Rapid isolation and cryo‐EM characterization of synaptic vesicles from mammalian brain

Synaptic vesicles (SVs) store and release neurotransmitters at chemical synapses. Precise regulation of SV trafficking, exocytosis and endocytosis is crucial for neural transmission. Biochemical characterization of SVs, which is essential for research into neurotransmitter uptake and release, requires effective in vitro isolation methods. Here, we describe an improved and simple purification protocol for isolating SVs from mouse brain within 6 h, achieving a yield of approximately 0.4 mg of SVs per single brain. The use of track‐etch membrane filtration and iodixanol cushion ensured the uniform morphology of SVs and low contaminants in the sample. Cryo‐electron microscopy was used to show that the in vitro isolated SVs retained intact membrane‐associated proteins, and observation of SVs in hippocampal neurons using cryo‐electron tomography confirmed the abundance of protein coating. Thus, our protocol allows effective isolation of SVs from small volumes of mammalian brain tissue, and the properties of the isolated SVs are close to those in vivo, making them suitable for biochemical analysis.

Synaptic vesicles (SVs) store and release neurotransmitters at chemical synapses. Precise regulation of SV trafficking, exocytosis and endocytosis is crucial for neural transmission. Biochemical characterization of SVs, which is essential for research into neurotransmitter uptake and release, requires effective in vitro isolation methods. Here, we describe an improved and simple purification protocol for isolating SVs from mouse brain within 6 h, achieving a yield of approximately 0.4 mg of SVs per single brain. The use of track-etch membrane filtration and iodixanol cushion ensured the uniform morphology of SVs and low contaminants in the sample. Cryoelectron microscopy was used to show that the in vitro isolated SVs retained intact membrane-associated proteins, and observation of SVs in hippocampal neurons using cryo-electron tomography confirmed the abundance of protein coating. Thus, our protocol allows effective isolation of SVs from small volumes of mammalian brain tissue, and the properties of the isolated SVs are close to those in vivo, making them suitable for biochemical analysis.
Synaptic vesicles (SVs) are special organelles at the presynaptic terminal that store and release neurotransmitters. Neurotransmitters encased in SVs are released into the synaptic cleft through SVs exocytosis, where they are recognized by corresponding receptors on the postsynaptic membrane to accomplish the transmission of neural signals [1,2]. Reinternalization and recycling of SVs follow the exocytosis.
Comprising specific compartments in the mammalian brain, the important role of SVs is structurally and functionally predominated via complicated protein/component and interactions. Neurotransmitter transporters are transmembrane proteins that are distributed on SVs and responsible for diverse substrates [3], such as vesicular glutamate transporters (vGLUT1-3), vesicular acetylcholine transporter (vAChT) and vesicular GABA transporter (vGAT). Vacuolar-type H+-ATPase (v-ATPase) resides in SVs to maintain an electrochemical gradient. Additionally, various SV-associated proteins also orchestrate vesicle behavior at the presynaptic side, including trafficking, as well as exo-and endocytosis. For example, the SNARE complex is well characterized with respect to facilitating membrane fusion [4]. Recent studies based on SV mass spectrometry provided a complicated composition, and more than 1000 proteins were identified [5]. Although many proteins have been identified, little is known about the membrane structure of SVs, such as the concentration Abbreviations cryo-EM, cryo-electron microscopy; cryo-ET, cryo-electron tomography; Rab3A, Ras-related protein Rab-3A; SEC, size exclusion chromatography; SV2C, synaptic vesicle glycoprotein 2C; SV, synaptic vesicle; vAChT, vesicular acetylcholine transporter; vGAT, vesicular GABA transporter; vGLUT, vesicular glutamate transporter; VPP, Volta phase plate. of membrane-associated proteins and the surface density of important proteins. Thus, in vitro purification of SVs is beneficial for biochemical analysis and comprehensive understanding the regulation of SVs.
Synaptic vesicles are spheroidal organelles of 40-50 nm in diameter and are highly abundant in brain tissue. SV-associated proteins contribute approximately 5% of the total content of the mammalian central nervous system [6]. Current isolation protocols based on multistep centrifugation and column chromatography [6][7][8] or beads based immunoaffinity [9,10] have been reported and directed a successful production of SVs from various tissues or cultured cells. However, the purity and yield are uncompromising. Here, we developed a rapid and simple purification strategy with effective filtration and gentle centrifugation to obtain considerable SVs with high purity. The stable SVs also allowed a suitable cryo-electron microscopy (cryo-EM) sample preparation and exhibited an intact, proteinrich membrane structure. In situ morphological observation of SVs via cryo-electron tomography (cryo-ET) confirmed that our protocol enables a rapid and highquality obtention of SVs from the mammalian resources.

Materials and methods
All animal experiments were approved by Institutional Animal Ethics Committee of Huazhong University of Science and Technology (License No. SYXK2021-0057).

Isolation of synaptic vesicles from mouse brain
Synaptic vesicles were isolated from the mouse brain. Here, we present an improved and simple purification protocol for isolating SVs from mouse brain within 6 h, which finally allows approximately 0.4 mg of SVs per single brain.
The detailed procedures: 1 Remove the brain from one mouse (6-8-

Negative staining
Synaptic vesicles sample was applied to a carbon-coated grid and stained with 2% uranyl acetate. Imaging was performed using a Tecnai Spirit electron microscope (FEI, Hillsboro, OR, USA) operated at 120 kV.

Cryo-EM sample preparation
For cryo-EM, using a Vitrobot Mark IV (FEI), the SV sample was applied to a glow discharged, Quantifoil R1. Hippocampal neurons were removed from P0 mice and were digested with papain for 30 min in a water bath at 37°C. Next, the tissues were gently washed twice with trituration solution and mechanically triturated 15-20 times with a 1-mL tip to obtain a single cell suspension (avoiding bubbles) [trituration solution (total 5 mL): Glutamax 0.025 mL, fetal bovine serum 0.5 mL, 10% BSA/HBSS 0.5 mL, DNase 0.05 mL in HBSS]. The sample were centrifuged at 200 g for 5 min at 4°C and the supernatant was discarded. Dissociated cells were suspended in 1 mL of plating medium (Dulbecco's modified Eagle's medium with 10% fetal bovine serum, catalog. no. 11995065; Gibco) and applied to the cell counter for counting (trypan blue 1 : 1). Dilute the cell suspension with plating medium and seed onto the coverslip with gold EM grids at a low density of 70-100 K cells per dish. Eight hours after seeding, the plating medium was replaced by the culture medium. The culture medium was NeuroBasal (Gibco) supplemented with 2% B27 (Invitrogen, Carlsbad, CA, USA) plus 0.5 mM Glutamax (Invitrogen). Subsequently, half of the culture medium was replaced with fresh culture medium every 3-4 days. Primary neurons were maintained in incubators at 37°C in 5% CO 2 until 14-18 days in vitro. Imaging of cryo-EM was performed with a Titan Krios electron microscope (FEI) operated at 300 kV. Tilt-series were collected in a dose-symmetric tilting scheme from À51°to +51°with a step size of 3°using SERIALEM [11]. Tilt-series were collected at a magnification of 53 0009, corresponding to a pixel size of 2.58 A. The total dose per tilt series was approximately 60 e À Á A À2 .

Immunoblotting
Samples were prepared with SDS loading buffer and resolved by 12% SDS/PAGE, then transferred for 1 h to 0.45-mm poly(vinylidene difluoride) membranes. The membranes were blocked with 5% milk in TBST (i.e. Tris-buffered saline-Tween 20), then incubated with antibodies in 3% milk TBST for 1 h at room temperature. Antibodies were used: SV2C (catalog. no. ab33892; Abcam, Cambridge, UK), vGlut1 (catalog. no. ab227805; Abcam) and Rab3A (catalog. no. ab3335; Abcam). Following incubation, membranes were washed three times with TBST and incubated with the corresponding secondary antibodies in 3% milk. After the membranes were washed three times again with TBST, they were visualized using enhanced chemiluminescence. For the gold nanoparticles immunoassay, V-ATPase E1 Polyclonal Antibody (catalog. no. PA5-29899; Thermo Fisher; Waltham, MA, USA) and 10 nm BSA gold tracer (catalog. no. 25486; Electron Microscopy Sciences) were used to mark the samples.

Rapid and simple purification of synaptic vesicles
The isolation of SVs facilitates accurate biochemical characterization and investigations based on in vitro experiments, which are essential for understanding neurotransmitter transmission [12,13]. Several methods have been reported for the direct purification of SVs from the mammalian brain [6,7,14]; however, these methods usually require complicated purification processes, and the yield and quality of SVs need to be improved. To this end, we have optimized and established a simple protocol to obtain high-purity SVs within 6 h or less and without the need for specific instruments (Fig. 1). The protocol allows the extraction of up to 0.4 mg of qualified SVs from a single mouse brain. The obtained SVs exhibited complete morphology and a stable state in the buffer; additionally, efficient removal of organelles or cell debris and proteins/protein complexes provided a suitable sample for electron microscopy observation.
Synaptosomes contain numerous SVs, and suitable hypotonic conditions can promote the final yield. In our protocol, secondary homogenization of synaptosomes after static swelling greatly improved the level of SV release. To effectively remove cell debris and other membrane organelles, the sample was subject to a filter. Our protocol introduces a track-etch membrane with a completely identical pore size that strictly limits large components, enabling us to easily obtain SVs with a relatively uniform diameter, thus allowing for the omission of density gradient centrifugation. To remove small particles of soluble protein/protein complexes in the sample, SVs were enriched to the bottom of the centrifuge tube by ultracentrifugation. In this procedure, we found at least two problems ( Fig. 2A): (a) Although most of the protein contaminants were removed, the collected SVs still contained ribosomal particles, and (b) the pellet enriched by ultracentrifugation caused irreversible aggregation of vesicles and mechanical damage. To further improve the purity of SVs, we introduced a buffer-layer at the bottom of the centrifuge tube. For this step, 15% iodixanol was determined and glycerol was also the alternative substitute, which achieved the best balance between the yield and the purity of SVs.
Because size exclusion chromatography (SEC) is a widely used strategy for the final purification of SVs in various reported methods [6,14], we also performed a comparison. We found that, if the starting sample contains large membrane-structured components, the gel filtration column will be blocked, which raises a higher requirement for the sample to be separated. The column pressure and shearing force of the stationary phase resulted in the elution effect of SV surface  Fig. 1. Rapid workflow for isolation of SVs from the mouse brain. Detailed procedures are reported in the Materials and methods. proteins (Fig. 2B), which is unfavorable for maintaining the natural shape of SVs and subsequent biochemical analysis. Based on negative staining, the purified SVs that followed our protocol retained proteins at the membrane; however, vesicles from SEC purification had a smooth surface (Fig. 2B,C). Additionally, the synaptic vesicle protein levels of synaptic vesicle glycoprotein 2C (SV2C), vesicular glutamate transporter 1 (vGlut1) and Ras-related protein Rab-3A (Rab3A) from SEC and Iodixanol layer purification indicated marked differences in immunoblotting (Fig. 2D). To further assess the integrity of SVs, we performed a goldnanoparticles immunoassay against v-ATPase and the negative staining results exhibited efficient labeling (Fig. 2E). The measurement of the diameter of purified SVs also showed a major distribution (75%) at 30-50 nm (Fig. 2F).
Taken together, we describe here a simple and effective purification strategy that allows the eligible isolation of SVs from limited animal sources with uncomplicated experimental equipment.

Characterization of isolated synaptic vesicles via cryo-EM
Heavy-metal staining is a rapid method for observing cellular ultrastructure, although fixation and dehydration of vesicles limit the understanding of molecular details [15,16]. Recently, the development of cryo-EM hardware and software has extended the range of observations of biological samples and the application for protein complexes has established high-resolution structural processing [17]. Thus, we prepared a cryo-EM sample of in vitro purified SVs from the mammalian brain. The vitrified, fully hydrated imaging provided an ellipsoidal and spherical shape of SVs (Fig. 3A). To further visualize morphological information of SVs, we applied cryo-ET to our purified sample, and the generated tomograms showed a crowded distribution of membrane-associated proteins to SVs (Fig. 3B). This observation is consistent with the recently reported proteomics of SVs [5], indicating that complicated protein interaction and regulation are required for diversified functions of SVs at presynaptic terminal. However, the heterogeneous distribution of coating proteins also limits effective identification of the copy number and type for these proteins, which may require hardware and structural algorithm optimization.
Next, we investigated whether morphological information of in vitro purified SVs represents a real state in cells. Mouse hippocampal neurons were isolated and cultured at 13-17 days in vitro, and thinner synapse regions allowed us to prepare eligible cryo-ET samples (Fig. 3C). Consistent with the above results, vesicles at the presynaptic terminal also exhibited distinct coating proteins. Volta phase plate (VPP) has been reported to enhance contrast and signal-to-noise ratio of transmission electron microscopy records [18]; as expected, more structural details of SVs were provided with the utilization of VPP (Fig. 3D). Additionally, we also observed the interconnection of SVs at the active zone, consistent with previous reports [19,20], and 'short filaments' of approximately Volta phase plate

Cryo-ET NS (A) (B)
Cryo-ET Cryo-EM 8-12 nm were clearly visualized between two vesicles (Fig. 3E). Surprisingly, in our in vitro purified SV, the connected SVs can also be found in well-dispersed transmission electron microscopy images (Fig. 3E). Thus, although there are many successful protocols for the purification of SVs in vitro, the integrity of vesicles is required for subsequent biochemical characterization, and the rapid isolation method that we proposed here can effectively retain the comprehensive configuration of intact SVs.

Discussion
Neurotransmitter release is crucial for neural information delivery. Neurotransmitters are released from the Ca 2+ ions-stimulated presynaptic terminal and captured by the postsynaptic side receptors to complete signal transmission [1,2]. Synaptic vesicles are compartments stored with various neurotransmitters in the brain. Comprising the distal connection of neurons, synapses can be considered as a partially independent release system, which has high requirements for kinetics regulation of SVs. The exocytosis of SVs, including docking, priming and fusion, also requires precise control to respond to rapid stimuli. Various proteins/components and complicated interactions are the main participants in the above events.
Purification of SVs in vitro is essential for understanding intracellular processes, which allows for quantitative investigation by biochemical methods. In the past decades, many isolation protocols of vesicles have been established from cultured cells and mammalian brain, and detailed procedures have been modified to meet different needs [6,7,9]. However, multistep purifications reduce the yield and cause unpredictable damage to the vesicles. Here, we provide a simple method for isolating SVs from mouse brain, which can obtain qualified biological samples for biochemical characterization without density gradient centrifugation or column chromatography. Based on the observation of transmission electron microscopy, the isolated SVs retained intact shape and are close to the real situation in the synapse. Moreover, we also show that the isolated SVs were covered by rich proteins, indicating that SVs have unknown functions at the presynaptic terminal that need precise regulation. This observation was also verified by the cryo-ET from neurons. Specifically, the filamentous connector between SVs has been proposed for the mobility and release of vesicles [19,21,22]. Of our isolated SVs, the connected vesicles were also reserved in vitro following rapid and gentle purification, indicating adequate strength of the connectors. Although the records from cryo-EM provide abundant protein details, the heterogeneous profiles of vesicles-associated proteins also bring limitations to high-resolution structural assigned. Isolated SVs from our protocol will be suitable for comprehensive proteomic dissection, as well as for further analysis of proteins on SVs.