Role of Junctin Protein Interactions in Cellular Dynamics of Calsequestrin Polymer upon Calcium Perturbation*

Background: In vitro studies have reported reversible calsequestrin polymerization and depolymerization. Results: Live cell imaging analysis revealed Ca2+-dependent decondensation of calsequestrin speckles, consistent with in vitro microscopic data. Conclusion: Calsequestrin depolymerization by calcium depletion requires coexistence of junctin. Significance: The role of calsequestrin in intracellular calcium homeostasis was explored. Calsequestrin (CSQ), the major intrasarcoplasmic reticulum calcium storage protein, undergoes dynamic polymerization and depolymerization in a Ca2+-dependent manner. However, no direct evidence of CSQ depolymerization in vivo with physiological relevance has been obtained. In the present study, live cell imaging analysis facilitated characterization of the in vivo dynamics of the macromolecular CSQ structure. CSQ2 appeared as speckles in the presence of normal sarcoplasmic reticulum (SR) Ca2+ that were decondensed upon Ca2+ depletion. Moreover, CSQ2 decondensation occurred only in the stoichiometric presence of junctin (JNT). When expressed alone, CSQ2 speckles remained unchanged, even after Ca2+ depletion. FRET analysis revealed constant interactions between CSQ2 and JNT, regardless of the SR Ca2+ concentration, implying that JNT is an essential component of the CSQ scaffold. In vitro solubility assay, electron microscopy, and atomic force microscopy studies using purified recombinant proteins confirmed Ca2+ and JNT-dependent disassembly of the CSQ2 polymer. Accordingly, we conclude that reversible polymerization and depolymerization of CSQ are critical in SR Ca2+ homeostasis.

acidic residues (60%) within the 63 carboxyl-terminal amino acids. Although human CSQ1 and CSQ2 are the products of different genes, the proteins display 65% sequence identity, are acidic, and share one glycosylation site. However, CSQ2 contains a 31-amino acid extension at its carboxyl terminus (residues 361-391) comprising 71% acidic residues, which significantly contributes to its ability to bind large amounts of Ca 2ϩ (13), as well as a second glycosylation site. CSQ exists either as a monomer or as a wide range of polymers, depending on the Ca 2ϩ concentration. The extended N terminus of CSQ establishes a front-to-front dimer interface through arm exchange. The C terminus of CSQ, the most negative region, is critical in forming tetrameric and higher-order linear polymers and capturing Ca 2ϩ in the back-to-back interface (14 -16). Wang et al. (17) demonstrated that knockdown of CSQ2, but not CSQ1, led to reduced Ca 2ϩ storage and release in C2C12 skeletal muscle myotubes. They also found a significant reduction in both the sarco/endoplasmic reticulum Ca 2ϩ ATPase (SERCA) and the RyR1 proteins in the CSQ2 knockdown myotubes. Thus, we chose to evaluate CSQ2 instead of CSQ1 for the experiments with C2C12 cells.
Although in vitro studies have reported reversible CSQ polymerization and depolymerization, no direct in vivo evidence has been obtained to date. Earlier electron microscopy analyses have revealed a fibrous array assumed to be Ca 2ϩbound CSQ in the SR at the junctional membrane (18). Crosslinking studies have also demonstrated that the majority of CSQ in the SR forms CSQ-CSQ complexes, supporting the physiological relevance of these polymers (14,19). Here, we transiently overexpressed a fluorescent protein-tagged CSQ2, along with JNT, and examined the in vivo dynamics of the macromolecular structure of the CSQ complex using live cell imaging. Our data provide in vivo evidence that the occurrence of CSQ2 depolymerization is synchronous with Ca 2ϩ depletion and that this conformational change requires coexpression of JNT.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-C2C12, a murine skeletal muscle cell line, was grown in Dulbecco's modified Eagle's medium (DMEM; Invitrogen). Media were supplemented with heat-inactivated 10% fetal bovine serum (Invitrogen), non-essential amino acids, 50 units/ml penicillin, and 50 units/ml streptomycin (Invitrogen). Cells were grown at 37°C and 5% CO 2 and transfected with Lipofectamine reagent (Invitrogen). The amounts of expression plasmids used are specified in the figure legends, and equivalent amounts of plasmid DNA were maintained with a relevant empty vector.
Plasmids and Lentiviral Infection-Human CSQ2 cDNA with a C-terminal yellow fluorescent protein (YFP) fusion was inserted into pLenti M1.4-MCMV, and that containing a C-terminal hemagglutinin (HA) tag was inserted into pcDNA3 vectors for the expression of CSQ2-YFP fusion and CSQ2-HA proteins in mammalian cells, respectively. Mouse JNT cDNA with a C-terminal cyan fluorescent protein (CFP) sequence was inserted into pLenti M1.4-MCMV for the expression of JNT-CFP fusion in mammalian cells. VSV-G-pseudotyped lentivirus was produced in HEK293T cells cotransfected with pLenti M1.4 construct, pLp1, pLp2, and pVSV-G. Culture medium was replaced with serum-free DMEM at 12 h after transfection. Supernatant fractions containing lentivirus particles were collected 48 h later, filtered through a 0.45-m filter unit, and used to infect C2C12 cells at 60% confluence in the presence of 4 g/ml of Polybrene or stored at Ϫ70°C.
Imaging-Imaging was carried out using an inverted confocal microscope (LSM 510 META and LIVE 5, Carl Zeiss) with a 40ϫ objective. To monitor cytosolic Ca 2ϩ levels, cells were loaded with 2 M fluo-4-acetoxymethylester (Fluo-4-AM) (Molecular Probes) in physiological salt solution (150 mM NaCl, 4 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 5 mM glucose, 5 mM HEPES) for 30 min at 37°C. Fluo-4 was excited with a 488-nm laser line and fluorescence was acquired at wavelengths of 505-530 nm. Cells were treated with thapsigargin (Calbiochem), 2,5-di-(tert-butyl)-1,4-benzohydroquinone (Sigma), cyclopiazonic acid (Sigma), or isoproterenol (Sigma) for intracellular Ca 2ϩ perturbation. Fluorescence resonance energy transfer (FRET) was measured with the acceptor photobleaching method using CFP and YFP fusion proteins as donor and acceptor, respectively. This method estimates the release of donor quenching after acceptor photobleaching, thus providing a measurement of close colocalization between the fusion proteins. CFP and YFP were excited with the 405-and 514-nm lines of an argon laser, respectively. Emitted fluorescence was collected at 420 -480 nm for CFP and 530 -600 nm for YFP. Repeated scans with unattenuated 514 nm illumination were used to photobleach YFP. FRET efficiency was estimated using the following equation: (F CFP Postbleach Ϫ F CFP Prebleach)/ F CFP Postbleach ϫ 100%.
Recombinant CSQ2 and JNT Protein Purification-Human cardiac CSQ2 (NP_001223) (amino acids 20 -399) and mouse JNT (AAG40812) without the transmembrane domain (amino acids 68 -207) were cloned into pET-28b vector (Novagen) with an N-terminal His 6 tag for affinity purification. For amplification of recombinant CSQ2 and JNT sequences, the following synthetic oligonucleotide primers were used: forward, 5Ј-GTA GAA CGC AGA ACT CTC ATA TGA AGA GAA CTC AC-3Ј; and reverse, 5Ј-CCC CCC CTC GAG GTC GAC CTC GAG CTA TTC ATC ATC ATC GTC ATC-3Ј for CSQ2; forward, 5Ј-GTG GCT GTC CAT ATG TTT GAC TTG GTC GAT TAT GAA G-3Ј, and reverse, 5Ј-GCC CTT GCT CAC CTC GAG TCA GTT CTT TCT CTT C-3Ј for JNT. The forward primer introduced a restriction enzyme site for NdeI (underlined), whereas the reverse primer introduced sites for XhoI (underlined) and the stop codon (bold). PCR was performed using Pfu DNA polymerase (SolGent) with the following cycling profile: initial denaturation at 95°C for 2 min, 30 cycles of denaturation at 95°C for 1 min, annealing at 50°C for 1 min, elongation at 72°C for 1 min, and final extension at 72°C for 10 min. PCR products digested with NdeI and XhoI were inserted into the vector using the corresponding restriction sites. The resulting constructs were verified by DNA sequencing. Prior to the cloning procedure, the first mutation was generated within CSQ2 cDNA to eliminate an undesirable NdeI site by converting CATATG to CTTATG, using the QuikChange II technique (Stratagene). Cells were grown at 37°C to an optical density (A 600 ) of ϳ0.6, and protein was induced with 0.2 mM isopropyl ␤-D-thiogalactopyranoside for 3 h. Next, cells were suspended in 20 mM Tris-Cl (pH 8.0), 500 mM NaCl, and 0.5% Triton X-100 and lysed with sonication. Proteins were subjected to affinity purification with Ni 2ϩ affinity chromatography, and the His 6 tag was removed by incubation with thrombin for 12 h at 4°C. Proteins were subsequently purified using a gel filtration column (HiLoad 16/60, Amersham Biosciences) in final buffer (20 mM HEPES, 150 mM NaCl, 2 mM DTT, pH 7.5) with the FPLC system (Amersham Biosciences).
Atomic Force Microscopy (AFM)-Purified CSQ2 alone and the CSQ2-JNT mixture (10 M each) were dialyzed against polymerization-inducing buffer (5 mM CaCl 2 , 150 mM NaCl, 2 mM DTT, 20 mM HEPES, pH 7.5) for 5 h at room temperature. An aliquot of each sample was deposited on film, dried under vacuum, and washed with distilled water. AFM measurements were performed with a Dimension 3100 scanning probe microscope (Digital Instruments Veeco) in tapping mode. Silicon nitride cantilevers were used as square pyramidal tips with a scan rate of 1 Hz for a 500 ϫ 500-nm 2 scale image. The resonance frequency was 320 kHz, and the force constant was 20 newtons/m. Raw AFM data were processed using a flattening/ low pass fitting process.
Electron Microscopy (EM)-Purified CSQ2 and CSQ2-JNT mixtures (10 M each) were incubated in polymerization-inducing buffer with or without 5 mM Ca 2ϩ , and the supernatant fractions were isolated by centrifugation. Aliquots (5 l) of the reactions were applied to glow-discharged carbon-coated grids and negatively stained with 2% (w/v) uranyl acetate solution. Micrographs were recorded with a CCD camera (2k, Gatan) using a Tecnai F20 field emission gun electron microscope operated at 200 kV (FEI).

SR Ca 2ϩ Depletion Induces Decondensation of CSQ2 Speckles-
To investigate the in vivo dynamics of CSQ2 oligomerization in response to perturbation of the Ca 2ϩ concentration, CSQ2 was ectopically expressed with a fluorescent protein tag in C2C12 myoblasts, and the real-time fluorescence image was examined with confocal microscopy. As CSQ interacts with JNT or TRN (4,5) and undergoes assembly into the RyR complex, JNT was cotransfected with CSQ. CSQ appeared as speckles in myoblasts, consistent with a previous finding (20). Ca 2ϩ perturbation was performed by treating cells with thapsigargin, which inhibits SERCA, resulting in SR Ca 2ϩ depletion within a few minutes and a corresponding transient Ca 2ϩ increase in the cytoplasm (21). Upon coexpression with JNT, CSQ speckles became diffuse in live cells after thapsigargin treatment (Fig. 1A, lower panel; supplemental Movies S1 and S2). Sequential decondensation progression was additionally evident in the high-resolution images of cells fixed at intervals after thapsigargin treatment (Fig. 1B). Speckle shapes began to change at 5 min after thapsigargin treatment and became fully diffuse in 30 min (Fig. 1B). However, upon expression of CSQ only, speckles remained unchanged following thapsigargin treatment (Fig.  1A, upper panel). This result suggests that JNT assists in the decondensation of CSQ speckles that occurs upon Ca 2ϩ depletion. We examined SR Ca 2ϩ release by monitoring fluorescence of Fluo-4 placed in the cytoplasm after thapsigargin treatment. Cells coexpressing CSQ and JNT showed similar patterns of Ca 2ϩ transients upon thapsigargin treatment as untransfected control cells (supplemental Fig. S1), implying Ca 2ϩ release from SR to the cytoplasm. We speculated that Ca 2ϩ release from CSQ induced by thapsigargin promotes changes in the speckle shape. To evaluate the state of CSQ2, we performed native gel electrophoresis of cell lysates expressing CSQ2 (supplemental Fig. S2). As expected, CSQ2 was present in the form of a variety of polymers in a normal condition and was also found in the depolymerized form, which revealed a high content of monomeric CSQ2 following thapsigargin treatment. This finding is consistent with previous in vitro experiments showing that lower environmental Ca 2ϩ concentrations trigger release of Ca 2ϩ from the CSQ polymer, with concomitant disruption of its polymeric structure (3,14). Other SERCA inhibitors, such as cyclopiazonic acid (22) and 2,5-t-butylhydroquinone (23), also affect decondensation of CSQ speckles (Fig. 2, B and C). However, caffeine, an RyR channel opener, did not alter the CSQ speckle shape. We suggest that the transient Ca 2ϩ decrease caused by caffeine treatment is not sufficient (in regard to concentration or duration) to induce major conformational changes in the CSQ polymer but that the irreversible SR Ca 2ϩ depletion as a result of treatment with SERCA inhibitors is enough to do so. As JNT is necessary for this process, we examined the stoichiometry of molar interactions for CSQ speckle decondensation. The molar ratio of JNT to CSQ was important in the thapsigargin response. Specifically, when the JNT:CSQ molar ratio approached 1 (Fig. 3), speckle decondensation progressed clearly. However, at significantly lower concentrations of JNT than CSQ (for example, 1:3), speckle decondensation was marginal (Fig. 3). This result is in keeping with previous in vitro experiments showing that loss of interactions with JNT (or TRN) induces compact folding of CSQ at high Ca 2ϩ concentrations (5). At higher JNT levels than CSQ (for example, 3:1), inefficient speckle decondensation was observed (Fig. 3). Accordingly, we sug-gest that stoichiometric interactions of CSQ and JNT are necessary for Ca 2ϩ -dependent CSQ depolymerization.
CSQ2 Interacts with JNT-To examine the interactions between exogenously expressed CSQ and JNT, C2C12 cells were cotransfected with HA-tagged CSQ2 and CFP-tagged JNT expression plasmids, and coimmunoprecipitation assays were performed. CSQ2 was detected in immunoblots of JNT-CFP immunoprecipitates with the anti-GFP antibody (Fig. 4A). In reverse order, JNT was probed in CSQ2-HA immunoprecipitates with the anti-HA antibody (Fig. 4B). Thapsigargin treatment did not influence this interaction significantly (Fig. 4B,  lane 3 versus lane 4), suggesting that the levels of CSQ-JNT interactions are maintained, even after disruption of CSQ2 speckles at lower Ca 2ϩ concentrations. To determine whether CSQ and JNT proteins colocalize in intact cells, we utilized FRET technology. CSQ2 and JNT were expressed as N-terminal fusions with YFP and CFP, respectively. In an acceptor photobleaching experiment, YFP and CFP fluorescence was monitored before and after YFP was bleached in a region of interest. If CFP and YFP are in close proximity, donor (CFP) fluorescence should increase in the region where the acceptor (YFP) is   bleached. JNT-CFP was distributed throughout the cytoplasm, whereas CSQ2-YFP was localized in part of the cytoplasm as speckles (Fig. 4C, upper panels). After bleaching of the acceptor fluorophore (CSQ2-YFP), which led to almost complete removal of fluorescence intensity (Fig. 4C, lower right panel), increased fluorescence of donor molecules (JNT-CFP) was detected in the region of acceptor bleaching (Fig. 4C, lower left panel, yellow to red pseudocolor). Quantitative analysis revealed that CSQ2-YFP lost its yellow fluorescence by ϳ80% after bleaching, whereas JNT-CFP gained cyan fluorescence by ϳ40% (Fig. 4C). This result indicates that the two proteins are in close proximity to each other near the speckle positions. To determine whether the interactions between CSQ and JNT proteins remain after decondensation of CSQ speckles, FRET was performed after thapsigargin treatment in cells coexpressing CSQ2-YFP and JNT-CFP. Consistent with immunoprecipitation data, thapsigargin hardly influenced FRET outcome (Fig.  4D), suggestive of constant interactions between CSQ and JNT, irrespective of the Ca 2ϩ and polymerization states.
Macromolecular Structure of Polymerized CSQ2 in Vitro-Data from the present study showed that CSQ2 speckle decondensation occurs in a Ca 2ϩ -and JNT-dependent manner in skeletal muscle cells. We examined this issue in vitro using purified proteins. First, we tested the solubility of polymers generated with different concentrations of CSQ2 and JNT. Purified CSQ2 was combined with JNT at different molar ratios or left unmixed. Polymerization proceeded in buffer containing 5 mM Ca 2ϩ , as described in the previous study (15). After centrifugation, precipitates and supernatants were separated. Supernatants (Fig. 5A, Soluble I) were further prepared for negative staining electron microscopic analysis (see below). To establish whether the precipitation (polymerization) reaction is reversible, precipitates (Fig. 5A, Pellet I) were incubated in buffer without Ca 2ϩ . Resolubilized fractions (Soluble II) were separated from insoluble precipitates (Pellet II) by centrifugation, and the protein concentrations were measured (Fig. 5B). In vitro Ca 2ϩfree conditions may mimic the in vivo conditions of SR Ca 2ϩ depletion ( Figs. 1 and 2) that induce solubilization (decondensation) of CSQ speckles. Polymers obtained from equimolar mixtures of CSQ and JNT exhibited highest solubility with the lowest residual precipitates after Ca 2ϩ removal (Fig. 5B, fourth bar, compare gray with dark parts) among those obtained from mixtures of various molar ratios of CSQ and JNT. The solubilization efficiency gradually decreased as the ratio diverged from equimolar concentrations. A low proportion of polymers generated from CSQ only was converted to the soluble form (Ͻ30%), whereas those from the 1:1 CSQ-JNT mixture were susceptible to solubilization (up to 90%). This result is in agreement with our in vivo data showing that CSQ-JNT stoichiometry is critical for speckle decondensation. Second, residual precipitates (Pellet II) of the CSQ polymers in the presence or absence of JNT were observed under AFM performed to analyze macromolecular structures. In the absence of JNT, significant amounts of large CSQ aggregates (heights of up to 150 nm) were observed in the AFM image (Fig. 6A). In contrast, in the presence of JNT, smaller, evenly sized aggregates appeared (Fig. 6B). This finding is compatible with data from the solubility test (Fig. 5B), indicating that JNT mediates solubilization (destabilization) of the CSQ polymer, both in vitro and in vivo.
Third, we examined the oligomeric states of the CSQ molecule in the absence or presence of JNT using EM. Protein solutions were incubated with or without 5 mM Ca 2ϩ , and their supernatant fractions (Soluble I) were isolated by centrifugation to remove insoluble aggregates, mounted, and negatively stained for EM. Previous studies showed that CSQ exists mainly in dimeric, tetrameric, and polymeric states in the presence of 5 mM Ca 2ϩ (16). Consistent with earlier findings, we observed dimer-, tetramer-, and oligomer-like structures in EM images (Fig. 7, B and C). After magnification (Fig. 7, insets), CSQ appeared as a propeller-shaped particle (diameter, 60 -80 Å) containing three to four globular domains in the presence of 5 mM Ca 2ϩ . We assume that the particle corresponds to a frontto-front dimer because dimers having three or four globular domains can be imaged, depending on the direction (angle) of attachment to the grid surface (Fig. 7B). The repetitive domain appears to be a thioredoxin-like domain (diameter ϳ30 Å), previously identified in x-ray crystallography analyses (14). However, no evident structure was observed in the absence of Ca 2ϩ (Fig. 7A). Previous intrinsic fluorescence, circular dichroism, Raman spectroscopy, NMR, and proteolytic digestion analyses performed to determine the physicochemical properties of CSQ revealed that CSQ is a mostly unfolded random coil at low ionic strength but that it folds into a compact structure with increasing ion concentrations (24 -26). This result has been replicated in our EM experiments showing disruption of the thioredoxin-like domain in the absence of Ca 2ϩ (Fig. 7A). Upon incubation of equal amounts of CSQ and JNT (Fig. 7C), the total number of propeller-shaped particles decreased, and the appearance of the high-order structure (chain-like) that prevailed in the CSQ-only sample also decreased (Fig. 7, B versus C). Our results consistently suggest that interactions between JNT and CSQ compete with the intermolecular interactions among CSQ, thereby facilitating disassembly or preventing the formation of CSQ polymers.

DISCUSSION
Reversibility is a typical mechanism to maintain homeostasis in biological systems. The bidirectional reactions of CSQ polymerization and depolymerization may thus be critical in Ca 2ϩ homeostasis. Previous studies have proposed that positive cooperative binding of Ca 2ϩ by CSQ is concomitant with a shift in equilibrium from the soluble to the crystalline forms of CSQ (27), which folds into a more compact structure (28 -30). Thus, CSQ polymerization appears beneficial for Ca 2ϩ deposit. Based on the crystal structures, various lengths of CSQ polymer are expected to form with increasing Ca 2ϩ concentrations (14 -16) (PDB 1SJI). These CSQ polymers tend to depolymerize upon lowering of the free Ca 2ϩ concentration (15,(31)(32)(33). In addition, during FRET analysis of CSQ2 polymerization in cells cotransfected with CFP-and YFP-tagged CSQs, depletion of the endoplasmic reticulum Ca 2ϩ stores with ionomycin  resulted in a marked decrease in FRET efficiency, suggestive of CSQ depolymerization (31). Stopped-flow analysis showed that intrinsic CSQ fluorescence decreases with decreasing concentrations of Ca 2ϩ from 1 mM to 0.1-100 M, indicative of depolymerization (32). However, there is no direct evidence of Ca 2ϩ -dependent CSQ depolymerization in vivo with physiological relevance. Live cell imaging analysis (supplemental Movies S1 and S2 and Fig. 1) suggested JNT-dependent depolymerization of the CSQ polymer in vivo after Ca 2ϩ depletion. CSQ polymer underwent slow conformational changes upon SR Ca 2ϩ depletion using a SERCA inhibitor, and this was not synchronized with the twitches of normal, millisecond-long excitation-contraction coupling. We speculate that complete depolymerization does not occur during excitation-contraction coupling but that partial Ca 2ϩ release from CSQ and a corresponding partial depolymerization do occur, which is followed by a rebound in Ca 2ϩ and the repolymerization of CSQ. In contrast, under certain pathological conditions, virtually all of the Ca 2ϩ becomes lost from CSQ, and this may lead to the full depolymerization of the CSQ polymer without the possibility of reconstruction. In cardiac muscle, it has been shown that the lower intraluminal SR Ca 2ϩ concentration characteristic of heart failure results in abnormal CSQ trafficking (34). Irrevers-ible SR Ca 2ϩ depletion also occurs in malignant hyperthermia (35), a syndrome evident in mice lacking CSQ (36).
Our data strongly suggest that direct molecular interactions between CSQ and JNT are necessary for CSQ-mediated Ca 2ϩ control in SR. JNT-null cardiomyocytes display increased frequency of Ca 2ϩ sparks, indicative of increased SR leak, the phenotype of which is associated with arrhythmia (37). JNT-overexpressing cardiomyocytes display depressed contractility and reduced Ca 2ϩ in both transient and SR load (38,39). Consistently, cardiac overexpression of canine JNT in mice results in decreased contractility, impaired relaxation, and hypertrophy (40,41). Experiments from the current study disclosed that upon overexpression of JNT, interactions between JNT and CSQ increase, inhibiting polymerization of CSQ and leading to reduced SR Ca 2ϩ load and reduced spontaneous SR Ca 2ϩ release. On the other hand, upon JNT ablation, the bridge between CSQ and RyR may be abolished, resulting in desynchronized SR Ca 2ϩ release (42). We showed that the molar ratio of JNT-CSQ is critically important in Ca 2ϩ -CSQ dynamics by ectopic coexpression in C2C12 cells (Fig. 3). When the JNT level was significantly higher or lower than the CSQ level, depolymerization barely occurred (Fig. 3), possibly due to abnormal SR Ca 2ϩ load and CSQ mislocalization, respectively. Pre-  viously, CSQ condensation was observed in L6 myoblast and HeLa cells transfected with CSQ, which remained unchanged after depletion of endoplasmic reticulum Ca 2ϩ stores (20), possibly due to imbalance in the levels of CSQ relative to JNT. We replicated the necessity for JNT in CSQ depolymerization reactions in vitro in experiments using the purified proteins (Figs. 5-7). Reversible CSQ polymerization and depolymerization were dependent on the Ca 2ϩ concentration and JNT-CSQ ratios, consistent with in vivo data (Figs. 1-3).
The stoichiometry of the RyR complex has been studied in detail. FKBP12 (or FKBP12.6), calmodulin, TRN, and JNT are each known to interact with the RyR subunit in a 1:1 manner (5,(43)(44)(45). Both transgenic and knock-out mouse models of CSQ2 have demonstrated significantly decreased TRN and JNT protein levels (46,47). TRN knock-out mice also have significantly decreased CSQ2 and JNT protein levels (48). These results suggest a potential interdependence between these three proteins. Furthermore, CSQ function is known to depend on the stoichiometry of the association with the RyR-TRN-JNT-CSQ macromolecular complex (49). Here, we found that the stoichiometric interaction of CSQ2 with JNT was critical for the CSQ2 depolymerization reaction in vivo and have suggested that the optimal ratio of JNT-CSQ expression for this reaction is 1:1 (protein levels). However, endogenous JNT is present at lower levels than endogenous CSQ under physiological conditions (49 -51). A possible explanation, which may reconcile the difference between physiological conditions and our experimental conditions, is seen in Fig. 8. It is likely that some JNT molecules remain free, whereas others interact with polymerized CSQ. Thus, the final interaction of JNT and CSQ may be at a ratio less than 1:1. This suggestion is in line with our in vivo imaging data showing that ectopically expressed JNT is diffusely expressed, whereas that of CSQ is speckled (Fig. 4, C and D). Thus, the limited concentrations of endogenous JNT in vivo might be necessary and sufficient to play key roles in the reversible dynamics of Ca 2ϩ -CSQ polymerization and depolymerization, which are located adjacent to RyR.