Co-expression of Metabotropic Glutamate Receptor 7 and N-type Ca2+ Channels in Single Cerebrocortical Nerve Terminals of Adult Rats
2003; Elsevier BV; Volume: 278; Issue: 26 Linguagem: Inglês
10.1074/jbc.m211471200
ISSN1083-351X
AutoresCarmelo Millán, Enrique Castro, Magdalena Torres, Ryuichi Shigemoto, José Sánchez‐Prieto,
Tópico(s)Ion channel regulation and function
ResumoThe modulation of calcium channels by metabotropic glutamate receptors (mGluRs) is a key event in the fine-tuning of neurotransmitter release. Here we report that, in cerebrocortical nerve terminals of adult rats, the inhibition of glutamate release is mediated by mGluR7. In this preparation, the major component of glutamate release is supported by P/Q-type Ca2+ channels (72.7%). However, mGluR7 selectively reduced the release component that is associated with N-type Ca2+ channels (29.9%). Inhibition of P/Q channels by mGluR7 is not masked by the higher efficiency of these channels in driving glutamate release when compared with N-type channels. Thus, activation of mGluR7 failed to reduce the release associated with P/Q channels when the extracellular calcium concentration, ([Ca2+]o), was reduced from 1.3 to 0.5 mm. Through Ca2+ imaging, we show that Ca2+ channels are distributed in a heterogeneous manner in individual nerve terminals. Indeed, in this preparation, nerve terminals were observed that contain N-type (31.1%; conotoxin GVIA-sensitive) or P/Q-type (64.3%; agatoxin IVA-sensitive) channels or that were insensitive to these two toxins (4.6%). Interestingly, the great majority of the responses to l-AP4 (95.4%) were observed in nerve terminals containing N-type channels. This specific co-localization of mGluR7 and N-type Ca2+-channels could explain the failure of the receptor to inhibit the P/Q channel-associated release component and also reveal the existence of specific targeting mechanisms to localize the two proteins in the same nerve terminal subset. The modulation of calcium channels by metabotropic glutamate receptors (mGluRs) is a key event in the fine-tuning of neurotransmitter release. Here we report that, in cerebrocortical nerve terminals of adult rats, the inhibition of glutamate release is mediated by mGluR7. In this preparation, the major component of glutamate release is supported by P/Q-type Ca2+ channels (72.7%). However, mGluR7 selectively reduced the release component that is associated with N-type Ca2+ channels (29.9%). Inhibition of P/Q channels by mGluR7 is not masked by the higher efficiency of these channels in driving glutamate release when compared with N-type channels. Thus, activation of mGluR7 failed to reduce the release associated with P/Q channels when the extracellular calcium concentration, ([Ca2+]o), was reduced from 1.3 to 0.5 mm. Through Ca2+ imaging, we show that Ca2+ channels are distributed in a heterogeneous manner in individual nerve terminals. Indeed, in this preparation, nerve terminals were observed that contain N-type (31.1%; conotoxin GVIA-sensitive) or P/Q-type (64.3%; agatoxin IVA-sensitive) channels or that were insensitive to these two toxins (4.6%). Interestingly, the great majority of the responses to l-AP4 (95.4%) were observed in nerve terminals containing N-type channels. This specific co-localization of mGluR7 and N-type Ca2+-channels could explain the failure of the receptor to inhibit the P/Q channel-associated release component and also reveal the existence of specific targeting mechanisms to localize the two proteins in the same nerve terminal subset. One function of presynaptic voltage-gated Ca2+ channels is to initiate the rapid influx of Ca2+ in nerve terminals, triggering the exocytotic release of neurotransmitters (1Llinás R. Sugimori M. Silver R.B. Science. 1992; 256: 677-679Google Scholar). Group III metabotropic glutamate receptors (mGluRs 4,6,7, and 8) 1The abbreviations used are: mGluRs, metabotropic glutamate receptors; l-AP4, l-(+)-2-amino-4-phosphonobutyrate; ω-CgT-GVIA, ω-conotoxin-GVIA; ω-Aga-IVA, ω-agatoxin-IVA; 4AP, 4-amino-pyridine; VGLUT1 and 2, vesicular glutamate transporters 1 and 2; GAD-65, glutamic acid decarboxylase; [Ca2+]cyt, cytosolic free calcium concentration; [Ca2+]o, extracellular calcium concentration; HBM, HEPES buffer medium; TBS, Tris-buffered saline; TTx, tetrodotoxin; GABA, γ-aminobutyric acid. 1The abbreviations used are: mGluRs, metabotropic glutamate receptors; l-AP4, l-(+)-2-amino-4-phosphonobutyrate; ω-CgT-GVIA, ω-conotoxin-GVIA; ω-Aga-IVA, ω-agatoxin-IVA; 4AP, 4-amino-pyridine; VGLUT1 and 2, vesicular glutamate transporters 1 and 2; GAD-65, glutamic acid decarboxylase; [Ca2+]cyt, cytosolic free calcium concentration; [Ca2+]o, extracellular calcium concentration; HBM, HEPES buffer medium; TBS, Tris-buffered saline; TTx, tetrodotoxin; GABA, γ-aminobutyric acid. are located in glutamatergic nerve terminals where they act as autoreceptors mediating the feedback inhibition of glutamate release (2Forsythe I.D. Clements J.D. J. Physiol. 1990; 429: 1-16Google Scholar, 3Gereau IV, R.W. Conn P.J. J. Neurosci. 1995; 15: 6879-6889Google Scholar, 4Herrero I. Vázquez E. Miras-Portugal M.T. Sánchez-Prieto J. Eur. J. Neurosci. 1996; 8: 700-709Google Scholar, 5O′Connor V. El Far O. Bofill-Cardona E. Nanoff C. Freissmuth M. Karschin A. Airas J.M. Betz H. Boehm S. Science. 1999; 286: 1180-1184Google Scholar). Through their coupling to Gi/o proteins, activation of these receptors inhibits adenylyl cyclase (6Conn P.J. Pin J.P. Annu. Rev. Pharmacol. Toxicol. 1997; 37: 205-237Google Scholar). However, the inhibition of evoked glutamate release has primarily been associated with a reduction in the activity of the Ca2+ channels that govern the exocytotic process (7Trombley P.Q. Westbrook G.L. J. Neurosci. 1992; 12: 2043-2050Google Scholar, 8Sahara Y. Westbrook G.L. J. Neurosci. 1993; 13: 3041-3050Google Scholar, 9Takahashi T. Forsythe I.D. Tsujimoto T. Barnes-Davies M. Onodera K. Science. 1996; 274: 594-597Google Scholar, 10Vázquez E. Sánchez-Prieto J. Eur. J. Neurosci. 1997; 9: 2009-2018Google Scholar, 11Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 14092-14101Google Scholar, 12Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 47796-47803Google Scholar). In cerebrocortical nerve terminals, the release of glutamate is predominantly controlled by P/Q-type Ca2+ channels, although a significant but less important contribution comes from N-type Ca2+ channels. However, we have recently found that the activation of mGluR7 only inhibits the N-type Ca2+ channel coupled release (11Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 14092-14101Google Scholar). A number of factors might influence the specific coupling between mGluR7 and the Ca2+ channels that control glutamate release. These include structural determinants in the mGluRs (13Perroy J. Gutierrez G.J. Coulon V.R. Bockaert J. Pin J.-P. Fagni L. J. Biol. Chem. 2001; 276: 45800-45805Google Scholar), the G proteins (14García D.E. Li B. García-Ferreiro R.E. Hernández-Ochoa E.O. Yan K. Gautam N. Caterall W.A. Macki K. Hille B. J. Neurosci. 1998; 18: 9163-9170Google Scholar), or the Ca2+ channels (15Zhang J.-F. Ellinor P.T. Aldrich R.W. Tsien R.W. Neuron. 1996; 17: 991-1003Google Scholar, 16Simen A.A. Lee C.C. Simen B.B. Bindokas V.P. Miller R.J. J. Neurosci. 2001; 21: 7587-7597Google Scholar) as well as the differential coupling of the Ca2+ channels to the release process (17Qian J. Noebels J.L. J. Neurosci. 2001; 21: 3721-3728Google Scholar, 18Millán C. Sánchez-Prieto J. Neurosci. Lett. 2002; 330: 29-32Google Scholar). An additional factor that should also be considered when contemplating the selective inhibition of the N-type Ca2+ channel by mGluR7 is that of the co-localization of the receptor and target Ca2+ channel at nerve terminals. In synapses of the central nervous system, the release of glutamate is supported by the P/Q- and N-types of Ca2+ channels (17Qian J. Noebels J.L. J. Neurosci. 2001; 21: 3721-3728Google Scholar, 19Luebke J.I. Dunlap K. Turner T.J. Neuron. 1993; 11: 895-902Google Scholar, 20Mintz I.M. Sabatini B.L. Regehr W.G. Neuron. 1995; 15: 675-688Google Scholar, 21Wheeler D.B. Randall A. Tsien R.W. J. Neurosci. 1996; 16: 2226-2237Google Scholar, 22Reid C.A. Clements J.D. Bekkers J.M. J. Neurosci. 1997; 17: 2738-2745Google Scholar). Thus, in view of the heterogeneous distribution of Ca2+ channels in these terminals that may express Ca2+ channels of the N-type, P/Q-type, or a combination of both types (12Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 47796-47803Google Scholar, 22Reid C.A. Clements J.D. Bekkers J.M. J. Neurosci. 1997; 17: 2738-2745Google Scholar, 23Reuter H. Neuron. 1995; 14: 773-779Google Scholar), the specific co-expression of mGluR7 and N-type channels in the same nerve terminal is a distinct possibility. In this paper we have used Ca2+ imaging in single nerve terminals from the cerebral cortex of adult rats to analyze the effects of Ca2+ channel toxins that specifically inhibit the N- and P/Q-types of Ca2+ channels. As a result, we found that the nerve terminals in this preparation contain either P/Q-(64.3% agatoxin IVA-sensitive) or N-type Ca2+ channels (31.1% conotoxin-GVIA-sensitive) or Ca2+ channels resistant to both toxins (4.6%). Interestingly, the responses to mGluR7 activation with high concentrations of the agonist l-(+)-2-amino-4-phosphonobutyrate (l-AP4) were almost exclusively found in terminals containing N-type Ca2+ channels (95.4%). The remaining responses (4.6%) were located in terminals that express P/Q-type channels. The highly specific distribution of mGluR7 in nerve terminals that express N-type Ca2+ channels explains the absence of P/Q-type Ca2+ channel modulation associated with mGluR7-mediated inhibition of glutamate release in this preparation. Furthermore, these results reveal the existence of specific targeting mechanisms that ensure that mGluR7 and N-type Ca2+ channels co-exist in the same subset of nerve terminals. Synaptosomal Preparation—The cerebral cortex was isolated from adult male Wistar rats (2–3 months), and the synaptosomes were purified on discontinuous Percoll (Amersham Biosciences) gradients as described previously (24Dunkley P.R. Jarvie P.E. Heath J.W. Kidd G.J.E. Rostas J.A.P. Brain Res. 1986; 372: 115-129Google Scholar, 25Wang J.K.T. Walaas S.I. Sihra T.S. Aderem A. Greengard P. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 2253-2256Google Scholar). Following the last centrifugation step at 22,000 × g for 10 min, the synaptosomes were resuspended in 8 ml of HEPES buffer medium (HBM; 140 mm NaCl, 5 mm KCl, 5 mm NaHCO3, 1.2 mm NaH2PO4, 1 mm MgCl2, 10 mm glucose, and 10 mm HEPES, pH 7.4), and the protein content was determined by the Biuret method. Finally, 1 mg of the synaptosomal suspension was diluted in 8 ml of HBM and spun at 3,000 × g for 10 min. The supernatant was discarded, and the pellet containing the synaptosomes was stored on ice. Under these conditions, the synaptosomes remain fully viable for at least 4–6 h as judged by the extent of KCl- and 4-aminopyridine (4AP)-evoked glutamate release. Glutamate Release—Glutamate release was assayed by on-line fluorometry. Synaptosomal pellets were resuspended in HBM (0.67 mg/ml) and preincubated at 37 °Cfor1hinthe presence of 16 μm bovine serum albumin that served to bind any free fatty acids released from synaptosomes during the preincubation. A 1-ml aliquot was transferred to a stirred cuvette, and the release of glutamate was measured as described previously (26Nicholls D.G. Sihra T.S. Sánchez-Prieto J. J. Neurochem. 1987; 49: 50-57Google Scholar). The Ca2+-dependent release was calculated by subtracting the release obtained during a 5-min period of depolarization at 200 nm free [Ca2+] from the release at 1.33 mm CaCl2. Ca2+Imaging Showing Ca2+Responses in Single Synaptosomes— Synaptosomes in HBM (2 mg/ml) with 16 μm bovine serum albumin were preincubated with 5 μm fura-2 AM and 1.33 mm CaCl2 for 40 min. The synaptosomal suspension was attached to a polylysine-coated coverslip for another hour. Synaptosomes were illuminated alternately at 340 and 380 nm for 0.8 s through a 100× objective with the aid of a monochromator (Kinetic Imaging Ltd., Brombrough, UK), and the fluorescence emitted from the nerve terminals was collected through a band-pass filter centered at 510 nm. The video images were obtained using a slow scan CCD camera (Hamamatsu C4880) operating at 12-bit digitalization (4096 levels), and the output from the camera was stored by the computerized imaging system Lucida 3.0 (Kinetic Imaging, Ltd.). Ca2+ images were analyzed as described previously (11Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 14092-14101Google Scholar, 12Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 47796-47803Google Scholar). For [Ca2+]cyt measurements, synaptosomes were stimulated by 10-s pulses (indicated in each graph by a bar) of 30 mm KCl or 1 mm 4AP in the absence and presence of pharmacological agonists or Ca2+ channel toxins. Antibody Controls—The affinity-purified rabbit polyclonal antibody against mGluR7a and the affinity-purified guinea pig polyclonal antibodies against mGluR4a, mGluR7a, and mGluR8a used here have been described elsewhere (27Shigemoto R. Kinoshita A. Wada E. Nomura S. Ohishi H. Takada M. Flor J.P. Neki A. Abe T. Nakanishi S. Mizuno N. J. Neurosci. 1997; 17: 7503-7522Google Scholar). The monoclonal anti-synaptophysin antibody is commercially available (Sigma), and the rabbit polyclonal antibodies against the vesicular glutamate transporters 1 and 2 (VGLUT1 and VGLUT2) were obtained from Synaptic Systems (Göttingen, Germany). The rabbit anti-glutamate decarboxylase (GAD-65) polyclonal antisera were purchased from Chemicon International (Temecula, CA). The rabbit anti-α1B (Cav2.2, N-type) and anti-α1A (Cav2.1, P/Q-type) subunits of voltage-dependent calcium channels were from (Alomone Laboratories, Jerusalem, Israel). As control for the immunochemical reactions, primary antibodies were omitted from the staining procedure, whereupon no immunoreactivity resembling that obtained using the specific antibodies was detected. Immunohistochemical Procedures—The synaptosomes were allowed to attach to polylysine-coated coverslips for an hour and then fixed for 5 min in 4% paraformaldehyde in 0.1 m phosphate buffer, pH 7.4. Following several washes with 0.1 m phosphate buffer, pH 7.4, the synaptosomes were pre-incubated in 10% normal goat serum diluted in 50 mm Tris buffer (pH 7.4) containing 0.9% NaCl (TBS) with 0.2% Triton X-100, for 1 h. They were then incubated for 24 h with the appropriate primary antisera diluted in TBS with 1% normal goat serum and 0.2% Triton X-100 as follows: mGluR7a (1 μg/ml); mGluR4a (1 μg/ml); mGluR8a (1 μg/ml); VGLUT1 (1:1000); VGLUT2 (1:1000); GAD-65 (1:1000); synaptophysin (1:1000); α 1B (Cav2.2, N-type) (1:200); and α 1A (Cav 2.1, P/Q-type) (1:200). After washing in TBS, the synaptosomes were incubated for2hina goat anti-rabbit antibody coupled to the cyanine-derived fluorochrome Cy2, a goat anti-rabbit antibody coupled to the cyanine-derived fluorochrome Cy3, a goat anti-mouse antibody coupled to Cy3, or a donkey anti-guinea pig antibody coupled to fluorescein. The secondary antibodies were diluted 1:200 in TBS (Jackson ImmunoResearch, West Grove, PA). After several washes in TBS, the coverslips were mounted with Fluoromount (Serva, Germany) and the synaptosomes were viewed with a Nikon Diaphot microscope equipped with a 100× objective, a mercury lamp light source, and fluorescein-rhodamine Nikon filter sets. Chemicals—ω-Conotoxin-GVIA (ω-CgT-GVIA) and ω-agatoxin-IVA (ω-Aga-IVA) were supplied by Peptide Institute, Inc. (Osaka, Japan). l-AP4 was from Tocris Cookson (Bristol, UK). Fura-AM was from Molecular Probes (Eugene, OR). NADP+, glutamate dehydrogenase, and all other reagents were from Sigma. The Actions of l-AP4 in Cerebrocortical Nerve Terminals from Adult Rats Are Mediated by mGluR7—We have reported previously that the preparation of cerebrocortical nerve terminals from young rats is highly enriched in markers of glutamatergic terminals such as VGLUT1 and VGLUT2 (12Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 47796-47803Google Scholar). To determine whether a change in neurotransmitter content occurs with development, we have also characterized the preparation of nerve terminals from adult rats. Thus, in this preparation, immunochemistry was performed with antisera against VGLUT1 and VGLUT2 to label glutamatergic terminals against GAD-65, used as a marker of the GABAergic terminals, and against the vesicular protein synaptophysin. Among the particles that contained synaptophysin (2,405 particles from eight fields), 59.5 ± 4.5% and 20.3 ± 2.4% (mean ± S.E.; n = 8) were immunoreactive for VGLUT1 (Fig. 1, A–C) and VGLUT2 (Fig. 1, D–F), respectively. In this preparation, the GABAergic terminals amounted to 23.8 ± 2.0% (n = 7) (Fig. 1, G–I). Therefore, it appears that, in a similar manner to the preparation from young rats, ∼80% of the nerve terminals of the cerebrocortical preparation from adult rats are glutamatergic. We demonstrated previously that the cerebrocortical nerve terminals from adult rats express mGluR7 (11Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 14092-14101Google Scholar). To establish whether other l-AP4-sensitive mGluRs are also present in this preparation, nerve terminals were double-labeled with antisera against synaptophysin and the l-AP4 sensitive mGluRs (subtypes 4, 7, and 8) expressed in the brain. Among the particles that contained synaptophysin (1,312 particles from eight fields), 29.1 ± 3.5% were immunopositive to mGluR7a (mean ± S.E.; Fig. 2, A–C), whereas only 1.2 ± 0.6% (2,017 particles from 10 fields) and 1.7 ± 0.6% (1,229 particles from 10 fields) were immunoreactive for mGluR4a and mGluR8a, respectively (Fig. 2, D–I). These data indicate that the responses to l-AP4 in the preparation of cerebrocortical nerve terminals from adult rats are almost exclusively mediated by mGluR7. l-AP4 Only Inhibits the N-type Ca2+Channel-coupled Release Component in Adult Rats—The Ca2+-dependent release of glutamate from isolated nerve terminals (synaptosomes) can be evoked by the addition of the K+-channel blocker 4AP, which induces tetrodotoxin-sensitive depolarizations (28Tibbs G.R. Barrie A.P. Van-Mieghem F.J.E. McMahon H.T. Nicholls D.G. J. Neurochem. 1989; 53: 1693-1699Google Scholar). In the adult rat preparation of cerebrocortical nerve terminals, this release was inhibited by high concentrations of the group III mGluR agonist l-AP4 through the activation of mGluR7 (11Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 14092-14101Google Scholar). Because this inhibition of release is the result of a reduced influx of Ca2+ into the nerve terminals (11Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 14092-14101Google Scholar), we sought to determine which type of Ca2+ channel was being inhibited by the receptor. To this end, glutamate release was measured in the presence of toxins that specifically inhibited the different Ca2+ channel types. The Ca2+-dependent release of glutamate after 5 min of depolarization with 4-aminopyridine was 3.90 ± 0.14 nmol of glutamate/mg (±S.E.; n = 5). Blocking N-type Ca2+ channels with ω-CgT-GVIA and P/Q type Ca2+ channels with ω-Aga-IVA reduced this release to 2.73 ± 0.15 (n = 4) and 1.06 ± 0.21 (n = 4) nmol/mg, respectively (Fig. 3A). In the absence of toxins, l-AP4 reduced the release to 2.93 ± 0.10 (n = 5) nmol/mg. The presence of this mGluR agonist did not further reduce the release when co-applied with ω-CgT-GVIA (2.65 ± 0.11, n = 4, nmol/mg). However, in the presence of ω-Aga-IVA, l-AP4 produced an additional decrease in glutamate release to 0.33 ± 0.15 nmol/mg (n = 4, Fig. 3B), indicating that the mGluR agonist only acted upon the component of glutamate release coupled to N-type Ca2+ channels. Thus, l-AP4 and ω-CgT-GVIA inhibited release to a similar extent as did the two Ca2+ channel toxins added together (Fig. 3B). l-AP4 Does Not Inhibit the P/Q-type Ca2+Channel-coupled Release Component Even at Low [Ca2+]o—In other preparations, it has been shown that l-AP4-sensitive mGluRs can inhibit P/Q-type Ca2+ currents (9Takahashi T. Forsythe I.D. Tsujimoto T. Barnes-Davies M. Onodera K. Science. 1996; 274: 594-597Google Scholar, 29Perroy J. Prezeau L. De Waard M. Shigemoto R. Bockaert J. Fagni L. J. Neurosci. 2000; 20: 7896-7904Google Scholar). Thus, it seems unlikely that the failure of l-AP4 to inhibit the P/Q-type channel-coupled release is due to the inability of the receptor to modulate this type of Ca2+ channel. Alternatively, it is possible that the inhibition of P/Q channel-coupled release is being masked by the high efficiency with which P/Q channels drive glutamate release (17Qian J. Noebels J.L. J. Neurosci. 2001; 21: 3721-3728Google Scholar, 18Millán C. Sánchez-Prieto J. Neurosci. Lett. 2002; 330: 29-32Google Scholar). If this were the case, although depressed, the Ca2+ current at the release site controlled by P/Q channels would still be sufficient to support the release of the vesicular pool associated with this channel population. The existence of a differential coupling of release to P/Q- and N-type channels was revealed by the different sensitivity of the release components to changes in the extracellular concentration of Ca2+, [Ca2+]o. Thus, Fig. 4 shows how the relative contribution of N-type channels to glutamate release is reduced by decreasing [Ca2+]o and that it is virtually abolished at 0.5 mm [Ca2+]o. Indeed, at such low [Ca2+]o, the release of glutamate is fully governed by P/Q-type Ca2+ channels. However, at 0.5 mm [Ca2+]o, the mGluR agonist l-AP4 failed to significantly reduce the release of glutamate (2.84 ± 0.16 and 2.78 ± 0.19 in the presence and absence of l-AP4, respectively; n = 3; Fig. 5, A and B). This suggests that the high efficiency coupling of P/Q channels to glutamate release is not responsible for the inability of l-AP4 to inhibit this release component.Fig. 5l-AP4 fails to reduce the release component associated to P/Q-type Ca2+ channels at low [Ca2+]o.A, the Ca2+-dependent release of glutamate evoked by 1 mm 4AP was determined in the absence (control) and presence of 1 mm l-AP4 at 0.5 mm [Ca2+]o. B, bar diagrams show the glutamate release after 5 min of depolarization with 1 mm 4AP in the presence and absence of 1 mm l-AP4. Results are means ± S.E. of three experiments obtained from the same number of synaptosomal preparations.View Large Image Figure ViewerDownload (PPT) Nerve Terminals from the Cerebral Cortex of Adult Rats Contain Either N- or P/Q-type Ca2+Channels—The Ca2+ channels that control glutamate release are distributed heterogeneously in nerve terminals (17Qian J. Noebels J.L. J. Neurosci. 2001; 21: 3721-3728Google Scholar, 21Wheeler D.B. Randall A. Tsien R.W. J. Neurosci. 1996; 16: 2226-2237Google Scholar, 22Reid C.A. Clements J.D. Bekkers J.M. J. Neurosci. 1997; 17: 2738-2745Google Scholar). These terminals may contain N-type, P/Q-type, or a combination of both types of Ca2+ channels (12Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 47796-47803Google Scholar). Thus, another possible explanation of the inability of l-AP4-sensitive mGluRs to modulate the P/Q-type Ca2+ channel is that these two proteins do not co-localize in the same nerve terminal. To determine whether or not this may be the case, we analyzed the Ca2+ images obtained from single terminals loaded with fura-2. In a typical field (as shown in Fig. 6A), individual particles were abundant, although clusters were also observed. The particles ranged from 0.5 to 1.2 μm in diameter and were typically circular, displaying strong fluorescence when loaded with fura-2 (Fig. 6A). A total of 1,060 fura-2-load particles that responded to 4AP and KCl from four fields similar to that shown in Fig. 6A were analyzed, and the basal [Ca2+]cyt was seen to be 110.4 ± 1.1 nm (n = 30). More than 95% of the fura-2-loaded particles responded to a 10-s application of 4AP at 1 mm or 30 mm KCl. The Ca2+ responses provoked by depolarization with 4AP and KCl were transient, ranging from 600 to 800 nm, but the responses to 4AP lasted far longer than those by KCl. In addition, the Ca2+ responses evoked by 4AP were completely abolished by the Na+-channel blocker tetrodotoxin (TTx), whereas those induced by KCl were insensitive to TTx (Fig. 6B). The sensitivity of the 4AP-induced Ca2+ responses to TTx indicates the existence of Na+ channels that mediate repetitive firing in synaptosomes, consistent with data showing that the great majority of the particles obtained in the Percoll-purified synaptosomal preparation are nerve terminals (24Dunkley P.R. Jarvie P.E. Heath J.W. Kidd G.J.E. Rostas J.A.P. Brain Res. 1986; 372: 115-129Google Scholar). To determine the type of Ca2+ channels involved in the Ca2+ responses in this preparation, the intensity of the fura-2 signals was determined in the presence of toxins that specifically block the different Ca2+ channel subtypes. For these experiments, ω-CgT-GVIA was used to block N-type channels, ω-Aga-IVA to block P/Q type channels, and both toxins together were also applied. A total of 3,803 nerve terminals from seven fields were analyzed. In a subpopulation of nerve terminals (31.1 ± 1.4%), the Ca2+ responses were largely reduced by ω-CgTGVIA but not ω-Aga-IVA (Fig. 7A), indicating that these nerve terminals contain N-type channels. In another subpopulation of nerve terminals (64.3 ± 2.4%), the Ca2+ responses were inhibited by ω-Aga-IVA but not ω-CgT-GVIA, indicating the presence of P/Q-type channels (Fig. 7B). Finally, a minor subpopulation of terminals (4.6 ± 1.3%) showed responses that were largely insensitive to both toxins (Fig. 7C). No Ca2+ responses mediated by a combination of N and P/Q channels were found in this preparation, in contrast to that reported in young rats where a large subpopulation of nerve terminals (42.6%) contained a mix of both types of Ca2+ channels (12Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 47796-47803Google Scholar). mGluRs with Low Affinity for l-AP4 Are Exclusively Located in Terminals That Express N-type Ca2+Channels—To determine which type of Ca2+ channels are expressed in the nerve terminals where l-AP4 reduces the Ca2+ response, we analyzed the effect of this mGluR agonist on fura-2-loaded particles in the presence of the Ca2+ channels toxins. A total of 4,323 fura-2-loaded particles obtained from six fields were analyzed. Interestingly, it was found that in the great majority of the nerve terminals (95.4 ± 1.0%) wherein a reduced Ca2+ response in the presence of l-AP4 was exhibited, the Ca2+ response was also blocked by ω-CgT-GVIA but not ω-Aga-IVA (Fig. 8A). Only in a minority of the nerve terminals that responded to l-AP4 (4.6 ± 0.6%) were the Ca2+ responses blocked by ω-Aga-IVA (Fig. 8B). To further assess the preferential localization of mGlu7 receptors in nerve terminals expressing N-type Ca2+ channels, nerve terminals were double-labeled with antisera against mGluR7 and the α1B (Cav2.2, N-type) and α1A (Cav2.1, P/Q-type) subunits of voltage-dependent calcium channels. Among the nerve terminals immunopositive to mGluR7 (351 terminals from seven fields) 94.1 ± 1.5% also expressed N-type Ca2+ channels (Fig. 9, A–C), whereas only 5.5 ± 0.8% expressed mGluR7 and the P/Q-type of Ca2+ channels (384 terminals from seven fields) (Fig. 9, D–F). These results indicate that mGluR7 and N-type Ca2+ channels coexist in the same nerve terminals, therefore providing an explanation for the absence of P/Q-type Ca2+ channel modulation by this receptor in glutamate release. By using Ca2+ imaging and immunocytochemistry in single nerve terminals from the cerebral cortex of adult rats, we have characterized the mGluRs that control the release of glutamate in this preparation and the Ca2+ channel subtypes modulated by these receptors. Here we show that, although the major component of Ca2+ influx and glutamate release is associated with P/Q-type channel activation, mGluR7 is selectively expressed in N-type Ca2+-channels containing nerve terminals wherein it inhibits the release component associated with these channels. Ca2+Channels in Nerve Terminals Change with Development—In nerve terminals from adult animals, the role that Ca2+ channels play in the control of Ca2+ entry matches that which they play in glutamate release, suggesting that most of the Ca2+ channels are coupled to the release process. Thus, N- and P/Q-type Ca2+ channels mediate Ca2+ entry in 31.1 ± 1.4 and 64.3 ± 2.4% of the nerve terminals, respectively, and these same Ca2+ channels support 29.9 ± 3.8 and 72.7 ± 5.4% of total glutamate release. These data are consistent with the predominant influence of P/Q channels in supporting glutamate release when compared with that of N-type Ca2+ channels in several regions from the brain (17Qian J. Noebels J.L. J. Neurosci. 2001; 21: 3721-3728Google Scholar, 19Luebke J.I. Dunlap K. Turner T.J. Neuron. 1993; 11: 895-902Google Scholar). However, this parallel between the role of Ca2+ channels in Ca2+ influx and in glutamate release does not exist in young rats. In preparations from young rats, Ca2+ entry is controlled by N-type, P/Q-type, or both types of Ca2+ channels in 47.5, 3.9, and 42.6% of the nerve terminals, respectively (12Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 47796-47803Google Scholar). In contrast, in these terminals N- and P/Q-type Ca2+ channels support 33.6 ± 2.1 and 62.1 ± 2.5% of the total glutamate release, respectively (18Millán C. Sánchez-Prieto J. Neurosci. Lett. 2002; 330: 29-32Google Scholar). These data indicate that, during development, there is an important redistribution of the Ca2+ channel subtypes in nerve terminals. The first developmental change we have observed is the disappearance in adult animals of the subpopulation of terminals with a mix of N and P/Q channels that is seen in young animals. The second developmental change is that, despite the prominent role of N-type channels in the control of Ca2+ entry in young rats, their contribution to glutamate release is similar to that found in adult animals. These last data are also consistent with observations from studies of synaptic transmission in pyramidal neurons from the visual cortex wherein no changes are observed in the contribution of N-type channels to glutamate release during development (30Iwasaki S. Momiyama A. Uchitel O.D. Takahashi T. J. Neurosci. 2000; 20: 59-65Google Scholar). Therefore, the general decline in the contribution of N-type channels to synaptic transmission during development observed at many synapses (30Iwasaki S. Momiyama A. Uchitel O.D. Takahashi T. J. Neurosci. 2000; 20: 59-65Google Scholar, 31Scholtz K.P. Miller R.J. J. Neurosci. 1995; 15: 4612-4617Google Scholar) does not seem to apply to cortical synapses. Because Ca2+ imaging detects Ca2+ influx in both glutamatergic and GABAergic terminals, the marked decline in the contribution of N channels to GABAergic neurotransmission during development (30Iwasaki S. Momiyama A. Uchitel O.D. Takahashi T. J. Neurosci. 2000; 20: 59-65Google Scholar) may explain, at least in part, the differences in the role exerted by N-type channels on Ca2+ entry and glutamate release in young animals. Alternatively, the lack of parallelism between Ca2+ entry and glutamate release during early developmental stages may reflect the fact that some N-type Ca2+ channels are not coupled to glutamate release but rather to other processes such as cell migration (32Komuro H. Rakic P. Science. 1993; 260: 95-97Google Scholar) or synaptogenesis (33Vigers A.J. Pfenninger K H. Dev. Brain. Res. 1991; 60: 197-203Google Scholar). Interestingly, these same channels seem to be lost in mature animals. A similar developmental change has been observed in the Calyx of Held. In this study, synaptic transmission was supported by P/Q-, N- and R-type Ca2+ channels in young animals (34Wu L.-P. Westenbroek R.E. Borst J.G.G. Caterall W.A. Sakmann B. J. Neurosci. 1999; 19: 726-736Google Scholar); however, the contribution of the Ca2+ channels that are less efficient in driving release (N- and R-type channels) was lost during development. Consequently, in older animals synaptic transmission was only supported by P/Q channels (30Iwasaki S. Momiyama A. Uchitel O.D. Takahashi T. J. Neurosci. 2000; 20: 59-65Google Scholar). The Group III mGluRs That Control Glutamate Release Also Change with Development—In addition to the redistribution of Ca2+ channels in nerve terminals, developmental changes in the function of group III mGluRs also occurs. Thus, whereas the inhibition of glutamate release in cerebrocortical nerve terminals from young rats is mediated by mGluR4 and mGluR7, (12Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 47796-47803Google Scholar), release inhibition in adult terminals is primarily mediated by the low affinity, l-AP4-sensitive mGluR7 (11Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 14092-14101Google Scholar). This is consistent with immunocytochemical studies wherein a restricted expression of mGluR4 in the cerebral cortex of adult animals has been described (35Corti C. Aldegheri L. Somogyi P. Ferraguti F. Neuroscience. 2002; 110: 403-420Google Scholar). Moreover, functional studies in CA1 from the hippocampus have shown that the robust inhibition of synaptic transmission by low concentrations of l-AP4 found in young animals virtually disappears in adult animals (36Baskys A. Malenka R.C. J. Physiol. 1991; 444: 687-701Google Scholar). During development, a profound remodeling in presynaptic terminals occurs. This remodeling affects the population of nerve terminals co-expressing N and P/Q type channels and harboring the majority of the mGluR4 receptors (12Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 47796-47803Google Scholar). In addition, mGluR7, which is largely distributed in N (70%) and in N and PQ termini (30%), becomes almost exclusively restricted to N termini (96%) in adult animals. Thus, the redistribution of Ca2+ channels in terminals together with developmental changes in group III mGluR expression may contribute to a remodeling of presynaptic modulation. Co-localization of mGluR7 and N-type Ca2+Channels—The glutamate receptor, mGluR7, has been shown to reduce the P/Q-type Ca2+ currents in cerebellar granule cells (29Perroy J. Prezeau L. De Waard M. Shigemoto R. Bockaert J. Fagni L. J. Neurosci. 2000; 20: 7896-7904Google Scholar). In cerebrocortical nerve terminals from young rats, mGluR7 also inhibits P/Q Ca2+ currents (12Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 47796-47803Google Scholar) as well as the release component associated with these channels (18Millán C. Sánchez-Prieto J. Neurosci. Lett. 2002; 330: 29-32Google Scholar). However, activating mGluR7 failed to modulate the release coupled to P/Q channels in terminals from adult animals, because the receptor rarely (4%) co-localizes with this type of Ca2+ channel. In fact, mGluR7 was almost exclusively found in nerve terminals expressing N-type Ca2+ channels (96%), consistent with the ability of this receptor to reduce N-type Ca2+ currents and, consequently, the associated glutamate release (11Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 14092-14101Google Scholar). The localization of mGluR7 in nerve terminals containing N-type Ca2+ channels does not seem to be the result of a random distribution of the receptor. Terminals with N-type Ca2+ channels represent only a minor subset of the nerve endings in the cerebrocortical preparation from adult rats. In fact, the majority of the terminals in this preparation contain P/Q-type Ca2+ channels, as seen both by imaging Ca2+ and the predominant role of these Ca2+ channels in the control of glutamate release (4Herrero I. Vázquez E. Miras-Portugal M.T. Sánchez-Prieto J. Eur. J. Neurosci. 1996; 8: 700-709Google Scholar, 11Millán C. Luján R. Shigemoto R. Sánchez-Prieto J. J. Biol. Chem. 2002; 277: 14092-14101Google Scholar). The specific co-localization of mGluR7 and N-type channels therefore requires the existence of precise mechanisms of targeting to place both proteins in the same nerve terminal. The correct localization of synaptic proteins seems to be the result of a two-step process, i.e. an initial targeting to an axonal or somatodendritic domain based on intrinsic sorting signals followed by clustering at presynaptic and postsynaptic sites that depend on cell-to-cell contacts (37Craig A.M. Banker G. Annu. Rev. Neurosci. 1994; 17: 267-310Google Scholar). In this context, molecular determinants involved in axonal targeting have been identified in the C-terminal of mGluR7 (38Stowell J.N. Craig A.M. Neuron. 1999; 22: 525-536Google Scholar), which might explain the presynaptic location of mGluR7 seen in vivo (27Shigemoto R. Kinoshita A. Wada E. Nomura S. Ohishi H. Takada M. Flor J.P. Neki A. Abe T. Nakanishi S. Mizuno N. J. Neurosci. 1997; 17: 7503-7522Google Scholar). However, additional mechanisms would be required to explain the nearly exclusive localization of mGluR7 in N-type Ca2+ channel-containing nerve terminals. The target cell also contributes to a specific expression of synaptic proteins. Thus, in the hippocampal CA1 region, pyramidal cell terminals presynaptic to mGluR1a-expressing interneurons have at least a 10-fold higher level of presynaptic mGluR7 than terminals making synapses with pyramidal cells and other types of interneurons (39Shigemoto R. Kulik A. Robert J.D.B. Ohishi H. Nusser Z. Kaneko T. Somogy P. Nature. 1996; 381: 523-525Google Scholar). Similarly, in hippocampal slices it has been found that a mGluR sensitive to l-AP4 reduces synaptic transmission only at nerve terminals contacting CA1 interneurons (40Scanziani M. Gahwiler B. Charpak S. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 12004-12009Google Scholar). Thus, it is likely that the specific colocalization of mGluR7 and N-type Ca2+ channels is the result of the axonal distribution of mGluR7 (38Stowell J.N. Craig A.M. Neuron. 1999; 22: 525-536Google Scholar) in conjunction with a target cell-mediated stabilization of the receptor in the subset of terminals bearing N-type Ca2+ channels. A similar mechanism of targeting seems to control the distribution of the N-type Ca2+ channel subunits at nerve terminals. Recently, it has been shown that the final distribution of presynaptic N-type Ca2+ channels in rat hippocampal neuronal cultures depends on neuronal contacts and synapse formation (41Maximov A. Bezprozvanny I. J. Neurosci. 2002; 22: 6939-6952Google Scholar). Thus, in immature neurons the α1B-1 (Cav2.2a) splice variant of the N-type Ca2+ channel is diffusely distributed, but in mature neurons it is exclusively recruited to a presynaptic location by means of interactions with the modular adaptor proteins Mint1 and CASK in mature neurons (41Maximov A. Bezprozvanny I. J. Neurosci. 2002; 22: 6939-6952Google Scholar). When analyzing the distribution of mGluR7 and Ca2+ channels within nerve terminals, we not only found that most mGluR7 (96%) is in terminals that contain N-type Ca2+ channels, but also that most nerve terminals that contain N-type Ca2+ (>85%) also express mGluR7. It is therefore likely that the synaptic targeting of the two proteins is coordinated or shares a common mechanism. In this respect, C-terminal mediated interactions with PDZ domain adaptor proteins are involved in the targeting of mGluR7 and N-type Ca2+ channels at presynaptic terminals. The interaction of mGluR7 with the adaptor protein PICK1 has been shown to be important both for clustering of this receptor at presynaptic terminals (42Boudin H. Doan A. Xia J. Shigemoto R. Huganir R.L. Worley P. Craig A.M. Neuron. 2000; 28: 485-497Google Scholar) and for receptor function (43Dev K.K. Nakajima Y. Kitano J. Braithwaite S.P. Henley J.M. Nakanishi S. J. Neurosci. 2000; 20: 7252-7257Google Scholar, 44Perroy J. El Far O. Pin J.-P. Betz H. Bockaert J. Fagni L. EMBO J. 2002; 17: 2990-2999Google Scholar). Synaptic clustering of N-type Ca2+ channels also depends on interactions with PDZ and SH3 domain binding motifs of the adaptor proteins Mint1 and CASK, respectively, (45Maximov A. Sudhof T.C. Bezprozvanny I. J. Biol. Chem. 1999; 274: 24453-24456Google Scholar). In addition, protein kinase C phosphorylates mGluR7 (46Nakajima Y. Yamamoto T. Nakayama T. Nakanishi S. J. Biol. Chem. 1999; 24: 27573-27577Google Scholar) and N-type Ca2+ channels (47Zamponi G.W. Bourinet E. Nelson D. Nargeot J. Snutch T.P. Nature. 1997; 385: 442-446Google Scholar). Finally, because both mGluR7 (38Stowell J.N. Craig A.M. Neuron. 1999; 22: 525-536Google Scholar) and N-type Ca2+ channels (48Bennet M.K. Calakos N. Scheller R.H. Science. 1992; 257: 255-259Google Scholar) are located at the active zone, and, given the functional coupling between mGluR7 and N-type Ca2+ channels, it seems likely that these two proteins form part of the same presynaptic multiprotein complex. We thank M. Sefton for editorial assistance.
Referência(s)