Glycine receptor immunoreactivity in rat and human cerebral cortex
Elke Naas 1, Karl Zilles 1, Hannes Gnahn 2., Heinrich Betz 3, Cord-Michael Becker 3 and Hannsj6rg Schr6der 1
llnstitut 1 flir Anatomie, Universitiit zu Ki~ln, Cologne (ER.G.), 2Neurologische Klinik, Technische Universitlit, Miinchen (ER.G.) and 3 Zentrum flir M olekulare Biologie, Universitat Heidelberg, Heidelberg ( E R. G. )
The distribution of the inhibitory glyeine receptor was studied in rat and human cerebral cortex using a monoclonal antibody (MAb 4a) directed against the ligand-binding subunit. Significant amounts of glycine receptor antigen were found in forebrain structures such as caw datum and neocortex, although cortical levels were significantly below those seen in spinal cord. Immunohistoehemically, glyeine receptors were preferentially localized to the apical dendrites of pyramidal neurons in layers III and V. Ultrastructurally, these sites corresponded to synaptic neuronal contacts. Immunoreactivity was found in neuronal perikarya, dendrites and postsynaptic membranes which may corre- spond to sites of intracellular synthesis, transport and membrane incorporation of the glycine receptor. These immunological data corrobo- rate previous pharmacological studies suggesting the existence of glycinergic transmission in mammalian cerebral cortex.
INTRODUCTION
Glycine is an es tabl ished inhibi tory neuro t ransmi t te r
in mammal ian spinal cord and bra ins tem 3'22. The firing
of spinal motoneurons is reduced upon release of glycine
from interneurons , result ing from chlor ide-dependent
synaptic inhibi t ion 22. lmmunocytochemica l studies have
shown that spinal motoneurons bear strychnine-sensit ive
inhibi tory glycine receptors 9 (GlyRs) preferent ia l ly at
postsynaptic membranes which are opposed to presyn-
A t t empt s have been made to discover glycinergic sys- tems at supramedul la ry levels 2'3'8'22'29"35'37. Recent ly , us-
ing immunocytochemis t ry van den Pol and G6rcs 37 dem-
ons t ra ted the presence of glycine and a GlyR-assoc ia ted
93 k D a prote in 3° in ra t olfactory bulb. However , only
sparse informat ion is available about glycinergic neuro-
transmission in the cerebra l cortex. Moreover , recep tor
au torad iography using [SH]strychnine fai led to reveal
glycinergic antagonist binding sites both in rat a9 and hu- man 28 cerebra l cortex.
Topical appl icat ion of the giycinergic antagonist strych-
nine onto the cerebra l cortex is known to produce con-
vulsions 15"35. The strychnine concentrat ions used, how-
ever, are beyond the margins of selectivity of this
receptor antagonist and thus, the involvement of inhib-
i tory GlyRs vs. y-aminobutyr ic acid ( G A B A ) receptors
is still a subject of controversy 1°'35. Also , Arak i and co-
workers 4 using a monoclonal an t ibody failed to detect
distinct cortical immunoreact iv i ty against the 93 k D a
G l y R protein.
In this repor t we invest igated selected neocort ical ar-
eas of man and ra t using the m o n o d o n a l ant ibody ( M A b )
4a which is d i rec ted against the l igand-binding a subunit of the GIyR 6'25'26'33. Our da ta provide convincing evi-
dence for the existence of GlyRs in rat and human ce-
rebral cortex.
MATERIALS AND METHODS
Tissue preparation Human cortical specimens were obtained either at the occasion
of neurosurgical tumor extirpation from the margins of the tumor infiltration zone (n = 2, (a) male, 27 years, temporal cortex, (b) female, 68 years, temporal cortex) or at autopsy (n = 1, male, 55 years; acute cardiac failure; post mortem-delay (PM): 1 h; occipi- tal cortex). Specimens for immunocytochemistry were fixed by im- mersion in Zamboni's solution 3s for 7-10 days (el. ref. 32). For Western blotting and immunoassay, human autopsy brain (frontal cortex; n = 1, male, 60 years; PM: 6 h; acute cardiac failure) and spinal cord (n = 1, female, 30 years, fatal accident; PM: 6 h) were dissected, and samples were frozen immediately in liquid nitrogen and stored at -70 °(2. Animal studies were conducted on 15 male Wistar rats (200-300 g) whose brains were processed for immuno- cytochemistry as described previously 32. Briefly, rats were perfused via the left ventricle in deep tribromethanol anesthesia with 300 ml
* Present address: Aiblinger Anger 6, D-8017 Ebersberg bei Miinchen, ER.G. Correspondence: H. Schr6der, Institut I fOr Anatomie, Universit~lt K61n, Joseph Stelzmann-Strafle 9, D-5000 K6in 41, ER.G. Fax: (49) (221) 4785318.
140
Zamboni's solution 3a. Brains were removed, postfixed for 2 h and subsequently rinsed in 0.1 M phosphate buffer, pH 7.4 (PB). For the present study, specimens of the occipital cortex were used. For biochemical investigations, animals were decapitated, parietal ce- rebral cortex and spinal cord immediately removed, frozen in liq- uid nitrogen, and stored at -70 °C.
lmmunoperoxidase procedure From both human and rat specimens, cortical vibratome sections
(40-50/~m) were prepared and treated for visualization of immu- noreactive sites as described previously 32. Briefly, sections were preincubated in PB containing 20% normal swine serum (NSS) for 1 h at room temperature (RT). Subsequently, the sections were rinsed in PB (2 x 10 min) and incubated with the GIyR antibody MAb 4a 6'26 for 48-72 h at 4 °(2 (dilution 1:100 in PB containing 1% NSS). Afterwards, sections were washed in PB (2 x 10 min). A biotinylated anti-mouse Ig was used as secondary antibody (1:50, 1 h, RT). After another wash in PB (2 x 10 min), the samples were incubated in a streptavidin-peroxidase complex (1:50, 1 h, RT) followed by a wash in PB. Visualization of the immunopre- eipitate was obtained by immersion of the sections in a solution containing 5 mg/10 mi diaminobenzidine (DAB, Sigma) and 130 /all10 ml 0.3% hydrogen peroxide (Merck) in PB (10-15 min, RT). The reaction was stopped by transferring the specimens in PB. The sections were mounted onto glass slides using Depex as mounting medium. For electron microscopy, sections from 5 rats and of both human biopsy samples were osmicated (2% OsO 4, 15 min), dehy- drated and subsequently flat-embedded in Durcupan. Semithin and ultrathin sections were prepared using a Reiehert ultramicrotome OmU2. For control purposes, sections otherwise treated as de- scribed above were incubated omitting (i) primary antibody or (ii) biotinylated Ig, or (iii) were exclusively incubated with the strepta- vidin-peroxidase complex to exclude unspecific adherence of the detection systems to the sections. Controls revealed negative re- suits. Photomicrographs were taken using an Olympus Vanox pho- tomicroscope and a Zeiss EM 9 S-2.
Preparation of membranes and [ZH]strychnine binding assay Crude membrane fractions were prepared from human and rat
CNS tissue as described TM. Briefly, tissue was homogenized in 50 mM Tris-HCl, pH 7.4, containing protease inhibitors, and washed
twice by centrifugation at 50,000 g for 30 min. Membranes were finally suspended in 25 mM potassium phosphate, pH 7.4, contain- ing 200 mM KCI. Protein contents of the suspended membranes were measured according to Larson et al. 19. Glycine-displaceable [3H]strychnine binding to crude CNS membrane fractions (100/~g protein) was determined in triplicate by filtration assay 5.
Western blotting and dot receptor immunoassay (DORA) Immunoquantitation of GlyR was performed by DORA as de-
scribed previously 7. Briefly, membrane proteins solubilizcd in 0.5% (w/v) deoxycholate in Tris-buffered saline containing 20% (v/v) methanol were adsorbed onto nitrocellulose filters. These sheets were subsequently reacted with MAb 4a at a dilution of 1:100, and antimouse immunoglobulin coupled to horseradish peroxidase (di- luted 1:7,500) 7 . Enzyme activity of triplicate samples was quanti- fied photometrically. Values were corrected for nonspecific anti- body binding as determined using either rat liver membranes which do not contain GlyR as a control or by omitting first antibody. GlyR-immunoreactivity in samples of rat spinal cord exceeded the range of the assay and therefore these samples were used at an ap- propriate dilution. Western blots were conducted using an alkaline phosphatase-coupled second antibody 7.
RESULTS
Biochemistry To characterize the human GIyR antigen, Western
blot analysis was conducted on membrane proteins from the grey m a t t e r o f h u m a n spinal cord. H e r e , M A b 4a
s ta ined a faint band of 48 k D a appa ren t mo lecu l a r
weight . A band o f ident ica l e l ec t rophore t i c mobi l i ty bu t
m o r e in tense s taining b e c a m e appa ren t wi th tissue f rom
the t r igemina l m o t o r nucleus. These po lypep t ides corre-
sponded in e l ec t rophore t i c mobi l i ty to the l igand-bind-
ing (a ) subuni t o f G l y R f r o m rat spinal cord 5'12"25 (Fig.
1A). Wi th cor t ica l m e m b r a n e s , M A b 4a also r evea l ed a
NV CX SC RSC
92.5 m I l l 66.2 - -
45.0
31.0 m d.f.
A
Q 0 5.00O
_>
i ,=1
E
l a
,ID
B
4.000
3.000
2.000
1.00O
0.000
/ / / ; oO ° / / /
Fig. 1. Immunoanalysis of GlyR in crude membrane fractions from human and rat CNS regions. A: Western blot of human membrane preparations from the trigeminal motor nucleus (NV), cerebral cortex (CX), and cervical spinal cord anterior horn (SC). Rat spinal cord membranes (RSC) served as a control. MAb 4a stained a band of 48 kDa (each lane) and, in some of the samples, a proteolytic fragment of 42 kDa (CX, RSC). B: immunoquantitation by DORA of GlyR in human (white boxes) and rat (black boxes) CNS areas. Immunore- activity of MAb 4a with crude membrane fractions is expressed as AOD/10/ag of protein -+ S.E.M. of a triplicate determination; nucl. V, trigeminal motor nucleus.
faint band of 48 kDa, but in addition, a prominent poly- peptide of ~-- 42 kDa. Proteolytic fragments of identical size have been observed previously with tryptic digests 5'12 of the GIyR a subunit. Thus, the lower molecular weight band of the human cortical preparation most likely re- sulted from partial proteolysis due to post mortem delay in freezing the tissue. Subtle differences in a subunit electrophoretic mobility are indicative of developmental GlyR isoforms in rat spinal cord 6. Isoform heterogene- ity might also exist between human brain regions. How- ever, considerable proteolysis of the human samples pre- cluded reliable conclusions on the postmortem material.
GIyR content was quantified in human cervical and total rat spinal co rd by both giycine-displaceable [3H]strychnine binding and DORA employing MAb 4a (Table I). Both methods revealed similar results. In the human, specific GlyR content in the anterior horn of the grey matter accounted for about 15% of the value seen with rat spinal cord. In the pooled intermediary part and dorsal horn of the grey matter, receptor content was about 2- to 3-fold lower whereas no GlyR was detect- able in the ventral white matter (Table I). These obser- vations are compatible with autoradiographic data on GlyR distribution 28.
Comparison of GIyR antagonist binding activity and immunoreactivity of individual membrane preparations has led to identification of a low-affinity receptor isoform expressed during early stages of postnatal development of rodent spinal cord 6. However, the relative binding activity of the GIyR antigen, detected here, i.e. the ra- tio of [3H]strychnine and MAb 4a binding, did not dif- fer significantly between samples from adult rat and hu- man spinal cord (Table I). Thus, the MAb 4a epitope, GIyR ct subunit molecular weight and antagonist bind- ing properties are conserved between human and rodent GlyRs 5'25. MAb 4a-based immunomethods therefore should allow a reliable analysis of GIyR distribution in the human CNS.
Quantitation of GlyR levels by DORA was extended
141
to other human and rat CNS areas (Fig. 1B). Specific immunoreactivity of human membrane samples was high- est with the trigeminal motor nucleus and reached about 30% of the GlyR density determined for rat spinal cord. Whereas GlyR contents were similar in human caudate and spinal cord ventral grey matter, they were lower, but still reliably detectable, in human cortical membranes. Rat cortex likewise contained significant amounts of GlyR antigen (Fig. 1B). These data indicate that GlyRs are present, although at low abundance, in vertebrate cortical and diencephalic areas. A comparison of cortex and spinal cord reveals a caudo-rostral gradient of GlyR antigen in both man and rat. The latter, however, is much steeper in the rodent (see Fig. 1B).
Immunocytochemistry
Light microscopy. Upon immunostaining of tissue sec- tions of rat cortex with MAb 4a, the apical dendrites of pyramidal perikarya in layers III and V were preferen- tially labelled (Figs. 2A, 3A). Immunoreactive dendrites could be followed from their origins at the border be- tween layers V I N and IV/III traversing the superficial laminae perpendicularly to the pial surface and branch- ing at the transition zone of layers IfII. Only a few rather faintly stained perikarya, mainly in layers III and V, were observed in rat (Fig. 3A).
In human cortex, perikaryal immunostaining was more pronounced (Figs. 2B, 3B). Predominantly pyramidal neurons of layers III and V were decorated. Often the zone of origin of the apical dendrites could be seen in pyramidal neurons (Fig. 3B). Besides pyramidal perikarya, a few round or ovoid neuronal perikarya in layers VI and IV were immunoreactive in both species. Axons, however, did not display any labelling in either species.
Electron microscopy. At the ultrastructural level, im- munostained dendrites were the predominant character- istic, often with immunoreactive postsynaptic densities (Figs. 4,5). Immunoprecipitate was also seen in the neu-
TABLE I
Binding of [3H]strychnine and MAb 4a to membrane preparations from human and rat spinal cord
Crude membrane fractions prepared from the indicated areas of human cervical and total rat spinal cord were analyzed for both [3H]strychnine binding and MAb 4a-immunoreactivity. Values represent means of tripficate determinations +- S.E.M. For each membrane preparation the specific binding of radioligand per unit receptor antigen is also shown (ratio A:B). n.d., not detectable; values not above background.
CNS region [3H]Strychnine bound MAb 4a-immunoreactivity Ratio A:B (A) (pmollmg) (B) (AODHO I~g) (fmolldOD)
Human anterior horn Human intermediary and posterior grey matter Human ventral white matter Rat spinal cord
Fig. 2. Survey microphotograph of GlyR-immtmoreactivity in human and rat temporal cortex. MAb 4a-immunostaining. A: rat. As in the human cortex (cf. Fig. 2B), mainly pyramidal neurons are labeled. However, here label is confined almost exclusively to the apical dendrites. B: human. Note predominant labefing of pyramidal perikarya and the proximal portions of the apical dendrites.
143
<.--
Fig . 3. Detail microphotographs of GlyR-immunoreactivity in hu- man and rat temporal cortex. MAb 4a-immunostaining. A: rat cor- tex. Layer II/III. At higher magnification, numerous immunoreae- five traversing dendrites can be seen. Only few immunolabeled neuronal perikarya are present. Bar = 20/~m. B: human temporal cortex. Layer V. At higher magnification, intense immunolabeling of pyramidal perikarya and apical dendrites is seen.
ronal perikarya, however, to a much lower extent. In perikarya, immunodeposits decorated nearly exclusively
the Golgi apparatus; rarely portions of the endoplasmic reticulum were stained. Other subcellular perikaryal
compartments including the nucleus were consistently
devoid of immunostaining. In longitudinal and transver- sal dendritic profiles, patchy immunoprecipitate was as-
sociated with the microtubular system. Immunostaining of dendrites appeared selective since numerous unstained
longitudinally- or cross-sectioned dendrites were encoun-
tered. Immunostained axonal profiles were not observed. At the terminal arborizations of immunostained den-
drites, often regions occurred where immunoreactive
postsynaptic sites were opposed to presynaptic boutons (Figs. 4,5). In these cases the immunoprecipitate was
regularly decorating the postsynaptic density (Figs. 4,5). Opposing presynaptic sites always displayed accumula-
tions of clear round vesicles. Besides immunoreactive
terminations, numerous unreactive synapses with clear vesicle-equipped presynaptic complexes were seen (cf.
Fig. 5). No presynaptic immunostaining was found,
Fig. 4. Immunoelectron microscopy of GIyR 4a-immunoreactivity. Human temporal cortex. A presynaptic site characterized by an ac- cumulation of clear round vesicles (V) is opposed to a postsynaptic membrane heavily stained by immunoperoxidase (arrows). (D) La- beled dendrite. Bar = 1/~m.
144
Fig. 5. GIyR 4a-immunoreactivity in rat occipital cortex. A synap- tie complex is shown which displays two terminals with clear round vesicles (V). These are in contact with a dendritic profile bearing an immunoreactive (arrows) and an unreactive postsynaptic density (arrowheads). Bar = 1 gin.
DISCUSSION
Here we have used a monoclonal antibody, MAb 4a, raised against the rat GlyR to demonstrate the presence of GIyR antigen in human and rat forebrain. This method was successful since the antigenic epitope MAb 4a is conserved in all GlyR a subunit variants identified in human as well as in rodent CNS 33'34. Indeed, sequence
comparison of human and rat GIyR cDNAs reveals a complete conservation of the epitope for this antibody 33.
GIyR antigen is present in rat and human cortex. Cor- tical levels, however, are particularly low in human brain, which displays a general reduction in GlyR content as compared to rat (cf. ref. 7). By contrast, immunocy- tochemistry revealed a rather high density of cortical GlyR-immunoreactive structures. Compared to biochem- ical studies on homogenized CNS tissue, this discrepancy may be explained by a higher resolution of morpholog- ical techniques recognizing small structures densely packed with MAb 4a-antigen. In fact, MAb 4a-antigen is present at a high concentration in a distinct subpopu- lation of cortical cells predominantly comprised of layer II/III and V pyramidal neurons.
At the light microscopical level, MAb 4a-reactivity re- flects the laminar and cellular distribution of glycinocep- tive neurons. At the electron microscopical level, MAb 4a-immunoprecipitate is thought to represent postsynap- tic structures 2"3~ as well as intracellular sites of GIyR protein synthesis, transport and assembly 16. Ultrastruc- turaUy, a preponderance of immunolabeled dendrites correlates well to the preferential staining of apical den- drites of pyramidal ceils observed by light microscopy.
Immunoreactive synapses had an asymmetric appear- ance which, at first sight, contrasts current assumptions about the morphology of putative glycinergic synapses. When analysing the shape of immunoperoxidase-labeled synapses, however, one should bear in mind that the electron-dense peroxidase product might precipitate be- yond the postsynaptic membrane area 2'2°.
Our data revealing the presence of GlyR in cerebral cortex do not match with autoradiographic studies em- ploying [3H]strychnine that failed to detect significant cortical antagonist binding 2s'39. Although this discrep- ancy may be due to the low resolution of the latter, an- other explanation appears more compelling: [3H]strych- nine autoradiography depends on a functional high- affinity binding domain of the GlyR molecule. Recently a GlyR isoform of low antagonist binding affinity has been described in spinal cord of newborn rats which is revealed only by MAb 4a but not by conventional [3H]strychnine binding assays. This neonatal receptor is distinct from the well characterized GlyR isoform of high antagonist binding affinity present in adult spinal cord 6. This heterogeneity of GlyR proteins correlates to cDNA variants of the GIyR a subunit which have been identi- fied in rodent and human CNS 1'13'16'17"34. In contrast to
the strychnine-sensitive a l subunit predominant in spi- nal cord of adult animals, functional GlyR channels characterized by a low sensitivity to strychnine are ob- served when the neonatal a2* subunit of rat spinal cord is expressed in Xenopus laevis oocytes 17. These data are complemented by the observation that, upon in vitro translation of poly(A) ÷ RNA from rat CNS in Xenopus oocytes, up to 3 distinct mRNA species can be identi- fied that generate glycine-gated chloride currents 1. These mRNA species appear to correspond to a subunit vari- ants of the GIyR which have recently been identified by cDNA cloning 13'17'18 (Y. Maulet and H. Betz, unpub-
lished observations). One of these predominates in new- born rat spinal cord and is also found, although at low levels, in adult rat cortex 1. Together, these data are con- sistent with a persistence in cortex of a neonatal type re- ceptor isoform which, due to its low strychnine-binding affinity, eludes autoradiographical detection. However, strychnine-resistant GlyRs from human CNS have not been identified at the eDNA level. Recombinant expres- sion of the two human a subunit variants known pro- duces strychnine-sensitive GIyR channels 13.
Our immunological data are corroborated by different lines of evidence indicating the presence in cerebral cor- tex of a presynaptic glycinergic system 3'8'11"29. Glycine
concentrations quantified in rat and human cerebral cor- tex 3 are approximately 1/4 of that measured in spinal cord 22. Also a stimulated release of glycine from cortical slices has been observed ~4'21 but is still a matter of con-
siderable debate. In at least one report the authors failed to elicit significant glycine release upon potassium depo- larization 23. Moreover, Na+-dependent glycine uptake into cortical preparations has been reported 24. Finally, numerous pharmacological studies have emphasized that strychnine-sensitive inhibition exists in cortex although some of these observations might be due to interaction of high concentrations of the alkaloid with postsynaptic GABA A receptors 1°'31. However, our immunological findings together with the presently available data
REFERENCES
1 Akagi, H. and Miledi, R., Heterogeneity of glycine receptors and their RNAs in rat brain and spinal cord, Science, 242 (1988) 270-273.
2 Altschuler, R., Betz, H., Parakkal, M.H., Reeks, K.A. and Wenthold, R.J., Identification of glycinergic synapses in the co- chlear nucleus through immunocytochemical localization of the postsynaptic receptor, Brain Research, 369 (1986) 316-320.
3 Aprison, M.H. and Daly, E.C., Biochemical aspects of trans- mission at inhibitory synapses: the role of glycine, Adv. Neuro- chem., 3 (1978) 203-294.
4 Araki, T., Yamano, M., Murakami, T., Wanaka, A., Betz, H. and Tohyama, M., Localization of glycine receptors in the rat central nervous system: an immunocytochemical analysis using monoclonal antibody, Neuroscience, 25 (1988) 613-624.
5 Becker, C.-M., Hermans-Borgmeyer, I., Schmitt, B. and Betz, H., The glycine receptor deficiency of the mutant mouse spas- tic: evidence for normal glycine receptor structure and localiza- tion, J. Neurosci., 6 (1986) 1358-1364.
6 Becker, C.-M., Hoch, W. and Betz, H., Glycine receptor het- erogeneity in rat spinal cord during development, EMBO J., 7 (1988) 3717-3726.
7 Becker, C.-M., Hoch, W. and Betz, H., Sensitive immunoas- say shows selective association of peripheral and integral mem- brane proteins of the inhibitory glycine receptor complex, J. Neurochem., 53 (1989) 124-131.
8 Bernardi, G., Cherubini, E., Marciani, M.G. and Stanzione, P., The inhibitory action of glycine on rat cortical neurons, Neuro- sci. Lea., 12 (1979) 335-338.
9 Betz, H. and Becker, C.-M., The mammalian glycine receptor: biology and structure of a neuronal chloride channel protein, Neurochem. Int., 13 (1988) 137-146.
10 Biscoe, T.J. and Curtis, D.R., Strychnine and cortical inhibi- tion, Nature, 214 (1967) 914-915.
11 Fagg, G.E., Jordan, C.C. and Webster, R.A., Descending fi- bre-mediated release of endogenous glutamate and glycine from the perfused spinal cord in vivo, Brain Research, 158 (1978) 158-170.
12 Graham, D., Pfeiffer, F. and Betz, H., Photoafiinity-labelling of the glycine receptor of rat spinal cord, Eur. J. Biochem., 131 (1983) 519-525.
13 Grenningloh, G., Schmieden, V., Schofield, P., Seeburg, P.H., Siddique, T., Mohandas, T.K., Becker, C.-M. and Betz, H., Alpha subunit variants of the human glycine receptor: primary structures, functional expression and chromosomal localization of the corresponding genes, EMBO J., 9 (1990) 771-776.
14 Hammerstad, J.P., Murray, J.E. and Cutler, R.W.P., Efflux of amino acid neurotrausmitters from rat spinal cord slices. II. Factors influencing the electrically induced efflux of [ 14C]glycine and [3H]GABA, Brain Research, 35 (1971) 357-367.
15 Hill, R.G., Simmonds, M.A. and Straughan, D.W., Convulsive properties of d-tubocurarine and cortical inhibition, Nature, 240 (1972) 51-52.
16 Hoch, W., Betz, H. and Becker, C.-M., Primary cultures of
145
strongly support an involvement of GlyRs in cortical neurotransmission 5.
Acknowledgements. We are greatly indebted to the late Mrs. M. Heusehel and to I. Wolters for skillful and expert technical assis- tance. We wish to thank Ch. Hoffmann and P. MOiler for help with the electron microscopical work, U. Koch for the photographic work and G. Kr/tuhsle for help in preparing the manuscript. This study was supported by the Deutsche Forschungsgemeinsehaft (Grants Sehr 283/4-1 and SFB 317).
mouse spinal cord express the neonatal isoform of the inhibi- tory glycine receptor, Neuron, 3 (1989) 339-348.
17 Kuhse, J., Schmieden, V. and Betz, H., A single amino acid exchange alters the pharmacology of neonatal rat glycine recep- tor subunit, Neuron, 5 (1990) 867-873.
18 Kuhse, J., Schmieden, V. and Betz, H., Identification and func- tional expression of a novel figand binding subunit of the inhib- itory glycine receptor, J. Biol. Chem., 265 (1990) 22317-22320.
19 Larson, E., Howlett, B. and Jagendorf, A., Artificial reductant enhancement of the Lowry method for protein determination, Anal. Biochem., 155 (1986) 243-248.
20 Lentz, T.L. and Chester, J., Localization of acetylcholine re- ceptors in central synapses, J. Cell Biol., 75 (1977) 258-267.
21 L6pez-Colom6, A., Tapia, R., Salceda, R. and Pasantes-Mo- rales, H., K+-stimulated release of labeled gamma-aminobu- tyrate, glycine and taurine in slices of several regions of rat central nervous system, Neuroscience, 3 (1978) 1069-1074.
22 McGeer, EL., Eceles, J.C. and McGeer, E.G., Inhibitory amino acid neurotransmitters. In P.L. McGeer, J.C. Eccles and E.G. McGeer (Eds.), Molecular Neurobiology of the Mammalian Brain, 2nd edn., Plenum, New York, 1987, pp. 197-234.
23 Mulder, A.H. and Snyder, S.H., Potassium-induced release of amino acids from cerebral cortex and spinal cord slices of the rat, Brain Research, 76 (1974) 297-308.
24 Neal, M.J., The uptake of [14C]glycine by slices of mammalian spinal cord, Z Physiol., 215 (1971) 103-117.
25 Pfeiffer, F., Graham, D. and Betz, H., Purification by affinity chromatography of the glycine receptor of rat spinal cord, J. Biol. Chem., 257 (1982) 9389-9393.
26 Pfeiffer, F., Simler, R., Grenningloh, G. and Betz, H., Mono- clonal antibodies and peptides mapping reveal structural simi- larities between the subunits of the glycine receptors of rat spi- nal cord, Proc. Natl. Acad. Sci. U.S.A., 81 (1984) 7224-7227.
27 Price, D.L., Stocks, A., Griffin, J.W., Young, A. and Peck, K., Giycine specific synapses in rat spinal cord. Identification by electron microscope autoradiography, J. Cell. Biol., 68 (1976) 389-395.
28 Probst, A., Cort6s, R. and Palacios, J.M., The distribution of glycine receptors in the human brain. A light microscopic an- toradiographic study using [3H]strychnine, Neuroscience, 17 (1986) 11-35.
29 Pycock, C.J. and Kerwin, R.W., The status of glycine as a su- praspinal neurotransmitter, Life Sci., 28 (1981) 2679-2686.
30 Schmitt, B., Knaus, P., Beeker, C.-M. and Betz, H., The M r 93000 polypeptide of the postsynaptic glycine receptor complex is a peripheral membrane protein, Biochemistry, 26 (1987) 805- 811.
31 Seholfield, C.N., Antagonism of ~,-aminobutyric acid and mus- cimol by picrotoxin, bicuculline, strychnine, bemegride, lepta- zol, d-tubocurarine and theophylline in the isolated olfactory cortex, Naunyn-Schmiedeberg's Arch. Pharmacol., 318 (1982) 274-280.
32 Schr6der, H., Zilles, K., Maelicke, A. and Haj6s, F., Immu- nohisto- and -cytochemical localization of cortical nicotinic re- ceptors in rat and man, Brain Research, 502 (1989) 287-295.
146
33 Schr6der, S., Hoch, W., Becker, C.-M., Grenningloh, G. and Betz, H., Mapping of antigenic epitopes on the a l subunit of the inhibitory glycine receptor, Biochemistry, 30 (1991) 42-47.
34 Sontheimer, H., Becker, C.-M., Pritehett, D.B., Schofield, P.R., Grenningloh, G., Kettenmann, H., Betz, H. and Seeburg, P.H., Functional chloride channels by mammalian expression of rat glycine receptor subunit, Neuron, 2 (1989) 1491-1497.
35 Tremblay, N., Warren, R. and Dykes, R.W., The effects of strychnine on neurons in cat somatosensory cortex and its in- teraction with the inhibitory amino acids, glycine, taurine and //-alanine, Neuroscience, 26 (1988) 745-762.
36 Triller, A., Cluzeaud, E, Pfeiffer, E, Betz, H. and Korn, H.,
Distribution of glycine receptors at central synapses. An immu- noelectronmicroscopy study, J. Cell. BioL, 101 (1985) 683-688.
37 van den Pol, A.N. and G6rcs, T., Glycine and glycine receptor immunoreactivity in brain and spinal cord, J. Neurosci., 8 (1988) 472-492.
38 Zamboni, L. and DeMartino, C., Buffered picric acid-formal- dehyde: a new rapid fixative for electron microscopy, J. Cell BioL, 35 (1967) 148A.
39 Zarbin, M.A., Wamsley, J.K. and Kuhar, M.J., Glycine recep- tor: light microscopic autoradiographic localization with [3H]strychnine, J. Neurosci., 1 (1981) 532-547.
Note added in proof Recent in situ-hybridization data reveal expression of GlyR a2* subunit mRNA in the cerebral cortex of the adult rat (Malosio, M-L.,
Marqu~ze-Pouey, B., Kuhse, J. and Betz, H., Widespread expression of glycine receptor subunit mRNAs in the adult and developing rat brain, EMBO J., in press).
