Abstract

ATP is known to be coreleased with glutamate at certain central synapses. However, the nature of its release is controversial. Here, we demonstrate that ATP release from cultured rat hippocampal neurons is sensitive to RNAi-mediated knockdown of the recently identified vesicular nucleotide transporter (VNUT or SLC17A9). In the intact brain, light microscopy showed particularly strong VNUT immunoreactivity in the cerebellar cortex, the olfactory bulb, and the hippocampus. Using immunoelectron microscopy, we found VNUT immunoreactivity colocalized with synaptic vesicles in excitatory and inhibitory terminals in the hippocampal formation. Moreover, VNUT immunolabeling, unlike that of the vesicular glutamate transporter VGLUT1, was enriched in preterminal axons and present in postsynaptic dendritic spines. Immunoisolation of synaptic vesicles indicated presence of VNUT in a subset of VGLUT1-containing vesicles. Thus, we conclude that VNUT mediates transport of ATP into synaptic vesicles of hippocampal neurons, thereby conferring a purinergic phenotype to these cells.

Introduction

Extracellular ATP has a multifaceted role as a signaling molecule in intercellular communication. In the nervous system, glia-derived ATP has been widely implicated as an extracellular messenger and as a precursor of adenosine in glia-to-glia and glia-to-neuron interaction (Koizumi et al. 2005). It is furthermore known that ATP may act as an excitatory neurotransmitter in several brain regions, including the hippocampus (Pankratov et al. 1998, 2006; Mori et al. 2001; Fields and Burnstock 2006). Considerable effort has been devoted to attempts at clarifying the mechanism of release of ATP from glial cells (Stout et al. 2002; Kang et al. 2008; Liu, Sabirov, et al. 2008; Liu, Touchiev, et al. 2008), although controversy regarding this issue remains (Hamilton and Attwell 2010). By contrast, how ATP is released from central neurons has not been well studied. A vesicular nature of neuronal ATP release was first indicated by observations of calcium-dependent release of ATP from whole-brain synaptosomes (White 1978) and several types of brain slice preparation, including hippocampal slices (Wieraszko et al. 1989; Cunha et al. 1996). Furthermore, ATP is taken up by isolated synaptic vesicles (Gualix et al. 1999). Recently, a vesicular nucleotide transporter (VNUT) capable of transporting ATP into vesicles was identified (Sawada et al. 2008). VNUT was shown to be present in the brain (Sawada et al. 2008), but its cellular and subcellular distribution is not known.

ATP exerts its transmitter effects through activation of ionotropic (P2X) and metabotropic (P2Y) receptors. Both P2X and P2Y receptors are widely distributed in neural tissue (Nörenberg and Illes 2000; Fields and Burnstock 2006). P2X receptor-mediated excitatory synaptic transmission has been demonstrated in several regions of the central nervous system (Edwards et al. 1992; Bardoni et al. 1997; Pankratov et al. 1998, 2002; Mori et al. 2001). In the hippocampus (Pankratov et al. 1998, 2006), as well as in cortex (Pankratov et al. 2002, 2003, 2007), results from electrophysiological studies have suggested that ATP is coreleased with glutamate. A recent study showed that ATP is released during stimulation of glutamatergic neuronal pathways (Jourdain et al. 2007). However, it is not clear whether ATP and glutamate reside in the same nerve terminals, let alone in the same synaptic vesicle pools.

Identification of cellular elements that may release ATP has previously been hampered by lack of ATP-specific antibodies. However, such information can now be obtained using antibodies against VNUT. Here, we use anatomical, biochemical, and functional approaches to investigate the vesicular localization and release of ATP in hippocampal neurons.

Materials and Methods

Expression and Purification of Mouse VNUT

A cDNA-encoding mouse VNUT (mVNUT) was cloned (Sawada et al. 2008). Recombinant baculovirus containing mVNUT cDNA were constructed using the Bac-to-Bac baculovirus expression system (Invitrogen) according to the manufacturer's protocol. mVNUT cDNA was amplified by polymerase chain reaction (PCR) using the primers 5′-CACCATGCCATCCCAGCGCTCTAGC-3′ and 5′-TTAGAGTCCTCATGAGTGG-3′. High Five cells (1 × 107 cells/10 cm dish) were grown at 27 °C in Express Five medium (Invitrogen) supplemented with 2 mM L-glutamine and 10 μg/mL gentamicin. Cells were infected by recombinant baculoviruses at a multiplicity of infection of 2, cultured for 48 h for High Five cells and harvested for membrane preparation. High Five cells (1–2 × 108 cells) were suspended in a 20 mM Tris–HCl buffer (pH 8.0) containing 0.1 M potassium acetate, 10% glycerol, 0.5 mM dithiothreitol, 10 μg/mL pepstatin A, and 10 μg/mL leupeptin and disrupted by sonication with a TOMY UD200 tip sonifier. Cell lysates were centrifuged at 700 × g for 10 min to remove debris, and the resultant supernatant was centrifuged at 160 000 × g for 1 h. The pellet (membrane fraction) was suspended in buffer containing 20 mM 3-(N-morpholino)propanesulfonic acid (MOPS)–Tris (pH 7.0), 10% glycerol, 10 μg/mL pepstatin A, and 10 μg/mL leupeptin to a concentration of approximately 1.5 mg protein/mL. The membrane fraction was solubilized with 2% octylglucoside. After centrifugation at 260 000 × g for 30 min, the supernatant was added to 1 mL of Ni-NTA Superflow resin (Qiagen). The resin was incubated for 4 h at 4 °C and washed with 10 mL of 20 mM MOPS–Tris (pH 7.0) containing 5 mM imidazole, 20% glycerol, and 1% octylglucoside in a column. mVNUT was eluted from the resin with 3 mL of the same buffer containing 60 mM imidazole. The eluate containing purified mVNUT was stored at −80 °C until use.

Reconstitution and ATP Transport

Reconstitution of purified mVNUT was carried out by the freeze–thaw method as described previously (Sawada et al. 2008). In brief, 10 μg mVNUT was mixed with liposomes (0.5 mg asolectin, Sigma type II-S), frozen at −80 °C. After 5 min, the mixture was thawed quickly and diluted 60-fold with 20 mM MOPS–Tris (pH 7.0) with 0.5 mM dithiothreitol, 0.15 M sodium acetate, and 5 mM magnesium acetate. Reconstituted proteoliposomes were sedimented by centrifugation at 160 000 × g for 1 h at 4 °C and suspended in 0.4 mL of the same buffer. For ATP transport assay, reconstituted proteoliposomes (0.5 μg protein per assay) were suspended in 500 μL of 20 mM MOPS–Tris (pH 7.0) containing 5 mM magnesium acetate, 4 mM KCl, and 0.15 M potassium acetate and incubated for 2 min at 27 °C. Valinomycin was added to give a final concentration of 2 μM, and the mixture was incubated for an additional 2 min. The assay was initiated by addition of 0.1 mM [α-32P]ATP (3.7 GBq/mmol). Aliquots (130 μL) were taken at the times indicated and centrifuged through a Sephadex G-50 (fine) spin column at 760 × g for 2 min. Radioactivity and protein concentration of the eluate were measured. To determine GTP or ADP transport, [α-32P]GTP (3.7 GBq/mmol) or [2,8-3H]ADP (0.37 GBq/mmol) was used as substrates instead of ATP.

RNAi-Mediated VNUT Knockdown

Neuronal cultures were prepared from the hippocampus of male Wistar rats (E17). Briefly, after dissection, the hippocampi were dissociated by treatment with trypsin (0.25% for 15 min at 37 °C) in the presence of DNase (0.01%). Cells (5 × 105 cells/35 mm dish) were plated on poly-L-lysine and laminin-coated cell culture dish and grown in Neurobasal medium (Invitrogen) containing B27 supplement (Invitrogen). The culture medium was exchanged every 2–3 days. Experiments were performed between 18 and 21 days in culture. For gene knockdown, HiPerFect transfection reagent (Qiagen) was used for transfection of 10 nM AllStars negative control small interfering RNA (siRNA) or rat SLC17A9 siRNA: UAUUCGAGAGAAUGUCACG. ATP secretion was assayed after 48 h incubation.

Stimulation of Neuronal Cultures and Measurement of ATP Release

ATP secretion was assayed as described (Sawada et al. 2008). Briefly, the cells (5 × 105 cells/35 mm dish) were incubated in 1.5 mL of medium consisting of 128 mM NaCl, 1.9 mM KCl, 1.2 mM KH2PO4, 1.3 mM MgSO4, 26 mM NaHCO3, 2.4 mM CaCl2, 10 mM glucose, 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)/Tris (pH 7.4), and 0.2% bovine serum albumin (BSA) for 30 min at 37 °C. Then, the cultured neurons were incubated under control conditions (1.9 mM KCl) or stimulated in depolarizing conditions (50 mM KCl) for 20 min (37 °C). The effect of preincubation with control and VNUT siRNA (48 h, see above) and tetanus toxin (15 μg/mL, 36 h), as well as the effect of low Ca2+ (0 mM Ca2+ or addition of EGTA-AM [50 μM]) and the Ca2+ ionophore A23186 (5 μM) on extracellular ATP concentration was measured under control and depolarizing conditions. Assay medium (1 mL) was collected under various experimental conditions, kept on ice, and 50 μL aliquots used for measuring ATP concentration with a bioluminescence method using a kit (Invitrogen). A 50 μL aliquot of the luciferin–luciferase reaction medium was added to each cuvette. The resulting light signal was measured with BLR-101C Luminescence Reader (Aloka).

Real-Time PCR

RNA was purified from cultured hippocampal neurons (8 × 106 cells) using RNeasy kit (Qiagen) according to the manufacturer's instructions. Real-time quantitative PCR was performed using SYBR Premix Ex Taq II2 (TARARA BIO, Japan) containing the double-stranded DNA-binding fluorescent probe Sybr Green and all necessary components except primers. Quantitative PCR conditions included an initial denaturation step of 94 °C for 10 min followed by 40 cycles of 94 °C for 15 s, and 55 °C for 15 s. Standards and samples were analyzed in triplicate. The following primers were used: VNUT, TGTGGTAGGCGTGTGTCTAG (forward), AGGTTGCTGACGATGGCCAC (reverse).

Antibodies

Site-specific polyclonal antibodies against mVNUT were prepared by repeatedly injecting GST-fusion polypeptides encoding LMQPIPEETRKTPSAAAEDTRWSRPECQAWTGILLLGTCLLYCARVTMPVCTVAMSQDFGWNKKEAGIVLS SFFWGYCLTQVVGGHLGDR (L8–R97) (Sawada et al. 2008). Antibodies against V-ATPase subunit A and synaptotagmin 1 (kindly supplied by Prof. Masami Takahashi, Kitasato University, Japan) have been described (Shoji-Kasai et al. 1992; Moriyama et al. 1995). A mouse anti-synaptophysin antibody (clone SY-38; Wiedenmann and Franke 1985) was obtained from PROGEN (Heidelberg, Germany). A mouse monoclonal antibody against MAP2A-C (clone HM-2) was purchased from Sigma. Two guinea pig anti-VGLUT1 antibodies from Millipore (AB5905) and Synaptic Systems (cat. no. 135 304; Göttingen, Germany), one rabbit anti-VGLUT1 (Synaptic Systems; cat. no. 135 302), one mouse anti-VGLUT1 (UC Davis/NIH NeuroMab Facility; clone N28/9), and a VGAT antibody raised in guinea pig (Synaptic Systems; cat. no. 131 004) were used. AB5905 has been used extensively for immunohistochemical detection of VGLUT1 (e.g., Persson et al. 2006). The VGLUT1 and VGAT antibodies from Synaptic Systems detect bands of appropriate sizes by immunoblotting of rat brain tissue (manufacturer's specification; L. Ormel, V. Gundersen, unpublished data) and have previously been successfully used for postembedding immunogold labeling (Zander et al. 2010). The mouse anti-VGLUT1 from NeuroMab detects a single band at ∼50 kDa (manufacturer's specification) and has been extensively used for immunohistochemistry (e.g., Linhoff et al. 2009). Alexa Fluor 488-labeled goat anti-rabbit IgG and Alexa Fluor 568- or Alexa Fluor 555-labeled goat anti-mouse IgG were obtained from Invitrogen (Molecular Probes). Biotinylated donkey anti-rabbit Ig was from GE Healthcare, whereas 10 nm colloidal gold-conjugated goat anti-rabbit IgG, 5 nm colloidal gold-conjugated goat F(ab)2 anti-rabbit IgG, and 15 nm colloidal gold-conjugated goat anti-guinea pig IgG were obtained from British Biocell International (Cardiff, UK).

Western Blotting on Cultured Neurons

The cells (2 × 106 cells) were suspended in the 20 mM MOPS–Tris (pH 7.0) containing 5 mM ethylenediaminetetraacetic acid (EDTA), 0.25 M sucrose, 10 μg/mL pepstatin A, and 10 μg/mL leupeptin and disrupted by sonication with TOMY UD200 tip sonifier. Cell lysates were centrifuged at 700 × g for 10 min to remove debris, and the resultant supernatant was centrifuged at 160 000 × g for 1 h. The pellet was suspended in the same buffer, and proteins were dissolved with sample buffer containing 10% sodium dodecyl sulfate (SDS) and 10% β-mercaptoethanol. Polyacrylamide gel electrophoresis (PAGE) in the presence of SDS and western blotting were performed as described (Sawada et al. 2008).

Immunofluorescence Miscroscopy on Cultured Neurons

Indirect immunofluorescence microscopy was performed as previously described (Sawada et al. 2008). Cultured cells on poly-L-lysine-coated coverslips were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 30 min. After washing with PBS, cells were incubated for 15 min in PBS containing 0.2% saponin, 2% normal goat serum, and 0.5% BSA. The specimens were incubated with the primary antibodies (anti-VNUT diluted 1:200 and anti-synaptophysin diluted 1:50 or anti-synaptotagmin diluted 1:500) in PBS containing 0.5% BSA for 1 h at room temperature. Samples were washed 4 times with PBS and then reacted with Alexa Fluor 568-labeled anti-mouse IgG (2 μg/mL) or Alexa Fluor 488-labeled anti-rabbit IgG (4 μg/mL) for 1 h at room temperature. Finally, immunoreactivity was examined in an Olympus FV300 confocal laser microscope.

Tissue Preparation for Light and Electron Microscopy

Adult Wistar rats were anesthetized using pentobarbital (60 mg/kg, intraperitoneally) and transcardially perfused with PBS containing 4% paraformaldehyde (for immunoperoxidase and immunofluorescence labeling; 3 rats) or 4% paraformaldehyde and 0.1 or 0.5% glutaraldehyde (for electron microscopy; 3 rats). Animal use was in accordance with the guidelines of the Norwegian Committees on Animal Experimentation (Norwegian Animal Welfare Act and European Communities Council Directive of 24 November 1986-86/609/EEC). Rat hippocampal tissue was embedded for postembedding immunogold labeling essentially as described (Bergersen et al. 2008). Briefly, tissue encompassing the hippocampal CA1 region and the dentate gyrus was dissected from the brain and cryoprotected in graded solutions of glycerol. The tissue was subsequently freeze-substituted and embedded at low temperature in Lowicryl HM20. Ultrathin sections were cut and collected on Ni mesh grids (one section per grid).

Immunoperoxidase and Immunofluorescence Labeling of Brain Tissue

Coronal or parasagittal 40 μm sections of the brain were cut on a freezing microtome, and immunohistochemistry was performed on free-floating sections according to standard immunofluorescence and immunoperoxidase protocols. The sections were incubated overnight, for 36 or 60 h at room temperature in anti-VNUT antibody diluted 1:1000 or 1:1500 in PBS with 3% normal goat serum, 0.5% BSA, and 0.5% Triton X-100. In immunofluorescence experiments, the primary antibody solution was supplemented with mouse anti-VGLUT1 (NeuroMab) diluted 1:500 or 1:1500 or mouse anti-MAP2 diluted 1:500. For immunofluorescent detection, sections were incubated in goat anti-rabbit Alexa Fluor 488 and goat anti-mouse Alexa Fluor 555 (Invitrogen). For immunoperoxidase detection, the sections were incubated sequentially in biotinylated donkey anti-rabbit and streptavidin-horseradish peroxidase. Vector SG (Vector Laboratories, Burlingame, CA) or 3,3′-diaminobenzidine (Sigma) was used as chromogen. Confocal microscopy of immunofluorescence labeling was performed using a 63× oil immersion objective with a Zeiss LSM 5 Pa confocal unit mounted on a Zeiss Axoplan 2 microscope. Micrographs of immunoperoxidase-labeled sections were obtained with a 20× objective in a Zeiss Axioplan 2 microscope or a 100× oil immersion objective in a Nikon Optiphot2 microscope. A Zeiss SteREO Lumar. V12 stereomicroscope was used to acquire low-power whole-brain micrographs. For presentation, brightness and contrast were adjusted using Adobe Photoshop CS2.

Postembedding Immunogold Labeling

A modified standard postembedding immunogold protocol was used (Larsson and Broman 2005; Bergersen et al. 2008). Ultrathin sections were incubated in drops in a humid chamber and rinsed by dipping 3 × 5 s in beakers containing the rinsing solution. The sections were incubated in 50 mM glycine in Tris-buffered saline (0.9% NaCl; pH 7.4) with 0.1% Triton X-100 (TBST) for 10 min. After rinsing in TBST and incubation in blocking buffer (10 min), the sections were incubated in primary antibody in blocking buffer overnight at room temperature. In VNUT single-immunolabeling and VNUT/VGLUT1 double-immunolabeling experiments, TBST containing 2% human serum albumin was used as blocking buffer; for VNUT/VGAT double immunolabeling, the blocking buffer was supplemented with 5% normal goat serum. The sections were subjected twice to a rinsing step consisting of dipping in TBST and a 10-min incubation in TBST, which was followed by incubation in blocking buffer (10 min) and secondary antibody solution (2 h). After rinsing in H2O, the sections were counterstained with uranyl acetate and lead citrate (the latter omitted when using 5 nm gold) prior to examination in a FEI Tecnai 12 or G2 Spirit electron microscope. The rabbit anti-VNUT antibody was used at a concentration of 1:300 for all experiments except the VNUT/VGAT double-immunolabeling experiment, where the concentration was 1:200. The 2 guinea pig anti-VGLUT1 antibodies (Synaptic System and Millipore) were used in combination, both at a concentration of 1: 2000, while guinea pig anti-VGAT was used at a concentration of 1:200. Gold-conjugated secondary antibodies were diluted in blocking buffer. In single VNUT immunogold labeling experiments, goat anti-rabbit conjugated to 10 nm gold (at a concentration of 1:20) was used, whereas in double-labeling experiments, the secondary antibodies were goat F(ab)2 anti-rabbit conjugated to 5 nm gold (1:40) and goat anti-guinea pig conjugated to 15 nm gold (1:20). Preabsorption of the VNUT antibody using the immunizing peptide at a concentration of 40 μg/mL substantially reduced the immunogold labeling, similar to what has been previously observed by western blot and immunofluorescence microscopy (Sawada et al. 2008; Iwatsuki et al. 2009; Haanes and Novak 2010).

Quantitative Analysis of Immunogold Labeling

Neuropil from the outer two-thirds of molecular layer of the dentate gyrus and from CA1 stratum radiatum as well as inhibitory terminals in the dentate granule and CA1 pyramidal cell layers were subject to quantitative analysis of single VNUT immunogold labeling in sections from 3 different rats. Sections through the dentate granule cell layer from 2 rats were used for quantitative analysis of double VNUT/VGLUT1 and VNUT/VGAT immunogold labeling. For each animal, analysis of all regions was performed in the same section. Random electron micrographs containing profiles of interest were obtained at a magnification of 43 000× using a Veleta digital camera. A plugin for ImageJ (http://rsb.info.nih.gov/ij/) was used to outline the plasma membrane of the profile and record the locations of the center of each gold particle in the profile; in dendritic spines, the postsynaptic density (PSD) was also outlined. Random points were placed over the micrograph. Recorded coordinates were submitted to a program written in Python (http://www.python.org) for computation of particle densities and particle–membrane and particle–particle distances (Larsson and Broman 2005). The source code of the ImageJ plugin and the Python program is available at http://www.neuro.ki.se/broman/maxl/software.html. Mitochondria and PSDs were excluded from the analysis. Gold particle density over basal lamina of the capillary endothelium was used as a measure of nonspecific background immunolabeling. Electron micrographs used for figures were adjusted for brightness and contrast in Adobe Photoshop CS2.

Immunoisolation of Synaptic Vesicles

Male Wistar rats (4–5 weeks old) were anesthetized with halothane and decapitated. This procedure was in accordance with the guidelines of the Norwegian Committees on Animal Experimentation as above. The forebrains were quickly removed, put on ice, and homogenized (10 strokes at 450 rpm) in 0.32 M sucrose. After centrifugation of the homogenate at 1000 × g for 10 min, the supernatant (S1) was collected and centrifuged at 21 000 × g for 30 min. The resulting pellet (P2) was resuspended in ice-cold H2O, homogenized (7 strokes at 450 rpm), and centrifuged at 21 000 × g for 30 min. The supernatant (S3) from this centrifugation was supplemented at a proportion of 9:1 with 1 M dipotassium tartrate in 0.25 M HEPES (pH 7.4) and centrifuged at 120 000 × g for 60 min. The resulting pellet (P4) resuspended in 0.32 M sucrose was used as starting material for immunoisolation of synaptic vesicles.

For immunoisolation, protein A-coated Dynabeads (Invitrogen) were incubated with primary antibodies diluted 1:4 (guinea pig anti-VGLUT1 from Synaptic Systems) or 1:9 (anti-VNUT) in 0.1 M phosphate buffer (pH 8; PB) at room temperature for 2 h. After washing in PB, the beads were incubated in 1% BSA in PB at room temperature for 1 h. Vesicle suspension prepared as above was added to the beads, and this mixture was incubated overnight at 4 °C. The beads were put on a magnet and the supernatant (material not bound to the beads) collected and diluted 1:1 in 1% SDS and 5 mM EDTA in 10 mM phosphate buffer (pH 7.4). After washing, material bound to the beads was eluted by incubation of the beads in 1% SDS and 5 mM EDTA in 10 mM phosphate buffer (pH 7.4) at room temperature for 15 min, followed by brief centrifugation and collecting of the eluate on magnet. The supernatants and eluates were subjected to SDS–PAGE (3 μg protein per lane) and western blotting according to standard protocol. The blots were incubated overnight at 4 °C in primary antibody (rabbit anti-VGLUT1 (from Synaptic Systems) or anti-VNUT both diluted 1:1000) in 1% BSA in Tris-buffered saline (pH 7.4) with 0.05% Tween 20.

Results

Functional Characterization of mVNUT

Mammalian VNUT orthologues show some interspecies sequence variation in the N-terminal domain but are rather highly conserved in other regions (Sawada et al. 2008). We first assessed whether rodent VNUT exhibits similar functional properties to the human orthologue that was previously characterized (Sawada et al. 2008). Heterologously expressed mVNUT was purified and reconstituted into liposomes (Fig. 1A). As expected, liposomes possessing mVNUT accumulated [α-32P]ATP in a manner dependent on Δψ (inside positive; established using the K+ ionophore valinomycin) (Fig. 1B; Supplementary Table S1) but not on ΔpH (Supplementary Table S1). mVNUT-mediated ATP uptake into liposomes exhibited a Km of 230 μM and a Vmax of 23.8 nmol/min/mg of protein (Fig. 1C). The values of these parameters were somewhat lower than for human VNUT (Sawada et al. 2008). The intracellular ATP concentration in dorsal root ganglion nerve cells and the squid giant axon has been estimated to be ∼1 to 5 mM (Mullins and Brinley 1967; Fukuda et al. 1983). Thus, although ATP concentrations in different compartments of central neurons and activity-dependent fluctuations of cytosolic ATP levels are unknown, it is likely that mVNUT-mediated vesicular transport is saturated during basal conditions. Similar to human VNUT, [α-32P]ATP uptake by mVNUT was inhibited by ATP, ADP, GTP, and UTP, as well as by diadenosine triphosphate and the ATP analogues AMP–PNP and ATP-γ-s. ATP transport was less sensitive to the presence of adenosine or adenine (Fig. 1D). Like human VNUT, mVNUT was dependent on Cl ions but not on divalent cations (Supplementary Fig. S1). Both 4,4′-diisothiocyanatostilbene-2,2′-disulphonic acid and Evans Blue suppressed ATP uptake by mVNUT. Finally, similar to human VNUT (Sawada et al. 2008), mVNUT was found to transport ADP and GTP (Supplementary Fig. S1).

Figure 1.

Functional characterization of mVNUT. (A) Purification of mVNUT. Purified mVNUT (10 μg protein) was run on SDS-gel electrophoresis in the presence of SDS and stained with Coomassie brilliant blue. The positions of molecular standards are indicated. (B) Time course of [α-32P]ATP uptake. Na+-trapped proteoliposomes were suspended in the presence or absence of valinomycin. Upon the addition of [α-32P]ATP, samples were taken at the indicated times, and radioactivity taken up by the proteoliposomes was counted. (C) Dose dependence of [α-32P]ATP uptake. ATP uptake at 1 min was determined under the indicated ATP concentrations. (D) [α-32P]ATP uptake was determined in the presence of various nucleotides at 1 mM. Each experiment was repeated twice. Error bars indicate standard error of the mean. **P < 0.01, ***P < 0.001.

Figure 1.

Functional characterization of mVNUT. (A) Purification of mVNUT. Purified mVNUT (10 μg protein) was run on SDS-gel electrophoresis in the presence of SDS and stained with Coomassie brilliant blue. The positions of molecular standards are indicated. (B) Time course of [α-32P]ATP uptake. Na+-trapped proteoliposomes were suspended in the presence or absence of valinomycin. Upon the addition of [α-32P]ATP, samples were taken at the indicated times, and radioactivity taken up by the proteoliposomes was counted. (C) Dose dependence of [α-32P]ATP uptake. ATP uptake at 1 min was determined under the indicated ATP concentrations. (D) [α-32P]ATP uptake was determined in the presence of various nucleotides at 1 mM. Each experiment was repeated twice. Error bars indicate standard error of the mean. **P < 0.01, ***P < 0.001.

VNUT-Mediated ATP Release from Hippocampal Neurons

Expression of VNUT in cultured rat hippocampal neurons was detected by western blot analysis and indirect immunofluorescence (Fig. 2A,B). RNAi-mediated knockdown of VNUT in hippocampal neuronal cultures robustly decreased VNUT mRNA and protein levels as determined by real-time PCR analysis, western blot and immunocytochemistry (Fig. 2A–C). VNUT immunofluorescence showed a somatoneuritic staining pattern, only partly overlapping with the presynaptic terminal markers synaptophysin and synaptotagmin 1 (Fig. 2B), suggesting that VNUT was located in several neuronal compartments (see below). VNUT knockdown attenuated K+-evoked but not basal ATP release (Fig. 2D). K+-evoked ATP release was abolished in the absence of extracellular Ca2+ or in the presence of the cell-permeable Ca2+ chelator EGTA-AM. Tetanus toxin prevented K+-evoked ATP release (Fig. 2D,E). Furthermore, the Ca2+ ionophore A23187 induced ATP release from hippocampal neuronal cultures (Fig. 2D). These observations suggest that evoked release of ATP from hippocampal neurons occurs by exocytosis of vesicles loaded with ATP by a VNUT-mediated mechanism.

Figure 2.

VNUT- and Ca2+-dependent ATP release from hippocampal neurons. (A) Western blot detection of VNUT in homogenate of cultured hippocampal neurons. RNAi-mediated knockdown strongly attenuated the VNUT immunoreactive band, whereas synaptophysin, synaptotagmin, and v-ATPase immunoreactivities were unchanged. (B) The VNUT immunofluorescence (green) observed in soma and neurites of cultured hippocampal neurons was attenuated by RNAi-mediated knockdown. Synaptophysin (red) and synaptotagmin 1 (red) were used as presynaptic terminal markers. Insets, higher magnification showing colocalization (yellow) of VNUT (green) and synaptophysin (red) or synaptotagmin 1 (red). Arrowheads indicate colocalized immunoreactivity for VNUT and synaptophysin/synaptotagmin 1. Arrow indicates a VNUT+ punctum that is not colocalized with synaptotagmin 1 immunolabeling. Differential interference contrast microscopy (DIC) shows the morphology of cultured neurons. Scale bar: 10 μm. (C) Effect of RNAi-mediated knockdown of VNUT on evoked ATP release from cultured hippocampal neurons. Left panel: Quantitative analysis for VNUT mRNA levels was performed by real-time PCR. VNUT expression was reduced to ∼20% of original expression by RNAi in cultured neurons. Right panel: Neuronal ATP release after 20-min incubation upon K+ stimulation from control siRNA-treated or VNUT siRNA-treated cultured hippocampal neurons. K+-evoked ATP release was strongly impaired in RNAi-treated cultures. (D) ATP release from cultured hippocampal neurons was dependent on exocytotic mechanisms. K+-evoked ATP release was attenuated in the absence of extracellular Ca2+, in the presence of the cell-permeable Ca2+ buffer EGTA-AM (50 μM). The calcium ionophore A23186 (5 μM) also elicited ATP release. (E) Neurons were treated in the presence or absence of 15 μg/mL tetanus toxin (TeTN) (Invitrogen) for 36 h, after which KCl-mediated ATP release was assayed. Each experiments were repeated twice. Error bars indicate standard error of the mean. **P < 0.01, ***P < 0.001, 2-way analysis of variance followed by Bonferroni post hoc test.

Figure 2.

VNUT- and Ca2+-dependent ATP release from hippocampal neurons. (A) Western blot detection of VNUT in homogenate of cultured hippocampal neurons. RNAi-mediated knockdown strongly attenuated the VNUT immunoreactive band, whereas synaptophysin, synaptotagmin, and v-ATPase immunoreactivities were unchanged. (B) The VNUT immunofluorescence (green) observed in soma and neurites of cultured hippocampal neurons was attenuated by RNAi-mediated knockdown. Synaptophysin (red) and synaptotagmin 1 (red) were used as presynaptic terminal markers. Insets, higher magnification showing colocalization (yellow) of VNUT (green) and synaptophysin (red) or synaptotagmin 1 (red). Arrowheads indicate colocalized immunoreactivity for VNUT and synaptophysin/synaptotagmin 1. Arrow indicates a VNUT+ punctum that is not colocalized with synaptotagmin 1 immunolabeling. Differential interference contrast microscopy (DIC) shows the morphology of cultured neurons. Scale bar: 10 μm. (C) Effect of RNAi-mediated knockdown of VNUT on evoked ATP release from cultured hippocampal neurons. Left panel: Quantitative analysis for VNUT mRNA levels was performed by real-time PCR. VNUT expression was reduced to ∼20% of original expression by RNAi in cultured neurons. Right panel: Neuronal ATP release after 20-min incubation upon K+ stimulation from control siRNA-treated or VNUT siRNA-treated cultured hippocampal neurons. K+-evoked ATP release was strongly impaired in RNAi-treated cultures. (D) ATP release from cultured hippocampal neurons was dependent on exocytotic mechanisms. K+-evoked ATP release was attenuated in the absence of extracellular Ca2+, in the presence of the cell-permeable Ca2+ buffer EGTA-AM (50 μM). The calcium ionophore A23186 (5 μM) also elicited ATP release. (E) Neurons were treated in the presence or absence of 15 μg/mL tetanus toxin (TeTN) (Invitrogen) for 36 h, after which KCl-mediated ATP release was assayed. Each experiments were repeated twice. Error bars indicate standard error of the mean. **P < 0.01, ***P < 0.001, 2-way analysis of variance followed by Bonferroni post hoc test.

Localization of VNUT in the Brain

Light Microscopy

We proceeded to assess the localization of VNUT in the hippocampus in vivo. In the light microscope, immunoperoxidase histochemistry showed that VNUT was widely distributed throughout the brain (Fig. 3). In particular, the cerebellum and the olfactory bulb were heavily immunolabeled. Considerable immunolabeling was also observed in the hippocampus and the superficial superior colliculus (Fig. 3). In the cerebellum, strong immunoreactivity was evident in the somatodendritic compartment of Purkinje cells, but labeling was also prominent throughout the molecular layer (Fig. 3B). VNUT immunolabeling was present in all hippocampal layers and the dentate gyrus. Somewhat stronger somatic or perisomatic immunolabeling was found in the principal cell layers (Fig. 3C,D). Furthermore, some interneuronal somata throughout the hippocampal formation were strongly immunoreactive. Confocal microscopy of sections immunofluorescence labeled for VNUT showed punctate immunoreactivity intermingled with a more diffuse staining in the hippocampus (Fig. 3E). In hippocampal sections, double labeled for VNUT and VGLUT1, VNUT+ and VGLUT1+ puncta were frequently colocalized. However, VNUT+ and VGLUT1+ puncta that did not coincide with punctate staining for the other transporter were also common (Fig. 3E). Furthermore, punctate VNUT immunofluorescent labeling was often localized to MAP2-immunolabeled dendrites in hippocampal sections double labeled for these proteins (Fig. 3F).

Figure 3.

Immunohistochemical labeling of VNUT in rat brain. (A–D) Immunoperoxidase labeling. (A) Parasagittal section showing strong immunoperoxidase labeling in the olfactory bulb, cerebellar cortex, and hippocampus. (B) Cerebellar cortex. ml, molecular layer; pc, Purkinje cell layer; gc, granule cell layer. (C) Dentate gyrus. ml, molecular layer; gc, granule cell layer; hi, hilus. (D) CA1. so, stratum oriens; py, pyramidal cell layer; sr, stratum radiatum. Inset shows a portion of proximal stratum radiatum (located as indicated by the dashed frame in the large micrograph) imaged using a 100× objective, showing punctate VNUT staining. Micrographs shown in B–D were obtained at 20× magnification. (E) Confocal images of VNUT (green) and VGLUT1 (magenta) immunofluorescent labeling in CA1 stratum radiatum. Some VNUT+ and VGLUT1+ puncta colocalize (white; examples indicated by arrowheads), but VNUT+/VGLUT1 (simple arrow) and VNUT/VGLUT+ (double arrow) puncta are also observed. (E) Confocal micrographs of VNUT (green) and MAP2 (magenta) immunofluorescence in CA1 stratum radiatum. Two VNUT+ puncta that localize to a MAP2-immunolabeled dendrite are indicated by arrowheads. Scale bar in D: 50 μm, valid for BD (except inset in D). Scale bar in E: 2 μm, valid also for F.

Figure 3.

Immunohistochemical labeling of VNUT in rat brain. (A–D) Immunoperoxidase labeling. (A) Parasagittal section showing strong immunoperoxidase labeling in the olfactory bulb, cerebellar cortex, and hippocampus. (B) Cerebellar cortex. ml, molecular layer; pc, Purkinje cell layer; gc, granule cell layer. (C) Dentate gyrus. ml, molecular layer; gc, granule cell layer; hi, hilus. (D) CA1. so, stratum oriens; py, pyramidal cell layer; sr, stratum radiatum. Inset shows a portion of proximal stratum radiatum (located as indicated by the dashed frame in the large micrograph) imaged using a 100× objective, showing punctate VNUT staining. Micrographs shown in B–D were obtained at 20× magnification. (E) Confocal images of VNUT (green) and VGLUT1 (magenta) immunofluorescent labeling in CA1 stratum radiatum. Some VNUT+ and VGLUT1+ puncta colocalize (white; examples indicated by arrowheads), but VNUT+/VGLUT1 (simple arrow) and VNUT/VGLUT+ (double arrow) puncta are also observed. (E) Confocal micrographs of VNUT (green) and MAP2 (magenta) immunofluorescence in CA1 stratum radiatum. Two VNUT+ puncta that localize to a MAP2-immunolabeled dendrite are indicated by arrowheads. Scale bar in D: 50 μm, valid for BD (except inset in D). Scale bar in E: 2 μm, valid also for F.

Electron Microscopy—VNUT Single Immunogold Labeling

To delineate the ultrastructural distribution of VNUT immunoreactivity in neuropil, we used postembedding immunogold labeling and electron microscopy. As we had demonstrated VNUT-dependent release of ATP from cultured hippocampal neurons, we focused here on the hippocampal region. Qualitative examination revealed immunolabeling of a variety of neuronal compartments. Gold particles–signaling VNUT were located over synaptic vesicles in nerve terminals forming asymmetric synaptic specializations with postsynaptic spines (presumed excitatory synapses; Fig. 4A–C) as well as in terminals forming symmetric axosomatic synapses (presumed inhibitory synapses) (Fig. 4D,E). Quantitative analysis of the immunolabeling in presumed excitatory terminals was performed in the outer two-thirds of the molecular layer of the dentate gyrus (MLo) and in the stratum radiatum (SR) of CA1. Labeling in inhibitory terminals was analyzed in the dentate granule and CA1 pyramidal cell layers (GCL and PCL, respectively). The pattern of immunolabeling was very similar between the examined regions. In the MLo, 44 ± 10% (n = 3 animals; mean ± standard deviation) of terminals forming asymmetric synapses were immunopositive for VNUT (Table 1). In the SR, 43 ± 5% (n = 3 animals) of such terminals contained VNUT immunoreactivity. VNUT immunolabeling was detected in 55 ± 27% (n = 2 animals) of presumed inhibitory terminals in the GCL and in 74 ± 6% (n = 2) of presumed inhibitory terminals in the CA1 PCL (Table 1). However, this apparent difference between GCL and PCL inhibitory terminals was partly attributed to differences in overall immunolabeling in the analyzed tissue sections. Many unmyelinated preterminal axons exhibited immunoreactivity for VNUT (Fig. 4F–H). In single ultrathin sections, 15 ± 4% and 13 ± 5% of transversely cut unmyelinated axons in MLo and SR, respectively, were immunopositive (Table 1). Myelinated axons were also often strongly VNUT immunoreactive (Fig. 4F,G), but the immunolabeling in such axons was not subject to quantitative analysis. VNUT immunogold labeling was present not only in presynaptic profiles but also in a proportion (33 ± 8% in MLo and 30 ± 6% in SR) of postsynaptic dendritic spines (Fig. 4I–K and Table 1). In both the MLo and the SR, the density of gold particles–signaling VNUT was higher in unmyelinated axons than in excitatory nerve terminals and dendritic spines (Table 1 and Fig. 4L). In dendritic spines, VNUT gold particles were observed in the vicinity of small vesicle-like structures (insets in Fig. 4I,J). However, in order to increase immunogold sensitivity, the tissue used here was fixed with low-concentration glutaraldehyde, treated with uranyl acetate rather than osmium tetroxide and embedded at low temperature in a methacrylate resin. Because this leads to suboptimal morphological preservation as compared to conventional embedding protocols, vesicular profiles in dendritic spines were often not readily discernible (Fig. 4K).

Table 1

Summary of VNUT immunogold labeling

 Animals Total n profiles Total profile area (μm2% Labeled Labeling density (μm−2
Dentate gyrus 
    Excitatory terminals 242 38.3 44 ± 10 6.1 ± 1.3 
    Inhibitory terminals 90 20.1 55 ± 27 6.7 ± 4.8 
    Dendritic spines 262 26.1 33 ± 8 9.1 ± 1.3 
    Unmyelinated axons 391 6.0 15 ± 4 24.2 ± 9.4 
CA1 
    Excitatory terminals 178 26.3 43 ± 5 8.1 ± 1.4 
    Inhibitory terminals 67 23.1 74 ± 6 9.1 ± 0.2 
    Dendritic spines 198 16.8 30 ± 6 9.5 ± 0.6 
    Unmyelinated axons 342 4.0 13 ± 5 19.2 ± 4.2 
 Animals Total n profiles Total profile area (μm2% Labeled Labeling density (μm−2
Dentate gyrus 
    Excitatory terminals 242 38.3 44 ± 10 6.1 ± 1.3 
    Inhibitory terminals 90 20.1 55 ± 27 6.7 ± 4.8 
    Dendritic spines 262 26.1 33 ± 8 9.1 ± 1.3 
    Unmyelinated axons 391 6.0 15 ± 4 24.2 ± 9.4 
CA1 
    Excitatory terminals 178 26.3 43 ± 5 8.1 ± 1.4 
    Inhibitory terminals 67 23.1 74 ± 6 9.1 ± 0.2 
    Dendritic spines 198 16.8 30 ± 6 9.5 ± 0.6 
    Unmyelinated axons 342 4.0 13 ± 5 19.2 ± 4.2 

Note: Quantitative data were obtained from the outer two-thirds of the dentate molecular layer or CA1 stratum radiatum, except for inhibitory terminals, which were sampled from the dentate granule cell layer or the CA1 pyramidal cell layer. Labeling percentages and summed densities were calculated for each animal and are given as mean ± standard deviation. Mitochondria and the PSD of spines were excluded from the analysis.

Figure 4.

Postembedding immunogold labeling of VNUT in the dentate gyrus and CA1. (A–C) Examples of VNUT-immunopositive terminals (t) establishing asymmetric axospinous synapses in the dentate molecular layer (A,B) or CA1 stratum radiatum (C). Note that the labeling is concentrated over synaptic vesicles. (D,E) Examples of VNUT-immunolabeled terminals forming symmetric axosomatic synapses in the granule cell layer. (F–H) Unmyelinated (ax) and myelinated (my) axons exhibiting strong VNUT immunoreactivity in CA1 stratum radiatum (F) and the dentate molecular layer (G,H). In F, t indicates an immunolabeled terminal. (I–K) Examples of VNUT-immunopositive dendritic spines (s) postsynaptic to presumed excitatory terminals (t) in the dentate molecular layer (I,J) and CA1 stratum radiatum (K). (L) VNUT labeling density in presumed excitatory terminals, dendritic spines, and unmyelinated axons of the molecular layer of the dentate gyrus (DG) and CA1 stratum radiatum (see also Tables 1). The summed gold particle density (number of gold particles/μm2) was determined for each sample and normalized against background labeling density over endothelial basal lamina. Error bars indicate standard error of the mean. *P < 0.05, **P < 0.01, ***P < 0.001, one-way analysis of variance followed by Bonferroni post hoc test (n = 3 animals).

Figure 4.

Postembedding immunogold labeling of VNUT in the dentate gyrus and CA1. (A–C) Examples of VNUT-immunopositive terminals (t) establishing asymmetric axospinous synapses in the dentate molecular layer (A,B) or CA1 stratum radiatum (C). Note that the labeling is concentrated over synaptic vesicles. (D,E) Examples of VNUT-immunolabeled terminals forming symmetric axosomatic synapses in the granule cell layer. (F–H) Unmyelinated (ax) and myelinated (my) axons exhibiting strong VNUT immunoreactivity in CA1 stratum radiatum (F) and the dentate molecular layer (G,H). In F, t indicates an immunolabeled terminal. (I–K) Examples of VNUT-immunopositive dendritic spines (s) postsynaptic to presumed excitatory terminals (t) in the dentate molecular layer (I,J) and CA1 stratum radiatum (K). (L) VNUT labeling density in presumed excitatory terminals, dendritic spines, and unmyelinated axons of the molecular layer of the dentate gyrus (DG) and CA1 stratum radiatum (see also Tables 1). The summed gold particle density (number of gold particles/μm2) was determined for each sample and normalized against background labeling density over endothelial basal lamina. Error bars indicate standard error of the mean. *P < 0.05, **P < 0.01, ***P < 0.001, one-way analysis of variance followed by Bonferroni post hoc test (n = 3 animals).

Moreover, the distance from the center of each VNUT gold particle to the plasma membrane was determined in excitatory terminals, postsynaptic dendritic spines, and unmyelinated axons in the MLo and SR. The frequency distributions of such distances were compared with the distribution of distances between the center of points randomly distributed over the above mentioned profiles and the plasma membrane. In excitatory nerve terminals, VNUT gold particles were located significantly closer to the plasma membrane than random points (Fig. 5). The largest fraction of VNUT gold particles was situated within 60 nm from the plasma membrane. However, in unmyelinated axons, the reverse pattern was observed: the proportion of gold particles found immediately below the membrane was significantly smaller than would be expected if the distribution was random (Fig. 5). In dendritic spines, there was no significant difference between the distribution of VNUT immunogold labeling relative to the plasma membrane and that of random points.

Figure 5.

Subcellular distribution of VNUT immunolabeling in neuronal compartments of the neuropil in the molecular layer of the dentate gyrus (DG) and the CA1 stratum radiatum. Shown are the distributions of the distance of the center of each VNUT-signaling gold particle and of random points to the plasma membrane in presumed excitatory terminals, dendritic spines and unmyelinated axons (n = 3 animals). Error bars indicate standard error of the mean. **, P < 0.01, ***, P < 0.001, 2-way analysis of variance followed by Bonferroni post hoc test.

Figure 5.

Subcellular distribution of VNUT immunolabeling in neuronal compartments of the neuropil in the molecular layer of the dentate gyrus (DG) and the CA1 stratum radiatum. Shown are the distributions of the distance of the center of each VNUT-signaling gold particle and of random points to the plasma membrane in presumed excitatory terminals, dendritic spines and unmyelinated axons (n = 3 animals). Error bars indicate standard error of the mean. **, P < 0.01, ***, P < 0.001, 2-way analysis of variance followed by Bonferroni post hoc test.

Electron Microscopy—VNUT–VGLUT1 Double Immunogold Labeling

In tissue double immunogold labeled for VNUT and VGLUT1, we observed many terminals that labeled for both vesicular transporters (Fig. 6A; Supplementary Table S2). About half (51 ± 3%; n = 2 animals) of the VGLUT1-immunopositive nerve terminals in the MLo also contained VNUT labeling. This was in accordance with the proportion of presumed excitatory terminals that was VNUT immunopositive in single-labeling experiments. Moreover, the same axonal enrichment of VNUT immunogold particles was observed as in single-labeled sections. By contrast, VGLUT1 immunoreactivity was very sparse in unmyelinated axons as compared with excitatory terminals (Fig. 6B). Next, we determined the distance between the center of each VGLUT1-signaling gold particle to the center of the nearest VNUT gold particle (Fig. 6C). Given a lateral resolution of the immunogold labeling of about 25 nm (determined by the radius of the gold particle and the dimension of the antibody bridge) (Bergersen et al. 2008) and a synaptic vesicle diameter of ∼40 nm, 2 gold particles located closer together than 90 nm could signal antigens on the same vesicle. A minority (15 ± 0%; n = 2 animals) of VGLUT1–VNUT interparticle distances was less than 90 nm (Fig. 6C). However, in nerve terminals, where synaptic vesicles are spaced close together, the immunogold method cannot precisely ascribe a gold particle to a particular synaptic vesicle. Nevertheless, the VGLUT1–VNUT interparticle distances were significantly shorter than distances between randomly distributed points and VNUT gold particles (Fig. 6C), suggesting that vesicles possessing VNUT and vesicles with VGLUT1 to some extent colocalize in the terminal. Importantly, the majority (85 ± 0%; n = 2 animals) of VNUT and VGLUT1 immunogold particles were localized above 90 nm from each other (Fig. 6C) and thus cannot signal VNUT and VGLUT1 in the same vesicle. Analysis of the perpendicular distance from gold particles representing VNUT and VGLUT1 to the plasma membrane in double-labeled excitatory nerve terminals showed that VNUT immunogold particles were significantly closer to the plasma membrane than were gold particles indicating VGLUT1 (Fig. 6D). Further this suggests that VNUT and VGLUT1 are segregated into different vesicular pools.

Figure 6.

Colocalization of VNUT and VGLUT1 in double immunogold–labeled hippocampal tissue (A) A terminal (t) double labeled for VGLUT1 (large gold particles) and VNUT (small gold particles; arrowheads) apposing an immunonegative spine (s) in the dentate molecular layer. Scale bar: 200 nm. (B) Ratio of VNUT and VGLUT1 immunogold labeling densities in unmyelinated axons versus excitatory nerve terminals in the dentate molecular layer (n = 2 rats). Error bars indicate standard error of the mean. *P < 0.05, Student's t-test. Scale bar in K: 200 nm, valid for A–K. (C) Boxplot (0, 25, 50, 75, and 100 percentiles) of distances of VGLUT1 gold particles (n = 279 particles) and of random points (n = 1848 points) to the nearest VNUT gold particle (n = 125 particles). ***P < 0.001, Mann–Whitney U-test. Similar observations were made in 2 animals. (D) The distributions of the distance of the center of each VNUT or VGLUT1 indicating gold particle to the plasma membrane in VNUT+/VGLUT1+ terminals. The P value indicates significant difference in distance distributions between VNUT and VGLUT1 immunolabeling as determined by χ2 test. Analysis of 2 animals yielded similar results.

Figure 6.

Colocalization of VNUT and VGLUT1 in double immunogold–labeled hippocampal tissue (A) A terminal (t) double labeled for VGLUT1 (large gold particles) and VNUT (small gold particles; arrowheads) apposing an immunonegative spine (s) in the dentate molecular layer. Scale bar: 200 nm. (B) Ratio of VNUT and VGLUT1 immunogold labeling densities in unmyelinated axons versus excitatory nerve terminals in the dentate molecular layer (n = 2 rats). Error bars indicate standard error of the mean. *P < 0.05, Student's t-test. Scale bar in K: 200 nm, valid for A–K. (C) Boxplot (0, 25, 50, 75, and 100 percentiles) of distances of VGLUT1 gold particles (n = 279 particles) and of random points (n = 1848 points) to the nearest VNUT gold particle (n = 125 particles). ***P < 0.001, Mann–Whitney U-test. Similar observations were made in 2 animals. (D) The distributions of the distance of the center of each VNUT or VGLUT1 indicating gold particle to the plasma membrane in VNUT+/VGLUT1+ terminals. The P value indicates significant difference in distance distributions between VNUT and VGLUT1 immunolabeling as determined by χ2 test. Analysis of 2 animals yielded similar results.

Electron Microscopy—VNUT–VGAT Double Immunogold Labeling

Double immunogold labeling for VNUT and VGAT demonstrated coexistence of these transporters in many presumed inhibitory terminals (Fig. 7). Quantitative analysis of the VNUT/VGAT double immunolabeling revealed no clear colocalization between VNUT and VGAT immunolabeling or any difference between VNUT and VGAT immunogold labeling with respect to their distribution relative to the plasma membrane.

Figure 7.

Colocalization of VNUT and VGAT immunolabeling in inhibitory nerve endings in the dentate granule cell layer in double immunogold–labeled tissue. A terminal (t) double labeled for VGAT (large gold particles) and VNUT (small gold particles; arrowheads) forming a symmetric synapse (arrow) with a cell body (cb). Scale bar: 200 nm.

Figure 7.

Colocalization of VNUT and VGAT immunolabeling in inhibitory nerve endings in the dentate granule cell layer in double immunogold–labeled tissue. A terminal (t) double labeled for VGAT (large gold particles) and VNUT (small gold particles; arrowheads) forming a symmetric synapse (arrow) with a cell body (cb). Scale bar: 200 nm.

Isolated Synaptic Vesicles

To determine whether VNUT and VGLUT1 may exist in the same synaptic vesicle, first, we used VNUT and VGLUT1 antibodies to immunoisolate VNUT- and VGLUT1-containing vesicles from crude synaptic vesicles obtained from whole forebrain. Then, we performed western blots to examine the presence of VGLUT1 and VNUT in the resulting eluates (the immunoisolated vesicle fraction) and supernatant (the vesicle fraction left after immunoisolation) (Fig. 8). Both the VNUT-immunoisolated vesicles and the non-VNUT vesicle fraction contained VGLUT1 (Fig. 8), suggesting that some, but not all VGLUT1-containing vesicles possess VNUT. VNUT immunoreactivity was not observed in the non-VNUT vesicle fraction, indicating that the VNUT immunoisolation efficiently depleted VNUT-containing vesicles from the crude synaptic vesicle fraction. Furthermore, VNUT was not found in VGLUT1 immunoisolated vesicles or in the corresponding supernatant. The failure to detect VNUT in vesicles isolated with anti-VGLUT1 is difficult to fully explain. However, it may be attributed partly to a low abundance of VNUT in VGLUT1-containing vesicles in combination with masking of VNUT antigenicity by other proteins of similar electrophoretic mobility. Moreover, the VNUT antibodies appear to have a relatively low affinity or titer, resulting in comparatively low sensitivity of the immunoblotting procedure for this antigen. This suggests that VNUT is restricted to a small proportion of VGLUT1-containing vesicles.

Figure 8.

Western blots of VGLUT1 and VNUT of forebrain vesicle fractions immunoisolated using VNUT or VGLUT1 antibody-coupled magnetic beads. Antigens that were targets for immunoisolation are indicated below the blots, whereas the antigens detected by immunoblotting are indicated to the left of each blot. eluate, material isolated with beads coupled to antibody reactive to the respective antigen. super, supernatant material that did not bind beads coupled to the respective antibody. The blots are representative of at least 3 separate immunoisolation experiments.

Figure 8.

Western blots of VGLUT1 and VNUT of forebrain vesicle fractions immunoisolated using VNUT or VGLUT1 antibody-coupled magnetic beads. Antigens that were targets for immunoisolation are indicated below the blots, whereas the antigens detected by immunoblotting are indicated to the left of each blot. eluate, material isolated with beads coupled to antibody reactive to the respective antigen. super, supernatant material that did not bind beads coupled to the respective antibody. The blots are representative of at least 3 separate immunoisolation experiments.

Discussion

The present study shows that VNUT is present in synaptic vesicles in a considerable proportion of glutamatergic and inhibitory nerve terminals in the hippocampal formation, that VNUT is localized to subpopulations of VGLUT1-containing vesicles, and that vesicular ATP release from cultured hippocampal neurons depends on VNUT. Thus, our data suggest that VNUT may confer a purinergic transmitter phenotype to neurons by establishing a vesicular pool of ATP that is releasable by regulated exocytosis (Sawada et al. 2008). This supports previous results suggesting that ATP is released by way of exocytosis from neurons (White 1978).

The quantitative data on the subcellular localization of VNUT in hippocampal neuropil presented here were obtained using the postembedding immunogold labeling technique. This method has distinct advantages for such analysis as compared with preembedding immunolabeling methods, including clear-cut measurement of labeling and a comparatively unbiased access to different epitopes in the tissue. On the other hand, the sensitivity of postembedding immunolabeling is relatively low, which may result in false-negative profiles. Therefore, the fractions of VNUT-immunopositive profiles of different types reported here are most likely underestimates of the true proportions of profiles containing antigen recognized by the antibody. This is especially true for small profiles such as the transversely cut unmyelinated axons and dendritic spines. Another potential problem with the postembedding immunogold technique, as with other immunochemical methods, is that of nonspecific background labeling. However, RNAi-mediated knockdown of VNUT or preabsorption with the peptide against which the VNUT antibody was raised, strongly attenuate VNUT immunolabeling as judged by western blot, light microscopic immunocytochemistry, and postembedding immunogold histochemistry, indicating that the observed VNUT labeling is specific (present study; Sawada et al. 2008; Iwatsuki et al. 2009; Haanes and Novak 2010). Furthermore, the nonrandom distribution of immunogold labeling, the clear association with synaptic vesicles and the distinct pattern of axonal and terminal labeling as compared to that of VGLUT1 and VGAT suggest that the postembedding VNUT immunogold labeling observed here reflected specific labeling.

Purinergic excitatory postsynaptic currents (EPSCs), indicative of quantal ATP release, have been described in several CNS regions (Edwards et al. 1992; Bardoni et al. 1997; Pankratov et al. 2002, 2003, 2007), including in the hippocampus (Pankratov et al. 1998, 2006; Mori et al. 2001). Furthermore, electrically evoked ATP release has been reported in the hippocampus. However, although this ATP was believed to at least partly originate from nerve terminals, the precise source could not be ascertained (Wieraszko et al. 1989; Cunha et al. 1996). The present findings complement and extend these studies by providing evidence at the ultrastructural level for presynaptic storage of transmitter ATP at hippocampal synapses and by identifying VNUT as the protein responsible for such storage. Moreover, previous immunogold studies have shown that there are postsynaptic purinergic receptors at the hippocampal synapses examined here (Rubio and Soto 2001; Tonazzini et al. 2007).

Whether ATP is always colocalized with other neurotransmitters in synaptic vesicles or is transported into a distinct subset of ATP-only vesicles is unclear. Pankratov et al. (2006, 2007) have shown that a proportion of low-amplitude miniature EPSCs in CA1 or neocortex layer 2/3 pyramidal cells is mediated by P2X but not glutamate receptors. Nevertheless, stimulation of single glutamatergic fibers often resulted in mixed glutamate/ATP-mediated EPSCs. These results are in line with the existence of a vesicular population containing ATP in glutamatergic terminals. However, whether ATP is located in the same vesicle pool as glutamate could not be definitely determined. The observations presented here suggest that ATP is stored in only a subset of glutamatergic VGLUT1-possessing vesicles. First, only a weak spatial relation was detected between VNUT-signaling immunogold particles and those indicating VGLUT1, although this could be attributed to inadequate sensitivity of the immunogold labeling. Second, in excitatory nerve terminals, VNUT immunolabeling was enriched near the plasma membrane (possibly indicating preferred labeling of docked vesicles), whereas no such pattern was seen for VGLUT1 labeling. Indeed, the distribution of the distance of VNUT gold particles to the plasma membrane was significantly different from that of VGLUT1. Third, whereas VGLUT1 is present at low levels in unmyelinated preterminal axons, such axons showed high levels of VNUT immunogold labeling compared with nerve terminals. Although VNUT may be present also in axons that do not express VGLUT1, it is not likely to explain the observed difference in axonal enrichment of VNUT versus VGLUT1 immunolabeling. Thus, the distinct axonal localization of VNUT immunolabeling as compared with that of VGLUT1 clearly indicates that these transporters may be differentially distributed between different vesicular populations. Thus, the present analysis of immunogold labeling suggests that VNUT is at least partly segregated from VGLUT1+ vesicles. Whether VNUT is present on a subpopulation of VGLUT1+ vesicles could not be firmly determined from the immunogold data. However, our immunoisolation experiments suggest that some, but far from all, VGLUT1+ vesicles also possess VNUT, whereas a considerable proportion of VNUT+ vesicles also contains VGLUT1. Taken together, our observations indicate that VNUT likely is present in a subpopulation of VGLUT1+ vesicles and that many VGLUT1+ vesicles do not have VNUT. Analysis of immunogold labeling suggests that a similar partial segregation exists between VGAT- and VNUT-containing vesicles in inhibitory terminals. Notably, whether some VNUT-containing vesicles in VGLUT1- or VGAT-expressing neurons lack either of the latter transporters and thus store ATP but not glutamate or ?-aminobutyric acid could not be inferred from the present data.

The VNUT immunoreactivity in preterminal axons was preferentially located to the axoplasm as opposed to near the plasma membrane, suggesting that VNUT-containing vesicles in the axon are undergoing axonal transport. However, such vesicles may be releasable under some conditions. It has been suggested that axonal activity induces release of ATP from the axons (Stevens and Fields 2000) possibly via an exocytotic mechanism (Thyssen et al. 2010; but see Fields and Ni 2010), activating purinergic receptors on NG2 glia and astrocytes (Ishibashi et al. 2006; Hamilton et al. 2008, 2010). This forms an interesting parallel to vesicular glutamate, which has recently been shown to be liberated from unmyelinated axons onto NG2 glia in the corpus callosum (Kukley et al. 2007; Ziskin et al. 2007).

Endogenously released ATP can activate presynaptic P2X receptors which induces or enhances glutamate release at many synapses (Gu and MacDermott 1997; Khakh and Henderson 1998; Nakatsuka and Gu 2001; Khakh et al. 2003; Rodrigues et al. 2005; Sperlágh et al. 2007; Khakh 2009), whereas presynaptic P2Y receptors may negatively modulate transmitter release (Rodrigues et al. 2005). Furthermore, adenosine derived from extracellular ATP may bidirectionally modulate transmitter release via presynaptic adenosine receptors (Gonçalves and Queiroz 2008). The ATP involved in presynaptic modulation could originate from, for example, astrocytes, neighboring nerve terminals, or the terminal itself. Indeed, in a parallel study, evidence of VNUT-mediated release of ATP from astrocytes has been obtained (Y. Moriyama, S. Koizumi, unpublished data), suggesting that also astrocytes are capable of exocytotic ATP release (see Pascual et al. 2005). An additional potential source of transmitter ATP that has received little attention is postsynaptic dendrites. In fact, we observed VNUT immunoreactivity in association with vesicular structures in postsynaptic dendritic spines. In line with this, in the cell body of dorsal root ganglion neurons it has been suggested that small and clear vesicles exocytose ATP, leading to activation of surrounding satellite cells (Zhang et al. 2007). The function of the postsynaptically localized VNUT needs to be addressed in future studies. However, it is tempting to speculate that certain spines have the ability to release ATP that may act as a retrograde signal to modulate presynaptic transmitter release.

In conclusion, we have found that hippocampal neurons in the rat are capable of releasing ATP transported into synaptic vesicles by VNUT. Furthermore, VNUT is broadly expressed in the brain, including in the cerebellar cortex, the olfactory bulb, and the hippocampus. At the ultrastructural level, VNUT is present in a wide variety of neuronal compartments in the hippocampal neuropil, including excitatory and inhibitory nerve terminals, postsynaptic dendrites and preterminal axons. Thus, multiple neuronal sources of ATP may contribute to a complex pattern of presynaptic and postsynaptic purinergic modulation of synaptic transmission.

Supplementary Material

Supplementary material can be found at: http://www.cercor.oxfordjournals.org/

Funding

This work was supported by the Norwegian Research Council and by a Grant-in-Aid from the Japanese Ministry of Education, Science, Sport and Culture.

Conflict of Interest : None declared.

References

Bardoni
R
Goldstein
PA
Lee
CJ
Gu
JG
MacDermott
AB
ATP P2X receptors mediate fast synaptic transmission in the dorsal horn of the rat spinal cord
J Neurosci
 , 
1997
, vol. 
17
 (pg. 
5297
-
5304
)
Bergersen
LH
Storm-Mathisen
J
Gundersen
V
Immunogold quantification of amino acids and proteins in complex subcellular compartments
Nat Protoc
 , 
2008
, vol. 
3
 (pg. 
144
-
152
)
Cunha
RA
Vizi
ES
Ribeiro
JA
Sebastiao
AM
Preferential release of ATP and its extracellular catabolism as a source of adenosine upon high- but not low-frequency stimulation of rat hippocampal slices
J Neurochem
 , 
1996
, vol. 
67
 (pg. 
2180
-
2187
)
Edwards
FA
Gibb
AJ
Colquhoun
D
ATP receptor-mediated synaptic currents in the central nervous system
Nature
 , 
1992
, vol. 
359
 (pg. 
144
-
147
)
Fields
RD
Burnstock
G
Purinergic signalling in neuron-glia interactions
Nat Rev Neurosci
 , 
2006
, vol. 
7
 (pg. 
423
-
436
)
Fields
RD
Ni
Y
Nonsynaptic communication through ATP release from volume-activated anion channels in axons
Sci Signal
 , 
2010
, vol. 
3
 pg. 
ra73
 
Fukuda
J
Fujita
Y
Ohsawa
K
ATP content in isolated mammalian nerve cells assayed by a modified luciferin-luciferase method
J Neurosci Methods
 , 
1983
, vol. 
8
 (pg. 
295
-
302
)
Gonçalves
J
Queiroz
G
Presynaptic adenosine and P2Y receptors
Handb Exp Pharmacol
 , 
2008
, vol. 
184
 (pg. 
339
-
372
)
Gu
JG
MacDermott
AB
Activation of ATP P2X receptors elicits glutamate release from sensory neuron synapses
Nature
 , 
1997
, vol. 
389
 (pg. 
749
-
753
)
Gualix
J
Pintor
J
Miras-Portugal
MT
Characterization of nucleotide transport into rat brain synaptic vesicles
J Neurochem
 , 
1999
, vol. 
73
 (pg. 
1098
-
1104
)
Haanes
KA
Novak
I
ATP storage and uptake by isolated pancreatic zymogen granules
Biochem J
 , 
2010
, vol. 
429
 (pg. 
303
-
311
)
Hamilton
N
Vayro
S
Kirchhoff
F
Verkhratsky
A
Robbins
J
Gorecki
DC
Butt
AM
Mechanisms of ATP- and glutamate-mediated calcium signalling in white matter astrocytes
Glia
 , 
2008
, vol. 
56
 (pg. 
734
-
749
)
Hamilton
N
Vayro
S
Wigley
R
Butt
AM
Axons and astrocytes release ATP and glutamate to evoke calcium signals in NG2-glia
Glia
 , 
2010
, vol. 
58
 (pg. 
66
-
79
)
Hamilton
NB
Attwell
D
Do astrocytes really exocytose neurotransmitters?
Nat Rev Neurosci
 , 
2010
, vol. 
11
 (pg. 
227
-
238
)
Ishibashi
T
Dakin
KA
Stevens
B
Lee
PR
Kozlov
SV
Stewart
CL
Fields
RD
Astrocytes promote myelination in response to electrical impulses
Neuron
 , 
2006
, vol. 
49
 (pg. 
823
-
832
)
Iwatsuki
K
Ichikawa
R
Hiasa
M
Moriyama
Y
Torii
K
Uneyama
H
Identification of the vesicular nucleotide transporter (VNUT) in taste cells
Biochem Biophys Res Commun
 , 
2009
, vol. 
388
 (pg. 
1
-
5
)
Jourdain
P
Bergersen
LH
Bhaukaurally
K
Bezzi
P
Santello
M
Domercq
M
Matute
C
Tonello
F
Gundersen
V
Volterra
A
Glutamate exocytosis from astrocytes controls synaptic strength
Nat Neurosci
 , 
2007
, vol. 
10
 (pg. 
331
-
339
)
Kang
J
Kang
N
Lovatt
D
Torres
A
Zhao
Z
Lin
J
Nedergaard
M
Connexin 43 hemichannels are permeable to ATP
J Neurosci
 , 
2008
, vol. 
28
 (pg. 
4702
-
4711
)
Khakh
BS
ATP-gated P2X receptors on excitatory nerve terminals onto interneurons initiate a form of asynchronous glutamate release
Neuropharmacology
 , 
2009
, vol. 
56
 (pg. 
216
-
222
)
Khakh
BS
Gittermann
D
Cockayne
DA
Jones
A
ATP modulation of excitatory synapses onto interneurons
J Neurosci
 , 
2003
, vol. 
23
 (pg. 
7426
-
7437
)
Khakh
BS
Henderson
G
ATP receptor-mediated enhancement of fast excitatory neurotransmitter release in the brain
Mol Pharmacol
 , 
1998
, vol. 
54
 (pg. 
372
-
378
)
Koizumi
S
Fujishita
K
Inoue
K
Regulation of cell-to-cell communication mediated by astrocytic ATP in the CNS
Purinergic Signal
 , 
2005
, vol. 
1
 (pg. 
211
-
217
)
Kukley
M
Capetillo-Zarate
E
Dietrich
D
Vesicular glutamate release from axons in white matter
Nat Neurosci
 , 
2007
, vol. 
10
 (pg. 
311
-
320
)
Larsson
M
Broman
J
Different basal levels of CaMKII phosphorylated at Thr286/287 at nociceptive and low-threshold primary afferent synapses
Eur J Neurosci
 , 
2005
, vol. 
21
 (pg. 
2445
-
2458
)
Linhoff
MW
Laurén
J
Cassidy
RM
Dobie
FA
Takahashi
H
Nygaard
HB
Airaksinen
MS
Strittmatter
SM
Craig
AM
An unbiased expression screen for synaptogenic proteins identifies the LRRTM protein family as synaptic organizers
Neuron
 , 
2009
, vol. 
61
 (pg. 
734
-
749
)
Liu
HT
Sabirov
RZ
Okada
Y
Oxygen-glucose deprivation induces ATP release via maxi-anion channels in astrocytes
Purinergic Signal
 , 
2008
, vol. 
4
 (pg. 
147
-
154
)
Liu
HT
Toychiev
AH
Takahashi
N
Sabirov
RZ
Okada
Y
Maxi-anion channel as a candidate pathway for osmosensitive ATP release from mouse astrocytes in primary culture
Cell Res
 , 
2008
, vol. 
18
 (pg. 
558
-
565
)
Mori
M
Heuss
C
Gahwiler
BH
Gerber
U
Fast synaptic transmission mediated by P2X receptors in CA3 pyramidal cells of rat hippocampal slice cultures
J Physiol
 , 
2001
, vol. 
535
 (pg. 
115
-
123
)
Moriyama
Y
Yamamoto
A
Yamada
H
Tashiro
Y
Tomochika
K
Takahashi
M
Maeda
M
Futai
M
Microvesicles isolated from bovine posterior pituitary accumulate norepinephrine
J Biol Chem
 , 
1995
, vol. 
270
 (pg. 
11424
-
11429
)
Mullins
LJ
Brinley
FJ
Jr.
Some factors influencing sodium extrusion by internally dialyzed squid axons
J Gen Physiol
 , 
1967
, vol. 
50
 (pg. 
2333
-
2355
)
Nakatsuka
T
Gu
JG
ATP P2X receptor-mediated enhancement of glutamate release and evoked EPSCs in dorsal horn neurons of the rat spinal cord
J Neurosci
 , 
2001
, vol. 
21
 (pg. 
6522
-
6531
)
Nörenberg
W
Illes
P
Neuronal P2X receptors: localisation and functional properties
Naunyn Schmiedebergs Arch Pharmacol
 , 
2000
, vol. 
362
 (pg. 
324
-
339
)
Pankratov
Y
Castro
E
Miras-Portugal
MT
Krishtal
O
A purinergic component of the excitatory postsynaptic current mediated by P2X receptors in the CA1 neurons of the rat hippocampus
Eur J Neurosci
 , 
1998
, vol. 
10
 (pg. 
3898
-
3902
)
Pankratov
Y
Lalo
U
Krishtal
O
Verkhratsky
A
Ionotropic P2X purinoreceptors mediate synaptic transmission in rat pyramidal neurones of layer II/III of somato-sensory cortex
J Physiol
 , 
2002
, vol. 
542
 (pg. 
529
-
536
)
Pankratov
Y
Lalo
U
Krishtal
O
Verkhratsky
A
P2X receptor-mediated excitatory synaptic currents in somatosensory cortex
Mol Cell Neurosci
 , 
2003
, vol. 
24
 (pg. 
842
-
849
)
Pankratov
Y
Lalo
U
Verkhratsky
A
North
RA
Vesicular release of ATP at central synapses
Pflugers Arch
 , 
2006
, vol. 
452
 (pg. 
589
-
597
)
Pankratov
Y
Lalo
U
Verkhratsky
A
North
RA
Quantal release of ATP in mouse cortex
J Gen Physiol
 , 
2007
, vol. 
129
 (pg. 
257
-
265
)
Pascual
O
Casper
KB
Kubera
C
Zhang
J
Revilla-Sanchez
R
Sul
JY
Takano
H
Moss
SJ
McCarthy
K
Haydon
PG
Astrocytic purinergic signalling coordinates synaptic networks
Science
 , 
2005
, vol. 
310
 (pg. 
113
-
116
)
Persson
S
Boulland
JL
Aspling
M
Larsson
M
Fremeau
RT
Edwards
RH
Storm-Mathisen
J
Chaudhry
FA
Broman
J
Distribution of vesicular glutamate transporters 1 and 2 in the rat spinal cord, with a note on the spinocervical tract
J Comp Neurol
 , 
2006
, vol. 
497
 (pg. 
683
-
701
)
Rodrigues
RJ
Almeida
T
Richardson
PJ
Oliveira
CR
Cunha
RA
Dual presynaptic control by ATP of glutamate release via facilitatory P2X1, P2X2/3, and P2X3 and inhibitory P2Y1, P2Y2, and/or P2Y4 receptors in the rat hippocampus
J Neurosci
 , 
2005
, vol. 
25
 (pg. 
6286
-
6295
)
Rubio
ME
Soto
F
Distinct localization of P2X receptors at excitatory postsynaptic specializations
J Neurosci
 , 
2001
, vol. 
21
 (pg. 
641
-
653
)
Sawada
K
Echigo
N
Juge
N
Miyaji
T
Otsuka
M
Omote
H
Yamamoto
A
Moriyama
Y
Identification of a vesicular nucleotide transporter
Proc Natl Acad Sci U S A
 , 
2008
, vol. 
105
 (pg. 
5683
-
5686
)
Shoji-Kasai
Y
Yoshida
A
Sato
K
Hoshino
T
Ogura
A
Kondo
S
Fujimoto
Y
Kuwahara
R
Kato
R
Takahashi
M
Neurotransmitter release from synaptotagmin-deficient clonal variants of PC12 cells
Science
 , 
1992
, vol. 
256
 (pg. 
1821
-
1823
)
Sperlágh
B
Heinrich
A
Csölle
C
P2 receptor-mediated modulation of neurotransmitter release-an update
Purinergic Signal
 , 
2007
, vol. 
3
 (pg. 
269
-
284
)
Stevens
B
Fields
RD
Response of Schwann cells to action potentials in development
Science
 , 
2000
, vol. 
287
 (pg. 
2267
-
2271
)
Stout
CE
Costantin
JL
Naus
CC
Charles
AC
Intercellular calcium signalling in astrocytes via ATP release through connexin hemichannels
J Biol Chem
 , 
2002
, vol. 
277
 (pg. 
10482
-
10488
)
Thyssen
A
Hirnet
D
Wolburg
H
Schmalzing
G
Deitmer
JW
Lohr
C
Ectopic vesicular neurotransmitter release along sensory axons mediates neurovascular coupling via glial calcium signalling
Proc Natl Acad Sci U S A
 , 
2010
, vol. 
107
 (pg. 
15258
-
15263
)
Tonazzini
I
Trincavelli
ML
Storm-Mathisen
J
Martini
C
Bergersen
LH
Co-localization and functional cross-talk between A1 and P2Y1 purine receptors in rat hippocampus
Eur J Neurosci
 , 
2007
, vol. 
26
 (pg. 
890
-
902
)
White
TD
Release of ATP from a synaptosomal preparation by elevated extracellular K+ and by veratridine
J Neurochem
 , 
1978
, vol. 
30
 (pg. 
329
-
336
)
Wiedenmann
B
Franke
WW
Identification and localization of synaptophysin, an integral membrane glycoprotein of Mr 38,000 characteristic of presynaptic vesicles
Cell
 , 
1985
, vol. 
41
 (pg. 
1017
-
1028
)
Wieraszko
A
Goldsmith
G
Seyfried
TN
Stimulation-dependent release of adenosine triphosphate from hippocampal slices
Brain Res
 , 
1989
, vol. 
485
 (pg. 
244
-
250
)
Zander
JF
Munster-Wandowski
A
Brunk
I
Pahner
I
Gomez-Lira
G
Heinemann
U
Gutierrez
R
Laube
G
Ahnert-Hilger
G
Synaptic and vesicular coexistence of VGLUT and VGAT in selected excitatory and inhibitory synapses
J Neurosci
 , 
2010
, vol. 
30
 (pg. 
7634
-
7645
)
Zhang
X
Chen
Y
Wang
C
Huang
LY
Neuronal somatic ATP release triggers neuron-satellite glial cell communication in dorsal root ganglia
Proc Natl Acad Sci U S A
 , 
2007
, vol. 
104
 (pg. 
9864
-
9869
)
Ziskin
JL
Nishiyama
A
Rubio
M
Fukaya
M
Bergles
DE
Vesicular release of glutamate from unmyelinated axons in white matter
Nat Neurosci
 , 
2007
, vol. 
10
 (pg. 
321
-
330
)