Calcium-Dependent Glutamate Release During Neuronal Development and Synaptogenesis: Different Involvement of &O-Agatoxin IVA- and O-Conotoxin GVIA-Sensitive Channels

Proc. Natl. Acad. Sci. USA Vol. 92, pp. 6449-6453, July 1995 Neurobiology

Calcium-dependent glutamate release during neuronal development and synaptogenesis: Different involvement of &o-agatoxin IVA- and o-conotoxin GVIA-sensitive channels

CLAUDIA VERDERIO*, SILVIA Coco*, GUIDO FUMAGALLIt, AND MICHELA MATTEOLI*t *Consiglio Nazionale delle Ricerche Center of Cytopharmacology and Bruno Ceccarelli Center, Department of Pharmacology, University of Milan, Milan, Italy; and tDepartment of Pharmacology, University of Verona, Verona, Italy Communicated by Vittorio Erspamer, University of Rome La Sapienza, Rome, Italy, March 15, 1995

ABSTRACT Hippocampal neurons maintained in primary MATERIALS AND METHODS culture recycle synaptic vesicles and express functional glutamate receptors since early stages of neuronal development. By Hippocampal Cell Culture. Primary neuronal cultures were analyzing glutamate-induced cytosolic calcium changes to sense prepared from the hippocampi of 18-day fetal rats as described presynaptically released neurotransmitter, we demonstrate that (8, 9). Briefly, hippocampi were dissociated by treatment with the ability of neurons to release glutamate in the extracellular trypsin (0.25% for 15 min at 37°C), followed by trituration with space is temporally coincident with the property of synaptic a fire-polished Pasteur pipette. Dissociated cells were plated vesicles to undergo exocytotic-endocytotic recycling. Neuronal on a poly(L-lysine)-treated glass coverslips in minimum essen- differentiation and maturation ofsynaptic contacts coincide with tial medium (MEM) with 10% horse serum at densities ranging a change in the subtype of calcium channels primarily involved from 10,000 cells per cm2 to 20,000 cells per cm2. After a few in controlling neurosecretion. Whereas o-agatoxin IVA- hours, coverslips were transferred to dishes containing a sensitive channels play a role in controlling neurotransmitter monolayer of cortical glial cells (10), so that they were sus- secretion at all stages of neuronal differentiation, ao-conotoxin pended over the glial cells but not in contact with them (9). GVIA-sensitive channels are primarily involved in mediating Cells were maintained in MEM (GIBCO) without serum, glutamate release at early developmental stages only. supplemented with 1% HL1 (Ventrex Laboratories, Portland, ME), 2 mM glutamine, and 1 mg of bovine serum albumin per Synaptic transmission is the process by which signals are ml (neuronal medium). transferred from a neuron to a target cell. At the nerve [Ca2+], Measurements with Fura-2. Neurons were loaded terminal synaptic vesicles undergo exocytotic-endocytotic re- for 30 min at 37°C with 2-4 ,M fura-2 pentakisacetoxymethyl cycling and at each cycle they release neurotransmitter in the ester in Krebs-Ringer-Henseleit solution (KRH) buffered extracellular space (1, 2). Released neurotransmitter binds to with Hepes (150 mM NaCl/5 mM KCl/1 mM MgSO4/2 mM postsynaptic receptors that are strategically localized at CaCl2/10 mM glucose/10 mM Hepes-NaOH, pH 7.4), washed postsynaptic sites and that in turn translate the external signals in the same solution to allow deesterification of the dye, and into a postsynaptic response (3). The anatomical relation transferred to the heated stage of the microscope, where between pre- and postsynaptic compartments is the basis on temperature was maintained at 35°C throughout the experi- which the point-to-point signaling typical of synaptic transmis- ment. Cells under different experimental conditions (see sion relies. below) were observed with an inverted Zeiss IM35 microscope However, several lines of evidence indicate that the func- equipped with a calcium-imaging unit (11). Fluorescence tionality of pre- and postsynaptic components is not dependent images were obtained by alternately illuminating the cells at on their spatial organization at synaptic sites. It has been 345 nm and 380 nm and emission filtered with a 418-nm demonstrated recently that synaptic vesicles present in devel- long-pass filter; images (345 nm or 380 nm) consisted of 0.125 oping processes of cultured hippocampal neurons are able to megabyte (MB) each and were acquired rhythmically at 2 Hz undergo exocytotic-endocytotic recycling (4, 5), with mecha- frequency. Fluorescence images were collected with an inten- nisms that appear to be already calcium-dependent (6). More- sified charge-coupled device camera (Hamamatsu Photonics, over, functional glutamate receptors are expressed already at Middlesex, NJ) and the camera output was analyzed by a digital stages preceding synapse formation (7). image processor (Argus 100; Hamamatsu Photonics) where We now investigate whether synaptic vesicle recycling before video frames were digitized and integrated in real time on four synaptogenesis is associated with neurotransmitter release. To 0.5-MB memory boards. The digital data were then transferred this aim, we used a biological assay that allows us to perform at high rate via a connecting board placed on the VME bus of an on-line analysis of glutamate released by hippocampal Argus 100 into a Motorola 68020-based host computer and neurons by analyzing the cytosolic calcium ([Ca2+]j) changes stored in two 300-MB hard disks. Background and calibration induced by N-methyl-D-aspartate (NMDA) receptor activa- images were similarly acquired at the two wavelengths; the tion. We show that hippocampal neurons are able to release calculation of Ca2+ concentration was carried out pixel by pixel glutamate by a calcium-dependent mechanism even before on pairs of corresponding 345-nm and 380-nm images according synapse formation. Moreover, we demonstrate that c-agatoxin to Grynkiewicz et at (12). The time resolution of the measure- IVA (w-Aga-IVA) and w-conotoxin GVIA (co-CTx-GVIA) have different effects on blocking glutamate release at various Abbreviations: [Ca2+]i, cytosolic calcium concentration; NMDA, N- developmental stages, indicating that formation and matura- methyl-D-aspartate; CNQX, 6-cyano-7-nitroquinoxaline-2,3-dione; tion of synaptic contacts are associated with changes in the APV, D-2-amino-5-phosphonopentanoic acid; w-Aga-IVA, w-aga- subtypes of voltage-activated calcium channels controlling toxin IVA; co-CTx-GVIA, w-conotoxin GVIA; BoNT/F, toxin from presynaptic glutamate release. Clostridium botulinum of F serotype; PDC, L-trans-pyrrolidine-2,4- dicarboxylic acid. 4To whom reprint requests should be addressed at: Consiglio Nazio- The publication costs of this article were defrayed in part by page charge nale delle Ricerche Center of Cytopharmacology and B. Ceccarelli payment. This article must therefore be hereby marked "advertisement" in Center, Department of Medical Pharmacology, University of Milan, accordance with 18 U.S.C. §1734 solely to indicate this fact. via Vanvitelli 32, 20129 Milan, Italy. 6449 Downloaded by guest on September 30, 2021 6450 Neurobiology: Verderio et aL Proc. Natl. Acad. Sci. USA 92 (1995) ments was 1 s. All temporal plots were obtained from neuronal A cell bodies. APV CNQX Experimental Treatments. Incubations were carried out either in 2 ml of KRH or in 2 ml of its Mg2+-free version, which

contained no MgCl2 and was supplemented with 1 ,uM glycine. - Application of NMDA (25 ,tM) or glutamate (20 j,M) was performed in a medium free of Mg2> ions. Incubations with Ca2+ channel blockers (w-Aga-IVA and cw-CTx-GVIA) were performed at 35°C for 20-25 min in KRH before the Mg2+- 'Us- BoNT/F IGlutamate free/glycine treatment. Drugs were maintained in the incubation medium for the entire length of the experiment. Incuba- + 400- tions with toxin from Clostridium botulinum of F serotype (BoNT/F) were performed at 30°C for 2 hr. For the experi- 100 p J f - Mg2- free/Gly ments aimed at accumulating glutamate in the extracellular medium, coverslips were maintained in a fixed amount of KRH -C Cd2+ NMDA NMDA (1 or 2 ml) in the presence of the glutamate uptake inhibitor 700- Cd2^ L-trans-pyrrolidine-2,4-dicarboxylic acid (PDC) (20 ,tM). The collected medium was divided in two aliquots, one of which 400- was supplemented with the non-NMDA glutamate receptor antagonist 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) (20 100- -_____J_ Mg2+ free/Gly j.tM). The same neuronal culture was then exposed subse- min quently to the two aliquots of medium. In other experiments, the same coverslip was first incubated for 1 hr with 1 ml of FIG. 1. Activation of NMDA receptors by presynaptic release of KRH and then incubated with a new 1 ml of KRH containing glutamate in neurons rich in synaptic contacts. (A) Temporal plot of the calcium channel blocker Cd2+ (100 ,tM). Before challeng- [Ca]2+i changes induced by the Mg2+-free/glycine medium in the cell ing a new coverslip with the two aliquots of conditioned body of a 14-day-old hippocampal neuron. [Ca]2+ responses are medium in sequence, Cd2> was added to the Cd2+-free aliquot inhibited by 100 ,uM D-2-amino-5-phosphonopentanoic acid (APV) (final concentration, 100 ,uM). but not by 10 ,uM CNQX. (B) [Ca2+]i responses to the Mg2+-free/ Materials. Most of the chemicals used were purchased from glycine medium in the soma of a 10-day-old neuron before and after Tocris Neuramin (Bristol, U.K.). Fura-2 was from Calbiochem. incubation (2 hr) with 7 nM BoNT/F. Note that [Ca2+], changes are International and totally prevented by toxin treatment. The presence of BoNT/F does c-Aga-IVA was purchased from Peptides not impair the neuronal ability to respond to 20 t,M glutamate. (C) w-CTx-GVIA was from Bachem. [Ca2+]i responses to Mg2+-free/glycine medium in the cell body of a 12-day-old neuron in the presence or in the absence of 200 AM Cd2+ RESULTS Cd2+ strongly inhibits the response to the Mg2+-free/glycine medium and reduces only partially the postsynaptic response to 25 ,tM NMDA. Mg2+-Free/Glycine Protocol as a Tool To Detect Glutamate Note that the kinetics of [Ca2+]i responses to the Mg2+-free/glycine Release from Hippocampal Neurons. To study glutamate protocol may vary depending on the neuron examined. The bars under release from hippocampal neurons, we devised a biological each curve indicate incubation in the Mg2+-free/glycine medium. assay that allows us to analyze [Ca2+], changes induced by endogenously released glutamate. This method is based on the mate (16). To test this possibility, we incubated neuronal use of solutions free of Mg2+ ions and supplemented with 1 cultures (donor cells) in a fixed volume (1 ml) of complete ,LM glycine, an experimental protocol that is known to relieve medium for 1 hr at 37°C in the presence of the glutamate the block of NMDA receptors and to allosterically facilitate its uptake inhibitor PDC (20 ,tM). A fixed aliquot of the condi- activation by glutamate (13, 14). When fura-2-loaded, fully tioned medium was then used to stimulate other neuronal cells. differentiated hippocampal neurons rich in synaptic contacts A [Ca>2 ]i response was indeed evoked in neuronal cells (10- to 20-day-old neurons) were exposed to this medium, exposed to the PDC-conditioned medium (number of exper- >98% of the cells (n = 87; number of experiments = 17) iments = 5) (Fig. 2A). At difference with the responses elicited exhibited large increases in the levels of intracellular calcium by the Mg2+-free/glycine medium, the effects of the PDC- (Fig. IA and ref. 15). The intensity and the kinetics of the conditioned medium were blocked by the non-NMDA antag- [Ca2+], changes at the level of the soma were highly repro- onist CNQX (Fig. 2B). This indicates that higher concentra- ducible when the same cell was repeatedly challenged with the tions of endogenously released glutamate were required to Mg2+-free/glycine medium. On the other hand, the levels and activate the non-NMDA receptor-dependent mechanisms for the kinetics of the [Ca2+], increases differed greatly between Ca2+ influx. neurons, the response being either persistent or transient To provide a direct demonstration that activation of NMDA depending on the cell examined. receptors following the Mg2+-free/glycine treatment was pro- The effects of the Mg2+-free/glycine medium were blocked duced by presynaptic vesicular release of glutamate, neuronal by 100 ,tM APV, a specific NMDA antagonist, and not by 10 cultures were preincubated for 2 hr in the presence of BoNT/F ,uM CNQX, a non-NMDA receptor blocker (Fig. 1A); these (7 nM) before challenging them with the Mg2+-free/glycine data are in agreement with published results (ref. 15) and medium. BoNT/F is a metalloendoprotease that potently indicate that the [Ca2>], changes are triggered by NMDA inhibits presynaptic vesicle exocytosis (for a review see ref. 17). receptor activation. Since the Mg2+-free/glycine medium also We found that this treatment completely prevented the [Ca]2+, activates an increase in spontaneous electrical activity (15), at changes induced by the Mg2+-free/glycine protocol without least part of the [Ca2+] changes were probably due to ion affecting the neuronal ability to respond to exogenous gluta- influx through voltage-operated Ca2+ channels activated by mate (Fig. 1B). NMDA receptor-induced depolarization. To test the hypothesis that the release of glutamate is at least The need to remove Mg2+ ions from the extracellular in part due to Ca2+-dependent mechanisms, 100-200 ,AM medium in order to detect a neuronal response to released Cd2+, a nonspecific calcium channel blocker, was added to the glutamate could be accounted for by the low levels of gluta- Mg2+-free/glycine medium. Under this condition, the [Ca>2+] mate present at the steady state in the extracellular medium responses were almost completely abolished (n = 19; number and by the low sensitivity of non-NMDA receptors to gluta- of experiments = 4) (Fig. 1C). The effect was not primarily due Downloaded by guest on September 30, 2021 Neurobiology: Verderio et at Proc. Natl. Acad. Sci. USA 92 (1995) 6451

A B A 400- APV CNQX 300 250- 1500 r ______Mg2+free/Gly 100 - . IL ro 500B Cd2+ NMDA NMDA 1C D 500 -~

300 - 300' 100 ~ Mg2+-free/Gly min 100 H min FIG. 3. Activation of NMDA receptors by release of glutamate in neurons before synaptogenesis. (A) Temporal plot of [Ca2+]i changes FIG. 2. Temporal plots of [Ca2+]i changes in single mature neurons induced by the Mg2+-free/glycine medium in the cell body of a single, exposed to media conditioned by exposure to 15-day-old neurons early-stage (2-day-old) neuron. [Ca2+]i responses are inhibited by 100 under various experimental treatments. (A) KRH containing the ,LM APV but not by 10 ,uM CNQX. (B) Temporal analysis of [Ca2+]i glutamate uptake inhibitor PDC (20 ,uM) and conditioned by a 1-hr responses to Mg2+-free/glycine in the soma of a single 3-day-old incubation. (B) KRH/PDC medium conditioned as in A and then neuron in the presence or in the absence of 200 ,uM Cd2+. Cd2+ supplemented with 20 ,uM CNQX. (C) KRH/PDC medium condi- completely blocks the response to the Mg2+-free/glycine medium but tioned as in A and then supplemented with 100 ,uM cadmium. (D) does not interfere with the neuronal response to 25 ,iM NMDA. KRH/PDC medium containing 100 ,uM cadmium and conditioned by a 1-hr incubation. Note that the elicited response is blocked by the Fig. 4 A-C indicate that pretreatment of 10- to 20-day-old non-NMDA antagonist CNQX. Glutamate release is strongly inhib- neurons with 300 nM c-CTx-GVIA only partially inhibited the ited when the conditioning cells are incubated in the presence of the responses to the Mg2+-free/glycine medium (n = 28; number Ca2+ channel blocker Cd2+. of experiments = 8). By contrast, in all 2- to 6-day-old neurons examined (n = 17; number of experiments = 7), w-CTx-GVIA to interference with NMDA receptors since Cd2+ only slightly inhibited almost completely (80-100% inhibition) the re- inhibited [Ca]2+] responses to direct application of NMDA sponses to the Mg2+-free/glycine medium (Fig. 4 D-F). In (Fig. IC). Additional evidence that the release of glutamate is both cases, the toxin did not affect the response induced by Ca2+-dependent came from PDC-conditioned medium stud- direct application of NMDA (not shown). ies: here the responses were greatly inhibited when the donor The effects of w-Aga-IVA were less strikingly modified cells were incubated for 1 hr in both PDC and Cd2+ (compare during neuronal differentiation. Before and after synaptogen- Fig. 2 C and D) (number of experiments = 3). esis (n = 24 and 33; number of experiments = 7 and 9, These findings therefore validate the Mg2+-free/glycine respectively), w-Aga-IVA substantially inhibited the response experimental protocol as a tool to sense presynaptic release of to the Mg2+-free/glycine medium (Fig. 4 C and F); however, glutamate from cultured hippocampal neurons. in a large number of cells, the responses were less potently Release of Glutamate Before Synaptogenesis. We then inhibited by the toxin before synaptogenesis (see an extreme asked the question of whether the exocytotic-endocytotic case of minimal efficacy in Fig. 4D). Also in the case of recycling of synaptic vesicles taking place in hippocampal co-Aga-IVA, the responses to NMDA were not affected by the neurons before synaptogenesis (4, 5) was already associated toxin (not shown). with glutamate release. Neuronal cultures at early stages of differentiation and before extensive synaptogenesis (2- to 6-day-old neurons) were exposed to the Mg2+-free/glycine DISCUSSION medium. In this case an increase in [Ca2+], could be detected We have used a calcium-imaging technique to study release of in most of the neurons (60%, n = 19). This response was glutamate from cultured hippocampal neurons at different selectively blocked by APV but not by CNQX (Fig. 3A). The developmental stages. This approach, based on the detection kinetics of the responses were found to be different in distinct and measurement of [Ca2+], changes induced by NMDA neurons and, in general, they appeared to be slower as receptor activation, allows the on-line monitoring of glutamate compared to the [Ca]2+ changes induced by the Mg2+-free/ release from a low-density culture. glycine medium in more differentiated cells. The Mg2+-free/glycine protocol has been widely reported in We found that even at early. developmental stages, the the literature because it relieves the block of NMDA receptor response to the Mg2+-free/glycine treatment was blocked by and allosterically facilitates its activation by glutamate (13, 14). the calcium channel blocker Cd>2 (Fig. 3B), indicating that a Exposure to the Mg2+-free/glycine medium produces large Ca2+-dependent mechanism for release of glutamate is active fluctuations in [Ca2+], in synaptically connected hippocampal in hippocampal neurons before synaptogenesis. No direct neurons (ref. 15 and this study). The lack of response to effect of Cd2+ could be detected on neuronal responses to Mg2+-free/glycine medium in the presence of APV indicates NMDA (Fig. 3B). that [Ca>2]1 increases are the result of calcium influx through Different Involvement of w-Aga-IVA- and w-CTx-GVIA- the activated NMDA receptors. Since a slight reduction of Sensitive Channels in Mediating Glutamate Release at Different [Ca>2]i response to maximal doses of NMDA is elicited in Developmental Stages. Different types of calcium channels me- some neurons in the presence of Cd>, we cannot exclude the diate central synaptic transmission. In adult hippocampus, w-Aga- possibility that part of the Ca>2 influx induced by the Mg2+- IVA-sensitive and c-CTx-GVIA-sensitive channels mediate free/glycine medium occurs also through voltage-activated neurotransmitter release (18-20). We therefore investigated Ca2+ channels. These Ca2+ channels are most likely dihydro- the effects of these two toxins on the responses induced by the pyridine-sensitive Ca2+ channels since no effect on the neu- Mg2+-free/glycine protocol in neurons before and after syn- ronal response to NMDA was ever seen by preincubation with aptogenesis. either co-Aga-IVA or w-CTx-GVIA. Downloaded by guest on September 30, 2021 6452 Neurobiology: Verderio et at Proc. Natl. Acad. Sci. USA 92 (1995)

A D 0)-CTx-GVIA o-CTx-GVIA o,-Aga-IVA w-Aga-IVA -Aga- IVA 300 (o-CTx-GVIA

M-+ - I-ee/Gly Mg2,- free/Gly v Illillfl .F mlill

r, 11 5 E (0)-Aga- IVA 350 - o-CTx-GVIA

100()-' A-~ Mg2- M& t-rec/Glv frec/Gly

l"')T C *o-CTx-GiVIA I100 -r F E O -Ag.a-IVA 8() 8() + I 0-20 datys ()ld 2-6 days old

60 6() r-

-d 40 40 Iz 20 -t

0) - .-- 0-20 21-40 41 60 61 80 81-10() 0-20 21-4() 41-60 6) 8o 81-100 4( inihibiti c1c inhibition

FIG. 4. Different involvement of co-Aga-IVA- and o-CTx-GVIA-sensitive Ca2+ channels in mediating glutamate release from neurons at different developmental stages. (A and B) Temporal plots of [Ca2+]i responses to the Mg2+-free/glycine medium in the somata of two 14-day-old neurons, rich in synaptic contacts. [Ca2+], changes are partially reduced by pretreatment with 300 nM w-CTx-GVIA (A) and almost completely inhibited by pretreatment with 100 nM w-Aga-IVA (B). Combination of the two toxins completely inhibited the [Ca2+]j response (A). (D and E) Temporal plots of [Ca2+]i responses to the Mg2+-free/glycine medium in the somata of two 3-day-old, developing neurons: in this extreme case, [Ca2+], changes are moderately reduced by pretreatment with 100 nM w-Aga-IVA (D) and almost completely inhibited by pretreatment with 300 nM w-CTx-GVIA (E). Combination of the two toxins completely inhibited the [Ca2+]i response (D). (C and F) Frequency distribution of degrees of inhibition produced by c-CTx-GVIA and w-Aga-IVA treatments in neurons before (F) and after (C) the establishment of synaptic contacts. On the abscissa classes of response inhibition (percentage) are reported.

The neuronal response to the Mg2+-free/glycine medium is release and postsynaptic receptors) and, in particular, to the due to glutamate released from presynaptic vesicles. This is progressive appearance of NMDA receptors. indicated by two lines of evidence. (i) The neuronal response Once validated the Mg2+-free/glycine protocol as a tool to to the Mg2+-free/glycine protocol is completely inhibited by monitor vesicular release of glutamate from hippocampal treatment with BoNT/F, which strongly inhibits neurotrans- neurons, we used this assay to investigate neurotransmitter mitter release by cleaving the synaptic vesicle protein synap- secretion before synaptogenesis. In the case of isolated grow- tobrevin/VAMP2 (for a review, see ref. 17). (ii) The effect of ing neurons the classic electrophysiological approach (21) the Mg2+-free/glycine medium is strongly reduced by toxins or based on the detection of postsynaptic potentials, which closely procedures known to interfere with the functionality of Ca2+ reflect neurotransmitter secretion from presynaptic terminals, channels relevant for neurotransmission in hippocampus (19, is not possible because of the lack of "bona fide" postsynaptic 20). Since synaptically coupled hippocampal neurons present elements. On the other hand, the fluorimetric enzymatic larger synaptic potentials when exposed to the Mg2+-free/ approach currently used for monitoring glutamate release glycine medium (15), the endogenous glutamate release we from synaptosomes (for a review, see ref. 22) is not sensitive detect with this assay is most likely a consequence of an enough to detect the small amount of glutamate released by increased spontaneous electrical activity. Taken together, developing hippocampal neurons (F. Valtorta, C.V., and M.M., these results indicate that neuronal responses to the Mg2+_ unpublished observations). The previous finding that cultured free/glycine protocol are produced by activation of NMDA hippocampal neurons express functional glutamate NMDA re- receptors by glutamate released from presynaptic terminals ceptors already at very early stages of neuronal development (7) through a synaptic vesicle-mediated, Ca2+-dependent mecha- opened the possibility to apply the Mg2+-free/glycine protocol as nism. The removal of Mg2> from the extracellular medium and a suitable tool to investigate neurotransmitter secretion during the addition of the allosteric modulator glycine allow NMDA neuronal development. receptors to be activated even by a low concentration of We report here that the ability of hippocampal neurons to glutamate. Non-NMDA receptor activation requires higher release glutamate by a Ca2+-dependent mechanism is not concentrations of glutamate such as those that accumulate in dependent on the formation and maturation of synaptic con- the medium during 1-hr incubation in the presence of the tacts. Our data complement results obtained by a parallel line glutamate uptake inhibitor PDC. The hetereogeneity of the of investigation indicating that the exocytotic-endocytotic responses to the Mg2+-free/glycine protocol between cultures recycling of synaptic vesicles is Ca2+-dependent already before at different developmental stages and even among neurons in synaptogenesis (6). the same culture could be related to different degrees of The early activation of Ca2'-dependent mechanisms for neuronal maturation (i.e., distance between sites of glutamate neurosecretion raised the question of whether different sub- Downloaded by guest on September 30, 2021 Neurobiology: Verderio et at Proc. Natl. Acad. Sci. USA 92 (1995) 6453 types of Ca21 channels were involved in mediating glutamate inhibition of evoked [Ca2+]i responses could be revealed in release at different developmental stages. Recent electrophysi- different neuronal regions after treatment with c-CTx-GVIA ological studies have demonstrated that at least two types of and cw-Aga-IVA. Ca2+ channels, the w-Aga-IVA- and co-CTx-GVIA-sensitive channels, mediate synaptic transmission in the mammalian We acknowledge Drs. F. Clementi and E. Sher (University of central nervous system (18-20). o-CTx-GVIA- and w-Aga- Milan) for critically reading the manuscript and Dr. C. Montecucco IVA-sensitive channels might correspond to the Ca2+ chan- for the gift of BoNT/F. This work was supported by the Consiglio nels previously classified as N type (23) and P type (18, 19), Nazionale delle Ricerche (to M.M.) and by Telethon Italy (to G.F.). even if the possibility exists that w-Aga-IVA-sensitive channels 1. De Camilli, P. & Jahn, R. (1990)Annu. Rev. Physiol. 52,625-645. are represented by a heterogeneous population of channels 2. Bennett, M. K. & Scheller, R. H. (1994)Annu. Rev. Biochem. 63, (20). In the adult hippocampus (19) as well as in cultured, 63-100. synaptically coupled hippocampal neurons (this study), a pre- 3. Hall, Z. W. & Sanes, J. R. (1993) Cell 72/Neuron 10, Rev. Suppl. dominant role in neurotransmission appears to be played by 99-121. co-Aga-IVA-sensitive channels. We show now that co-CTx- 4. Matteoli, M., Takei, K., Perin, M. S., Sudhof, T. C. & De Camilli, GVIA-sensitive Ca2+ channels as well are crucially involved in P. (1992) J. Cell Biol. 117, 849-861. controlling glutamate release from hippocampal neurons be- 5. Mundigl, O., Matteoli, M., Daniell, L., Thomas-Reetz, A., Met- calf, A., Jahn, R. & De Camilli, P. (1993) J. Cell Biol. 122, fore the establishment and maturation of synaptic contacts. 1207-1221. These results indicate that the molecular mechanisms respon- 6. Kraszewski, K., Mundigl, O., Daniell, L., Verderio, C., Matteoli, sible for Ca2+-dependent release of glutamate are differen- M. & De Camilli, P. (1995) J. Neurosci., in press. tially activated in neurons at different developmental stages. 7. Verderio, C., Coco, S., Fumagalli, G. & Matteoli, M. (1994) w-CTx-GVIA-sensitive channels have been described to be J. Cell Biol. 126, 1527-1536. functional and to play a role during neuronal differentiation 8. Banker, G. A. & Cowan, W. M. (1977) Brain Res. 126, 379-425. since they mediate spontaneous elevation of [Ca2+], in undif- 9. Bartlett, W. P. & Banker, G. A. (1984)J. Neurosci. 4, 1944-1953. ferentiated developing neurons (24). A selective role of w-CTx- 10. Booher, J. & Sensenbrenner, M. (1972) Neurobiology 2, 97-105. GVIA-sensitive channels in 11. Grohovaz, F., Zacchetti, D., Clementi, E., Lorenzon, P., Mel- directing migration of immature dolesi, J. & Fumagalli, G. (1991) J. Cell Biol. 113, 1341-1350. neurons before the establishment of their synaptic circuits has 12. Grynkiewicz, G., Poenie, H. & Tsien, R. Y. (1985) J. Biol. Chem. been demonstrated in mouse cerebellar slices, indicating that 260, 3440-3450. the early expression of w-CTx-GVIA-sensitive channels may 13. Ascher, P. & Nowak, L. (1988) J. Physiol. (London) 399, 247-266. be an essential requisite to the initiation and the execution of 14. Benveniste, M., Clements, J., Vyklicky, L. & Mayer, M. L. (1990) neuronal movement (25). Our demonstration that w-CTx- J. Physiol. (London) 428, 333-357. GVIA-sensitive channels may control neurotransmitter re- 15. Abele, A. E., Scholz, K. P., Scholz, W. P. & Miller, R. J. (1990) lease during neuronal development opens the possibility that Neuron 2, 413-419. 16. Kiskin, N. I., Krishtal, 0. A. & Tsyndrenko, A. Y. (1986) Neu- the early activation of the neuronal secretory machinery rosci. Lett. 63, 225-230. modulates the interactions between the growth cone and its 17. Huttner, W. B. (1993) Nature (London) 365, 104-105. immediate environment. A role of secreted neurotransmitters 18. Turner, T. J., Adams, M. E. & Dunlap, K. (1992) Science 258, as developmental signals has been widely proposed (26, 31). 310-313. Neurotransmitters may act as specific chemoattractants for 19. Takahashi, T. & Momiyama, T. (1993) Nature (London) 366, 156. growth cone guidance, suggesting that localized secretion of 20. Wheleer, D. B., Randall, A. D. & Tsien, R. W. (1994) Science neuromediators from the target cell in the embryo may 264, 107-111. influence the direction of neurite growth during the develop- 21. Del Castillo, J. & Katz, B. (1956) Prog. Biophys. Chem. 6, 121-170. ment of the nervous 22. Nicholls, D. G. (1989) J. Neurochem. 52, 331-341. system (27). 23. Sher, E. & Clementi, F. (1991) Neuroscience 42, 301-307. In parallel with neuronal differentiation, w-CTx-GVIA- 24. Holliday, J. & Spitzer, N. C. (1990) Dev. Biol. 141, 13-23. sensitive channels become less involved in controlling neuro- 25. Komuro, H. & Rakic, P. (1992) Science 257, 806-809. transmitter release. This reduction could be accounted for by 26. Meier, E., Hertz, L. & Schousboe, A. (1991) Neurochem. Int. 19, a down-regulation in the channel expression. Alternatively, a 1-15. specific localization of distinct subtypes of Ca2+ channels in 27. Zheng, J. Q., Felder, M., Connor, J. A. & Poo, M. M. (1994) different districts of neuronal plasmalemma could be hypoth- Nature (London) 368, 140-144. esized. In mature neurons, a different distribution of co-CTx- 28. Jones, 0. T., Kunze, D. L. & Angelides, K. J. (1989) Science 244, 1189-1193. GVIA- and dihydropyridine-sensitive subtypes of calcium 29. Westenbroek, R. L., Ahlijanian, M. K. & Catterall, W. A. (1990) channels has been demonstrated (28-30), sustaining therefore Nature (London) 347, 281-284. the possibility that distinct domains of mature hippocampal 30. Benke, T. A., Jones, 0. T., Collingridge, G. L. & Angelides, K. J. neuron membranes are endowed with specific Ca2+ channel (1993) Proc. Natl. Acad. Sci. USA 90, 7819-7823. subtypes. It will be interesting to investigate whether a selective 31. Mattson, M. P. (1988) Brain Res. Rev. 13, 179-212. Downloaded by guest on September 30, 2021