Biochemical Journal

Research article

S165F mutation of junctophilin 2 affects Ca2+ signalling in skeletal muscle

Jin Seok Woo , Ji-Hye Hwang , Jae-Kyun Ko , Noah Weisleder , Do Han Kim , Jianjie Ma , Eun Hui Lee

Abstract

JPs (junctophilins) contribute to the formation of junctional membrane complexes in muscle cells by physically linking the t-tubule (transverse-tubule) and SR (sarcoplasmic reticulum) membranes. In humans with HCM (hypertrophic cardiomyopathy), mutations in JP2 are linked to altered Ca2+ signalling in cardiomyocytes; however, the effects of these mutations on skeletal muscle function have not been examined. In the present study, we investigated the role of the dominant-negative JP2-S165F mutation (which is associated with human HCM) in skeletal muscle. Consistent with the hypertrophy observed in human cardiac muscle, overexpression of JP2-S165F in primary mouse skeletal myotubes led to a significant increase in myotube diameter and resting cytosolic Ca2+ concentration. Single myotube Ca2+ imaging experiments showed reductions in both the excitation–contraction coupling gain and RyR (ryanodine receptor) 1-mediated Ca2+ release from the SR. Immunoprecipitation assays revealed defects in the PKC (protein kinase C)-mediated phosphorylation of the JP2-S165F mutant protein at Ser165 and in binding of JP2-S165F to the Ca2+ channel TRPC3 (transient receptor potential cation canonical-type channel 3) on the t-tubule membrane. Therefore both the hypertrophy and altered intracellular Ca2+ signalling in the JP2-S165F-expressing skeletal myotubes can be linked to altered phosphorylation of JP2 and/or altered cross-talk among Ca2+ channels on the t-tubule and SR membranes.

  • Ca2+ homoeostasis
  • junctional membrane complex
  • junctophilin 2
  • muscle hypertrophy
  • protein kinase C (PKC)
  • transient receptor potential cation canonical-type channel 3 (TRPC3)

INTRODUCTION

During EC (excitation–contraction) coupling in striated muscle (cardiac and skeletal muscle), the DHPR (dihydropyridine receptor) Ca2+ channels on t-tubule (transverse-tubule) membranes are activated by membrane depolarization and allow RyR (ryanodine receptor) Ca2+ channels on the SR (sarcoplasmic reticulum) to release Ca2+ from the SR into the cytoplasm to induce muscle contraction [13]. In addition to the central theme of EC coupling, TRPC3 (transient receptor potential canonical-type cation channel 3), which is a Ca2+ channel on the t-tubule membrane required for extracellular Ca2+ entry, is necessary for the full skeletal EC coupling gain (duration and maintenance) [36]. The junctional membrane complex, the diad and triad junctions, where the t-tubule and SR membranes are juxtaposed, provides the structural context for proper arrangement of Ca2+ channels and efficient EC coupling [7,8].

JPs (junctophilins) contribute to the formation of junctional membrane complexes in excitable cells by interacting with plasma (or t-tubule in muscle cells) membranes via their N-terminal MORN (membrane occupation and recognition nexus) motifs and with endoplasmic reticulum (or SR in muscle cells) via the C-terminal transmembrane domain [3,79]. Four JP subtypes have been identified: JP1 is expressed in skeletal muscle; JP2 is expressed abundantly in all muscle types (skeletal, cardiac and smooth muscle); and JP3 and JP4 are expressed in the brain [911]. JP1- and JP2-knockout mice die shortly after birth or during the embryonic stage respectively, due to deformation of the junctional membrane complexes and disrupted Ca2+ homoeostasis [7,8]. JP1-overexpressing transgenic mice develop additional subcellular tubular structures rolled up together with SR membranes without any change in cellular morphology [12]. JP2 is known to interact with TRPC3 via a region (amino acids 143–234 of JP2) in skeletal muscle [5,13]. The functional relationship between JP and Ca2+ homoeostasis has also been examined. Suppression of JP1 and JP2 in skeletal myotubes leads to uncontrolled intracellular Ca2+ release and decreased extracellular Ca2+ entry [14], whereas increased JP2 expression in TRPC3-knockdown mouse skeletal myotubes was associated with a decreased level of Ca2+ transients through RyR1 [4].

HCM (hypertrophic cardiomyopathy), a significant cause of sudden cardiac death, is a genetic disorder related to mutations of muscle genes such as those encoding actin, myosin, tropomyosin, troponin, Z-disc proteins and others [15]. It has also been reported that JP2 is associated with the development of HCM. For example, down-regulation of JP2 has been observed in hypertrophic and dilated cardiomyopathic mouse models [16] and in a pressure-overload-induced cardiac hypertrophic rat model [17]. In HCM patients, five JP2 single mutants (S101R, Y141H, S165F, R436C and G505S) have been identified [18,19]. Among these human HCM-associated JP2 mutants, Y141H and S165F significantly disrupt Ca2+ signalling when expressed in the cardiac myocyte cell line HL-1 [18]. Interestingly, Ser165 in JP2 is the only putative phosphorylation site for PKC (protein kinase C) [20].

Until now, the effects of JP2 mutations on skeletal muscle function have not been examined. In the present study, we examined the effects of overexpressing JP2-S165F on Ca2+ signalling in skeletal muscle and found that JP2-S165F had a dominant-negative inhibition over the native JP2. This inhibitory action resulted in several consequences, including significant myotube hypertrophy and decreases in both in K+-induced (i.e. skeletal EC coupling) and caffeine-induced SR Ca2+ release (RyR1-dependent Ca2+ release) without a concomitant change in the SR Ca2+ content. Furthermore, we suggest that defects in the PKC-dependent phosphorylation state of JP2-S165F, which resulted in a decreased binding to TRPC3, play a mechanistic role in these effects.

EXPERIMENTAL

Materials

FBS (fetal bovine serum), F-10 nutrient mixture, L-glutamine, penicillin/streptomycin and bFGF (basic fibroblast growth factor) were obtained from Invitrogen. A monoclonal anti-RyR1 antibody (against purified RyRa and RyRb from chicken skeletal muscle) was provided by Dr J. Airey and Dr J. Sutko (Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA, U.S.A.) and was used at a 1:5000 dilution in the immunoblot assay. Matrigel was obtained from BD Biosciences. CPA (cyclopiazonic acid), caffeine, KCl, GF-109203X, horse serum and collagen were obtained from Sigma–Aldrich. An anti-phosphoserine antibody (used at 20 μg of antibody per mg of protein for the immunoprecipitation assay) was obtained from Chemicon International. The anti-JP2 antibody (used at 1:2000 in the immunoblot assay) and anti-TRPC3 antibody (used at 1:800 for immunoblotting and 1:250 for co-immunoprecipitation) were obtained from Santa Cruz Biotechnology. Anti-calsequestrin antibody (used at 1:2000 in the immunoblot assay), antiDHPR antibody (against the α1 subunit of DHPR; used at 1:500 in the immunoblot assay) and anti-SERCA1 (sarcoplasmic/endoplasmic reticulum Ca2+-ATPase 1) antibody (used at 1:1000 in the immunoblot assay) were obtained from Affinity BioReagents. Protein G–Sepharose 4 Fast Flow affinity beads were obtained from Amersham Biosciences. FuGENE6 was obtained from Roche. Fura-2-AM (fura-2-acetoxymethyl ester) and fluo-4-AM (fluo-4-acetoxymethyl ester) were obtained from Invitrogen.

Cell culture and expression of wild-type JP2, JP2-S165F and JP2-Y141H

Primary wild-type myoblasts were derived from mouse skeletal muscle as described previously [21]. All surgical interventions and presurgical and postsurgical animal care was provided in accordance with the Laboratory Animals Welfare Act, the Guide for the Care and Use of Laboratory Animals and the Guidelines and Policies for Rodent Survival Surgery provided by the IACUC (Institutional Animal Care and Use Committee) for the College of Medicine, The Catholic University of Korea, Republic of Korea. Primary wild-type myoblasts were cultured on 10-cm dishes coated with collagen in growth medium (F-10 nutrient mixture) containing 20% (v/v) FBS, 100 units/ml penicillin, 100 mg/ml streptomycin, 2 mM L-glutamine and 20 nM bFGF, at 37 °C in a 5% CO2 incubator. For differentiation into myotubes, wild-type myoblasts were re-plated either on to 10-cm dishes (to prepare myotube lysate) or 96-well plates (for single myotube Ca2+ imaging experiments) coated with Matrigel. When cells reached ~ 70% confluence, the growth medium was replaced with differentiation medium [low-glucose DMEM (Dulbecco's modified Eagle's medium) containing 5% (w/v) heat-inactivated horse serum] without growth factors, and the cells were placed in an 18% CO2 incubator to induce differentiation. After 4 days of culture in differentiation medium, myotubes were transfected with cDNA encoding wild-type JP2, JP2-S165F or JP2-Y141H in an expression vector (pCMS-RFP; a mixture of 30 μl of FuGENE6 and 20 μg cDNA per 10-cm dish, or the same ratio of components in the wells of a 96-well plate) for 4 h; construction of the expression vectors was described previously [18]. Myotubes were either imaged or disrupted to prepare whole myotube lysates at 36 h post-transfection, at which time approx. 60% of myotubes had been transfected, as estimated by the RFP (red fluorescent protein) signal. Only myotubes containing more than four nuclei were selected for use in single myotube Ca2+ imaging experiments. When used, 10 μM GF-109203X was dissolved in the differentiation medium (from a 20 mM stock dissolved in methanol) for 30 min prior to preparation of the whole myotube lysates [22,23]. Methanol (0.05%) alone had no effect on the phosphorylation of JP2 proteins.

Single myotube Ca2+ imaging experiments and measurement of myotube diameter

The myotubes expressing either wild-type JP2 or JP2-S165F (cultured on 96-well plates) were loaded with 5 μM fluo-4-AM (or fura-2-AM for measurement of resting cytosolic Ca2+ levels) in imaging buffer [25 mM Hepes, pH 7.4, 125 mM NaCl, 5 mM KCl, 2 mM KH2PO4, 2 mM CaCl2, 6 mM glucose, 1.2 mM MgSO4 and 0.05% BSA (fraction V)] at 37 °C for 45 min. Each well of the 96-well plate was then washed three times with 200 μl of imaging buffer. The myotubes were transferred on to an inverted stage microscope (Nikon Eclipse TS100) equipped with a Nikon ×40 oil-immersion objective (numerical aperture 1.30). The microscope was modified to incorporate a pair of separate three-dimensional micromanipulators on either side of the vertical post that holds the condenser. Images of the myotubes were captured using a high-speed monochromator with a 75-W Xenon lamp (FSM150Xe; Bentham Instruments) as the light source and a 12-bit charge-coupled device camera (Pixelfly II Standard 200XS; DVC). Before starting the single myotube Ca2+ imaging experiments, images of myotubes were captured for size comparison (especially diameter). Fluo-4 in the myotubes was excited at 494 nm and fluorescence emission was measured at 516 nm. Results were displayed and analysed using InCyt Im1™ image acquisition and analysis software (version 5.29; Intracellular Imaging). Fura-2 in the myotubes was excited at 340 nm and 380 nm, and fluorescence emission was measured at 510 nm using high-speed InCyt Im2™ image acquisition and analysis software (version 5.29; Intracellular Imaging). Intracellular Ca2+ concentrations were calculated as described by Grynkiewicz et al. [24] using 225 nM as the Ca2+–fura-2 dissociation constant. Caffeine or KCl was dissolved in the imaging buffer and applied via an auto-perfusion system with an eight-channel perfusion pipette (AutoMate Scientific). CPA was dissolved in DMSO (<0.05%) and manually applied to myotubes. DMSO (0.05%) alone had no effect on Ca2+ transients. To examine the effects of GF-109203X on the KCl or caffeine response and the resting cytoplasmic Ca2+ level in myotubes transfected with wild-type JP2, the myotubes were pre-treated with 10 μM GF-109203X in the imaging buffer (20 mM stock dissolved in methanol) for 10 min [23]. Methanol (0.05%) alone had no effect on Ca2+ transients. Myotubes transfected with wild-type JP2 and not treated with GF-109203X were used as controls. The three-dimensional micromanipulators were used to position a pair of capillaries precisely into each well; one capillary served as an inlet for delivery of the desired medium and the other as an outlet for removal of excess medium via vacuum suction. Thus the amount of medium in each well was kept constant. The perfusion outlet was positioned approx. 1.5 mm above the imaged myotubes to allow rapid perfusion of the imaged area. To analyse the Ca2+ transient obtained from the Ca2+ imaging experiments, both the peak amplitude and peak area (results not shown), which exhibited similar increases or decreases, were considered. For the relatively long-term Ca2+ transients induced by CPA treatment, the peak area was analysed.

Immunoprecipitation assay

For the immunoprecipitation assay [25], myotubes were solubilized in lysis buffer [10 mM Tris/HCl, pH 7.4, 1% (v/v) Triton X-100, 1 mM Na3VO4, 10% (v/v) glycerol, 150 mM NaCl, 5 mM EDTA and protease inhibitors (1 μM pepstatin, 1 μM leupeptin, 1 mM PMSF, 20 mg/ml aprotinin and 1 μM trypsin inhibitor)] overnight at 4 °C with gentle mixing (by adding 300 μl of lysis buffer to myotubes in a 10-cm culture dish). The solubilized lysate was then centrifuged at 1500 g for 30 s to remove insoluble matter. Solubilized myotube samples (800 μg of total protein) were incubated with anti-phosphoserine or anti-TRPC3 antibody overnight at 4 °C, followed by incubation with Protein G–Sepharose beads for 4 h at 4 °C. Beads were washed five times with lysis buffer without Triton-X. Bound proteins were eluted by boiling in SDS sample buffer and were any analysed by SDS/PAGE (10% gels) followed by immunoblotting with anti-TRPC3 or anti-JP2 antibody.

Immunoblotting

For the immunoblot assays, the proteins on the gel were transferred on to a PVDF membrane at 100 V for 2 h. The membranes were blocked using 5% (w/v) non-fat milk dissolved in PBS for 1 h, incubated with primary antibody, washed three times with PBST (PBS containing 0.1% Tween 20) and then incubated with a secondary horseradish peroxidase-conjugated anti-(mouse Ig) or anti-(rabbit Ig) antibody at a dilution of 1:20000 for 45 min at room temperature (24 °C). The membranes were washed three times with PBST and developed using SuperSignal ultrachemiluminescent substrate (Pierce).

Statistical analysis

Results are presented as the means±S.E.M. for the number of myotubes indicated (values were normalized to the mean value of myotubes from corresponding controls). Significant differences were analysed using the paired Student's t test algorithm in the GraphPad InStat program. Differences were considered to be significant at P<0.05. Graphs were prepared using Origin.

RESULTS

S165F mutation of JP2 induces hypertrophy in skeletal myotubes

Ser165 of JP2 is located in the region between two groups of MORN motifs (Figure 1). We expressed a mutant form of JP2, JP2-S165F, which is know to be associated with HCM in humans, in mouse primary skeletal myotubes as described in the Experimental section. To compare myotube size, images of untransfected myotubes and myotubes expressing wild-type JP2 or JP2-S165F were obtained just prior to single myotube Ca2+ imaging experiments (this condition is more similar to the native conditions than following methanol fixation) (Figure 2). A significant increase in diameter was found in JP2-S165F-expressing myotubes compared with wild-type controls (increased by 1.53±0.14-fold); untransfected myotubes did not show a significant change in diameter compared with myotubes transfected with wild-type JP2 (Figure 2 and Table 1). These results indicate that a single mutation in JP2 at Ser165 induced a hypertrophic response in skeletal myotubes, which is similar to the response observed in the cardiac muscle of HCM patients.

Figure 1 Schematic diagram of the primary sequence of mouse JP2

JP2 has eight MORN motifs (shown in dark grey) and a C-terminal transmembrane domain (TM; shown in light grey). The location of Ser165, which is associated with human HCM, is indicated.

Figure 2 Increased diameter of JP2-S165F-expressing myotubes

The diameters of the thickest part of wild-type JP2-expressing (Wild type), untransfected and JP2-S165F-expressing (S165F) myotubes were measured. Representative diameters of myotubes are indicated by white arrows. Scale bar, 100 μm. The histogram shows the diameters normalized to the mean value of diameters from myotubes transfected with wild-type JP2. JP2-S165F-expressing myotubes showed increases in diameters. There was no significant change in the diameter of myotubes due to the expression of wild-type JP2. Results are means±S.E.M. of independent myotubes as presented in Table 1. *P<0.05 compared with wild-type.

View this table:
Table 1 Properties of J2-S165F-expressing myotubes

Results were normalized to the mean value of those from myotubes expressing wild-type JP2 (expect the resting cytosolic Ca2+ level) and are means±S.E.M. for the number of myotubes shown in parentheses. *P<0.05 compared with wild-type.

Hypertrophic JP2-S165F-expressing skeletal myotubes have altered Ca2+ signalling

We next examined Ca2+ signalling in the hypertrophic JP2-S165F-expressing myotubes. To measure cytosolic Ca2+ transients per unit area in JP2-S165F-expresing myotubes during EC coupling, 80 mM KCl was applied to the myotubes loaded with fluo-4 to induce membrane depolarization (Figure 3). Myotubes transfected with wild-type JP2 were used as controls. JP2-S165F-expressing myotubes showed a significant decrease in cytosolic Ca2+ transients compared with wild-type controls (0.67±0.11-fold; Figure 3 and Table 1). However, the expression level of the α1 subunit of DHPR, which is the Ca2+-conducting pore-forming subunit of DHPR [1,2,26], was not significantly changed (see Figure 6A), suggesting that the EC coupling gain is reduced by the JP2 mutation at Ser165. Interestingly, the activation slope of the transients in JP2-S165F-expressing myotubes was approximately half that of wild-type controls (0.51±0.07-fold; Figure 3 and Table 1).

Figure 3 Decreased EC coupling gain in hypertrophic JP2-S165F-expressing myotubes

KCl (80 mM) was applied to JP2-S165F-expressing myotubes loaded with fluo-4. JP2-S165F-expressing myotubes showed a significant decrease in cytosolic Ca2+ transients per unit area in response to KCl. Myotubes expressing wild-type JP2 were used as controls. Activation slopes, which were fitted using a linear equation, are represented by a dotted line (the vertical is represented by a solid line). The two histograms show the peak amplitude (left-hand histogram) or activation slope (right-hand histogram) normalized to the mean value of myotubes from wild-type controls. Results are means±S.E.M. for independent myotubes as presented in Table 1. *P<0.05 compared with wild-type.

There are two possible explanations for the reduced EC coupling gain in the hypertrophic JP2-S165F-expressing myotubes: the interaction between RyR1 and DHPR could be impaired; or RyR1 activity might be directly affected, which then results in reduced EC coupling gain. The first possibility has no physiological relevance to cardiac EC coupling because a paired arrangement of RyRs and DHPRs is not a necessary factor for cardiac EC coupling and the JP2-S165F mutation was first found in the cardiac tissue of HCM patients. To test the latter possibility, we measured the Ca2+ transients per unit area in response to 40 mM caffeine, a direct RyR1 agonist (Figure 4). JP2-S165F-expressing myotubes showed significant reductions in Ca2+ transients compared with wild-type controls (0.76±0.10-fold; Figure 4 and Table 1). As seen with KCl-mediated depolarization, JP2-S165F-expressing myotubes had significantly slower activation in response to caffeine (0.55±0.18-fold; Figure 4 and Table 1). These results suggest that reduced RyR1 activity partially and/or fully explains the reduced EC coupling gain in JP2-S165F-expressing myotubes.

Figure 4 Decreased RyR1 activity in hypertrophic JP2-S165F-expressing myotubes

Caffeine (40 mM), a direct agonist of RyR1, was applied to JP2-S165F-expressing myotubes loaded with fluo-4. Myotubes expressing wild-type JP2 were used as controls. JP2-S165F-expressing myotubes showed significantly reduced cytosolic Ca2+ transients per unit area in response to caffeine. Activation slopes were fitted using a linear equation and are represented by a dotted line (the vertical is represented by a solid line). Histograms are shown for peak amplitude (left-hand histogram) or activation slope (right-hand histogram) normalized to the mean value of those from wild-type controls. Results are means±S.E.M. for independent myotubes as presented in Table 1. *P<0.05 compared with wild-type.

As the Ca2+ transient would be expected to be larger if the diameter of myotubes increased, net decreases in Ca2+ transients in response to KCl or caffeine would possibly be larger than those shown in Figures 3 and 4. The increases in diameter would also be expected to slow the activation slope of Ca2+ transients in response to both KCl and caffeine because of diffusion effects, which is in accordance with our results (i.e. decreased activation slopes in Figures 3, 4 and Table 1).

As endogenous expression of JP2 is abundant in skeletal muscle, both the JP2-S165F-overexpression-mediated hypertrophy and the modification of Ca2+ signalling in the skeletal myotubes suggests that the JP2-S165F mutant protein has a dominant-negative effect over native JP2.

Hypertrophic JP2-S165F-expressing skeletal myotubes have increased resting cytoplasmic Ca2+ levels

To determine the mechanism underlying the reduced Ca2+ signalling, including the reduced RyR1 activity and the reduced EC coupling gain, we first tested the possibility that the reduced Ca2+ transient in response to caffeine was caused by reduced Ca2+ storage in the SR. The SR Ca2+ content in JP2-S165F-expressing myotubes was measured directly by treatment with 10 μM CPA, which is an inhibitor of SERCA1, the SR Ca2+-uptake pump, and thus allows for measurement of the SR Ca2+ content [2,27,28]. Myotubes transfected with wild-type JP2 were used as controls. There was no significant difference in the SR-stored Ca2+ level between wild-type and JP2-S165F-expressing myotubes loaded with fluo-4 (Figure 5A) or fura-2 (results not shown) in the presence of extracellular Ca2+ ions. In addition, the expression level of SERCA1 or calsequestrin (an SR Ca2+-buffering protein [1,2]) remained unchanged (Figure 5B). These results indirectly suggest that changes in SR Ca2+ content or refilling are not major contributors to the reduced Ca2+ signalling.

Figure 5 No change in SR Ca2+ content, but increased resting cytoplasmic Ca2+ levels in hypertrophic JP2-S165F-expressing myotubes

(A) SR Ca2+ in JP2-S165F-expressing myotubes loaded with fluo-4 was depleted by treatment with 10 μM CPA. Myotubes expressing wild-type JP2 were used as controls. Ca2+ transients per unit area induced by CPA are shown in the histogram (the peak area was normalized to the mean value of those from wild-type controls). Small peaks at the moment of CPA treatment were artifacts of manual CPA injection. There was no change in SR Ca2+. (B) Each solubilized myotube lysate (20 μg of total protein) was subjected to an immunoblot (IB) assay with anti-calsequestrin antibody (upper panel) or anti-SERCA1 antibody (lower panel). There was no significant difference in the expression level of calsequestrin or SERCA1 between the wild-type and JP2-S165F samples. (C) Resting cytoplasmic Ca2+ levels per unit area in JP2-S165F-expressing myotubes loaded with fura-2 normalized to the mean value of those from wild-type controls. The resting Ca2+ level of cytoplasm was increased in JP2-S165F-expressing myotubes. Results are means±S.E.M. for independent myotubes as presented in Table 1. *P<0.05 compared with wild-type. Ab, antibody.

Next, cytosolic Ca2+ levels in resting myotubes were examined. JP2-S165F-expressing myotubes maintained significantly higher (an approx. 50% increase) resting Ca2+ levels per unit area than wild-type controls (102.61±11.16 nM compared with 69.05±6.86 nM; Figure 5C and Table 1). Substantially, the absolute amount of total Ca2+ in the cytoplasm of a single myotube expressing JP2-S165F was much greater than that of a wild-type myotube. As JP2-S165F-expressing myotubes were hypertrophic, the absolute cytosolic volume of the hypertrophic myotubes (in terms of three-dimensional status) was much greater than that of wild-type controls. Therefore it is possible that elevated cytosolic Ca2+ might contribute to the reduced Ca2+ signalling in JP2-S165F-expressing myotubes.

Finally, to rule out the possibility that the level of RyR1 expression was reduced in the JP2-S165F-expressing myotubes, an immunoblot assay was performed on each solubilized myotube lysate using anti-RyR1 antibody (Figure 6A, left-hand panel, middle image). As for the α1 subunit of DHPR (Figure 6A, left-hand panel, bottom image), there were no significant difference in the amount of RyR1s between wild-type and JP2-S165F-expressing myotubes. Thus changes in the expression level of RyR1 or DHPR in the JP2-S165F-expressing myotubes are not contributors to the reduced Ca2+ signalling.

Figure 6 Defects in PKC-mediated phosphorylation of JP2-S165F at Ser165 and in the binding of JP2-S165F to TRPC3

(A) Each solubilized myotube lysate (30 μg of total protein) was subjected to an immunoblot (IB) assay with the anti-DHPR antibody (left-hand panel, bottom), anti-RyR1 antibody (left-hand panel, middle) or anti-JP2 antibody (left-hand panel, top). There was no significant difference in the expression level of DHPR or RyR1. Transfection of wild-type JP2 or JP2-S165F into myotubes increased the expression of JP2 protein (both endogenous wild-type JP2s and exogenously expressed wild-type or mutated JP2) approx. 2-fold compared with untransfected myotubes. Intensities of the JP bands were normalized to the mean value of those from untransfected myotubes and are shown as a histogram (right-hand panel). (B) Immunoprecipitation (IP) of each myotube lysate (800 μg of total protein) with the anti-phosphoserine antibody followed by immunoblotting of the immuoprecipitates with anti-JP2 antibody showed that phosphorylation of JP2 protein in the JP2-S165F sample was significantly less than that in the wild-type sample (left-hand panel). Unlike the JP2-S165F sample, the amount of phosphorylated JP2 protein in the JP2-Y141H sample was nearly identical with that in the wild-type sample. Pre-treatment of myotubes expressing wild-type JP2 with GF-109203X (Wild type+GF) reduced the phosphorylation level of JP2 proteins in the wild-type sample to that of the JP2-S165F sample. Band intensities were normalized to the mean value of those from the wild-type sample and are shown as histograms (right-hand panel). (C) Co-immunoprecipitation of each myotube lysate (800 μg of total protein) with the anti-TRPC3 antibody followed by immunoblotting of the immunoprecipitates with anti-JP2 antibody showed that only a trace amount of JP2 protein in the JP2-S165F sample was co-immunoprecipitated with TRPC3 (left-hand panel). Pre-treatment of myotubes expressing wild-type JP2 with GF-109203X (Wild type+GF) almost abolished the interaction between JP2 and TRPC3. Intensities of JP bands were normalized to the mean value of those from the wild-type sample and are shown as histograms (right-hand panel). (D) To examine the phosphorylation level of endogenous JP2s at serine residues by PKC, myotube lysate (800 μg of total protein) from untransfected myotubes (−) and untransfected myotubes pre-treated with GF-109203X (+) was subjected to immunoprecipitation with the anti-phosphoserine antibody, followed by immunoblotting of the immunoprecipitates with the anti-JP2 antibody (left-hand panel). Approx. 60% of endogenous JP2 was phosphorylated at serine residues by PKC. Band intensities were normalized to the mean value of those from the GF-109203X-untreated sample and are shown as a histogram (right-hand panel). *P<0.05 compared with untransfected myotubes, wild-type or the GF-109203X-untreated sample. Ab, antibody.

Ser165 of JP2 is a PKC phosphorylation site

As Ser165 is the only putative PKC phosphorylation site in JP2 [20], we determined whether the PKC-mediated phosphorylation contributes to altered Ca2+ signalling in the hypertrophic JP2-S165F-expressing skeletal myotubes. To confirm that the amount of JP2 protein was similar in wild-type JP2 (both endogenous wild-type JP2 and exogenously expressed JP2) and JP2-S165F samples, the solubilized myotube lysates were subjected to an immunoblot assay using the anti-JP2 antibody (Figure 6A, left-hand panel, top image). There was no significant difference in the total amount of JP2 protein between wild-type and JP2-S165F-transfected samples (both samples contained approximately twice as much JP2 protein as untransfected samples). Immunoprecipitation with the anti-phosphoserine antibody, followed by immunoblotting of the immunoprecipitates with the anti-JP2 antibody showed that the amount of phosphorylated JP2 protein in the JP2-S165F sample was much less than that in the wild-type sample (Figure 6B). On the other hand, transfection of myotubes with JP2-Y141H, another JP2 mutant associated with human HCM (and not expected to interfere with PKC-mediated phosphorylation), did not alter the amount of phosphorylated JP2 protein. Pre-treatment of myotubes expressing wild-type JP2 with 10 μM GF-109203X, which is a potent PKC inhibitor [29], reduced the amount of phosphorylated JP2 protein by approx. 40% compared with untreated wild-type controls (Figure 6B). Taken together, the JP2 mutation at Ser165 blocks the PKC-mediated phosphorylation of JP2 at Ser165. These results suggest that Ser165 is probably a site for JP2 phosphorylation by PKC. However, it is likely that other protein kinases might also participate in phosphorylation of JP2 because approx. 40% of the JP2 protein was still detected by the anti-phosphoserine antibody after treatment with GF-109203X (Figure 6B). A recent genomic study predicted that 15 additional serine residues are possible sites of phosphorylation by protein kinases other than PKC [20].

Lack of phosphorylation of JP2 at Ser165 disrupts JP2–TRPC3 binding

We have found previously that JP2 physically interacts with TRPC3 in skeletal muscle [5,13]. To determine whether the PKC-mediated phosphorylation of JP2 at Ser165, which is located in the region which is responsible for the interaction (from amino acids 143–234 of JP2), affects the interaction between JP2 and TRPC3, co-immunoprecipitation with the anti-TRPC3 antibody followed by immunoblotting of the immunoprecipitates with the anti-JP2 antibody was performed on each solubilized myotube lysate (Figure 6C). Only a trace amount of JP2 protein was co-immunoprecipitated with TRPC3 in the JP2-S165F sample compared with the wild-type sample; this might be attributable to endogenous JP2 given that we determined approx. 60% of endogenous JP2 was phosphorylated at serine residues by PKC (Figure 6D). In addition, pre-treatment of myotubes expressing wild-type JP2 with 10 μM GF-109203X almost abolished the interaction between JP2 and TRPC3 (Figure 6C). It is possible that the JP2-S165F mutation (with a lack of phosphorylation at Ser165) induces a significant conformational change in the protein, which could simply interrupt the physical interaction between JP2 and TRPC3. Indeed, secondary structure modelling of wild-type JP2 and the JP2-S165F mutant using the secondary structure prediction program Jpred3 [30] predicted different secondary structures features for wild-type JP2 and JP2-S165F [13]. Although it would be more relevant to compare predicted structures for wild-type JP2 when it is phosphorylated at Ser165 with unphosphorylated wild-type JP2 and/or the JP2-S165F mutant protein, tools or software to allow us to predict changes in secondary or tertiary structure of peptides or proteins by phosphorylation are not available.

In addition to the biochemical and in silico approaches described above, we examined the functional effects of GF-109203X on the Ca2+ transients of skeletal myotubes in response to KCl or caffeine and on resting cytosolic Ca2+ level. KCl or caffeine were applied to myotubes transfected with wild-type JP2 and pre-treated with 10 μM GF-109203X (Figure 7). Myotubes expressing wild-type JP2 and not treated with GF-109203X were used as controls. Ca2+ transients in response to both KCl and caffeine were significantly decreased by GF-109203X (by approx. 50%; Figure 7 and Table 2), as seen in the myotubes expressing JP2-S165F in Figures 3 and 4. The GF-109203X-treated myotubes also had significantly slower activation in response to both KCl and caffeine, and higher resting Ca2+ level than controls not treated with GF-109203X (Table 2). These results strongly support the hypothesis that a lack of PKC-mediated phosphorylation of JP2-S165F at Ser165, and the subsequent loss of binding between JP2-S165F and TRPC3, could be the mechanism for inducing the abnormal Ca2+ signalling and hypertrophy in JP2-S165F-expressing skeletal myotubes.

View this table:
Table 2 Effects of GF-109203X on myotubes expressing wild-type JP2

Results were normalized to the mean value of those from GF-109203X-untreated wild-type controls (expect the resting cytosolic Ca2+ level) as described in the legend of Figure 7 and are means±S.E.M. for the number of myotubes shown in parentheses. *P<0.05 compared with wild-type myotubes not treated with GF-109203X.

Figure 7 Decreased EC coupling gain and RyR1 activity in myotubes transfected with wild-type JP2 treated with GF-109203X

KCl or caffeine was applied to myotubes transfected with wild-type JP2 and pre-treated with GF-109203X (Wild type + GF) as described in Figures 3 and 4. Myotubes transfected with wild-type JP2 and untreated with GF-109203X were used as controls (Wild type). Pre-treatment with GF-109203X induced a significant decrease in cytosolic Ca2+ transients in response to (A) KCl or (B) caffeine. The activation slopes were fitted using a linear equation and is represented by a dotted line (the vertical is represented by a solid line). (C) Histogram showing the for peak amplitude normalized to the mean value of those from GF-109203X-untreated (−) wild-type controls. Results are means±S.E.M. for independent myotubes as presented in Table 2. Normalized activation slope is also presented in Table 2. *P<0.05 compared with wild-type myotubes not treated with GF-109203X.

In summary, hypertrophy and altered intracellular Ca2+ signalling (i.e. reduced RyR1 activity, reduced EC coupling gain and slower activation of RyR1 and EC coupling) were observed in JP2-S165F-expressing skeletal myotubes. Possible causes of the altered intracellular Ca2+ signalling in hypertrophic JP2-S165F-expressing skeletal myotubes are elevated cytosolic Ca2+ levels, defective PKC-mediated phosphorylation of JP2-S165F at Ser165 and disruption of the binding between JP2-S165F and TRPC3. Therefore the alterations in intracellular Ca2+ signalling might be linked to altered phosphorylation of JP2 and/or altered cross-talk among the Ca2+ channels (i.e. RyR1, DHPR and TRPC3) on t-tubule and SR membranes. This might ultimately result in skeletal myotube hypertrophy and the finding provides insight into the subcellular molecular mechanisms for the JP2-linked human HCM progression.

DISCUSSION

To examine the role of JP2 in skeletal muscle and to elucidate the subcellular molecular mechanisms for the effect of a JP2 mutation (the S165F mutation that is linked to human HCM) on skeletal muscle, we examined the dominant-negative role of JP2-S165F in mouse primary skeletal myotubes. We found increases in myotube diameter and abnormalities in Ca2+ signalling, including reduced RyR1 activity and reduced EC coupling gain. In addition, we observed an increase in resting cytosolic Ca2+ levels, defective PKC-mediated phosphorylation of JP2-S165F at Ser165 and disruption of the interaction of JP2-S165F with TRPC3. It is not presently clear what the in vivo effects of this mutation in skeletal muscle are or whether the in vitro mechanisms associated with the JP2-S165F mutation described in the present study are similar to those in vivo. However, the present study suggests that JP2 plays important roles in maintaining Ca2+ homoeostasis of skeletal muscle and provides insight into possible mechanisms by which this mutant causes cardiac hypertrophy in humans bearing this mutation.

Elevated cytosolic Ca2+ could result in the abnormal Ca2+ signalling observed in JP2-linked hypertrophic skeletal myotubes

It is known that various hypertrophic stimuli elevate intracellular Ca2+, which mediates the progression of cardiac hypertrophy via various routes, for example by activation of the Ca2+-dependent protein phosphatase, calcineurin [31]. The relationship between RyR1 activity and cytosolic Ca2+ concentration is bell-shaped (i.e. activation and then inactivation by increasing Ca2+ concentrations), and caffeine is known to increases the sensitivity of RyR1 to Ca2+ [26,32,33]. As hypertrophic JP2-S165F-expressing skeletal myotubes also showed elevated cytosolic Ca2+ levels in the resting state, the elevated cytosolic Ca2+ might shift the RyR1s/Ca2+ bell-shaped curve ‘to the left’ by partly occupying high-affinity activatory Ca2+-binding sites on RyR1s. In this scenario, elevated cytosolic Ca2+ ions could exert an inhibitory action on RyR1 by rapidly and effectively occupying the low-affinity inhibitory Ca2+-binding sites, as well as the remaining high-affinity Ca2+-binding sites, on RyR1s in hypertrophic JP2-S165F-expressing skeletal myotubes activated by caffeine. Therefore the number of inactivated RyR1s would rapidly increase, which would result in a reduced RyR1 response to caffeine in hypertrophic JP2-S165F-expressing skeletal myotubes. This mechanism might also explain the defects in Ca2+ homoeostasis observed previously in HL-1 cardiac myocytes expressing JP2-S165F [18]; single channel recordings and [3H]ryanodine-binding experiments have shown that RyR2 also exhibits a bi-phasic response to cytoplasmic Ca2+ level (maximal activation at 10−6–10−5 M and maximal inhibition at ~ 10−2 M Ca2+) [34].

Defects in phosphorylation(s) of JP2 could result in the abnormal Ca2+ signalling observed in JP2-linked hypertrophic skeletal myotubes

We identified that the lack of PKC-mediated phosphorylation of JP2-S165F at Ser165 is a potential mechanism which explains the relationship between Ser165 and reduced Ca2+ signalling. Ser165 of JP2 is located in the joining region between the two groups of MORN motifs, which interact with the t-tubule membranes to form junctional membrane complexes and allow DHPR, RyR1 and TRPC3 to be positioned close enough to concert efficient EC coupling [3,4,7,8]. Thus the mutation at Ser165, and subsequent lack of PKC-mediated phosphorylation at Ser165, might alter the structure of the junctional membrane complexes, resulting in reduced EC coupling gain in the hypertrophic JP2-S165F-expressing skeletal myotubes. Indeed, reduced EC coupling gain due to dispersion of the junctional membrane complexes in the absence of altered Ca2+ channel expression or SR Ca2+ content has been described in a rat model of heart failure, in hypertrophic rat myocytes and in mouse cardiac arrhythmias [3537]. Regarding the level of JP2 expression and hypertrophy, decreased JP2 expression was detected during the early stages of hypertrophy in the rat [17] and in murine hypertrophic and dilated cardiomyopathies [16].

TRPC3 could be related to the abnormal Ca2+ signalling observed in JP2-linked hypertrophic skeletal myotubes

Another potential mechanism for the relationship between Ser165 and the reduced Ca2+ signalling in hypertrophic JP2-S165F-expressing skeletal myotubes is via disruption of the normal JP2–TRPC3 interaction. We reported previously that the TRPC3 Ca2+ channel is required for full skeletal EC coupling gain (i.e. duration and maintenance) due to the functional interaction between TRPC3 and RyR1 [4]. In addition, we have shown that JP2 binds to TRPC3 in skeletal muscle [5,13]. In the present study, we found that PKC-mediated phosphorylation of JP2 at Ser165 is involved in the interaction between JP2 and TRPC3 in skeletal muscle. Taken together, it is probable that the lack of PKC-mediated phosphorylation in the JP2-S165F mutant interferes with the binding of JP2-S165F to TRPC3, which could potentially impair the functional interaction and subsequently reduce both RyR1 activity and the EC coupling gain, as seen in TRPC3-knockdown skeletal myotubes [4]. In this way, TRPC3 might be involved in the progression of hypertrophy in skeletal myotubes. This hypothesis is supported by cardiac studies: TRPC3-overexpressing transgenic mice showed increased calcineurin/NFAT (nuclear factor of activated T-cell) activation and cardiomyopathy [38]; TRPC3-mediated Ca2+ entry contributed to angiotensin II-induced rat cardiac hypertrophy [39]; and expression of hypertrophy-associated genes (A- and B-type natriuretic peptides) was decreased in TRPC3 knockdown neonatal rat ventricular myocytes [40]. Ectopic expression of TRPC3 in neonatal rat ventricular myocytes also stimulated cardiomyocyte hypertrophy [41]. Among the seven PKC isoforms, PKCα and PKCβ are involved in human HCM and both PKC isoforms are hyper-activated in human heart failure (but not in patients with end-stage heart failure) [42,43].

AUTHOR CONTRIBUTION

Jianjie Ma and Eun Hui Lee designed the research. Jin Seok Woo, Ji-Hye Hwang and Eun Hui Lee performed the research. Do Han Kim and Jianjie Ma contributed new reagents and analytical tools. Jin Seok Woo, Jae-Kyun Ko and Eun Hui Lee analysed the results. Jin Seok Woo and Eun Hui Lee wrote the paper. Noah Weisleder and Do Han Kim edited the paper.

FUNDING

This work was supported by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD) [grant number KRF-2008-331-E00011 (to E.H.L.)]; by the National Research Foundation of Korea (NSP) funded by the Ministry of Education, Science and Technology (MEST) [grant number 20090065569 (to D.H.K.)]; and by the National Institutes of Health, U.S.A. (to J.M.).

Abbreviations: bFGF, basic fibroblast growth factor; CPA, cyclopiazonic acid; DHPR, dihydropyridine receptor; EC, excitation–contraction; FBS, fetal bovine serum; fluo-4-AM, fluo-4-acetoxymethyl ester; fura-2-AM, fura-2-acetoxymethyl ester; HCM, hypertrophic cardiomyopathy; JP, junctophilin; MORN, membrane occupation and recognition nexus; PBST, PBS containing 0.1% Tween 20; PKC, protein kinase C; RyR, ryanodine receptor; SERCA1, sarcoplasmic/endoplasmic reticulum Ca2+-ATPase 1; SR, sarcoplasmic reticulum; TRPC3, transient receptor potential canonical-type cation channel 3; t-tubule, transverse tubule

References

View Abstract