HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 48, 46374-46384, November 29, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/48/46374    most recent M203032200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Desai, A. Articles by Buonanno, A. Search for Related Content PubMed PubMed Citation Articles by Desai, A. Articles by Buonanno, A.

Analysis of Transcriptional Regulatory Sequences of the N-Methyl-D-aspartate Receptor 2A Subunit Gene in Cultured Cortical Neurons and Transgenic Mice^*

Anand Desai, Dorothy Turetsky, Kuzhalini Vasudevan, and Andres Buonanno^§

From the Section of Molecular Neurobiology, NICHD, National Institutes of Health, Bethesda, Maryland 20892-4480

Received for publication, March 28, 2002, and in revised form, September 25, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The postnatal appearance and up-regulation of the NR2A subunit of the N-methyl-D-aspartate receptor contributes to the functional heterogeneity of the receptor during development. To elucidate the molecular mechanisms that regulate the neural and developmental specific expression of NR2A, an upstream ~9-kb region of the gene harboring the promoter was isolated and characterized in transgenic mice and transfected cortical neurons. Transgenic mouse lines generated with luciferase reporter constructs driven by either 9 or 1 kb of upstream sequence selectively transcribe the transgene in brain, as compared with other non-neural tissues. Reporter luciferase levels in dissociated cultures made from these mice are over 100-fold greater in neuronal/glial co-cultures than in pure glial cultures. Analysis of NR2A 5'-nested deletions in transfected cultures of cortical neurons and glia indicate that while sequences residing upstream of 1079 bp augment NR2A neuronal expression, sequences between 486 and 447 bp are sufficient to maintain neuronal preference. An RE1/NRSE element is not necessary for NR2A neuron specificity. Furthermore, comparison of the 5'-deletion constructs in cortical neurons grown for 5, 8, 11, or 14 days in vitro indicate that sequences between 1253 and 1180 bp are necessary for maturational up-regulation of NR2A. Thus, different cis-acting sequences control the regional and temporal expression of NR2A, implicating distinct regulatory pathways.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The targeting of ion channels and neurotransmitter receptors to their correct cellular location is crucial for the proper development and connectivity of the nervous system. The first level of the intricate spatio-temporal organization required to form the central nervous system is ensuring that genes are transcribed in the appropriate cell types. Although studies on a small number of central nervous system-specific genes have significantly contributed to the identification of cis- and trans-acting elements necessary for neuron-specific transcription (1-4), the expression profiles of different neuronal genes are very diverse. This diversity provides a unique opportunity to identify novel mechanisms that confine gene expression to selective cell populations during development.

Glutamate is the major excitatory neurotransmitter in the mammalian central nervous system, and its signals are transduced by activation of either G-protein-coupled metabotropic receptors or ligand-gated ion channels, referred to as ionotropic receptors. Three families of ionotropic receptors are recognized based on their pharmacological agonist preference for N-methyl-D-aspartate (NMDA),¹ -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), or kainate. NMDA receptors are especially interesting because of their high Ca²⁺ permeability (5, 6) and unique gating properties. In addition to ligand binding, postsynaptic membrane depolarization is necessary to remove a Mg²⁺ ion residing in the channel that blocks ion flux at neuronal resting potentials (7, 8). The requirement for concomitant ligand binding and postsynaptic membrane depolarization makes the NMDA receptor an ideal coincidence detector of synaptic function. Consistent with these properties, NMDA receptors have been implicated in the induction of long term potentiation (9) and in the refinement of synaptic connections during development (10-13). The overstimulation of NMDA receptors can result in excitotoxic neuronal death (reviewed in Ref. 14).

NMDA receptors are composed of multiple subunits, which are encoded by three families of genes: NR1 (Grin1), NR2 (Grin2), and NR3 (Grin3). The obligatory NR1 subunit is encoded by a single gene (15, 16) but is alternatively spliced to produce eight distinct isoforms (17). NR2 subunits are encoded by four separate genes (NR2A-NR2D) (Grin2a-Grin2d) (18-21). These subunits, when hetero-oligomerized with NR1, largely determine the electrophysiological and pharmacological properties of the receptor such as single channel conductance, inactivation kinetics, Mg²⁺ sensitivity, and affinity for a variety of competitive and non-competitive antagonists (reviewed in Refs. 22 and 23). The recently described NR3 subunits are encoded by two genes and have been termed modulatory subunits because they down-regulate receptor function when co-assembled into NR1/NR2 receptors (24, 25).

In addition to the different physiological properties that NR2 subunits impart on NMDA receptors, their expression profiles differ during development. Whereas NR2B and NR2D mRNAs are already present during fetal development, NR2A and NR2C begin to be expressed postnatally (26, 27). In the telencephalon, where NR2C and NR2D are minimally expressed and NR2B is ubiquitously present from mid-gestation onward, the postnatal appearance and dramatic up-regulation of NR2A is a major determinant of receptor heterogeneity in maturing neurons. The timing of NR2A up-regulation and the faster inactivation kinetics this subunit imparts on the NMDA channel led to the idea that NR2A-containing receptors may define critical periods (28, 29). Although a recent report (30) found this view inconsistent with critical periods associated with the refinement of somatosensory maps, the appearance of NR2A-containing receptors may play a role in defining critical periods in visual cortex. In contrast to NR2B, NR2A selectively localizes at synapses (31, 32), suggesting a unique role in synaptic transmission. In fact, NR2A knock-out mice have a marked reduction in long term potentiation in hippocampal CA1 pyramidal neurons and manifest behavioral deficits in the Morris water maze (33, 34). Given the importance of NR2A, it is of considerable interest to elucidate the pathways that modulate its expression.

Multiple mechanisms are utilized to regulate the cellular and subcellular levels of NR2A-containing receptors. In addition to the developmental increase in NR2A steady-state mRNA and protein levels (18, 26, 35-38), expression of NR2A-containing receptors may be translationally and post-translationally controlled. In Xenopus oocytes, secondary mRNA structure directly upstream of the translation initiation site blocks NR2A translation (39). A recent report by Quinlan et al. (40) suggests that translation of dendritic NR2A mRNA may contribute to the rapid pharmacological changes in NMDA receptors observed when dark-reared rats are exposed to light. Similar to AMPA receptors, the surface expression of NMDA receptors can be modulated by insertional dynamics (40-43). Although all of these levels of regulation interact to determine the number of functional NR2A-containing receptors on the cell surface, the proper cellular and temporal transcription of the NR2A gene is a necessary prerequisite for further levels of regulation. With this in mind, we set out to explore the molecular mechanisms involved in the neuronal specific expression and developmental up-regulation of NR2A transcription.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Library Screening and Construct Generation-- To screen genomic sequences mapping to the 5'-end of the NR2A gene, an RT-PCR fragment was amplified from 8-week-old mouse forebrain RNA using primers located 233 bp upstream (5'-AACTATGTGGAGAGAGGCTGC-3') and 25 bp downstream (5'-AGGTCCAGTAGCCCAG-3') of the NR2A translation start site. This 258-bp ³²P-labeled fragment was used to screen an SVJ mouse genomic bacterial artificial chromosome (BAC) library (Genome Systems, St. Louis, MO). Initially, three NR2A-positive BAC clones were isolated. Two of these proved to correspond to the NR2A gene when analyzed on Southern blots probed with the end-labeled upstream primer. Further restriction mapping analyses and Southern blotting using additional oligonucleotide probes demonstrated that both BAC clones encompass similar regions of the NR2A gene. Restriction fragments of ~9 kb (ApaI) and 1 kb (MluI/ApaI), containing the NR2A 5'-UTR sequence, were excised from one of the BACs and subcloned into pBSK(+) (Stratagene). The shorter construct was used to determine the sequence (Applied Biosystems PRISM kit) of the putative core promoter region on an Applied Biosystems PRISM 310 DNA sequencer.

To analyze NR2A transcription regulatory sequences, the 9- and 1-kb genomic fragments were subcloned into the promoter-less luciferase reporter vector pGL-3 Basic; a 33-bp intronic region at the 3'-ends of both fragments had to be removed prior to subcloning. An oligonucleotide primer that generates an artificial XhoI site at the 3'-exonic boundary and a 5'-primer that contains the native MluI site were used to generate an MluI/XhoI PCR fragment that was subcloned into pGL-3 Basic to generate 1253/210 NR2A luciferase (Fig. 3A). The ~9-kb ApaI fragment was also subcloned into pGL-3 Basic, cut with MluI and XhoI to remove ~1 kb of 3'-sequence, and then ligated with the MluI/XhoI PCR fragment to generate the intron-less 9.2-kb/210 NR2A luciferase construct. Subsequent nested 5'-deletions of the NR2A genomic sequence were generated by PCR, using primers that contained 5' MluI and 3' XhoI restriction sites. All PCR products were verified by sequencing and then directionally cloned into pGL3-Basic.

5'-RACE-- A modified rapid amplification of 5' cDNA ends (5'-RACE) protocol, which selects for 5'-capped mRNAs (Ambion, FirstChoice^TM RLM-RACE kit), was used to define the transcription initiation sites for NR2A using recommendations from the manufacturer. Briefly, 1-5 µg of total RNA from DIV 14-15 mixed murine cortical cultures was treated with calf intestinal phosphatase followed by tobacco acid pyrophosphatase to remove the cap structure. T4 RNA ligase was used to ligate the 45-bp RNA adapter, and the RNA was reverse-transcribed with Moloney murine leukemia virus-reverse transcriptase and an NR2A-specific primer (+1/+25) (5'-AGGTCCAGTAGCCCAGTCTGCCCAT-3'). cDNAs were amplified with nested primers to the RNA adapter (5'-GCTGATGGCGATGAATGAACACTG-3') and an antisense NR2A-specific primer bridging the exon 2/exon 3 boundary (28/9) (5'-CTTAGGGTCCCTGTAGCCGG-3'). The first PCR was re-amplified using a nested inner RNA adapter primer (Ambion, RLM-RACE kit) and one of three antisense NR2A-specific primers: the exon 2/exon 3 boundary primer (above), exon 1/exon 2 boundary primer (220/201) (5'- AGCAGGGCTCGCAGCCTCTC-3'), or a primer in the middle of exon 1 (822/802) (5'-TTGGAGCTCTGGTCCGCCTGC-3'). PCR was carried out using Advantage GC Genomic polymerase mix (Clontech), using the following conditions: denaturing at 94 °C for 3 min, 25 cycles amplification 94 °C for 30 s, 60 °C for 30 s, 72 °C for 60 s, and final extension at 72 °C for 7 min. PCR products from the second amplification were resolved on 2% agarose gels, transferred to GeneScreen Plus (PerkinElmer Life Sciences), and analyzed on Southern blots probed with specific nested NR2A primers (230/208, 5'-CTCGCAGCCTCTCTCACATAGTT-3', or 837/818, 5'-CCTGCAGCCGCCCACCCCGG-3'). The NR2A transcription initiation sites were mapped by sequencing the PCR products that were cloned in pGEM-T Easy (Promega) with T7 and SP6 primers using the Applied Biosystems PRISM dRhodamine Terminator Cycle kit.

RNase Protection Assay-- The RNase protection assay (RPA) was performed using the Ambion RPAII kit. Antisense ³²P-radiolabeled probes were synthesized with T3 or T7 RNA polymerase from four subclones of NR2A extending ~1 kb of upstream sequence generated by subcloning the following fragments into pBSK: MluI 1253 to SmaI 1011, SmaI 1011 to SacI 805, SacI 805 to BamHI 673, and BamHI 506 to SacI 289. A non-radiolabeled NR2A sense RNA (MluI to ApaI, 1078 bp) was synthesized in parallel for use in control experiments. All transcripts were gel-purified on 5% acrylamide, 8 M urea denaturing gels to isolate full-length cRNA and sense probes. 20 µg of total adult mouse forebrain or liver RNA were hybridized with ~2-6 × 10⁵ cpm. of probe in 80% formamide, 100 mM sodium citrate, pH 6.4, 300 mM sodium acetate, pH 6.4, and 1 mM EDTA (12 h at 42 °C), treated with 0.5 units of RNase A and 20 units of RNase T1 for 30 min at 37 °C and resolved on 5% acrylamide, 8 M urea denaturing gels. Gels were dried prior to autoradiography.

Northern Blot Analysis of in Vitro Luciferase Expression-- Primary murine cortical cultures containing both neurons and glia were transfected on DIV 11 (as described below) with the luciferase reporter constructs 9.2-kb/210 or 1253/210 NR2A (described above and Fig. 3A). On DIV 14 cultures were harvested and RNA-purified by ultracentrifugation on a CsCl cushion. 20 µg of total RNA was loaded per lane on a 1.2% agarose/MOPS/formaldehyde gel, transferred by electroblotting, and probed with a ³²P-labeled cDNA probe corresponding to the ~1-kb NcoI/AvaI fragment of luciferase, and washed at high stringency. Blots were subsequently exposed to Kodak x-ray film at 70 °C for 5-7 days.

Generation and Analysis of Transgenic Mouse Lines-- To isolate DNA fragments free of plasmid sequences for the generation of transgenic mice, the luciferase reporter vectors 9.2-kb/210 and 1253/210 NR2A (described above) were digested with ApaI/SalI or MluI/SalI, respectively. Fragments were resolved on 0.8% agarose gels and purified by electroelution. Transgenic lines were generated and propagated in FVB/N mice using methods described previously (44). Putative founders and their offspring were screened for transgene integration by Southern blot analysis of tail DNA using a luciferase-specific probe. For luciferase activity analysis, tissues from 6-week-old transgenic mice were collected and homogenized in reporter lysis buffer (Promega) supplemented with a mixture of protease inhibitors (Roche Molecular Biochemicals). Homogenates were centrifuged at 12,000 × g for 10 min at 4 °C to remove debris, and the supernatants were collected for analysis. Luciferase activities were assayed on a Berthold Lumat LB9507 luminometer; activities were normalized to total protein concentration as determined by a Bio-Rad Protein Assay. Samples from wild-type mice were used to assess background values.

Cell Culture-- Mixed cortical cultures, containing both neurons and glia, were prepared as described previously (45). Briefly, dissociated neocortices from E15 mice were resuspended in Eagle's minimal essential medium (MEM, Earle's salts) supplemented with 20 mM glucose, 2 mM glutamine, 5% fetal bovine serum, and 5% horse serum (HS), and plated onto a previously established glial monolayer at a density of 0.5 × 10⁶ cells per pit (in a 24-well plate). Medium was changed after 1 week to MEM supplemented with 20 mM glucose, 2 mM glutamine, and 10% HS, as well as cytosine arabinoside (final concentration 10 µM) to inhibit cell division. Subsequently, cultures were fed once more with 10% HS medium and moved into serum-free medium on DIV 12.

Glial monolayers were prepared from dissociated neocortices of postnatal day 1-2 mice. Cells were plated in Earle's MEM supplemented with 20 mM glucose, 2 mM glutamine, 10% fetal bovine serum, 10% HS, and 10 ng/ml epidermal growth factor at a density of 2 hemispheres per 24-well plate. Plates were coated with poly-D-lysine (100 ng/ml) and laminin (4 ng/ml) on the day prior to the dissection. In experiments where mixed cultures were compared with pure glial cultures, a single glial dissection was used to generate all plates. From the time of neuronal plating, glial and mixed cultures were fed in parallel and treated identically.

Transient Transfection-- Cells were transfected using a modified calcium phosphate technique based on that of Xia et al. (46). DNA purity is a crucial factor in obtaining fine precipitates and good neuronal transfection efficiency, which was typically >2%. All constructs were purified using a spheroplast protocol to reduce lipopolysaccharide contamination (47), followed by alkaline lysis and two rounds of CsCl gradient purification. For most experiments cortical cultures plated in 24-well plates were transfected on DIV 11. Approximately 1 h prior to transfection, cultures were washed into high glucose Dulbecco's modified Eagle's medium (Invitrogen, 11995-073) containing 4 mM kynurenic acid and returned to the incubator to re-equilibrate. Kynurenic acid was included in all transfections performed after DIV 7 to minimize neuronal damage cause by endogenous glutamate release. To form the precipitate, 13.3 µg of double CsCl gradient purified DNA was diluted into 180 µl of H₂O and mixed with 20 µl of 2.5 M CaCl₂. Equimolar amounts of DNA were used in all conditions; for constructs smaller than 9.2-kb/210 NR2A luciferase pBluescript was added for mass adjustment. This DNA solution was then added dropwise to an equal volume (200 µl) of 2× HEPES-buffered saline (HeBS), pH 7.12, and left at room temperature for 25 min to form a precipitate. 2× HeBS contains 275 mM NaCl, 10 mM KCl, 1.4 mM Na₂HPO₄, 15 mM D-glucose, and 42 mM HEPES, brought to pH using NaOH. 45 µl of the DNA/calcium phosphate precipitate was then added dropwise to each well, and cells were returned to the 37 °C, 5% CO₂ incubator for 10 min. The DNA precipitate was gently centrifuged onto the cells by spinning plates for 1 min at 300 rpm (17 × g) in a Savant Cell Culture Centrifuge at room temperature. Plates were returned to the incubator for an additional hour. Transfection was terminated by washing three times with Dulbecco's modified Eagle's medium to remove excess precipitate and returning the cells to growth media (MEM supplemented with 20 mM glucose, 2 mM glutamine, and 3% HS) containing 4 mM kynurenic acid. Cultures were switched into serum-free media containing 4 mM kynurenic acid on DIV 12, and subsequently harvested on DIV 17 in 1× passive lysis buffer (Promega). For developmental analysis, cultures were transfected on DIV 3 without kynurenic acid (which was not necessary at this early date) and harvested either on DIV 5, 8, 11, or 14. Luciferase activities were measured using a Promega assay kit and a Berthold Lumat LB 9507 luminometer.

Immunocytochemistry-- Mixed cortical cultures were transfected with the 9.2-kb/210 NR2A luciferase construct on DIV 6 as described above and were processed for immunocytochemistry on DIV 14. Cells were fixed in 4% paraformaldehyde for 30 min at room temperature, permeabilized in 0.25% Triton X-100, and incubated for 1 h at room temperature in TSA blocking buffer (TSA Fluorescence Kit, PerkinElmer Life Sciences) to block nonspecific binding. Cultures were then incubated overnight at 4 °C in anti-luciferase antibody (Promega, Madison, WI) and anti-MAP2 antibody (Roche Molecular Biochemicals) both diluted 1:500 in TSA blocking buffer. Afterward cells were washed thoroughly and incubated for 45 min at room temperature in CY3 goat anti-mouse IgG (1:200 dilution, Jackson Immunochemicals, West Grove, PA) and Alexa 488 donkey anti-goat IgG (1:400 dilution, Molecular Probes, Eugene, OR) to visualize staining. Nuclei were counterstained by a 10-min incubation in 1 ng/ml Hoechst 33258.

Real Time PCR-- Total RNA was isolated from mixed cortical cultures on DIV 5, 8, 11, or 14, and used as a template for cDNA synthesis (Ambion RETROscript kit, using random decamers). Real time PCR was performed on a LightCycler (Roche Molecular Biochemicals) using SYBR green to measure accumulation of double-stranded product. A 401-bp NR2A product was generated in 1 mM MgCl₂ using a two-step protocol (40 cycles, 90 °C for 0 s, and 72 °C for 80 s) and the following specific primers: NR2A sense (5'-CGCGGATCCGCCTGAGAATGTGGACTTCC-3') and antisense (5'-CCGGAATTCTGTTCTGTGACCAGTCCTGC-3'). Primer pairs specific for NR2B and L7 (48) were used as controls. Because the NR2B and L7 primers had lower melting temperatures, amplification was carried out in 2 mM MgCl₂ using a three-step protocol (40 cycles, 95 °C for 0 s, 58 °C for 10 s, and 72 °C for 30 s). The NR2B and L7 products were 624 and 352 bp, respectively, and PCR primers were as follows: NR2B sense (5'-GCATTTGCCACAATGAGAAGAA-3'), NR2B antisense (5'-CACAGTCATAGAGCCCATCAAT-3'), L7 sense (5'-AGATGTACCGCACTGAGATCC-3'), and L7 antisense (5'-ACTTACCAAGCGACCGAGCAA-3'). Raw fluorescence curves (Fig. 7, A and B) were analyzed using LightCycler software (Roche Molecular Biochemicals) and a second derivative maximum method. Melting curve analysis for NR2A and NR2B confirmed that products obtained with RNA from culture samples corresponded to those synthesized from plasmid template. In all cases PCR products were analyzed on 2% agarose gels to confirm that a single band of the correct molecular weight was obtained.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cloning of NR2A Genomic Sequences and Mapping of Transcriptional Start Sites-- A 258-bp RT-PCR fragment, containing part of the 5'-UTR and extending 25 bp into the NR2A coding sequence, was used to screen an SVJ mouse BAC library for clones harboring genomic sequences mapping to the upstream region of the gene (see "Experimental Procedures"). Initially, three NR2A sequence-positive BAC clones were isolated. Southern blot analysis using NR2A-specific ³²P-labeled oligonucleotide probes demonstrated that two of these clones harbored similar sequences that encompassed the 5'-UTR of the gene and extended upstream (data not shown). Restriction fragments of ~9 kb (ApaI) and 1 kb (MluI/ApaI) of genomic DNA containing NR2A 5'-UTR sequence were subcloned into pBluescript, and the shorter construct was sequenced (Fig. 1).

View larger version (81K): [in this window] [in a new window]   Fig. 1.   Nucleotide sequence of the 5'-end of the mouse NR2A gene. Nucleotide sequence of the NR2A gene extending from 1253 (MluI site) to 23 bp downstream (+26) of the translation initiation codon (boxed); location of introns are indicated. Underlined sequences denote primers used for reverse transcription and PCR in RNA ligase-mediated 5'-RACE (note two primers are split by introns). Restriction enzyme sites utilized to generate probes for the RNase protection assay are shown in boldface type and labeled. The arrows mark the 5'-boundaries of sense () and antisense () primers used to generate deletion constructs. GenBankTM accession number AF493152.

Three independent methods were utilized to examine transcription initiation in the NR2A promoter as follows: 1) 5'-RNA ligase-mediated RACE (5'-RLM-RACE), 2) RNase protection assay (RPA), and 3) Northern blot analysis of in vitro luciferase expression. Initially, a conventional PCR was utilized to map NR2A transcription initiation sites, but because of the multiple bands obtained by this method and reports of secondary structure in NR2A 5'-UTR sequences (39), we used RLM-RACE. This method selects for full-length capped mRNA (see "Experimental Procedures"). RLM-RACE was performed using total RNA purified from DIV 15 murine cortical cultures. To increase the specificity of the RACE and ensure that no cap sites were missed, an NR2A-specific primer encompassing the translation initiation site was used for reverse transcription (+1/+25, with all numbering relative to translation initiation = +1), and the initial PCR was performed using a nested primer (28/9). To maximize for the detection of longer transcripts, the second PCR was carried out with one of the following three different non-coding NR2A-specific primers: 28/9 (as above), 220/201, or 822/802 (see Fig. 1). The products of the second PCR for each of the three non-coding primers are shown in Fig. 2A. As shown in lane 5, a parallel control sample prepared by omitting the tobacco acid pyrophosphatase (to remove the cap structure and make it accessible for primer ligation) did not yield products detectable by ethidium bromide staining. A Southern blot of the same gel, probed with internal NR2A oligonucleotides, demonstrates that all of the strong ethidium bands in lanes 2-4 are NR2A-specific and originate from capped transcripts (Fig. 2B). PCR products for all three primer sets were subcloned into pGEM-T Easy, and 12 clones from the 28/9 primer, 38 clones from the 220/201 primer, and 20 clones from the 822/802 primer were picked and sequenced to identify individual sites of transcription initiation. Of the 70 individual clones picked, 68 yielded NR2A sequences. The clones generated with non-coding primers 28/9 and 220/201 yielded an overlapping set of start sites ranging from 802 to 240 with prominent clusters between 802 and 713, and 461 and 425 (Fig. 2E). Although the non-coding primer 28/9 was specifically picked to allow amplification of transcripts initiating downstream of 220, no start sites were found in this area. The clones generated with non-coding primer 822/802 had start sites between 1199 and 1019, again with a noticeable cluster between 1047 and 1019 (Fig. 2E). Thus the NR2A gene, like other glutamate receptor genes (49), has several clustered transcription initiation sites. Based on the above results, and the comparison of genomic and cDNA sequences, we conclude that the NR2A gene has 2 introns in the region corresponding to the 5'-UTR.

View larger version (54K): [in this window] [in a new window]   Fig. 2.   Mapping of NR2A transcription initiation sites. A, mapping of NR2A transcription initiation sites using RNA ligase-mediated 5'-RACE. An ethidium bromide-stained agarose gel shows products from the second RLM-RACE PCR, using NR2A-specific non-coding primers located at 28/9 (lane 2), 220/201 (lane 3), or 822/802 (lanes 4 and 5). As a control, RNA in lane 5 was not treated with tobacco acid pyrophosphatase (to remove the cap structure) and thus was not accessible for adaptor ligation and subsequent amplification. Lane 1 shows DNA size markers (arrow indicates 500 bp). B, autoradiogram of a Southern blot of the gel shown in A, hybridized with nested NR2A-specific oligonucleotide probes. C, autoradiogram of representative RPAs. Assays were performed with 20 µg of total RNA from adult mouse forebrain (lane 3) or liver (lanes 2, 4, and 5). Lane 1 shows brain RNA hybridized with the MluI/SmaI cRNA NR2A probe (but not treated with RNase A/T1) to indicate integrity of the probe in absence of RNase treatment (FL indicates the full-length 304-bp probe). Treatment of samples with RNase A/T1 generated numerous protected fragments, as well as a band corresponding to the fully protected probe (FP), with forebrain RNA (lane 3). In contrast, the probe is fully digested in the liver sample (lane 2). Lane 5 shows an RPA of liver RNA spiked with a trace amount of in vitro synthesized and unlabeled sense NR2A RNA. In addition to the fully protected fragment, several smaller bands are observed in the spiked sample, which are absent in the liver control sample (lane 4). The (T + C) DNA sequencing ladder to the right was used as size reference (lane M). D, Northern blot of luciferase transcripts driven by NR2A promoter constructs. 20 µg of total RNA isolated from primary murine cortical cultures, transiently transfected with the 1253-kb/210 NR2A (lanes 2 and 3) or 9.2-kb/210 NR2A (lanes 4 and 5) luciferase constructs (see Fig. 3A). Lane 1, a control sample of luciferase transcripts driven by a troponin I promoter in skeletal muscle (single band runs at 1.95 kb). Two independent dissections are denoted by 1 or 2 below the lanes. Approximate size markings are shown to the right. E, schematic of sequences surrounding the NR2A promoter region, determined by the alignment of genomic and cDNA sequences. Introns are depicted in white. Arrowheads mark areas of transcription initiation determined by RLM-RACE, and ATG indicates the translational start site.

RPA was used next to confirm the NR2A transcription initiation sites identified by RLM-RACE. To scan the ~1-kb NR2A upstream region flanking the initiation ATG using RPA, ³²P-labeled cRNA probes were synthesized from four different subclones extending through this region and hybridized to adult mouse RNA (see "Experimental Procedures"). As shown in a representative experiment using the MluI/SmaI probe (Fig. 2C), no protected products were observed with the negative control sample using liver RNA (lane 2), indicating that RNase digestion is complete and the probe does not self-protect. In contrast, numerous protected products were observed with forebrain RNA (lane 3); similar results were obtained in RPAs performed with 3 of the 4 cRNA probes (not shown). In an attempt to determine whether these multiple bands originate from the protection of capped NR2A mRNAs or result from secondary RNA structure, liver RNA was spiked with traces of in vitro synthesized NR2A sense transcripts (see "Experimental Procedures"). Instead of a predicted single product, numerous bands corresponding to products obtained with forebrain RNA were observed (lanes 5), suggesting that a number of the protected fragments in the forebrain RNA sample originate from RNA secondary structure and do not represent bona fide transcription initiation sites. The SacI/BamHI and BamHI/SacI probes had similar problems with secondary structure (data not shown), and indeed the RNA folding program mfold predicts the formation of extensive stem/loop structures throughout the 1253/210 region of the NR2A gene. These results indicate that technical limitations prevent the accurate mapping of NR2A start sites using RPA. However, the RPA analysis did yield two interesting pieces of information. First, the SmaI/SacI probe, which encompassed an area that did not show complications of secondary structure, yielded a single band representing the 210-bp fully protected fragment (data not shown), which confirms the RACE results showing no transcription initiation sites in this area. Second, whereas the three downstream probes (SmaI/SacI, SacI/BamHI, and BamHI/SacI) show strong fully protected bands indicating the initiation of transcripts upstream of 1101, 805, and 506 (data not shown), a relatively small amount of the MluI/SmaI probe was fully protected (Fig. 2C), suggesting that the majority of transcription initiation sites are located between 1253 and 210 of the NR2A gene.

Northern blot analysis was utilized as an alternative experimental approach, not limited by RNA secondary structure, to investigate the possibility of other transcription initiation sites upstream of the ~1-kb area examined by RLM-RACE and RPA. Primary murine cortical cultures were transfected with either 9.2-kb/210 NR2A luciferase or 1253/210 NR2A luciferase constructs (see Fig. 3A and below), allowed to mature, and harvested for RNA. As shown in Fig. 2D, a Northern blot probed with a labeled luciferase cDNA fragment hybridizes with transcripts that run between ~2.0 and 2.8 kb; a control sample containing luciferase transcripts driven by a muscle troponin I promoter in skeletal muscle gave a single band of 1.95 kb. The range of sizes observed for the luciferase transcripts driven by NR2A regulatory sequences, compared with the control, is consistent with the multiple, clustered initiation sites found by RACE. The fact that there is no observable difference in the size of luciferase transcripts between the 9- and 1-kb transfections also supports the RPA data suggesting that the majority of NR2A transcription initiation sites are contained in the 1253/210 fragment.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Tissue specificity of NR2A luciferase constructs in transgenic mice. A, schematic of the 9.2-kb/210 NR2A and 1253/210 NR2A luciferase constructs used to generate transgenic mice (not drawn to scale). B, luciferase reporter levels were analyzed in several brain regions and non-neuronal tissues isolated from five independent transgenic lines (BU2219, BU2202, BU2211, BU2197, and BU2196) harboring the 9.2-kb/210 NR2A construct. C, luciferase expression in two independent 1253/210 NR2A transgenic lines (BU3002 and BU3022) also showed a strong neural preference. Luciferase activity was normalized to µg of total protein.

Tissue Specificity of NR2A Genomic Sequences in Transgenic Mice-- The NR2A gene is specifically expressed in neurons (27, 50-53). To determine whether the ~9-kb genomic ApaI fragment contained the regulatory elements necessary for neuronal specific expression, transgenic mice were generated with this NR2A genomic fragment. To generate this construct (9.2-kb/210 NR2A), intronic sequences at the 3'-end of the ApaI fragment were removed, and the remaining fragment was subcloned into the promoterless luciferase reporter vector pGL-3 Basic (Fig. 3A). Of 12 tail-positive founder lines, 5 displayed whole brain luciferase activities that were clearly above the background levels found in wild-type mice. Analysis of tissue extracts made from 6-week-old transgenic mice from each of these 5 lines showed that the 9.2-kb/210 NR2A sequence conferred central nervous system specificity (Fig. 3B). All transgenes expressed luciferase activity in cortex, hippocampus, colliculus, and cerebellum, with a marked preference for neuronal tissues compared with non-neuronal tissues. To further delineate regulatory sequences important for neural specificity of the NR2A gene, we generated 1253/210 NR2A luciferase transgenic mice (Fig. 3A). As observed with the 9.2-kb/210 NR2A transgenes, the 1253/210 NR2A luciferase construct was preferentially expressed in neural tissues (Fig. 3C). With both fragments some lines had appreciable luciferase activity in heart or adrenal gland; moreover, individual transgenic lines showed relative differences in luciferase activity among the distinct brain regions. These results were not totally unexpected given the precedence that the random site of transgene integration frequently affects expression patterns. Similar observations have been reported for transgenic mice generated with other neural specific promoters (-aminobutyric acid, type A, 6 (54), CaMKII (55, 56), NR2C (57), and NR2B²; for a more extensive discussion see Ref. 58). Transgenic mice generated with constructs containing locus control regions and insulators are less prone to manifest these problems (59).

Neuronal Specificity of NR2A Genomic Sequences in Cortical Cultures-- Having established that the upstream region of the NR2A gene directs brain-specific luciferase expression in transgenic mice, transcription from NR2A constructs was further analyzed in primary murine cortical cultures to determine its cellular specificity and to map more precisely sequences necessary for neuron-specific transcription. Initially, mixed cortical cultures containing both neurons and glia and pure glial cultures were prepared from 9.2-kb/210 NR2A luciferase transgenic mice (line BU2197). When neurons derived from E15 transgenic mice were plated on a confluent glial monolayer prepared from wild-type mice, these cultures displayed high levels of luciferase activity (3.19 × 10⁵ ± 3.4 × 10⁴ S.E. relative light units; n = 4 separate dissections). In stark contrast, when glial monolayers were prepared from BU2197 mice luciferase levels were ~700-fold lower (4.34 × 10² ± 5.35 × 10¹ S.E. relative light units, significantly different from neurons, p < 0.001, t test; n = 3 separate dissections). These results strongly suggested that the central nervous system-specific expression observed in these mice resulted from selective transcription in neurons.

As an alternate approach, mixed cortical cultures were transiently transfected with the 9.2-kb/210 NR2A luciferase construct using a modified CaPO₄ technique (see "Experimental Procedures"). Double immunostaining for luciferase (Fig. 4A) and the neuron-specific marker MAP2 (Fig. 4B) demonstrated that all cells expressing luciferase were neurons (overlay, Fig. 4C). Counterstaining with Hoechst 33258 labeled all nuclei and revealed the presence of numerous glia in the field of view (Fig. 4D). These results demonstrate that the upstream sequences of the NR2A gene are sufficient to direct neuron specificity both in vivo and in vitro, either as integrated or episomal DNA.

View larger version (51K): [in this window] [in a new window]   Fig. 4.   The 9.2-kb/210 NR2A construct drives neuron-specific expression in transiently transfected cortical neurons. Primary murine cortical cultures, containing both neurons and glia, were transiently transfected with the 9.2-kb/210 NR2A luciferase construct and processed for immunocytochemistry. Fluorescent photomicrographs of the same field immunostained for luciferase (A) and MAP2 (B). C, an overlay of luciferase and MAP2 immunostaining demonstrates that all the luciferase-positive cells are neurons. D, counterstaining with Hoechst 33258 reveals neuronal and glial nuclei in the cultures. (Original magnification = 20×; bar = 100 µm.)

Deletional Analysis of the NR2A Upstream Regulatory Region-- A series of NR2A 5'-nested deletion constructs were generated and tested in transfected cortical neurons to identify the cis-acting sequences that confer neuron-specific transcription (see Fig. 5). Because analysis of 1253/210 NR2A luciferase transgenic mice indicated that sequences between 1253 and 210 were sufficient to confer nervous system specificity, we focused further deletional analysis in this area. Mixed cortical cultures containing neurons and glia, and pure glial cultures, were transfected 11 days after the plating of the neurons (DIV 11) in the presence of 4 mM kynurenic acid to minimize neuronal damage caused by endogenous glutamate release. Cultures were maintained in antagonist until DIV 17, when cells were collected and extracts prepared for analysis of luciferase activity (Fig. 6). Consistent with the neuron-specific luciferase expression seen by immunocytochemistry, the 9.2-kb/210 NR2A construct was ~30-fold more active in mixed cultures than in glial cultures. As shown in Fig. 6, a general decrement in luciferase activity was observed in the neuronal cultures as 5'-sequences were removed, most markedly in the deletions removing sequences from 9.2 kb to 1253 bp and from 1133 to 1079 bp. Nevertheless, all deletion constructs between 9.2 kb and 486 bp were expressed at higher levels in the mixed neuronal cultures than in glial cultures (statistically significant, p < 0.05 two-way ANOVA with Tukey post-test). Deletions that disrupted or removed the most downstream cluster of transcription initiation sites (constructs 447/210 and 368/210 NR2A) decreased neuronal luciferase activities to the point where they were indistinguishable from those in pure glial cultures. In addition, two other constructs lacking the downstream end of exon 1 (817/437 and 618/437) had similar luciferase activity to their 210 counterparts (not significantly different, two-way ANOVA), suggesting that sequences between 437 and 210 are not necessary for neuronal expression. To ensure that glial cultures were efficiently transfected, a luciferase reported vector driven by the viral promoter SV40 (pGL-3 SV40) was included in experiments and shown to be expressed at high levels in glial cells (Fig. 6). Taken together, the results suggest that 40 bp of NR2A 5'-sequence (486/447) is sufficient to confer neural selectivity in transiently transfected cortical neurons.

View larger version (36K): [in this window] [in a new window]   Fig. 5.   NR2A luciferase reporter constructs used in transient transfections. Schematics of NR2A 5'-nested deletion constructs used for transient transfection of cortical cultures. All NR2A (Grin2a) sequences were subcloned into pGL-3 Basic, a promoter-less luciferase reporter vector. The resulting constructs are named for the 5'- and 3'-boundaries of NR2A (Grin2a) sequence; +1 corresponds to the translation initiation site. pGL-3 SV40 is a positive control luciferase vector containing the SV40 promoter and enhancer. White boxes represent sequence upstream of the 5'-most transcription initiation site (240), and black boxes depict the common 5'-UTR sequence, and the luciferase gene is shown in light gray.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Mapping of NR2A genomic sequences necessary for neuronal specificity. Activity of NR2A luciferase reporter constructs was compared between mixed cortical cultures, containing neurons and glia (black bars), and pure glial cultures (gray bars). Cultures were transiently transfected with various NR2A luciferase reporter constructs 11 days after the neuronal plating and harvested on DIV 17 to assay luciferase activity. Because NR2A sequences were subcloned into the promoter-less luciferase reporter vector pGL-3 Basic, all values were normalized to the amount of luciferase activity seen in sister cultures transfected with equimolar amounts of pGL-3 Basic. The bars represent the mean ± S.E. of 3 to 27 independent experiments, each done in quadruplicate. An asterisk indicates that luciferase values were significantly different between neuronal and glial cultures for a given construct (p < 0.05, two-way ANOVA with Tukey post-test).

Analysis of NR2A 5'-Deletion Constructs during Maturation of Cortical Neurons-- During brain development, there is a dramatic increase in NR2A expression (26, 27, 35, 36, 38, 60). To determine whether our cortical cultures displayed this up-regulation, the quantitative technique of real time RT-PCR was utilized. RNA was prepared from mixed cortical cultures grown for 5, 8, 11, or 14 DIV. Real time PCR fluorescence curves used for quantification of NR2A and NR2B transcripts from a representative experiment are shown in Fig. 7, A and B, respectively. Levels of NR2A and NR2B mRNAs were measured in cultures grown for various days in vitro and normalized to the value observed on DIV 5 (Fig. 7C); the ribosomal subunit L7 served as a control for RNA amounts (48). Whereas NR2A mRNA levels increased by >7-fold between DIV 5 and 14, neither NR2B nor L7 transcripts showed a significant difference. Thus, the striking developmental up-regulation of NR2A mRNA seen in vivo is recapitulated during the maturation of these mouse cortical cultures, which is consistent with previous observations in other culture systems (61, 62).

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Endogenous NR2A mRNA levels increase during the maturation of murine cortical cultures. Total RNA was isolated from mixed cortical cultures grown between 5 and 14 DIV, reversed-transcribed, and analyzed by real time PCR. Representative fluorescence curves used for measuring the relative levels of NR2A (A) and NR2B (B) transcripts, in cultures grown for 5, 8, 11, and 14 days in vitro (DIV). C, quantification of NR2A, NR2B, and ribosomal subunit protein L7 relative mRNA levels in maturing mixed cortical cultures. Bars represent the mean ± S.E. of 3-5 different dissections. An asterisk indicates that RNA levels were significantly different from DIV 5 levels for a given message (p < 0.05, two-way ANOVA with Tukey post-test).

To identify the cis-elements involved in this maturational up-regulation of NR2A, mixed cortical cultures were transfected with NR2A deletion constructs on DIV 3 and harvested on various subsequent days for luciferase analysis. On DIV 5 all NR2A constructs showed a similar, low level of activity (Fig. 8). As cultures aged, a trend toward increased luciferase activity was observed with all constructs; however, the rate of change was much faster with the two longest constructs. Only cultures transfected with 9.2-kb/210 NR2A or 1253/210 NR2A showed statistically significant differences in luciferase activity levels at DIV 14, compared with DIV 5 (p < 0.05, two-way ANOVA followed by Tukey test). In addition to being transcribed at higher rates than other constructs, 9.2-kb/210 and 1253/210 NR2A were also significantly different from each other. Taken together, these findings indicate that sequences between 9.2 kb and 1253 bp contribute to the maturational increase of NR2A expression and that elements residing between 1253 and 1180 bp are crucial for this up-regulation.

View larger version (26K): [in this window] [in a new window]   Fig. 8.   Identification of DNA sequences that contribute to the developmental up-regulation of NR2A during maturation of cortical neurons. NR2A reporter constructs were transiently transfected into mixed cortical cultures on DIV 3, and cells were harvested on DIV 5, 8, 11, or 14 to analyze luciferase activity. All values were normalized to the amount of luciferase activity seen in sister cultures transfected with pGL-3 Basic. The bars represent the mean + S.E. of 4-9 independent experiments, each done in quadruplicate. On DIV 14, luciferase values for 9.2-kb/210 and 1253/210 NR2A were significantly different from all other constructs (*). On DIV 11, 9.2-kb/210 NR2A was significantly different from constructs 1180/210 and 817/210 (#), and 1253/210 NR2A was significantly different from 817/210 ($). On DIV 8, 9.2-kb/210 NR2A significantly differed from construct 817/210 (&). In all cases p < 0.05 by two-way ANOVA with Tukey post-test.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We have isolated and characterized ~9 kb of upstream regulatory sequence of the NR2A gene and demonstrated that this region is sufficient to confer neuronal specific expression in transgenic mice. Mapping of the intron/exon structure within the proximal 1.2 kb of this sequence reveals two non-coding exons and places the translation start site in the third exon; this structure is similar to that found in the evolutionarily related NR2B gene (63). Several transcription initiation sites were mapped by RACE analysis between 1199 and 240 (relative to the AUG that initiates translation), with three prominent clusters located between 1047 and 1019, 802 and 713, and 461 and 425. No start sites were found downstream of 240; thus this region of the 5'-UTR is common to all transcripts. Although we cannot rule out the possibility of other minor transcription initiation sites upstream of this ~1-kb region, examination of the length of luciferase transcripts in transfected cortical cultures indicates that other major initiation sites are not present in the upstream 9-kb region. The presence of multiple transcription initiation sites in the NR2A gene is typical of characterized glutamate receptor subunits, all of which have multiple clustered transcriptional start sites (63-68), which are sometimes used differentially in different brain regions (67). The NR2A gene is somewhat exceptional, however, in the spacing of these sites throughout an entire kilobase of sequence. Moreover, this area is highly conserved between species, with 92% identity between mouse and rat (GenBank^TM accession number gi17977857) and ~80% identity between mouse and human (GenBank^TM accession number gi22068014), an unusually high degree of homology for sequences residing upstream of a translational start site. These features suggest that a novel mechanism controls NR2A transcription initiation. Given the slightly >200-bp spacing between the three major initiation clusters, which is reminiscent of the ~200 bp winding around and stretching between nucleosomes, it is tempting to speculate that some aspect of chromatin structure is involved.

Because the 9.2-kb/210 NR2A luciferase construct was able to direct neuronal expression both in transgenic mice, and in transiently transfected cortical cultures, we constructed a series of sequential 5'-deletion constructs to isolate the cis-elements important for neuron-specific expression. These constructs were then compared in primary cortical cultures containing both neurons and glia or pure glial monolayers. Whereas neuronal specificity (as measured by neuron to glia ratio) decreased with increases in the size of deletions, the loss of specificity predominantly resulted from the loss of activity in the neuronal cultures. No differences in luciferase expression were observed in glial cultures for any of the deletion constructs. The major (and statistically significant) drops in neuronal activity occurred between 9.2 kb and 1253 bp, between 1133 and 1079 bp, and between 486 and 447 bp. Because the deletions from 1133 to 1079 bp and from 486 to 447 bp both remove some transcription start sites, we cannot rule out the possibility that the decrements in neuronal activity observed are due to the direct loss of initiation sites. However, over the ~1 kb harboring transcriptional start sites (1199 to 240 bp), there appears to be little correlation between loss of initiation sites and loss of activity (in particular, two clusters of initiation sites are lost between 1078 and 617 bp with no concomitant loss of luciferase activity). Thus, while sequences in the 5'-most 8 kbs of this regulatory region and between 1133 and 1079 bp contribute to NR2A activity, neuronal specificity can be maintained by elements residing between 486 and 447 bp.

Presently, the best understood mechanism underlying neuronal specific gene expression is the transcriptional silencing of neuronal genes in non-neuronal cell types. A 21-bp neuron-restrictive silencer, known as RE1/NRSE, was originally described in genes encoding the type II voltage-gated sodium channel (1) and SCG10 (2). Transcription is silenced by the interaction of the RE1/NRSE cis-element with a repressor protein, known as REST/NRSF, which is expressed in non-neuronal cells (3, 4). Several genes expressed preferentially in neurons, such as synapsin (69), Na,K-ATPase (70), choline acetyltransferase (71), and the -aminobutyric acid, type A, 2 subunit (72), harbor functional RE1/NRSE elements (see Ref. 73 for other examples). RE1/NRSE or RE1/NRSE-like sequences have been described in the transcriptional regulatory regions of glutamate receptor subunits. The RE1/NRSE sequences in the NR1 and GluR2 (Gria2) genes contribute to a 2-4-fold transcriptional silencing in glia and other non-neuronal cell types (67, 74). In contrast, an RE1/NRSE-like sequence (76% homologous) in the NR2B gene does not affect transcription in non-neuronal tissues in vivo (63). Functional analysis of the RE1/NRSE-like sequence in the NR2C gene (65) has not been reported. An RE1/NRSE-like sequence (TTCAGCACCAAGGTTGCGCGC) present at position 989 in the NR2A regulatory region does not appear to affect transcription in glial cultures or C2C8 myoblasts (data not shown). This sequence is ~86% homologous to the RE1/NRSE consensus with mismatches in 3 of 13 critical bases (73); a site with a similar degree of homology is found at position 427. As discussed above, none of the individual constructs are statistically different from each other in glial cultures. These results indicate that the neuronal specificity observed with the NR2A constructs expressed in transgenic mice and transfected neurons results from transcriptional activation selectively in neurons, rather than by non-neuronal silencing. Given the combinatorial nature of transcriptional regulation, it is likely that the multiple elements residing in the distal 8-kb region, between 1133 and 1079 bp and between 486 and 447 bp, cooperate to increase neuronal specificity.

Transcription driven by the two longest constructs used in this study, 9.2-kb/210 and 1253/210 NR2A, was markedly increased during the maturation of cortical cultures. Although the developmental increase in steady-state NR2A mRNA levels had been reported previously (26, 27), to our knowledge, this is the first demonstration that transcription is important for NR2A up-regulation. To identify potential cis-acting regulatory elements in the region crucial for the developmental up-regulation of NR2A, a computer search of the MatInspector/TRANSFAC transcription factor data base was performed with sequences residing between 1253 and 1180. The search identified a potential CRE-like element (TGACATCA) at position 1195; a result that is particularly interesting given the role of CREB in activity-dependent gene regulation (reviewed in Refs. 75 and 76), and recent literature (32, 77) suggesting that expression of NR2A-containing NMDA receptors is regulated by activity. The CRE-like element at position 1195 is a variant found in numerous promoters, which was shown by CASTing to preferentially bind CREB/c-Jun heterodimers. However, when this multimerized element was transfected into the human choriocarcinoma cell line JEG-3, it functioned as a relatively poor transactivating site (78). Future deletion and linker scanning experiments through the 1253 to 1180 region will be necessary to explore the possible role of this CRE-like sequence, and other novel elements, required for the developmental regulation of the NR2A gene in neurons.

The identification of the NR2A promoter region and the characterization of sequences that contribute to the neuron- and development-specific expression of the NR2A subunit represent a significant step toward understanding the transcriptional regulation of this important gene. Elucidating the mechanisms involved in NR2A transcriptional activation, as well as other processes regulating the expression of NR2A-containing NMDA receptors, will contribute to our understanding of how the nervous system responds to developmental and environmental cues to regulate NMDA receptor heterogeneity.

	ACKNOWLEDGEMENTS

We thank Daniel Abebe for technical assistance and Dr. Detlef Vullhorst for the Northern blot and many helpful discussions.

	Note Added in Proof

While this paper was in revision, a preliminary characterization of the 5' flanking and 5' untranslated regions of the rat NR2A gene was reported. It showed the same intron/exon boundary and a transcription start site mapping to our most downstream cluster (Richter, M., Suau, P., and Ponte, I. (2002) Gene (Amst.) 295, 135-142).

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EBI Data Bank with accession number(s) AF493152.

Both authors contributed equally to this work.

^§ To whom correspondence should be addressed: LDN, NICHD, Bldg. 49, Rm. 5A38, 49 Convent Dr., Bethesda, MD 20892-4480. Tel.: 301-496-3298; Fax: 301-496-9939; E-mail: buonanno@helix.nih.gov.

Published, JBC Papers in Press, September 27, 2002, DOI 10.1074/jbc.M203032200

² M. Sasner and A. Buonanno, unpublished results.

	ABBREVIATIONS

The abbreviations used are: NMDA, N-methyl-D-aspartate; AMPA, -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; DIV, days in vitro; ANOVA, analysis of variance; RT, reverse transcriptase; RACE, rapid amplification of cDNA ends; RLM-RACE, 5'-RNA ligase-mediated RACE; RPA, RNase protection assay; MOPS, 4-morpholinepropanesulfonic acid; MEM, minimal essential medium; HS, horse serum; UTR, untranslated region; CRE, cAMP-response element; CREB, cAMP-response element-binding protein; BAC, bacterial artificial chromosome; RE1/NRSE, restrictive element 1/neuron-restrictive silencer element; REST/NRSF, RE1 silencing transcription factor/neuron-restrictive silencer factor.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES