- Email: [email protected]
Glutamate receptor editing in the mammalian hippocampus and avian neurons
Molecular Brain Research 48 Ž1997. 37–44
Research report
Glutamate receptor editing in the mammalian hippocampus and avian neurons Deborah L. Lowe a , Klaus Jahn b, Dean O. Smith b
c,)
a Neurosciences Training Program, UniÕersity of Wisconsin, Madison, WI 53706, USA Physiologisches Institut der Technischen UniÕersitat 80802 Munchen, Germany ¨ Munchen, ¨ ¨ c Pacific Biomedical Research Center, UniÕersity of Hawaii, Honolulu, HI 96822, USA
Accepted 14 January 1997
Abstract RNA editing determines receptor kinetics and permeability of glutamate receptors. This post-transcriptional modification alters single nucleotides within an RNA transcript changing the codon specified by the genome resulting in the incorporation of a different amino acid, profoundly affecting the properties of the protein subunit. We have studied the three sites subject to RNA editing within the kainate-specific subunit GluR6 in the mammalian hippocampus to determine developmental changes and cell-specific variation in editing. GluR6, when measured in the whole rat hippocampus, is predominantly expressed in the unedited form at E18, with a gradual progression to the edited form during the 1st post-natal week, and remains stable from P8 through 30 months. Individual neurons from P0 through P8 rat hippocampal slices analyzed with single-cell PCR show predominant expression of fully edited GluR6, unlike the population profile. In contrast, single astrocytes from P0 hippocampal cultures show that the most common variant is partially edited. Thus, editing in neurons and glia differs, and this difference accounts for part of the disparity between single-neuron and whole-hippocampus data. Editing in astrocytes is affected by conditions in the external environment, as purified astrocytes fail to edit GluR6, although editing occurs in astrocytes from hippocampal cultures. The homogeneity of GluR6 editing between species was also determined by comparing editing in avians and mammals. Genomic and cDNA analysis of chick glutamate receptors demonstrates avian editing of GluR2 but not GluR6. Keywords: RNA editing; Hippocampus; Glutamate receptor; Development; Chick; Single-cell PCR
1. Introduction The process of cellular development and maturation is regulated by alterations in cell gene expression. In the nervous system, genetic expression is controlled transcriptionally through positive and negative regulation and posttranscriptionally via alternative splicing and RNA editing w8,20x. Glutamate receptors are the only protein in the nervous system known to be edited w17,22,33x. RNA editing in glutamate receptors involves an enzymatic-base modification at a single nucleotide in an RNA transcript w23,30x. Thus, an amino acid not specified by the genome is incorporated into the protein subunit significantly altering a receptors functional properties. Receptor characteristics determined by editing include desensitization, rectification and ionic permeability. ) Corresponding author. University of Hawaii, 2444 Dole Street, Bachman 204, Honolulu, HI 96822, USA. Fax: q1 Ž808. 956-8061; E-mail: [email protected]
In the AMPA-receptor family, desensitization kinetics are controlled by editing in the glutamate receptor subunits ŽGluR. 2 through 4 at a site preceding the putative fourth transmembrane region w22x. Ionic permeability w5,12,15x and rectification w35x are determined by GluR2 editing at a site in the pore region w33x. In the kainate-receptor family ionic permeability and receptor kinetics are determined by editing of GluR6 w7x. Editing occurs in two residues of the putative first transmembrane region ŽTM1. resulting in an isoleucine ŽI. to valine ŽV. conversion ŽIrV site. and a tyrosine ŽY. to cysteine ŽC. conversion ŽYrC site., and at one residue of the pore region where a glutamine ŽQ. to arginine ŽR. conversion occurs ŽQrR site. w17x. GluR5 is also edited at the QrR site, but its function is unknown w17,33x. Editing of GluR6 is complex, as each site is edited with a different efficiency creating eight possible molecular variants. Editing in both the kainate- and AMPA-receptor families is developmentally regulated, as unedited receptors appear at early stages of development and are less com-
0169-328Xr97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 9 - 3 2 8 X Ž 9 7 . 0 0 0 7 2 - 7
38
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
mon in adult animals w3,5,17,22x. This regulation indicates that RNA editing could play a role in altering kinetic and ionic properties of glutamate receptors at critical points in proliferation, migration and synapse formation in the developing nervous system. In order to understand how GluR6 editing is related to cell differentiation and development, one must establish the extent of editing at all three sites on individual transcripts, as receptor phenotype is a function of editing at each site. The techniques employed in studies thus far allow only single sites to be examined at one time, making it impossible to know the proportion of GluR6 variants that occur in each of the eight possible forms. A thorough analysis of GluR6 editing requires sequencing of individual transcripts from regions of the CNS and from individual neurons to determine the extent of editing in a given region and to define the molecular constructs responsible for receptor characteristics measured in single cells. To address the question of GluR6 editing in development, we have determined the relative proportions of GluR6 variants present in whole rat hippocampus from E18 through 30 months of age, from individual neurons in hippocampal slices from P0 through P8 and from astrocytes in hippocampal cultures. We have found that editing in single neurons and single astrocytes during the 1st post-natal week does not reflect population measures. We also report neuron-glia differences in the extent to which GluR6 transcripts are edited and cell-culture effects on astrocyte editing. The conservation of glutamate receptor editing between mammalian and avian genes was determined, as RNA editing of GluR2 is absent in fish w19x. We found that glutamate receptor editing in avians is not conserved for all subunits. 2. Materials and methods 2.1. Hippocampal neuronsr glia culture for single-cell astrocyte PCR Cell cultures were prepared from hippocampi of P0 or P8 rats. Hippocampal neurons were mechanically dissociated following incubation with papain Ž40 Urml, 60 min, 328C; Worthington Biochemicals, Freehold, NJ.. Dissociated cells were plated on collagen coated 35 mm petri dishes ŽCollaborative Biomedical, Bedford, MA. at a final density of 10 5 –10 6 cellsrplate. Cells were incubated at 378C in an humidified atmosphere of 5% CO 2 and 95% air in minimal essential media ŽMEM. ŽGibco, Gaithersburg, MD. supplemented with 10% horse serum ŽGibco., 10% fetal bovine serum ŽGibco., glucose Ž6 mgrml., penicillin Ž50 Urml. and streptomycin Ž50 mgrml. for 1–2 days. Prior to obtaining cytoplasm from single astrocytes via whole-cell patch-clamp, the cells were washed with divalent-free phosphate-buffered saline and incubated with 0.05% trypsin for 3–5 min to detach neurons and expose
the underlying bed of glia. The dish was then perfused with normal extracellular saline Žin mM: 125 NaCl, 5 KCl, 1 MgCl 2 , 10 Hepes, 2 CaCl 2 and 10 glucose, pH 7.3 w290 mOsmx. to remove the enzyme and detached neurons. 2.2. Chick motoneurons Chick a-motoneurons were prepared using techniques described in detail by Rosenheimer and Smith w28x. Lumbar spinal cords from 6.5 d embryonic chicks were dissected and freed of their meninges and dorsal root ganglia. Following incubation at 378C with 0.05% trypsin ŽGibco. and 0.005% DNase I ŽSigma, St. Louis, MO., the cells were dissociated mechanically. Motoneuron-enriched fractions were generated on the basis of their buoyant density in a metrizamide solution w32x. 2.3. Purified type-1 astrocytes Type-1 astrocytes were purified from hippocampi of six P2 rats. Cell suspensions were prepared by the method described in Williams et al. w36x for macroglia dissociation from P2–12 brains. Modifications to the technique included use of 1% collagenase ŽWorthington Biochemicals. for 20 min at 378C for the initial enzyme digestion and no further use of collagenase following trypsinization. Tissue was placed in Dulbecco’s modified Eagle’s medium plus 10% fetal calf serum ŽDMEM-F10. and mechanically dissociated with a 5 ml pipette, a 21 G needle and finally a 23 G needle. Cells were plated at a density of 10 7 cellsr75 cm2 collagen-coated flask and incubated at 378C in a humidified atmosphere of 5% CO 2 and 95% air. These cultures were fed every 3 days with DMEM-F10. After 8 days in culture, type-1 astrocytes were separated from oligodendrocytes, neurons, macrophages and type-2 astrocytes by overnight shaking Ž100 rpm. on a rotary platform at 378C as described by Levison and McCarthy w21x. Cytosine arabinoside Ž2 = 10y5 M. was added 1 and 3 days after shaking to achieve a purity of 95% type-1 astrocytes Ž95% GFAP-positive.. 2.4. RNA extraction and GluR6 amplification from rat hippocampi, purified astrocytes and single cells mRNA was extracted from P0 through 30 month whole rat hippocampi, a population of dissociated chick amotoneurons and a population of purified type-1 astrocytes with the Micro-Fast Track mRNA Isolation Kit ŽInvitrogen, San Diego, CA.. Total RNA was isolated from single neurons in hippocampal slices and single glia in hippocampal cultures by aspirating the cytoplasm from single cells after attaining whole-cell patch-recording conditions. Cytoplasm from single cells was expelled from the patch-recording electrode into a sterile microcentrifuge tube with positive pressure and combined with the following for reverse transcription ŽRT.: Random hexamers Žfinal
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
w1.25 m Mx., 1 m l 10 = RT buffer, 20 U Rnase inhibitor, 50 U reverse transcriptase, deoxyribonucleotide triphosphates – dATP, dCTP, dGTP and dTTP Žfinal w0.75 m Mx., MgCl 2 final w5 mMx and 5 m l 10 mM Tris–HCl, pH 8.3. RT was performed at 428C for 30 min followed by 5 min at 998C and held at 48C until polymerase chain reaction ŽPCR.. PCR was then performed with GluR6-specific primers Žsense: 5X-gcgaattcggacaatggaatggaatggttc-3X ; antisense: 5X-agcggatccgtatacgaagaaatgatgat-3X w33x. and glial fibrillary acidic protein ŽGFAP.-specific primers Žsense: 5Xcaatctcacacaggacctcggc-3X ; antisense: 5X -ataccactcctctgtctcttg-3X . at a final w0.07 m Mx. Eight m l 10 = PCR Buffer, 66 m l DEPC-treated H 2 0, MgCl 2 final w3 mMx and 2.5 U of Ampli-Taq DNA polymerase ŽPerkin Elmer, Branchburg, NJ. were added for a final volume of 100 m l. PCR cycling was as follows: 958C, 2 min, 20 step cycles Ž958C, 1 min; 568C, 1 min; 728C, 1 min., 20 step cycles Ž958C, 1 min; 568C, 1 min; 728C, 1 min with a 5 s extensionrcycle. and 728C for 7 min. RT of GluR6 from whole-hippocampal mRNA and type-1 astroctye mRNA was performed with the GluR6specific antisense primers used in single-cell amplifications. The sense primer specific for GluR6 was added before PCR. RT and PCR were carried out under the same cycling conditions used for single-cell amplification. GluR2, GluR5 and GluR6 were amplified from mRNA of chick a-motoneurons. Sense and antisense primer pairs for cDNA amplification were: 5X-gcgaattcaggaaatgacacgtctgggc-3X and 5X-agcggatccgtgtaggaggagattatgat-3X for GluR2; 5X-gcgaattcttcacaccctacgagtggtataacc-3X and 5Xagcggatcccccacattttctcataggtggag-3X for GluR5. GluR6 primers were listed under single-cell PCR w33x. Antisense primers were used for RT and sense primers were used for PCR with the same protocol used for single cells.
39
2.6. Whole-cell patch conditions and cell identification Transverse hippocampal slices of 150–180 m m were cut from brains of Sprague Dawley rats ŽP0–P8. using a vibratome. Neurons from CA1, CA3 and the stratum radiatum were visually identified under Nomarski optics w13x. Identification was based on morphology and location in the slice. Neurons from CA1 and CA3 had pyramid-shaped cell bodies and were located opposite the dentate gyrus. Neuronal identity was further verified by the absence of GFAP amplification. Glial cells were from P0 hippocampal cultures and were visualized with a 40 = water immersion objective. Astrocytes from hippocampal cultures were identified on the basis of morphology. Cells had either a flattened, epithelioid shape with a few processes or were stellate. Glial identity was verified by positive results for GFAP amplification. GFAP was positively identified by southern blot analysis. Cells were patch-clamped in the whole-cell configuration. Patch pipettes had a tip resistance of 3–5 M V and were filled with 5 m l of the following diethyl pyrocarbonate-treated, autoclaved solution: 150 mM KCl and 10 mM Tris–HCl Žbuffered to pH 7.4 with NaOH w300 mOsmx.. The cells were continuously perfused with oxygenated saline: Žin mM. 125 NaCl, 2.5 KCl, 1.0 MgCl 2 , 0.4 CaCl 2 , 26 NaHCO 3 , 1.25 NaH 2 PO4 and 10 glucose Žbuffered to pH 7.3 with NaOH w320 mOsmx.. The silver wire connected to the patch electrode was dipped in bleach before every recording to remove residual mRNA from the previous cell. Patch pipettes were pulled from borosilicate glass tubing Ž2.0 mm outer diameter, 0.5 mm wall thickness. that had been baked overnight at 2008C. 2.7. Quantification of GluR editing from cDNA and chick genomic DNA
2.5. Genomic DNA amplification Genomic DNA from E20 chicken liver was prepared by standard procedures w31x. DNA was amplified by PCR in 100 m l reactions containing: 50 mM KCl, 10 mM Tris–HCl ŽpH 8.3., 2.5 mM MgCl 2 , 0.02 m M each of dATP, dCTP, dGTP and dTTP, 1.5 m M each of sense and antisense primers and 2.5 U Ampli-Taq DNA polymerase ŽPerkinElmer.. Reactions were run with the cycling used for cDNA amplification. Genomic amplifications were performed with the following sense and antisense primer pairs: 5X-gcgaattccttgcagttgctccactggct-3X ; 5X-agcggatccgagcatgttacctggctatg-3X for GluR6 TM1; 5X-gcgaattcgtttagtccctatgagtggtata-3X and 5X-agcggatccgtatacgaagaaatgatgat-3X for GluR6 TM2; 5X-gcgaattcttcacaccctacgagtggtataacc-3X and 5X-agcggatccaaccggtgtaccttg-3X for GluR5 TM2, and 5-gcgaattcagaagtccaaaccaggag-3X and 5X-agcggatccatgaatatccacttgagac-3X specific for GluR2 TM2 w17,33x. Each primer spans an intron-exon border with the exception of the antisense primer for GluR6 TM2 and sense primer for GluR5 TM2.
The GluR6-, GluR5- and GluR2-specific primers contained restriction enzyme sites for EcoRI and BamHI to allow for cloningrsequencing analysis of GluR transcripts. PCR product from single cells, cell populations and genomic DNA was digested with EcoRI and BamHI ŽGibco., precipitated and gel-purified. Purified product was directionally ligated into M13mp18 or 19 RF-DNA ŽGibco.. Ligation reactions were transformed and positive plaques Žbluerwhite screening. were processed and sequenced ŽSequenase Version 2.0 DNA Sequencing Kit; Amersham, Arlington Heights, IL.. Sequence information determined the proportion of edited variants present in single cells and cell populations and genomically encoded chick sequence ŽFig. 1.. 2.8. Statistical analysis Differences in GluR6 editing as a function of age for whole-hippocampus and single-cell data were determined by single-factor analysis of variance. Data from whole
40
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
Ž F1,7 s 23.7, P - 0.0002. and unedited Ž F1,7 s 25.1, P 0.0001. proportions of GluR6. Post-hoc analyses were not performed, as there were no variability measures at E18, P0 and P2. The profile of GluR6 editing in whole rat hippocampus is an average measure of editing occurring in the heterogeneous cell populations that comprise the hippocampus. To determine if population measures of GluR6 editing reflect those in single cells within this region, we used single-cell PCR to construct a profile of GluR6 editing for neurons in hippocampal slices from CA1, CA3 and the stratum radiatum and in astrocytes from hippocampal cultures. 3.2. Glur6 editing in single hippocampal neurons
Fig. 1. Experimental protocol for analysis of glutamate receptor editing. RNA was extracted from single cells with whole-cell patch clamp or cell populations with standard techniques. BamHI and EcoRI restriction enzyme sites within the GluR-specific primers allowed for directional cloning of transcripts into M13. Positive clones were identified by blue-white screening of transformed cells and sequenced.
hippocampus and single cells at a given age were compared with Student’s t-tests. The number of sequences required to determine the dominant form of GluR6 present in single cells was arrived at by comparing each successive set of 10 sequences collected for whole hippocampus at 30 months. The dominant form of GluR6 was the same in each group of 10; thus, 10 sequences were sufficient to make this determination. Before analysis, editing data from cell populations and single cells were converted to proportions.
Analysis of GluR6 transcript editing in single neurons from P0 through P8 hippocampal slices illustrates that editing in neurons from CA1, CA3 and stratum radiatum is strikingly different from the population profile ŽFig. 3.. First, editing in single neurons at P0 is nearly complete Ž72%. with remaining transcripts being partially edited and no transcripts in the unedited form. This profile reflects that found in whole hippocampus at P8 and not at P0. In P0 single hippocampal neurons, the proportion of GluR6 transcripts in the edited Ž t s 1.6; P - 0.05., partially edited Ž t s 0.7; P - 0.05. and unedited Ž t s 1.0; P - 0.05. variants does not significantly differ from P8 whole hippocampus. Second, there is no change in GluR6 editing in single neurons from P0 through P8. Fully edited GluR6 was always the dominant transcript present in single neurons from CA1, CA3 and the stratum radiatum during this period, unlike the dramatic shift from dominant expression
3. Results 3.1. Glur6 editing in whole hippocampus In whole rat hippocampus, the dominant form of GluR6 transcripts changes from partially edited at E18 to fully edited from P8 through 30 months ŽFig. 2.. At E18 55% of GluR6 transcripts sequenced were partially edited, however; by P8 the proportion of partially edited transcripts had decreased to 15%. The proportion of fully edited transcripts present in whole hippocampus changed in an opposite manner. At E18 only 13% of transcripts were fully edited, while this proportion increased to 83% by P8. The proportion of unedited transcripts decreased in a manner similar to partially edited transcripts in that their numbers were much greater at E18 Ž33%. than at P8 Ž2%.. The shift from partial editing of GluR6 to full editing of GluR6 within the whole hippocampus appeared to be a gradual developmental change occurring during the 1st post-natal week. There was a significant effect of age for the fully edited Ž F1,7 s 20.2, P - 0.0005., partially edited
Fig. 2. GluR6 editing in whole-rat hippocampus from E18 through 30 months. Edited values represent the percentage of transcripts sequenced in which nucleotide substitutions occurred at the IrV, YrC and QrR sites. Unedited values represent the percentage of transcripts in which no nucleotide substitutions occurred. Partial values represent the percentage at which one or two of the possible nucleotide substitutions occurred Ž6 variants ŽTM1, TM1 and TM2.: I,C;Q, V,Y;Q, V,C;Q, I,Y;R, I,C;R, V,Y;R.. E18, P0 and P2 values are from the combined RNA of 3 hippocampi, so no variability measures are possible. P6 through 30 month values are the mean"S.E.M. of editing in three separate animals per age. The number of transcripts sequenced per age was 47"9.
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
Fig. 3. The extent of GluR6 editing in CA1, CA3 and stratum radiatum neurons from P0 through P8 hippocampal slices. Each set of values represents the percentage of transcripts found in the fully edited, partially edited Ž6 variants. or unedited forms from single cells in P0 through P8 hippocampal slices. The number of cells included at each age is listed above the bar graphs. P0 data include four CA1 neurons, one CA3 neuron and one neuron from the stratum pyramidal. P2 data include one CA1 neuron, one CA3 neuron and one neuron from the stratum pyramidal. P6 data include one CA1 neuron and one CA3 neuron. P8 data include one CA1 neuron, four CA3 neurons and one neuron from the stratum pyramidal. Values represent the mean"S.E.M. of all cells collected at each age Ž ns9 sequencesrcell.. Each cell was from a different animal. For statistical analysis all cells at each age were combined.
of partially edited GluR6 in P0 hippocampus to fully edited GluR6 in P8 hippocampus. A third difference is the lack of unedited transcripts in single neurons at all ages measured. With the exception of a few unedited transcripts at P8, no unedited transcripts were found at any other age. The marked differences between the developmental profile of GluR6 editing in whole hippocampus and single neurons indicates that editing in other neuronal populations or glia must account for these disparities. 3.3. Glur6 editing in single astrocytes and purified astrocytes from hippocampus Editing in single astrocytes from hippocampal cultures was quite different from that found in single neurons from hippocampal slices. Fig. 4 illustrates editing of GluR6 in single astrocytes from hippocampal cultures and purified type-1 astrocytes. The majority of transcripts in three single GFAP-positive astrocytes from P0 hippocampal cultures are partially edited Ž80%.. This is significantly different from the proportion of partially edited transcripts in single neurons at P0 Ž t s 11.8; P - 0.05.. In contrast to reports of a lack of GluR6 editing in glial cell lines of cortical origin w27x, we found astrocytes in hippocampal cultures edit GluR6. This difference could be a function of tissue origin Žcortex vs. hippocampus. or
41
Fig. 4. GluR6 editing from single astrocytes in hippocampal cultures and purified type-1 astrocyte cultures. Each set of values represents the percentage of transcripts Žmean"S.E.M.. found in the fully edited, partially edited Ž6 variants. or unedited forms from single astrocytes at P0 Ž ns number of cells.. Each cell was from a different culture. Eleven transcripts were sequenced per cell. Values for purified type-1 astrocytes were determined from a population of cells.
culture conditions. This led us to determine editing in a neuron-free culture of purified hippocampal type-1 astrocytes ŽFig. 4.. In this culture, 100% of the GluR6 transcripts sequenced were unedited at the IrV, YrC and QrR sites Ž10 of 10 transcripts. indicating that culturing astrocytes in the absence of the normal complement of hippocampal cells may alter the post-transcriptional regulation of GluR6. 3.4. Complement of GluR6 isoforms in neurons Õs. glia Single neurons and astrocytes not only differ in the most common form of GluR6 expressed, but also in the order of prevalence of partially edited transcripts present. Table 1 summarizes the relative proportions of GluR6 variants in whole hippocampus, single neurons and single astrocytes at P0. The most common partially edited form in neurons was V,C;Q ŽTM1 fully edited, TM2 unedited., while the most common form in glia was I,Y;R ŽTM1 Table 1 The breakdown of edited GluR6 variants present in combined data from whole P0 hippocampus, single P0 neurons and single P0 astrocytes
IYQ VYQ ICQ IYR VCQ ICR VYR VCR
P0 hippocampus Ž ns 42, 3 animals.
P0 neurons Ž ns 57, 6 cells.
P0 astrocytes Ž ns 34, 3 cells.
Ž15. 36% 0 Ž3. 7% Ž9. 21% Ž4. 10% Ž3. 7% 0 Ž8. 19%
0 Ž1. 2% 0 Ž1. 2% Ž10. 17% Ž3. 5% Ž2. 2% Ž41. 72%
Ž3. 9% Ž1. 3% Ž3. 9% Ž11. 32% Ž2. 6% Ž8. 23% Ž2. 6% Ž4. 12%
n, number of transcripts.
42
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
unedited, TM2 edited.. A second neuron-glia difference was that the partially edited transcripts I,C;Q and I,Y;Q were not found in P0 neurons while all partially edited forms were found in P0 astrocytes. Differences in the complement of GluR6 variants present could lead to phenotypically dissimilar KA receptors in neurons and glia. Editing in whole hippocampus at P0 did not represent a summation of editing found in single neurons and glia at this age. At P0, the dominant form of GluR6 in the whole hippocampus was I,Y;Q ŽTM1 unedited; TM2 unedited. differing from both P0 glia and P0 neurons. Editing in other cell types such as dentate granule neurons and oligodendrocytes might account for this discrepancy. 3.5. Glutamate receptor editing in chicken To determine the homogeneity of glutamate receptor editing between species, we analyzed genomic DNA and cDNA of chick GluR2, GluR5 and GluR6. Genomic analysis of GluR6 revealed that the equivalent of the rat edited form is encoded by the chick genome ŽGluR6ŽV,C;R... cDNA analysis of E6.5 chick motoneurons and spinal cord supported this finding as 100% of transcripts Ž36r36. sequenced were V,C;R. This difference does not extend to GluR2 and GluR5 where the chick genomic sequence is the same as the rat. cDNA analysis demonstrated 100% editing of the QrR site Ž6 of 6 transcripts. of GluR2 in chick motoneurons at E6.5. The edited state of GluR5, however, is unknown because we were unable to amplify GluR5 from chick motoneurons, spinal cord and brain at E6.5 to E11.5. We conclude that GluR5 is not expressed in these structures at these ages. Genomic DNA and cDNA sequenced from chick have 99% homology with rat in TM1 and the pore region. This high homology is similar to that reported for pigeon sequence w26x.
4. Discussion 4.1. GluR6 editing in hippocampus The predominant molecular form of GluR6 RNA in the developing mammalian hippocampus follows a gradual progression from unedited and partially edited variants at E18 to fully edited at P8, with the fully edited form dominant through 30 months. This is consistent with the QrR editing analysis of embryonic brain and structures of the adult hippocampus done by Bernard and Khrestchatisky w3x in which QrR editing in GluR6 was 54% in E18 brain and 70–85% in CA1, CA3 and dentate gyrus of the adult hippocampus. Data from neuronal single-cell PCR conflicts with this profile. At P0, the dominant form of GluR6 in neurons from CA1, CA3 and stratum radiatum is fully edited, indicating that analysis of the whole hippocampus early in development does not accurately depict GluR6 editing in single cells during this period. This discrepancy
can be attributed, in part, to differences in GluR6 editing in neuronal and glial populations at this age. Glial single-cell PCR indicates that the dominant form of GluR6 in astrocytes at P0 is partially unedited, in agreement with the population profile. The ratio of neurons to glia in the developing hippocampus is unknown. It is estimated that glia comprise one half the volume of the adult nervous system w18x. If we assume this to be true of the hippocampus at P0 then combining the single-neuron and single-glia data in a 1 : 1 fashion should create an editing profile similar to P0 whole hippocampus. Combining P0 single-neuron data and P0 astrocyte data yields an editing profile where 49% of GluR6 variants are fully edited, 48% of GluR6 variants are partially edited and 3% are unedited. These proportions are different from P0 hippocampus where 19% are fully edited, 45% are partially edited and 36% are unedited. If GluR6 editing is a developmentally regulated phenomenon, oligodendrocytes and dentate granule neurons are likely sources of the unedited variants. Both cell types continue to proliferate into post-natal life unlike CA1 neurons where neuronal number at P0 is the same as 1 month w4,10x. Glial progenitor cells could also contribute as they have been shown to express unedited GluR6 in vitro w27x. Differences in GluR6 editing between the population profile and single-neuron profile do not persist throughout the life span. By P8, GluR6 editing in whole hippocampus is not significantly different from single-neuron editing. Three possible explanations for this phenomenon are: Ž1. GluR6 expression in hippocampal neurons from CA1, CA3 and non-pyramidal neurons increases to the extent that they express much greater levels of GluR6 than other cell populations so the population profile reflects their expression; Ž2. GluR6 expression is transient in the cell populations contributing the partially edited and unedited variants so they do not contribute to the population profile at all ages; and Ž3. editing of GluR6 increases with maturation so cell populations that mature post-natally eventually edit a large proportion of their transcripts. Support for this third possibility is provided by editing analysis of a single P8 astrocyte from a hippocampal culture Ždata not shown.. The dominant form of GluR6 in this cell was fully edited Ž50%.. The functional significance of GluR6 editing in neurons and glia in this study was not examined. However, both rectification and Ca2q conductance of kainate receptors are determined by GluR6 editing. Ruano et al. w29x have linked rectification properties of kainate receptors in single cultured hippocampal neurons with QrR site editing. Linear I–V relationships were found when QrR was predominantly edited ŽR., while strong inward rectification was common among cells when QrR was unedited ŽQ.. In this study, the QrR site was predominantly edited in single neurons and astrocytes. Therefore, native kainate receptors in these cells would be expected to have linear I–V relationships.
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
In studies of recombinant GluR6, Burnashev et al. w7x found Ca2q conductance to be inversely related to editing at the QrR site. Cells expressing recombinant GluR6 receptors edited at the QrR site had barely detectable Ca2q currents regardless of editing at the IrV and YrC sites. Cells expressing receptors unedited at the QrR site had larger Ca2q currents when the IrV and YrC sites were also unedited than when these sites were edited. In this study, the majority of transcripts were edited at the QrR site indicating that native kainate receptors would conduct a negligible amount of Ca2q. The relationship between GluR6 editing, receptor kinetics and Ca2q conductance makes editing an effective means by which a cell can alter its receptor phenotype in response to external stimuli or, as the result of endogenous events, signal a cell’s state of maturation. Kainate receptors have been identified on pre-synaptic membranes in the CA1 region of the hippocampus w9x and dorsal root ganglion w14x. Pre-synaptic expression of Ca2q conducting kainate receptors could elevate basal levels of Ca2q in the pre-synaptic terminal, thus, facilitating neurotransmitter release w16x. During development, this could alter the probability that a given pre-synaptic terminal will form a functional contact and that this contact will be maintained.
43
tal switch in kainate and AMPA receptor editing in rats most likely alters receptor characteristics in a manner that is advantageous for cell survival and maturation. The absence of GluR6 editing in chick indicates the aspects of development for which this switch is necessary in rat do not exist in chick or are performed by another mechanism. This is not the first instance of novel glutamate receptor characteristics or expression in chick when compared to mammals. Neurons from the chick nucleus magnocellularis express an AMPA receptor that is Ca2q permeable and linearly or outwardly rectifying w25x. This is not what is predicted from expression studies of AMPA subunits where Ca2q permeability is linked to receptors with inward rectification w15x. A second difference between the mammalian CNS and avian CNS has been reported in chick Bergmann glia. In the mammalian CNS, Burnashev et al. w6x found that Bergmann glia cells do not express the AMPA subunit GluR2 ŽGluR2ŽR. precludes Ca2q permeability.. Mammalian Bergmann glia possess an AMPA receptor that is Ca2q permeable and inwardly rectifying w24x as predicted by expression studies. Ottiger et al. w26x found that Bergmann glia cells in the avian CNS express GluR2. The subunit composition and characteristics of AMPA receptors in avian Bergmann glia should, thus, differ from their mammalian counterparts.
4.2. GluR6 editing in glia Single-cell PCR of P0 astrocytes from hippocampal cultures indicate that astrocytes edit GluR6 when cultured with neurons. In contrast, purified hippocampal type-1 astrocytes do not edit GluR6. The absence of editing in purified type-1 astrocytes and glial cell lines leads one to suspect that culture conditions can modify the editing process. Chemical or electrical signals between neurons and the bed of glia upon which they grow could be essential for editing activity. Glia and neurons have been shown to modify each other’s characteristics. Neuronal activity up-regulates gene expression in astroglia w34x, while astrocytes modify the appearance of transient potassium currents in hippocampal pyramidal neurons w37x. In addition, in co-cultures of purified type-1 astrocytes and retinal ganglion cells the density of ion channels expressed by type-1 astrocytes more closely approximates those found in P10 tissue prints from optic nerve w1x. Purified cultures could be used to determine if growth factors, neurotransmitters, electrical activity or other environmental cues regulate or modulate GluR6 editing. A precedent for environmental regulation of RNA editing exists for mammalian apolipoprotein B, which is modified by diet and hormones w2,11x. 4.3. RNA editing in chicken The mechanism by which GluR6 mRNA is edited in rats and humans appears not to be necessary for chick as the edited form is genomically encoded. The developmen-
Acknowledgements This work was supported by grants from the National Institutes of Health ŽNS13600. and the Spinal Cord Research Foundation.
References w1x Barres, B.A., Koroshetz, W.J., Chun, L.L. and Corey, D.P., Ion channel expression by white matter glia: the type-1 astrocyte, Neuron, 5 Ž1990. 527–544. w2x Baum, C.L., Teng, B. and Davidson, N.O., Apolipoprotein B messenger RNA editing in the rat liver: modulation by fasting and refeeding a high carbohydrate diet, J. Biol. Chem., 31 Ž1990. 19263–19270. w3x Bernard, A. and Khrestchatisky, M., Assessing the extent of RNA editing in the TM2 regions of GluR5 and GluR6 kainate receptors during rat brain development, J. Neurochem., 62 Ž1994. 2057–2060. w4x Boss, B.D., Turlejski, K., Stanfield, B.B. and Coean, W.M., On the numbers of neurons in fields CA1 and CA3 of the hippocampus of Sprague-Dawley and Wistar rats, Brain Res., 406 Ž1987. 280–287. w5x Burnashev, N., Monyer, H., Seeburg, P.H. and Sakmann, B., Divalent ion permeability of AMPA receptor channels is dominated by the edited form of a single subunit, Neuron, 8 Ž1992. 189–198. w6x Burnashev, N., Khodorova, A., Jonas, P., Helm, P.J., Wisden, W., Monyer, H., Seeburg, P.H. and Sakmann, B., Calcium-permeable AMPA-kainate receptors in fusiform cerebellar glial cells, Science, 256 Ž1992. 1566–1570. w7x Burnashev, N., Zhou, Z., Neher, E. and Sakmann, B., Fractional calcium currents through recombinant GluR channels of the NMDA, AMPA and kainate receptor subtypes, J. Physiol., 485.2 Ž1995. 403–418.
44
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
w8x Chan, L. and Seeburg, P.H., RNA editing, Sci. Am. Sci. Med., 2 Ž1995. 68–77. w9x Chittajallu, R., Vignes, M., Dev, K.K., Barnes, J.B., Collingridge, L. and Henley, J.M., Regulation of glutamate release by presynaptic kainate receptors in the hippocampus, Nature, 379 Ž1996. 78–81. w10x Crain, B., Cotman, C., Taylor, D. and Lynch, G., A quantitative electron microscopic study of synaptogenesis in the dentate gyrus of the rat, Brain Res., 63 Ž1973. 195–204. w11x Davidson, N.O., Carlos, R.C. and Lukaszewicz, A.M., Apolipoprotein B mRNA editing is modulated by thyroid hormone analogs but not growth hormone administration in the rat, Mol. Endocrinol., 4 Ž1990. 779–785. w12x Dingledine, R., Hume, R.I. and Heinemann, S.F., Structural determinants of barium permeation and rectification in non-NMDA glutamate receptor channels, J. Neurosci., 12 Ž1992. 4080–4087. w13x Edwards, F.A., Konnerth, A., Sakmann, B. and Takahashi, T., A thin slice preparation for patch clamp recordings from neurones of the mammalian central nervous system, Pfluger’s Arch., 414 Ž1989. ¨ 600–612. w14x Huettner, J.E., Glutamate receptor channels in rat DRG neurons: activation by kainate and quisqualate and blockade of desensitization by Con A, Neuron, 5 Ž1990. 255–266. w15x Hollmann, M., Hartley, M. and Heinemann, S., Ca2q permeability of KA-AMPA-gated glutamate receptor channels depends on subunit composition, Science, 252 Ž1991. 851–853. w16x Katz, B. and Miledi, R., The role of calcium in neuromuscular facilitation, J. Physiol., 195 Ž1968. 481–492. w17x Kohler, M., Burnashev, N., Sakmann, B. and Seeburg, P.H., Deter¨ minants of Ca2q permeability in both TM1 and TM2 of high affinity kainate receptor channels: diversity by RNA editing, Neuron, 10 Ž1993. 491–500. w18x Kuffler, S.W., Nicholls, J.G. and Martin, R.A., From Neuron to Brain, Sinauer, Sunderland, MA, 1984. w19x Kung, S., Wu, Y. and Chow, W., Characterization of two fish glutamate cDNA molecules: absence of RNA editing at the QrR site, Mol. Brain Res., 35 Ž1996. 119–130. w20x Laurie, D.J. and Seeburg, P.H., Regional and developmental heterogeneity in splicing of the rat brain NMDAR1 mRNA, J. Neurosci., 14 Ž1994. 3180–3194. w21x Levison, S.W. and McCarthy, K.D., Astroglia in culture. In Culturing NerÕe Cells, MIT Press, Cambridge, MA, 1991. w22x Lomeli, H., Mosbacher, J., Melcher, T., Hoger, T., Geiger, J.R.P., Kuner, T., Monyer, H., Higuchi, M., Bach, A. and Seeburg, P.H., Control of kinetic properties of AMPA receptor channels by nuclear RNA editing, Science, 266 Ž1994. 1709–1713. w23x Melcher, T., Maas, S., Herb, A., Sprengel, R., Seeburg, P.H. and Higuchi, M., A mammalian RNA editing enzyme, Nature, 379 Ž1996. 460–463.
w24x Muller, T., Moller, T., Berger, T., Schnitzer, J. and Kettenmann, H., ¨ ¨ Calcium entry through kainate receptors and resulting potassiumchannel blockade in bergmann glial cells, Science, 256 Ž1992. 1563–1566. w25x Otis, T.S., Raman, I.M. and Trussell, L.O., AMPA receptors with high Ca2q permeability mediate synaptic transmission in the avian auditory pathway, J. Physiol., 482.2 Ž1995. 309–315. w26x Ottiger, H., Gerfin-Moser, A., Del Principe, F., Dutly, F. and Streit, P., Molecular cloning and differential expression patterns of avian glutamate receptor mRNAs, J. Neurochem., 64 Ž1995. 2413–2426. w27x Puchalski, R.B., Louis, J.C., Brose, N., Traynelis, S.F., Egebjerg, J., Kukekov, V., Wenthold, R.J., Rogers, S.W., Lin, F., Moran, T., Morrison, J.H. and Heinemann, S.F., Selective RNA editing and subunit assembly of native glutamate receptors, Neuron, 13 Ž1994. 131–147. w28x Rosenheimer, J. and Smith, D.O., Age-related increase in soluble and cell surface-associated neurite-outgrowth factors from rat muscle, Brain Res., 509 Ž1990. 309–320. w29x Ruano, D., Lambolez, B., Rossier, J., Paternain, A.V. and Lerma, J., Kainate receptor subunits expressed in single cultured hippocampal neurons: molecular and functional variants by RNA editing, Neuron, 14 Ž1995. 1009–1017. w30x Rueter, S.M., Burns, C.M., Coode, S.A., Mookherjee, P. and Emeson, R.B., Glutamate receptor RNA editing in vitro by enzymatic conversion of adenosine to inosine, Science, 267 Ž1995. 1491–1494. w31x Sambrook, J., Fritsch, E.F. and Maniatis, T., Molecular Cloning: A Laboratory Manual, 2nd edn., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1989. w32x Schnarr, R.L. and Schaffner, A.E., Separation of cell types from embryonic chicken and rat spinal cord: characterization of motoneuron-enriched fraction, J. Neurosci., 1 Ž1981. 204–217. w33x Sommer, B., Kohler, M., Sprengel, R. and Seeburg, P.H., RNA ¨ editing in brain controls a determinant of ion flow in glutamate-gated channels, Cell, 67 Ž1991. 11–19. w34x Steward, O., Torre, E.R., Tomasulo, R. and Lothman, E., Neuronal activity up-regulates gene expression, Proc. Natl. Acad. Sci. USA, 88 Ž1991. 6819–6823. w35x Verdoorn, T.A., Burnashev, N., Monyer, H., Seeburg, P.H. and Sakmann, B., Structural determinants of ion flow through recombinant glutamate receptor channels, Science, 252 Ž1991. 1715–1718. w36x Williams, B.P., Abney, E.R. and Raff, M.C., Macroglial cell development in embryonic rat brain: studies using monoclonal antibodies, fluorescence activated cell sorting, and cell culture, DeÕ. Biol., 112 Ž1985. 126–134. w37x Wu, R.L. and Barish, M.E., Astroglial modulation of transient potassium current development in cultured mouse hippocampal neurons, J. Neurosci., 14 Ž1994. 1677–1687.
Research report
Glutamate receptor editing in the mammalian hippocampus and avian neurons Deborah L. Lowe a , Klaus Jahn b, Dean O. Smith b
c,)
a Neurosciences Training Program, UniÕersity of Wisconsin, Madison, WI 53706, USA Physiologisches Institut der Technischen UniÕersitat 80802 Munchen, Germany ¨ Munchen, ¨ ¨ c Pacific Biomedical Research Center, UniÕersity of Hawaii, Honolulu, HI 96822, USA
Accepted 14 January 1997
Abstract RNA editing determines receptor kinetics and permeability of glutamate receptors. This post-transcriptional modification alters single nucleotides within an RNA transcript changing the codon specified by the genome resulting in the incorporation of a different amino acid, profoundly affecting the properties of the protein subunit. We have studied the three sites subject to RNA editing within the kainate-specific subunit GluR6 in the mammalian hippocampus to determine developmental changes and cell-specific variation in editing. GluR6, when measured in the whole rat hippocampus, is predominantly expressed in the unedited form at E18, with a gradual progression to the edited form during the 1st post-natal week, and remains stable from P8 through 30 months. Individual neurons from P0 through P8 rat hippocampal slices analyzed with single-cell PCR show predominant expression of fully edited GluR6, unlike the population profile. In contrast, single astrocytes from P0 hippocampal cultures show that the most common variant is partially edited. Thus, editing in neurons and glia differs, and this difference accounts for part of the disparity between single-neuron and whole-hippocampus data. Editing in astrocytes is affected by conditions in the external environment, as purified astrocytes fail to edit GluR6, although editing occurs in astrocytes from hippocampal cultures. The homogeneity of GluR6 editing between species was also determined by comparing editing in avians and mammals. Genomic and cDNA analysis of chick glutamate receptors demonstrates avian editing of GluR2 but not GluR6. Keywords: RNA editing; Hippocampus; Glutamate receptor; Development; Chick; Single-cell PCR
1. Introduction The process of cellular development and maturation is regulated by alterations in cell gene expression. In the nervous system, genetic expression is controlled transcriptionally through positive and negative regulation and posttranscriptionally via alternative splicing and RNA editing w8,20x. Glutamate receptors are the only protein in the nervous system known to be edited w17,22,33x. RNA editing in glutamate receptors involves an enzymatic-base modification at a single nucleotide in an RNA transcript w23,30x. Thus, an amino acid not specified by the genome is incorporated into the protein subunit significantly altering a receptors functional properties. Receptor characteristics determined by editing include desensitization, rectification and ionic permeability. ) Corresponding author. University of Hawaii, 2444 Dole Street, Bachman 204, Honolulu, HI 96822, USA. Fax: q1 Ž808. 956-8061; E-mail: [email protected]
In the AMPA-receptor family, desensitization kinetics are controlled by editing in the glutamate receptor subunits ŽGluR. 2 through 4 at a site preceding the putative fourth transmembrane region w22x. Ionic permeability w5,12,15x and rectification w35x are determined by GluR2 editing at a site in the pore region w33x. In the kainate-receptor family ionic permeability and receptor kinetics are determined by editing of GluR6 w7x. Editing occurs in two residues of the putative first transmembrane region ŽTM1. resulting in an isoleucine ŽI. to valine ŽV. conversion ŽIrV site. and a tyrosine ŽY. to cysteine ŽC. conversion ŽYrC site., and at one residue of the pore region where a glutamine ŽQ. to arginine ŽR. conversion occurs ŽQrR site. w17x. GluR5 is also edited at the QrR site, but its function is unknown w17,33x. Editing of GluR6 is complex, as each site is edited with a different efficiency creating eight possible molecular variants. Editing in both the kainate- and AMPA-receptor families is developmentally regulated, as unedited receptors appear at early stages of development and are less com-
0169-328Xr97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 9 - 3 2 8 X Ž 9 7 . 0 0 0 7 2 - 7
38
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
mon in adult animals w3,5,17,22x. This regulation indicates that RNA editing could play a role in altering kinetic and ionic properties of glutamate receptors at critical points in proliferation, migration and synapse formation in the developing nervous system. In order to understand how GluR6 editing is related to cell differentiation and development, one must establish the extent of editing at all three sites on individual transcripts, as receptor phenotype is a function of editing at each site. The techniques employed in studies thus far allow only single sites to be examined at one time, making it impossible to know the proportion of GluR6 variants that occur in each of the eight possible forms. A thorough analysis of GluR6 editing requires sequencing of individual transcripts from regions of the CNS and from individual neurons to determine the extent of editing in a given region and to define the molecular constructs responsible for receptor characteristics measured in single cells. To address the question of GluR6 editing in development, we have determined the relative proportions of GluR6 variants present in whole rat hippocampus from E18 through 30 months of age, from individual neurons in hippocampal slices from P0 through P8 and from astrocytes in hippocampal cultures. We have found that editing in single neurons and single astrocytes during the 1st post-natal week does not reflect population measures. We also report neuron-glia differences in the extent to which GluR6 transcripts are edited and cell-culture effects on astrocyte editing. The conservation of glutamate receptor editing between mammalian and avian genes was determined, as RNA editing of GluR2 is absent in fish w19x. We found that glutamate receptor editing in avians is not conserved for all subunits. 2. Materials and methods 2.1. Hippocampal neuronsr glia culture for single-cell astrocyte PCR Cell cultures were prepared from hippocampi of P0 or P8 rats. Hippocampal neurons were mechanically dissociated following incubation with papain Ž40 Urml, 60 min, 328C; Worthington Biochemicals, Freehold, NJ.. Dissociated cells were plated on collagen coated 35 mm petri dishes ŽCollaborative Biomedical, Bedford, MA. at a final density of 10 5 –10 6 cellsrplate. Cells were incubated at 378C in an humidified atmosphere of 5% CO 2 and 95% air in minimal essential media ŽMEM. ŽGibco, Gaithersburg, MD. supplemented with 10% horse serum ŽGibco., 10% fetal bovine serum ŽGibco., glucose Ž6 mgrml., penicillin Ž50 Urml. and streptomycin Ž50 mgrml. for 1–2 days. Prior to obtaining cytoplasm from single astrocytes via whole-cell patch-clamp, the cells were washed with divalent-free phosphate-buffered saline and incubated with 0.05% trypsin for 3–5 min to detach neurons and expose
the underlying bed of glia. The dish was then perfused with normal extracellular saline Žin mM: 125 NaCl, 5 KCl, 1 MgCl 2 , 10 Hepes, 2 CaCl 2 and 10 glucose, pH 7.3 w290 mOsmx. to remove the enzyme and detached neurons. 2.2. Chick motoneurons Chick a-motoneurons were prepared using techniques described in detail by Rosenheimer and Smith w28x. Lumbar spinal cords from 6.5 d embryonic chicks were dissected and freed of their meninges and dorsal root ganglia. Following incubation at 378C with 0.05% trypsin ŽGibco. and 0.005% DNase I ŽSigma, St. Louis, MO., the cells were dissociated mechanically. Motoneuron-enriched fractions were generated on the basis of their buoyant density in a metrizamide solution w32x. 2.3. Purified type-1 astrocytes Type-1 astrocytes were purified from hippocampi of six P2 rats. Cell suspensions were prepared by the method described in Williams et al. w36x for macroglia dissociation from P2–12 brains. Modifications to the technique included use of 1% collagenase ŽWorthington Biochemicals. for 20 min at 378C for the initial enzyme digestion and no further use of collagenase following trypsinization. Tissue was placed in Dulbecco’s modified Eagle’s medium plus 10% fetal calf serum ŽDMEM-F10. and mechanically dissociated with a 5 ml pipette, a 21 G needle and finally a 23 G needle. Cells were plated at a density of 10 7 cellsr75 cm2 collagen-coated flask and incubated at 378C in a humidified atmosphere of 5% CO 2 and 95% air. These cultures were fed every 3 days with DMEM-F10. After 8 days in culture, type-1 astrocytes were separated from oligodendrocytes, neurons, macrophages and type-2 astrocytes by overnight shaking Ž100 rpm. on a rotary platform at 378C as described by Levison and McCarthy w21x. Cytosine arabinoside Ž2 = 10y5 M. was added 1 and 3 days after shaking to achieve a purity of 95% type-1 astrocytes Ž95% GFAP-positive.. 2.4. RNA extraction and GluR6 amplification from rat hippocampi, purified astrocytes and single cells mRNA was extracted from P0 through 30 month whole rat hippocampi, a population of dissociated chick amotoneurons and a population of purified type-1 astrocytes with the Micro-Fast Track mRNA Isolation Kit ŽInvitrogen, San Diego, CA.. Total RNA was isolated from single neurons in hippocampal slices and single glia in hippocampal cultures by aspirating the cytoplasm from single cells after attaining whole-cell patch-recording conditions. Cytoplasm from single cells was expelled from the patch-recording electrode into a sterile microcentrifuge tube with positive pressure and combined with the following for reverse transcription ŽRT.: Random hexamers Žfinal
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
w1.25 m Mx., 1 m l 10 = RT buffer, 20 U Rnase inhibitor, 50 U reverse transcriptase, deoxyribonucleotide triphosphates – dATP, dCTP, dGTP and dTTP Žfinal w0.75 m Mx., MgCl 2 final w5 mMx and 5 m l 10 mM Tris–HCl, pH 8.3. RT was performed at 428C for 30 min followed by 5 min at 998C and held at 48C until polymerase chain reaction ŽPCR.. PCR was then performed with GluR6-specific primers Žsense: 5X-gcgaattcggacaatggaatggaatggttc-3X ; antisense: 5X-agcggatccgtatacgaagaaatgatgat-3X w33x. and glial fibrillary acidic protein ŽGFAP.-specific primers Žsense: 5Xcaatctcacacaggacctcggc-3X ; antisense: 5X -ataccactcctctgtctcttg-3X . at a final w0.07 m Mx. Eight m l 10 = PCR Buffer, 66 m l DEPC-treated H 2 0, MgCl 2 final w3 mMx and 2.5 U of Ampli-Taq DNA polymerase ŽPerkin Elmer, Branchburg, NJ. were added for a final volume of 100 m l. PCR cycling was as follows: 958C, 2 min, 20 step cycles Ž958C, 1 min; 568C, 1 min; 728C, 1 min., 20 step cycles Ž958C, 1 min; 568C, 1 min; 728C, 1 min with a 5 s extensionrcycle. and 728C for 7 min. RT of GluR6 from whole-hippocampal mRNA and type-1 astroctye mRNA was performed with the GluR6specific antisense primers used in single-cell amplifications. The sense primer specific for GluR6 was added before PCR. RT and PCR were carried out under the same cycling conditions used for single-cell amplification. GluR2, GluR5 and GluR6 were amplified from mRNA of chick a-motoneurons. Sense and antisense primer pairs for cDNA amplification were: 5X-gcgaattcaggaaatgacacgtctgggc-3X and 5X-agcggatccgtgtaggaggagattatgat-3X for GluR2; 5X-gcgaattcttcacaccctacgagtggtataacc-3X and 5Xagcggatcccccacattttctcataggtggag-3X for GluR5. GluR6 primers were listed under single-cell PCR w33x. Antisense primers were used for RT and sense primers were used for PCR with the same protocol used for single cells.
39
2.6. Whole-cell patch conditions and cell identification Transverse hippocampal slices of 150–180 m m were cut from brains of Sprague Dawley rats ŽP0–P8. using a vibratome. Neurons from CA1, CA3 and the stratum radiatum were visually identified under Nomarski optics w13x. Identification was based on morphology and location in the slice. Neurons from CA1 and CA3 had pyramid-shaped cell bodies and were located opposite the dentate gyrus. Neuronal identity was further verified by the absence of GFAP amplification. Glial cells were from P0 hippocampal cultures and were visualized with a 40 = water immersion objective. Astrocytes from hippocampal cultures were identified on the basis of morphology. Cells had either a flattened, epithelioid shape with a few processes or were stellate. Glial identity was verified by positive results for GFAP amplification. GFAP was positively identified by southern blot analysis. Cells were patch-clamped in the whole-cell configuration. Patch pipettes had a tip resistance of 3–5 M V and were filled with 5 m l of the following diethyl pyrocarbonate-treated, autoclaved solution: 150 mM KCl and 10 mM Tris–HCl Žbuffered to pH 7.4 with NaOH w300 mOsmx.. The cells were continuously perfused with oxygenated saline: Žin mM. 125 NaCl, 2.5 KCl, 1.0 MgCl 2 , 0.4 CaCl 2 , 26 NaHCO 3 , 1.25 NaH 2 PO4 and 10 glucose Žbuffered to pH 7.3 with NaOH w320 mOsmx.. The silver wire connected to the patch electrode was dipped in bleach before every recording to remove residual mRNA from the previous cell. Patch pipettes were pulled from borosilicate glass tubing Ž2.0 mm outer diameter, 0.5 mm wall thickness. that had been baked overnight at 2008C. 2.7. Quantification of GluR editing from cDNA and chick genomic DNA
2.5. Genomic DNA amplification Genomic DNA from E20 chicken liver was prepared by standard procedures w31x. DNA was amplified by PCR in 100 m l reactions containing: 50 mM KCl, 10 mM Tris–HCl ŽpH 8.3., 2.5 mM MgCl 2 , 0.02 m M each of dATP, dCTP, dGTP and dTTP, 1.5 m M each of sense and antisense primers and 2.5 U Ampli-Taq DNA polymerase ŽPerkinElmer.. Reactions were run with the cycling used for cDNA amplification. Genomic amplifications were performed with the following sense and antisense primer pairs: 5X-gcgaattccttgcagttgctccactggct-3X ; 5X-agcggatccgagcatgttacctggctatg-3X for GluR6 TM1; 5X-gcgaattcgtttagtccctatgagtggtata-3X and 5X-agcggatccgtatacgaagaaatgatgat-3X for GluR6 TM2; 5X-gcgaattcttcacaccctacgagtggtataacc-3X and 5X-agcggatccaaccggtgtaccttg-3X for GluR5 TM2, and 5-gcgaattcagaagtccaaaccaggag-3X and 5X-agcggatccatgaatatccacttgagac-3X specific for GluR2 TM2 w17,33x. Each primer spans an intron-exon border with the exception of the antisense primer for GluR6 TM2 and sense primer for GluR5 TM2.
The GluR6-, GluR5- and GluR2-specific primers contained restriction enzyme sites for EcoRI and BamHI to allow for cloningrsequencing analysis of GluR transcripts. PCR product from single cells, cell populations and genomic DNA was digested with EcoRI and BamHI ŽGibco., precipitated and gel-purified. Purified product was directionally ligated into M13mp18 or 19 RF-DNA ŽGibco.. Ligation reactions were transformed and positive plaques Žbluerwhite screening. were processed and sequenced ŽSequenase Version 2.0 DNA Sequencing Kit; Amersham, Arlington Heights, IL.. Sequence information determined the proportion of edited variants present in single cells and cell populations and genomically encoded chick sequence ŽFig. 1.. 2.8. Statistical analysis Differences in GluR6 editing as a function of age for whole-hippocampus and single-cell data were determined by single-factor analysis of variance. Data from whole
40
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
Ž F1,7 s 23.7, P - 0.0002. and unedited Ž F1,7 s 25.1, P 0.0001. proportions of GluR6. Post-hoc analyses were not performed, as there were no variability measures at E18, P0 and P2. The profile of GluR6 editing in whole rat hippocampus is an average measure of editing occurring in the heterogeneous cell populations that comprise the hippocampus. To determine if population measures of GluR6 editing reflect those in single cells within this region, we used single-cell PCR to construct a profile of GluR6 editing for neurons in hippocampal slices from CA1, CA3 and the stratum radiatum and in astrocytes from hippocampal cultures. 3.2. Glur6 editing in single hippocampal neurons
Fig. 1. Experimental protocol for analysis of glutamate receptor editing. RNA was extracted from single cells with whole-cell patch clamp or cell populations with standard techniques. BamHI and EcoRI restriction enzyme sites within the GluR-specific primers allowed for directional cloning of transcripts into M13. Positive clones were identified by blue-white screening of transformed cells and sequenced.
hippocampus and single cells at a given age were compared with Student’s t-tests. The number of sequences required to determine the dominant form of GluR6 present in single cells was arrived at by comparing each successive set of 10 sequences collected for whole hippocampus at 30 months. The dominant form of GluR6 was the same in each group of 10; thus, 10 sequences were sufficient to make this determination. Before analysis, editing data from cell populations and single cells were converted to proportions.
Analysis of GluR6 transcript editing in single neurons from P0 through P8 hippocampal slices illustrates that editing in neurons from CA1, CA3 and stratum radiatum is strikingly different from the population profile ŽFig. 3.. First, editing in single neurons at P0 is nearly complete Ž72%. with remaining transcripts being partially edited and no transcripts in the unedited form. This profile reflects that found in whole hippocampus at P8 and not at P0. In P0 single hippocampal neurons, the proportion of GluR6 transcripts in the edited Ž t s 1.6; P - 0.05., partially edited Ž t s 0.7; P - 0.05. and unedited Ž t s 1.0; P - 0.05. variants does not significantly differ from P8 whole hippocampus. Second, there is no change in GluR6 editing in single neurons from P0 through P8. Fully edited GluR6 was always the dominant transcript present in single neurons from CA1, CA3 and the stratum radiatum during this period, unlike the dramatic shift from dominant expression
3. Results 3.1. Glur6 editing in whole hippocampus In whole rat hippocampus, the dominant form of GluR6 transcripts changes from partially edited at E18 to fully edited from P8 through 30 months ŽFig. 2.. At E18 55% of GluR6 transcripts sequenced were partially edited, however; by P8 the proportion of partially edited transcripts had decreased to 15%. The proportion of fully edited transcripts present in whole hippocampus changed in an opposite manner. At E18 only 13% of transcripts were fully edited, while this proportion increased to 83% by P8. The proportion of unedited transcripts decreased in a manner similar to partially edited transcripts in that their numbers were much greater at E18 Ž33%. than at P8 Ž2%.. The shift from partial editing of GluR6 to full editing of GluR6 within the whole hippocampus appeared to be a gradual developmental change occurring during the 1st post-natal week. There was a significant effect of age for the fully edited Ž F1,7 s 20.2, P - 0.0005., partially edited
Fig. 2. GluR6 editing in whole-rat hippocampus from E18 through 30 months. Edited values represent the percentage of transcripts sequenced in which nucleotide substitutions occurred at the IrV, YrC and QrR sites. Unedited values represent the percentage of transcripts in which no nucleotide substitutions occurred. Partial values represent the percentage at which one or two of the possible nucleotide substitutions occurred Ž6 variants ŽTM1, TM1 and TM2.: I,C;Q, V,Y;Q, V,C;Q, I,Y;R, I,C;R, V,Y;R.. E18, P0 and P2 values are from the combined RNA of 3 hippocampi, so no variability measures are possible. P6 through 30 month values are the mean"S.E.M. of editing in three separate animals per age. The number of transcripts sequenced per age was 47"9.
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
Fig. 3. The extent of GluR6 editing in CA1, CA3 and stratum radiatum neurons from P0 through P8 hippocampal slices. Each set of values represents the percentage of transcripts found in the fully edited, partially edited Ž6 variants. or unedited forms from single cells in P0 through P8 hippocampal slices. The number of cells included at each age is listed above the bar graphs. P0 data include four CA1 neurons, one CA3 neuron and one neuron from the stratum pyramidal. P2 data include one CA1 neuron, one CA3 neuron and one neuron from the stratum pyramidal. P6 data include one CA1 neuron and one CA3 neuron. P8 data include one CA1 neuron, four CA3 neurons and one neuron from the stratum pyramidal. Values represent the mean"S.E.M. of all cells collected at each age Ž ns9 sequencesrcell.. Each cell was from a different animal. For statistical analysis all cells at each age were combined.
of partially edited GluR6 in P0 hippocampus to fully edited GluR6 in P8 hippocampus. A third difference is the lack of unedited transcripts in single neurons at all ages measured. With the exception of a few unedited transcripts at P8, no unedited transcripts were found at any other age. The marked differences between the developmental profile of GluR6 editing in whole hippocampus and single neurons indicates that editing in other neuronal populations or glia must account for these disparities. 3.3. Glur6 editing in single astrocytes and purified astrocytes from hippocampus Editing in single astrocytes from hippocampal cultures was quite different from that found in single neurons from hippocampal slices. Fig. 4 illustrates editing of GluR6 in single astrocytes from hippocampal cultures and purified type-1 astrocytes. The majority of transcripts in three single GFAP-positive astrocytes from P0 hippocampal cultures are partially edited Ž80%.. This is significantly different from the proportion of partially edited transcripts in single neurons at P0 Ž t s 11.8; P - 0.05.. In contrast to reports of a lack of GluR6 editing in glial cell lines of cortical origin w27x, we found astrocytes in hippocampal cultures edit GluR6. This difference could be a function of tissue origin Žcortex vs. hippocampus. or
41
Fig. 4. GluR6 editing from single astrocytes in hippocampal cultures and purified type-1 astrocyte cultures. Each set of values represents the percentage of transcripts Žmean"S.E.M.. found in the fully edited, partially edited Ž6 variants. or unedited forms from single astrocytes at P0 Ž ns number of cells.. Each cell was from a different culture. Eleven transcripts were sequenced per cell. Values for purified type-1 astrocytes were determined from a population of cells.
culture conditions. This led us to determine editing in a neuron-free culture of purified hippocampal type-1 astrocytes ŽFig. 4.. In this culture, 100% of the GluR6 transcripts sequenced were unedited at the IrV, YrC and QrR sites Ž10 of 10 transcripts. indicating that culturing astrocytes in the absence of the normal complement of hippocampal cells may alter the post-transcriptional regulation of GluR6. 3.4. Complement of GluR6 isoforms in neurons Õs. glia Single neurons and astrocytes not only differ in the most common form of GluR6 expressed, but also in the order of prevalence of partially edited transcripts present. Table 1 summarizes the relative proportions of GluR6 variants in whole hippocampus, single neurons and single astrocytes at P0. The most common partially edited form in neurons was V,C;Q ŽTM1 fully edited, TM2 unedited., while the most common form in glia was I,Y;R ŽTM1 Table 1 The breakdown of edited GluR6 variants present in combined data from whole P0 hippocampus, single P0 neurons and single P0 astrocytes
IYQ VYQ ICQ IYR VCQ ICR VYR VCR
P0 hippocampus Ž ns 42, 3 animals.
P0 neurons Ž ns 57, 6 cells.
P0 astrocytes Ž ns 34, 3 cells.
Ž15. 36% 0 Ž3. 7% Ž9. 21% Ž4. 10% Ž3. 7% 0 Ž8. 19%
0 Ž1. 2% 0 Ž1. 2% Ž10. 17% Ž3. 5% Ž2. 2% Ž41. 72%
Ž3. 9% Ž1. 3% Ž3. 9% Ž11. 32% Ž2. 6% Ž8. 23% Ž2. 6% Ž4. 12%
n, number of transcripts.
42
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
unedited, TM2 edited.. A second neuron-glia difference was that the partially edited transcripts I,C;Q and I,Y;Q were not found in P0 neurons while all partially edited forms were found in P0 astrocytes. Differences in the complement of GluR6 variants present could lead to phenotypically dissimilar KA receptors in neurons and glia. Editing in whole hippocampus at P0 did not represent a summation of editing found in single neurons and glia at this age. At P0, the dominant form of GluR6 in the whole hippocampus was I,Y;Q ŽTM1 unedited; TM2 unedited. differing from both P0 glia and P0 neurons. Editing in other cell types such as dentate granule neurons and oligodendrocytes might account for this discrepancy. 3.5. Glutamate receptor editing in chicken To determine the homogeneity of glutamate receptor editing between species, we analyzed genomic DNA and cDNA of chick GluR2, GluR5 and GluR6. Genomic analysis of GluR6 revealed that the equivalent of the rat edited form is encoded by the chick genome ŽGluR6ŽV,C;R... cDNA analysis of E6.5 chick motoneurons and spinal cord supported this finding as 100% of transcripts Ž36r36. sequenced were V,C;R. This difference does not extend to GluR2 and GluR5 where the chick genomic sequence is the same as the rat. cDNA analysis demonstrated 100% editing of the QrR site Ž6 of 6 transcripts. of GluR2 in chick motoneurons at E6.5. The edited state of GluR5, however, is unknown because we were unable to amplify GluR5 from chick motoneurons, spinal cord and brain at E6.5 to E11.5. We conclude that GluR5 is not expressed in these structures at these ages. Genomic DNA and cDNA sequenced from chick have 99% homology with rat in TM1 and the pore region. This high homology is similar to that reported for pigeon sequence w26x.
4. Discussion 4.1. GluR6 editing in hippocampus The predominant molecular form of GluR6 RNA in the developing mammalian hippocampus follows a gradual progression from unedited and partially edited variants at E18 to fully edited at P8, with the fully edited form dominant through 30 months. This is consistent with the QrR editing analysis of embryonic brain and structures of the adult hippocampus done by Bernard and Khrestchatisky w3x in which QrR editing in GluR6 was 54% in E18 brain and 70–85% in CA1, CA3 and dentate gyrus of the adult hippocampus. Data from neuronal single-cell PCR conflicts with this profile. At P0, the dominant form of GluR6 in neurons from CA1, CA3 and stratum radiatum is fully edited, indicating that analysis of the whole hippocampus early in development does not accurately depict GluR6 editing in single cells during this period. This discrepancy
can be attributed, in part, to differences in GluR6 editing in neuronal and glial populations at this age. Glial single-cell PCR indicates that the dominant form of GluR6 in astrocytes at P0 is partially unedited, in agreement with the population profile. The ratio of neurons to glia in the developing hippocampus is unknown. It is estimated that glia comprise one half the volume of the adult nervous system w18x. If we assume this to be true of the hippocampus at P0 then combining the single-neuron and single-glia data in a 1 : 1 fashion should create an editing profile similar to P0 whole hippocampus. Combining P0 single-neuron data and P0 astrocyte data yields an editing profile where 49% of GluR6 variants are fully edited, 48% of GluR6 variants are partially edited and 3% are unedited. These proportions are different from P0 hippocampus where 19% are fully edited, 45% are partially edited and 36% are unedited. If GluR6 editing is a developmentally regulated phenomenon, oligodendrocytes and dentate granule neurons are likely sources of the unedited variants. Both cell types continue to proliferate into post-natal life unlike CA1 neurons where neuronal number at P0 is the same as 1 month w4,10x. Glial progenitor cells could also contribute as they have been shown to express unedited GluR6 in vitro w27x. Differences in GluR6 editing between the population profile and single-neuron profile do not persist throughout the life span. By P8, GluR6 editing in whole hippocampus is not significantly different from single-neuron editing. Three possible explanations for this phenomenon are: Ž1. GluR6 expression in hippocampal neurons from CA1, CA3 and non-pyramidal neurons increases to the extent that they express much greater levels of GluR6 than other cell populations so the population profile reflects their expression; Ž2. GluR6 expression is transient in the cell populations contributing the partially edited and unedited variants so they do not contribute to the population profile at all ages; and Ž3. editing of GluR6 increases with maturation so cell populations that mature post-natally eventually edit a large proportion of their transcripts. Support for this third possibility is provided by editing analysis of a single P8 astrocyte from a hippocampal culture Ždata not shown.. The dominant form of GluR6 in this cell was fully edited Ž50%.. The functional significance of GluR6 editing in neurons and glia in this study was not examined. However, both rectification and Ca2q conductance of kainate receptors are determined by GluR6 editing. Ruano et al. w29x have linked rectification properties of kainate receptors in single cultured hippocampal neurons with QrR site editing. Linear I–V relationships were found when QrR was predominantly edited ŽR., while strong inward rectification was common among cells when QrR was unedited ŽQ.. In this study, the QrR site was predominantly edited in single neurons and astrocytes. Therefore, native kainate receptors in these cells would be expected to have linear I–V relationships.
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
In studies of recombinant GluR6, Burnashev et al. w7x found Ca2q conductance to be inversely related to editing at the QrR site. Cells expressing recombinant GluR6 receptors edited at the QrR site had barely detectable Ca2q currents regardless of editing at the IrV and YrC sites. Cells expressing receptors unedited at the QrR site had larger Ca2q currents when the IrV and YrC sites were also unedited than when these sites were edited. In this study, the majority of transcripts were edited at the QrR site indicating that native kainate receptors would conduct a negligible amount of Ca2q. The relationship between GluR6 editing, receptor kinetics and Ca2q conductance makes editing an effective means by which a cell can alter its receptor phenotype in response to external stimuli or, as the result of endogenous events, signal a cell’s state of maturation. Kainate receptors have been identified on pre-synaptic membranes in the CA1 region of the hippocampus w9x and dorsal root ganglion w14x. Pre-synaptic expression of Ca2q conducting kainate receptors could elevate basal levels of Ca2q in the pre-synaptic terminal, thus, facilitating neurotransmitter release w16x. During development, this could alter the probability that a given pre-synaptic terminal will form a functional contact and that this contact will be maintained.
43
tal switch in kainate and AMPA receptor editing in rats most likely alters receptor characteristics in a manner that is advantageous for cell survival and maturation. The absence of GluR6 editing in chick indicates the aspects of development for which this switch is necessary in rat do not exist in chick or are performed by another mechanism. This is not the first instance of novel glutamate receptor characteristics or expression in chick when compared to mammals. Neurons from the chick nucleus magnocellularis express an AMPA receptor that is Ca2q permeable and linearly or outwardly rectifying w25x. This is not what is predicted from expression studies of AMPA subunits where Ca2q permeability is linked to receptors with inward rectification w15x. A second difference between the mammalian CNS and avian CNS has been reported in chick Bergmann glia. In the mammalian CNS, Burnashev et al. w6x found that Bergmann glia cells do not express the AMPA subunit GluR2 ŽGluR2ŽR. precludes Ca2q permeability.. Mammalian Bergmann glia possess an AMPA receptor that is Ca2q permeable and inwardly rectifying w24x as predicted by expression studies. Ottiger et al. w26x found that Bergmann glia cells in the avian CNS express GluR2. The subunit composition and characteristics of AMPA receptors in avian Bergmann glia should, thus, differ from their mammalian counterparts.
4.2. GluR6 editing in glia Single-cell PCR of P0 astrocytes from hippocampal cultures indicate that astrocytes edit GluR6 when cultured with neurons. In contrast, purified hippocampal type-1 astrocytes do not edit GluR6. The absence of editing in purified type-1 astrocytes and glial cell lines leads one to suspect that culture conditions can modify the editing process. Chemical or electrical signals between neurons and the bed of glia upon which they grow could be essential for editing activity. Glia and neurons have been shown to modify each other’s characteristics. Neuronal activity up-regulates gene expression in astroglia w34x, while astrocytes modify the appearance of transient potassium currents in hippocampal pyramidal neurons w37x. In addition, in co-cultures of purified type-1 astrocytes and retinal ganglion cells the density of ion channels expressed by type-1 astrocytes more closely approximates those found in P10 tissue prints from optic nerve w1x. Purified cultures could be used to determine if growth factors, neurotransmitters, electrical activity or other environmental cues regulate or modulate GluR6 editing. A precedent for environmental regulation of RNA editing exists for mammalian apolipoprotein B, which is modified by diet and hormones w2,11x. 4.3. RNA editing in chicken The mechanism by which GluR6 mRNA is edited in rats and humans appears not to be necessary for chick as the edited form is genomically encoded. The developmen-
Acknowledgements This work was supported by grants from the National Institutes of Health ŽNS13600. and the Spinal Cord Research Foundation.
References w1x Barres, B.A., Koroshetz, W.J., Chun, L.L. and Corey, D.P., Ion channel expression by white matter glia: the type-1 astrocyte, Neuron, 5 Ž1990. 527–544. w2x Baum, C.L., Teng, B. and Davidson, N.O., Apolipoprotein B messenger RNA editing in the rat liver: modulation by fasting and refeeding a high carbohydrate diet, J. Biol. Chem., 31 Ž1990. 19263–19270. w3x Bernard, A. and Khrestchatisky, M., Assessing the extent of RNA editing in the TM2 regions of GluR5 and GluR6 kainate receptors during rat brain development, J. Neurochem., 62 Ž1994. 2057–2060. w4x Boss, B.D., Turlejski, K., Stanfield, B.B. and Coean, W.M., On the numbers of neurons in fields CA1 and CA3 of the hippocampus of Sprague-Dawley and Wistar rats, Brain Res., 406 Ž1987. 280–287. w5x Burnashev, N., Monyer, H., Seeburg, P.H. and Sakmann, B., Divalent ion permeability of AMPA receptor channels is dominated by the edited form of a single subunit, Neuron, 8 Ž1992. 189–198. w6x Burnashev, N., Khodorova, A., Jonas, P., Helm, P.J., Wisden, W., Monyer, H., Seeburg, P.H. and Sakmann, B., Calcium-permeable AMPA-kainate receptors in fusiform cerebellar glial cells, Science, 256 Ž1992. 1566–1570. w7x Burnashev, N., Zhou, Z., Neher, E. and Sakmann, B., Fractional calcium currents through recombinant GluR channels of the NMDA, AMPA and kainate receptor subtypes, J. Physiol., 485.2 Ž1995. 403–418.
44
D.L. Lowe et al.r Molecular Brain Research 48 (1997) 37–44
w8x Chan, L. and Seeburg, P.H., RNA editing, Sci. Am. Sci. Med., 2 Ž1995. 68–77. w9x Chittajallu, R., Vignes, M., Dev, K.K., Barnes, J.B., Collingridge, L. and Henley, J.M., Regulation of glutamate release by presynaptic kainate receptors in the hippocampus, Nature, 379 Ž1996. 78–81. w10x Crain, B., Cotman, C., Taylor, D. and Lynch, G., A quantitative electron microscopic study of synaptogenesis in the dentate gyrus of the rat, Brain Res., 63 Ž1973. 195–204. w11x Davidson, N.O., Carlos, R.C. and Lukaszewicz, A.M., Apolipoprotein B mRNA editing is modulated by thyroid hormone analogs but not growth hormone administration in the rat, Mol. Endocrinol., 4 Ž1990. 779–785. w12x Dingledine, R., Hume, R.I. and Heinemann, S.F., Structural determinants of barium permeation and rectification in non-NMDA glutamate receptor channels, J. Neurosci., 12 Ž1992. 4080–4087. w13x Edwards, F.A., Konnerth, A., Sakmann, B. and Takahashi, T., A thin slice preparation for patch clamp recordings from neurones of the mammalian central nervous system, Pfluger’s Arch., 414 Ž1989. ¨ 600–612. w14x Huettner, J.E., Glutamate receptor channels in rat DRG neurons: activation by kainate and quisqualate and blockade of desensitization by Con A, Neuron, 5 Ž1990. 255–266. w15x Hollmann, M., Hartley, M. and Heinemann, S., Ca2q permeability of KA-AMPA-gated glutamate receptor channels depends on subunit composition, Science, 252 Ž1991. 851–853. w16x Katz, B. and Miledi, R., The role of calcium in neuromuscular facilitation, J. Physiol., 195 Ž1968. 481–492. w17x Kohler, M., Burnashev, N., Sakmann, B. and Seeburg, P.H., Deter¨ minants of Ca2q permeability in both TM1 and TM2 of high affinity kainate receptor channels: diversity by RNA editing, Neuron, 10 Ž1993. 491–500. w18x Kuffler, S.W., Nicholls, J.G. and Martin, R.A., From Neuron to Brain, Sinauer, Sunderland, MA, 1984. w19x Kung, S., Wu, Y. and Chow, W., Characterization of two fish glutamate cDNA molecules: absence of RNA editing at the QrR site, Mol. Brain Res., 35 Ž1996. 119–130. w20x Laurie, D.J. and Seeburg, P.H., Regional and developmental heterogeneity in splicing of the rat brain NMDAR1 mRNA, J. Neurosci., 14 Ž1994. 3180–3194. w21x Levison, S.W. and McCarthy, K.D., Astroglia in culture. In Culturing NerÕe Cells, MIT Press, Cambridge, MA, 1991. w22x Lomeli, H., Mosbacher, J., Melcher, T., Hoger, T., Geiger, J.R.P., Kuner, T., Monyer, H., Higuchi, M., Bach, A. and Seeburg, P.H., Control of kinetic properties of AMPA receptor channels by nuclear RNA editing, Science, 266 Ž1994. 1709–1713. w23x Melcher, T., Maas, S., Herb, A., Sprengel, R., Seeburg, P.H. and Higuchi, M., A mammalian RNA editing enzyme, Nature, 379 Ž1996. 460–463.
w24x Muller, T., Moller, T., Berger, T., Schnitzer, J. and Kettenmann, H., ¨ ¨ Calcium entry through kainate receptors and resulting potassiumchannel blockade in bergmann glial cells, Science, 256 Ž1992. 1563–1566. w25x Otis, T.S., Raman, I.M. and Trussell, L.O., AMPA receptors with high Ca2q permeability mediate synaptic transmission in the avian auditory pathway, J. Physiol., 482.2 Ž1995. 309–315. w26x Ottiger, H., Gerfin-Moser, A., Del Principe, F., Dutly, F. and Streit, P., Molecular cloning and differential expression patterns of avian glutamate receptor mRNAs, J. Neurochem., 64 Ž1995. 2413–2426. w27x Puchalski, R.B., Louis, J.C., Brose, N., Traynelis, S.F., Egebjerg, J., Kukekov, V., Wenthold, R.J., Rogers, S.W., Lin, F., Moran, T., Morrison, J.H. and Heinemann, S.F., Selective RNA editing and subunit assembly of native glutamate receptors, Neuron, 13 Ž1994. 131–147. w28x Rosenheimer, J. and Smith, D.O., Age-related increase in soluble and cell surface-associated neurite-outgrowth factors from rat muscle, Brain Res., 509 Ž1990. 309–320. w29x Ruano, D., Lambolez, B., Rossier, J., Paternain, A.V. and Lerma, J., Kainate receptor subunits expressed in single cultured hippocampal neurons: molecular and functional variants by RNA editing, Neuron, 14 Ž1995. 1009–1017. w30x Rueter, S.M., Burns, C.M., Coode, S.A., Mookherjee, P. and Emeson, R.B., Glutamate receptor RNA editing in vitro by enzymatic conversion of adenosine to inosine, Science, 267 Ž1995. 1491–1494. w31x Sambrook, J., Fritsch, E.F. and Maniatis, T., Molecular Cloning: A Laboratory Manual, 2nd edn., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1989. w32x Schnarr, R.L. and Schaffner, A.E., Separation of cell types from embryonic chicken and rat spinal cord: characterization of motoneuron-enriched fraction, J. Neurosci., 1 Ž1981. 204–217. w33x Sommer, B., Kohler, M., Sprengel, R. and Seeburg, P.H., RNA ¨ editing in brain controls a determinant of ion flow in glutamate-gated channels, Cell, 67 Ž1991. 11–19. w34x Steward, O., Torre, E.R., Tomasulo, R. and Lothman, E., Neuronal activity up-regulates gene expression, Proc. Natl. Acad. Sci. USA, 88 Ž1991. 6819–6823. w35x Verdoorn, T.A., Burnashev, N., Monyer, H., Seeburg, P.H. and Sakmann, B., Structural determinants of ion flow through recombinant glutamate receptor channels, Science, 252 Ž1991. 1715–1718. w36x Williams, B.P., Abney, E.R. and Raff, M.C., Macroglial cell development in embryonic rat brain: studies using monoclonal antibodies, fluorescence activated cell sorting, and cell culture, DeÕ. Biol., 112 Ž1985. 126–134. w37x Wu, R.L. and Barish, M.E., Astroglial modulation of transient potassium current development in cultured mouse hippocampal neurons, J. Neurosci., 14 Ž1994. 1677–1687.