INTRODUCTION

The development and maintenance of a functional neuromuscular junction require that expression of all the molecular components be temporally and spatially regulated at the nerve-muscle contact. For example, the acetylcholine receptor (AChR)¹ is highly enriched at the crests of the postjunctional folds at a concentration 3 orders of magnitude higher than in the extrasynaptic membrane (1, 2). There are at least three mechanisms that contribute to control of AChR subunit gene expression. Myogenic factors, including MyoD, myogenin, myf5, and MRF4, appear to play an important role in the regulation of AChR gene expression during the myogenic differentiation program by binding directly to the E-box element located in sequences upstream of the transcription initiation site of AChR subunit genes (3-6). This regulation is critical for muscle-specific expression of AChR subunit genes. Second, electrical activity of muscle fibers, resulting from motor nerve firing, appears to down-regulate transcription of AChR subunit genes in extrasynaptic nuclei (7, 8). The molecular mechanism underlying transcriptional regulation by electrical activity is unclear. It has been suggested that the reduced rates of transcription may result from suppression of the transcriptional activation mediated by members of the MyoD family (9). Third, elevated transcription in those nuclei localized to the synaptic region of the muscle contributes to the maintenance of a high level of AChR mRNAs at the neuromuscular junction (9, 10). Such synapse-specific transcription is believed to be mediated by ARIA, the trophic factor utilized by motoneurons to stimulate AChR synthesis (11-13). ARIA binds to and activates its receptor, one or two members of the ErbB family of protein tyrosine kinases (14-17). We (17) and Tansey et al. (18) recently found that ARIA-induced AChR gene expression requires activation of the MAP kinase signal pathway.

To characterize further the molecular mechanisms of synapse-specific transcription, we have analyzed the promoter of the AChR  subunit. The mRNA coding for the  subunit is restricted to junctional areas from the outset of its expression (19). The synaptic localization of  mRNAs is not affected by denervation, which produces a reappearance of other AChR subunit mRNAs in extrajunctional areas (19, 20). Therefore a study of the  subunit promoter activity in response to ARIA would provide a better understanding how expression of  subunit is regulated during development. In addition, the  subunit promoter is an excellent model for the study of synapse-specific transcription. In this report, we provide evidence that a 10-nucleotide element in the sequence upstream of the transcription initiation site, termed ARE, is required for ARIA-induced  subunit expression in muscle cells. We also demonstrate that ARE binds specifically to a nuclear protein(s) in a manner dependent on ARIA stimulation of muscle cells and phosphorylation of nuclear proteins.

EXPERIMENTAL PROCEDURES

The recombinant ARIA (rHRG177-244), a peptide of HRG₁ residues 177-244) was generously provided by Dr. Mark Sliwkowski (21). Calf intestinal phosphatase was purchased from Boehringer Mannheim. All other chemicals, including okadaic acid, were purchased from Sigma.

Deletion mutants of the  subunit promoter were generated by polymerase chain reaction (PCR) using the p3500-nlacZ plasmid as a template. The upstream primers with a HindIII site were 5-GGG AGA AGC TTC TGA ATC TCA CTC TCA GC (Primer 1) for 416-Luc, 5-GGG AGA AGC TTC TCT CCT GAG ATG ACA GG for 170-Luc, 5-GGG AGA AGC TTG GGG CAG CTG CCT CCC CC for 78-Luc, and 5-GGG AGA AGC TTG GCA GAG GAT for 45-Luc. The downstream primer (Primer 2) was 5-GGG AGA AGC TTG AGG GAA CAG G with a HindIII site. Point and internal deletion mutants of the  subunit promoter were generated by the method of PCR overlap extension (22). The p3500-nlacZ plasmid was used as a template in the first PCR reaction using Vent DNA polymerase (New England Biolabs). The sense oligonucleotides of the internal overlapping primers (with mutated bases underlined) were 5-GGG ACA GGG GCA GCT G (position 89/68) for the ets mutant, 5-ATG GGG CAG CTG CCT CA CCC CCA CAG CAG GGG C (position 79/43) for the AP-2 mutant, 5-CTG CCT CCC CCA CCC CAG CGG CAG AGG ATT AGG TGA (position 70/28) for the ARE mutant, 5-CTG CCT CCC CCA CCC GGC AGA GGA TTA GGT GA (position 70/28) for the ARE deletion mutant, and 5-GGT GAC AGT CCC CAA ACC TAG CCT AAC ACC CTC CTC CCC TTC ACA (position 33/+18) for the N-box mutant. The two overlapping DNA fragments from the first PCR reaction were used as templates in the second PCR reaction with Primers 1 and 2. After digestion with HindIII, the PCR products were subcloned in pGL2-Basic (Promega) upstream of the luciferase gene. The authenticity and orientation of the synthetic promoter fragments were verified by DNA sequencing.

Cell Culture, Transfection Procedures, and Luciferase and -Galactosidase Assays

C2C12 cells were cultured as described previously (17). C2C12 myoblasts at approximately 50% confluence were co-transfected with an  subunit promoter-luciferase transgene (20 µg of DNA) and a control plasmid (1 µg of DNA) (pCMV, which encodes -galactosidase), by the standard calcium phosphate technique (23). Luciferase assay was performed using a kit from Promega following the manufacturer's instruction. -Galactosidase activity was determined as described previously (17). Luciferase activity of transgenes was normalized to -galactosidase activity to correct for variations in transfection efficiency.

C2C12 cells were washed in phosphate-buffered saline and homogenized in a buffer containing 10 mM HEPES, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 0.2 mM phenylmethylsulfonyl fluoride, and 0.5 mM dithiothreitol. Cell nuclei were pelleted by centrifugation at 4,000 rpm for 15 min at 4 °C. The pellets were extracted in DNA binding buffer containing 12% glycerol, 12 mM HEPES-NaOH, pH 7.9, 4 mM Tris/HCl, pH 7.9, 60 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, 0.1% Nonidet P-40, 0.2 mM phenylmethylsulfonyl fluoride, 1 µg/ml pepstatin, 1 µg/ml leupeptin, and 2 µg/ml aprotinin. Nuclear extract protein (10 µg) was incubated with 2 µg of nonspecific competitor poly(dI·dC), 4.5 µg of bovine serum albumin, and 1 ng of ³²P-labeled double-stranded oligonucleotide probe (~10,000 cpm) in a final volume of 15 µl of DNA binding buffer. After incubation at 30 °C for 15 min, the reaction was stopped by the addition of 5 µl of loading buffer and run onto a 4% acrylamide gel (24). The radiolabeled probes were 5-CCC CAC AGC AGG GG for ARE or 5-CTA GCC CGG AAC TAA for N-box (25). In some experiments, unlabeled double-stranded oligonucleotides were used as a competitor, including the ARE or N-box wild type, ARE mutant (5-CCA AGC AGC GTT GG), or SRE wild type (5-CCC CAT ATT AGG GG).

RESULTS AND DISCUSSION

Localization of the ARIA-responsive Promoter Activity in 34 Nucleotides between 78 and 45 in the 5-Flanking Region of the  Subunit Gene

To determine the minimum length of the 5 regulatory region required for promoter activity in response to ARIA, we generated transgenes containing a series of deletion mutants in the 5-flanking region of the  subunit gene. Transgenes were transfected in C2C12 muscle cells, and promoter activity was characterized by luciferase assay in ARIA-stimulated C2C12 myotubes. Expression of the 416-Luc transgene, which contains 416 nucleotides upstream from the transcription initiation site, was increased by ARIA stimulation in a concentration-dependent manner (Fig. 1). The maximal response (around 2-fold) was achieved with 1 nM ARIA, which agrees well the our previous observation using p3500nlacZ in which expression of -galactosidase is driven by the 3.5-kilobase 5-flanking sequence of the  subunit gene (17). These results suggest that the promoter element(s) required for the ARIA response is contained within 416 nucleotides upstream of the transcription initiation site. Therefore, the 416-Luc transgene was considered as equivalent to wild type for the purposes of this study. The ARIA response was specific to the  subunit promoter, since the promoter of rat skeletal muscle voltage-sensitive sodium channel subtype 2 (skm-2) did not respond to ARIA stimulation (Fig. 1C). Promoter activities of the transgene deletion constructs 170-Luc, 78-Luc, and 45-Luc in myotubes in response to 1 nM ARIA are shown in Fig. 1C. Seventy-eight nucleotides upstream from the transcription initiation site of the  subunit gene was sufficient to confer an ARIA response. This is in agreement with previous observations that 150 and 83 nucleotides upstream of the transcription initiation site of the  subunit gene are sufficient to respond to ARIA (15) and to confer preferential synaptic expression (25), respectively. The transgene containing 45 nucleotides, however, failed to respond to ARIA, even at concentrations up to 10-fold higher than required for a maximal response with the wild type transgene (Fig. 1D). The difference in the promoter activities between 78-Luc and 45-Luc indicated that the element required for ARIA response could be localized in the 34 nucleotides between 78 and 45.

Requirement of 34 nucleotides (position 78/45) in the  subunit promoter for ARIA-stimulated  subunit gene expression.

Identification of a 10-Nucleotide Element (ARE) Required for ARIA-induced  Subunit Expression

We previously demonstrated that ARIA-stimulated AChR gene expression requires MAP kinase activation (17). Major transcription factors mediating MAP kinase action in mammalian cells include ternary complex factor proteins (for the ets element) (26) and SRF (for the SRE element) (27). Anticipating that the ARE-responsive element might be localized between 78 and 45, we identified and selectively mutated a number of sites in this region, concentrating on elements that could potentially bind transcription factors regulated by the MAP kinase pathway. Among the sites mutated, a ten-nucleotide element (CCA CAG CAG G, position 55/46) was included. This element was chosen because it resembles SRE, having two cytidines on the 5 end and two guanosines on the 3 end (27). However, it differs from SRE in that the intervening 6 nucleotides have two cytidines and one guanosine instead of purely adenosines or thymidines as in the consensus SRE. Replacement of adenosines or thymidines with cytidines or guanosines in the internal 6 nucleotides dramatically decreases or diminishes SRE binding activity (28). In addition, the putative ets site (CAG GAT), which we identified at 83/78, the putative AP-2 element (position 64/54), and the N-box element (11/6), which was recently found to be required for synapse-specific expression (28), were also mutated. All mutants were made in the 416-Luc construct. While ets and AP-2 mutants diminished basal expression, they had no effect on ARIA-stimulated expression. By contrast, mutation of the 10-nucleotide SRE-like element attenuated greatly or abolished the  promoter activity in response to ARIA (Fig. 2). Myotubes possessing this mutant transgene failed to respond to concentrations of ARIA sufficient to activate the wild type transgene (Fig. 3A). However, an increase in mutant transgene expression was observed at 10 nM ARIA. These data strongly suggested that this 10-nucleotide element may be required for the ARIA-induced  subunit gene expression. We named this element ARE for RIA-esponsive lement. To confirm that ARE is required for the ARIA response, we tested the promoter activity of an ARE deletion mutant (ARE.). As shown in Fig. 3B, the ARE. mutant promoter failed to confer ARIA response even at 10 nM of ARIA. While this work was in progress, Duclert et al. (25) reported that the N-box is crucial for synaptic expression of AChR  subunit gene. Using an in vivo expression assay in intact muscles, the authors observed that 5 deletion mutants (83 and 75) of  subunit promoter provided approximately the same level of expression, whereas 63 and 52 5 deletion mutants gave a decreased expression and the 36 nucleotide construct did not express. This observation prompted that authors to suggest "the presence of an activating element" in this region. In addition, site-directed mutation in this region diminished synaptic expression of the  subunit gene (25). We believe that ARE may be the activating element in this region crucial for synaptic expression.

Requirement of ARE for ARIA-stimulated  subunit gene expression.

ARE.

Electrophoretic mobility shift analyses were performed to determine if ARE was able to interact with nuclear proteins in ARIA-stimulated myotubes. Using [-³²P]ATP-labeled double-stranded oligonucleotides containing the ARE element as a probe, we detected a prominent complex on the autoradiogram of a 4% acrylamide gel (Fig. 4A). The interaction between the ARE probe and nuclear binding protein was specific. First, formation of the complex was competitively inhibited by a 10- or 100-fold excess of unlabeled double-stranded wild type ARE oligonucleotides. In contrast, ARE mutant oligonucleotides required a much higher concentration to disrupt the complex formation. Second, 100-fold excess of unlabeled double-stranded oligonucleotides containing a typical SRE element had little, if any, inhibitory effect on the ARE nuclear protein binding complex (Fig. 4A), suggesting that ARE interacted with a protein different from those that interact with SRE. Interestingly, formation of the specific complex was enhanced by ARIA stimulation of C2C12 myotubes. Treatment of myotubes with ARIA increased the binding activity by at least 2-fold (Fig. 4B). These results suggested that ARE interacted with a nuclear protein(s).

Knowing that activation of MAP kinase is required for ARIA-induced AChR gene expression (17, 18), we attempted to determine whether the ARE nuclear binding activity was regulated by MAP kinase activity. Myotubes were treated with PD98059, an inhibitor of MAP kinase kinase, which is able to disrupt MAP kinase activation and ARIA-induced AChR gene expression in C2C12 cells (17). The ARE binding activity was decreased in PD98059-treated myotube nuclear extract (Fig. 4B). Incubation of myotube nuclear extract with calf intestinal phosphatase to dephosphorylate nuclear proteins attenuated the ARE binding activity in a concentration-dependent manner (Fig. 4C). There appeared to be a high level of endogenous phosphatase activity in myotube nuclear extract. Preincubation of myotube nuclear extract at 30 °C for 15 min alone, without the addition of exogenous phosphatase, was able to decrease ARE binding activity (Fig. 4C). Moreover, the preincubation-induced decrease in ARE nuclear binding activity was blocked by 2 nM okadaic acid, a specific inhibitor of serine/threonine phosphatases (Fig. 4C). Thus the ARE binding activity may derive from a phosphoprotein that has an enhanced DNA binding activity in the phosphorylated state. The phosphatase activity in C2C12 myotubes may be due to protein phosphatase 2A, a phosphatase which is specifically inhibited by okadaic acid at a nanomolar concentration (29).

N-box Is Not Required for ARIA-stimulated  Subunit Gene Expression in C2C12 Cells

Recently, a cis-element of 6 nucleotides (11/6) in the  subunit promoter was identified and implicated in compartmentalized expression of AChR genes at the neuromuscular junction (25). We found that although mutating the N-box dramatically decreased basal expression of the transgene, the N-box mutant promoter was still able to respond to ARIA stimulation (Fig. 2C). Using the N-box probe, we detected a binding activity in the C2C12 myotube nuclear extract (Fig. 5A), confirming the previous observation (25). This binding activity was specific, since it was inhibited by excess unlabeled double-stranded N-box oligonucleotides (Fig. 5A). Stimulation of C2C12 cells with ARIA, or with ARIA in the presence of PD98059, did not appear to affect the N-box binding activity (Fig. 5B). These results suggest that N-box may not be crucial for ARIA-induced  subunit gene expression in C2C12 cells.

In summary, we have identified ARE, a 10-nucleotide element (CCA CAG CAG G, position 55/46) in the  subunit promoter, which was required for ARIA-induced expression of  subunit gene in C2C12 muscle cells. ARE was able to interact with a nuclear protein(s). Furthermore, interaction between ARE and the nuclear protein was enhanced by ARIA stimulation of muscle cells and seemed to depend on phosphorylation of nuclear proteins. Experiments to elucidate the identity of the ARE binding protein are currently in progress.