Light and electron immunocytochemical localization of AMPA-selective glutamate receptors in the rat brain.

THE JOURNAL OF COMPARATIVE NEUROLOGY 318:329354 (1992)

Light and Electron Immunocytochemical Localization of AMPA-Selective Glutamate Receptors in the Rat Brain RONALD S. PETRALIA AND ROBERT J. WENTHOLD Laboratory of Neurochemistry, National Institute on Deafness and Other Communication Disorders, National Institutes of Health, Bethesda, Maryland 20892

ABSTRACT Since four AMPA-type excitatory amino acid receptor subunits have been cloned recently, it is now possible to localize these important molecules in the nervous system. A comprehensive study of AMPA receptor immunocytochemistry was carried out on vibratome sections of rat brain, which were immunolabeled with antibodies made against peptides corresponding to the C-terminal portions of AMPA-receptor subunits: GluR1, GluR2/3, and GluR4. Labeling was most prominent in forebrain structures such as the olfactory bulb and tubercle, septal nuclei, amygdaloid complex, hippocampus, induseum griseum, habenula, and interpeduncular nucleus, and in the cerebellum. Different patterns of immunolabeling were evident with the antibodies to the four subunits, with marked contrast between densely and lightly stained structures with antibody to GluR1, widespread dense staining with antibody to GluR2/3, and moderate staining with antibody to GluR4. In the parietal cortex, some non-pyramidal neurons were more densely stained than pyramidal cells with antibodies to GluR1. Neurons of the main olfactory bulb, other than granule cells, were most densely stained with antibody to GluR1. In the cerebellum, Bergmann glia were densely stained with antibodies to GluRl and 4, while neurons, other than granule cells, were most densely stained with antibody to GluR2/3. Immunolabeling patterns of all antibodies were consistent with that of previous in situ hybridization histochemistry studies and with the overall pattern of 3H-AMPA binding. Electron microscopy of thin sections taken from immunolabeled vibratome sections of hippocampus and cerebral cortex showed staining which was restricted mainly to postsynaptic densities and adjacent dendritoplasm, and to neuron cell body cytoplasm. We saw no convincing examples of stained presynaptic terminals, and only limited evidence of glial staining, excepting Bergmann glia. Key words: ultrastructure, hippocampus, cerebellum, cerebral cortex, olfactorybulb

Excitatory amino acids are among the most widely distrib- al., '90) respond to AMPA, kainate, and glutamate when uted brain neurotransmitters and also play key roles in expressed in oocytes and cultured cells, with responses developmental and adult synaptic plasticity and neurogene- blocked by CNQX (Boulter et al., '90; Keinanen et al.,'90; sis (Monaghan et al., '89; Cambray-Deakin et al., '90; Nakanishi et al., '90). These subunits appear to be inteCollingridgeand Singer, '90; Mayer and Miller, '90). Excita- grated into a heteromeric receptor complex (Keinhen et tory amino acid receptors include several forms (Monaghan al., '90; Nakanishi et al., '90; Sommer et al., '90; Wenthold et al., '89), although a definitive classification of these et al., '92), with the GluR2 subunit being most important in receptors is not yet possible (e.g., Patneau and Mayer, '91). determining ion flux (Verdoorn et al., '91) including that of Tentatively, they can be divided into G-protein coupled Ca2+(Hollmann et al., '91). Other types of non-NMDA metabotropic receptors (Masu et al., '91; Houamed et al., receptors including GluR5, GluR6, and the putative recep'91) and ionotropic receptors including NMDA (Monaghan tor KA1 are pharmacologically different and may include et al., '88; Monaghan, '91) and non-NMDA subtypes. Non-NMDA receptors include several cloned subunits of Accepted December 20,1991. which four are known to bind 'H-AMPA, with binding Please address correspondence and reprint requests to Dr. Ronald S. inhibited by CNQX (Keinanen et al., '90). These four Petralia, NIH, NIDCD, Bldg. 36, Rm.5D08, 9000 Rockville Pike, Bethesda, subunits, GluR1-4 (also known as Glum-D; Keinanen et MD 20892. o 1992 WILEY-LISS, INC.

330

R.S. PETRALIA AND R.J. WENTHOLD

kainate receptors and kainate binding proteins (Bettler et al., '90; Egebjerget al., '91; Werner et al., '91). Distribution of AMPA-bindingreceptors in the brain has been examined previously with 3H-AMPAbinding studies (Monaghan et al., '84; Rainbow et al., '84; Olsen et al., '87; Insel et al., '90) and in situ hybridization histochemistry studies of mRNA to GluR1-4 (Bettler et al., '90; Boulter et al., '90; Gall et al., '90; Keinanen et al., '90; PellegriniGiampetro et al., '91).However, comparison of the distributions indicated by the two methods reveal some discrepancies (Keinanen et al., 'go), which could result from either differential localization of receptor mRNA and the translated receptor molecules (i.e., in situ hybridization histochemistry), or from a lack of 3H-AMPAbinding to some forms of the functional receptor complexes. Thus, since binding studies with 3H-AMPA may be inadequate to represent the complete localization of GluR1-4 receptors in the brain, immunocytochemical localization studies are

required for assessment of their overall distribution, as well as for the distribution of individual subunits. More importantly, immunocytochemistry is required for good resolution of receptor distribution at both the light and electron microscope levels and is the only technique which can provide data on synaptic localization of receptors. However, at present, only preliminary immunolabeling studies have been published (Heinemann et al., '91). Further, no ultrastructural immunolabeling has been completed for AMPA receptors. In fact, with the exception of kainate binding protein (Dechesne et al., '90; Somogyi et al., '901,very little data has been published on ultrastructural localization of any putative excitatory amino acid receptor (Bobryshev et al., '89). In this study, we describe the immunocytochemicallocalization of GluR1, 2/3, and 4 in the rat brain, employing antibodies to synthetized oligopeptidesbased on C-terminal amino acid sequences of GluR1-4. A pre-embedding method

Abbreviations

aa Ab A0 aP AOB As Ar As BG B1 BS

c1

c3 cc cg Cn CP DC DG DL

DS ds EC En EP Fa FH Fr G1 Go GP Gr IC ic IG I0 IP

t L

LA LH LP LS LV mCP MF MG MH Mi mi ml Mo

Anterior commissure, anterior Accumbens n., core Anterior olfactory n. Anterior commissure, posterior Accessory olfactorybulb Aqueduct Arcuate hypothalamic n. Astrocyte-likecells Bergmann glial process Basolateral amygdaloid n. Bed n. of the stria terminalis Field CA1 ofAmmon's horn Field CA3 of Ammon's horn Corpus callosum Cingulate cortex Cuneiform n. Caudate-putamen Dorsal cochlear n. Dentate gyrus Lateral geniculate n., dorsal part Specialized spines of CA3 pyramidal cell apical dendrites Dorsal spinocerebellar and olivocerebellar tracts External cuneate n. Entorhinal cortex External plexiform layer Facial nucleus Forelimb/hindlimb areas of the cortex Frontal cortex Glomerular layer Golgi cell Globus pallidus Granule cell layer of olfactory bulb; granular layer of cerebellum Inferior colliculus Internal capsule Induseum griseum Inferior olive Internal plexiform layer Interpeduncular n. Interposed cerebellar n. Lugaro cell Lateral arnygdaloid n. Lateral habenula Lateral posterior thalamic n. Lateral superior olive Lateral ventricle Middle cerebellar peduncle Mossy fiber terminals Medial geniculate n. Medial habenula Mitral cell Mitochondrionin labeled dendrite Medial lemniscus Molecular layer

NP No1 Nu 01 ox P PC

P1 Pd Pf pg Pi Pj Pn Pr Pre PY rf RN Rt RTg S SA XP

Sf SG

S1 SNc SN SOg

so

sl, s2, s3 s5 s5

St

m sv T1 Tu

uc

uns VA

VL VP VTg WM ZI 3 7 8 10 12 111

w

Non-pyramidal cell Nucleolus Nucleus Occipital cortex, area 1 Optic chiasm Pyramidal cell Posterior cortical amygdaloid n. Parietal cortex, area 1 Pyramidal cell apical dendrite Parafascicular thalamic n. Periglomerular cell Piriform cortex Purkinje cell Pontine nuclei Perirhinal cortex possible presynaptic labeling Pyramidal tract Rhinal fissure Red n. Reticular thalamic (reticulothalamic) n. Recticulotegmental n. pons Subiculum Short axon cell of granule cell layer Superior cerebellar peduncle Septofimbrial n. Superficial gray layer of superior colliculus Medial nucleus of the solitary tract Substantia nigra, compact part Substania nigra, lateral and reeticular parts Subfornical organ Stratum oriens synapses contacting same dendrite Spinal n. of the trigeminal nerve Spinal trigeminal tract Subthalamic n. Superior vestibular n. synaptic vesicles Temporal cortex, area 1 Olfactory tubercle Unlabeled neuron cell body Unstained synapse Ventral cochlear n., anterior part Lateral geniculate n., ventral part Ventral cochlear n., posterior part Ventral tegmental n. White matter Zona incerta oculomotor n. Facial nerve or its root Vestibulocochlearnerve Dorsal motor nucleus of the vagus Hypoglossal n. Third ventricle Fourth ventricle

IMMUNOCYTOCHEMISTRY OF GLUTAMATE RECEPTORS is used to determine the immunolabeling of a large number of brain structures. In addition, we present the first electron microscope studies of the localization of these receptor subunits at synapses.

331

Immunocytochemistry for electron microscopy

Sections used for pre-embedding electron microscope immunocytochemistry were prepared as described for light microscopy. Following the DAB reaction, selected sections were fixed in 1%osmium tetroxide in PBS for 1 hour and then washed in PBS for 1hour. Sections were dehydrated MATERIALS AND METHODS in a graded series of ethanols, followed by propylene oxide, Antibody production and purification overnight infiltration in Poly/BED 812 resin (Polysciences, Synthetic peptides corresponding to C-terminal se- Inc., Warrington, PA), flat embedding, and polymerization quences (Keinanen et al., '90) of GluRl (SHSSGMPL- at 60°C for 24 hours. Embedded tissue was mounted on GATGL), GluR2/3 (EGYNVYGIESVKI),and GluR4 (RQSS- resin stubs and sectioned at 60 to 90 nm (75 nm average) on GLAVIASDLP) were produced commercially, conjugated to an LKB Ultratome IV ultramicrotome. Sections were examBSA with glutaraldehyde, and injected into rabbits as ined unstained in a JEOL JEM-100CX I1 transmission described previously (Wenthold et al., 'gob, '92). The electron microscope at 60 kV. resulting antisera, Ab7A and 7B for GluR1, Ab25 for GluR2/3 and Ab22 for GluR4, were affinity-purified by using immobilized peptides that were the same as (Ab7A, Areas surveyed Ab25) or part of (MPLGATGL for Ab7B; LAVIASDLP for The brains of seven animals were surveyed. Coronal Ab22) the peptides used as antigens. Peptides were coupled sections were taken through the following regions: main (10 mg/ 1.5 ml) to CH- (Ab7,Ab22) or AH-Sepharose (Ab25) resins, which were incubated with antisera in phosphate- olfactory bulb (3/PW1; i.e., number of animals/figure from buffered saline (PBS; pH 7.4) for 1.5 hours at room Paxinos and Watson, '86); anterior olfactory nucleus (11 temperature and then washed with 30 ml PBS. Bound PW5); caudate-putamenlsubfornicalorgan (4/PW21-22); antibody was eluted with 0.2 M glycine, pH 2.5, neutralized lateral geniculate nucleus/ habenula nucleus (4/PW32-34); with 1M potassium phosphate, and the antibodies with 1% superior colliculus/red nucleuslinterpeduncular nucleus BSA were dialyzed against PBS. All antibodies were shown (4/PW42-43); inferior colliculus (4lPW51); facial genu and to be specific for their receptor subunits by Western blot nerve (4/PW57-59);facial nucleusldorsal cochlear nucleus analysis of membranes of cultured cells transfected with (1/PW63-64); and inferior olive (4IPW68-70).Sagittal secGluR1,2,3or 4 cDNAs (Wenthold et al., '92). tions of the cerebellum were studied in all seven animals. Whole-brain sagittal sections (three animals) were taken at Tissue preparation up to six levels (PW80-85).Additional sagittal sections of Young male Sprague-Dawley rats (125-250 g) were anes- the cochlear nuclei (three animals) and cerebellum (four thetized with 20% urethane in H,O (1 m1/100 g body animals) were examined. Identification of many structures weight) and perfused transcardially. Animals were perfused was corroborated by examination of corresponding sagittal with about 100 ml of 0.12 M phosphate buffer, pH 7.2 and transverse sections stained with cresyl violet, provided (Friedrich and Mugnaini, '81), at room temperature, fol- by Dr. W. Bruce Warr (Boys Town Inst., Omaha, NE). lowed by perfusion with about 500 ml of ice-cold fixative. Some coronal sections in the region of the lateral genicuThe fixativewas 4% paraformaldehyde in 0.12 M phosphate late nucleus/habenula nucleus (1/PW32-35) were probuffer (pH 7.2) with or without 0.1% glutaraldehyde for cessed for electron microscopy. The lateral portion of the light/electron microscope or light microscope study, respec- hippocampus or adjacent portion of the cerebral cortex tively. Brain and eyes were removed and postfixed in the (probably parietal cortex, area 1)was mounted on its side same fixative for 1hour at 4"C, stored overnight at 4°C in and thin sections were taken from the lateral edge. Thus, PBS, and then sectioned in cold PBS with a microslicer the resulting thin sections were 50 pm wide and spanned (Pelco DTK-3000W) at 50 km. One percent agarose in PBS the dorsoventral height of the hippocampus or cortex, was used as a supportive material for sectioning of eyes. respectively. Thin sections of the hippocampus included the CA3 region for all four antibodies (Ab7A, Ab7B for GluR1, Immunocytochemistryfor light microscopy Ab25 for GluR2/3, and Ab22 for GluR4), as well as the All incubations were carried out on a shaker at medium CA1-2 region for the GluR2/3 and 4 antibodies. speed in 6 or 24 well tissue culture dishes (clusters). Identification of brain structures was based mainly on Sections were first placed in 10%normal goat serum in PBS Paxinos and Watson ('86). Only structures which could be for 1hour and then incubated in the primary antibody in identified with certainty in the vibratome sections were PBS overnight at 4°C. Antibodies were used in a wide range included in the Results and Table 1. Some large structural of concentrations, but most common were 2.4 pg/ml for complexes (e.g., the amygdala, septum, hypothalamus, thalAb7A (GluRl), 1 kg/ml for Ab7B, 2.8 pglml for Ab25 (GluR2/3),and 1.2 kg/ml forAb22 (GluR4).Next, in 1hour amus, etc.) extend beyond the sections examined, as defined intervals, sections were washed three times in PBS, placed by the corresponding coronal and sagittal sections from in biotinylated secondary antibody (Vectastain kit), washed Paxinos and Watson ('86) described above, and our descripthree times in PBS, placed in ABC reagent (avidinlbiotinl tions of immunolabeling can be considered valid for the peroxidase), and washed in PBS three times. Sections were portion examined only. Unless indicated otherwise, structreated with 3',3-diaminobenzidine tetrahydrochloride tures in Table 1 were assigned a relative value based upon (DAB; 10 mg/20 ml PBS + 5 k1 hydrogen peroxide),washed the overall staining, i.e., including that of both cell bodies in PBS, and mounted on slides using 1%gelatin in 25% and neuropil. This value was relative to the range of ethanol. Air-dried slides were placed in xylene, and cover- staining intensity from light to dense, seen with that antibody throughout the brain. slips were attached with Permount.

332

R.S. PETRALIA AND RJ. WENTHOLD

TABLE 1. Localization of AMPA Receptor Subunits in the Rat Brain, Wed From 0-4

TABLE 1 (continued)

GluRl GluR2/3 GluR4 I. Forebrain A. Isocortex 1. Frontalcortex 3 3 2. Forelimbihindlimb area 3.5 3.5 3 3.5 3. Temporal cortex, Area 1 (prim. aud.), 3 4. Occipital cortex, Area 1 3 3 5. Medial occipital area 2 3 3.5 6. Parietal cortex, Area 1 (primary somatosensorylneuron cell bodies) I 3 3.5 I1 1 3 I11 1 3 N 2 3 V 3 3.5 VI 3 3 VI (deep layer) 3 3.5 White matter 0 1 B. Allwortex 1 Cingulate cortex, Area 1,2 2.5 3 2. Perirhinal cortex 4 3 3. Insular cortex 3.5 3.5 C. Olfactory regions 1. Anterior olfactory n. 3.5 3.5 2. Olfactory tubercle 3.5 3.5 3. Piriform cortex 3.5 4 4. Accessory olfactory bulb Vomeronasal nerve layer 2 2.5 Processes of glomerular layer 3 3 External plexiform layer (EPL) 2.5 3.5 Neuron cell bodies in EPL 2 3.5 Internal plexiform laver 1 1 Granule cells 1.5 2 5. Main olfactory bulb (ventroposterior region stains slightly lighter) Olfactory nerve 3 3 Processes of glomerular layer 4 4 Periglomerular cells 3.5 3 3 External plexiform layer (EPL) 3.5 Neuron cell bodies in EPL 3 Mitral cells 2.5 3 Internal plqiform layer 2 1.5 Granule cells-superficial 2 2 Granule cellsdeep 2.5 2 Short axon cells of granule c. layer 3 2 1 1 Ependymal layer D. Hippocampal formation (cortex) 1. Entorhinal area 3.5 4 4 2. Suhiculum 3.5 3. CA1 3.5 4 S. lacunosum-moleculare Stratum radiatum 3 3.5 4 4 Pyramidal layer Stratum oriens 4 3.5 4. cA2,3,4 3.5 3 Molecular layer 4 4 Pyramidal layer 4 3.5 Stratum oriens 5. Dentate a t u s 3.5 2.5 Molecular layer 4 4 Granular layer 2.5 Polymorph layer 2.5 4 4 6. Induseum griseum E. Amygdala 3 4 1. Medial nucleus 4 2. Anterolateral amygdalohippocampal a 4 4 3. Posterior cortical nucleus 3.5 3.5 4. Lateral nucleus 3 4 5. Basolateral nucleus 3.5 6. Basomedial nucleus 3 3.5 F. Septum 1. Lateral nucleus, pars dorsalis 4 4 2. Bed nucleus, stria terminalis 3.5 3.5 3.5 3. Septofimbrial nucleus 3 3.5 3 4. Subfornical organ G. Basalganglia 1. Caudate-putamen 3 3 3 3 2. Nucleus accumbens 2 1.5 3. Glohus pallidus 4. Ventral pallidum 3 2 5. Substantia innominata 3 2 2.5 6. Subthalamic nucleus 3 7. Substantia nigra Compact part 3 3 2.5 Reticular, lateral parts 3 H. Thalamus 1. Medial hahenula 4 3.5 2. Lateral habenula 3 3.5 3. Intermediodorsal n. 2 2 4. Centromedial n. 2 2.5 5. Centrolakal n. 2 2.5 2.5 6 Lateral posterior n. 1.5

2 2.5 2.5 2.5 2.5 2.5

1.5 1.5 2 2.5 2 2 1 2 3 2 2.5 2.5 3

2.5 2.5 3 1.5 1 2.5 3 3 1.5 3.5

-

2.5 2 2.5 2.5 1.5 1.5 3 3 2.5 2 3 2 2 3.5 2 2 3.5 2 3 3.5 3.5 3.5 2.5 3 2.5 2.5 2.5 2 3 2 2 1.5 2 2 2 3 2.5 3.5

3 2 2 2 2

-

-

GluRl GluR213 g1ur4 7. Posterior paraventricular n. 8. Ventroposterior n.

2.5

3

2.5

Medial Lateral 9. Posterior n. 10. Medial geniculate n. 11. Lateral geniculate n. Dorsal part Ventral part 12. Parafascicular n. 13. Reticulothalamic n. 14. Zona incerta I. Hypothalamus 1. Supraoptic n. 2. Posterior n. 3. Arcuaten. 4. Ventromedial n. 5. Jcn.-median eminencelinfundibulum J. Retina 1. Ganglion cell layer 2. Inner plexiform layer 3. Inner nuclear layer 4. Outer plexiform layer 5. Outer nuclear layer 6. Rods 7. Cones X. Whitematter 1. Corpus callosum 2. Internal capsule 3. Optic chiasm 4. Anterior commissure Anterior Posterior 11. Brainstem A. Sensory 1. Superior colliculus Superficial gray Optic nerve layer Intermediate gray Intermediate white Deep gray Lge. multipolar neurons, lat. lay. 3 2. Vth cranial nerve Mesencephalic n. V neurons Spinal n. V Oral part Interpolar part Caudal part 3. External Cuneate n. 4. Auditory nuclei Cochlear nuclei Anteroventral Posteroventral Dorsal, layers 1.2 N. trapezoid body Superior olive Lateral Medial Periolivary nuclei Superior Dorsal and ventral N. lateral lemniscus, dorsal, ventral Inferior mlliculus 5. Vestibular nuclei Medial n. Lateral n. Superior n. Spinal n. 6 . Medial solitary n. B. Motor 1. Oculomotor n. (111) 2. Abducens n. (VI) 3. M0torn.V 4. Facial n. (VII) 5. Hypoglossal n. 6. Dorsal motor n. vaws (X) C, Reticular core, incl. central gray 1. Ventral tegmental n. 2. Reticulotegmental n. 3. Cuneiform n. 4. Pontine nuclei 5. Interpeduncular n. 6. Redn. 7. Inferior olive Medial Principal and dorsal 8. Reticular formation Pontine n.,oral part Pamucellular

1.5 2

2.5

2

2.5

2.5

1

2 2

2 2.5

2.5 3 2.5 3 2.5

2.5 3 2.5 3 2.5

2.5 2.5 2 3

2 2.5 3.5 3.5 2.5

3.5 3 3.5

3 2.5 3.5 3 3

3 1 3

4

3

2.5 1.5

3 3.5 3.5 3

1

-

2

2 3 3 2 1

3

-

0 0 0

1 1 1

1 0.5 1

0.5

1

1

2

0.5 1.5

2 1 2

2.5 2 2 1.5 2.5 3

1.5 2.5

2.5

3.5

2.5

2 2 2 3

3

2.5 2.5 3

3 2.5 2 3

1.5 2 3 2.5

2.5 3 2.5 3

2.5 3 2.5 3

2.5 2.5

3 3

3 3

2 2 2 2

3 3 3 3

2.5 3 2.5 2.5

2 2 2 2 2

3 3

2.5 2.5 2.5

3 2.5

2.5

3 1.5 2 2.5 1.5

3.5

1 '

1

2 2

4

3.5

3

2

2 1 1.5 1

2

1.5

3 3

3 3 3 3 2.5 2.5

2 3 2 2 3.5 3

3 3 3 3 4 3

2.5 2.5 2 3 3.5 3

3 3

3.5 3

2.5

1.5 2.5

3

2 2

3 3

a

9

2

333

IMMUNOCYTOCHEMISTRY OF GLUTAMATE RECEPTORS TABLE 1 (continued)

GluRl GluR213 GluR4 9. Median raphe n. (pons) 10. Nucleus of Roller D. WhiteMatter 1. Dorsal spinocerebellartract 2. Tract n. V 111. Cerebellum A. Deep nuclei 1. Interposed 2. Lateral B. Cortex 1. Molecular layer 2. Granular layer 3. Bergmann glia 4. Purkinje cell bodies 5. Purkinje cell dendrites 6. Granule cells 7. Molecular layer cells 8. Golgi cells 9. Lugamcells 10. Whitematter

2 1.5

3 3

2.5 2.5

0 0

2 2

1 1

2 1.5

3 3

3

3.5 2 4

2.5 2 0

1 1

3 3

3 2.5 3 1 1

0.5 -

2 1

2

2

2

2 1

0

2.5

2 -

0.5

Controls We substituted PBS for the primary antibody for some sections in every run in the procedure. In addition, whole brain sagittal sections fixed in paraformaldehyde/glutaraldehyde were used for a preadsorption control, in which each of the four antibodies was incubated with its corresponding peptide antigen for 7 hours. In addition, Ab7A and Ab7B were incubated for 31 hours in a second test. Following centrifugation, each preadsorbed antibody solution was applied to sections which were processed for immunocytochemistry along with experimental sections. Also, PBS control sections were run for electron microscopy of both hippocampus and cerebral cortex.

Electrophoresis and immunoblotting Brains were removed from urethane-anesthetized rats which were perfused through the heart with PBS. Tissue from kidney, liver, and skeletal muscle were removed from rats anesthetized with CO, and decapitated without perfusion. Twenty-five micrograms of protein from each organ and seven regions of the brain (olfactory bulb, cerebral cortex, hippocampus, superior colliculus, inferior colliculus, cerebellum, and hindbrain) were subjected to SDS-PAGE ( 4 2 0 % polyacrylamide; Laemmli, ’70). Immunoblots on nitrocellulose membranes (Towbin et al., ’79) were treated with 5% instant powdered milk in Tris-buffered saline (TBS; 50 mM Tris, 150 mM NaCl, pH 7.4) with 0.05% Tween-20, incubated with antibodies at concentrations of 3 Fg/ml for Ab7A, 0.3 pg/ml for Ab25, and 1.5 bg/ml for Ab22, and detected with an alkaline phosphatase-conjugated second antibody (Kirkegaard and Perry Laboratories, Inc., Gaithersburg, MD). Prestained protein standards, obtained from Bethesda Research Laboratories, included myosin, phosphorylase B, bovine serum albumin, and ovalbumin, migrating at M, = 200,000, 104,000, 66,000, and 42,000, respectively.

RESULTS Controls Immunoblot analyses with all antibodies produced a single major band migrating with a M, = 108,000 for all regions of the brain with little or no trace in the peripheral organs (Fig. 1).Antibody to GluRl showed the densest band for the olfactory bulb, cerebral cortex, hippocampus, and

cerebellum, with a lighter major band for superior and inferior colliculi, and the lightest band for the hindbrain (Fig. la). In comparison, antibody to GluR213 produced a dense major band for all brain structures except the hindbrain (Fig. lb). In contrast to both GluRl and 2/3, antibody to GluR4 produced only a light, diffuse major band for each brain region except the cerebellum (Fig. lc). Also, with antibody to GluR4, the peripheral organs showed a major band migrating very close to the position of the prestained bovine serum albumin standard. This band may represent a blood serum component not present in the brain samples, which were perfused with PBS. Light and electron microscope controls in which PBS was substituted for the primary antibody were all unstained. Peptide preadsorption controls for GluR2/3 and 4 were unstained. The 7 hour preadsorption control for GluRl had a slight stain in the Purkinje cells of the cerebellum, some neurons in all layers (includingMitral cells) of the olfactory bulb, cerebral cortex, especially the frontal and occipital cortex, area 1, and in many brainstem nuclei. Other structures such as the hippocampus and the Bergmann glia of the cerebellum were completely unstained. Most of this staining was eliminated with the 31 hour preadsorption, while the corresponding experimental sections stained normally. Since the GluRl antibody Ab7A gave the best staining and its preadsorption control staining was less pronounced than with Ab7B, descriptions of staining densities in the affected structures are based upon results with Ab7A. Evidence of this non-specific staining was seen occasionally in certain nuclei in glutaraldehyde-fixedtissue with antibody Ab7B. Most notable was a non-specific staining, including staining of many cell nuclei, of neurons in the anteroventral cochlear nucleus, which normally stained very lightly with Ab7B and Ab7A on paraformaldehyde-fixed tissue and with Ab7A on paraformaldehydel glutaraldehyde-fixedtissue.

Light microscopy-general Overall pattern of staining differed for the three antibodies (GluR1, 2/3, 4) and was consistent with the pattern revealed in immunoblot analysis. Immunolabeling with antibodies to GluRl showed the greatest contrast between densely stained and lightly stained structures, and this was most evident in the forebrain. Immunolabeling of most structures was dense with antibodies to GluR2/3, while most structures stained moderately and diffusely with antibodies to GluR4 (Figs. 2-7). Specific staining was limited to the neurons or distributed throughout the neuropil, or both. For practical purposes, “neuronal staining” refers to staining of the cell body, excluding nucleus, and the major dendrites, which could be traced from the cell body, while “neuropilar staining” includes processes not traced to specific cell bodies and the unresolvable matrix in between cells. Structures in which there was a significant staining of small processes and matrix are described as having a “diffuse neuropilar stain.” Typically the matrix contained numerous fine puncta which often increased in number relative to the overall staining intensity.

Cortex In coronal sections of the parietal cortex, area 1(Figs. 2a, 3a,; PW33; Zilles and Wree, ’85), immunolabeling with


334

A B C D E

F G H

I J

antibody to GluRl resulted in the highest contrast between the deep (V-VI) and superficial layers (11-IV). This difference was due mainly to a denser diffuse neuropilar staining in V-VI for GluR1. In contrast, neuropilar staining was equally dense in layers 11-VI with antibodies to GluR2/3 and 4, with staining significantly lighter with antibody to GluR4. All antibodies stained lightly the region corresponding to the relatively cell-sparse border between IV and V. Immunolabeling with antibody to GluRl was densest in scattered small and medium non-pyramidal cells (Fig. 8a). These varied in shape from fusiform to round to irregular and usually had 3-4 main dendrites. Those of layers 11-IV tended to be smaller and were often of the vertical-fusiform type. The large pyramidal cells of layer V were evident with antibody to GluR1, but staining was slightly less dense than in the non-pyramidal cells. In addition, numerous cells in all layers including pyramidal and non-pyramidal were stained very lightly. In contrast, immunolabeling with antibody to GluR2/3 was densest in the pyramidal cells of layer V (Fig. 8b). Also, staining of pyramidal cells of layers I1 and I11 was almost as dense as that of layer V. A moderate staining was evident in numerous non-pyramidal cells of layers IV and VI. In addition, neurons of the deepest part of layer VI, just adjacent to the white matter, were densely stained using antibody to GluR2/3. In sagittal sections with antibody to GluR2/3 (PW82), this deep region of VI was most densely stained in the forelimb/hindlimb area, in which stained neurons of the deep region of VI were mainly horizontal cells. Cellular staining for antibody to GluR4, in coronal sections of parietal cortex, layer 1, was similar to that of antibody to GluR2/3, although much lighter (Fig. 812). Overall staining of isocortex and allocortex was highest for antibody to GluR2/3, with staining for antibody to GluRl similar but slightly less and antibody to GluR4 much less. Frontal cortex was often the lightest staining area in sections of isocortex (Figs. 6, 7). Increased density in the forelimb/hindlimb area (Fig. 2) was especially noticeable in many sagittal sections, where it could be contrasted to the adjacent frontal and occipital cortex (Figs. 6, 7). Insular cortex showed the greatest difference in staining intensity among the three antibodies.

Olfactory regions Overall, the olfactory regions stained densely with all antibodies (Figs. 6, 7, 9). In the main olfactory bulb, the external plexiform layer could be delineated easily due to a diffuse neuropilar labeling most evident for antibodies to GluR2/3 and 4 (Fig. 9b,c). The processes of the glomerular layers were very distinct with antibodies to GluRl and 2/3. All neurons, excluding granule cells, stained most densely with antibody to GluR1. Neurons of the external plexiform layer were distinguishable with antibody to GluRl only (Fig. 9a,d). We did not distinguish between short axon cells and periglomerular cells proper among the periglomerular

Fig. 1. Immunoblot analyses of sodium dodecyl sulfate (SDS)gels of rat brain tissues, using anti-peptide antibodies to GluRl (a),GluR2/3 (b),and GluR4 (c).A, kidney; B, liver; C, skeletal muscle; D, olfactory bulb; E, cerebral cortex; F, hippocampus; G, superior colliculus; H, inferior colliculus;I, hindbrain; J, cerebellum. Twenty-fivemicrograms of protein were applied to each lane. Blots were extensivelydeveloped to reveal all minor bands. Arrows show the positions of apparent molecular weights of prestained standards myosin (H chain), M, = 200,000; phosphorylase b, M, = 104,000; bovine serum albumin, M, = 66,000; ovalbumin,M. = 42,000.

IMMUNOCYTOCHEMISTRY OF GLUTAMATE RECEPTORS

Figs. 2-7 (on following pages) Low magnification survey of sections immunolabeled with antibodies to GluRl (a),213 (b),and 4 ( c ) .Contrast and density have been adjusted to best reflect the real differences in immunolabeling among the three antibodies for each region of the brain. For abbreviations, see list. Fig. 2. Coronal sections at the level of the caudate-putamenisubfornicalorgan (PW21-22). ~6.7.

335

336


Fig. 3. Coronal sections at the level of the lateral geniculate nucleusihabenula nucleus (PW32-34). ~ 6 . 7 .


Fig. 4. Coronal sections at the level of the superior colliculus/red nucleus/interpeduncular nucleus (PW42-43).x6.7.

337


338

Fig. 5. Coronal sections at the level of the inferior colliculus ( a x ; PW51), facial genu (d-f; PW57-591, and inferior olive (g-i; PW68-70). GluRl (a,d,g);GluR2/3 (b,e,h);GhR4 k,f,i). x6.7.

Hippocampus

and Ribak, '85). Nonpyramidal neurons were difficult to identify within the pyramidal and granular layers due to the high cell density (Fig. 10). The stratum oriens of all CA regions showed a diffuse neuropilar staining and contained a few non-pyramidal neurons, especially in the CA3-4 regions (Fig. 10). In contrast, staining in the molecular layer or stratum radiatum of CA regions delineated the large branching processes of the pyramidal cell apical dendrites, with only occasional non-pyramidal neurons evident (Fig. 10). Immunolabeling for antibody to GluR4 showed the same general pattern as for the other antibodies, but was much lighter. Neuron cell bodies of the CA regions were difficult to distinguish due to dense packing and moderate staining with antibody to GluR4. In contrast the more densely stained apical dendrites of the pyramidal cells were always distinctive, as with the other antibodies. A few scattered small cells which appeared to be astrocytes were seen in the stratum oriens with antibody to GluR4.

All regions of the hippocampus (Brown and Zador, '90) stained very densely when immunolabeled with antibodies to GluRl and 2/3 (Figs. 3a,b, 4a,b, lOa,b,d,e).In addition to pyramidal cells, nonpyramidal cells throughout the hippocampus were stained, with those of the polymorph layer (hilus) of the dentate gyrus being the most distinctive (Amaral, '78; Rib& and Seress, '83; Ribak et al., '85; Seress

Most regions of the amygdala (Krettek and Price, '78; as well as Paxinos and Watson, '86) stained densely with all antibodies (Fig. 3).This was due mainly to dense neuropilar staining, although neuron somas were stained also. The lateral nucleus had the lightest staining with antibodies to

cells (Fig. 9a,d; Mori, '87; Shepherd and Greer, '90). Granule cells were densest with antibody to GluR4 and in the deep layers with antibody to GluR2/3. Thin vertical processes were stained in the granule cell layer with antibody to GluR2/3 but could not be traced to particular cell bodies. The medial ventroposterior region of the main olfactory bulb was stained lighter overall than other regions, but the staining pattern of all regions was similar. The accessory olfactory bulb was the lightest staining structure of the olfactory regions (Fig. 6). Unlike the main olfactory bulb, staining of neurons of the external plexiform layer was evident with all antibodies and distinctly denser with antibody to GluR2/3. The pattern of staining of these neurons differed from the pattern of the staining of the mitral cells of the main olfactory bulb, in spite of the presumed relationship between these cell types (Mori, '87).

Amygdala

339


Fig, 6. Sagittal sections at the level of the superior and inferior colliculi (PW80-82). ~ 4 . 8 .

GluRl and 4, of the regions listed in Table 1. In contrast, staining of this nucleus with antibody to GluR2/3 was similar to that of the other nuclei.

Septum Nuclei were more densely stained when immunolabeled with antibodies to GluRl and 213 than with antibody to GluR4 (Figs. 2, lla-c). This difference in staining density was greatest in the lateral nucleus which consisted mainly of a plexus of numerous, densely staining processes interspersed with scattered cell bodies with antibody to GluRl (Fig. lla), with staining density slightly less with antibody

to GluR2/3 (Fig. l l b ) . Only cell bodies and major dendritic processes were evident with antibody to GluR4 (Fig. 1lc).

Basal ganglia Structures in this group could be divided into 1) those in which overall staining was moderately dense for all three antibodies (substantia nigra, pars compacta; Figs. 4, 6); 2) those which stained most densely when immunolabeled with antibody to GluR2/3 (globus pallidus, ventral pallidum, substantia innominata, and substantia nigra, rostra1 and lateral parts; Figs. 2b, 4b, 6b); and 3) those which stained moderately dense when immunolabeled with anti-


340

Fig. 7. Sagittal sections at the level of the cochlear nuclei (PW84-85). The cerebellum and attached cochlear nuclei were positioned farther from the forebrain than normal during the mounting procedure. x4.8.

bodies to GluRl and 2 / 3 , only (caudate-putamen, nucleus accumbens, and subthalamic nucleus; Figs. 2 , 3, 6, 7, lld-0. However, when only neurons were considered, i.e., excluding the neuropil, some nuclei had a different staining pattern. Thus, neurons in the substantia nigra stained moderately dense with antibodies to GluR1 and 213 only. In the caudate-putamen, a dense staining of most neurons was found with antibodies to GluR2/3 only (Fig. l l e ) . The dense staining of the caudate-putamen with antibody to GluRl was due largely to the dense neuropilar staining as well as to scattered densely stained neurons (Fig. l l d ) . There was no apparent morphological difference between the densely and lightly staining neurons. Some of the densely staining neurons were clearly members of neuron aggregates (Paskevich et al., '911, although dense staining of two adjoining neurons of an aggregate was uncommon. A

similar pattern of staining with antibody to GluRl was seen in the nucleus accumbens.

Thalamus The habenula stained densely with all antibodies (Fig. 31, due to staining of both the neurons and neuropil. Most other nuclei in the thalamus had a light to moderate staining with all the antibodies (Fig. 3). Overall staining was lightest with antibody to GluR1, as evident in the posterior nucleus and medial geniculate nucleus. Exceptions were the ventral part of the lateral geniculate nucleus and the reticulothalamic nucleus (Spreafico et al., '911, in which staining was moderately dense with all antibodies. The reticulothalamic nucleus was the densest stained thalamic nucleus (excepting the habenula) with antibody to GluR4.


341

Fig. 8. Coronal sections of parietal cortex, area 1, layer V and adjacent portions of layers IV (top of micrograph) and VI (bottom of micrograph), immunolabeled with antibodies to GluRl (a),213 (b),and 4 ( c ) .NP, non-pyramidal cell; P, pyramidal cell; Pd, pyramidal cell apical dendrite. X 230.

Hypothalamus

Forebrain-white matter

Overall staining in the hypothalamic nuclei was low to moderate with all antibodies, with antibody to GluRl giving the lightest staining. Exceptions were the dense staining in the supraoptic nucleus and the junction of the median eminence and infundibulum when immunolabeled with antibody to GluR2/3 and in the ventromedial nucleus with antibodies to GluRl and 2/3. Also, the arcuate nucleus was densely stained with all antibodies (Fig. 3). The dense staining in the supraoptic, ventromedial, and arcuate nuclei was due to staining of both neuropil and neuron cell bodies.

Little or no staining was found in most white matter structures immunolabeled with antibody to GluRl (Figs. 2a, 3a, 6a). A few small cells, probably astrocytes, were stained in the optic chiasm with antibody to GluR1. Stained astrocyte-like cells were evident in the optic chiasm and anterior commissure with antibody to GluR4. White matter immunolabeled with antibodies to GluR2/3 or 4 contained some lightly stained structures, including undefined processes and occasional cell bodies. The increased staining of the posterior part of the anterior commissure compared to the anterior part was due to diffuse staining.

Retina The immunolabeling of cells with all antibodies was denser in the ganglion cell and inner nuclear layers than in the outer nuclear layer. However, staining of the ganglion cell bodies was obscured slightly by an “edge effect.”

Brainstem-sensory The superior colliculus stained lightly overall (Figs. 4, 6). The slightly denser staining in the superficial gray was due to a diffuse neuropilar stain, since staining of the neurons

Fig. 9. Coronal sections of the main olfactory bulb, immunolabeled with antibodies to GluRl (a,d,e), 213 (b), and 4 ( c ) .EP, external plexiform layer; G1, glomerular layer; Gr, granule cell layer; IP, internal

plexiform layer; Mi, mitral cell; Pg, periglomerular cell; SA, short axon cells of granule cell layer; (arrowheads),neuron cell bodies of EP. a-c, x 165; d,e, ~ 4 6 5 .


Fig. 10. Coronal sections of the CA1 (a-c) and CA3 (d-fl regions of the hippocampus immunolabeled with antibodies to GluRl (a,d), 213 (b,e), and 4 (c,f). NP, non-pyramidal cell; P, pyramidal cell; Pd, pyramidal cell apical dendrite; SO, stratum oriens. ~ 2 8 0 .

343

Figure 11

IMMUNOCYTOCHEMISTRY OF GLUTAMATE RECEPTORS in this layer was similar to that of neurons of other layers. The large multipolar neurons of the lateral deep gray (Tokunaga and Otani, '76) were densest with antibodies to GluR2/3 and 4. Smaller neurons were not studied. Most sensory nuclei stained lightest when immunolabeled with antibody to GluRl and densest with 2/3 (Figs. 5, 7). Notable exceptions included the external cuneate (Fig. 5g-i) in which staining with all antibodies was moderately dense; the medial solitary nucleus which stained moderately with all antibodies (Figs. 5g-i, llg-i); and the outer two layers of the dorsal cochlear nucleus (Fig. 71, which stained more intensely with antibody to GluRl than with 213 and 4. The denser staining of the external cuneate (for all antibodies) and dorsal cochlear (GluR1) nucleus was found in both neurons and neuropil. In the medial solitary nucleus, the neuron cell bodies were moderately stained when immunolabeled with antibody to GluRl and 213 and lightly stained with antibody to GluR4. However, these neurons were most evident with antibody to GluRl due to the lighter staining neuropil.

345

to be glia), while the dorsal spinocerebellar tract contained scattered, lightly stained, undefined processes only. In contrast, many nerve fibers appeared to be stained in both tracts when immunolabeled with antibody to GluR2/3. In addition, a few small cell bodies were stained in the tract of nerve V.

Cerebellum

Bergmann glia of the molecular layer stained densely when immunolabeled with antibodies for GluRl (Fig. 12a) and GluR4 (Fig. 12c,d). Usually, Bergmann glia were not labeled with antibody to GluR2/3 (Fig. 12b) although there was occasional, light staining of the latter in limited and possibly random regions of the cerebellum. Purkinje cell bodies and dendrites stained densely with antibody to GluR2/3 only (Fig. 12b,e).Medium-sizedneurons including those of the molecular layer (Fig. 12b; basketlstellate), Golgi cells (Fig. 12b), and Lugaro cells (Fig. 12e; horizontal cells in the outer granular layer; Aoki et al., '86; Sahin and Hockfield, '90) stained light to moderate when immunolabeled with antibody to GluR2/ 3. Occasionalexamples of the Brainstem-motor latter two cell types were identified with antibody to GluR1, The overall staining pattern was similar to that of the while no distinctive medium-sized neurons were identified sensory nuclei, i.e., denser staining with antibodies to with antibody to GluR4. The white matter had little or no GluR2/3 and 4 than to GluRl (Figs. 4, 5 , 6). The level of stain with antibody to GluR1, while some nerve fibers or staining was similar for all antibodies in the oculomotor astrocyte-like cells stained with antibodies to GluR2/3 or 4, (Fig. 4) and facial nuclei (Fig. 6), only. The medial solitary respectively (Fig. 12a-c) . nucleus (sensory) was stained slightly denser than the Electron microscopy-hippocampus adjacent dorsal motor nucleus of the vagus (forms the ventral boundary of the medial solitary nucleus; see Staining of all sections was limited to the edges of the Bystrzycka and Nail, '851, hypoglossal nucleus (motor), and section (Fig. 13; see Piekut, '86). Staining was consistently nucleus of Roller (reticular), with antibody to GluRl (Fig. present in small dendrites and Type 1 synapses with 5g, Ilg). The opposite pattern was seen with antibodies to asymmetric densities with all antibodies, although less GluR2/3 and 4 (Fig. 5h,i). This increased density of the common and with lighter staining with antibody to GluR4 latter 3 nuclei was due largely to groups of densely stained (Fig. 14). Typically, the stain filled the dendritoplasm, neuron cell bodies and larger processes of the neuropil, as coating the microtubules and mitochondria, which were illustrated for the dorsal motor nucleus of the vagus (Fig. never stained (Fig. 14a,d). Staining was more spotty in the llh,i). larger dendrites and neuron cell bodies (Fig. 13). Often, staining covered patches of rough endoplasmic reticulum, Brainstem-reticularcore cytoplasmic matrix, and outer nuclear envelope. Cell bodies Unlike the sensory and motor nuclei, structures which of the pyramidal layer were most densely stained when stained more densely when immunolabeled with antibody immunolabeled with antibody to GluR2/3 (Fig. 13) and to GluRl than with 4 were common in the reticular core. slightly less with antibody to GluR1. Cells of the CA1-2 These included the reticulotegmental nucleus (Fig. 5a-c), region stained lightly with antibody to GluR4, while those inferior olive (Fig. 5g-i), and red nucleus (Fig. 4), of which of CA3 had little specific staining. the last was equally dense with all antibodies. The densest Synaptic labeling was limited to postsynaptic densities, staining structure was the interpeduncular nucleus (Fig. 4) adjacent to unstained synaptic clefts and presynaptic termiwhich was filled with a plexus of densely staining processes, nals which were filled with round or round and pleomorphic in addition to having scattered, dense-staining neurons and vesicles (Fig. 14). Occasional unstained Type 1 synapses, dense-staining, diffuse neuropilar staining. most common with antibody to GluR4, were easily distinguished from stained synapses in most cases (Fig. 14c).We Brainstem-white matter did not identify the source of small and medium-sized The dorsal spinocerebellar (Fig. 5g) and Vth nerve (Fig. stained synapses, which were found in all regions and with 5d,g) tracts were unstained when immunolabeled with all antibodies. In the CA3 region, postsynaptic densities of antibody to GluR1. With antibody to GluR4 the tract of active zones of the specialized mossy fiber terminals, originerve V contained numerous small stained cells (presumed nating from the granule cells of the dentate gyms (e.g., Sandler and Smith, '91), were stained for all antibody types (Figs. 13a; 14d,e). These showed the typical mossy fiber terminal structure with multiple active zones along the Fig. 11. Coronal sections of dorsocaudal portion of lateral septal giant specialized spines of the CA3 pyramidal cell apical nucleus, rostral to the sections in Figure 2 (a-c), caudate-putamen dendrites (Blackstadand Kjaerheim, '61; Amaral and Dent, (d-f), and at level of inferior olive (g-i), showing the full dorsoventral '8 1; Gaarskjaer, '86). The unstained presynaptic terminal extent of the dorsal motor nucleus of the vagus (10)and an adjacent portion of the medial nucleus of the solitary tract (Sl).GluRl (a,d,g); was very large and irregular and filled with synaptic vesicles GluR2/3 (b,e,h);GluR4 (c,f,i). Dorsal is to top and lateral to left in all as well as one or more large, dense-cored vesicles. Staining of postsynaptic densities of mossy fiber terminals and micrographs. x228.

346


Fig. 12. Sagittal sections of the cerebellar cortex immunolabeled with antibodies to GluRl (a),213 (b,e);and 4 (c,d).As, astrocyte-like cells; BG, Bergmann glial processes; Go, Golgi cell; Gr, granular layer;

L, Lugaro cell; Mo, molecular layer; Pj, Purkinje cell body; WM, white matter; (small arrow), Purkinje cell dendrite; (asterisk), Bergmann glia cell body; (arrowhead),molecular layer cell. a-c, x230; d,e, ~ 7 2 5 .

adjacent apical dendrites was light with antibodies to GluR4. There was little evidence of immunolabeling in other structures. While we saw no definitive, dense labeling of presynaptic terminals, there was occasional light staining in presynaptic terminals (Fig. 14b) and staining in axons. No definitive presynaptic or nonsynaptic staining of the cell membrane was discerned although there was occasional staining of the cell membrane adjacent to unidentified cell processes. Many stained processes could not be identified

and could be either dendritic spines or fine glial processes. Perivascular processes, probably representing glia, were stained occasionally (GluR4). However, no good example of stained oligodendrocytes, astrocytes, or other glial cells was found.

Electron microscopy-cerebral cortex The staining pattern was similar to that found in the hippocampus with most stain limited to dendrites and the


Fig. 13. Electron micrograph of cell bodies of CA3 (a)and CA1-2 (b) regions from the same section of hippocampus, immunolabeled with antibodies to GluR213. Cell bodies near the surface of the vibratome (microslicer) section (outer side margins of micrographs) show heavy labeling specific to the cytoplasm, with no staining in the nucleus (Nu).

347

Cell bodies deeper within the vibratome section show little or no labeling (UC). Dorsal is to the top. DS, specialized spines of CA3 pyramidal cell apical dendrites; MF, mossy fiber terminal; Nol, nucleolus; (arrowhead), stained synapse. x 10,000.

348


Figure 14

IMMUNOCYTOCHEMISTRY OF GLUTAMATE RECEPTORS postsynaptic densities of small and medium-sized Type 1 synapses (Fig. 15). Some stained synapses contained perforated densities (Fig. 15d).As in the hippocampus, many fine stained processes could not be identified as to neuronal or glial origin. However, at least one example was found of a stained glial process which enveloped a synapse possessing a stained postsynaptic density (Fig. 15e; GluR4).

DISCUSSION Comparison to in situ hybridization histochemistry studies of AMPA receptor subunits The overall pattern of AMPA receptor subunit distribution in our study resembles that seen in previous in situ hybridization studies (Bettler et al., '90; Boulter et al., '90; Gall et al., '90; Keinanen et al., '90; Pellegrini-Giampetroet al., '91). Also, the immunolabeling pattern resembles results of in situ hybridization studies on the two alternative splice variants for GluR1-4 (flip and flop; Sommer et al., '90; Monyer et al., '91), since our antibodies are made from peptides based on the carboxy termini and therefore would be expected to recognize both flip and flop versions of GluR subunits. As in all of these studies, labeling is densest in many forebrain structures, such as the hippocampus and olfactory bulb, and in the cerebellum. The greatest difference between densely and lightly stained structures is found with antibody to GluR1, while immunolabelingdistribution with antibodies to GluR2/3, i.e., the combined distributions of GluR2 and 3, is the most extensive, and that of GluR4 is the most diffuse (compare to Fig. 4 of Keiniinen et al., '90). In the parietal cortex, we found that layers II-IV were stained notably lighter than V-VI with antibody to GluRl. Similarly, Pellegrini-Giampetro et al. ('9 1)noted a slightly lower hybridization with probe to GluRl in layers III-IV of the parietal cortex, and Keinben et al. ('90) noted lower hybridization in III-IV throughout the cortex. Also, lower hybridization is evident in these layers with probe to GluRl in Figure 1of Gall et al. ('90). In our study, all layers were densely labeled with antibody to GluR2/3, which also produced the overall densest staining of neocortical neurons. In comparison, Keinanen et al. ('90) noted strong hybridization in all layers and Pellegrini-Giampetro et al. ('91) described the highest hybridization of neocortical neurons, with probe to GluR2. In the olfactory regions, our findings match the high levels of hybridization with probe to GluRl seen in the anterior olfactory nucleus, piriform (olfactory) cortex, and periglomerular and mitral cells of the olfactory bulb (Gall et al., '90). Other similar findings

Fig. 14. Electron micrographsof hippocampus immunolabeled with antibodies to GluRl (CA3; a,d,e), 2/3 (CA1-2;b), and 4 (CA1-2; c ) . Labeled postsynaptic densities (arrowheads) appose small and medium sized presynaptic terminals, which may contact the same dendrite (a; synapses sl, s2, and s3). Mossy fiber terminals (MF) form synapses on the immunolabeled, specialized spines (DS;e) and rarely the main shaft (d) of CA3 pyramidal cell apical dendrites. Region in b (GluR2/3) is about 25 pm ventral to the cell bodies of CA1-2 in Figure 13b. mi, mitochondrion in labeled dendrite; pre, possible presynaptic labeling (rare); sv, synaptic vesicles; uns, unstained synapse; (small arrow), dense-cored vesicle of mossy fiber terminal;(largearrow),active zone of mossy fiber terminal;(asterisk),point where base of dendritic spine and presynaptic mossy fiber terminal overlap in the plane of the section. ~40,000.

349

include substantial hybridization in the inferior olive, interpeduncular nucleus, and some septal nuclei with probe to GluRl (Gall et al., '90)' the caudate-putamen and amygdala with probe to GluRl (Gall et al., '90; Keinanen et al., '90) and 2/3 (Keinanen et al., 'go), and low levels in the globus pallidus (Gall et al., '90; Pelligrini-Giampetroet al., '91) and thalamus (Boulter et al., '90; KeinSinen et al., '90; Pellegrini-Giampetro et al., '91). The presence of moderately dense staining in the ventral lateral geniculate nucleus matches the level of hybridization with probe to GluRl found by Gall et al. ('90). Labeling of the reticulothalamic nucleus was significantly denser than most other thalamic nuclei with antibody to GluR4, as noted for probe hybridization by Bettler et al. ('90) and Keinanen et al. ('90). In our study, the habenula stained densely with all antibodies, as described for the hybridization with probe to GluRl by Gall et al. ('901, for probe to GluR4 by Bettler et al. ('901, and as evident for probes to GluRl and 4 in Figure 4 of Keinanen et al. ('90; see also Fig. 4 of Boulter et al., '90 and Fig. 10A of Heinemann et al., '91). When the cerebellum was immunolabeled with antibodies to GluRl and 4, staining was strongest in the Bergmann glia as in the in situ hybridization studies. We found dense staining in the Purkinje cells with antibodies to GluR2/3 and relatively light staining with antibodies to GluRl and 4. In comparison, both Keinanen et al. ('90) and PellegriniGiampetro et al. ('91) found high levels of hybridization in Purkinje cells with probes to GluR1,2, and 3, while hybridization in Purkinje cells with probes to GluR4 was apparently not definitive in Keinanen et al. ('90) and Bettler et al. ('90). Staining of other cell types in the cerebellum including granule cells was similar to that described in the in situ hybridization studies. The light staining that we found in the white matter of the cerebellum and other regions with antibodies to GluR2/3 and 4 is consistent with hybridization patterns evident in the published micrographs of other studies (Fig. 4 of Keinanen et al., '90; Fig. 6 of PellegriniGiampetro et al., '91). However, there are a few differencesbetween the distribution of immunolabeling and in situ hybridization. Highest immunolabeling in the retina with antibody to Glum was in the ganglion cell and inner nuclear layers, as described for hybridization studies in 21-day postnatal rats (Bettler et al., '90). However, we did not discern the stronger labeling in the amacrine cells of the inner nuclear layer noted by these authors. The difference may reflect age-related changes in expression, as described for numerous structures of the brain (Bettler et al., '90; Monyer et al., '91; Pellegrini-Giampetroet al., '91). In our studies, the entorhinal cortex stained densely with all antibodies. In comparison, Keiniinen et al. ('90) reported low levels with probe to GluR1, although their Figure 4 illustrates a strong but somewhat restricted hybridization pattern. The only major discrepancy between our study and the published in situ hybridization studies is our staining of both pyramidal and nonpyramidal cells in the CA3 region with antibody to GluR4. Bettler et al. ('90) and Keinanen et al. ('90) found very low hybridization for probe to GluR4 in the CA3 region. Sommer et al. ('90) found the flop variant of GluR4 expressed in the nonpyramidal cells of the CA3 region, but did not find either flip or flop in the pyramidal cells of this region. We suggest that the definite although light staining in the pyramidal cells of the CA3 region with antibody to GluR4 may represent either receptor molecules produced from a level of mRNA so low that it can not be detected

350


Fig. 15. Electron micrographs of cerebral cortex immunolabeled with antibodies to GluRl (a,d),213 (b), and 4 (c,e).mi, mitochondrion in labeled dendrite; sv, synaptic vesicles; uns, unstained synapse; (arrowhead), stained synaptic density; (arrow), perforated density; (asterisks), glial process wrapping a stained synaptic terminal. ~40,000.

IMMUNOCYTOCHEMISTRY OF GLUTAMATE RECEPTORS above background or the detection of additional forms of GluRs, not yet identified,that have a C-terminal amino acid sequence identical to that of GluR4. Also, it is possible that the staining of pyramidal cells of CA3 with antibody to GluR4 is nonspecific, although our controls were completely negative and specific localization was seen at the ultrastructural level.

Comparison to 3H-AMPAbinding studies Studies on the distribution of 3H-AMPA binding in the rat brain (Monaghanet al., '84;Rainbow et al., '84; Olsen et al., '87; Insel et al., '90) show many similarities to our results as well as to the in situ hybridization results (discussed in Keinanen et al., '90). Examples include a very dense binding in the induseum griseum (Monaghan et al., '84); dense binding in the anterior olfactory nucleus, lateral nucleus of the septum, pars dorsalis (Monaghanet al., '841, piriform cortex (Monaghan et al., '84; Insel et al., '901, basolateral arnygdala nucleus (Insel et al., 'go), and subiculum (Rainbowet al., '84); moderate binding in the isocortex (Monaghan et al., '84; Olsen et al., '87; Insel et al., '90); and light binding in the medial geniculate nucleus (Monaghan et al., '84; Rainbow et al., '84). Some nuclei which stained densely in our study, such as the interpeduncular nucleus, were not examined in the 'H-AMPA binding studies in the rat brain, although a dense binding of 3H-AMPA in the latter nucleus was noted in an electric fish (Maler and Monaghan, '91). Notable differences between our results and the 3HAMPA binding results on the rat brain include moderate binding in the olfactory bulb (Monaghan et al., '84; Insel et al., '90) and light binding in the cerebellum (Monaghan et al., '84; Rainbow et al., '84; see also Henley and Barnard; ,911, compared to the dense staining seen in our study and strong hybridization seen in the in situ hybridization studies. Nevertheless, in both structures the pattern of labeling in the 3H-Ah4PAbinding studies is similar to ours with the densest binding in the external plexiform layer of the olfactory bulb (Insel et al., '90) and the molecular layer ofthe cerebellum (Monaghan et al., '84). Another difference is a light binding in the habenula (Monaghan et al., '84), which was densely stained in our study. Perhaps these differences reflect in part multiple receptor populations labeled by AMPA with varying affinities (Olsen et al., '87). Binding in the hippocampus was very dense in all the 3H-AMPA studies, although it was consistently higher in the CA1 region than in the CA3 region. Thus, there are discrepancies among the immunolabeling,in situ hybridization, and 3H-AMPAbinding studies for the hippocampus. Neither explanation suggested by Keinanen et al. ('go), i.e., either presynaptic autoreceptors or formation of pharmacologicallyuncharacterized receptors, seems to be adequate to account for all these differences.

Ultrastructural localization of AMPA receptors Receptors in the neuron can be classified into internal and surface membrane receptors, as described for acetylcholine receptors (Pestronk, '85). Immunolabeling with receptor antibodies in the cell body and dendrites was common in our study and has been described in numerous other studies (reviewed by Wenthold et al., '90a). Typically, immunolabeling in our study had a spotty distribution throughout the cytoplasm with notable accumulations on the outer membrane of mitochondria and nucleus and on

351

microtubules,as described for other neurotransmitter receptors (Strader et al., '83; Aoki et al., '87; Juiz et al., '89). The vacant appearance of some of the finer stained processes, including some possible glial processes, has been illustrated in other studies (e.g., Strader et al., '83; Aoki et al., '87; Juiz et al., '89) and presumably reflects the effect of the immunolabeling procedures on the sparse cytoplasm of the processes (Peters et al., '91). While it is probable that at least part of the internal labeling represents receptors in transport to and from the cell membrane, many of these molecules may never associate with the cell surface (Pestronk, '85; also reviewed in Jacob et al., '86). However, it is not known if the latter represent defective receptors, a reserve pool (Jacob et al., '86) or fundionly different molecules (Pestronk, '85). In contrast to the widespread localization of internal receptors in our study, plasma membrane receptors seemed to be localized at the postsynaptic densities. We saw no definitive evidence of nonsynaptic localization of receptors in the plasma membrane, although their existence has been suggested (Fagg and Matus, '84; Jacob et al., '86; Dechesne et al., '90). Synaptic labeling in our study was found in the postsynaptic density and seemed to be confined to the intracellular side of the membrane. A similar distribution is seen with some antibodies to other types of receptors including glycine, GABAIbenzodiazepine, and P-adrenergic receptors (reviewed by Wenthold et al., '90a). While this may be indicative of an intracellular location for the antigenic site (C-terminal) on the functional receptor molecule, it is equally possible that dense staining in the postsynaptic density represents receptor molecules not yet integrated into the postsynaptic membrane. In this case, a smaller number of extracellular antigenic sites of the membrane-bound receptor molecules may not be evident with our technique, but may be revealed in future studies with immunogold techniques. Typically, the presynaptic terminal adjacent to an immunolabeled postsynaptic density was unlabeled and contained round or round and pleomorphic vesicles. While this variation may represent a range of shape of a single vesicle type, it is possible that there are different vesicle types associated with different fiber inputs, as described for glutamatergic mossy fiber terminals in the rat cerebellum (Hamori et al., '90). Other indications of immunolabeling of multiple inputs was the variety of sizes and morphology of labeled synapses, including specialized synapses such as mossy fiber terminals of the hippocampus and synapses with perforated densities or discrete glial wrappings in the cerebral cortex. Perforated densities, associated with input plasticity (e.g., Geinisman et al., '86; Peters, '87; Geinisman et al., '891, and specialized glial wrappings, uncommon in the cerebral cortex (Peters et al., '911, may be indicativeof specificinputs, as described for glial-wrapped synapses in the piriform cortex (Haberly and Behan, '83;Haberly, '90). We found no definitive evidence of presynaptic labeling. The rare examples of lightly stained presynaptic terminals may represent a subset of terminals with presynaptic AMPA receptors. However, their presynaptic membranes appeared unlabeled, and it is more likely that these terminals were stained nonspecifically. This is consistent with previous work which supports only the presence of the 3H-kainatebinding (Griesser et al., '82; Repressa et al., '87; Miller et al., '90) and L-AP4 antagonized (Cotman et al., '86) types of possible presynaptic excitatory amino acid receptors. Assuming this is true, then the occasionalimmu-

352


nolabeling in axons seen at both light and electron microscope levels may represent either an accidental acquisition as suggested for neuronal nicotinic receptors (Swanson et al., '87) or a functionally different molecule as discussed above.

Patterns of distribution of AMPA receptor subunits It is evident that AMPA receptor subunits have a wide distribution in the brain with the greatest densities being in the cerebellum and in structures involved in rhinencephalic circuitry (Carpenter and Sutin, '83), including the lateral mammillary bodies (Gall et al., '901, olfactory bulb and tubercle, septal nuclei, amygdaloid complex, hippocampus, stria terminalis, induseum griseum, habenula, and interpeduncular nucleus. The presence of different subunits in the postsynaptic density of the mossy fiber terminal probably reflects the heteromeric nature of the AMPA receptor complex (Keinanen et al., '90; Nakanishi et al., '90; Sommer et al., '90; Wenthold et al., '92). However, our results may represent different types of mossy fiber terminals instead. Future studies using postembedding immunolabeling of thin sections should clarify whether different subunits colocalizein the same synapse. Another possibility is that different subunits are expressed in the same neuron but are localized at different terminals, as suggested for GABA receptors (Juiz et al., '89). Since it has been demonstrated that different neural inputs can activate different glutamate receptors (NMDA and non-NMDA) on the same neuron (Mooney and Konishi, '911, the same could be true for different subunits of AMPA receptors. Probably the best place to study this phenomenon is along neurons on which different neural inputs are segregated into discrete termination zones, such as the pyramidal cells of the hippocampus (Bayer, '85).

AMPA receptors in nonneuronal cells Western blots showed little evidence of GluR1-4 outside of the central nervous system. However, these antibodies, in particular the antibody to GluR4, labeled lower molecular weight proteins in kidney, liver, and skeletal muscle. Since these proteins are much smaller than the GluR subunits, their labeling is probably due to crossreactivity and does not reflect the presence of the receptor in these tissues. Within the brain, labeled cells included neurons, Bergmann glia of the cerebellum and astrocyte-like cells in several regions. The function of glial glutamate receptors is currently the subject of intense research (e.g., Usowicz et al., '89; Cornell-Bell et al., '90a,b; Jensen and Chiu, '91). While we found only light staining in astrocyte-like cells with antibody to GluR4, receptors for GluRl and 4 localized preferentially on the Bergmann glia compared to other structures in the cerebellum in our study and the in situ hybridization studies. Other studies have found a similar preferential localization of kainate binding protein in Bergmann glia of chicks and fish (Somogyi et al., '901, the a2 subunit of the GABAAreceptor in Bergmann glia of bovine cerebellum (Wisden et al., '891, and excitatory sulfurcontaining amino acids in Bergmann glia and astrocytic endfeet in the rat cerebellum (Cubnod et al.,'90; Streit et al., '91). Thus, these glia may play an active role in the neural circuitry of the cerebellar cortex.

ACKNOWLEDGMENTS We thank Dr. J. Fex, Dr. C. Hunter, Dr. E. Wada, and Dr. K. Wada for critically reading the manuscript and N. Alvanzo, M. Dennie and T. Vu for assistance in the preparation of the manuscript.

LITERATURE CITED Amaral, D.G. (1978) A Golgi study of cell types in the hilar region of the hippocampus in the rat. J. Comp. Neurol. 182:851-914. Amaral, D.G., and J.A. Dent (1981) Development of the mossy fibers of the dentate gyrus. 1. A light and electron microscopic study of the mossy fibers and their expansions. J. Comp. Neurol. 19551-86. Aoki, C., T.H. Joh, and V.M. Pickel (1987) Ultrastructural localization of p-adrenergic receptor-lie immunoreactivity in the cortex and neostriatum of rat brain. Brain Res. 437:264-282. Aoki, E., R. Semba, and S. Kashiwamata (1986) New candidates for GABAergic neurons in the rat cerebellum: An immunoqtochemical study with anti-GABAantibody. Neurosci. Lett. 68:267-271. Bayer, S.A. (1985) Hippocampal region. In G. Paxinos (ed): The Rat Nervous System,Vol. 1. New York: Academic Press, pp. 335-352. Bettler, B., J. Boulter, I. Hermans-Borgmeyer, A. O'Shea-Greenfield, E.S. Deneris, C. Moll, U. Borgmeyer, M. Hollmann, and S. Heinemann (1990) Cloning of a novel glutamate receptor subunit, GluR5: Expression in the nervous system during development. Neuron 5583-595. Blackstad, T.W., and A. Kjaerheim (1961) Special axo-dendritic synapses in the hippocampal cortex: Electron and light microscopic studies on the layer of mossy fibers. J. Comp. Neurol. 11 7:133-157. Bobryshev, I.V., E.A. Orlova, I.V. Balabanov,AS. Kuznetsov, T.M. Smirnova, and S.A. Dambinova (1989) The study of distribution of glutamate receptors in the motor cerebral cortexof the rats byimmunoelectronomicroscopictechnique [in Russian]. Tsitologiia 31~176-181. Boulter, J., M. Hollmann, A. O'Shea-Greenfield, M. Hartley, E. Deneris, C. Maron, and S. Heinemann (1990) Molecular cloning and functional expression of glutamate receptor subunit genes. Science 249:1033-1037. Brown, T.H., and A.M. Zador (1990) Hippocampus. In G.M. Shepherd (ed): The Synaptic Organization of the Brain, 3rd Ed. New York: Oxford University Press, pp. 346-388. Bystrzycka, E.K., and B.S. Nail (1985) Brain stem nuclei associated with respiratory, cardiovascular and other autonomic functions. In G. Paxinos (ed): The Rat Nervous System, Vol. 2. New York Academic Press, pp. 95-110. Cambray-Deakin, M.A., A.C. Foster, and R.D. Burgoyne (1990) The expression of excitatory amino acid binding sites during neuritogenesis in the developingrat cerebellum. Dev. Brain Res. 54:265-271. Carpenter, M.B., and J. Sutin (1983) Human Neuroanatomy, 8th Ed. Baltimore: Williams and Wilkins. Collingridge,G.L., and W. Singer (1990) Excitatoryamino acid receptors and synaptic plasticity. Trends Pharmacol. Sci. 11:290-296. Cornell-Bell, A.H., S.M. Finkbeiner, M.S. Cooper, and S.J. Smith (1990a) Glutamate induces calcium waves in cultured astrocytes: Long-range glial signaling. Science247:470473. Cornell-Bell, A.H., P.G. Thomas, and S.J. Smith (1990b) The excitatory neurotransmitter glutamate causes filopodia formation in cultured hippocampal astrocytes. Glia 3~322-334. Cotman, C.W., J.A. Flatman, A.H. Ganong, and M.N. Perkins (1986) Effects of excitatory amino acid antagonists on evoked and spontaneous excitatory potentials in guinea-pig hippocampus. J. Physiol. 378~403415. Cubnod, M., K.Q. Do, P. Grandes, P. Morino, and P. Streit (1990) Localization and release of homocysteic acid, an excitatory sulfur-containing amino acid. J. Histochem. Cytochem. 38:1713-1715. Dechesne, C.J., M.D. Oberdorfer, D.R. Hampson, K.D. Wheaton, A.J. Nazarali, G. Goping, and R.J. Wenthold (1990)Distribution of a putative kainic acid receptor in the frog central nervous system determined with monoclonal and polyclonal antibodies: Evidence for synaptic and extrasynaptic localization. J. Neurosci. 10:479490. Egebjerg, J., B. Bettler, I. Hermans-Borgmeyer, and S. Heinemann (1991) Cloningof a cDNA for a glutamate receptor subunit activated by kainate but not AMPA. Nature 351:745-748. Fagg, G.E., and A. Matus (1984) Selective association of N-methyl aspartate and quisqualate types of L-glutamate receptor with brain postsynaptic densities. Proc. Natl. Acad. Sci. U S A . 81 ~6876-6880. Friedrich, V.L., and E. Mugnaini (1981) Electron microscopy: Preparation of

IMMUNOCYTOCHEMISTRY OF GLUTAMATE RECEPTORS neural tissues for electron microscopy. In L. Heimer and M.J. Robards (eds): Neuroanatomical Tract-Tracing Methods. New York Plenum Press, pp. 345-375. Gaarskjaer, F.B. (1986) The organization and development of the hippocampal mossy fiber system. Brain Res. Rev. 1k335-357. Gall, C., K. Sumikawa, and G. Lynch (1990) Levels of mRNA for a putative kainate receptor are affected by seizures. Proc. Natl. Acad. Sci. U.S.A. 87:7643-7647. Geinisman, Y., F. Morrell, and L. de Toledo-Morrell (1989) Perforated synapses on double-headed dendritic spines: A possible structural substrate of synaptic plasticity. Brain Res. 480:326-329. Geinisman, Y.,L.de Toledo-Morrell, and F. Morrell(1986) Aged rats need a preserved complement of perforated axospinous synapses per hippocampal neuron to maintain good spatial memory. Brain Res. 398:266-275. Griesser, CA.V., M. Cuenod, and H. Henke (1982) Kainic acid receptor sites in the cerebellum of nervous, Purkinje cell degeneration, reeler, staggerer and weaver mice mutant strains. Brain Res. 246:265-271. Haberly, L.B. (1990) Olfactory cortex. In G.M. Shepherd (ed): The Synaptic Organization of the Brain, 3rd Ed. New York Oxford University Press, pp. 317-345. Haberly, L.B., and M. Behan (1983) Structure of the piriform cortex of the opossum. 111. Ultrastructural characterization of synaptic terminals of association and olfactory bulb afferent fibers. J. Comp. Neurol. 219:448460. H h o r i , J., J. TakAcs, and P. Petrusz (1990) Immunogold electron microscopic demonstration of glutamate and GABA in normal and deafferented cerebellar cortex: Correlation between transmitter content and synaptic vesicle size. J.Histochem. Cytochem. 38:1767-1777. Heinemann, S., B. Bettler, J. Boulter, E. Deneris, G. Gasic, M. Hartley, M. Hollman, T.E. Hughes, A. O'Shea-Greenfield, and S. Rogers (1991) The glutamate receptor gene family. In B.S. Meldrum, F. Moroni, P. Simon, and J.H. Woods (eds): Excitatory Amino Acids. New York Raven Press, pp. 109-133. Henley, J.M., and E.A. Barnard (1991) Comparison of solubilized kainate and a-amino-3-hydroxy-5-methyliixazalepropionatebinding sites in chick cerebellum. J. Neurochem. 56:702-705. Hollmann, M., M. Hartley, and S. Heinemann (1991) Caa+permeability of KA-AMPA-gated glutamate receptor channels depends on subunit composition. Science2522451-853. Houamed, K.M., J.L. Kuijper, T.L. Gilbert, B.A. Haldeman, P.J. O'Hara, E.R. Mulvihill, W. h e r s , and F.S. Hagen (1991) Cloning, expression, and gene structure of a G proteincoupled glutamate receptor from rat brain. Science252:1318-1321. Insel, T.R., L.P. Miller, and R.E. Gelhard (1990) The ontogeny of excitatory amino acid receptors in rat forebrain-I. N-methyl-D-aspartate and quisqualate receptors. Neuroscience 3Z31-43. Jacob, M.H., J.M. Lindstrom, and D.K. Berg (1986) Surface and intracellular distribution of a putative neuronal nicotinic acetylcholine receptor. J. Cell Biol. 103:205-214. Jensen, A.M., and S.Y. Chiu (1991) Differential intracellular calcium responses to glutamate in type 1and type 2 cultured brain astrocyh. J. Neuroscience 11:1674-1684. Juiz, J.M., R.H. Helfert, R.J. Wenthold, A.L. De Blas, and R.A. Altachuler (1989) Immunocytochemical localization of the GABA,/benzodiazepine receptor in the guinea pig cochlear nucleus: Evidence for receptor localization heterogeneity. Brain Res. 504:173-179. Keiniinen, K., W. Wisden, B. Sommer, P. Werner, A. Herb, T.A. Verdoorn, B. Sakmann, and P.H. Seeburg (1990) A family of AMPA-selective glutamate receptors. Science249:556-560. Krettek, J.E., and J.L. Price (1978) A description of the amygdaloid complex in the rat and cat with observations on intra-amygdaloid axonal connections. J. Comp. Neurol. 178:255-280. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4.Nature 227:680-685. Maler, L., and D. Monaghan (1991) The distribution of excitatory amino acid binding sites in the brain of an electric fish, Aptemnotus leptorhynchus. J. Chem. Neuroanat. 4:39-61. Masu,M., Y.Tanabe, K. Tsuchida, R. Shigemoto, and S. Nakanishi (1991) Sequence and expression of a metabotropic glutamate receptor. Nature 349:760-765. Mayer, M.L., and R.J. Miller (1990) Excitatory amino acid receptors, second messengers and regulation of intracellular Caa+in mammalian neurons. Trends Pharmacol. Sci. llr254-260. Miller,L.P.,A.E. Johnson,R.E. Gelhard,andT.R.Inse1(1990)Theontogeny of excitatory amino acid receptors in the rat forebrain-I1 Kainic acid receptors. Neuroscience 35:45-51.

353

Monaghan, D.T. (1991) Agonist- and antogonist-preferring N-methyl-Daspartate receptors and independent activation of their ion channels. In B.S. Meldrum, F. Moroni, R.P. Simon, and J.H. Woods (eds):Excitatory Amino Acids. NewYork: Raven Press, pp. 20>211. Monaghan, D.T., R.J. Bridges, and C.W. Cotman (1989) The excitatory amino acid receptors: Their classes, pharmacology, and distinct properties in the function of the central nervous system. Annu. Rev. Pharmacol. Toxicol. 29:365402. Monaghan, D.T., D. Yao, and C.W. Cotman (1984) Distribution of ['HI AMPA binding sites in rat brain as determined by quantitative autoradiography. Brain Res. 324~160-164. Monaghan, D.T., H.J. Olverman, L. Nguyen, J.C. Watkins, and C.W. Cotman (1988) Two classes of N-methyl-D-aspartaterecognition sites: Differential distribution and differential regulation by glycine. Proc. Natl. Acad. Sci. U.S.A. 85:983&9840. Monyer, H., P.H. Seeburg, and W. Wisden (1991) Glutamate-operated channels: Developmentally early and mature forms arise by alternative splicing. Neuron 6:799-810. Mooney, R., and M. Konishi (1991) T w o distinct inputs to an avian song nucleus activate different glutamate receptor subtypes on individual neurons. Proc. Natl. Acad. Sci. U.S.A. 88:40754079. Mori, K. (1987) Membrane and synaptic properties of identified neurons in the olfactory bulb. Prog. Neurobiol. 29:275-320. Nakanishi, N., N.A. Shneider, and R. Axel (1990) A family of glutamate receptor genes: Evidence for the formation of heteromultimeric receptors with distinct channel properties. Neuron 5569-581. Olsen, R.W., 0. Szamraj, and C.R. Houser (1987) [$HI AMPA binding to glutamate receptor subpopulations in rat brain. Brain Res. 402~243-254. Paskevich, P.A., H.K. Evans, and V.B. Domesick (1991) Morphological assessment of neuronal aggregates in the striatum of the rat. J. Comp. Neurol. 305:361-369. Patneau, D.K., and M.L. Mayer (1991) Kinetic analysis of interactions between kainate and AMPA: Evidence for activation of a single receptor in mouse hippocampal neurons. Neuron 6335-798, Paxinos, G., and C. Watson (1986) The Rat Brain in Stereotaxic Coordinates, 2nd Ed. New York AcademicPress. Pellegrini-Giampietro, D.E., M.V.L. Bennett, and R.S. Zukin (1991)Differential expression of three glutamate receptor genes in developingrat brain: An in situ hybridization study. Proc. Natl. Acad. Sci. U.S.A. 88:41574161. Pestronk, A. (1985) Intracellular acetylcholine receptors in skeletal muscles of the adult rat. J. Neurosci. 51111-1117. Peters, A. (1987) Synaptic specificity in the cerebral cortex. In G.M. Edelman, W.E. Gall, and W.M. Cowan (eds): Synaptic Function. New York John Wiley & Sons,pp. 373-397. Peters, A., S.L. Palay, and H. deF. Webster (1991) The Fine Structure of the Nervous System, 3rd Ed. New York: Oxford University Press. Piekut, D.T. (1986) Ultrastructural characteristics of peptidergic neurons using pre-embedding immunocytochemical methods. Am. J. Anat. 175: 197-216. Rainbow, T.C., C.M. Wieczorek, and S. Halpain (1984) Quantitative autoradiography of binding sites for PHI AMPA, a structural analogue of glutamic acid. Brain Res. 309:173-177. Represa, A., E. Tremblay, and Y. Ben-Ari (1987) Kainate binding sites in the hippocampal mossy fibers: Localization and plasticity. Neuroscience 20:739-748. Ribak,C.E., and L. Seress (1983)Five types of basket cell in the hippocampal dentate gyrus: A combined Golgi and electron microscopic study. J. Neurocytol. 12:577-597. Ribak, C.E., L. Seress, and D.G. Amaral(1985) The development, ultrastructure and synaptic connections of the mossy cells of the dentate gyrus. J. Neurocytol. 14:835-857. Sahin, M., and S. Hockfield (1990) Molecular identification of the Lugaro cell in the cat cerebellar cortex. J.Comp. Neurol. 301575-584. Sandler, R., and A.D. Smith (1991) Coexistence of GABA and glutamate in mossy fiber terminals of the primate hippocampus: A n ultrastructural study. J. Comp. Neurol. 303:177-192. Seress, L., and C.E. Ribak (1985) A combined Golgi-electron microscopic study of non-pyramidal neurons in the CAI area of the hippocampus. J. Neurocytol. 14:717-730. Shepherd, G.M., and C.A. Greer (1990) Olfactory Bulb. In G.M. Shepherd (ed):The Synaptic Organization of the Brain, 3rd Ed. New York Oxford University Press, pp. 133-169. Sommer, B., K. Keinhen, T.A. Verdoorn, W. Wisden, N. Burnashev, A. Herb, M. Kohler, T. Takagi, B. Sakmann, and P.H. Seeburg (1990) Flip

354 and flop:A cell-specific functional switch in glutamate-operated channels of the CNS. Science249:1580-1585. Somogyi, P., N. Eshhar, V.I. Teichberg, and J.D.B. Roberts (1990)Subcellular localization of a putative kainate receptor in Bergmann glial cells using a monoclonal antibody in the chick and fish cerebellar cortex. Neuroscience 359-30. Spreafico, R., G. Battaglia, and C. Frassoni (1991) The reticular thalamic nucleus (RTN) of the rat: Cytoarchitectural, Golgi immunocytochemical, and horseradish peroxidase study. J. Comp. Neurol. 304:478-490. Strader, C.D., V.M. Pickel, T.H. Joh, M.W. Strohsacker, R.G.L. Shorr, R.J. Lefkowitz, and M.G. Caron (1983) Antibodies to the P-adrenergic receptor: Attenuation of catecholamine-sensitive adenylate cyclase and demonstration of postsynaptic receptor localization in brain. Proc. Natl. Acad. Sci. U.S.A. 8Ot1840-1844. Streit, P., P. Grandes, P. Morino, P. Tschopp, and M. Cuenod (1991) Immunohistochemical localization of glutamate and homocysteate. In B.S. Meldrum, F. Moroni, R.P. Simon, and J.H. Woods (eds): Excitatory Amino Acids. New York: Raven Press, pp. 61-67. Swanson, L.W., D.M. Simmons, P.J. Whiting, and J. Lindstrom (1987) Immunohistochemical localization of neuronal nicotinic receptors in the rodent central nervous system. J. Neurosci. 7t3334-3342. Tokunaga, A., and K. Otani (1976) Dendritic patterns of neurons in the rat superior collieulus. Exp. Neurol. 52189-205. Towbin, H., T. Staehelin, and J. Gordon (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proc. Natl. Acad. Sci. U.S.A. 76:43504354. Usowicz, M.M., V. Gallo, and S.G. Cull-Candy (1989) Multiple conductance

R.S. PETRALIA AND R.J. WENTHOLD channels in type-2 cerebellar astrocytes activated by excitatory amino acids. Nature 339:380-383. Verdoorn, T.A., N. Burnashev, H. Monyer, P.H. Seeburg, and B. Sakmann (19911 Structural determinants of ion flow through recombinant glutamate receptor channels. Science 252:1715-1 718. Wenthold, R.J., R.A. Altschuler, and D.R. Hampson (1990a) Immunoeytochemistry of neurotransmitter receptors. J. Electr. Micr. Tech. 158196. Wenthold, R.J., C. Hunter, K. Wada, and C.J. Dechesne (1990b) Antibodies to a C-terminal peptide of the rat brain glutamate receptor subunit, GluR-A, recognize a subpopulation of AMPA binding sites but not kainate sites. FEBS Lett. 276:147-150. Wenthold, R.J., N. Yokotani, K. Doi, and K. Wada (1992) Immunochemicalcharacterization of the non-NMDA glutamate receptor using subunitspecific antibodies: Evidence for a hetero-oligomeric structure in rat brain. J. Biol. Chem. 267:501-507. Werner, P., M. Voigt, K. Keinanen, W. Wisden, and P.H. Seeburg (1991) Cloning of a putative high-affinity kainate receptor expressed predominantly in hippocampal CA3 cells. Nature 351:742-744. Wisden, W., L.A. McNaughton, M.G. Darlison, S.P. Hunt, and E.A. Barnard (1989) Differential distribution of GABA,, receptor mRNAs in bovine cerebellum-localization of a2 mRNA in Bergmann glia layer. Neurosci. Lett. 106:7-12. Zilles, K., and A. Wree (1985) Cortex: Areal and laminar structure. In G. Paxinos (ed):The Rat Nervous System, Vol. 1. New York: Academic Press, pp. 375-415.

Immunocytochemical localization of kynurenine aminotransferase in the rat striatum: a light and electron microscopic study.

Immunocytochemical localization of glutamate decarboxylase in rat spinal cord.

Immunocytochemical localization of glutamate decarboxylase in rat substantia nigra.

Immunocytochemical localization of argininosuccinate synthetase in the rat brain.

Light and electron microscopic immunocytochemical localization of glutamine synthetase in the superficial pineal gland of the rat.

Immunocytochemical localization of protein kinase C subspecies in the rat spinal cord: light and electron microscopic study.

Immunocytochemical localization of endopeptidase 24.15 in rat brain.

Immunocytochemical localization of phosphatase inhibitor-1 in rat brain.

Immunocytochemical localization of aromatase in the brain.

Cellular localization of a metabotropic glutamate receptor in rat brain.

Immunocytochemical localization of cyclic GMP: light and electron microscope evidence for involvement of neuroglia.

Co-localization of brain corticosteroid receptors in the rat hippocampus.

Immunocytochemical localization of muscarinic acetylcholine receptors in the rat endocrine pancreas.

Effect of colchicine on the immunohistochemical localization of somatostatin in the rat brain: light and electron microscopic studies.

Ultrastructural immunocytochemical observations on the localization, metabolism and transport of glutamate in normal and ischemic brain tissue.

Immunocytochemical localization of calretinin in the forebrain of the rat.

Immunocytochemical localization of octopamine in the central nervous system of Limulus polyphemus: a light and electron microscopic study.

Immunocytochemical localization of the alpha subspecies of protein kinase C in rat brain.

Immunocytochemical localization of cholinergic terminals in the region of the nucleus basalis magnocellularis of the rat: a correlated light and electron microscopic study.

The immunocytochemical localization of angiotensinogen in the rat ovary.

Presynaptic localization of histamine H3-receptors in rat brain.

The characterization and localization of the glutamate receptor subunit GluR1 in the rat brain.

A GABA immunocytochemical study of rat motor thalamus: light and electron microscopic observations.

Immunocytochemical localization of fibronectin (LETS proteins) on the surface of L6 myoblasts: light and electron microscopic studies.