Cdk5 is essential for synaptic vesicle endocytosis

Synaptic vesicle endocytosis (SVE) is triggered by calcineurin-mediated dephosphorylation of the dephosphin proteins. SVE is maintained by the subsequent rephosphorylation of the dephosphins by unidentified protein kinases. Here, we show that cyclin-dependent kinase 5 (Cdk5) phosphorylates dynamin I on Ser 774 and Ser 778 in vitro, which are identical to its endogenous phosphorylation sites in vivo. Cdk5 antagonists and expression of dominant-negative Cdk5 block phosphorylation of dynamin I, but not of amphiphysin or AP180, in nerve terminals and inhibit SVE. Thus Cdk5 has an essential role in SVE and is the first dephosphin kinase identified in nerve terminals.     

The in vivo phosphorylation sites in multiple isoforms of amphiphysin I from rat brain nerve terminals

Amphiphysin I (amphI) is dephosphorylated by calcineurin during nerve terminal depolarization and synaptic vesicle endocytosis (SVE). Some amphI phosphorylation sites (phosphosites) have been identified with in vitro studies or phosphoproteomics screens. We used a multifaceted strategy including 32P tracking to identify all in vivo amphI phosphosites and determine their relative abundance and potential relevance to SVE. AmphI was extracted from 32P-labeled synaptosomes, phosphopeptides were isolated from proteolytic digests using TiO2 chromatography, and mass spectrometry revealed 13 sites: serines 250, 252, 262, 268, 272, 276, 285, 293, 496, 514, 539, and 626 and Thr-310. These were distributed into two clusters around the proline-rich domain and the C-terminal Src homology 3 domain. Hierarchical phosphorylation of Ser-262 preceded phosphorylation of Ser-268, -272, -276, and -285. Off-line HPLC separation and two-dimensional tryptic mapping of 32P-labeled amphI revealed that Thr-310, Ser-293, Ser-285, Ser-272, Ser-276, and Ser-268 contained the highest 32P incorporation and were the most stimulus-sensitive. Individually Thr-310 and Ser-293 were the most abundant phosphosites, incorporating 16 and 23% of the 32P. The multiple phosphopeptides containing Ser-268, Ser-276, Ser-272, and Ser-285 had 27% of the 32P. Evidence for a role for at least one proline-directed protein kinase and one non-proline-directed kinase was obtained. Four phosphosites predicted for non-proline-directed kinases, Ser-626, -250, -252, and -539, contained low amounts of 32P and were not depolarization-responsive. At least one alternatively spliced amphI isoform was identified in synaptosomes as being constitutively phosphorylated because it did not incorporate 32P during the 1-h labeling period. Multiple phosphosites from amphI-co-migrating synaptosomal proteins were also identified, including SGIP (Src homology 3 domain growth factor receptor-bound 2 (Grb2)-like (endophilin)-interacting protein 1), AAK1, eps15R, MAP6, alpha/beta-adducin, and HCN1. The results reveal two sets of amphI phosphosites that are either dynamically turning over or constitutively phosphorylated in nerve terminals and improve understanding of the role of individual amphI sites or phosphosite clusters in synaptic SVE.     

Effect of Aging on Tyrosine Hydroxylase Protein Content and the Relative Number of Dopamine Nerve Terminals in Human Caudate

This study examined the effect of aging on the relative number of dopamine (DA) nerve terminals in human caudate nucleus, their content of tyrosine hydroxylase (TH) protein, and the relative abundance of TH monomers with different molecular weights. Preliminary studies on brain tissue cryopreservation, performed with rat striatum, indicated that intact synaptosomes can be prepared from fresh tissue slowly frozen in 0.32 M sucrose with 5% dimethyl sulfoxide and then thawed rapidly prior to synaptosome preparation. Synaptosomes were prepared in this manner from postmortem caudate nucleus tissue obtained from normal humans 1 month to 63 years of age. To determine the relative number of DA nerve terminals for each individual, dopaminergic synaptosomes were selectively labeled with a monoclonal antibody to TH and quantified by fluorescence-activated cell sorting. To determine the relative amount of TH protein for each individual, the concentration of TH protein in the same synaptosomal preparations was determined using immunoblots. Our results suggest that caudate TH levels plateau soon after birth and tend to remain relatively stable during aging, since no changes in either the relative number of TH-containing nerve terminals or the concentration of TH protein were found in subjects 15-63 years of age. In light of previous studies showing an age-related loss of DA cell bodies, these findings suggest that remaining DA neurons compensate to maintain caudate levels of TH protein and TH-containing nerve terminals. Immunoblot studies identified three forms of TH monomer (60.6, 61.7, and 65.1 kDa), indicating that mRNAs coding for high molecular mass forms of TH may be actively translated in human brain. No age-related differences in the relative abundance of these forms were found.     

Regulation of cytosolic free calcium concentration by intrasynaptic mitochondria.

By the use of digitonin permeabilized presynaptic nerve terminals (synaptosomes), we have found that intrasynaptic mitochondria, when studied "in situ," i.e., surrounded by their cytosolic environment, are able to buffer calcium in a range of calcium concentrations close to those usually present in the cytosol of resting synaptosomes. Adenine nucleotides and polyamines, which are usually lost during isolation of mitochondria, greatly improve the calcium-sequestering activity of mitochondria in permeabilized synaptosomes. The hypothesis that the mitochondria contributes to calcium homeostasis at low resting cytosolic free calcium concentration ([Ca2+]i) in synaptosomes has been tested; it has been found that in fact this is the case. Intrasynaptic mitochondria actively accumulates calcium at [Ca2+]i around 10(-7) M, and this activity is necessary for the regulation of [Ca2+]i. When compared with other membrane-limited calcium pools, it was found that depending on external concentration the calcium pool mobilized from mitochondria is similar or even greater than the IP3- or caffeine-sensitive calcium pools. In summary, the results presented argue in favor of a more prominent role of mitochondria in regulating [Ca2+]i in presynaptic nerve terminals, a role that should be reconsidered for other cellular types in light of the present evidence.     

How is the cytoplasmic calcium concentration controlled in nerve terminals?

1. The ability of intraterminal organelles to sequester calcium and buffer the cytoplasmic free Ca2+ concentration ([Ca2+]i) has been investigated in isolated mammalian presynaptic nerve terminals (synaptosomes). A combination of biochemical and morphological methods has been used. 2. When the plasmalemma of synaptosomes is disrupted by osmotic shock or saponin, Ca from the medium can be sequestered by two types of intraterminal organelles in the presence of ATP. 2. Typical mitochondrial poisons (e.g., oligomycin, azide and 2,4-dinitrophenol) block the Ca uptake into one type of organelle (mitochondria); the second type of organelle, which has a higher affinity for Ca (half-saturation congruent to 0.35 microM Ca2+) is spared by the mitochondrial poisons. 4. When the "leaky" synaptosomes are incubated in media containing oxalate, and then fixed and prepared for electron microscopy, electron-dense deposits are observed in the intraterminal mitochondria and smooth endoplasmic reticulum (SER). Mitochondrial poisons block the formation of the deposits in the mitochondria, but spare the SER. 5. X-ray microprobe analysis demonstrates that these deposits contain Ca. 6. Experiments with the Ca-sensitive metallochromic indicator, arsenazo III, demonstrate that the intraterminal organelles in the "leaky" synaptosomes can buffer Ca2+ in the medium to below 5 X 10(-7) M. With small (physiological) Ca loads, the Ca2+ is effectively buffered (to < 5 X 10(-7) M) even in the presence of mitochondrial poisons. 7. The data indicate that the SER in presynaptic terminals may play an important role in helping to buffer the Ca that normally enters during neuronal activity.     

Adenine nucleotides in cholinergic transmission: presynaptic aspects.

1. Isolated nerve terminals (T-sacs and synaptosomes) prepared from the purely cholinergic Torpedo electric organ have been studied for their ability to incorporate and metabolise [2-3H] adenosine and to degrade 5'-AMP to adenosine. 2. A temperature-dependent, saturable uptake system for adenosine was found with kinetic properties similar to nucleoside transport systems in other cells. The uptake system in Torpedo nerve terminals was inhibited by 2'-deoxyadenosine, a known inhibitor of adenosine transport. 3. Intraterminal adenosine is rapidly metabolised to a number of products including AMP, ADP and ATP. 4. Isolated nerve terminals contain considerable 5'-nucleotidase activity, most of which resides on the outer face of the external membrane. The Km of the enzyme is congruent to 5 micron and it is inhibited by a phosphonate analogue of ADP, alpha-beta-methylene-ADP. It is suggested that this 5'-nucleotidase plays an important role in the production of adenosine from a nucleotide pool in the synaptic cleft.     

Identification of a Synaptic Vesicle Antigen (Mr 86,000) Conserved Between Torpedo and Rat

Antisera were raised in guinea pigs to synaptic vesicles purified from the electric organ of Torpedo marmorata. In cholinergic nerve terminals from Torpedo the major antigens identified had Mr 300,000-150,000, 86,000, and 18,000. The Mr 86,000 antigen was conserved between Torpedo and rat, where it is neuron-specific and concentrated in nerve terminals. When rat brain synaptosomes are subfractionated the antigen is associated with synaptic vesicles. The antigen is not found in the cytoskeleton and in the vesicle-free cytosol. Immunohistochemical localization of the antigen in rat shows it to be associated with synapses in diaphragm, cerebellum, hippocampus, and cerebral cortex. The staining pattern of the antigen indicates that the antigen is not cholinergic-specific. The function of the Mr 86,000 antigen remains to be identified.     

Axonal and presynaptic RNAs are locally transcribed in glial cells.

In the last few years, the long-standing opinion that axonal and presynaptic proteins are exclusively derived from the neuron cell body has been substantially modified by the demonstration that active systems of protein synthesis are present in axons and nerve terminals. These observations have raised the issue of the cellular origin of the involved RNAs, which has been generally attributed to the neuron soma. However, data gathered in a number of model systems indicated that axonal RNAs are synthesized in the surrounding glial cells. More recent experiments on the perfused squid giant axon have definitively proved that axoplasmic RNAs are transcribed in periaxonal glia. Their delivery to the axon occurs by a modulatory mechanism based on the release of neurotransmitters from the stimulated axon and on their binding to glial receptors. In additional experiments on squid optic lobe synaptosomes, presynaptic RNA has been also shown to be synthesized locally, presumably in nearby glia. Together with a wealth of literature data, these observations indicate that axons and nerve terminals are endowed with a local system of gene expression that supports the maintenance and plasticity of these neuronal domains.     

Involvement of Different Types of Voltage-Sensitive Calcium Channels in the Presynaptic Regulation of Noradrenaline Release in Rat Brain Cortex and Hippocampus

Abstract: Transmitter release at the nerve terminal is mediated by the influx of Ca²⁺ through voltage-sensitive calcium channels (VSCCs). Many types of VSCCs have been found in neurons (T, N, L, and P), but uncertainty remains about which ones are involved in neuronal excitation-secretion coupling. Specific ligands for the L- and N-type VSCCs were used to determine which of these subtypes might be involved in the K⁺-evoked [³H]noradrenaline release from superfused rat brain cortical and hippocampal synaptosomes. In cortical presynaptic terminals the 1,4-dihydropyridine agonist Bay K 8644 enhanced the K⁺ (15 m M )-evoked [³H]noradrenaline release. This effect was reversed by the 1,4-dihydropyridine antagonists nimodipine and nitrendipine. The L-type VSCC ligands had no effect on hippocampal synaptosomes. In contrast, the N-type VSCC blocker ω-conotoxin markedly reduced the K⁺-evoked [³H]noradrenaline release in nerve terminals from both regions. Inhibition was greater in hippocampal synaptosomes. When applied together the inhibitory actions of nimodipine and ω-conotoxin were approximately additive. These findings indicate that both L- and N-type VSCCs participate in noradrenaline release in rat brain cortex and suggest that noradrenergic terminals in the two regions examined may have distinct populations of VSCCs: L type in cortex and N type in hippocampus.     

Presynaptic Mechanisms Underlying the γ-Aminobutyric Acid-Evoked Receptor-Independent Release of [3H]Norepinephrine in Rat Hippocampus

The effects of gamma-aminobutyric acid (GABA) on the spontaneous efflux of [3H]norepinephrine ([3H]NE) were studied in synaptosomes prepared from rat hippocampus and prelabelled with [3H]NE. It had been observed previously that, when synaptosomes were exposed in superfusion to GABA, the basal release of the tritiated catecholamine was enhanced, apparently with no involvement of the known GABA receptors. The mechanisms underlying this effect have now been investigated. The potency of GABA as a releaser of [3H]NE was decreased by lowering the Na+ content of the superfusion medium, and its effect disappeared at 23 mM Na+. The GABA-induced [3H]NE release was counteracted by the GABA uptake inhibitor N-(4,4-diphenyl-3-butenyl)nipecotic acid (SKF 89976A), but it was unaffected by the NE uptake blockers desmethylimipramine and nisoxetine. The GABA-induced release of [3H]NE was Ca2+-dependent and tetrodotoxin-sensitive. The data support the hypothesis that GABA provoked [3H]NE release by a novel mechanism which involves penetration into the noradrenergic nerve terminals through a GABA carrier located on the NE terminals themselves. This uptake process might be electrogenic and provoke depolarization of the nerve terminals, causing an exocytotic release of [3H]NE.     

The regulation of cytosolic pH in isolated presynaptic nerve terminals from rat brain

Cytosolic pH (pHi) was measured in presynaptic nerve terminals isolated from rat brain (synaptosomes) using a fluorescent pH indicator, 2',7'-bis(carboxyethyl)-5,6-carboxyfluorescein (BCECF). The synaptosomes were loaded with BCECF by incubation with the membrane-permanent acetoxy-methyl ester derivative of BCECF, which is hydrolyzed by intracellular esterases to the parent compound. pHi was estimated by calibrating the fluorescence signal after permeabilizing the synaptosomal membrane by two different methods. Synaptosomes loaded with 15-90 microM BCECF were estimated to have a pHi of 6.94 +/- 0.02 (mean +/- standard error; n = 54) if the fluorescence signal was calibrated after permeabilizing with digitonin; a similar value was obtained using synaptosomes loaded with 10 times less BCECF (6.9 +/- 0.1; n = 5). When the fluorescence signal was calibrated by permeabilizing the synaptosomal membrane to H+ with gramicidin and nigericin, pHi was estimated to be 7.19 +/- 0.03 (n = 12). With the latter method, pHi = 6.95 +/- 0.09 (n = 14) when the synaptosomes were loaded with 10 times less BCECF. Thus, pHi in synaptosomes was approximately 7.0 and could be more precisely monitored using the digitonin calibration method at higher BCECF concentrations. When synaptosomes were incubated in medium containing 20 mM NH4Cl and then diluted into NH4Cl-free medium, pHi immediately acidified to a level of approximately 6.6. After the acidification, pHi recovered over a period of a few minutes. The buffering capacity of the synaptosomes was estimated to be approximately 50 mM/pH unit. Recovery was substantially slowed by incubation in an Na-free medium, by the addition of amiloride (KI = 3 microM), and by abolition of the Nao/Nai gradient. pHi and its recovery after acidification were not affected by incubation in an HCO3-containing medium; disulfonic stilbene anion transport inhibitors (SITS and DIDS, 1 mM) and replacement of Cl with methylsulfonate did not affect the rate of recovery of pHi. It appears that an Na+/H+ antiporter is the primary regulator of pHi in mammalian brain nerve terminals.     

Dopamine receptor regulation of Ca2+ levels in individual isolated nerve terminals from rat striatum: comparison of presynaptic D1-like and D2-like receptors

We have directly observed the effects of activating presynaptic D1-like and D2-like dopamine receptors on Ca2+ levels in isolated nerve terminals (synaptosomes) from rat striatum. R-(+)-SKF81297, a selective D1-like receptor agonist, and (-)-quinpirole, a selective D2-like receptor agonist, induced increases in Ca2+ levels in different subsets of individual striatal synaptosomes. The SKF81297- and quinpirole-induced effects were blocked by R-(+)-SCH23390, a D1-like receptor antagonist, and (-)-sulpiride, a D2-like receptor antagonist, respectively. SKF81297- or quinpirole-induced Ca2+ increases were inhibited following blockade of voltage-gated calcium channels or sodium channels. In a larger subset of synaptosomes, quinpirole decreased baseline Ca2+. Quinpirole also inhibited veratridine-induced increases in intrasynaptosomal Ca2+ level. Immunostaining confirmed the presynaptic expression of D1, D5, D2 and D3 receptors, but not D4 receptors. The array of neurotransmitter phenotypes of the striatal nerve endings expressing D1, D5, D2 or D3 varied for each receptor subtype. These results suggest that presynaptic D1-like and D2-like receptors induce increases in Ca2+ levels in different subsets of nerve terminals via Na+ channel-mediated membrane depolarization, which, in turn, induces the opening of voltage-gated calcium channels. D2-like receptors also reduce nerve terminal Ca2+ in a different but larger subset of synaptosomes, consistent with the predominant presynaptic action of dopamine in the striatum being inhibitory.     

Localization of vasopressin in synaptic vesicles of extra-hypothalamic rat brain

The subcellular localization of vasopressin (VP) from extra-hypothalamic areas of rat brain was investigated by measuring its distribution (a) along a continuous sucrose gradient; (b) during the preparation of isolated nerve endings (synaptosomes) and (c) during the preparation of synaptic vesicles.Quite large amounts of vasopressin are isolated in the same fractions as mitochondria, as well as synaptosomes. Osmotic rupture of membrane bound organelles in the homogenate results in the vasopressin being measured largely in the fraction containing synaptic vesicles. These results would suggest that vasopressin could be released by nerve terminals which is consistent with the hypothesis that it may have a neurotransmitter/neuromodulator function in the CNS.     

Ca<Superscript>2+</Superscript> Signalling in Brain Synaptosomes Activated by Dinucleotides

Diadenosine polyphosphates are a family of dinucleotides formed by two adenosines joined by a variable number of phosphates. Diadenosine tetraphosphate, Ap₄A, diadenosine pentaphosphate Ap₅A, and diadenosine hexaphosphate, Ap₆A, are stored in synaptic vesicles and are released upon nerve terminal depolarization. At the extracellular level, diadenosine polyphosphates can stimulate presynaptic dinucleotide receptors. Responses to diadenosine polyphosphates have been described in isolated synaptic terminals (synaptosomes) from several brain areas in different animal species, including man. Dinucleotide receptors are ligand-operated ion channels that allow the influx of cations into the terminals. These cations reach a threshold for N- and P/Q-type voltage-dependent calcium channels, which become activated. The activation of the dinucleotide receptor together with the activation of these calcium channels triggers the release of neurotransmitters. The ability of Ap₅A to promote glutamate, GABA or acetylcholine release has been recently described by the present authors in rat midbrain synaptosomes.     

Cytosolic O-glycosylation is abundant in nerve terminals

Phosphorylation plays a key role in regulating growth cone migration and protein trafficking in nerve terminals. Here we show that nerve terminal proteins contain another abundant post-translational modification: beta-N-acetylglucosamine linked to hydroxyls of serines or threonines (O-GlcNAc(1)). O-GlcNAc modifications are essential for embryogenesis and mounting evidence suggests that O-GlcNAc is a regulatory modification that affects many phosphorylated proteins. We show that the activity and expression of O-GlcNAc transferase (OGT) and N-acetyl-beta-D-glucosaminidase (O-GlcNAcase), the two enzymes regulating O-GlcNAc modifications, are present in nerve terminal structures (synaptosomes) and are particularily abundant in the cytosol of synaptosomes. Numerous synaptosome proteins are highly modified with O-GlcNAc. Although most of these proteins are present in low abundance, we identified by proteomic analysis three neuron-specific O-GlcNAc modified proteins: collapsin response mediator protein-2 (CRMP-2), ubiquitin carboxyl hydrolase-L1 (UCH-L1) and beta-synuclein. CRMP-2, which is involved in growth cone collapse, is a major O-GlcNAc modified protein in synaptosomes. All three proteins are implicated in regulatory cascades that mediate intracellular signaling or neurodegenerative diseases. We propose that O-GlcNAc modifications in the nerve terminal help regulate the functions of these and other synaptosome proteins, and that O-GlcNAc may play a role in neurodegenerative disease.     

Regulation of histamine release in rat hypothalamus and hippocampus by presynaptic galanin receptors.

The effect of galanin, a peptide present in a subpopulation of histaminergic neurons emanating from the rat posterior hypothalamus, was investigated on K(+)-evoked [3H]histamine release in slices and synaptosomes from rat cerebral cortex, striatum, hippocampus and hypothalamus. Porcine galanin (0.3 microM) significantly inhibited histamine release induced by 25 mM K+ in slices from hypothalamus and hippocampus, but not from cerebral cortex and striatum, i.e., only in regions in which a colocalization of histamine and galanin has been described. The inhibitory effect of galanin was concentration dependent, with an EC50 value of 5.8 +/- 1.9 nM. The maximal inhibition was of 30-40% in hypothalamic and hippocampal slices depolarized with 25 mM K+. The galanin-induced inhibition observed in hypothalamic slices was not prevented in the presence of 0.6 microM tetrodotoxin and also occurred in hippocampal and hypothalamic synaptosomes, strongly suggesting the activation by galanin of presynaptic receptors located upon histaminergic nerve endings. The maximal inhibitory effect of galanin in slices or synaptosomes was lower than that previously reported for histamine acting at H3-autoreceptors, possibly suggesting that not all histaminergic axon terminals, even in the hypothalamus and hippocampus, are endowed with galanin receptors. It increased progressively in hypothalamic and hippocampal synaptosomes as the strength of the depolarizing stimulus was reduced. It is concluded that galanin modulates histamine release via presynaptic receptors, presumably autoreceptors located upon nerve terminals of a subpopulation of cerebral histaminergic neurons.     

Introduction of Impermeant Molecules into Synaptosomes Using Freeze/Thaw Permeabilization

Brief freezing as a means of transiently permeabilizing synaptosomes was explored. Rat brain synaptosomes frozen and thawed in the presence of 5% dimethyl sulfoxide, a cryoprotectant, were shown to release, in a calcium-dependent manner, previously accumulated [3H]norepinephrine and [14C]acetylcholine in response to elevated [K+]. In addition, synaptosomes subjected to freeze/thaw were shown to retain their ability to exhibit resting protein phosphorylation, as well as stimulated protein phosphorylation occurring in response to calcium influx. Brief freezing of synaptosomes in the presence of [gamma-32P]ATP and either the catalytic subunit of cyclic AMP-dependent protein kinase or calcium/calmodulin-dependent protein kinase II rendered the synaptosomal interior accessible to these agents, as reflected by the phosphorylation of substrate proteins, such as synapsin I, which reside within the nerve terminal. Inclusion of inhibitors of these protein kinases during freeze/thaw blocked synaptosomal protein phosphorylation, indicating that the inhibitors were also introduced. After freezing, the synaptosomes resealed rapidly and spontaneously, as shown by the inability of any of the agents to elicit an effect on phosphorylation when added at the end of the freezing period. The permeabilization procedure should contribute to an understanding of the functional roles of phosphoproteins, and of their associated protein kinases and protein phosphatases, in nerve terminals.     

Potassium- and capsaicin-induced release of agmatine from spinal nerve terminals

Agmatine (decarboxylated arginine) was originally identified in the CNS as an imidazoline receptor ligand. Further studies demonstrated that agmatine antagonizes NMDA receptors and inhibits nitric oxide synthase. Intrathecally administered agmatine inhibits opioid tolerance and hyperalgesia evoked by inflammation, nerve injury, and intrathecally administered NMDA. These actions suggest an anti-glutamatergic role for agmatine in the spinal cord. We have previously reported that radiolabeled agmatine is transported into spinal synaptosomes in an energy- and temperature-dependent manner. In the present study, we demonstrate that agmatine is releasable from purified spinal nerve terminals upon depolarization. When exposed to either elevated potassium or capsaicin, tritiated agmatine (but not its precursor L-arginine or its metabolite putrescine) is released in a calcium-dependent manner. Control experiments confirmed that the observed release was specific to depolarization and not due to permeabilization of or degradation of synaptosomes. That capsaicin-evoked stimulation results in agmatine release implicates the participation of primary afferent nerve terminals. Radiolabeled agmatine also accumulates in purified spinal synaptosomal vesicles in a temperature-dependent manner, suggesting that the source of releasable agmatine may be vesicular in origin. These results support the proposal that agmatine may serve as a spinal neuromodulator involved in pain processing.     

Presynaptic Glutamate Receptors Regulate Noradrenaline Release from Isolated Nerve Terminals

The wide-ranging neuronal actions of excitatory amino acids, such as glutamate, are thought to be mediated mainly by postsynaptic N-methyl-D-aspartate (NMDA) and non-NMDA receptors. We now report the existence of presynaptic glutamate receptors in isolated nerve terminals (synaptosomes) prepared from hippocampus, olfactory bulb, and cerebral cortex. Activation of these receptors by NMDA or non-NMDA agonists, in a concentration-dependent manner, resulted in Ca(2+)-dependent release of noradrenaline from vesicular transmitter stores. The NMDA-stimulated release was potentiated by glycine and was blocked by Mg2+ and selective NMDA antagonists. In contrast, release stimulated by selective non-NMDA agonists was blocked by 6-cyano-7-nitroquinoxaline-2,3- dione, but not by Mg2+ or NMDA antagonists. Our data suggest that the presynaptic glutamate receptors can be classified pharmacologically as both the NMDA and non-NMDA types. These receptors, localized on nerve terminals of the locus ceruleus noradrenergic neurons, may play an important role in interactions between noradrenaline and glutamate.