A 102 base pair sequence of the nicotinic acetylcholine receptor delta-subunit gene confers regulation by muscle electrical activity

The muscle nAChR is a pentameric integral membrane protein that mediates communication between nerve and muscle (Schuetze and Role, 1987). Prior to innervation nAChRs are distributed throughout the muscles surface. However, after innervation they are concentrated beneath the neuromuscular junction (NMJ). The loss of receptors from extrajunctional regions of the muscle fiber is largely a result of nerve-induced muscle activity suppressing nAChR gene expression (Goldman et al., 1988; Merhe and Komhauser, 1989). This activity-dependent regulation is reversible, since denervation of adult skeletal muscle causes these receptors to reappear throughout the muscle fiber’s surface (Goldman et al., 1985; Merlie et al., 1984; Shieh et al., 1987).

The molecular mechanism by which muscle activity controls nAChR gene expression is not well understood. One approach to this problem is to identify the nAChR genes and characterize DNA sequences conferring activity-dependent regulation on their expression. This approach has resulted in the cloning of the chick α- and δ-, and mouse γ- and δ-subunit gene promoters (Klarsfeld et al., 1987; Wang et al., 1990; Gardner et al., 1987; Baldwin and Burden, 1988). However, convenient systems for characterizing activity-dependent gene expression have been lacking. To date, transgenic animals have provided the only system for studying electrical regulation of nAChR promoter activity. These studies have shown that electrical regulation of the chick α subunit promoter is mediated by sequences located within 850 bp 5’ to position +20 of the chick α-subunit gene (Merhe and Komhauser, 1989).

In order to rapidly analyze nAChR promoters for activity-dependent regulation we have identified a rat primary muscle culture system whose nAChR genes respond to muscle electrical activity in a fashion similar to that observed in vivo. We report here the cloning of the rat nAChR δ-subunit gene promoter and the use of this system to map a minimal 102 bp sequence that allows muscle electrical activity to regulate expression from this promoter. In addition, we show that this same DNA retains tissue and developmental stage-specific expression.

Approximately 5 × 10_⁵ phage from a Fisher rat genomic DNA library (Stratagene) were screened with a rat nAChR δ-subunit, _³²P-labelled, cDNA probe. A 7 kb fragment, containing exon 1, was subcloned into the pBS SK(+) vector (Stratagene) and characterized by Southern blot analysis for orientation and position of exon 1. A combination of restriction enzyme digestions and Southern blotting analysis resulted in the cloning of a 1.3 kb fragment of DNA containing part of the first intron and exon 1 at its 3’ end. This clone is called BSSK(+) δ-l.

Unidirectional deletions were generated by either exonuclease m or Bal 31 digestions of restricted DNA (Maniatis et al., 1982) Single-stranded DNA was isolated from phagemids according to manufactures directions (Stratagene) and sequenced using the dideoxy chain termination method of Sanger et al, 1977.

δ-gene sequences were subcloned into the pXPl or pXP2 expression vectors 5’ to the luciferase reporter gene (Nordeen, 1988). Internal controls were the mouse creatine kinase (MCK) promoter (Jaynes et al., 1986) driving chloramphenicol acetyltransferase expression (Luckow and Schutz, 1987) (MCKCAT), the SV40 promoter (Gorman et al., 1982) driving β-galactosidase expression (SV40βgal) or the CMV promoter (Boshart et al., 1985) driving CAT expression

Primary rat muscle cell cutures were prepared as previously described (Goldman et al., 1991). Cells were plated on collagen coated culture dishes at a density of 2000 cells/mm_² and grown in 20 mM HEPES, 2% chick embryo extract, 10% fetal calf serum, 10% horse serum, and 25 μg/ml gentamycin. Cultures were placed in a 37°C incubator with 8% CO2 Generally, by day 4 myotubes have formed and the cultures were treated with 3 μg/ml cytosine arabinoside for 24 hours to inhibit fibroblast proliferation. Cultures were either treated with TTX (10 μM) or stimulated on day 5 using previously described conditions (Goldman et al., 1988). Briefly, stimulating electrodes were immersed in the culture medium and the muscles stimulated chronically for 3-7 days in 100 Hz trains, 1 sec duration, applied once every 100 sec Stimulus pulses within trains were of alternating polarity, their duration was 0.5 msec, and their strength was 8 mA.

The mouse muscle cell line, C2C12, was cultured as previously described (Evans et al., 1987).

NIH 3T3 cells were grown in DMEM supplemented with 10% calf serum and 25 μg/ml gentamycin.

Transfection of muscle cells in culture was performed by CaPO4 precipitation of DNA as previously described (Graham and Van der Eb, 1973), with the addition of a 2 min glycerol shock after 4 hours (Parker and Stark, 1979). Cells were harvested 48-72 hours later. Luciferase, chloramphenicol acetyltransferase (CAT) and β-galactosidase assays were performed as previously described (Brasier et al, 1989; Neumann et al., 1987; Bradford, 1976)

RNA from cell culture and tissue was isolated using the guanidinium isothiocyanate procedure as previously described (Chirgwin et al, 1979).

A 25 nucleotide long ohgodeoxynbonucleotide complementary to a portion of exon 1 (+48 to +72) of the δ-subunit gene (see Fig. 1) was used for primer extensions as previously described (Maniatis et al., 1982).

RNase protection assays were carried out as previously described (Goldman and Staple, 1989). The delta clone BSSKδ-1087/+96 linearized with PvuII generates a 440 bp probe which spans most of exon 1 and extends 344 bp upsteam of the transcriptional start site. The muscle creatine kinase clone, BSSKII(+) MCK, is a 292 nucleotide long BglII/ HindIIIfragment corresponding to nucleotides 857-1149 of the mouse muscle creatine kinase subclone pMCKm36 (Jaynes et al., 1986). The MCK probe protects two fragments in the RNase protection assay that reflects slight sequence differences between mouse and rat creatine kinase RNAs

The δ-subunit promoter region was isolated from a rat genomic library as described in Materials and methods. DNA sequencing revealed subclone BSSK(+) δl is 1,317 nucleotides long and contains part of the first intron and all of exon 1 at its extreme 3’ end (Fig. 1). The nucleotides corresponding to the start site of transcription were mapped by RNase protection and primer extension experiments (Fig. 2). The primer extension experiments showed denervated muscle and primary rat myotubes initiate ó-subunit gene transcription at the A and T residues marked with an asterisk at positions +1 and +2, respectively, in Fig. 1. In addition, denervated muscle also initiated transcription at position −1. Consistent with these results were those obtained with RNase protections. Since RNase A only cleaves 3’ to pyrimidine residues and the probe is an antisense sequence, the high relative molecular mass band on the gel is consistent with transcripts starting at nucleotide −1 (RNase cleavage 3’ to the C residue complementary to nucleotide −5 in Fig. 1) and the intermediate band is consistent with transcripts starting at nucleotides +1 and +2 (RNase cleavage 3’ to the T residue complementary to nucleotide +1 in Fig. 1). The lowest band is a result of probe secondary structure (data not shown).

No canonical TATA or CAAT boxes were found upstream of the start site of transcription, although the start site is centered at an 8 nucleotide stretch that is identical to the 8 central nucleotides of the murine terminal deoxynucleotidyltransferase (TdT) genes initiator sequence (Smale and Baltimore, 1989). This latter sequence has been implicated to play a role in basal promotion and proper initiation of transcription.

Comparison of the rat δ-subunit genes 5’ flanking sequence with that of the published mouse and chick δ-subunit sequences (Baldwin and Burden, 1988; Wang et al., 1990) showed about 90% identity with the mouse sequence for the 246 nucleotides upstream of the start site of transcription. In contrast, the chick sequence showed little homology with the rat sequence except for two regions: one centered around postions −22 of the rat (relative to transcriptional start site) and −110 of the chick (relative to translation inititiator ATG), which contains a consensus sequence for binding myogenic regulators and shows a 12/14 bp match; and the second centered around positions −147 of the rat and −192 of the chick, which contains part of the SHUE box (Crowder and Merlie, 1988) and is identical for 8 nucleotides.

Upstream of the transcriptional start site, centered around nucleotides −550 and −223, are two regions of DNA that contain 15 and 12 CA repeats, respectively. CA repeats have also been found in the mouse δ- and γ-subunit genes. Finally, there is a 27 bp palindromic region centered around base pair −314 (dashed underline in Fig. 1).

We have found that muscle activity appropriately regulates nAChR gene expression in rat primary muscle cells. This effect is most pronounced when one compares cells stimulated with extracellular electrodes to those treated with the sodium channel blocker, 1T X. Specifically, the α-, β-, γ- and δ-subunits of the nAChR are all suppressed by electrical activity, while the electrically insensitive e-subunit and muscle creatine kinase genes are unaffected (K. G. Chahine and D. Goldman, unpublished data). Fig. 3A, shows the δ-subunit RNA is suppressed nearly 10-fold by electrical stimulation, while the creatine kinase gene (MCK) is slightly increased. Therefore, with a system in which to study muscle electrical activity established we isolated the δ-subunit promoter and began to identify cis-acting elements that mediate this effect.

The first intron and most of exon 1, including the initiator ATG, were removed from the 3’ end of BSSK(+) δ-l by digestion with Bal 31 exonuclease. One of the clones generated from this digestion, BSSK(+) δ1087/4, had its 3’ end at nucleotide +11 and extended 1098 nucleotides in the 5’ direction as determined by DNA sequencing (Fig. 1). This clone served as the parent for all further deletions. Unidirectional deletions were created from the 5’ end of this DNA using ExoIII. The precise 5’ end of each of these deletions was determined by DNA sequencing. Clones were named according to the number of 5’ nucleotides remaining upstream of the start site of transcription. These deleted prompters were subcloned into the pXP luciferase expression vector and transfected into rat primary myotube cultures that were either stimulated to contract with extracellular electrodes or kept inactive by addition of TTX for 3-5 days. Following transfection, the cells were returned to the stimulator or retreated with TLX and harvested 48-72 hours later. Transfection efficiency was normalized by co-transfecting cultures with MCKCAT. The MCK promoter like the endogenous gene is not suppressed by muscle electrical activity (unpublished data).

These experiments showed pXPδ1087/4 contained all the necessary elements to. confer electrical regulation on promoter activity (−1087/4 in Fig. 3B). The minimal cis-acting region that allowed promotion and still retained electrical regulation was 102 bp upstream of the transcriptional start site (Fig. 3B). Deletion of an additional 20 bp from the 5’ end of pXPδ102/4 resulted in loss of promoter activity. Within these 20 bp resides a putative AP-2 consensus sequence which may be required for promoter activity. These experiments also indicated that there is a negative element located between positions −1087 and −550. Removal of this element increased the activity of the promoter about 4-fold in TTX-treated cells. Additional deletions have shown this element to reside between nucleotides −677 and −550 (data not shown). There was a gradual decrease in promoter activity down to deletion pXPδ102/4. Further deletions resulted in a dramatic loss of promoter activity. Deletion pXPδ102/4 retained approximately 65% of the activity observed in pXPδ550/4.’

The muscle nAChR genes are specifically expressed in muscle cells after myoblasts fuse into multinucleated myotubes. Thus these genes are expressed in a tissue and developmental stage-specific manner. We assessed whether our various δ-subunit promoter deletions retained tissue and developmental stage-specific expression by comparing their promoter activity in transfected C2 myoblasts and myotubes, and NIH 3T3 cells (Fig. 4). Transfection efficiency was normalized to a co-transfected SV40β-gal expression vector. These experiments revealed that all δ-promoter deletions that contained a minimum of 82 bp upstream of the start site of transcription (pXPδ82/4) had a significantly higher level of expression in myotubes than in myoblasts or NIH3T3 cells (Fig. 4). However, this level of expression is significantly lower than constructs containing additional 5’ DNA.

As was observed in transfected primary muscle cultures a negative element is located between nucleotides −1087 and −550. This element is promiscuous in that it is active in all cell lines and primary cultures tested. Interestingly, δ-promoter activity increased over 20-fold in C2 myoblasts and myotubes when the negative element was removed as compared to a 6-fold increase in NIH3T3 cells. These results are similar to those observed with the chick δ-promoter (Wang et al., 1990).

Finally, we noticed two significant differences concerning δ-promoter activity in the C2 muscle cell line compared to primary muscle cultures. First the negative element had a larger effect on δ-promoter activity in the C2 cell line than it did in the primary rat muscle cultures. Therefore, its removal increased promoter activity over 20-fold in C2 cultures (Fig. 4), but only by about 3-fold in primary muscle cultures (Fig. 3). Second, the largest 5’ deletion tested that still gave significant promoter activity in the C2 cells was pXPδ82/4 (Fig. 4), while in primary cells this construct was at background levels of expression (Fig. 3). This latter result was not due to differences in transfection efficiency (data not shown).

We have presented data showing that rat primary muscle cultures can be used to study mechanisms by which muscle electrical activity regulates nAChR gene expression. In order to characterize the DNA sequences conferring activity-dependent regulation on these genes we cloned and sequenced the δ-subunit genes 5’ flanking DNA (Fig. 1). When δ-promoter/luciferase expression vectors were introduced into rat primary muscle cultures that were either stimulated with extracellular electrodes or made inactive by TTX-treatment, the δ-promoter conferred activity-dependent regulation on luciferase expression (Fig. 3). A series of deletions from the 5’ end of the δ-promoter identified a minimal 102 bp sequence that was sufficient for conferring this regulation on promoter activity. Further deletions resulted in loss of promoter activity, perhaps due to deletion of a putative AP2 site (Fig. 1). Expression from the MCKCAT vector used as an internal control, like the endogenous gene, was not suppressed by muscle electrical activity (data not shown).

One significant difference between the endogenous and transfected genes are the degree of regulation by muscle electrical activity. Generally, we can suppress the endogenous δ-RNA about 10-fold by electrical stimulation, yet the transfected δ-promoter is only suppressed about 4-fold (Fig. 3). Although this difference may indicate additional post-transcriptional controls regulating δ-RNA levels, we believe it may be accounted for by the observation that following calcium phosphate transfection the primary muscle cultures experienced a 12-16 hour period of inactivity. This would result in a higher level of promoter activity in stimulated cells causing a reduction in the observed effects of TTX. We have tried other transfection techniques, such as DEAE-dextran and lipofection, without success.

It is interesting that as we deleted DNA from the 5’ end of the δ-promoter its activity was consistently affected to a greater extent in TTX-treated cells than in stimulated cells (Fig. 3B). Perhaps the proteins that bind these upstream regulatory elements interact with a second protein that is expressed at low levels or not at all in active muscle. This would be consistent with the observation that ongoing protein synthesis is required to maintain the elevated level of nAChR expression seen in inactive muscle (Duclert et al., 1990).

The nAChR α, β, γ, and δ-genes are all suppressed by muscle electrical activity. The sequences mediating this effect may be common among these different genes. We compared the electrically responsive 102 bp sequence of the rat δ-promoter to the other nAChR gene sequences. Except for the mouse and chick δ-promoters the other nAChR promoters showed little sequence similarity to this 102 bp region. However, it is important to note that within this 102 bp region at position −22 is a consensus sequence binding site for myogenic regulators (Lassar et al., 1991). Both the chick δ-(sense strand) and a-promoter (antisense strand) sequences show a 12/14 and 10/14 match with the rat δ-promoter sequence in this region. Myogenic regulatory sequences are found in all the nAChR promoters characterized to date (Baldwin and Burden, 1988; Gardner et al., 1987; Klarsfeld et al., 1987; Wang et al., 1990). Although these sequences play an important role in activating muscle specific genes during development (Wright et al., 1989) they may also participate in other regulatory mechanisms.

That this myogenic regulatory sequence at position −22 might participate in developmental regulation of nAChR δ-subunit expression is supported by the observation that pXPδ82/4 is expressed at significantly higher levels in C2 myotubes than myoblasts or NIH3T3 cells (Fig. 4). Albeit this level of expression is about 5-fold lower than pXPδ102/4. The reason for this latter decrease in expression is likely due to removal of a putative AP2 consensus sequence located within the 20 bp deleted from pXPδ102/4 (Fig. 1).

Since the amount of expression drops as we delete DNA from pXPδ218/4 (Figs 3 and 4), it is likely that additional sequences involved in tissue and developmental stage-specific expression exist within this region of DNA. Consistent with this interpretation, one finds pXPδ218/4 contains an additional myogenic regulatory sequence (position −218) and 2 additional AP2 sequences (positions −211 and −104) compared to pXPδ824. The reason we do not detect pXPδ82/4 expression in primary muscle cultures (Fig. 3) but do detect it in C2 myotubes may simply reflect a difference in the level of expression of trarw-acting factors in these two sources of muscle cells.

Surprisingly, we find a low but above background level of expression of various δ-promoter constructs in NIH3T3 cells (Fig. 4). The reason for this is not known. Clearly, the endogenous gene is not expressed in these cells. One possibility is that there are additional cis-acting sequences required for tissue specific expression. Alternatively, as has been found for the skeletal α-actin promoter (Yisraeli et al., 1986), appropriate DNA methylation may be important for completely suppressing δ-promoter activity in non-muscle cells.

There are many parallels between myogenesis and muscle inactivity: (1) they both cause induction of nAChR gene expression (Evans et al., 1987); (2) myogenic regulators are induced in a developmental stage-specific manner (Wright et al., 1989) and suppressed by muscle activity (Eftimie et al., 1991) as are the nAChR genes (Evans et al., 1987; Goldman et al., 1988), and (3) DNA elements conferring developmental stage-specific expression and regulation by electrical activity have been localized to the same 102 bp region of DNA that contains a consensus binding site for myogenic regulators (this report). Therefore, it is tempting to speculate that this latter sequence participates in both induction of muscle specific genes during myotube formation and responsiveness of the nAChR genes to muscle activity. Since the trans-acting protein(s) that binds this regulatory sequence may be one of many hetero-dimers (Lassar et al., 1991), one might be able to obtain differential effects (developmental versus electrical) depending on which dimer is bound to this site.

We thank Supama Bhalla for help with DNA sequencing, Dr. Jeffrey Chamberlain for the MCK promoter and cDNA fragment and we are grateful to Linda Harper for her help in isolating rat primary muscle cells and maintaining cell cultures. We would also like to thank members of the MHRI electronics shop for their help in setting-up stimulators for muscle stimulation and the MHRI computing facility for help with quantitating autoradiograms.