Cloning and characterization of a novel transcriptional repressor of the nicotinic acetylcholine receptor delta-subunit gene.

We have identified a negative cis-acting regulatory element in the nicotinic acetylcholine receptor delta-subunit gene's promoter. This element resides within a previously identified 47-base pair activity-dependent enhancer. Proteins that bind this region of DNA were cloned from a lambdagt11 innervated muscle expression library. Two cDNAs (MY1 and MY1a) were isolated that encode members of the Y-box family of transcription factors. MY1/1a RNAs are expressed at relatively high levels in heart, skeletal muscle, testis, glia, and specific regions of the central nervous system. MY1/1a are nuclear proteins that bind specifically to the coding strand of the 47-base pair enhancer and suppress delta-promoter activity in a sequence-specific manner. These results suggest a novel mechanism of repression by MY1/1a, which may contribute to the low level expression of the delta-subunit gene in innervated muscle. Finally, the gene encoding MY1/1a, Yb2, maps to the mid-distal region of mouse chromosome 6.

We have identified a negative cis-acting regulatory element in the nicotinic acetylcholine receptor ␦-subunit gene's promoter. This element resides within a previously identified 47-base pair activity-dependent enhancer. Proteins that bind this region of DNA were cloned from a gt11 innervated muscle expression library. Two cDNAs (MY1 and MY1a) were isolated that encode members of the Y-box family of transcription factors. MY1/1a RNAs are expressed at relatively high levels in heart, skeletal muscle, testis, glia, and specific regions of the central nervous system. MY1/1a are nuclear proteins that bind specifically to the coding strand of the 47-base pair enhancer and suppress ␦-promoter activity in a sequence-specific manner. These results suggest a novel mechanism of repression by MY1/1a, which may contribute to the low level expression of the ␦-subunit gene in innervated muscle. Finally, the gene encoding MY1/1a, Yb2, maps to the mid-distal region of mouse chromosome 6.
Characterizing the molecular mechanisms that regulate the expression and distribution of synaptic proteins is central to understanding synapse formation and modification. As a model synapse, the neuromuscular junction provides an ideal system to study these intricate mechanisms. Genes encoding muscle nicotinic acetylcholine receptors (nAChR) 1 have been used as probes to delineate mechanisms by which the presynaptic motor neuron regulates postsynaptic muscle protein expression. These studies have shown that nerve-induced muscle electrical activity suppresses expression of the embryonic-type (␣␤␥␦) nAChR genes (reviewed by Hall and Sanes (1993)).
The mechanism mediating this activity-dependent control of gene expression is not well understood. Recent studies have identified E-box sequences (CANNTG) in the nAChR gene promoters that are necessary for activity-dependent regulation of gene expression (Tang et al., 1994;Bessereau et al., 1994;Su et al., 1995). These sequences are known to bind helix-loop-helix proteins of the MyoD family of transcription factors. Myogenin is a strong candidate for binding to these sequences and mediating developmental stage-specific and activity-dependent expression, since myogenin knockout mice do not express embryonic-type ␣and ␥-subunit genes (Hasty et al., 1993). In addition, overexpression of these factors results in increased nAChR promoter activity in nonmuscle cells (Gilmour et al., 1991;Prody and Merlie, 1991;Bessereau et al., 1993;Berberich et al., 1993). However, the ␦-subunit gene appears to be unique in that mice lacking myogenin still express this gene in embryonic muscle fibers (Hasty et al., 1993), suggesting that other E-box binding proteins may be involved in its regulation.
Since several other muscle-specific genes have E-box sequences but are not regulated by muscle activity (Chahine et al., 1992;Chahine et al., 1993;Merlie et al., 1994;Gilmour et al., 1995;Su et al., 1995), it is likely that activity-dependent control of gene expression involves additional cis-acting elements that may work in conjunction with the E-box sequences. Identification of these sequences will provide probes for the proteins that bind them.
Using in vivo expression assays, we have recently identified these additional sequences in the nAChR ␦-subunit gene promoter. 2 A 47-bp enhancer (nucleotides Ϫ1 to Ϫ47 in Chahine et al. (1992)) was identified that contained all the necessary elements to confer activity-dependent expression onto a heterologous promoter. In addition to an E-box, this sequence contains a region with similarity to the SV40 core enhancer, SVCE (Khoury and Gruss, 1983). Point mutations in either the E-box or SVCE sequence block the ability of the 47-bp enhancer to activate a heterologous promoter in response to muscle denervation.
When considering mechanisms mediating activity-dependent control of nAChR gene expression, it is important to consider both activation of these genes upon denervation and suppression of these genes in the innervated state. We report here on a mutation in the SVCE of the ␦-promoter that results in increased promoter activity in innervated muscle, suggesting that this sequence acts as a repressor and may participate in maintaining low levels of ␦-subunit gene expression following muscle innervation. Using this DNA sequence as a probe for proteins that may mediate this repression, we cloned and characterized cDNAs encoding muscle Y-box proteins, MY1 and MY1a, which suppress ␦-promoter activity. Although these proteins were identified by their ability to bind to the ␦-promoter's SVCE sequence, they appear to show a preference for binding the coding strand of the 47-bp activity-dependent enhancer. MY1/1a are members of the Y-box family of transcription factors . MY1a lacks 207 nucleotides * This work was supported by grants from the National Institutes of Health (to D. G. and M. B.), Muscular Dystrophy Association (to D. G.), and the Lucille P. Markey Charitable Trust (to D. G. and M. B.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence ( found in MY1, suggesting they are derived from the same gene by alternative splicing. MY1 RNAs are most abundant in skeletal muscle, heart, testis, glia, and specific central nervous system neurons.

MATERIALS AND METHODS
In Vivo Expression Assays-Direct DNA injection into innervated and denervated rat skeletal muscle was performed as described previously (Wolff et al., 1991). 2 Briefly, approximately 1-month-old rats were anesthetized with ether and the left lower hindlimb was denervated by removing a 3-mm section of sciatic nerve just below the hip. Four to five days post-denervation, DNA solutions were injected into innervated and denervated extensor digitorum longus muscles of anesthetized rats using a 100-l Hamilton syringe with a 0.5-inch-long, 27-gauge needle that was fitted with a Williams collar exposing 2-3 mm of the needle tip.
DNA, for injection into muscle, was purified twice on CsCl gradients. Prior to injection, 150 -200 g of CMV/CAT along with 150 -200 g of one of the ␦-550 expression constructs were mixed and ethanol-precipitated twice. The final precipitate was rinsed two times with 70% ethanol and dried briefly. This DNA was resuspended at 10 mg/ml in 150 mM NaCl, and approximately 10 l was injected slowly into each of four different locations along the length of the extensor digitorum longus muscle.
Expression Library Screening-A gt11 expression library, prepared from rat diaphragm muscle, was kindly provided by Dr. John Merlie (Washington University). Approximately 5 ϫ 10 5 plaques were screened using a radioactive, concatenated double-stranded oligomer corresponding to nucleotides Ϫ51 to Ϫ31 of the rat ␦-subunit gene's 5Ј flanking DNA (Chahine et al., 1992). Prior to screening, fusion proteins were induced by isopropyl 1-thio-␤-D-galactopyranoside and denatured/renatured according to the protocol of Vinson et al. (1988).
Oligonucleotides-The following oligonucleotides were used in these studies. Although both coding (ϩ) and non-coding (Ϫ) strands were synthesized, only the coding strand is shown: SVCE, CTCTTCTTTC-CAAACCCCTA; E-box, TAAGCCGCCAGCACCTGTCCC; oligonucleotide 3, TCTTTCCAAACCCCTAAGCCGCCAGCACCTGTCCCCTTGC-TTGCCTCA; start site, GATCTTTGCTTGCCTCATTCCACAGCCG. Oligonucleotides were purified by fractionation through a denaturing 12% polyacrylamide gel and eluted into 0.5 M sodium acetate, 1 mM EDTA, 0.1% SDS, pH 8.0. Single-stranded probes were end-labeled using polynucleotide kinase and purified over Sephadex G50 columns. Double-stranded probes were generated by mixing a particular radiolabeled single-stranded probe with a 5-fold molar excess of unlabeled complementary strand in hybridization buffer (10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 10 mM MgCl 2 , 1 mM EDTA), heated to 90°C for 2 min, and then slowly cooled to room temperature.
DNA Sequencing-Recombinant phage inserts were subcloned into BSSK(ϩ) vector (Stratagene). Unidirectional deletions for use in DNA sequencing were created using exonuclease III digestion of restriction endonuclease-treated DNA (Sambrook et al., 1989). Nucleotide sequence was obtained using the dideoxy-mediated termination method (Sanger et al., 1977). Both strands of the DNA were sequenced.
RNase Protection Assays-Total RNA was isolated from various tissues using the method of Chirgwin et al., (1979). RNase protection assays were carried out as described previously (Saccomanno et al., 1992). Following hybridization of RNA with appropriate probes, RNase T2 was used to digest away single-stranded RNA. RNase-resistant hybrids were analyzed on 6% polyacrylamide, 8 M urea gels. Although not shown in all RNase protection figures, we routinely included RNA standards for size estimation of protected bands. These standards were in vitro transcribed RNA of 630, 480, and 400 nucleotides. Following electrophoresis, gels were dried and exposed to x-ray film.
In Situ Hybridization-In situ hybridization assays were performed as described previously (Goldman et al., 1991). 35 S-Labeled sense and antisense RNA probes were generated by run-off transcription (Melton et al., 1984) using [ 35 S]uridine 5Ј-(␣-thio)triphosphate (40 mCi/ml; Amersham Corp.) and either T3 or T7 polymerase. Probes were used at a concentration of 50,000 cpm/l of hybridization solution. Sense-strand probes consistently gave background signals.
Cell Culture-C2 muscle cells were grown in growth media as described previously (Evans et al., 1987). Neuronal cell lines SK-N-SH and NG108, the glial cell line C6, NIH 3T3 cells, and the hepatic cell line HEPG-2 were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. The PC12 cell line was grown in Dulbecco's modified Eagle's medium with 5% fetal calf serum and 10% horse serum.
Nuclear and Cytoplasmic Extracts-Extracts were prepared from either cultured cell lines or rat tissue. Phosphate-buffered salinewashed cells were suspended in 5 ϫ the packed cell volume of hypotonic buffer (10 mM Tris-HCl, pH 7.8, 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM DTT, 0.2 mM PMSF) and centrifuged at 3,000 rpm for 5 min. The resuspended cell pellet was allowed to swell on ice for 10 min in 3 ϫ the packed cell volume of hypotonic buffer and then homogenized with 10 -30 strokes. Efficient lysis was monitored by staining with trypan blue. Lysed cells were centrifuged at 4,000 rpm for 15 min. The supernatant, containing cytoplasmic proteins, was combined with an equal volume of 2 ϫ Laemmli sample buffer (Laemmli, 1970) and boiled for 5 min. The nuclear pellet was combined with an equal volume of 2 ϫ Laemmli sample buffer, boiled for 5 min, and sheared by passage through 21-and 25-gauge needles. Soluble nuclear extracts were prepared as described above, except the nuclear pellet was resuspended in an equal volume of high salt buffer (final concentration: 20 mM Hepes, 25% glycerol, 1.5 mM MgCl 2 , 0.42 M KCl, 0.2 mM EDTA, 0.2 mM PMSF, 0.5 mM DTT) and stirred at 4°C for 30 min. The sample was then spun at 14,500 rpm for 30 min, and the supernatant, containing extracted nuclear proteins, dialyzed against 20 mM HEPES (pH 7.9), 20% glycerol, 100 mM KCl, 0.2 mM EDTA, 0.2 mM PMSF, and 0.5 mM DTT for 4 h. Following dialysis the sample was centrifuged at 14,500 rpm for 20 min to remove precipitated material. The protein concentration of the supernatant was assayed using Bradford reagent (Bio-Rad), and the supernatant was stored in small aliquots at Ϫ80°C.
Expression of Recombinant Protein in Escherichia coli-Fusion proteins were generated using the pET-16b vector (Novagen). Recombinant plasmids were generated that contained MY1 or MY1a cloned into the XhoI site of the pET-16b vector. Constructs were prepared by restricting MY1 and MY1a with EagI and filling-in recessed ends with Klenow. Inserts were gel-purified and subcloned into the Klenow filled-in XhoI site of pET-16b. This resulted in the generation of a recombinant plasmid that would express a fusion protein containing 10 histidine residues at the NH 2 terminus of MY1/MY1a, which lack their first 18 amino acids. The generation of appropriate fusions was confirmed by DNA sequencing. Recombinant pET-16b vectors were used to generate fusion proteins in BL21(DE3)pLysS host (Novagen) according to manufacturer's directions. Recombinant proteins, containing histidine residues at their amino terminus, were purified according to manufacturer's directions (Novagen) using a metal chelation resin with nickel bound to it. Purified proteins were analyzed by SDS-PAGE.
Antipeptide Antisera-The multiple antigen peptide, TAIKKNNPR-KYLRSVG (encoded by nucleotides 520 -567; Fig. 2), was synthesized by Research Genetics and used to generate antisera in rabbits. Approximately 0.5 mg of peptide, emulsified in Freund's complete adjuvant, was injected into New Zealand White rabbits with two booster injections (0.5 mg peptide in Freund's incomplete adjuvant) at 3-week intervals. Antisera was collected at 8, 12, and 15 weeks post-injection.
Antibody Purification-Briefly, 450 mg of 6-aminohexanoic acid nhydroxysuccinimide ester-Sepharose 4B (Sigma) was washed and swelled in 100 ml of 1 mM HCl. MY1 recombinant protein, approximately 0.5 mg, was dissolved in 1.7 ml of 0.1 M NaHCO 3 , 0.5 M NaCl, pH 8.3. The dissolved protein and Sepharose 4B were mixed and stirred for 2 h at room temperature. The mixed slurry was transferred to a minicolumn (5-ml syringe) and the column resin washed with 10 volumes of 0.1 M NaHCO 3 , 0.5 M NaCl, pH 8.3, followed by five volumes of the same buffer containing 1 M ethanolamine. The column was allowed to sit for 2 h in the above buffer and then washed with three volumes of 0.1 M sodium acetate, 0.5 M NaCl, pH 4.2 buffer, followed by three volumes of 0.1 M sodium borate, 0.5 M NaCl, pH 8.2, buffer. The cycling of low and high pH buffers was repeated two more times. The column was then washed first with two volumes of 6.0 M guanidine-HCl followed by three volumes of distilled water and then equilibrated with 150 mM KPO 4 , pH 8.2. Subsequently, the column resin was washed and equilibrated with 150 mM NaPO 4 , 0.9% NaCl, 0.3% bovine serum albumin, and 0.1% bacitracin, pH 8.2, and incubated in 150 mM NaPO 4 , pH 8.2. Approximately, 0.5 ml of antibody containing serum was added to the resin and mixed at room temperature for 2 h. The resin was then returned to the mini-column and washed with 10 volumes of 150 mM NaPO 4 , pH 8.2. Bound antibody was eluted with 5 ml of 0.1 M glycine-HCl, pH 2.3 and collected in 3 ml of 3 M Tris-HCl, pH 8.6, 0.1% bovine serum albumin and mixed. The sample was first dialyzed against 100 mM Tris-HCl, pH 8, 2 M NaCl, followed by overnight dialysis in phosphate-buffered saline.
Expression Vectors-The wild-type ␦-promoter construct, pXP ␦-550, described previously (Chahine et al., 1992), contains sequences Ϫ550 to ϩ11 relative to the transcriptional start site. Scanner-linker mutations (slm) were generated by creating a series of 5Ј and 3Ј deletions on this construct using exonuclease III. These deleted DNAs were sequenced in order to define their 5Ј and 3Ј ends. ␦-550 slm Ϫ44/Ϫ29 has nucleotides Ϫ44 through Ϫ29 deleted, which were replaced with a 17-bp linker. The sequence of the linker is CAGATCTCGAGCTCCAC. ␦ 550 ⌬Ϫ52/Ϫ5 has nucleotides Ϫ52 through Ϫ5 deleted. This deletion mutant deleted an E-box sequence and a region with similarity to the SV40 core enhancer.
Transfections-SK-N-SH and NIH 3T3 cell cultures were transfected using Lipofectamine reagent according to Life Technologies, Inc. protocol. Briefly, each 60-mm cell culture dish was transfected with 0.5 g of test plasmid (pXP ␦ 550 or ␦ 550 ⌬Ϫ52/Ϫ5), 0. 5 g of CMV/CAT, and 3.0 g of pCMV/MY1 or pCMV vector without an insert. Forty-eight hours after transfection, the cells were harvested and assayed for luciferase and CAT activity as described previously (Brasier et al., 1989;Neumann et al., 1987).
Electrophoretic Mobility Shift Assays-The recombinant protein was incubated with 1.5 g of poly(dI)-(dC) in 10% glycerol, 15 mM HEPES (pH 7.8), 0.1 mM EDTA, 0.01% Nonidet P-40, and 0.5 mM DTT and the indicated concentration of competitors for 5 min at room temperature. Labeled DNA or RNA was then added and allowed to incubate for 20 min at 37°C. Samples were then run on a 4% polyacrylamide gel, 0.5 ϫ TBE (1 ϫ TBE is 89 mM Tris, 65 mM boric acid, 2.5 mM EDTA, pH 8.5) and electrophoresed at 250 V at 4°C. Gels were fixed, dried, and exposed to x-ray film.
Chromosomal Localization in the Mouse-The Jackson Laboratory Backcross Panel BSS (the N 2 progeny of ((C57BL/6y ϫ SPRET/Ei)F1 ϫ SPRET/Ei)) was used for chromosomal mapping (Rowe et al., 1994). A genomic Southern blot of all 94 backcross progeny was prepared from 3 g of each DNA digested with BglII which had shown a restriction fragment length polymorphism between Mus spretus (Ϸ11 kb, major band; Ϸ11.5 kb, minor band) and C57Bl/6J (Ϸ9 kb, major band; 8.5 kb, minor band) (not shown). The blot was hybridized with a random primed MY1 insert probe (SmaI fragment, nucleotides 360-1288), washed with high (0.2 ϫ SSC) stringency, and exposed to autoradiography. The observed segregation pattern was compared to approximately 1500 loci that had been previously typed in this cross by using Mapmanager, and its position determined by minimizing double crossovers (Rowe et al., 1994). 3

Scanner Linker Mutagenesis Identifies a Negative Regulatory
Element in the nAChR ␦-Subunit Gene Promoter-Direct injection of plasmid DNA into muscle allowed us to assay wild type and mutant nAChR ␦-promoter activity in vivo. Using this assay, we identified a scanner linker mutation (␦-550 slm Ϫ44/ Ϫ29) that resulted in increased promoter activity in innervated (approximately 3-fold increase over wild-type) and denervated (1.6-fold increase over wild-type) muscle (Fig. 1). In addition, this mutation resulted in a smaller induction of ␦-promoter activity after muscle denervation (3.7-fold for the mutant ver-sus 6.2-fold for the wild-type; Fig. 1). Scanner linker mutation Ϫ44/Ϫ29 largely disrupts the ␦-promoter's SVCE-like sequence. These results suggest that the SVCE-like sequence of the ␦-promoter participates in maintaining a low level of ␦-subunit gene expression in innervated muscle.
Cloning and DNA Sequence of MY1 and MY1a-To identify proteins that may bind to this repressor element, a doublestranded oligonucleotide (SVCE) corresponding to nucleotides Ϫ51 to Ϫ31 (Chahine et al., 1992) was prepared and used to screen an expression library created from innervated rat muscle RNA. After screening approximately 5 ϫ 10 5 phage, 15 positive hybridizing plaques were purified and analyzed by restriction enzyme digestion, Southern blots, and DNA sequencing. This analysis indicated that these clones were similar and could be divided into two classes. The longest inserts from these two classes of clones were subcloned into the BSSK(ϩ) vector and their DNA sequence determined.
Analysis of the DNA sequences showed that clones MY1 and MY1a are 1847 and 1639 nucleotides long, respectively, excluding their poly(A) tails (Fig. 2). The only difference in DNA sequence between these two clones is that MY1a lacks 207 nucleotides that correspond to nucleotides 736 -942 in MY1 (dashed line in Fig. 2). MY1 and MY1a contain open reading frames that encode proteins of 423 and 354 residues, respectively (Fig. 2). The 3Ј untranslated region of these clones contains two polyadenylation signal sequences, of which the most 3Ј is followed by a poly(A) tail.
Comparison of the DNA sequence or deduced amino acid sequence with corresponding sequences in the GenBank data base indicate that these two clones are most similar to a family of proteins containing a cold shock domain, referred to as Y-box binding proteins . Within this family of proteins MY1/1a are most similar to dbpA and RYB-a (approx. 81% at the nucleotide level) (Sakura et al., 1988 and Ito et al., FIG. 1. Scanner linker mutations identify a suppressor sequence in the nAChR ␦-subunit gene's promoter. Either wild-type (␦-550) or mutant (␦-550 slm Ϫ44/Ϫ29) ␦-promoter expression constructs were injected into innervated and denervated rat extensor digitorum longus muscle to study their in vivo expression. The expression vectors were coinjected with CMV/CAT for normalization. One week following DNA injections, muscles were harvested and luciferase and CAT activities determined. Bar graphs represent the mean luciferase activity from 3-4 injections normalized to CAT activity; error bars represent standard error of the mean.

1994).
Comparison of the deduced amino acid sequence of MY1/1a with the Y-box family of binding proteins indicates that the conserved cold shock domain is the most highly conserved region between members of this family of proteins . Within the cold shock domain (amino acids 138 -245, encoded by nucleotides 412-735) is a putative RNA binding domain (amino acids 156 -163, encoded by nucleotides 466 -489) (Landsman, 1992). The predicted isoelectric point for these proteins is 10.8, which is consistent with their high arginine content (approximately 12%). These arginine residues are clustered in the carboxyl half of the protein. In addition, these proteins contain a high percentage of proline residues (12%).
Tissue Distribution of MY1 and MY1a RNAs-RNase protection assays were used to identify MY1/1a RNAs in various tissues (Fig. 3A). Probes were prepared from a subclone of MY1, which contains nucleotides 1-1310. This DNA was linearized with PvuII and used to generate an antisense RNA probe complementary to nucleotides 829-1310. The probe contained 120 nucleotides complementary to MY1, which are not present in MY1a, and therefore allowed us to distinguish between these two RNAs. The upper, fully protected band (approximately 480 nucleotides) in the RNase protection assay presumably represents MY1 RNA while the lower, partially protected band (approximately 380 nucleotides) represents MY1a RNA (Fig. 3). Relatively high levels of MY1a RNA were detected in heart, skeletal muscle, and lung, while relatively high levels of MY1 RNA were found in retina. Although not shown, we also detected relatively high levels of these RNAs in testis. We were unable to detect any MY1/1a RNA expression in liver tissue.
To confirm that the two protected RNA bands do indeed correspond to MY1 and MY1a RNA, probes were generated that contained different amounts of sequence present in MY1 but deleted in MY1a. These probes allowed us to map the RNA sequence that was responsible for generating this doublet RNase protection pattern. The probes used in this experiment are diagrammed in Fig. 3B. It is clear from the RNase protection pattern that the difference in the two protected RNAs result from the 207 nucleotides present in MY1, but lacking in MY1a, since once this region is deleted from the probe a single protected fragment is observed (Fig. 3B).
MY1/1a expression in brain and retina may reflect restricted expression to a few specific cell types or general low level expression throughout these tissues. To investigate this further, we used in situ hybridization to assay for these RNAs in retina and brain (Fig. 4). This analysis showed relatively high levels of MY1/1a RNA in cells of the pia, layer 2 of the cortex, cerebellum, and glia, but lower levels in various neurons located throughout the brain. In layer 2 of the cortex, highest expression was observed in the motor area. The cerebellum showed high expression in all three cell layers and in glia. In The conserved cold shock domain is located between the arrows depicted beneath the sequence. Asterisks beneath the sequence identify a conserved RNA binding domain. The dashed line beneath the sequence identifies residues present in MY1 but deleted in MY1a. Polyadenylation signal sequences are in bold italics in the 3Ј untranslated region of the sequence.

FIG. 3. Tissue distribution of MY1/MY1a RNAs.
A, tissue distribution of MY1/MY1a RNAs as determined by RNase protection assays. Ten micrograms of total RNA was hybridized with 5 ϫ 10 5 cpm of MY1 antisense RNA probe 3 (see B) prior to RNase T2 digestion. B, various deleted antisense RNA probes identify the upper and lower bands in RNase protection assays as representing MY1 and MY1a, respectively. 1-3 above the lanes represents the probe used in that particular experiment. the retina, most of the expression is confined to the inner segment of the photoreceptors.
We also surveyed a number of cell lines for MY1/1a expression (Fig. 5). We observed a high level of expression in the muscle C2 and glial C6 cell lines, with relatively lower levels of expression in the NG108 and PC12 cell lines. Interestingly, we were unable to detect any MY1/MY1a RNA in the SK-N-SH cell line. Very low levels were identified in the hepatic HepG2 cell line, although a partially protected band is observed, which may represent an alternatively spliced or related gene product.
Antibodies to MY1/1a Identify Nuclear Proteins in C2 Muscle Cells-Members of the Y-box family of proteins have been shown to be involved in both regulated gene expression and RNA translation. In order to begin to define a role for MY1/1a, we investigated their cellular distribution. Nuclei and cytoplasm were isolated from various cell lines and fractionated on 10% polyacrylamide-SDS gels. We used SK-N-SH cell cytoplasm and nuclei as a negative control in these experiments since these cells contain no detectable MY1/1a RNA and presumably protein. Western blots were probed with affinity-purified antibody generated against a synthetic peptide corresponding to MY1 as described under "Materials and Methods" (Fig. 6). This analysis identified a doublet of proteins, approximately 32 kDa in molecular mass, that are absent from SK-N-SH cell extracts but present specifically in nuclei of cells expressing MY1/1a RNA. This same protein doublet is detected in a 0.4 M KCl extract of C2 nuclei (data not shown) implying it is a soluble nuclear protein. Two other sets of proteins were also detected: a 34-kDa protein present in nuclei of SK-N-SH cells and HepG2 cells and a 45-kDa protein present in both cytoplasm and nuclei of all cells assayed. These latter two proteins are unlikely to represent MY1/1a since they are present in SK-N-SH cells, which lack MY1/1a RNA. Preincubation of the primary antibody with the peptide used to generate it resulted in no signal, indicating that all the bands shown are a result of specific binding by the primary antibody.
Although MY1 and MY1a are predicted to have molecular masses of 36 and 28 kDa, respectively, it is likely that these highly basic proteins are migrating anomalously on SDS-PAGE. Therefore, based on the identification of a doublet of approximately 32 kDa by Western blot and the absence of this doublet from SK-N-SH cells, we conclude that this doublet represents MY1/1a. The 34-and 45-kDa proteins identified on the Western blot likely represent other members of the Y-box family of proteins that are related to MY1/1a. This antibody cross-reactivity may be expected since the antibody was generated against a peptide whose sequence is in the cold shock domain, which is expected to be conserved among various members of the Y-box family of proteins.  2B) followed by digestion with RNase T2. Samples were fractionated on 6% polyacrylamide, 8 M urea gels, which were subsequently dried and exposed to x-ray film overnight at room temperature.

MY1/1a Suppresses ␦-Subunit
Promoter Activity-Due to the fact that MY1/1a is highly expressed in muscle tissue, we chose to examine its effect on ␦-promoter activity in a cell line that does not contain detectable levels of this protein. We thus overexpressed MY1/1a in SK-N-SH cells and assayed for their ability to regulate the activity of the nAChR ␦-subunit gene's promoter. MY1 and MY1a both reduced ␦-promoter activity to less than 30% of that found in the absence of these proteins (data shown for MY1 in Fig. 7). However, deletion of the 47-bp activity-dependent enhancer (␦ 550 ⌬Ϫ52/Ϫ5) completely relieved suppression of ␦-subunit promoter activity by MY1/1a (Fig. 7).
The pCMV vector, without MY1/1a insert, had no effect on ␦-promoter activity.In addition, MY1/1a did not suppress CMV/ CAT or pXP⑀-321 activity (data not shown). Furthermore, we repeated these experiments using NIH 3T3 cells, which unlike SK-N-SH cells, endogenously express low levels of MY1/1a. As shown in Fig. 7, similar results were obtained in 3T3 cells except that MY1 overexpression resulted in about a 50% decrease in ␦-subunit promoter activity.
Interestingly, compared to wild-type ␦-promoter activity, the deletion mutant (␦ 550 ⌬Ϫ52/Ϫ5) showed about 1.5-fold higher activity in 3T3 cells (Fig. 7). This is consistent with the fact that the deleted sequence (spanning Ϫ5 to Ϫ52) contains the repressor element. In SK-N-SH cells, however, the lowered expression of ␦ 550 ⌬Ϫ52/Ϫ5, compared to the wild-type ␦ 550, may indicate that the deleted sequence also harbors a neuron-specific positive regulatory element.
MY1/1a Preferentially Binds to Single-Stranded DNA-To characterize the binding of MY1/1a to nucleic acids, we performed gel electrophoretic mobility shift assays. In all these experiments, MY1 and MY1a bound in an identical manner. Therefore, data are presented for only one of the proteins. Comparison of recombinant MY1/1a (60 and 120 ng) binding to SVCE, E-box, oligonucleotide 3, and start-site double-stranded oligonucleotides revealed promiscuous binding as illustrated in Fig. 8 (data shown for SVCE and E-box oligonucleotides, left panel). This binding was competed with approximately a 50fold excess of unlabeled double-stranded oligonucleotide (data not shown). In contrast, when single-stranded oligonucleotide probes were used in the gel shift assay, specific binding was revealed (Fig. 8, middle panels). In these experiments MY1/1a (1.2, 3.6, and 10.8 ng) preferred to bind to the coding (ϩ) strand of the SVCE or E-box oligonucleotide, but did not bind to the (Ϫ) non-coding strand. In addition, note that to generate a band shift we required approximately 50-fold less protein when us-ing single-stranded DNA versus double-stranded DNA. This binding to single-stranded DNA was blocked by including an excess of unlabeled (ϩ) strand oligonucleotide in the binding FIG. 6. MY1/MY1a cDNAs encode nuclear proteins. Nuclear and cytoplasmic proteins were isolated and fractionated on SDS-polyacrylamide gels. Western blots were probed with affinity-purified antibody and signals detected using the enhanced chemiluminescence system (Amersham). Preincubation of antibody in peptide used to generate the antibody resulted in no signal. C, cytoplasm; N, nuclei.
FIG. 7. MY1 overexpression suppresses nAChR ␦-subunit promoter activity. SK-N-SH and NIH 3T3 cells were cotransfected with a pCMV expression vector containing or lacking a MY1 cDNA insert, a CMV/CAT expression vector, and the test plasmid, i.e. pXP wild-type nAChR ␦-promoter expression vector, ␦ 550 (left panels) or deletion mutant ␦ 550 ⌬Ϫ52/Ϫ5 (right panels). The cells were harvested and assayed for luciferase and CAT activity 48 h after transfection. Experiments were repeated a minimum of three times. Bar graphs represent the average of triplicate transfections normalized to CAT activity; error bars are Ϯ standard deviation.
FIG. 8. MY1/MY1a recombinant proteins specifically bind pyrimidine-rich single-stranded DNA. Various oligonucleotides were end-labeled and used as probes for MY1/MY1a binding in electrophoretic mobility shift assays. Although the data shown are results obtained with MY1 binding, identical results were obtained when MY1a was used in these experiments. Binding to the double-stranded (ds) probes used 60 and 120 ng of MY1/MY1a recombinant protein, while binding to the single-stranded (ss) probes used 1.2, 3.6, and 10.8 ng of recombinant MY1/MY1a protein. Triangles above the figure represent increasing protein used in that experiment. Following electrophoresis, gels were dried and exposed to x-ray film with intensifying screen at Ϫ80°C overnight. ϩ represents the coding strand, and Ϫ represents the non-coding strand. reaction. No competition was observed using similar amounts of (Ϫ) strand oligonucleotide.
Oligonucleotide 3 (which contains sequences present in both the SVCE and E-box oligonucleotides) bound MY1/1a better than any of the other single-stranded probes tested (Fig. 8,  right panels). A complete shift of the oligonucleotide 3 (ϩ) strand probe was generated with 3.6 -10.8 ng of protein, which only caused a partial shift of the SVCE or E-box (ϩ) strand oligonucleotides and the oligonucleotide 3 (Ϫ) strand oligonucleotide. In addition, we were able to identify three different complexes generated by binding these proteins to oligonucleotide 3. The increased binding of MY1/1a to either strand of the oligonucleotide 3 probe compared to the corresponding strands of the SVCE or E-box oligonucleotides suggest that the additional sequences found in oligonucleotide 3 facilitate MY1/1a binding.
Mapping Yb2, the Gene Encoding MY1/1a-To investigate if any mouse mutations in MY1/1a exist, we chromosomally mapped the gene encoding MY1/1a using the Jackson backcross panel BSS (Rowe et al., 1994). This gene has been assigned the name Yb2 (Fig. 9). No recombinants were found between Yb2, the microsatellite marker D6Mit220, and the Kcna1 gene of the shaker-type potassium channel gene cluster ( Fig. 9) (Dietrich et al., 1994;Lock et al., 1994). One recombinant each placed Yb2 distal to D6Mit218 and proximal to D6Mit25 (Fig. 9). Our results place Yb2 in the mid-distal region of mouse chromosome 6 (Elliot and Moore, 1994). Under high stringency hybridization conditions, two bands that always co-segregated in the cross were observed. It is thus a single gene that codes for MY1/1a. DISCUSSION We report here the identification of a mutation in the nAChR ␦-subunit gene's promoter that results in increased promoter activity in innervated muscle (Fig. 1), suggesting a mechanism of repression. Proteins participating in this repression were cloned and found to be members of the Y-box family of nucleic acid-binding proteins. These proteins are referred to as MY1 and MY1a (muscle Y-box proteins 1 and 1a) and were shown to decrease ␦-promoter activity in a sequence-specific manner (Fig. 7). The fact that these proteins bind to a previously identified 47-bp activity-dependent enhancer of the nAChR ␦-subunit gene (Fig. 8), and require sequences within this enhancer for their action (Fig. 7), suggests a role for these proteins in activity-dependent regulation of the nAChR ␦-subunit gene.
Y-box binding proteins have historically been defined by their ability to bind to an inverted CCAAT box in DNA. A number of these proteins have recently been cloned from mammalian cDNA libraries and include YB-1, EF1a, MSY1, p50, MUSY-1, and dbpA (Didier et al., 1988;Ozer et al., 1990;Tafuri et al., 1993;Evdokimova et al., 1995;Sakura et al., 1988;Wolffe et al., 1992). These proteins share an 80-amino acid sequence (referred to as the cold shock domain) with each other and with the E. coli cold shock protein, CS 7.4 . The cold shock domain appears to participate in nucleic acid binding Bouvet et al., 1995).
The Y-box family of proteins carry out diverse functions ranging from transcriptional to translational controls. For instance, YB-1, a Y-box protein, has recently been shown to repress transcription of human major histocompatibility class II genes (Ting et al., 1994). Some other members of this family have been implicated in activating transcription from the Rous sarcoma virus , HLA class II (Didier et al., 1988), and the hst gene promoters (Hasan et al., 1994). In these cases the Y-box binding protein recognizes an inverted CCAAT sequence in the double-stranded DNA. Likewise, Y-box proteins from Xenopus laevis, FRGY1 and FRGY2, also are capable of activating transcription from promoters containing an inverted CCAAT sequence . However, there are also reports that certain Y-box proteins prefer to bind pyrimidine-rich single-stranded DNA (Kolluri et al., 1992). Based on these studies, it has been proposed that these proteins participate in regulating gene expression via binding to single-stranded regions of DNA that are complementary to those participating in an intramolecular DNA triplex structure (Kolluri et al., 1992;Horwitz et al., 1994).
In addition to transcriptional regulation, Y-box proteins MSY1, p50, and FRGY2 also participate in regulating translation by binding and sequestering cytoplasmic RNAs from the translational machinery (Tafuri et al., 1993;Evdokimova et al., 1995;Bouvet and Wolffe, 1994). Therefore, Y-box binding proteins represent a family of proteins with nucleic acid-binding properties that have important implications for DNA and RNA expression.
The Y-box family members we report here, MY1 and MY1a, are identical except that MY1a lacks 207 nucleotides found in MY1 at positions 736 -942 (Fig. 2). These data suggest MY1/1a RNAs are derived from the same gene by alternative splicing. The significance of expressing these two alternatively spliced forms is not clear. Most tissues examined express both of these RNAs with similar ratios (Fig. 3). They both suppressed ␦-promoter activity in transfection experiments (Fig. 7), and they both bound DNA with a similar affinity (data not shown).
Most interesting were our experiments which showed that MY1/1a overexpression was able to suppress ␦-promoter activity, while a mutant ␦-promoter expression construct containing a deletion spanning the 47-bp activity-dependent enhancer (␦ 550 ⌬Ϫ52/Ϫ5) completely relieved this suppression (Fig. 7). We chose to use this deletion in the expression studies since our DNA binding assays suggested that MY1/1a bound best to the 47-bp enhancer sequence (Fig. 8). These results suggest that MY1/1a may normally contribute to maintaining a low level of ␦-promoter activity in innervated muscle, consistent with the high level of MY1/1a RNAs found in this tissue. However, because we detected high levels of MY1/1a RNAs in C2 myo- FIG. 9. Genetic map of mouse chromosome 6 near Yb2. The gene encoding MY1/1a (assigned the name Yb2) was mapped in the BSS panel of the Jackson M. spretus backcross (28). Other genes that were also mapped to this region in this cross are shown. The genetic distance in centimorgans (Ϯ standard error) and, in parentheses, the number of recombinant animals over the total number of animals scored are shown. The one recombinant between D6Mit218 and Yb2 was not scored for Kcna1, which places Kcna1 simultaneously on top of D6Mit218 and Yb2. The position of homologous loci to human chromosomes is shown to the right. tubes (Fig. 5) and denervated muscle (data not shown), it is likely that other post-transcriptional mechanisms contribute to the regulation of functional MY1/1a proteins.
It is interesting that MY1/1a binds with highest affinity to the coding (ϩ) strand of the nAChR ␦-subunit promoter's 47-bp activity-dependent enhancer. Based on quantitating the band shifts obtained using single and double-stranded oligonucleotides and the different amount of protein used to generate a band shift, we estimate at least a 200-fold difference in binding affinity between double-and single-stranded SVCE oligonucleotide. Whether these proteins prefer a particular binding site could not be determined by the limited number of binding studies reported here. However, based on the oligonucleotides used in these studies and the strand preference displayed by MY1/1a, it appears that these proteins prefer pyrimidine-rich DNA sequences.
Pyrimidine-rich sequences have been proposed to mediate Y-box family member YB-1 binding to single-stranded DNA of the c-Myc and ␥-globin gene promoters (Kolluri et al., 1992;Horwitz et al., 1994). In addition, YB-1 has been reported to promote or stabilize single-strandedness in the major histocompatibility class II DRA promoter (MacDonald et al., 1995). However, in this latter case there is no evidence for this protein preferring pyrimidine-rich sequences. In addition, these investigators showed that the DNA sequences responsible for this single-stranded binding activity are different from those responsible for double-stranded binding (inverted CCAAT box). These latter studies suggest that YB-1 binding to the DRA promoter results in single-stranded regions that prevent binding of other trans-activators necessary for activation of DRA expression.
If MY1/1a acts by a similar mechanism, we predict that binding of MY1/1a to the ␦-subunit promoter would promote formation of single-stranded regions and reduce activation by other transcriptional regulators that normally bind to the double-stranded 47-bp enhancer, such as the E-box binding MyoD family members. We are currently testing this possibility.
Furthermore, we have demonstrated that the gene encoding MY1/1a, Yb2, is located in the mid-distal region of mouse chromosome 6. The glial and muscular expression of this gene suggests that it might be involved in neuromuscular diseases. However, the only known mouse mutation that maps into this region is scruffy (Scr), a mutation affecting the coat (Beechey and Searle, 1992), in which MY1/1a is unlikely to be involved. In addition, the genetic mapping to mouse chromosome 6 can be used to predict its homologous location in the human genome. Several genes that map near Yb2 (Elliot and Moore, 1994) map to human 12p13 where the human homologue of Yb2 is thus expected to lie. No human neuromuscular diseases are known to be linked to 12p13. The position of Yb2 is clearly distinct from the position of four unlinked loci, Yb1a-Yb1d, on mouse chromosomes 3, 7, 8, and 9 that represent the genes and possibly pseudogenes of YB-1, a different Y-box protein (Spitkovsky et al., 1992). In contrast to these studies on Yb1, we detected only a single locus for Yb2 and no evidence for related or pseudogenes.
In conclusion, we report the first characterization of a negative cis-acting DNA element in the nAChR ␦-subunit gene's promoter that binds Y-box family members MY1 and MY1a. The fact that the cis-acting element these proteins bind to is also involved in mediating activity-dependent regulation of the ␦-subunit gene suggests that these proteins participate in this regulation by keeping ␦-promoter activity low in innervated muscle. The preferential binding of MY1/1a to the coding strand of the 47-bp activity dependent enhancer suggests a novel mechanism of repression that may involve stabilization of single-stranded DNA regions. Finally, the expression of MY1/1a in the central nervous system suggests a regulatory role for these proteins within specific neurons. Whether or not these proteins also serve as repressors in these neurons remains to be determined.