HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 52, 50734-50748, December 27, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/52/50734    most recent M206738200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Schäfer, M. K.-H. Articles by Erickson, J. D. Search for Related Content PubMed PubMed Citation Articles by Schäfer, M. K.-H. Articles by Erickson, J. D. Social Bookmarking                      What's this?

Molecular Cloning and Functional Identification of Mouse Vesicular Glutamate Transporter 3 and Its Expression in Subsets of Novel Excitatory Neurons^*

Martin K.-H. Schäfer^§, Hélène Varoqui^§¶, Norah Defamie^¶, Eberhard Weihe, and Jeffrey D. Erickson^¶**

From the  Department of Molecular Neuroscience, Institute of Anatomy and Cell Biology, Philipps University Marburg, D-35033 Marburg, Germany and ^¶ Neuroscience Center and Departments of  Ophthalmology and ^** Pharmacology, Louisiana State University Health Sciences Center, New Orleans, Louisiana 70112

Received for publication, July 8, 2002, and in revised form, September 23, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

We have cloned and functionally characterized a third isoform of a vesicular glutamate transporter (VGLUT3) expressed on synaptic vesicles that identifies a distinct glutamatergic system in the brain that is partly and selectively promiscuous with cholinergic and serotoninergic transmission. Transport activity was specific for glutamate, was H⁺-dependent, was stimulated by Cl ion, and was inhibited by Rose Bengal and trypan blue. Northern analysis revealed higher mRNA levels in early postnatal development than in adult brain. Restricted patterns of mRNA expression were observed in presumed interneurons in cortex and hippocampus, and projection systems were observed in the lateral and ventrolateral hypothalamic nuclei, limbic system, and brainstem. Double in situ hybridization histochemistry for vesicular acetylcholine transporter identified VGLUT3 neurons in the striatum as cholinergic interneurons, whereas VGLUT3 mRNA and protein were absent from all other cholinergic cell groups. In the brainstem VGLUT3 mRNA was concentrated in mesopontine raphé nuclei. VGLUT3 immunoreactivity was present throughout the brain in a diffuse system of thick and thin beaded varicose fibers much less abundant than, and strictly separated from, VGLUT1 or VGLUT2 synapses. Co-existence of VGLUT3 in VMAT2-positive and tyrosine hydroxylase -negative varicosities only in the cerebral cortex and hippocampus and in subsets of tryptophan hydroxylase-positive cell bodies and processes in differentiating primary raphé neurons in vitro indicates selective and target-specific expression of the glutamatergic/serotoninergic synaptic phenotype.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Glutamate, the major excitatory neurotransmitter in the mammalian brain, is packaged in synaptic vesicles in glutamatergic nerve terminals prior to its regulated release from neurons by exocytosis (1, 2). Like other classical neurotransmitters, vesicular uptake and storage of glutamate is dependent upon the proton electrochemical gradient established by the vacuolar-type H⁺-ATPase (3-5). The accumulation of glutamate from the cytoplasm into vesicles is mediated by a selective low affinity transport system (6, 7) and requires low concentrations of chloride ion for maximal activity (8, 9).

Recently, the brain-specific Na⁺-dependent inorganic phosphate transporter (10) and differentiation-associated Na⁺-dependent inorganic phosphate transporter (11) have been established as two isoforms of the vesicular glutamate transporter (VGLUT1¹ and VGLUT2) by genetic (12), electrophysiological (13, 14), and biochemical methods (13-18). Two sets of glutamatergic neurons in the central nervous system have now been defined with VGLUT1 being the dominant system in telencephalic regions and VGLUT2 predominantly expressed in the diencephalic regions of the brain and in the spinal cord (11, 16-22). VGLUT1 and VGLUT2 define two sets of synaptic vesicles (13, 14), and their pathway-specific expression (e.g. VGLUT1 in learning and memory and VGLUT2 in sensory processing, i.e. nociception) in distinct sets of synapses along the pathway indicates that they may have unique intrinsic activity, trafficking patterns, or regulatory properties (16-18). Their combined distribution pattern can account for most, if not all, of the presumed glutamatergic neurons in the central nervous system.

The expression of a vesicular glutamate transporter is necessary and may be sufficient for exocytotic release of glutamate from a nerve terminal (13). Recently, the human genome project has revealed the presence of a highly related sequence to VGLUT1 and VGLUT2. The existence of a third isoform of the vesicular glutamate transporter may therefore define a novel excitatory phenotype in the brain. Glutamate is a likely candidate for the substrate of this transporter. Aspartate is also released from some excitatory nerve endings (23), and neither VGLUT1 nor VGLUT2 recognize aspartate as a substrate. Considerable evidence also exists to suggest that some monoaminergic (24) and cholinergic neurons (25-27) have high concentrations of glutamate that might be important for excitatory neurotransmission. Monoaminergic neurons in primary culture also exhibit mixed chemical neurotransmitter phenotypes and appear to co-release glutamate. The fast excitatory responses are apparently blocked by glutamate receptor antagonists and are followed by slower inhibitory responses mediated by 5-HT or dopamine in postsynaptic electrophysiologic recordings at autapses (28, 29).

We report the cloning of a novel cDNA from newborn mouse brain with high sequence identity to VGLUT1 and VGLUT2, and we describe its functional properties following transient transfection of cultured neuroendocrine cells. Northern analysis was performed to examine the relative abundance of mRNA expression in developing whole brain compared with VGLUT1 and VGLUT2. Antibodies were raised against the N and C terminus of the mouse homologue and were used to examine the presence of protein on synaptic vesicles and the distribution pattern in the mouse and rat brain compared with VGLUT1 and VGLUT2 by immunohistochemistry. In addition, double-confocal scanning fluorescence microscopy and double in situ hybridization histochemistry (ISHH) with classical neurochemical phenotypic markers were also used to identify aminergic neurons where co-existence with vesicular glutamate storage might occur. Finally, dissociated primary cultures of raphé nuclei were examined by double fluorescence deconvolution microscopy of VGLUT3 and the serotonin-synthesizing enzyme tryptophan hydroxylase (TrypH) to determine whether a subpopulation of developing serotoninergic neurons exhibit a mixed classical neurotransmitter phenotype.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cloning of VGLUT3 and Plasmid Constructions-- We searched a translated mouse expressed sequence tag (EST) data base with rat VGLUT2 amino acid sequence, and we found two ESTs (GenBank^TM accession numbers BB650431 and BB396899) that were highly homologous to VGLUT1 and VGLUT2. Specific forward and reverse primers containing EcoRI sites were then designed (5'-ggaattcctccatcgccccccgttcaaaatgcaattt and 5'-ggaattccagttgtaagctgaggtgaagccaga) to amplify a cDNA containing the protein coding region of the mouse homologue by Pfu Turbo DNA polymerase (Stratagene) from postnatal day 1 mouse brain cDNA. This product was sequenced on both strands by the ThermoSequenase cycle sequencing kit (Amersham Biosciences). VGLUT3 cDNA was subcloned into pCDNA3.1 (Invitrogen) for functional studies. For the preparation of glutathione S-transferase fusion proteins, sequences corresponding to the coding region of the terminal 56 hydrophilic amino acids of VGLUT3 (** in Fig. 1) were amplified by PCR from the VGLUT3 cDNA. The following primers (5'-cgggatccgaactcaaccacgagactttcgtaagt and 5'-ggaattcttaggatgtttctgagaagtctccttcggc) were designed to contain BamHI and EcoRI sites that facilitated cloning into the bacterial expression vector pGEX-KT (APBiotech). For the preparation of riboprobes, sequences corresponding to the non-conserved C terminus were amplified by PCR from the VGLUT3 cDNA. The following primers (5'-cgggatccgaactcaaccacgagactttcgtaagt and 5'-ggaattccagttgtaagctgaggtgaagccaga) were designed to contain BamHI and EcoRI sites that facilitated cloning into pBluescript (Stratagene) to drive expression from T7 (sense) and T3 promoters (antisense). This sequence (205 bp) corresponds to the portion of the C-terminal sequence used for GST fusion protein and antibody production and also includes some 3'-non-coding sequences of VGLUT3.

Antibody Preparation and Characterization-- Polyclonal antibodies were raised against the N terminus of mouse VGLUT3 in rabbits and the C terminus of mouse VGLUT3 in guinea pigs. A cysteine-linked N-terminal peptide (MPFKAFDTFKEKILKP-C) containing amino acids that are specific for VGLUT3 was coupled to keyhole limpet hemocyanin (Pierce), mixed with Freund's adjuvant, and used as immunogen in White New Zealand rabbits. Antibodies were affinity-purified using the immunogenic peptide and SulfoLink kit (Pierce). The C-terminal fusion protein was prepared from the recombinant GST fusion plasmid prepared above as described previously (18). Samples (10 mg) were dialyzed overnight against 10 mM Tris, adjusted to pH 6.0 with HCl, and then mixed with 10 ml of colloidal gold (d = 20 nm, Sigma) under stirring at room temperature. This solution was centrifuged (20,000 × g; 15 min), and the reddish pellet was resuspended in a solution containing dialyzed GST-VGLUT3 fusion protein (2 mg/ml), mixed with Freund's adjuvant, and used as the immunogen in guinea pigs. A similar immunogold immunization protocol has successfully produced extremely high titer polyclonal antibodies (30).

Antisera and affinity-purified antibodies were first screened by immunocytochemistry in transfected cells to assess the relative titers and cross-reactivities of VGLUT3 antibodies toward VGLUT1 and VGLUT2. CV-1 fibroblasts were plated onto collagen (type IV)-coated 2-well chamber slides (5 × 10⁴ cells/well) and transfected using the vaccinia virus/bacteriophage T7 hybrid system (see below) with plasmids containing VGLUT1, VGLUT2, VGLUT3, or mock controls. After 18 h, the cells were fixed and processed as described (31). Antigens were visualized under bright field after incubation with rabbit or guinea pig VGLUT antibodies (1:4000), followed by horseradish peroxidase-conjugated anti-rabbit or anti-guinea pig secondary antibodies (Vectastain Elite Kit; Vector Laboratories) and reaction with the peroxidase substrate 3'3-diaminobenzidine.

Specificity of the VGLUT3 antibodies and cross-reactivity of the VGLUT1 and VGLUT2 antibodies toward VGLUT3 were also assessed by Western blot. Post-nuclear homogenates from probe-sonicated PC12 cells (10 µg of protein) were resuspended in sample buffer containing 62 mM Tris-HCl, pH 6.8, 1 mM EDTA, 10% glycerol, 5% SDS, and 50 mM dithiothreitol, fractionated by SDS-PAGE through an 8% polyacrylamide gel, and electrotransfered onto nitrocellulose membrane (Hybond-ECL; Amersham Biosciences). Following a 30-min preincubation in TBS (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Tween 20) containing 5% non-fat dry milk, the blots were incubated for 3 h at room temperature with primary antibodies in TBS, 1% bovine serum albumin. For preadsorption studies, VGLUT3 peptide or GST fusion protein (1 µM; final concentration) was added during the incubation period with the primary antibodies. The VGLUT1, VGLUT2, and VGLUT3 fusion protein antibodies (1:3000 to 1:10,000) and the VGLUT3 anti-peptide antibody (1:3000) were detected using horseradish peroxidase-conjugated anti-rabbit or anti-guinea pig IgG secondary antibodies (Sigma), respectively, and enhanced chemiluminescent reagents (SuperSignal WestPico substrate; Pierce) followed by exposure to Hyperfilm ECL (Amersham Biosciences). Rabbit and guinea pig antibodies raised against different epitopes of mouse VGLUT3 sequence were completely cross-reactive with rat as identical distribution patterns were observed using these antibodies in both species (see below).

Northern Analysis-- RNA was purified from whole mouse brain at various developmental ages (E18, P1, P7, P14, P26, and P60) by guanidine isothiocyanate extraction and ultracentrifugation through a cesium trifluoroacetic acid cushion followed by a single round of oligo(dT)-cellulose chromatography (32). Ten micrograms of poly(A)⁺ RNA were electrophoresed through formaldehyde-agarose gels, electroblotted onto a nylon membrane, and hybridized with ³²P-labeled oligonucleotide probes in buffer containing 4× SSC (600 mM NaCl, 60 mM sodium citrate, pH 7.0); 50% formamide; 5× Denhardt's solution (0.1% Ficoll, 0.1% polyvinylpyrrolidone, 0.1% bovine serum albumin (BSA)); 0.25 mg/ml tRNA for 18 h at 45 °C. Blots were washed in 1× SSC, 0.1% SDS at 60 °C. Four VGLUT3 oligonucleotide probes from non-conserved regions of the mouse VGLUT3 sequence were used simultaneously for hybridization as follows: 5'-gctcagactggccaagttggccccttctcctatgcttgtctctatg from the N terminus; 5'-ctttgtaagatgcccagcgagtctcccacggcattcttcactccttcc from the predicted lumenal loop between TMD V and VI; 5'-ctcgtggttgagttctgtctcctcagctaactcatcttggtcaatgattcc; and 5'-gtaagctgaggtgaagccagacatttaggatgtttctgagaagtctccttcg from the C terminus and 3'-non-coding region. Specific probes against the N-terminal sequences of mouse VGLUT1 (5'-ccgccagcttccgaaactcctcctgccggaactccatggtggctgtgc; derived from mouse EST data base GenBank^TM accession numbers BE950784, BB649955, and BB650431) and VGLUT2 (5'-cttaccgtcctctgtcagctcgatggtctctcggttgtcctgcttcttc; GenBank^TM accession number NM080853) were also prepared. Oligonucleotide probes were labeled to a similar specific activity using terminal transferase (Roche Molecular Biochemicals) and [-³²P]dATP (6000 Ci/mmol; PerkinElmer Life Sciences). The VGLUT probes were hybridized sequentially to the developmental Northern blot of whole mouse brain mRNA beginning with VGLUT3 and then VGLUT2 which was followed by VGLUT1. Oligonucleotides were stripped from the blot by incubation at 60 °C in 95% formamide containing 10 mM Tris and 1 mM EDTA, pH 8.0, for 30 min followed by a thorough wash in 1× SSC. Blots were placed against an image plate for various times and developed with a Typhoon 8600 PhosphorImager (Amersham Biosciences).

ISHH-- Frozen serial sections through the mouse brain and spinal cord were cut to 14-µm thickness on a cryostat, thaw-mounted on adhesive slides, and stored at 70 °C until subjected to the prehybridization procedure as described (33). Tissues were post-fixed on the slide by immersion in ice-cold 4% phosphate-buffered formaldehyde solution for 1 h and then washed in 10 mM phosphate-buffered saline (PBS), pH 7.4, three times for 10 min each and once in PBS containing 0.4% Triton X-100 at room temperature. After a short rinse in distilled water, the sections were washed in 0.1 M triethanolamine, pH 8.0 (Sigma), followed by a second wash in 0.1 M triethanolamine, pH 8.0, containing acetic anhydride (0.25% v/v) for 10 min at room temperature. After incubation in 2× SSC, sections were dehydrated in graded alcohols (50 and 70%) and air-dried.

Radioactive probes were generated by in vitro transcription using ³⁵S-UTP and ³⁵S-CTP as radioactive label and diluted in hybridization buffer (3× SSC, 50 mM NaPO₄, 10 mM dithiothreitol, 1× Denhardt's solution, 0.25 mg/ml yeast tRNA, 10% dextran sulfate, and 50% formamide) to a final concentration of 5 × 10⁴ dpm/µl, and 30-50 µl of hybridization solution was applied to each section, and slides were coverslipped and incubated for 14 h at 60 °C in a humid chamber. Slides were washed in 2× SSC and 1× SSC for 20 min each followed by incubation in RNase A buffer (10 mM Tris, pH 8.0, 0.5 M NaCl, 1 mM EDTA) containing 20 µg/ml RNase A and 1000 units/ml RNase T1 (Roche Molecular Biochemicals) for 30 min at 37 °C. The slides were washed at room temperature in 1, 0.5, and 0.2× SSC for 20 min each at 60 °C in 0.2× SSC for 60 min and at room temperature in 0.2× SSC and distilled water for 10 min each. Tissues were dehydrated in 50 and 70% 2-propanol, air-dried, and then exposed to x-ray film for 20-40 h. Autoradiographic detection of ³⁵S was performed with NTB-2 nuclear emulsion (Eastman Kodak). After an exposure time of 14 days, slides were developed. Bright and dark field microscopic analysis was performed using the Olympus AX-70 microscope.

Double Labeling ISHH-- Detection of two different RNA transcripts in the same tissue section was performed with radioactive- and nonradioactive-labeled probes as published previously with some modifications (34). A 678-bp digoxigenin-labeled riboprobe for rat VAChT (GenBank^TM accession number NM031663; nucleotides 1863-2540, 90% identity to mouse VAChT) was generated by in vitro transcription with a digoxigenin labeling mix containing 10 mM each of ATP, CTP, and GTP; 6.5 mM UTP, and 3.5 mM dioxigenin-11-UTP (Roche Molecular Biochemicals). For VMAT2 mRNA detection, a 269-bp DNA fragment of the rat VMAT2 cDNA corresponding to nucleotides 233-500 (GenBank^TM accession number NM013031) was subcloned into pCRII (Invitrogen). Transcription with SP6 yielded probes in antisense orientation (identity to mouse VMAT2 = 94%). After hydrolysis, probes were purified by sodium acetate precipitation and then added to the appropriate radioactive hybridization solution to a final concentration of 0.1 µg/ml. Hybridization and washing procedures were performed as described above. For the detection of nonradioactive hybrids, slides were equilibrated to buffer A (100 mM Tris and 150 mM NaCl, pH 7.5) containing 0.05% Tween 20 (Merck). Blocking was performed by incubation for 1 h in blocking buffer (buffer A containing 2% normal lamb serum). Alkaline phosphatase-conjugated anti-digoxigenin Fab fragments (Roche Molecular Biochemicals) were diluted to 1 unit/ml in blocking buffer. After the slides were rinsed with buffer A, the diluted antibody was applied for 1 h at room temperature. Excessive antibody was removed by two 15-min washes in buffer A. Slides were equilibrated in buffer B (100 mM Tris and 100 mM NaCl, 50 mM MgCl₂, pH 9.4) containing 0.05% Tween 20, and then the color reaction took place in buffer B containing 0.2 mM 5-bromo-4-chloro-3-indolyl phosphate and 0.2 mM nitro blue tetrazolium salt (Roche Molecular Biochemicals). The reaction was stopped by washing the slides in double-distilled water for 24 h. For detection of ³⁵S-labeled probes, slides were covered with K5 photoemulsion (Ilford) diluted 1:1 in water for 14 days in the dark.

Immunohistochemistry-- Mice and rats were anesthetized with ketamine/xylazine and transcardially perfused with PBS containing procaine-HCl (5 g/liter) followed by Bouin Hollande fixative as described (35, 36). The perfused brains were removed, dissected into anterior, middle, and posterior parts, and post-fixed for 24-48 h in Bouin Hollande fixative. After dehydration in a graded series of 2-propanol solutions, tissues were embedded in Paraplast Plus (Merck). Deparaffinized serial sections were subjected to antigen retrieval by heating at 92-95 °C for 15 min in 0.01 M citrate buffer (pH 6). Nonspecific binding sites were blocked with 5% BSA in PBS followed by an avidin-biotin blocking step (avidin-biotin blocking kit, Boehringer Ingelheim). For single peroxidase immunostaining, sections were alternately incubated with the primary affinity-purified polyclonal rabbit (1 µg/ml) or polyclonal guinea pig (diluted 1:4000) antibodies raised against VGLUT3 overnight at 18 °C followed by an additional incubation for 2 h at 37 °C. Polyclonal rabbit or guinea pig antiserum against VGLUT1 (diluted 1:4000) and VGLUT2 (diluted 1:8000) was used to identify the two dominant excitatory systems in the brain in adjacent sections. A polyclonal sheep antiserum against tyrosine hydroxylase (diluted 1:1000; Chemicon) was applied to identify catecholaminergic cell groups and terminals. Polyclonal rabbit (37) or guinea pig (Chemicon) antibodies against VMAT2 (diluted 1:2000) were applied on adjacent sections to detect serotoninergic and histaminergic cell groups and terminals in addition to catecholaminergic cell groups and fibers/terminals stained for TH (22). After several washes in distilled H₂O followed by rinsing in 50 mM PBS, species-specific biotinylated secondary antibodies (Dianova) were applied for 45 min at 37 °C. Following another series of washes, sections were incubated for 30 min with the AB reagents (Vectastain Elite ABC Kit). Immunoreactions were visualized with 3',3-diaminobenzidine (Sigma) enhanced by the addition of 0.08% ammonium nickel sulfate (Fluka) resulting in a dark blue staining. The specificity of VGLUT3 immunostaining was tested by preadsorption with the antigens used as the immunogen (1 µM final concentration).

Confocal Laser Scanning-- Confocal laser scanning double immunofluorescence analysis was performed as described previously (38). The specificity of VGLUT3 immunostaining obtained with the crude and immunity-purified antisera and with the guinea pig antiserum was tested by preadsorption with the peptide used as the immunogen (10 µM) and with the fusion protein (10 µM), respectively. Immunoreactions obtained were preabsorbed with the respective homologous antigen (data not shown). For double labeling, sections were incubated overnight at room temperature with a mixture of polyclonal affinity-purified rabbit (4.5 µg/ml) or guinea pig VGLUT3 (diluted 1:600), the polyclonal rabbit or guinea pig VGLUT1, and VGLUT2 antibodies (diluted 1:400 or 1:800), the polyclonal rabbit or guinea pig VMAT2 antibodies (diluted 1:200), a polyclonal goat antibody (diluted 1:400), and a polyclonal rabbit antibody (diluted 1:150) against the vesicular acetylcholine transporter (VAChT) (39). TH was detected with the polyclonal sheep antibody diluted 1:100. Immunoreactivities were visualized with indocarbocyanine (Cy3)-conjugated species-specific secondary antibodies (Dianova) diluted 1:200 and applied for 45 min at 37 °C, resulting in a red-orange fluorescence labeling or with biotinylated IgG (Dianova) diluted 1:200, applied for 45 min at 37 °C, followed by incubation with Alexa 488-conjugated streptavidin diluted 1:200 (MoBiTec) for 2 h at 37 °C, resulting in a green fluorescence. Furthermore, double visualization was achieved by combining indodicarbocyanine (Cy5)-labeled species-specific secondary antisera diluted 1:200 (Dianova) with species-specific secondary biotinylated antisera diluted 1:200 and Alexa 488-conjugated streptavidin diluted 1:200 (MoBiTec), or by combining Alexa 488-labeled species-specific secondary antisera (MoBiTec) with species-specific secondary biotinylated antisera and Alexa 647-conjugated streptavidin diluted 1:200 (MoBiTec). Sections were analyzed with the Olympus Fluoview confocal laser scanning microscope (Olympus Optical Co.) and documented as false color confocal images.

Primary Raphé Nuclei Cultures and Double Fluorescence Deconvolution Microscopy-- Rat embryos (E15 ± 1 day) were obtained from pregnant Sprague-Dawley females, and the brainstem region containing both rostral and caudal raphé nuclei was removed, and serum-free cultures were prepared by standard procedures (40, 41). Dissected tissue was placed in Hepes-buffered Ringer solution (130 mM NaCl, 4 mM KCl, 1 mM MgCl₂, 1.5 mM CaCl₂, 10 mM Hepes, 10 mM dextrose, 3 mM NaOH, pH 7.3) and then digested for 15 min with papain solution (Hepes-buffered Ringer solution with 1.5 mM CaCl₂, 0.5 mM EDTA, 9 units/ml papain, 0.2 mg/ml cysteine). Digested tissue was washed 2 times in trituration solution (10% fetal bovine serum in DMEM with 5 units/ml DNase I) and then triturated mechanically 3 times by fire-polished Pasteur pipettes, centrifuged at 1000 rpm for 10 min at 4 °C, redissociated in starter medium (serum-free neurobasal medium supplemented with B27, 0.5 mM glutamine, 25 µM glutamate), counted, and plated in a 2-well chamber slide at a density of 20,000 cells per cm² in cultivating medium (starter medium without glutamate). Cells were plated in two-chamber glass slides precoated with polyornithine (4 µg/ml; Sigma) and analyzed between 2 and 10 days in vitro (DIV) for expression of neurochemical phenotypic markers. Cultures were maintained at 37 °C in a 5% CO₂ atmosphere and fed by replacing half of the medium on day 4 and each day thereafter.

For immunostaining, the culture medium was removed from the cells on ice and directly fixed with ice-cold acetone (9 min), rinsed with PBS, and incubated with blocking buffer (0.1% Triton, 1% BSA, 5% normal goat serum (Sigma)) in PBS at room temperature for 1 h. All subsequent steps were conducted at room temperature. The primary antibodies used are as follows: affinity-purified guinea pig VGLUT3 (2.5 µg/ml), affinity-purified rabbit VMAT2 (2.5 µg/ml), and mouse anti-tryptophan hydroxylase (WH3; Sigma) (1:1000) to identify serotoninergic cell groups and varicosities (42). Secondary antibodies used were directly coupled to Alexa Fluorophores 488 or 594 (Molecular Probes). Co-localization of red and green signals was not due to bleed-through due to the small overlap in wavelength spectra between Alexa 488/594 as other phenotypes (TrypH-positive, VGLUT3-negative cells; TrypH-negative, VGLUT3-positive cells) are clearly present in this raphé brainstem culture. Images were obtained from a Leica DVRA2 deconvolution fluorescence microscope and deconvolved using SlideBook deconvolution software (Intelligent Imaging Innovations) with a "nearest neighbors" deconvolution algorithm. All images were obtained from approximately the middle plane of the cells.

Transient Infection//Transfection of PC12 Cells-- Rat PC12 cells were maintained at 37 °C in an atmosphere of 90% air, 10% CO₂ in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum, 5% heat-inactivated horse serum, penicillin (100 units/ml), streptomycin (50 µg/ml), and glutamine (4 mM). PC12 cells (6 × 10⁶) were plated in 10-cm dishes precoated with polyornithine (4 µg/ml). The following day, the cells were rinsed with PBS and infected with a recombinant vaccinia virus encoding bacteriophage T7 RNA polymerase at a multiplicity of infection of 20 for 2 h as described (18). The medium was removed, and the cells were transfected with DMEM (2 ml) containing T7 promoter-bearing plasmid cDNA (2 µg/ml) and LipofectAMINE 2000 (1:100) for 6 h, and fresh medium (5 ml) without serum was added. After 16-18 h the cells were harvested and vesicle membranes prepared. Transfection efficiency (>90%) was generally monitored in a parallel 10-cm dish of PC12 cells transfected with the lacZ gene using a histochemical stain for -galactosidase as described (18, 43).

Preparation of Vesicle Membranes from PC12 Cells-- Control and VGLUT3-expressing PC12 cells (~20 plates each) were rinsed with PBS, collected by scraping, rinsed again in PBS, and homogenized with a ball-bearing device (11 µM clearance, 20 strokes) in 3 ml of ice-cold buffer containing 0.32 M sucrose, 10 mM Hepes, pH 7.4, 1 mM phenylmethylsulfonyl fluoride, 6 µg/ml leupeptin, 5 µg/ml aprotinin, and 5 µg/ml pepstatin. The resulting homogenates were cleared by successive centrifugation at 2000 × g for 10 min and 10,000 × g for 20 min to remove nuclei, mitochondria, and cell debris. The supernatant was sedimented by centrifugation at 400,000 × g for 45 min. The resulting membrane pellet was resuspended in 2 ml of 0.32 M sucrose, 10 mM Hepes, pH 7.4, containing protease inhibitors. Protein was measured by the Bradford assay (Bio-Rad) using bovine serum albumin as the standard.

Vesicular L-[³H]Glutamate Transport Assay-- For L-[³H]glutamate uptake assay, aliquots (100 µl) of membranes containing 100 µg of protein were mixed with uptake buffer (50 µl) containing 110 mM potassium tartrate, 20 mM Hepes, pH 7.4, in the presence or absence of Cl ion or non-substrate inhibitors and incubated at 32 °C for 2 min. Following preincubation, a 4× solution (50 µl) containing 20 mM Mg²⁺-ATP (neutralized with KOH to pH 7.4) and 200 µM glutamate plus 1 µM L-[³H]glutamate (42.9 Ci/mmol, PerkinElmer Life Sciences) in the presence and absence of various inhibitors were added. The final concentration of MgATP was 5 mM and glutamate was 50 µM. For kinetic analysis, various concentrations of unlabeled glutamate (312 µM to 10 mM) were added to the radiolabeled mix, and uptake was terminated after 3 min. For competition experiments, unlabeled amino acids or neurotransmitters were added at 4× (with appropriate adjustments in osmolarity) with MgATP and L-[³H]glutamate. For each condition, membranes from mock-transfected PC12 cells and PC12 cells expressing VGLUT3 were always analyzed in parallel. The uptake reactions were stopped by placing tubes in ice water, and the samples vacuum were filtered through GF/F glass fibers and washed with 6 ml of ice-cold uptake buffer containing 10 mM MgSO₄. Radioactivity bound to the filters was solubilized in 1 ml of 1% SDS followed by addition of 8 ml of EcoScint scintillation fluid and quantitated by liquid scintillation counting. Experiments performed in duplicate or triplicate were repeated two to four times. K_m values were determined by nonlinear regression (Prism3; Graphpad Software).

Synaptic Vesicle Purification-- Synaptic vesicles were purified from rat forebrain using a modified procedure of Jahn and co-workers (44). Briefly, 15 g of mouse whole brain was frozen and pulverized in liquid nitrogen to a fine powder. The powder was resuspended in 100 ml of sucrose buffer containing 0.32 M sucrose, 10 mM Tris-HCl, pH 7.4, 1 mM phenylmethylsulfonyl fluoride, 6 µg/ml leupeptin, 5 µg/ml aprotinin, and 5 µg/ml pepstatin and homogenized in a tight-fitting glass-Teflon homogenizer (9 strokes, 900 rpm). The homogenate was centrifuged for 10 min at 2000 × g and then at 47,000 × g followed by centrifugation at 120,000 × g for 2 h. The supernatant (20 ml) was layered onto a cushion (5 ml) of 0.6 M sucrose, 10 mM Tris-HCl, pH 7.4, and centrifuged at 260,000 × g for 2 h. The pellet was resuspended in 3 ml of sucrose buffer (without inhibitors) and cleared by centrifugation at 27,000 × g for 10 min.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Structural Characteristics of a Homologue of VGLUT1 and VGLUT2-- The open reading frame of the mouse cDNA clone contains a predicted protein sequence of 601 amino acids. Sequence identity of this protein to VGLUT1 (rat and human) and VGLUT2 (mouse, rat, and human) is ~71% (black balls in Fig. 1). Secondary structure predictions of VGLUT1 and VGLUT2 have included 8-10 putative transmembrane domains (TMDs) (10, 11, 16), but these are not as easy to model as VMATs and VAChT which are predicted to have 12 TMDs (45). In a 10-TMD model, predicted by hydropathy profile analysis as shown in Fig. 1, each membrane spanning region is contained within a single exon determined from the presence of conserved sequences of the mouse homologue within the contigs of VGLUT1, VGLUT2, and the human orthologue from the human genome project (NCBI Blast). TMD VI (16) and the highly conserved stretch of hydrophobic amino acids (containing a glutamic acid residue) in the lumenal loop between TMD VII and VIII may be too short to cross the bilayer and instead be re-entrant loops as are known to occur in plasma membrane glutamate transporters (46, 47) and receptors (48). A highly conserved glycosylation site is predicted to lie in the intraluminal loop between TMDs I and II, a characteristic feature of vesicular transporters such as VAChT and VMATs. At least two highly conserved charged amino acid residues are predicted to lie within TMDs of the VGLUTs. In the mouse homologue these include Arg¹⁰⁶ in TMD 1 and Arg³³⁹ in TMD VII. A membrane-embedded arginine residue might directly participate in the binding of glutamate as has been demonstrated for plasma membrane glutamate transporters (49) and receptors (50). Numerous consensus site sequences for phosphorylation by various kinases also exist (not shown). Extensive divergence between the sequences of the VGLUT isoforms is apparent in the N and C termini that are predicted to face the cytoplasm. Within these regions the sequence identity of the mouse homologue with VGLUT2 is somewhat greater than with VGLUT1. Sequences used to generate antibodies against this homologue were therefore carefully selected and included the 15 terminal amino acids in the N terminus (* in Fig. 1) for generation of antibodies in rabbits and the 56 terminal amino acids in the C terminus (** in Fig. 1) for generation of antibodies in guinea pigs. Sequences used for antibody production have little or no sequence identity with VGLUT1 or VGLUT2 (gray). The N terminus (amino acids 1-15) is identical to the human homologue predicted from the DNA sequence on chromosome 12 (GenBank^TM accession number AJ459241; Locus, HSA459241); however, the C terminus displays only 65% identity within the terminal 56 amino acids.

View larger version (40K): [in this window] [in a new window]   Fig. 1.   Primary amino acid sequence and predicted secondary structure of mouse VGLUT3. Ten putative transmembrane domains (TMD I-X) and potential sites for N-linked glycosylation (pronged branches) are indicated. Black indicates amino acids that are identical between VGLUT3, VGLUT2 (mouse, rat and human), and VGLUT1 (rat and human). Gray indicates amino acids conserved between mouse VGLUT2 and VGLUT3. White indicates amino acids unique to VGLUT3. * and ** denote the end of the N-terminal VGLUT3 peptide and the beginning of the C-terminal VGLUT3-GST fusion protein used for antibody production, respectively.

Mouse Homologue Functions as a Vesicular Glutamate Transporter (VGLUT3)-- We utilized a transient transfection procedure in rat PC12 cells to produce high levels of VGLUT3 in light membrane fractions as described previously for rat VGLUT2 (18). When postnuclear homogenates were prepared and analyzed by Western blot, a predominant band of the expected molecular weight was observed in PC12 cells transfected with this cDNA and not from mock-transfected cells (see Fig. 3B). Uptake of L-[³H]glutamate (50 µM) was examined in enriched vesicle preparations from PC12 in the presence of Mg²⁺-ATP (5 mM) and KCl (4 mM), conditions known to be optimal for glutamate transport by synaptic vesicles isolated from brain (8, 9). Glutamate uptake into vesicles from VGLUT3-transfected cells was generally two to three times greater than that seen with mock-transfected controls (Fig. 2, A and C). The initial rate of glutamate uptake (Fig. 2B) measured during the linear portion of the time course (3 min) was saturable with an apparent K_m of 1.3 mM (n = 2). The apparent affinity (K_m) of VGLUT3 for glutamate was similar to the other isoforms and well below the concentrations (10 mM) expected in glutamatergic neurons and in many characterized neurons expressing VGLUT3 in vivo (see below). Uptake was dependent upon exogenous ATP and was abolished by bafilomycin and N-ethylmaleimide, which are specific inhibitors of the vacuolar-type H⁺-ATPase (data not shown). Low, physiologically relevant concentrations of Cl enhanced the VGLUT2-mediated uptake of glutamate by ~200-300% (Fig. 2C), which is also a characteristic feature of VGLUT1 and VGLUT2. Vesicular glutamate uptake was blocked by the proton ionophore FCCP (50 µM) indicating that transport is dependent upon the H⁺-electrochemical gradient (Fig. 2C). Inclusion of Cl or FCCP with mock-transfected samples showed no effect (Fig. 2C). Transport observed in transfected cells was somewhat selective for L-glutamate, compared with the D-isoform, and GABA (10 mM), acetylcholine (1 mM), and serotonin (50 µM) did not interfere with the uptake of L-[³H]glutamate (Fig. 2D). Glutamate uptake was inhibited by L- and D-aspartate when present at 10 mM (Fig. 2D) but not at 5 mM (data not shown). At 5 mM, the L- and D-isoforms of glutamate inhibited [³H]glutamate uptake by 75 and 58%, respectively. Because VGLUT3 greatly prefers glutamate to aspartate, as do VGLUT1 and VGLUT2, vesicular loading and release of aspartate at nerve terminals may operate by a unique mechanism. Uptake was significantly reduced in the presence of 2 µM trypan blue or Rose Bengal, inhibitors of vesicular glutamate uptake in synaptic vesicles (51, 52). These results provide compelling evidence that this mouse homologue of VGLUT1 and VGLUT2 functions as an H⁺-dependent vesicular transporter highly specific for glutamate, and we propose to name the protein VGLUT3.

View larger version (36K): [in this window] [in a new window]   Fig. 2.   Vesicular glutamate transport by VGLUT3 in transiently transfected PC12 cells. A, time course of glutamate (50 µM) uptake in preparations of light membranes from VGLUT3-transfected () and mock-transfected () PC12 cells. B, saturation analysis of glutamate uptake (300 µM to 10 mM) by VGLUT3. VGLUT3-specific uptake (mock-subtracted) is represented and saturates at 3-4 mM. Inset, Lineweaver-Burk analysis of the initial velocity determines a Km of 1.3 mM. C, transport is stimulated by Cl ion and dependent upon the H+ electrochemical gradient. Uptake of glutamate (50 µM) was measured in mock-transfected (open bars) and VGLUT3-transfected (gray bars) PC12 membrane preparations in the absence (Cl) or in the presence of 4 mM KCl (+Cl) or 4 mM KCl + 50 µM carbonyl cyanide p-trifluoromethoxyphenylhydrazone (+Cl + FCCP). Mock values are unchanged by either treatment. D, VGLUT3 is specific for glutamate. Uptake of glutamate (50 µM, 5 min) was measured in 4 mM Cl-containing medium in the absence or presence of 10 mM L- or D-glutamate (Glu), L- or D-aspartate (Asp), or GABA, 50 µM serotonin (5-HT), 1 mM acetylcholine (ACh) or 2 µM Rose Bengal (RB), or 2 µM trypan blue (TB). Con, control.

Antibody Specificity and Expression of VGLUT3 on Synaptic Vesicles-- Transfected cells expressing high levels of the three VGLUT isoforms were used to demonstrate complete specificity of the immunoreaction provided by VGLUT3 polyclonal antibodies used in the immunohistochemical analysis in vivo and the immunocytochemical analysis in vitro. Here, constructs encoding rat VGLUT1, rat VGLUT2, or mouse VGLUT3 cDNA were used. CV-1 fibroblasts expressing VGLUT1, VGLUT2, or VGLUT3 and mock-transfected cells were examined by immunocytochemical staining with both rabbit and guinea pig anti-VGLUT3 antibodies. Specific immunostaining was observed by VGLUT3 antibodies only in cells that expressed VGLUT3 indicating that little background was present and no cross-reactivity with VGLUT1 or VGLUT2 was observed (Fig. 3A). Homogenates from PC12 cells expressing VGLUT1, VGLUT2, or VGLUT3 and mock-transfected cells were also examined by Western blot (Fig. 3B). Both the N-terminal anti-peptide antibody and the C-terminal fusion protein antibody labeled a broad band of ~55 kDa, corresponding to VGLUT3. The higher molecular weight species that were observed with all VGLUT antibodies may be a result of dimerization or due to nonspecific aggregation or processing in vaccinia-infected cells. VGLUT3 antibodies are isoform-specific as demonstrated by the absence of signal in VGLUT1- and VGLUT2-expressing lanes. Specific VGLUT3 immunoreactivity was lost if 1 µM peptide or fusion protein was included with the primary antibodies during the incubation (data not shown). Similarly, our guinea pig antibodies directed against VGLUT1 and VGLUT2 do not recognize VGLUT3 (Fig. 3B). Identical results were observed using our rabbit VGLUT1 and VGLUT2 antibodies (data not shown, see Ref. 18).

View larger version (89K): [in this window] [in a new window]   Fig. 3.   Antibody specificity and synaptic vesicle localization. A, mock-transfected (mock), rat VGLUT1 (1), rat VGLUT2 (2), mouse VGLUT3 (3)- expressing PC12 cells were immunostained with guinea pig anti-VGLUT3 C-terminal fusion protein (1:4000; top) or with rabbit anti-VGLUT3 N-terminal peptide (1:4000; bottom). Ab, antibody. B, quadruplicate Western blots of VGLUT1, VGLUT2, and VGLUT3-expressing PC12 cell homogenates were probed with guinea pig antisera directed against a C-terminal fusion protein from VGLUT1 (1:4000), VGLUT2 (1:4000), or VGLUT3 (1:10,000) and an affinity-purified rabbit antisera (0.6 µg/ml) directed against an N-terminal peptide from VGLUT3. Each antiserum only detects the appropriate isoform. C, purified synaptic vesicles from mouse forebrain blotted and probed with VGLUT3 guinea pig antiserum (1:10,000) revealed a single protein band () that was not seen in the presence of 1 µM GST-VGLUT3 (C terminus) fusion protein (+).

To determine whether VGLUT3 is present on brain synaptic vesicles, we examined highly purified synaptic vesicle preparations from whole mouse brain by Western blot using the guinea pig VGLUT3 antiserum (Fig. 3C). A single VGLUT3 immunoreactive band (~60 kDa) required a considerably longer exposure under film (1 h) to observe than did VGLUT1 and VGLUT2 bands (<min exposure) in contrast to the identical exposure time used in the blots shown in Fig. 3B. This suggests that the abundance of VGLUT3 containing synaptic terminals in the brain is low. Specificity of the VGLUT3 immunoreaction was confirmed by the absence of labeling when samples were preincubated with 1 µM fusion protein (Fig. 3C).

VGLUT3 mRNA Transcripts by Northern Blot-- A single ~4.9-kb mRNA transcript hybridizing to VGLUT3 was detected in whole mouse brain preparations of poly(A)⁺ mRNA in early postnatal development with somewhat higher levels at postnatal day 7 (P7) compared with adult whole mouse brain (Fig. 4). In contrast, VGLUT1 transcripts were relatively low between E18 and P7, and VGLUT2 appeared relatively constant. VGLUT3 transcripts were extremely low in abundance (4 oligonucleotides/6 days under PhosphorImager plate) compared with VGLUT1 (several hours) or VGLUT2 (overnight exposure) where bands could be clearly visualized using only one radiolabeled oligonucleotide. No signals were detected in mRNA prepared from kidney or testis, but a VGLUT3-related mRNA species of similar size and abundance was present in liver (data not shown).

View larger version (90K): [in this window] [in a new window]   Fig. 4.   Northern analysis of VGLUT expression in developing whole brain. The expression of VGLUT1, VGLUT2, and VGLUT3 mRNA in whole mouse brain at embryonic day 18 (E18) and postnatal ages (P1, P7, P14, P26, P60, in days) reveals a differential distribution pattern. For VGLUT3, an ~4.9-kb transcript is observed at all ages with somewhat higher levels seen at P7. The position of the 28 S ribosomal band is shown on the right.

Distribution of VGLUT3 mRNA in the Rodent Brain-- For the cellular localization of VGLUT3 gene expression in the mouse central nervous system, a specific ³⁵S-labeled riboprobe (205 bp) was used that was complementary to the non-conserved region of the C terminus and a portion of the 3'-untranslated sequence of mouse VGLUT3 cDNA. Longer probes for VGLUT3 containing putative TMDs were not found suitable for a specific localization of VGLUT3 transcripts due to long stretches of identical amino acids shared with both VGLUT1 and VGLUT2. The lack of any hybridization signal using the corresponding sense probes further underlines the specificity of the hybridization experiments with our VGLUT3-specific probes (data not shown).

After a 20-h exposure to x-ray film, the VGLUT3 probes produced specific hybridization signals with a unique neuronal distribution pattern that was different from that reported previously (19, 20) for VGLUT1 and VGLUT2 mRNA. For the characterization of mRNA distribution at the cellular level, nuclear emulsion-coated slides developed after 2-3 weeks were examined under the light microscope (Figs. 5 and 6). Glial cells, including astrocytes and microglia, did not appear to be labeled with the VGLUT3-specific probes.

View larger version (112K): [in this window] [in a new window]   Fig. 5.   Mapping of VGLUT3 mRNA expression in coronal sections through the mouse brain. Dark field photographs of emulsion-coated autoradiograms in rostral to caudal orientation (A-F) that were hybridized with a 35S-labeled riboprobe specific for mouse VGLUT3 mRNA illustrate the unique expression pattern of the VGLUT3 gene in distinct brain regions. Numerous scattered neurons were observed in the neostriatum and ventral pallidum (A and B). Note the highest labeling intensity in the mesencephalic raphé nuclei (E). Lower levels of VGLUT3 mRNA was expressed by thalamic (C and D), hypothalamic (D), and brainstem nuclei including the NTS, the raphé obscurus ncl., and the spinal trigeminal ncl. (F). ac, anterior commissure; aq, aqueduct; BST, bed nucleus of stria terminalis; cc, corpus callosum; CA1, field CA1 of Ammon's horn; CA3, field CA3 of Ammon's horn; CBL, cerebellum; CM, central medial thalamic ncl.; CPu, caudate putamen; CTX, neocortex; DG, dentate gyrus; DR, dorsal raphé ncl.; GP, globus pallidus; LHA, lateral hypothalamic area; MnR, median raphé ncl.; NTS, nucleus of the solitary tract; Sp5, spinal ncl. of the trigeminal tract; SON, supraoptic ncl.; 3V, third ventricle; PVA, anterior periventricular thalamic ncl.; Rob, ncl. raphé obscurus; VP, ventral pallidum; VPM, medial ventroposterior thalamic ncl.; Xi, xiphoid thalamic ncl. Exposure time of autoradiograms, 2 weeks.

View larger version (38K): [in this window] [in a new window]   Fig. 6.   Localization of VGLUT3 mRNA expressing neurons in discrete brain nuclei. High resolution dark field photomicrographs illustrating the distribution of VGLUT3 mRNA expressing neurons in coronal sections through the basal forebrain (bregma 0.6 mm) (A), the hippocampal formation (B), the posterior hypothalamus (bregma 1.5 mm) (C), and the lower brainstem (bregma 7.4 mm) (D). Relatively large neurons expressing the highest levels of VGLUT3 mRNA, in comparison to other VGLUT3-positive neurons, are scattered throughout the striatum and ventral pallidum (A). The medial subdivision of the anterios bed ncl. of stria terminalis contains many densely packed VGLU T3 mRNA-positive neurons. In the hippocampal formation VGLUT3 mRNA-positive neurons are preferentially located and most numerous in the CA3 and hilus region (B). Note that pyramidal neurons and granule cells of the dentate gyrus do not express VGLUT3 mRNA. In hypothalamus neurons weakly labeled for VGLUT3 mRNA are scattered close to the periventricular regions and in the lateral hypothalamic area as well (C). Many neurons located in the medial subdivision of the NTS exhibit moderate hybridization signals for VGLUT3 mRNA (D). Note the absence of VGLUT3 mRNA in the hypoglossal ncl. (12) and in the area postrema (AP). Asterisk in D labels the aqueduct. ac, anterior commissure; aq, aqueduct; AP, area postrema; BST, bed nucleus of stria terminalis; CA1, field CA1 of Ammon's horn; CA3, field CA3 of Ammon's horn; CPu, caudate putamen; DG, dentate gyrus; LHA, lateral hypothalamic area; NTS, nucleus of the solitary tract; 3V, third ventricle; 12, hypoglossal ncl. Scale bar, A-D, 200 µm.

VGLUT3 mRNA expression was restricted to very few brain areas. Highest levels of VGLUT3 mRNA were observed in scattered neurons of the basal forebrain (Fig. 5, A and B) the hippocampal formation (Fig. 5, C and D), and in the majority of neurons of the mesencephalic raphé nuclei (Fig. 5E). Low to moderate VGLUT3 mRNA expression levels were observed in the neocortex (Fig. 5, B-D), the posterior hypothalamus (Fig. 5D), and few brainstem nuclei including the ncl. of the solitary tract (NTS), the parabrachial ncl., the spinal trigeminal ncl., and the caudal raphé nuclei (Fig. 5F). In the forebrain, numerous strongly labeled cells scattered throughout the caudate putamen and the ventral pallidum were detected (Fig. 5, A and B). The distribution pattern and the relatively large size of 25-50 µm of VGLUT3 mRNA expressing neurons in the caudate-putamen suggest that they are cholinergic interneurons (see below). Apart from the neostriatum, the bed nucleus of stria terminalis contained numerous VGLUT3 mRNA-positive perikarya densely packed within its medial anterior subdivision (Figs. 5B and 6A). A band of weakly labeled VGLUT3 mRNA expressing neurons was observed in the frontal, parietal, and temporal cortex that was located in the lower lamina IV and upper lamina V (Fig. 5, B-D) suggesting the presence of glutamatergic excitatory interneurons in neocortex. In the hippocampal formation, scattered VGLUT3 mRNA-positive neurons were predominantly found in the stratum radiatum and oriens of CA3 and CA1 region as well as in the hilus area of the dentate gyrus (Figs. 5, C and D, and 6B). Single neurons scattered throughout the pyramidal and granule cell layers were also detected strongly suggesting the expression of VGLUT3 by hippocampal interneurons with excitatory function. Pyramidal neurons of the Ammon's horn and granule cells of the dentate gyrus did not express VGLUT3 mRNA. In the diencephalon, cells expressing weak hybridization signals for VGLUT3 were observed in thalamic midline nuclei such as the paraventricular, the intermediodorsal, central medial, and reuniens and xiphoid thalamic ncl. (Fig. 5, C and D). In the hypothalamus, scattered VGLUT3 mRNA-positive neurons were detected in the periventricular region of the third ventricle, in the medial subdivision and in the lateral hypothalamic area (Fig. 6C). The highest VGLUT3 mRNA expression levels and highest density of VGLUT3 mRNA expressing cells were observed in the mesencephalic raphé nuclei especially in the dorsal and median raphé ncl. (Fig. 5E). Some VGLUT3 mRNA expressing neurons were located in the interpeduncular ncl. Other brainstem areas exhibiting VGLUT3 mRNA expression included the nucleus of the solitary tract (NTS) (Fig. 6D). Here VGLUT3 mRNA-expressing cells were exclusively located in the medial subdivision, an area where serotonin-immunoreactive neurons were described previously (53). VGLUT3 mRNA could not be detected in the area postrema. Singular VGLUT3 mRNA-containing neurons were present in the caudal medulla particularly scattered through the spinal trigeminal nucleus (Fig. 5F) and the reticular formation. Neurons in the spinal cord did not express VGLUT3 mRNA (data not shown).

General Pattern of VGLUT3 Immunoreactivity in Rodent Brain-- The vast majority of VGLUT3 immunoreactivity was found in a diffuse system of beaded varicose nerve fibers exhibiting low to moderate densities throughout the forebrain and hindbrain in both mouse and rat brains. Glial cells, ependymal cells, and endothelial cells did not exhibit specific staining for VGLUT3 (data not shown). Two types of varicose fiber staining for VGLUT3 were found: 1) thin varicose fibers with small varicosities, and 2) somewhat thicker varicose fibers with large varicosities. Typically, strands of varicose fibers of variable length and thickness of varicosities and corresponding cross-profiles were encountered. VGLUT3-positive varicosities, however, displayed strong staining indicating that levels of VGLUT3 in a given neuronal process may be quite high.

The VGLUT3 System Is Distinct from VGLUT1 and VGLUT2 Pathways in Rodent Brain-- In order to answer the question whether VGLUT3-operated synapses in vivo represent a novel separate entity of the glutamatergic phenotype distinct from the VGLUT1 and VGLUT2 pathways, alternate staining for the three VGLUT isoforms on adjacent sections and confocal double immunofluorescence for VGLUT3 and VGLUT1 or VGLUT2 was performed. This revealed that VGLUT3 immunostaining was much less abundant than and clearly separated from staining for VGLUT1 and VGLUT2, respectively, throughout the forebrain, hindbrain, and cerebellum. Absence of VGLUT3 immunoreactivity from VGLUT1 or VGLUT2 synapses is exemplarily demonstrated for the cerebral cortex (see Fig. 10, A-F). Thus, the VGLUT3 system is a glutamatergic entity on its own and neither part of the VGLUT1 nor the VGLUT2 system.

Regional Patterns of VGLUT3 Immunoreactivity in Rodent Brain-- The hypothalamus was the only brain region where robust cell body immunostaining for VGLUT3 was detectable both in mouse and rat. VGLUT3 immunopositive cell groups accumulated especially in the lateral hypothalamus where also VGLUT3 mRNA-positive cells were found. These magnocellular VGLUT3-positive neurons were embedded between VGLUT3-positive varicose fibers (Fig. 7, A and C). In addition some single cells juxtaposed to the third ventricle midline region did also stain for VGLUT3. VGLUT3 cell body and varicose fiber staining was fully preabsorbed by the antigens against which the antisera were raised both in mouse and rat indicating the specificity of the immunoreactions. VGLUT3 immunoreactivity in the hypothalamus was separated from catecholaminergic TH-positive cell bodies and fibers (Fig. 7B). From continuity of VGLUT3-positive varicose fiber strands with VGLUT3-positive cell bodies, it appeared that a substantial part of the hypothalamic VGLUT3 innervation originates from these local neurons implicating the existence of a new intrinsic VGLUT3-operated glutamatergic hypothalamic system. Extrahypothalamic projection sites of these hypothalamic VGLUT3 neurons need to be identified.

View larger version (168K): [in this window] [in a new window]   Fig. 7.   Distribution of VGLUT3 in rat hypothalamus. The hypothalamus represents the only region with VGLUT3-immunoreactive-labeled cell bodies. Low resolution photographs illustrate that most of the VGLUT3-immunoreactive large cell bodies are located in the perifornical region of the lateral hypothalamic area (LHA in A). Few scattered neurons of smaller size are also present in the periventricular region of the third ventricle (3V in A). VGLUT3-immunoreactive neurons and fibers were negative for tyrosine hydroxylase (TH) (B). High magnification of the lateral hypothalamic area reveals magnocellular neurons of the lateral hypothalamic area expressing strong VGLUT3-immunoreactivity that are surrounded by VGLUT3-immunoreactive neuronal processes and long varicose fibers (C). The median eminence (ME) receives VGLUT3-immunoreactive input from VGLUT3-immunoreactive varicose fibers (D). Punctate VGLUT3-immunoreactivity is present mainly in the external layer of the median eminence, but few VGLUT3-immunoreactive puncta are also present in the internal layer. Scale bars, A and B, 200 µm; C, 50 µm; D, 20 µm.

In all other brain regions, VGLUT3 immunostaining was only present in varicose fibers and their cross-section profiles and was virtually undetectable in cell bodies in the specific regions where VGLUT3 mRNA was found. Although every brain region received at least some VGLUT3 innervation, as minimal as it may be, distinct differences in the density of VGLUT3 innervation to the various target areas were apparent. Thus, low to moderate numbers of VGLUT3-positive varicose fibers were seen throughout the frontal, parietal, temporal, and occipital cortex with highest densities in the cingulate, piriform, and entorhinal cortex. They occasionally formed close appositions to regional blood vessels. A low to moderate number of VGLUT3 terminals was seen in the hippocampus including CA1-3, dentate gyrus, and hilar regions.

Some selective preference of relatively dense VGLUT3 innervation to the limbic and basal forebrain system system was noted. The amygdala complex received VGLUT3 input with preferentially higher density in the central basomedial nucleus region than in other parts. The bed nucleus of the stria terminalis and the nucleus accumbens also contained considerable numbers of VGLUT3 varicose fibers.

In the basal ganglia, the densities of VGLUT3 varicose fibers varied from very low in the caudate putamen to low in the globus pallidum externum and moderate in globus pallidum internum and ventral pallidum.

In the thalamus, dorsal midline regions such as the paraventricular thalamic nucleus and adjacent areas contained higher densities of varicose fibers than other thalamic regions. Circumventricular organs such as the subfornical organ, the median eminence (Fig. 7D), and the area postrema received moderate VGLUT3 input.

In the mesencephalon, VGLUT3-positive varicosities and fibers were concentrated in the peri-aqueductal gray. The mesencephalic, pontine, and medullary serotoninergic raphé nuclei contained moderate numbers of VGLUT3-positive varicose terminals (see Fig. 9B). In contrast to VGLUT3 mRNA, VGLUT3 immunoreactivity was undetectable in cell bodies of raphé nuclei in vivo. The NTS and the parabrachial nucleus, autonomic brainstem regions where VGLUT3 mRNA expressing cell groups were found, had more VGLUT3 input than other brainstem areas such as the cholinergic peduncolopontine nuclei and the noradrenergic locus ceruleus. VGLUT3 varicose fibers were extremely sparse in the cerebellum. The spinal cord gray matter was supplied by low to moderately dense and diffusely distributed VGLUT3-positive varicose fiber strands.

The VGLUT3 System Coincides with Intrinsic Striatal Cholinergic Neurons but Is Distinct from Other Cholinergic Pathways-- By using double ISHH experiments with radiolabeled probes against VGLUT3 mRNA and digoxigenin-labeled probes against the VAChT as a marker for cholinergic striatal interneurons, we could demonstrate that VGLUT3 mRNA and VAChT mRNA were co-expressed (Fig. 8, A and B), indicating that cholinergic striatal interneurons have a glutamatergic phenotype. This is underscored by the immunocytochemical co-localization of VGLUT3 in some VAChT-positive varicose fibers in the striatum (Fig. 8, C-E). However, given the overall paucity of VGLUT3 innervation, VGLUT3 immunoreactivity was only present in a very minor subpopulation of VAChT-positive varicose fibers in the striatum. This indicates that the glutamatergic synaptic phenotype in the striatum may be rather cryptic as compared with the abundant cholinergic synaptic phenotype.

View larger version (87K): [in this window] [in a new window]   Fig. 8.   Co-existence of VGLUT3 and VAChT in cholinergic striatal interneurons. High power micrograph of an emulsion-dipped autoradiogram illustrating the co-expression of VGLUT3 RNA transcripts (detected with radiolabeled probes) and the VAChT mRNA in two cholinergic interneurons of caudate-putamen (A). Cholinergic projection neurons located in the area of the diagonal band of Broca do not express VGLUT3 mRNA (B). Ultra-thin sections (6 µM) showing VGLUT3-immunoreactivity (C) coexists with VAChT-immunoreactivity (D) in varicose fibers that are relatively sparse and represent a minor subset of the abundant cholinergic innervation in caudate putamen (E). Detection of VGLUT3 in Alexa 488 mode (green) and VAChT in CY5 (red) excludes false co-positives because of 100% separation of wavelength spectrum. Scale bars, A and B, 20 µm. DIG, digoxigenin.

In contrast, VGLUT3 mRNA was absent from all other VAChT mRNA expressing cholinergic cell groups in the basal forebrain including the diagonal band of broca and the basal nucleus of Meynert (Fig. 8B). Consequently, VGLUT3 immunoreactivity was absent from cholinergic synapses in the cortex (see Fig. 10, M-O) and hippocampus, the known projection areas of the basal forebrain cholinergic system. Furthermore, cholinergic VAChT-positive motoneurons and other cholinergic cell groups in the brainstem areas were also devoid of positive hybridization signals for VGLUT3. In concordance, VGLUT3 immunoreactivity was absent from VAChT-positive synapses originating from the brainstem cholinergic projection systems. Thus, VGLUT3 has a unique expression preference to the cholinergic striatal interneuron and is absent from the cholinergic projection systems throughout the brain. Nonetheless, scattered VGLUT3 mRNA expressing neurons that were negative for VAChT were observed in cholinergic basal forebrain areas (Fig. 5, A and B) in conjunction with high numbers of VGLUT3-positive/VAChT -negative varicose fibers.

The VGLUT3 System Coincides with a Subpopulation of Serotoninergic Synapses but Is Distinct from Catecholaminergic Synapses-- The expression patterns of VGLUT3 in brainstem raphé neurons revealed by ISHH strongly suggest that the rostral serotoninergic neurons originally described by Dahlström and Fuxe (54) as areas B7 to B9 express VGLUT3 and therefore are of a mixed glutamatergic/serotoninergic phenotype. Hybridization on adjacent sections and double ISHH experiments on identical sections with probes against the vesicular monoamine transporter 2 (VMAT2) and TH revealed an overlapping distribution pattern of VGLUT3 mRNA and VMAT2 mRNA (Fig. 9A), but not of VGLUT3 mRNA and TH mRNA, in the dorsal raphé and median raphé nuclei (data not shown). This indicates that VGLUT3 mRNA is expressed in raphé serotoninergic neurons, as TH-negative VMAT2-positive neurons in the raphé are known to be serotoninergic (5). Therefore, serotoninergic synapses in raphé projection areas are operated by VMAT2. Thus, VMAT2 can be used as a marker of serotoninergic synapses provided that TH is absent (5). Confocal double immunofluorescence for VGLUT3 and VMAT2 revealed that VGLUT3 was present in a subpopulation of both thick and thin VMAT2-positive varicose fibers in the cortex (Fig. 10, G-I) and in the hippocampus but not in the caudate putamen or globus pallidum (data not shown). In addition, VGLUT3-positive fibers that were negative for VMAT2 were present in cortex (Fig. 10, G and H) and hippocampus. As catecholaminergic synapses are also operated by VMAT2, we tested whether VGLUT3 was expressed in TH-positive synapses and fibers. Confocal double immunofluorescence revealed absence of VGLUT3 from TH-positive terminals or fibers throughout the brain. Segregation of VGLUT3-positive varicose fibers and TH innervation is demonstrated exemplarily for the cortex (Fig. 10, J-L).

View larger version (100K): [in this window] [in a new window]   Fig. 9.   Double in situ hybridization and double immunofluorescence confocal analysis for VGLUT3 and VMAT2 in median raphé nuclei. A, high power micrograph of an emulsion-dipped autoradiogram illustrating the co-expression of VGLUT3 RNA transcripts (detected with radiolabeled probes) and the VMAT2 detected with a digoxigenin-labeled probe. Note that a few neurons express VGLUT3 mRNA only. B, VGLUT3 immunoreactivity is confined to terminals, whereas VMAT2 staining is present in both neuronal cell bodies and terminals. Note the segregation of VGLUT3 and VMAT2 immunoreactive fibers in this region. Scale bars, A, 20 µm; and B, 10 µm. DIG, digoxigenin.

View larger version (68K): [in this window] [in a new window]   Fig. 10.   Characterization of VGLUT3-immunopositive fibers in the rat neocortex. False color micrographs of confocal images from double immunofluorescence for VGLUT3 (green) (A, D, G, J, and M), VGLUT1 (B), VGLUT2 (E), VMAT2(H), TH (K), and VAChT (N) (red) in sections through the rat neocortex. VGLUT3 varicose fibers are mutually exclusive to VGLUT1 or VGLUT2 immunoreactivities (A-C and D-F). VGLUT3-stained thick and thin varicose fibers partly coincide with VMAT2 staining coincide with VMAT2 staining (G-I) but are negative for TH (J-L). VMAT2-negative populations of VGLUT3-positive varicosities are also present. Note the sparse innervation of the neocortex and the varicose nature of the VGLUT3-positive fibers typical for serotoninergic terminals. Interestingly, VAChT-positive fibers do not co-stain for VGLUT3 (M-O) underscoring that cholinergic afferents originating in forebrain projection neurons do not co-express VGLUT3.

In light of the ISHH analysis showing absence of VGLUT3 mRNA from noradrenergic cells in the local ceruleus, from dopaminergic neurons in the substantia nigra and ventral tegmental area, and from other catecholaminergic cell groups in the diencephalon and brainstem, we conclude that the cortical and hippocampal VGLUT3/VMAT2 co-positive system most likely stems from dorsal and medial raphé nuclei and represents a novel neural projection subsystem co-coded for glutamatergic and serotoninergic neurotransmission. This points to a specific link of VGLUT3-operated glutamatergic neurotransmission to subsets of serotoninergic raphé pathways related to the regulation of mood and depression. Whereas a majority of VGLUT3 raphé neurons may express VMAT2 mRNA, VGLUT3 mRNA was also detectable in a minor population of raphé neurons that was VMAT2 -negative suggesting the existence of a non-serotoninergic VGLUT3-coded raphé neuronal system too. In comparison to the rostral areas, the caudal serotoninergic raphé nuclei exhibited much weaker hybridization signals for VGLUT3 (Fig. 5F). As VGLUT3 mRNA was demonstrated to be absent from the histaminergic tuberomamillary nucleus, we can exclude that VGLUT3 is present in histaminergic cortical or hippocampal projections that also have to be taken into account as VMAT2-operated synapses (5).

VGLUT3 Is Expressed in a Subset of Serotoninergic and Non-serotoninergic Neurons in Primary Raphé Cultures-- In order to examine whether VGLUT3 protein is found in all serotoninergic raphé neurons and if it is found in all the axonal processes and varicosities from a single cell body, we examined primary raphé brainstem cultures (E15; 2-10 DIV) by double fluorescence deconvolution microscopy with a bona fide serotoninergic marker, TrypH. Detection of both VGLUT3 and TrypH in cell bodies was observed in primary raphé cultures (2-6 DIV). Co-staining with other neurochemical phenotypic markers enabled identification of two populations of VGLUT3-positive neurons: 1) a subset of TrypH-and VMAT2-positive neurons (Fig. 11, A and C), and 2) a population of TrypH- and VMAT2-negative neurons (Fig. 11B). The visualization of cell bodies with vesicular transporter antibodies in developing primary cultured neurons enables the identification of the origin of the numerous and extensive beaded processes that are seen in both serotoninergic and non-serotoninergic VGLUT3-containing cell types. In serotoninergic axons, VGLUT3 was enriched in the fine axonal processes and varicosities that exit from various branching points (intersections) along the main TrypH-positive varicose fiber (Fig. 11A, inset).

View larger version (38K): [in this window] [in a new window]   Fig. 11.   Characterization of VGLUT3-immunopositive cell bodies and fibers in cultured raphé neurons. False color micrographs of deconvolved images from double immunofluorescence for VGLUT3 (green; A-C) and TrypH (red; A and B) and VMAT2 (red; C) in rat primary raphé cultures (DIV = 6). A, serotoninergic neuron and axonal arborization are co-positive for TrypH and VGLUT3. A single long varicose fiber strand can be seen leaving the cell body (*), making a loop and then bifurcating (inset, 4). Note accumulation of VGLUT3 at intersections (1-5) along a single TrypH-positive varicose fiber and the VGLUT3-enriched fine processes that originate from these junctions (1-4). Single staining reveals detectable levels of TrypH in fine processes and VGLUT3 in the central varicose fiber strand. B, serotoninergic neurons negative for VGLUT3 protein and non-serotoninergic VGLUT3-positive neurons define two additional raphé chemical phenotypes. C, VGLUT3 co-localizes with VMAT2 in a subpopulation of raphé neurons (yellow). Note VGLUT3-positive and VMAT2-positive neuron with multiple varicose processes.

The non-serotoninergic neurons expressing VGLUT3 in brainstem raphé cultures may represent a local VGLUT3 system or may project to other areas of the brain including the hypothalamus and limbic regions innervated by both serotoninergic and non-serotoninergic dorsal raphé neurons. VGLUT3 staining was segregated from VGLUT2 labeling, which was also found to label a large subset of neurons in this culture preparation (data not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Here we have cloned and characterized a third isoform of the VGLUT3 that displays similar functional properties to VGLUT1 and VGLUT2 but exhibits a discrete and selective expression pattern in rodent brain. Transport activity by VGLUT3 is highly specific for glutamate, H⁺-dependent, stimulated by Cl ion, and inhibited by the dyes Rose Bengal and trypan blue. These are distinguishing characteristics of the glutamate transport system that is present on synaptic vesicles purified from the brain (see Ref. 2) and that have been found for VGLUT1 and VGLUT2 (13-18). We also show that VGLUT3 identifies a novel excitatory system in the brain that is partly an entity on its own and partly selectively promiscuous with the classical transmitters acetylcholine and serotonin. Although the three isoforms of VGLUT may display subtle differences in intrinsic activity (i.e. turnover), their anatomical distinction is striking and suggests that specialized modes of regulation and intracellular trafficking might be uniquely associated with the three classes of excitatory neurons that express these genes.

VGLUT3 Expression Defines Novel Subsets of Excitatory Neurons-- High levels of VGLUT3 mRNA are found in discrete and dispersed populations of neurons throughout the forebrain, many of which resemble the pattern displayed by interneurons. Many described interneurons are inhibitory and use GABA or glycine as neurotransmitters or are cholinergic (55). The existence of VGLUT3 mRNA in interneurons in the CA3/4 region and dentate hilus of the hippocampus and in the layer IV-V region of the cerebral cortex would suggest that these neurons affect excitatory transmission in a manner opposite to inhibitory GABAergic interneurons that have been described there (38). Excitatory interneurons expressing VGLUT3 might exert a feed-forward control of the excitability of projection neurons and would have significant impact on modulation of cortical and hippocampal circuits and cognitive processing in these regions. VGLUT1 and VGLUT2-expressing interneurons are also found in many brain areas (see Ref. 18). Glutamatergic interneurons may be critical to generate and control the oscillatory rhythms, pacemaker patterns, and final output of projection neurons (56-59) and may play key roles in degenerative disease (60), mental illness (61, 62), and epilepsy (63).

An intriguing example of VGLUT3 presence in interneurons is demonstrated by its selective expression in cholinergic striatal interneurons, as shown by double ISHH and in synaptic varicosities demonstrated by double fluorescence confocal laser microscopy. Previous reports (64, 65) have concluded that excitatory postsynaptic potentials evoked in neostriatal neurons by local stimulation in slices were due entirely to excitatory amino acids. Our results indicate that the excitatory nature of these neurons is in part a consequence of vesicular storage and release of glutamate due to selective VGLUT3 expression. Cholinergic striatal interneurons attend in the correct processing of information flow arising from the cortex and are activated during the induction phase of striatal plasticity leading to striatal long term potentiation in projection neurons (66). Whereas all cholinergic striatal interneurons express VGLUT3 mRNA, few VAChT-positive terminals actually contain VGLUT3 protein. Thus, the cholinergic interneuron system in the striatum may be highly selective and plastic in the recruitment of glutamate as a co-transmitter by regulated expression and subsequent sorting of VGLUT3 to sites of synaptic release.

VGLUT3 in the forebrain may also be expressed in excitatory projection neurons. Most VGLUT3-immunoreactive neurons in the lateral hypothalamus appear to be magnocellular neurons and may partially coincide with the melanin concentrating hormone containing neurons known to display wide projections to many areas of the brain including limbic and cortical areas where they are proposed to participate with serotoninergic systems in the regulation of sleep and feeding behavior (21, 67). Hypothalamic VGLUT3 neurons are also located in a region considered an "aggression area" in the rat (68). Thus, the hypothalamic VGLUT3 system may represent an excitatory link in such pathways. In addition, VGLUT3-operated neurons from the dorsal raphé nuclei are likely to terminate in the hypothalamus and provide the final excitatory loop from brainstem autonomic centers to the hypothalamus and limbic autonomic areas. Bed nucleus of striata terminalis projections expressing VGLUT3 mRNA are also thought to innervate the hypothalamus providing a direct excitatory link from the limbic system.

An Excitatory System in a Subset of Serotoninergic Neurons-- VGLUT3 is expressed in a subset of serotoninergic projection neurons with highest densities of cell bodies in the midline region of the mesopontine raphé system that has projections to many areas of the forebrain. Noteworthy, the caudal medullary raphé neurons including the raphé obscurus and pallidus expressed much lower levels of VGLUT3 mRNA possibly indicating different states of activity of the serotoninergic system with respect to co-coding for glutamatergic co-transmission. An extensive diversity in chemical phenotype is a hallmark of serotoninergic neurons in vivo and in vitro, and it is not clear whether all raphé neurons (i.e. phenotypes) project to all areas of the brain (28, 69). However, most serotoninergic axons are of considerable length, issue abundant collaterals, and branch profusely with most parts of the neuroaxis, giving rise to literally billions of varicose nerve endings (70, 71). Co-localization of VGLUT3 with VMAT2 and not with TH or VAChT in the neocortex would indicate that vesicular glutamate and 5-HT storage occurs in the same fiber systems originating from a distinct raphé neuron. Glutamate immunoreactivity has been localized in a majority of medullary rat 5-HT neurons (72) that would support VGLUT3 activity and vesicular glutamate storage. Extracellularly evoked glutamatergic synaptic potentials have been recorded in mesopontine raphé slices (73) suggesting exocytotic release from VGLUT3-loaded vesicles. These excitatory postsynaptic potentials are followed by serotoninergic inhibitory responses. Similar biphasic postsynaptic potentials are recorded in the brain and spinal cord (74, 75).

Glutamatergic Co-transmission with Classical Neurotransmitters-- The co-transmission concept put forward by Hökfelt et al. (76) indicates that Ca²⁺-dependent co-release of classical transmitters and neuropeptides occurs at all synapses within a given neuron. At the molecular level, this relies on the expression of the specific genes that define the chemical and peptidergic phenotype and at the cellular level, on appropriate targeting of the secretory organelles that contain the classical neurotransmitter or peptide to synaptic release sites. In the case of excitatory neurons, the most reliable marker for the glutamatergic phenotype is a vesicular glutamate transporter. Our current results provide molecular evidence for the concept of distinct pathway- and subset-specific classical transmitter co-existence as VGLUT3 is co-expressed with VMAT2 and TrypH in subsets of serotoninergic neurons of the mesopontine raphé and with VAChT in cholinergic striatal interneurons. Apparently, very distinct neuronal populations encode the potential for dual classical transmitter release. Expression of VGLUT3 protein in a subset of cholinergic and serotoninergic neurons in the adult brain in vivo indicates that vesicular storage of glutamate also occurs at these synapses, and this likely reflects a role in subset-specific neurotransmitter co-release amplifying versatility of target- and pathway-specific neurotransmission. VGLUT3 is expressed in serotoninergic raphé neurons that project to the cerebral cortex and hippocampus but not to most other areas of the brain including the raphé neurons themselves (via collaterals), suggesting a specialized role for glutamatergic/serotoninergic co-transmission in the cortical regions. Conventional or "junctional" serotoninergic synapses are a rarity in cerebral cortex (77, 78) suggesting that glutamate release from non-junctional varicosities (79) upon axonal depolarization serves a neuro-hormonal role, in this case activating. Interestingly, small serotoninergic fibers in the cerebral cortex are vulnerable to axonal degeneration following 3,4-methylenedioxymethamphetamine (MDMA, ecstasy) exposure, whereas large beaded varicosities are spared (80). Although VGLUT3 protein is not present in all VMAT2-positive processes in the cerebral cortex and is not in all serotoninergic cell bodies and processes in culture, it does appear to co-localize with VMAT2 in both thin and thicker varicose fibers in vivo and in vitro. Interestingly, VGLUT3-enriched fine varicose fibers are seen originating from branch points along a thicker TrypH co-positive varicose fiber in raphé cultures. The accumulation of VGLUT3 at these intersections may represent important sites for sorting to axonal varicosities. Serotoninergic reinnervation following MDMA toxicity may be the result of axon arborizations of spared varicosities or arise from new processes (collaterals) from the same cell (80). Serotonin varicosities have the intrinsic capacity to form a junctional synapse (81), or an arborization, that may be largely influenced by epigenetic factors in their territory of innervation. A role for VGLUT3 and glutamatergic co-transmission in the plasticity of axonal arborization in raphé cortical projections is possible that may also be important during synapse remodeling and development (79, 82, 83).

Glutamatergic co-transmission with classical transmitters may provide unique control of the synaptic response generated in the postsynaptic neuron. If VGLUT3 is expressed on small clear vesicles known to be present besides large dense core vesicles storing and releasing serotonin (69, 84, 85), classic transmission could be sequential rather than co-temporal depending on the frequency or intensity-dependent release properties of the two vesicle populations (86). Co-release from the same or different synaptic vesicles may provide co-temporal release with glutamate as the fast excitatory part and serotonin the postexcitatory "brake." A gain in excitation may be provided by co-release of glutamate and acetylcholine. A reason for co-temporal or sequential co-transmission of two classical transmitters may be to increase versatility and fine-tuning of the transmission machinery especially in highly adaptive systems such as the raphé projections contributing to regulating stress, emotions, and other vegetative behavior. Plastic wiring may need more messenger repertoires than solid stereotype wiring like motor and sensory programs that once learned are more firmly ingrained. Differential neuropeptide repertoires in these three classes of glutamatergic synapses are likely to further contribute to an unprecedented chemical and functional diversity of glutamatergic neurotransmission with major relevance for the pathophysiology and pharmacotherapy of many neurological and neuropsychiatric diseases.

	ACKNOWLEDGEMENTS

The assistance of M. Zibuschka, E. Rodenberg, P. Sack, B. Wiegand, H. Hlawaty, M. Schneider, M. Kuebel, and L. Davis and the photographic expertise of H. Schneider are gratefully acknowledged.

	Addendum

After this paper was prepared for submission and review, two related studies appeared reporting rat (87) and human (88) VGLUT3.

	FOOTNOTES

^* This work was supported by German Research Foundation Grants SFB 297, BMB+F 01GS01118, and BMB+F 01 GG 9818/0 (to E. W. and M. K.-H. S.) and by National Institutes of Health Grants 1P29RR16816 (to H. V.) and NS36936 (to J. D. E.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EBI Data Bank with accession number(s) AF510321.

^§ Both authors contributed equally to this work.

To whom correspondence should be addressed: Neuroscience Center, Louisiania State University Health Sciences Center, 2020 Gravier St., Ste. D, New Orleans, LA 70112. E-mail: jerick@lsuhsc.edu.

Published, JBC Papers in Press, October 15, 2002, DOI 10.1074/jbc.M206738200

	ABBREVIATIONS

The abbreviations used are: VGLUT, vesicular glutamate transporter; DMEM, Dulbecco's modified Eagle's medium; ISHH, in situ hybridization histochemistry; GST, glutathione S-transferase; TrypH, tryptophan hydroxylase; BSA, bovine serum albumin; TMD, transmembrane domain; GABA, -aminobutyric acid; VAChT, vesicular acetylcholine transporter; TH, tyrosine hydroxylase; DIV, days in vitro; 5-HT, serotonin; PBS, phosphate-buffered saline; FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone; ncl., nucleus.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				D. Balschun, D. Moechars, Z. Callaerts-Vegh, B. Vermaercke, N. Van Acker, L. Andries, and R. D'Hooge Vesicular Glutamate Transporter VGLUT1 Has a Role in Hippocampal Long-Term Potentiation and Spatial Reversal Learning Cereb Cortex, July 2, 2009; (2009) bhp133v1. [Abstract] [Full Text] [PDF]

				S. Grewal, N. Defamie, X. Zhang, S. De Gois, A. Shawki, B. Mackenzie, C. Chen, H. Varoqui, and J. D. Erickson SNAT2 Amino Acid Transporter Is Regulated by Amino Acids of the SLC6 {gamma}-Aminobutyric Acid Transporter Subfamily in Neocortical Neurons and May Play No Role in Delivering Glutamine for Glutamatergic Transmission J. Biol. Chem., April 24, 2009; 284(17): 11224 - 11236. [Abstract] [Full Text] [PDF]

				L. A. Parra, T. Baust, S. El Mestikawy, M. Quiroz, B. Hoffman, J. M. Haflett, J. K. Yao, and G. E. Torres The Orphan Transporter Rxt1/NTT4 (SLC6A17) Functions as a Synaptic Vesicle Amino Acid Transporter Selective for Proline, Glycine, Leucine, and Alanine Mol. Pharmacol., December 1, 2008; 74(6): 1521 - 1532. [Abstract] [Full Text] [PDF]

				R. C. Reyes and V. Parpura Mitochondria Modulate Ca2+-Dependent Glutamate Release from Rat Cortical Astrocytes J. Neurosci., September 24, 2008; 28(39): 9682 - 9691. [Abstract] [Full Text] [PDF]

				J. A. Mendez, M.-J. Bourque, G. D. Bo, M. L. Bourdeau, M. Danik, S. Williams, J.-C. Lacaille, and L.-E. Trudeau Developmental and Target-Dependent Regulation of Vesicular Glutamate Transporter Expression by Dopamine Neurons J. Neurosci., June 18, 2008; 28(25): 6309 - 6318. [Abstract] [Full Text] [PDF]

				B. J. Maher and G. L. Westbrook Co-Transmission of Dopamine and GABA in Periglomerular Cells J Neurophysiol, March 1, 2008; 99(3): 1559 - 1564. [Abstract] [Full Text] [PDF]

				L.-E. Trudeau and R. Gutierrez On Cotransmission & Neurotransmitter Phenotype Plasticity Mol. Interv., June 1, 2007; 7(3): 138 - 146. [Abstract] [Full Text] [PDF]

				S. Geisler, C. Derst, R. W. Veh, and D. S. Zahm Glutamatergic Afferents of the Ventral Tegmental Area in the Rat J. Neurosci., May 23, 2007; 27(21): 5730 - 5743. [Abstract] [Full Text] [PDF]

				E. Hinoi, T. Takarada, K. Uno, M. Inoue, Y. Murafuji, and Y. Yoneda Glutamate Suppresses Osteoclastogenesis through the Cystine/Glutamate Antiporter Am. J. Pathol., April 1, 2007; 170(4): 1277 - 1290. [Abstract] [Full Text] [PDF]

				D. Moechars, M. C. Weston, S. Leo, Z. Callaerts-Vegh, I. Goris, G. Daneels, A. Buist, M. Cik, P. van der Spek, S. Kass, et al. Vesicular Glutamate Transporter VGLUT2 Expression Levels Control Quantal Size and Neuropathic Pain. J. Neurosci., November 15, 2006; 26(46): 12055 - 12066. [Abstract] [Full Text] [PDF]

				W. Zhou, L.-W. Fu, S. C. Tjen-A-Looi, Z.-l. Guo, and J. C. Longhurst Role of glutamate in a visceral sympathoexcitatory reflex in rostral ventrolateral medulla of cats Am J Physiol Heart Circ Physiol, September 1, 2006; 291(3): H1309 - H1318. [Abstract] [Full Text] [PDF]

				E. Hrabovszky, I. Kallo, G. F. Turi, K. May, G. Wittmann, C. Fekete, and Z. Liposits Expression of Vesicular Glutamate Transporter-2 in Gonadotrope and Thyrotrope Cells of the Rat Pituitary. Regulation by Estrogen and Thyroid Hormone Status Endocrinology, August 1, 2006; 147(8): 3818 - 3825. [Abstract] [Full Text] [PDF]

				S. De Gois, M. K.-H. Schafer, N. Defamie, C. Chen, A. Ricci, E. Weihe, H. Varoqui, and J. D. Erickson Homeostatic Scaling of Vesicular Glutamate and GABA Transporter Expression in Rat Neocortical Circuits J. Neurosci., August 3, 2005; 25(31): 7121 - 7133. [Abstract] [Full Text] [PDF]

				Y. Zilberter, T. Harkany, and C. D. Holmgren Dendritic Release of Retrograde Messengers Controls Synaptic Transmission in Local Neocortical Networks Neuroscientist, August 1, 2005; 11(4): 334 - 344. [Abstract] [PDF]

				N. R. Wilson, J. Kang, E. V. Hueske, T. Leung, H. Varoqui, J. G. Murnick, J. D. Erickson, and G. Liu Presynaptic Regulation of Quantal Size by the Vesicular Glutamate Transporter VGLUT1 J. Neurosci., June 29, 2005; 25(26): 6221 - 6234. [Abstract] [Full Text] [PDF]

				A. Lavin, L. Nogueira, C. C. Lapish, R. M. Wightman, P. E. M. Phillips, and J. K. Seamans Mesocortical Dopamine Neurons Operate in Distinct Temporal Domains Using Multimodal Signaling J. Neurosci., May 18, 2005; 25(20): 5013 - 5023. [Abstract] [Full Text] [PDF]

				G. Z. Mentis, F. J. Alvarez, A. Bonnot, D. S. Richards, D. Gonzalez-Forero, R. Zerda, and M. J. O'Donovan Noncholinergic excitatory actions of motoneurons in the neonatal mammalian spinal cord PNAS, May 17, 2005; 102(20): 7344 - 7349. [Abstract] [Full Text] [PDF]

				S. Winter, I. Brunk, D. J. Walther, M. Holtje, M. Jiang, J.-U. Peter, S. Takamori, R. Jahn, L. Birnbaumer, and G. Ahnert-Hilger G{alpha}o2 Regulates Vesicular Glutamate Transporter Activity by Changing Its Chloride Dependence J. Neurosci., May 4, 2005; 25(18): 4672 - 4680. [Abstract] [Full Text] [PDF]

				G. Salazar, B. Craige, R. Love, D. Kalman, and V. Faundez Vglut1 and ZnT3 co-targeting mechanisms regulate vesicular zinc stores in PC12 cells J. Cell Sci., May 1, 2005; 118(9): 1911 - 1921. [Abstract] [Full Text] [PDF]

				N. K. Mueller, S. Di, C. M. Paden, and J. P. Herman Activity-Dependent Modulation of Neurotransmitter Innervation to Vasopressin Neurons of the Supraoptic Nucleus Endocrinology, January 1, 2005; 146(1): 348 - 354. [Abstract] [Full Text] [PDF]

				E. Hrabovszky, G. Wittmann, G. F. Turi, Z. Liposits, and C. Fekete Hypophysiotropic Thyrotropin-Releasing Hormone and Corticotropin-Releasing Hormone Neurons of the Rat Contain Vesicular Glutamate Transporter-2 Endocrinology, January 1, 2005; 146(1): 341 - 347. [Abstract] [Full Text] [PDF]

				B. G. Croft, G. D. Fortin, A. T. Corera, R. H. Edwards, A. Beaudet, L.-E. Trudeau, and E. A. Fon Normal Biogenesis and Cycling of Empty Synaptic Vesicles in Dopamine Neurons of Vesicular Monoamine Transporter 2 Knockout Mice Mol. Biol. Cell, January 1, 2005; 16(1): 306 - 315. [Abstract] [Full Text] [PDF]

				H. Hioki, F. Fujiyama, K. Nakamura, S.-X. Wu, W. Matsuda, and T. Kaneko Chemically Specific Circuit Composed of Vesicular Glutamate Transporter 3- and Preprotachykinin B-producing Interneurons in the Rat Neocortex Cereb Cortex, November 1, 2004; 14(11): 1266 - 1275. [Abstract] [Full Text] [PDF]

				R. T. Fremeau Jr., K. Kam, T. Qureshi, J. Johnson, D. R. Copenhagen, J. Storm-Mathisen, F. A. Chaudhry, R. A. Nicoll, and R. H. Edwards Vesicular Glutamate Transporters 1 and 2 Target to Functionally Distinct Synaptic Release Sites Science, June 18, 2004; 304(5678): 1815 - 1819. [Abstract] [Full Text] [PDF]

				K. Nakamura, K. Matsumura, T. Hubschle, Y. Nakamura, H. Hioki, F. Fujiyama, Z. Boldogkoi, M. Konig, H.-J. Thiel, R. Gerstberger, et al. Identification of Sympathetic Premotor Neurons in Medullary Raphe Regions Mediating Fever and Other Thermoregulatory Functions J. Neurosci., June 9, 2004; 24(23): 5370 - 5380. [Abstract] [Full Text] [PDF]

				T. Harkany, C. Holmgren, W. Hartig, T. Qureshi, F. A. Chaudhry, J. Storm-Mathisen, M. B. Dobszay, P. Berghuis, G. Schulte, K. M. Sousa, et al. Endocannabinoid-Independent Retrograde Signaling at Inhibitory Synapses in Layer 2/3 of Neocortex: Involvement of Vesicular Glutamate Transporter 3 J. Neurosci., May 26, 2004; 24(21): 4978 - 4988. [Abstract] [Full Text] [PDF]

				S. M. Wojcik, J. S. Rhee, E. Herzog, A. Sigler, R. Jahn, S. Takamori, N. Brose, and C. Rosenmund An essential role for vesicular glutamate transporter 1 (VGLUT1) in postnatal development and control of quantal size PNAS, May 4, 2004; 101(18): 7158 - 7163. [Abstract] [Full Text] [PDF]

				V. Montana, Y. Ni, V. Sunjara, X. Hua, and V. Parpura Vesicular Glutamate Transporter-Dependent Glutamate Release from Astrocytes J. Neurosci., March 17, 2004; 24(11): 2633 - 2642. [Abstract] [Full Text] [PDF]

				Y. Moriyama and A. Yamamoto Glutamatergic Chemical Transmission: Look! Here, There, and Anywhere J. Biochem., February 1, 2004; 135(2): 155 - 163. [Abstract] [Full Text] [PDF]

				L. Cheng, C.-L. Chen, P. Luo, M. Tan, M. Qiu, R. Johnson, and Q. Ma Lmx1b, Pet-1, and Nkx2.2 Coordinately Specify Serotonergic Neurotransmitter Phenotype J. Neurosci., November 5, 2003; 23(31): 9961 - 9967. [Abstract] [Full Text] [PDF]

				M. Hayashi, R. Morimoto, A. Yamamoto, and Y. Moriyama Expression and Localization of Vesicular Glutamate Transporters in Pancreatic Islets, Upper Gastrointestinal Tract, and Testis J. Histochem. Cytochem., October 1, 2003; 51(10): 1375 - 1390. [Abstract] [Full Text] [PDF]

				B. Mackenzie, M. K.-H. Schafer, J. D. Erickson, M. A. Hediger, E. Weihe, and H. Varoqui Functional Properties and Cellular Distribution of the System A Glutamine Transporter SNAT1 Support Specialized Roles in Central Neurons J. Biol. Chem., June 20, 2003; 278(26): 23720 - 23730. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 277/52/50734    most recent M206738200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Schäfer, M. K.-H. Articles by Erickson, J. D. Search for Related Content PubMed PubMed Citation Articles by Schäfer, M. K.-H. Articles by Erickson, J. D. Social Bookmarking                      What's this?