Address correspondence and reprint requests to Dr Guang Bai, Department of Oral & Craniofacial Biological Sciences, University of Maryland Dental School, 666 W. Baltimore Street, Baltimore, MD 21201, USA. E-mail: GNB001@dental.umaryland.edu
To understand the genetic mechanism controlling the expression of the NMDA subtype of glutamate receptors during neuronal differentiation, we studied activation of the N-methyl-D-aspartate receptor subunit 1 (NR1) gene and the role of the repressor element-1 (RE1) element in NR1 promoter activation. Following neuronal differentiation of P19 embryonic carcinoma cells, the NR1 transcription rate and mRNA level were significantly increased, while the nuclear level of the repressor RE1 silencing transcription factor (REST)/neuron-restriction silencer factor (NRSF) was reduced. Nuclear REST/NRSF from undifferentiated cells formed a large complex with the NR1 RE1 element. While this complex was significantly reduced after the differentiation, REST/NRSF from differentiated cells formed a new, faster migrating complex. In transient transfections, deletion of the RE1 element increased activity of the 5.4-kb NR1 promoter sixfold in undifferentiated cells, but only induced approximately 1.4-fold increase in differentiated cells. Forced expression of REST/NRSF in differentiated cells suppressed the promoter, while forced expression of a dominant-negative REST/NRSF induced promoter activity as well as the mRNA of the NR1 gene in undifferentiated cells. In stable transfectants, the wild-type promoter showed a robust increase in activity following differentiation in a pattern similar to the NR1 mRNA increase. Conversely, the promoter lacking the RE1 element showed only a moderate increase. Our data suggest that the NR1 gene up-regulation during neuronal differentiation is controlled by its promoter activation, which is largely determined by the interaction between the RE1 element and the repressor REST/NRSF.
Neuronal differentiation is a process through which postmitotic neuronal precursors become mature by gaining neuronal functions and phenotypes. This maturation results from the expression of many groups of genes at different stages of the differentiation following a designed genetic program, the mechanism of which is largely unknown (Bertrand et al. 2002). It has been well documented that many neurotransmitter receptor genes, such as GABA receptors (Siegel 1998), glutamate receptors (Monyer et al. 1994) and dopamine receptors (Chen and Weiss 1991; Gurevich et al. 1999), undergo robust up-regulation during the neuronal differentiation characteristic of the neonatal period in the developing brain. Of the glutamate receptors, the functional N-methyl-D-aspartate (NMDA) subtype expresses postnatally, and plays a crucial role in the continuation of differentiation by supporting cell survival and synaptogenesis (Haberny et al. 2002). There are three families of NMDA receptor genes in mammalian brain: N-methyl-D-aspartate receptor subunit 1 (NR1), NR2A-D and NR3A-B. The product of the NR1 gene is essential for functional NMDA receptors and can be found in almost all neurons of the central nervous system (Haberny et al. 2002). In rodents, its mRNA level appears at a low level in the embryonic brain from gestational day 14 and robustly increases in the developing brain within three weeks of birth (Watanabe et al. 1992; Akazawa et al. 1994; Monyer et al. 1994). The increase in the steady-state level of the NR1 mRNA may be caused by an increase in the transcription rate and/or decrease in the mRNA degradation rate. So far, which mechanism controls the change of the NR1 mRNA level during the neonatal period is largely unknown. This question also remains unanswered for all other genes encoding the glutamate receptors.
Recent studies have revealed that transcription factors, such as NeuroD (Lee et al. 1995) and neurogenins (Anderson 1999), play important roles in neurogenesis and neuronal differentiation (Bertrand et al. 2002). Following expression of these factors, neurogenesis and neuronal differentiation are initiated and neuronal genes are expressed. These studies suggest that activation of transcription may be a mechanism controlling mRNA increase during neuronal differentiation. This concept is also supported by evidence obtained from cell culture models of neuronal differentiation. For example, over-expression of transcription factors that are activated at the early stage of neuronal differentiation of P19 cells induced neuronal differentiation of PC12 pheochromocytoma cells (Itoh et al. 1997; Noma et al. 1999; Mandolesi et al. 2002). P19 cells are a pluripotent, embryonic carcinoma cell line and, after appropriate induction by retinoic acid (RA), can enter neurogenesis and then neural differentiation to form neurons and glial cells. Many neuronal genes are found in differentiated P19 cells that form functional neurons (Bain et al. 1994; MacPherson and McBurney 1995; Queva et al. 1998). Interestingly, NR1 mRNA, protein and functional NMDA receptors are also found in differentiated P19 cells (MacPherson et al. 1997), so that this cell line may serve as an ideal model to study the NR1 gene up-regulation during neuronal differentiation.
We previously identified a 3-kb promoter of the rat NR1 gene and observed that a negative cis-element, repressor element 1 (RE1) (also named neuron-restriction silencer element, NRSE), located in the 5′ untranslated region participates in tissue specific regulation (Bai et al. 1998). Potential RE1 elements were found in more than 324 genes in one survey of the Celera mouse database (Roopra et al. 2001) and 1047 sites were found in another survey of the human and mouse genomes (Lunyak et al. 2002). Many of these genes are neuronal and their RE1 elements interact with the trans-acting factor, RE1 silencing transcription factor (REST)/neuron-restriction silencer factor (NRSF) (Schoenherr et al. 1996; Roopra et al. 2001). This factor is a 116-kDa zinc-finger protein that migrates at 210-kDa when fully glycosylated, and contains two repressor domains at the N- and C-termini. After binding to DNA, REST/NRSF recruits several cofactors via its repressor domains and negatively regulates the transcription rate of a cognate gene (Roopra et al. 2001). However, it has also been reported that the RE1 element can activate the promoter of the neuronal nicotinic acetylcholine receptor β-2-subunit gene in cultured PC12 and neuroblastoma cells (Bessis et al. 1997) as well as the promoter of the dynamin 1 gene in neuroblastoma cells (Yoo et al. 2001). The REST/NRSF mRNA was found to be at a high level in neuroprogenitors of the embryonic nervous system, but at a low level in the brain from the late prenatal stage to the adult in rodents (Chong et al. 1995; Schoenherr and Anderson 1995). In cell lines, down-regulation of the REST/NRSF mRNA was observed after neuronal differentiation (Nishimura et al. 1996; Palm et al. 1999). However, whether the change of mRNA level results in relevant alteration of REST/NRSF protein level and in turn its ability to regulate neuronal genes is not well documented. In addition, whether the regulation of REST/NRSF expression contributes to the activation of the NR1 gene during neuronal differentiation remains largely unknown. In the present study, we first studied the mechanism underlying NR1 mRNA induction and increase during neuronal differentiation, using nuclear run-on and reporter gene techniques in a cellular model of neuronal differentiation. Then, we investigated the contribution of the RE1 cis-element and the REST/NRSF protein to the activation of a 5.4-kb NR1 promoter. Our data indicate that during neuronal differentiation the NR1 transcription rate and mRNA level in P19 cells are significantly activated, and that the RE1 element, the REST/NRSF protein and their interaction play an important role in activation of the NR1 gene.
RA is purchased from Eastman Kodak (New York, NY, USA). A monoclonal antibody against the N-terminal sequence of the REST/NRSF (12C11) was kindly provided by Dr D. J. Anderson. DNA plasmid of pB-ζ1 harboring the mouse ζ1 (NR1–1a) cDNA was from Dr M. Mishina, and pB-GAPDH bearing the mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA was from Dr D. Rowitch. Expression construct of human REST/NRSF (plasmid REEX1) was a kind gift of Dr G. Mandel and a dominant-negative mouse NRSF (plasmid mtNRSF) was from Dr D. J. Anderson.
Isolation of 5.4-kb promoter sequence of the NR1 gene and reporter gene construction A PromoterFinder DNA Walking kit (Clontech, Palo Alto, CA, USA) was utilized to clone the genomic sequence upstream of the previously isolated 3-kb NR1 promoter (Bai and Kusiak 1993). A rat genomic library was provided by Clontech (Palo Alto, CA, USA). An NR1 promoter specific primer, GB25 (5′-TTGGCTCAGGGTCAGCCTAGGGATAC), was designed to be 149 bp downstream of the 5′ end of the previously identified 3-kb promoter (Bai and Kusiak 1993). This design should guarantee the fidelity of the newly isolated upstream sequence. PCR products obtained were sequenced as described previously (Bai and Kusiak 1993). Four clones were sequenced and compared to exclude the possibility of mutations caused by the DNA polymerase. DNA sequence of the newly isolated fragment has been deposited in GeneBank with the accession number of AY157515.
Construction of the 5.4-kb promoter-luciferase fusion gene was accomplished as follows. A 4.5-kb sequence was excised from the 5.4-kb promoter at a 5′-Sma1 site in the vector and at the Kpn1 site at the position of − 919 (Bai and Kusiak 1993). This fragment was inserted into the Kpn1 site of pGL3basic (Promega, Madison, WI, USA) with a ligation-fill-in-ligation strategy to form the construct pNR1-5.4/919. Restriction enzyme digestion was used to identify the correct orientation. Then, the sequence of − 919/− 1, prepared from pNRL919 (Bai and Kusiak 1995) with Kpn1 and Hind3, was inserted into pNR1-5.4/919 at Kpn1-Hind3 sites to generate pNRL5.4. The mutant with the deletion of the RE1 site was produced by inserting the − 919/− 1 sequence from pNRL3029ΔRE1 (Bai and Kusiak 1997) into pNR1-5.4/919 as described above to form pNRL5.4ΔRE1.
Nuclei were isolated from P19 cells using a method described by Blum (1989). The isolated nuclei were resuspended in storage buffer B [50 mm Tris-HCl, pH 8.3, 5 mm MgCl2, 0.1 mm EDTA, 16% (v/v) glycerol] and saved at − 80°C until needed. Nascent RNAs were synthesized from 107 nuclei in the presence of 0.5 mm of ATP, CTP and GTP each, RNasin at 1 unit/µL (Promega) and 100 µCi of α-32P-UTP (800 Ci/mmol, ICN, Aurora, OH, USA) at 30°C for 30 min in the buffer described by Blum (1989). Then, RNA was extracted with a guanidinium isothiocyanate/phenol method (Chomczynski and Sacchi 1987). Newly synthesized and radiolabeled RNA was broken down with 0.2 m NaOH at room temperature for 5 min, and then neutralized with 1 m Tris base solution. RNA fragments were precipitated with ethanol and dissolved in DEPC-treated dH2O. Plasmid DNA of pB-ζ1 was linearized with Hind3, pB-GAPDH with EcoR1 and pBSKII as control vector with EcoR1. Linearized DNA was blotted to nitrocellulose membrane at 5 µg per spot after denaturation by 0.1 m NaOH at room temperature for 5 min. DNA on the spot was prehybridized with ExpressHyb (Clontech) at 64°C for 30 min in a 2-mL cryogenic vial. The nascent RNA fragments were added to DNA blot at 4 × 106 cpm/mL and incubated at 64°C for 2 h. The blots were then washed with 15 mL of 2 × saline sodium citrate buffer (SSC), 0.05% (w/v) sodium dodecyl sulfate (SDS) at room temperature four times and 0.2 × SSC, 0.05% SDS at 42°C for 30 min twice. The membranes were exposed to X-ray film or a PhosphorImager plate (Typhoon, Amersham, Piscataway, NJ, USA).
Isolation of RNA and quantitative RT-PCR
Total cellular RNA free of genomic DNA contamination was extracted from cultured cells with an Absolutely RNA kit following the manufacturer's instructions (Stratagene, La Jolla, CA, USA). To quantify the NR1 mRNA level, we developed a homologous competitive RT-PCR method. We created a homologous competitor by inserting a 77-bp unspecific sequence into mouse ζ1 (NR1–1a) cDNA (Yamazaki et al. 1992) at the Avr2 site in the sequence encoded by exon 16. This sequence was inserted into Bluescript SKII (Stratagene) to form plasmid pBζ1–77. The upstream primer for PCR was GB448 (5′-GTCCTCTGCCATGTGGTTTT), homologous to exon 15 sequences, and the downstream primer was GB447 (5′-GTCATTGATGCCTGTGATGC), complementary to sequences downstream of the insertion. Full length mRNA of the NR1 gene was synthesized from linearized pBζ1–77 by RNA polymerase T3 with an in vitro RNA transcription protocol as described previously (Bai and Kusiak 1995). The homologous competitor was mixed, at 2.5 × 10−19 molar, with 2 µg of total RNA for each sample and subject to reverse-transcription (RT) reaction with a random primer and Superscript II (Invitrogen, Carlsbad, CA, USA) at 42°C for 30 min. One twentieth of the RT products went to a 50-µL PCR reaction catalyzed by ExTaq DNA polymerase (TaKaRa/PanVera, Madison, WI, USA) in the presence of 1 unit of PerfectMatch reagent (Stratagene). The reaction was cycled 32 times with a program of 20 s at 94°C, 30 s at 58°C and 30 s at 72°C. Other PCR reactions either by less or by more cycles were tested and the 32 cycles generated the maximal difference of products in amount. In order to gain more signal, in some experiments, the RT products were amplified for NR1 cDNA with primers of GB435 (5′-TGTCTCCTACACAGCTGGCTT) and GB436 (5′-GCAGCGTCGTCCTCGCTTGCA) using conditions above for 35 cycles.
GAPDH mRNA was measured by cycling 1/20 RT reaction with primers of GB450 (5′-ACCACAGTCCATGCCATCAC) and GB 449 (5′-TCCACCACCCTGTTGCTGTA), which are able to amplify GAPDH cDNA from human, rat and mouse under following conditions: 20 s at 94°C, 30 s at 58°C and 30 s at 72°C for 25 cycles. The predicted PCR product is 452 bp for the mouse GAPDH cDNA.
RT-PCR of mouse REST/NRSF mRNA was carried out with 1/10 RT reaction used for NR1 RT-PCR and with primer GB477 (5′-GACTCATCTAACGCGACACATGCGG) and GB478 (5′-GCATGTCGGGTCACTTCATGCTGATT) for PCR. ExTaq DNA polymerase was used to amplify the DNA. The reaction was cycled 35 times with a program of 20 s at 94°C, 30 s at 58°C and 30 s at 72°C. One fifth of reaction of each sample was fractionated on 1% (w/v) agarose plus 2% low-melt agarose in TAE buffer in the presence of 10 µg/mL ethidium bromide. The gel was scanned and analyzed as described previously (Liu et al. 2001) using a Kodak CF440 system and D1 software (PerkinElmer, Boston, MA, USA).
Cell culture, transfection and reporter gene assay
The P19 embryonic carcinoma cell line was cultured in α-minimum essential medium (MEM) (Sigma) supplemented with 10% fetal bovine serum (FBS) and differentiated with RA as described previously (Krainc et al. 1998). Briefly, P19 cells were treated with 0.3 µm RA for 48 h and then cultured on Petri dish in the presence of a 2nd dose of RA (30 µm) for 24 h to form aggregates. Aggregated cells were trypsinized and platted back to cell culture dishes in the absence of RA to initiate neural differentiation. In some experiments, 24 h later, Ara-C at 5 µg/mL was added to cells for 4 consecutive days with medium change every day to produce neuron-rich culture. Time of RA treatment was counted from the first day when the RA was added. Transfection of the cells was conducted with lipofectamin2000 following manufacturer's instructions (Invitrogen). For forced expression of wild-type or dominant-negative REST, expression construct and the NR1 promoter-luciferase construct at a ratio of 4:1 were cotransfected into P19 cells. For transient transfections, LacZ cDNA driven by the human cytomegolovirus (CMV) promoter was used to normalize transfection efficiency. Transfectants were harvested for reporter gene assay for undifferentiated cells 24 h and for differentiated cells 48 h after transfection. To interfere with the endogenous NR1 expression, construct mtNRSF expressing a dominant-negative NRSF/REST (Chen et al. 1998) was transfected into cells. In general, 3 µg DNA for cells on 10-cm2 culture surface was used. Twenty-four hours after transfection, cells are harvested and total cellular RNA was extracted for RT-PCR. To establish stable tranfectants, reporter gene constructs were cotransfected with pcDNA3, that contains a neomycin resistant gene, and cells were screened by treatment with 200 ng/mL of geneticin (G418, Invitrogen) to remove cells not stably transfected. The reporter gene constructs pNRL5.4 and pNRL5.4ΔRE1 were used to establish the stable cell lines P19NR1-5.4 and P19NR1-5.4ΔRE1, respectively. Cells were differentiated as described above for different time.
Activity of reporter gene luciferase and β-galactosidase was measured by chemiluminescence methods as described previously (Liu et al. 2001), except that cells were lysed with lysis buffer provided by the Galacto-Star kit (AB/Tropix, Foster City, CA, USA). Protein concentration in the lysates of stable transfectants was assayed by the Brandt method (Bio-Rad, Hercules, CA, USA). Luciferase activity was normalized to activity of cotransfected β-galactosidase in transient transfections, and to protein concentration in stable transfectants. Copy numbers of the NR1 promoter-luciferase gene in stable transfectants of P19 cells were measured by a slot blot analysis of genomic DNA on the basis of a standard curve of pGL3-basic DNA as described previously (Bai and Kusiak 1997).
Electrophoresis mobility shift assay (EMSA)
EMSA was performed as described in previous studies (Bai and Kusiak 1995). Nuclear proteins were extracted from cells as described previously (Bai and Kusiak 1995). The long DNA probe encompassing the NR1 RE1 site was formed by annealing an upper-strand primer, GB428 (5′-GCCAAACACGCTTCAGCACCTCGGACAGCA), with a lower-strand primer, GB429 (5′-GCGCGGCGGATGCTGTCCGAGGTGCTGAAG) to form uneven ends. The short probe consisting of the 20-bp consensus of the RE1 element was formed first by annealing a pair of primers with uneven ends: GB537 (5′-GCTTCAGCACCTCGGAC) as the upper strand and GB538 (5′-GCTGTCCGAGGTGCTGA) as the lower strand. Then the annealed primers were labeled by filling in the uneven ends with the Klenow enzyme in the presence of α-32P-dCTP (> 3000Ci/mmol, ICN). EMSA experiments were conducted as described previously with 10 µg of nuclear proteins per reaction (Bai and Kusiak 1995). In some experiments nuclear proteins were incubated before the radio-labeled probe was added to the reaction with one of the following reagents: unlabeled probe or non-specific probe at 50 times the concentration of the labeled probe or 1 µL of antibody against REST/NRSF or Sp1. Radioactivity associated with each shifted band was recorded by a PhosphoImager and analyzed by Kodak D1 Image Analysis software as described previously (Liu et al. 2001).
Immunoblot analysis and immunocytochemistry
Immunoblot analysis of nuclear REST/NRSF was performed as described previously (Liu et al. 2001). Briefly, 20 µg of nuclear proteins or 40 µg of total cell lysates were denatured and fractionated on a 10% (w/v) NuPAGE gel in a running buffer containing SDS (Invitrogen). To indicate relative size of protein species and blotting efficiency, prestained BenchMark (Invitrogen) was loaded. Sizes of the prestained protein standards on the NuPAGE gel were corrected on the basis of those obtained from MultiMarker provided by the manufacturer (Invitrogen). Proteins were then transferred to a polyvinylidene difluoride membrane and analyzed with a REST/NRSF antibody (1 : 200 dilution) (12C11, provided by Dr D. J. Anderson). Immunocomplexes were visualized by an enhanced chemiluminescence reagent (Pierce, Rockford, IL, USA) and recorded by exposure to X-ray film. Signals associated with bands of interest were quantitatively analyzed with Kodak D1 software as stated previously (Liu et al. 2001).
For immunocytochemistry of NeuN or glial fibrillary acidic protein (GFAP), P19 cells were plated on collagen-coated coverslips and fixed with 4% paraformaldehyde-PBS for 4 min at room temperature. Cells were permeabilized with 0.7% triton X-100 and incubated with antibody against NeuN or GFAP at 1: 8000 dilution (Chemicon, Temecula, CA, USA). The second antibody was conjugated with Cy2 (Jackson ImmunoResearch Laboratory, West Grove, PA, USA). For counterstaining, cells were incubated with 20 µg/mL of propidium iodide (PI) for 5 min at room temperature. Cy2 was visualized by green fluorescence and PI by red fluorescence with a Nikon E8000 microscopy system equipped with an RT slide on CCD camera. Images of both Cy2 and PI were taken for each view of the cells. Twenty random views were pictured for cells obtained from every time point after differentiation. The positive-staining cells were counted by a combination of Photoshop software and NIH Image software. In each view, between 200 and 400 cells were counted. Mean values ± SE were calculated from all 20 views of each time point.
The mRNA level and transcription rate of the NR1 gene were increased in P19 cells during neuronal differentiation
RA treatment promotes P19 cells to enter neurogenesis and then neuronal differentiation. Neurons begin to appear 4 days after the initiation of treatment and continue differentiation for more than 10 days (MacPherson and McBurney 1995). Previous studies revealed that P19 cells treated with RA for 8 days (Ray and Gottlieb 1993) or 12 days (MacPherson et al. 1997) showed a detectable level of the NR1 mRNA in comparison to untreated cells. In order to obtain the complete profile of NR1 mRNA expression during the neurogenesis and differentiation, we treated P19 cells with RA for varying numbers of days. In one group of experiments Ara-C was added 4 days after RA treatment to inhibit growth of proliferative cells. Figure 1(aii) shows one example of cells 8 days after RA treatment in comparison with cells before the treatment (Fig. 1ai). In cells without Ara-C treatment, non-neuronal cells predominated 8 days after RA treatment (Fig. 1aiii). Next we measured NR1 mRNA in the cells treated with RA for different times. To quantify the NR1 mRNA level, we established a homologous competitive RT-PCR protocol that included an in vitro synthesized mouse NR1–1a mNRA containing a 77-bp insert in the common region of all NR1 splice variants is a competitor (refer Fig. 1b for schematic illustration). This competitor not only functioned in the PCR reaction, but also controlled the RT efficiency of individual RNA samples that can be contaminated by trace amounts of inhibitory chemicals from the extraction process (Freeman et al. 1999). Using this protocol, we found that the increase in the NR1 mRNA level started 4 days after treatment and continued until 9 days after RA treatment in cells treated with Ara-C (Fig. 1c,d). In cells without additional Ara-C treatment, there was no detectable NR1 mRNA under the experimental conditions (after 32 cycles of PCR) we used (data not shown). To ensure the quality of RT and PCR, we examined the presence of a house-keeping gene, GAPDH, in the RNA samples and found that its mRNA was present in all samples at stable levels (Fig. 1c). These results suggest that the increase in the NR1 mRNA may be associated with neurons during differentiation. To further support this concept, we examined the relative population of neurons in the culture treated with Ara-C by staining neuronal cells with an antibody against the neuronal marker NeuN and labeling all nuclei with propidium iodide (MacPherson et al. 1997). We found that the percentage of NeuN positive cells stayed relatively stable: 5 days after RA, 40.66% ± 2.38 (mean ± SE); 8 days, 43.95% ± 2.43; 10 days, 58.63% ± 2.49. In comparison to an 18% increase in NeuN positive cells in the culture during the same time period, the increase in NR1 mRNA was more than fivefold. Therefore, we believe that the NR1 mRNA is associated with neurons and that the NR1 mRNA level per neuron is increased following neuronal differentiation.
Next, to explore the mechanism underlying this increase, we performed nuclear run-on experiments. Using nuclei isolated from P19 cells that had been treated with RA for 4 and 8 days, we observed that the NR1 transcription rate was significantly increased during RA-induced neuronal differentiation while the transcription of the GAPDH gene was relatively stable (Fig. 1e).
Promoter activity in differentiating and undifferentiated P19 cells
The transcription rate of a given gene is genetically controlled by its promoter and regulatory elements located in a region that spans the 5′ end of the gene and its upstream sequence. The size of this region can vary from hundreds to thousands of base pairs of DNA sequence (Blackwood and Kadonaga 1998). We have previously studied a 3-kb sequence containing the NR1 promoter, and an RE1 element in the region encoding the 5′ untranslated region. In order to evaluate the role of the RE1 element in regulating the promoter and its possible requirement for other upstream cis-elements (Lonnerberg et al. 1996; Mieda et al. 1997), we isolated an additional 2.4-kb sequence upstream of the previously studied 3-kb sequence using a genomic PCR approach. To clone more 5′-flanking sequences, a gene-specific PCR primer was designed to be 149-bp downstream of the 5′ end of the previously identified 3-kb promoter. Four clones of PCR products were sequenced and all had sequence overlapping the previous 3-kb promoter. Within this newly isolated region reside consensus sequences for several transcription factors, including SRF and nuclear factor-kappa B (Fig. 2a, and Genebank database with accession number of AY157515). We linked this newly isolated 2.4-kb fragment to the previously isolated 3-kb sequence to form a 5.4-kb NR1 promoter.
As the first step to examine the role of the RE1 element in controlling the NR1 promoter, we deleted a 17-bp core sequence of the RE1 from the previously studied 3-kb promoter and the 5.4-kb NR1 promoter and compared promoter activity of these mutants with the wild-type promoter in driving the transcription of a luciferase reporter gene. In transient transfections of undifferentiated P19 cells, the 5.4-kb promoter showed much stronger promoter activity than the 3-kb promoter. In addition, deletion of the RE1 site significantly increased activity of both promoter fragments and the increase in the 5.4-kb promoter is more than sixfold over that of the wild-type promoter (Fig. 2b). These results suggest that the 5.4-kb fragment harbors more regulatory elements than the 3-kb fragment. Therefore, we utilized the 5.4-kb fragment in all subsequent studies. In cells differentiated with RA for 8 days and enriched for neurons by Ara-C treatment, the RE1 mutant showed only about 1.4-fold increase in activity (Fig. 2b). These data suggested that the RE1 element suppresses the promoter in non-neuronal P19 cells, but not in differentiated P19 cells.
The REST/NRSF protein is known as the major transcription factor interacting with the RE1 element and, as suggested by results in Fig. 2(b), may inhibit the NR1 promoter in undifferentiated P19 cells. To confirm this possibility, we employed a dominant-negative (dn) REST/NRSF to interfere with the function of REST/NRSF in P19 cells, and then examined whether this interference could de-repress the NR1 promoter. As shown in Fig. 2(c), forced expression of the dnREST/NRSF robustly increased activity of the cotransfected 5.4-kb NR1 promoter, while there was no significant change of the promoter activity after wild-type REST/NRSF was cotransfected. These results suggest that REST/NRSF plays an important role in suppression of the NR1 promoter in undifferentiated P19 cells. To further test if this suppression may be a key factor for silencing of the NR1 gene, we transfected undifferentiated P19 cells with the dnREST/NRSF expression construct and measured the NR1 mRNA using an RT-PCR protocol. As can be seen in Fig. 2(e), the NR1 mRNA was significantly induced by overexpression of dnREST/NRSF, but not by transfected vector DNA. In contrast, the mRNA levels of the house-keeping gene, GAPDH, remained stable in transfected cells. These data strongly suggest that the REST/NRSF protein plays an important role in the silencing of the NR1 gene in undifferentiated P19 cells.
Epi-chromosome expression of transiently transfected genes may not accurately represent the dynamic regulation of a promoter located in a chromosome. To overcome this problem, we stably transfected the promoter-reporter gene constructs into the chromosomes of P19 cells. We established stable transfectants P19NR1-5.4, carrying the 5.4-kb promoter-luciferase gene, and P19NR1-5.4ΔRE1, bearing a 5.4-kb NR1 promoter with the RE1 element deleted. Using a slot blot analysis of genomic DNA, we found that the copy number of integrated promoter-reporter gene is similar in each cell line (1.03 for P19NR1-5.4 and 1.21 for P19NR1-5.4ΔRE1). Under undifferentiated conditions, however, the P19NR1-5.4ΔRE1 cell line had a much higher luciferase activity (24 493 ± 877.8 unit/µg protein, mean ± SE) than the P19NR1-5.4 cell line (2700 ± 475 units/µg protein from six independent experiments). We differentiated these stable transfectants into neural cells with RA, measured luciferase activity and normalized it to protein content. The normalized luciferase activity was related to values obtained on day 0 of RA treatment. As shown in Fig. 2(d), relative activity of the wild-type promoter increased robustly following 4 days of RA treatment and became stabilized following 6 days of treatment. The maximal increase was about 73-fold over that in undifferentiated cells and the luciferase activity remained stable as long as the neurons remained viable. In contrast, the RE1 mutant showed only approximately a 10-fold maximal increase during the same time period, indicating that the promoter activity of the mutant does not need significantly further relieving. These data suggest that the RE1 element plays a key role in the up-regulation of the NR1 promoter during neuronal differentiation.
Nuclear REST/NRSF and its binding on the NR1 promoter were down-regulated in P19 cells during neuronal differentiation
After defining the roles of the RE1 element and the REST/NRSF protein in the regulation of the NR1 promoter activity, we investigated the interaction of the REST/NRSF protein with this DNA cis-element in P19 cells. Although mRNA has been detected for the REST/NRSF gene in P19 cells (Palm et al. 1999), one report failed to find nuclear binding activity on the NR1 RE1 element with a short DNA probe consisting only of the 20-bp consensus of the NR1 RE1 element (Okamoto et al. 1999). Our results above indicated that the RE1 element is involved in the regulation of the NR1 promoter in the P19 cells. Therefore we tested whether undifferentiated P19 cells have REST/NRSF binding activity on the NR1 promoter. In EMSA experiments, we tested this possibility using a radioactively labeled probe covering the 20-bp sequence of the NR1 RE1 element plus an additional 10-bp flanking sequence at both ends. As shown in Fig. 3(a), this probe formed a slowly migrating complex with nuclear proteins extracted from P19 cells. We named it complex A. This binding was specifically competed out by unlabeled probe, but not by an unrelated Sp1 consensus. Importantly, this complex was super-shifted by a monoclonal antibody against the REST/NRSF. In a comparison, a short probe consisting of 20-bp consensus of the NR1 RE1 site failed to produce any binding complex under the same conditions even after overexposure of the gel to X-film for 2 days with an intensifying screen at − 80°C or 17 h to a PhosphorImage plate (Fig. 3b). These results demonstrate that a REST/NRSF-like protein is present in undifferentiated P19 cells and forms a binding complex with the NR1 RE1 element.
Next, we differentiated P19 cells with RA plus Ara-C treatment, prepared crude nuclear proteins at different times after the treatment, and examined the capability of these proteins to bind the NR1 RE1 element. As can be seen in Fig. 3(c), binding complex A was largely reduced in cells after differentiation. Interestingly, a new band, named complex B, appeared in differentiated cells and showed a much faster migration than the complex A. This new binding is stronger than the complex A in differentiated cells, but weaker than the complex A seen in undifferentiated cells. To clarify the identity of proteins involved in this new complex, prior to probe addition, we preincubated a REST/NRSF antibody with the nuclear proteins of cells treated with RA for 4 days. As shown in Fig. 3(d), complex B was totally disrupted by the antibody, but not by an Sp1 antibody that had successfully super-shifted Sp1 protein in our previous studies (Bai and Kusiak 1995). These results suggested to us that REST/NRSF is involved in the formation of two different complexes depending on the state of the P19 cells.
It has been reported that the REST/NRSF gene has alternative splice variants with different C-terminal truncations that may retain a low DNA binding capability (Lee et al. 2000). Since the antibody used above is against the N-terminus of the protein, it is able to bind all variants, thus interfering with the binding or migration of the complexes in EMSA if these variants exist. To test the effects of neuronal differentiation on the expression level and pattern of the REST/NRSF protein, we performed immunoblot analysis of nuclear proteins using the same antibody. It is apparent that the expression level of nuclear REST/NRSF protein was largely reduced after differentiation for 4 days and there is no significant change at subsequent time points (Fig. 3e). As shown in Fig. 3(e), no new band appears in the fractionation range that covers all sizes of the possible REST/NRSF variants with approximately around 50-kDa (Palm et al. 1999). By analysis of the density associated with each band in the X-ray film image of immunoblot (Fig. 3e), we found that the level of the nuclear REST/NRSF protein in differentiated cells was less than 9% of that in undifferentiated cells. In contrast, as measured by a PhosphorImager, radioactivity associated with the complex A, in EMSA experiments, was reduced to less than 6% of that in undifferentiated cells. However, analysis of the newly formed complex B revealed that it has radioactivity about 12–26% of the complex A obtained from undifferentiated cells (Fig. 3c). Therefore, the reduction of the REST/NRSF protein level after differentiation is relatively more than that in DNA binding activity revealed in EMSA. In another immunoblot, the REST/NRSF in the total cell lysates showed the same change as that in nuclear protein, although all signals were weaker (data not shown). The weaker signal may be caused by the dilution of nuclear proteins by other cytosolic proteins. These results suggest a down-regulation of the REST/NRSF protein in the entire cell population after neuronal differentiation. In a converse strategy, we cotransfected NR1 promoter constructs with an REST/NRSF expression construct or the vector, pcDNA3, into neuronally differentiated P19 cells. Our results indicated that activity of the 5.4-kb promoter was attenuated by the REST/NRSF construct to 52% of that cotransfected with vector (Fig. 3f). As a control, the expressed REST/NRSF showed no effect on the promoter with the RE1 site deleted.
To support our observation that REST/NRSF is expressed in P19 cells, we further examined the presence of its mRNA in cells treated with RA for 4, 8 and 11 days. The RNA samples used were the same for detection of the NR1 mRNA shown in Fig. 1(c,d). Using RT-PCR, we observed that P19 cells express a high level of REST/NRSF mRNA and this expression was not significantly changed after neuronal differentiation (Fig. 3g). As a control, the RNA extracted from adult mouse brains showed a relatively weaker signal. It has been reported that four out of five REST/NRSF splice variants have sequence insertions or deletions between exon V and VI of the REST/NRSF gene. A neuronal exon resides in this region and produces 16- or 28-bp insertion to the wild-type mouse REST/NRSF mRNA (Palm et al. 1999). To clarify whether the fast migrating REST-containing band involves products of splicing variants, we designed PCR primers to encompass the boundary between exon V and VI, and therefore to be able to detect all 2 neuronal splice variants and 2 additional variants as previously shown for rat tissues (Palm et al. 1998). The PCR product was expected to be 105 bp without insertion or deletion. It will be 121 bp with 16-bp insertion or 133 bp with the 28-bp insertion of the neuronal exon (Palm et al. 1999). However, we did not observe any larger PCR product appearing in P19 cells before or after differentiation. These results agree with those obtained from immunoblot analysis that P19 cells did not express splice variants of REST/NRSF even after neuronal differentiation (Fig. 3e).
NR1 mRNA appears in the CNS of the rat fetus at embryonic day 14, when neurogenesis is underway and remains at a low level until birth (Watanabe et al. 1992; Akazawa et al. 1994; Monyer et al. 1994; Bayer et al. 1995). In the developing brain, during neonatal period, NR1 mRNA increases robustly. In our studies of differentiating P19 cells, we observed that NR1 mRNA increased 4 days after RA treatment. During this time P19 cells become neuroprogenitors and begin to express the neuronal marker nestin (MacPherson and McBurney 1995; Lin et al. 1996). NR1 mRNA was robustly up-regulated in P19 cells after RA treatment from 5 to 8 days, when neuronal differentiation begins and continues. These results are consistent with the developmental up-regulation of NR1 mRNA in rodents as well as with the demonstrated NR1 message levels in differentiated P19 cells (Ray and Gottlieb 1993; MacPherson et al. 1997; Okamoto et al. 1999). Although both transcription rate and mRNA stability may significantly contribute to the steady-state level of mRNA, few studies have been performed to elucidate the mechanism underlying the developmental increase in mRNA level of neuronal genes. Transcription activation may be a major mechanism of neuronal mRNA up-regulation in development as many transcription factors have been found to play important roles in the initiation and continuation of neurogenesis and neuronal differentiation. However, their effects on individual neuronal promoters remain mostly unknown. Additionally, studies of transgenic mice carrying promoter-reporter chimeras demonstrate, in some cases, such as the neuronal nicotinic acetylcholine receptor β2-subunit, an expression patter similar to the mRNA change in the developing brain (Bessis et al. 1997; Kallunki et al. 1998). In the present study we took advantage of the P19 cell model of neuronal differentiation and demonstrated that the NR1 gene transcription rate and promoter are activated by neuronal differentiation. Our findings strongly support the concept that transcription activation is a major mechanism underlying the up-regulation of the NR1 mRNA in differentiating neurons.
Recently, studies of neuronal gene expression revealed a negative regulatory mechanism by which the RE1 element interacts with the active suppressor REST/NRSF. Because REST/NRSF mRNA is expressed at a high level in the embryonic brain and becomes reduced in the developing brain during the neonatal period, it was proposed that regulation of this protein plays a key role in activation of the neuronal genes during the brain development (Chong et al. 1995; Schoenherr and Anderson 1995). However, there is a lack of evidence demonstrating a change of REST/NRSF DNA binding and protein level in the brain at the embryonic stage later than pretnatal 10 day (Chen et al. 1998), or in adult rodents brain although a change in mRNA levels was detected (Palm et al. 1998). In addition, each of the target genes of REST/NRSF has distinct expression patterns during the brain development. For example, SCG10 mRNA appears in the early embryonic brain, peaks at E19 and decreases gradually by P30 (Sugiura and Mori 1995) while the NR1 mRNA remains at a low level in embryonic brain and undergoes a robust increase in the brain during the neonatal period (Watanabe et al. 1992; Akazawa et al. 1994; Monyer et al. 1994). Constitutive expression of REST/NRSF or overexpression of a dominant-negative REST/NRSF protein in the developing CNS only resulted in expression changes of a limited number of neuronal genes, not including the NR1 gene (Chen et al. 1998; Paquette et al. 2000). Studies of transgenic animals revealed that the neuronal L1 promoter lacking the RE1 element produced a mixed enhancement and repression of reporter gene expression in different regions of the postnatal brain (Kallunki et al. 1998). Similar results were reported for the neuronal nicotinic acetylcholine receptor β2-subunit promoter in the embryonic brain at E13.5 (Bessis et al. 1997). These studies suggest a complex role of this negative regulator in controlling developmental activation of neuronal genes. The NR1 gene contains an RE1 site downstream of multiple transcription start sites and also undergoes a significant up-regulation during the brain development. The role of the RE1 site in the regulation of the NR1 gene during neuronal differentiation is an open and interesting question. We found that mutation of this site dramatically increased promoter activity in undifferentiated P19 cells. P19 cells retain many features of embryonic stem cells (MacPherson and McBurney 1995) that express a high level of REST/NRSF mRNA (Chong et al. 1995; Schoenherr and Anderson 1995). In the present study, we demonstrated that P19 cells not only expressed high levels of REST/NRSF mRNA, but also REST/NRSF protein in the nucleus, and that this protein bound the NR1 RE1 site. Most importantly, we observed that forced expression of a dominant-negative REST/NRSF interfered with NR1 silencing, and induced NR1 promoter activity as well as NR1 mRNA in undifferentiated P19 cells (Fig. 2e). Consistent with our findings, Lietz et al. (Lietz et al. 2003) reported that P19 cells contain functional REST/NRSF capable of binding to the newly identified RE1 element in the first intron of the human synaptophysin gene. Therefore, it is reasonable that the NR1 promoter in undifferentiated P19 cells is also suppressed by REST/NRSF via interaction with the RE1 element.
Our studies of the 5.4-kb NR1 promoter in stable transfectants of P19 cells treated with RA provide a complete profile of the promoter activity during neuronal differentiation. We observed a coordinated pattern of NR1 promoter activation and mRNA up-regulation in the cells. We also noted that nuclear REST/NRSF protein and DNA binding activity were down-regulated 2 days ahead of significant changes in the NR1 promoter activity and mRNA level (Figs 1–3). This timing lag may be the result of several possible mechanisms. The first one may be that there is a requirement for a positive mechanism of promoter activation during neuronal differentiation and that the cognate transcription activator(s) are not sufficiently expressed during this lag period. For example, it has been demonstrated that the interaction of the NR1 promoter with the general transcription factor, Sp1 and MASH1, is required in differentiating P19 cells (Okamoto et al. 2002). The contribution of other upstream cis-elements on the activation of the NR1 promoter is currently under investigation in our laboratory. These cis-elements include, but are not limited to, NFkappaB and AP1 (Bai et al. unpublished data; Zhuang et al. 2002). The second possible mechanism is that the REST/NRSF becomes an activator after neuronal differentiation through interaction with the RE1 element or with other cis-element(s) on the promoter. This possibility is supported by results obtained from studies of other neuronal genes: In transgenic animals, the RE1 element in the neuronal nicotinic acetylcholine receptor β2-subunit promoter or the neuronal L1 promoter drives an increase in reporter gene transcription in mature neurons of selective brain regions (Bessis et al. 1997; Kallunki et al. 1998). During Xenopus development, suppressing REST/NRSF expression with an antisense technique or interfering with REST/NRSF function with a dominant-negative mutant suppressed the expression of the neuronal sodium channel NaV1.2 instead of inducing its expression (Armisen et al. 2002). In rat NG108 neuronal cells, deletion of the RE1 element from the promoter of the neuronal corticotrophin-releasing hormone gene resulted in an up-regulation by the over-expressed REST/NRSF protein, suggesting that REST/NRSF may act positively on an unknown DNA cis-element in this promoter (Seth and Majzoub 2001). In our studies, an NR1 promoter lacking the RE1 element underwent a relatively mild increase in activity following neuronal differentiation, during which the nuclear REST/NRSF protein level was reduced, but not completely abolished. What is the role of the REST/NRSF protein in differentiated neurons and whether the RE1 cis-acting element may have a positive effect on the NR1 promoter are interesting questions currently under further investigation in our laboratory.
In our studies, we observed that the NR1 RE1 element bound nuclear proteins and formed a large complex (band A, Fig. 3a). While this large complex was reduced in neuronally differentiated cells, a new and faster migrating complex appeared and this complex contains REST proved by antibody interference. Although EMSA is performed in a non-denaturing gel by which the size of a given protein cannot be determined, in general, the migration speed still indicates the size and complexity of the binding complex. Therefore, the fast migration of this new complex may be caused by either a smaller REST/NRSF protein or less complicated composition. Multiple splice isoforms of the REST/NRSF gene have been found in human, rat and mouse tissues (Palm et al. 1998; Palm et al. 1999). Two neuronal isoforms were found to carry small insertions that cause translational frameshifts and produce c-terminally truncated proteins (Palm et al. 1999). These truncated proteins named REST4 and REST5 are able to interfere with the function of the full-length protein via a competition mechanism (Shimojo et al. 1999). However, in differentiated P19 cells, we did not detect truncated proteins that have molecular weight approximately of 46 kDa revealed by in vitro translation experiments (Palm et al. 1999), nor their mRNA (Fig. 3e,g). On the basis of our data, we exclude the possibility that products of splice isoforms interact with the NR1 promoter and up-regulate the NR1 promoter by counteracting REST/NRSF in differentiated P19 cells. In view of a stable steady-state level of REST/NRSF mRNA versus a significant down-regulation of the REST/NRSF protein in neuronally differentiated P19 cells, one may wonder whether there is an increased degradation rate of the REST/NRSF protein and the small DNA-protein complex is the result of degraded REST/NRSF protein. Though we did not measure the turnover rate of the REST/NRSF protein, the possibility that the complex B is the product of degraded protein may be excluded. This concept is supported by the following observations: First, the REST/NRSF antibody shifted complex A and B in EMSA experiments, suggesting the REST/NRSF proteins in both complexes at least share the N-terminus that is the binding epitope of this antibody. However, the same antibody detected only a full-length REST/NRSF protein and not any other smaller product in an immunoblot analysis. As discussed above, differentiated P19 cells still express full-length REST/NRSF protein that may bind the NR1 promoter. Altered complexity may then be the cause of the new, fast migrating complex. In support of our explanation, the binding of engineered REST/NRSF always shows the fast migrating pattern (for example, see Fig. 5, Thiel et al. 1998). Recently, it has been found that REST/NRSF must recruit multiple cofactors to the promoter in order to regulate transcription. A working model proposed that this recruitment takes place via protein–protein interaction of both N- and C-termini of the REST/NRSF with protein cofactors (Roopra et al. 2001). For example, the REST/NRSF N-terminus binds mSin3A and its C-terminus binds Co-REST. Studies of chromatin remodeling in fibroblasts with trichostatin A (TSA) induced NR1 mRNA, suggesting that protein acetylation is required for the action of REST/NRSF on the NR1 promoter (Naruse et al. 1999). Interestingly, a recent report demonstrated that the recruitment of cofactors by REST/NRSF is promoter specific. For example, the SCG10 promoter is sensitive to histone deacetylation and can be induced by TSA treatment while the Na channel type II promoter is insensitive to TSA treatment and requires CoREST and methylation of CpG islands (Lunyak et al. 2002). The NR1 promoter is of GC-rich and can be induced by TSA treatment in fibroblastoma or P19 cells (Bai et al. unpublished data). Based on these observations, whether both CoREST and histone deacetylation complexes are involved in the REST/NRSF related regulation of the NR1 promoter is an important question that requires further study.
In the present study, we have explored the mechanism that controls the developmental up-regulation of the NR1 gene using P19 cells as a model of neuronal differentiation. In view of the fact that all previous studies of the gene expression of glutamate receptors are limited to either mRNA level or promoter activity, our studies generate the first evidence that an increase in the transcription rate is the mechanism underlying NR1 gene activation during neuronal differentiation. Our studies further revealed that during neuronal differentiation the NR1 promoter became activated and up-regulated so as to be a major driving force of the increase in transcription rate. We then demonstrated that an RE1 consensus sequence located in this promoter plays an important role in the suppression of NR1 transcription in undifferentiated cells or neuroprogenitors, while this suppression was reversed in differentiating cells as well as by RE1 mutation or by a dominant-negative REST/NRSF in undifferentiated P19 cells. We also observed that after neuronal differentiation the expression and binding of REST/NRSF to the NR1 RE1 element were significantly down-regulated. Interestingly, we note that the binding pattern is also changed after neuronal differentiation, and its possible mechanism and significance are under further investigation in our laboratory.
We thank Dr D. J. Anderson (California Institute of Technology) for kindly providing the antibody against NRSE/REST and construct expressing mouse dominant-negative NRSF, Dr G. Mandel (State University of New York) for construct expressing human REST/NRSF protein, Dr D. Rowitch (Dana-Farber Cancer Institute/Harvard Medical School) for GAPDH cDNA and Dr M. Mishina (University of Tokyo) for mouse NR1 cDNA. GB is supported by NIH grant NS38077.