Neurotransmitter diffuses out of a synaptic vesicle into the synaptic cleft

Release of neurotransmitter (Fig. 7.4)

Once loaded with neurotransmitter, synaptic vesicles are docked at the presynaptic membrane awaiting release. Docking takes place at active zones. These consist of multi-protein complexes that tether the synaptic vesicle to the presynaptic membrane and contain high concentrations of voltage-gated calcium channels. This leads to brisk calcium influx in response to axon terminal depolarization, since the concentration of calcium is 10,000 times higher in the extracellular fluid. The focal rise in free calcium causes a number of synaptic vesicles to fuse with the presynaptic membrane, emptying their contents into the synaptic cleft by exocytosis.

Release is said to be ‘packeted’ (or quantized) since the total amount of transmitter entering the synaptic cleft is a whole-number multiple of the amount stored in a single vesicle. Transmitter release is regulated by presynaptic autoreceptors which exert negative feedback.

B Sequences of Synaptogenesis

The respective sequences of appearance of the pre- and postsynaptic membranes, and the differential roles of the membrane thickenings and synaptic vesicles in the elaboration of the presynaptic contact area are depicted in Fig. 80.

When both pre- and postsynaptic densities are present early in development, as in Fig. 8a, they appear as continuous plaques (Aghajanian and Bloom, 1967; Jones and Revell, 1970a,b) which, later in development, undergo gradual differentiation and focalization to produce the typical paramembranous densities of the adult synaptic junction. Early workers repeatedly stressed the presence of membrane thickenings prior to the appearance of synaptic vesicles (Glees and Sheppard, 1964; Hamori and Dyachkova, 1964; Meller, 1964; Bunge et al., 1967; Larramendi, 1969; Tennyson, 1970; Alley, 1973; Stelzner et al., 1973). In all these instances, however, caution is necessary to ensure that the membrane specializations actually are precursors of synaptic junctions rather than early components of desmosome-like contacts. This also applies to the idea that puncta adhaerenta are precursors of synaptic junctions (McGraw and McLaughlin, 1980).

An alternative, therefore, is Fig. 8b with synaptic vesicles preceding membrane thickenings. While the presence of presynaptically located vesicles may not constitute an obligatory feature of immature synapses, it remains a helpful one and should not be lightly discarded. Unfortunately, once the presence of vesicles becomes an essential diagnostic feature of early synapses, it also to some extent prejudges the question of whether vesicles or membrane thickenings appear first in synaptic development.

A third alternative is Fig. 8c, in which increases in vesicle numbers and maturation of the paramembranous densities are simultaneous events (Ochi, 1967; Bodian, 1968; Jones and Revell, 1970a,b). For instance, Jones and Revell noted in rat cerebral cortex that vesicles are present when both synaptic membranes are plaquelike and undifferentiated, suggesting that synaptic vesicles appear with (or even after) undifferentiated membrane thickenings but before their differentiation into recognizable dense projections. Comparable observations have been made in other situations, including chick retina (McLaughlin, 1976) and kitten spinal trigeminal nucleus (Dunn and Westrum, 1978).

The discussion this far has assumed that the pre- and postsynaptic membranes appear concurrently (Aghajanian and Bloom, 1967; Jones and Revell, 1970a,b; Adinolfi, 1972a,b; Bloom, 1972). However, this is far from a universal phenomenon. In some studies, as shown in Fig. 8d, the postsynaptic density appears as a well-established entity before the presynaptic region with its dense projections (Poppe et al., 1973; Jones et al., 1974; West and Del Cerro, 1976; Hinds and Hinds, 1976). This phenomenon is best seen in E-PTA-stained material and has so far been described in developing guinea pig cerebral cortex (Jones et al., 1974) and in some synaptic junctions in fetal rat cerebellum (West and Del Cerro, 1976). It is possible that this appearance reflects features of the E-PTA staining reaction, in that it is not staining the presynaptic components of at least some of the immature synapses. Even if this is the case, however, other evidence points to the legitimacy of “naked” postsynaptic sites during early neuronal development.

The converse arrangement, whereby the presynaptic membrane appears prior to the postsynaptic, is far less frequently encountered (Fig. 8e). A commonly quoted illustration of this is the membrane-vesicle clusters described in the cervical spinal cord of Xenopus laevis embryos (Hayes and Roberts, 1973). A similar sequence of events has been described as part of the pattern for photoreceptor synaptogenesis in developing chick retina (McLaughlin, 1976).

These diverse developmental patterns reflect the range of ways by which pre- and postsynaptic junctional elements establish their adult organization. It is not known whether a particular pattern characterizes a particular synaptic population, or whether similar synapses are put together in different ways under differing environmental pressures.

It has long been argued that axodendritic synapses are formed prior to axo-somatic ones (Voeller et al., 1963; Jones, 1975), the synaptic vesicles being spherical rather than flattened (Oppenheim and Foelix, 1972). More recent studies substantiate these general principles in a wide range of situations, e.g., chick optic tectum (McGraw and McLaughlin, 1980), human cervical spinal cord (Okado, 1980), and mouse spinal motor neurons (Vaughn et al., 1977). Occasional exceptions do occur however, including the study of Smolen and Raisman (1980) on rat superior cervical ganglion. In this instance, axosomatic synapses are present transiently for the first 2 weeks of life. It is possible that such synapses may facilitate dendritic development by providing directive cues to the location of the developing synaptogenic field (Vaughn et al., 1977).

Early axodendritic synapses may be of two types: those with asymmetrical junctions and spherical vesicles, and those with symmetrical junctions and spherical vesicles (Devon and Jones, 1981). Neocortical development is characterized by a decrease in the symmetrical/spherical group. This trend may represent a developmental continuum of symmetrical/spherical to asymmetrical/spherical synapses (Bunge et al., 1967; Cragg, 1972; Hinds and Hinds, 1976; Dunn and Westrum, 1978), although the possibility that the asymmetrical/axodendritic terminals represent an intermediate synaptic type cannot be excluded (Adams and Jones, 1982a). The recognition of the symmetrical/spherical terminal type may provide a means of distinguishing immature terminals in adult, and especially in retarded adult, cortical tissue (Devon and Jones, 1981). In this connection it would be interesting to know whether these symmetrical/spherical terminals correspond to the immature type E synaptic junction described by Dyson and Jones (1976a) in E-PTA-stained cortical material.

Terminals with spherical vesicles are repeatedly described as being the first to appear, with flattened or pleomorphic vesicles appearing in later-developing terminals, e.g., in chick optic tectum (McGraw and McLaughlin, 1980), human cervical spinal cord (Okado, 1980), and rabbit superior colliculus and visual cortex (Mathers et al., 1978).

Quantitative approaches to the morphological categorization of developing synaptic junctions are currently confined to the characterization of E-PTA-stained junctions. The quantitative morphogenetic scheme of Dyson and Jones (1976a) assigns synaptic junctions to five categories, A–E, on the basis of variations in the organization of their presynaptic densities. Of these descriptive characteristics, type A represents a mature junctional form with well-defined and discrete dense projections, whereas, at the other end of the scale, type E is an immature form lacking recognizable presynaptic densities; types B–D probably represent intermediate forms (Fig. 3).

The approach to synaptogenesis illustrated by this study gives some idea of the manner in which synaptic ultrastructure can be traced at varying stages during maturation and adult life. By permitting quantitation of the contribution a particular morphogenetic form makes to the population at developmental intervals, it provides a model system for assessing the maturity of both individual junctions and synaptic populations.

Structure and Mechanism of Action of BTs:

BTs are synthesized as single-chain polypeptides and are activated by proteolytic enzymes in a cleaving process (Fig. 18-4). The cleaved heavy chain is responsible for high-affinity docking of the neurotoxin to the presynaptic nerve terminal receptor, enabling the internalization of the bound toxin into the cell.50 The light chain is a zinc-dependent endopeptidase that cleaves membrane proteins responsible for docking acetylcholine vesicles on the inner side of the nerve terminal membrane. The cleavage of these proteins irreversibly precludes fusion of the vesicles with the nerve membrane, thereby preventing release of neurotransmitters into the neuromuscular junction.51

Type A BT appears to be the most potent of the subtypes, and, when injected clinically, has the longest duration of action. While type B and F have limited clinical use, and others are the subject of further study, the multiple differences thus far observed suggest that the subtypes are not interchangeable.

Pharmacological effect of BTs occurs in three stages: binding, internalization, and proteolysis.52

Synapses

Effective messages move along an axon and are transferred to another cell, occurring at a synapse, a specialized area where a neuron communicates with another cell. Synapses are also referred to as synaptic junctions. At a synapse, information moves from the presynaptic neuron to the postsynaptic neuron. Synapses can involve various postsynaptic cells. An example is the neuromuscular junction, a synapse in which the postsynaptic cell is a skeletal muscle fiber. There are more than 100 trillion synapses in the human brain alone. The types of synapses in the CNS are summarized in Table 2.4.

Table 2.4. Types of synapses in the central nervous system.

Classification

There are two types of synapses: electrical and chemical, with unique functions. At electrical synapses, there is direct physical contact between cells. Presynaptic and postsynaptic membranes of cells are joined at gap junctions (see Fig. 2.6A). Lipid areas of nearby membranes are separated by just two nanometers (nm), kept in position by essential membrane proteins called connexons. Pores formed by these proteins allow ions to pass between the cells. Changes in membrane potential of one cell cause local currents to affect another. An electrical synapse propagates action potentials between cells efficiently and quickly because of this. In adults, electrical synapses are not common in the CNS or PNS. They occur in brain areas such as the vestibular nuclei, involved in balance, also in the eyes, and in one or several pairs of PNS ganglia (ciliary ganglia). They are also present in embryonic structures.

Figure 2.6. (A) Electrical synapse; (B) Chemical synapse.

In a chemical synapse (Fig. 2.6B), one neuron signals another. This uses the axon terminal of the presynaptic neuron, which sends the message, and the postsynaptic neuron, which receives it. The cells are separated by the narrow synaptic cleft. A presynaptic cell is most often a neuron. Specialized receptor cells create synaptic connections with dendrites. Postsynaptic cells are neurons or other forms of cells. Communications between neurons occur at synapses on dendrites, cell bodies, or along axons of receiving cells. Axoaxonic synapses are between axons of two neurons. Axosomatic synapses have junctions, at an axon terminal of a neuron, and the cell body of another neuron. In an axodendritic synapse, synaptic contact is between an axon terminal of one neuron and a dendrite of another. Chemical synapses are very common compared to electrical synapses. Communications across chemical synapses occur from presynaptic membranes to postsynaptic membranes, and not in the reverse direction.

A neuromuscular junction is a synapse between a neuron and a skeletal muscle cell. At a neuroglandular junction, a neuron regulates activities of a secretory cell. Neurons innervate many other types of cells, such as adipocytes. Axon terminals of presynaptic cells release neurotransmitters into the synaptic cleft. The neurotransmitters are within synaptic vesicles in the axon terminal. Synaptic vesicles contain a small collection, or quanta, of neurotransmitter. When the terminal is depolarized from an action potential of its “parent” axon, there is a calcium influx that leads to phosphorylation of proteins called synapsins. After this phosphorylation, the vesicle links to the presynaptic membrane that faces the synaptic cleft, and the neurotransmitter is released. Another method of neurotransmitter release is based on transporter molecules, which usually act by taking up neurotransmitters from the synaptic cleft.

Each postsynaptic cell has its own type of axon terminal. A round axon terminal is present if a postsynaptic cell is another neuron. At a neuromuscular junction, the axon terminal is of much more complex. Axon terminal structures include mitochondria and thousands of vesicles containing neurotransmitters. Axon terminals reabsorb breakdown molecules of neurotransmitters at the synapse, then reproduce the neurotransmitters. Axon terminals also constantly receive neurotransmitters created by the cell body, along with enzymes and lysosomes, via anterograde flow.

C Nucleation of Synaptic Scaffolds by Adhesion Molecules

The mechanical stability of synapses and the precise apposition of pre- and postsynaptic membrane domains implies that adhesion molecules might bridge the synaptic cleft and thereby keep both sides of the synapse in register. Ultrastructural studies revealed the presence of different sites of cell-cell adhesion at the synapse—including the linkage across the synaptic cleft, as well as at puncta adherentia that flank the synaptic cleft. A third site of cellular adhesion at central synapses are the end feet of glial cells that tightly enclose the synaptic junctions.

Most attention in recent years has been focused on neuronal adhesion molecules that directly bridge the synaptic cleft and that might instruct the assembly of synaptic structures. Several such adhesion molecules have been identified. The synaptogenic activity of these adhesive factors is thought to rely on the nucleation of cytoplasmic scaffolds at the pre- and postsynaptic side, which subsequently facilitate the recruitment of other cytoplasmic and cell surface molecules to an incipient synaptic connection.

In vitro experiments with cultured neurons revealed that exposing axonal or dendritic processes to specific adhesive triggers is sufficient to induce pre- and postsynaptic structures. One extensively studied adhesion system with synapse-inducing function is the neuroligin-neurexin complex. Neuroligins and neurexins are heterophilic adhesion proteins that form a trans-synaptic complex (Ichtchenko et al. 1995; Nguyen and Südhof 1997; Song et al. 1999). Cytoplasmic sequences in neuroligin-1 target this protein exclusively to the synaptic sites in dendrites and ensure exclusion from the axons (Dresbach et al. 2004; Iida et al. 2004; Rosales et al. 2005). This strict polarity of neuroligin-1 provides this heterophilic adhesion system with the required directionality for the assembly of asymmetric synaptic structures. Contact of axons with neuroligin-1 ectopically expressed in a nonneuronal cell induces clustering of axonal neurexins and the cytoplasmic scaffolding molecule CASK (Scheiffele et al. 2000; Dean et al. 2003). These neuroligin-induced axonal contacts then mature into functional presynaptic release sites, including active zone components and clusters of synaptic vesicles. Conversely, contact of dendrites with neurexin-1 expressing cells leads to the formation of neuroligin clusters and postsynaptic structures, such as the accumulation of the scaffolding protein PSD95 and NMDA receptors (Graf et al. 2004). These experiments suggest that the neuroligin-neurexin adhesion system can nucleate pre- and postsynaptic structures. To a large extent these structures are thought to assemble around scaffolding components that bind to the cytoplasmic tails of neuroligins and neurexins (see Figure 4.1). For example, postsynaptic neuroligin-1 binds to PSD95, which in turn can recruit NMDA-receptors and other scaffolding components such as GKAP, Homer, and Shank (Irie et al. 1997; Meyer et al. 2004; Prange et al. 2004; Chih et al. 2005).

Interestingly, the neuroligin-2 isoform, which differs in its cytoplasmic sequences from neuroligin-1, is found primarily at GABAergic synapses. Extracellular aggregation of neuroligin-2 results in not only the recruitment of PSD95, but also of gephrin, a scaffolding protein specific to inhibitory synapses (Graf et al. 2004; Varoqueaux et al. 2004). Selective interaction of excitatory and inhibitory postsynaptic scaffolding proteins with different neuroligin cytoplasmic tail sequences therefore might contribute to the nucleation of specific postsynaptic scaffolds. However, interactions between extracellular domains that may contribute to this process have not been ruled out.

Similar experiments revealed synaptogenic activities for the cell adhesion molecule SynCAM, a homophilic Ig-domain protein (Biederer et al. 2002). Interestingly, SynCAM and neuroligin-1 differ in their cytoplasmic interactions with scaffolding proteins. These differences may explain the differential recruitment of NMDA and AMPA-type glutamate receptors to neuroligin-1- and SynCAM-induced synapses, respectively (Sara et al. 2005). Neuroligin-1 recruits PSD95 and NMDA-receptors through the type I PDZ-binding motif in neuroligin, whereas SynCAM contains a type II PDZ-binding motif that contributes to the recruitment of functional AMPA-receptors to synapses. The importance of these selective scaffolding interactions was confirmed in experiments showing that a chimaeric protein, which includes neuroligin-1 containing the cytoplasmic sequences of SynCAM, is capable of stimulating AMPA receptor recruitment (Sara et al. 2005).

Another presumptive cell adhesion molecule that has been specifically linked to the synaptic recruitment of AMPA receptors through interaction with scaffolding molecules is Dasml (Shi et al. 2004). However, the extracellular binding partners of this protein are currently unknown.

It is likely that several adhesion molecules that act through mechanisms similar to the neuroligin-neurexin complex and SynCAM will cooperate in the nucleation of synaptic scaffolds and the subsequent recruitment of neurotransmitter receptors. Candidate molecules include the nectins, a family of homophilic Ig-domain proteins that are similar to SynCAM (Satoh-Horikawa et al. 2000; Mizoguchi et al. 2002). The integration of multiple adhesive signals likely occurs at the level of the cytoplasmic scaffold. There is substantial crosstalk between different scaffolding components; for example, the CASK/Mint/MALS complex that binds to neurexins is linked to beta-catenin, the cytoplasmic binding partner of cadherins (see Figure 4.1; Hata et al. 1996; Perego et al. 2000; Bamji et al. 2003). Such crosstalk might be employed for the sequential recruitment of different adhesive factors to a growing adhesion site as has been observed for cadherins and nectins in epithelial cells (Takahashi et al. 1999; Tachibana et al. 2000).

When does Ca2+ enter the presynaptic element in response to a presynaptic spike?

In order to record the presynaptic Ca2+ current, the presynaptic membrane is clamped at VH = −80 mV by means of two whole-cell electrodes. To isolate the presynaptic Ca2+ current, voltage-dependent Na+ and K+ currents are blocked with TTX and tetraethlyammonium chloride (TEA), respectively. An action potential waveform is injected into the presynaptic terminal through one of the whole-cell electrodes. It evokes a presynaptic Ca2+ current that is recorded by the second whole-cell electrode. Recordings show that Ca2+ influx is tightly associated with the repolarizing phase of the action potential (Figure 7.4): it is essentially a tail current (see Appendix 5.2) that activates shortly after the peak of the action potential and ends before repolarization is complete. It has a peak amplitude of 2.6 ± 0.2 nA and a half-width of about 350 ms. The delay between the beginning of the action potential and that of Ca2+ current is about 500 μs at 23–24°C.

Acetylcholine Is Degraded and Then Recycled

Each action potential fuses perhaps 300 or so synaptic vesicles with the presynaptic membrane. Continuous excitation and subsequent fusion of these large numbers of vesicles with the presynaptic membrane would expand the area of the membrane and deplete it of vesicles. To avoid this, the presynaptic cell rapidly recycles the vesicles. Through an elaborate mechanism involving specialized proteins such as clathrin and dynamin, vesicles are pinched off and endocytosed. The empty vesicles are refilled with acetylcholine through a transport channel that exchanges acetylcholine for H+ ions. The H+ ions are accumulated within the vesicles through a vacuolar H-ATPase in the vesicle membrane.

Acetylcholine is also partially recycled. Acetylcholine is degraded to acetate and choline in the synapse through the action of acetylcholinesterase. The acetate is not recycled, but the choline is taken back up into the presynaptic cell through a transporter on the presynaptic cell membrane. In the cell, acetylcholine is resynthesized from acetyl CoA and choline through the enzyme choline acetyltransferase. The events occurring in the axon terminal that initiate and shut off neuromuscular transmission are shown in Figure 3.6.4.

Types of synapses

There are two types of synapses, electrical and chemical, which function in unique ways. It is important to understand the differences between these two types.

Electrical synapses

At electrical synapses, the cells have direct physical contact. The presynaptic and postsynaptic membranes of the two cells are joined at gap junctions (see Fig. 1.6A). The lipid parts of adjacent membranes are separated by only 2 nm. They are positioned by essential membrane proteins called connexons. Pores, formed by these proteins, allow ions to pass between the cells. Changes in membrane potential of one cell result in local currents affecting the other cell, much as if they shared a common membrane. Therefore, an electrical synapse propagates action potentials between the cells efficiently and quickly.

Fig. 1.6. (A) Electrical synapse. (B) Chemical synapse.

In adults, electrical synapses are not common in both the CNS and PNS. They occur in certain brain areas such as the vestibular nuclei, involved in balance, in the eyes, and in one or more pairs of PNS ganglia (ciliary ganglia). They also occur in certain embryonic structures.

Introduction

Chemical signaling at a synapse occurs when a synaptic vesicle fuses with the presynaptic membrane in response to calcium influx through voltage-gated calcium channels. Vesicle fusion results in the formation of a fusion pore through which neurotransmitter packaged in the vesicle can escape into the synaptic cleft. The released neurotransmitter then diffuses through the synaptic cleft and binds to ligand-gated receptors on the postsynaptic membrane, which then lead to current flow across the postsynaptic membrane. To understand synaptic transmission, it is necessary to determine the spatiotemporal pattern of channel opening. However, physical limitations preclude direct observation of the dynamics of neurotransmitter release and receptor activation at single synapses. With the vast amount of biophysical and structural data that are now available, it is possible to make constrained models with sufficient biological realism that can be used to answer the following questions about the generation of synaptic responses:

How does the structure of the synapse shape the response?

How does the biophysics of receptor activation contribute to the response?

What is the influence of the kinetics of release of neurotransmitter and its reuptake or degradation?

We first discuss modeling methods that can be used for the study of synapses and then discuss the application of these methods to the study of excitatory transmission at the frog neuromuscular junction and at central glutamatergic synapses.

Synapse Structure

A synapse is most commonly understood as an individual contact between electron-dense pre- and postsynaptic membranes. This is most generally a small contact area with rapid diffusional loss of neurotransmitter from the synapse, although the synapse may have a more complex topography in which the neurotransmitter is temporarily trapped. An extreme example of the latter may be found at the unipolar brush cell in the cerebellum, but similar examples may be found in sensory inputs (e.g., to the olfactory bulb). Synaptic responses from recordings between pre- and postsynaptic neurons typically represent a summed synaptic current from a number of such release sites.

The structure of the axon may profoundly affect presynaptic inhibitory mechanisms. Although many synapses exist as varicosities along a single axon trunk, others are found at the terminations of elaborate axon branches (Figure 1). The complexity of these structures may impede the action potential invasion of distal synaptic active zones. Depending on the location within the structure of the axon, presynaptic inhibition might therefore selectively reduce or completely prevent action potential depolarization of active zones within distinct axon branches.