- Email: [email protected]
Calmodulin, synchronous and asynchronous release of neurotransmitter
Camp. Biochem. Physiol. Vol. 82A, Printed in Great Britain
No. 1, pp. 7-11, 1985
6
0300-9629/85$3.00+ 0.00 1985 Pergamon Press Ltd
MINIREVIEW CALMODULIN, SYNCHRONOUS AND ASYNCHRONOUS RELEASE OF NEUROTRANSMITTER S. J. PUBLICOVER Department of Zoology and Comparative Physiology, University of Birmingham, P.O. Box 363, Birmingham B15 2TT, UK. Telephone: (021) 472-1301 (Received 13 December 1984) Abstraa-Evidence collected from studies on a wide range of secretory cells suggests that calmodulin may play an important role in stimulus-secretion coupling. Work on synaptosomes, central synaptic preparations and chromafin cell preparations indicates that calmodulin probably also acts as the intracellular Ca’+-receptor for secretion in neuronal cells, Ca’ +-binding resulting in activation of protein kinases and phosphorylation of certain secretory vesicle proteins. Studies on the effects of calmodulinbinding drugs at peripheral synapses have given surprising results, particularly the finding that evoked
(synchronous) transmitter reiease is not suppressed by calmodulin inhibition, though asynchronous release can be markedly inhibited. It is suggested that the insensitivity of synchronous release to drug treatment is due to the fact that only vesicle-bound calmoduiin is involved in this form of transmitter secretion. Asynchronous release, however, involves recruitment of cytosolic calmodulin and can therefore be inhibited by calmodulin-binding drugs.
INTRODUCTION
(CaM) is a Ca*+-binding protein which apparently occurs in all eukaryotic cells and which upon binding of- Ca”+ becomes activated and in turn activates a “target” or “effector” molecule (Cheung, 1980). CaM can exist free in the cytoplasm, attaching to and activating its target upon elevation of [Ca’+], , or it may be bound in a Ca*+-independent manner, only becoming active in the presence of elevated levels of Ca2+ (Vincenzi, 1981). Recently this protein (CaM) has been proposed as the possible target for Ca*+ in a range of secretory cells, mainly due to observations on the effects of CaM-interacting drugs (Naccache et al., 1980; Douglas and Nemeth, 1982; Schettini et al., 1983; Fray et al., 1984). It appears that a widespread feature of Ca*+-mediated stimulus-secretion coupling may be activation through CaM and CaM-stimulated enzymes. Concurrently, a number of studies have been carried out on various synaptic preparations with a view to determining whether CaM is the intracellular target for Ca’+ during excitation-secretion coupling at the presynaptic nerve terminal. The data that have been obtained are generally in accord with the hypothesis that CaM is the intracellular Ca2+ target, but certain anomalies exist, which will be discussed below. Calmodulin
SYNAPTOSOMFS AND SYNAPTIC
CENTRAL PREPARATIONS
The majority of studies on CaM and transmitter release have used mammalian synaptosomes or other preparations derived from central synapses. A considerable amount of data has been collected by De Lorenzo and his coworkers (De Lorenzo and Freedman, 1977; De Lorenzo et al., i979; De Lorenzo,
1980, 1981; Burke and De Lorenzo, 1982; De Lorenzo, 1982) and might be summarized as follows: (i) CaM is present in presynaptic cytoplasm and in synaptic vesicle preparations. (ii) CaM regulates Ca*+-stimulated phosphorylation of a number of synaptic vesicle and synaptic membrane proteins, particularly those of molecular weight 52-54,000 and 60-64,000. (iii) Depolarization-induced Ca’ +-influx stimulates phosphorylation of several vesicle proteins in intact synaptosomes, especially those that are known to be phosphorylated in the presence of CaM (see ii). (iv) Interaction of isolated synaptic vesicle and synaptic membrane preparations is regulated by Ca2+-CaM, and is accompanied by neurotransmitter release and protein phosphorylation. (v) K +.-induced and A23187-induced release of transmitters from synaptosomes are inhibited by the CaM-binding drug tr~fluoperazine (TFP) and by the drugs phenytoin and diazepam, inhibitors of CaMstimulated protein kinases. Protein phosphorylation in whole synaptosome and synaptic vesicle preparations is similarly inhibited. (vi) Ca*+ and CaM stimulate activity of a synaptic tubulin kinase, such that part of the Ca2+-stimulated protein phosphoryIation (ii) is of synaptic tubulin, a process that results in the formation of insoluble tubulin fibrils. Corroborative data has been published by a number of other laboratories. For instance, calmodulin binds to synaptic vesicles and stimulates endogenous synaptic vesicle protein kinase (Moskowitz et al., 1983a,b; Callaghan et a!., 1980), depolarization of synaptosomes induces phosphorylation of specific proteins (Krueger et al., 1977), depolarizationdependent protein phosphorylation is inhibited by
S. J. PUBLIC-OVER
8
CaM-inhibitors (Robinson and Dunkley, 1983) and release of transmitters from K ’ -stimulated synaptosomes and from K+-stimulated and ionophorestimulated hypothalamic nerve endings is inhibited by CaM-binding drugs (Leung and Collard, 1983; Drouva ef ul., 1984). Interestjng~y, it would appear that Ca’+-CaM stimulated phosphoryiation of presynaptic membrane proteins also occurs in preparations derived from ganglia of the mollusc Aplysiu (Novak-Hofer and Levitan, 1984). The accumulation of these data have led to the “CaM hypothesis of synaptic transmission”, which states that presynaptic CaM plays an important role in the process of excitationsecretion coupling through the Ca2+-CaM activation of protein kinases, phosphorylation of tuhulin possibly being of primary importance (De Lorenzo, 1980, 1981, 1982). CHR~~~AFFIN
CELLS
The para-neuronal adrenal medullary (chromailin) cell also provides a useful model for the study of cells the secretion in neuronal-type and Ca’+-requirement for catecholamine release from chromaffin cell preparations suggests a marked similarity to normal neuronal transmitter release (Baker and Knight, 1978; Konings and De Potter, 1981; Kesteven and Knight, 1982). Investigations into the possible involvement of CaM in excitation-secretion coupling in these cells have produced data that strongly support the idea that CaM is the intracellular Ca2+-receptor. Calmodulin binds to and stimuIates phosphorylation in isolated chroma~n granule membranes (Burgoyne and Geisow, 1981; Geisow et al., 1982) binding at a number of sites in both Ca2+-dependent and Ca’+-independent fashion (Geisow and Burgoyne, 1983) and studies on secretion of catecholamines from whole cell preparations have shown that release is inhibited by the CaMbinding drug TFP in a manner that suggests that the drug does not suppress Ca’+-currents at the plasma membrane, but acts at a point subsequent to Ca’+ mobilization (Baker and Knight, 1981: Burgoyne er al., 1982; Keningsberg et al., 1982; Clapham and Neher, 1984). It thus appears that Ca*+-CaM stimulated protein phosphorylation may be a part of the excitation-secretion coupling mechanism in these cells in a way that directly parallels the proposed function of CaM in the release of transmitters from synaptosomes (De Lorenzo, 1980, 198 I, 1982). PERIPHERAL
r
SYNAPSES
Since central and peripheral synapses appear to be essentially similar in their functioning, the evidence concerning the role of CaM in transmitter release at central synapses (see above) would suggest that the protein might play a similar role in the release process at peripheral synapses. However, electrophysiological studies on the effects of CaM-binding drugs on transmitter release at neuromuscular quanta1 synapses have produced data that are difficult to reconcile with this theory (Duncan, 1983). The phenothiazine TFP has been shown to inhibit tetanically induced and K+-induced asynchronous release of transmitter (Fig. I; Publicover, 1983), but either
II
Ca ImM Mg 2mM
Ca 1.8mb.4
Ma 0
Fig. I Effect of trifluoperazine on the stimulation of MEPP frequency induced by 10 mM K+-saline at two different extracellular calcium concentrations. MEPP frequency was measured first in saline containing 2.5 mM K- and then after perfusion of the bath with saline containing 10mM K’. Stimulation of release is expressed as the percentage increase in MEPP frequency. Plain bars represent control experiments and shaded bars represent experiments carried out in the presence of trifluoperazine. Each bar represents the mean of four or five separate experiments ( &SE of mean). From Publicover (1983).
slightly stimulates (at low quanta1 content), or has no effect whatsoever on evoked release of transmitter in both frog and chick preparations (Cheng et af., 198 1: Publicover, 1983). Calmidazolium (R24571, which suppresses K+-induced transmitter release from synaptosomes; Leung and Collard, 1983) markedly inhibits spontaneous transmitter release and will also inhibit K+-induced release at high concentrations, though this effect is sometimes masked by secondary stimuIatory effects of the drug. However, similarly to TFP, R24571 has either no effect or a slight stimulatory action on evoked release (Publicover and Sahaf, 1985; Sahaf and Publicover. 1985). Of particular interest is the fact that application of 2-5 x 10m7M R24571 (a concentration at which the drug should be acceptably CaM-specific), causes a marked reduction in spontaneous release whilst evoked release is either stimulated or unaltered (Fig. 2; Sahaf and Publicover, 1985). Since the inhibitory effect of the drug on MEPP frequency is not sensitive to removal of extracellular Ca*+, an intracellular action, possibly via CaM, would seem to be a distinct possibility, and yet no inhibitory effect of the drug on evoked release is observed. Of the drugs that are known to interact with GaM only chlorpromazine has been shown to inhibit evoked transmitter release, and this effect is seen at 5 x 10m6 M (Argov and Yaari, 1979) a concentration at which the drug should not interact with CaM (Landry ef LII.,
Calmodulin and transmitter release
control
ImV
L
1mV
4s
L 20ms
R2457 1
Fig. 2. Effect of R24571 (2 x IW’M) on spontaneous and evoked release of transmitter. The preparation was bathed in saline containing 0.3 mM Ca** and LOmM Mgz+. Upper pane& control; lower panel. after treatment with R24571 for 20 min. Evoked release shows 12 su~~mpo~d EPPs. Note that MEPP frequency is reduced by drug treatment (control = 30.9 + 2.3 MEPPs min-‘; after R24571 = 19.1 & 1.0 MEPPs min-‘; P
1981). It thus seems sensible to accept the suggestion of the authors that this inhibitory effect results from suppression of Ca *+-influx during the action potential (Argov and Yaari, 1979). Furthermore, a number of studies by other authors have revealed no inhibitory effect of chiorpromazine (up to 201;rM) on evoked transmitter release, though a small rise in the quanta1 content of the EPP may occur (Quastel et al., 1972; Boucher and Katz, 19’77;Sahaf and Publicover, unpubi~shed data}. it would therefore appear that those drugs that are known to interact with CaM are able to suppress only asynchronous release of transmitter, whether spontaneous or artificially induced, and that true evoked release of transmitter at the neuromuscular junction is not impaired by application of these compounds. A’tl the data gathered from central synaptic preparations and chromaffin cells (see above) also refer to stimulation of asynchronous release and not stimulus-evoked synchronous transmitter release. The data available at present thus suggest that there may be some fundamental difference between the inductian of synchronous and asynchronous release of transmitter such that only asynchronous release is susceptible to treatment with Cam-bindjng drugs. Since synchronous quanta1 transmitter release is the primary form of release in civo such a conclusion would seem to throw doubt on the idea the CaM acts as the intracellular Ca*+-receptor in nerve terminals (Duncan, 1983). A suggestion for the resolution of this probiem is presented below.
CALMODULIN IN SYNCHRONOUS AND ASYNCHRONOUS RELEASE OF TRANSMITTER It has recently been demonstrated that binding of CaM to cholinergic synaptic vesicles derived from etasmobranch electric organ takes place by at least two distinct paths, one of which is essentiahy Ca’+-ind~~ndent and one of which is Ca2+-dependent and is inhibited by trifluo~razine (Hooper and Kelly, 1984a,b). Binding of CaM to chroma~n granule membranes shows similar Ca2*-dependent and -independent fractions (Geisow and Burgoyne, 1983). Since CaM-binding drugs attach to the free Ca2+-CaM complex and thus do not affect CaM that is already bound to its target protein in a Ca*‘-independent manner (such as in skeletal muscle phosphorylase kinase; Vincenzi, 1981) only the Ca2+-dependent fraction of this CaM-binding will show susceptibility to inhibition by these compounds. Furthermore, since binding of Ca’+-CaM to its receptors is liable to be a relatively slow process (Vincenzi et a&, 1980; Wang and Sharma, 1980) such that CaM-activated processes will be damped (Vincenzi, 1981) it seems likely that the rapid kinetics of st~m~Ius-s~retion coupling at serve terminals (minimum latency < = 1.6 msec; Daytner and Gage, 1980) could only be achieved through CaM if the protein were already bound to its target. Were this the case, release could be rapidly effected by binding of Ca2+ to the CaM-effector complex (Vincenzi, 1981). The result of this would be that the rapid
S. .I. P~BLIC~VER
10
release Fig. 3. Diagramatic representation of suggested actions of Ca” and CaM (0) in induction of synchronous and asynchronous transmitter release. Transmitter release is a function of the intracellular Ca” concentration [Ca2+],. Upon stimulation the level is elevated by influx of Ca2+ from outside the cell. Pathway 1 indicates direct binding of Ca” to CaM on the surface of synaptic vesicles (and possibly at other sties) allowing a rapid response to the elevation of [Ca? ‘1, occurring upon stimulation. Pathway 2 indicates recruitment of cytosolic CaM, which attaches to the vesicle membrane (and possibly other sites) upon binding of Ca”. This pathway 15 probahly slow compared to pathway i and is susceptible to the attachment of CaMbinding drugs (CBDs) to the Ca”-CdM complex. Evoked (synchronous) release of transmitter is rapid. involves only pathway I and is thus insensitive to CaM-binding drugs. Asynchronous release, whether spontaneous or stimulated, invjolves both pathways I and 2 and can thus be partially blocked by treatment with CaM-binding drugs.
evoked release system responsible for synchronous transmitter release (recorded as the EPP in neuromuscular preparations) would be insensitive to treatment with CaM-binding drugs (Fig. 3, pathway 1). However, asynchronous release (both spontaneous and stimulated) might well include a significant fraction of transmitter secretion induced by Ca’+-dependent binding of CaM to synaptic vesicles (Fig. 3, pathway 2). This form of release would thus show a sensitivity to drug treatment though release could never be totally inhibited, a conclusion that is in accord with the fact that even the highest concentrations of ham-binding drugs can produce only 25%50:;; inhibition of asynchronous release (De Lorenzo, 1982; Leung and Collard, 1983; Publicover, 1983; Publicover and Sahaf, 1985). HYPOTHESIS It is therefore suggested that CaM does act as the intracellular target for Ca:+ in excitation-secretion coupling at presynaptic nerve terminals, but that rapid synchronous release involves only CaM that is already bound to vesicles in a Cal+-independent manner, making this form of transmitter secretion insensitive to blockade by CaM-binding drugs (Fig. 3). Asynchronous release involves both Ca*‘-dependent and Ca’+-independent binding of CaM to receptors on the synaptic vesicles, and can thus be partially blocked by treatment with these drugs (Fig. 3). In this way it is possible to reconcile the apparently contradictory data obtained from
studies on CaM and transmitter release mechanisms and to conceive of a genera1 mechanism of stimulus-secretion coupling. probably involving phosphorylation of secretory vesicle proteins via the activation of CaM-stimulated protein kinases. Further studies to determine whether synchronous transmitter release is insensitive to the fzdll range of CaM-binding drugs are clearly now required. It is interesting to note that a dramatic reduction of EPP quanta1 content at the frog neuromuscular junction can be achieved by treatment with phenytoin, an inhibitor of CaM-stimulated kinases (Yaari et ul., 1979) though this effect may well reflect inhibition of Ca?+-fluxes at the presynaptic membrane (Pincus and Weinfeld, 1984).
REFERENCES
Argov Z. and Yaari Y. (1979) The action of chlorpromazine at an isolated cholinergic synapse. Brrtipr Res. 164, Z-236. Baker P. F. and Knight D. E. (1978) Calcium-dependent exocytosis in bovine adrenal medullary cells with leaky plasma membranes. Nuture, Lond. 276, 620-622. Baker P. F. and Knight D. E. (1981) Calcium control of exocytosis and endocytosis in bovine adrenal medullary eelis. Phil. Truns. R. .Soc. tomi. B 296, 83-103. Boucher S. D. and Katz N. L. (1977) Effects of severai “membrane stabilizing” agents on frog neuromuscular junction. Eur. /. Pharmac. 42, 1139-l 145. Burgoyne R. D. and Geisow M. .I. (1981) Specific binding of “51-calmodulin to and protein phosphorylation in adrenal chromaffin granule membranes. FEB.5 Levi. 131, 127-131. Burgoyne R. D., Geisow M. J. and Barron .I. (1982) Dissection of stages in exocytosis in the adrenal chromaffin cell with use of trifluoperazine. Proc. R. SW. Land. B 216, 1I I-I 15. Burke B. E. and De Lorenzo R. J. (1982) Ca’+ and calmodulin-dependent phosphorylation of endogenous synaptic vesicle tubulin by a vesicle-bound calmodulin kinase system. f. ~~ur~che~~. 38, 120%1218. Callaghan J. B., Dunn L. A. and Lovenberg W. (1980) Calcium-regulated phosphorylation in synaptosomal cytosol: Dependence on calmodulin. Proc. Nuti Acad. Sci. U.S.A. 77, 5812-5186. Cheng K.-C., Lambert J. J., Henderson E. G., Smilowitz H. and Epstein P. M. (1981) Postsynaptic inhibition of neuromuscular transmission by trifluoperazine. J. Pharmac. exp. Tlzer. 217, 44-50. Cheung W. Y. (1980) Calmodulin plays a pivotal role in cellular regulation. Science 207, 19-27. Clapham D. E. and Neher E. (1984) Trifluoperazine reduces inward ionic currents and secretion by separate mechanisms in bovine chromaffin cells. J. Plr_r.yiol. Lond. 353, 541-564. Daytner N. B. and Gage P. W. (lY80) Phasic secretion of a~etyl~holine at a ~mmalian neuromuscular junction. J. Phyriol. Land. 303, 299-3 14. De Lorenzo R. J. (1980) Role of calmodulin in neurotransmitter release and synaptic function. Ann. N. Y. Acad. Sri. 356, 92-109. De Lorenzo R. J. (1981) The calmodulin hypothesis of neurotransmission. Cell Colciu,n 2, 3655385. De Lorenzo R, J. (1982) Calmodulin in neurotransmitter release and synaptic function. Fkdn Proc. 41, 226552272. De Lorenzo R. J. and Freedman S. D. (1977) Caiciumdependent phosphorylation of synaptic vesicle proteins and its possible role in mediating neurotransmitter release and vesicle function. Biochem. hiophys. Rrs. Comrnun. 77, 1036-1043.
Calmodulin and transmitter release De Lorenzo R. J., Freedman S, D., Yohe W. B. and Maurer S. C. (1979) Stimulation of Ca*+-dependent neurotransmitter release and presynaptic nerve terminal protein phosphorylation by calmodulin and a calmodulinlike protein isolated from synaptic vesicles. Proc. Null Acad. Sci. U.S.A. 76, 1838-1842. Douglas W. W. and Nemeth E. F. (1982) On the calcium receptor activating exocytosis: inhibitory effects of calmodulin interacting drugs on rat mast cells. J. Physiol. Lond. 323, 229.--244.
Drouva S. V., Epelbaum J., Laplante E. and Kordon C. (1984) Calmodulin involvement in the Ca2+dependent release of LHRH and SRIF in &ro. Neuroendocrinology 38, !89-192.
Duncan C. J. (1983) Role of calcium in triggering the release of transmitters at the neuromuscular junction. CeN Catcium 4, 171-193. Fray J. C. S., Lush D. J. and Valentine A. N. D. (1983) Possible role of calmodulin in renin secretion from isolated rat kidneys and renal cells: Studies with trifluoperazine. J. P&s&l. Land. 343, 447-454. Geisow M. J. and Buraovne R. D. (1983) R~ru~tment of cytosolic proteins to-a-secretory &anuie membrane depends on Cal+-calmodulin. Nafure, Land. 301, 432-435. Geisow M. J., Burgoyne R. D. and Harris A. (1982) Interaction of calmodulin with chromaffin granule membranes. FEDS Le;t. 143, 69-72. Hooper J. E. and Kelly R. B. (1984a) Calcium-dependent calmodu~in binding to cholinergic synpatic vesicles. J. biol. Chem. 259, 141-147. Hooper J. E. and Kelly R. B. (1984b) Calmodulin is tightly associated with synaptic vesicles independent of calcium. J. biol. Chem. 259, 148-153. Keningsberg R. L., Cote A. and Trifaro J. M. (1982) Trifluoperazine, a calmoduiin inhibitor, blocks secretion in cultured chromaffin cells at a step distal from calcium entry. Neuroscience 7, 2277-2286. Kesteven N. T. and Knight D. E. (1982) Transient changes of intracellular free Ca’+ associated with catecholamine secretion in isolated bovine adrenal medullary cells. J. Physiot. Land. 328, 57-58P.
Konings F. and De Potter W. (1984) In oitro interaction between bovine adrenal medullary cell membranes and chromaffin granules: specific control by Ca’+. NaunynSchm~edeberg’.~ Archs Fharm~s. 317, 97-99.
Krueger B., Forn J. and Greengard P. (1977) Depolarization-induced phosphorylation of specific proteins, mediated by calcium influx, in rat brain synaptosomes. J. biot. Chem. 252, 2164-2773. Landry Y., Amellal M. and Ruckstuhl M. (1981) Can calmoduiin inhibitors be used to probe calmodulin effects? &o&em. Phurmoe. 39, 203 I-2032. Leung M. T. K. and Collard K. J. (1983) The effect of trifluoperazine and R24571 on the K+-evoked release of 5-hydroxytryptamine from superfused synaptosomes. Neuropharmacology
9, 1095-1099.
11
Moskowitz N., Glassman A., Ores C., Schook W. and Puszkin S. (1983a) Phosphorylation of brain synaptic and coated vesicles by endogenous Ca’+/calmodu!in-and CAMP-devendent orotein kinases. J. Neurochem. 40. 711-718. * a Moskowitz N., Schook W., Beckenstein K. and Puszkin S. (1983b) Preliminary characterization of svnaotic vesicle/ calmodulin interaction. Naccache P. H., Molski T. F. P., Alobaidi T., Becker E. L., Showe!! H. J. and Sha’aff R. I. (1980) Calmodulin inhibitors block neutrophil degranulation at a step distal from the mobilization of calcium. Biochem. biophys. Res. I
Commun. 97, 62-68.
Novak-Hofer I. and Levitan I. B. (1983) Ca~+/calmodulin-regulated protein phosphorylation in the Aptysia nervous system. J. Neurosci. 3, 473-48 I, Pincus J. H. and Weinfeld H. M. (1984) Acetylcholine release from synaptosomes and phenytoin action. Brain Res. 296, 3 13.--317. Publicover S. J. (1983) Presynaptic action of trifluoperazine at the frog neuromuscular junction. Naunvn Schm~edeberg’s Archs Pharmac. 322, 83-88.
Publicover S. J. and Sahaf Z. Y. (1985) Calmodulininhibitors and transmitter release at the frog neuromuscular junction. In Calcium, Neuronat Function and Transmitler Release (Edited by Katz B. and Rahamimoff R.). Martinus Nyhoff, Boston (in press). Quastel D. M. J., Hackett J. T. and Okamoto K. (1972) Can. J. Phys. Pharmac. 50, 279-284.
Robinson P. J. and Dunkley P. R. (1983) The effect of fluohenazine and depolarisation on synaptosomal protein phbsphorylation. Neurosci. Letl. Suipt.-11, S7l. _ Sahaf Z. Y. and Publicover S. J. (19851 Dual effect of calmidazolium (R24571) on transmitter release at the frog neuromuscular junction. (In preparation.) Vincenzi F. F. (I 98 I) Calmoduiin pharmacology. Cell Cat&m 2, 387-409. Vincenzi F. F., Hinds T. R. and Raess B. U. (1980) Calmodulin and the plasma membrane calcium pump. Ann. N.Y. Acad. Sri. 356, 232-244.
Wang J. H. and Sharma R. K. (1980) On the mechanism of activation of cyclic nucleotide phosphodiesterase by calmodulin. Ann. k.Y. Acad. Sci. j56, 190-20. Yaari Y., Pincus J. H. and Argov Z. (1981) Phenytoin and transmitter release at the neuromuscular junction of the frog. Brain. Res. 217, 119-129. Note added in proof
Since the submission of this paper Keningsberg and Trifiro have demonstrated that introduction of anti-CaM antibodies into cultured bovine chromaffin cells causes a marked suppression of both ACh-induced and K*induced noradrenaline secretion: (Microinjection of calmodulin antibodies into cultured chromaffin cells blocks catecholamine release in response to stimulation. Neuroscience 14, 335-347, 1985).
No. 1, pp. 7-11, 1985
6
0300-9629/85$3.00+ 0.00 1985 Pergamon Press Ltd
MINIREVIEW CALMODULIN, SYNCHRONOUS AND ASYNCHRONOUS RELEASE OF NEUROTRANSMITTER S. J. PUBLICOVER Department of Zoology and Comparative Physiology, University of Birmingham, P.O. Box 363, Birmingham B15 2TT, UK. Telephone: (021) 472-1301 (Received 13 December 1984) Abstraa-Evidence collected from studies on a wide range of secretory cells suggests that calmodulin may play an important role in stimulus-secretion coupling. Work on synaptosomes, central synaptic preparations and chromafin cell preparations indicates that calmodulin probably also acts as the intracellular Ca’+-receptor for secretion in neuronal cells, Ca’ +-binding resulting in activation of protein kinases and phosphorylation of certain secretory vesicle proteins. Studies on the effects of calmodulinbinding drugs at peripheral synapses have given surprising results, particularly the finding that evoked
(synchronous) transmitter reiease is not suppressed by calmodulin inhibition, though asynchronous release can be markedly inhibited. It is suggested that the insensitivity of synchronous release to drug treatment is due to the fact that only vesicle-bound calmoduiin is involved in this form of transmitter secretion. Asynchronous release, however, involves recruitment of cytosolic calmodulin and can therefore be inhibited by calmodulin-binding drugs.
INTRODUCTION
(CaM) is a Ca*+-binding protein which apparently occurs in all eukaryotic cells and which upon binding of- Ca”+ becomes activated and in turn activates a “target” or “effector” molecule (Cheung, 1980). CaM can exist free in the cytoplasm, attaching to and activating its target upon elevation of [Ca’+], , or it may be bound in a Ca*+-independent manner, only becoming active in the presence of elevated levels of Ca2+ (Vincenzi, 1981). Recently this protein (CaM) has been proposed as the possible target for Ca*+ in a range of secretory cells, mainly due to observations on the effects of CaM-interacting drugs (Naccache et al., 1980; Douglas and Nemeth, 1982; Schettini et al., 1983; Fray et al., 1984). It appears that a widespread feature of Ca*+-mediated stimulus-secretion coupling may be activation through CaM and CaM-stimulated enzymes. Concurrently, a number of studies have been carried out on various synaptic preparations with a view to determining whether CaM is the intracellular target for Ca’+ during excitation-secretion coupling at the presynaptic nerve terminal. The data that have been obtained are generally in accord with the hypothesis that CaM is the intracellular Ca2+ target, but certain anomalies exist, which will be discussed below. Calmodulin
SYNAPTOSOMFS AND SYNAPTIC
CENTRAL PREPARATIONS
The majority of studies on CaM and transmitter release have used mammalian synaptosomes or other preparations derived from central synapses. A considerable amount of data has been collected by De Lorenzo and his coworkers (De Lorenzo and Freedman, 1977; De Lorenzo et al., i979; De Lorenzo,
1980, 1981; Burke and De Lorenzo, 1982; De Lorenzo, 1982) and might be summarized as follows: (i) CaM is present in presynaptic cytoplasm and in synaptic vesicle preparations. (ii) CaM regulates Ca*+-stimulated phosphorylation of a number of synaptic vesicle and synaptic membrane proteins, particularly those of molecular weight 52-54,000 and 60-64,000. (iii) Depolarization-induced Ca’ +-influx stimulates phosphorylation of several vesicle proteins in intact synaptosomes, especially those that are known to be phosphorylated in the presence of CaM (see ii). (iv) Interaction of isolated synaptic vesicle and synaptic membrane preparations is regulated by Ca2+-CaM, and is accompanied by neurotransmitter release and protein phosphorylation. (v) K +.-induced and A23187-induced release of transmitters from synaptosomes are inhibited by the CaM-binding drug tr~fluoperazine (TFP) and by the drugs phenytoin and diazepam, inhibitors of CaMstimulated protein kinases. Protein phosphorylation in whole synaptosome and synaptic vesicle preparations is similarly inhibited. (vi) Ca*+ and CaM stimulate activity of a synaptic tubulin kinase, such that part of the Ca2+-stimulated protein phosphoryIation (ii) is of synaptic tubulin, a process that results in the formation of insoluble tubulin fibrils. Corroborative data has been published by a number of other laboratories. For instance, calmodulin binds to synaptic vesicles and stimulates endogenous synaptic vesicle protein kinase (Moskowitz et al., 1983a,b; Callaghan et a!., 1980), depolarization of synaptosomes induces phosphorylation of specific proteins (Krueger et al., 1977), depolarizationdependent protein phosphorylation is inhibited by
S. J. PUBLIC-OVER
8
CaM-inhibitors (Robinson and Dunkley, 1983) and release of transmitters from K ’ -stimulated synaptosomes and from K+-stimulated and ionophorestimulated hypothalamic nerve endings is inhibited by CaM-binding drugs (Leung and Collard, 1983; Drouva ef ul., 1984). Interestjng~y, it would appear that Ca’+-CaM stimulated phosphoryiation of presynaptic membrane proteins also occurs in preparations derived from ganglia of the mollusc Aplysiu (Novak-Hofer and Levitan, 1984). The accumulation of these data have led to the “CaM hypothesis of synaptic transmission”, which states that presynaptic CaM plays an important role in the process of excitationsecretion coupling through the Ca2+-CaM activation of protein kinases, phosphorylation of tuhulin possibly being of primary importance (De Lorenzo, 1980, 1981, 1982). CHR~~~AFFIN
CELLS
The para-neuronal adrenal medullary (chromailin) cell also provides a useful model for the study of cells the secretion in neuronal-type and Ca’+-requirement for catecholamine release from chromaffin cell preparations suggests a marked similarity to normal neuronal transmitter release (Baker and Knight, 1978; Konings and De Potter, 1981; Kesteven and Knight, 1982). Investigations into the possible involvement of CaM in excitation-secretion coupling in these cells have produced data that strongly support the idea that CaM is the intracellular Ca2+-receptor. Calmodulin binds to and stimuIates phosphorylation in isolated chroma~n granule membranes (Burgoyne and Geisow, 1981; Geisow et al., 1982) binding at a number of sites in both Ca2+-dependent and Ca’+-independent fashion (Geisow and Burgoyne, 1983) and studies on secretion of catecholamines from whole cell preparations have shown that release is inhibited by the CaMbinding drug TFP in a manner that suggests that the drug does not suppress Ca’+-currents at the plasma membrane, but acts at a point subsequent to Ca’+ mobilization (Baker and Knight, 1981: Burgoyne er al., 1982; Keningsberg et al., 1982; Clapham and Neher, 1984). It thus appears that Ca*+-CaM stimulated protein phosphorylation may be a part of the excitation-secretion coupling mechanism in these cells in a way that directly parallels the proposed function of CaM in the release of transmitters from synaptosomes (De Lorenzo, 1980, 198 I, 1982). PERIPHERAL
r
SYNAPSES
Since central and peripheral synapses appear to be essentially similar in their functioning, the evidence concerning the role of CaM in transmitter release at central synapses (see above) would suggest that the protein might play a similar role in the release process at peripheral synapses. However, electrophysiological studies on the effects of CaM-binding drugs on transmitter release at neuromuscular quanta1 synapses have produced data that are difficult to reconcile with this theory (Duncan, 1983). The phenothiazine TFP has been shown to inhibit tetanically induced and K+-induced asynchronous release of transmitter (Fig. I; Publicover, 1983), but either
II
Ca ImM Mg 2mM
Ca 1.8mb.4
Ma 0
Fig. I Effect of trifluoperazine on the stimulation of MEPP frequency induced by 10 mM K+-saline at two different extracellular calcium concentrations. MEPP frequency was measured first in saline containing 2.5 mM K- and then after perfusion of the bath with saline containing 10mM K’. Stimulation of release is expressed as the percentage increase in MEPP frequency. Plain bars represent control experiments and shaded bars represent experiments carried out in the presence of trifluoperazine. Each bar represents the mean of four or five separate experiments ( &SE of mean). From Publicover (1983).
slightly stimulates (at low quanta1 content), or has no effect whatsoever on evoked release of transmitter in both frog and chick preparations (Cheng et af., 198 1: Publicover, 1983). Calmidazolium (R24571, which suppresses K+-induced transmitter release from synaptosomes; Leung and Collard, 1983) markedly inhibits spontaneous transmitter release and will also inhibit K+-induced release at high concentrations, though this effect is sometimes masked by secondary stimuIatory effects of the drug. However, similarly to TFP, R24571 has either no effect or a slight stimulatory action on evoked release (Publicover and Sahaf, 1985; Sahaf and Publicover. 1985). Of particular interest is the fact that application of 2-5 x 10m7M R24571 (a concentration at which the drug should be acceptably CaM-specific), causes a marked reduction in spontaneous release whilst evoked release is either stimulated or unaltered (Fig. 2; Sahaf and Publicover, 1985). Since the inhibitory effect of the drug on MEPP frequency is not sensitive to removal of extracellular Ca*+, an intracellular action, possibly via CaM, would seem to be a distinct possibility, and yet no inhibitory effect of the drug on evoked release is observed. Of the drugs that are known to interact with GaM only chlorpromazine has been shown to inhibit evoked transmitter release, and this effect is seen at 5 x 10m6 M (Argov and Yaari, 1979) a concentration at which the drug should not interact with CaM (Landry ef LII.,
Calmodulin and transmitter release
control
ImV
L
1mV
4s
L 20ms
R2457 1
Fig. 2. Effect of R24571 (2 x IW’M) on spontaneous and evoked release of transmitter. The preparation was bathed in saline containing 0.3 mM Ca** and LOmM Mgz+. Upper pane& control; lower panel. after treatment with R24571 for 20 min. Evoked release shows 12 su~~mpo~d EPPs. Note that MEPP frequency is reduced by drug treatment (control = 30.9 + 2.3 MEPPs min-‘; after R24571 = 19.1 & 1.0 MEPPs min-‘; P
1981). It thus seems sensible to accept the suggestion of the authors that this inhibitory effect results from suppression of Ca *+-influx during the action potential (Argov and Yaari, 1979). Furthermore, a number of studies by other authors have revealed no inhibitory effect of chiorpromazine (up to 201;rM) on evoked transmitter release, though a small rise in the quanta1 content of the EPP may occur (Quastel et al., 1972; Boucher and Katz, 19’77;Sahaf and Publicover, unpubi~shed data}. it would therefore appear that those drugs that are known to interact with CaM are able to suppress only asynchronous release of transmitter, whether spontaneous or artificially induced, and that true evoked release of transmitter at the neuromuscular junction is not impaired by application of these compounds. A’tl the data gathered from central synaptic preparations and chromaffin cells (see above) also refer to stimulation of asynchronous release and not stimulus-evoked synchronous transmitter release. The data available at present thus suggest that there may be some fundamental difference between the inductian of synchronous and asynchronous release of transmitter such that only asynchronous release is susceptible to treatment with Cam-bindjng drugs. Since synchronous quanta1 transmitter release is the primary form of release in civo such a conclusion would seem to throw doubt on the idea the CaM acts as the intracellular Ca*+-receptor in nerve terminals (Duncan, 1983). A suggestion for the resolution of this probiem is presented below.
CALMODULIN IN SYNCHRONOUS AND ASYNCHRONOUS RELEASE OF TRANSMITTER It has recently been demonstrated that binding of CaM to cholinergic synaptic vesicles derived from etasmobranch electric organ takes place by at least two distinct paths, one of which is essentiahy Ca’+-ind~~ndent and one of which is Ca2+-dependent and is inhibited by trifluo~razine (Hooper and Kelly, 1984a,b). Binding of CaM to chroma~n granule membranes shows similar Ca2*-dependent and -independent fractions (Geisow and Burgoyne, 1983). Since CaM-binding drugs attach to the free Ca2+-CaM complex and thus do not affect CaM that is already bound to its target protein in a Ca*‘-independent manner (such as in skeletal muscle phosphorylase kinase; Vincenzi, 1981) only the Ca2+-dependent fraction of this CaM-binding will show susceptibility to inhibition by these compounds. Furthermore, since binding of Ca’+-CaM to its receptors is liable to be a relatively slow process (Vincenzi et a&, 1980; Wang and Sharma, 1980) such that CaM-activated processes will be damped (Vincenzi, 1981) it seems likely that the rapid kinetics of st~m~Ius-s~retion coupling at serve terminals (minimum latency < = 1.6 msec; Daytner and Gage, 1980) could only be achieved through CaM if the protein were already bound to its target. Were this the case, release could be rapidly effected by binding of Ca2+ to the CaM-effector complex (Vincenzi, 1981). The result of this would be that the rapid
S. .I. P~BLIC~VER
10
release Fig. 3. Diagramatic representation of suggested actions of Ca” and CaM (0) in induction of synchronous and asynchronous transmitter release. Transmitter release is a function of the intracellular Ca” concentration [Ca2+],. Upon stimulation the level is elevated by influx of Ca2+ from outside the cell. Pathway 1 indicates direct binding of Ca” to CaM on the surface of synaptic vesicles (and possibly at other sties) allowing a rapid response to the elevation of [Ca? ‘1, occurring upon stimulation. Pathway 2 indicates recruitment of cytosolic CaM, which attaches to the vesicle membrane (and possibly other sites) upon binding of Ca”. This pathway 15 probahly slow compared to pathway i and is susceptible to the attachment of CaMbinding drugs (CBDs) to the Ca”-CdM complex. Evoked (synchronous) release of transmitter is rapid. involves only pathway I and is thus insensitive to CaM-binding drugs. Asynchronous release, whether spontaneous or stimulated, invjolves both pathways I and 2 and can thus be partially blocked by treatment with CaM-binding drugs.
evoked release system responsible for synchronous transmitter release (recorded as the EPP in neuromuscular preparations) would be insensitive to treatment with CaM-binding drugs (Fig. 3, pathway 1). However, asynchronous release (both spontaneous and stimulated) might well include a significant fraction of transmitter secretion induced by Ca’+-dependent binding of CaM to synaptic vesicles (Fig. 3, pathway 2). This form of release would thus show a sensitivity to drug treatment though release could never be totally inhibited, a conclusion that is in accord with the fact that even the highest concentrations of ham-binding drugs can produce only 25%50:;; inhibition of asynchronous release (De Lorenzo, 1982; Leung and Collard, 1983; Publicover, 1983; Publicover and Sahaf, 1985). HYPOTHESIS It is therefore suggested that CaM does act as the intracellular target for Ca:+ in excitation-secretion coupling at presynaptic nerve terminals, but that rapid synchronous release involves only CaM that is already bound to vesicles in a Cal+-independent manner, making this form of transmitter secretion insensitive to blockade by CaM-binding drugs (Fig. 3). Asynchronous release involves both Ca*‘-dependent and Ca’+-independent binding of CaM to receptors on the synaptic vesicles, and can thus be partially blocked by treatment with these drugs (Fig. 3). In this way it is possible to reconcile the apparently contradictory data obtained from
studies on CaM and transmitter release mechanisms and to conceive of a genera1 mechanism of stimulus-secretion coupling. probably involving phosphorylation of secretory vesicle proteins via the activation of CaM-stimulated protein kinases. Further studies to determine whether synchronous transmitter release is insensitive to the fzdll range of CaM-binding drugs are clearly now required. It is interesting to note that a dramatic reduction of EPP quanta1 content at the frog neuromuscular junction can be achieved by treatment with phenytoin, an inhibitor of CaM-stimulated kinases (Yaari et ul., 1979) though this effect may well reflect inhibition of Ca?+-fluxes at the presynaptic membrane (Pincus and Weinfeld, 1984).
REFERENCES
Argov Z. and Yaari Y. (1979) The action of chlorpromazine at an isolated cholinergic synapse. Brrtipr Res. 164, Z-236. Baker P. F. and Knight D. E. (1978) Calcium-dependent exocytosis in bovine adrenal medullary cells with leaky plasma membranes. Nuture, Lond. 276, 620-622. Baker P. F. and Knight D. E. (1981) Calcium control of exocytosis and endocytosis in bovine adrenal medullary eelis. Phil. Truns. R. .Soc. tomi. B 296, 83-103. Boucher S. D. and Katz N. L. (1977) Effects of severai “membrane stabilizing” agents on frog neuromuscular junction. Eur. /. Pharmac. 42, 1139-l 145. Burgoyne R. D. and Geisow M. .I. (1981) Specific binding of “51-calmodulin to and protein phosphorylation in adrenal chromaffin granule membranes. FEB.5 Levi. 131, 127-131. Burgoyne R. D., Geisow M. J. and Barron .I. (1982) Dissection of stages in exocytosis in the adrenal chromaffin cell with use of trifluoperazine. Proc. R. SW. Land. B 216, 1I I-I 15. Burke B. E. and De Lorenzo R. J. (1982) Ca’+ and calmodulin-dependent phosphorylation of endogenous synaptic vesicle tubulin by a vesicle-bound calmodulin kinase system. f. ~~ur~che~~. 38, 120%1218. Callaghan J. B., Dunn L. A. and Lovenberg W. (1980) Calcium-regulated phosphorylation in synaptosomal cytosol: Dependence on calmodulin. Proc. Nuti Acad. Sci. U.S.A. 77, 5812-5186. Cheng K.-C., Lambert J. J., Henderson E. G., Smilowitz H. and Epstein P. M. (1981) Postsynaptic inhibition of neuromuscular transmission by trifluoperazine. J. Pharmac. exp. Tlzer. 217, 44-50. Cheung W. Y. (1980) Calmodulin plays a pivotal role in cellular regulation. Science 207, 19-27. Clapham D. E. and Neher E. (1984) Trifluoperazine reduces inward ionic currents and secretion by separate mechanisms in bovine chromaffin cells. J. Plr_r.yiol. Lond. 353, 541-564. Daytner N. B. and Gage P. W. (lY80) Phasic secretion of a~etyl~holine at a ~mmalian neuromuscular junction. J. Phyriol. Land. 303, 299-3 14. De Lorenzo R. J. (1980) Role of calmodulin in neurotransmitter release and synaptic function. Ann. N. Y. Acad. Sri. 356, 92-109. De Lorenzo R. J. (1981) The calmodulin hypothesis of neurotransmission. Cell Colciu,n 2, 3655385. De Lorenzo R, J. (1982) Calmodulin in neurotransmitter release and synaptic function. Fkdn Proc. 41, 226552272. De Lorenzo R. J. and Freedman S. D. (1977) Caiciumdependent phosphorylation of synaptic vesicle proteins and its possible role in mediating neurotransmitter release and vesicle function. Biochem. hiophys. Rrs. Comrnun. 77, 1036-1043.
Calmodulin and transmitter release De Lorenzo R. J., Freedman S, D., Yohe W. B. and Maurer S. C. (1979) Stimulation of Ca*+-dependent neurotransmitter release and presynaptic nerve terminal protein phosphorylation by calmodulin and a calmodulinlike protein isolated from synaptic vesicles. Proc. Null Acad. Sci. U.S.A. 76, 1838-1842. Douglas W. W. and Nemeth E. F. (1982) On the calcium receptor activating exocytosis: inhibitory effects of calmodulin interacting drugs on rat mast cells. J. Physiol. Lond. 323, 229.--244.
Drouva S. V., Epelbaum J., Laplante E. and Kordon C. (1984) Calmodulin involvement in the Ca2+dependent release of LHRH and SRIF in &ro. Neuroendocrinology 38, !89-192.
Duncan C. J. (1983) Role of calcium in triggering the release of transmitters at the neuromuscular junction. CeN Catcium 4, 171-193. Fray J. C. S., Lush D. J. and Valentine A. N. D. (1983) Possible role of calmodulin in renin secretion from isolated rat kidneys and renal cells: Studies with trifluoperazine. J. P&s&l. Land. 343, 447-454. Geisow M. J. and Buraovne R. D. (1983) R~ru~tment of cytosolic proteins to-a-secretory &anuie membrane depends on Cal+-calmodulin. Nafure, Land. 301, 432-435. Geisow M. J., Burgoyne R. D. and Harris A. (1982) Interaction of calmodulin with chromaffin granule membranes. FEDS Le;t. 143, 69-72. Hooper J. E. and Kelly R. B. (1984a) Calcium-dependent calmodu~in binding to cholinergic synpatic vesicles. J. biol. Chem. 259, 141-147. Hooper J. E. and Kelly R. B. (1984b) Calmodulin is tightly associated with synaptic vesicles independent of calcium. J. biol. Chem. 259, 148-153. Keningsberg R. L., Cote A. and Trifaro J. M. (1982) Trifluoperazine, a calmoduiin inhibitor, blocks secretion in cultured chromaffin cells at a step distal from calcium entry. Neuroscience 7, 2277-2286. Kesteven N. T. and Knight D. E. (1982) Transient changes of intracellular free Ca’+ associated with catecholamine secretion in isolated bovine adrenal medullary cells. J. Physiot. Land. 328, 57-58P.
Konings F. and De Potter W. (1984) In oitro interaction between bovine adrenal medullary cell membranes and chromaffin granules: specific control by Ca’+. NaunynSchm~edeberg’.~ Archs Fharm~s. 317, 97-99.
Krueger B., Forn J. and Greengard P. (1977) Depolarization-induced phosphorylation of specific proteins, mediated by calcium influx, in rat brain synaptosomes. J. biot. Chem. 252, 2164-2773. Landry Y., Amellal M. and Ruckstuhl M. (1981) Can calmoduiin inhibitors be used to probe calmodulin effects? &o&em. Phurmoe. 39, 203 I-2032. Leung M. T. K. and Collard K. J. (1983) The effect of trifluoperazine and R24571 on the K+-evoked release of 5-hydroxytryptamine from superfused synaptosomes. Neuropharmacology
9, 1095-1099.
11
Moskowitz N., Glassman A., Ores C., Schook W. and Puszkin S. (1983a) Phosphorylation of brain synaptic and coated vesicles by endogenous Ca’+/calmodu!in-and CAMP-devendent orotein kinases. J. Neurochem. 40. 711-718. * a Moskowitz N., Schook W., Beckenstein K. and Puszkin S. (1983b) Preliminary characterization of svnaotic vesicle/ calmodulin interaction. Naccache P. H., Molski T. F. P., Alobaidi T., Becker E. L., Showe!! H. J. and Sha’aff R. I. (1980) Calmodulin inhibitors block neutrophil degranulation at a step distal from the mobilization of calcium. Biochem. biophys. Res. I
Commun. 97, 62-68.
Novak-Hofer I. and Levitan I. B. (1983) Ca~+/calmodulin-regulated protein phosphorylation in the Aptysia nervous system. J. Neurosci. 3, 473-48 I, Pincus J. H. and Weinfeld H. M. (1984) Acetylcholine release from synaptosomes and phenytoin action. Brain Res. 296, 3 13.--317. Publicover S. J. (1983) Presynaptic action of trifluoperazine at the frog neuromuscular junction. Naunvn Schm~edeberg’s Archs Pharmac. 322, 83-88.
Publicover S. J. and Sahaf Z. Y. (1985) Calmodulininhibitors and transmitter release at the frog neuromuscular junction. In Calcium, Neuronat Function and Transmitler Release (Edited by Katz B. and Rahamimoff R.). Martinus Nyhoff, Boston (in press). Quastel D. M. J., Hackett J. T. and Okamoto K. (1972) Can. J. Phys. Pharmac. 50, 279-284.
Robinson P. J. and Dunkley P. R. (1983) The effect of fluohenazine and depolarisation on synaptosomal protein phbsphorylation. Neurosci. Letl. Suipt.-11, S7l. _ Sahaf Z. Y. and Publicover S. J. (19851 Dual effect of calmidazolium (R24571) on transmitter release at the frog neuromuscular junction. (In preparation.) Vincenzi F. F. (I 98 I) Calmoduiin pharmacology. Cell Cat&m 2, 387-409. Vincenzi F. F., Hinds T. R. and Raess B. U. (1980) Calmodulin and the plasma membrane calcium pump. Ann. N.Y. Acad. Sri. 356, 232-244.
Wang J. H. and Sharma R. K. (1980) On the mechanism of activation of cyclic nucleotide phosphodiesterase by calmodulin. Ann. k.Y. Acad. Sci. j56, 190-20. Yaari Y., Pincus J. H. and Argov Z. (1981) Phenytoin and transmitter release at the neuromuscular junction of the frog. Brain. Res. 217, 119-129. Note added in proof
Since the submission of this paper Keningsberg and Trifiro have demonstrated that introduction of anti-CaM antibodies into cultured bovine chromaffin cells causes a marked suppression of both ACh-induced and K*induced noradrenaline secretion: (Microinjection of calmodulin antibodies into cultured chromaffin cells blocks catecholamine release in response to stimulation. Neuroscience 14, 335-347, 1985).