- Email: [email protected]
Stretch activation of GTP-binding proteins in C2C12 myoblasts
Experimental Cell Research 292 (2004) 265 – 273 www.elsevier.com/locate/yexcr
Stretch activation of GTP-binding proteins in C2C12 myoblasts Craig B. Clark, a Nathan L. McKnight, a,b and John A. Frangos a,b,* a
Department of Bioengineering, University of California San Diego, La Jolla, CA 92093-0142, USA b La Jolla Bioengineering Institute, La Jolla, CA 92037, USA Received 31 December 2002, revised version received 24 September 2003
Abstract Mechanical stimulation has been proposed as a fundamental determinant of muscle physiology. The mechanotransduction of strain and strain rate in C2C12 myoblasts were investigated utilizing a radiolabeled GTP analogue to detect stretch-induced GTP-binding protein activation. Cyclic uniaxial strains of 10% and 20% at a strain rate of 20% s 1 rapidly (within 1 min) activated a 25-kDa GTPase (183 F 17% and 186 F 19%, respectively), while 2% strain failed to elicit a response (109 F 11%) relative to controls. One, five, and sixty cycles of 10% strain elicited 187 F 20%, 183 F 17%, and 276 F 38% increases in activation. A single 10% stretch at 20% s 1, but not 0.3% s 1, resulted in activation. Insulin activated the same 25-kDa band in a dose-dependent manner. Western blot analysis revealed a panel of GTP-binding proteins in C2C12 myoblasts, and tentatively identified the 25-kDa GTPase as rab5. In separate experiments, a 40-kDa protein tentatively identified as Gai was activated (240 F 16%) by 10% strain at 1 Hz for 15 min. These results demonstrate the rapid activation of GTP-binding proteins by mechanical strain in myoblasts in both a strain magnitude- and strain rate-dependent manner. D 2003 Elsevier Inc. All rights reserved. Keywords: C2C12; Myoblast; Uniaxial strain; Strain rate; Mechanotransduction; Photoaffinity
Introduction Mechanical loading and deformation play an important role in the physiology of a variety of tissues, with extensive research in both animal and cell culture models identifying a rapidly expanding list of cellular responses. Mechanotransduction—translation of a mechanical stimulus into an intracellular biochemical processes—has been the subject of intensive study, yet many questions remain as to the underlying cellular mechanisms and the physical stimuli which drive them. Muscle provides an ideal tissue for the study of mechanotransduction, as it experiences a wide range of strains during normal use. In vivo, muscle is affected by a variety of stimuli including neural interactions, membrane depolarization, intracellular calcium flux, metabolite transport or depletion, hormonal influences, as well as the stresses and mechanical strains due to both external loading and myosin-generated tension. While each of these stimuli influences muscle performance and adaptation, me-
* Corresponding author. La Jolla Bioengineering Institute, 505 Coast Boulevard South, La Jolla, CA 92037-4616. Fax: +1-858-456-7540. E-mail address: [email protected] (J.A. Frangos). 0014-4827/$ - see front matter D 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.yexcr.2003.09.017
chanical strain has been proposed as a fundamental determinant of muscle physiology [16]. The strains under loading (strain being a normalized measure of stretch; where Strain = Deformation / Initial Length) and the cyclic strains during locomotion impose a spectrum of mechanical stimuli to direct the muscular phenotype, with strains during normal locomotory movements typically F10% of resting length, at strain rates of 0– 700% [17,18,23,47]. Strain rate (the rate at which the deformation is applied; where Strain Rate = Strain / Duration) varies by orders of magnitude and may therefore be an important parameter in modulating the mechanoresponse of cells and tissues. A host of candidates have been proposed for the mechanotransduction of physical forces including adhesion molecules, kinases, cytoskeletal elements, ion channels, and guanine nucleotide-binding regulatory proteins (GTP-binding proteins) [3,11,22]. The diverse family of GTP-binding proteins play a central role in signal transduction [27,36], and are also sensitive to physical forces, being activated by both fluid shear and mechanical strain [3,19,20]. GTP-binding proteins therefore provide an ideal candidate for the study of the mechanical stimulation of skeletal muscle. In the present study, the mechanotransduction of strain and the sensitivity to strain rate were investigated using
266
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
C2C12 myoblasts, a cell line morphologically and functionally similar to satellite cells [34]. A uniaxial stretch device [6] applied defined cyclic strain to C2C12 myoblast cultures, while incorporation of a radiolabeled GTP analogue allowed specific and sensitive detection of GTPbinding protein activation. Stretch activation was detected in GTP-binding proteins, tentatively identified as the small GTPase rab5 and a heterotrimeric Gai subunit. Activation was found to be both strain magnitude- and strain ratedependent.
Materials and methods Cell culture and uniaxial stretch C2C12 murine skeletal muscle myoblasts (ATCC) were kept as subconfluent cultures in DMEM (Irvine Scientific) supplemented with 10% fetal calf serum (FCS; Biocell). For all stretch and agonist experiments, cells were seeded at 20,000 cells cm 2 and the experiments run 48 h after seeding, at which time the cells had formed a confluent monolayer but had not begun fusing into myotubes. Standard 35-mm-diameter-tissue culture dishes (Falcon) seeded at 20,000 cells cm 2 were used for agonist and inhibitor studies. Cultures were rinsed twice with serum-free media and switched to reduced serum (1%) media 8 h before the start of experiments. Serum content was reduced to provide quiescent cultures in terms of minimizing the basal GTPbinding protein activity, allowing improved signal-to-noise under stimulated (agonist or mechanical stretch) conditions. No visible changes in cell morphology were noted after the 8 h in 1% serum, with all cultures maintained as a confluent layer of mononucleated cells. Cells cultured on compliant membranes were subjected to cyclic uniaxial strain in a strain device described in Clark et al. [6]. Briefly, cells were cultured within a sterile enclosure on 4.8 3.7 0.010 cm in thick silicone membranes (Specialty Manufacturing) which were held at each end with stainless steel clamps. Membranes were coated before cell seeding with type I collagen (Sigma) to a final concentration of 6 Ag cm 2. A stepper motor (Anaheim Automation) applied cyclic strain to four membranes simultaneously, with a triangular displacement waveform featuring a constant rate of strain for both stretch and relaxation, starting from an initial membrane length of 48.3 mm. Controls for all stretch experiments consisted of cells seeded on membranes and treated identically, but not subjected to stretch. Experiments were carried out in a temperature-controlled hood at 37jC, the cultures flushed with humidified and sterile filtered 5% CO2/95% air. Photolabeling of GTP-binding proteins The 32P-labeled, nonhydrolyzable, photoreactive GTP analogue, azidoanalide-GTP (AAGTP, specific activity
25 – 50 mCi mmol 1, Affinity Labeling Technologies) was utilized to detect GTP-binding protein activation [30]. Briefly, photoaffinity experiments were performed in a darkened room utilizing a safelight, with all equipment and reagents maintained at 37jC within the temperaturecontrolled hood for the duration of the experiments. The nucleotide was suspended in serum-free DMEM or PBS along with digitonin (15 AM, Calbiochem) in 150-Al total volume and 10– 15 ACi AAGTP per sample. Digitonin is a detergent that permeabilized the cell membrane and facilitated loading of the AAGTP label into the cells. As such, the perturbation of the lipid bilayer may nonspecifically activate certain intracellular signaling pathways, and therefore identically treated sham (not stretched) controls were utilized to account for any nonspecific activation effects. Media were aspirated from the cell cultures, the AAGTP label added for the selected incubation period, gently rinsed with the appropriate buffer (serum-free DMEM or PBS), and the bout of stretch applied. Rinsing and aspiration of the excess media and label before applying stretch minimize any potential effects of increased AAGTP uptake due to stretch, as well as eliminate potential fluid shear effects due to fluid movement. Following stretch, a UV lamp (UVP, model UVG-54; 254 nm wavelength) was placed over the stretch chamber approximately 10 cm above the membranes, and the cells subjected to UV light for 60 s. Ice-cold PBS with 4 mM dithiothreitol (DTT, Calbiochem) was added after the initiation of UV to reduce nonspecific cross-linking of the photolabel. Samples were immediately harvested in 2 electrophoresis sample buffer (62.5 mmol Tris –HCl, pH 6.8, 2% w/v SDS, 10% glycerol, 20% v/v h-mercaptoethanol, and 0.1% w/v bromophenol blue, Sigma), boiled for 5 min, sheared with a 21 gauge needle, and stored at 80jC until analyzed. In select cultures, insulin (Sigma) was added at various concentrations. Following preincubation in serum-free media, media were aspirated and the AAGTP label incubated for 15 min. Excess label was then gently rinsed with serum-free media and aspirated, and insulin (10 and 100 nM, 1 AM) in DMEM added for a subsequent 5-min incubation. Dishes were then transferred onto ice, subjected to UV, and harvested as described above. For the serum response experiments, the AAGTP label in DMEM was added along with the appropriate serum concentration (0– 20%) without digitonin for 60-min incubation before UV. Electrophoresis, autoradiography, and Western blot analysis Equal volumes of sample were separated by SDS-PAGE on 12% acrylamide running/5% stacking gels. Gels were stained with Coomasie brilliant blue (Sigma) to verify equal loading of protein, then dried and film (Fuji) exposed for 2– 14 days for autoradiography. Radiographs were scanned and
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
267
band intensities analyzed (Quantity One, BioRad), with values reported as a percentage of controls from within each experiment. Western blot analysis was performed for selected experiments, with AAGTP-labeled samples harvested in 2 sample buffer and run on 12% gels as described above. Gels were then equilibrated in transfer buffer (192 mM glycine, 25 mM Tris, 20% methanol) for 15 min and electrotransferred onto PVDF membranes (ImmobilonP, Millipore). Following transfer, the membranes were rinsed, air dried, and film (Fuji) exposed for 2 –14 days. Following autoradiography, the membranes were cut into strips corresponding to lanes of the gel and incubated overnight at 4jC in blocking buffer [immunoblot buffer (IB—10 mM Tris pH 7.4, 100 mM NaCl, 0.1% tween-20) with 5% nonfat milk]. Primary rabbit polyclonal antibodies against rhoA, ras, Gai/o/t/z, Gas, Gaq/11, (Santa Cruz), and rab4, rab5 (StressGen) were incubated at 1:1000 dilution in blocking buffer for 1 h at room temperature, rinsed six times in IB, followed by 30 min incubation in horseradish peroxidase-conjugated secondary antibody (anti-rabbit, New England Biolabs) at 1:1000 dilution in blocking buffer, and six additional rinses in IB. Bands were visualized using the SuperSignal (Pierce) chemiluminescent kit per manufacturer’s instructions. Autoradiographs and Western blots from the same PVDF membrane were then overlaid and aligned for band comparison and identification. Statistics All results are presented as the mean F SE, normalized with respect to dedicated controls from each experiment. Data were first evaluated by analysis of variance (ANOVA), followed by the Student – Newman– Keuls (SNK) test for multiple comparisons, with P < 0.05 considered statistically significant.
Results Photoaffinity labeling of GTP-binding proteins As a means of activating and identifying a spectrum of GTP-binding proteins, fetal calf serum was added to serumstarved cultures along with AAGTP for 1 h, with a panel of GTP-binding proteins being revealed (Fig. 1). Activation with 20% serum relative to the BSA (serum-free) controls ranged from 380 F 9% (mw 33 kDa) to 165 F 4% (mw 38 kDa), with the protein activation levels appearing to cluster into two distinct groups. As an additional control, 50-AM digitonin was added for 3 min to BSA controls following the 1-h label incubation period to investigate nonspecific activation. Average intensity after treatment with digitonin for the six bands marked by arrows in Fig. 1A was 99 F 13% relative to non-digitonin-treated samples (data not
Fig. 1. Long-term (60 min) incubation with AAGTP revealed a panel of GTP-binding proteins, while serum activated select isoforms in a dosedependent manner. (A) Typical radiograph is shown for samples incubated with 0 – 20% serum. Note the two bands on right are BSA (0% serum) controls, with digitonin (50 AM) added after label incubation. (B) Activation levels are plotted normalized to BSA controls, with apparent molecular weights (mw) corresponding to the radiograph bands next to each plot. Mean F SE for n = 3 (error bars shown on top trace only for clarity).
shown), indicating that the use of digitonin as a permeabilizing agent had negligible effect on GTP-binding protein activation. A series of control experiments were performed to demonstrate specificity of the AAGTP label to G proteins. The nonhydrolyzable GDP analogue GDPhS irreversibly binds to GTP-binding proteins, inhibiting subsequent nucleotide exchange and photolabel incorporation. Preincubation with 300 or 900 AM GDPhS for 8 h significantly
268
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
reduced subsequent AAGTP binding in the 25-kDa band to 56 F 9% and 34 F 8% of controls, respectively (Fig. 2A). The addition of non-radiolabeled GTP at 50 or 200 AM during AAGTP incubation competitively reduced label
Fig. 3. Stimulation with insulin for 5 min following AAGTP loading activated the 25-kDa GTPase protein in a dose-dependent manner. Typical radiograph is shown for the three insulin concentrations (upper panel). Radiograph band intensities are plotted, normalized to control (lower panel). Mean F SE for n = 4. GTPase activity (25-kDa) correlated significantly (P < 0.001) with insulin concentration as determined by Spearman rank correlation test.
incorporation to 64 F 12% and 49 F 9% of controls, respectively (Fig. 2B). Similar values of inhibition were seen for other molecular weight GTP-binding proteins (data not shown). Insulin added for 5 min following AAGTP incubation elicited a dose-dependent activation of the 25-kDa band over the range of concentrations tested (10 and 100 nM, and 1 AM) (Fig. 3), with 103 F 6%, 153 F 8%, and 192 F 17% activation relative to control, respectively.
Fig. 2. (A) Preincubation with 300 or 900 AM GDPhS (8 h) inhibited nucleotide exchange and binding of the radiolabeled probe to 56 F 9% and 34 F 8% of control, respectively. Typical radiograph showing the 38-, 33-, and 25-kDa bands (upper panel). Plot of radiograph band intensities for 25-kDa (typical for all bands) normalized to control (lower panel). Mean F SE for n = 4, *P < 0.01 vs. control. (B) Addition of unlabeled 50 or 200 AM GTP during AAGTP incubation competitively reduced binding of radiolabeled probe in C2C12 myoblasts to 64 F 12% and 49 F 9% of control, respectively. Typical radiograph showing the 25-kDa band (upper panel). Plot of radiograph band intensities, normalized to control (lower panel). Mean F SE for n = 3 (*P < 0.01 for all comparisons).
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
269
rate constant. Cyclic strain of 10% at a strain rate of 20% s 1 (1 Hz) was applied for 1, 5, and 60 cycles, with UV irradiation applied 60 s after the initiation of stretch in each experiment. This treatment activated the 25-kDa protein, with 1, 5, and 60 cycles of stretch resulting in increased AAGTP binding of 187 F 20%, 183 F 17%, and 276 F 38%, respectively, relative to controls (Fig. 5). Finally, the strain magnitude and number of cycles were held constant, as a single cycle of 10% strain applied over 1 min (‘slow’ strain rate = 0.3% s 1) was compared to a single cycle of 10% strain over 1 s (‘fast’ strain rate = 20% s 1). The single ‘fast’ cycle activation was 187 F 20% of control, while the single ‘slow’ cycle was not significantly different from controls (107 F 13%) (Fig. 6). Activation of a heterotrimeric G protein by 15 min of stretch Longer-term stretch experiments utilized AAGTP label suspended in CF-DPBS followed by 15 min of continuous
Fig. 4. Application of five cycles of 10% or 20% uniaxial strain activated the 25-kDa protein in myoblasts. (A) Displacement waveforms for the applied uniaxial strain are shown, all applied at a strain rate of 20% s 1. (B) Typical radiograph showing the 25-kDa band activation after the indicated stretch regime. (C) Radiograph band intensities are plotted, normalized to sham control for 2%, 10%, and 20% strain. Application of five cycles of 10% or 20% uniaxial strain activated the 25-kDa band 183 F 17% and 186 F 19%, respectively, while activation by 60 cycles of 2% strain (109 F 11%) was not different from control. Mean F SE for n = 6 (*P < 0.05 vs. 2% and control).
Activation of a small GTPase by 1 min of cyclic stretch Activation of GTP-binding proteins by mechanical strain was investigated utilizing C2C12 myoblasts cultured on silicone membranes, and subjected to uniaxial strain applied at various strain magnitudes, strain rates, and number of cycles. A set of experiments was performed across a range of strain magnitudes, all applied at the same strain rate (20% s 1). Five cycles of stretch at strains of 10% and 20% elicited activation of the 25kDa protein 183 F 17% and 186 F 19% relative to controls, respectively, while activation by 60 cycles of 2% strain was not significantly different from control values (109 F 11%, Fig. 4), suggesting a threshold of activation by strain. In the next set of experiments, the number of cycles was varied while holding the strain magnitude and strain
Fig. 5. Application of 1, 5, or 60 cycles of 10% uniaxial strain at a strain rate of 20% s 1 resulted in increasing GTPase protein activation. (A) Typical radiograph is shown, with the 25-kDa band activated by the indicated number of uniaxial strain cycles. (B) Radiograph band intensities are plotted, normalized to sham control, with 1, 5, or 60 cycles of 10% uniaxial strain at 20% s 1 activating the 25-kDa band 187 F 20%, 183 F 17%, and 276 F 38% relative to control, respectively. Mean F SE for n = 6 (*P < 0.05 vs. sham; **P < 0.01 vs. sham, P < 0.05 vs. 1 or 5 cycles).
270
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
at their appropriate molecular weights in the C2C12 myoblasts. Overlay and alignment of the Western blots and autoradiograph revealed alignment of bands labeled by both the antibody and AAGTP photoaffinity probe. The three heterotrimeric subunits matched the three radiograph bands between 38 and 45 kDa, and rab5 aligned with the prominent 25-kDa band. A faint band was also seen on the radiograph that aligns with the rhoA immunoblot. While this method was sufficient to tentatively identify the 25-kDa band as rab5 and the 40-kDa band as Gai, it can only positively exclude a specific protein from being associated with a particular autoradiograph band. Other GTP-binding proteins which co-mi-
Fig. 6. Stimulation of GTPase protein activity was sensitive to strain rate. (A) Stretch protocol of 10% strain applied over 1 s (‘fast’ 20% s 1—solid line) or 1 min (‘slow’ 0.3% s 1—dashed line) is shown. Application of a single ‘fast’ cycle resulted in GTPase protein activation, while one cycle of ‘slow’ strain did not significantly increase activation. A typical radiograph is shown in Fig. 5A. (B) Relative band intensities are plotted, normalized to sham control, with ‘fast’ and ‘slow’ activated 187% F 20% and 107 F 13%, respectively. Mean F SE for n = 6 (*P < 0.05 for all comparisons).
cyclic stretch. The application of 900 cycles of 10% strain at 20% s 1 activated a 40-kDa band 240 F 16% relative to sham controls (Fig. 7). This protocol of label incubation and stretch also activated the 25-kDa band (198 F 7% relative to control) similar to the previous 1-min protocols in DMEM. Identification of GTP-binding proteins Electroblotted membranes were first exposed for autoradiography, and then incubated with primary antibodies against rab4, rab5, rhoA, Gaq/11, Gai, and GaS (Fig. 8). The Western blot confirmed the presence of each isoform
Fig. 7. Loading of AAGTP in CF-DPBS followed by 900 cycles of 10% strain (20% s 1) indicated activation of both 40- and 25-kDa GTP-binding proteins. (A) Typical radiograph showing the 40- and 25-kDa bands for control (C) vs. stretch (S). (B) Plot of radiograph band intensities, normalized to control. Mean F SE for n = 6 (*P < 0.05 vs. sham control).
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
Fig. 8. Autoradiography and Western blot analysis identified GTP-binding protein isoforms in C2C12 myoblasts. Western blot against rab4, rab5, rhoA, Gaq/11, Gai, and GaS revealed the presence of these isoforms (upper panel). In the lower panel, the Western blot image is false colored (white) and overlaid onto an autoradiograph from the same transfer membrane. Alignment correlated bands labeled by both the antibody and the AAGTP photoaffinity probe. Rab5 aligned with the prominent 25-kDa band, and the three heterotrimeric subunits matched the three radiograph bands between 38- and 45-kDa. Note that the membrane was cut in the middle of lane 2 so that rab4 and rab5 each stained half of the lane to demonstrate the difference in molecular weight.
grate with rab5 and Gai may be responsible for the activation noted in the experiments above.
Discussion The present study demonstrated the novel activation of a 25-kDa GTPase and a 40-kDa heterotrimeric Ga subunit by mechanical stretch in C2C12 myoblasts. Activation of the small GTPase was further shown to be dependent on the magnitude of strain, the number of cycles, and the rate at which strains were applied. Application of uniaxial strain to cultures of unfused C2C12 cells provides a model for both myoblasts during development as well as mononucleated satellite cells which reside along the muscle fibers beneath the basal lamina [33,34]. Mechanical strain influences muscle at the level of these mononucleated cells, as strain imposed during development and use drives cell proliferation and fusion into myofibers [9,10,34]. In adult muscle, satellite cells are activated during muscle stretch or damage, gaining the ability to fuse to each other to form new fibers or to fuse with existing fibers [9,10,34]. Coupled with the recent work on stretch-induced satellite cell activation
271
[40,41], the activation of GTP-binding proteins by stretch demonstrated in this study supports the hypothesis that myoblasts sense and respond to mechanical stimuli. When the strain rate was held at a constant 20% s 1, a 25-kDa GTPase was activated at strains of 10% and 20%, but not at strains of 2%, suggesting a threshold of straindependent activation. Application of one or five cycles of 10% strain at 20% s 1 produced equivalent activation of the 25-kDa GTPase, with 60 cycles eliciting a further increase in activation, indicating not only a dose response with the number of cycles, but highlighting the extreme sensitivity of GTP-binding protein activation to even a single cycle of strain. Comparing the GTP-binding protein activation by a single ‘fast’ (one cycle of 10% strain applied over one second) to a single ‘slow’ (10% strain over 1 min) cycle confirmed that cells can sense and differentially respond to the rate at which strains are applied (strain rate). This observation suggests that activation is due to dynamic straining of the cells rather than simple tension, as the ‘slow’ protocol applied the deformation to the cells for much greater duration than the ‘fast’ protocol, yet did not stimulate GTP-binding protein activity. Previous cell culture studies have addressed the strain rate response of cells, as activation of heterotrimeric G proteins in cardiac fibroblasts [20] and mitogen-activated protein kinase phosphorylation in smooth muscle [28] were shown to be strain rate sensitive. Modulation by strain rate was seen during myotube formation in vitro, as extremely slow elongation (0.5% h 1) induced alignment along the axis of stretch [42] while cyclic stretch at higher strain rates (20% s 1) aligned the myotubes perpendicular to the stretch axis [43]. In contrast, strain rate did not affect the orientation response of C2C12 myotubes to stretch [7]. However, the rates imposed in the latter study (2.26% day 1 and 1.77% day 1) were designed to mimic murine long bone growth rates and are significantly different from those utilized in the present work. The strains and strain rates applied in the current study were chosen to be representative of the physiologic range of skeletal muscle during normal locomotion. Satellite cells reside between the basal lamina and the sarcolemma of myofibers and participate in the hypertrophy of muscle during normal growth and after myofiber injury. While a variety of factors have been identified which can stimulate growth of cultured muscle [45], the activation of satellite cells after skeletal muscle injury is associated with hepatocyte growth factor (HGH) [39] through autocrine release of HGH induced by cyclic stretch [41]. It was further demonstrated this HGH release is dependent on stretch-stimulated nitric-oxide synthase (NOS) activity [40]. This provides one example where stretch triggers specific mechano-sensitive signal transduction pathways in the satellite cells, as opposed to inducing a response (in this case HGH release) via a nonspecific mechanism such as stretch-induced injury and permeabilization.
272
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
Both heterotrimeric G proteins and the small GTPase proteins have been implicated in numerous aspects of muscle physiology. In culture, myoblast proliferation, differentiation, and fusion were Pertussis toxin (PTx)-sensitive in the BC3H1 myoblast cell line [24], indicating the involvement of Gai. Federov et al. [14] indicated the Ghg subunits dissociated from the activated Gai also participate in FGF-mediated myoblast proliferation process. For the small GTPases, constitutively activated ras inhibited skeletal muscle differentiation [31], while myoblasts expressing a constitutively activated rac1 fused yet the resulting myotubes failed to form ordered sarcomeres [15]. The small GTPase rho is also active in muscle differentiation [46], and the expression of MyoD, myogenin, and myocyte enhancer binding factor 2 (MEF2) were enhanced by rho activity and suppressed by rho inhibition [5,38]. Evidence of mechanical activation of heterotrimeric G proteins in skeletal muscle comes indirectly from studies such as the PTx inhibition of stretch-induced cyclooxygenase activity, PGF2a production, phospholipase D activation, and cell growth in differentiated avian myoblasts [44]. The present study demonstrates the rapid activation of a 40-kDa G protein by mechanical strain in skeletal muscle. Both insulin treatment and stretch of C2C12 myoblasts activated a 25-kDa GTPase tentatively identified as rab5. A major function of insulin is to stimulate glucose uptake and metabolism, with skeletal muscle the principal tissue site for insulin-mediated glucose uptake [2]. Myoblasts and myotubes in culture also respond to insulin [29]. In skeletal muscle, insulin stimulated ras [26], as well as rho and rac [32], though their lower molecular weights (20 – 21 kDa) exclude these proteins as candidates for the insulin-stimulated 25-kDa band seen in the present study. The rab family of small GTPases has a size range (24 – 25 kDa) which correlates with the 25-kDa band seen in Fig. 8, and has demonstrated insulin sensitivity, with the isoforms rab4 [32,35], rab5 [8], and rab5B [1] all activated by insulin. The exocytosis of vesicle-bound glucose transporters to the plasma membrane is directed by these rab isoforms [1], and although both insulin [4,35] and mechanical activity in skeletal muscle [13,21] activate transporter mobilization, activation of rab by mechanical activity had not been previously demonstrated. Mechanical stretch represents a novel activation mechanism for the small GTPase, and supports the role of passive mechanical strain in activating fundamental processes within skeletal muscle cells. It also suggests a possible mechanism not only for crosstalk between receptor tyrosine kinases and GTP-binding proteins [12,32], but also the interaction between ligand- and stretchactivated events. The longer (15 min) stretch experiments revealed the activation of a 40-kDa band, with the Western blot alignment tentatively identifying the protein as a Gai isoform. Gai3 activation has been previously identified in HUVEC stimulated by fluid shear [19], while Gai1 was activated by mechanical strain in cardiac fibroblasts [20] indicating
possible isoform specificity utilized by different cell types. The antibody used in the present study is common to the Gai family of isoforms, and the specific isoform (Gai1, Gai2, or Gai3) involved in the C2C12 mechanoresponse remains to be identified. The use of calcium-free buffer rather than DMEM in these experiments allowed detection of the Ga subunit, presumably by altering the kinetics of GDP release or GTP/AAGTP binding, which are Ca2+ and Mg2+ sensitive [25,37]. While only two GTP-binding proteins were identified as responding to mechanical strain, additional GTP-binding protein isoforms may participate in the mechanoresponse, but may have a poor affinity for the AAGTP probe or are activated at levels too low to be detected by the assay under the experimental conditions utilized in this study. In summary, photoaffinity labeling of GTP-binding proteins in C2C12 myoblasts identified a panel of enzymes, and indicated the rapid activation by mechanical strain of proteins tentatively identified as rab5 and a heterotrimeric Gai subunit. Furthermore, GTP-binding protein activation was dependent not only on the magnitude of the strain, but on the rate at which the strain was applied.
Acknowledgments This study was supported in part by NIH Grant HL 40696. C.B. Clark expresses his appreciation to the Achievement Rewards for College Scientists and the Robert M. Golden Foundation for their support.
References [1] S. Bao, J. Zhu, W.T. Garvey, Cloning of rab GTPases expressed in human skeletal muscle: studies in insulin-resistant subjects, Horm. Metab. Res. 30 (1998) 656 – 662. [2] A.D. Baron, G. Brechtel, P. Wallace, S.V. Edelman, Rates and tissue sites of non-insulin- and insulin-mediated uptake in humans, Am. J. Physiol. 255 (1988) E764 – E769. [3] F. Berthiaume, J.A. Frangos, Flow-induced prostacyclin production is mediated by a pertussis toxin-sensitive G protein, FEBS Lett. 308 (1992) 277 – 279. [4] D.M. Calderhead, K. Kitagawa, G.E. Lienhard, G.W. Gould, Translocation of the brain-type glucose transporter largely accounts for insulin stimulation of glucose transport in BC3H-1 myocytes, Biochem. J. 269 (1990) 597 – 601. [5] G. Carnac, M. Primig, M. Kitzmann, P. Chafey, D. Tuil, N. Lamb, A. Fernandez, RhoA GTPase and serum response factor control selectively the expression of MyoD without affecting Myf5 in mouse myoblasts, Mol. Biol. Cell 9 (1998) 1891 – 1902. [6] C.B. Clark, T.J. Burkholder, J.A. Frangos, Uniaxial strain system to investigate strain rate regulation in vitro, Rev. Sci. Instrum. 72 (2001) 2415 – 2422. [7] A.M. Collinsworth, C.E. Torgan, S.N. Nagda, R.J. Rajalingam, W.E. Kraus, G.A. Truskey, Orientation and length of mammalian skeletal myocytes in response to a unidirectional stretch, Cell Tissue Res. 302 (2000) 243 – 251. [8] M. Cormont, E. van Obberghen, M. Zerial, Y. LeMarchand-Brustel, Insulin induces a change in rab5 subcellular localization in adipocytes
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
[9]
[10]
[11] [12]
[13]
[14]
[15]
[16]
[17]
[18]
[19] [20]
[21]
[22] [23]
[24]
[25] [26]
[27]
[28]
independently of phosphotidylinositol 3-kinase activation, Endocrinology 137 (1996) 3408 – 3415. K.C. Darr, E. Schultz, Hindlimb suspension suppresses muscle growth and satellite cell proliferation, J. Appl. Physiol. 67 (1987) 1827 – 1834. K.C. Darr, E. Schultz, Exercise-induced satellite cell activation in growing and mature skeletal muscle, J. Appl. Physiol. 63 (1987) 1816 – 1821. P.F. Davies, Flow-mediated endothelial mechanotransduction, Physiol. Rev. 75 (1995) 519 – 560. J. Dorrestijn, F.J. van Bussel, J.A. Maasen, D.S. Gomes de Mesquita, Early steps in insulin action, Arch. Physiol. Biochem. 106 (1998) 269 – 289. G.J. Etgen, A.R. Memon, G.A. Thompson, J.L. Ivy, Insulin- and contraction-stimulated translocation of GTP-binding proteins and GLUT4 protein in skeletal muscle, J. Biol. Chem. 268 (1993) 20164 – 20169. Y.V. Fedorov, N.C. Jones, B.B. Olwin, Regulation of myogenesis by fibroblast growth factors requires beta-gamma subunits of pertussis toxin-sensitive G proteins, Mol. Cell. Biol. 18 (1998) 5780 – 5787. R. Gallo, M. Serafini, L. Castellani, G. Falcone, S. Alema, Distinct effects of Rac1 on differentiation of primary avian myoblasts, Mol. Biol. Cell 10 (1999) 3137 – 3150. G. Goldspink, Changes in muscle mass and phenotype and the expression of autocrine and systemic growth factors by muscle in response to stretch and overload, J. Anat. 194 (1999) 323 – 334. G.E. Goslow, H.J. Seerherman, C.R. Taylor, M.N. McCutchin, N.C. Heglund, Electrical activity and relative length changes of dog limb muscles as a function of speed and gait, J. Exp. Biol. 94 (1981) 15 – 42. G.E. Goslow, R.M. Reinking, D.G. Stuart, The cat step cycle: hind limb joint angles and muscle lengths during unrestrained locomotion, J. Morph. 141 (1973) 1 – 41. S.R.P. Gudi, C.B. Clark, J.A. Frangos, Fluid flow rapidly activates G proteins in human endothelial cells, Circ. Res. 79 (1996) 834 – 839. S.R.P. Gudi, A.A. Lee, C.B. Clark, J.A. Frangos, Equibiaxial strain and strain rate stimulate early activation of G proteins in cardiac fibroblasts, Am. J. Physiol. 274 (1998) C1424 – C1428. J. Ihlemann, T. Ploug, Y. Hellsten, H. Galbo, Effect of tension on contraction-induced glucose transport in rat skeletal muscle, Am. J. Physiol. 277 (1999) E208 – E214. D.E. Inbger, Tensegrity: the architectural basis of cellular mechanotransduction, Annu. Rev. Physiol. 59 (1997) 575 – 599. R.S. James, J.D. Altringham, D.F. Goldspink, The mechanical properties of fast and slow skeletal muscles of the mouse in relation to their locomotory function, J. Exp. Biol. 198 (1995) 491 – 502. D.J. Kelvin, G. Simard, H.H. Tai, T.P. Yamaguchi, J.A. Connolly, Growth factors, signaling pathways, and the regulation of proliferation and differentiation in BC3H1 muscle cells: a pertussis toxinsensitive pathway is involved, J. Cell Biol. 108 (1989) 159 – 167. M. Kjeldgaard, J. Nyborg, B.F.C. Clark, The GTP binding motif: variations on a theme, FASEB J. 10 (1996) 1347 – 1368. D. LeRoith, P.C. Sampson, C.T. Roberts, How does the mitogenic insulin-like growth factor 1 receptor differ from the insulin receptor? Horm. Res. 41 (1994) 74 – 78. I.G. Macara, K.M. Lounsbury, S.A. Richards, C. McKiernan, D. Bar-Sagi, The ras superfamily of GTPases, FASEB J. 10 (1996) 625 – 630. N.L. McKnight, J.A. Frangos, Strain rate mechanotransduction in
[29]
[30] [31]
[32] [33]
[34] [35]
[36] [37]
[38]
[39]
[40]
[41]
[42] [43]
[44]
[45]
[46]
[47]
273
aligned human vascular smooth muscle cells, Ann. Biomed. Eng. 31 (2003) 239 – 249. R.M. Palmer, M.G. Thompson, T.R.M. Knott, G.P. Campbell, A. Thom, K.S. Morrison, Insulin and insulin-like growth factor-I responsiveness and signalling mechanisms in C2C12 satellite cells: effect of differentiation and fusion, Biochim. Biophys. Acta 1355 (1997) 167 – 176. T. Pfeuffer, GTP-binding proteins in membranes and the control of adenylate cyclase activity, J. Biol. Chem. 252 (1977) 7224 – 7234. M.B. Ramocki, M.A. Whie, S.F. Konieczny, E.J. Taparowsky, A role for ralGDS and a novel ras effector in the ras-mediated inhibition of skeletal myogenesis, J. Biol. Chem. 273 (1998) 17696 – 17701. M.A. Rizzo, G. Romero, The role of G proteins in insulin signaling, J. Basic Clin. Physiol. Pharmacol. 9 (1998) 167 – 195. T.A. Robertson, M.A. Maley, M.D. Grounds, J.M. Papadimitriou, The role of macrophages in skeletal muscle regeneration with particular reference to chemotaxis, Exp. Cell Res. 207 (1993) 321 – 331. E. Schultz, K.M. McCormick, Skeletal muscle satellite cells, Rev. Physiol. Biochem. Pharmacol. 123 (1994) 213 – 257. H. Shibata, W. Omata, I. Kojima, Insulin stimulates guanine nucleotide exchange on rab4 via a wortmanin-sensitive signaling pathway in rat adipocytes, J. Biol. Chem. 272 (1997) 14542 – 14546. M.I. Simon, M.P. Strathmann, N. Gautam, Diversity of G proteins in signal transduction, Science 252 (1991) 802 – 808. I. Simon, M. Zerial, R.S. Goody, Kinetics of interaction of rab5 and rab7 with nucleotides and magnesium ions, J. Biol. Chem. 271 (1996) 20470 – 20478. H. Takano, I. Komuro, T. Oka, I. Shiojima, Y. Hiroi, T. Mizuno, Y. Yazaki, The rho family of G proteins plays a critical role in muscle differentiation, Mol. Cell. Biol. 18 (1998) 1580 – 1589. R. Tatsumi, J.E. Anderson, C.J. Nevoret, O. Halevy, R.E. Allen, HGF/ SF is present in normal adult skeletal muscle and is capable of activating satellite cells, Dev. Biol. 194 (1998) 114 – 128. R. Tatsumi, A. Hattori, Y. Ikeuchi, J.E. Anderson, R.E. Allen, Release of hepatocyte growth factor from mechanically stretched skeletal muscle satellite cells and role of pH and nitric oxide, Mol. Biol. Cell 13 (2002) 2909 – 2918. R. Tatsumi, S.M. Sheehan, H. Iwasaki, A. Hattori, R.E. Allen, Mechanical stretch induces activation of skeletal muscle satellite cells in vitro, Exp. Cell Res. 267 (2001) 107 – 114. H.H. Vandenburgh, Dynamic mechanical orientation of skeletal myofibers in vitro, Dev. Biol. 93 (1982) 438 – 443. H.H. Vandenburgh, P. Karlisch, Longitudinal growth of skeletal myotubes in vitro in a new horizontal mechanical cell stimulator, In Vitro Cell Dev. Biol. 25 (1989) 607 – 616. H.H. Vandenburgh, J. Shansky, R.L. Solerssi, J. Chromiak, Mechanical stimulation of skeletal muscle increases prostaglandin F2a production, cyclooxygenase activity, and cell growth by a pertussis toxin sensitive mechanism, J. Cell. Physiol. 163 (1995) 285 – 294. J. Vierck, B. O’Reilly, K. Hossner, J. Antonio, K. Byrne, L. Bucci, M. Dodson, Satellite cell regulation following myotrauma caused by resistance exercise, Cell Biol. Int. 24 (2000) 263 – 272. L. Wei, W. Zhou, J.D. Croissant, F.E. Johansen, R. Prywes, A. Balasubramanyam, R.J. Schartz, RhoA signaling via serum response factor plays an obligatory role in myogenic differentiation, J. Biol. Chem. 273 (1998) 30287 – 30294. F.E. Zajac, M.S. Zomlefer, W.S. Levine, Hindlimb muscular activity, kinetics, and kinematics of cats jumping to their maximum achievable heights, J. Exp. Biol. 91 (1981) 73 – 86.
Stretch activation of GTP-binding proteins in C2C12 myoblasts Craig B. Clark, a Nathan L. McKnight, a,b and John A. Frangos a,b,* a
Department of Bioengineering, University of California San Diego, La Jolla, CA 92093-0142, USA b La Jolla Bioengineering Institute, La Jolla, CA 92037, USA Received 31 December 2002, revised version received 24 September 2003
Abstract Mechanical stimulation has been proposed as a fundamental determinant of muscle physiology. The mechanotransduction of strain and strain rate in C2C12 myoblasts were investigated utilizing a radiolabeled GTP analogue to detect stretch-induced GTP-binding protein activation. Cyclic uniaxial strains of 10% and 20% at a strain rate of 20% s 1 rapidly (within 1 min) activated a 25-kDa GTPase (183 F 17% and 186 F 19%, respectively), while 2% strain failed to elicit a response (109 F 11%) relative to controls. One, five, and sixty cycles of 10% strain elicited 187 F 20%, 183 F 17%, and 276 F 38% increases in activation. A single 10% stretch at 20% s 1, but not 0.3% s 1, resulted in activation. Insulin activated the same 25-kDa band in a dose-dependent manner. Western blot analysis revealed a panel of GTP-binding proteins in C2C12 myoblasts, and tentatively identified the 25-kDa GTPase as rab5. In separate experiments, a 40-kDa protein tentatively identified as Gai was activated (240 F 16%) by 10% strain at 1 Hz for 15 min. These results demonstrate the rapid activation of GTP-binding proteins by mechanical strain in myoblasts in both a strain magnitude- and strain rate-dependent manner. D 2003 Elsevier Inc. All rights reserved. Keywords: C2C12; Myoblast; Uniaxial strain; Strain rate; Mechanotransduction; Photoaffinity
Introduction Mechanical loading and deformation play an important role in the physiology of a variety of tissues, with extensive research in both animal and cell culture models identifying a rapidly expanding list of cellular responses. Mechanotransduction—translation of a mechanical stimulus into an intracellular biochemical processes—has been the subject of intensive study, yet many questions remain as to the underlying cellular mechanisms and the physical stimuli which drive them. Muscle provides an ideal tissue for the study of mechanotransduction, as it experiences a wide range of strains during normal use. In vivo, muscle is affected by a variety of stimuli including neural interactions, membrane depolarization, intracellular calcium flux, metabolite transport or depletion, hormonal influences, as well as the stresses and mechanical strains due to both external loading and myosin-generated tension. While each of these stimuli influences muscle performance and adaptation, me-
* Corresponding author. La Jolla Bioengineering Institute, 505 Coast Boulevard South, La Jolla, CA 92037-4616. Fax: +1-858-456-7540. E-mail address: [email protected] (J.A. Frangos). 0014-4827/$ - see front matter D 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.yexcr.2003.09.017
chanical strain has been proposed as a fundamental determinant of muscle physiology [16]. The strains under loading (strain being a normalized measure of stretch; where Strain = Deformation / Initial Length) and the cyclic strains during locomotion impose a spectrum of mechanical stimuli to direct the muscular phenotype, with strains during normal locomotory movements typically F10% of resting length, at strain rates of 0– 700% [17,18,23,47]. Strain rate (the rate at which the deformation is applied; where Strain Rate = Strain / Duration) varies by orders of magnitude and may therefore be an important parameter in modulating the mechanoresponse of cells and tissues. A host of candidates have been proposed for the mechanotransduction of physical forces including adhesion molecules, kinases, cytoskeletal elements, ion channels, and guanine nucleotide-binding regulatory proteins (GTP-binding proteins) [3,11,22]. The diverse family of GTP-binding proteins play a central role in signal transduction [27,36], and are also sensitive to physical forces, being activated by both fluid shear and mechanical strain [3,19,20]. GTP-binding proteins therefore provide an ideal candidate for the study of the mechanical stimulation of skeletal muscle. In the present study, the mechanotransduction of strain and the sensitivity to strain rate were investigated using
266
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
C2C12 myoblasts, a cell line morphologically and functionally similar to satellite cells [34]. A uniaxial stretch device [6] applied defined cyclic strain to C2C12 myoblast cultures, while incorporation of a radiolabeled GTP analogue allowed specific and sensitive detection of GTPbinding protein activation. Stretch activation was detected in GTP-binding proteins, tentatively identified as the small GTPase rab5 and a heterotrimeric Gai subunit. Activation was found to be both strain magnitude- and strain ratedependent.
Materials and methods Cell culture and uniaxial stretch C2C12 murine skeletal muscle myoblasts (ATCC) were kept as subconfluent cultures in DMEM (Irvine Scientific) supplemented with 10% fetal calf serum (FCS; Biocell). For all stretch and agonist experiments, cells were seeded at 20,000 cells cm 2 and the experiments run 48 h after seeding, at which time the cells had formed a confluent monolayer but had not begun fusing into myotubes. Standard 35-mm-diameter-tissue culture dishes (Falcon) seeded at 20,000 cells cm 2 were used for agonist and inhibitor studies. Cultures were rinsed twice with serum-free media and switched to reduced serum (1%) media 8 h before the start of experiments. Serum content was reduced to provide quiescent cultures in terms of minimizing the basal GTPbinding protein activity, allowing improved signal-to-noise under stimulated (agonist or mechanical stretch) conditions. No visible changes in cell morphology were noted after the 8 h in 1% serum, with all cultures maintained as a confluent layer of mononucleated cells. Cells cultured on compliant membranes were subjected to cyclic uniaxial strain in a strain device described in Clark et al. [6]. Briefly, cells were cultured within a sterile enclosure on 4.8 3.7 0.010 cm in thick silicone membranes (Specialty Manufacturing) which were held at each end with stainless steel clamps. Membranes were coated before cell seeding with type I collagen (Sigma) to a final concentration of 6 Ag cm 2. A stepper motor (Anaheim Automation) applied cyclic strain to four membranes simultaneously, with a triangular displacement waveform featuring a constant rate of strain for both stretch and relaxation, starting from an initial membrane length of 48.3 mm. Controls for all stretch experiments consisted of cells seeded on membranes and treated identically, but not subjected to stretch. Experiments were carried out in a temperature-controlled hood at 37jC, the cultures flushed with humidified and sterile filtered 5% CO2/95% air. Photolabeling of GTP-binding proteins The 32P-labeled, nonhydrolyzable, photoreactive GTP analogue, azidoanalide-GTP (AAGTP, specific activity
25 – 50 mCi mmol 1, Affinity Labeling Technologies) was utilized to detect GTP-binding protein activation [30]. Briefly, photoaffinity experiments were performed in a darkened room utilizing a safelight, with all equipment and reagents maintained at 37jC within the temperaturecontrolled hood for the duration of the experiments. The nucleotide was suspended in serum-free DMEM or PBS along with digitonin (15 AM, Calbiochem) in 150-Al total volume and 10– 15 ACi AAGTP per sample. Digitonin is a detergent that permeabilized the cell membrane and facilitated loading of the AAGTP label into the cells. As such, the perturbation of the lipid bilayer may nonspecifically activate certain intracellular signaling pathways, and therefore identically treated sham (not stretched) controls were utilized to account for any nonspecific activation effects. Media were aspirated from the cell cultures, the AAGTP label added for the selected incubation period, gently rinsed with the appropriate buffer (serum-free DMEM or PBS), and the bout of stretch applied. Rinsing and aspiration of the excess media and label before applying stretch minimize any potential effects of increased AAGTP uptake due to stretch, as well as eliminate potential fluid shear effects due to fluid movement. Following stretch, a UV lamp (UVP, model UVG-54; 254 nm wavelength) was placed over the stretch chamber approximately 10 cm above the membranes, and the cells subjected to UV light for 60 s. Ice-cold PBS with 4 mM dithiothreitol (DTT, Calbiochem) was added after the initiation of UV to reduce nonspecific cross-linking of the photolabel. Samples were immediately harvested in 2 electrophoresis sample buffer (62.5 mmol Tris –HCl, pH 6.8, 2% w/v SDS, 10% glycerol, 20% v/v h-mercaptoethanol, and 0.1% w/v bromophenol blue, Sigma), boiled for 5 min, sheared with a 21 gauge needle, and stored at 80jC until analyzed. In select cultures, insulin (Sigma) was added at various concentrations. Following preincubation in serum-free media, media were aspirated and the AAGTP label incubated for 15 min. Excess label was then gently rinsed with serum-free media and aspirated, and insulin (10 and 100 nM, 1 AM) in DMEM added for a subsequent 5-min incubation. Dishes were then transferred onto ice, subjected to UV, and harvested as described above. For the serum response experiments, the AAGTP label in DMEM was added along with the appropriate serum concentration (0– 20%) without digitonin for 60-min incubation before UV. Electrophoresis, autoradiography, and Western blot analysis Equal volumes of sample were separated by SDS-PAGE on 12% acrylamide running/5% stacking gels. Gels were stained with Coomasie brilliant blue (Sigma) to verify equal loading of protein, then dried and film (Fuji) exposed for 2– 14 days for autoradiography. Radiographs were scanned and
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
267
band intensities analyzed (Quantity One, BioRad), with values reported as a percentage of controls from within each experiment. Western blot analysis was performed for selected experiments, with AAGTP-labeled samples harvested in 2 sample buffer and run on 12% gels as described above. Gels were then equilibrated in transfer buffer (192 mM glycine, 25 mM Tris, 20% methanol) for 15 min and electrotransferred onto PVDF membranes (ImmobilonP, Millipore). Following transfer, the membranes were rinsed, air dried, and film (Fuji) exposed for 2 –14 days. Following autoradiography, the membranes were cut into strips corresponding to lanes of the gel and incubated overnight at 4jC in blocking buffer [immunoblot buffer (IB—10 mM Tris pH 7.4, 100 mM NaCl, 0.1% tween-20) with 5% nonfat milk]. Primary rabbit polyclonal antibodies against rhoA, ras, Gai/o/t/z, Gas, Gaq/11, (Santa Cruz), and rab4, rab5 (StressGen) were incubated at 1:1000 dilution in blocking buffer for 1 h at room temperature, rinsed six times in IB, followed by 30 min incubation in horseradish peroxidase-conjugated secondary antibody (anti-rabbit, New England Biolabs) at 1:1000 dilution in blocking buffer, and six additional rinses in IB. Bands were visualized using the SuperSignal (Pierce) chemiluminescent kit per manufacturer’s instructions. Autoradiographs and Western blots from the same PVDF membrane were then overlaid and aligned for band comparison and identification. Statistics All results are presented as the mean F SE, normalized with respect to dedicated controls from each experiment. Data were first evaluated by analysis of variance (ANOVA), followed by the Student – Newman– Keuls (SNK) test for multiple comparisons, with P < 0.05 considered statistically significant.
Results Photoaffinity labeling of GTP-binding proteins As a means of activating and identifying a spectrum of GTP-binding proteins, fetal calf serum was added to serumstarved cultures along with AAGTP for 1 h, with a panel of GTP-binding proteins being revealed (Fig. 1). Activation with 20% serum relative to the BSA (serum-free) controls ranged from 380 F 9% (mw 33 kDa) to 165 F 4% (mw 38 kDa), with the protein activation levels appearing to cluster into two distinct groups. As an additional control, 50-AM digitonin was added for 3 min to BSA controls following the 1-h label incubation period to investigate nonspecific activation. Average intensity after treatment with digitonin for the six bands marked by arrows in Fig. 1A was 99 F 13% relative to non-digitonin-treated samples (data not
Fig. 1. Long-term (60 min) incubation with AAGTP revealed a panel of GTP-binding proteins, while serum activated select isoforms in a dosedependent manner. (A) Typical radiograph is shown for samples incubated with 0 – 20% serum. Note the two bands on right are BSA (0% serum) controls, with digitonin (50 AM) added after label incubation. (B) Activation levels are plotted normalized to BSA controls, with apparent molecular weights (mw) corresponding to the radiograph bands next to each plot. Mean F SE for n = 3 (error bars shown on top trace only for clarity).
shown), indicating that the use of digitonin as a permeabilizing agent had negligible effect on GTP-binding protein activation. A series of control experiments were performed to demonstrate specificity of the AAGTP label to G proteins. The nonhydrolyzable GDP analogue GDPhS irreversibly binds to GTP-binding proteins, inhibiting subsequent nucleotide exchange and photolabel incorporation. Preincubation with 300 or 900 AM GDPhS for 8 h significantly
268
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
reduced subsequent AAGTP binding in the 25-kDa band to 56 F 9% and 34 F 8% of controls, respectively (Fig. 2A). The addition of non-radiolabeled GTP at 50 or 200 AM during AAGTP incubation competitively reduced label
Fig. 3. Stimulation with insulin for 5 min following AAGTP loading activated the 25-kDa GTPase protein in a dose-dependent manner. Typical radiograph is shown for the three insulin concentrations (upper panel). Radiograph band intensities are plotted, normalized to control (lower panel). Mean F SE for n = 4. GTPase activity (25-kDa) correlated significantly (P < 0.001) with insulin concentration as determined by Spearman rank correlation test.
incorporation to 64 F 12% and 49 F 9% of controls, respectively (Fig. 2B). Similar values of inhibition were seen for other molecular weight GTP-binding proteins (data not shown). Insulin added for 5 min following AAGTP incubation elicited a dose-dependent activation of the 25-kDa band over the range of concentrations tested (10 and 100 nM, and 1 AM) (Fig. 3), with 103 F 6%, 153 F 8%, and 192 F 17% activation relative to control, respectively.
Fig. 2. (A) Preincubation with 300 or 900 AM GDPhS (8 h) inhibited nucleotide exchange and binding of the radiolabeled probe to 56 F 9% and 34 F 8% of control, respectively. Typical radiograph showing the 38-, 33-, and 25-kDa bands (upper panel). Plot of radiograph band intensities for 25-kDa (typical for all bands) normalized to control (lower panel). Mean F SE for n = 4, *P < 0.01 vs. control. (B) Addition of unlabeled 50 or 200 AM GTP during AAGTP incubation competitively reduced binding of radiolabeled probe in C2C12 myoblasts to 64 F 12% and 49 F 9% of control, respectively. Typical radiograph showing the 25-kDa band (upper panel). Plot of radiograph band intensities, normalized to control (lower panel). Mean F SE for n = 3 (*P < 0.01 for all comparisons).
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
269
rate constant. Cyclic strain of 10% at a strain rate of 20% s 1 (1 Hz) was applied for 1, 5, and 60 cycles, with UV irradiation applied 60 s after the initiation of stretch in each experiment. This treatment activated the 25-kDa protein, with 1, 5, and 60 cycles of stretch resulting in increased AAGTP binding of 187 F 20%, 183 F 17%, and 276 F 38%, respectively, relative to controls (Fig. 5). Finally, the strain magnitude and number of cycles were held constant, as a single cycle of 10% strain applied over 1 min (‘slow’ strain rate = 0.3% s 1) was compared to a single cycle of 10% strain over 1 s (‘fast’ strain rate = 20% s 1). The single ‘fast’ cycle activation was 187 F 20% of control, while the single ‘slow’ cycle was not significantly different from controls (107 F 13%) (Fig. 6). Activation of a heterotrimeric G protein by 15 min of stretch Longer-term stretch experiments utilized AAGTP label suspended in CF-DPBS followed by 15 min of continuous
Fig. 4. Application of five cycles of 10% or 20% uniaxial strain activated the 25-kDa protein in myoblasts. (A) Displacement waveforms for the applied uniaxial strain are shown, all applied at a strain rate of 20% s 1. (B) Typical radiograph showing the 25-kDa band activation after the indicated stretch regime. (C) Radiograph band intensities are plotted, normalized to sham control for 2%, 10%, and 20% strain. Application of five cycles of 10% or 20% uniaxial strain activated the 25-kDa band 183 F 17% and 186 F 19%, respectively, while activation by 60 cycles of 2% strain (109 F 11%) was not different from control. Mean F SE for n = 6 (*P < 0.05 vs. 2% and control).
Activation of a small GTPase by 1 min of cyclic stretch Activation of GTP-binding proteins by mechanical strain was investigated utilizing C2C12 myoblasts cultured on silicone membranes, and subjected to uniaxial strain applied at various strain magnitudes, strain rates, and number of cycles. A set of experiments was performed across a range of strain magnitudes, all applied at the same strain rate (20% s 1). Five cycles of stretch at strains of 10% and 20% elicited activation of the 25kDa protein 183 F 17% and 186 F 19% relative to controls, respectively, while activation by 60 cycles of 2% strain was not significantly different from control values (109 F 11%, Fig. 4), suggesting a threshold of activation by strain. In the next set of experiments, the number of cycles was varied while holding the strain magnitude and strain
Fig. 5. Application of 1, 5, or 60 cycles of 10% uniaxial strain at a strain rate of 20% s 1 resulted in increasing GTPase protein activation. (A) Typical radiograph is shown, with the 25-kDa band activated by the indicated number of uniaxial strain cycles. (B) Radiograph band intensities are plotted, normalized to sham control, with 1, 5, or 60 cycles of 10% uniaxial strain at 20% s 1 activating the 25-kDa band 187 F 20%, 183 F 17%, and 276 F 38% relative to control, respectively. Mean F SE for n = 6 (*P < 0.05 vs. sham; **P < 0.01 vs. sham, P < 0.05 vs. 1 or 5 cycles).
270
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
at their appropriate molecular weights in the C2C12 myoblasts. Overlay and alignment of the Western blots and autoradiograph revealed alignment of bands labeled by both the antibody and AAGTP photoaffinity probe. The three heterotrimeric subunits matched the three radiograph bands between 38 and 45 kDa, and rab5 aligned with the prominent 25-kDa band. A faint band was also seen on the radiograph that aligns with the rhoA immunoblot. While this method was sufficient to tentatively identify the 25-kDa band as rab5 and the 40-kDa band as Gai, it can only positively exclude a specific protein from being associated with a particular autoradiograph band. Other GTP-binding proteins which co-mi-
Fig. 6. Stimulation of GTPase protein activity was sensitive to strain rate. (A) Stretch protocol of 10% strain applied over 1 s (‘fast’ 20% s 1—solid line) or 1 min (‘slow’ 0.3% s 1—dashed line) is shown. Application of a single ‘fast’ cycle resulted in GTPase protein activation, while one cycle of ‘slow’ strain did not significantly increase activation. A typical radiograph is shown in Fig. 5A. (B) Relative band intensities are plotted, normalized to sham control, with ‘fast’ and ‘slow’ activated 187% F 20% and 107 F 13%, respectively. Mean F SE for n = 6 (*P < 0.05 for all comparisons).
cyclic stretch. The application of 900 cycles of 10% strain at 20% s 1 activated a 40-kDa band 240 F 16% relative to sham controls (Fig. 7). This protocol of label incubation and stretch also activated the 25-kDa band (198 F 7% relative to control) similar to the previous 1-min protocols in DMEM. Identification of GTP-binding proteins Electroblotted membranes were first exposed for autoradiography, and then incubated with primary antibodies against rab4, rab5, rhoA, Gaq/11, Gai, and GaS (Fig. 8). The Western blot confirmed the presence of each isoform
Fig. 7. Loading of AAGTP in CF-DPBS followed by 900 cycles of 10% strain (20% s 1) indicated activation of both 40- and 25-kDa GTP-binding proteins. (A) Typical radiograph showing the 40- and 25-kDa bands for control (C) vs. stretch (S). (B) Plot of radiograph band intensities, normalized to control. Mean F SE for n = 6 (*P < 0.05 vs. sham control).
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
Fig. 8. Autoradiography and Western blot analysis identified GTP-binding protein isoforms in C2C12 myoblasts. Western blot against rab4, rab5, rhoA, Gaq/11, Gai, and GaS revealed the presence of these isoforms (upper panel). In the lower panel, the Western blot image is false colored (white) and overlaid onto an autoradiograph from the same transfer membrane. Alignment correlated bands labeled by both the antibody and the AAGTP photoaffinity probe. Rab5 aligned with the prominent 25-kDa band, and the three heterotrimeric subunits matched the three radiograph bands between 38- and 45-kDa. Note that the membrane was cut in the middle of lane 2 so that rab4 and rab5 each stained half of the lane to demonstrate the difference in molecular weight.
grate with rab5 and Gai may be responsible for the activation noted in the experiments above.
Discussion The present study demonstrated the novel activation of a 25-kDa GTPase and a 40-kDa heterotrimeric Ga subunit by mechanical stretch in C2C12 myoblasts. Activation of the small GTPase was further shown to be dependent on the magnitude of strain, the number of cycles, and the rate at which strains were applied. Application of uniaxial strain to cultures of unfused C2C12 cells provides a model for both myoblasts during development as well as mononucleated satellite cells which reside along the muscle fibers beneath the basal lamina [33,34]. Mechanical strain influences muscle at the level of these mononucleated cells, as strain imposed during development and use drives cell proliferation and fusion into myofibers [9,10,34]. In adult muscle, satellite cells are activated during muscle stretch or damage, gaining the ability to fuse to each other to form new fibers or to fuse with existing fibers [9,10,34]. Coupled with the recent work on stretch-induced satellite cell activation
271
[40,41], the activation of GTP-binding proteins by stretch demonstrated in this study supports the hypothesis that myoblasts sense and respond to mechanical stimuli. When the strain rate was held at a constant 20% s 1, a 25-kDa GTPase was activated at strains of 10% and 20%, but not at strains of 2%, suggesting a threshold of straindependent activation. Application of one or five cycles of 10% strain at 20% s 1 produced equivalent activation of the 25-kDa GTPase, with 60 cycles eliciting a further increase in activation, indicating not only a dose response with the number of cycles, but highlighting the extreme sensitivity of GTP-binding protein activation to even a single cycle of strain. Comparing the GTP-binding protein activation by a single ‘fast’ (one cycle of 10% strain applied over one second) to a single ‘slow’ (10% strain over 1 min) cycle confirmed that cells can sense and differentially respond to the rate at which strains are applied (strain rate). This observation suggests that activation is due to dynamic straining of the cells rather than simple tension, as the ‘slow’ protocol applied the deformation to the cells for much greater duration than the ‘fast’ protocol, yet did not stimulate GTP-binding protein activity. Previous cell culture studies have addressed the strain rate response of cells, as activation of heterotrimeric G proteins in cardiac fibroblasts [20] and mitogen-activated protein kinase phosphorylation in smooth muscle [28] were shown to be strain rate sensitive. Modulation by strain rate was seen during myotube formation in vitro, as extremely slow elongation (0.5% h 1) induced alignment along the axis of stretch [42] while cyclic stretch at higher strain rates (20% s 1) aligned the myotubes perpendicular to the stretch axis [43]. In contrast, strain rate did not affect the orientation response of C2C12 myotubes to stretch [7]. However, the rates imposed in the latter study (2.26% day 1 and 1.77% day 1) were designed to mimic murine long bone growth rates and are significantly different from those utilized in the present work. The strains and strain rates applied in the current study were chosen to be representative of the physiologic range of skeletal muscle during normal locomotion. Satellite cells reside between the basal lamina and the sarcolemma of myofibers and participate in the hypertrophy of muscle during normal growth and after myofiber injury. While a variety of factors have been identified which can stimulate growth of cultured muscle [45], the activation of satellite cells after skeletal muscle injury is associated with hepatocyte growth factor (HGH) [39] through autocrine release of HGH induced by cyclic stretch [41]. It was further demonstrated this HGH release is dependent on stretch-stimulated nitric-oxide synthase (NOS) activity [40]. This provides one example where stretch triggers specific mechano-sensitive signal transduction pathways in the satellite cells, as opposed to inducing a response (in this case HGH release) via a nonspecific mechanism such as stretch-induced injury and permeabilization.
272
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
Both heterotrimeric G proteins and the small GTPase proteins have been implicated in numerous aspects of muscle physiology. In culture, myoblast proliferation, differentiation, and fusion were Pertussis toxin (PTx)-sensitive in the BC3H1 myoblast cell line [24], indicating the involvement of Gai. Federov et al. [14] indicated the Ghg subunits dissociated from the activated Gai also participate in FGF-mediated myoblast proliferation process. For the small GTPases, constitutively activated ras inhibited skeletal muscle differentiation [31], while myoblasts expressing a constitutively activated rac1 fused yet the resulting myotubes failed to form ordered sarcomeres [15]. The small GTPase rho is also active in muscle differentiation [46], and the expression of MyoD, myogenin, and myocyte enhancer binding factor 2 (MEF2) were enhanced by rho activity and suppressed by rho inhibition [5,38]. Evidence of mechanical activation of heterotrimeric G proteins in skeletal muscle comes indirectly from studies such as the PTx inhibition of stretch-induced cyclooxygenase activity, PGF2a production, phospholipase D activation, and cell growth in differentiated avian myoblasts [44]. The present study demonstrates the rapid activation of a 40-kDa G protein by mechanical strain in skeletal muscle. Both insulin treatment and stretch of C2C12 myoblasts activated a 25-kDa GTPase tentatively identified as rab5. A major function of insulin is to stimulate glucose uptake and metabolism, with skeletal muscle the principal tissue site for insulin-mediated glucose uptake [2]. Myoblasts and myotubes in culture also respond to insulin [29]. In skeletal muscle, insulin stimulated ras [26], as well as rho and rac [32], though their lower molecular weights (20 – 21 kDa) exclude these proteins as candidates for the insulin-stimulated 25-kDa band seen in the present study. The rab family of small GTPases has a size range (24 – 25 kDa) which correlates with the 25-kDa band seen in Fig. 8, and has demonstrated insulin sensitivity, with the isoforms rab4 [32,35], rab5 [8], and rab5B [1] all activated by insulin. The exocytosis of vesicle-bound glucose transporters to the plasma membrane is directed by these rab isoforms [1], and although both insulin [4,35] and mechanical activity in skeletal muscle [13,21] activate transporter mobilization, activation of rab by mechanical activity had not been previously demonstrated. Mechanical stretch represents a novel activation mechanism for the small GTPase, and supports the role of passive mechanical strain in activating fundamental processes within skeletal muscle cells. It also suggests a possible mechanism not only for crosstalk between receptor tyrosine kinases and GTP-binding proteins [12,32], but also the interaction between ligand- and stretchactivated events. The longer (15 min) stretch experiments revealed the activation of a 40-kDa band, with the Western blot alignment tentatively identifying the protein as a Gai isoform. Gai3 activation has been previously identified in HUVEC stimulated by fluid shear [19], while Gai1 was activated by mechanical strain in cardiac fibroblasts [20] indicating
possible isoform specificity utilized by different cell types. The antibody used in the present study is common to the Gai family of isoforms, and the specific isoform (Gai1, Gai2, or Gai3) involved in the C2C12 mechanoresponse remains to be identified. The use of calcium-free buffer rather than DMEM in these experiments allowed detection of the Ga subunit, presumably by altering the kinetics of GDP release or GTP/AAGTP binding, which are Ca2+ and Mg2+ sensitive [25,37]. While only two GTP-binding proteins were identified as responding to mechanical strain, additional GTP-binding protein isoforms may participate in the mechanoresponse, but may have a poor affinity for the AAGTP probe or are activated at levels too low to be detected by the assay under the experimental conditions utilized in this study. In summary, photoaffinity labeling of GTP-binding proteins in C2C12 myoblasts identified a panel of enzymes, and indicated the rapid activation by mechanical strain of proteins tentatively identified as rab5 and a heterotrimeric Gai subunit. Furthermore, GTP-binding protein activation was dependent not only on the magnitude of the strain, but on the rate at which the strain was applied.
Acknowledgments This study was supported in part by NIH Grant HL 40696. C.B. Clark expresses his appreciation to the Achievement Rewards for College Scientists and the Robert M. Golden Foundation for their support.
References [1] S. Bao, J. Zhu, W.T. Garvey, Cloning of rab GTPases expressed in human skeletal muscle: studies in insulin-resistant subjects, Horm. Metab. Res. 30 (1998) 656 – 662. [2] A.D. Baron, G. Brechtel, P. Wallace, S.V. Edelman, Rates and tissue sites of non-insulin- and insulin-mediated uptake in humans, Am. J. Physiol. 255 (1988) E764 – E769. [3] F. Berthiaume, J.A. Frangos, Flow-induced prostacyclin production is mediated by a pertussis toxin-sensitive G protein, FEBS Lett. 308 (1992) 277 – 279. [4] D.M. Calderhead, K. Kitagawa, G.E. Lienhard, G.W. Gould, Translocation of the brain-type glucose transporter largely accounts for insulin stimulation of glucose transport in BC3H-1 myocytes, Biochem. J. 269 (1990) 597 – 601. [5] G. Carnac, M. Primig, M. Kitzmann, P. Chafey, D. Tuil, N. Lamb, A. Fernandez, RhoA GTPase and serum response factor control selectively the expression of MyoD without affecting Myf5 in mouse myoblasts, Mol. Biol. Cell 9 (1998) 1891 – 1902. [6] C.B. Clark, T.J. Burkholder, J.A. Frangos, Uniaxial strain system to investigate strain rate regulation in vitro, Rev. Sci. Instrum. 72 (2001) 2415 – 2422. [7] A.M. Collinsworth, C.E. Torgan, S.N. Nagda, R.J. Rajalingam, W.E. Kraus, G.A. Truskey, Orientation and length of mammalian skeletal myocytes in response to a unidirectional stretch, Cell Tissue Res. 302 (2000) 243 – 251. [8] M. Cormont, E. van Obberghen, M. Zerial, Y. LeMarchand-Brustel, Insulin induces a change in rab5 subcellular localization in adipocytes
C.B. Clark et al. / Experimental Cell Research 292 (2004) 265–273
[9]
[10]
[11] [12]
[13]
[14]
[15]
[16]
[17]
[18]
[19] [20]
[21]
[22] [23]
[24]
[25] [26]
[27]
[28]
independently of phosphotidylinositol 3-kinase activation, Endocrinology 137 (1996) 3408 – 3415. K.C. Darr, E. Schultz, Hindlimb suspension suppresses muscle growth and satellite cell proliferation, J. Appl. Physiol. 67 (1987) 1827 – 1834. K.C. Darr, E. Schultz, Exercise-induced satellite cell activation in growing and mature skeletal muscle, J. Appl. Physiol. 63 (1987) 1816 – 1821. P.F. Davies, Flow-mediated endothelial mechanotransduction, Physiol. Rev. 75 (1995) 519 – 560. J. Dorrestijn, F.J. van Bussel, J.A. Maasen, D.S. Gomes de Mesquita, Early steps in insulin action, Arch. Physiol. Biochem. 106 (1998) 269 – 289. G.J. Etgen, A.R. Memon, G.A. Thompson, J.L. Ivy, Insulin- and contraction-stimulated translocation of GTP-binding proteins and GLUT4 protein in skeletal muscle, J. Biol. Chem. 268 (1993) 20164 – 20169. Y.V. Fedorov, N.C. Jones, B.B. Olwin, Regulation of myogenesis by fibroblast growth factors requires beta-gamma subunits of pertussis toxin-sensitive G proteins, Mol. Cell. Biol. 18 (1998) 5780 – 5787. R. Gallo, M. Serafini, L. Castellani, G. Falcone, S. Alema, Distinct effects of Rac1 on differentiation of primary avian myoblasts, Mol. Biol. Cell 10 (1999) 3137 – 3150. G. Goldspink, Changes in muscle mass and phenotype and the expression of autocrine and systemic growth factors by muscle in response to stretch and overload, J. Anat. 194 (1999) 323 – 334. G.E. Goslow, H.J. Seerherman, C.R. Taylor, M.N. McCutchin, N.C. Heglund, Electrical activity and relative length changes of dog limb muscles as a function of speed and gait, J. Exp. Biol. 94 (1981) 15 – 42. G.E. Goslow, R.M. Reinking, D.G. Stuart, The cat step cycle: hind limb joint angles and muscle lengths during unrestrained locomotion, J. Morph. 141 (1973) 1 – 41. S.R.P. Gudi, C.B. Clark, J.A. Frangos, Fluid flow rapidly activates G proteins in human endothelial cells, Circ. Res. 79 (1996) 834 – 839. S.R.P. Gudi, A.A. Lee, C.B. Clark, J.A. Frangos, Equibiaxial strain and strain rate stimulate early activation of G proteins in cardiac fibroblasts, Am. J. Physiol. 274 (1998) C1424 – C1428. J. Ihlemann, T. Ploug, Y. Hellsten, H. Galbo, Effect of tension on contraction-induced glucose transport in rat skeletal muscle, Am. J. Physiol. 277 (1999) E208 – E214. D.E. Inbger, Tensegrity: the architectural basis of cellular mechanotransduction, Annu. Rev. Physiol. 59 (1997) 575 – 599. R.S. James, J.D. Altringham, D.F. Goldspink, The mechanical properties of fast and slow skeletal muscles of the mouse in relation to their locomotory function, J. Exp. Biol. 198 (1995) 491 – 502. D.J. Kelvin, G. Simard, H.H. Tai, T.P. Yamaguchi, J.A. Connolly, Growth factors, signaling pathways, and the regulation of proliferation and differentiation in BC3H1 muscle cells: a pertussis toxinsensitive pathway is involved, J. Cell Biol. 108 (1989) 159 – 167. M. Kjeldgaard, J. Nyborg, B.F.C. Clark, The GTP binding motif: variations on a theme, FASEB J. 10 (1996) 1347 – 1368. D. LeRoith, P.C. Sampson, C.T. Roberts, How does the mitogenic insulin-like growth factor 1 receptor differ from the insulin receptor? Horm. Res. 41 (1994) 74 – 78. I.G. Macara, K.M. Lounsbury, S.A. Richards, C. McKiernan, D. Bar-Sagi, The ras superfamily of GTPases, FASEB J. 10 (1996) 625 – 630. N.L. McKnight, J.A. Frangos, Strain rate mechanotransduction in
[29]
[30] [31]
[32] [33]
[34] [35]
[36] [37]
[38]
[39]
[40]
[41]
[42] [43]
[44]
[45]
[46]
[47]
273
aligned human vascular smooth muscle cells, Ann. Biomed. Eng. 31 (2003) 239 – 249. R.M. Palmer, M.G. Thompson, T.R.M. Knott, G.P. Campbell, A. Thom, K.S. Morrison, Insulin and insulin-like growth factor-I responsiveness and signalling mechanisms in C2C12 satellite cells: effect of differentiation and fusion, Biochim. Biophys. Acta 1355 (1997) 167 – 176. T. Pfeuffer, GTP-binding proteins in membranes and the control of adenylate cyclase activity, J. Biol. Chem. 252 (1977) 7224 – 7234. M.B. Ramocki, M.A. Whie, S.F. Konieczny, E.J. Taparowsky, A role for ralGDS and a novel ras effector in the ras-mediated inhibition of skeletal myogenesis, J. Biol. Chem. 273 (1998) 17696 – 17701. M.A. Rizzo, G. Romero, The role of G proteins in insulin signaling, J. Basic Clin. Physiol. Pharmacol. 9 (1998) 167 – 195. T.A. Robertson, M.A. Maley, M.D. Grounds, J.M. Papadimitriou, The role of macrophages in skeletal muscle regeneration with particular reference to chemotaxis, Exp. Cell Res. 207 (1993) 321 – 331. E. Schultz, K.M. McCormick, Skeletal muscle satellite cells, Rev. Physiol. Biochem. Pharmacol. 123 (1994) 213 – 257. H. Shibata, W. Omata, I. Kojima, Insulin stimulates guanine nucleotide exchange on rab4 via a wortmanin-sensitive signaling pathway in rat adipocytes, J. Biol. Chem. 272 (1997) 14542 – 14546. M.I. Simon, M.P. Strathmann, N. Gautam, Diversity of G proteins in signal transduction, Science 252 (1991) 802 – 808. I. Simon, M. Zerial, R.S. Goody, Kinetics of interaction of rab5 and rab7 with nucleotides and magnesium ions, J. Biol. Chem. 271 (1996) 20470 – 20478. H. Takano, I. Komuro, T. Oka, I. Shiojima, Y. Hiroi, T. Mizuno, Y. Yazaki, The rho family of G proteins plays a critical role in muscle differentiation, Mol. Cell. Biol. 18 (1998) 1580 – 1589. R. Tatsumi, J.E. Anderson, C.J. Nevoret, O. Halevy, R.E. Allen, HGF/ SF is present in normal adult skeletal muscle and is capable of activating satellite cells, Dev. Biol. 194 (1998) 114 – 128. R. Tatsumi, A. Hattori, Y. Ikeuchi, J.E. Anderson, R.E. Allen, Release of hepatocyte growth factor from mechanically stretched skeletal muscle satellite cells and role of pH and nitric oxide, Mol. Biol. Cell 13 (2002) 2909 – 2918. R. Tatsumi, S.M. Sheehan, H. Iwasaki, A. Hattori, R.E. Allen, Mechanical stretch induces activation of skeletal muscle satellite cells in vitro, Exp. Cell Res. 267 (2001) 107 – 114. H.H. Vandenburgh, Dynamic mechanical orientation of skeletal myofibers in vitro, Dev. Biol. 93 (1982) 438 – 443. H.H. Vandenburgh, P. Karlisch, Longitudinal growth of skeletal myotubes in vitro in a new horizontal mechanical cell stimulator, In Vitro Cell Dev. Biol. 25 (1989) 607 – 616. H.H. Vandenburgh, J. Shansky, R.L. Solerssi, J. Chromiak, Mechanical stimulation of skeletal muscle increases prostaglandin F2a production, cyclooxygenase activity, and cell growth by a pertussis toxin sensitive mechanism, J. Cell. Physiol. 163 (1995) 285 – 294. J. Vierck, B. O’Reilly, K. Hossner, J. Antonio, K. Byrne, L. Bucci, M. Dodson, Satellite cell regulation following myotrauma caused by resistance exercise, Cell Biol. Int. 24 (2000) 263 – 272. L. Wei, W. Zhou, J.D. Croissant, F.E. Johansen, R. Prywes, A. Balasubramanyam, R.J. Schartz, RhoA signaling via serum response factor plays an obligatory role in myogenic differentiation, J. Biol. Chem. 273 (1998) 30287 – 30294. F.E. Zajac, M.S. Zomlefer, W.S. Levine, Hindlimb muscular activity, kinetics, and kinematics of cats jumping to their maximum achievable heights, J. Exp. Biol. 91 (1981) 73 – 86.