Abstract
Sphingosine-1-phosphate (S1P), a bioactive sphingolipid concentrated in the brain, is essential for normal brain functions, such as learning and memory and feeding behaviors. Sphingosine kinase 1 (SphK1), the primary kinase responsible for S1P production in the brain, is abundant within presynaptic terminals, indicating a potential role of the SphK1/S1P axis in presynaptic physiology. Altered S1P levels have been highlighted in many neurologic diseases with endocytic malfunctions. However, it remains unknown whether the SphK1/S1P axis may regulate synaptic vesicle endocytosis in neurons. The present study evaluates potential functions of the SphK1/S1P axis in synaptic vesicle endocytosis by determining effects of a dominant negative catalytically inactive SphK1. Our data for the first time identify a critical role of the SphK1/S1P axis in endocytosis in both neuroendocrine chromaffin cells and neurons from mice of both sexes. Furthermore, our Ca2+ imaging data indicate that the SphK1/S1P axis may be important for presynaptic Ca2+ increases during prolonged stimulations by regulating the Ca2+ permeable TRPC5 channels, which per se regulate synaptic vesicle endocytosis. Collectively, our data point out a critical role of the regulation of TRPC5 by the SphK1/S1P axis in synaptic vesicle endocytosis.
SIGNIFICANCE STATEMENT Sphingosine kinase 1 (SphK1), the primary kinase responsible for brain sphingosine-1-phosphate (S1P) production, is abundant within presynaptic terminals. Altered SphK1/S1P metabolisms has been highlighted in many neurologic disorders with defective synaptic vesicle endocytosis. However, whether the SphK1/S1P axis may regulate synaptic vesicle endocytosis is unknown. Here, we identify that the SphK1/S1P axis regulates the kinetics of synaptic vesicle endocytosis in neurons, in addition to controlling fission-pore duration during single vesicle endocytosis in neuroendocrine chromaffin cells. The regulation of the SphK1/S1P axis in synaptic vesicle endocytosis is specific since it has a distinguished signaling pathway, which involves regulation of Ca2+ influx via TRPC5 channels. This discovery may provide novel mechanistic implications for the SphK1/S1P axis in brain functions under physiological and pathologic conditions.
- calcium
- cell-attached capacitance recordings
- live-cell imaging
- Sphingosine-1-phosphate
- synaptic vesicle endocytosis
- TRPC5
Introduction
Synaptic transmission, mediated by exocytosis of synaptic vesicles, is sustained by subsequent endocytosis, an essential process to recapture and reuse membranes that have fused with the plasma membrane from exocytosis (Saheki and De Camilli, 2012; Rizzoli, 2014; Soykan et al., 2016). Synaptic vesicle endocytosis is thus indispensable for normal brain functions (Saheki and De Camilli, 2012; Rizzoli, 2014; Soykan et al., 2016), and deficiency in endocytic machinery has been highlighted in many neurologic disorders including multiple sclerosis, Huntington's disease, Parkinson's disease, and Alzheimer's disease (Cataldo et al., 2000; Schreij et al., 2016; Alsaqati et al., 2018; Kyung et al., 2018; Helbig et al., 2019; McAdam et al., 2020; Pensalfini et al., 2020; Barron et al., 2021; Kesharwani et al., 2021). Nevertheless, the precise molecular mechanisms for synaptic vesicle endocytosis are still far from being fully understood (Chanaday et al., 2019; Rennick et al., 2021).
Sphingosine-1-phosphate (S1P), a bioactive sphingolipid concentrated in the brain (Edsall and Spiegel, 1999; Karunakaran and van Echten-Deckert, 2017), is essential for normal brain functions. For example, S1P in hippocampus is critical for spatial learning as assessed by water maze test (Kanno et al., 2010) or T-maze test (Weth-Malsch et al., 2016), while S1P in hypothalamus regulates energy homeostasis by controlling feeding behaviors (Silva et al., 2014). On the other hand, altered S1P metabolism has been linked to many neurologic diseases (Ghasemi et al., 2016; Karunakaran and van Echten-Deckert, 2017; Di Pardo and Maglione, 2018). In addition, fingolimod (FTY720), a S1P analog approved by Food and Drug Administration to treat multiple sclerosis (Chiba and Adachi, 2012; Ali et al., 2013), may slow down pathologic changes in the brain tissue from animal models for Alzheimer's disease (Karunakaran and van Echten-Deckert, 2017; Yin et al., 2021). These important functions of S1P within the brain under physiological and pathologic conditions reflect a role of S1P in synaptic transmission. Indeed, it has been reported that S1P modulates synaptic strength and plasticity (Kanno et al., 2010; Kempf et al., 2014; Weth-Malsch et al., 2016).
Sphingosine kinase 1 (SphK1) is the primary kinase responsible for S1P production in the brain (Fukuda et al., 2003; Bryan et al., 2008). The presynaptic SphK1 abundance (Kajimoto et al., 2007; Chan et al., 2012) indicates an involvement of the SphK1/S1P axis in synaptic transmission via presynaptic mechanisms. In line with this idea, the SphK1/S1P axis has been reported to be critical for synaptic vesicle exocytosis in neurons from both Caenorhabditis elegans and mice (Kajimoto et al., 2007; Chan et al., 2012). In non-neuronal cells, SphK1 has been found to be enriched on physiologically occurring early endocytic intermediates and a disruption of the SphK1/S1P axis induces significant endocytic defects (Shen et al., 2014; Young et al., 2016; Lima et al., 2017). In mouse and human brains, defects of synaptic vesicle endocytosis and altered S1P levels have been concurrently identified as pathologic features for neurodegenerative diseases including Alzheimer's disease and Huntington disease (Couttas et al., 2014; Pirhaji et al., 2016; Schreij et al., 2016; Lee et al., 2018; Di Pardo et al., 2019; McAdam et al., 2020; Pensalfini et al., 2020). However, it remains unknown whether the SphK1/S1P axis may regulate synaptic vesicle endocytosis in neurons.
In the present study, we evaluate potential functions of the SphK1/S1P axis in synaptic vesicle endocytosis by determining effects of a dominant negative catalytically inactive SphK1 (SphK1DN) in neurons and chromaffin cells (Pitson et al., 2000; Bonhoure et al., 2006; Gomez-Brouchet et al., 2007; Z.J. Jiang et al., 2019b). Based on data from synaptophysin-pHluorin (sypHy) based live-cell imaging in neurons and cell-attached capacitance recordings in chromaffin cells, we for the first time identify a critical role of the SphK1/S1P axis in neuronal endocytosis. Furthermore, our Ca2+ imaging indicates that the SphK1/S1P axis may be important for presynaptic Ca2+ increases during prolonged stimulations via Ca2+ permeable TRPC5 channels, which per se regulate synaptic vesicle endocytosis. Altogether, our data point out an importance of the SphK1/S1P axis in synaptic vesicle endocytosis via TRPC5 channels in neurons.
Materials and Methods
Cultures of neurons, chromaffin cells, and HEK 293 cells
In this study, newborn pups of either sexes (postnatal day 0) from C577BL/6J (RRID: IMSR JAX:000664) mouse mating cages were used and cared in accordance with the guidelines of the National Institutes of Health, as approved by the Animal Care and Use Committee of the University of Illinois at Chicago (approval number of 19-189). Neuronal culture from cortex was prepared as describe previously (Vevea and Chapman, 2020; Z.J. Jiang et al., 2021). Briefly, cortical neurons were isolated and gently dissociated with papain (Worthington; LS003127) and mechanical disruption before plated on glass coverslips (Assistant; 41001118) coated with poly-D-lysine (Sigma; P0899). Neurons were maintained in Neurobasal-A (Thermo Fisher Scientific; 10888-022) medium supplemented with B-27 (2%, Thermo Fisher Scientific; 17504001), Glutamax (2 mm, Invitrogen; 35050061), and penicillin/streptomycin at 37°C in 5% CO2 humidified incubator. To minimize potential consequences from previously reported roles of the SphK1/S1P axis or TRPC5 in neuronal development (Greka et al., 2003; Mizugishi et al., 2005; Davare et al., 2009; Puram et al., 2011), lentiviral transduction in the present study were exclusively performed in neurons at days in vitro (DIV)11–DIV12 when the initial wave of synaptogenesis is complete (Bassani et al., 2012; Wang et al., 2017; Ribeiro et al., 2019; Griggs et al., 2021; Nawalpuri et al., 2021); accordingly, live-cell imaging experiments were conducted on neurons at DIV16–DIV20.
Chromaffin cells in culture, prepared from adrenal glands of newborn pups with either sex as previously described (Gong et al., 2005; L.H. Yao et al., 2012, 2013), were maintained at 37°C in 5% CO2 humidified incubator and used within 4 d for electrophysiology. Lentiviral transduction was conducted at DIV0, and cell-attached capacitance recordings was performed 72–96 h after infection.
HEK 293 cells were cultured and maintained in DMEM supplemented with 10% fetal bovine serum (FBS), 0.1 mm MEM nonessential amino acids, 6 mm L-glutamine, 1 mm MEM sodium pyruvate, 1% Pen-Strep, and 500 µg/ml Geneticin antibiotic. Cells, which typically reached 80% confluence every 3 d, were treated with trypsin-EDTA and passaged with a 1:10 ratio for the general maintenance. Cells dissociated were seeded onto 24-well culture plates and transfected with Lipofectamine LTX Reagent (Thermo Fisher Scientific; catalog #15338100), at 40–50% confluency. Transfected cells were maintained at 37°C in a 5% CO2 humidified incubator for 48–72 h before being trypsinized and seeded to PDL-coated coverslips for whole-cell patch clamping recordings.
Cloning of DNA constructs
Plasmids containing sypHy and SynGCAMP6f were kind gifts from Dr. L. Lagnado lab (Granseth et al., 2006) and Dr. T. A. Ryan lab (de Juan-Sanz et al., 2017), respectively. The lentiviral vector pCDH-EF1-MCS-T2A-GFP was from System Biosciences. pCDH-EF1-SphK1DN-T2A-GFP, pCDH-EF1-SphK1WT (N-terminal FLAG tag), pCDH-EF1-SphK1DN (N-terminal FLAG tag), pCDH-SYN1-sypHy and pCDH-SYN1-SynGCAMP6f were previously generated in our lab (Z.J. Jiang et al., 2019b, 2021). The protein-coding cDNA clone of mouse TRPC5, clone ID 22 067 (BC112972), was purchased from transOMIC technologies. To obtain pCDH-SYN1-SphK1DN-T2A-sypHy or pCDH-SYN1-SphK1DN-T2A-synGCaMP6f, SphK1DN was amplified by PCR, together with sypHy or synGCaMP6f, and then enzyme-digested and ligated into multiple cloning sites before and after T2A in pCDH-SYN1-MCS-T2A-MCS (a tool vector we have created in lab). Thereafter, pCDH-SYN1-SphK1DN-T2A-synaptophysin-mCherry was created by replacing pHluorin in pCDH-SYN1-SphK1DN-T2A-sypHy with mCherry via restriction digest and ligation. To obtain pCDH-SYN1-sypHy-T2A-TRPC5DN, TRPC5WT was first amplified by PCR and then enzyme-digested and ligated between the BspEI and PmeI sites after T2A in pCDH-SYN1-sypHy-T2A-MCS, which was created previously in the lab from pCDH-SYN1-sypHy by deleting the stop codon in sypHy. TRPC5DN mutation (LFW-AAA; Strübing et al., 2003) was then made by QuikChange PCR (Stratagene), with pCDH-SYN1-sypHy-T2A- TRPC5WT as the template, thus creating pCDH-SYN1-sypHy-T2A-TRPC5DN. Full length SphK1, TRPC5 and their mutants, together with sypHy and synGCaMP6f, amplified by PCR were verified by Sanger sequencing.
Lentiviral productions
Low-passage HEK 293 cells were co-transfected with 10 μg pCDH lentiviral vectors containing genes of interests, along with 7-μg packaging vector psPAX2 and 3 μg envelope vector pMD2.VSVG, using the polyethylenimine-mediated transfection method. The supernatant containing lentiviral particles was collected and filtered through 0.45 μm filter to remove cell debris at 48 and 72 h. Lentiviral particles, concentrated by PEG-it precipitation kits from System Bioscience, were re-suspended in cold PBS and stored at –80°C.
Drugs and reagents
Stock solutions were prepared by dissolving the following molecules: S1P (Sigma-Aldrich, catalog #S9666) at 1 mm in methanol; FTY720 (Sigma-Aldrich, catalog #SML0700) at 100 μm in ddH2O, CAY10444 (Cayman Chemicals, catalog #10005033) at 1 mm in dimethyl formamide, and W146 (Cayman Chemicals, catalog #10009109) at 0.2 mm in methanol. Stock solutions were then diluted to final concentrations into culture medium for neurons, or bath solution for HEK293 cells. Neurons in culture were co-preincubated with CAY10444 (10 μm), a selective S1P3 antagonist, and W146 (4 μm), a selective S1P1 antagonist, for 1 h before overnight 1 nm FTY720 incubation. During electrophysiological recordings of whole-cell TRPC5 currents, 5 μm S1P is directly applied extracellularly with a pressurized local perfusion system (VC3, ALA scientific instruments) placed ∼40 μm away from the patched cell.
Immunocytochemistry
Neurons in culture were fixed with 4% paraformaldehyde for 10 min at 37°C. Fixed neurons were washed twice with PBS and then permeabilized and quenched for 15 min in PBS with 0.2% saponin and 50 mm ammonium chloride at room temperature. Samples were next incubated for 1 h in blocking buffer (PBS with 5% normal donkey serum, 4% bovine serum albumin (BSA), 0.02% sodium azide (NaN3), and 0.04% saponin) at room temperature. Primary antibodies were either diluted 1:500 (mouse anti-FLAG, clone M2, Sigma-Aldrich F1804, RRID:AB_262044) or 1:1000 (rabbit anti- synaptophysin, Synaptic Systems catalog #101002, RRID: AB_887905) in blocking buffer and incubated with neurons overnight at 4°C. Fluorescent dye-conjugated secondary antibodies (donkey anti-mouse IgG secondary antibody, Alexa Fluor 555: A-31570, RRID:AB_2536180; donkey anti-rabbit IgG secondary antibody, Alexa Fluor 488: A-21206, RRID:AB_2535792; Thermo Fisher Scientific) were diluted at 1:1000 in the blocking buffer, and incubated with cells for 1 h at room temperature. Images were collected on an Olympus IX51 microscope with a 60× oil immersion objective and captured by a Prime BSI Scientific CMOS (sCMOS) camera (Teledyne Photometrics) controlled by μManager software (https://micro-manager.org/). A Semrock TRITC-B filter cube (543/22-nm excitation, 593/40-nm emission, 562-nm dichroic mirror) was used for Alexa Fluor 555 signals, with a Semrock GFP-3053B filter cube (473/31-nm excitation, 520/35-nm emission, 495-nm dichroic mirror) for Alexa Fluor 488 signals.
Pearson's correlation coefficient (PCC) of FLAG to synaptophysin was measured using Just Another Colocalization Plugin (JACoP) plugin in Fiji (Bolte and Cordelières, 2006). Two to three fields of view with an absence of cell bodies were randomly selected from each coverslip, and PCCs were analyzed using Costes' automatic threshold in JACoP. Data are pooled from coverslips of two independent cultures.
Live-cell imaging in neurons
Experiments were performed on an Olympus IX51 microscope with a 60× oil immersion objective at room temperature except 35–37°C in Figure 2E,F, with a GFP-3035B filter cube (472/30-nm excitation, 520/35-nm emission, 495-nm dichroic mirror; Semrock) as the filter set. Neurons on coverslips were continuously perfused at a flow rate of ∼1 ml/min with bath solution (130 mm NaCl, 2.8 mm KCl, 2 mm CaCl2, 1 mm MgCl2, 10 mm glucose, 10 mm HEPES; pH 7.4 and ∼310 mOsm). To stimulate neurons, we used Chamlide magnetic perfusion chamber (Live Cell instrument) equipped with two parallel platinum electrodes (embedded within the perfusion chamber with 7-mm spacing), which serve as field electrodes to deliver electric field stimulation at 10 V/cm with 1-ms duration to induce action potentials (APs) firing in neurons; 20 μm CNQX (Tocris Bioscience) and 50 μm D-AP5 (Tocris Bioscience) were included in the bath solution to prevent recurrent activities.
To achieve high quantum efficiency, we used the Prime BSI Scientific CMOS (sCMOS) camera (Teledyne Photometrics) controlled by μManager software (micro-manager.org) for capturing images. To achieve a sampling rate of 100 Hz for experiments of single AP stimulations and minimize photobleaching for experiments of train stimulations, the electrical stimulation via a programmable isolated high-powered stimulator (AM Systems, Model 4100), illumination via λ 421 Optical Beam Combining System (Sutter Instruments) and data acquisition via camera were simultaneously controlled by TTL output signals via Igor software (RRID: SCR_000325, WaveMetrics). For Ca2+ imaging experiments using SynGCaMP6f, the light intensity of illumination was set at 40% in combination with a 10-ms exposure time for single AP experiments stimulation or 20% in combination with a 100-ms exposure time for train stimulations. Images were captured at a 1- or 2-s interval with 100-ms exposure time at 20% illumination during sypHy experiments.
To measure reacidification rate of newly formed endocytic vesicles using sypHy as described previously (Z.J. Jiang et al., 2021), neurons on coverslip were rapidly superfused with a valve controlled pressurized perfusion system (ALA Scientific). The perfusing buffer was switched between the standard pH 7.3 bath solution described above and pH 5.5 buffer with the HEPES substituted with equimolar MES, with electrical stimulation time-locked to frame acquisition and buffer changes. Specifically, surface fluorescence of pHluorin is quenched by the first application of acidic buffer (pH 5.5) for 6s to establish the baseline, and the quenched fluorescence is recovered by washing off the acidic solution. To isolate pHluorin signal from newly endocytosed SVs, acidic buffer (pH 5.5) is applied for a 30-s duration 2 s right after the cessation of stimulation train. Newly endocytosed alkaline pool of pHluorin is detached from the plasma membrane and thus resistant to quenching by the applied acidic buffer, and the decay of this portion of fluorescence to the baseline represents acidification of newly endocytosed vesicles. Therefore, reacidification rate of newly endocytosed vesicles is estimated by an exponential fit of the fluorescent signal decay from the first to the last point during the 30-s duration of acidic buffer application.
To measure vesicles exocytosis as previously described (Balaji and Ryan, 2007; Baumgart et al., 2015), boutons were stimulated in the presence or absence of 1 μm bafilomycin A1 (Cayman, catalog #11038), which eliminated interference of endocytosis by blocking reacidification of endocytosed vesicles. Tyrode's solution containing 50 mm NH4Cl were perfused at the end of each test to collapse the pH gradient across vesicle membrane thus revealing the total pool.
Image analysis
Images from sypHy experiments were analyzed in ImageJ (RRID: SCR_000415; http://rsb.info.nih.gov/ij) using custom-written plugins (http://rsb.info.nih.gov/ij/plugins/time-series.html) and Igor software using custom-written procedures. Photobleaching was <2% for all images included in our analysis and thus was not corrected. To analyze images from sypHy experiments, regions of interest (ROIs) of identical size (4 × 4 pixels) were placed in the center of individual synaptic boutons reacting to stimuli, and fluorescence changes were tracked throughout the image stack. In general, 10–25 ROIs were analyzed from each coverslip for sypHy experiments. With Igor, fluorescent intensities of individual ROIs before stimuli were averaged as baseline (F0), the fluorescent changes (ΔF) were normalized to the baseline (F0) as ΔF/F0 or the fluorescence changes induced by NH4Cl (ΔFNH4Cl) as ΔF/ΔFNH4Cl for exocytosis, or to the peak of fluorescent increase (ΔFmax) as ΔF/ΔFmax for endocytosis. Normalized ΔF of individual ROIs were taken as independent replicates (n) for each sypHy experiment, acquired from 7–12 coverslips of two to four cultures.
To analyze data for SynGCaMP6f experiments, ROIs were created as previously reported in Fiji (https://fiji.sc/; Vevea and Chapman, 2020; Z.J. Jiang et al., 2021). Briefly, ROIs (>10 pixels), defined by a series of image subtractions and thresholding, were used to measure fluorescence changes of image stacks. Next, fluorescent intensities of individual ROIs before stimuli were averaged as baseline (F0), and the fluorescent changes (ΔF) were normalized to the baseline (F0) as ΔF/F0 in Igor. The independent replicates (n) for each SynGCaMP6f study, pooled from two to four cultures, was measured as the average of normalized ΔF of all ROIs from each coverslip.
Cell-attached capacitance recordings and analyses of endocytic fission-pore
Cell-attached capacitance recordings of membrane capacitance and conductance were performed as described previously (L.H. Yao et al., 2012, 2013). Fire polished pipettes had a typical resistance of ∼2 megaohms in the bath solution. The bath solution contained 130 mm NaCl, 5 mm KCl, 2 mm CaCl2, 1 mm MgCl2, 10 mm HEPES-NaOH, and 10 mm HEPES-NaOH; the pH was adjusted to 7.3 with NaOH. As mentioned previously (L.H. Yao et al., 2012, 2013), capacitance steps and fission-pore durations were considered as reliably detected for step sizes >0.2 fF, and smaller capacitance steps are not included in the analysis. Endocytic events were recorded right after gigaseal formation, since we often detected action currents in the membrane current trace as shown previously (Dernick et al., 2005), which indicates firing of action potentials likely induced by mechanic stimulus of patching in chromaffin cells. The number of endocytic events per patch detected in the cell-attached recordings is counted as the total number of downward capacitance steps within the first 5 min of recordings (L.H. Yao et al., 2012, 2013). The fission-pore Gp for endocytic events was calculated as Gp(Im) =
Whole-cell recordings of TRPC5 currents in HEK 293 cells
TRPC5 currents were monitored in HEK 293 cells using the whole-cell configuration as previously reported (Xu et al., 2006; Gomis et al., 2008; Blair et al., 2009). The bath solution contained 130 mm NaCl, 5 mm KCl, 2 mm CaCl2, 1 mm MgCl2, 10 mm HEPES-NaOH, and 10 mm glucose; the pH was adjusted to 7.3 with NaOH, and the osmolarity was ∼310 mmol/kg. The solution in the whole-cell pipette contained 130 mm Cs-MeSO3, 5 mm Na2ATP, 0.1 mm Na3GTP, 2 mm MgCl2, 1.3 mm CaCl2, 2 mm EGTA-CsOH, and 10 mm HEPES; the pH was adjusted to 7.3 with CsOH, and the osmolarity was adjusted to 290 mmol/kg. The voltage ramp protocol holds cells at −60 mV, steps to –100 mV for 40 ms, and then ramps to +100 mV over 500 ms, holding at +100 mV for 40 ms before stepping back to −60 mV, and this protocol is repeated every 5 s during recordings. Series resistance was monitored by applying a voltage step of 10 mV. Current signals were filtered at 3 kHz and digitized at 20 kHz. Only recordings with an initial pipette-membrane seal resistance >2 GΩ were included. Current traces and time courses were analyzed with customized macro for Igor software.
Statistical analysis
Data were tested for normal distribution and, if necessary, log-transformed for the fission-pore duration in Figures 3E to fulfill the criteria of normal distributions for Student's t tests. All statistical analyses were performed with Prism (RRID: SCR_002798, GraphPad).
No statistical method was used to predetermine sample size, but our samples sizes are comparable to those reported in previous studies (Fernández-Alfonso and Ryan, 2004; Granseth et al., 2006; L.H. Yao et al., 2012, 2013; Soykan et al., 2017; Z.J. Jiang et al., 2021). Cells were randomly assigned into control or experimental groups, and data were expressed as mean ± SEM, with all values from quantification plotted as gray dots in the bar graph. Statistical analysis was performed with unpaired two-tailed Student's t test, except paired two-tailed Student's t test in Figure 8H and Newman of one-way ANOVA in Figures 4B,D, 7B, and 8B. All analyses were performed by an investigator blinded to the experimental condition. N numbers, full statistics, and p values were reported and listed in the related figure legends and Extended Data Table 1-1.
Extended Data Table 1-1
The list of statistical analyses for all figures. Download Table 1-1, DOCX file.
Results
SphK1DN inhibits the readily release Pool (RRP) during exocytosis in neurons
SphK1, a kinase essential for S1P production in central nervous system, is concentrated within presynaptic terminals in neurons (Fukuda et al., 2003; Kajimoto et al., 2007; Bryan et al., 2008; Chan et al., 2012). A dominant-negative mutant, SphK1DN (G82D) is reported to significantly inhibit S1P production (Lan et al., 2011; Pitson et al., 2002; Qi et al., 2015). In neuroendocrine chromaffin cells, we have previously demonstrated that downregulation of S1P production by SphK1DN induces significant defects in vesicle exocytosis (Z.J. Jiang et al., 2019b). In neurons, SphK1DN expressed by lentivirus displays a similar presynaptic localization as wild-type (WT) SphK1 (SphK1WT; Fig. 1A,B), which is not surprising as this SphK1DN mutant only harbors a single amino acid substitution in its kinase catalytic domain (Pitson et al., 2000; Alemany et al., 2007). We next analyzed effects of SphK1DN on synaptic vesicle recycling in cortical neuron cultures using sypHy (Granseth and Lagnado, 2008; Royle et al., 2008). SphK1DN significantly reduced the fluorescence increase of sypHy signals induced by 100 APs at 20-Hz stimulations (p < 0.001; Fig. 1C,D), indicating a role of SphK1 in synaptic vesicle exocytosis. It is aware that such an increase in sypHy fluorescence likely represents a net outcome of synaptic vesicle exocytosis and endocytosis (Fernández-Alfonso and Ryan, 2004; Voglmaier et al., 2006). To eliminate any interference of endocytic signals, the increase in sypHy signals during stimulations were measured in the presence of the H+ ATPase inhibitor bafilomycin A1 (Granseth et al., 2006; Z.J. Jiang et al., 2021). To our expectation, the increase of sypHy signals was also inhibited by SphK1DN under this condition, thus confirming a role of SphK1 kinase activity in synaptic vesicle exocytosis (p < 0.001; Fig. 1E,F).
SphK1DN inhibits RRP during exocytosis. A, Representative micrographs of immunofluorescence of synaptophysin and exogenously expressed SphK1WT or SphK1DN show similar presynaptic localizations of these two SphK1 versions. Scale bar: 5 μm. B, No significant differences in the Pearson's correlation coefficient of SphK1WT or SphK1DN to synaptophysin (SphK1WT: n = 10 fields of view from 4 coverslips; SphK1DN: n = 10 fields of view from 4 coverslips, p = 0.6367). C, Normalized changes of sypHy fluorescence in response to a 100 APs at 20-Hz stimulation in control or SphK1DN-expressing neurons. D, Bar graph comparing the peak of normalized fluorescence increases in control or SphK1DN-expressing neurons (control: n = 167 boutons; SphK1DN: n = 166 boutons). E, F, In the presence of bafilomycin A1, time course (E) and bar graph (F) of increases in the sypHy signals induced by a 100 APs at 20-Hz stimulation in control or SphK1DN-expressing neurons. Fluorescence increases induced by electrical stimulations are normalized to that by NH4Cl application (control: n = 128 boutons; SphK1DN: n = 131 boutons). G, In the presence of bafilomycin A1, time course of normalized sypHy fluorescence increases induced by a 600 APs at 10-Hz stimulation in control and SphK1DN-expressing neurons. H, Bar graph showing no effects of SphK1DN on the time constant of TRP depletion kinetics, which is measured by exponentially fitting sypHy increases during stimulations (p = 0.6656). I, Bar graph showing SphK1DN has no influences on TRP size (p = 0.9057; control: n = 145 boutons; SphK1DN: n = 147 boutons). J, In the presence of bafilomycin A1, time course of normalized sypHy fluorescence increases induced by a 40 APs at 20-Hz stimulation in control and SphK1DN-expressing neurons. K, Bar graph showing a reduction in the RRP size by SphK1DN (control: n = 89 boutons; SphK1DN: n = 89 boutons). L, In the presence of bafilomycin A1, time course of fluorescence increases of sypHy induced by two successive 40 APs at 20-Hz stimulations with a time interval of 5 s in control and SphK1DN-expressing neurons. Fluorescence changes are normalized to those induced by the first stimulation. M, Bar graph showing a smaller recovery of RRP in neurons expressing SphK1DN (control: n = 129 boutons; SphK1DN: n = 121 boutons). **p < 0.01 and ***p < 0.001, unpaired two-tailed Student's t test.
We next determined effects of SphK1DN on the total recycling pool (TRP) and readily releasable pool (RRP), two pools that are essential in determining the amount of synaptic vesicle exocytosis (S.H. Kim and Ryan, 2010; Alabi and Tsien, 2012; Baumgart et al., 2015). As reported previously, TRP was evaluated as fluorescence increases of sypHy in response to stimulations of 600 APs at 10 Hz (Fig. 1G; Ryan and Smith, 1995; Z. Hua et al., 2011). Our results showed that SphK1DN did not alter either the size or the depletion kinetics of TRP (p > 0.05; Fig. 1H,I). On the other hand, RRP was evaluated as sypHy increases in response to 40 APs at 20-Hz stimulation (Fig. 1J; Murthy and Stevens, 1999; Schikorski and Stevens, 2001; Li et al., 2005; Granseth et al., 2006; Granseth and Lagnado, 2008). Additionally, the replenishment of RRP was estimated as fluorescent increase induced by the second 40 APs at 20-Hz stimulation 5 s after RRP depletion by the first 40 APs at 20-Hz stimulation, as described previously (Granseth and Lagnado, 2008). SphK1DN significantly reduced RRP size (p < 0.001; Fig. 1J,K) and slowed down RRP replenishment (p < 0.01; Fig. 1L,M). Collectively, this set of experiments indicate that SphK1 kinase activity may be critical for synaptic vesicle exocytosis by regulating RRP rather than TRP within presynaptic terminals.
SphK1DN slows down synaptic vesicle endocytosis in neurons
To determine any potential roles of the SphK1/S1P axis in synaptic vesicle endocytosis, we examined effects of SphK1DN on the decay kinetics of sypHy signals on the cessation of stimulations (Fig. 2A). SphK1DN substantially increased the time constant of sypHy fluorescence decay (p < 0.001; Fig. 2B), with no alternation in the reacidification kinetics of newly formed endocytic vesicles (p > 0.05; Fig. 2C,D), indicating an importance of SphK1 kinase activity in synaptic vesicle endocytosis. In addition, we observed an increase in the endocytic time constant induced by SphK1DN at the physiological temperature of 35–37°C (p < 0.001; Fig. 2E,F), thus reiterating an importance of SphK1 kinase activity in synaptic vesicle endocytosis in neurons.
Synaptic vesicle endocytosis is impaired in SphK1DN-expressing neurons. A, Normalized fluorescence changes of sypHy signals in control and SphK1DN-expressing neurons. B, Bar graph showing an increase in the endocytic time constant by SphK1DN (control: n = 121 boutons; SphK1DN: n = 117 boutons). C, Normalized sypHy signals with two perfusion periods with pH 5.5 MES buffer as indicated in the trace (gray bars). Newly endocytosed vesicles “trapped” during acidic buffer perfusion starting 2 s after 100 APs at 20 Hz. D, Bar graph comparing the reacidification rate of endocytic vesicles, which was obtained by exponential fittings of fluorescence decay, indicates no change in this parameter by SphK1DN (control: n = 117 boutons; SphK1DN: n = 115 boutons, p = 0.4198). E, Normalized changes of sypHy signals in control and SphK1DN-expressing neurons at 35–37°C. F, Bar graph comparing the endocytic time constant shows an increase in the endocytic time constant induced by SphK1DN at physiological temperature (control: n = 101 boutons; SphK1DN: n = 100 boutons). ***p < 0.001, unpaired two-tailed Student's t test.
SphK1DN prolongs fission-pore duration during endocytosis in chromaffin cells
To understand whether the SphK1/S1P axis is involved in the modulation of vesicle fission, a critical and arguably rate limiting step during endocytosis (Haucke et al., 2011; Saheki and De Camilli, 2012; Kononenko and Haucke, 2015; Kaksonen and Roux, 2018), we analyzed effects of SphK1DN on the fission-pore kinetics in neuroendocrine chromaffin cells by monitoring single vesicle endocytosis with cell-attached capacitance recordings as described previously (Fig. 3A; L.H. Yao et al., 2012; Varga et al., 2020; Z.J. Jiang et al., 2021). SphK1DN induced a 17.8% reduction in the number of endocytic events, although this reduction was not significant (Fig. 3B). Meanwhile, the capacitance size of endocytic vesicles was comparable between these two groups (Fig. 3C), indicating that SphK1 may not be critical in controlling the size of endocytic vesicles. Additionally, SphK1DN had no obvious effect on the fission-pore conductance (Gp; Fig. 3D), implying that SphK1 may not be essential in determining the geometry of the tubular membrane neck during endocytosis. However, SphK1DN significantly increased the fission-pore duration (p < 0.001; Fig. 3E), indicating an essential role of SphK1 kinase activity in vesicle fission during endocytosis.
SphK1DN slows down vesicle fission during endocytosis in neuroendocrine chromaffin cells. A, Representative endocytic events, as membrane conductance (Re), membrane capacitance (Im) and the fission-pore conductance (Gp), in control and SphK1DN-expressing cells. B–E, The number of endocytic events (control: n = 83 cells; SphK1DN: n = 69 cells, p = 0.2501; B), the capacitance Cv of endocytic vesicles (p = 0.8041; C), and the fission-pore Gp (p = 0.7110; D) were statistically comparable between control and SphK1DN-expressing cells; the log transformed fission-pore duration (E) was significantly increased in SphK1DN-expressing cells (control: n = 36 events; SphK1DN: n = 48 events). ***p < 0.001, unpaired two-tailed Student's t test.
S1P receptors are crucial for the SphK1/S1P axis-mediated actions in synaptic vesicle exocytosis but not endocytosis
FTY720, a clinical oral treatment of relapsing remittent multiple sclerosis (Chiba and Adachi, 2012; Ali et al., 2013), is a synthetic S1P analog (Nofer et al., 2007). We thus examined whether FTY720 could rescue exocytic and endocytic defects of synaptic vesicles induced by SphK1DN. Neurons in culture expressing SphK1DN were incubated with FTY720 at a low concentration of 1 nm overnight to mimic in vivo applications in animals (Stessin et al., 2012; Di Menna et al., 2013; Zhang et al., 2017). FTY720 treatments significantly enhanced the fluorescence increase of sypHy in response to 100 APs at 20-Hz stimulations (Fig. 4A,B) and reduced the endocytic time constant of sypHy signal decay after the cessation of stimulations (Fig. 4C,D), suggesting that FTY720 may be capable of rescuing SphK1DN induced exocytic and endocytic defects in neurons. These data emphasize an importance of S1P in SphK1DN-mediated actions in synaptic vesicle exocytosis and endocytosis, thus emphasizing critical roles of the SphK1/S1P axis in synaptic vesicle recycling within presynaptic terminals. Notably, it has been reported that the S1P-induced effect on synaptic transmission at presynaptic terminals is mostly mediated through activation of S1P3 and/or S1P1 receptors (Kajimoto et al., 2007; Sim-Selley et al., 2009; Kanno and Nishizaki, 2011; Riganti et al., 2016; Willems et al., 2016). Hence, we next examined whether co-preincubations of W146, a specific S1P1 receptor inhibitor (Sanna et al., 2006; Ammar et al., 2013; Riganti et al., 2016), and CAY10444, a specific S1P3 receptor inhibitor (Koide et al., 2002; Pyne and Pyne, 2011; Riganti et al., 2016), would be able to block FTY720-mediated rescues we have observed. W146 and CAY10444 treatments indeed abolished FTY720-mediated rescue on the increase of sypHy fluorescence induced by stimulations (Fig. 4A,B), suggesting that S1P1/S1P3 receptors may mediate the role of the SphK1/S1P axis in synaptic vesicle exocytosis. On the other hand, treatments with W146 and CAY10444 had no detectable effects on the FTY720-mediated rescue on the endocytic time constant on the cessation of stimulations (Fig. 4C,D), indicating that S1P1/S1P3 receptors may be dispensable for the regulation of the SphK1/S1P axis in synaptic vesicle endocytosis.
S1P1/S1P3 receptors are critical for the FTY720-mediated rescue on synaptic vesicle exocytosis but not endocytosis in SphK1DN-expressing neurons. A, In the presence of bafilomycin A1, synaptic vesicle exocytosis induced by electrical stimulations was measured by normalizing the fluorescence increase induced by electrical stimulation to that by NH4Cl perfusion in SphK1DN-expressing neurons under variable conditions (control, FTY720, or FTY720/W146/CAY10444). FTY720 is a clinical S1P analog; W146 and CAY10444 are specific antagonists for S1P1 and S1P3 receptors, respectively. B, Bar graph comparing synaptic vesicle exocytosis in SphK1DN-expressing neurons between these three groups as shown in A (control: n = 95 boutons; FTY720: n = 85 boutons; FTY720 and S1PR antagonists: n = 106 boutons). One-way ANOVA followed by Tukey's post hoc test: FTY720 versus control (*p < 0.05); FTY720 and S1PR antagonists versus FTY720 (***p < 0.001). C, Normalized changes of sypHy signals in SphK1DN-expressing neurons under different conditions (control, FTY720, or FTY720 and S1P1/S1P3 inhibitors). D, Bar graph comparing the endocytic time constant in SphK1DN-expressing neurons shows that FTY720 treatment accelerates synaptic vesicle endocytosis independent of S1PR inhibitors (control: n = 162 boutons; FTY720: n = 159 boutons; FTY720 and S1PR inhibitors: n = 178 boutons). One-way ANOVA followed by Tukey's post hoc test: FTY720 versus control (**p < 0.01); FTY720 and S1PR inhibitors versus FTY720 (p = 0.2759; ns: not statistically significant).
The SphK1/S1P axis is important for presynaptic Ca2+ levels during prolonged stimulations
Previous studies in non-neuronal cells have implied important roles of S1P in modulating intracellular Ca2+ signaling (Ghosh et al., 1990; Itagaki and Hauser, 2003; Spiegel and Milstien, 2003; Xu et al., 2006). Since Ca2+ is a key regulating factor for synaptic vesicle endocytosis (Cousin, 2000; Wu et al., 2009; Wu and Wu, 2014; Chanaday et al., 2019; Z.J. Jiang et al., 2021), it is possible that a regulation of presynaptic Ca2+ signaling may account for the action of the SphK1/S1P axis in synaptic vesicle endocytosis we have observed. To test this possibility, we examined effects of SphK1DN on presynaptic Ca2+ signaling using synGCaMP6f (GCaMP6f fused with synaptophysin), a presynaptic Ca2+ reporter with fast speed and high sensitivity (Chen et al., 2013; Kyung et al., 2017; Brockhaus et al., 2019; Z.J. Jiang et al., 2021). SphK1DN did not alter the increase of synGCaMP6f signals induced by single AP stimulation (Fig. 5A,B; p > 0.05), indicative of no change in Ca2+ influx via VGCCs (S.H. Kim and Ryan, 2013; Brockhaus et al., 2019; Z.J. Jiang et al., 2021). It should be noted that a deflection of the SynGCaMP6f fluorescence signal was observed in the decay phase at around 110 ms following single AP stimulation (Fig. 5A). We suggest that this deflection may be because of, but not limited to, that single AP fluorescence responses of individual ROIs peak within the 110-ms period after stimulus, with variability in the time of the signal to peak (Fig. 6A–D), and its presence has no impact on the exponential decay of the averaged fluorescent signal (Fig. 6E,F). On the other hand, the increase in synGCaMP6f signals induced by 100 APs at 20-Hz stimulations was significantly reduced by SphK1DN (p < 0.05; Fig. 5C,D), suggesting an importance of the SphK1/S1P axis in regulating presynaptic Ca2+ influx during prolonged stimulations. This conclusion is further supported by our observation that FTY720 enhanced the increase in synGCaMP6f signals induced by 100 APs at 20-Hz stimulations in SphK1DN-expressing neurons (p < 0.001), which is independent of S1P1/S1P3 receptors (Fig. 7). Presynaptic Ca2+ influx on prolonged stimulations has been well documented to be essential for synaptic vesicle endocytosis (C.K. Yao et al., 2009, 2017; Wu and Wu, 2014; Z.J. Jiang et al., 2021), our data thus indicate that the SphK1/S1P axis may regulate synaptic vesicle endocytosis by modulating presynaptic Ca2+ levels during prolonged stimulations.
SphK1DN impairs presynaptic Ca2+ increase in response to prolonged stimulations rather than single AP stimulation. A, Normalized synGCaMP6f increases induced by single AP in control or SphK1DN-expressing neurons. Gray lines are mean fluorescence changes of individual ROIs from each coverslip, with black and red line as the averaged response of individual coverslips in control and SphK1DN group, respectively. Raw data for Ca2+ increases induced by single AP are presented in Figure 6. B, Bar graph comparing peak synGCaMP6f values as shown in A reveals no change in Ca2+ signals by SphK1DN (control: n = 21 coverslips; SphK1DN: n = 20 coverslips, p = 0.4865). C, Normalized synGCaMP6f increases induced by 100 APs at 20-Hz stimulations in control or SphK1DN-expressing neurons. Gray lines are mean fluorescence changes of individual ROIs from each coverslip, with black and red line as the averaged response of individual coverslips in control and SphK1DN group, respectively. D, Bar graph comparing peak synGCaMP6f values as shown in C demonstrates a defective Ca2+ signals by SphK1DN (control: n = 27 coverslips; SphK1DN: n = 27 coverslips). In addition, this SphK1DN-induced presynaptic Ca2+ reductions can be rescued by FTY720 application, independent of S1P1/S1P3 receptors (Fig. 7). *p < 0.05, unpaired two-tailed Student's t test.
Raw data for Ca2+ increases induced by single AP in control and SphK1DN-expressing neurons. A, Time courses of normalized synGCaMP6f increases induced by single AP in control neurons. Gray lines are mean fluorescence changes of individual ROIs from each coverslip, with the black line as the averaged response of individual coverslips. As shown on top of the traces, peak Ca2+ signal induced by 1 AP from individual ROIs occurred within 110 ms, ranging from the first to 11th frame after stimulus as shown with color coded diagram (upper right). B, Color coded pie charts showing frame location distributions of Ca2+ peaks for all ROIs in coverslips “a” (left) and “b” (right). Along with each pie chart, two color coded traces from each coverslip showed representative ROIs with Ca2+ peaked at the fourth and sixth frame. C, Time courses of normalized synGCaMP6f increases induced by single AP in SphK1DN-expressing neurons. Gray lines are mean fluorescence changes of individual ROIs from each coverslip, with the red line as the averaged response of individual coverslips. As shown on top of the traces, peak Ca2+ signal induced by 1 AP from individual ROIs occurred within 110 ms, ranging from the first to 11th frame after stimulus as shown with color coded diagram (upper right). D, Color coded pie charts showing frame location distributions of Ca2+ peaks from all ROIs in coverslips “a*” (left) and “b*” (right) from SphK1DN group. Along with each pie chart, two color coded traces from each coverslip showed representative ROIs with Ca2+ peaked at the second and seventh frame. E, Truncated (above) and full (below) time courses of mean synGCaMP6f signals induced by single AP of coverslip “a” or “b” from control group, with decay phases fitted by single exponential functions (dotted gray lines). F, Truncated (above) and full (below) time courses of mean synGCaMP6f signals induced by single AP of coverslip “a*” or “b*” from SphK1DN group, with decay phases of each mean Ca2+ signal fitted by single exponential functions (dotted gray lines).
FTY720 rescues presynaptic Ca2+ increase in response to prolonged stimulations in SphK1DN-expressing neurons, independent of S1P1/S1P3 receptors. A, Normalized synGCaMP6f increases induced by 100 APs at 20-Hz stimulations in SphK1DN-expressing neurons under variable conditions (control, FTY720, or FTY720/W146/CAY10444). Gray lines are mean fluorescence changes of individual ROIs from each coverslip, with red, blue, and yellow line as the averaged response of individual coverslips from control, FTY720, and FTY720/W146/CAY10444 group, respectively. B, Bar graph comparing peak synGCaMP6f values as shown in A demonstrates that FTY720 treatment increases Ca2+ influx in SphK1DN-expressing neurons, which is not affected by the pretreatment of S1PR inhibitors (control: n = 22 coverslips; FTY720: n = 20 coverslips; FTY720/W146/CAY10444: n = 20 coverslips). All experiments are performed in synapses expressing the SphK1DN mutant, one-way ANOVA followed by Tukey's post hoc test: FTY720 versus control (***p < 0.001); FTY720 and S1PR inhibitors versus FTY720 (p = 0.4651; ns: not statistically significant).
S1P modulates presynaptic Ca2+ levels via TRPC5 channels
It is reported that the TRPC5 channel, a Ca2+ permeable member of canonical transient receptor potential channel (TRPC) superfamily (Clapham, 2003; Venkatachalam and Montell, 2007), is located to presynaptic terminals (Greka et al., 2003; Nichols et al., 2007; Schwarz et al., 2019). Additionally, TRPC5 channels may be involved in maintaining presynaptic Ca2+ levels during stimulation trains (Schwarz et al., 2019). Notably, S1P is a physiological agonist for Ca2+ influx via TRPC5 channels (Xu et al., 2006; Venkatachalam and Montell, 2007; Naylor et al., 2016). It is therefore possible that the regulation of TRPC5 by the SphK1/S1P axis may account for the SphK1DN induced presynaptic Ca2+ defect during prolonged stimulations we have observed (Fig. 5C,D). We next explored this possibility by examining effects of SphK1DN on presynaptic Ca2+ levels in neurons expressing a dominant negative nonconducting TRPC5 (TRPC5DN), which was created by site-specific mutagenesis of WT TRPC5 (TRPC5WT; LFW-AAA) as reported (Greka et al., 2003). The nonconducting and dominant negative properties of this TRPC5DN channel was confirmed by whole-cell patch recordings in HEK 293 cells (Fig. 8A,B). Furthermore, ionic currents of TRPCWT channels were significantly enhanced by exogenously applied S1P (p < 0.001; Fig. 8G,H), confirming the S1P-mediated regulation of TRPC5 as reported by previous Ca2+ imaging studies (Xu et al., 2006; Naylor et al., 2016). Consistent with a previous study on hippocampal neurons using transgenic knock-out (KO) and knock-in mice (Schwarz et al., 2019), TRPC5DN induced a significant decrease in presynaptic Ca2+ levels in response to 100 APs at 20-Hz stimulations (p < 0.05; Fig. 8E,F) rather than single AP stimulation (p > 0.05; Figs. 8C,D, 9), implying an importance of TRPC5 in maintaining presynaptic Ca2+ levels during prolonged stimulations. Moreover, in contrast to an inhibition of presynaptic Ca2+ influx in control neurons (Fig. 5C,D), SphK1DN had no obvious effects on presynaptic Ca2+ levels in response to 100 APs at 20-Hz stimulations in TRPC5DN-expressing neurons (p > 0.05; Fig. 8I,J), pointing out that Ca2+ influx via TRPC5 channels may be critical for the role of the SphK1/S1P axis in presynaptic Ca2+ signaling during prolonged stimulations.
Ca2+ influx via TRPC5 is necessary for the action of the SphK1/S1P axis on presynaptic Ca2+ levels in responses to prolonged stimulations. A, Representative whole-cell currents recorded in HEK293 cells expressing TRPC5WT (black), TRPC5DN (green), or TRPC5DN and TRPC5WT (red), in response to a voltage ramp from −100 to +100 mV. B, Quantifications showing substantial reductions of whole-cell currents in cells expressing TRPC5DN or TRPC5DN/TRPC5WT, as compared with TRPC5WT, at both –100 and +100 mV (TRPC5WT: n = 16 cells; TRPC5DN: n = 18 cells; co-transfection: n = 15 cells; one-way ANOVA followed by Tukey's post hoc test, TRPC5DN or TRPC5DN/TRPC5WT vs TRPC5WT: ***p < 0.001). C, Normalized increases of synGCaMP6f fluorescence induced by single AP in control or TRPC5DN-expressing neurons. Gray lines are mean fluorescence changes of individual ROIs from each coverslip, with black and red line as the averaged response of individual coverslips from control and TRPC5DN groups, respectively. Raw data for Ca2+ increases induced by single AP are presented in Figure 9. D, Bar graph comparing peak synGCaMP6f values as shown in C reveals no change in Ca2+ signals in TRPC5DN-expressing neurons (control: n = 20 coverslips; TRPC5DN: n = 19 coverslips, p = 0.4639). E, Normalized SynGCaMP6f increases induced by 100 APs at 20 Hz in control or TRPC5DN-expressing neurons. Gray lines are mean fluorescence changes of individual ROIs from each coverslip, with black or red line as the averaged response of individual coverslips from control or TRPC5DN group. F, Bar graph comparing peak Ca2+ values as shown in E demonstrates a significant decrease of Ca2+ signals in TRPC5DN-expressing neurons (control: n = 22 coverslips; SphK1DN: n = 22 coverslips). G, Representative whole-cell TRPC5 currents before (black) and after (red) 5 μm S1P application, recorded from a TRPC5WT-expressing HEK293 cell, in response to a voltage ramp from −100 to +100 mV. H, Quantifications showing significant increases of whole-cell TRPC5 currents by S1P application (red vs black) at both +100 mV (top) and −100 mV (bottom; n = 16 cells; paired t test, ***p < 0.001). I, Normalized fluorescence increases of SynGCaMP6f signals induced by 100 APs at 20 Hz from TRPC5DN or SphK1DN/TRPC5DN-expressing neurons. Gray lines are mean fluorescence changes of individual ROIs from each coverslip, with black or red line as the averaged response of individual coverslips from TRPC5DN or SphK1DN/TRPC5DN-expressing neurons. J, Bar graph comparing peak Ca2+values as shown in I demonstrates that SphK1DN had no significant effects on Ca2+ signals in TRPC5DN-expressing neurons (TRPC5DN: n = 24 coverslips; SphK1DN/TRPC5DN: n = 24 coverslips, p = 0.4856). *p < 0.05, unpaired two-tailed Student's t test in D, F, and J.
Raw data for Ca2+ increases induced by single AP in control and TRPC5DN-expressing neurons. A, Time courses of normalized synGCaMP6f increases induced by single AP in control neurons. Gray lines are mean fluorescence changes of individual ROIs from each coverslip, with the black line as the averaged response of individual coverslips. As shown on top of the traces, peak Ca2+ signal induced by 1 AP from individual ROIs occurred within 110 ms, ranging from the first to 11th frame after stimulus as shown with color coded diagram (upper right). B, Color coded pie charts showing frame location distributions of Ca2+ peaks for all ROIs in coverslips “c” (left) and “d” (right). Along with each pie chart, two color coded traces from each coverslip showed representative ROIs with Ca2+ peaked at the fourth and sixth frame. C, Time courses of normalized synGCaMP6f increases induced by single AP in TRPC5DN-expressing neurons. Gray lines are mean fluorescence changes of individual ROIs from each coverslip, with the red line as the averaged response of individual coverslips. As shown on top of the traces, peak Ca2+ signal induced by 1 AP from individual ROIs occurred within 110 ms, ranging from the first to 11th frame after stimulus as shown with color coded diagram (upper right). D, Color coded pie charts showing frame location distributions of Ca2+ peaks from all ROIs in coverslips “c*” (left) and “d*” (right) from TRPC5DN group. Along with each pie chart, two color coded traces from each coverslip showed representative ROIs with Ca2+ peaked at the second and seventh frame. E, Truncated (above) and full (below) time courses of mean synGCaMP6f signals induced by single AP of coverslip “c” or “d” from control group, with decay phases of Ca2+ signals fitted by single exponential functions (dotted gray lines). F, Truncated (above) and full (below) time courses of mean synGCaMP6f signals induced by single AP of coverslip “c*” or “d*” from TRPC5DN group with decay phases of mean Ca2+ signals fitted by single exponential functions (dotted gray lines).
TRPC5 is critical for the role of the SphK1/S1P axis in synaptic vesicle endocytosis
Consistent with critical roles of TRPC5 channels in mediating Ca2+ influx on train stimulations (Fig. 8E,F), TRPC5DN induced a significant increase in the time constant of sypHy signal decay (p < 0.05; Fig. 11A,B), implying an importance of Ca2+ influx via TRPC5 in synaptic vesicle endocytosis. On the other hand, there was no change in sypHy fluorescence increase induced by electric stimulations in TRPC5DN-expressing neurons as compared with control neurons (p > 0.05; Fig. 11C,D). We reasoned that any potential reductions in sypHy increase induced by TRPC5DN could be masked by a comparable reduction in endocytosis during stimulation, since a significant defect in synaptic vesicle endocytosis induced by TRPC5DN has been observed as shown in Figure 11A,B. Indeed, in the presence of bafilomycin to block vesicle reacidification because of endocytosis, TRPC5DN induced a small, although nonsignificant, reduction in the increase of pHluorin signals induced by 100 APs at 20-Hz stimulations (p > 0.05; Fig. 10). Collectively, our data indicate that Ca2+ influx via TRPC5 may be specifically required for synaptic vesicle endocytosis. Given the dependence on Ca2+ influx via TRPC5 for the role of the SphK1/S1P axis in presynaptic Ca2+ levels (Fig. 8I,J), we then analyzed effects of SphK1DN on synaptic vesicle endocytosis in TRPC5DN-expressing neurons. In contrast to an increase in the time constant of sypHy signal decay after the cessation of stimulations induced by SphK1DN expression in control neurons (Fig. 2A,B), SphK1DN induced no detectable changes in this parameter in TRPC5DN-expressing neurons (Fig. 11E,F), indicating that Ca2+ via TRPC5 channels may be necessary for the SphK1/S1P axis-mediated action in synaptic vesicle endocytosis. However, similar as SphK1DN expression in control neurons (Fig. 1C,D), SphK1DN reduced the increase in pHluorin signals induced by 100 APs at 20 Hz in TRPC5DN-expressing neurons (p < 0.001; Fig. 11G,H), indicating that the role of the SphK1/S1P axis in exocytosis may be independent of TRPC5. Taken together, our data point out that the regulation of TRPC5 channels by the SphK1/S1P axis may be especially crucial for synaptic vesicle endocytosis in neurons.
TRPC5DN has no significant effect on synaptic vesicle exocytosis in neurons. A, In the presence of bafilomycin A1, time courses of averaged increases in the sypHy signals induced by 100 APs at 20-Hz stimulations in control or TRPC5DN-expressing neurons. Fluorescence increases induced by electrical stimulations are normalized to that by NH4Cl to minimize potential impacts from sypHy expression variations. B, Bar graph showing comparable peak values of normalized fluorescence increases in control or TRPC5DN-expressing neurons (control: n = 143 boutons; TRPC5DN: n = 139 boutons; p = 0.1096). Unpaired two-tailed Student's t test. ns: not statistically significant.
TRPC5 is required for the role of the SphK1/S1P axis in synaptic vesicle endocytosis in neurons. A, Normalized changes of sypHy signals in response to 100 APs at 20-Hz stimulations in control and TRPC5DN-expressing neurons. B, Bar graph showing an increase in the endocytic time constant in TRPC5DN-expressing neurons (control: n = 163 boutons; TPRC5DN: n = 154 boutons). C, Time courses of averaged increases in the sypHy signals induced by 100 APs at 20-Hz stimulations in control or TRPC5DN-expressing neurons. Fluorescence increases induced by electrical stimulations are normalized to that by NH4Cl perfusion at the end of each trial to minimize potential impacts of sypHy expression variations. The time constant (τ) of pHluorin signal decay for the averaged response of control or TRPC5DN group was obtained by fitting the fluorescence decay with a single exponential function (gray lines). D, Bar graph showing comparable peak values of normalized fluorescence increases in control or TRPC5DN-expressing neurons (control: n = 142 boutons; TRPC5DN: n = 118 boutons; p = 0.1002). In the presence of bafilomycin, TRRPC5DN induces a small, but nonsignificant, reduction in exocytosis (Fig. 10). E, Normalized changes of sypHy fluorescence in TRPC5DN or SphK1DN/TRPC5DN-expressing neurons. F, Bar graph showing that SphK1DN had no detectable effects on the endocytic time constant in TRPC5DN-expressing neurons (TPRC5DN: n = 150 boutons; SphK1DN/TRPC5DN: n = 150 boutons, p = 0.6463). G, Time courses of averaged increases in the sypHy signals induced by a 100 APs at 20-Hz stimulation in TRPC5DN or SphK1DN/TRPC5DN-expressing neurons. Fluorescence increases induced by electrical stimulations are normalized to that by NH4Cl perfusion at the end of each trial. The time constant (τ) of pHluorin signal decay for the averaged response of TRPC5DN or SphK1DN/TRPC5DN group was obtained by fitting the fluorescence decay with a single exponential function (gray lines). H, Bar graph showing that SphK1DN causes a significant reduction of exocytosis in TRPC5DN-expressing neurons (TRPC5DN: n = 115 boutons; SphK1DN/TRPC5DN: n = 113 boutons). *p < 0.05, ***p < 0.001, unpaired two-tailed Student's t test.
Discussion
By showing an endocytic defect induced by SphK1DN (Fig. 2), we demonstrate an essential role of the SphK1/S1P axis in synaptic vesicle endocytosis in neurons. This conclusion is reiterated by the evidence that the SphK1DN induced endocytic defect can be rescued by a clinical S1P analog FTY720 (Fig. 4C,D). Meanwhile, Ca2+ imaging shows that SphK1DN induces a decreased Ca2+ level during prolonged stimulations in control neurons (Fig. 5), indicating an importance of the SphK1/S1P axis in synaptic vesicle endocytosis by regulating presynaptic Ca2+ signaling. Furthermore, SphK1DN induced no detectable effects on either presynaptic Ca2+ influx or synaptic vesicle endocytosis in TRPC5DN-expressing neurons, implying that Ca2+ influx via TRPC5 may be necessary for the SphK1/S1P axis-mediated action. Collectively, our data point out that regulations of Ca2+ influx via TRPC5 channels by the SphK1/S1P axis may be important for synaptic vesicle endocytosis in neurons. On the other hand, SphK1DN induces a defect in synaptic vesicle exocytosis (Fig. 1), which can also be rescued by exogenous FTY720 application (Fig. 4A,B), confirming critical roles of the SphK1/S1P axis in synaptic vesicle exocytosis as reported (Kajimoto et al., 2007; Chan et al., 2012; Riganti et al., 2016; Z.J. Jiang et al., 2019b). Moreover, this FTY720-mediated rescue is abolished by S1P receptor inhibitors (Fig. 4A,B), suggesting that S1P receptors may be responsible for the action of the SphK1/S1P axis in synaptic vesicle exocytosis in neurons. In view of different downstream molecular pathways identified in our study, the SphK1/S1P axis is inferred to regulate synaptic vesicle exocytosis and endocytosis through distinct mechanisms.
The SphK1/S1P axis regulates RRP during exocytosis
Our data showed that synaptic vesicle exocytosis induced by 100 APs at 20-Hz stimulations is reduced by SphK1DN (Fig. 1C–F), indicating a role of the SphK1/S1P axis in synaptic vesicle exocytosis. This conclusion agrees with previous studies in neurons from C. elegans (Chan et al., 2012), mouse neuroendocrine cells (Pan et al., 2006; Z.C. Jiang et al., 2019a) and hippocampal neurons (Kajimoto et al., 2007; Kanno et al., 2010; Riganti et al., 2016), frog neuromuscular junctions (Brailoiu et al., 2002), and zona glomerulosa cells (Brizuela et al., 2007). It is acknowledged that two vesicle pools, TRP and RRP, are essential in determining the amount of exocytosis on stimulations (S.H. Kim and Ryan, 2010; Alabi and Tsien, 2012; Baumgart et al., 2015). Our imaging data reveal that SphK1DN reduces RRP size and slows down RRP replenishment (Fig. 1J–M) with no obvious effects on TRP size or its depletion kinetics (Fig. 1G–I), indicating a physiological role of the SphK1/S1P axis in regulating the size and replenishment of RRP during exocytosis. Consistently, a previous study implies that exogenous S1P addition may increase the size of RRP, measured as sucrose evoked EPSCs (Riganti et al., 2016). The RRP regulation by the SphK1/S1P axis may reflect its function in vesicle docking and priming, since FTY720 has been implied to affect docking/priming process in primary rat cultured astrocytes, a non-neuronal cell (Trkov et al., 2012). Alternatively, the SphK1/S1P may control RRP by modulating localization of synapsin I (Riganti et al., 2016), a major synaptic vesicle associated phosphoprotein regulating RRP size (Baldelli et al., 2007).
The SphK1/S1P axis promotes vesicle endocytosis in neuronal cells
By showing that SphK1DN slowed down sypHy fluorescence decay after the cessation of stimulations at both room and physiological temperatures (Fig. 2), our data provide the first evidence for critical roles of the SphK1/S1P axis in synaptic vesicle endocytosis in neurons. Additionally, single vesicle endocytosis monitored using cell-attached capacitance recordings in chromaffin cells reveals an importance of the SphK1/S1P axis in vesicle fission during endocytosis (Fig. 3). Our findings thus expand the reported roles of the SphK1/S1P axis in endocytic trafficking in non-neuronal cells (Yonamine et al., 2011; Shen et al., 2014; Young et al., 2016; Lima et al., 2017) to neuronal endocytosis. Based on the documented importance of endocytosis for the synaptic vesicle reavailability in RRP (Pyle et al., 2000; Granseth and Lagnado, 2008; Cheung et al., 2010; Rizzoli, 2014; Kyung et al., 2018; Y. Kim et al., 2020), the role of the SphK1/S1P axis in synaptic vesicle endocytosis may account for the SphK1DN-induced disruption in RRP replenishment we have observed (Fig. 1L,M).
The regulation of TRPC5 by the SphK1/S1P axis is critical for synaptic vesicle endocytosis
TRPC5DN reduced Ca2+ increase induced by 100 APs at 20 Hz (Fig. 8E,F) rather than single AP (Fig. 8C,D), suggesting critical roles of TRPC5 in presynaptic Ca2+ signaling during prolonged stimulations. Our results are in line with a previous study using transgenic KO and knock-in mice, demonstrating an importance of TRPC5 channels, independent of VGCCs, in presynaptic Ca2+ influx during stimulation trains (Schwarz et al., 2019). Furthermore, TRPC5DN slowed down sypHy signal decay after stimulations (Fig. 11A,B), indicating that Ca2+ influx via TRPC5 channels may be critical for synaptic vesicle endocytosis in neurons. It is of note that previous studies have identified that Ca2+ influx via TRPM7 (Z.J. Jiang et al., 2021), another TRP superfamily member (Clapham, 2003), and Flower (C.K. Yao et al., 2017), a Ca2+ permeable protein with Ca2+ selectivity filter similar as TRPV5 and TRPV6 (C.K. Yao et al., 2009), may be crucial for synaptic vesicle endocytosis in mammalian neurons. Along this line, it would be interesting to find out the importance of other members of TRP superfamily, which are Ca2+ permeable and widely expressed in the brain (Ramsey et al., 2006; Venkatachalam and Montell, 2007; Sawamura et al., 2017; Koivisto et al., 2022), in synaptic vesicle endocytosis in the mammalian brain. On the other hand, Ca2+ influx via TRPC5 has been previously shown to be important for short-term plasticity of fast glutamatergic synapses (Schwarz et al., 2019). Given the well-established significance of synaptic vesicle endocytosis in short-term synaptic plasticity (Zucker and Regehr, 2002; Granseth and Lagnado, 2008; Hosoi et al., 2009; Regehr, 2012; Y. Hua et al., 2013; Z.J. Jiang et al., 2021), our discovery suggests that the previously reported functions of TRPC5 in short-term plasticity may reflect its role in synaptic vesicle endocytosis in neurons.
The present study using electrophysiological recordings and previous Ca2+ imaging assays in HEK293 cells have testified the role of S1P as a physiological activator for TRPC5 channels (Xu et al., 2006; Naylor et al., 2016). Our Ca2+ imaging studies in neurons shows that SphK1DN induces a reduction in presynaptic synGCaMP6f signal increase in response to 100 APs at 20-Hz stimulations in control (Fig. 5C,D) but not TRPC5DN-expressing neurons (Fig. 8I,J), indicating that the SphK1/S1P axis may be crucial for presynaptic Ca2+ signaling via regulating TRPC5 channels during prolonged stimulations. This finding is likely to provide novel mechanistic insights into the role of S1P, an enigmatic signaling molecule, in Ca2+ homeostasis in the mammalian brain (Spiegel and Milstien, 2003; Karunakaran and van Echten-Deckert, 2017). In parallel, SphK1DN slows down sypHy fluorescence decay after the cessation of stimulations in control (Fig. 2A,B) but not TRPC5DN-expressing neurons (Fig. 11E,F), indicating an importance of TRPC5 in the SphK1/S1P axis-mediated regulation of synaptic vesicle endocytosis. Collectively, our data point out regulations of Ca2+ influx via TRPC5 by the SphK1/S1P axis may be critical for synaptic vesicle endocytosis. This discovery may thus provide novel mechanistic implications for parallel involvements of TRPC5 and S1P in brain functions under physiological and pathologic conditions and pain-related behaviors in mice and human (Camprubí-Robles et al., 2013; Hong et al., 2015, 2020; Weth-Malsch et al., 2016; Bröker-Lai et al., 2017; Di Pardo et al., 2017; Karunakaran and van Echten-Deckert, 2017; Hill et al., 2018; Sadler et al., 2021).
Implications to neurologic disorders
The SphK1/S1P axis and synaptic vesicle endocytosis have been linked to the same group of neurologic disorders, including Parkinson's disease (Pyszko and Strosznajder, 2014), Huntington's disease (Di Pardo et al., 2019), Alzheimer's disease (Ceccom et al., 2014; Couttas et al., 2014), and multiple sclerosis (Subei and Cohen, 2015; Mao-Draayer et al., 2017). By identifying for the first time a role of the SphK1/S1P axis in synaptic vesicle endocytosis, our study indicates that impaired regulations of the SphK1/S1P axis in synaptic vesicle endocytosis may contribute to pathophysiological mechanisms of these neurologic disorders (Couttas et al., 2014; Pirhaji et al., 2016; Schreij et al., 2016; Lee et al., 2018; Di Pardo et al., 2019; McAdam et al., 2020; Pensalfini et al., 2020). Meanwhile, it is worth mentioning that the SphK1/S1P axis has recently been identified as a key regulator for autophagy in neurons (Moruno Manchon et al., 2015, 2016), and dysfunction of autophagy is identified in many neurologic diseases (Menzies et al., 2017; Malik et al., 2019; Klionsky et al., 2021). Indeed, there is an interdependency of autophagy and synaptic vesicle endocytosis within presynaptic terminals (Limanaqi et al., 2018; Overhoff et al., 2021; Yang et al., 2022). As the SphK1/S1P axis may modulate the coordination between autophagy and endocytic membrane trafficking in mouse embryonic fibroblasts (Young et al., 2016), it would be interesting to understand whether the SphK1/S1P axis may regulate the interplay between autophagy and synaptic vesicle endocytosis within presynaptic terminals in the future.
Footnotes
This work was supported by National Institutes of Health Grant R01NS110533 (to L.-W.G.). We thank Dr. L. Lagnado for sypHy and Dr. T. A. Ryan for SynGCaMP6f.
The authors declare no competing financial interests.
- Correspondence should be addressed to Liang-Wei Gong at lwgong{at}uic.edu
