To better understand the role of the transient expression of ryanodine receptor (RyR) type 3 (RyR3) on Ca2+ homeostasis during the development of skeletal muscle, we have analyzed the effect of expression levels of RyR3 and RyR1 on the overall physiology of cultured myotubes and muscle fibers. Dyspedic myotubes were infected with RyR1 or RyR3 containing virions at 0.2, 0.4, 1.0, and 4.0 moieties of infection (MOI), and analysis of their pattern of expression, caffeine sensitivity, and resting free Ca2+ concentration ([Ca2+]r) was performed. Although increased MOI resulted in increased expression of each receptor isoform, it did not significantly affect the immunopattern of RyRs or the expression levels of calsequestrin, triadin, or FKBP-12. Interestingly, myotubes expressing RyR3 always had significantly higher [Ca2+]r and lower caffeine EC50 than did cells expressing RyR1. Although some of the increased sensitivity of RyR3 to caffeine could be attributed to the higher [Ca2+]r in RyR3-expressing cells, studies of [3H]ryanodine binding demonstrated intrinsic differences in caffeine sensitivity between RyR1 and RyR3. Tibialis anterior (TA) muscle fibers at different stages of postnatal development exhibited a transient increase in [Ca2+]r coordinately with their level of RyR3 expression. Similarly, adult soleus fibers, which also express RyR3, had higher [Ca2+]r than did adult TA fibers, which exclusively express RyR1. These data show that in skeletal muscle, RyR3 increases [Ca2+]r more than RyR1 does at any expression level. These data suggest that the coexpression of RyR1 and RyR3 at different levels may constitute a novel mechanism by which to regulate [Ca2+]r in skeletal muscle.
- ryanodine receptor
- calcium release
- ryanodine binding
- muscle fibers
although the role of ryanodine receptor (RyR) type 1 (RyR1) in Ca2+ homeostasis of skeletal muscle has been well documented, the role of RyR3 in this tissue remains elusive. Its coexpression with RyR1, the predominant skeletal isoform, and the broad (<1–50%) variability in expression levels of RyR3 in muscle fibers of different species have been the source of some of the main difficulties in conducting a comprehensive study and in understanding the role of RyR3 in Ca2+ homeostasis. RyR1 and, to a lesser extent, RyR3 have been characterized extensively, both biochemically and electrophysiologically. The results of these studies have revealed that although these channels share several common functional properties (4, 16, 33), they appear to differ in many respects, such as regulation by channel modulators [i.e., Ca2+, Mg2+, caffeine, and 4-chloro-m-cresol (4-CmC)] and subcellular organization (11, 27). The extent to which the differences between these RyR isoforms contribute to Ca2+ signaling in skeletal muscle is not fully understood.
The use of animal models null for all RyR isoforms has been of particular help in uncovering the specific role that each of the RyR isoforms plays in normal muscle physiology. Studies of cells from RyR1- or RyR3-knockout mice have shown that unlike the absence of RyR1, the lack of RyR3 does not result in alteration of electrically evoked Ca2+ release or the contractile function of adult mammalian skeletal muscle fibers (3, 34). In addition, in mouse diaphragm, one of the mammalian muscles that has the highest expression levels of RyR3, there have been no alterations in excitation-contraction (EC) coupling or in the ability of this muscle to contract detected in mature fibers from null mice (1, 5, 9). Moreover, expression of recombinant RyR3 in dyspedic 1B5 myotubes, which do not express any of the RyR isoforms, failed to restore either electrically stimulated or K+-induced Ca2+ release (13, 26), suggesting that RyR3 has virtually no role in initiating or maintaining skeletal EC coupling.
Significant functional differences between RyR3 and RyR1 are observed in their sensitivity to caffeine and 4-CmC, two known direct RyR agonists. Studies in dyspedic 1B5 myotubes have shown that cells expressing RyR3 have a lower threshold and EC50 for caffeine than do cells expressing RyR1 (13, 26), a property that was also observed when these two isoforms were expressed in human embryonic kidney (HEK)-293 cells (28). To the contrary, myotubes expressing RyR1 have a significantly higher sensitivity to 4-CmC than do those expressing RyR3, which has almost no 4-CmC response at all, even at millimolar concentrations (12, 13). These differences suggest that in situ both RyR1 and RyR3 have a characteristic and differential sensitivity to direct agonists.
Despite very low levels of RyR3 in adult mammalian muscle relative to RyR1, a transient increase in expression of RyR3 during postnatal muscle development has been reported (3, 35). Although such variation in expression levels of RyR3 is likely to cause a significant effect on Ca2+ signaling in skeletal muscle, no systematic studies have been performed to correlate these two events. Studies in neonatal mouse skeletal myotubes null for RyR3 have shown that the presence of RyR3 in wild-type (wt) myotubes amplified the Ca2+-induced Ca2+ release (CICR) signal induced by RyR1 (40) and improved the electrical and caffeine-evoked muscle contracture (3, 7). In addition, expression of RyR3 has been associated with augmented spontaneous Ca2+ activity in muscle fibers and cultured myotubes, as well as with increased frequency and size of Ca2+ sparks (6, 7, 37, 38). Although these studies did not define the role of RyR3 in muscle cells, they clearly suggested that the expression of RyR3 in mammalian skeletal muscle, although small compared with RyR1, could play a significant role in Ca2+ homeostasis. The gaps in information regarding the function of RyR3 stress the need for a comprehensive understanding of the functional properties of RyR3 and the role of RyR3 in Ca2+ homeostasis in skeletal muscle.
In the present study, we sought to further characterize the functional differences between RyR1 and RyR3 in myotubes and muscle fibers. Using recombinant RyRs expressed in dyspedic 1B5 myotubes as well as in neonatal muscle fibers, we assessed the effects of variation in expression levels of each isoform on the physiology of cultured myotubes and muscle fibers. Our results show that expression of either isoform had a profound effect on myoplasmic resting free Ca2+ and caffeine sensitivity of the myotubes. Expression of RyR3 always resulted in myotubes with resting free Ca2+ concentration ([Ca2+]r) that was significantly higher, as well as EC50 for caffeine that was significantly lower, than myotubes expressing RyR1. Similarly, in both postnatal and adult skeletal muscle fibers that expressed RyR3, myoplasmic [Ca2+] increased proportionally to the levels of RyR3. These results show that both caffeine sensitivity and increased [Ca2+]r induced by the expression of RyR1 or RyR3 are not the result of variations in expression level of the channels but arise from intrinsically distinct properties that are characteristic of each isoform.
MATERIALS AND METHODS
Cell culture, infection, and Ca2+ imaging.
Dyspedic 1B5 myoblasts (which lack expression of all RyRs) were cultured on Matrigel-coated (BD Biosciences, Franklin Lakes, NJ), 96-well plates (Opticlear; Costar Corning, Acton, MA) in growth medium and differentiated at 37°C in 10% CO2-5% O2 for 5–6 days in DMEM (GIBCO, Grand Island, NY) supplemented with 2% heat-inactivated horse serum. Differentiated myotubes were infected for 2 h with 5, 10, 25, or 100 × 103 herpes simplex virus type 1 virion particles containing either wt-RyR1 or wt-RyR3 cDNA. The number of virion particles corresponded to moieties of infection (MOI) of 0.2, 0.4, 1.0, and 4.0, respectively (36). To avoid intrinsic variability in titer observed between different viral batches, all experiments were performed using a single, large-pooled batch of virus for RyR1 and RyR3. Ca2+ imaging was performed 36–48 h postinfection during the stable phase of transduced protein expression in the myotubes. At that time, myotubes were loaded with 5 μM fluo-4 AM (Molecular Probes, Eugene, OR) for 30 min at 37°C in imaging buffer containing (in mM) 125 NaCl, 5 KCl, 2 CaCl2, 1.2 MgSO4, 6 glucose, and 25 HEPES, pH 7.4, supplemented with 0.05% BSA. After extensive washing with the imaging buffer to remove any excess fluo-4 AM, Ca2+ imaging was performed by collecting the fluorescence signal at 490–500 nm at 15 frames/s with an intensified 12-bit digital charge-coupled device (Mega12; Stanford Photonics, Stanford, CA). Data were analyzed using the QED camera plug-in package (QED Imaging, Pittsburgh, PA).
Crude membranes were prepared from six to eight 100-mm plates of 1B5 myotubes 36 h after transduction with virion particles. Each plate was infected with either 1.0, 2.0, 5.0, or 20.0 × 106 virion particles, which corresponded to MOI equivalent to those used in the 96-well plate format for imaging. Myotubes were harvested as described previously (26).
Neonatal tibialis anterior (TA) muscles were dissected from mice at postnatal days 5, 10, and 15, while adult TA and soleus muscles were harvested at postnatal day 45. All collected muscles were quick-frozen in liquid N2. Microsomal vesicles were prepared by homogenization in a Polytron cell disrupter (Brinkmann Instruments, Westbury, NY) in buffer consisting of 20 mM HEPES, pH 7.0, supplemented with 250 mM sucrose and protease inhibitor cocktail (Complete; Roche, Indianapolis, IN). Whole homogenates from cultured cells and muscles were centrifuged at 1,500 g, and the supernatants were collected and recentrifuged at 100,000 g. The collected membranes were resuspended in 250 mM sucrose and 20 mM HEPES, pH 7.4; frozen in liquid N2; and stored at −80°C. Junctional sarcoplasmic reticulum (JSR) vesicles used as controls were obtained from rabbit skeletal muscle according to the method described by Hidalgo et al. (15).
Gel electrophoresis and immunoblotting.
SDS-polyacrylamide gel electrophoresis (18) was performed on proteins from crude homogenates as described above. Immunoblots were incubated with monoclonal antibody (MAb) 34C (Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA), which recognizes both RyR1 and RyR3 isoforms, or a polyclonal anti-RyR3 antibody that specifically recognizes RyR3 (gift from Dr. T. Murayama). Identification of FK506-binding protein (FKBP-12) was achieved using polyclonal antibody PA1-026A, while sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA)1 and triadin were detected with MAb MA3-911 and MA3-927 (Affinity Bioreagents, Golden, CO), respectively. Calsequestrin was detected with a COOH-terminal polyclonal antibody that recognizes both skeletal and cardiac calsequestrin (gift from Dr. L. Jones). Membranes were then incubated with either anti-mouse or anti-rabbit secondary antibody conjugated to horseradish peroxidase (Jackson ImmunoResearch, West Grove, PA). Immunoreactive proteins were developed with SuperSignal enhanced chemiluminescence substrate (Pierce Biotechnology, Rockford, IL).
[3H]ryanodine binding assay.
Maximal high-affinity [3H]ryanodine binding to crude membrane extracts (0.1 mg/ml) was performed at equilibrium in the presence of 1 M KCl, 100 μM free Ca2+, 20 mM HEPES, pH 7.2, and 10 nM [3H]ryanodine (PerkinElmer Life Sciences, Boston, MA) in the presence of a cocktail of protease inhibitors (Complete). Caffeine dose-response measurements were performed at 0.5 M KCl, 20 mM HEPES, pH 7.4, and 10 nM [3H]ryanodine, with free [Ca2+] adjusted to 500 nM (2). Nonspecific binding was determined by incubating the vesicles with 5 μM unlabeled ryanodine. Separation of bound and free ligand was performed by rapid filtration through GF/B glass fiber filters (Whatman, Middlesex, UK) using a cell harvester (Brandel, Gaithersburg, MD). Filters were washed with ice-cold wash buffer containing 20 mM Tris·HCl, pH 7.4, and placed into vials with 5 ml of scintillation cocktail. The [3H]ryanodine remaining on the filters was quantified using liquid scintillation spectrometry.
Resting free Ca2+ measurements.
[Ca2+]r measurements in cells were performed in myotubes differentiated in 96-well plates as described above. Determination of [Ca2+]r in muscle fibers were performed in situ to avoid muscle hypoxia during measurement. Mice were anesthetized with ketamine and xylazine (50 and 7 mg/kg, respectively). TA and soleus muscles were exposed by dissection from the surrounding connective tissue and then were immersed in Hanks' balanced salt solution (GIBCO, Carlsbad, CA) supplemented with either 2 or 10 mM CaCl2.
Double-barreled Ca2+-selective microelectrodes were prepared as described previously (19). Microelectrodes were assembled using ETH129 resin, which has greater sensitivity to low [Ca2+] compared with ETH1001 (20). Electrodes were calibrated individually, and only those electrodes that had a Nernstian response between pCa3 and pCa7 (28.5 mV/pCa U at 24°C) were used experimentally. To prevent artifacts due to changes in the sensitivity of the microelectrode, individual Ca2+-selective microelectrodes were used for only five or six determinations. If the postexperiment recalibration curve did not agree within 2.5 mV/decade of the initial calibration between pCa6 and pCa8, the data were discarded. Single-myotube impalements were performed using an inverted microscope (Axiovert 10) at ×160 magnification. Because the plasma membrane resting potential was acquired simultaneously with the Ca2+ potential during the entire experiment, any cell damage due to the impalement could be detected quickly. Therefore, all recordings in which Vm was less negative than −55 were discarded (normal range in myotubes is −60 to −55 when measured with 3 M KCl microelectrodes). To rule out the possibility of any potential contribution of extracellular Ca2+ leakage to the measured intracellular [Ca2+]r, a series of control determinations were performed in muscle fibers with two different concentrations of extracellular Ca2+. The potentials from both the 3 M KCl barrel (Vm) and the Ca2+ barrel (VCae) were acquired with high-impedance amplifiers (WPI FD-223). Vm was subtracted electronically from the VCae to obtain a differential signal (VCa) that represented the myoplasmic [Ca2+]r. Vm and VCa potentials were filtered at 200 Hz (WPI-LPF-30) and were analyzed using AxoGraph version 4.8 software.
Ca2+ imaging data were analyzed as described previously (25). Briefly, to compare data from different experiments, fluorescent Ca2+ transients induced by any given caffeine concentration were normalized to the peak amplitude of the maximal response to caffeine from the same cell. Caffeine EC50 for each condition was calculated by fitting a logistic dose-response curve to individual dose responses. Statistically significant differences among the data were evaluated using one-way Kruskal-Wallis ANOVA (nonparametric) (GraphPad Software, San Diego, CA). Data are presented as means±SD or as means±SE.
Infection and RyR expression level.
To evaluate the impact of RyR expression level on myotube function, we infected dyspedic 1B5 myotubes with increasing numbers of virion particles carrying recombinant RyR1 or RyR3 cDNA. Four different virus concentrations were assayed over a 20-fold range of infection capacity (MOI 0.2, 0.4, 1.0, and 4.0). To assess the effect of cDNA copy number on the expression levels of RyRs, the infected myotubes were fixed and immunostained with MAb 34C. The evaluation of the number of positive cells/well, performed with a ×20 lens objective, showed that increasing the number of virions resulted in an increased number of cells expressing either RyR1 or RyR3 (Fig. 1A). The nonlinear relationship between MOI and the percentage of positive cells observed for both RyR1- and RyR3-expressing myotubes suggests that at the highest MOI, myotubes have a high likelihood of being infected by multiple viral particles. Figure 1A also shows that, on the basis of an average of 750 myotubes/well, the percentage of myotubes infected by RyR1-containing viruses was not significantly different from the percentage of those infected with RyR3-containing viruses at equivalent MOI (6.6±1.6%, 16.9±4.7%, 42.8±7.7%, and 80.7±12.5%, respectively, for RyR1; 11.9±3.1%, 27.3±5.6, 56.1±4.4%, and 94.2±8.2%, respectively, for RyR3), showing that there were no significant differences (P > 0.05) in transduction efficiency between these two viruses.
Western blot analysis of crude membrane preparations confirmed that increasing the number of RyR1 or RyR3 virions resulted in equivalent increases in the amount of receptor protein detected using MAb 34C (Fig. 1B). To quantify the amount of RyRs expressed by the myotubes at any MOI, we performed total [3H]ryanodine binding analysis. Figure 1C shows that, consistent with both the increased number of infected cells and the increased amount of receptor protein detected using Western blot analysis, RyRs containing vesicles had a stepwise increase in specific [3H]ryanodine binding. Whereas total binding of RyR1-expressing vesicles increased from 0.076±0.013 to 0.842±0.090 pmol [3H]ryanodine/mg total protein between MOI 0.2 and 4.0, the RyR3-containing vesicles increased [3H]ryanodine binding from 0.082±0.033 to 1.181±0.058 pmol [3H]ryanodine/mg total protein. Interestingly, while myotubes infected with RyR3 had a linear correlation between [3H]ryanodine binding and the amount of virus used for infection at all ranges tested (R2 = 0.999), the relationship was linear in RyR1-expressing myotubes only from MOI 0.2 to 1.0 but did not maintain linearity at MOI 4.0 (Fig. 1C, inset), suggesting that expression of RyR1 by the cell might be more tightly regulated than expression of RyR3.
Immunopattern of expressed RyRs.
To evaluate whether the increased levels of RyR expression had any effects on the subcellular organization of the expressed receptors, we analyzed the immunopattern of expression of myotubes infected with >10-fold differences in MOI. Myotubes were differentiated on coverslips and infected with either RyR1 or RyR3 at MOI 0.4 and 4.0, stained with MAb 34C, and analyzed with a ×100 oil-immersion lens objective. As previously reported (27), three main immunopatterns of expression were readily identified in all groups of cells. These patterns included 1) a punctate pattern, in which the immunostaining appeared as defined regular foci of fluorescence, 2) a reticular pattern, defined as a network-like distribution with no recognizable foci of fluorescence, and 3) a punctate-reticular pattern, a combination of the two preceding patterns. Analysis of immunostained myotubes revealed that the three patterns of expression were always detected, regardless of the amount of virus used for infection (Table 1). At MOI 0.4, the punctate pattern was always predominant, accounting for 88.0% of the population of RyR1-expressing myotubes and 84.2% of RyR3-expressing cells. The punctate-reticular pattern, on the other hand, accounted for 10.4% of the RyR1-expressing myotubes and 12.7% of RyR3 population. The reticular pattern represented only 1.6% of RyR1- and 3.1% of RyR3-expressing cells. In myotubes infected at MOI 4.0, the punctate immunopattern of expression was still predominant for both isoforms, accounting for >60% of the cell population. However, at MOI 4.0, there was an increase in the percentage of punctate-reticular and reticular immunopatterns of expression in both RyR1- and RyR3-expressing cells. This was particularly apparent in cells overexpressing RyR3 (MOI 4.0), in which the punctate-reticular immunopattern was twofold and the reticular pattern was fourfold the percentage observed in myotubes infected at MOI 0.4. In addition, in myotubes overexpressing RyR3, the percentage was twofold that of the punctate-reticular and reticular immunopatterns in myotubes expressing RyR1 at similar MOI (Table 1).
Caffeine dose-response curve in myotubes expressing RyR1 or RyR3.
As a means of evaluating whether increased expression levels of RyRs had an effect on Ca2+ release, we tested the sensitivity of the RyR-expressing myotubes to caffeine, which is a direct agonist of RyR. Infected myotubes were loaded with 5 μM fluo-4 AM and exposed to increasing caffeine concentrations (0.1–40 mM). Figure 2 shows representative fluorescent images of caffeine dose responses of myotubes infected at MOI 0.2, 0.4, 1.0, and 4.0. Analysis of the fluorescence images showed that RyR1-expressing myotubes had a response threshold at between 0.5 and 1.0 mM caffeine and maximal Ca2+ release responses at 20 mM caffeine. There was no difference in the caffeine dose-response profile in RyR1-expressing cells as MOI was increased. By contrast, RyR3-expressing cells had a lower threshold for caffeine than did the cells expressing RyR1 at all MOI tested. The threshold for caffeine-induced Ca2+ release in RyR3-expressing cells was between 0.1 and 0.5 mM caffeine, and maximal Ca2+ release was induced at 3 mM caffeine. In addition, at MOI 4.0, unlike RyR1-expressing cells, RyR3 infected myotubes had a significant reduction in their threshold for caffeine to <0.1 mM (Fig. 2A).
Caffeine-induced Ca2+ transients were normalized to the Ca2+ transient induced by the maximal caffeine concentration and then quantified, followed by calculation of EC50 values. Analysis of average EC50 values for cells demonstrated that there was no statistically significant difference in EC50 among cells expressing RyR1 infected at different MOI (Fig. 2B and Table 2). In contrast to this observation, myotubes expressing RyR3 had a stepwise shift to the left in caffeine sensitivity as MOI increased (Fig. 2B), which was statistically significant at 1.0 and 4.0 MOI (P < 0.01) (Table 2). These results suggest that variation in expression levels of RyR3 may have a variable impact on Ca2+ signaling and that RyR1 activity is highly regulated, regardless of expression level. In addition, these results showed that regardless of expression level, RyR1-expressing cells are always less sensitive than RyR3-expressing cells to caffeine.
To further examine the difference in caffeine sensitivity observed between RyR1 and RyR3, we performed [3H]ryanodine binding analysis on crude membrane preparations from myotubes infected at MOI 1.0 to verify the differential effect of caffeine on channel activity. To ensure an appropriate level of channel activation, all binding assays were performed in the presence of 500 nM free Ca2+ and 0.5 M KCl and exposed to increasing concentrations of caffeine (0.5–50 mM). As shown in Fig. 3, under these conditions, RyR1-containing vesicles showed a significantly higher EC50 for caffeine than did the vesicles expressing RyR3 (4.87±0.35 and 1.93±0.36 mM, respectively; P < 0.001), verifying that RyR3 has an intrinsically higher sensitivity to caffeine. In addition, a clear difference between these isoforms was evident in the Hill slope (nH = 1.6 vs. 2.3 for RyR1 vs. RyR3, respectively), suggesting that caffeine activation has a more cooperative action on RyR3 than on RyR1. This, too, supports the notion that these two isoforms exhibit a differential response with regard to caffeine sensitivity. These results are also consistent with those obtained in intact cells and suggest that the differences in caffeine sensitivity between RyR1 and RyR3 represent a unique property of each isoform and that this observation is not the consequence of differences in the expression levels of each isoform.
Expression levels of other SR proteins.
To test whether the increase in expression levels of RyRs induced by increased concentrations of virion particles have any effect on the expression levels of other Ca2+-regulatory proteins, we evaluated the expression levels of FKBP-12, triadin, calsequestrin, and SERCA1 in all groups of cells. Western blot analysis of crude membrane preparations of myotubes infected with RyRs at all tested MOI showed that changing the expression level of RyR1 or RyR3 did not result in noticeable alterations in expression of FKBP-12, triadin, or calsequestrin (Fig. 4). Interestingly, a slight reduction in SERCA1 expression was observed in both RyR1- and RyR3-expressing myotubes as the expression levels of RyR increased. Densitometric analysis of Western blots revealed that whereas cells infected with RyR1 showed reductions in SERCA1 expression from 15.2±2.7% to 35.7±6.5% between 0.4 and 4.0 MOI (relative MOI, 0.2), the RyR3-expressing myotubes showed a reduction of SERCA1 expression from 10.6±4.3% to 41.4±16.2% for the same range of MOI (n = 3; P > 0.05 vs. RyR1). These data confirm that infection of myotubes with increasing concentrations of cDNA-containing virions resulted in the increased expression levels of RyRs without inducing a major reshaping of the triadic protein machinery.
RyR expression levels and [Ca2+]r.
To test whether the amount of RyRs expressed in myotubes had any impact on the myoplasmic [Ca2+]r, we measured [Ca2+]r in each group of cells using Ca2+ microelectrodes. Figure 5 summarizes the results of [Ca2+]r measurement in myotubes infected with each RyR isoform at various virus titers. These data show that [Ca2+]r was ∼110 nM in myotubes expressing RyR1, with small variation among individual cells, and that the average [Ca2+]r remained constant as MOI was increased. Even at MOI 4.0, a level at which SERCA1 expression was reduced by 35%, [Ca2+]r was not different from that at MOI 0.2, with mean [Ca2+]r levels of 113±8.8 and 116.8±7.6 nM for MOI 0.2 and 4.0, respectively. By comparison, myotubes expressing RyR3 consistently had higher mean [Ca2+]r levels and greater variability in individual [Ca2+]r measurements than did the cells infected with RyR1 at equivalent MOI. Although there was no significant difference in the average [Ca2+]r observed among the groups of cells infected with RyR3 virus from MOI 0.2 to 1.0, there was an upward trend in the mean [Ca2+]r (192.4±24 and 213.6±31 nM at MOI 0.2 and 1.0, respectively; P > 0.05) and increased intermyotube variability. In myotubes infected with RyR3 virions at MOI 4.0, a significant increase in [Ca2+]r was observed (336.6±46 nM; P < 0.001) compared with cells infected at MOI 1.0. These results demonstrate that 1) regardless of the level of expression, myotubes expressing RyR3 always had a mean [Ca2+]r higher than did myotubes expressing RyR1 (∼192–336 vs. ∼110 nM) and 2) overexpression of RyR3, but not of RyR1, resulted in an increase in average myoplasmic [Ca2+]r, supporting the hypothesis that RyR1 activity is negatively regulated in skeletal muscle cells (23).
Coexpression of RyR1 and RyR3 in skeletal muscles.
With the exception of diaphragm and soleus muscles, significant expression of RyR3 has not been detected in adult murine skeletal muscles. However, coexpression of RyR1 and RyR3 does occur in peripheral skeletal muscles during early stages of postnatal development, when changes in expression levels of both isoforms have been reported for several muscles (3, 8, 35). Whereas there is a sustained increase in RyR1 expression in fast-twitch murine skeletal muscle from embryonic day 18 to postnatal day 15, RyR3 is only transiently upregulated during the postnatal period, reaching a peak around postnatal day 15, after which its expression declines to undetectable levels in adult muscle (postnatal days 45–50). On the basis of the facts that 1B5 myotubes expressing RyR3 had increased [Ca2+]r levels and that this effect was related to the amount of expression of RyR3, we hypothesized that variation in [Ca2+]r levels could occur in developing skeletal muscle during the period when the coexpression of RyR3 and RyR1 takes place. Likewise, we hypothesized that myoplasmic resting [Ca2+]r levels also would be higher in adult muscles still expressing RyR3 compared with those that did not. To evaluate this hypothesis, we measured [Ca2+]r in tibialis anterior (TA) muscles from anesthetized mice in situ 5, 10, and 15 days after birth and compared these data with in situ measurements in the TA and soleus muscles of adult mice (45–49 days old).
To analyze the expression levels of RyR1 and RyR3 at each stage of development, Western blot analysis was performed in studied muscles collected immediately after Ca2+ measurements. Crude membrane preparations from each muscle were obtained and analyzed with MAb 34C and anti-RyR3 antibodies to assess the total RyRs and RyR3, respectively. Figure 6A shows that although 34C antibody detected significant expression levels of total RyRs at all stages of TA muscle development, the adult muscles demonstrated only a small increase in total expression compared with the neonatal muscles. No significant variability in total RyR expression was evident between muscles from postnatal days 5 and 15 samples. However, there were significant variations in RyR3 expression among all tested stages. Figure 6A demonstrates that there was a trace of RyR3 expression in TA muscles at postnatal day 5, followed by stepwise increases in RyR3 expression at postnatal days 10 and 15 (3.5- and 4.5-fold increases, respectively, relative to 5-day-old muscle). There was no detectable RyR3 expression observed in adult muscles. Although this result is consistent with previous findings showing no detectable expression of RyR3 in microsome preparations of adult TA muscle (3), very low levels of RyR3 expression have been reported with the use of Western blot analysis and in situ hybridization in adult mouse and rat TA muscles (8).
There was a significant amount of expression of RyR3 in crude membrane vesicles from soleus (slow twitch) muscle, which was previously shown to express higher levels of RyR3 than the TA muscle. The comparison of RyR expression shown in Fig. 6B demonstrates that although MAb 34C was able to detect significant expression of RyR1 in both types of adult muscle, the anti-RyR3 antibody revealed detectable levels of RyR3 only in soleus muscle. Control lanes exposed to a polyclonal antibody against GAPDH indicated that equivalent levels of protein were loaded in each lane, suggesting that the expression of total RyR was significantly higher in TA muscle and that if there was any RyR3 expression, those levels were below the detectable levels for this particular antibody.
Effects of coexpression of RyR1 and RyR3 on [Ca2+]r.
Figure 7A summarizes the measurements of [Ca2+]r collected from several individual fibers of 6–10 muscles per group. As expected, a significant increase in [Ca2+]r was detected as the level of RyR3 expression increased in TA fibers. This increase was transient in nature because it declined as the animals reached maturity. Although no differences in average [Ca2+]r were evident between fibers from postnatal day 5 and adult muscles (107.3±4.8 and 111.8±3.0 nM, respectively; P > 0.05), a significant increase in average [Ca2+]r was observed in fibers from postnatal days 10 and 15 muscles (163.2±12.7 and 221.5±9.3 nM, respectively) compared with fibers from postnatal day 5 and adult muscles. On the other hand, measurement of [Ca2+]r in adult soleus muscle showed that [Ca2+]r in this muscle was significantly higher than that in adult TA fibers (166.4±8.6 vs. 111.8±3.0 nM, respectively; P < 0.005). Interestingly, just as in 1B5 myotubes, there was a wide variation in individual [Ca2+]r measurements, which we attributed to interfiber differences in the expression of RyR3. There was a large variation in individual [Ca2+]r measurements in TA fibers at postnatal days 10 and 15 and, although to a lesser extent, in adult soleus muscle, suggesting that there was a broad range of expression of RyR3 from fiber to fiber as suggested by Flucher et al. (14).
To rule out any contribution of extracellular Ca2+ to the [Ca2+]r, with Ca2+ microelectrodes we performed paired determinations in which the extracellular Ca2+ in the bath solution was increased from 2 to 10 mM. As shown in Fig. 7A, the [Ca2+]r values determined in adult and 15-day-old fibers at high levels of extracellular Ca2+ were not different from those detected at lower Ca2+ levels. This result suggests that the increased changes in [Ca2+]r observed in developing TA fibers were not the result of extracellular Ca2+ drawn into the cell with the microelectrode or of Ca2+ leakage through the plasma membrane around the microelectrode. Parallel determination of the membrane potential showed a sustained increase in resting potential from postnatal day 5 to adult TA fibers (−59.9 to −83.6 mV) and adult soleus fibers (−80.4 mV), a result that is consistent with reports of membrane hyperpolarization during postnatal development of mouse extensor digitorum longus fibers (39).
Although the role of RyR1 in Ca2+ homeostasis in skeletal muscles is clear, the reasons why RyR3 is transiently expressed during development and the role that RyR3 plays in muscle function are still uncertain. Because it is almost impossible to study the independent functioning of RyR1 and RyR3 except in null models, in which one or both isoforms are artificially expressed and most frequently overexpressed, it is also important to understand the consequences of overexpression on the function of skeletal muscle cells.
In the present study, we have analyzed the effects of expression levels of RyR1 and RyR3 on Ca2+ response in 1B5 myotubes. It was established that over a broad range of RyR expression levels, there were very few differences in the overall phenotype of cultured myotubes expressing either isoform. The expression of either isoform had a substantial impact on caffeine sensitivity (both restored it) and [Ca2+]r (both increased it) compared with null myotubes. In addition, the caffeine sensitivity and [Ca2+]r measured in RyR3-expressing myotubes were significantly higher than the same parameters measured in RyR1-expressing cells. We also observed increases in [Ca2+]r in situ in postnatal days 10 and 15 TA muscle fibers and in adult soleus fibers. This difference appeared to be closely correlated with the expression levels of RyR3 observed at each stage.
Expression of RyR3 and its effect on myoplasmic [Ca2+]r.
As viral MOI increased, we observed not only that a larger number of cells were transduced but also that a higher number of myotubes were infected by more than one viral particle, and thus more cells had the potential to overexpress their respective proteins. However, our Ca2+ microelectrode measurements demonstrated that RyR3-expressing myotubes had a cytosolic [Ca2+]r that, on average, was ≥2-fold that in myotubes expressing RyR1. Because total [3H]ryanodine binding was determined to be identical between RyR1- and RyR3-expressing cells at between MOI 0.2 and 1.0, it seems unlikely that the difference in [Ca2+]r resulted from the different levels of RyRs being expressed. Instead, this result suggests that the level of [Ca2+]r must strongly depend on the unique characteristics of each RyR isoform. On the basis of our measurements of resting free Ca2+ (Figs. 5 and 7), it is also clear that RyR3-expressing cells, in addition to having consistently higher [Ca2+]r than cells expressing only RyR1, both 1B5 myotubes and muscle fibers demonstrated wider dispersion of [Ca2+]r than did the cells expressing RyR1. It is our hypothesis that this phenomenon could be caused by the variable expression levels of RyR3 in individual myotubes or myofibers. This is consistent with immunohistochemical studies of developing murine muscle using anti-RyR3 antibodies, which have shown that there was an induction followed by a reduction of RyR3-expressing fibers during postnatal muscle development and that this induction followed by reduction does not take place gradually and simultaneously in all fibers, but instead occurs rapidly in some fibers and very slowly in others (14). Alternatively, the high variability of interfiber and myotube [Ca2+]r values may suggest that cells expressing RyR3 are variably hyperactive, constantly releasing puffs of Ca2+ into the cytosol. This hypothesis is consistent with studies of Ca2+ spark activity of RyR3-expressing cells that have indicated this to be the case. Interestingly, the high dispersion in individual [Ca2+]r measurements observed in our study in both myotubes and RyR3-expressing muscle fibers is in agreement with the report that embryonic myotubes that express RyR3 have considerably more variability in the size and kinetics of their Ca2+ sparks than do adult cells (7). This finding is also consistent with the fact that our [Ca2+]r data do not appear to reveal a defined population of fibers with either high or low [Ca2+]r as would be expected from muscles in which only some fibers express RyR3 and the others express only RyR1 (14). Significantly, at MOI 1.0, RyR3-expressing myotubes had [Ca2+]r values between 178 and 345 nM, a range similar to the 136–363 nM concentration observed in 15-day-old TA fibers, which suggests that in terms of Ca2+ regulation, RyR3 in 1B5 myotubes demonstrated behavior similar to its behavior under physiological conditions. The importance of this result is that it 1) rules out the possibility that high [Ca2+]r detected in myotubes may be the result of toxic expression of RyR3 and 2) suggests that induction of high [Ca2+]r associated with RyR3 expression may be physiologically relevant.
Caffeine sensitivity of RyR1- and RyR3-expressing myotubes.
It was reported previously that caffeine sensitivity of RyRs is highly Ca2+ dependent (30, 32, 41). However, it is unlikely that the higher [Ca2+]r induced by RyR3 was the only factor responsible for the differential caffeine sensitivity observed between RyR3- and RyR1-expressing cells in this study. This is supported by the [3H]ryanodine binding data in the present study, which demonstrated that there were significant differences in caffeine sensitivity between RyR1 and RyR3 even when the free [Ca2+] was fixed at a constant suboptimal level (500 nM). Nevertheless, under this condition, a significant difference in affinity to Ca2+ activation between both isoforms could also explain the divergence in caffeine sensitivity observed between RyR1 and RyR3, both in vivo and in vitro. Although it was previously suggested that RyR1 and RyR3 have a significant difference in Ca2+ sensitivity, these reports are somehow contradictory (see below). In this regard, [3H]ryanodine binding studies conducted under the same experimental conditions assayed in the present work indicated that RyR1 and RyR3 have very little difference in sensitivity for Ca2+ activation (Voss A., unpublished data). Thus it is highly unlikely that the difference in caffeine sensitivity in the present study is just the result of differential Ca2+ affinity between the two isoforms. Nevertheless, it is likely that the higher [Ca2+]r in RyR3-expressing cells was responsible for at least some of the increase in caffeine sensitivity observed in these cells.
The hypothesis that both an innate difference in caffeine sensitivity between the two isoforms and the increase in [Ca2+]r are responsible for the large difference in sensitivity observed in intact cells is supported by the fact that under the conditions used for binding, the difference in caffeine EC50 for RyR1 and RyR3 was only 2.5-fold (4.87±0.35 vs. 1.93±0.36 mM), which was significantly smaller than the 4.4-fold difference observed in intact cells infected at MOI 1.0 (3.54±0.34 vs. 0.79±0.19 mM). These results are in agreement with our previous findings (12, 13, 26) and those of others (6, 28, 29). On the basis of these results, it can be inferred that the differences in caffeine sensitivity observed in myotubes expressing RyR1 and RyR3 come from a true intrinsic difference between the two isoforms. Our data, however, are not in agreement with those of Murayama and Ogawa (23), who reported no difference in caffeine sensitivity between purified RyR1 and RyR3 isolated from bovine diaphragm. It is likely that this discrepancy arises from differences in the experimental conditions used in the two studies. The conditions the Murayama and Ogawa study were strongly activating, and thus it is possible that they any difference between the two isoforms may have been masked to further activation. Alternatively, the RyR3 isoform expressed in diaphragm may correspond to a different splice variant than the one expressed in uterus or in mouse skeletal muscle that were used in our present study. Recent studies in rabbit tissues have demonstrated that multiple splice variants of RyR3 are expressed in uterus and diaphragm and that some of these variants have significantly reduced caffeine sensitivity. Thus tissue-specific expression of RyR3 splice variants might offer an explanation for the heterogeneity of caffeine response of RyR3 in different tissues and cells (17).
Functional differences between RyR1 and RyR3.
Several studies have revealed significant functional differences between RyR1 and RyR3, particularly in their Ca2+ and caffeine sensitivity. Several lines of evidence seem to support that RyR3 is less sensitive than RyR1 to Ca2+ inhibition (4, 16, 23, 33), although different experimental approaches have led various investigators to very different conclusions regarding its sensitivity to Ca2+ activation. Using single-channel analysis, Sonnleitner et al. (33) reported two populations of Ca2+ channels in SR vesicles from diaphragm muscle, one with properties similar to those of RyR1 and another, which they suggested was RyR3, that was less sensitive than RyR1 to Ca2+ activation. Consistent with this finding, [3H]ryanodine binding studies in purified receptors from diaphragm have consistently shown that RyR3 has a significantly lower sensitivity than RyR1 to Ca2+ activation (21–23). However, this difference seems to be reversed when the receptors are fused into lipid bilayers in which Ca2+ EC50 for activation measured for RyR3 was found to be lower than that of RyR1 (21, 24). Studies of RyR3 from mouse parotid acini (10) and rabbit uterine RyR3 heterologously expressed in HEK-293 cells (4) have suggested that RyR3 has the same apparent sensitivity as RyR1 for Ca2+ activation. Whether these seemingly large differences in observed RyR3 Ca2+ sensitivity are due simply to different experimental conditions used in these studies or whether they reflect real intrinsic differences of various splice variants of RyR3 has yet to be resolved. A detailed analysis of the receptor splice variants at the single-channel level of RyR3 should be able to provide a conclusive answer.
In this regard, a detailed study of single-channel properties of the recombinant RyR3 used in the present work showed that although overall RyR3 shared several similarities with RyR1, at near-[Ca2+]r (200 nM) levels, RyR3 exhibited significantly longer mean open time (4) and an increase in subconductance behavior compared with RyR1 (13). These differences in gating kinetics between the two isoforms are consistent with the apparent higher Ca2+ leakage rates, which would lead to the higher [Ca2+]r that we observed in RyR3-expressing cells. This difference is also consistent with the fact that RyR3-expressing cells have a higher frequency and greater magnitude of Ca2+ sparks than do cells expressing only RyR1 (37, 38). Likewise, in studies of primary myotubes null for RyR3, a similar contribution of RyR3 to the magnitude (6, 7) and duration (31) of the Ca2+ sparks has been described, suggesting that important differences between the two isoforms exist with regard to Ca2+ release behavior. Our data support this hypothesis and suggest that this difference in spontaneous activity could be one of the major contributors to the difference in myoplasmic free Ca2+ that we observed between RyR1- and RyR3-expressing cells.
In summary, the results of this study demonstrate that the expression level of RyR3 in myotubes has a significant effect on myoplasmic free [Ca2+]r. These findings suggest that the coexpression of RyR3 in vivo at various levels might underlie a tightly regulated mechanism by which muscles fine-tune cytosolic free [Ca2+]. Such control would facilitate the diversity of cellular responses that muscle cells undergo during their early development.
This work was supported by National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant P01 AR-47605 (to P. D. Allen).
We thank Dr. Y. Wang and R. Hirsh for providing technical expertise and help in virus packaging; Dr. S. R. W. Chen and Dr. T. Murayama for kindly providing RyR3 cDNA and anti-RyR3 antibody, respectively; and Dr. A. Shtifman for critical reading of the manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society