Characterization of two alternatively spliced forms of a metabotropic glutamate receptor in the central nervous system of the rat

Neuroscience Vol. 60, No. 2, pp. 325 336, 1994

~

Pergamon

0306-4522(94)E0014-U

Elsevier ScienceLtd Copyright © 1994 IBRO Printed in Great Britain. All rights reserved 0306-4522/94 $7.00 + 0.00

CHARACTERIZATION OF TWO ALTERNATIVELY SPLICED FORMS OF A METABOTROPIC GLUTAMATE RECEPTOR IN THE CENTRAL NERVOUS SYSTEM OF THE RAT D. R. HAMPSON,*t E. THERIAULT,~ X.-P. HUANG,* P. KRISTENSEN,§D. S. PICKERING,* J. E. FRANCK¶ and E. R. MULVIHILLI] *Faculty of Pharmacy, University of Toronto, 19 Russell Street, Toronto, Ontario, Canada M5S 2S2 :[:Playfair Neuroscience Unit, Toronto Western Hospital, Toronto, Canada §CNS Division, Novo Nordisk A/S, Malov, Denmark ¶Department of Neurological Surgery, University of Washington, Seattle, Washington, U.S.A. r4ZymoGenetics Inc., Seattle, Washington, U.S.A. Abstract--Amplification of complementary DNA by the polymerase chain reaction and anti-peptide antibodies were used to characterize the expression of two alternatively spliced forms of a metabotropic glutamate receptor (mGluRlct and mGluRlfl) in the central nervous system of the rat. Polymerase chain reaction analysis showed that mGluRl~ was the predominate of the two forms in the cerebellum, diencephalon, mesencephalon, olfactory bulb and brainstem, while mGluRlfl was the major form present in the hippocampus. Approximately equal amounts of the two receptors were expressed in the cerebral cortex, septum and striatum. Immunochemical analyses of the two receptors were conducted in the rat cerebellum and hippocampus. An mGluRl~t-specific antibody labelled a protein with a relative molecular weight of 146,000 on immunoblots of the hippocampus and cerebellum. Immunoblot analysis of the developmental expression of mGluRlct in the hippocampus and cerebellum demonstrated that in both structures, the levels of mGluRlct were at or near their maximum levels in the adult brain. In contrast, two mGluR 1fl-specific antibodies failed to detect mGluR I fl on immunoblots of brain tissue, thus precluding an immunocytochemical analysis of this receptor. Although low levels of a higher-molecular weight protein, possibly a dimeric form of mGluRlfl were seen with one of the mGluRlfl-specific antibodies, we hypothesize that some of the mGluRlfl present in brain tissue may undergo proteolytic cleavage of the carboxy terminus. Immunocytochemical analysis of mGluR 1ct showed that very high levels of this receptor were expressed in Purkinje cell bodies and dendrites. In the granule cell layer, some Golgi neurons were immunostained. The granule cells were not labelled. In the hippocampus, mGluRlct immunoreactivity was present in interneurons of the stratum oriens and the dentate hilar region. Double-labelling studies demonstrated that these interneurons were also immunopositive for the neuropeptide somatostatin. The presence of mGluRlct in cells of the hippocampus that are associated with the release of somatostatin, suggest that this receptor could play a role in regulating hippocampal excitability in both normal and epileptic tissues.

The identification of a family of metabotropic glutamate receptors (mGluRs) by expression cloning ~°'~2and homology screening 1' 14.24and the subsequent functional characterization of these cloned receptors has demonstrated that members of this receptor group are coupled to several second-messenger systems including phospholipase C and adenylyl cyclase.L 2. 17 The elucidation of the nucleotide and amino acid sequences of the m G l u R s has also provided the opportunity to produce subtype specific probes that can be used to study the structures and distributions of these receptors in detail. These studies may, in turn, provide further insight into another tTo whom correspondence should be addressed. Abbreviations: BHK, baby hamster kidney cells; InsP, phos-

phoinositol; mGluR, metabotropic glutamate receptor; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; SDS-PAGE, sodium dodecylsulphate polyacrylo amide gel electrophoresis. 325

relatively unknown aspect of mGluRs, namely the characterization of the biological roles or functions of these molecules in the CNS. The first m G l u R to be cloned has been termed m G l u R 1~.~°' J: Subsequently, several additional alternatively spliced forms of m G l u R l ~ have been identified. In the rat CNS the m G l u R l ~ polypeptide is composed of 1199 amino acids with a calculated molecular weight of 133,000. The alternatively spliced forms, termed mGluRlf124 and m G l u R l c ~9 are truncated versions of m G l u R l ~ with modified carboxy termini. The amino acid sequences of the three proteins are identical from the amino terminus to asparagine 887 where the sequences diverge. Thus the carboxy termini are different in each of the variants; m G l u R l ~ has 313, m G l u R l f l has 20 and m G l u R l c has l0 amino acids that are not present in the other alternatively spliced forms of mGluR1. It has been demonstrated that both m G l u R l f l ~7

326

D.R. HAMPSONel al.

a n d m G l u R I c ~9 can couple to the phosphoinositol (InsP) signal t r a n s d u c t i o n pathway. However, several differences in the properties o f these alternatively spliced forms c o m p a r e d to m G l u R l ~ were observed. m G l u R l ~ a n d mGluR1/3 expressed in baby h a m s t e r kidney cells ( B H K 570 cells) displayed differences in the subcellular distributions, agonist affinities and pertussis toxin sensitivities, ~7 while m G l u R l c showed a different pattern of intracellular calcium release c o m p a r e d to m G l u R l ~ . L~ These results suggest that the three receptors may be associated with different physiological functions in the nervous system. The distribution of m R N A for m G l u R l in the rat C N S has been analysed via in situ hybridization h i s t o c h e m i s t r y / l High levels o f m G l u R l m R N A were reported to be present in m a n y areas of the C N S including in the cerebellum, h i p p o c a m p u s , olfactory bulb and thalamus. The reported distribution of m G l u R 1 7 receptor protein has been determined using an anti-peptide a n t i b o d y raised against the carboxy terminus of m G l u R l ~ . It A l t h o u g h the pattern of i m m u n o s t a i n i n g generally corresponded to the distribution of the m G l u R 1 m R N A , several i m p o r t a n t exceptions to this pattern were observed. F o r example, in the h i p p o c a m p u s , the m R N A was a b u n d a n t in the granule, pyramidal and n o n - p y r a m i d a l cells, whereas the expression of the mGluRlc~ receptor protein appeared to be restricted primarily to n o n - p y r a m i d a l cells in the hilus and stratum oriens. In the present study, we have examined the expression a n d distribution of mGluRlc~ and m G l u R l / t in the rat CNS using polymerase chain reaction ( P C R ) amplification of m R N A and receptor subtypespecific antibodies directed against peptides corresponding to a unique sequence in the carboxy termini. Because of the discrepancies in the reported distributions of the m R N A and the i m m u n o s t a i n i n g patterns of m G l u R l ~ noted above, we have also re-examined the distribution of m G l u R 1~ i m m u n o r e activity using a n a n t i b o d y raised against a different epitope on m G l u R l ~ t h a n was used previously. EXPERIMENTAL PROCEDURES

Polvmerase chain reaction analysis

Two/lg of polyA ~ RNA purified from various regions of rat brain were used in each cDNA synthesis reaction, cDNA was prepared using Moloney murine leukemia virus reverse transcriptase primed with 20 pmol of downstream primer (see below) in a reaction mixture containing 45 mM Tris (pH 8.3), 68mM KC1, 15mM dithiothreitol, 9 m M MgC1 z, 0.08 mg/ml bovine serum albumin and 1.8 mM of each deoxynucleotide for 60min at 3T'C. The entire cDNA reaction was used for PCR amplification after diluting threefold with water and the addition of a total of 60 pmol of each primer (i.e. an additional 40 pmol of the downstream primer and 60pmol of the upstream primer). The downstream primer (5' AGGCCGTCTCGTTGGTCTTCA) corresponds to base pair (bp) 3126 in mGluR 1c~and bp 3211 in mGluR 1,6; the upstream primer (5" CCTGGGGTGCATGTTTACTCC) corresponds to bp 2854 in both forms. Two units of Taq polymerase were added after denaturation and equilibration at the annealing temperature. Amplification was carried out as follows: 40 cycles at 95, 56 and 7 2 (1 min each).

Production and puri[4cation o[" antibodies

Because of the homology that exists between various members of the mGluR family, careful consideration must be given to the choice of peptide sequences for the generation of subtype-specific antibodies. Thus for antibody production, peptides of unique sequences near the carboxy terminus of mGluRl~ (A4 antibody, sequence= EFVYEREGNTEEDEL) or at the carboxy terminus of mGluR1/~ (GS1 antibody : KKPGAGNAKKRQPEFS; GS2 antibody, sequence = PEFSPSSQCPSAHAQL) were used as immunogens. The peptides were purified by highperformance liquid chromatography and the structures were confirmed by mass spectrometry. Antisera were raised in rabbits by immunizing them with peptides conjugated to keyhole limpet hemocyanin by crosslinking with gluteraldehyde (100 200 mg injected subcutaneously). Injections were made every three weeks and antisera were collected 10 days after the third injection. The A4 antibody was affinity purified from crude antiserum using the A4 peptide and the ProtOn kit no. 1 from Multiple Peptides Systems (San Diego, CA). The GS1 antibody was affinity-purified by conjugating the GSI peptide to AH-amino Sepharose (Pharmacia) using carbodiimide (BioRad) and the GS2 antibody was affinity-purified by conjugating the GS2 peptide to thiopropyl Sepharose (Pharmacia). Electrophoresis and immunoblotting

Sodium dodecylsulphate polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting procedures were carried out as described previously by Hampson et al? A goat anti-rabbit IgG conjugated to alkaline phosphatase (Promega) was used as the secondary antibody. Brain tissue samples were prepared for SDS-PAGE by homogenizing freshly dissected tissues in SDS sample buffer (62raM Tris-HC1, pH 6.8, containing 2% SDS and 10% glycerol) with dithiothreitol added to a final concentration of 100 raM. All samples were heated at 100°C for 2 min prior to SDS-PAGE. Crude membrane preparations from BHK 570 cells expressing mGluRl~ or mGluR1/~ were prepared for SDS-PAGE in the same manner. The production and characterization of these cell lines has been described by Pickering et al. ~7 For quantitative analyses of the developmental expression of mGluRlc~, the immunoblots were scanned with a microcomputer imaging device from Imaging Research Inc., St Catherines, Ontario. lmmunocvtochemistry

Wistar rats (200 300 g; Charles River Inc., Montreal) were deeply anaesthetized with sodium pentobarbital and perfused transcardially with 400 ml of cold 0.12 M phosphate buffer, pH 7.2, followed by 500-700 ml of cold 4% paraformaldehyde in phosphate buffer. The brains were then immersed in fixative for 2-3 h and subsequently placed in phosphate-buffered saline (PBS) overnight at 4°C. Sections (20-30/~m) were cut on a Vibratome. After rinsing in PBS, the tissue sections were incubated in blocking buffer (PBS containing 2% bovine serum albumin and 1% normal goat serum) for 1 h followed by incubation with purified antibodies diluted in blocking buffer for 16 h at 4~C and l h at 22'~C. In some experiments, the sections were treated with 0.1% Triton X-100 for 1 h or overnight to permeabilize the tissue. After incubation with primary antibody, the sections were washed three times in PBS, incubated with biotinylated anti-rabbit secondary antibody (Vectastain Kit, Vector Laboratories) diluted in blocking buffer for 2 h at 25°C, washed three times in PBS and then incubated with the Vectastain reagent for 1 h. After washing three times in PBS, the sections were incubated with diaminobenzidine, washed two times with PBS and then incubated with 1% gelatin in 25% ethanol for 5 rain prior to mounting. After air drying, the sections were dehydrated and defatted using ethanol and xylene. For peptide blocking controls, the primary

327

Distribution of metabotropic glutamate receptors antibodies were incubated with the tissue sections (or immunoblots) in the presence of the appropriate peptide (1 #g/ml). For double-labellingstudies, biotinylated goat anti-rabbit IgG (Vector Laboratories; diluted 1:500 in blocking buffer) and avidin-Texas Red (Vector; 1:25) were used to detect the mGluRl~t antibody. After elution of these antibodies with 0.2 M glycine HCl buffer, pH 2.4, the sections were incubated with the anti-somatostatin antibody (SS 309 diluted 1:1000, a gift from Dr R. Benoit). Biotinylated Protein-A (1:400) and avidin AMCA (1:25, both from Vector) were used to detect the anti-somatostatin primary antibody.

were present in approximately equal amounts in the cerebral cortex, septum and striatum. Although it may be difficult to accurately quantitate the absolute amount of nucleic acids via PCR, we believe that these results are an accurate representation of the relative amounts of these mRNAs since identical amounts of mRNA purified by the same method were used and the same primers were used to amplify both RNAs.

Antibody specificity and immunoblot analyses RESULTS

Polymerase chain reaction analysis Analysis of the amplified products from the PCR reactions showed the presence of two bands of 378 and 294 bp in all of the brain regions examined (Fig. 1). These bands correspond to the expected sizes of the PCR products for mGluRlfl and mGluRlct respectively. The smaller protein, mGiuRlfl, is represented by a larger PCR product because the mGluRlfl mRNA has an 85 bp in-frame insertion at an intron/exon splice site?4 All regions examined showed the presence of both RNAs. The mGluRlct was the predominant form found in the cerebellum, diencephalon, mesencephalon, olfactory bulb and brainstem, while mGluRlfl RNA was the predominant form expressed in the hippocampus. Both RNAs

a

b

c

de

Antibodies directed against unique amino acid sequences at (mGluRlfl) or near (mGluRlct) the carboxy termini of the two receptors were made by injecting rabbits with peptides conjugated to keyhole limpet hemocyanin. Recombinant receptors expressed in BHK 570 cells were used to demonstrate the lack of cross-reactivity of the mGluRlct-specific antibody (A4) for mGluRlfl and the lack of cross-reactivity of the mGluRlfl antibodies (GS1 and GS2) for mGluRlct (Fig. 2). Moreover, based on an alignment of reported sequences of the other mGluRs (mGluRIc and mGluR2~i), it appears extremely unlikely that the A4, GS1 and GS2 antibodies would cross-react with other mGluRs. The anti-mGluRl~-specific antibody labelled a protein with a relative molecular weight of 154,000 on immunoblots of BHK 570 cells expressing mGluRlct

f

g

h

i

653 394 298 154

Fig. 1. PCR amplification of mGluRl~ and mGluRlfl cDNA from rat brain. One-fifth of each PCR reaction was separated on a 3% agarose gel. The upper band (378 bp) in lanes b-j corresponds to mGluRlfl and the lower band (294 bp) corresponds to mGluRlcc Lane a, molecular size markers (bp); b, cerebral cortex ; c, cerebellum; d, hippocampus; e, diencephalon; f, mesencephalon; g, striatum; h, septum; i, olfactory bulb; j, brainstem.

j

328

D.R. HAMPSONet al.

(Fig. 2a). A major immunoreactive protein at 146,000 mol. wt was observed in the cerebellum; this band was also seen in the hippocampus, but at a much lower intensity than the corresponding band in the cerebellum. In addition to the 146,000 mol.wt protein, the mGluRl~ antibody also labelled two less intense bands at 159,000 and 130,000 mol. wt in the cerebellum and hippocampus. Although the identity of these bands is unknown, the higher band could be a more heavily glycosylated form of mGluRl~, while the lower band may be a proteolytic product. All of the bands were eliminated when the blots were incubated in the presence of the A4 peptide (data not shown). Both of the mGtuR1/t-specific antibodies (GS1 and GS2) labelled a 96,000 mol. wt protein in the BHK 570 cells expressing mGluRlfl (Fig. 2b, c). In addition to the mGluRlfl monomer at 96,000mol. wt, both mGluRlfl antibodies also labelled a 186,000 mol. wt protein in the BHK 570 cells expressing mGluRlfl. Although a reducing reagent (dithiothreitol, 100 mM) was present in these samples, it is possible that the 186,000mol. wt protein is a mGluRl/3 dimer since it is immunoreactive with both of the mGluR 1fl-specific antibodies and its calculated molecular weight is approximately double that of the mGluRlfl monomer. Although the mGluR 1/~-specific antibodies labelled mGluRlfl on immunoblots of transfected cells, these antibodies failed to detect the mGluRlfl monomer in brain tissue (Fig. 2b, c). However, in the hippocampus, faint labelling of a protein that comigrated with the proposed mGluRlfl dimer was seen with the GS2 antibody. The GS2 antibody also labelled a band at 78,000 mol. wt in the hippocampus (Fig. 2b, lane 3); this band was not labelled with the GS1 antibody. The labelling of the 96,000 and 186,000mol. wt bands by GS1 and GS2 on immunoblots of transfected BHK 570 cells and the 78,000 and 186,000mol. wt bands labelled by the GS2 antibody in the hippocampus were blocked by incubating the blots in the presence of the appropriate inhibitory peptide (data not shown). The reason for the inability to detect the mGluRlfl monomer in brain tissue is not known, but it may be related to proteolytic processing of the receptor (see further discussion below). The inability to detect mGluRlfl on immunoblots of brain tissue samples precluded further immunocytochemical analyses of mGluRlfl.

Developmental expression of metabotropic glutamate receptor l~t Tissue samples of rat hippocampus and cerebellum obtained from animals at various postnatal ages were prepared for SDS-PAGE and immunoblot analyses. In these experiments, two individual sets of tissue samples from two animals at each age were analysed. The mGluRl~ band at 146,000mol. wt was quantirated by densitometric scanning of the immunoblots. In the cerebellum, the expression of mGluRl~ was very low on postnatal day 1 and increased steadily until adulthood (Fig. 3a). The levels of m G l u R l a were at least 20-fold higher in the adult rat cerebellum compared to postnatal day 1. In the hippocampus, the levels of mGluRl~ were substantially lower compared to the cerebellum and a slightly different pattern of expression was observed. Peak levels of expression occurred at about two weeks postnatal and then appeared to remain constant or decline slightly into adulthood (Fig. 3b).

Immunocytochemical analys& t~f metabotropic glutamate receptor 1 On tissue sections of the cerebellum, intense staining of Purkinje cell bodies and dendrites was observed with the mGluRl~ antibody (Fig. 4a). Pretreatment of the tissue sections with Triton X-100 greatly enhanced the staining in the molecular layer (Fig. 4b). Immunostaining was also observed in some Golgi neurons in the granule cell layer, although the staining intensity was not as great as that seen with Purkinje neurons. lmmunostainingwith the A4 antibody was eliminated by incub~ltingtissue sections of the cerebellum (Fig. 4c) or the hippocampus (data not shown) in the presence of the A4 peptide. No specific staining of the granule cell bodies was observed on tissue sections (Fig. 4a, b) or on immunoblots of cultured granule cells with the mGluRlct antibody (unpublished observation). The mGluR 1~t-specific antibody showed a restricted distribution of immunoreactivity on tissue sections of the rat hippocampus. Intense staining was observed in cells of the stratum oriens of CA1 (Fig. 5a, b), where immunoreactive cell bodies a~adfibers in this layer were distributed along the entire CA 1 area from the subiculure to area CA3. In addition to the cells in the stratum oriens, a population of cells in the hilus were also labelled (Fig. 5c, d). The distribution of the cells labelled by the mGluRl~ antibody in the hippocampus resembled a

Fig. 2. Immunoblots showing labelling with mGluRla- specific antibody (a) and the mGluRlfl-specific antibodies (b, c). Each lane contains 2(V30#g of protein. (a) A4 antibody: lane 1, mGluRl~t expressed in BHK 570 cells; lane 2, mGluRlfl in BHK 570 cells; lane 3, hippocampus; lane 4, cerebellum. (b) GS2 antibody and (c) GSI antibody: lane 1, BHK 570 cells expressing mGluRlfl; lane 2, mGluRl~t in BHK 570 cells; lane 3, hippocampus; lane 4, cerebellum. The asterisk indicates the position of the proposed mGluRlfl dimer. Molecular weight markers indicated in each panel are myosin (200,000mol. wt), phosphorylase b (97,000 tool. wt) and bovine serum albumin (67,000tool. wt).

4~h

D.R. HAMPSONel al.

330 2O "~ °v,,l

15

o-1 io 5

o

2 4 9 1216202842Adult

Age (days) Fig. 3. Quantitativc immunoblot analysis of the postnatal expression of mGluRla in the rat cerebellum and hippocampus. Tissue samples were obtaincd on the days indicated, subjected to SDS-PAGE, transferred to nitrocellulose, labelled with the mGluRl~ antibody and scanned with a microcomputer imaging device. (a) Histogram of blots of rat cerebellum;each lane contained 30/~g protein; (b) hippocampal samples; each lane contained 50 #g of protein. Data are the averages + S.E.M. from three to four immunoblots using two separate sets of tissue samples. population of interneurons that contain somatostatin. 23Therefore, we attempted to ascertain whether or not the mGluRl~-positive cells were somatostatinergic neurons. The pattern of labelling seen with the mGluRlc~ antibody in both the stratum oriens (Fig. 6a) and the hilus was also observed with antisomatostatin antibody (Fig. 6b). Double-labelling studies using the mGluRl~ antibody, the antisomatostatin antibody and fluorescence-labelled second antibodies confirmed that the cells in the stratum oriens and the hilus that were immunostained with the m G l u R l a antibody, were also specifically labelled with the anti-somatostatin antibody (Fig. 6c, d). DISCUSSION

Analysis o f metabotropic glutamate receptor 1/3 in the central ner~'ous system One of the major findings of this study was the observation that there appears to be a discrepancy

between the levels mGluRlfl mRNA and the levels of mGluR1/3 protein in the rat CNS. In our PCR experiments, we were able to amplify cDNAs of the predicted sizes for mGluRl~ and mGluRl/3 from all brain areas examined. However, in immunoblot analyses using antibodies directed against unique amino acid sequences in the carboxy termini of mGluRl~ and mGluRl/3, we were able to consistently detect mGluRl~ protein in several regions of the brain, while we were unable to detect a protein corresponding to mGluRl/3. The results of the PCR analysis demonstrated that the hippocampus possesses relatively high levels of mGluRlfl mRNA and lower amounts of m G l u R l e mRNA. The indication that the hippocampus contains relatively little mGluRl~ was supported by our immunoblot and immunocytochemical analyses. Since we detected mGluRl~ on immunoblots of the hippocampus, it follows that sufficient mGluRl// should be present to detect it on immunoblots. In the mGluR1/3 protein, only the carboxy terminal 20 amino acids are unique to this protein; the preceding 886 amino acids are also present in mGluRl~ and lc. Thus the possibilities, in terms of generating mGluRl/3specific antibodies, are limited. We produced two antibodies against peptide sequences spanning this carboxy terminal region and showed that they specifically recognize mGluRlfl, but not mGluRl~, on immunoblots of transfected cell membranes. The robust labelling of mGluR1/3 on immunoblots of transfected cells indicates that the inability to detect a protein in the same region of the gel (approximately 100,000 tool. wt) in the brain samples was not due to the to the inability of the two antibodies to recognize the denatured receptor on immunoblots. One potential explanation for this discrepancy may be that a portion of the mGluRl/3 protein undergoes proteolytic processing of its carboxy terminus. This processing could be in the form of proteolytic cleavage of the carboxy terminal residues or post-translational modification of these residues resulting in the lack of antibody recognition. Direct-support for this idea stems from recent studies using antibodies raised against the amino terminal domain ofmGluR1. On immunoblots of rat brain tissues, prominent expression of a 96,000-100,000 tool. wt protein in various areas of the rat CNS including the hippocampus was observed using this antibody. 5 It is possible that this protein may be mGluRIc or a proteolytic product of mGluRl~ and/or mGluRl/3. Since recent studies of recombinant truncated versions of mGluRl~ and lc have indicated that an intact carboxy terminus is not required for functional activity (see Refs 7. 18 and

Fig. 4. mGluRl~ immunoreactivity in the cerebellum. (a) Intense immunostaining of Purkinje neurons. (b) Higher-power magnification of a tissue section treated with Triton X-100. Note that after treatment with Triton X-100, the staining in the molecular layer intensifieswhile the staining in the granule cell layer decreases. (c) Peptide blocking control. Arrowheads in a and b indicate immunoreactive Golgi cells. G, granule cell layer, M, molecular layer. Scale bars = 350 #m (a), 250 ~m (b) and 600 pm (c).

I

Z

?ig. 5. lmmunostaining of the hippocampus using the mGluRlzt antibody. (a) Low-power magnification of the hippocampus, arrowhead denotes extensive staining of neuronal cell bodies md processes in the stratum oriens): (b) higher magnification of the immunoreactivity in the stratum oriens of CAI revealing the band of mediolaterally oriented stained processes: medium c} and high Idt magnifications of the cells and processes in the hilus. G. granu]e cell layer: H. hi l,us: PY, pyramidal cell layer: SO. stratum oriens. Scale bars = 5001Lm (a), 200ltm (b, c) and

-v

t.3

Distribution of metabotropic glutamate receptors Pickering and Hampson, unpublished observations), it is possible that processed versions of mGluR1 in the CNS may have important physiological roles. Although no protein of the expected molecular weight was detected with either of the mGluRlflspecific antibodies in brain samples, one of these antibodies (GS2) did label a faint band that co-migrated with the postulated mGluRlfl dimer. Evidence indicating that this protein may be a mGluRlfl dimer includes: (i) it was labelled by the mGluRIB-specific antibody; (ii) it was present in both transfected BHK 570 cells and rat brain tissue; and (iii) the relative molecular weight on SDS-PAGE (186,000 mol. wt) was about twice that of the mGluRlfl monomer (96,000 mol. wt). The reason for the inability of the other mGluRl~-specific antibody, GS1, to label the 186,000 mol. wt protein in the hippocampus is not known, although the GS1 antibody did label this protein in transfected BHK 570 cells. The observation that the high-molecular weight form persists in

333

the presence of SDS and a reducing reagent indicates that if this protein is in fact a dimer, the monomeric components must be tightly coupled. Conceivably, the dimeric form ofmGluRlfl could be more resistant to proteolytic cleavage; this may explain why we detected an immunoreactive band at 186,000 mol. wt but not the 96,000 mol. wt monomer in the hippocampus. Studies with other G-protein-coupled receptors such as muscarinic acetylcholine3 and serotonin ~5 receptors have also provided evidence for the existence of receptor dimers.

lmmunoblot analyses and developmental expression of metabotropic glutamate receptor lot The mGluRlct-specific antibody labelled a 146,000 mol. wt protein in the rat cerebellum; this band, albeit at a much lower intensity, was also observed in the hippocampus. The lower molecular weight of mGluRlc( expressed in brain (146,000 mol. wt) compared to that observed in BHK 570 cells

Fig. 6. Co-localization of mGluRl~t and somatostatin immunoreactivity in the hippocampus. (a) Immunostaining of mGluRl~ in area CA 1 showing clearly stained cell bodies and dendrites lying within a dense network of thinly varicose, presumably axonal processes. (b) Somatostatin-positive cells in CA1. Note that only the cell bodies and the occasional primary dendrite are stained. Fluoresence photomicrographs of a tissue section of the hilus showing double-labelling of cells and processes with the mGluR1ct. antibody (c), and the labelling of cell bodies with an anti-somatostatin antibody (d). A, alveus; SO, stratum oriens. Scale bars = 80/tm (a, b) and 40#m (c, d).

334

D.R. HAMPSONel al.

(154,000mol. wt) and the fact that both relative molecular weights were higher than the predicted molecular weight (133,000mol. wt), could be explained by differential post-translational processing (e.g., glycosylation). The expression of mGluRl~ in the rat cerebellum showed a progressive but gradual increase from birth through to adulthood. In the hippocampus, peak levels occurred around the second week after birth and subsequently remained constant or showed a slight decrease in adulthood. In both structures the pattern of protein expression did not correspond to the pattern of glutamate-stimulated InsP turnover: glutamate-stimulated lnsP turnover apparently reaches peak levels at about days 6 and 9 in the rat cerebellum and hippocampus, respectively, and subsequently declines to very low levels in the adult brain. ~6 Our findings indicate that peak levels of mGluRl~ receptor protein are present in the adult brain when minimal mGluR-mediated lnsP hydrolysis is seen. A similar finding was reported by Condorelli et al., ~ who examined mGluR17 mRNA levels in rat brain. Although the reason for the lack of correspondence in the receptor levels and second messenger production is not clear, it is possible that the diminution in glutamate-stimulated InsP hydrolysis may be due to a reduction in, or an uncoupling from, one of the other downstream components of the system, for example a specific G-protein or phospholipase C isotype. Alternatively, it is also possible that the glutamate-stimulated InsP hydrolysis in these brain regions is mediated by another mGluR subtype with a different developmental profile. Distribution o f metabotropie glutamate receptor 1~ in the cerebellum

The most intense immunoreactivity on tissue sections observed in the present study was the labelling of the Purkinje neurons with the mGluRl~ antibody. Both the cell bodies and the dendritic trees were densely immunoreactive with the mGluR 1~ antibody. Our results with the mGluRla-specific antibody in the cerebellum and hippocampus concur with the distributions reported by Martin et al., L~ who examined the distribution of mGluRl7 using an antibody raised against a different epitope in the carboxy terminus of mGluRl~. Studies using in situ hybridization techniques to investigate the distribution of mGluRl mRNA in the rat CNS have reported that very high levels of mRNA are present in Purkinje cells, with lower levels in the granule cell layer and in scattered cells in the molecular layer. 12'2J In these studies, an EcoRI-SacI fragment of mGluRl~ was used as the probe. Since the sequences of mGluRlc~, mGluRl/~ and mGluRlc are identical in this region, the probe probably hybridized to all three receptor mRNAs. Results from our immunocytochemical experiments indicate the high levels of mRNA hybridization observed in Purkinje cell bodies and dendrites, is in fact due primarily

to mGluRl~, while the moderate levels of hybridization seen in the granule cell layer and in stellate/basket cells in the molecular layer could be due to the presence of mGluRl/~ and/or mGluRlc. The observations that high levels of mGluRl~ and the InP receptor ~ are expressed in Purkinje cells, combined with the observation that mGluRlc~ efficiently couples to the lnP system in Xenopus oocytes "~'~'~and BHK 570 cells. ~7 provides further evidence that the lnP system is a major signal transduction pathway for mGluRl~ in Purkinje cells. Metabotropic glutamate receptor distributions in the hippocampus

In the hippocampus, mGluRl~ immunoreactivity was confined to two relatively small groups of interneurons in the hilus and stratum oriens. The distribution of mGluRl~ that we observed in these regions resembles the distribution of a subpopulation of inhibitory interneurons that are immunopositive for somatostatin. ~ The results of our double-labelling experiments confirmed that the cells in the hilus and stratum oriens that possess mGluRl~ receptors also contain somatostatin. The mGluRl~ receptor was present in cell bodies, axons and some dendrites, while somatostatin immunoreactivity was confined primarily to cell bodies. In addition to our results demonstrating the colocalization of mGluRl~ and somatostatin, it has also been reported that somatostatin-containing neurons in the hilus express calcium-permeable kainate receptors. 8 Although these cells appear to be non-GABAergic interneurons whose physiological function is not entirely clear, the possible co-localization of both mGluRl~ and kainate receptors on somatostatin-containing neurons could have important implications for some types of epileptic seizures. For example, Vezzani et al. 25 have shown that the levels of somatostatin in the hippocampus are elevated after kindling; they have postulated that this increase may reflect an endogenous compensatory mechanism to control neuronal hyperexcitabitity. The possibility that mGluRl~ and ionotropic glutamate receptors are present on these cells suggests that glutamate receptors may be involved in controlling the release of somatostatin from these interneurons. However, the co-localization of these two glutamate receptors, both of which have the capability of increasing intracellular calcium levels, could have deleterious consequences for these neurons in terms of an increased susceptibility to calcium-induced degeneration. Data from both animal models6'22 and human studies2° have demonstrated that the somatostatin-containing neurons in the hippocampus are exceptionally susceptible to degeneration induced by epileptic seizures. Moreover, the finding that the seizure-induced degeneration of these cells apparently precedes the degeneration of pyramidal neurons22 indicates that these interneurons may play a critical role in the etiology of epilepsy. In the

Distribution of metabotropic glutamate receptors light of these observations, further studies on the potential relationship between the activation of glutamate receptors and the release of somatostatin and other neuropeptides appears warranted. CONCLUSIONS Messenger R N A coding for m G l u R l f l was detected in all brain regions examined including the hippocampus. Despite the relatively high levels of m G l u R l f l m R N A in this structure, we were unable to detect intact monomeric m G l u R l f l protein on immunoblots of either the hippocampus or the cerebellum using two different mGluRlfl-specific antibodies whose epitopes spanned the unique amino acid sequence in the carboxyl terminus of this receptor. We hypothesize that the carboxyl terminus of m G l u R l f l may be post-translationally modified, possibly by proteolytic cleavage.

335

Immunocytochemical analyses with an m G l u R l ~ specific antibody demonstrated that very high levels of this receptor are present in cerebellar Purkinje neurons. In the hippocampus, high levels of mGluRl~ immunoreactivity were present in interneurons of the hilus and the stratum oriens/alveus. Double-labelling experiments demonstrated that these cells also contained somatostatin immunoreactivity. The presence of metabotropic glutamate receptors on neurons containing somatostatin may be important in the generation and control of hippocampal seizure activity. Acknowledgements--The authors thank Drs A. Baskys,

P. S. Pennefather, R. Sloviter, R. J. Wenthold and J. M. Wojtowicz for helpful comments. This work was supported by grants to D.R.H. from the Natural Sciences and Engineering Research Council of Canada. D.S.P. was supported by a Postdoctoral Fellowship from NSERC.

REFERENCES

1. Abe T., Sugihara H., Nawa H., Shigemoto R., Mizuno N. and Nakanishi S. (1992) Molecular characterization of a novel metabotropic glutamate receptor mGluR5 coupled to inositol phosphate/Ca signal transduction. J. biol. Chem. 267, 13,361-13,368. 2. Armori I. and Nakanishi S. (1992) Signal transduction and pharmacological characteristics ofa metabotropic glutamate receptor in transfected CHO cells. Neuron 8, 757-765. 3. Chidiac P. and Wells J. W. (1992) Effects of adenyl nucleotides and carbachol on cooperative interactions among G-proteins. Biochemistry 31, 10,908-10,921. 4. Condorelli D F., Dell'Albani P., Amico C., Casabona G., Genazzani A., Sorting A. and Nicoletti F. (1992) Developmental profile of metabotropic receptor mRNA in rat brain. Molec. Pharmac. 41,660~64. 5. Fotuhi M. Sharp A. H., Glatt C. E., Hwang P. M., von Krosigk M., Snyder S. H. and Dawson T. M. (1993) Differential localization of phosphoinositide-linked metabotropic glutamate receptor (mGluR1) and the inositol 1,4,5-triphosphate receptor in rat brain. J. Neurosci. 13, 2001-2012. 6. Freund T. F., Ylinen A., Miettinen R., Pitkanen A. Lahtinen H., Bainbridge K. G. and Riekkinen P. J. (1991) Pattern of neuronal death in the rat hippocampus after status epilepticus. Relationship to calcium binding protein content and ischemic vulnerability. Brain Res. Bull. 28, 27-38. 7. Gabellini N., Manev R. M., Candeo P., Favaron M. and Manev H. (1993) Carboxyl domain of glutamate receptor directs its coupling to metabolic pathways. NeuroReport 4, 531-534. 8. Gibbons S. J. and Miller R. J. (1992) Somatostatin containing dentate hilar neurons in primary culture express Ca permeable kainate receptors. Soc. Neurosci. Abstr. 613.11. 9. Hampson D. R., Wheaton K. D., Dechesne C. J. and Wenthold R. J. (1989) Identification and characterization of the ligand binding subunit of a kainic acid receptor using monoclonal antibodies and peptide mapping. J. biol. Chem. 264, 13,329-13,335. 10. Houamed K. M., Kuijper J. L., Gilbert T. L., Haldeman B. A., O'Hara P. J., Mulvihill E. R., Almers W. and Hagan F. S. (1991) Cloning, expression and gene structure of a G protein-coupled glutamate receptor from rat brain. Science 252, 1318-1321. 1I. Martin L. J., Blackstone C. D., Huganir R. L. and Price D. L. (1992) Cellular localization ofa metabotropic glutamate receptor in rat brain. Neuron 9, 259-270. 12. Masu M., Tanabe Y., Tsuchida K., Shigemoto R. and Nakanishi S. (1991) Sequence and expression of a metabotropic glutamate receptor. Nature 349, 760 765. 13. Nakagawa T., Shiota C., Okano H. and Mikoshiba K. (1991) Differential localization of alternative spliced transcripts encoding 1,4,5-triphosphate receptors in mouse cerebellum and hippocampus: an in situ hybridization study. J. Neurochem. 57, 1807-1810. 14. Nakajima Y., Iwakabe H., Akazawa C., Nawa H., Shigemoto R., Mizuno N. and Nakanishi S. (1993) Molecular charaterization of a novel retinal metabotropic glutamate receptor mGluR6 with high agonist selectivity for L-2-amino-4phosphonobutyrate. J. biol. Chem. 268, 11868-11873. 15. Ng G. YK., George S. R., Zastawny R. L., Caron M., Bouvier M., Dennis M. and O'Dowd B. F. (1993) Human serotonint Breceptor expression in Sf9 cells : phosphorylation, palmitoylation and adenylyl cyclase inhibition. Biochemistry 32, 11,727-11,733. 16. Palmer E., Nangel-Taylor K., Krause J. D., Roxas A. and Cotman C. W. (1990) Changes in excitatory amino acid modulation of phosphoinositide metabolism during development. Devl Brain Res. 51, 132-134. 17. Pickering D. S., Thomsen C., Suzdak P. D,, Fletcher E. J., Robitaille R., Salter M. W., MacDonald J. F., Huang X.-P. and Hampson D. R. (1993) Characterization of two alternatively spliced forms of a metabotropic glutamate receptor. J. Neurochem. 61, 85-92. 18. Pin J.-P., Ahern S. and Jolly C. (1993) Metabotropic glutamate receptors : differences from other G-protein coupled receptors. J. Neurochem. 61, Suppl. S117D.

336

D.R. HAMPSONel u/.

19. Pin J.-P., Waeber C., Prezeau L., Bockaert J. and Heinemann S. F. (1992) Alternative splicing generates metabotropic glutamate receptors inducing different patterns of calcium release in Xenopus oocytes. Proc. nam. Aead. Sci. 89, 10,331 10~335. 20. Robbins R. J., Brines M. L., Kim J. H., Adrian T., deLanerolle N., Welsh S. and Spencer D. D. (1991) A selective loss of somatostatin in the hippocampus of patients with temporal lobe epilepsy. Ann. Neurol. 29, 325 332. 21. Shigemoto R., Nakanishi S. and Mizuno N. (1992) Distribution of the mRNA for a metabotropic glutamate receptor (mGluRl) in the central nervous system: an in situ hybridization study in adult and developing rat. J. comp. Neurol. 322, 121 135. 22. Sloviter R. S. (1987) Decreased hippocampal inhibition and a selective loss of interneurons in experimental epilepsy. Science 235, 73-76. 23. Sloviter R. S. and Nilaver G. (1987) lmmunocytochemical localization of GABA-, cholecystokinin-, vasoactive intestinal polypeptide- and somatostatin-like immunoreactivity in the area dentata and hippocampus of the rat. J. comp. Neurol. 256, 42 60. 24. Tanabe Y., Masu M., Ishii T., Shigemoto R. and Nakanishi S. (1992) A family of metabotropic glutamate receptors. Neuron 8, 169 179. 25. Vezzani A., Monno A., Rizzi M., Galli A., Barrios M. and Samanin R. (1992) Somatostatin release is enhanced in the hippocampus of partially and fully kindled rats. Neuroscience 51, 41 46. (Accepted 22 Not,ember 1993)