Abstract
Phospholipids (PLs) are asymmetrically distributed at the plasma membrane. This asymmetric lipid distribution is transiently altered during calcium-regulated exocytosis, but the impact of this transient remodeling on presynaptic function is currently unknown. As phospholipid scramblase 1 (PLSCR1) randomizes PL distribution between the two leaflets of the plasma membrane in response to calcium activation, we set out to determine its role in neurotransmission. We report here that PLSCR1 is expressed in cerebellar granule cells (GrCs) and that PLSCR1-dependent phosphatidylserine egress occurred at synapses in response to neuron stimulation. Synaptic transmission is impaired at GrC Plscr1−/− synapses, and both PS egress and synaptic vesicle (SV) endocytosis are inhibited in Plscr1−/− cultured neurons from male and female mice, demonstrating that PLSCR1 controls PL asymmetry remodeling and SV retrieval following neurotransmitter release. Altogether, our data reveal a novel key role for PLSCR1 in SV recycling and provide the first evidence that PL scrambling at the plasma membrane is a prerequisite for optimal presynaptic performance.
- calcium-regulated exocytosis
- compensatory endocytosis
- lipid scrambling
- neurons
- neurotransmission
- pHluorin
- PLSCR1
Significance Statement
During calcium-regulated exocytosis, phospholipids (PLs) like phosphatidylserine (PS) undergo dynamic remodeling. Phospholipid scramblase 1 (PLSCR1) belongs to a family of proteins able to randomize lipids at the cell surface in response to intracellular Ca2+ increases. Whether PLSCR1 and PS egress have a role during neurotransmission is unknown. We show that PLSCR1 expression is restricted to specific brain regions capable of sustaining neurotransmission during high firing rates. In the absence of PLSCR1, synaptic transmission is impaired, and both PS egress and synaptic vesicle endocytosis are hindered. This study highlights the pivotal role of PLSCR1 in regulating optimal presynaptic performance by redistributing PL at the plasma membrane to control compensatory endocytosis.
Introduction
Phospholipids (PLs) at the plasma membrane are asymmetrically distributed by energy-dependent transporters, which move phosphatidylethanolamine (PE) and phosphatidylserine (PS) from the extracellular to the cytoplasmic leaflet (flippases) and, conversely, phosphatidylcholine from the cytoplasmic to the extracellular leaflet (floppases). PS and PE are therefore virtually absent from the outer leaflet of the plasma membrane (Zachowski, 1993; Kobayashi and Menon, 2018; Clarke et al., 2020). This homeostatic and asymmetric lipid distribution is disrupted in several biological processes resulting in PS exposure at the extracellular leaflet. Cell surface PS exposure is massive and irreversible during events such as apoptosis or platelet activation which is required for thrombosis (Bevers and Williamson, 2016). In contrast during regulated exocytosis in either immune or neuroendocrine cells, PS egress is transient and strictly restricted to sites of membrane fusion (Martin et al., 2000; Vitale et al., 2001; Kato et al., 2002; Malacombe et al., 2006; Ory et al., 2013; Rysavy et al., 2014; Audo et al., 2017).
Whether massive or limited, the loss of plasma membrane asymmetry is mainly due to the activation of PL scramblases which randomize PLs across the plasma membrane (Daleke, 2003; Kobayashi and Menon, 2018). Phospholipid scramblase 1 (PLSCR1), originally isolated from erythrocytes, was the first scramblase identified able to reproduce lipid scrambling in response to Ca2+ increases when introduced into liposomes (Bassé et al., 1996; Zhou et al., 1997). PLSCR1 also controls PS egress in response to mast cell and chromaffin cell stimulation (Kato et al., 2002; Amir-Moazami et al., 2008; Smrž et al., 2008; Ory et al., 2013). However, preventing PS egress by suppressing PLSCR1 expression leads to different functional outcomes. For example, the release of secretory granule contents by exocytosis is reduced in mast cells (Amir-Moazami et al., 2008) but unaltered in chromaffin cells (Ory et al., 2013) which instead show defective compensatory endocytosis, preventing secretory granule protein retrieval (Ory et al., 2013). Therefore, the leaflet organization of the lipids at the plasma membrane and/or PLSCR1 controls exocytosis and/or endocytosis in secretory cells. However, the mechanism of lipid scrambling by PLSCR1 is currently a matter of debate. This is because PLSCR1 is a single-pass transmembrane protein that cannot form a channel to transport PLs and the deletion of Plscr1 gene in mice does not alter plasma membrane PS egress either in activated platelets or in response to apoptosis (Zhou et al., 2002). Nonetheless, the identification of additional scramblases, namely, TMEM16F and XKR-8 controlling extracellular PS exposure during platelet activation (Suzuki et al., 2010; Yang et al., 2012) and apoptosis (Suzuki et al., 2013), respectively, provides evidence that, despite their common ability to collapse lipid asymmetry, scramblases can regulate distinct mechanisms at the plasma membrane.
Neuronal communication relies on the release of neurotransmitters at synapses, which occurs by Ca2+-dependent exocytosis of synaptic vesicles (SVs), resulting in the insertion of SV membrane into the presynaptic plasma membrane. To support high rates of release during synaptic transmission, SV components must be retrieved rapidly by compensatory endocytosis to preserve plasma membrane homeostasis, to clear exocytic sites for the recruitment of release-ready SVs and to replenish the releasable SV pool (Wu et al., 2014; Maritzen and Haucke, 2018; Chanaday et al., 2019). Activity-dependent SV fusion and retrieval are spatially and temporally coupled within nerve terminals, and dysfunction in either process is increasingly associated with pathologies including epilepsy, autism, or intellectual disability (Bonnycastle et al., 2021). The molecular mechanism coupling endocytosis to exocytosis have been extensively studied in neurons (Maritzen and Haucke, 2018; Bolz et al., 2023; Wu and Chan, 2024). However, although exocytosis and endocytosis imply successive reorganization of PLs belonging to separate membrane compartments, little attention has been paid to the role of plasma membrane PLs in these events.
The present study sought to determine whether SV recycling relied on PLSCR1 and whether neurotransmission was influenced by its loss. Through a combination of approaches including immunofluorescence on primary cerebellar cultures, electron microscopy (EM), and electrophysiology on acute cerebellar slices from both wild-type (Plscr1+/+) and Plscr1 knock-out mice (Plscr1−/−), we revealed that PLSCR1 activity at synapses is essential for maintaining efficient neurotransmission. Moreover, optical live-cell imaging of pHluorin probes revealed that compensatory endocytosis was altered in Plscr1−/− neurons. Taken together, our data unveil a pivotal role of PLSCR1 in mediating stimulation-dependent PS egress at excitatory synapses. Furthermore, our study provides the first evidence of the impact of PLSCR1-dependent PS egress on facilitating SV recycling and sustaining synaptic transmission during periods of high-frequency stimulation.
Materials and Methods
Animals and genotyping
Neuronal cells and brain tissues were obtained from Plscr1−/− and wild-type mice. Plscr1+/− mice were purchased from Cryopréservation, Distribution, Typage et Archivage animal housed and raised at Chronobiotron UMS 3415. All mice were bred, handled, and maintained in agreement with European Council Directive 86/609/EEC and resulting French regulation. Animal work in Edinburgh was performed in accordance with the UK Animal (Scientific Procedures) Act 1986, under Project and Personal License Authority, and was approved by the Animal Welfare and Ethical Review Body at the University of Edinburgh (Home Office project license to M. Cousin, 70/8878). Animals were killed by Schedule 1 procedures in accordance with UK Home Office Guidelines. Animals used in France were handled according to French regulations. Postnatal Day (P) 6 to P7 mice pups were killed by anesthetic overdose, with death confirmed via the destruction of the brain, and adult mice were killed by cervical dislocation. Plscr1+/+ and Plscr1−/− mice were maintained as heterozygous breeding pairs and genotyped by duplex PCR using 5′-CTACACTGACCTTTAATCAGAGCAG-3′, 5′-CCATGTCTGCCCAAGTTCACTCTC-3′, and 5′-GCAGCGCATCGCCTTCTATC-3′ primers to detect the presence of a 261 bp or a 311 bp fragment for the wild-type or mutant allele, respectively. The PCR program was set at 95°C for 3 min: 30 cycles (95°C for 30 s, 60°C for 30 s,72°C for 30 s) followed by 10 min elongation at 72°C.
Cell culture, transfection, and plasmids
Cerebellar granule cell (GrC) cultures were prepared as previously described (Cheung and Cousin, 2011). Briefly, primary cultures of cerebellar neurons were prepared from the cerebella of 6–7-d-old C57BL/6J Plscr1−/− pups and their littermate controls. After removal, cerebella were placed in an HBSS/Hepes solution (HH, Table 1), minced up with tweezers, and then incubated in a trypsin solution (2 T; Table 1) at 37°C for 20 min. After digestion with trypsin, an equal volume of a neutralization (N) solution (Table 1) was added to the suspension, and the sample was centrifuged at 150 × g for 60 s. The supernatant was removed, and the pellet was suspended in 1 ml of N solution before being triturated using a 1,000 µl pipette until the solution reached homogeneity. The cell suspension was centrifuged at 340 × g for 2 min, and the pellet was resuspended in a 1.5 ml of DMEM/FCS medium (DFM; Table 1) and plated as one spot/well at a density of 5–10 × 106 cells/coverslip coated with poly-D-lysine (20 µg/ml) diluted in boric acid (100 mM), pH 8.5. After 1 h, wells were flooded with Neurobasal growth medium (Table 1) containing 25 mM KCl. The following day, cultures were further supplemented with 1 µM cytosine β-D-arabinofuranoside to inhibit glial cell proliferation. Four or five days after seeding, cells were transfected with Lipofectamine 2000 as described by Nicholson-Fish et al. (2015). Briefly, cells were preincubated in 2 ml of MEM (Thermo Fisher Scientific) in 10% CO2 for 30 min at 37°C and then transfected for 2 h with a complex containing 2 µl of Lipofectamine and 1 µg of the indicated plasmids/well. Cells were subsequently washed with MEM before replacement with conditioned Neurobasal Media. Cells were imaged 48 h post-transfection.
Solutions for CGN preparation
Synaptophysin–pHluorin (Syp-pH) was a gift from Prof. L. Lagnado (University of Sussex), and GFP-PLSCR1 plasmids were previously described (Granseth et al., 2006; Ory et al., 2013). The mCherry-PLSCR1 constructs were obtained by PCR using the mouse Plscr1 as a template. Amplified fragments were subcloned in pmCherry-C1 vectors between BglII and EcoRI restriction sites using forward 5′-CAGATCTGAAAACCACAGCAAGCAAAC-3′ and reverse primer 5′-GGAATTCTTACTGCCATGCTCCTGATC-3′.
Immunofluorescence, confocal microscopy, and image analysis
Neurons were fixed with ice-cold 4% (w/v) paraformaldehyde in phosphate-buffered saline (PBS) for 10 min at room temperature and permeabilized with 0.1% v/v Triton X-100 in PBS for 4 min. The cells were washed with PBS and blocked with 3% BSA in PBS and 5% goat serum (Sigma-Aldrich, G9023) for 1 h before being incubated with primary antibodies in 3% BSA in PBS for 1 h. Then cells were labeled with secondary antibodies coupled to fluorescent Alexa Fluor dyes (see Table 2 for antibody list and references). Actin was stained with tetramethyl-rhodamine B coupled phalloidin (TMR-phalloidin, Table 2), and nuclei were stained with 1 µg/ml Hoechst (Thermo Fisher Scientific, 33342). Cells were observed under a confocal microscope equipped with continuous laser emitting at 405, 488, 561, and 633 nm (SP5, Leica Microsystems), using a 63× objective (NA 1.4). Fluorophore emission at individual wavelength was sequentially acquired and performed with an acousto-optical beam set to 415–470, 498–520, 571–610, and 643–720 nm depending on the fluorophore detected.
List and reference of reagent used
Extended data
Statistical analysis report of all experiments. Download Extended Data, XLSX file.
Annexin-A5 (AnxA5) staining was performed as described before (Ory et al., 2013). GrCs were hyperpolarized for 10 min in an imaging buffer containing the following (in mM): 170 NaCl, 3.5 KCl, 0.4 KH2PO4, 20 TES [N-Tris (hydroxyl-methyl)-methyl-2-aminoethane-sulfonic acid], 5 NaHCO3, 5 glucose, 1.2 MgCl2, and 1.3 CaCl2, pH 7.4. Neurons were then maintained for 10 min at 37°C in the presence of Alexa Fluor 647-conjugated AnxA5 (1/50, BioLegend) in a resting solution (Locke’s solution) containing the following (in mM), 140 NaCl, 4.7 KCl, 2.5 CaCl2, 1.2 KH2PO4, 1.2 MgSO4, 11 glucose, and 15 HEPES, pH 7.2, or in a stimulation solution containing 50 mM K+ (Locke’s solution containing 50 mM KCl and 89 mM NaCl). Neurons were then fixed and counterstained for 30 min with TMR-phalloidin. To identify PS egress sites, cells were incubated in a culture medium containing 30 mM KCl and both Alexa Fluor 647-conjugated AnxA5 and anti-Synaptotagmin1 (Syt1) antibody directed against the luminal domain. Cells were then fixed, and Syt1 antibodies revealed with a goat anti-rabbit antibody conjugated to Alexa Fluor 555. AnxA5 and Syt1 staining were observed under a confocal microscope (SP5, Leica Microsystems). Image analyses were performed with the Icy software. The HKmeans thresholding method (Manich et al., 2020) was used to segment actin staining, and mean AnxA5 intensity was computed in the resulting region of interest (ROI) delineating the neuronal shape. AnxA5 mean intensity was obtained from 10 fields of view per each independent experiments (n = 3).
Preparation of synaptosomal fraction
Synaptosomal fractions were prepared as previously described (Dunkley et al., 2008). Briefly, cerebella from adult mice were collected in an isotonic sucrose/EDTA buffer (0.32 M sucrose, 1 mM EDTA, 5 mM Tris–HCl; homogenizing buffer), pH 7, and dissociated by applying 13 even strokes with a dounce homogenizer. After centrifugation of the homogenates (1,000 × g, 10 min, 4°C), the supernatant (S1, crude extract) was kept on ice, and the pellet was resuspended in homogenizing buffer, subjected to additional 17 even strokes, and centrifuged (1,000 × g, 10 min, 4°C). The pellet was discarded, and the supernatant S2 was pooled with S1. Protein concentration was measured and adjusted with an ice-cold homogenizing buffer to 4–5 mg/ml. Crude extract was loaded onto a discontinuous Percoll gradient [3, 10, 15, and 23% Percoll (v/v)] and centrifuged at 31,000 × g for 5 min, at 4°C. Each fraction was individually recovered, and Fractions 3 and 4 enriched in synaptosomes were pooled. Synaptosomes were then diluted in an ice-cold homogenizing buffer and centrifuged at 20,000 × g for 30 min at 4°C to concentrate the synaptosomes into the pellet and remove Percoll. Synaptosomes were resuspended in a lysis buffer (cell extraction buffer: 10 mM Tris–HCl, 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM NaF, 20 mM Na4P2O7, 2 mM Na3VO4, 1% Triton X-100, 10% glycerol, 0.1% SDS, 0.5% deoxycholate; Invitrogen, FNN0011), pH 7.4, supplemented with protease and phosphatase inhibitor cocktail (Sigma-Aldrich, P8340), and further subjected to Western blot analysis.
Western blotting
Cells and tissues were lysed in a cell extraction buffer with protease and phosphatase inhibitor cocktail (Sigma-Aldrich, P8340,). Cell lysates were cleared at 20,000 × g for 10 min at 4°C, and protein concentration is determined using Bio-Rad protein assay. Proteins (20 µg total) were separated on Novex 4–12% Bis-Tris gels (Thermo Fisher Scientific) and transferred to nitrocellulose membrane (Bio-Rad Laboratories, 1704156). Blots were blocked for 1 h at room temperature in Tris-buffered saline containing 5% (w/v) milk powder (fat free) and 0.1% Tween 20 (TBST; 0.1% Tween, 150 mM NaCl, 10 mM Tris–HCl), pH 7.5, and probed with the anti-PLSCR1, anti-Syt1, or anti-PSD95 antibodies (Table 2). After three washes, blots were incubated with the corresponding secondary antibodies coupled to HRP (Table 2). When possible and when primary antibodies were from different species, blots were incubated with 0.02% sodium azide in TBST for 1 h at room temperature to inactivate HRP. Blots were then processed to reveal a second protein on the same blot. Detection was carried out with Prime Western Blotting System (Thermo Fisher Scientific), and immunoreactive bands were imaged using the Amersham Imager 680 RGB camera system (GE healthcare Life Sciences). Values were normalized to the corresponding β-actin protein levels.
Acute slice preparation and electrophysiological recordings
Slice preparation and electrophysiology recording were performed as previously described (Doussau et al., 2017). Acute horizontal cerebellar slices were prepared from male and female C57Bl/6 mice of both genotype (Plscr1+/+ and Plscr1−/−) aged 18–30 d. Mice were anesthetized by isoflurane inhalation (4%) and decapitated. The cerebellum was dissected out in an ice-cold artificial cerebrospinal fluid (ACSF) bubbled with carbogen (95% O2, 5% CO2) and containing the following (in mM):120 NaCl, 3 KCl, 26 NaHCO3, 1.25 NaH2PO4, 2.5 CaCl2, 2 MgCl2, 10 glucose, and 0.05 minocyclin. Slices were then prepared (Microm HM650V) in an ice-cold solution containing the following (in mM): 93 N-methyl-D-glucamine, 2.5 KCl, 0.5 CaCl2, 10 MgSO4, 1.2 NaH2PO4, 30 NaHCO3, 20 HEPES, 3 Na-pyruvate, 2 thiourea, 5 Na-ascorbate, 25 D-glucose, and 1 kynurenic acid (Zhao et al., 2011). Slices (300 µm thick) were maintained in a bubbled ACSF medium (see above) at 34°C until their use for experiments.
After at least 1 h of recovery at 34°C, a cerebellar slice was transferred to a recording chamber. In order to block inhibitory transmission, postsynaptic plasticity, GABAB, and endocannabinoid signaling, slices were continuously perfused with bubbled ACSF containing blockers of GABAA, GABAB, NMDA, CB1, and mGluR1 receptors. To do so, we added the following antagonists (see Table 2 for references): 100 µM picrotoxin, 10 µM CGP52432 [3-([(3,4-dichlorophenyl)-methyl]amino)propyl(diethoxymethyl)phosphinic acid], 100 µM D-AP5 [D-(−)-2-amino-5-phosphonopentanoic acid], 1 µM AM251 [1-(2,4-dichlorophenyl)-5-(4-iodophenyl)-4-methyl-N-(piperidin-1-yl)-1H-pyrazole-3-carboxamide], and 2 µM JNJ16259685[(3,4-dihydro-2H-pyrano[2,3-b]quinolin-7-yl)-(cis-4-methoxycyclohexyl)-methanone]. Recordings were made at 34°C in Purkinje cells (PCs) located in the vermis. PCs were visualized using infrared contrast optics on an Olympus BX51WI upright microscope. Whole-cell patch–clamp recordings were obtained using a MultiClamp 700A amplifier (Molecular Devices). Pipette (2.5–3 MΩ resistance) capacitance was canceled, and series resistance (Rs) between 5 and 8 mΩ was compensated at 80%. Rs was monitored regularly during the experiment, and the recording was stopped when Rs changed significantly (>20%). PCs were held at −60 mV. The intracellular solution for voltage-clamp recording contained the following (in mM): 140 CsCH3SO3, 10 phosphocreatine, 10 HEPES, 5 QX314-Cl, 10 BAPTA, 4 Na-ATP, and 0.3 Na-GTP. Beams of parallel fibers were stimulated extracellularly using a monopolar glass electrode filled with ACSF, positioned at least 100 µm away from the PC to ensure a clear separation between the stimulus artifact and excitatory postsynaptic currents (EPSCs). Pulse trains were generated using an Isostim A320 isolated constant current stimulator (World Precision Instruments) controlled by a WinWCP freeware (John Dempster, Strathclyde Institute of Pharmacy and Biomedical Sciences, University of Strathclyde). The synaptic currents evoked in PCs were low-pass filtered at 2 KHz and sampled at 20–50 KHz (National Instruments).
Cerebellum slices and plasma membrane sheet preparation for transmission EM
Wild-type (n = 3) and Plscr1−/− (n = 3) mice were anesthetized with a mixture of ketamine (100 mg/kg) and xylazine (5 mg/kg) and transcardiacally perfused with a 0.1 M phosphate buffer, pH 7.3, containing 2% paraformaldehyde and 2.5% glutaraldehyde. The 2-mm-thick slices were cut from the cerebellum and postfixed in 1% glutaraldehyde in phosphate buffer overnight at 4°C. The slices were then immersed for 1 h in OsO4 0.5% in a phosphate buffer. The 1 mm3 blocks were cut in the cerebellum, dehydrated, and processed classically for embedding in araldite and ultramicrotomy. Ultrathin sections were counterstained with uranyl acetate. Fields of view were randomly selected in ultrathin sections from several blocks (one section/block) from each mouse.
Membrane sheets were prepared and processed as described previously (Gabel et al., 2015; Delavoie et al., 2021). In brief, carbon-coated Formvar films on nickel electron grids were inverted onto unstimulated or stimulated GrCs (Locke’s buffer containing 50 mM KCl) incubated with gold-conjugated AnxA5 for 10 min. To prepare membrane sheets, we applied pressure to the grids for 25 s; then we lifted the grids so that the fragments of the upper cell surface adhered to the grids. These membrane sheets were then fixed in 2% paraformaldehyde for 10 min at 4°C, blocked in PBS with 1% BSA and 1% acetylated BSA, and incubated with anti-GFP and anti-vGlut1 antibodies (Table 2) overnight at 4°C. Then the membranes were washed six times with PBS and incubated 3 h with 25 nm gold particle-conjugated goat anti-rabbit IgG and 10 nm gold particle-conjugated goat anti-guinea pig IgG (Table 2). These membrane portions were fixed in 2% glutaraldehyde in PBS, postfixed with 0.5% OsO4, dehydrated in a graded ethanol series, treated with hexamethyldisilazane (Sigma-Aldrich), and air-dried. Transmission electron microscope (7500; Hitachi) equipped with a camera (C4742-51-12NR; Hamamatsu Photonics) was used for acquisition. Twenty to forty fields of view were randomly taken, and membrane patches labeled with vesicular glutamate transporter1 (vGlut1), GFP, and/or AnxA5 were counted.
Morphometric analysis of EM slices
SVs, pre- and postsynaptic membranes, and boutons were manually selected or delineated using the Icy bioimaging software. The density of SVs in the boutons and the shortest distance between SV and the presynaptic membrane were calculated. Synapse densities per field were also calculated, and the mean distance of synaptic cleft were computed.
Synaptophluorin live-cell imaging
Imaging was performed with DIV6 to DIV7 GrCs as reported by Nicholson-Fish et al. (2015). GrCs were removed from the culture medium 48–72 h post-transfection and repolarized in an imaging buffer (see above, Immunofluorescence, confocal microscopy, and image analysis) for 15 min. Coverslips were then mounted in an imaging chamber (Warner Instruments, RC-21BRFS) supplied with a pair of embedded platinum wires allowing connection to a low impedance field stimulator (Digitimer, D330-MultiStim System). The chamber was placed on the stage of an inverted microscope (Zeiss AxioObserver Z1) equipped with either a sCMOS camera (Orca-FLASH4, Hamamatsu Photonics) or a Zeiss AxioCam506 and a control of focus (Zeiss Definite Focus) to prevent drift. During recording, cells were continuously perfused with an imaging buffer at 37°C. Syp-pH–transfected neurons were visualized with a Zeiss Plan Apochromat 40× oil-immersion objective (NA 1.4) at 475 ± 28 nm excitation wavelength using 525 ± 50 nm emission filter (LED light source SpectraX, Lumencor) or exciter 450–490 nm, beam splitter 495 nm, emitter 500–550 nm, Colibri 7 LED light source (Carl Zeiss). When specified, cotransfected mCherry expression was visualized using 555 ± 28 nm excitation and 605 ± 70 nm emission filters or exciter 538–652 nm, beam splitter 570 nm, and emitter 570–640 nm. Neurons were stimulated with a train of action potentials (400 action potentials delivered at 40 Hz; 100 mA, 1 ms pulse width) and the acquisition sequence driven by the MetaMorph software (Molecular Devices) or Zeiss Zen2 software at 0.5 Hz frequency. At the end of the recording, cells were challenged with an alkaline imaging buffer (50 mM NH4Cl substituted for 50 mM NaCl) to reveal total pHluorin fluorescence. To protect neurons from photodamage and subsequently minimize fluorescent protein photobleaching, we restricted light exposure (using low light power and a 0.5 Hz frequency). Photobleaching was evaluated by monitoring fluorescence stability in regions unaffected by stimulation (cell body when in the field of view or neuron extension without synapses if absent cell body). Photobleaching measurements appeared limited and more importantly comparable between conditions and accounts for roughly 12 ± 2% of the fluorescence decay at 120 s. If, for any reason and despite automatic focus control, visible drift occurred during recordings, the field was excluded from the analysis. Quantification of the time-lapse series was performed using the Time Series Analyzer plugin for ImageJ, and only synapses that responded to action potential stimulation were selected for the analysis. A circular ROI (1.8 μm diameter) was drawn around each spot characterized by a sudden rise in fluorescence. ROI was centered on the maximum fluorescence of the spot. The pHluorin fluorescence change in each spot was calculated as ΔF / F0, and n refers to the number of individual cells examined.
Syt1 antibody uptake assay
Plscr1+/+ and Plscr1−/− neurons were incubated for 30 min in a culture medium (containing 25 mM KCl) in the presence of antibodies directed against the Syt1 luminal domain. Following incubation, neurons were washed and then fixed with ice-cold 4% PFA in PBS. The bound anti-Syt1 antibodies were visualized using Alexa Fluor 555-conjugated goat anti-rabbit secondary antibodies. The neuronal network was outlined using the hierarchical K-means thresholding method applied to the Syt1 staining, and mean intensity within the resulting region of interest was computed. The signal intensity from Plscr1−/− neurons was normalized to that of Plscr1+/+ neurons (±SEM). Experimental n refers to the number of individual cells examined.
Statistical analysis
Analysis was performed by using Microsoft Excel and GraphPad Prism. All data passed the normality test (Shapiro–Wilk test) and variance equality. A Student’s t test was performed for comparison between two datasets. Data were analyzed by one-way analysis of variance (ANOVA) with Sidak’s multiple-comparison test when greater than two datasets. To compare fluorescence response over time or data with more than one variable, two-way ANOVA with Tukey’s multiple-comparison post hoc test was performed. All data are reported as mean ± standard error of the mean (SEM). To attest for similar SV distribution, the Kolmogorov–Smirnov test was applied. Significance was set as *p < 0.05, **p < 0.01, and ***p < 0.001, ns, non significant.
Databases
Preliminary information about Plscr1 distribution in the mouse brain were obtained from the Allen Brain Atlas (https://mouse.brain-map.org/experiment/show/632487) and the Human Brain Atlas databases (https://www.proteinatlas.org/ENSG00000188313-PLSCR1/brain).
Results
PLSCR1 is expressed in cerebellar GrCs and localizes at synapses
PLSCR1 is expressed in a wide range of tissue including the brain (Wiedmer et al., 2000; Zhou et al., 2005). Mining in mouse brain databases revealed that the Plscr1 transcript is barely detectable by in situ hybridization (Allen Brain Atlas; Lein et al., 2007). However, it is expressed in the olfactory bulb, the cerebellum, the pons, and the medulla when using the more sensitive next-generation sequencing (Human Protein Atlas; Sjöstedt et al., 2020). To validate the expression profile of the PLSCR1 gene product in a specific area of the mouse brain, we dissected different brain areas to perform Western blots on tissue extracts from Plscr1+/+ adult mice. As a negative control, we used cerebellar extracts from Plscr1−/−. We confirmed that PLSCR1 is abundant in the cerebellum, with expression also observed in the olfactory bulb and midbrain albeit at lower levels. However, expression is barely detectable in all the other brain regions investigated (Fig. 1A). We therefore focused our studies on the cerebellum, which displayed the highest expression of PLSCR1 in the mouse brain.
PLSCR1 is enriched at synapses of GrCs of the cerebellum. A, Immunodetection of PLSCR1 protein by a Western blot in the olfactory bulb (Olf. Bulb), cortex [prefrontal (Ctx Pre), parietal (Ctx Par), and occipital (Ctx Occ)], striatum, hippocampus (Hippoc.), midbrain, pons, medulla, and cerebellum (Cereb.) from Plscr1+/+ mice. As a control, the cerebellum of Plscr1−/− mice was used. Blots with short and longer exposure are shown. Actin is shown as a loading control. B, Immunodetection of PLSCR1 protein by a Western blot in cultured GrCs. C, PLSCR1 is enriched in synaptosomes. Cerebella were homogenized, cell body is removed by centrifugation, and the resultant supernatant is layered on a discontinuous Percoll gradient. Fractions were collected and probed for PLSCR1, the presynaptic marker Synaptotagmin1 (Syt1), and the postsynaptic marker PSD95. F2 corresponds to the fraction enriched in membranes, and fractions F3 and F4 enriched in synaptosomes were pooled and compared with crude extract (CE) loaded on gradient. D, Confocal microscopy of GrCs transfected with expression vector coding for GFP-PLSCR1 and labeled for vGlut1 and Syt1. The mask of synapses containing the three markers is shown. E, Top, The principle of plasma membrane sheet preparation. Carbon-coated Formvar films on nickel EM grids (1) were inverted onto neurons grown on poly-L-lysine–coated coverslips, and pressure is applied to the grid (2). Grids are then lifted (3), leaving fragments of the plasma membrane (blue) and synapses containing docked SVs (red) on the grids (4). Bottom left, Confocal images of GrCs stained for vGlut1 (red) and Synaptophysin (green). Bottom right, A representative electron micrograph of GrCs plasma membrane sheets labeled for vGlut1 (10 nm gold particles, arrow). F, An electron micrograph of plasma membrane sheets prepared from Plscr1−/− GrCs expressing GFP-PLSCR1. Immunolabeling of GFP and vGlut1 were revealed with 25 nm (green arrows) and 10 nm (red arrows) gold particles, respectively.
Since GrCs account for 95% of cerebellum cells (D’Mello et al., 1993; Krämer and Minichiello, 2010), we first probed PLSCR1 expression in primary cultures of GrCs displaying high homogeneity. Western blot experiments revealed that PLSCR1 was expressed in GrCs cultured from Plscr1+/+ mice but, as expected, absent from Plscr1−/− GrC culture (Fig. 1B). We next analyzed PLSCR1 subcellular distribution. First, we prepared synaptosomes from the mouse cerebellum to determine whether PLSCR1 was localized to synaptic terminals (Dunkley et al., 2008). Compared with crude brain homogenate, PLSCR1 is enriched in synaptosomal fractions containing the post- and presynaptic markers PSD95 and Syt1, respectively (Fig. 1C). To further investigate the presence of PLSCR1 at synapses, we performed immunofluorescence on GrCs. Because of the lack of reliable antibodies to detect endogenous mouse PLSCR1 by immunofluorescence, we exogenously expressed GFP-tagged PLSCR1 and compared its distribution with specific markers of the presynaptic terminal, the vGlut1, and Syt1. GFP-PLSCR1 was found in GrC axons and colocalized with vGlut1 and Syt1, indicating that it is also present at the synapse (Fig. 1D). To further explore PLSCR1 localization, we performed immunogold EM analysis of native plasma membrane sheets of GrCs (Fig. 1E). Sheets were obtained by tearing off the plasma membrane at the dorsal surfaces of cells to examine, at high resolution, the associated molecules (Wilson et al., 2000; Gabel et al., 2015). Synaptic components were retained by this method, since confocal analysis of recovered membrane pieces were stained for Synaptophysin and vGlut1. EM analysis using secondary antibodies coupled to 10 nm gold particles showed vGlut1 clusters on membrane patches of various sizes (Fig. 1E). Plasma membrane sheets from Plscr1−/− neurons expressing GFP-PLSCR1 were analyzed. Secondary antibodies coupled to 10 and 25 nm gold particles were used to reveal both the expression of vGlut1 and GFP-PLSCR1, respectively (Fig. 1F). Gold particles of different sizes were found distributed over most membrane patches. In addition, GFP-PLSCR1 was localized close to vGlut1, with the majority of membrane patches containing vGlut1 also containing GFP-PLSCR1 (73% ± 13%; n = 5 experiments). Altogether, our data indicate that GFP-PLSCR1 is associated with the plasma membrane of GrCs and is most likely enriched at the presynaptic terminal where vGlut1-positive SVs are located.
To determine whether Plscr1 deletion could modify synaptic organization and/or characteristics, we performed morphometric analysis of sections from the Plscr1−/− and Plscr1+/+ cerebellum observed by transmission EM (Fig. 2A). GrC to PC synapses from Plscr1−/− mice showed no significant differences in either synapse density per slice or SV density in synaptic boutons compared with Plscr1+/+ mice (Fig. 2B). We measured the minimal distance separating each SV from the synaptic membrane facing the postsynaptic density but found no significant differences in either the distribution of SVs (Fig. 2C) or the number of docked SVs (Fig. 2D). We did however note a slight decrease in synaptic length (Fig. 2E). The synaptic cleft, measured as the space separating the pre- and postsynaptic membranes, remained unchanged (Fig. 2E). Consequently, the absence of PLSCR1 has negligible impact on the structural organization of GrC-PC synapses, suggesting that PLSCR1 does not influence developmental processes in the cerebellum.
Morphometric analysis of cerebellum slices observed by transmission EM. A, Representative electron micrographs of the cerebellum from Plscr1+/+ and Plscr1−/− mice showing synapses between GrC and PC dendrites (asterisks). Scale bar, 100 nm. B, A number of synapses in each field of view were counted (5 fields of 4 µm2; n = 3 mice), and a number of SVs were counted and reported on the surface of a bouton for Plscr1+/+ and Plscr1−/− mice (±SEM). C, The shortest distance between SVs and the presynaptic membrane facing the postsynaptic density was calculated. The graph represents the distribution of mean number of SVs found according to their distance to the synapse. D, A graph representing cumulative distribution of docked vesicles at GrC-PC synapse (distance, <50 nm). E, The presynaptic and postsynaptic membranes were manually delineated. The length of the presynapse and the mean distance between the pre- and postsynapse were calculated (synaptic cleft size). Statistical significance was assessed using unpaired t test or Kolmogorov–Smirnov test for distribution analysis. ns, non significant; *p < 0.05. [p = 0.79 and p = 0.788 (B); p = 0.7453 (C); p = 0.996 (D); p = 0.0162 and p = 0.2157 (E)]. More than 210 synapses from three mice were analyzed for each genotype
GrC stimulation triggers PLSCR1-dependent PS egress at synapses
Rapid activity-dependent transbilayer movement of PLs, mainly PS and PE, occurs in synaptosomes prepared from the electric organ of the electric ray, Narke japonica (Lee et al., 2000). We also demonstrated PLSCR1-dependent PS egress in chromaffin cells during exocytosis (Ory et al., 2013). We therefore asked whether stimulation of GrCs could lead to disruption of the plasma membrane asymmetry and whether PLSCR1 was required for that process. To detect PS exposed to the extracellular leaflet of the plasma membrane, living GrCs were stimulated with a 50 mM KCl solution containing fluorescent AnxA5, which selectively binds to PS (Andree et al., 1990). GrCs were then fixed and stained for F-actin to delineate neuronal processes and to quantify AnxA5 staining. Compared with unstimulated neurons, depolarization of Plscr1+/+ GrCs induced an increase in AnxA5 staining (Fig. 3A,B). Interestingly, the increase in AnxA5 staining was abrogated in stimulated Plscr1−/− GrCs (Fig. 3A,B) indicating that PLSCR1 was required for the activity-dependent translocation of PS to the extracellular leaflet.
PLSCR1-dependent PS egress occurs at synapses. A, Cultured GrCs from Plscr1+/+ or Plscr1−/− mice were stimulated for 10 min with 50 mM KCl or maintained under resting condition in the presence of 1 µg/ml fluorescent AnxA5 (magenta). Cells were fixed, counterstained for nuclei (blue) and actin (yellow), and observed under a confocal microscope. B, Quantification of AnxA5 mean intensity of Plscr1+/+ or Plscr1−/− neurons stimulated for 10 min with 50 mM KCl or maintained under resting condition [10 fields of view per experiments (±SEM), n = 3 independent experiments]. C, A scheme of Syt1 staining assay. Living Plscr1+/+ or Plscr1−/− GrCs were incubated in culture medium containing 25 mM KCl; anti-Syt1 antibodies were directed against the luminal domain of Syt1 and fluorescent AnxA5 (green). Cells were fixed, and anti-Syt1 antibodies were revealed with fluorescent secondary antibody without cell permeabilization to reveal Syt1 at the plasma membrane (red). Cells were observed under a confocal microscope, and intensity profile along the depicted line is shown. D, A representative electron micrograph of plasma membrane sheets prepared from GrCs stimulated for 10 min in the presence of gold-conjugated AnxA5. Labeling of AnxA5 and vGlut1 was revealed with 15 nm (green arrows) and 10 nm (red arrows) gold particles, respectively. Enlargements are shown in Insets 1 and 2. Statistical significance was assessed using two-way ANOVA with a Holm–Sidak post hoc test. ns, non significant; **p < 0.01; ***p < 0.001. At least 21 fields of view from three independent experiments were analyzed for each condition.
AnxA5 staining organized as discrete spots along neuronal processes (Fig. 3A) suggesting that PS egress could preferentially occur at the synapse. To test this hypothesis, we took advantage of anti-Syt1 antibodies that recognize the luminal domain of the integral SV protein Syt1, which is transiently accessible from the extracellular space upon SV fusion with the plasma membrane (Fig. 3C). Incubating living GrCs with both fluorescent AnxA5 and anti-Syt1 antibodies showed that AnxA5 and Syt1 staining partially overlapped indicating that PS egress occurs at active synapses (Fig. 3C). We also prepared plasma membrane sheets of stimulated neurons to determine more precisely where PS egress occurs. Thus, Plscr1+/+ GrCs were stimulated in the presence of gold-coupled AnxA5, and the distribution of AnxA5 was compared with vGlut1 on membrane sheets. As shown in Figure 3D, clusters of AnxA5 gold particles were found in membrane patches close to vGlut1 staining. Altogether, these data indicate that activity-dependent PS egress occurs at synapses and that this egress requires expression of PLSCR1.
GFP-PLSCR1 restores PS egress in Plscr1−/− GrCs
To confirm that PLSCR1 is essential for PS egress, we performed rescue experiments. GFP-tagged PLSCR1 protein was expressed in Plscr1−/− GrCs, and PS egress in response to KCl-dependent stimulation was analyzed on plasma membrane sheets. We observed GFP-PLSCR1 in membrane patches containing both vGlut1 and gold-coupled AnxA5 in response to cell stimulation (Fig. 4A). In addition, the percentage of membrane patches containing both vGlut1 and AnxA5 was comparable between Plscr1+/+ neurons and Plscr1−/− neurons expressing GFP-PLSCR1 and increased in response to stimulation. In contrast, no increase was observed in Plscr1−/− GrCs (Fig. 4A). These data confirm that PLSCR1 is required for activity-dependent PS egress at synapses.
Overexpression of PLSCR1 restores PS egress in Plscr1−/− GrCs. A, A representative electron micrograph of plasma membrane sheets prepared from Plscr1−/− GrCs expressing GFP-PLSCR1. Neurons were stimulated for 10 min with 50 mM KCl (S) or maintained under resting condition (R) in the presence of gold-conjugated AnxA5 (15 nm beads, blue arrows). Immunolabeling of vGlut1 and GFP was revealed with 10 nm (red arrows), and 25 nm (green arrows) gold particles, respectively. The graph represents the percentage of synapses (membrane patches vGlut1-positive) closely associated with AnxA5 beads. Each point represents the mean of 2 (PLSCR1+/+) or 3 independent experiments (PLSCR1−/−). In each experiment, 40–50 images per conditions were quantified. ns, non significant; ***p < 0.001.
PLSCR1 is required for evoked synaptic transmission
A key unsolved question is whether PLSCR1activity is involved in neurotransmission. In the cerebellar cortex, GrCs convey high-frequency information (several hundreds of Hz) to PCs and molecular layer interneurons. GrC synapses stand out from most other synapse types by the striking ability to recruit reluctant SVs within milliseconds to sustain the release of glutamate at such extreme frequencies. This phenomenon underlies a large facilitation of glutamate release during paired-pulse stimulation elicited at high frequency (Miki et al., 2016; Doussau et al., 2017). GrCs can also be endowed with specific mechanisms optimizing the recycling of used SVs as the high facilitation of glutamate release observed during high-frequency trains is maintained for hundreds of milliseconds (Doussau et al., 2017). Hence, to test whether PLSCR1 is required for these presynaptic properties of GrCs, we prepared acute cerebellar slices from Plscr1+/+ and Plscr1−/− littermates in which we recorded EPSCs in PCs evoked by GrC axon stimulation (parallel fibers, PFs; Fig. 5A) with either twin stimuli (paired-pulse stimulation at 50 Hz) or high-frequency trains. As expected in Plscr1+/+ mice, amplitudes of EPSCs increased after the first stimulus (Fig. 5B). Furthermore, excitatory neurotransmission was facilitated, and this facilitation was maintained during tens of stimuli (Fig. 5C). In contrast, in Plscr1−/− mice, paired-pulse facilitation was strongly reduced (mean paired-pulse ratio: 2.03 ± 0.7 for Plscr1+/+ vs 1.59 ± 0.07 for Plscr1−/−; t test; p = 0.001; n = 7 for Plscr1+/+ and n = 8 for Plscr1 −/−). In addition, facilitation during the high-frequency train was only transient, disappearing rapidly after the first 10 stimuli (Fig. 5C).
PLSCR1 is required to sustain synaptic transmission at high firing rates. A, A schematic representing the connectivity in the cerebellar cortex and the position of the stimulating and recording electrodes. (GrC, granule cell; PC, Purkinje cell; PF, parallel fiber; ML, molecular layer). B, Left, Representative EPSCs evoked by paired-pulse stimulation (50 Hz) of PFs recorded in Plscr1+/+ and Plscr1−/− mice acute slices (black and red traces, respectively). Thick lines correspond to averaged EPSCs of at least 10 successive EPSCs (thin lines). Lower traces correspond to normalized averaged EPSCs. Right, Box plots showing the values of the paired-pulse ratio obtained in Plscr1+/+ and Plscr1−/− acute slice. White and blue lines correspond to mean and median values, respectively (±SEM). C, Left, representative recording traces evoked by PF stimulations at 100 Hz (50 pulses) and recorded in Plscr1+/+ and Plscr1−/− mice acute slices (black and red traces, respectively). Right, Mean values of normalized EPSC amplitude elicited by trains of stimulation at 100 Hz and recorded in Plscr1+/+ and Plscr1−/− mice acute slices (black and red traces, respectively, ± SEM)
The absence of observable abnormalities in the ultrastructure of Plscr1−/− GrC synapses (as shown in Fig. 2) suggests that the presynaptic dysfunctions observed are unlikely to be caused by fewer SVs in GrC boutons or major anatomical defects. Instead, it is more likely that PLSCR1 plays a key role in facilitating the rapid recruitment of reluctant vesicles (exocytosis) and/or in the recycling of previously used SVs (endocytosis).
PLSCR1 is required for SV retrieval
To ascertain whether the effects observed on neurotransmission in the absence of PLSCR1 were due to altered SV exocytosis or endocytosis, we performed real-time monitoring of the genetically encoded Syp-pH reporter. Syp-pH is widely used to report SV recycling since it responds to local pH changes occurring during SV exocytosis and endocytosis (Granseth et al., 2006; Nicholson-Fish et al., 2015; Jäpel et al., 2020). In resting conditions, as pHluorin faces the SV lumen, the fluorescence of Syp-pH is quenched by the acidic pH of the SV. Upon exocytosis, Syp-pH is exposed to the neutral extracellular medium, resulting in fluorescence unquenching, the quantification of which represents an assessment of the extent of synaptic exocytosis. After termination of stimulation, Syp-pH fluorescence intensity decreases rapidly as it is retrieved by endocytic SVs, which quickly reacidify. As SV endocytosis is the rate-limiting step in this fluorescence decrease, measurement of poststimulus Syp-pH fluorescence decay is a reliable indicator of compensatory endocytosis (Sankaranarayanan and Ryan, 2000; Atluri and Ryan, 2006; Granseth et al., 2006; Egashira et al., 2015). Plscr1+/+ or Plscr1−/− GrCs were transfected with Syp-pH and stimulated with a train of high-frequency action potentials (40 Hz, 10 s) to evoke robust exocytosis (Nicholson-Fish et al., 2015). Analysis of the fluorescence intensity profiles showed a fast rise in fluorescence upon stimulation for both genotypes (Fig. 6A,B) and a progressive return to the fluorescence baseline for Plscr1+/+ GrCs. Although no significant differences in the amount of Syp-pH visiting the cell surface was observed (Fig. 6C), fluorescence decay was severely delayed in Plscr1−/− GrCs (Fig. 6B,D), suggesting that SV exocytosis was unaltered but that endocytosis was impaired in the absence of PLSCR1. To confirm the requirement for PLSCR1 in SV endocytosis, we performed rescue experiments in Plscr1−/− GrCs using exogenous expression of the mCherry-tagged version of PLSCR1. Expression of exogenous PLSCR1 significantly restored the decay of fluorescence intensity (Fig. 6B,D), consistent with the role of PLSCR1 in compensatory endocytosis.
PLSCR1 is required for efficient SV endocytosis during high-frequency stimulation of GrCs. A, Representative images of GrCs transfected with Syp-pH and stimulated with a train of 400 action potentials delivered at 40 Hz. Still images show Plscr1+/+ and Plscr1−/− GrCs before stimulation, 30 and 180 s after stimulation. The inset shows a representative time course of Syp-pH variation at a single synapse. B, Average normalized time course of fluorescence response of Syp-pH represented as ΔF / F0 ± SEM (5 independent preparations, n ≥ 15 neurons by condition) in Plscr1+/+, Plscr1−/− GRCs, and Plscr1−/− GrCs transfected with mCherry-PLSCR1. C, Mean peak of Syp-pH response during stimulation. Maximum Syp-pH intensity after stimulation was normalized to the total amount of Syp-pH at synapse obtained after incubation of neurons with 50 mM NH4Cl (±SEM). D, Mean normalized intensity of Syp-pH fluorescence remaining after 2 min imaging (±SEM). ns, non significant; **p < 0.01. E, Representative confocal images of anti-Syt1 antibody uptake experiments. Plscr1+/+ or Plscr1−/− GrC neurons were incubated in a culture medium containing 25 mM KCl and antibodies directed against the luminal domain of Syt1 for 30 min. Neurons were fixed and stained for the presence of Syt1 at the plasma membrane. F, Quantification of Syt1 fluorescence normalized to Plscr1+/+ (3 independent experiments, 10 fields of view per experiment; *p < 0.05)
To further probe the role of PLSCR1 in SV endocytosis, we took advantage of GrC neuronal culture conditions, which uses mild depolarization (25 mM KCl) to maintain spontaneous synaptic activity required for GrC maturation and survival (D’Mello et al., 1993; Lawrie et al., 1993). To detect the spontaneous release of SVs and a potential subsequent defect in endocytosis, GrCs were incubated with a culture medium containing antibodies directed against the luminal domain of Syt1. To reveal the amount of antibody remaining at the cell surface after continuous SV turnover, we added fluorescent secondary antibodies without cell permeabilization and quantified their fluorescence intensity. In agreement with pHluorin experiments, surface Syt1 was increased by ∼50% in Plscr1−/− GrCs as compared with that in Plscr1+/+ (Fig. 6E,F) providing further evidence that endocytosis of SV proteins was perturbed in the absence of PLSCR1.
Discussion
To sustain neurotransmission during elevated neuronal activity, a major reorganization of the neuronal plasma membrane has to occur to integrate the SV at the exocytic site and to subsequently retrieve lipids and proteins to replenish the vesicular pool (Binotti et al., 2021). While experimental evidence points to a major role of phosphoinositides (PIs) and their metabolism in this process, the role of PLs such as PS and the regulation of their localization have remained unexplored. Here, we provide the first evidence that PLSCR1 randomizes PS at synapses during heightened neuronal activity and that the resulting loss of plasma membrane asymmetry is dispensable for SV exocytosis but necessary for compensatory SV endocytosis. As a consequence of impaired vesicular pool replenishment, synaptic transmission at GrC to PC synapses in cerebellar acute slices displays accelerated depletion in the absence of PLSCR1. Therefore, we propose that PLSCR1 activity controls compensatory endocytosis in GrCs and contributes to sustain neurotransmission at high frequencies.
PLSCR1 and plasma membrane asymmetry disruption
Immunofluorescence and immunogold EM approaches revealed that (1) PLSCR1 is enriched at synaptic terminals and (2) PLSCR1 is required for activity-dependent PS egress at these synapses. This is the first evidence that PLSCR1 promotes shuffling of the lipid PS in neurons. PLSCR1 has been involved in lipid mixing both in vitro (Rayala et al., 2014) and in immune and chromaffin cells during regulated exocytosis (Kato et al., 2002; Ory et al., 2013). Our observation that PS egress is restricted to synapses indicates that PLSCR1 is probably activated locally as has been reported in chromaffin cells (Ory et al., 2013). In these cells, PS egress occurred only in the vicinity of exocytosis sites, despite PLSCR1 being homogeneously distributed at the plasma membrane (Ory et al., 2013). The low affinity of PLSCR1 for Ca2+ (Stout et al., 1998) together with the activity described here occurring only after neurotransmission suggests that PLSCR1 may only be activated when intracellular Ca2+ reaches a critical concentration. Typically, low-affinity Ca2+ sensors must be localized close to the exocytic site, since the Ca2+ burst is spatially constrained within a microdomain at the presynaptic active zone (Neher and Sakaba, 2008). Our finding that PLSCR1 is required for SV endocytosis during periods of high activity suggests that its activity-dependent triggering of PS redistribution must occur in or directly adjacent to the active zone. This latter region is termed the periactive zone and is where SV endocytosis is proposed to occur (Gad et al., 1998; Watanabe et al., 2013). It is particularly relevant in high-frequency synapses, where rapid and repetitive stimulation leads to the accumulation of intracellular Ca2+ (Delvendahl et al., 2015). Thus, the localized activation of PLSCR1 and the ensuing PS redistribution play a crucial role in modulating synaptic function, especially in scenarios involving high-frequency synaptic activity.
How can the local loss of plasma membrane asymmetry control compensatory endocytosis?
In the absence of PLSCR1, PS translocation to the extracellular leaflet is blocked, and SV endocytosis, as measured by the retrieval of Syp-pH and the amount of stranded Syt1 at the plasma membrane, is retarded dramatically. It is important to note that, while our analysis focuses on PS distribution to visualize membrane asymmetry changes, PLSCR1 itself is not selectively active toward PS but can redistribute various lipid types at the cell surface. Nonetheless, the unique physicochemical properties of PS render it an attractive lipid for influencing membrane dynamics. Indeed, PS, along with phosphatidic acid (PA) and PIs is an anionic PLs thought to contribute to the negative charge of the inner leaflet of the plasma membrane. Compared to PA and PIs, which represent only a minor fraction of the PLs present at the plasma membrane, PS is the predominant PL, and its redistribution from one leaflet to the other should alter the biophysical properties of the plasma membrane (Yeung et al., 2008; Puchkov and Haucke, 2013). Depending on the timing of lipid scrambling relative to SV fusion, a local decrease of negative charges by local PS depletion may, for example, reduce the repulsive forces of opposing membrane to favor fusion between SVs and the plasma membrane (Davletov and Montecucco, 2010). However, the amount of Syp-pH visiting the plasma membrane during activity was unaffected in Plscr1−/− GrCs, suggesting that exocytosis was unaltered. In agreement, large dense-core vesicle fusion was unaltered in chromaffin cells from Plscr1−/− mice (Ory et al., 2013).
An alteration in local membrane charge is not predicted to impact SV endocytosis. This is because accumulation of negatively charged PLs such as PtdIns(4,5)P2 increases the recruitment of cytosolic proteins with positively charged residues (Kay and Fairn, 2019; Clarke et al., 2020) to favor membrane bending (Micheva et al., 2001; Puchkov and Haucke, 2013). In contrast, PL scrambling may modify plasma membrane fluidity. Enriched in tightly packed sphingolipids, the outer plasma membrane leaflet is highly ordered and rigid (Gupta et al., 2020; Lorent et al., 2020). Insertion of SVs that are highly enriched in cholesterol will most likely modify the lipid organization of the presynaptic plasma membrane (Takamori et al., 2006). Indeed, membrane cholesterol content affects lipid lateral self-diffusion, confines SV proteins, and limits their diffusive behavior (Mercer et al., 2011; Dason et al., 2014; Byczkowicz et al., 2018; Wilson et al., 2020). PS and cholesterol are codistributed in the inner leaflet of the plasma membrane, and the local decrease of PS may also lead to a concomitant decrease in cholesterol content (Maekawa and Fairn, 2015) helping to fluidize the leaflet and clear the active zone. Alternatively, changes in PS distribution may unlock protein function. For example, PS shapes the transmembrane domain of synaptogyrin to bend membrane and generate SVs of homogeneous size. To achieve this, PS must be present simultaneously on both leaflets to induce a structural change in synaptogyrin, a function that is intrinsic to PLSCR1 (Yu et al., 2023). It remains to be seen whether proteins involved in endocytosis may have the same properties, and the precise role of PLs egress needs to be further explored.
PLSCR1 and GrC neurotransmission
Synaptic plasticity allows for fast adaptation of synaptic strength to neuronal network activity, which is critical for information processing. This comes in two main modes: (1) functional plasticity, which changes the strength of the synapse by modifying signal transmission, and (2) structural plasticity, which alters the number and/or shape of synapses, resulting in differential connectivity between neurons (Caroni et al., 2012). Morphometric analysis of Plscr1−/− cerebellum slices showed no major alteration within presynaptic terminals, suggesting that PLSCR1 may impair functional, rather than structural, plasticity. In the cerebellar cortex, GrC-PC synapses have to transmit sensorimotor information elicited at extreme frequencies (several hundred of Hz to kHz; Chadderton et al., 2004). To do so, these synapses are endowed with specific presynaptic processes allowing a fast and sustained boost (facilitation) of the release of neurotransmitters during bursts of high-frequency activities. The strikingly large facilitation that occurs during paired-pulse facilitation and during the first phase of high-frequency trains is underpinned by the ultrafast recruitment of a pool of reluctant SVs that can only be mobilized by high-frequency activities (Miki et al., 2016; Doussau et al., 2017). This phenomenon is affected in Plscr1−/− GrC terminals, indicating that PLSCR1 activity control directly or indirectly the size of the reluctant pool and/or its recruitment. The presence of PLSCR1 is also required for the fast mobilization of recycling SVs that support the maintenance of a high level of facilitation during prolonged stimulations. Whether SV mobilization defects are directly or indirectly related to the slowing down of SV endocytosis needs to be clarified. Furthermore, we cannot exclude alternative mechanisms underlying the facilitation defect as altered clearance of SV cargo from release sites or an increase in release probability resulting from compensatory mechanisms.
Interestingly, we detected PLSCR1 in the cerebellum and the olfactory bulb, two brain regions known to sustain high-frequency stimulation. In contrast, PLSCR1 is barely detected in other brain areas. PLSCR1 appears therefore to be a good candidate to play a specific role in high-frequency synaptic transmission by scrambling PLs at synapse, a fast way to modify plasma membrane properties.
Footnotes
This work was financially supported by the Agence Nationale pour la Recherche (“LipidTrans4NeuroTraffic,” No. ANR-19-CE16-0012-01) to S.G.; by a fellowship from la Fondation pour la Recherche Médicale (No. FDT202106013135) and a travel grant from The Company of Biologists (No. JCSTF1911344) to M.C.; and by the Wellcome Trust Investigator Award to M.A.C. (204954/Z/16/Z). For open access, we have applied a CC-BY public copyright license to any accepted manuscript version arising from this submission. INSERM is providing salary to S.C.G., S.G., and N.V. We acknowledge the Plateforme Imagerie In Vitro at CNRS UPS3256, Dr. Jean-Daniel Fauny (IBMC, UPR3512) for his technical assistance with live-cell imaging microscopy setup, and the animal facility of Chronobiotron (UMS 3415). We are grateful to Charlotte Caquineau and Audrey Groh for their technical assistance.
↵*S.G. and S.O. contributed equally to this work.
The authors declare no competing financial interests.
- Correspondence should be addressed to Petra Tóth at petra.toth{at}inci-cnrs.unistra.fr, Nicolas Vitale at vitalen{at}inci-cnrs.unistra.fr, Stéphane Gasman at sgasman{at}inci-cnrs.unistra.fr, or Stéphane Ory at ory{at}inci-cnrs.unistra.fr.
