engineering methods, EVs have been modified by chemical and physical complexation with functional biomaterials [10]. For example, pH responsiveness [11], surface charge [12,13], magnetic-responsiveness [14], and targeting ability [15] have been added to EVs.
Liposomes are excellent candidates for EV modification because liposomal components (e.g. functional lipids or encapsulated drugs) can be introduced to EVs through membrane fusion. EVs and liposomes have been fused by simple mixing [16], freeze-thaw processes [17], and poly(ethylene glycol) addition [18]. However, the fusion efficiency is sometimes low and there are concerns involving damage to hybrid EV membranes or contamination with exogenous polymer. Thus, there is a need for a more rapid and efficient membrane fusion pathway that does not compromise the integrity of hybrid EVs. Furthermore, it is important to analyze the fusion process at the single-particle level because the membrane composition of EVs is heterogeneous and the fusion behavior is likely to be different for each EV [19,20].
Our research focuses on viral fusogenic proteins to generate EV-liposome hybrids. Envelope viruses infect host cells by fusogenic proteins that have rapid environmentally responsive fusion functions (e.g. influenza hemagglutinin and human T-cell leukemia virus envelope protein) [21]. We recently developed Sf9 insect cell-derived EVs on which functional membrane proteins were presented using a baculovirus-expression system [22]. Programmed cell death 1 (PD-1) membrane proteins were expressed on the EV surface while maintaining the ability to bind to ligand proteins and ligand-expressing cancer cells. In addition, the baculoviral envelope protein gp64 was expressed on the surface of PD-1-expressing EVs (PD-1 EVs). Gp64 is known to exhibit a membrane fusion function under acidic conditions and is critical in baculovirus infection and budding [23]. The gp64 present in virus-like particles from Sf9 insect cells is able to mediate membrane fusion with giant liposomes [24][25][26]. Although this fusion method is useful for the preparation of hybrid proteoliposomes for biomedical applications [25,26], the fusion behavior is not fully understood at singleparticle level. Here, we report a detailed evaluation of pH-responsive fusion of small liposomes and Sf9derived EVs equipped with fusogenic gp64 by lipid mixing and single-particle analyses.

Construction of recombinant baculoviruses
Recombinant baculoviruses encoding PD-1 mutant or Connexin 43 (Cx43)-enhanced green fluorescence protein (EGFP) were constructed using the Bac-to-Bac Baculovirus Expression System, as described previously [22]. Briefly, pFastBac1 plasmid carrying the membrane protein sequence of interest was transformed into DH10Bac Escherichia coli containing Bacmid and a helper plasmid encoding a transposase gene. The target gene located between Tn7 transposon sequences was transposed into the Bacmid. Colonies containing the recombinant bacmid were identified by blue/white selection and the Bacmid was isolated using a PureLink HiPure Plasmid Miniprep Kit (Invitrogen). Sf9 cells were transfected with the recombinant Bacmid using Cellfectin II Reagent (Invitrogen) and incubated at 27°C for 5 days. The supernatant containing P1 viruses was collected and the viral concentration was amplified three times. The viral titers were then determined using a BacPAK Baculovirus Rapid Titer Kit (Takara Bio USA, Inc., Palo Alto, CA, USA).

Isolation of EVs
Sf9-derived EVs were isolated using a previously reported method [22]. Briefly, 4.0 9 10 5 Sf9 cellsÁmL À1 were maintained in Sf-900 III SFM medium (Invitrogen) overnight at 27°C. The budded virus (BV) suspension was added at multiplicity of infection of 0.5 and incubated at 27°C for 96 h. The culture medium was centrifuged at 500 g for 5 min and 2000 g for 10 min at 4°C, followed by 0.22-µm filtration. The supernatant was ultra-centrifuged at 100 000 g for 70 min at 4°C and the resultant pellet was resuspended in phosphate-buffered saline. The suspension was ultra-centrifuged at 40 000 g for 30 min at 4°C in a stepwise sucrose density gradient [10%, 15%, 20%, 25%, and 30% sucrose (w/v) in phosphate-buffered saline buffer]. The upper fraction (containing EVs) and lower fraction (containing BVs) were collected separately. The EV protein concentration was estimated using a Pierce BCA Protein Assay Kit (Thermo Scientific, Waltham, MA, USA).

Liposome preparation
DOPC, DOPS, NBD-DOPE, Rho-DOPE, and Cy5-DOPE were mixed in chloroform into glass microtubes at various molar ratios. The solutions were evaporated under flowing argon gas, resulting in the formation of lipid film. The films were placed in a desiccator in vacuo overnight to completely remove chloroform. The film was hydrated by adding 250 µL of buffer [20 mM CH 3 COOH/CH 3 COONa (pH 4.5) or 10 mM Tris-HCl (pH 7.5)] and incubated overnight at 27°C. The suspension was extruded through a 100-nm pore polycarbonate membrane using a mini-extruder (Avanti Polar Lipids). The lipid concentration was measured using the Phospholipid C-Test (Wako, Osaka, Japan), which quantifies choline produced by phospholipase D activity.

Transmission electron microscopy (TEM)
Suspensions of EVs, BVs or hybrid EVs were placed on a copper grid coated with a formvar membrane for 5 min. After suspension removal, 1% phosphotungstic acid solution was placed on the grid for 5 min and then removed. Samples were observed using an HT7700-TEM (Hitachi, Tokyo, Japan) at an accelerating voltage of 100 kV.

Nanoparticle tracking analysis
The size distributions of EVs, liposomes, or hybrid EVs were measured using a NanoSight LM10 (NanoSight, Amesbury, UK). Samples were measured with a 532 nm wavelength laser at 25°C and analyzed using NANOSIGHT NTA, version 2.3 (NanoSight).

Western blotting
EV samples solubilized with sodium dodecyl sulfate buffer were separated by 12.5% polyacrylamide gel and transferred to poly(vinylidene difluoride) membranes. The membranes were blocked with PVDF Blocking Reagent for Can Get Signal (Toyobo Co., Ltd, Osaka, Japan), then probed with primary antibody to gp64 (Santa Cruz Biotechnology), PD-1 (Abcam, Cambridge, UK) or Cx43 (BD Transduction Laboratories, KY, USA) diluted at 1 : 1000 in Can Get Signal Solution 1 (Toyobo Co., Ltd) at 4°C overnight. After they had been washed in Tris-buffered saline with Tween, the bands were identified by reaction with horseradish peroxidase-conjugated goat anti-mouse IgG (Santa Cruz Biotechnology) diluted at 1 : 2000 in Can Get Signal Solution 2 (Toyobo Co., Ltd) at room temperature for 1 h. After a second round of washing in Tris-buffered saline with Tween, the membranes were reacted with ECL Western Blotting Detection Reagents (GE Healthcare, Milwaukee, WI, USA) and the band signals were visualized using LAS-4000 (GE Healthcare).

Statistical analysis
Experimental data were statistically evaluated using oneway factorial analysis of variance followed by Tukey's multiple comparisons test, or a two-tailed Welch's t-test. An adjusted P < 0.05 was considered statistically significant. All statistical analyses were performed using PRISM, version 9 (GraphPad Software Inc., San Diego, CA, USA).

Results and Discussion
Lipid mixing assay of membrane fusion between liposomes and fusogenic EVs Ultra-centrifugation was used to collect EVs from Sf9 insect cells that had been infected with recombinant baculoviruses. Basic characterization of Sf9-derived EVs by TEM revealed successful isolation of vesiclelike nanoparticles (Fig. S1A). Western blotting analysis confirmed the expression of the gp64 viral fusogenic protein on EVs (Fig. S1B). The EV fusion ability was evaluated by a lipid mixing assay based on FRET [27]. Membrane fusion of EVs and FRET liposomes containing 7-nitro-2-1,3-benzoxadiazol-4-yl (NBD) and rhodamine fluorescence was monitored on the basis of NBD fluorescence recovery by dilution of lipids (Fig. 1A). When FRET liposomes and EVs were mixed under acidic conditions (pH 4.5), NBD fluorescence recovered rapidly beginning 1 min after mixing; it eventually recovered to approximately 17% of the maximum fluorescence value (Fig. 1B, blue dots). By contrast, NBD fluorescence did not show any recovery under neutral conditions (Fig. 1B, green squares) or under acidic conditions in the absence of EVs (Fig. S2). Moreover, greater NBD fluorescence recovery efficiency was observed as the final concentration of EVs increased (Fig. 1C). To evaluate the specificity of gp64 fusion, an inhibition experiment was carried out using an anti-gp64 antibody. As shown in Fig. 1D, the elimination of FRET under acidic conditions was suppressed depending on the concentration of anti-gp64 antibody. By contrast, when an isotype control antibody was used, FRET elimination was not suppressed (Fig. 1D). These results indicated that the gp64 viral protein expressed on EVs could be activated under acidic conditions and induce membrane fusion with liposomes.

Nanoparticle tracking analysis and TEM observation of hybrid EVs
To investigate the fusion of EVs and liposomes, the particle size distribution of hybrid EVs was determined by nanoparticle tracking analysis. The respective mean particle sizes of EVs and liposomes before mixing were 160 nm and 164 nm (Fig. 2A). The particle sizes of EVs and liposomes mixed under acidic conditions increased and demonstrated a broad distribution (mean 252 nm) compared to EVs and liposomes mixed under neutral conditions (Fig. 2B). In addition, the number of detected particles decreased under acidic conditions. Because gp64 on EVs can mediate self-fusion between EVs, the particle size distribution was measured after incubation of PD-1 EVs under acidic conditions (Fig. S3). Compared with neutral conditions ( Fig. 2A), the mean particle size increased slightly (184 nm). Although this may indicate self-fusion between EVs, the change in the particle size distribution was not as drastic as that of the mixture of EVs and liposomes (Fig. 2B). The fusion between nanoparticles was directly observed by TEM (Fig. 2C). Under acidic conditions, hemifusion intermediates and fully fused large vesicles were observed. By contrast, under neutral conditions, no interparticle There are several methods for directly visualizing the fusion of nanoparticles, such as electron microscopic observation of EVs and liposomes labeled with gold nanoparticles [18], or realtime observation using high-speed atomic force microscopy [28,29]. Therefore, to confirm the formation of hybrid EVs, it is necessary to show the complexation of components of EVs and liposomes at the single-particle level.

Single-particle analysis of membrane fusion by IFC
Next, the complexation of EVs and liposomes at the single-particle level was evaluated using IFC, a highthroughput analysis method that combines the functions of flow cytometry and fluorescence microscopy. Because of its high sensitivity, IFC enables fluorescence analysis of EVs at the single-particle level, which has been difficult thus far [30,31]. Sf9-derived EVs were labeled with CFSE to evaluate their complexation with Cy5-labeled liposomes. CFSE serves to label proteins inside of EVs; it fluoresces following hydrolysis by esterase in EVs. CFSE-labeled EVs were successfully detected by IFC, whereas they were absent in control groups (Fig. S4). CFSE-labeled EVs and Cy5 liposomes were mixed under various conditions and particle populations were analyzed by IFC using gating strategy shown in Fig. S5. Under acidic conditions, there was a greater proportion of particles in the CFSE and Cy5 double-positive region, following changes in the EV-liposome ratio (Fig. 3A,B). By contrast, under neutral conditions, the proportion of particles present in the double-positive region was low (Fig. S6, S7). The proportion of double-positive particles is considered the fusion efficiency at the  (Fig. 3D). Furthermore, Cy5 fluorescence intensity in the double-positive region increased under acidic conditions, depending on liposome concentration (Fig. S8). This result indicates that the surface modification of hybrid EVs at single-particle level can be controlled by changing the EV-liposome ratio. Finally, Sf9-derived EVs expressing enhanced green fluorescence protein-conjugated Cx43 (i.e. Cx43-EGFP) were used to evaluate the co-localization of EV and liposome membranes. Cx43 is a four-pass transmembrane protein and a component of gap junctions [25]. Cx43-EGFP EVs were isolated from Sf9 cells infected with recombinant baculoviruses; the expression of gp64 and Cx43 on Cx43-EGFP EVs was analyzed by western blotting (Fig. S1B). Membrane fusion of Cx43-EGFP EVs and Cy5 liposomes was visualized at the single-particle level by IFC. The particle populations shifted toward the EGFP and Cy5 double-positive region depending on changes in the EV-liposome ratio under acidic conditions (Fig. 4A,B;  Fig. S9, S10). EGFP and Cy5 fluorescence spots were co-localized under acidic conditions, but not under neutral conditions (Fig. 4C,D). Moreover, liposome concentration-dependent Cy5 fluorescence transfer to Cx43-EGFP EVs suggested the potential to freely perform surface engineering of hybrid EVs (Fig. S11). These results suggest that EVs and small liposomes fused under acidic conditions to form hybrid EV particles in which membrane protein components were transferred. Though the method of fusion between EVs and liposomes has been verified by various approaches, some problems such as contamination of exogenous materials (e.g. polyethylene glycol) and a long fusion reaction time (e.g. 2-12 h) remain [16,18]. By contrast, the fusion method in the present study is a very simple and rapid process of incubation for 30 min under acidic conditions. Furthermore, single-particle analysis of membrane fusion by IFC enabled the determination of detailed EVliposome fusion efficiency and surface composition of hybrid EVs, which have not been clarified so far.

Conclusions
In conclusion, we have developed hybrid EVs by membrane fusion of small liposomes and EVs equipped with baculoviral fusogenic proteins. The gp64mediated fusion of EVs and liposomes was analyzed at the single-particle level, which revealed controlled introduction of liposome components into EVs. The achievement of highly efficient EV-liposome membrane fusion and the evaluation methods used in the present study will support further applications of EV engineering in biomedicine. We expect that the hybrid EVs will serve as new drug delivery carriers that achieve cell membrane fusion by means of fusogenic gp64 and active targeting by means of functional membrane proteins such as PD-1 and Cx43.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article.    Fig. S4. Dot plots and representative fluorescence images determined by IFC of CFSE-labeled EVs (A) and control samples (B-D). Addition of CFSE to bovine serum albumin (BSA) protein with esterase activity showed very high background fluorescence, although ultra-filtration purification removed most background fluorescence. Fig. S5. Gating strategy for detection of fluorescence nanoparticles by IFC. Plots were obtained in the 1 µM liposomes (pH 4.5) condition; the gating process was similar for other conditions. (A) Removal of speed beads using channels 1 (bright-field) and 6 (side scatter). (B) Removal of fluorescent noise for channels 2 and 5. Finally, 10 000 particles were acquired and analyzed in the R2 region.