Coupling muscle electrical activity to gene expression via a cAMP-dependent second messenger system.

Embryonic-type nicotinic acetylcholine receptor (nAChR) gene expression is regulated by muscle activity. The mechanism by which this activity is transduced to the genome is not known. We have addressed this issue by using a rat primary muscle cell culture system that mimics the in vivo effects of muscle activity on nAChR expression. We report here that the suppression of nAChR gene expression by muscle activity can be reversed by increasing intracellular cAMP levels. This effect is specific to the embryonic-type receptor genes. Electrically insensitive genes such as those encoding the adult-type nAChR epsilon-subunit and creatine kinase are not up-regulated by cAMP. In addition, muscle inactivity caused either by tetrodotoxin or denervation increases cAMP levels and protein kinase A activity, consistent with their proposed role in mediating nAChR gene expression. Finally, we report that this same mechanism appears to regulate other genes, such as those encoding the tetrodotoxin-insensitive sodium channel, MyoD, and myogenin which, like the nAChR, are regulated by muscle electrical activity. Based on these results it is proposed that muscle electrical activity is coupled to gene expression via a cAMP-dependent second messenger system.

The language of the nervous system is electrical. Information flows from one neuron to another by changing ion conductances across the cells membrane. The coupling of these changes between two neurons is often performed by neurotransmitters and their receptors. A central issue in neurobiology concerns the potential of a cell to respond to changes in electrical activity by altering its transcriptional and posttranscriptional repertoire. That such transcriptional changes do take place is illustrated by the depolarization-induced changes in expression of the genes encoding proenkephalin (Nguyen et al., 1990), c-fos (Bartel et al., 1989), and the muscle nicotinic acetylcholine receptor (nAChR)' (Klarsfeld and Changeux, 1985;Goldman et al., 1988). This latter protein and the vertebrate neuromuscular junction have provided * This work was supported by grants from the National Institutes of Health and the Muscular Dystrophy Association (awarded to D. G.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. §To whom correspondence should be addressed University of Michigan, Mental Health Research Institute, 205 Zina Pitcher PI., Ann Arbor, MI 48109.
During development of the neuromuscular junction, the level, distribution, and subunit composition of the nAChR change (for review see Schuetze and Role, 1987). Embryonictype receptors (az@yd) are distributed throughout the noninnervated embryonic muscle fiber. However, after innervation an c-subunit replaces the y-subunit, and these adult-type receptors (aZ@t6) are locally expressed beneath the neuromuscular junction. The loss of embryonic-type receptors from extrajunctional regions of the muscle fiber during synaptogenesis is largely a result of nerve-induced electrical activity suppressing expression of their genes. This is exemplified by the fact that one can prevent the induction of embryonic-type nAChR gene expression in extrajunctional regions of denervated rat muscle by electrically stimulating the muscle with extracellular electrodes (Goldman et al., 1988). In contrast to the electrically sensitive a-, 0-, y-, and &subunit genes, the adult-type c-subunit gene does not appear to be regulated by muscle activity Goldman et al., 1991).
The mechanism by which muscle electrical activity regulates nAChR gene expression has received much attention recently. Cloning of the a-subunit gene and promoter  has resulted in the identification of an 850base pair sequence that contains elements conferring activitydependent regulation on the reporter gene chloramphenicol acetyltransferase (Merlie and Kornhauser, 1989). Our laboratory has recently identified a 102-base pair sequence of the &-subunit gene that mediates regulation by muscle electrical activity (Chahine et al., 1992).
Second messenger pathways coupling electrical activity to altered patterns of nAChR gene expression have been explored using chick primary muscle cultures. These experiments have focused on the role of protein kinase C (PKC) in mediating the effects of muscle depolarization since this enzymatic activity has been reported to increase in active versus inactive muscle fibers (Cleland et al., 1989). In this system pharmacological stimulation of PKC abolished the increase in a-subunit mRNA caused by tetrodotoxin (TTX), whereas inhibition of this enzyme caused an increase in a-subunit mRNA (Klarsfeld et al., 1989).
We have investigated the mechanism by which muscle activity regulates nAChR RNA levels by employing rat primary muscle cells in culture that can be induced to contract by stimulating them with extracellular electrodes. Unlike that found in the chick system, we report that muscle electrical activity mediates its effect on embryonic-type nAChR subunit gene expression through a CAMP-dependent second messenger system. Furthermore, we show that this same second messenger system can similarly regulate the expression of other muscle specific genes that are known to be regulated by muscle activity.

EXPERIMENTAL PROCEDURES
Cell Culture and Electrical Stimuhtion-Rat skeletal myoblasts were isolated from the hind legs of day 17-19 Sprague-Dawley rat fetuses as described previously (Chahine et al., 1992). Cultures were stimulated on day 4 using conditions described previously (Goldman et al., 1988). Briefly, stimulating electrodes were immersed in the culture medium, and muscles were stimulated chronically for 3-7 days in 100-Hz trains, I-s duration, applied once every 100 s. Stimulus pulses within trains were of alternating polarity, their duration was 0.5 ms, and their strength was 8 mA. Pharmacological agents were generally added on day 7-8, except for TTX which was usually added when cells were stimulated. Forskolin and isobutyl-1-methylxanthine (IBMX) were prepared in 95% ethanol and used at a final concentration of 10 and 500 pM, respectively. The cAMP analog cpt-CAMP was prepared in Dulbecco's modified Eagle's medium and used at a final concentration of 0.25 mM. TTX was prepared in water and used at a final concentration of 3 gg/ml. Soleus Muscle Denervations and Diaphragm Organ Culture-Male Sprague-Dawley rats were anesthetized with ether. The sciatic nerve innervating the left lower hind limb muscles was sectioned in the upper thigh. The right hind limb served as an innervated control. At various times following muscle denervation, innervated (right) and denervated (left) soleus muscles were rapidly removed in the cold and divided longitudinally in half. One half was used for isolating RNA for RNase protection assays, and the other half was used to prepare extracts for cAMP and protein kinase assays.
Diaphragms with some adjoining ribs were dissected from 21-35day-old Sprague-Dawley rats. Isolated diaphragms were pinned by the adjoining ribs to culture dishes filled with Sylgard 184 silicone (Dow Corning). Primary muscle culture medium (described above) was added until the diaphragms were covered, and the dish was placed on a rocker to improve aeration. Diaphragms were maintained in 8% COz at 37 "C. Pharmacological agents were added to the diaphragms within 5 min of their isolation.
RNA Isolation-RNA was isolated from primary muscle cultures as described previously (Chomczynski and Sacchi, 1987) except that after the final precipitation the RNA pellet was resuspended in 10 mM Tris, pH 8.0, and digested with proteinase K (50 pg/ml) in 0.5% sodium dodecyl sulfate at 50 "C for 30 min, followed by pheno1:chloroform extraction and ethanol precipitation. RNA was isolated from rat soleus and diaphragm muscles using the guanidinium isothiocyanate procedure as described previously (Chirgwin et al., 1979;Goldman et al., 1985).
RNase Protection Assay and Probes-RNase protection assays were carried out as described previously (Goldman and Staple, 1989). Ten micrograms of total RNA was used in the assay. Autoradiograms from RNase protection experiments were quantitated by densitometry. Various exposures were used to confirm that the signal was in the linear range of the film. Densitometric values were normalized either to creatine kinase RNA levels, which were not regulated by muscle activity, or to total RNA. 32P-Labeled antisense RNA probes were prepared by runoff transcription of linearized vectors harboring either cDNA or genomic DNA. The nAChR a-subunit probe contains the complete 240 nucleotides of exon 8 flanked by about 50 nucleotides of intron on its 3' end and approximately 310 nucleotides of intron on its 5' end. The nAChR @subunit probe is a BglII/SmaI fragment of rat cDNA (328 (Goldman and Tamai, 1989). This probe spans nucleotides 1044-1273 of the published rat cDNA . The nAChR y-subunit probe is an EcoRIIBamHI subclone of RIG13-1 (Goldman and Staple, 1989). This probe spans nucleotides 91-651 of the published rat cDNA . The nAChR &subunit probe was derived from BSSK 6-1087/+96, which, upon linearization with PuuII, generates a 440-base pair probe that spans most of exon 1 and extends 344 nucleotides upstream of the transcriptional start site (Chahine et al., 1992). This probe identifies two &subunit RNA transcripts representing different transcriptional start sites (Chahine et al., 1992). The nAChR e-subunit probe was derived from cDNA e19-1-1 as described previously (Goldman et al., 1991). This probe spans exons 10, 11, and 12 of the c gene. The muscle creatine kinase probe was generated from BSSKII(+) muscle creatine kinase as described previously (Chahine et al., 1992). This probe spans nucleotides 857-1149 of the mouse muscle creatine kinase subclone pMCKm36 (Jaynes et al., 1986). The muscle creatine kinase probe protects two fragments in the RNase protection assay that reflects slight sequence differences between mouse and rat creatine kinase RNAs. The TTX-insensitive a-sodium channel (a-SC) probe was generated from a 3"untranslated portion of the a-SC cDNA, HSC3UT.' This probe is a 194-nucleotide-long HindIII/XhoI fragment of the sodium channel a-subunit. It encompasses nucleotides 6177-6371 of the published SKM2 cDNA (Kallen et al., 1990). The myogenin probe was generated from an EcoRI/SacI subclone of the myogenin cDNA (Wright et al., 1989). This probe spans nucleotides 1-173 of the published myogenin cDNA (Wright et al., 1989). The MyoD probe was derived from a HindIII-linearized BSSK(+) vector harboring the MyoD cDNA. This probe spans nucleotides 1202-1785 of the published MyoD cDNA (Davis et al., 1987).
cAMP Measurements and CAMP-dependent Protein Kinase Assays-Intracellular cAMP levels were determined usinga cAMP assay kit from Amersham Corp. This is a competition assay in which unlabeled cAMP (intracellular) competes with radiolabeled cAMP for binding to a protein that has a high affinity and specificity for CAMP. Extracts were prepared as follows. First, tissue culture cells were washed twice with phosphate-buffered saline, harvested in 0.3 ml of 10 mM NaP04, pH 7, 1 mM EDTA, 1 mM dithiothreitol, 0.25 M sucrose, and sonicated on ice. Ethanol extracts were prepared from a 0.15-ml aliquot according to manufacturer's directions; soluble material was dried under vacuum, resuspended in buffer, and 0.05 ml was used in the cAMP assay. Second, rat soleus muscles were rapidly removed in the cold, cut into small pieces with a razor blade, and sonicated in 0.5 ml of 4 mM EDTA, pH 8, containing 50 p~ IBMX, at 4 "C. Samples were then centrifuged at 15,000 rpm for 5 min and 0.05 ml used in the cAMP assay. All values reported were measured in the linear range of the assay.
CAMP-dependent protein kinase activity was assayed as described by Clegg et al. (1987). Both free (-CAMP) and total protein kinase A (+CAMP) were measured although we only are reporting changes in the free activity. Protein concentrations were determined using the method of Bradford (1976).

Electrical Activity Suppresses nAChR RNA Expression in Rat Primary
Muscle Cultures-To study the regulation of nAChR genes by electrical activity we used an in vitro rat primary muscle culture system that was responsive to electrical stimulation via extracellular electrodes. In this system myotubes began to spontaneously contract within a few days after their formation. However, this response was generally heterogeneous and often did not last more than a few days.
We induced a more uniform and long lasting effect by stimulating the cells with extracellular electrodes. Under these conditions, we were able to keep muscles uniformly contracting for more than 7 days. This was ample time to observe the influence muscle activity had on nAChR expression. We have not found any differences in total protein or RNA levels between stimulated and TTX-treated myotubes.
RNase protection assays were used to determine the effects of muscle electrical activity on nAChR RNA levels. Myotubes were either stimulated with extracellular electrodes or treated with T T X for a minimum of 5 days before harvesting. These experiments showed low to undetectable levels of receptor RNAs in stimulated muscle compared with cells whose activity was blocked with T T X (Fig. 1). In addition, other musclespecific RNAs whose expression is regulated by muscle activity, such as those encoding the TTX-insensitive sodium channel (a-SC) (Offord and Catterall, 1989;Kallen et al., 1990), MyoD and myogenin (Eftimie et al., 1991) are also regulated by muscle activity in our primary culture system (Fig. 1). Quantitation of this effect revealed a range between a 3.5-fold (myogenin) and over a 50-fold (nAChR a-subunit) difference in RNA levels in stimulated compared with TTX-treated cells (Fig. 1B). We have found that these RNAs are affected by muscle activity to different extents in different experiments (for example compare stimulated and TTX levels in Fig. 2).
However, we rarely see differences less than 3-fold for muscles that have been stimulated or TTX-treated for more than 5 days.
The specificity of this effect is illustrated by the finding E. Baracchini, unpublished data. Filled bars represent stimulated cells; stippled bars represent TTX-treated cells. The (star) in the alpha graph indicates that the level of this RNA in stimulated cells was helow the limits of detection. Although the &subunit RNA probe detects two RNAs representing different start sites of transcription (Chahine et al., 1992), only one of these RNAs is shown in this figure. However, both of these transcripts were regulated by muscle activity in a similar manner.

CAMP-dependent Regulation
that those genes that are relatively insensitive to muscle activity, such as those encoding the nAChR 6-subunit and creatine kinase, are expressed a t similar levels in stimulated and TTX-treated cells (Fig. 1).
Activity-dependent Regulation of RNA Expression Can Re Reversed by Raising Intracellular cAMP Levels-There are many reports that cAMP can increase expression of nAChRs in cultured muscle cells (Betz and Changeux, 1979;Blosser and Appel, 1980;McManaman et al., 1982;Fontaine et al., 1987). Based on these studies we investigated the role cAMP may play in mediating the effect of electrical activity on nAChR RNA expression. Cultured myotubes were stimulated or treated with T T X for 3-4 days prior to perturbing their cAMP levels. Stimulated cells were then treated with one of the following drugs for a minimum of 3 days: 1) forskolin, an activator of adenylate cyclase; 2) cpt-CAMP, a cAMP analog; and 3) IBMX, a phosphodiesterase inhibitor. Controls were treated with vehicle alone. Stimulated cells, in the presence or absence of drug, were monitored microscopically every 6-12 h to ensure that they were contracting throughout the course of the experiment. None of the drugs was found to inhibit muscle activity at the concentrations used in these experiments.
RNase protection assays were used to monitor the effect of increasing cAMP on nAChR, a-SC, myogenin, MyoD, and muscle creatine kinase RNA levels in stimulated muscle cells (Fig. 2). There were no significant differences in total RNA isolated from these muscle cultures. These experiments revealed that pharmacological agents that increased cAMP levels caused an increase in those RNAs known to be regulated by muscle electrical activity in uiuo. Similar to the effects of muscle electrical activity, these drugs had only a small effect on the nAChR 6-subunit and muscle creatine kinase RNAs. The relatively large decrease in muscle creatine kinase RNA upon the addition of cpt-CAMP likely reflects a nonspecific J , , + , . , effect of this drug since it was not observed with other drugs that increase CAMP.
Quantitation of these RNase protection assays revealed that genes responding to muscle electrical activity generally responded best to cpt-CAMP, followed by forskolin and then IBMX (Fig. 2R). This pattern of drug potency likely reflects their ability to perturb cAMP levels. Since cAMP stimulates PKA activity, we assayed the effect these drugs had on PKA activity in stimulated muscle cultures. As predicted, these experiments showed the same order of potency as revealed by the RNase protection experiments (Fig. 2C).
Based on the above results one would predict that the levels of cAMP and PKA activity should be higher in an inactive (TTX or denervated) uerw.9 active (stimulated or innervated) muscle. Fig. 3 presents the results of such a determination. Primary muscle cell cultures were either stimulated or treated with T T X for 4 days. Extracts prepared from TTX-treated cells contained 100% more cAMP and 65% more PKA activity than did extracts prepared from stimulated cells (Fig. 3 A ) . In addition, when we assayed these activities in denervated muscle we found an increase in both cAMP and PKA activity within 20 h of denervation (Fig. 3 R ) .
We confirmed that our assay was specific for PKA activity by including a specific inhibitory peptide, PKI (Smith et ai., 1990), in the enzymatic reaction. This resulted in complete inhibition of protein kinase activity (data not shown). Therefore, the assay we use is specific for CAMP-dependent protein kinase.
If cAMP mediates the effects of muscle electrical activity on gene expression, one might suspect that pharmacological agents that increase cAMP levels would change the time Reported PKA values are fold ahove stimulated or innervated muscles. Experiments were repeated four times, and error bars are f the standard deviation. Samples used to assay for cAMP and PKA were also used to assay for nAChR and creatine kinase RNAs, to confirm that the stimulation and TTX treatments were effect,ive. These RNAs were regulated as shown in course of expression of these genes upon muscle denervation. We tested this by determining the effect cpt-CAMP had on the induction of nAChR, a-SC, MyoD, myogenin, and muscle creatine kinase RNAs at early times following muscle denervation. Rat diaphragms were placed into organ culture and immediately treated with cpt-CAMP or buffer. Twelve hours later diaphragms were harvested and RNA isolated. RNase protection assays showed that after 12 h of cpt-CAMP treatment there was a significant induction in nAChR cr-and 6subunit, n-SC, and MyoD RNAs compared with the control treated diaphragm (Fig. 4). No effect was observed on the rate of appearance of the nAChR t -and &subunits, myogenin,

FIG. 5. Cyclic AMP and
TTX increase nAChR a-subunit heteronuclear RNA. RNnse protection assavs using an n-suhunit JiNA prohe that contains exon 8 and part of the flanking introns (see "Experimental Procedures"). Shown is the portion of the autoradiogram corresponding to complete protection of the prohe by heteronuclear RNA. For these experiments primary muscle cells were stimulated with extracellular electrodes for 4 days prior to the addition of cpt-CAMP or TTX. Stimulation was continued in the presense of drugs for an additional 4 days after which RNA was isolated and analyzed hy RNase protection experiments. and muscle creatine kinase RNAs, and the nAChR y-suhunit RNA was undetectable at these times.

Cyclic AMI' and TTX Increase nAChR tr-subunit Heteronuclear RNA in Stirnulatpd Muscle-In vioo,
muscle activity mediates its effect by suppressing embryonic-t-ype nAChR gene transcription. However, cAMP has been reported to stimulate acetylcholine receptor expression via a post-transcriptional mechanism in mouse fibroblast cultures (Green et al., 1991). T h e effects of TTX and CAMP-inducing pharmacological agents in our primary muscle culture system may be transcriptional or post-transcriptional. We addressed this issue by RNase protection assays to determine if raising intracellular cAMP increased the amount of heteronuclear nsubunit encoding RNA. These experiments employed a radiolabeled a-subunit RNA probe containing sequences spanning exon 8 and extending into its flanking introns to detect heteronuclear RNA (see "Experimental Procedures" for a description of the probe). Only a completely protected probe was assayed in these experiments, which would correspond to protection of the probe by heteronuclear cr-subunit RNA. We found that stimulated cells treated with the cAMP analog cpt-CAMP or T T X have increased levels of heteronuclear (1subunit RNA (Fig. 5).

Inactive Muscle Increases nAChR, n-SC, Mvol), and Myogenin RNAs Independent of Protein Synthesis-The effects
of blocking muscle activity on nAChR gene expression are a relatively slow process (Goldman et al., 1988). Perhaps this slow response reflects the requirement for synthesis of new proteins. We tested whether protein synthesis is required to induce nAChR RNAs and those encoded by other genes responding to muscle electrical activity in our in vitro culture system. Muscle cells were stimulated to contract with extracellular electrodes for 4 days. Cells were then treated with either TTX or T T X p l u s cycloheximide for 24 h. Cycloheximide was used a t a concentration of 10 pglrnl, which corresponded to approximately 99% inhibition of protein synthesis as assayed by incorporation of radiolabeled methionine into trichloroacetic acid precipitable proteins. Following drug treatment cells were harvested and RNA isolated for RNase protection assays. These experiments showed little effect of the protein synthesis inhibitor on induction of these RNAs (Table I).

DISCtJSSION
The experiments reported here were undertaken to identify second messenger systems mediating the effects o f muscle electrical activity on nAChR gene expression. These experiments employed a rat primary muscle culture system that could be maintained in an actively contracting state hy stimulating them with extracellular electrodes. This stimulation paradigm had no adverse effects on the cells as judged by cell viability, total RNA and protein levels, and regulation of gene expression. Those genes whose expression had been shown to
be regulated by muscle activity in vivo were also found to be regulated by this activity in vitro (Fig. 1). In addition, musclespecific genes that did not respond significantly to muscle activity in vivo likewise did not respond to this activity in vitro.
One difference between the in vitro and in vivo system is that generally in vivo muscle denervation results in a larger increase in RNA expression than was found between active and inactive muscle cultures. Although we do not know the reason for this difference, two possibilities exist: 1) the nerve may exert an additional influence, besides muscle activity, that suppresses gene expression, as has been suggested for the y-subunit gene (Witzemann et al., 1991); and 2) the muscle cells in culture are at a developmental stage in which these genes are being induced, and the induced state activating gene expression and the muscle activity depressing gene expression act against each other to give a level of expression that is higher than that seen in the adult innervated muscle fiber. This latter effect would diminish the difference between active and inactive muscles grown in culture.
Having established a system that responded to muscle activity, we then pursued the role various second messenger systems might play in mediating these effects. There are numerous reports in the literature indicating that cAMP may up-regulate nAChR protein synthesis (Betz and Changeux, 1979;Blosser and Appel, 1980;McManaman et al., 1982;Fontaine et al., 1987). Consistent with these reports, we found that pharmacological agents that increased cAMP levels also increased RNA levels derived from the embryonic-type nAChR genes and other genes known to be regulated by muscle electrical activity (Fig. 2). This effect manifested itself most dramatically in electrically stimulated muscle cells, implying that cAMP levels increased in an inactive/denervated muscle compared with an active/innervated muscle cell. Indeed when one measures cAMP and PKA activity in either active and inactive primary muscle cultures or innervated and denervated adult skeletal muscle one finds a significant increase in these activities in the inactive/denervated cells (Fig.  3).
If cAMP levels mediate the activity-dependent expression of muscle genes, such as those encoding nAChRs, why do we not see a 50-fold or larger increase in RNA levels as is reported for innervated/denervated muscle i n vivo (Merlie et al., 1984;Goldman et al., 1985)? Although the answer to this question is not known it may reflect the fact that our in vitro experiments are done on muscle cells that are stimulated to contract. Therefore, we might be observing the sum of two antagonistic processes: 1) muscle activity down-regulating RNA levels by an additional CAMP-independent second messenger pathway; and 2) pharmacological increases in cAMP causing an increase in these RNAs. Since one would not expect both of these pathways to be activated at the same time in innervated or denervated muscle, maximal differences in gene expression would be observed i n vivo.
Consistent with our i n vitro experiments are the results obtained with diaphragms in organ culture. Diaphragms were chosen for these experiments since they are a thin muscle, which would likely facilitate accessibility of fibers to drugs added to the medium. These experiments showed that by pharmacologically increasing cAMP levels at an early time following muscle denervation one can increase the rate at which specific RNAs appear in response to this denervation (Fig. 4). Although in general our results turned out as expected, we were not able to show an increase in the rate of appearance of the nAChR 0-subunit and myogenin RNAs, both of which have been shown to respond to muscle activity. The lack of an effect on the p-subunit RNA is not surprising since in vivo, in the absence of stimulation with extracellular electrodes, this RNA is already expressed at a relatively high level that does not change much upon muscle denervation (Goldman et al., 1988). The lack of a significant effect on the rate of appearance of the myogenin RNA may also not be too surprising since, unlike MyoD and nAChR RNAs, myogenin RNA is induced 9 h following muscle denervation (Eftimie et al., 1991). Thus at the time point analyzed this RNA may have already reached its maximal rate of induction.
Consistent with cAMP mediating the effects of muscle activity on nAChR gene expression is the observation that cAMP levels rise during myoblast fusion and decrease after muscle innervation Montague, 1974, 1975;Novak et al., 1972;Smith, 1980;Moriyama et al., 1976). In addition, our data (Fig. 3) and those of others (Carlsen, 1975;Hopkins and Manchester, 1982) show that muscle denervation increases cAMP levels. These results indicate that muscle denervation in the adult may recapitulate some of the events taking place during development. Therefore, it is likely that non-innervated myotubes, generated either from fusion of myoblasts or denervation of adult muscle, express their embryonic-type receptor genes at a high level as a result of increased levels of cAMP/PKA activity.
This conclusion is different from that proposed by investigators using a chick primary muscle culture system. In this case cAMP was assigned to mediating the neurotrophic effect of muscle innervation, and PKC activity was proposed to be involved in mediating the effects of muscle activity (Fontaine et al., 1987;Klarsfeld et al., 1989;Laufer et al., 1991). In this system suppression of PKC activity resulted in an 8-10-fold increase in a-subunit RNA (Klarsfeld et al., 1989). Recent experiments, using intact chick muscle, also indicate a role for PKC in coupling muscle activity with nAChR gene expression (Huang et al., 1992). We performed similar experiments on rat primary muscle cultures and found little affect of these drugs on embryonic-type nAChR gene expre~sion.~ In addition, investigators working with the chick system do not have a gene, like the rat e , that is predominantly regulated by neurotrophic controls. Therefore, experiments designed to distinguish neurotrophic from electrical mechanisms of control in the chick system are indirect.
Another difference that we found between the rat and chick system was the effect of inhibiting protein synthesis on induction of nAChR gene expression in response to TTX treatment. In the chick system it appears that new protein synthesis is required for denervation-induced increases in nAChR K. G . Chahine and D. Goldman, manuscript in preparation.
RNA (Duclert et al., 1990;Tsay et al., 1990), whereas in our rat system this appeared not to be the case (Table I). However, until we measure RNA turnover in the presence and absence of cycloheximide we cannot conclude that muscle inactivity induces nAChR and other RNAs in the absence of new protein synthesis in the rat system.
The reason for the above mentioned differences between the two systems is not clear but may simply represent alternative transduction pathways employed by these two species. If this is the case they must converge on a common pathway since transgenic mice created with the chicken a-subunit promoter regulate transgene expression upon muscle denervation similar to the endogenous mouse gene (Merlie and Kornhauser, 1989).
Based on the data presented here the following model emerges for regulation of embryonic-type nAChR, TTX-insensitive sodium channel (Sherman et al., 1985), myogenic regulator (MyoD and myogenin) gene expression, and perhaps other genes, such as aceylcholinesterase, (Rubin, 1985) that are regulated in a similar fashion by muscle electrical activity. This model is based in part on the observation that there is an inverse relationship of calcium-dependent adenylate cyclase and phosphodiesterase activity mediated by intracellular calcium (Piascik et al., 1980). Low concentrations of calcium preferentially activated adenylate cyclase, whereas high concentrations of calcium preferentially activated phosphodiesterase. In our model, muscle activity causes release of calcium from the sarcoplasmic reticulum. The increased intracellular calcium levels, via calmodulin, stimulate cAMP phosphodiesterases which maintain a low level of cAMP and therefore protein kinase A activity in innervated or stimulated muscle. Low intracellular calcium levels resulting from muscle denervation stimulate cAMP production by inhibiting phosphodiesterase activity but stimulating adenylate cyclase activity. In addition, muscle denervation stimulates phosphorylation of the PKA type I1 regulatory subunit which decreases its affinity for the catalytic subunit (Sayers et al., 1988;Rymond and Hofmann, 1982). The combination of increased cAMP levels and phosphorylation of the regulatory subunit of PKA would lead to increased PKA activity in the denervated muscle. Since the nAChR promoters do not contain obvious CAMPresponsive elements, it is likely that these genes are activated indirectly by cAMP/PKA. The steps that follow activation of PKA are not clear but may involve activation of the myogenic regulators MyoD and myogenin. Since these transcription factors are involved in regulating the expression of a wide range of muscle-specific genes, many of which are not electrically sensitive, one might expect these transcription factors to interact with other proteins that confer genetic specificity.
Finally, it should be noted that at this stage we have only tested this model by characterizing the role cAMP plays in mediating the effects of muscle activity on gene expression. Future work will focus on the molecular mechanisms by which cAMP levels change in response to muscle activity and the mechanisms by which these changed cAMP levels are transduced to the genome.