Chemical transmission between nerve cells involves multiple steps

Until the late nineteenth century, many physiologists believed that there were direct physical connections between nerves and that an impulse from one nerve was communicated to another through a direct physical connection. However, studies by Golgi, Ramon y Cajal and others convinced many histologists that most connections, which we now know as synapses, were close but not continuous. The pioneering work of Oliver and Shäfer, of Langley and of Elliot beginning in the 1890s provided data that raised the possibility of chemical transmission between nerves. Chemical transmission was convincingly demonstrated in the historic experiments of Otto Loewi. He electrically stimulated the vagus nerve of an isolated frog heart to decrease the strength and rate of contractions. The bathing solution caused a decrease in the strength and rate of contractions when subsequently applied to a second heart. We now know that the inhibition was caused by the neurotransmitter acetylcholine (ACh) that had been released by the nerve terminals of the vagus nerve. (See Davenport [1] for an entertaining and excellent review of the early history of chemical transmission.)

Chemical transmission is the major means by which nerves communicate with one another in the nervous system. The pre- and postsynaptic events are highly regulated and subject to use-dependent changes that are the basis for plasticity and learning in the CNS. Although direct electrical connections also occur, these account for transmission of information between nerves only in specialized cases.

Chemical transmission requires the following steps:

1.

Synthesis of the neurotransmitter in the presynaptic nerve terminal.

2.

Storage of the neurotransmitter in secretory vesicles.

3.

Regulated release of neurotransmitter in the synaptic space between the pre- and post-synaptic neurons.

4.

The presence of specific receptors for the neurotransmitter on the postsynaptic membrane, such that application of the neurotransmitter to the synapse mimics the effects of nerve stimulation.

5.

A means for termination of the action of the released neurotransmitter.

An overview of some of the processes involved in synaptic transmission is shown in Figure 10-1. Many of the processes are discussed below or in other chapters of this book. Many different types of substances are neurotransmitters. “Classical” neurotransmitters, such as ACh (Chap. 11) and norepinephrine (NE) (Chap. 12), are low-molecular-weight substances that have no other function but to serve as neurotransmitters. The predominant excitatory neurotransmitter in the brain, glutamate, and the inhibitory neurotransmitter in the spinal cord, glycine, are common and essential amino acids (Chaps. 15 and 16). They can function as neurotransmitters because the membranes of secretory vesicles in glutamatergic and glycinergic nerve terminals have specific transport systems that concentrate and store these amino acids so that they can be released by exocytosis in a highly regulated manner. Aminergic neurotransmitters, ACh and GABA, the predominant inhibitory amino acid in brain, also enter synaptic vesicles through specific transport proteins. Synaptic vesicles have an acidic interior, pH ~5.5, which is maintained by a vacuolar-type, proton-translocating ATPase (see Chap. 5). The uptake of low-molecular-weight neurotransmitters is coupled via the transporters to the electrochemical H⁺ gradient (for reviews, see [2,3]).

Figure 10-1

Depolarization opens voltage-sensitive Ca^{2 +} channels in the presynaptic nerve terminal. The influx of Ca²⁺ and the resulting high Ca²⁺ concentrations at active zones on the plasma membrane trigger the exocytosis of small synaptic vesicles that store (more...)

Neurotransmitter release is a highly specialized form of the secretory process that occurs in virtually all eukaryotic cells

The fundamental similarity between the events in the nerve terminal that control neurotransmitter release and the ubiquitous vesicular trafficking reactions in all eukaryotic cells is described in Chapter 9. This similarity has important implications for the biochemistry of synaptic transmission. Many of the proteins essential for constitutive secretion and endocytosis in yeast and mammalian cells are similar to those involved in the presynaptic events of synaptic transmission (Chap. 9).

Peptides and proteins can also be released from nerve terminals. Their biosynthetic and storage processes are similar to those in other protein secretory cells [4]. They utilize the endoplasmic reticulum, Golgi and trans-Golgi network, which are present in the cell body but not in the nerve terminal. The peptide- and protein-containing vesicles must be transported into the nerve terminal by axonal transport (Chap. 28). Examples of peptide neurotransmitters include substance P, thyrotropin-releasing hormone (TRH), vasopressin, oxytocin, enkephalins and endorphins (endogenous opiate-like agonists), vasoactive intestinal peptide (VIP) and luteinizing hormone—releasing hormone (LHRH) (Chap. 18). The interior of mature secretory granules in neuroendocrine cells in the regulated pathway has a pH of 5.3 to 5.5, similar to the pH in synaptic vesicles. The low pH influences intravesicular protein processing and the conformation of the stored proteins, as well as the transport of other substances into the granules.

A variety of methods have been developed to study exocytosis

Neurotransmitter and hormone release can be measured by electrical effects of released neurotransmitter or hormone on postsynaptic membrane receptors, such as the neuromuscular junction (NMJ) (see below), and directly by biochemical assay. Another direct measure of exocytosis is the increase in membrane area due to the incorporation of the secretory granule or vesicle membrane into the plasma membrane. This can be measured by increases in membrane capacitance (C _m). C_m is directly proportional to membrane area and is defined as: C _m = QA _m/V, where C _m is the membrane capacitance in farads (F), Q is the charge across the membrane (coulombs), V is voltage (volts) and A _m is the area of the plasma membrane (cm²). The specific capacitance, Q/V, is the amount of charge that must be deposited across 1 cm² of membrane to change the potential by 1 volt. The specific capacitance is mainly determined by the thickness and dielectric constant of the phospholipid bilayer membrane and is similar for intracellular organelles and the plasma membrane. It is approximately 1 μF/cm². Therefore, the increase in plasma membrane area due to exocytosis is proportional to the increase in C _m.

The electrophysiological technique used to measure changes in membrane capacitance is the patch clamp [5,6] in the whole-cell recording mode, where the plasma membrane patch in the pipet is ruptured. In another configuration of the patch clamp, the plasma membrane patch is maintained intact. In this case, small currents due to the opening of individual channels can be measured in the membrane patch. The whole-cell patch clamp technique establishes a high resistance seal between the glass rim of the micropipet and the plasma membrane that allows low noise, high sensitivity, electrical measurements across the entire plasma membrane. An example of the use of membrane capacitance to measure exocytosis in chromaffin cells is shown in Figure 10-2 [7].

Figure 10-2

Secretory events monitored by simultaneous amperometric (I _amp ) and capacitance (C _m ) measurements demonstrate typical patterns of release. A: Configuration of the recording setup. B: shows a wide amperometric response composed of multiple spikes due (more...)

Sensitive electrochemical techniques have also been developed to directly measure the release of oxidizable neurotransmitters such as catecholamines (CAs) and serotonin (5-hydroxytryptamine; 5-HT). Current flows in the circuit when the potential of the electrode is positive enough to withdraw electrons from, that is, oxidize, the released neurotransmitter. The technique is very sensitive and readily detects the release of individual quanta of neurotransmitter resulting from the fusion of single secretory granules (Fig. 10-2).

A variety of different types of tissue preparations are used to study neurosecretion and synaptic transmission. A classical preparation is the frog NMJ (discussed below). The brain slice has been used for many years for biochemical studies of CNS metabolism and is a useful preparation for electrophysiological studies of synaptic transmission in the CNS. Slices can be oriented to maintain the local neuronal circuitry and can be thin, ~0.3 mm, to minimize anoxia. The transverse hippocampal slice is widely used as an electrophysiological preparation to study synaptic plasticity (see Chap. 50). Primary cultures of neurons from selected CNS areas and sympathetic ganglia are also frequently used. They permit excellent visual identification of individual neurons and control of the extracellular milieu, but the normal neuronal connections are disrupted.

Gentle homogenization of brain tissue results in suspensions of intracellular organelles and pinched-off nerve terminals, synaptosomes. Homogenization shears off nerve terminals from axons, especially in brain regions with clearly defined anatomical layers, such as the cerebral cortex and hippocampus. Synaptosomes can be partially separated from other organelles by centrifugation techniques. Each of these remarkable structures is 0.5 to 1.0 μm in diameter, contains hundreds of synaptic vesicles and one or more mitochondria and is often associated with postsynaptic membrane fragments. Synaptosomes remain functional for several hours and can be used to study biochemical events, including energy and Ca²⁺ metabolism, neurotransmitter synthesis, transport and secretion. A related preparation is the neurosecretosome, from the posterior pituitary. These nerve terminals originate in the hypothalamus and contain vasopressin and oxytocin in large dense core granules. They are obtained in high purity from the neurohypophysis, which does not contain cell bodies. Neurosecretosomes are somewhat larger than synaptosomes and can be used for biochemical and patch clamp studies.

Several types of cells related to sympathetic neurons can be maintained and studied in tissue culture. Adrenal medullary chromaffin cells have the same precursor cells as postganglionic sympathetic neurons. These excitable neuroendocrine cells store, in large dense core granules called chromaffin granules, epinephrine or NE, together with ATP, and a variety of proteins (chromogranins, opiate peptides and precursors and dopamine β-hydroxylase). Relatively pure primary cultures can be prepared by collagenase digestion of bovine adrenal glands followed by cell-purification techniques. Various aspects of neurotransmitter metabolism and secretion have been extensively studied with these cells. They are amenable to both biochemical and electrophysiological experiments. A clonal cell line, PC12, is derived from a rat pheochromocytoma, a tumor of the adrenal medulla. Upon incubation with nerve growth factor (NGF), PC12 cells differentiate within days into neurons with axons and terminals (Chap. 19). Thus, they are used not only for biochemical and secretion studies but also for investigation of neuronal differentiation.

The neuromuscular junction is a well-defined structure that mediates the release and postsynaptic effects of acetylcholine

The first detailed studies of synaptic transmission were performed at the NMJ. The NMJ is a beautiful example of how structure and function are intimately entwined. The myelinated axon originating from the motor neuron in the spinal cord forms unmyelinated terminals that run longitudinally along the muscle fiber. Specialized transverse release sites, or active zones, occur periodically along the terminals and are oriented opposite invaginations of the postsynaptic membrane (Fig. 10-3). There are approximately 300 active zones per NMJ. The active zones in the nerve terminal display a cloud of clear vesicles, 50 to 60 nm in diameter, that contain ACh (Fig. 10-3). There are approximately 500,000 vesicles in all of the active zones at one NMJ. It is estimated that on the average a vesicle contains 20,000 ACh molecules. A small subset of the vesicles is attached in rows to the presynaptic membrane (Fig. 10-3A, B). These are thought to be docked vesicles that are able to undergo exocytosis upon Ca²⁺ influx. In freeze fracture, these rows coincide with rows of intramembrane particles that may be Ca²⁺ channels (Fig. 10-3C). Ca²⁺ entry that occurs upon stimulation of the nerve causes exocytosis that is seen as pits in freeze-fracture micrographs (Fig. 10-3D) or as “omega” figures in thin-section electron microscopy (Fig. 10-4). The vesicle membranes in the nerve terminal are recycled by endocytosis (see below).

Figure 10-3

Synaptic membrane structure. A: Entire frog neuromuscular junction (NMJ, left) and longitudinal section through a portion of the nerve terminal (right). Arrows indicate planes of cleavage during freeze-fracture. B: Three-dimensional view of presynaptic (more...)

Figure 10-4

High-magnification (×145,000) view of freeze-substituted neuromuscular junctions in a muscle frozen during the abnormally large burst of acetylcholine release that is provoked by a single nerve stimulus of 2 mM 4-aminopyridine, in this case delivered (more...)

The postsynaptic membrane opposite release sites is also highly specialized, consisting of folds of plasma membrane containing a high density of nicotinic ACh receptors (nAChRs). Basal lamina matrix proteins are important for the formation and maintenance of the NMJ and are concentrated in the cleft. Acetylcholinesterase (AChE), an enzyme which hydrolyzes ACh to acetate and choline to inactivate the neurotransmitter, is associated with the basal lamina (see Chap. 11).

Quantal analysis defines the mechanism of release as exocytosis

Stimulation of the motor neuron causes a large depolarization of the motor end plate. In 1952, Fatt and Katz [8] observed that spontaneous potentials of approximately 1 mV occur at the motor end plate. Each individual potential change has a time course similar to the much larger evoked response of the muscle membrane that results from electrical stimulation of the motor nerve. These small spontaneous potentials were therefore called miniature end-plate potentials (MEPPs). Because the MEPPs are reduced by the nicotinic antagonist d-tubocurarine and increased in amplitude and duration by the AChE inhibitor prostigmine, it was concluded that they are initiated by the release of ACh. Because the potential changes are too large to be accounted for by the interaction of individual molecules of ACh with the end plate, Fatt and Katz [8] postulated that they reflect the release of packets, or quanta, of ACh molecules from the nerve terminal.

A “curious effect” was observed by Fatt and Katz [8]: when the Ca²⁺ concentration is reduced and the Mg²⁺ increased, the evoked end-plate potential (EPP) is diminished without altering the size of the spontaneous MEPPs. With sufficiently low Ca²⁺, the evoked EPP is similar in size to MEPPs and varies in a stepwise manner. A single nerve impulse results in either no EPP or EPPs the approximate size of one, two, three or more MEPPs in an apparently random manner. The results of this type of experiment are shown in Figure 10-5. The frequency histogram shows that the amplitudes of evoked potentials are clustered in multiples of the mean spontaneous MEPP value. Statistical analysis [12] demonstrates that the release is a random process described by a Poisson distribution. Each event is unaffected by the preceding events. The model assumes n release sites capable of responding to a nerve impulse, each with a probability, p, of releasing a quantum of ACh. The mean number of quanta (m), or quantal content, released per nerve impulse is m = np. For a Poisson distribution, p must be small, <0.05, and n large, >100. The probability of evoked release of x quanta is P _x = (m ^x/x!)e ^{− m}. (See Martin [13] for a review of the Poisson distribution in the analysis of synaptic transmission.)

Figure 10-5

Comparison of the amplitudes of the spontaneous miniature end-plate potentials and the evoked end-plate potentials indicates that transmitter is released in quantal packages that are fixed in amplitude but variable in number. A: Intracellular recording (more...)

One critical test for the validity of the Poisson distribution as a description of release in the presence of reduced Ca²⁺ was the excellent agreement of two measures of m. One was derived empirically:

The other was derived from the Poisson equation and the observed probability of no response, or failures, upon nerve stimulation:

A more stringent test of the model is its ability to predict the histogram in Figure 10-5.

The quantal size m differs for different types of synapses. For a single impulse at the NMJ, 100 to 300 quanta are released. The large number of quanta that are released during a single impulse reflects the need for a large safety factor in the all-or-none response of muscle contraction. Where integration of inputs is important, quantal size is much less. At single terminals in sympathetic ganglia, at inhibitory and excitatory inputs on spinal motor neurons and at individual boutons of cultured hippocampal neurons, m is 1 to 3.

Ca²⁺ is necessary for transmission at the neuromuscular junction and other synapses and plays a special role in exocytosis

In most cases in the CNS and PNS, chemical transmission does not occur unless Ca²⁺ is present in the extracellular fluid. Katz and Miledi [14] elegantly demonstrated the critical role of Ca²⁺ in neurotransmitter release. The frog NMJ was perfused with salt solution containing Mg²⁺ but deficient in Ca²⁺. A twin-barrel micropipet, with each barrel filled with either 1.0 M CaCl₂ or NaCl, was placed immediately adjacent to the terminal. The sodium barrel was used to depolarize the nerve terminal electrically and the calcium barrel to apply Ca²⁺ ionotophoretically. Depolarization without Ca²⁺ failed to elicit an EPP (Fig. 10-6A). If Ca²⁺ was applied just before the depolarization, EPPs were evoked (Fig. 10-6B). In contrast, EPPs could not be elicited if the Ca²⁺ pulse immediately followed the depolarization (Fig. 10-6C). EPPs occurred when a Ca²⁺ pulse as short as 1 msec preceded the start of the depolarizing pulse by as little as 50 to 100 μsec. The experiments demonstrated that Ca²⁺ must be present when a nerve terminal is depolarized in order for neurotransmitter to be released.

Figure 10-6

Synaptic transmission requires that Ca²⁺ be present during the action potential. The effects of iontophoretic pulses of Ca²⁺ on end-plate response are shown. Depolarizing pulses (P) and Ca²⁺ were applied from a double-barrel micropipet to a small part (more...)

The normal extracellular Ca²⁺ concentration is approximately 2 mM. The basal cytosolic Ca²⁺ concentration is 0.1 μM or less. In nerve terminals, the rise of intracellular Ca²⁺ caused by depolarization of the plasma membrane opens voltage-sensitive Ca²⁺ channels. Ca²⁺ influx and the resultant rise in the cytosolic Ca²⁺ concentration adjacent to release sites along the plasma membrane trigger exocytosis. The sites of exocytosis are closely associated with Ca²⁺ channels (Fig. 10-3C). Ca²⁺ channels may, in fact, be components of multimeric protein complexes involved in exocytosis. Intracellular [Ca²⁺] immediately adjacent to Ca²⁺ channels is probably in the range 50 to 100 μM [17–19]. It is this high Ca²⁺ concentration that triggers exocytosis. Neuroendocrine cells, such as chromaffin cells from the adrenal medulla, also release hormones, such as epinephrine and opioid peptides, upon Ca²⁺ influx through membrane channels. It is thought that in this type of cell, release sites are usually not closely associated with Ca²⁺ channels and that [Ca²⁺] in the 0.5 to 10 μM range can trigger exocytosis. It should be noted that other types of cells, such as exocrine cells (for example, pancreatic acinar cells), also release stored protein by exocytosis upon a rise in cytosolic Ca²⁺. In many cases, Ca²⁺ is released from intracellular stores by inositol trisphosphate (IP₃), which is generated by the hormonal activation of G protein—linked receptors that activate phosphoinositide-specific phospholipase C (PI-PLC) (Chap. 21). In this case, extracellular Ca²⁺ sustains secretion by refilling the intracellular, IP₃-sensitive Ca²⁺ stores rather than by directly triggering secretion.

Presynaptic events during synaptic transmission are rapid, dynamic and interconnected

The time between Ca²⁺ influx and exocytosis in the nerve terminal is very short. At the frog NMJ at room temperature, 0.5 to 1 msec elapses between the depolarization of the nerve terminal and the beginning of the postsynaptic response. In the squid giant synapse, recordings can be made simultaneously in the presynaptic nerve terminal and in the postsynaptic cell. Voltage-sensitive Ca²⁺ channels open toward the end of the action potential. The time between Ca²⁺ influx and the postsynaptic response is 200 μsec (Fig. 10-7). Recent studies of synaptic transmission between CNS neurons using optical methods to record presynaptic events indicate a delay of only 60 μsec between Ca²⁺ influx and the postsynaptic response at 38°C [20].

Figure 10-7

The delay between Ca²⁺ influx into the nerve terminal and the postsynaptic response is brief. The temporal relationships between the Ca²⁺ current and the action potential in the nerve terminal and the postsynaptic response in the squid giant synapse are (more...)

The short delays between Ca²⁺ influx and exocytosis have important implications for the mechanism of fusion of synaptic vesicles (Chap. 9). In this short time, a synaptic vesicle cannot move significantly and must be already at the release site. From the diffusion constant of Ca²⁺ in squid axoplasm, one calculates that Ca²⁺ could diffuse only 850 Å, somewhat greater than the diameter of a synaptic vesicle. Therefore, in fast synapses, release sites must be close to the Ca²⁺ channels that trigger exocytosis. Vesicles are exposed to [Ca²⁺] of a few hundred micromolar near the mouth of the channels.

The supply of synaptic vesicles in the nerve terminal is limited. With continuous stimulation of the NMJ, the number of quanta released can exceed by many fold the number available in the nerve terminal. Possible transport of secretory vesicles from the cell body would be much too slow to maintain fast synaptic transmission in the terminal. Instead, the synaptic vesicle membrane which fuses with the plasma membrane is rapidly recycled via clathrin-mediated endocytosis (reviewed in [21]) (see Chap. 9). Hence, the vesicle membrane is a reusable container for neurotransmitter storage and exocytosis. The process of membrane recycling at the nerve terminal is closely related to the general process of endocytosis that occurs in non-neuronal cells. Strong evidence for this process came from electron micrographs of horseradish peroxidase uptake from the extracellular medium into the nerve terminal of the frog NMJ following nerve stimulation [23,24]. Endocytosis is dispersed along the membrane away from active zones. It was originally proposed that clathrin-coated vesicles bud from the plasma membrane, lose their triskelion clathrin coat and fuse to an intermediary endosomal compartment, from which new synaptic vesicles bud. Synaptic vesicles then take up neurotransmitter and recycle to release sites. However, recent studies suggest an alternative pathway that bypasses the intermediate endosomal compartment. It would allow more rapid endocytic recycling of the synaptic vesicle membrane. Clathrin-coated vesicles bud from the plasma membrane, become uncoated, take up neurotransmitter and recycle to the plasma membrane. Strong stimulation of the nerve terminal may cause invaginations of the plasma membrane, from which clathrin-coated vesicles can also bud (see Chap. 2).

The development of amphipathic fluorescent dyes that label endocytic vesicles has permitted the study of endocytosis in nerve terminals in real time [25]. The probe FM1–43 equilibrates between the aqueous phase and the membrane but is not membrane-permeant. The plasma membrane becomes fluorescent (Fig. 10-8). Upon endocytosis, the labeled membrane is internalized. When removed from the extracellular medium, the dye is retained by the endocytic vesicles but lost from the plasma membrane. Endocytic vesicles are transformed into synaptic vesicles containing FM1–43. Importantly, recycled synaptic vesicles lose the probe upon exocytosis.

Figure 10-8

The probe FM1–43 was used to visualize endocytosis and exocytosis at the neuromuscular junction (NMI). A: Structure of the amphipathic membrane probe FM1–43. B: Labeling of the plasma membrane by FM1–43 in the extracellular medium. (more...)

This technique has permitted the dynamics of the exocytic/endocytic cycle to be investigated. At the NMJ, a complete cycle of exocytosis and endocytosis requires approximately 1 min. The recycled vesicles mix homogeneously with, and have the same probability of again undergoing exocytosis as, unlabeled vesicles. A single nerve impulse releases 0.1% of the recycling pool. The dynamics of recycling are similar in cultured hippocampal neurons [26]. Endocytosis follows approximately 20 sec after exocytosis. The transformation of the endocytic vesicle into a functioning synaptic vesicle requires about 15 sec. About 0.5% of the recycling pool is released by a nerve impulse. This corresponds to approximately one vesicle per synaptic bouton.

The importance of endocytosis for the normal function of the nerve terminal is demonstrated by the shibire mutant of Drosophila. When these mutant flies, bearing a temperature-sensitive allele, are exposed to high temperature, they become paralyzed within 1 min but rapidly recover when returned to the permissive temperature. Electron microscopy demonstrates that the paralysis results from a block of synaptic vesicle endocytosis at the NMJ. The shibire allele encodes dynamin, a GTPase that is essential for the fission of the endocytic bud from the plasma membrane.

There are important differences between fast synaptic transmission at nerve terminals and the release of proteins and peptides from nerve terminals and neuroendocrine cells

Because fast synaptic transmission involves recycling vesicles, the neurotransmitter must be replenished locally. Thus, fast synaptic transmission uses neurotransmitters such as ACh, glutamate, GABA, glycine, dopamine (DA) and NE, all of which can be synthesized within the nerve terminal or transported rapidly across the nerve terminal plasma membrane. In contrast, proteins are inserted into secretory granules in the cell body. The secretory granules must then be transported by fast axonal transport into the nerve terminal, a process that can take many hours or days depending on the distance of the nerve terminal from the soma (see Chap. 28). Nerve terminals that are specialized for fast synaptic transmission may have peptidergic granules as well as the recycling vesicles of fast synaptic transmission. For example, nerve terminals can contain VIP as well as ACh, enkephalin as well as NE and substance P as well as 5-HT. The peptidergic granules are usually far less numerous than the smaller vesicles involved in fast exocytosis and are not localized at active zones (Fig. 10-1). The exocytosis of protein-containing granules in nerve terminals may be closely related to exocytosis of protein-containing granules in endocrine and exocrine cells. While a single nerve impulse will release vesicles at active zones, exocytosis of peptidergic granules in nerve terminals can require multiple or high-frequency stimulation. This may reflect the need for sustained elevations of Ca²⁺ that extend into the interior of the nerve terminal. Peptides and proteins released from nerves may have slower and longer-lasting effects on postsynaptic cells than fast neurotransmitters and can modulate the response to fast neurotransmitters.

Discrete steps in the regulated secretory pathway can be defined in neuroendocrine cells

The rapid presynaptic events of synaptic transmission produce closely coordinated exocytosis and endocytosis. Insights into the steps involved in the exocytotic limb of the pathway have come from the studies of kinetics of secretion of protein-containing granules from adrenal chromaffin cells and PC12 cells. Adrenal chromaffin cells are excitable and contain a large number of secretory or chromaffin granules. In bovine chromaffin cells, a neuronal-type nAChR and voltage-sensitive Ca²⁺ channels permit Ca²⁺ entry, which stimulates secretion. The intracellular milieu of chromaffin cells can be directly controlled by extracellular solutions in cells with plasma membranes rendered leaky by the detergent digitonin, by streptolysin-O [27] or by mechanically disrupting the plasma membrane by passage of the cells through a steel cylinder partially blocked by a precision steel bearing [28]. PC12 cells contain far fewer granules than adrenal chromaffin cells, and many are closely associated with the plasma membrane. An analysis of the effects of Ca²⁺, ATP and temperature suggests that ATP hydrolysis occurs before Ca²⁺ is able to cause secretory granules to secrete. Two distinct Ca²⁺-dependent steps have been identified. One triggers exocytosis with maximal effects at 100 to 300 μM Ca²⁺, whereas the other enhances the ability of ATP to prime secretion, with maximal effects at approximately 1 μM Ca²⁺. Electrophysiological studies have identified additional steps associated with the triggering of exocytosis that may reflect the dynamics of inter-related pools of granules [29].

What is the function of ATP in secretion? While protein phosphorylation can modulate the secretory response, there is compelling evidence that the effect of ATP in priming involves other processes. ATP is necessary for the function of N-ethylmaleimide-sensitive factor (NSF). This protein is an ATPase that acts as a molecular chaperone to dissociate complexes of the SNARE proteins VAMP (synaptobrevin), syntaxin and SNAP-25 (see Chap. 9). This may permit their subsequent reassociation as part of the exocytotic response. Another function of ATP in priming exocytosis is the maintenance of the polyphosphoinositides, phosphatidylinositol 4,5-bisphosphate (PIP₂) and phosphatidylinositol 4-phosphate (PIP), by phosphorylation of lipid precursors via phosphatidylinositol 4-kinase and PIP kinase [30–32]. Interestingly, phosphatidylinositol 4-kinase is an integral membrane protein of chromaffin granules. The polyphosphoinositides appear to function in the priming step not as precursors for the formation of IP₃ and diacylglycerol (DAG) but, rather, in some other capacity (see Chap. 21). PIP₂ binds specifically to the vesicle or granule proteins synaptotagmin and rabphilin 3a. PIP₂ also regulates numerous proteins that control the cytoskeleton, such as profilin, gelsolin, scinderin and myosin I. Therefore, the polyphosphoinositides on the secretory granule membrane may coordinate the function in secretion of several secretory granule proteins or may modulate dynamic changes in the cytoskeletal network that are important for exocytosis.