Search

Transmembrane alpha-amino-3-hydroxyl-5-isoxazolepropionate (AMPA) receptor regulatory proteins (TARPs) play pivotal roles in AMPA receptor trafficking and gating. Here we examined cellular and subcellular distribution of TARP gamma-8 in the mouse brain. Immunoblot and immunofluorescence revealed the highest concentration of gamma-8 in the hippocampus. Immunogold electron microscopy demonstrated dense distribution of gamma-8 on the synaptic and extrasynaptic surface of hippocampal neurons with very low intracellular labeling. Of the neuronal surface, gamma-8 was distributed at the highest level on asymmetrical synapses of pyramidal cells and interneurons, whereas their symmetrical synapses selectively lacked immunogold labeling. Then, the role of gamma-8 in AMPA receptor expression was pursued in the hippocampus using mutant mice defective in the gamma-8 gene. In the mutant cornu ammonis (CA)1 region, synaptic and extrasynaptic AMPA receptors on dendrites and spines were severely reduced to 35-37% of control levels, whereas reduction was mild for extrasynaptic receptors on somata (74%) and no significant decrease was seen for intracellular receptors within spines. In the mutant CA3 region, synaptic AMPA receptors were reduced mildly at asymmetrical synapses in the stratum radiatum (67% of control level), and showed no significant decrease at mossy fiber-CA3 synapses. Therefore, gamma-8 is abundantly distributed on hippocampal excitatory synapses and extrasynaptic membranes, and plays an important role in increasing the number of synaptic and extrasynaptic AMPA receptors on dendrites and spines, particularly, in the CA1 region. Variable degrees of reduction further suggest that other TARPs may also mediate this function at different potencies depending on hippocampal subregions, input sources and neuronal compartments.

Following cell surface receptor activation, the alpha subunit of the Gq subclass of GTP-binding proteins activates the phosphoinositide signalling pathway. Here we examined the expression and localization of Gq protein alpha subunits in the adult mouse brain by in situ hybridization and immunohistochemistry. Of the four members of the Gq protein alpha subunits, Galphaq and Galpha11 were transcribed predominantly in the brain. The highest transcriptional level of Galphaq was observed in cerebellar Purkinje cells (PCs) and hippocampal pyramidal cells, while that of Galpha11 was noted in hippocampal pyramidal cells. Antibody against the C-terminal peptide common to Galphaq and Galpha11 strongly labelled the cerebellar molecular layer and hippocampal neuropil layers. In these regions, immunogold preferentially labelled the cytoplasmic face of postsynaptic cell membrane of PCs and pyramidal cells. Immunoparticles were distributed along the extra-junctional cell membrane of spines, dendrites and somata, but were almost excluded from the junctional membrane. By double immunofluorescence, Galphaq/Galpha11 was extensively colocalized with metabotropic glutamate receptor mGluR1alpha in dendritic spines of PCs and with mGluR5 in those of hippocampal pyramidal cells. Together with concentrated localization of mGluR1alpha and mGluR5 in a peri-junctional annulus on PC and pyramidal cell synapses (Baude et al. 1993, Neuron, 11, 771-787; Luján et al. 1996, Eur. J. Neurosci., 8, 1488-1500), the present molecular-anatomical findings suggest that peri-junctional stimulation of the group I metabotropic glutamate receptors is mediated by Galphaq and/or Galpha11, leading to the activation of the intracellular effector, phospholipase Cbeta.

The synaptic localization of the N-methyl-D-aspartate (NMDA) type glutamate receptor (GluR) channel is a prerequisite for synaptic plasticity in the brain. We generated mutant mice carrying the carboxy-terminal truncated GluR epsilon2 subunit of the NMDA receptor channel. The mutant mice died neonatally and failed to form barrelette structures in the brainstem. The mutation greatly decreased the NMDA receptor-mediated component of hippocampal excitatory postsynaptic potentials and punctate immunofluorescent labelings of GluR epsilon2 protein in the neuropil regions, while GluR epsilon2 protein expression was comparable. Immunostaining of cultured cerebral neurons showed the reduced punctate staining of the truncated GluR epsilon2 protein at synapses. These results suggest that the carboxy-terminal region of the GluRepsilon2 subunit is important for efficient clustering and synaptic localization of the NMDA receptor channel.

Multiple epsilon subunits are major determinants of the NMDA receptor channel diversity. Based on their functional properties in vitro and distributions, we have proposed that the epsilon 1 and epsilon 2 subunits play a role in synaptic plasticity. To investigate the physiological significance of the NMDA receptor channel diversity, we generated mutant mice defective in the epsilon 2 subunit. These mice showed no suckling response and died shortly after birth but could survive by hand feeding. The mutation hindered the formation of the whisker-related neuronal barrelette structure and the clustering of primary sensory afferent terminals in the brainstem trigeminal nucleus. In the hippocampus of the mutant mice, synaptic NMDA responses and longterm depression were abolished. These results suggest that the epsilon 2 subunit plays an essential role in both neuronal pattern formation and synaptic plasticity.

The NMDA (N-methyl-D-aspartate) receptor channel is important for synaptic plasticity, which is thought to underlie learning, memory and development. The NMDA receptor channel is formed by at least two members of the glutamate receptor (GluR) channel subunit families, the GluR epsilon (NR2) and GluR zeta (NR1) subunit families. The four epsilon subunits are distinct in distribution, properties and regulation. On the basis of the Mg2+ sensitivity and expression patterns, we have proposed that the epsilon 1 (NR2A) and epsilon 2 (NR2B) subunits play a role in synaptic plasticity. Here we show that targeted disruption of the mouse epsilon 1 subunit gene resulted in significant reduction of the NMDA receptor channel current and long-term potentiation at the hippocampal CA1 synapses. The mutant mice also showed a moderate deficiency in spatial learning. These results support the notion that the NMDA receptor channel-dependent synaptic plasticity is the cellular basis of certain forms of learning.

Neuron-specific enolase (NSE) occurs in mature neurons and paraneurons. We have isolated the genomic clone coding for rat NSE and clarified its gene structure. In order to analyze the regulatory sequence in the 5'-upstream region and introns, we carried out transient expression experiments of NSE genomic DNA fragments fused to chloramphenicol acetyltransferase (CAT) gene which were transfected into several cultured cells. The used cells were primary cultured rat neurons, PC12, neuroblastoma 35, neuroblastoma 103, C6, primary cultured rat glial cells and HeLa cells. The promoter sequence (190 bp) upstream to the transcription initiation site was important in the expression of CAT gene in these cells. From the experiments with external and internal deletion mutants of the fusion gene, the cis-acting regulatory region responsible for the enhanced expression of the CAT activity in the primary cultured neuron and PC12 cells was found to be localized at upstream 500 bp sequence of the intron 1 and 1.5 kbp upstream sequence of the transcription initiation site. In the upstream important sequences, there were the nearest sequences for AP-1 binding motif, AP-2 binding element, SP-1 binding sequence, cAMP response element, half site of glucocorticoid receptor (GRE) binding sequence, half site of thyroid hormor receptor (TR) or retinoic acid receptor (RAR) binding sequence and MTF-1 binding sequence. Furthermore, Octamer-6 binding motifs also were found. In the intron 1, 5' end upstream 50 bp and downstream 100 bp were the most important sequences. We found the nearest sequences for cAMP response element, E2F binding sequence, early growth response (EGR)-1 binding motif, half site of TCF-1 binding sequence and a neuron-specific element-like sequence in the intron 1.

The primary structure of a putative subunit of the mouse glutamate receptor channel, designated as the delta 2 subunit, has been deduced by cloning and sequencing the cDNA. The delta 2 subunit has four putative transmembrane segments characteristic for neurotransmitter-gated ion channels, and shares 56% amino acid sequence identity with the delta 1 subunit of the mouse glutamate receptor channel and 14-24% identity with the subunits of the AMPA-, kainate- or NMDA-selective glutamate receptor channel. RNA blot and in situ hybridization analyses show that the delta 2 subunit mRNA is localized in cerebellar Purkinje cells. Furthermore, immunoblot and immunohistochemical analyses suggest that the delta 2 subunit protein is actually expressed in vivo in Purkinje neurons. The selective localization of the delta 2 subunit in Purkinje cells may imply a role of the delta 2 subunit in Purkinje cell-specific function such as the cerebellar LTD.

Two novel subunits of the mouse NMDA receptor channel, the epsilon 2 and epsilon 3 subunits, have been identified by cloning and expression of complementary DNAs. The heteromeric epsilon 1/zeta 1, epsilon 2/zeta 1 and epsilon 3/zeta 1 NMDA receptor channels exhibit distinct functional properties in affinities for agonists and sensitivities to competitive antagonists and Mg2+ block. In contrast to the wide distribution of the epsilon 1 and zeta 1 subunit messenger RNAs in the brain, the epsilon 2 subunit mRNA is expressed only in the forebrain and the epsilon 3 subunit mRNA is found predominantly in the cerebellum. The epsilon 1/zeta 1 and epsilon 2/zeta 1 channels expressed in Xenopus oocytes, but not the epsilon 3/zeta 1 channel, are activated by treatment with 12-O-tetradecanoylphorbol 13-acetate. These findings suggest that the molecular diversity of the epsilon subunit family underlies the functional heterogeneity of the NMDA receptor channel.

The primary structure of the mouse glutamate receptor beta 2 subunit has been deduced by cloning and sequencing cDNA. The beta 2 subunit has structural characteristics common to the subunits of glutamate-gated ion channels. Expression of the cloned cDNA in Xenopus oocytes yields functional glutamate receptor channels selective for kainate.

The glutamate receptor (GluR) channel plays a key part in brain function. Among GluR channel subtypes, the NMDA (N-methyl-D-aspartate) receptor channel which is highly permeable to Ca2+ is essential for the synaptic plasticity underlying memory, learning and development. Furthermore, abnormal activation of the NMDA receptor channel may trigger the neuronal cell death observed in various brain disorders. A complementary DNA encoding a subunit of the rodent NMDA receptor channel (NMDAR1 or zeta 1) has been cloned and its functional properties investigated. Here we report the identification and primary structure of a novel mouse NMDA receptor channel subunit, designated as epsilon 1, after cloning and sequencing the cDNA. The epsilon 1 subunit shows 11-18% amino-acid sequence identity with rodent GluR channel subunits that have been characterized so far and has structural features common to neurotransmitter-gated ion channels. Expression from cloned cDNAs of the epsilon 1 subunit together with the zeta 1 subunit in Xenopus oocytes yields functional GluR channels with high activity and characteristics of the NMDA receptor channel. Furthermore, the heteromeric NMDA receptor channel can be activated by glycine alone.

The presence and primary structure of a novel subunit of the mouse glutamate receptor channel, designated as gamma 2, have been revealed by cloning and sequencing the cDNA. The gamma 2 subunit has structural characteristics common to the neurotransmitter-gated ion channel family and shares a high amino acid sequence identity with the rat KA-1 subunit, thus constituting the gamma subfamily of the glutamate receptor channel. Expression of the gamma 2 subunit together with the beta 2 subunit in Xenopus oocytes yields functional glutamate receptor channels selective for kainate.

Functional analysis of AMPA-selective glutamate receptor channels expressed in Xenopus oocytes from cloned cDNAs has shown that homomeric channels formed by the GluR1 subunit are permeable to Ca2+, whereas heteromeric channels composed of the GluR1 and GluR2 subunits show little permeability. Furthermore, substitution of glutamine for arginine in putative transmembrane segment M2 of the GluR2 subunit makes the heteromeric channels permeable to Ca2+. These results suggest that the GluR2 subunit plays a key role in keeping AMPA-selective glutamate receptor channels essentially impermeable to Ca2+ and that the critical determinant is the positively charged residue in M2 segment.

The structure of rat muscle-specific enolase (beta beta enolase) gene was determined. It comprises 12 exons of various lengths (59-223 bp) spanning about 6 kbp and its exon-intron organization is similar to that of neuron-specific enolase (gamma gamma enolase) gene. A transcriptional start site was identified by a combination of S1 nuclease mapping and primer extension analyses. In the 5'-flanking region we found a TATA-box-like sequence and several MyoD-binding motifs. The in vitro cell free transcription of the truncated genomic DNA fragment using HeLa cell extract showed that the transcription start site has been correctly identified and the promoter sequences work well.

The complete amino acid sequences of two mouse glutamate receptor subunits (GluR1 and GluR2) have been deduced by cloning and sequencing the cDNAs. Xenopus oocytes injected with mRNA derived from the GluR1 cDNA exhibit current responses both to kainate and to quisqualate as well as to glutamate, whereas oocytes injected with mRNA derived from the GluR2 cDNA show little response. Injection of oocytes with both the mRNAs produces current responses larger than those induced by the GluR1-specific mRNA and the dose-response relations indicate a positively cooperative interaction between the two subunits. These results suggest that kainate and quisqualate can activate a common glutamate receptor subtype and that glutamate-gated ionic channels are hetero-oligomers of different subunits.

The uptake of 14C-GABA and 14C-choline by cerebral cortex slices did not show any change in rats which showed neurological signs 7 days after the last 7 methylmercury injections (10 mg/kg/day). However, 14C-GABA uptake by dorsal ganglia greatly decreased, although 14C-choline uptake did not decrease in dorsal ganglia. 14C-GABA uptake by cerebellum also was somewhat decreased. The degree of inhibition of 14C-GABA uptake by dorsal ganglia increased each day after the commencement of methylmercury injection. The decrease of 3H-GABA uptake by dorsal ganglia of methylmercury-injected rats was confirmed by autoradiography. Autoradiographic and inhibitor studies showed labelled GABA accumulated in the satellite glial cells. In vitro addition of methylmercury (10(-4)-10(-5) M) equally inhibited the uptake of GABA and choline by brain and dorsal ganglia slices. These studies may reveal a possible mechanism of methylmercury neuropathy.

Neuron-specific (gamma gamma) enolase (NSE) is an isoenzyme form of glycolytic enzyme, enolase. We isolated genomic clones for NSE and clarified NSE gene structures. The NSE-gene spanned about 9 kb and consisted of twelve exons and eleven introns. Multiple transcriptional start points were identified by a combination of S1 nuclease mapping and primer extension analysis. In the 5'-flanking region we found a TATA-like sequence TCTATAGGC which was only partially homologous to the consensus sequence, but we did not find a CAAT box. The sequence in the immediate 5'-flanking region was of a relatively high G + C content and contained GC-box-like clusters that did not correspond to the typical GC box. In addition, we found seven classes of the repeated sequences. In the introns 1, 5 and 10 there were tandem repeats (GT)33, (GT)21 and (GT)24, respectively. The 3' end contains a single polyadenylation site and an identifier sequence 2 kb downstream from the poly(A)-addition site. The in vitro cell-free transcription of the truncated genomic DNA fragment using HeLa cell extract showed that the transcription start points have been correctly identified and the putative promoter sequences appear to be functional.

The nucleotide sequence of rat muscle-specific enolase cDNA was determined by sequencing three cDNA clones encoding this enolase isozyme. The nearly full-length cDNA consists of 13-bp 5'- and 84-bp 3'-noncoding regions and a poly(A) tail in addition to a 1302-bp coding region encoding a polypeptide composed of 434 amino acid residues. The deduced primary structure of this enolase isozyme is about 80% similar to those determined previously for rat neuron-specific and non-neuronal enolase isozymes. Southern blot analysis suggested strongly the existence of a single copy of the muscle-specific enolase gene per haploid genome. The mRNA for this enolase isozyme was detected in rat skeletal muscle on day 1 after birth and its level increased rapidly during 10-30 days after birth without any change in its size (1500 bases).

The nucleotide sequence for alpha alpha enolase (non-neuronal enolase: NNE) of rat brain and liver was determined from recombinant cDNA clones. The sequence was composed of 1722 bp which included the 1299 bp of the complete coding region, the 108 bp of the 5'-noncoding region and the 312 bp of the 3'-noncoding region containing a polyadenylation signal. In addition, the poly(A) tail was also found. A potential ribosome-binding site was located 30 nucleotides upstream to the initiation codon in the 5'-noncoding region. The amino acid sequence deduced from the nucleotide sequence was 433 amino acids in length and showed very high homology (82%) to the amino acid sequence of gamma gamma enolase (neuron-specific enolase: NSE), although the nucleotide sequence showed slightly lower homology (75%). The size of NNE mRNA was approximately 1800 bases by Northern transfer analysis and much shorter than that of NSE mRNA (2400 bases) indicating a short 3'-noncoding region. A dot-blot hybridization and Northern transfer analysis of cytoplasmic RNA from the developing rat brains using a labeled 3'-noncoding region of cDNA (no homology between NSE and NNE) showed a decrease of NNE mRNA at around 10 postnatal days and then a gradual increase to adult age without changes of mRNA size. Liver mRNA did not show any significant change during development.

We characterized the brain-specific gene coding for rat S-100 protein beta-subunit and found three "brain identifier (ID)" elements, which have been proposed to regulate the gene expression in rat brain. The nucleotide sequences of these elements corresponded well with that of the consensus ID element and were clearly different from those of "ID-like" elements in rat beta B1-crystallin gene, etc. ID elements were also observed in the flanking regions of rat neuron-specific enolase and cholecystokinin genes, which were expressed in the neuronal cells. Direct repeats were observed in the regions flanking ID elements.

In order to search for the relationship between the structure and the function of the nervous system, the spinal cord provides suitable material. We devised a procedure for isolation of large ventral and small dorsal horn neurons. Then we examined the gamma-aminobutyric acid (GABA) system in both neurons isolated using our procedure. Glutamic decarboxylase (GAD) activity in dorsal horn neurons was much higher than that in ventral horn neurons. Further, GABA uptake activity by the dorsal horn neurons was also somewhat higher than that by the ventral horn neurons, although some properties of GABA uptake were found to be almost the same in both neuronal fractions. However, we could not find any difference of GABA-alpha-ketoglutarate transaminase (GABA-T) activity between the dorsal and ventral horn neurons. These results suggest that GAD and GABA uptake may be indicators for cell specificity to some extent.

Neuron-specific enolase (NSE), and non-neuronal enolase (NNE) which exists in many tissues including liver but is localized in glial cells within the nervous system, were synthesized in the rabbit reticulocyte cell-free translation system programmed with brain mRNAs. The in vitro synthesized NSE and NNE were indistinguishable from the two enzymes purified from rat brains. NSE mRNA activity was found only in brain RNAs, while NNE mRNA activity existed in brain RNAs as well as liver RNAs. In developing brains, the level of translatable NSE mRNA was low at the embryonic stage and at birth, increased rapidly from about 10 days postnatal, and reached the adult level, while that of NNE mRNA was high at the embryonic stage and at birth, followed by a slight decrease then a gradual rise to adult levels. These changes correlated with the developmentally regulated appearance and accumulation pattern of each of the two enzymes. These results suggest that the levels of NSE and NNE are controlled primarily by the level of each of the two translatable mRNAs. In developing livers, only the NNE mRNA activity was detected and its level generally paralleled the changes in the level of NNE.

The unique DNA expression value in the brain nuclear RNA was much higher than those from other tissues. The origin of this high complexity of brain nuclear RNA was examined using the [3H]DNA-RNA saturation hybridization technique. RNA was extracted from the isolated neuronal and glial nuclei. The experimental results may indicate that the high complexity in the brain nuclei does not result from the summation of many cell types but only from the neuronal nuclei.

We have isolated the ventral and dorsal horn neurons from the bovine spinal cord using our isolation procedure and determined choline acetyltransferase, acetylcholinesterase and choline uptake activity in these neurons. Choline acetyltransferase and choline uptake activity were higher in the ventral than in the dorsal horn neurons, showing their cell specificity to some extent, but acetylcholinesterase activity did not show any differences between the two cell types.