Presynaptic Receptors

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS

Richard J. Miller Department of Pharmacological and Physiological Sciences, The University of Chicago, Chicago, Illinois 60637; e-mail: [email protected]

KEY WORDS: transmitter release, Ca channels, potassium channels, ionotropic receptors, G proteins

ABSTRACT Activation of different types of G-protein-linked and ionotropic presynaptic receptors has been shown to regulate neurotransmitter release throughout the central and peripheral nervous systems. In the case of G-protein-linked receptors, three major mechanisms have been suggested: (a) inhibition of Ca channels in the nerve terminal; (b) the activation of presynaptic K channels, resulting in a reduction in the effectiveness of the action potential; and (c) direct modulation of one or more components of the neurotransmitter vesicle release apparatus. In the case of ionotropic presynaptic receptors, inhibition of release may be achieved through depolarization of the terminal and inactivation of Na and Ca channels. Activation of presynaptic ionotropic receptors that are appreciably Ca permeable can also enhance the release of transmitters as a result of their ability to raise [Ca]i in the terminal directly. Many transmitters employ several of these mechanisms, thus allowing considerable ﬂexibility in the presynaptic regulation of transmitter release.

INTRODUCTION Neurons communicate with each other through the release and subsequent action of neurotransmitters. The efﬁciency of the transmitter release process can be inﬂuenced by a number of factors that allow the strength of synaptic com- munication to be constantly modulated. Presynaptic receptors are particularly important in this regard. There are numerous studies in vertebrates and inver- tebrates demonstrating that the presynaptic terminal of a neuron may possess receptors for the same neurotransmitter that it releases. Activation of these

201 0362-1642/98/0415-0201$08.00

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

202 MILLER

receptors closes a feedback loop, which results in either the enhancement or inhibition of transmitter release. Such presynaptic receptors are usually known as autoreceptors. In addition, the presynaptic terminal can possess receptors for transmitters that are released from other neurons in the vicinity, and activation of these can also inﬂuence the release process. Such receptors are generally known as heteroreceptors. It is likely that activation of presynaptic receptors helps to shape the strength of neurotransmission at virtually every synapse (1–4). Thus, these receptors are of considerable importance both from the neurophysiological point of view and also as potential targets for therapeutic agents. How does the activation of presynaptic receptors produce changes in the release of neurotransmitters? In this article, I examine some of the major possi- bilities. For reasons of space, I restrict the discussion to the vertebrate nervous system.

MOLECULAR ASPECTS OF TRANSMITTER RELEASE In order to understand the way in which activation of presynaptic receptors modulates neurotransmitter release, it is important to review, at least in outline, some of the molecular events that underlie the release process (5–7). Neurotransmitter-containing vesicles are thought to be localized at special- ized release sites, or active zones, on the presynaptic membrane, in a state of readiness so that exocytosis can occur rapidly when required (5–8). Initially, vesicles are targeted (docked) at release sites, although the precise details as to how this happens are not well understood. Once docked, however, synaptic vesicles are primed so that they can respond by rapid exocytosis to the appropriate stimulus (normally a highly localized Ca inﬂux). The priming process involves the formation of a multiprotein complex that tethers the secretory vesicle to the plasma membrane in close juxtaposition to other important signaling elements. This complex involves proteins from the vesicle membrane and the presynaptic neuronal plasma membrane, as well as other elements that help to link the two together (5, 6). Some of the proteins also have properties that allow them to respond to the Ca signal that triggers exocytosis (9). Although a complete discussion of all the components of the complex is beyond the scope of this article, several of the components are particularly relevant. Initially, a core complex is formed between the plasma membrane proteins syntaxin 1A and SNAP-25 and the vesicle membrane protein synaptobrevin with a 1:1:1 stoichiometry (5–7, 10). These three proteins provide a high-afﬁnity receptor scaffold for the protein α-SNAP, and its arrival then allows the recruitment of NSF (N-ethylmaleimide-sensitive ATPase). ATP hydrolysis, together with dis- sociation of NSF/α-SNAP, may now cause partial fusion of the vesicle with the

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 203

plasma membrane, thus readying it for final release. Also brought into close juxtaposition at this point are the Ca channels that provide the final Ca trigger for exocytosis (10–15) and the putative Ca sensor molecule on the vesicle membrane—the protein synaptotagmin (16–18). Other elements that may be important in the final release of synaptic vesicles are the small GTP binding protein, rab3A (19–22), and the local generation of fusigenic inositol phospho- lipids that can bind to synaptotagmin (23). It is particularly interesting to consider the interaction of Ca channels with the rest of the complex. At least six different α1 pore-forming Ca channel subunits have been identified (24, 25). All of these appear to share the same basic architecture, as well as many basic functional properties. They also exhibit important differences in the way they are regulated that are consistent with their diverse functions. Not all Ca channels appear to be concerned with the regulation of neurotransmitter release. However, the channels formed by the α1B (N-type) and α1A (P/Q-type) subunits, together with their appropriate an- cillary subunits, are clearly important in this regard (4, 26). This being the case, it is interesting to note that both of these α1 subunits can form specific associations with the vesicle release complex (11–15). This interaction occurs between a specific region of the large intracellular loop connecting domains 2/3 of the α1-subunits and the presynaptic membrane protein syntaxin 1A. In par- ticular, binding appears to occur between the transmembrane region of syntaxin 1A and a specific region of the intracellular loop of the α1-subunit designated the synprint region (synaptic protein interaction site) (27–31). This region can also interact with SNAP-25 (27–31), and possibly with other members of the vesicle secretory complex as well. Indeed, this same loop, which contains the most divergent sequences between α1-subunits, appears to be generally significant in determining downstream targeting of Ca channels to important effector molecules. For example, the targeting of Ca channel α1 subunits to ryan- odine receptors in skeletal muscle is also provided by analogous interactions (32). How is transmitter release finally achieved? The sequence of events appears to be approximately as follows. Upon depolarization of the nerve terminal, the Ca channels that have been appropriately targeted to release sites open, rapidly elevating the local submembrane [Ca]i to values of >100 µM (5, 7). This Ca can then interact with the appropriate Ca-sensing molecule on the partially fused vesicle membrane (probably synaptotagmin), allowing the ultimate fusion of the vesicle and release of its contents. Ca simultaneously produces other effects on the complex of proteins assembled in association with the vesicle (9). For example, the interaction between Ca channels and syntaxin 1A is enhanced as the local Ca concentration climbs to about 10 µM or so but is then reduced at higher [Ca]i, which may facilitate final vesicle exocytosis (29, 30). The

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

204 MILLER

association between syntaxin 1A and Ca channels is not only important from the point of view of bringing these elements into close juxtaposition at release sites, but it also has important functional effects on the behavior of Ca channels, as I discuss below.

MECHANISMS OF PRESYNAPTIC INHIBITION How does activation of presynaptic receptors modulate the release process? Presynaptic receptors can be of several types, including members of both the G-protein-linked (metabotropic) (4, 33) and multisubunit ion channel (ionotropic) families (34). Activation of both types of receptors can regulate transmitter release in different circumstances. There are several major explanations of how presynaptic receptors could work.

1. Inhibition of Ca channels. We know that Ca influx through appropriately localized channels is the major trigger for the final release of synaptic vesicles. Thus, changes in Ca influx due to regulation of these channels should certainly influence release at some point (4). Indeed, as I discuss in detail below, activation of many G-protein-linked receptors has been shown to influence Ca channel activity.

2. Activation of presynaptic ion channels. Stimulation of a presynaptic receptor could activate a conductance in the nerve terminal (e.g. a K or Cl channel), thereby “shunting” the action potential in this region of the neuron (35–37). Activation of such a presynaptic conductance would reduce the duration and amplitude of the action potential in the presynaptic terminal, Ca influx would be reduced, and less transmitter would be released. Alternatively, activation of Na- or Cl-permeable presynaptic channels might depolarize the terminal sufficiently to inactivate voltage-dependent Na and Ca channels, thereby blocking propagation of the action potential into the terminal or its local influence (38).

3. Regulation of the vesicle release complex. As I have discussed above, the ﬁnal steps that lead to the release of synaptic vesicles involve a complex of proteins, including several Ca-sensitive components. Activation of presynaptic receptors could signal to one or more of these proteins and inﬂuence the release process at a point distal to Ca entry.

Evidence exists that all of these mechanisms are actually used. Indeed, some- times several strategies may be employed simultaneously.

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 205

G-PROTEIN-LINKED RECEPTORS Here, I first consider the ways in which G-protein-linked receptors might influence transmitter release. The greatest variety of presynaptic receptors likely fit into this category. Even in cases in which important presynaptic actions of a transmitter are produced by activation of ionotropic receptors (e.g. GABA), presynaptic G-protein-linked receptors for the same transmitter are also likely to be important. In virtually all cases, activation of presynaptic receptors of this type leads to a reduction in evoked transmitter release, thereby closing a negative feedback loop.

Regulation of Presynaptic Ca Channels THE MECHANISM OF Ca CHANNEL INHIBITION Ca influx is the ultimate trigger for the release of neurotransmitters. Although the relationship between Ca influx and release is complex (7), a reduction in Ca influx must, at some point, impact transmitter release. If Ca channels are to be targets for the inhibition of neurotransmitter release, it would be important for the relevant types of Ca channels to be inhibited. These would be the classes of Ca channels that have been shown to be localized at release sites and most closely connected with the release process (see above). In general the release of neurotransmitters appears to be under the influence of both N and P/Q types of Ca channels (4, 26, 34–65). The basic observations that support this contention concern the ability of toxins and drugs that selectively target these Ca channels (e.g. ω-conotoxin GVIA for N-type and ω-agatoxin IVA for P/Q-type) to inhibit transmitter release, whereas drugs and toxins that block L-type channels (e.g. dihydropyridines) are generally ineffective (4, 34–65). There are, naturally, some exceptions to this rule, and instances can be found where dihydropyridines inhibit release (46). Furthermore, in some circumstances Ca channels that have not yet been fully identified and characterized also seem to influence release (43). Nevertheless, the key roles of N and P/Q channels have been clearly illustrated in numerous instances and so will serve as the basis for discussion. Activation of many G-protein-linked receptors has been shown to produce inhibition of N- or P/Q-type Ca currents in whole-cell voltage-clamp experiments (e.g. 3, 4, 47–49). Several mechanisms can apparently transduce the signal between the receptor and these Ca channels. Let us start by discussing the pathway that has been most fully characterized. In this situation, activation of a receptor changes the gating properties of Ca channels so that they are no longer opened by depolarizations in the normal range, although very strong depolarizations can still produce channel opening (49–52). Depending on the paradigm employed, this results in slowing of the rate of current activation and/or steady state inhibition of the current. The kinetic slowing and inhibition

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

206 MILLER

can be relieved by a strongly depolarizing prepulse, and it has been suggested that this may be of physiological significance. For example, it raises the possibility that inhibition might be relieved during a rapid train of action potentials (53). Such a phenomenon would help to explain the “use dependence” com- monly associated with presynaptic inhibition (54, 55). Thus, it has frequently been demonstrated that presynaptic inhibition is less effective when neurons are firing more rapidly. However, when this proposal has been formally tested, the data indicate that Ca channel inhibition would not be significantly allevi- ated by firing rates in the normal range (56, 57). Thus, the use dependence of presynaptic inhibition probably has a different explanation (58). How is Ca channel inhibition produced? Activation of the receptor initially leads to activation of heterotrimeric G proteins. The first question to be an- swered therefore is how the G protein influences the behavior of the Ca channel. Given the fact that that no diffusible second messenger seems to be involved (47), it is likely that the G protein exerts its influence by directly binding to the channel. Experiments using heterologous expression systems suggest that it is the β/γ subunits of G proteins that are responsible for transducing the signal (59–61). Thus, the G-protein-mediated modulation of Ca channels is similar in this regard to the regulation of the GIRK family of K channels by receptors and G proteins—a process that also involves signal transduction via G-protein β/γ subunits (62). Indeed, several G-protein-mediated effects are now known to be produced in this way (63). Initial attempts to identify the point of interaction between β/γ subunits and Ca channels have suggested that the binding domain is located in the first intracellular loop that connects domains 1 and 2 of the α1 subunit, the major pore-forming subunit of the Ca channel (64, 65). Indeed, there is little doubt that β/γ subunits can interact with Ca channels at this site, as suggested by both mutagenesis and biochemical studies (64–66). This intracellular loop contains a motif (QXXER) that has been shown to be involved in the binding of G-protein β/γ subunits to several other effectors as well, including GIRK-like K channels (63). This motif is present in α1B, α1A, and α1E, the dihydropyridine- insensitive Ca channels, all of which seem to be modulated by G-protein β/γ subunits to some extent (59–61). On the other hand, the motif is absent from the dihydropyridine-sensitive Ca channels (α1C, α1D, and α1S), which do not seem to be directly modulated by G proteins (67, 68). Synthetic peptides from the 1/2 loop that include this sequence seem to reduce G-protein-mediated inhibition of α1A-orα1B-based Ca channels in the predicted manner, and, in some reports, mutations in this sequence interfere with the effects of G proteins. This supports the proposal that β/γ binding at this site mediates G-protein inhibition (64–66). Nevertheless, the story appears to be far from complete. One should consider, for example, experiments in which the 1/2 linker has been exchanged between

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 207

different α1 subunits. It has been shown that although α1A, α1B, and α1E are all inibited by G proteins, they differ considerably in terms of their sensitivity. Thus, α1B is the most sensitive and α1E the least (59–61, 67–69). Chimeric Ca channels have been tested in which the 1/2 linker from α1A or α1B has been inserted into an α1E background (70), the 1/2 linker from α1B has been inserted into an α1A background (64), or the 1/2 linker of α1A or α1C has been inserted into an α1B background (68). The results of these experiments are quite inconsistent with one another and range from the inserted 1/2 linker resulting in a large change in G-protein inhibition (64), a small change (70), or no change at all (68). Furthermore, β/γ subunits also seem to bind to more than one site on responsive Ca channel α1 subunits (65). Thus, DeWaard et al (65) demonstrated that there is a second binding site in the 1/2 linker that has an even higher affinity for the β/γ subunit. Interestingly, this site does not possess the QXXER binding motif. In addition, Zhang et al (68) demonstrated that portions of the α1 subunit that are outside the 1/2 linker also appear to be important in mediating G-protein inhibition. Consistent with this, a β/γ binding site has now been identified in the C-terminal tail of α1-subunits (71). Thus, the precise way in which β/γ binding is transduced into channel inhibition is still incompletely understood. The idea that the QXXER-containing binding site in the 1/2 linker is an important element in the β/γ effect is interesting for at least two reasons. First, this site overlaps with the binding site for the Ca channel β-subunit (72). This suggests that Ca channel β-subunits and G-protein β/γ subunits may be able to interfere with their respective interactions with Ca channel α1-subunits. Such a proposal is consistent with observations that the effects of Ca channel β-subunits and of G proteins on channel function are, in many ways, the opposite of one another (65, 73, 74). Secondly, Zamponi et al (64) have demonstated that this putative β/γ binding site is a substrate for protein kinase C–mediated phosphorylation. Phosphorylation at this site may be responsible for the observed uncoupling of receptor-mediated inhibition of N channels by phorbol esters (75) and may be of physiological significance in this regard. The specific localization and clustering of Ca channels at sites of neurotransmitter release (see above) may also have important implications for Ca channel regulation. For example, there is evidence that the interaction with the protein syntaxin 1A, in addition to helping to locate Ca channels to the appropriate place, has a functional effect. Coexpression studies have demonstrated that syntaxin 1A can exert an inhibitory effect on N- and P/Q-type Ca channels through an interaction between the membrane spanning domain of the syntaxin 1A molecule and the large intracellular loop connecting domains 2 and 3 of Ca channel α1 subunits (76, 77). This interaction may be modulated by other components of the vesicle docking complex, such as synaptotagmin and SNAP-25 (77, 78) or by phosphorylation (31). It has also recently been demonstrated

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

208 MILLER

that cleavage of syntaxin 1A using botulinum toxin prevents the regulation of N channels by G proteins (73). Such functional interactions between docking complex proteins and Ca channels may represent a mechanism by which these proteins not only help to juxtapose vesicles and the source of incoming Ca, but also ensure that channels behave in the appropriate manner, i.e. receptors will modulate only appropriately located channels. The mechanism of Ca channel modulation described above represents a tightly coupled arrangement that makes excellent sense in terms of organiz- ing the local feedback inhibition of release from the nerve terminal. It is clear, however, that other mechanisms of Ca channel regulation can occur (47). In these instances Ca current inhibition does not necessarily exhibit the kinetic slowing and voltage dependence described above. In some cases one or more diffusible second messengers have been shown to be involved in the signal transduction pathway, although the identities of these messengers have not been generally determined. A receptor may modulate Ca channels by recruit- ing several pathways involving the activation of more than one G protein. The case of bradykinin serves as a good example of this. It has been shown that bradykinin can inhibit Ca currents in dorsal root ganglion (DRG) neurons and in certain neuronal cell lines (80, 81). Belardetti and colleagues demonstrated that in the NG108-15 cell line, bradykinin inhibits Ca channels using at least three different pathways, two of which seem to require the participation of intracellular Ca (81). In addition, a third pathway seems to signal through activation of the heterotrimeric G-protein G13 and subsequently through small G proteins of the Rac1 type (82). It is likely that this then feeds into the MAP kinase (MAPK) pathway, but exactly what happens after that is unclear. The situation with bradykinin also highlights the fact that the receptor regulation of neuronal Ca channels is not necessarily solely concerned with the feedback inhibition of transmitter release. In DRG neurons, for example, the effects of bradykinin likely lead to a reduction in spike accommodation through effects on IKCa and subsequently to increased transmitter release in the spinal cord. In this way effects on Ca channels may underlie the hyperalgesic actions of bradykinin (80).

LOCALIZATION OF Ca CHANNEL INHIBITION The vast majority of electrophysiological studies on Ca channels have been carried out on cultured neurons or acutely isolated nerve cell bodies. In some instances attempts have been made to correlate effects on cell body Ca currents with transmitter release from the same neurons (83, 84). However, the question has naturally arisen as to whether observations in the cell body really reﬂect events occurring in the nerve terminal, which is, after all, the principal site of neurotransmitter release. Although it is much harder to perform studies on nerve terminals, data of various types

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 209

indicate that Ca channels in terminals are modulated by receptors in the same way as indicated above for the cell soma. Evidence suggests that there are different types of Ca channels in nerve terminals. The most direct evidence comes from studies using specific antibodies that indicate the presence of the α1A and α1B Ca channel subunits in nerve terminals (85, 86). Similar conclusions come from studies using fluorescently labeled toxins that target different channel types (87). In addition, many studies have examined the effects of specific drugs and toxins on transmitter release from intact neurons and synaptosomes (3, 26). The results of these studies generally show that either ω-conotoxin-GVIA or ω-agatoxin-IVA can block release, to varying degrees, in different preparations, whereas dihydropyridines are usually ineffective. Some studies have sought to determine how channels are distributed among different synapses. For example, Reuter examined this question at synapses between cultured rat hippocampal pyramidal neurons by using the dye FMI-43 to measure the exocytosis of vesicles from individual nerve terminals together with ω-conotoxin-GVIA and ω-agatoxin-IVA to block N and P/Q channels, respectively (88). Although both of the toxins used were partially effective at blocking evoked glutamatergic excitatory postsynaptic potentials (EPSPs) at these synapses, the two types of Ca channels proved to be arranged differently. Whereas P/Q channels were fairly evenly distributed at most nerve terminals, the distribution of N channels was much more heterogeneous. Thus, the ratio of channel types differed considerably at individual terminals. Although there are some quantitative differences between the results of this study and more conventional electrophysiological investigations (e.g. see 89), there is a general concensus that the distribution of Ca channels among terminals is not homo- geneous and that most terminals contain mixtures of channels rather than a single type. The use of selective channel blockers has also indicated that the efficiency of coupling of different types of Ca channels to transmitter release depends on the situation. For example, Mintz et al concluded that at the parallel fiber/Purkinje cell synapse in the cerebellum, P/Q channels were more “effi- ciently” coupled to transmitter release than were N channels (62); whereas at the CA3/CA1 synapses of the hippocampus, Wu & Saggau found that the two channel types were equally well coupled to transmitter release (90). Are Ca channels in nerve terminals modulated by receptors in the same way as channels in the cell body? This is not an easy question to answer, as it is obviously difficult to access nerve terminals in the same way as the soma. Nevertheless, such experiments have been performed by using advantageous preparations. Channel function in terminals has been assessed both electrophys- iologically and by measuring Ca flux directly through channels using imaging paradigms. At calyx-like giant synapses in chick ciliary ganglia (15, 79, 91) and the rat brain stem calyx of Held (92, 93), measurement of Ca currents in

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

210 MILLER

presynaptic terminals has been possible. Transmitter release has been found to be mostly dependent on N and P/Q channels, respectively, at these two synapses. Activation of presynaptic receptors for somatostatin (92), adenosine Al (91), or glutamate (mGluRs class 3) (92) inhibited the Ca currents and transmitter release in each instance. At rat sympathetic neuroeffector junctions between sympathetic neurons and cardiac myocytes, presynaptic Ca influx (mostly via N channels) at single nerve terminals was inhibited by neuropeptide Y (NPY), as was transmitter release at the same synapses (95). At populations of guinea- pig hippocampal CA3/CA1 synapses (96–99) and granule cell/parallel fiber– Purkinje cell synapses in the cerebellum (100), presynaptic Ca influx via N and P/Q channels, as well as the “residual” Ca influx, was inhibited by activation of adenosine A1, GABAB, or NPY receptors, whereas activation of muscarinic receptors specifically targeted Ca influx via N channels. In all of these instances transmitter release was also shown to be inhibited in parallel with inhibition of Ca flux. Results of studies on the hippocampus demonstrate that activation of adenosine A1 receptors and GABAB receptors inhibits Ca currents in the neuronal soma as well as transmitter release (83, 84). Thus, in the hippocampus, at any rate, a correlation exists among effects of A1 and GABAB agonists in the soma, the presynaptic terminal, and the release of transmitter. Regulation of Presynaptic K Channels Another mechanism by which presynaptic G-protein-linked receptors might influence transmitter release is through the activation of some conductance (e.g. a K or Cl channel) in the presynaptic terminal. Sufficient numbers of open channels could have the effect of “shunting” the incoming action potential so that it was less effective in depolarizing the terminal; thus, fewer Ca channels would open, thereby shortening the Ca spike, and less Ca influx would ensue (35–38). A variation on this theme is that activation of presynaptic currents would result in an actual failure of the action potential to invade the terminal. However, whether activation of presynaptic K currents by G-protein-linked receptors is really a widely used mechanism for producing presynaptic inhibition by this, or any other mechanism, is less clear—at least in the vertebrate nervous system. In contrast, the participation of K channels in the presynaptic regulation of transmitter release at invertebrate synapses is much better established (101). It is certainly the case that many G-protein-linked receptors can activate inwardly rectifying K conductances in neurons and possibly other types of K channels as well (42, 102). Such effects often underlie neurotransmitter-induced neuronal hyperpolarization/inhibition, of which there are many examples (103). The inwardly rectifying K channels involved are probably members of the GIRK/CIR family (104). Activation of these channels by neurotransmitters is probably an important method for controlling neuronal excitability (103).

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 211

Indeed, animals lacking GIRK-like channels are prone to seizures (105). The mechanism by which receptors activate GIRK-like channels has been shown to involve G-protein β/γ subunits that bind directly to one or more sites on the GIRK molecule (106). Thus, the molecular mechanisms involved appear to have much in common with the receptor-induced inhibition of Ca channels, discussed above. Indeed, at least one of the β/γ binding sites on the GIRK channel also contains the same motif (QXXER) as the β/γ binding site on the first intracellular loop of N and P/Q Ca channel α1 subunits (see above; 63). One study has suggested that the GIRK1 subunits in the hippocampus are predominantly found in the dendrites and dendritic spines of pyramidal cells, rather than in presynaptic terminals (107), although other studies with GIRKs 1–4 have also localized them to terminal fields (108–110). The possible role of K current activation in presynaptic inhibition has been mostly studied in the hippocampus, but evidence for such a mechanism here and elsewhere is limited. The type of experiment most frequently performed is to examine the effects of Ba. The idea here is that because GIRKs are inhibited by Ba, presynaptic inhibition should also be inhibited by Ba if these K channels are involved. In some studies, such as those on the effects of the GABAB receptor agonist baclofen in the hippocampus, this is the case (111). Thus, inhibition of GABAergic inhibitory postsynaptic potentials (IPSPs) by baclofen in the CA3 (but not in CA1) and by dopamine in the nucleus accumbens is blocked by Ba (112, 113). In contrast, the effects of baclofen on EPSPs, and of several other neurotransmitters on excitatory and inhibitory synaptic events throughout the hippocampus (and other parts of the brain), are unaffected by Ba (111, 113, 114, but see 115). Furthermore, Luscher et al (116) showed that presynaptic inhibition produced by GABAB and several other types of agonists was unaltered in GIRK2 knockout mice. In a slightly different example, Simmons & Chavkin demonstrated that κ-opioid agonist–mediated inhibition of neurotransmitter release from mossy fiber terminals in the hippocampus was unaffected by Ba but was inhibited by dendrotoxin, which is a blocker of the Shaker Kv1 class of K channels (117). Whether this latter mechanism is widely used in the nervous system and what the signal transduction pathway leading to channel activation might be is unclear at this time. In summary, data using K channel blockers have not really supported the idea that K channels play a wide-ranging role in mediating presynaptic inhibition. A further argument suggesting that effects on K channels may not be generally important comes from observations on the selectivity of the effects of receptors on different components of the presynaptic Ca flux. Thus, presynaptic Ca influx can be divided into different components based on the effects of different Ca channel blockers (see above). Activation of presynaptic receptors inhibits these different components to varying degrees. For example, in the

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

212 MILLER

hippocampus, activation of adenosine A1, GABAB, or muscarinic receptors seems to more effectively inhibit Ca influx through N rather than through P/Q channels (4, 96, 97, 118). At cerebellar parallel fiber/Purkinje cell synapses, activation of adenosine A1 or GABAB receptors preferentially targets N and P channels rather than the “residual” Ca influx (4, 98). A similar result was obtained for the effects of NPY at sympathetic neuroeffector junctions (95). In contrast, inhibition of the residual Ca influx by presynaptic metabotropic glutamate receptors was found to be favored at excitatory synapses in the nucleus of the solitary tract (119). It has been argued (e.g. in 4) that if the major effect of presynaptic receptor activation was to hyperpolarize the presynaptic terminal, then each component of the Ca influx should be equally inhibited. Because this was not found to be the case, it follows that, at least in the instances investigated to date, activation of a K conductance does not play a major role. In conclusion, considering the anatomical data on the localization of GIRK1, the data on Ba sensitivity, and the data on Ca influx, not much evidence links presynaptic inhibition at vertebrate synapses with the activation of K conductances. Direct Modulation of the Release Machinery A third possible way in which activation of presynaptic receptors might influence transmitter release is by directly influencing some component of the vesicle release/exocytosis complex. As I discussed above, the release complex contains many proteins that seem to be important in the docking, priming, and ultimate release of neurotransmitter vesicles. Activation of a receptor could alter the properties of one or more of these proteins and inhibit release in this manner. That this can indeed be the case was first demonstrated by Silinsky, studying transmission at frog motor neuron/skeletal muscle synapses (120). Silinsky observed that the evoked release of the transmitter (acetylcholine) at this synapse could be reduced by activation of adenosine A1 receptors acting by a presynaptic mechanism. Further analysis led to the conclusion that the major site of action of adenosine analogues was intracellular, i.e. beyond or downstream of the point of Ca entry, rather than on Ca entry per se. For example, Silinsky was able to increase the frequency of miniature end plate potentials (MEPPs) by using Ca containing liposomes to deliver Ca directly to the nerve terminal, thus bypassing the normal route of Ca entry. He also used the poly- valent cation La, which increases the frequency of release by some mechanism that is independent of Ca influx. As A1 agonists were still effective in reducing MEPP frequency, but not amplitude, in both of these cases, Silinsky concluded that they must produce their effects downstream of Ca entry. Moreover, that the effects of Ca itself could be blocked implied that the site of inhibition was actually downstream of the site of Ca action on the release process.

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 213

These observations have subsequently been confirmed in several more recent studies of both the vertebrate and invertebrate nervous systems. Scholz & Miller (121) studied the inhibitory effects of A1 agonists on evoked glutamate release at excitatory hippocampal pyramidal neuron synapses. Following the demonstration that A1 agonists blocked the evoked glutamatergic excitatory postsynaptic current (EPSC) at these synapses and the Ca current in the cell soma (83), these authors showed that the frequency but not the amplitude of miniature excitatory postsynaptic currents (mEPSCs) recorded in the presence of tetrodotoxin (TTX) and Cd (a nonselective blocker of all types of high- threshold Ca channels) was also blocked. All three of these effects were absent following pertussis toxin treatment. mEPSCs are usually thought to result from the spontaneous exocytosis of transmitter-containing vesicles occurring in the absence of Ca influx. Thus, these results also imply that A1 agonists can inhibit release at hippocampal nerve terminals downstream of Ca entry. Sim- ilar effects have been reported for several other presynaptic receptor agonists acting on mEPSCs and inhibitory postsynaptic currents (mIPSCs) in various regions of the hippocampus, including further studies on A1 agonists (111, 122–124) and also baclofen (GABAB) (123, 124), somatostatin (125), acetylcholine (muscarinic) (126), glutamate (metabotropic) (126, 127), and opioid (mu) agonists (128). In all of these instances, the authors demonstrated that receptor activation produced a reduction in the frequency of mEPSCs under conditions allowing them to conclude that release was inhibited at a point beyond Ca entry. Similar results have been reported in other parts of the brain as well (100, 113). As in the case of Silinsky’s original report, several studies have attempted to increase transmitter release by employing various agents that enhance secretion while bypassing Ca channels. These agents include polycations such as Gd (123) and ruthenium red (122) as well as the Ca ionophore ionomycin (123) and α-latrotoxin from black widow spider venom (123). In all of these instances, agonists at presynaptic receptors (µ-opioid, A1, and GABAB) were still effective at reducing the frequency of mEPSCs and mIPSCs, further supporting a site of action downstream of Ca entry. Nevertheless, not all investigations have reached the same conclusion. There are studies on the effects of mGluR (128), A1 (129), and GABAB (130) agonists as well as NPY (131) in the hippocampus that found no effect on mEPSC and mIPSC frequency, although each agonist strongly suppressed evoked transmitter release. Curi- ously, these studies all utilized slice preparations, whereas the former group of studies, reporting a downstream point of action, all utilized cultured neurons. Whether this is merely a coincidence or whether it indicates something more significant is unclear. In cerebellar slices, however, Dittman & Regehr (100) did show that a component of the presynaptic inhibition caused by baclofen at parallel fiber/Purkinje cell synapses involved a downsteam effect, and Nicola

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

214 MILLER

& Malenka (113) came to similar conclusions in their study of inhibition of EPSPs by dopamine in nucleus accumbens slices. Recent observations in hy- pothalamic slices have also noted a downstream component in the presynaptic inhibition of glutamate release produced by the peptide galanin in the arcuate nucleus (G Kinney & RJ Miller, unpublished observations). Results obtained on the mechanism of hormone secretion from different types of endocrine cells parallel the results discussed here for neurons. Just as with neurons, stimulation of different types of G-protein-linked receptors has often been shown to inhibit the evoked release of hormone-containing vesicles. As release in these cases is usually the result of Ca inﬂux through voltage-dependent Ca channels, there are obvious similarities between the two processes. Indeed, it has often been found that activation of receptors will inhibit Ca currents in these cells (132). Nevertheless, several studies have demonstrated that activation of receptors can still inhibit hormone release from “permeabilized” endocrine cells in which release is no longer dependent on Ca inﬂux via Ca channels. Thus, in these cases an inhibitory mechanism downstream of Ca entry has also been in- voked (133, 134). At any rate, whether one considers the situation in endocrine cells or neurons, nobody has any idea what the molecular basis for such effects might be. Recently, Krasnoperov et al (135) made an intriguing observation about the mechanism of action of α-latrotoxin (135). This toxin increases transmitter release by two mechanisms, one Ca dependent and one Ca independent. These mechanisms involve the binding of the toxin to two receptors. The Ca- independent effects of the toxin appear to be mediated by a G-protein-linked receptor that is a member of the extended secretin receptor family. It appears that this receptor can interact with syntaxin 1A and synaptotagmin. Thus, there is now a precedent for a direct interaction between a G-protein-linked receptor and protein components of the transmitter release apparatus. Might this be an archetype for how downstream regulation of release occurs? In keeping with this possibility, it has recently been demonstrated that synaptosomal muscarinic receptors can directly interact with several components of the release apparatus, including syntaxin, synaptotagmin, and SNAP-25 (136). Quantitative Aspects of Release The above discussion indicates that there is evidence for two or possibly three types of effects contributing to presynaptic inhibition produced by G-protein- linked receptors at vertebrate synapses. Some attempts have been made to try to estimate which effects are quantitatively the most important. Given the fact that little evidence exists supporting a role for K channels, except in certain instances, most of the arguments concern the relative importance of inhibitory mechanisms operating directly on or downstream of Ca entry. As might be expected, there is no consensus on the answer. In studies in which downstream

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 215

effects on transmitter release have not been observed, conclusions obviously favor inhibition of Ca channels as the sole mechanism of importance (128, 130). The opposing view has also been put forward. Thompson and colleagues (111, 124, 126) have demonstrated that although activation of several types of presynaptic receptors in the hippocampus reduced mEPSC and mIPSC frequency, Cd produced no effect under the same conditions. The same group also demonstrated through the use of selective toxins, such as botulinum toxin, that release leading to the production of mIPSCs and mEPSCs appears to involve the same proteins and mechanisms as evoked release (123). Hence, it is argued that inhibition of presynaptic Ca channels is unnecessary for presynaptic inhibition to occur and that it can be completely accounted for by downstream mechanisms. In support of this view are instances in which the activation of receptors produced quantitatively equivalent inhibition of evoked and mIPSCs and mEPSCs. However, even though it is clear that effects on Ca influx are not a sine quae non for presynaptic inhibition to occur, this does not necessarily mean that inhibition of Ca influx is not normally an important factor. Indeed, other estimates have concluded that although the downstream mechanism may play a role (4, 100), an effect on Ca influx is of primary importance, although precise quantitative estimates as to the relative contributions of the two mechanisms differ depending on the situation.

IONOTROPIC RECEPTORS In addition to the effects of G-protein-linked receptors, it is also clear that presynaptic ionotropic receptors exist and that these may be very influential in the regulation of neurotransmitter release. Evidence exists for presynaptic GABAA/C, nicotinic cholinergic, and diverse types of glutamate receptors in various parts of the nervous system whose activation can modulate transmitter release. All of these receptors are multisubunit ligand–gated ion channels. Receptor activation leads directly to the opening of ion channels in the presynaptic terminal, and the subsequent redistribution of ions influences transmitter release. There are several ways this could happen. For example, as discussed above, activation of an ion channel that is permeable to K or Cl may produce shunting of the action potential in the presynaptic terminal, resulting in less effective opening of Ca channels and subsequent inhibition of release. Alter- natively, and this is considered to be more likely, activation of channels leading to the redistribution of Cl or Na might depolarize the terminal, leading to inactivation of voltage-dependent Na and Ca channels and inhibition of the spread of the action potential into the terminal and the influx of Ca. On the other hand, if the channel activated is actually appreciably permeable to Ca, one might expect enhancement of transmitter release due to increased Ca influx

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

216 MILLER

directly into the terminal. This might subsequently be followed by an inhibition of release when stores of transmitter are exhausted. There is evidence that mixtures of all of these processes are actually utilized in practice. Although it is not a neurotransmitter, the drug capsaicin binds to a presynaptic receptor on small nonmyelinated sensory nociceptors, causing the opening of an ion channel that is permeable to both Ca and Na (137, 138). This opening results in increased neuronal excitability and Ca inﬂux, initially producing increased transmitter (substance P) release in the spinal cord and pain. However, when stores of transmitter are exhausted, hypoalgesia results. Indeed, the drug is used topically in the treatment of disorders such as shingles for this very reason.

PRESYNAPTIC GABAA/C RECEPTORS The first bona fide example of presynaptic inhibition was actually a reduction in Group 1A EPSPs in spinal motoneurons, described by Frank & Fuortes (139), that could not be accounted for by a postsynaptic mechanism. It has subsequently been determined that this effect is mediated by GABA and that similar GABA-mediated presynaptic inhibition is found throughout the nervous system (140). The receptors involved in the presynaptic inhibitory effects of GABA are of different types (36). G-protein-coupled baclofen-sensitive GABAB receptors clearly contribute to such effects at many synapses (see above). In addition, however, activation of ionotropic GABAA (141) and GABAC (142) receptors is believed to constitute a widely distributed mechanism that mediates presynaptic inhibition in, among other places, the hippocampus, striatum, substantia nigra, retina, spinal cord, and posterior pituitary (34). Such a role has also been suggested by the many anatomical and neurochemical studies on the effects of GABAA agonists in different neuronal preparations (34, 141, 142). The ionotropic GABA receptors constitute a family of multisubunit hetero- multimers containing diverse combinations of subunits. At this time, α1–6, β1–4, γ 1–4, δ, ε and ρ1–3 subunits have been identified (141–144). They are homologous to nicotinic cholinergic receptor subunits as well as to each other. GABAA receptors are thought to minimally contain αβγ, αβδ,orαβε subunits in what is probably a pentameric array. Numerous clinically important classes of drugs interact with these receptors, and the selectivity of these interactions is determined by receptor subunit combinations (141, 142). For example, the binding of benzodiazepines is dependent on the presence of the γ subunit (141, 142). Receptors containing ρ subunits are quite exceptional, however (142). First, it is thought that they may assemble into homooligomeric channels (GABAC receptors), and furthermore, they do not bind many of the drugs that define the pharmacology of classical GABAA receptors—they are not blocked by bicuculline, for example (142, 145).

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 217

The exact composition of the ionotropic GABA receptors that exist in nerve terminals is unknown. In the case of the terminals of retinal bipolar cells, pharmacological analysis has suggested that the receptors involved are GABAC, and so are presumably made up only of ρ subunits (142, 146). These subunits have also been found in other parts of the brain as well, but whether they are presynaptic is unknown (147). Recordings made directly from the terminals of posterior pituitary neurosecretory neurons have indicated the presence of GABAA receptors (38). The pharmacology and functional characteristics of these receptors suggest that they may have a different subunit composition from GABAA receptors found in other parts of neurons, and this may be true for presynaptic GABAA receptors in other parts of the nervous system as well (148, 149). Activation of GABAA/C receptors leads to the opening of anion-selective channels (141, 147, 150). The permeability of these channels to Cl and HCO3 presumably underlies their ability to produce presynaptic inhibition, although the exact mechanism involved is not entirely clear. Although it was initially believed that the effect of GABAA/C receptor activation resulted from action potential shunting, modeling studies (37) and direct recording from nerve terminals (38) have suggested that it may primarily result from depolarization of terminals and inactivation of Na and Ca channels.

PRESYNAPTIC NICOTINIC RECEPTORS A wealth of data from the neurochemical literature suggests that presynaptic nicotinic receptors exist in the CNS and that their activation enhances transmitter release (34, 151, 152). Anatomical data on the localization of these receptors support this contention in several instances (34). Recent studies have started to analyze the mechanism of action of nicotinic agonists. It is known that a large family of nicotinic receptor subunits exist in the nervous system, including families of α(2–9) and β(2–4) subunits (34, 151, 152). It is believed that nicotinic receptors in the brain are made up of heterologous combinations of α and β subunits, although in some cases (e.g. α7) homologous arrangements of subunits are likely to occur (153). By analogy with nicotinic receptors at the neuromuscular junction, it is also probable that receptors in the brain are pentameric (152). Receptors that contain α7 are highly permeable to Ca, although other subunit combinations also retain an appreciable Ca permeability (154–157). It is also clear that CNS receptors that contain the α7 (or α8 or α9) are blocked by α-bungarotoxin (α-βTX), whereas those that contain other α subunits are not (158). Pharmacological studies have indicated that both types of receptors may be involved in the regulation of transmitter release in different circumstances (34, 159, 160).

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

218 MILLER

In some instances stimulatory effects of nicotinic agonists on transmitter release have been reported to be inhibited by TTX (161–163). This is true, for example, for nicotinic receptors on GABAergic terminals in the chick lateral spiriform nucleus (162) and rat interpeduncular nucleus (161). The inhibitory effects of TTX have been interpreted as indicating that the receptors in these cases are actually “preterminal” and that the effects are exerted by triggering action potentials that subsequently invade the nerve terminal. On the other hand, several other reports have demonstrated stimulatory effects of nicotinic agonists that are resistant to TTX, and it is thought that the receptors in these cases are truly presynaptic, i.e. in the nerve terminal itself. Nicotinic agonists enhance mPSCs in the rat olfactory bulb (OB) (164), interpeduncular nucleus (IPN), lum- bar sympathetic ganglion (LSG) (165), and the CA3 region of the hippocampus (166). In these instances, block with α-βTX and/or antisense knockout of α7 subunits have indicated their participation in the response (164–166). Experi- ments in the CA3, IPN, and LSG also showed that nicotinic agonists increased [Ca]i in nerve terminals from which transmitter release occurred. In the CA3 it was demonstrated that nicotinic agonists were still effective in stimulating transmitter release and raising [Ca]i in mossy fiber terminals in the presence of Cd as well as TTX, indicating that at least a component of the Ca flux into these terminals occurred directly through nicotinic receptors rather than through voltage-sensitive Ca channels. Studies in the dorsolateral geniculate (DLG) (167, 168) and the ventrobasal complex (VB) (167), however, demonstrated that the stimulation of GABA release (mIPSCs) produced by nicotinic agonists was insensitive to α-βTX but was absent in β2 subunit knockout mice, indicating that the receptors involved probably contained β2 plus an α subunit such as α3 or α4 (167). Interestingly, the stimulation produced by nicotinic agonists was insensitive to Cd (which blocked the effects of high K) in the DLG but substantially blocked by Cd in the VB (167). Taken together, these results suggest that presynaptic nicotinic receptors of various types can stimulate the release of transmitters by causing direct Ca flux into presynaptic terminals and also causing influx through voltage-dependent Ca channels that open as a result of depolarization. The precise contributions of these two processes seems to differ according to the situation. One surprising observation, however, is that nicotinic agonists also increased the evoked release of transmitter in the VB and IPN (165, 167). This is interesting in view of the fact that if depolarization of the terminal had occurred in these situations, one would have expected to see an inhibition of evoked release as observed with presynaptic GABAA or kainate receptors (see below). At any rate, the facilitation of release produced by nicotinic agonists may be of considerable importance therapeutically in several neu- rodegenerative disorders such as Parkinson’s and Alzheimer’s disease in which

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 219

deﬁcits in transmitter release are clearly a problem. Similar considerations to those discussed for nicotinic receptors may also apply to P2X ATP receptors, which have been shown to regulate transmitter release in the spinal cord (169).

PRESYNAPTIC GLUTAMATE RECEPTORS Ionotropic glutamate receptors also constitute a large family of widely distributed subunits that can be organized into heterologous combinations of AMPA, kainate, and NMDA receptors (170–172). These receptors, which like nicotinic and GABAA receptors are likely to be pentamers, are ligand-gated cation channels that differ in their pharmacological speciﬁcity and in the details of their ion selectivity, particularly their permeability to Ca. It is clear that AMPA and NMDA receptors transduce excitatory transmission at many glutamate-utilizing synapses in the CNS (173). On the other hand, the physiological functions of kainate receptors in the CNS are far from clear (172, 173, but see 174, 175). Neurochemical data obtained using synaptosomal preparations from different parts of the brain have suggested that all three types of glutamate receptors may also act as presynaptic receptors. Activation of synaptosomal AMPA and NMDA receptors has been generally linked to the enhancement of transmitter release from different parts of the brain (176–180). However, the best data supporting a presynaptic receptor function come from recent studies on the role of kainate receptors in the hippocampus. In contrast to the stimulatory effects of AMPA-selective agonists, kainate inhibited the release of glutamate from hippocampal synaptosomes, and these effects were suggestive of the activation of kainate rather than AMPA receptors (181). Electrophysiological studies in the hippocampal slice demonstrated that kainate had two effects on NMDA receptor–mediated EPSCs evoked in the CA1 (181). Following a brief enhancement of the EPSC, a long-term depression was observed. Here again, the selectivity of the effect suggested activation of kainate rather than AMPA receptors. Recently, Clarke et al (182) developed two drugs that are selective for GluR5 kainate receptors: the agonist ATPA and the antagonist LY294486. The authors used these tools to demonstrate that activation of presynaptic GluR5- like kainate receptors regulated the release of GABA from terminals in the CA1, as was also observed in another recent study (183). Both kainate and ATPA inhibited evoked GABAergic IPSCs in CA1, and their effects were blocked by LY294486. Thus, in the hippocampus, presynaptic kainate receptors appear to regulate both excitatory and inhibitory synapses. Although the mechanism of action of kainate receptors is still unclear, depolarization of terminals leading to the inactivation of Na and Ca channels is likely.

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

220 MILLER

Visit the Annual Reviews home page at http://www.AnnualReviews.org.

Literature Cited

1. Langer SV. 1997. 25 years since the 15. Stanley EF. 1997. The calcium channel discovery of presynaptic receptors: pre- and the organization of the presynaptic sent knowledge and future perspectives. release face. Trends Neurosci. 20:404–9 Trends Pharmacol. Sci. 18:95–99 16. Geppert M, Goda Y, Hammer RE, Li C, 2. Nicoll RA, Alger BE. 1979. Presynaptic Rosahl TW, et al. 1994. Synaptotagmin inhibition: transmitter and ionic mecha- 1: a major Ca release sensor for trans- nisms. Int. Rev. Neurobiol. 21:217–58 mitter release at a central synapse. Cell 3. Miller RJ. 1990. The receptor mediated 79:717–27 regulation of calcium channels and neu- 17. Shao X, Cai L, Fernandez I, Zhang rotransmitter release. FASEB J. 4:3291– X, Sudhof TC, et al. 1997. Synapto- 300 tagmin—syntaxin interaction: the C2 4. Wu LG, Saggau P. 1997. Presynaptic in- domain as a Ca dependent electrostatic hibition of elicited neurotransmitter re- switch. Neuron 18:133–42 lease. Trends Neurosci. 20:204–12 18. Bommert K, Charlton MP, De Bello 5. Matthews G. 1996. Neurotransmitter re- WM, Chin GJ, Betz H, et al. 1993. Inhi- lease. Annu. Rev. Neurosci. 19:219–33 bition of neurotransmitter release by C2 6. Sudhof TC. 1995. The synaptic vesicle: domain peptides implicates synaptotag- a cascade of protein—protein interac- min in exocytosis. Nature 363:163–65 tions. Nature 375:645–53 19. Sudhof TC. 1997. Function of Rab3 7. Zucker RS. 1996. Exocytosis: a molec- GDP-GTP exchange. Neuron 18:519–22 ular and physiological perspective. Neu- 20. Geppert M, Goda Y, Stevens CF, Sudhof ron 17:1049–55 TC. 1997. The small GTP-binding pro- 8. Heuser JS, Reese TS. 1981. Structural tein Rab3A regulates a late step in synap- changes after neurotransmitter release at tic vesicle fusion. Nature 387:810–14 the frog neuromuscular junction. J. Cell 21. Geppert M, Bolshakov VY, Siegelbaum Biol. 88:564–80 SA, Takel K, De Camilli P, et al. 1994. 2 9. Bennett MK. 1997. Ca + and the regu- The role of Rab3Ainneurotransmitter lation of neurotransmitter release. Curr. release. Nature 369:493–97 Opin. Neurobiol. 7:316–22 22. Bean AJ, Scheller RH. 1997. Better late 10. De Bello WM, O’Connor V,Dresbach T, than never: a role for Rabs late in exo- Whiteheart SW, Wang SSH, et al. 1995. cytosis. Neuron 19:751–54 SNAP-mediated protein-protein interac- 23. Schiavo G, Gu QM, Prestwich GD, Soll- tions essential for neurotransmitter re- ner TH, Rothman JE. 1996. Calcium lease. Nature 373:626–30 dependent switching of the specificity 11. O’Connor VM, Shamotienko O, Grishin of phosphoinositide binding to synap- E, Betz H. 1993. On the structure of totagmin. Proc. Natl. Acad. Sci. USA the “synaptosecretosome.” FEBS Lett. 93:13327–32 326:255–60 24. Miller RJ. 1997. Calcium channels prove 12. Leveque C, El Far O, Martin-Moutot N, to be a real headache. Trends Neurosci. Sato K, Kato R, et al. 1994. Purification 20:189–92 2 of the N type calcium channel associ- 25. Snutch TP, Reiner PB. 1992. Ca + chan- ated with syntaxin and synaptotagmin. nels: diversity of form and function. J. Biol. Chem. 269:6306–12 Curr. Opin. Neurobiol. 2:247–53 13. Martin-Moutot N, Charvin N, Leveque 26. Dunlap K, Luebke JI, Turner TJ. 1995. C, Sato K, Nishiki T, et al. 1996. Interac- Exocytotic Ca channels in mammalian tion of SNARE complexes with P/Q type central neurons. Trends Neurosci. 18: calcium channels in rat cerebellar synap- 89–98 tosomes. J. Biol. Chem. 271:6567–70 27. Sheng ZH, Rettig J, Takahashi M, Cat- 14. Bennett MK, Calakos N, Scheller RH. terall WA. 1994. Identification of a syn- 1992. Syntaxin, a synaptic protein im- taxin binding site on N type calcium plicated in docking of synaptic vesi- channels. Neuron 13:1303–13 cles at presynaptic active zones. Science 28. Mochida S, Sheng ZH, Baker C, 257:255–59 Kobayashi H, Catterall, WA. 1996.

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 221

Inhibition of neurotransmission by pep- release at a cerebellar synapse. Neuron tides containing the synaptic protein in- 5:675–88 teraction site of N type Ca channels. 43. Doroshenko PA, Woppmann A, Mil- Neuron 17:781–88 janich G, Augustin GJ. 1997. Pharma- 29. Sheng ZH, Rettig J, Cook T, Catterall cologically distinct presynaptic calcium WA. 1996. Calcium dependent interac- channels in cerebellar excitatory and in- tion of N type calcium channels with the hibitory synapses. Neuropharmacology synaptic core complex. Nature 379:451– 36:865–72 54 44. Regehr WG, Mintz IM. 1994. Participa- 30. Reltig J, Heinemann C, Ashery V,Sheng tion of multiple calcium channel types in Z-H, Yokoyama CT, et al. 1997. Alter- transmission at single climbing fibers to ation of Ca dependence of neurotrans- Purkinje cell synapses. Neuron 12:605– 2 mitter release by disruption of Ca + 13 channel/syntaxin interaction. J. Neu- 45. Wu LG, Saggau P. 1995. Block of mul- rosci. 17:6647–56 tiple presynaptic calcium channel types 31. Yokoyama CT, Sheng Z-H, Catter- by ω-conotoxin-MVIIC at hippocampal all WA. 1997. Phosphorylation of the CA3 to CA1 synapses. J. Neurophysiol. synaptic protein interaction site on 73:1965–72 N-type calcium channels inhibits inter- 46. Heidelberger R, Matthews G. 1992. Cal- actions with SNARE proteins. J. Neu- cium influx and calcium current in sin- rosci. 17:6929–38 gle synaptic terminals of goldfish retinal 32. Catterall WA. 1991. Excitation-contrac- bipolar neurons. J. Physiol. 447:235–56 tion coupling in vertebrate skeletal mus- 47. Hille R. 1994. Modulation of ion channel cle: a tale of two calcium channels. Cell function by G-protein coupled receptors. 64:871–74 Trends Neurosci. 17:531–36 33. Miller RJ. 1990. The receptor mediated 48. Dolphin AC. 1995. Voltage dependent regulation of calcium channels and neu- calcium channels and their modulation rotransmitter release. FASEB J. 4:3291– by neurotransmitters and G-proteins. 300 Exp. Physiol. 80:1–36 34. McGehee DS, Role LW. 1996. Presy- 49. Jones SW, Elmslie KS. 1997. Trans- naptic ionotropic receptors. Curr. Opin. mitter modulation of neuronal calcium Neurobiol. 6:342–49 channels. J. Membr. Biol. 155:1–10 35. Gage PW. 1992. Activation and modula- 50. Bean BP. 1989. Neurotransmitter inhi- tion of neuronal K channels by GABA. bition of neuronal calcium currents by Trends Neurosci. 15:46–51 changes in channel voltage dependence. 36. Segev I. 1990. Computer study of presy- Nature 340:153–56 naptic inhibition controlling the spread 51. Patel PG, De Leon M, Reed RR, Dubel of action potentials into axonal termi- S, Snutch TP, et al. 1996. Elementary nals. J. Neurophysiol. 633:987–98 events underlying voltage dependent G- 37. Graham B, Redman S. 1994. A stim- protein inhibition of N-type calcium cur- ulation of action potentials in synaptic rents. Biophys. J. 71:2509–21 boutons during presynaptic inhibition. J. 52. Boland LM, Bean BP. 1993. Modu- Neurophysiol. 71:538–49 lation of N-type calcium channels in 38. Zhang SJ, Jackson MB. 1993. GABA bullfrog sympathetic neurons by luteini- activated chloride channels in secretory zing hormone releasing hormone: kinet- nerve endings. Science 259:531–34 ics and voltage dependence. J. Neurosci. 39. Wheeler DB, Randall A, Tsien RW. 13:516–33 1994. Roles of N type and Q type Ca 53. Elmslie KS, Kammermeier PJ, Jones channels in suporting hippocampal syn- SW. 1992. LHRH and GTP-γ -S mo- aptic transmission. Science 264:107–11 dify calcium current activation in bull- 40. Scholz KP, Miller RJ. 1995. Develop- frog sympathetic neurons. Neuron 5:75– mental changes in presynaptic calcium 80 channels coupled to glutamate release in 54. Budai D, Duckles SP. 1989. Opioid cultured hippocampal neurons. J. Neu- induced prejunctional inhibitory vaso- rosci. 15:4612–17 constriction in the rabbit ear artery: 41. Takahashi T, Momiyama A. 1993. Dif- α-adrenoceptor activation and exter- ferent types of calcium channels medi- nal calcium. J. Pharmacol. Exp. Ther. ate central synaptic transmission. Nature 251:497–501 3666:156–58 55. Grundemar L, Widmark E, Waldeck B, 42. Mintz IM, Sabatini BL, Regehr WG. Hakanson R. 1988. NPY: presynaptic 1995. Calcium control of transmitter inhibition of vagally induced contrac-

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

222 MILLER

tions in the guinea-pig trachea. Reg. Pep. 68. Zhang JF, Ellinor PT, Aldrich RW, Tsien 23:309–13 RW. 1996. Multiple structural elements 56. Penington NJ, Kelly JS, Fox AP. 1992. in voltage dependent Ca channels sup- Action potential waveforms reveal si- port their inhibition by G-proteins. Neu- multaneous changes in ICa and IK pro- ron 17:991–1003 duced by 5-HT in rat dorsal raphe 69. Toth PT, Shekter LR, Ma GH, Philipson neurons. Proc. R. Soc. London Ser. B LH, Miller RJ. 1996. Selective G-protein 248:171–79 regulation of neuronal calcium channels. 57. Toth PT, Miller RJ. 1995. Calcium and J. Neurosci. 16:4617–24 sodium currents evoked by action poten- 70. Page KM, Stephens GJ, Berrow NS, tial waveforms in rat sympathetic neu- Dolphin AC. 1997. The intracellular rones. J. Physiol. 485:43–57 loop between domains I and II of the 58. Miller RJ. 1991. The control of neu- β-type calcium channel confers aspects ronal Ca homeostasis. Prog. Neurobiol. of G-protein sensitivity to the E-type 37:255–85 calcium channel. J. Neurosci. 17:1330– 59. Ikeda S. 1996. Voltage dependent mod- 38 ulation of N-type calcium channels by 71. Qin N, Platino D, Olcese R, Stefani E, G-protein β/γ subunits. Nature 380: Birnbaumer L. 1997. Direct interaction 255–58 of Gβγ with a C-terminal Gβγ-binding 2 60. Herlitze S, Garcia DE, Mackie K, Hille domain of the Ca + channel α1 subunit B, Scheuer T, et al. 1996. Modulation of is responsible for channel inhibition by Ca channels by G-protein β/γ subunits. G protein coupled receptors. Proc. Natl. Nature 380:258–62 Acad. Sci. USA 94:8866–71 61. Shekter LR, Taussig R, Gillard SE, 72. Pragnell MP, DeWaard M, Mori Y, Tan- Miller RJ. 1997. Regulation of human abe T, Snutch TP, et al. 1994. Calcium neuronal calcium channels by G-protein channel β-subunit binds to a conserved β/γ subunits expressed in human embry- motif in the I-II cytoplasmic linker of the onic kidney 293 cells. Mol. Pharmacol. α1 subunit. Nature 368:67–70 52:282–91 73. Campbell V,Berrow NS, Fitzgerald EM, 62. Huang C, Slesinger PA, Casey PJ, Jan Brickley K, Dolphin AC. 1995. Inhibi- YN, Jan LY. 1995. Evidence that direct tion of the interaction of G-protein Go binding of Gβ/γ to the GIRK1 G-protein with calcium channels by the calcium inwardly rectifying K+ channel is im- channel β-subunit in rat neurones. J. portant for channel activation. Neuron Physiol. 485:365–72 15:1133–43 74. Olcese R, Quin N, Schneider T, Neely 63. Chen J, DeVivo M, Dingles J, Harry A, A, Wei X, et al. 1994. The amino termi- Li J, et al. 1995. A region of adenylyl cy- nus of a Ca channel β-subunit sets rates clase 2 critical for regulation by G-pro- of channel inactivation independently of tein subunits. Science 268:1166–69 the subunit’s effect on activation. Neuron 64. Zamponi G, Bourinet WE, Nelson D, 13:1433–38 Nargeot J, Snutch TP. 1997. Crosstalk 75. Swartz KJ, Merritt A, Bean BP,Lovinger between G-proteins and protein kinase DM. 1993. Protein kinase C modulates 2 C mediated by the calcium channel a1 glutamate receptor inhibition of Ca + subunit. Nature 385:442–46 channels and synaptic transmission. Na- 65. DeWaard M, Lui H, Walker D, Scott ture 361:165–68 VES, Gurnett CA, et al. 1997. Di- 76. Bezprozvanny J, Scheller RH, Tsien rect binding of G-protein β/γ complex RW.1995. Functional impact of syntaxin to voltage dependent calcium channels. on gating of N type and Q type calcium Nature 385:446–50 channels. Nature 378:623–26 66. Herlitze S, Hockerman GH, Scheuer T, 77. Wiser O, Bennett MK, Atlas D. 1996. Catterall WA. 1997. Molecular deter- Functional interaction of syntaxin and minants of inactivation of G-protein SNAP-25 with voltage-sensitive L and modulation in the intracellular loop con- N type Ca channels. EMBO J. 15:4100– necting domains I and II of the calcium 10 channel α1A subunit. Proc. Natl. Acad. 78. Wiser O, Tobi D, Trus M, Atlas D. 1997. Sci. USA 94:1512–16 Synaptotagmin restores kinetic proper- 67. Bourinet E, Soong TW, Stea A, Snutch ties of a syntaxin associated N type TP. 1996. Determinants of the G-protein voltage sensitive calcium channel. FEBS dependent opioid modulation of neu- Lett. 404:203–7 ronal calcium channels. Proc. Natl. 79. Stanley EF, Mirotznik RR. 1997. Cleav- Acad. Sci. USA 93:1486–91 age of syntaxin prevents G-protein

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 223

regulation of presynaptic calcium chan- 91. Yawo H, Chuma N. 1993. Preferen- nels. Nature 385:340–43 tial inhibition of ω-conotoxin sensitive 2 80. Bleakman D, Thayer SA, Glaum SR, presynaptic Ca + channels by adenosine Miller RJ. 1990. Bradykinin induced receptors. Nature 365:256–58 modulation of calcium signals in rat dor- 92. Takahashi T, Forsythe ID, Tsujimoto sal root ganglion neurons in vitro. Mol. T, Barnes-Davies M, Onodera K. 1996. Pharmacol. 38:785–96 Presynaptic calcium current modulation 81. Wilk-Blaszczak MA, Singer WD, Be- by a metabotropic glutamate receptor. lardetti F. 1996. Three distinct G-protein Science 274:594–98 pathways mediate inhibition of neuronal 93. Borst JGG, Helmchen F, Sakmann B. calcium current by bradykinin. J. Neuro- 1995. Pre and postsynaptic whole cell physiol. 76:3559–62 recordings in the medial nucleus of the 82. Wilk-Blaszczak MA, Singer WD, Quill trapezoid body of the rat. J. Physiol. T, Miller B, Frost JA, et al. 1997. 489:825–40 The monomeric G-proteins Racl and/or 94. Stanley EF, Cox C. 1991. Calcium chan- Cdc42 are required for the inhibition of nels in the presynaptic nerve terminal of voltage dependent calcium current by the chick ciliary ganglion giant synapse. bradykinin. J. Neurosci. 17:4094–100 Ann. NY Acad. Sci. 635:70–79 83. Scholz KP, Miller RJ. 1991. Analysis 95. Toth PT, Bindokas V, Bleakman D, of adenosine actions on calcium cur- Colmers WF, Miller RJ. 1993. Presynap- rents and synaptic transmission in cul- tic inhibition by neuropeptide Y is medi- tured rat hippocampal neurones. J. Phys- ated by reduced Ca influx at sympathetic iol. 435:373–93 nerve terminals. Nature 364:635–39 84. Scholz KP, Miller RJ. 1991. GABA-B 96. Wu LG, Saggau P. 1994. Adenosine in- 2 receptor mediated inhibition of Ca + hibits evoked synaptic transmission pri- currents and synaptic transmission in marily by reducing presynaptic calcium cultured rat hippocampal neurones. J. influx in area CA1 of the hippocampus. Physiol. 444:669–86 Neuron 12:1139–48 85. Sakarai T, Westenbroek RE, Rettig J, 97. Wu LG, Saggau P. 1995. GABA-B re- Hell JW, Catterall WA. 1996. Biochem- ceptor mediated presynaptic inhibition ical properties and subcellular distribu- in guinea-pig hippocampus is caused by 2 tion of the β1 and rbA isoforms of α1A a reduction of presynaptic Ca + influx. subunits of brain calcium channels. J. J. Physiol. 485:649–57 Cell Biol. 134:511–28 98. Qian J, Colmers WF, Saggau P. 1997. 86. Westenbroek RE, Hell JW, Warner C, Inhibition of synaptic transmission by Dubel SJ, Snutch TP, et al. 1992. Bio- neuropeptide Y in rat hippocampal area 2 chemical properties and subcellular dis- CA1: modulation of presynaptic Ca + tribution of an N-type calcium channel entry. J. Neurosci. 17:8169–77 α1 subunit. Neuron 7:1099–115 99. Qian J, Saggau P. 1997. Presynaptic in- 87. Jones OT, Kunze DL, Angelides KJ. hibition of synaptic transmission in the 1989. Localization and mobility of rat hippocampus by activation of mus- ω-conotoxin sensitive Ca channels in carinic receptors: involvement of presy- hippocampal CA1 neurons. Science naptic calcium influx. Br. J. Pharmacol. 244:1189–93 122:511–19 88. Reuter H. 1995. Measurements of ex- 100. Dittman JS, Regehr WG. 1996. Contri- ocytosis from single presynaptic nerve butions of calcium dependent and cal- terminals reveal heterogeneous inhibi- cium independent mechanisms to presy- 2 tion by Ca + channel blockers. Neuron naptic inhibition at a cerebellar synapse. 14:773–79 J. Neurosci. 16:1623–33 89. Reid CA, Clements JD, Bekkers JM. 101. Kretz R, Shapiro E, Bailey CH, Chen M, 1997. Non-uniform distribution of Ca Kandel ER. 1986. Presynaptic inhibition channel subtypes on presynaptic termi- produced by an identified presynaptic in- nals of excitatory synapses in hippocam- hibitory neuron. II. Presynaptic conduc- pal cultures. J. Neurosci. 17:2738–45 tance changes caused by histamine. J. 90. Wu L-G, Saggau P. 1994. Pharmaco- Neurophysiol. 55:131–46 logical identification of two types of 102. Chuang HH, Jan YN, Jan LY.1997. Reg- presynaptic voltage dependent calcium ulation of IRK3 inward rectifier K chan- channels at CA3-CA1 synapses of the nel by M1 acetylcholine receptors and hippocampus. J. Neurosci. 14:5613– intracellular magnesium. Cell 89:1121– 22 32

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

224 MILLER

103. Nicoll RA, Malenka RC, Kauer JA. synaptic transmission by distinct mecha- 1990. Functional comparison of neuro- nisms in the nucleus accumbens. J. Neu- transmitter receptor subtypes in mam- rosci. 17:5697–700 malian central nervous system. Physiol. 114. Lambert NA, Harrison NL, Teyler TJ. Rev. 70:514–65 1991. Baclofen induced disinhibition in 104. Krapivinsky G, Gordon EA, Wickman arch CA1 of rat hippocampus is resis- 2 B, Velimirovic L, Krapavinsky L, et al. tant to extracellular Ba +. Brain Res. 1995. The G protein gated atrial K+ 547:349–52 channel, IKACh, is a heteromultimer of 115. Scholfield CN, Steel L. 1988. Presynap- two inwardly rectifying K+ channel pro- tic K channel blockade counteracts the teins. Nature 374:135–41 depressant effect of adenosine in olfac- 105. Signorini S, Liao YJ, Duncan SA, Jan tory cortex. Neuroscience 24:81–91 LY, Stoffel M. 1997. Normal cerebel- 116. Luscher C, Jan LY, Stoffel M, Malenka lar development but susceptibility to RC, Nicoll RA. 1997. G-protein-coup- seizures in mice lacking G protein cou- led inwardly rectifying K+ channels pled, inwardly rectifying K+ channel (GIRKS) mediate postsynaptic but not GIRK2. Proc. Natl. Acad. Sci. USA 94: presynaptic transmitter actions in hip- 923–27 pocampal neurons. Neuron 19:687–95 106. Huang CL, Slesinger PA, Casey BJ, Jan 117. Simmons ML, Chavkin C. 1996. κ-opi- YN, Jan LY. 1995. Evidence that direct oid receptor activation of a dendrotoxin binding of Gβγ to the GIRK1 G protein sensitive potassium channel mediates gated inwardly rectifying K+ channel is presynaptic inhibition of mossy fiber important for channel activation. Neuron neurotransmitter release. Mol. Pharma- 15:1133–43 col. 50:80–85 107. Drake CT, Bausch SB, Milner TA, 118. Scholz KP, Miller RJ. 1996. Presynap- Chavkin C. 1997. GIRK1 immunoreac- tic inhibition at excitatory hippocam- tivity is present predominantly in den- pal synapses: development and role of drites, dendritic spines and somata in the presynaptic Ca channels. J. Neurophys- CA1 region of the hippocampus. Proc. iol. 76:39–46 Natl. Acad. Sci. USA 94:1007–12 119. Glaum SR, Miller RJ. 1995. Presynaptic 108. Murer G, Adelbrecht C, Lauritzen I, metabotropic glutamate receptors mod- Lesage F, Lazdunski M, et al. 1997. An ulate ω-conotoxin-GVIA-insensitive cal immunocytochemical study on the dis- cium channels in the rat medulla. Neu- tribution of two G-protein-gated inward ropharmacology 34:953–64 rectifier potassium channels (GIRK2 120. Silinsky EM. 1984. On the mechanism and GIRK4) in the adult rat brain. Neu- by which adenosine receptor activa- roscience 80:345–57 tion inhibits the release of acetylcholine 109. Lesage F, Guillemare E, Fink M, Duprat from motor nerve endings. J. Physiol. F, Heurteaux C, et al 1995. Molecu- 346:243–56 lar properties of neuronal G-protein- 121. Scholz KP, Miller RJ. 1992. Inhibition activated inwardly rectifying K+ chan- of quantal transmitter release in the ab- nels. J. Biol. Chem. 270:28660–67 sence of calcium influx by a G-protein 110. Ponce A, Bueno E, Kentros C, de Miera linked adenosine receptor at hippocam- EV-S, Chow A, et al. 1996. G-protein pal synapses. Neuron 8:1139–50 gated inward rectifier K+ channel pro- 122. Trudeau LE, Doyle RT, Emery DG, Hay- teins (GIRK1) are present in the soma don PG. 1996. Calcium independent ac- and dendrites as well as in nerve termi- tivation of the secretory apparatus by nals of specific neurons in the brain. J. ruthenium red in hippocampal neurons: Neurosci. 16:1990–2001 a new tool to assess modulation of presy- 111. Thompson SM, Capogna M, Scanziani naptic function. J. Neurosci. 16:46–54 M. 1993. Presynaptic inhibition in the 123. Capogna M, Gahwiler BH, Thompson hippocampus. Trends Neurosci. 16:222– SM. 1996. Presynaptic inhibition of cal- 27 cium dependent and independent release 112. Thompson SM, Gahwiler BH. 1992. elicited with ionomycin, gadolinium and Comparison of the actions of baclofen α-latrotoxin in the hippocampus. J. Neu- at pre and postsynaptic receptors in the rophysiol. 75:2017–28 rat hippocampus in vitro. J. Physiol. 124. Scanziani M, Capogna M, Gahwiler BH, 451:329–45 Thompson SM. 1992. Presynaptic inhi- 113. Nicola SM, Malenka RC. 1997. Dopa- bition of miniature excitatory synaptic mine depresses excitatory and inhibitory currents by baclofen and adenosine in

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 225

the hippocampus. Neuron 9:919–27 of the exocytic machinery. J. Physiol. 125. Boehm S, Betz H. 1997. Somatostatin 504:251–56 inhibits excitatory transmission at rat 137. Bleakman D, Brorson JD, Miller RJ. hippocampal synapses via presynaptic 1991. The effect of capsaicin on voltage receptors. J. Neurosci. 17:4066–75 gated calcium currents and calcium sig- 126. Scanziani M, Gahwiler BH, Thompson nals in cultured dorsal root of ganglion SM. 1995. Presynaptic inhibition of ex- cells. Br. J. Pharmacol. 101:423–32 citatory synaptic transmission by mus- 138. Caterina MJ, Shumacher MA, Tominaga carinic and metabotropic glutamate re- M, Rosen TA, Levine JD, Julius D. 1997. ceptor activation in the hippocampus: The capsaicin receptor: a heat activated 2 are Ca + channels involved. Neurophar- ion channel in the pain pathway. Nature macology 34:1549–57 289:816–24 127. Tyler EC, Lovinger DM. 1995. Meta- 139. Frank K, Fuortes MGF. 1957. Presy- botropic glutamate receptor modulation naptic and postsynaptic inhibition of of synaptic transmission in corticostri- monosynaptic reﬂexes. Fed. Proc. 16: atal co-cultures: role of calcium inﬂux. 39–40 Neuropharmacology 34:939–52 140. Nicoll RA, Alger BE. 1979. Presynaptic 128. Gerau RW, Conn PJ. 1995. Mul- inhibition: transmitter and ionic mecha- tiple presynaptic metabotropic gluta- nisms. Int. Rev. Neurobiol. 21:217–58 mate receptors modulate excitatory 141. McKernan RM, Whiting PJ. 1996. and inhibitory synaptic transmission in Which GABA-A receptor subtypes re- hippocampal area CA1. J. Neurosci. ally occur in the brain. Trends Neurosci. 15:6879–89 19:139–43 129. Prince DA, Stevens CF. 1992. Adeno- 142. Johnston GAR. 1996. GABA-C recep- sine decreases neurotransmitter release tors: relatively simple transmitter gated at central synapses. Proc. Natl. Acad. ion channels. Trends Pharmacol. Sci. Sci. USA 89:8586–90 17:319–23 130. Doze VA, Cohen GA, Madison DV. 143. Shingai R, Yanagi K, Fukashima T, 1995. Calcium channel involvement in Sakata K, Ogurusu T. 1996. Func- GABAB receptor mediated inhibition tional expression of GABA ρ3 recep- of GABA release in area CA1 of the rat tors in Xenopus oocytes. Neurosci. Res. hippocampus. J. Neurophysiol. 74:43– 26:387–90 53 144. Whiting PJ, McAllister G, Vasilatis D, 131. McQuision AR, Colmers WF. 1996. Bonnert D, Hearens RP,et al. 1997. Neu- Neuropeptide Y2 receptors inhibit the ronally restricted RNA splicing regu- frequency of spontaneous but not minia- lates the expression of a novel GABA- ture EPSC’s in CA3 pyramidal cells A receptor subunit conferring atypi- of rat hippocampus. J. Neurophysiol. cal functional properties. J. Neurosci. 70:3159–68 17:5027–37 132. Sharp DWG. 1996. Mechanisms of inhi- 145. Strata F, Cherubini E. 1994. Transient bition of insulin release. Am. J. Physiol. expression of a novel type of GABA 271:C1781–99 response in rat CA3 hippocampal neu- 133. Ullrich S, Prentki M, Wollheim CB. rones during development. J. Physiol. 1990. Somatostatin inhibition of Ca in- 480:493–503 duced insulin secretion in permeabilized 146. Matthews G, AyoubGS, Heidelberger R. HIT-Y15 cells. Biochem. J. 70:273– 1994. Presynaptic inhibition by GABA 76 is mediated via two distinct GABA re- 134. Ullrich S, Wollheim CB. 1988. GTP- ceptors with novel pharmacology. J. dependent inhibition of insulin secre- Neurosci. 14:1079–90 tion by epinephrine in permeabilized 147. Enz R, Brandstatter JH, Hartveit E, RINm5F cells. J. Biol. Chem. 263:8615– Wassle H, Bormann J. 1995. Expression 20 of GABA receptor ρ1 and ρ2 subunits 135. Krasnoperov VG, Bittner MA, Beavis in the retina and brain of the rat. Eur. J. R, Kuang Y,Salinkow KV,et al. 1997. α- Neurosci. 7:1495–501 Latrotoxin stimulates exocytosis by the 148. Minchin MC, Ennis C, Lattimer N, interaction with a neuronal G-protein White JF, White AC, et al. 1992. The coupled receptor. Neuron 18:925– GABA-A like autoreceptor is a pharma- 37 cologically novel GABA receptor. Adv. 136. Linial M, Ilouz N, Parnas H. 1997. Volt- Biochem. Psychopharmacol. 47:199– age dependent interactions between the 203 muscarinic Ach receptor and proteins 149. Zhang SJ, Jackson MB. 1995. Properties

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

226 MILLER

of the GABA-A receptor of the rat pos- rat interpeduncular nucleus. J. Neurosci. terior pituitary nerve terminals. J. Neu- 13:2680–88 rophysiol. 73:1135–44 162. McMahon LL, Yoon K-W, Chiapinelli 150. Staley KJ, Soldo BL, Proctor WR. 1995. VA. 1994. Nicotinic receptor activation Ionic mechanisms of neuronal excitation facilitates GABA-ergic neurotransmis- by inhibitory GABA-A receptors. Sci- sion in the avian lateral spiriform nu- ence 269:977–81 cleus. Neuroscience 59:689–98 151. Role LW, Berg DK. 1996. Nicotinic re- 163. Coggan JS, Paysan J, Conroy WG, Berg ceptors in the development and modula- DK. 1997. Direct recording of nicotinic tion of CNS synapses. Neuron 16:1077– responses in presynaptic nerve termi- 85 nals. J. Neurosci. 17:5798–806 152. Patrick J, Seguela P, Verrinos S, Amador 164. Alkondon M, Rocha ES, Maelicke A, M, Luetje C, et al. 1993. Functional Albuquerque E. 1996. Diversity of nico- diversity of neuronal nicotinic acetyl- tinic acetylcholine receptors in rat brain. choline receptors. Prog. Brain. Res. 98: V. α -Bungarotoxin sensitive nicotinic 113–20 receptors in olfactory bulb neurons 153. Rangwala F, Drisdel RC, Rakhilin S, and presynaptic modulation of gluta- Ko E, Atlun P, et al. 1997. Neuronal mate release. J. Pharmacol. Exp. Ther. a-bungarotoxin receptors differ struc- 278:1460–71 turally from other nicotinic acetylcho- 165. McGehee DS, Heath MJS, Gelber S, line receptors. J. Neurosci. 17:8201– Devay P, Role LW. 1995. Nicotine en- 13 hancement of fast excitatory synaptic 154. Verrino S, Amador M, Luitje CW, transmission in CNS by presynaptic re- Patrick J, Dani JA. 1992. Calcium modu- ceptors. Science 269:1692–96 lation and high calcium permeability of 166. Gray R, Rajan AS, Radcliffe KA, Yake- neuronal nicotinic acetylcholine recep- hiro M, Dani JA. 1996. Hippocam- tors. Neuron 8:127–34 pal synaptic transmission enhanced by 155. Verrino S, Rogers M, Radcliffe KA, low concentrations of nicotine. Nature Dani JA. 1994. Quantitative measure- 383:713–16 ment of calcium ﬂux through muscle and 167. Lena C, Changeux J-P. 1997. Role of Ca neuronal nicotinic acetylcholine recep- ions in nicotinic facilitation of GABA tors. J. Neurosci. 14:5514–24 release in mouse thalamus. J. Neurosci. 156. Seguela P, Wadiche J, Dineley-Miller 17:576–85 K, Dani JA, Patrick JW. 1993. Molec- 168. McMahon LL, Yoon, KW, Chiapinelli ular cloning, functional properties, and VA. 1994. Electrophysiological evi- distribution of rat brain α7: a nicotinic dence for presynaptic nicotinic receptors cation channel highly permeable to cal- in the avian ventral lateral geniculate nu- cium. J. Neurosci. 13:596–604 cleus. J. Neurophysiol. 71:826–29 157. Dani JA, Mayer ML. 1995. Struc- 169. Gu JG, MacDermott AB. 1997. Ac- ture and function of glutamate and tivation of ATP 2X receptors elicits nicotinic acetylcholine receptors. Curr. glutamate release from sensory neuron Opin. Neurobiol. 5:310–17 synapses. Nature 289:749–53 158. Zhang ZW, Vijayaraghavan S, Berg 170. Seeburg PH. 1993. The molecular bi- DK. 1994. Neuronal acetylcholine re- ology of mammalian glutamate recep- ceptors that bind α-bungarotoxin with tor channels. Trends Neurosci. 14:297– high afﬁnity function as ligand gated ion 303 channels. Neuron 12:167–77 171. Hollmann M, Heineman S. 1994. Cloned 159. Grady S, Marks MJ, Wonnacott S, glutamate receptors. Annu. Rev. Neu- Collins AC. 1992. Characterization rosci. 17:31–108 of nicotinic receptor mediated [3H]- 172. Lerma J, Morales M, Vincente MA, Her- dopamine release from synaptosomes reras O. 1997. Glutamate receptors of the prepared from mouse striatum. J. Neu- kainate type and synaptic transmission. rochem. 59:848–56 Trends Neurosci. 20:9–12 160. Kulak JM, Nguyen TA, Olivera BM, 173. Collingridge GL, Lester RAJ. 1989. Ex- McIntosh JM. 1997. α-conotoxin MII citatory amino acid receptors in the ver- blocks nicotine stimulated dopamine re- tebrate central nervous system. Pharma- lease in rat striatal synaptosomes. J. Neu- col. Rev. 40:145–95 rosci. 17:5262–70 174. Vignes M, Collingridge GL. 1997. The 161. Lena C, Changeux JP, Miller C. 1993. synaptic activation of kainate receptors. Evidence for “preterminal” nicotine re- Nature 388:179–82 ceptors on GABAergic axons in the 175. Castillo PE, Malenka RC, Nicoll RA.

P1: ARK/vks P2: ARK/plb QC: ARK February 17, 1998 9:9 Annual Reviews AR053-09

PRESYNAPTIC RECEPTORS 227

1997. Kainate receptors mediate a slow citatory amino acids in rat striatal synap- postsynaptic current in hippocampal tosomes. Neurochem. Int. 25:145–54 CA3 neurons. Nature 388:182–85 180. Patel DR, Croucher MJ. 1997. Evidence 176. Barnes JM, Dev KK, Henley JM. 1994. for a role of presynaptic AMPA recep- Cyclothiazide unmasks AMPA-evoked tors in the control of neuronal glutamate stimulation of [3H]-glutamate release release in the rat forebrain. Eur. J. Phar- from rat hippocampal synaptosomes. Br. macol. 332:143–51 J. Pharmacol. 113:339–41 181. Chittajallu R, Vignes M, Dev KK, 177. Montague PR, Gancayo CD, Winn MJ, Barnes JM, Collingridge GL, et al. Marchase RB, Friedlander MJ. 1994. 1996. Regulation of glutamate release Role of NO production in NMDA recep- by presynaptic kainate receptors in the tor mediated neurotransmitter release in hippocampus. Nature 379:78–81 cerebral cortex. Science 263:973–77 182. Clarke VRJ, Ballyk BA, Hoo KH, Man- 178. Pittaluga A, Thellang S, Maura G, Rai- delzys A, Pellizzari A, et al. 1997. A teri M. 1994. Characterization of two GluR5 kainate receptor that regulates in- central AMPA preferring receptors hav- hibitory synaptic transmission in the hip- ing distinct location, function and phar- pocampus. Nature 389:599–606 macology. Arch. Pharmacol. 349:555– 183. Rodriguez-Moreno A, Herrars O, Lerna 58 J. 1997. Kainate receptors presynapti- 179. Cherany A, Desce JM, Godeheu G, cally downregulate GABAergic inhibi- Glowinski J. 1994. Presynaptic control tion in the rat hippocampus. Neuron of dopamine synthesis and release by ex- 19:893–901