The role of glutamate transporters in synaptic transmission

Although analysis of the neuromuscular junction originally suggested that the response to release of a single vesicle filled with neurotransmitter is fixed, and hence represents the elemental "quantum" of synaptic transmission (109), considerable work has now shown that quantal size can change as a function of activity, contributing to such forms of plasticity as long-term potentiation (124). Nonetheless, the locus for this regulation is postsynaptic, and involves changes in receptor number or sensitivity. More recently, it has become clear that changes in quantal size can also reflect presynaptic changes in vesicle filling. It has long been appreciated that changes in the amount of neuromodulator released per vesicle can have profound consequences for the activation of receptors at a distance from the release site. Many G protein-coupled receptors have a relatively high affinity for their peptide and monoamine ligands, but the small amounts of ligand that actually impinge on receptors are not likely to saturate binding. As a result, the release of more modulator activates more receptors, and considerable attention has focused on the regulation of quantal size for monoamines, taking advantage of electrochemical detection to measure dopamine release directly and in real time (181). It has been less clear whether changes in vesicle filling with classical transmitters such as acetylcholine, GABA and glutamate make a difference in the postsynaptic response. These transmitters are generally released in close apposition to postsynaptic receptors, many of which are ionotropic and have a high affinity for ligand (such as NMDA receptors for glutamate). If receptors are normally saturated by the contents of a single vesicle, packaging more transmitter will have no effect on the postsynaptic response. If receptors are saturated, this will tend to reduce the variation in postsynaptic response. However, quantal size exhibits considerable variation, particularly at central synapses. Although this might result from variation in the distance of different synapses from the recording electrode, due to differences in electrotonic filtering, as well as variation in release probability and the number of receptors at each synapse, a number of observations have demonstrated that variation in quantal size is intrinsic to a single synapse. Focal stimulation of one bouton, or localized dendritic recording, both show variation in quantal size similar to that observed from electrical stimulation of release from multiple boutons (17, 63, 120, 121). Increased cytosolic glutamate in the presynaptic terminal also increases quantal size at the calyx of Held in the auditory pathway (99), prov iding additional evidence against receptor saturation. Remarkably, a single vesicle filled with glutamate fails to saturate lowaffinity AMPA receptors as well as high-affinity NMDA receptors (123, 132). Consistent with this, AMPA and NMDA responses are highly correlated at individual synapses, supporting a presynaptic locus for the variation. GABA receptors at many (but not all) inhibitory synapses also appear not to be saturated by a single vesicle (14, 67, 79). How can synaptic release fail to saturate receptors? Although the concentration of transmitter achieved in the synaptic cleft is high, the receptors are closely apposed to the release site, and many are of high affinity, the peak concentration of transmitter is very brief, so that only a few receptors become activated. Regardless of the precise explanation, changes in the amount of transmitter per vesicle are thus predicted to have a major influence on the postsynaptic response. The amount of neurotransmitter released from a synaptic vesicle may be controlled either before or after the fusion event. After fusion, premature closure of the pore may interrupt the full release of vesicle contents. Indeed, the exocytosis of large dense core vesicles frequently exhibits "kiss-and-run", but this mechanism remains controversial for small synaptic vesicles, and the topic has recently been reviewed elsewhere (60, 82). This review focuses on changes in quantal size before fusion with the plasma membrane, that involve direct changes in vesicle filling. It is important to note that the mechanism of vesicular release poses several inherent problems. Large amounts of transmitter per vesicle will result in the activation of more receptors, but high rates of firing will also deplete transmitter from the terminal unless it is actively replaced by, for example, recycling or biosynthesis. At the same time, vesicular transport is generally slow, and may limit refilling if vesicles recycle quickly, even at concentrations of cytosolic transmitter that saturate the transport mechanism. Subsaturating cytosolic concentrations will further slow refilling and release. However, low cytosolic concentrations may be important to prevent the oxidation and toxicity of monoamines such as dopamine (136), and this is compensated for by the ability of the vesicular monoamine transporter to generate an extremely large concentration gradient up to 105 higher in the lumen than the cytoplasm. Other classical transmitters including glutamate produce toxicity through a specific interaction with cell surface receptors, and can therefore be tolerated at higher levels in the cytoplasm. This reduces the magnitude of the concentration gradient required to fill vesicles with glutamate, and presumably also speeds filling. We will therefore consider now the factors that influence vesicle filling with glutamate, from its cytosolic concentration to the H+ electrochemical gradient that drives transport, glutamate transport itself, and finally, the physiological regulation of these mechanisms and their role in synaptic plasticity. The amount of transmitter achieved inside secretory vesicles indeed reflects the cytosolic concentration of transmitter, the driving force, transport into and non-specific leakage across the vesicle membrane.