Regulation of Na+/K+-ATPase by Nuclear Respiratory Factor 1

Background: NRF-1 regulates mediators of neuronal activity and energy generation. Results: NRF-1 transcriptionally regulates Na+/K+-ATPase subunits α1 and β1. Conclusion: NRF-1 functionally regulates mediators of energy consumption in neurons. Significance: NRF-1 mediates the tight coupling of neuronal activity, energy generation, and energy consumption at the molecular level. Energy generation and energy consumption are tightly coupled to neuronal activity at the cellular level. Na+/K+-ATPase, a major energy-consuming enzyme, is well expressed in neurons rich in cytochrome c oxidase, an important enzyme of the energy-generating machinery, and glutamatergic receptors that are mediators of neuronal activity. The present study sought to test our hypothesis that the coupling extends to the molecular level, whereby Na+/K+-ATPase subunits are regulated by the same transcription factor, nuclear respiratory factor 1 (NRF-1), found recently by our laboratory to regulate all cytochrome c oxidase subunit genes and some NMDA and AMPA receptor subunit genes. By means of multiple approaches, including in silico analysis, electrophoretic mobility shift and supershift assays, in vivo chromatin immunoprecipitation, promoter mutational analysis, and real-time quantitative PCR, NRF-1 was found to functionally bind to the promoters of Atp1a1 and Atp1b1 genes but not of the Atp1a3 gene in neurons. The transcripts of Atp1a1 and Atp1b1 subunit genes were up-regulated by KCl and down-regulated by tetrodotoxin. Atp1b1 is positively regulated by NRF-1, and silencing of NRF-1 with small interference RNA blocked the up-regulation of Atp1b1 induced by KCl, whereas overexpression of NRF-1 rescued these transcripts from being suppressed by tetrodotoxin. On the other hand, Atp1a1 is negatively regulated by NRF-1. The binding sites of NRF-1 on Atp1a1 and Atp1b1 are conserved among mice, rats, and humans. Thus, NRF-1 regulates key Na+/K+-ATPase subunits and plays an important role in mediating the tight coupling between energy consumption, energy generation, and neuronal activity at the molecular level.

The brain is one of the highest energy-demanding organs of the body (1,2), and much of the energy is used for repolarizing neuronal membrane potential after depolarization to re-establish the transmembrane ionic gradient for reactivation (3)(4)(5). This ion pumping activity is accomplished mainly by the Na ϩ / K ϩ -ATPase, which extrudes 3 Na ϩ and lets in 2 K ϩ for every ATP molecule hydrolyzed (6). This enzyme consumes the bulk of energy in the brain (7)(8)(9). Failure of Na ϩ /K ϩ -ATPase activity is implicated in the pathogenesis of several neurodegenerative disorders (10). The holoenzyme is composed of an ␣and a ␤-subunit. There are four types of ␣-subunits and three types of ␤-subunits that show tissue and/or cell specificity. Most neurons of the central nervous system express ␣1 (Atp1a1), ␣3 (Atp1a3), and ␤1 (Atp1b1) subunits, whereas glial cells express ␣2-, ␤2-, and ␤3-subunits (11)(12)(13)(14).
Whereas the demand for energy comes largely from Na ϩ / K ϩ -ATPase activity, the supply of energy in neurons comes almost entirely from oxidative phosphorylation in the mitochondria (4,5). Our earlier studies have shown that in the mammalian visual cortex and retina, cytochrome c oxidase (COX), 2 the terminal enzyme of the electron transport chain and a metabolic marker of neuronal activity (4), and Na ϩ /K ϩ -ATPase co-localize in the same regions that receive strong excitatory synaptic input, and both are down-regulated by impulse blockade in deprived visual cortical neurons (15)(16)(17). Hence, an important enzyme of the energy-generating machinery and a major energy-consuming enzyme of neurons are tightly coupled to neuronal activity (4,5). Recently, we found that the tight coupling between neuronal activity and energy metabolism extends to the molecular level in that the same transcription factor, nuclear respiratory factor 1 (NRF-1), regulates both energy metabolism (18,19) and synaptic transmission (20,21). NRF-1 itself is regulated by neuronal activity, and sustained activity is required for NRF-1 expression in cultured neurons and in vivo (22,23). Because the energy consumption by Na ϩ / K ϩ -ATPase is closely linked to energy production, and both are necessary to sustain neuronal activity, we hypothesize that these three processes are also tightly coupled at the molecular level.
The goal of the present study was to test our hypothesis that NRF-1 mediates the coupling of all three processes by regulating the expression of Na ϩ /K ϩ -ATPase subunits in neurons. Using multiple approaches, including in silico analysis, electrophoretic mobility shift (EMSA) and supershift assays, chromatin immunoprecipitation (ChIP), promoter mutational assays, RNA interference, and overexpression studies, we documented that NRF-1 has functional binding sites on the ␣1and ␤1-subunit of Na ϩ /K ϩ -ATPase in murine neurons. Furthermore, the binding sites are conserved among mice, rats, and humans.

EXPERIMENTAL PROCEDURES
Cell Culture-Murine Neuro-2a neuroblastoma (N2a) cells were obtained from the American Type Culture Collection (ATCC, CCL-131) and grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 50 units/ml penicillin, and 100 g/ml streptomycin (Invitrogen) at 37°C in a humidified atmosphere with 5% CO 2 .
In Silico Analysis of Promoters of Na ϩ /K ϩ -ATPase Subunit Genes-DNA sequences encompassing 1 kb upstream and 1 kb downstream of the transcription start point (TSP) of murine, rat, and human Na ϩ /K ϩ -ATPase subunit genes were obtained from the Genome Database in GenBank TM and aligned using MegAlign (DNAStar Lasergene) software. A putative NRF-1 core binding sequence with an invariant "GCA" core and flanking GC-rich regions (18) was searched using DNAStar Lasergene software. Regions of high homology and/or containing known NRF-1 binding sites were selected for experimental analyses.
Electrophoretic Mobility Shift Assays and Supershift Assays-EMSAs for NRF-1 interactions with putative binding elements on Na ϩ /K ϩ -ATPase subunit promoters were carried out as described previously (21,24) with minor modifications. Briefly, oligonucleotide probes with putative NRF-1 binding sites (Table 1; based on in silico analysis) were synthesized, annealed, and labeled by a Klenow fragment fill-in reaction with [␣-32 P]dATP (50 Ci/200 ng). Each labeled probe was incubated with 2 g of calf thymus DNA and 5 g of HeLa nuclear extract (Promega, Madison, WI) and processed for EMSA. Supershift assays were also performed, and in each reaction, 1-1.5 g of NRF-1-specific polyclonal antibody (Abcam, Cambridge, MA) was added to the probe/nuclear extract mixture and incubated for 20 min at room temperature. For competi-tion, a 100-fold excess of unlabeled oligonucleotide was incubated with nuclear extract before adding labeled or nonspecific oligonucleotide. Reaction mixtures were loaded onto 4% polyacrylamide gel and run at 200 V for 2.5 h in 0.25ϫ TBE buffer. Results were visualized by autoradiography. Rat cytochrome c with an NRF-1 binding site at position Ϫ172/Ϫ147 was designed as described previously (25) and used as a positive control. NRF-1 mutants with mutated sequences as shown in Table 1 were used as negative controls.
ChIP Assays-ChIP assays were performed similar to those described previously (26). Briefly, ϳ750,000 N2a cells were used for each immunoprecipitation and were fixed with 1% paraformaldehyde in PBS for 10 min at room temperature. Formaldehyde was neutralized using glycine, and cells were scraped and resuspended in a swelling buffer (5 mM PIPES, pH 8.0, 85 mM KCl, 1% Nonidet P-40, and protease inhibitors) and homogenized 10 times in a small pestle Dounce tissue homogenizer (7 ml). Nuclei were then isolated by centrifugation before being subjected to sonication. The sonicated lysate was immunoprecipitated with either 1 g of NRF-1 polyclonal antibody (Abcam) or 2 g of anti-nerve growth factor receptor p75 polyclonal goat antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Semiquantitative PCR was performed using 1 ⁄ 20 of precipitated chromatin. Primers targeting promoter sequences near the TSP of Na ϩ /K ϩ -ATPase subunit genes were designed (Table 2) as described previously (24). The Tfb2m (transcription factor B2 of mitochondria) promoter was used as a positive control, and exon 5 of the ␤-actin gene was used as a negative control ( Table 2). PCRs were carried out with the DreamTaq DNA polymerase (Fermentas, Glen Burnie, MD) with the following cycling parameters: 30-s denaturation at 94°C, 30-s annealing, and 20-s extension at 72°C (32-36 cycles/reaction). Because the proximal promoters of Na ϩ /K ϩ -ATPase subunit genes are very GC-rich, reaction conditions for some amplicons were optimized by adding betaine (Sigma) and/or DMSO (Sigma). PCR products were visualized on 2% agarose gels stained with ethidium bromide.
Promoter Mutagenesis Studies-Luciferase reporter constructs of Na ϩ /K ϩ -ATPase subunit promoters were made by PCR cloning the proximal promoter sequences using genomic DNA prepared from mouse N2a cells as template, digesting with KpnI and HindIII, and ligating the product directionally into pGL3 basic vector (Promega). Sequences of primers used for PCR cloning and mutagenesis primers are provided in Tables 3 and 4. Site-directed mutagenesis of putative NRF-1 binding site on each promoter were generated using specific primers. All constructs were verified by sequencing. Each promoter construct was transfected into N2a cells in a 24-well plate using Lipofectamine 2000. Each well received 0.6 g of reporter construct and 0.06 g of pRL-TK Renilla luciferase vector.
To further investigate the effect of KCl stimulation after mutating the NRF-1 binding site, transfected neurons were stimulated with 20 mM KCl in the culture media for 5 h as described previously (21). Cell lysates were then harvested and measured for luciferase activity using the Dual-Luciferase reporter system (Promega) according to the manufacturer's suggestions. For each sample, luciferase activity was normalized with Renilla luciferase activity. Data from six independent transfections were averaged for each promoter construct. p values were calculated from two-tailed t-tests, and p Յ 0.05 was considered significant.
NRF-1 Silencing Using Small Hairpin RNA (shRNA) Vectors-NRF-1 silencing was carried out using shRNAs against murine NRF-1 (GenBank TM accession number NM_010938) cloned into pLVTHM with H1 promoter and green fluorescent protein reporter (gift of Dr. P. Aebischer, Swiss Federal Institute of Technology). Four shRNA sequences were selected: 5Ј-GAAAGCTGCAAGCCTATCT-3Ј, 5Ј-GCCACAGGAG-GTTAATTCA-3Ј, 5Ј-GCATTACGGACCATAGTTA-3Ј, and 5Ј-AGAGCATGATCCTGGAAGA-3Ј. PLKO.1 control vector with the PAC gene (SHC003) was used to select the transfected cells. Empty vectors with scrambled shRNA served as negative controls. The basic gene cloning method was followed as described previously (18,27). N2a cells were plated in 6-well plates at a density of 5-8 ϫ 10 6 cells/well. They were transfected 1 day after plating with shRNA-containing plasmids (0.5 g of each sequence) or empty or scrambled vectors (1.5 g) using JetPrime reagent (Polyplus Transfection, New York, NY). Puromycin at 5 g/ml was added to the culture medium on the second day after transfection for the selection of transfected cells. Transfection efficiency for N2a cells ranged from 65 to 75%. However, puromycin selection effectively yielded almost 100% transfected cells. N2a cells transfected with shRNA were further stimulated with 20 mM KCl in the culture media for 5 h as described previously (22). Cells were harvested and processed for the isolation of RNA and protein.

NRF-1 Overexpression and Tetrodotoxin (TTX) Treatment-
The human NRF-1 cDNA clones were obtained from Open Biosystems (Lafayette, CO). The NRF-1 cDNA was cloned into pcDNA Dest40 vector using the Gateway multisite cloning kit (Invitrogen) and according to the manufacturer's instructions. N2a cells were plated in 6-well plates at a density of 5-8 ϫ 10 6 cells/well. They were transfected 3 days after plating with 2 g of NRF-1-containing plasmids using JetPrime reagent. Empty vectors were used as a control and at the same concentrations as vectors with the NRF-1 insert. Geneticin (500 g/ml) (Invitrogen) was added to the culture medium on the second day after transfection for the selection of transfected cells. Transfection efficiency for N2a cells ranged from 65 to 75%; however, Geneticin selection for 3 days yielded almost 100% transfected cells. N2a cells transfected with NRF-1 were treated with 0.4 M TTX in the culture media for 3 days. Cells were harvested and processed for the isolation of RNA and protein.
Real-time Quantitative PCR-Total RNA was isolated using the RNeasy kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. cDNA was synthesized using the iScript cDNA synthesis kit (Bio-Rad) according to the manufacturer's instructions. Real-time quantitative PCRs were carried out in a Bio-Rad iCycler using IQ SYBR Green Supermix (Bio-Rad) following the manufacturer's instructions. Primer sequences are shown in Table 5. PCR runs were as follows: hot start 2 min at 95°C, denaturation 10 s at 95°C, annealing 15 s according to the T m of each primer, and extension 10 s at 72°C for 15-30 cycles. Melt curve analyses verified the formation of single desired PCR product. Mouse Actb (␤-actin) and Gapdh were the internal controls, and the 2 Ϫ⌬⌬ C T method (28) was used to calculate the relative amount of transcripts.
Statistical Analysis-Significance among group means was determined by analysis of variance. Significance between two groups was analyzed by Student's t test. p values of 0.05 or less were considered significant.

RESULTS
In Silico Promoter Analysis and Homology-The proximal promoters of murine Na ϩ /K ϩ -ATPase subunit isoforms were analyzed in silico with DNA sequence 1 kb 5Ј upstream and 1 kb 3Ј of TSP. Our earlier studies have shown that besides its originally reported binding sequence (25), NRF-1 also binds to sequences with an invariant GCA core flanked by GC-rich regions (18,20,21,29). Based on this, Atp1a1, Atp1a3, and Atp1b1 all had the GCA sequence with flanking GC-rich regions.
In Vitro Binding of NRF-1 to Na ϩ /K ϩ -ATPase Subunit Promoters-EMSA for NRF-1 interactions were carried out in vitro using 32 P-labeled oligonucleotide probes to determine the specificity of NRF-1 binding to murine Na ϩ /K ϩ -ATPase promoters (Fig. 1). Rat cytochrome c promoter with a known NRF-1 site at positions Ϫ172/Ϫ147 (25) served as a positive control, and it formed specific DNA/NRF-1 shift and supershift complexes (Fig. 1, lanes 1 and 3, respectively). When an excess of unlabeled probe was added as a competitor, no shift band was formed (Fig. 1, lane 2). Atp1a1 and Atp1b1 each formed a specific DNA-protein shift complex when incubated with HeLa nuclear extract (Fig. 1, lanes 4 and 10) and a DNA-proteinantibody supershift complex with the addition of NRF-1 antibodies (Fig. 1, lanes 6 and 12). Competition with excess unlabeled probe eliminated the shift complex (Fig. 1, lanes 5 and  11). Competition with an excess of unlabeled oligonucleotides with mutated NRF-1 site was not able to compete with the shift band but outcompeted nonspecific bands (Fig. 1, lanes 7 and  13). Labeled oligonucleotides with a mutated NRF-1 site on Atp1a1 and Atp1b1 served as negative controls and did not show a shift (Fig. 1, lanes 8 and 14) or a supershift complex (Fig.  1, lanes 9 and 15). Atp1a3 did not form a shift or supershift complex (Fig. 1, lanes 16 -18).
In Vivo Interaction of NRF-1 with Na ϩ /K ϩ -ATPase Subunit Promoters-To verify in vivo interactions of NRF-1 with the promoters of Na ϩ /K ϩ -ATPase subunits, ChIP assays were performed. PCR was carried out with primers encompassing putative NRF-1 binding sites on chromatin immunoprecipitated from N2a cells. A 0.5 and 0.1% dilution of input chromatin was used as a standard to indicate the efficiency of the PCRs. The promoter of the Tfb2m gene with a known NRF-1 binding site was used as a positive control, whereas ␤-actin exon 5 (Actb) was used as a negative control. The promoters of Tfb2m, Atp1a1, and Atp1b1 each produced a band from DNA immunoprecipitated with anti-NRF-1 antibodies at a position identical to that of the genomic DNA control (input) (Fig. 2). On the other hand, Atp1a3 and Actb (␤-actin) yielded no bands (Fig. 2).
Investigation of Mutated NRF-1 Binding Sites on Promoter Activity-Based on the results obtained with ChIP primers and EMSA probes that formed NRF-1-specific complexes, site-directed mutations of these putative NRF-1 sites on Atp1a1, Atp1b1, and Cox6b promoters were constructed. As shown in Fig. 3, mutations in the NRF-1 binding sequence in the Cox6b promoter led to a significant decrease in the luciferase activity (p Ͻ 0.001). A mutation in the NRF-1 site of the Atp1b1 promoter led to a significant decrease in its activity (p Ͻ 0.001), but mutating the NRF-1 site in the Atp1a1 promoter increased its activity (p Ͻ 0.05).
To investigate if the expression of Cox6b and Atp1b1 genes responded to depolarizing stimulation, cells were treated with 20 mM KCl for 5 h, a regimen previously found to activate NRF-1 and COX gene expression (22,23). As shown in Fig. 3, depolarizing stimulation resulted in a significant increase in the activity of Cox6b, Atp1a1, and Atp1b1 promoters (p Ͻ 0.001 for all). Cox6b and Atp1b1 promoters with a mutated NRF-1 binding site failed to respond to depolarization stimulation, whereas Atp1a1 promoter with a mutated NRF-1 site showed an increase in its activity (p Ͻ 0.01 when compared with the mutant).
NRF-1 Silencing by RNA Interference-To determine the effect of reduced NRF-1 on the expression of Na ϩ /K ϩ -ATPase subunits, vectors expressing shRNA against four target sequences of NRF-1 mRNA were used. These vectors were previously found to silence NRF-1 expression in N2a cells (18).
TTX is a voltage-dependent Na ϩ channel blocker and blocks impulse activity in neurons. At 0.4 M, TTX has been shown to decrease the levels of COX subunit mRNAs as well as COX enzyme activity in vivo and in primary neurons (15,16,23). To evaluate the effect of impulse blockade on Na ϩ /K ϩ -ATPase subunits, N2a cells were treated with TTX for 3 days. TTX reduced the expression of NRF-1, Cox2, Cox7c, Tfb2m, Atp1a1, Atp1a3, and Atp1b1 (Fig. 5). This indicates an overall suppressive effect of TTX on Na ϩ /K ϩ -ATPase subunit gene expression in neurons as it had on other genes. On the other hand, reductions in the expression of NRF-1, Cox2, Cox7c, Tfb2m, and   FIGURE 3. Site-directed mutational analysis of the promoter elements of wild type (wt) and those with mutated NRF-1 binding site (mut) for Cox6b, Atp1a1, and Atp1b1 genes. Cox6b is one of 10 nucleus-encoded COX subunits and served as a positive control. Mutating the NRF-1 site on Cox6b and Atp1b1 promoters resulted in significant decreases in luciferase activity as compared with the wild type. NRF-1 binding site mutation in the Atp1a1 gene resulted in a significant increase in luciferase activity. KCl depolarization significantly increased promoter activity in all wild types but not in Cox6b and Atp1b1 promoters with a mutated NRF-1 site. However, mutating the NRF-1 site increased the activity of the Atp1a1 promoter. n ϭ 6 for each construct. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; ϫ, non-significant. All mutants and wild type ϩ KCl are compared with the wild type. All mutants ϩ KCl are compared with mutant. Error bars, S.E.  Cox2, Cox6b, and TFB1M served as positive controls. Of the three Na ϩ /K ϩ -ATPase subunits reported in neurons, Atp1b1 showed a significant decrease in its mRNA with NRF-1 silencing as compared with those with scrambled vectors. On the other hand, Atp1a1 showed an increase in its expression with NRF-1 shRNA. KCl depolarization increased the mRNA levels of all genes, including Atp1a1, but it did not affect transcript levels of Atp1b1, Cox2, Cox7c, and Tfb2m in cells that were transfected with NRF-1 shRNA. n ϭ 10 for each data point; *, p Ͻ 0.05; ***, p Ͻ 0.001; ϫ, non-significant when compared with controls. All NRF-1 shRNA and scrambled vector ϩ KCl are compared with scrambled vector. All NRF-1 shRNAs ϩ KCl are compared with NRF-1 shRNA. Error bars, S.E.  NRF-1, Cox2, Cox7c, Atp1a1, Atp1a3, and Atp1b1 were all reduced by TTX as compared with controls. NRF-1 overexpression in N2A cells significantly increased mRNA levels for NRF-1, Cox2, Cox7c, Tfb2m, and Atp1b1. However, it reduced the mRNA levels of Atp1a1, whereas Atp1a3 remained unaffected. Overexpression of NRF-1 rescued Cox2, Cox6c, Tfb2m, and Atp1b1 from TTX-induced suppression, but it further reduced the expression of Atp1a1. n ϭ 10 for each group. Asterisks, p values were compared with empty vectors (*, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001). #, p values were compared with empty vector ϩ TTX (#, p Ͻ 0.05; ##, p Ͻ 0.01; ###, p Ͻ 0.001). ϫ, p values were compared with NRF-1 overexpression (ϫ, p Ͻ 0.05; ϫϫ, p Ͻ 0.01; ϫϫϫ, p Ͻ 0.001). Error bars, S.E.

DISCUSSION
Using multiple approaches, the present study demonstrates that Na ϩ /K ϩ -ATPase subunits ␣1 and ␤1 are functionally regulated by NRF-1. Depolarizing stimulation induced the expression of these subunits, whereas impulse blockade down-regulated their expression. These observations suggest that the expression of Na ϩ /K ϩ -ATPase ␣1 and ␤1 are tightly coupled to neuronal activity. The binding sites of NRF-1 on ␣1 and ␤1 are conserved among mice, rats, and humans, consistent with conserved NRF-1 binding sites on glutamatergic receptor subunits and COX subunit gene promoters (18,20,21). The binding motif on the Na ϩ /K ϩ -ATPase ␤1-subunit promoter differs slightly from that of the traditional consensus sequence for NRF-1 (YGCGCAYGCGCR). However, the GCA core remained invariant. Previously, we have reported similar atypical NRF-1 binding sites with an invariant GCA core on subunits of COX (18), NMDA receptor (21), AMPA receptor (20), and NOS1 genes (29).
NRF-1 is a key transcription factor involved in the biogenesis of mitochondria by regulating genes important in energy generation. NRF-1 controls the expression of some subunits of the five respiratory chain complexes in mitochondria (31,32). It also regulates the expression of many other target genes (reviewed in Refs. 32 and 33). Thus, it is not surprising that NRF-1 knock-out is embryonically lethal (34). NRF-1 binds and activates the promoters of mitochondrial transcriptional factor A (TFAM), transcription specificity factors (TFB1M and TFB2M), and RNA-processing proteins required for mitochondrial DNA transcription and replication (32,35,36). Recently, we found that NRF-1 regulates all 10 nucleus-encoded COX subunit genes in neurons (18). Thus, together with its regulation of the three mitochondrially encoded COX subunits, NRF-1 coordinates the expression of all 13 COX subunits from the two genomes. We also found that NRF-1 regulates the expression of several glutamatergic receptor subunit genes, including NMDA receptor subunits 1 and 2B and AMPA receptor subunit 2 (20,21). In addition, NRF-1 regulates KIF17, a member of the kinesin protein superfamily that transports its cargo NR2B from cell bodies to dendrites in neurons (37). Hence, NRF-1 is vital to neurons, and NRF-1 itself is regulated by neuronal activity both in culture and in vivo (22,23). The present study documents that NRF-1 also plays an important role in regulating energy consumption in neurons by controlling the expression of key subunits of the Na ϩ /K ϩ -ATPase.
A major subunit of Na ϩ /K ϩ -ATPase regulated by NRF-1 is ␤1. This is a smaller glycosylated subunit of the holoenzyme and assembles with the larger ␣-subunit in the endoplasmic reticulum. Such association is a primary requisite before the release of Na ϩ /K ϩ -ATPase from the endoplasmic reticulum and its insertion into the plasma membrane (for a review see Refs. 38 and 39). The association of the ␤-subunit facilitates the correct membrane integration and packing of the ␣-subunit, which is necessary for its protection against cellular degradation (40), for its acquisition of functional properties (41), and for its routing to the plasma membrane (42). In tissues, including the brain, the activity of the Na ϩ /K ϩ -ATPase enzyme correlates with the level of ␤1 mRNA content (43). The abundance of ␤1 mRNA is significantly lower than that of ␣ mRNA in N2a cells (data not shown) and in many tissues, including the brain (44 -46), suggesting that the synthesis of the ␤-subunit is tightly controlled and that ␤ plays an important role in controlling the abundance of the enzyme. Hence, the positive regulation of ␤1-subunit expression by NRF-1 conceivably affects the regulation of the entire holoenzyme in neurons. The ␤-subunit mRNA and protein levels increased with KCl-induced depolarization and decreased with TTX-induced impulse blockade, indicating that they correlate with neuronal activity. Silencing of NRF-1 prohibited ␤1 from being up-regulated by KCl, whereas overexpression of NRF-1 rescued these transcripts from being down-regulated by TTX. The NRF-1 binding site on ␤1 is located in the 5Ј-untranslated region, 450 bp downstream of TSP. This binding site has the conserved GCA core motif for NRF-1 with partially conserved GC-rich flanking regions. Thus, NRF-1 plays a critical role in regulating the activity of the enzyme.
The larger ␣-subunit of the ␣␤ functional enzyme determines its major physiological characteristics, such as binding to ligand, hydrolysis of ATP, and transport of Na ϩ and K ϩ ions (39,47). The four ␣-subunits (␣1-␣4) are distributed in a tissue-specific manner, but their tissue-specific functions and regulation are poorly understood. ␣1 and ␣3 are the major subunits in neurons (11)(12)(13)(14). The expression of ␣1-subunit was shown earlier to be tightly regulated based on tissue and/or type and differentiation status of the cell. In the brain, ␣1 increases severalfold during development, but it remains constant in the skeletal muscles, kidney, and other organs (48,49). The present study documents a functional NRF-1 binding site on ␣1 but not on ␣3 promoter. Overexpression of NRF-1 led to a strong down-regulation of ␣1, whereas the silencing of NRF-1 caused a mild up-regulation, indicating that it is negatively regulated by NRF-1. Previous studies have identified several positive regulatory elements on the ␣1 gene promoter (50). Sp1 (specificity factor 1) is a major transcription factor driving the tissue-and cell type-specific or development-specific regulation of ␣1 (50 -52). In contrast to this positive regulation, negative regulatory elements that are cell type-specific have also been identified in FIGURE 6. Aligned partial sequences of Atp1a1 and Atp1b1 promoters from mice (M), rats (R), and humans (H) showed conserved putative NRF-1 binding sites. Atp1a1 has typical and Atp1b1 has atypical NRF-1 binding sites. Solid boxes, NRF-1 binding sites; highlighted areas, conserved bases within the NRF-1 sites.
the ␣1 promoter and are conserved among vertebrates (53). However, these putative suppressors have not been thoroughly characterized, and, to our knowledge, no work has been done thus far on neurons. The NRF-1 binding site on the ␣1 promoter is located in the region upstream of the previously identified positive regulatory elements (50) and is located in the region of promoter that contains putative negative regulatory elements (53). It is also conserved among humans, mice, and rats. It remains to be determined if NRF-1 acts alone or in concert with other transcription factor(s) or co-factor(s) for the negative regulation of ␣1.
Negative regulation of ␣1 by NRF-1 may provide a functional explanation for the presence of the two ␣ isoforms in neurons that may be biologically advantageous to neurons. Based on their Na ϩ , K ϩ , and ATP affinity (␣1 has greater affinity for Na ϩ and lower affinity for K ϩ and ATP than ␣3), it has been suggested that ␣1-containing enzyme operates under normal physiological conditions, whereas ␣3-containing enzyme plays a major role in restoring the membrane potential in highly active neurons (38, 54 -56). The negative regulation of ␣1 by NRF-1 may be advantageous to neurons during repeated firing of action potentials, when the activity of ␣3-containing enzyme is in greater demand. Our earlier studies and the present study have shown that, when neuronal activity is stimulated, NRF-1 expression increases (22,23), which may help to suppress ␣1 expression in favor of ␣3. However, KCl-induced up-regulation of ␣1 suggests that the expression of ␣1 may be under the control of multiple transcription factors and upstream signaling mechanisms, and NRF-1 may or may not interact with these factors in such mechanisms.
Thus, the present study documents that NRF-1 regulates Na ϩ /K ϩ -ATPase, a major energy consumer in neurons. Specifically, it positively regulates an important regulatory subunit, ␤1, and negatively regulates ␣1. In addition to our previously reported tight coupling between neuronal activity and energygenerating machinery, this report extends this tight coupling to energy consumption along with neuronal activity and energy production. Thus, the same transcription factor, NRF-1, can effectively coordinate the expression and function of Na ϩ /K ϩ -ATPase, glutamatergic receptors, and energy-generating machinery, thereby coordinating a controlled and stable interplay between energy utilization, synaptic transmission, and energy generation (Fig. 7).
It is possible that other sets of transcription factors also participate in the regulation of this tight coupling. Transcription factors, such as NRF-2 and Sp1, as well as transcription coactivators, such as peroxisome proliferator-activated receptor-␥ coactivator 1 (PGC-1) (23,24,35), may be involved by interacting directly or indirectly with NRF-1. Research is under way to explore these possibilities.