Protein Kinase C and the Regulation of Glutamate Exocytosis from ...

THEJOURNALOF BIOLOGICAL CHEMISTRY

Vol. 268, No. 28, Issue of October 5 , pp. 21060-21065,1993 Printed in U.S.A.

0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Protein Kinase C and the Regulation of GlutamateExocytosis from Cerebrocortical Synaptosomes* (Received for publication, March 16, 1993, and in revised form, May 17, 1993)

Eleanor T.Coffeys, Talvinder S . SihraO, and David G . Nichollsll From the Department of Biochemistry, University of Dundee, Dundee, Scotland, United Kingdom

The roleof protein kinase C (PKC) in the regulation in synaptosomes (3)and not when elevated KC1 is added to of transmitter glutamaterelease from rat cerebral cor- give a “clamped” depolarization of the terminals. This was tical synaptosomesis investigated. Two depolarization interpreted as indicating a locus for PKC at 4-AP-resistant protocols are used: first, elevated KCl, which produces K+ channels regulating the frequency or duration of these aclampeddepolarization,andsecond,4-aminopyri“action potentials’’ (2). dine, whichevokes spontaneous “action potentials” alRecently it has been reported that the effects of phorbol or K+ channels esters on this preparation can be mimicked by the metabolowing any potential modulation of Na+ to influence release. Although the PKC inhibitor Ro tropicglutamate agonist (1S,3R)-l-aminocyclopentane-1,331-8220 prevents both the depolarization-evoked and phorbol dibutyrate (PDBu)-evokedphosphorylation of &carboxylate (ACPD) in the presence of low concentrations the major presynaptic PKC substrate, myristoylated of arachidonic acid to increase the sensitivity of the presynalanine-rich C kinase substrate, it is without effect on apticPKC tothe generated diacylglycerol (4). Thusthis KC1-evoked Ca2+-dependent glutamate release. Ro 31- pathway could provide a positive feedback at cerebrocortical 8220 totally inhibits the Ca2+-dependent 4-aminopyr-glutamatergic synapses under conditions where phospholipase idine-evoked release of glutamate in the presence andAPis activated, asmay occur during synaptic potentiation(5). absence of PDBu and again decreases the phosphoryl- Even in the absence of phorbol ester or added glutamate, ation of myristoylated alanine-richC kinase substrate. KC1 depolarization of synaptosomes increases the phosphorylation state of a number of PKC-dependent proteins RO 31-8220 stronglyinhibitsthe4-aminopyridine(15). To establish the function of this “intrinsic” PKCactivity evokedincreasein[Ca2+]bothinthepresenceand absence of PDBu and antagonizes the PDBu enhance-on glutamate exocytosis, we examine the effects of a selective ment of depolarization. This indicates that PKC iso- PKCinhibitor Ro 31-8220 (capable of inhibiting PKC-a, forms activatable by PDBu and sensitive to Ro 31- PKC-PI, PKC-PII, PKC-7, and PKC-6)’ exploiting the two 8220 play no discernable role in Ca2+-secretion cou- depolarization strategies discussed above to localize possible pling per se in cerebral cortical glutamatergic nerve effects. We find that there is no PKC requirement for Ca2+terminals, but that the kinase plays a major role in exocytosis coupling itself, that MARCKS phosphorylation is regulating the depolarization of the terminal. not required for glutamate exocytosis, but that PKC modulation of ion channel activity revealed by 4-AP-dependent depolarization is more extensive than was previously considered, and is capable of controlling glutamate release over the The protein kinase C isoenzymes PKC-a,’ PKC-PII,and to full range from complete inhibition to maximal release. a lesser extent PKC-7 are present in presynaptic nerve terminals from mammalian central nervous system (1). The EXPERIMENTALPROCEDURES isolated nerve terminal (synaptosome) preparation is the simMaterials-Glutamate dehydrogenase (EC 1.4.1.3) wasfrom Sigma plest inwhich neurotransmitter exocytosis can be evoked and is thus suitable for the elucidation of the presynaptic role(s) and was dialyzed under pressure to remove contaminating glutamate. Fura-2 acetoxymethyl ester was from Calbiochem Novabiochem (Notof the kinase. We have previously reported (2) that enhance- tingham, United Kingdom). DiSCZ(5)was from Molecular Probes, ment of PKC activity by phorbol esters greatly increases the Eugene, OR. 32Pwas from Amersham International, Amersham, UK. Caz+-dependent release of transmitter glutamate, but only Ro31-8220was a gift from Dr. G . Lawton, Roche Products Ltd., when this is evoked by the K+channel inhibitor 4-aminopyr- Welwyn Garden City, Herts, UK. Other reagents were supplied by idine (4-AP), which induces spontaneous “action potentials’’ Sigma.

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 4 Supported by a grant from the British Medical Research Council (to T. S. S.and D. G . N.). 8 Present address: Department of Pharmacology, Royal Free Hospital School of Medicine, University of London, London NW3 2PF. n To whom correspondence should be addressed Dept. of Biochemistry, University of Dundee, Dundee DD1 4HN, Scotland, UK. Tel.: 44-382-307780;Fax: 44-382-200894. The abbreviations used are: PKC, protein kinase C; PDBu, 4pphorbol dibutyrate; 4-AP, 4-aminopyridine; ACPD, (lS,BR)-l-arninocyclopentane-l,3-dicarboxylate; MARCKS, myristoylated alaninerich C kinase substrate; DiSC2(5),3,3’-diethylthiadicarbocyanine iodide; [Ca2+Ic,cytoplasmic free Ca2+concentration.

’

Synaptosomal Preparation-Synaptosomes from the cerebral cortex of 6-week-old male Wistar rats were purified on Percoll gradients (6) with minor modifications (7). Protein was determined by the Bradford (8)method. The final synaptosomal pellet was resuspended in incubation buffer consisting of 140 mM NaCl, 5 mM KCl, 5 mM NaHC03, 1.2 mM NaH2P0,, 1 mM MgC12,lO mM glucose, and 10 mM Hepes (pH 7.4), a t 0.25 mg of protein/ml and preincubated for 5 min at 37 ‘C prior to theaddition of CaC12 (1mM) where indicated. Glutamate Release-Glutamate release was determined by continuous fluorometric assay as described previously (9) using a PerkinElmer LS5B spectrofluorimeter with a stirred, thermostatted cuvette attachment. Synaptosomal pellets were resuspended and preincubated in incubation buffer (see above) at 37 “C in the presence of NADP+ (1 mM), glutamate dehydrogenase (50 units/ml), and CaCL (1 mM) where indicated. Preincubation with Ro 31-8220, PDBu, or

21060

G. Lawton, personal communication.

Protein Kinase

C and Glutamate Release

dimethyl sulfoxide controls was carried out asdescribed in the figure legends. Depolarizing agents (30 mM KC1 or 1 mM 4-aminopyridine) were added where indicated, 2 min after Ro 31-8220 addition and 30 s after 4O-PDBu. Fluorescence was monitored (excitation and emission wavelengths, 340 and 460 nm, respectively), and 2-9 data points were collected. A glutamatestandard (2.5 nmol) was added at the end of each experiment. Data were analyzed using Lotus 1-2-3. [Ca2+]. Determination-Synaptosomes were preincubated in incubation medium containing additionally 5 p M fura-2-acetoxymethyl ester and 0.1 mM Ca2+for 30 min at 37 "C as previously described (10). After fura-2 loading, synaptosomes were centrifuged in an Eppendorf microcentrifuge (10,000 X g-) for 15 s, the pellet resuspended in incubation buffer a t 37 "C and transferred to a thermostatted cuvette in the Perkin-Elmer LS5B spectrofluorimeter. CaC1, (1 mM) was added after 5 min, and further additions were made as detailed in the legends. 505-nm emission was determined in response to alternate excitation at 340 and 380 nm at intervals of3.4 s. Cytoplasmic free Ca2+ concentration, [Ca2+],, was calculated (11) using Lotus 1-2-3. Membrane Potential-The cationic carbocyanine dye DiSCz(5) (12) was used to monitor synaptosomal membrane potential changes at excitation and emission wavelengths of 643 and 666 nm, respectively. Synaptosomes (0.25 mg/ml) were incubated with 1 mM CaC12 for 5 min with constant stirringina thermostatted cuvette a t 37 "C, followed by a 5-min equilibration with 4 p M DiSC2(5). Synaptosomal Phosphorylation-Synaptosomes (2 mg/ml) were labeled with [32P]orthophosphate(1mCi/ml) for 45 min in phosphatefree incubation buffer. Extrasynaptosomal [32P]orthophosphate was removed by centrifugation in an Eppendorf microcentrifuge (10,000 X g-) for 15 8 , and thesynaptosomes were resuspended in phosphatecontaining incubation medium a t a final protein concentrationof 0.6 mg/ml. Additions were made parallel to those for the fluorometric assays. Incubations were stopped by the addition of 0.16 volume of stopping buffer to give finalconcentrations of 1% (w/v) sodium dodecyl sulfate, 6.25 mM Tris (pH 6.8), 5% 2-mercaptoethanol, 10% glycerol (v/v), and 0.001% (w/v) bromphenol blue in the final sample. Samples were boiled for 10 min and chromatographed on 7.5% SDSpolyacrylamide gels. Gels were fixed, stained with Coomassie Blue, and destained. Where appropriate, the 75-kDa band, containing both the MARCKS 87-kDa protein and the synapsins were excised from the 7.5% wet gel, proteolyze with Staphyloccoccus aureus V8 protease, and rerun on a 15% SDS-polyacrylamide gel. 32P-Labeledproteins and peptides were visualized and quantified by autoradiography or phosphoimaging using a Molecular Dynamics program (Sunnyvale, CAI. Experiments were repeated with threeto four independent synaptosomal preparations, and the data were normalized to the phosphorylation signal obtained in controlconditions. RESULTS

Modulation of PKC Activity Has Different Effects on 4-APand KCI-induced Glutamate Release-We have previously reported that PDBu potentiates the release of glutamate from cortical synaptosomes evoked by 4-AP (2).Fig. L4 shows that not only does Ro 31-8220 reverse the stimulation of total glutamate release, but reduces release to below the control both in the presence and absence of PDBu. Analysis of the Caz+ dependence of this modulation reveals that the major effect is on the net Ca2+-dependentrelease (Fig. lC), where the initial rate of glutamate release can be enhanced &fold by PDBu and inhibited3-fold by Ro 31-8220. Thus PKC can modulate Ca2+-dependent glutamate release evoked by 4-AP over a 15-fold dynamic range in vitro. TOdetermine whether the locus for PKC action was purely on ion channels regulating 4-AP-evoked "action potentials," or whether PKC was additionally involved in the exocytotic process itself, a parallel series of experiments were performed with KC1-stimulated synaptosomes. PDBu causes little stimulation ofKC1-evoked glutamate release from guinea pig cortical synaptosomes, and such as does occur is Ca2+-independent (2). Fig. 2 shows a similar insensitivity for the rat synaptosomes. While the total and Ca2+-independentrelease is slightly enhanced by PDBu, the increased release is not sensitive to Ro 31-8220 and may therefore not be mediated

21061

A

1

I

I

I

I

Tsme llCOs/D#vI

FIG. 1. PDBu and Ro 31-8220modulation of 4-AP-evoked glutamate release. Glutamate release was evoked by the addition of 1 mM 4-AP in the presence or absence of 1 mM Ca2+.1 p M PDBu was added to theappropriate traces 30 s prior to depolarization; Con, 4-AP alone; Ro-31, 10 p~ Ro 31-8220 added 2 min prior to depolarization. Ro-Sl/PDBu, Ro 31-8220 and PDBu both present. A , total release; B, release in the absence of added Ca2+;C, net Ca2+-dependent release obtained by subtracting individual Ca2+-independent traces from paired total release traces. Tracesare computer-generated means of at least 4 independent experiments. Standard error of the means were too small to plot.

I

I

I

I

T8me : l O O ~ / l a ~ v l

FIG. 2. PDBu and Ro 31-8220modulation of KC1-evoked glutamate release. Glutamate release was evoked by the addition of 30 mM KC1 in the presence or absence of 1 mM Ca2+. Other additions were as in the legend to Fig. 1. A , total release; B, release in the absence of added Ca2+;C, net Ca2+-dependentrelease. Traces are computer-generated means of at least four independent experiments. Standard errorof the means were too small to plot.

via PKC. Ro 31-8220 does not reduce the control KC1-evoked release in contrast to itseffect on 4-AP- evoked release. Most dramatically, there isabsolutely no effect of either PDBu, Ro 31-8220, or a combination of both agents on the net Ca2+dependent release of glutamate (Fig. 2C). Thus 4-AP-stimulated glutamate release is entirely dependent on a Ro 318220-sensitive PKC for release while KC1-induced release is totally insensitive. PKC and Depolarization-induced Ca2+ Elevation-Fura-2 fluorescence determines the average bulk increase in cytoplasmic free Ca2+, [Ca2+IE,within the terminal population without detectingthe high local Ca2+which triggers glutamate exocytosis (13). It is thus a qualitative measure of Ca2+entry into the terminals. Fig. 3 confirms our earlier report that PDBu causes a significant enhancement of 4-AP-evoked [Ca2+Icelevation with no effect on the elevation in response to KCl. Ro 31-8220 causes an extensive inhibition of both the 4-AP-evoked and KC1-evoked, PDBu-stimulated fura-2 responses, consistent with the extensive inhibition of Ca2+dependent glutamate release. Surprisingly, however, Ro 318220 caused a 30% inhibition of the plateau fura-2 response to KCl. However, as the block of fura-2 response with Aga-

21062

Protein Kinase C and Glutamate Release 200

@-PDBu and Ro 31-8220"To confirm that the effects of PDBu and RO 31-8220 in situ are on PKC, the major PKC 5 substrate MARCKS was selected as a monitor of activity. Fig. T c c 4 shows a representative 7.5% SDS-polyacrylamide gel eleci .-0 150 trophoresis depicting the influence of KCl, Ro 31-8220, and > PDBu on the phosphorylation profile of synaptosomal pro-aa), teins. Since MARCKS does notseparate cleanly on one.-5 dimensional electrophoresis from synapsins Ia and Ib, 5'. 0 5 100 aureus V8 protease treatment of excised synapsin/MARCKS 0 bands was performed, and the resultant peptides were sepaa, rated by a second electrophoresis through a 15% gel (14, 15). h Three major bands are obtained by this treatment (Fig. 5A) V .2 50 which have been shown (15) to correspond, respectively, to a 0 35-kDa peptide encompassing two Ca2+/calmodulin-dependu ent protein kinase I1 (CaMKII) phosphorylation sites of syn0" apsin-I, a 13-kDa MARCKS-derived peptide and a 10-kDa band which is a composite of a MARCKS peptide and the 0 + + + + - protein kinase A/Ca2+/calmodulin-dependentprotein kinase 4AP - + f KC1 I (CaMKI) site of synapsin-I. Thus phosphorylation of the - + - + - + 13-kDa band can be used as a marker of PKC activity (14, PDBu + + " 15). RO-31 Visual inspection of Fig. 5A shows that phosphorylation of FIG. 3. PDBu and Ro 31-8220 modulation of 4-AP- and KC1-induced cytosolic free calcium elevation, [Ca''.],. Addi- the 13-kDa peptide is enhanced by KC1 and PDBu and that tions of 1 mM 4AP, 1 p~ PDBu, and/or 10 p~ Ro 31-8220 (Eo-31) these effects are reversed by Ro 31-8220 at theconcentration were made exactly as described in the legend to Fig. 1. The elevation used in the functional assays described above. The data from of [Caz'], relative to the pre-depolarizationvalue 200 s after addition multiple replicate experiments is expressed quantitatively in of KC1or 4-AP is shown. [Ca2+], prior to depolarization was 140 f Fig. 6. It is apparent that an elevated Ca" associated with 4.8 nm. Ears show the means and standard errorof the means of three independent experiments. A Student's two-tailed t test found depolarization does not further increase the PDBu-elevated that there was no difference at the 1% level of significance between MARCKS phosphorylation. This may indicate that the restKC1-stimulatedcalcium elevation in the presence and absence of ing [Ca2+],is sufficient to activate the Ca2+-dependentPKC. PDBu. Results obtained during 4-AP-evoked depolarization (Figs. 5B and 6) are similar to those following KC1. It should be TABLEI noted that in the 4-AP experiment the synapsin band was PDBu and Ro 31-8220modulation of 4-AP- and KC1-induced only partially excised from the first gel to increase the sensidepolarization tivity of the 13-kDa assay, thus changes in the synapsin 35Additions were made under conditions parallel to those described kDa band are notmeaningful. In contrast to KC1,4-AP alone in the legend to Fig. I. causes no enhancement of the MARCKS phosphorylation Relative change in DiSC2(5) fluostate. In control, 4-AP-stimulated and KC1-stimulated conrescence ditions Ro 31-8220 addition causes a dephosphorylation of 1 mM 4-AP 30 m M KC1 MARCKS, indicating that continuous phosphorylation and 76 dephosphorylation of MARCKS occurs in resting synaptoControl 100 100 somes. In viewof the ability of glutamate to mimic the Depolarization functional effects of PDBu in synaptosomes (4) the ability of +10 pM RO 31-8220 86.2 f 6.26 84.6 f 2.3 glutamate dehydrogenase plus NADP+, which partially re162.8 f 2.6 106.3 f 12.29 +1 pM PDBu moves extrasynaptosomal glutamate (the basis of the fluoro+lo pM RQ 31-8220 109.0 f 1.9 82.2 f 2.2 metric assay), to decrease the initial phosphorylation of the +llMPDBu 13-kDa peptide was monitored. In two experiments, exposure of synaptosomes to glutamate dehydrogenase for 10 min deGI, a toxin which abolishes calcium influx through the glucreased the intensity of the 13-kDa peptide in the tamate release coupled channel (52), was additive on the inhibition produced by Ro 31-8220, we can assume that this PhosphorImager by 18%. The band excised from the first electrophoresis included inhibitory effect is on a channel which is not coupled t o the totalsynapsin-I and thus the35-kDa band can be used as glutamate release since there is no inhibition of Ca2+-dependa marker of CaMKII activity. The KC1-evoked increase in ent glutamate release underthese conditions. It must be synapsin phosphorylation is not blocked by Ro 31-8220 (Fig. assumed that Ro 31-8220 affects a Ca2+pool or Ca2+regulatory pathway which is not coupled to glutamate release. It should 5 A ) showing that the inhibitor retains kinase specificity in be noted that individual calibrations for each trace allowed situ. for color quenching at 340 and 380 nm caused by both 4-AP DISCUSSION and Ro 31-8220. PKC and Transmitter Release-It is important when anaSynaptosomal Membrane Potentials Monitored with DiSC2(5)-Table I shows the effects of PDBu andRo 31-8220 lyzing the extensive literature on the modulation of transmiton the fluorescence of DiSC2(5), a qualitative monitor of ter release by PKC activators and inhibitors to distinguish between smallstimulations of10-30% inthe absence of membrane potential. PDBu causes a large increase in the mean depolarization evoked by 4-AP, which is reversed by Ro external Ca2+or with KC1-depolarized preparations and the 31-8220, while having an insignificant effect on the signal far larger degree of control (2-5-fold) found in brain slices subjected to repetitive electrical stimulation (16-20) or in subsequent to KC1-evoked depolarization. Phosphorylation of "MARCKS" 87-kDa Protein, Effects of brain slices (21, 22) or synaptosomes (2, 4,23) stimulated

-

v

L

"

"

'

Protein Kinase C and Glutamate Release

21063

MW kDa

FIG.4. Polyacrylamide gel electrophoresis of “P-prelabeled synaptosomes incubated in the presence of KC1, Ro 31-8220,or PDBu. Syn-

tr

aptosomes were “P-labeled as described under “Experimental Procedures.” Where indicated 10 p~ Ro 31-8220 and/ or 1 p~ PDBu were added 135 s prior to quenching, and 30 mMKC1 was added 15 s prior to quenching. The indicated band contains both MARCKS and synapsin I.

A

-

b 3 1

PDBU

i

MARCKS 13kDa SYNAPSIN 10kDa

B

1