Nitric-oxide Synthase Is a Mechanical Signal Transducer That Modulates Talin and Vinculin Expression*

Mechanical stimuli can cause changes in muscle mass and structure which indicate that mechanisms exist for transducing mechanical stimuli into signals that influence gene expression. Myotendinous junctions show adaptations to modified muscle loading which suggest that these are transcriptionally distinct domains in muscle fibers that may experience local regulation of expression of structural proteins that are concentrated at these sites. Vinculin and talin are cytoskeletal proteins that are highly enriched at myotendinous junctions that we hypothesize to be subject to local transcriptional regulation. Our findings show that mechanical stimulation of muscle cells in vivo and in vitro causes an increase in the expression of vinculin and talin that is mediated by nitric oxide. Furthermore, nitric oxide-stimulated increases in vinculin and talin expression occur through a protein kinase G-dependent pathway and therefore differ from other mechanisms through which nitric oxide has been shown previously to modulate transcription. Analysis of vinculin mRNA distribution in mechanically stimulated muscle fibers shows that the mRNA is highly concentrated at myotendinous junctions, which supports the hypothesis that myotendinous junctions are distinct domains in which the expression of cytoskeletal proteins is modulated by mechanical stimuli through a nitric oxide and protein kinase G-dependent pathway.

Mechanical stimuli can cause changes in muscle mass and structure which indicate that mechanisms exist for transducing mechanical stimuli into signals that influence gene expression. Myotendinous junctions show adaptations to modified muscle loading which suggest that these are transcriptionally distinct domains in muscle fibers that may experience local regulation of expression of structural proteins that are concentrated at these sites. Vinculin and talin are cytoskeletal proteins that are highly enriched at myotendinous junctions that we hypothesize to be subject to local transcriptional regulation. Our findings show that mechanical stimulation of muscle cells in vivo and in vitro causes an increase in the expression of vinculin and talin that is mediated by nitric oxide. Furthermore, nitric oxidestimulated increases in vinculin and talin expression occur through a protein kinase G-dependent pathway and therefore differ from other mechanisms through which nitric oxide has been shown previously to modulate transcription. Analysis of vinculin mRNA distribution in mechanically stimulated muscle fibers shows that the mRNA is highly concentrated at myotendinous junctions, which supports the hypothesis that myotendinous junctions are distinct domains in which the expression of cytoskeletal proteins is modulated by mechanical stimuli through a nitric oxide and protein kinase G-dependent pathway.
Adaptation of organisms to their environment relies largely on modifications in gene expression in response to changes in stimuli applied to constituent cells. The best understood mechanisms through which environmental signals produce changes in gene expression occur in systems in which the stimulus produces a change in the chemical environment surrounding the cell, for example through signaling information mediated by hormones, growth factors, or cytokines (e.g. Refs. 1 and 2). However, cells also respond to changes in their mechanical environment (3), so systems must exist for the transduction of mechanical information to chemical signals in the cell that in turn regulate the expression of specific genes.
Skeletal muscle cells, perhaps more than any other cell type, are profoundly responsive to their mechanical environment. For example, removal of normal mechanical stimuli from muscle cells in vivo can result in rapid changes in gene expression, enzyme activity, and protein synthesis and stability (4 -9).
Application of mechanical loads to myotubes in vitro yields similar cellular responses (10 -13). Although little is known of the mechanisms of mechanical signal transduction in muscle, previous investigations have provided evidence that soluble factors released by muscle cells experiencing loads in vitro may contribute to signaling an increase in protein synthesis (12). In addition, stretch-activated ion channels (14) could provide increases in the concentration of specific cytosolic ions that are capable of activating signaling pathways that can influence gene expression. However, the extent to which these systems of mechanical signal transduction may contribute to modifying the expression of specific proteins in muscle is unknown.
Myotendinous junctions (MTJs), 1 which are highly specialized sites of force transmission across the muscle cell membrane (15), are responsive to changes in their mechanical environment (16). Increased mechanical stimulation in vivo or in vitro causes an increased expression of structural proteins that are concentrated at these sites (13). The increased concentration at MTJs of the mRNA encoding proteins whose expression is increased at MTJs suggests that muscle cell nuclei near the MTJ preferentially transcribe those mRNAs.
The local regulation of mRNA and protein synthesis at MTJs that are experiencing modified loading suggests that mechanical signal transduction that regulates the expression of these molecules occurs at MTJs. Recently, neuronal nitric-oxide synthase (nNOS) was shown to be concentrated at MTJs (17), and both its activity and expression are positively regulated by mechanical loading of muscle in vivo and in vitro (18 -20). Because nitric oxide (NO) has been shown to modify gene expression in other systems (21)(22)(23)(24)(25), it is feasible, but untested, that increased [NO] that is generated during increased muscle loading could affect the expression of structural proteins. Furthermore, nitric oxide that is generated during increased muscle use has a short half-life (26), so its effects in muscle would be local and consistent with the expectation that gene expression in nuclei located near the MTJ would be most affected.
In this investigation, we test the hypothesis that NOS is a mechanical signal transducer in muscle that positively regulates the expression of structural proteins enriched at MTJs. We assay the effects of mechanical stimulation that are mediated by the NOS signaling pathway on the expression of talin and vinculin, which are cytoskeletal proteins that are highly enriched at MTJs and are links in a chain of proteins that couple thin filaments to integrin. Talin was selected for study because its expression is up-regulated by mechanical stimulation of muscle (13). Vinculin was selected because previous investigations have shown that it plays an important role in mechanical coupling in transmembrane assemblies of structural proteins (27).

EXPERIMENTAL PROCEDURES
In Vitro Mechanical Loading Protocol-C2C12 muscle cells were subjected to cyclic strains using a mechanical cell stimulator (Cell Kinetics, Providence, RI), which consists of a stainless steel plate containing wells in which the floor is a silastic membrane. Culture conditions were identical to those used previously (20). Myotubes were cyclically strained using a 6.7% mean deformation of the membrane and adherent cells. Five cycles of strain during a 20-s period followed by 10 s of no strain were applied to the cells, and then the strain cycle was repeated two more times, followed by a 30-min period of no strain. Mean strain rate was 3.4%⅐s Ϫ1 . The sequence was repeated 47 times, over the course of 24 h. Cultures were then inspected with an inverted microscope to confirm that the myotubes remained attached to the substratum. Control cultures were grown under identical conditions, but they were not subjected to loading.
Modulation of [NO] and cGMP-dependent Protein Kinase Activity in Vitro-NO concentration in vitro was modulated by the addition of NO donor or NOS inhibitors to cultures. In addition, potential NO-mediated events in target cells were blocked downstream of NO stimulation by inhibition of cGMP-dependent protein kinase (PKG) with KT5823. Cultures in which NOS activity was inhibited received 100 M N w -nitro-Larginine methyl ester (L-NAME) in 10% FBS in DMEM containing no phenol red immediately prior to the beginning of experimental loading protocols. Cultures not subjected to experimental inhibition of NOS activity were subjected to a media exchange at the same time as cultures receiving L-NAME.
Cultures that were stimulated with exogenous NO donor received 100 M S-nitroso-N-acetylpenicillamine (SNAP) in DMEM containing 10% FBS for 2 h prior to collection for analysis. PKG was inhibited in cultured myotubes by incubation in 2 M KT5823 for 2 h before analysis. Cells treated with both SNAP and KT5823 were incubated for 30 min with KT5823 only, followed by 2 h in SNAP and KT5823. NO concentration in culture media at the end of the experimental treatments was measured using previously described procedures (20).
In Vivo Mechanical Loading Protocol-Modified loading of rat hindlimb musculature in vivo was achieved using previously described procedures in which the hindlimb musculature experienced 10 days of unloading followed by 2 days of reloading by normal body weight (13,28). At the end of the period of unloading or reloading, animals were euthanized, and plantaris muscles were collected for analysis.
Modulation of [NO] in Vivo-Animals experiencing NOS inhibition during muscle reloading following unloading received water containing the NOS inhibitor L-NAME at 0.5 mg⅐ml Ϫ1 that was provided ad libitum starting 1 day prior to reloading. Control animals received untreated drinking water.
Northern Blots-RNA was isolated from whole plantaris muscles according to the technique of Chomczynski and Sacchi (29). Electrophoresis and hybridizations were performed as described previously (20). Transfer efficiency and uniformity of loading were checked by staining the membrane with methylene blue. Blots were hybridized with 32 P-labeled probes generated by random priming (Amersham Pharmacia Biotech). Following hybridization at 65°C, blots were washed with 0.05 M sodium phosphate, 0.75 M sodium chloride, 5 mM EDTA, and 0.1% SDS for 1 h and exposed to autoradiographic film.
Probes for Northern Hybridization-The following probes were used to hybridize Northern blots: 1) mouse talin cDNA (generous gift from Dr. D. J. G. Rees; University of Oxford, UK); 2) 28 S ribosomal subunit cDNA (30); 3) rat GAPDH cDNA (31); or 4) rat vinculin cDNA. Rat vinculin cDNA probe was synthesized by PCR of rat cDNA using the following primers: 1) 5Ј CTG GTG GAC GAG GCT AT 3Ј (upstream) and 2) 5Ј ATG TTT CCA GCC ACA GC 3Ј (downstream) which produce a 290-base pair product. The sequence of the primers was derived from the mouse cDNA sequence. 2 Briefly, RNA was isolated (21), and 1 g was used to make cDNA using random primers (10 ng/l) and Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc.) in 20 l at 42°C for 30 min. Following cDNA synthesis, 2 l of the reaction was used to perform PCR using TAQ polymerase (Promega), 2.5 mM MgCl 2 , 4% Me 2 SO, and 0.2 M dNTP (Amersham Pharmacia Biotech). Cycling conditions were 1 min at 94°C, 1 min at 56°C, and 1 min at 72°C for 35 cycles. Fifteen microliters of the PCR reaction was run to a 1% agarose gel, and the presence of the 290-base pair product was verified. The remaining 35 l of the reaction was placed on a G-50 microspin column (Amersham Pharmacia Biotech) to remove nucleotides and primers. The concentration of the purified reaction was determined, and 60 ng was used in a ligation reaction to pCR2.1 (Invitrogen).
In Situ Reverse Transcription Polymerase Chain Reaction of Single Muscle Fibers-Plantaris muscles were fixed in 2% paraformaldehyde in phosphate-buffered saline for at least 2 days at 6°C. Muscles were then rinsed in phosphate-buffered saline, and single muscle fibers were dissected from surrounding tissue, while keeping the MTJ region of the fibers intact. Fibers were treated with 2 mg/ml of trypsin at 37°C for 20 min before washing in sterile water followed by 1 min in 100% ethanol and air drying. Fibers were then incubated with DNase I for 4 h at 37°C, rinsed in water followed by 1 min in 100% ethanol and air-dried. The RT reaction was performed using SuperScript TM (Life Technologies, Inc.) according to the manufacturer's instructions. The PCR mixture contained 10 mM Tris-HCl, pH 8.3, 4.5 mM MgCl 2 , 200 M dNTPs, 0.1 unit/ml Taq, 16 M digoxigenin-11-2Ј-deoxyuridine-5Ј-triphosphate, 0.08% bovine serum albumin, and 1.0 M primers. The same primers were used for vinculin RT-PCR as were used for generating vinculin cDNA for Northern analysis. Primer specificity for rat muscle vinculin was confirmed by RT-PCR of rat muscle RNA using the same RT-PCR conditions as those used for in situ RT-PCR. PCR on isolated fibers was performed for 35 cycles (1 min 94°C, 1 min at 54°C, 1.5 min at 72°C; Ericomp Twinblock thermocycler). After completion of the PCR, the fibers were rinsed in ethanol and then in two changes of TBS (100 mM Tris-HCl, pH 7.6, 150 mM NaCl) for 10 min and then incubated in TBS containing 1% bovine serum albumin for 1 h followed by incubation for 1 h at room temperature in alkaline phosphatase-conjugated anti-digoxigenin diluted 1:700 diluted in TBS. After 3 washes in TBS, the sections were immersed in TBS at pH 9.5 and containing 5 mM MgCl 2 and 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium reagents to produce a precipitate at the reaction sites. Positive controls consisted of omitting DNase treatment so that myonuclei containing the targeted sequence would stain after performing a successful PCR reaction. Reverse transcriptase was omitted from negative controls.
Vinculin mRNA distribution was compared with the distribution of ␤-actin mRNA by in situ RT-PCR on single rat plantaris fibers using methods similar to those used for vinculin mRNA localization. The primers used to generate a ␤-actin 296-base pair product were 5Ј CGT TGA CAT CCG TAA AGA CCT CTA 3Ј (upper primer) and 5Ј TAA AAC GCA GCT CAG TAA CAG TCC G 3Ј (lower primer) (33). However, PCR on fibers used for ␤-actin localization used 30 cycles of PCR (94°C for 1 min followed by 58°C for 1 min and 72°C for 2 min).
mRNA Stability Assay-The possibility that SNAP treatments may reduce the rate of mRNA degradation rather than increasing transcription was tested by assaying for the rate of decline in concentration of vinculin and talin mRNA in cells treated with an inhibitor of transcription. C2C12 myotube cultures were incubated with 5 g/ml actinomycin D in 10% FBS in DMEM for 0, 2, 8, or 24 h. The cells were then collected for RNA isolation and Northern analysis for changes in vinculin and talin mRNA concentrations.
Densitometry-Relative mRNA concentrations were determined by scanning densitometry (Alpha Innotec) and expressed in arbitrary units. Comparisons were made only between those experimental and control samples that were run in the same gel and probed in the same autoradiograph. The significance of differences between the concentration of any given protein or mRNA in loaded and control samples was tested by the Mann-Whitney test, with confidence limit set at p Ͻ 0.05.

Mechanical Stimulation Increases Talin and Vinculin Expression in Myotubes via an NO-dependent Mechanism-
C2C12 myotubes subjected to mechanical stimulation by periodic cyclic loading in vitro for 24 h showed large increases in the concentration of both talin and vinculin mRNA by approximately 3-and 2.5-fold, respectively, compared with control myotubes not subjected to stimulation (Fig. 1). The presence of the NOS inhibitor L-NAME prevented the increases in talin and vinculin mRNA concentrations caused by mechanical stimulation of myotubes in vitro (Fig. 1). Measurement of NO concentration in the culture media of mechanically stimulated myotubes and unstimulated controls showed that stimulation nearly doubled NO production by myotubes, from 0.5 to 0.9 pmol NO/mg/min (Ϯ0.04; n ϭ 4).  (Table I). Inhibition of cGMP-dependent protein kinase with KT5823 prevented the increase in vinculin or talin mRNA stimulated by SNAP (Fig. 2). PKG inhibition by KT5823 in the presence of SNAP reduced vinculin mRNA concentration below that of control cultures that were not stimulated by SNAP. Application of KT5823 alone did not significantly decrease vinculin or talin mRNA concentrations (Fig. 2) which indicates that the KT5823 inhibition of the SNAP-stimulated increase of talin and vinculin mRNA does not occur through inhibition of a NO-independent positive regulator of their expression. However, KT5823 application to myotubes in the absence of SNAP resulted in a significant increase in vinculin mRNA concentration, which indicates that there must be other NO-independent, PKG-modulated controls on vinculin expression that remain unidentified. Increased Talin and Vinculin Expression during Muscle Loading Is Inhibited by NOS Inhibition-Increased loading of plantaris muscles by 2 days of normal weight-bearing following a 10-day period of unloading produced an increase in vinculin mRNA concentration to 2.6 times control and in talin mRNA to 1.8 times control concentrations ( Fig. 6; Table II). The increased concentrations of both talin and vinculin mRNA during increased muscle loading was greatly attenuated in animals that received L-NAME in their drinking water during the reloading period (Fig. 6). GAPDH mRNA concentration as a fraction of total mRNA decreased during muscle reloading, but L-NAME had no detectable influence on the effect of reloading on GAPDH mRNA concentration. Thus, NOS inhibition does not affect the changes in concentration of all mRNAs that are caused by muscle reloading.

Exogenous NO Stimulates Vinculin and Talin Expression via PKG-dependent Mechanism-Addition
Vinculin mRNA was primarily concentrated at MTJs in muscles experiencing increased loading (Fig. 7), as has been shown previously for talin mRNA (13). Negative control fibers showed no labeling at the MTJ following the in situ RT-PCR procedure. Positive control fibers showed strong labeling of nuclei following the in situ RT-PCR procedure (Fig. 7). In situ RT-PCR of ␤-actin mRNA showed no detectable differences in the concentration of that mRNA at MTJ and non-MTJ regions of the fibers on reloaded muscles (Fig. 8). Thus, the relatively high concentration of vinculin and talin mRNA at MTJs does not occur for all mRNA. RT-PCR showed that the primers and RT-PCR conditions used for in situ RT-PCR yielded products only of the predicted size for vinculin or ␤-actin (Fig. 9), indicating that the reaction products seen by in situ RT-PCR were specific for vinculin or ␤-actin mRNA. DISCUSSION Our findings show that the increases in talin and vinculin mRNAs that result from mechanical stimulation of muscle fibers are mediated by NO. Thus, NOS acts as a mechanical signal transducer in muscle cells and may couple modifications in the mechanical environment to changes in gene expression. Several previous investigations have reported that increased expression of structural proteins in skeletal muscle occurs when muscle is subjected to increased mechanical stimulation (9,13,34,35), but little is known of the mechanisms by which those stimuli from the mechanical environment are transduced into chemical information that can influence gene expression. Although NO synthesis has been shown to be positively regulated by mechanical stimulation of cells in bone (36), endothe- lial cells (37,38), and skeletal muscle (18 -20), and NO is a famously pleuripotent signaling molecule (26,39), the present findings are the first to show that it functions in modifying muscle gene expression in response to modified use.
The hypothesis that NOS could function as a mechanical signal transducer in skeletal muscle was supported by several previous investigations. First, nNOS is highly concentrated at MTJs (17), which are specialized for force transduction across the muscle cell membrane (15). This localization would place nNOS at an optimal site for sensing changes in the mechanical environment. Furthermore, increased skeletal muscle loading causes an increase both in nNOS expression (20) and activity (18,20), and NO has been shown to function as a signaling molecule in several cell types (39). Finally, the short half-life of NO (26) suggests that it would locally regulate changes in transcriptional activity. This local regulatory effect may have particular functional significance in skeletal muscle where the transcriptional activities of individual nuclei located in a single muscle fiber can function independently. This independent regulation of transcriptional activity of individual myonuclei in a single cell has been best demonstrated for myonuclei located near the neuromuscular junction in which the increased expression of acetylcholine receptor subunits can be regulated  Fig. 2, which is representative of the results from three experiments. Mean value obtained by densitometry for untreated cells in 10% FBS in DMEM was set at 100 and used to normalize other samples in each data set. A, muscles collected from animals after 10 days unloading by hindlimb suspension. B, muscles collected from animals after 10 days unloading by hindlimb suspension, followed by 2 days of reloading by normal weight-bearing while receiving drinking water ad libitum. C, muscles collected from animals after 10 days unloading by hindlimb suspension, followed by 2 days of reloading by normal weight-bearing while receiving drinking water containing L-NAME ad libitum.

TABLE II
Relative density of signals in Northern analysis of select mRNAs in plantaris muscles of animals experiencing modified muscle loading The Northern blot used for densitometry is shown in Fig. 6 and is representative of the results from three experiments. Mean value obtained by densitometry for unloaded muscle samples was set at 100 and used to normalize other samples in each data set. independently of non-neuromuscular junction nuclei in the same cell (40,41). The finding reported here that NO transduces mechanical signals to influence the expression of talin and vinculin and that NOS, NO, and the mRNAs of talin and vinculin are most concentrated at MTJs indicates that MTJs may also be a distinct nuclear domain in muscle cells in which NO plays a role in the local modulation of transcription. Other evidence indicating that MTJs may be distinct nuclear domains in muscle includes the findings that the mRNA for myosin heavy chain (34) and talin (13) increase in concentration at the MTJ during modified muscle use. The observations reported here indicate that NO-mediated modulation of talin and vinculin mRNA concentrations occurs primarily by influencing the rate of transcription, although the possibility that there is some NO-mediated influence on talin and vinculin mRNA stability cannot be excluded entirely. This conclusion is supported by the finding that NO stimulation of myotubes for 24 h caused an approximately 170% increase in vinculin mRNA and 260% increase in talin mRNA, although the half-lives of those mRNAs were approximately 27 and 12 h, respectively. If NO were causing the observed increases in mRNA concentrations reported here solely through RNA stabilization, the maximum increase that would be observed in 24 h stimulation would be approximately 44% for vinculin and 93% for talin, which would be insufficient to explain the majority of the NO-mediated increase in mRNA concentrations.
Although there are several mechanisms through which NO can function as a signaling molecule, to our knowledge the only previously identified mechanism through which it has been shown to influence transcription is through structural modification of transcription factors (23)(24)(25). For example, c-Fos, c-Jun, and NF-B binding affinities for target DNA sequences are all influenced by NO (23)(24)(25). Current evidence indicates that NO modifies transcription factors by S-nitrosylation of cysteines located in or near DNA binding domains (23)(24)(25) and that the NO-mediated effects are independent of guanylate cyclase activation (21). The present findings show that NO modulation of talin and vinculin expression is blocked by inhibition of PKG and therefore NO is not influencing transcription of talin and vinculin through direct interaction with transcription factors but rather through a cGMP-and PKG-dependent pathway. NO modulation of talin and vinculin expression also differs from the mechanisms through which it modulates the expression of genes regulated by AP-1 or NF-B (21)(22)(23)(24)(25) in that NO is a positive modulator of talin and vinculin but negatively regulates genes whose expression is influenced by AP-1 or NF-B binding.
Other mechanisms may also exist at MTJs for mechanical signal transduction that may influence the expression of genes for cytoskeletal proteins. The integrin-associated complex of proteins is highly concentrated at MTJs (15) and provides a prominent candidate system for this function because integrins have also been shown to function in signal transduction. Mechanical loads placed on integrins and associated proteins influence the assembly of complexes of structural proteins at the loading site (42)(43)(44) and activate kinases concentrated at those sites (42). Although integrin loading per se has not been shown to regulate the expression of genes encoding cytoskeletal proteins, integrin binding of extracellular matrix molecules can dramatically influence gene expression (e.g. Refs. 32, 45, and 46). The co-distribution of nNOS and the integrin complex and the role of both proteins in influencing the response of cells to changes in their mechanical environment support the possibility that future investigations will show synergistic interactions between the two systems when muscle cells respond to changes in their mechanical environment.