Quantal size fits central synaptic depression

In this issue of PNAS, Chen et al. (1) report the first direct electrophysiological recordings from single visualized excitatory synapses and further define the elementary properties of chemical synaptic transmission. They show that single synapses of hippocampal neurons typically release one quantum of neurotransmitter and that presynaptic regulation of the size of this quantum contributes to short-term synaptic depression.

Neural networks are defined by synaptic connections among neurons, typically chemical synapses, at which electrical signals are transmitted and received. The basic principles of chemical synaptic transmission were defined in pioneering studies of the neuromuscular junction and subsequently extended to synapses of the central nervous system (2–4). As illustrated in Fig. 1, axons often branch extensively before contacting postsynaptic targets where they form terminal compartments referred to as boutons. Presynaptic boutons contain many neurotransmitter-filled synaptic vesicles including a small, readily releasable pool docked at specialized plasma membrane structures called active zones (Fig. 1). After excitation of the presynaptic neuron, vesicle fusion and exocytosis of neurotransmitter is triggered by calcium influx through voltage-gated calcium channels. Subsequent activation of neurotransmitter-gated ion channels clustered within the postsynaptic density results in excitation or inhibition of the postsynaptic membrane. The vesicle hypothesis of neurotransmitter release provides a structural basis for the early observation that chemical synaptic transmission is quantal. One quantum likely corresponds to release of neurotransmitter from a single synaptic vesicle.

Fig. 1.

Overview of synaptic structure and identification of isolated visualized single synapses. A branched axon, presynaptic boutons, and single-synapse structure are illustrated. In cultured hippocampal neurons, most boutons form only one synapse. Chen et al. visualized synapses by loading synaptic vesicles with the membrane-associated fluorescent dye FM 1-43 (green). To avoid inclusion of adjacent boutons or synapses, spatially isolated FM 1-43 puncta exhibiting moderate fluorescence intensity were selected for single-synapse stimulation and analysis.

The functional strength of a synaptic connection, defined as the amplitude of the postsynaptic response to a presynaptic stimulus, exhibits a marked dependence on neuronal activity. This plasticity makes a fundamental contribution to neural function by storing a transient or sustained record of previous activity. Here our focus is on short-term synaptic plasticity in which synaptic strength is transiently increased (facilitation) or decreased (depression) during repetitive activity (5).

Chen et al. address both the quantal nature of synaptic transmission and its regulation during short-term synaptic plasticity. This study builds on a long history of intensive investigation by providing direct electrophysiological analysis of single excitatory synapses and consequently sheds new light on the elements of synaptic transmission. In considering single-synapse recordings, it is important to recognize that the term “synapse” refers to each assembly of a presynaptic active zone and postsynaptic density (Fig. 1). Thus, one axon may produce many boutons, and each bouton may form one or more synapses. Single-synapse recordings were achieved by first loading synaptic vesicles with a membrane-associated fluorescent dye to visualize vesicle clusters within presynaptic boutons (Fig. 1) and then using focal stimulation to excite a single isolated bouton while recording the resulting postsynaptic current. Cultured hippocampal neurons are well suited for these studies, because systematic ultrastructural analysis (6) indicates that ≈70% of their presynaptic boutons contain only a single synapse. Thus, focal stimulation of single boutons may be used to examine unitary synaptic activity. Analogous methods have been used to examine short-term synaptic plasticity at single inhibitory γ-aminobutyric acidergic boutons (7).

With the resolution provided by single-synapse electrophysiology, Chen et al. first addressed an important and controversial question regarding the elementary properties of synaptic transmission: Is neurotransmitter release from a single synapse limited to one quantum, or can multiple vesicles fuse in response to a presynaptic stimulus? The “one-vesicle hypothesis” (3) suggests that the synapse, and in the case of hippocampal neurons the entire bouton, may signal the postsynaptic membrane in a binary fashion by releasing predominantly 1 or 0 quanta. A thoughtful discussion of this hypothesis, the current debate regarding monovesicular versus multivesicular release, and the technical challenges on both sides of the issue may be found in a recent review by Stevens (4). The experiments of Chen et al. support the one-vesicle hypothesis by demonstrating that a maximum of one quantum was recorded at single synapses even under conditions producing a high probability of neurotransmitter release. These experiments provide direct confirmation of the monovesicular release observed previously by using minimal stimulation in hippocampal slices (8). It seems this mode of release is not universal, because electrophysiological studies have demonstrated robust multivesicular release at climbing fiber synapses in the cerebellum (9); however, Chen et al. provide an example of a predominantly monovesicular and thus binary mode of transmission at the single-synapse level. As discussed above, focal stimulation was restricted to isolated boutons exhibiting moderate-intensity FM 1-43 fluorescence. These were selected because regions of high-intensity fluorescence yielded multivesicular release, perhaps because they represent adjacent boutons or synapses. This observation may be relevant to recent imaging studies demonstrating multivesicular release at hippocampal synapses through quantitative analysis of postsynaptic calcium signals (10). Additional experiments are needed to address whether, as suggested by Chen et al., selection of putative single synapses on the basis of fluorescence intensity may help to reconcile the conflicting conclusions of these studies.

Chen et al. went on to study short-term plasticity at the single-synapse level by examining facilitation and depression elicited by paired-pulse stimulation. In these experiments, paired stimulation pulses were delivered to single synapses and the resulting excitatory postsynaptic currents (EPSCs) averaged over many trials. For both the first and second stimulus, these experiments examined the frequency of success (1 quantum) and failure (0 quanta), which define the probability of release from that synapse, as well as the quantal size. Under conditions promoting a high release probability, paired-pulse depression was observed consistently. It is important to note here that depression was first measured as a reduction in the median amplitude of the second EPSC relative to the first, including failures. Thus, a simple reduction in release probability (increased failures) with no change in quantal size would produce a corresponding reduction in the median EPSC amplitude. In fact, previous studies have attributed paired-pulse depression at these and other synapses to such a change in release probability, which may reflect depletion of the readily releasable pool and/or a reduced probability that vesicles within this pool will release neurotransmitter (5). Although Chen et al. confirmed a component of depression resulting from reduced release probability, they also made a surprising discovery. Analysis of median EPSC amplitudes excluding failures indicated that the EPSC evoked by the second stimulus reflected a presynaptic reduction in quantal size, suggesting a novel mechanism of short-term synaptic depression.

Excitatory postsynaptic currents evoked by a second stimulus reflect presynaptic reduction in quantal size.

The observed quantal size reduction during paired-pulse depression did not depend on whether the first stimulus succeeded or failed to elicit neurotransmitter release. Thus, postsynaptic receptor desensitization cannot account for this effect. Rather, the underlying mechanism seems to be presynaptic and also independent of prior synaptic vesicle fusion. Chen et al. emphasize that a variety of mechanisms may explain the observed quantal size reduction, including those regulating the quantity of neurotransmitter within a vesicle, such as vesicular glutamate transport (11), and the extent or kinetics of neurotransmitter release after vesicle fusion. With respect to the latter, partial release of vesicle contents is well documented in studies of regulated exocytosis from dense-core vesicles (12). This occurs during a mode of vesicle fusion referred to as “kiss and run,” in which a transient fusion pore allows release without full collapse of the vesicle into the plasma membrane (13, 14). Recent imaging of synaptic vesicle trafficking has provided the strongest and most direct evidence for kiss-and-run exocytosis at synapses (15, 16); however, it remains unclear whether this process may restrict neurotransmitter release from a single vesicle. These considerations fit within an emerging theme in the study of short-term synaptic plasticity involving the coordination and interaction of mechanisms operating in synaptic vesicle exocytosis and endocytosis (16–20). Future studies will further define the extent to which such interactions influence the size of the readily releasable vesicle pool, release probability, and quantal size. Surely the findings of Chen et al., in providing strong support for a predominantly monovesicular release mode and a role for quantal size regulation in short-term depression, will make a stimulating contribution.