TRIC-A Channel Maintains Store Calcium Handling by Interacting with Type 2 Ryanodine Receptor in Cardiac Muscle

Abstract

Rationale:

TRIC-A and TRIC-B are distributed to endoplasmic reticulum (ER)/sarcoplasmic reticulum (SR) intracellular Ca²⁺ stores. The crystal structure of TRIC has been determined, confirming the homo-trimeric structure of a potassium channel. While the pore architectures of TRIC-A and TRIC-B are conserved, the carboxyl-terminal tail domains (CTT) of TRIC-A and TRIC-B are different from each other. Aside from its recognized role as a counter-ion channel that participates in excitation-contraction coupling of striated muscles, the physiological function of TRIC-A in heart physiology and disease has remained largely unexplored.

Objective:

In cardiomyocytes, spontaneous Ca²⁺ waves, triggered by store overload-induced Ca²⁺ release (SOICR) mediated by the type 2 ryanodine receptor (RyR₂), develop extra-systolic contractions often associated with arrhythmic events. Here we test the hypothesis that TRIC-A is a physiologic component of RyR₂-mediated Ca²⁺ release machinery that directly modulates SOICR activity via CTT.

Methods and Results:

We show that cardiomyocytes derived from the TRIC-A^−/− mice display dysregulated Ca²⁺ movement across SR. Biochemical studies demonstrate a direct interaction between CTT-A and RyR₂. Modeling and docking studies reveal potential sites on RyR₂ that show differential interactions with CTT-A and CTT-B. In HEK293 cells with stable expression of RyR₂, transient expression of TRIC-A, but not TRIC-B, leads to apparent suppression of spontaneous Ca²⁺ oscillations. Ca²⁺ measurements using the cytosolic indicator Fura-2 and the ER luminal store indicator D1ER suggest that TRIC-A enhances Ca²⁺ leak across the ER by directly targeting RyR₂ to modulate SOICR. Moreover, synthetic CTT-A peptide facilitates RyR₂ activity in lipid bilayer reconstitution system, enhances Ca²⁺ sparks in permeabilized TRIC-A^−/− cardiomyocytes, and induces intracellular Ca²⁺ release after microinjection into isolated cardiomyocytes, whereas such effects were not observed with the CTT-B peptide. In response to isoproterenol stimulation, the TRIC-A^−/− mice display irregular electrocardiogram and develop more fibrosis than the WT littermates.

Conclusions:

In addition to the ion-conducting function, TRIC-A functions as an accessory protein of RyR₂ to modulate SR Ca²⁺ handling in cardiac muscle.

Graphical Abstract

TRIC channels were discovered in 2007. Aside from its recognized role as a counter-ion channel, the physiological function of TRIC-A in the heart has remained largely unexplored. Even with the recent resolution of the crystal structure of the TRIC channels, the interacting partners for TRIC have yet to be defined. This study uncovered a novel function for TRIC-A as an accessory protein that interacts with the RyR to modulate intracellular Ca²⁺ release in cardiomyocytes. The CTT domain of TRIC-A, an important portion of the channel that is left out of the crystal structure determination, constitutes an active motif that interacts with RyR. The vulnerability of the TRIC-A knockout mouse heart to handling of isoproterenol stimulation further support the important physiological function of TRIC-A in heart physiology and disease.

INTRODUCTION

In cardiac muscle, sarcolemma depolarization triggers the release of Ca²⁺ from the sarcoplasmic reticulum (SR) via Ca²⁺-induced Ca²⁺ release (CICR)^1–3. During this process, Ca²⁺ influx through the voltage-dependent Ca²⁺ channels on the sarcolemma activates the type 2 ryanodine receptor (RyR₂) channel located on the SR. Under certain pathological conditions, spontaneous Ca²⁺ release from the SR can take place in the absence of membrane excitation due to store overload induced Ca²⁺ release (SOICR)^4–10. SOICR may evoke propagating Ca²⁺ waves that can further result in delayed after-depolarizations, which may cause arrhythmias in heart failure.

The SR Ca²⁺ store of cardiomyocytes is maintained by uptake and release processes, both of which are electrogenic events. The release of Ca²⁺ through the RyR₂ channels will lead to the development of a negative potential inside the SR lumen, and this would further limit Ca²⁺ release from the SR if uncompensated. Likewise, SERCA-mediated uptake of Ca²⁺ into the SR would lead to the accumulation of a positive potential within the SR lumen, and that would tend to inhibit Ca²⁺-pumping function. Thus coordinated counter-ion movements are required to balance the SR membrane potential in order to maintain efficient Ca²⁺ release and uptake in cardiomyocytes.

In 2007, Takeshima and colleagues identified trimeric intracellular cation (TRIC) channels located on the SR and endoplasmic reticulum (ER) of multiple cell types¹¹. In human and mouse genomes, two isoforms of TRIC genes were identified: TRIC-A, a subtype predominantly expressed in SR of excitable cells, and TRIC-B, a ubiquitous subtype expressed in ER of all tissues^11–21. TRIC-A and TRIC-B appear to have different functions in Ca²⁺ signaling in excitable and non-excitable cells, since knocking out TRIC-B affects IP₃ receptor mediated Ca²⁺ release in airway epithelial cells, which results in respiratory defects and neonatal death with TRIC-B^−/− pups²¹. The TRIC-A^−/− mice, in addition to dysfunction of skeletal muscle²⁰, develop hypertension that is linked to defective Ca²⁺ sparks and spontaneous transient outward currents in arterial smooth muscle¹⁹. A common feature with the TRIC-A^−/− and TRIC-B^−/− mice is the development of Ca²⁺ overload inside the SR/ER of multiple tissues^{11, 19–21}. This Ca²⁺ overload can impact the function of SOICR, causing instability of Ca²⁺ movement across the SR membrane in muscle cells, which could further contribute to tissue dysfunctions associated with ablation of the TRIC genes.

While researchers over the past ten years have established that genetic ablations or mutations of TRIC channels are associated with hypertension, muscle dysfunction, respiratory defects, and brittle bone disease^11–21, the function of TRIC channels in heart physiology and disease has yet to be established. El-Ajouz et al²² reported dampened activity of RyR channels in skeletal muscle lacking TRIC-A, suggesting TRIC-A could play a modulatory role in control of RyR-mediated intracellular Ca²⁺ release. Recently, the crystal structure of TRIC has been determined, confirming the homo-trimeric structure of a potassium channel^23–26. While the pore architectures of TRIC-A and TRIC-B appear to be conserved, the carboxyl-terminal tail (CTT) domains of TRIC-A and TRIC-B are different from each other. These CTT domains show the flexible structure and potentially interfere with crystal formation, thus all available structural determinations of the TRIC channels were obtained without the CTT domain.

Here we report that cardiomyocytes derived from the TRIC-A^−/− mice show dysregulated Ca²⁺ movement across the SR membrane. Our biochemical and immunohistochemical studies show that TRIC-A can physically associate with RyR₂ via the CTT domain. Reconstitution studies in the lipid bilayer membrane, Ca²⁺ measurements in adult cardiomyocytes and heterologous cells demonstrate that TRIC-A, in addition to serving as a counter-ion channel, can interact with RyR₂ to modulate the intracellular Ca²⁺ homeostasis and consequently the operation of SOICR.

METHODS

The data, analytical methods, and other study materials will be provided upon reasonable request to the corresponding author.

Cardiomyocyte isolation from adult mice.

TRIC-A knock out (TRIC-A^−/−) mice used in this study were reported in our previous publication¹¹. Handling of mice followed IACUC guidelines from The Ohio State University. Isolated hearts from adult TRIC-A^−/− and wild type (WT) littermate mice (10–12 weeks of both genders) were perfused with a Langendorff apparatus at 37°C. The enzyme digestion step consisted of perfusing Tyrode’s solution containing 1 mg/ml collagenase (Type II, 300 U/mg; Worthington) and 0.1 mg/ml protease (Type XIV) for 6 min. Cardiomyocytes were dissociated from digested ventricles by gentle mechanical dissociation and used within 3 hours. The Tyrode’s solution contained (in mM) 136 NaCl, 5.4 KCl, 0.33 NaH₂PO₄, 1.0 MgCl₂, 10 glucose and 10 HEPES (pH 7.4).

Mice and isoproterenol-treatment protocol.

Alzet mini-osmotic pump (Alza Corp.) containing either isoproterenol (ISO), phosphate-buffered saline (PBS) as a control were surgically inserted subcutaneously in mice under isoflurane anesthesia. The mini-osmotic pump provides controlled delivery of ISO (60 mg/kg/day). In separate study, 80 mg/kg ISO was injected once intraperitoneally in mice under isoflurane anesthesia. 1 hour later, electrocardiogram (ECG) was measured using Animal Bio Amp and PowerLab 8/30, and analyzed by LabChart 7.0 (ADInstruments).

Hearts from the TRIC-A^−/− and WT mice were collected at the indicated times after ISO treatment, fixed in 10% formalin-containing PBS, and embedded in paraffin. Serial 4-μm heart sections were cut and stained with H & E and Masson’s trichrome. Blinding was not performed because the genotype of mice was already known before the experiments were conducted.

Ca²⁺ sparks and waves measurement in cardiomyocytes.

Intracellular Ca²⁺ sparks of intact cardiomyocytes isolated from adult mouse (4–8 months age) were recorded using a Zeiss780 confocal microscope with a 40× 1.42 NA oil immersion objective. Cardiomyocytes were loaded with Fluo-4-AM (2 μM) and then stimulated with a field stimulation of 0.5 Hz for 20 seconds in a normal Tyrode’s solution containing 1.8 mM Ca²⁺. Spontaneous Ca²⁺ spark activities were measured afterward. Experiments were conducted at room temperature (24–26 °C).

Isolated cardiomyocytes were loaded with 5 μM Fluo-4 AM (Thermo Fisher Scientific) in 1.8 mM Ca²⁺ physiological salt solution (PSS) (in mM): 130 NaCl, 5.6 KCl, 1 MgCl₂, 11 glucose, and 10 Hepes, pH 7.4 for 30 min at room temperature. Line scan images of Fluo-4 fluorescence intensity were acquired at a sampling rate of 2 ms per line using the Galvano scan mode of a Nikon A1R confocal microscope. During the 1 min line scan imaging, cardiomyocytes were subjected to an electric field stimulation of 0.2 Hz, 10 ms duration pulses of 6 V.

To examine the impact of store-overload induced intracellular Ca²⁺ release in the isolated cardiomyocytes, extracellular Ca²⁺ was increased from 1.8 mM to 10 mM in PSS containing 50 μM BTS to mitigate myocyte contraction. Serial x-y images of cardiomyocytes in 10 mM Ca²⁺ PSS were acquired at 130 ms per frame using the resonant scan mode of Nikon A1R confocal microscope. All image data were analyzed by Fiji-ImageJ.

Separate studies were performed with Ca²⁺ spark measurement in saponin-permeabilized cardiomyocytes, following the protocol of Lukyanenko and Gyorke²⁷ and Guo et al²⁸. Specifically, cardiomyocytes were permeabilized with saponin (0.01% for ~45 s). The MaxChelator program (URL: maxchelator.stanford.edu) developed by Bers et al²⁹ was used to prepare the solution for precise control of free cytosolic Ca²⁺. The cytosolic solution contained (mM): 120 potassium aspartate, 5 MgATP, 10 phosphocreatine, 0.03 Fluo-4 pentapotassium salt, 10 Hepes (pH 7.25) and 5 U/ml creatine phosphokinase, with 0.4 EGTA and 0.087 CaCl₂ which buffers free Ca²⁺ at 100 nM. The Fluo-4 dye was excited with the 488 nm line of a LED laser (0.30% of maximum power, Nikon-A1R), pinhole 2.0 (equivalent to 39.6 μm). The use of low laser-power avoids laser-induced damage to cardiomyocytes during line-scan imaging of Ca²⁺ sparks using the Nikon-A1R imaging system. During Ca²⁺ spark measurement, 0.001% saponin was continuously present in the cytosolic solution to avoid resealing of the sarcolemmal membrane. Quantification and characterization of Ca²⁺ sparks were performed using SparkMaster ImageJ program³⁰.

Plasmid construction.

The mouse TRIC-A or TRIC-B cDNA was cloned into pCMS-mRFP vector (Clontech). The pCMS-mRFP plasmid contained a red fluorescent protein (RFP) reporter gene driven under a separate SV40 promoter, allowing for selection of cells transfected with TRIC-A or TRIC-B using red fluorescence. For the investigation of the carboxyl-terminal tail (CTT) domain of TRIC on Ca²⁺ signaling, two chimeras were constructed, TRIC-AB and TRIC-BA. TRIC-AB contains a.a. 1–230 from TRIC-A plus a.a. 229–291 from TRIC-B. TRIC-BA contains a.a. 1–228 from TRIC-B plus a.a. 231–299 from TRIC-A. The cDNA coding sequences were chemically synthesized (Quintara Biosciences, CA) and cloned into the pCMS-mRFP vector to express the chimeric TRIC proteins in mammalian cells. All plasmid identities were confirmed by gene sequencing.

Baculovirus production and its infection to HEK293 cells.

Baculovirus carrying vesicular stomatitis virus G-protein (VSVG) on the virus envelope that effectively infects mammalian cells was produced as described previously³¹. cDNAs for TRIC-A and TRIC-B were cloned into the modified pFastBac1 vector (Invitrogen). Baculovirus was produced in Sf9 cells according to the manufacturer’s instructions (Invitrogen). P2 virus was used for the experiments. HEK293 cells inducibly expressing RyR₂ which were kindly provided by Dr. Nagomi Kurebayashi or control HEK293 cells were cultured in DMEM, 5% FBS, 1% penicillin, and streptomycin at 37°C, 5% CO₂. P2 virus solution was supplemented to the culture media to infect the virus and cells were observed at 36 hours post infection.

Ca^{2 +} imaging in HEK293 cells.

HEK293 cells with inducible expression of RyR₂ were provided by Dr. Wayne Chen and cultured in DMEM, 5% FBS, 1% penicillin, and streptomycin at 37°C, 5% CO₂⁵. pCMS-mRFP plasmids containing the various TRIC-A and TRIC-B cDNAs were transfected into HEK293 cells using lipofectamine reagent following manufacture’s instruction. 24 hours later, tetracycline (0.1 μg/ml) was added to the culture medium to induce RyR₂ expression. Measurements of cytosolic Ca^{2 +} were conducted at 16–18 hours post tetracycline induction. Cells were loaded with Fura-2-AM (5 μM) or Fluo-4 AM (2 μM) in 0 Ca^{2 +} Krebs–Ringer Hepes (KRH) buffer containing (mM): 125 NaCl, 5 KCl, 25 HEPES, 6 glucose, and 1.2 MgCl_2, pH 7.4 for 40 min at room temperature. All solutions were prepared with double-distilled H₂O (18.2 MΩ cm resistance). Cells were then continuously perfused with KRH solution containing different concentrations of CaCl₂ (0–2.0 mM) at room temperature. Fura-2 fluorescence was captured by a dual-wavelength excitation spectrofluorometer (Photon Technology International, Monmouth Junction, NJ) with fast switching of 340 nm and 380nm excitations. Fluorescence of Fluo-4 was excited at 488 nm and detected by a Zeiss 780 confocal microscope.

In separate studies, HEK-293 cells with inducible expression of RyR₂ were co-transfected with TRIC-A and D1ER cDNA, or TRIC-B and D1ER to determine the impact of TRIC-A/TRIC-B on the ER Ca²⁺ stores. Cells were perfused continuously at room temperature with KRH containing 2 mM CaCl₂. Tetracaine (2 mM) was used to block RyR₂ activity, and caffeine (10 mM) was used to activate RyR₂ to deplete the ER Ca²⁺ store. Cells were excited at 458 nm, and emissions of yellow fluorescent protein (YFP) and cyan fluorescent protein (CFP) from D1ER were captured every 0.5 s. The amount of fluorescence resonance energy transfer (FRET) was determined by the ratio of the emissions at 535 and 480 nm^{5, 32}.

Immunoblotting.

HEK293 cells expressing RyR₂ and TRIC-A were harvested and lysed in RIPA buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% Triton x-100, 1% sodium deoxycholate, 0.1% SDS, 1 mM EDTA) supplemented with a protease inhibitor cocktail (PIC, Sigma) for 10 min at 4°C. PIC contains (final concentration) AEBSF (1 mM), aprotinin (0.8 μM), bestatin (40 μM), E-64 (14 μM), leupeptin (20 μM) and pepstatin A (15 μM). Heart lysates derived from wild type mouse were used for positive control of TRIC-A and RyR₂. Mouse hearts were minced and re-suspended in PIC containing RIPA buffer. Heart protein extraction was done by three-cycle of freeze-thawing on dry ice and extracted with pestle homogenizer with disposable polypropylene pestles (Thermo Fisher Scientific). Cells or tissue lysates were centrifuged (20,000 × g, 15 min. at 4°C) and protein containing supernatant was collected. Protein concentration was determined using a Bio-Rad Protein Assay kit. 10 μg of protein were loaded on 12% or 5% SDS-PAGE and transferred onto a polyvinylidene difluoride membrane for detection of TRIC-A or RyR₂ expressions, respectively. The expression levels of TRIC-A or RyR₂ were visualized using specific antibodies anti-RyR₂ (34C) (1:3000) (Abcam, ab2868) and anti-TRIC-A (1:2000) (Proteintech Group, 19920–1-AP).

Immunofluorescence staining.

Isolated mouse cardiomyocytes were washed by phosphate buffered saline (PBS) and fixed in 4% phosphate-buffered paraformaldehyde (PFA) for 20 min at room temperature, and then permeabilized for 15 min with 0.2% Triton X-100 in PBS. Cells were blocked by 3% BSA for 40 min and then incubated with TRIC-A (Proteintech Group, 19920–1-AP) and RyR₂ antibodies (1:200, Abcam, ab2868) at 4°C overnight. Alexa 647 labeled goat anti-rabbit (Thermo Fischer A21244) and Alexa 546 labeled goat anti-mouse (Thermo Fisher A11003) secondary antibodies were applied and incubated for 1 h at 37 °C. After a brief wash, DAPI was used to stain the nucleus for 5 min and then washed by PBST 5 times. TRIC-A^−/− mice heart was used as negative control for TRIC-A antibody and IgG only staining was used for RyR₂ antibody staining that confirmed the specificity of both antibodies. Images were captured by Zeiss LSM 780 confocal microscope and analyzed by ImageJ.

GST-CTT-A protein expression and beads binding assay.

cDNA coding for the CTT domain of TRIC-A (a.a. 231–299) was cloned into the pGEX4T-1 bacteria expression plasmid. The pGEX-CTT-A plasmid was transformed into E. coli (DH5α), and expression of GST-CTT-A protein was induced with 0.3 mM isopropyl-β-D-thiogalactopyranoside (IPTG, Sigma) for 3.5 h at 30°C. Harvested bacterial pellets were re-suspended in a buffer containing 20 mM Tris-HCl, 200 mM NaCl, and 0.2 mM PMSF (pH 7.5) supplemented with 0.5% protease inhibitor cocktails (PIC, Sigma), 1 mg/ml lysozyme, and 1 mg/ml MgCl₂, and sonicated (Branson Ultrasonics) for 15 s for 5 cycles at 30% power output. The lysates were centrifuged twice at 15,000 rpm for 30 min at 4°C. The cleared supernatants were diluted 1:4 into the suspension buffer, mixed with Glutathione Sepharose Beads (Amersham) and incubated at 4°C overnight on an orbital shaker. The mixture was then loaded into a poly-prep chromatography column (Bio-Rad), allowed to precipitate by gravity, and then washed with 50 ml buffer containing 20 mM Tris-HCl, 200 mM NaCl, 0.2 mM PMSF and PIC, pH 7.5. C1148 cells with stable expression of RyR₁ were described previously^{33, 34}. Microsomal vesicles derived from C1148 cells or SR vesicles derived from rat heart (kindly provided by Dr. Michael Fill, Rush University, Chicago) were lysed in RIPA buffer supplemented with PIC, and loaded to either GST or GST-CTT-A bound beads and incubated at 4°C on an orbital shaker for 3 hours. The bound proteins were washed 4 times with buffer containing 0.5% NP-40, 20 mM Tris (pH 7.4) supplemented with PIC, eluted with sample loading buffer, and loaded onto SDS-PAGE gel and western blotted with anti-RyR (1:3000, Abcam, ab2868).

Channel activity assay on planar bilayer system.

SR vesicles isolated from rat hearts were used for lipid bilayer reconstitution studies of single RyR₂ channels^{35, 36}. The SR vesicles were kindly provided by Dr. Michael Fill at Rush Medical University²⁸. Planar lipid bilayers were formed with a lipid composition of phosphatidylethanolamine and phosphatidylserine (Avanti Polar Lipids, Birmingham, AL) (1:1 ratio, dissolved in decane, 50 mg/ml) across a 200 μm diameter hole in a polystyrene partition. CTT-A and CTT-B peptides were synthesized by GL Biochem (Shanghai, China) with >95% purity. SR vesicles and CTT-A or CTT-B peptides were added to one chamber using cis design, composed of 250 mM CsCl, 0.22 mM CaCl₂, 1 mM EGTA, 10 mM HEPES, pH 7.4. The other chamber, designated as trans, contained 50 mM CsCl, 10 mM HEPES, pH 7.4. Single channel data acquisition and analysis were conducted with pClamp software.

Microinjection of CTT-A and CTT-B peptide.

CTT-A and CTT-B peptides were dissolved in PBS solution at a concentration of 10 μM. Freshly isolated wild type mouse cardiomyocytes were loaded with 5 μM Fluo-4 AM. Microinjection pipettes made of borosilicate glass (Sutter Instrument Co.) were prepared using a micropipette puller (Model P-97, Sutter Instrument Co.). The microfilament inside the pipette ensures that the solution reaches the tip for microinjection. The tip pore of the pipette was adjusted to allow for delivery of 10–20 pico liter of peptide solution in 60 ms using a picospritzer (Parker Instrumentation) coupled to PatchMan micromanipulator (Eppendorf). Changes in intracellular [Ca]_i were monitored under a BioRad confocal microscope.

Modeling and Docking of CTT-A and CTT-B with RyR₂.

The atomic structures of CTT-A and CTT-B, were modeled using the I-TASSER server³⁷. The atomic structure of human RyR₂ (hRyR₂) has not been solved. However, the cryo-EM structures of porcine RyR₂ (pRyR₂) are available in the Protein Data Bank³⁸ (PDB ID: 5goa, open state; PDB ID: 5go9, closed state)³⁹. Due to the large size of RyR₂, traditional homology modeling cannot be used to model human RyR₂ using pRyR₂ structures. Instead, mutations were introduced in the structures of pRyR₂ (5goa and 5go9, respectively) following the hRyR₂ sequence, using UCSF Chimera⁴⁰. The structures of CTT-A and CTT-B were respectively docked onto the homology model of hRyR₂ using our in-house docking software, MDockPP^{41, 42}. The P1 domain, P2 domain, and channel domain were blocked from docking to reduce the enormous search space on the huge RyR₂ protein, as the peptides unlikely interact with these domains. Four potential binding sites of CTT-A/CTT-B were identified, and the representative binding mode in each site was optimized using the minimization tool of UCSF Chimera to remove intermolecular side chain clashes. The differential binding modes in the binding sites were further compared between the open and closed state of the hRyR₂ channel.

Data handling and statistical analysis.

D’Agostino & Pearson normality test was used to evaluate normality of the data. If normal, we used unpaired t-test for single comparison and Tukey’s test for multiple comparisons. If not normal, we used non-parametric test (Mann-Whitney test for single comparison or Dunn’s test for multiple comparisons) to derive the p-value. Before doing a Tuckey’s or Dunn’s test for multiple comparisons, ANOVA or Kruskal-Wallis (if not passed normality test) test was conducted for the whole group.

Data are represented as mean ± standard deviation. A value of P<0.05 was considered significant. All data were analyzed using Excel and Prism 8 software.

RESULTS

Abnormal Ca²⁺ signaling in cardiomyocytes derived from the TRIC-A^−/− mice.

We have generated knockout mice carrying the deletion of either TRIC-A or TRIC-B. While TRIC-A^−/− mice survived past their adolescent age, homozygous ablation of TRIC-B proved lethal as the TRIC-B^−/− mice died at the neonatal stage due to respiratory dysfunction²¹. Moreover, the aggravated embryonic fatality was observed with the TRIC-A^−/−TRIC-B^−/− mice, demonstrating the essential role of TRIC in development^{11, 17}. Using cardiomyocytes derived from the TRIC-A^−/− adult mice and littermate wild type (WT) controls (10–12 weeks age), we examined Ca²⁺ spark signaling properties following electrical pacing. As shown in Fig. 1A, following a 0.5 Hz electrical field stimulation for 20 s, spontaneous Ca²⁺ sparks were observed in the isolated cardiomyocytes upon the termination of the electric stimulation. Compared with the WT cardiomyocytes, the TRIC-A^−/− cardiomyocytes showed less frequent Ca²⁺ sparks (Fig. 1B, left). Interestingly, the amplitudes of individual Ca²⁺ sparks appear to be significantly higher in TRIC-A^−/− cardiomyocytes than those in WT cardiomyocytes (Fig. 1B, middle).

Fig. 1. Compromised Ca²⁺ signaling in cardiomyocytes derived from TRIC-A^−/− mice.

A. Confocal line scan images of Ca²⁺ sparks in isolated cardiomyocytes from WT and TRIC-A^−/− mice, after pacing by 0.5 Hz of electric field stimulation for 20 s. B. Ca²⁺ spark amplitude was higher (p=0.0091, 2-tailed t-test) and spark frequency was lower (p=0.0098, 2-tailed t-test) in TRIC-A^−/− cardiomyocytes, compared with WT cardiomyocytes. Caffeine-induced Ca²⁺ release was significantly higher (p=0.0487, 2-tailed t-test) in TRIC-A^−/− cardiomyocytes. C. Line-scan imaging of electric pacing induced intracellular Ca²⁺ transients, demonstrating the different patterns in WT and TRIC-A^−/− cardiomyocytes. Data from individual cardiomyocytes were plotted. Rising time (D) (P=0.0090, 2-tailed t-test), full-duration-half-maximum (FDHM) (E) (P= <0.0001, Mann-Whitney test), and peak amplitude of electric pacing induced Ca²⁺ transients (F) (P=0.7473, Mann-Whitney test) are plotted for TRIC-A^−/− (red) and WT (black) cardiomyocytes. G. Different propagating patterns of store-overload induced Ca²⁺ waves in WT and TRIC-A^−/− cardiomyocytes. Speed of Ca²⁺ wave propagation (H) (P=0.0001, 2-tailed t-test), percentage of oscillating cells (I) (p=0.0205, 2-tailed t-test), and interval between Ca²⁺ waves (J) (n=165 for WT, n=89 for TRIC-A^−/−, p=0.0047, Mann-Whitney test) were significantly larger in TRIC-A^−/− cardiomyocytes. * P<0.05, ** P<0.01, **** P<0.0001.

We also compared caffeine-induced Ca²⁺ release from the WT and TRIC-A^−/− cardiomyocytes. As illustrated in (Fig. 1B, right), the total caffeine-releasable Ca²⁺ pool from the SR was significantly higher in TRIC-A^−/− than in WT cardiomyocytes. This observation is consistent with our published data with skeletal muscle²⁰ and epithelial cells²¹ where the absence of TRIC-A or TRIC-B led to elevated Ca²⁺ storage inside the SR or ER.

Confocal line scan imaging of Fluo-4 fluorescence revealed different patterns of intracellular Ca²⁺ transients in WT and TRIC-A^−/− cardiomyocytes following electric field stimulation (Fig. 1C). There appeared to be a delayed onset of intracellular Ca²⁺ transients and a prolonged decaying phase of Ca²⁺ transients in the TRIC-A^−/− cardiomyocytes compared with those in the WT cardiomyocytes. On average, the rising time of Ca²⁺ transients was significantly longer in TRIC-A^−/− cardiomyocytes when compared with WT cardiomyocytes (Fig. 1D). The full-duration-half-maximal (FDHM) of Ca²⁺ transients was significantly longer in the TRIC-A^−/− cardiomyocytes (Fig 1E), whereas the amplitude of Ca²⁺ transients showed no significant difference (Fig. 1F) between the WT and TRIC-A^−/− cardiomyocytes. Studies from other investigators demonstrated that reduced RyR₂ channel activity could contribute to the delayed time-to-peak and prolonged decaying phase of electric pacing-induced Ca²⁺ release from cardiomyocytes^43–45. This suggests the possibility that ablation of TRIC-A in cardiomyocytes could lead to reduced function of the RyR₂ channel.

Using the resonant scan mode of the Nikon A1R confocal microscope, we performed continuous x-y imaging of spontaneous Ca²⁺ waves (at a rate of 130 ms per frame) in the isolated cardiomyocytes upon elevation of the extracellular Ca²⁺ from 1.8 mM to 10 mM. Such fast time-lapse imaging enabled us to examine the impact of TRIC-A ablation on the speed of Ca²⁺ wave propagation. As shown in Fig. 1G, the speed of Ca²⁺ wave propagation appeared to be faster in WT than in TRIC-A^−/− cardiomyocytes (see Online Movie I and II). Measurements from multiple experiments support the statistical difference in the speed of Ca²⁺ wave propagation between WT and TRIC-A^−/− cardiomyocytes (Fig. 1H). Moreover, the percentage of cells that displayed Ca²⁺-overload induced spontaneous Ca²⁺ waves was significantly lower in TRIC-A^−/− cardiomyocytes than that in WT cardiomyocytes (Fig. 1I). In Online Movie III and IV, one could see the different behavior of Ca²⁺ waves in WT and TRIC-A^−/− cardiomyocytes.

Isoproterenol treatment induces irregular ECG and cardiac fibrosis in TRIC-A^−/− mice.

Under basal condition, there was no clear difference in electrocardiograms (ECG) recorded in the TRIC-A^−/− mice and WT littermates (Fig. 2A). However, following a bolus intraperitoneal injection of isoproterenol (ISO, 80 mg/kg), irregular ECG was observed with the TRIC-A^−/− mice at 1 hour post ISO injection, whereas the WT mice displayed regular patterns of ECG. Analysis of the R-R intervals between the individual ECGs demonstrated large variations, confirming the irregularity of ISO-induced ECG with the TRIC-A^−/− mice (Fig. 2B).

Fig. 2. Irregular ECG and increased cardiac fibrosis in TRIC-A^−/− mice upon stimulation with isoproterenol.

A. Representative traces of ECG in WT and TRIC-A^−/− mice before and after treatment of ISO. B. Distribution of ECG R-R intervals at 1 hr after 80 mg/kg ISO injection. (P<0.0001, Mann-Whitney test) C. Trichrome stain of WT and TRIC-A^−/− heart at 2 weeks after treatment of PBS or ISO (60 mg/kg/day). D. Summary data with percentage of fibrosis area in WT and TRIC-A^−/− heart stained by Mason’s trichrome (P<0.001 across the groups, ANOVA test; WT+ISO vs TRIC-A^−/−+ISO: P<0.0001, Tukey’s test).

We next tested if ablation of TRIC-A could alter the response of the mice to chronic treatment with ISO. With chronic treatment of ISO (60 mg/kg/day for 2 weeks, delivered using an Alzet mini-osmotic pump), the TRIC-A^−/− heart would develop extensive fibrosis than the WT heart (Fig. 2C). Mason’s trichrome staining revealed significant increase in fibrosis in the heart tissue derived from the TRIC-A^−/− mice at 2 weeks post ISO treatment, compared with the WT mice (Fig. 2D).

Thus, chronic exposure to ISO likely caused increased death of cardiomyocytes and fibrotic remodeling, which may be linked to the overload of SR Ca²⁺ associated with ablation of TRIC-A. In response to ISO, SR Ca²⁺ overload may cause mitochondria Ca²⁺ toxicity and apoptosis in the TRIC-A^−/− cardiomyocytes. Further dissecting the role of TRIC-A in modulating the Ca²⁺ signaling crosstalk from SR to mitochondria will require dedicated effort in future studies.

Co-expression of TRIC-A and RyR₂ in HEK293 cells modulates SOICR.

We used HEK293 cells with tetracycline-inducible expression of RyR₂ to investigate the impact of TRIC-A and TRIC-B on RyR₂-mediated intracellular Ca²⁺ signaling^{5, 6} (Fig. 3A). In this model, the elevation of extracellular [Ca]_o lead to increased Ca²⁺ content inside the ER, which triggered the opening of the RyR₂ channel via its luminal Ca²⁺ sensing mechanism, leading to the appearance of SOICR. Previously, we showed that co-expression of RyR₁ and MG29 in Chinese hamster ovary (CHO) cells caused depletion of the ER Ca²⁺ store and consequently apoptotic cell death⁴⁶. Interestingly, we found that transient co-expression of RyR₂ and TRIC-A, but not TRIC-B, in HEK293 cells led to changes in cell morphology, and a large portion of the cells expressing TRIC-A and RyR₂ would die within 36 hours (Fig. 3B), suggesting cytotoxicity by potential perturbation of ER Ca²⁺ homeostasis.

Fig. 3. Effect of TRIC-A and TRIC-B on RyR₂-mediated SOICR in HEK293 cells.

A. Western blot of RyR₂ and TRIC-A in HEK293 cells with tetracycline-inducible expression of RyR₂. Mouse heart lysate was used as positive control of RyR₂ and TRIC-A expression. B. Representative images of HEK293 cells without RyR₂ expression (top) and with RyR₂ expression (bottom). Following baculovirus infection, only cells co-expressing RyR₂ and TRIC-A became round and detached from the plate (middle-bottom panel). Co-expression of RyR₂ and TRIC-B had no effect on cell viability. Scale bar, 200 μm. C. Line-scan image of Ca²⁺ oscillation, representing SOICR, in HEK293 cells expressing RyR₂ alone (non-red cell) and lack of Ca²⁺ oscillation in HEK293 cells co-expressing RyR₂ and TRIC-A (red cell). D. Representative Fura-2 Ca²⁺ measurement in HEK293 cells with stable expression of RyR₂ (blue trace), co-expression of TRIC-A and RyR₂ (red trace), and co-expression of TRIC-B and RyR₂ (green trace). E. Summary data with percentage of cells showing spontaneous Ca²⁺ oscillations in 0.5mM Ca²⁺. (P=0.0003 across groups, 2-way ANOVA; TRIC-A vs NT: P<0.0001, Sidak’s test). F. Co-expression of TRIC-A and RyR₂ alters the dependence of SOICR on extracellular Ca²⁺ concentration. Co-expression of TRIC-B and RyR₂ produced similar response of SOICR with changes of [Ca]_o, as cells transfected with pCMS-RFP vector control. Data from 4 individual experiments were used for the statistical analyses. (P<0.0001 across groups, 2-way ANOVA; RyR₂+TRIC-A vs RyR₂+TRIC-B: P<0.0001, Sidak’s test) ****P<0.0001.

For assaying the effect of TRIC-A in HEK293 cells, we used a dual-reporter plasmid with mRFP-expression cassette driven by a separate promoter. This allows selection of transfected vs non-transfected cells in the same dish (Fig. 3C). From the line-scan measurements, it is clear that cells co-expressing RyR₂ and TRIC-A (red labeled) did not show oscillating patterns of Ca²⁺ signaling, unlike those cells expressing RyR₂ alone, which often display Ca²⁺ oscillations. A representative trace of Ca²⁺ measurement in a HEK293 cell with co-expression of TRIC-A and RyR₂ (labelled red, Fig. 3D), illustrates the apparent absence of spontaneous Ca²⁺ oscillations. After treatment of tetracaine, a RyR₂ blocker, SOICR was diminished and cytosolic Ca²⁺ level was decreased. At the end of the recording, caffeine was introduced to the dish to test the release of Ca²⁺ via RyR₂. Ca²⁺ release from TRIC-A transfected cells triggered by caffeine confirmed the apparent suppression of SOICR is not due to the dysfunction of RyR₂ (Fig. 3D, top). This effect appears to be specific to TRIC-A, as co-expression with TRIC-B did not affect RyR₂-mediated SOICR in HEK293 cells (Fig. 3D, bottom). Summary data from multiple experiments is shown in Fig. 3E, indicating that TRIC-A, not TRIC-B, could modulate SOICR in HEK293 cells.

Studies from Chen and colleagues demonstrated that the frequency of RyR₂-mediated SOICR in HEK293 cells was dependent on the concentration of Ca²⁺ in the extracellular solution⁵. Here we tested if co-expression of TRIC-A and RyR₂ could modulate SOICR in a Ca²⁺-dependent manner. The data shown in Fig. 3F clearly suggest that the presence of TRIC-A could alter [Ca]_o-dependent activation of SOICR. Moreover, this effect is TRIC-A specific, since cells with co-expression of TRIC-B and RyR₂ showed similar response as those transfected with vector alone (as control).

To elucidate the mechanism of TRIC-A mediated modulation of SOICR, we used D1ER, a Ca²⁺ biosensor targeted to ER, to directly measure ER luminal Ca²⁺ level^{5, 32, 47}. We followed the approach developed by Bers and colleagues to assay the ER Ca²⁺ content using tetracaine to inhibit RyR₂ activity^{48, 49}. This will allow for testing to what extent co-expression of TRIC-A alters the Ca²⁺ leak rate across the ER. As shown in Fig. 4A, HEK293 cells co-expressing RyR₂ and TRIC-A displayed reduced ER Ca²⁺ content (measured with the D1ER probe), when compared with cells expressing RyR₂ alone. While cells expressing RyR₂ alone show oscillating patterns of ER Ca²⁺ content (Fig. 4A, blue trace) that mirrors the SOICR signal (Fig. 4B), co-expression with TRIC-A led to apparent suppression of the ER Ca²⁺ content below the oscillating threshold levels (Fig. 4A, red trace). In addition, we observed less Ca²⁺ release by direct caffeine treatment in TRIC-A expressing cells (Fig. 4B). This data further supports that TRIC-A reduced the ER Ca²⁺ content in the RYR₂ expressing HEK293 cells.

Fig. 4. Effect of TRIC-A on changes in ER Ca²⁺ load require interaction with RyR₂.

A. ER luminal Ca²⁺ measured by D1ER in cells expressing RyR₂ (blue), or RyR₂ + TRIC-A (red). 2 mM tetracaine inhibited ER Ca²⁺ oscillations and increased ER Ca²⁺ store and 10 mM caffeine caused depletion of ER Ca²⁺ store. (P=0.0042, 2-tailed t-test) B. Caffeine induced Ca²⁺ release from the ER store in cells expressing RyR₂ alone (blue) and in cells with co-expression of TRIC-A and RyR₂ (red) (p=0.0001, 2-tailed t-test) . C. Cytosolic Ca²⁺ release by ATP in HEK293 cells without RyR₂ expression showed similar response with or without TRIC-A expression. (P=0.2667, Mann-Whitney test) D. D1ER measurement of CPA (SERCA inhibitor) treatment showed that TRIC-A overexpression did not affect the total ER Ca²⁺ and the ER Ca²⁺ uptake in HEK293 cells (2 mM Ca²⁺: P=0.2627; 2 mM Ca²⁺+CPA: P=0.2695, 2-way ANOVA, Sidak’s test).

The effect of TRIC-A on ER Ca²⁺ handling likely require the participation of the RyR₂ channel, since transfection of TRIC-A in HEK293 cells (in the absence of RyR₂) did not affect the ER Ca²⁺ content as measured by ATP-induced Ca²⁺ release through IP₃R (Fig. 4C) nor cytotoxicity (Fig. 3B). Moreover, when blocking ER Ca²⁺ uptake by cyclopiazonic acid (CPA), the decline of luminal Ca²⁺ is not affected by TRIC-A (Fig. 4D). These results support that altered ER Ca²⁺ signaling by TRIC-A is RyR₂ dependent.

TRIC-A interact with RyR₂ through its carboxyl-terminal tail (CTT) domain.

We performed immunohistochemical (IHC) staining of TRIC-A and RyR₂ in cardiomyocytes derived from the WT mice. The clear overlap between TRIC-A and RyR₂ could be observed (Fig. 5A). Images were analyzed by Pearson and Manders coefficients and showed significant co-localization (Fig. 5B).

Fig. 5. Co-localization of TRIC-A and RyR₂ in mouse cardiomyocytes and CTT-A interaction with RyR.

A. Fluorescent staining shows TRIC-A and RyR₂ are highly co-localized together in mouse isolated cardiomyocytes. B. Images were analyzed by scatter plot, Pearson coefficient (R) and Mander’s coefficient (M1 and M2). Bright yellow color of the scatterplot across the line indicates stronger co-localization. A value of 1 for R, M1/M2 indicates perfect co-localization and 0 indicates no co-localization; a threshold of 0.5 usually indicates good co-localization. M1 represents the degree of green fluorescence (TRIC-A) overlay with red fluorescence (RyR₂); M2 represents the degree of red fluorescence overlay with green fluorescence. C. CTT-A and CTT-B diverge from each other. CTT-A contains a histidine-rich motif (orang) and a polylysine domain (purple), which are flanked by a hydrophobic domain (blue). Such structure is similar to the intracellular-loop joining repeats II and III of the L-type Ca²⁺ channel. Only the polylysine domain is present in CTT-B. D. GST-CTT-A can pull down RyR₁ from C1148 cells. E. GST-CTT-A can pull down RyR₂ from SR vesicle isolated from rat heart, and three other candidate proteins (labeled as a, b and c.). Western blot confirmed the pull-down of RyR₂ from heart SR vesicle. GST alone does not pull down RyR₂. F. TRIC-A could be immuno-precipitated with RyR₂ in HEK293 cells co-expressing RyR₂ and TRIC-A.

Topology analysis revealed that both TRIC-A and TRIC-B contained large CTT domains that reside in the cytosol (Fig. 5C). This CTT domain likely represents a flexible structure that interfered with crystallization and thus was not included in the structural determination of TRIC channels by Liu and colleagues⁵⁰. The CTTs of TRIC-A and TRIC-B diverge from each other. CTT-A contains a histidine-rich motif and a polylysine domain that are flanked by a hydrophobic domain. This structure is similar to the intracellular loop joining repeats II and III of the L-type Ca²⁺ channel, which has been shown to be a critical domain that regulates the activity of the SR Ca²⁺ release in muscle cells^51–53. Only the polylysine domain appears in the CTT-B domain.

To test if the divergent CTT-A constitutes a site for interaction with the RyR channel, we generated a GST-fusion peptide containing the CTT-A (a.a. 231–299) and CTT-B (a.a. 229–291). These recombinant peptides were purified from E. coli as GST-fusion proteins. Using a protein pull-down assay, we observed that CTT-A could interact with RyR₁ stably expressed in CHO cells^{54, 55} (Fig. 5D), and RyR₂ from rat heavy SR vesicles (Fig. 5E). Using antibodies against RyR₂, we could co-immunoprecipitate TRIC-A from HEK293 cells that co-express RyR₂ and TRIC-A (Fig. 5F). These interactions appear to be specific for CTT-A, as GST peptide alone could not pull down either RyR₁ or RyR₂. In addition to RyR₂, several other candidate proteins were pulled-down by CTT-A as marked by a, b and c in Fig. 5E. MALDI-MS identification revealed a-band contains SERCA, MPP1, and C2CD2; b-band contains TRIC-A, destrin, Rtn2, and proteolytic fragment derived from RyR₂; and the c-band contains mostly proteolytic fragment for RyR₂ and SERCA.

TRIC-A carboxyl-terminal tail peptide increases RyR₂ channel activity.

Based on these findings, we generated chimeric constructs of TRIC-AB or TRIC-BA. TRIC-AB contains a.a. 1–230 from TRIC-A plus a.a. 229–291 from TRIC-B. TRIC-BA contains a.a. 1–228 from TRIC-B plus a.a. 231–299 from TRIC-A. These constructs were co-expressed with RyR₂ in HEK293 cells. We found that replacement of the CTT domain in TRIC-A with CTT-B could alleviate the impact on SOICR (Fig. 6). Consistent with the data shown in Fig. 3, co-expression of TRIC-A with RyR₂ could reduce the spontaneous Ca²⁺ oscillation in HEK293 cells (Fig. 6A, red color), whereas co-expression of TRIC-B with RyR₂ had no effect on SOICR activity (Fig. 6A, green color). Co-expression of TRIC-AB with RyR₂ resulted in the elimination of the inhibitory effect of TRIC-A on SOICR (Fig. 6A, orange color). This finding supports the important function of CTT-A on RyR₂-mediated SOICR in HEK293 cells. Interestingly, co-expression of TRIC-BA with RyR₂ also had minimal effect on RyR₂-mediated SOICR (Fig. 6B, blue color), suggesting the possibility that CTT-A alone in the context of TRIC-B might lose its capability to interact with the RyR₂ channel.

Fig 6. CTT-A directly modulates RyR₂ channel function.

A. TRIC-A, TRIC-B, TRIC-AB and TRIC-BA were transiently transfected into HEK293 cells with stable expression of RyR₂. Represent traces of spontaneous Ca²⁺ oscillations in 2 mM Ca²⁺ were shown. B. Percentage of cells with SOICR was quantified. (P<0.0001 across groups, ANOVA; Control vs TRIC-A: P<0.0001; TRIC-A vs TRIC-AB: P=0.0009, Tukey’s test) C. Caffeine-induced Ca²⁺ release was used to quantify the impact of the various TRIC-A and TRIC-B constructs on ER Ca²⁺ stores. (P<0.001 for Kruskal-Wallis test for whole group; TRIC-A vs TRIC-AB: P<0.0001, TRIC-A vs Control: P<0.0001, Dunn’s test). D. Reconstitution of RyR₂ channels in lipid bilayer. Single channel activity was measured with 250 mM Cs (cis)/50 mM Cs (trans) and 1 μM free Ca²⁺ in the cis solution. Holding potential was −60 mV. Open probability of individual channels was calculated from a continuous recording of 5 min, and data from multiple measurements were averaged. Addition of 1 μM CTT-A to cis solution led to significant increase in the bursting pattern of RyR₂ channel, whereas addition of 10 μM TRIC-B-tail did not affect RyR₂ activity. E. Confocal line scan images of Ca²⁺ events measured by Fluo-4 after microinjection of CTT-A or CTT-B into isolated cardiomyocytes. F. Statistical analysis of RyR₂ channel open probabilities. (P<0.0001 across groups, ANOVA; TRIC-A vs Control: P<0.0001, TRIC-A vs TRIC-B: P<0.0001, Tukey’s test) G. CTT-A induced significantly more Ca²⁺ release events in cardiomyocytes than CTT-B and PBS does. (CTT-A vs PBS: P=0.0174; CTT-A vs CTT-B: P=0.0301, chi-square test)

We used 10 mM caffeine to induce activation of the RyR₂ channel as the assessment of the total ER Ca²⁺ load (Fig. 6C). Clearly, only TRIC-A had a significant impact on the caffeine response, whereas TRIC-B, TRIC-AB and RIC-BA had a negligible effect on ER Ca²⁺ content in HEK293 cells expressing RyR₂.

For direct evaluation of the CTT-A and CTT-B peptides on RyR₂ channel function, we reconstituted the RyR₂ channels from rat cardiac muscle into the lipid bilayer membrane. As shown in Fig. 6D, the addition of CTT-A to the cis-cytoplasmic solution significantly enhanced the RyR₂ channel activity (n=5). The effect appears to be specific for CTT-A, as the addition of CTT-B peptide did not result in significant changes in RyR₂ channel activity (n=4). Data from multiple experiments are summarized in Fig. 6F.

To further evaluate the functional effect of CTT-A and CTT-B peptides on Ca²⁺ signaling in cardiomyocytes, we performed microinjection of CTT-A or CTT-B peptides into isolated mouse cardiomyocytes (Fig. 6E). We found that microinjection of CTT-A peptide elicited significantly more intracellular Ca²⁺ events as compared to those cardiomyocytes injected with CTT-B peptide (Fig. 6G).

CTT-A modulates Ca²⁺ spark activity in cardiomyocytes derived from TRIC-A^−/− cardiomyocytes.

We followed the protocol of Lukyanenko and Gyorke²⁷ to characterize the Ca²⁺ spark activity in saponin-permeabilized cardiomyocytes derived from wild type and TRIC-A^−/− mice (Fig. 7A). Saponin permeabilization of the sarcolemma allows access of CTT-A and CTT-B to the intracellular organelles of the cardiomyocytes. Starting from a basal condition of 100 nM free Ca²⁺ in the cytosolic solution, we observed the significantly lower frequency of spontaneous Ca²⁺ sparks in the TRIC-A^−/− cardiomyocytes compared with the wild type cardiomyocytes (Fig. 7B). For testing the effect of CTT-A or CTT-B on Ca²⁺ spark signaling, the free cytosolic Ca²⁺ was reduced to 50 nM, which led to lower activity of Ca²⁺ spark events in the TRIC-A^−/− cardiomyocytes (Fig. 7C). When CTT-A (10 μM) was added to the cytosolic solution, transient elevation of Ca²⁺ spark activity was observed in the early phase (within 3 min after CTT-A addition), which is followed by a decline of Ca²⁺ spark activity during the later phase (starting from ~3 min following CTT-A addition) (Fig. 7C, left). In the presence of CTT-A, we frequently observed Ca²⁺ spark events that occur in a consecutive streaming pattern of opening-closings, reflecting increased RyR₂ channel activity. Such streaming patterns of Ca²⁺ spark events were rarely observed under control conditions and did not appear to change with either early or late treatment of CTT-B (10 μM) (Fig. 7C, right). Diary plot of Ca²⁺ spark activity over the 10 min time period remained steady under the control condition of 50 nM free Ca²⁺ in the cytosol (Fig. 7D, left). Clearly, the addition of CTT-A led to transient stimulation of the spontaneous Ca²⁺ spark events (Fig 7D, middle). In contrast, CTT-B did not have a measurable impact on Ca²⁺ spark activity (Fig 7D, right). Data from multiple experiments were summarized in Fig. 7E, showing a significant increase in CTT-A-induced transient Ca²⁺ spark activity which did not occur with CTT-B. Consistent with the streaming pattern of Ca²⁺ spark events observed with CTT-A treatment that may cause depletion of the SR Ca²⁺ store (Fig. 7C, left), quantitative analysis revealed significant reduction of Ca²⁺ spark amplitude in the presence of CTT-A, but not CTT-B (Fig. 7F). The effect of CTT-A on Ca²⁺ sparks measured from the saponin-permeabilized cardiomyocytes derived from in the TRIC-A^−/− mice is reversible, as washout of CTT-A led to the recovery of Ca²⁺ spark activities to the basal level (Supplemental Online Figure I).

Fig. 7. CTT-A enhances Ca²⁺ spark activities in permeabilized TRIC-A^−/− cardiomyocytes.

A. Line-scan images of Ca²⁺ sparks in saponin-permeabilized wild type (top) and TRIC-A^−/− cardiomyocytes (bottom). B. Dot plot of spontaneous Ca²⁺ spark activities in wild type and TRIC-A cardiomyocytes at 100 nM free cytosolic Ca²⁺. (P =0.0147, 2-tailed t-test) . C. Left panel shows distinct streaming-pattern of Ca²⁺ sparks observed in TRIC-A^−/− cardiomyocytes following 1 min application of 10 μM CTT-A (early), with apparent suppression of Ca²⁺ sparks at 5 min after CTT-A treatment (late). Treatment with CTT-B (10 μM) did not induce measurable changes in Ca²⁺ spark activity (right panels). All Ca²⁺ sparks were measured with a cytosolic solution containing 50 nM free Ca²⁺. D. Diary plots of Ca²⁺ spark activities under control condition of 50 nM free Ca²⁺ (left), with addition of CTT-A (middle) and CTT-B (right). E. Scattered plots of Ca²⁺ spark events from individual cardiomyocytes. Two separate groups of experiments were performed with CTT-A and CTT-B, all with cytosolic free Ca²⁺ buffered at 50 nM. Early – indicates 0–3 min following CTT-A/B treatment; late – indicates 3–6 min following CTT-A/B treatment. (P<0.0001 across groups, ANOVA; 50nM Ca²⁺ vs early: P=0.0130, early vs late: P=0.0002, Tukey’s test) F. Averaged value of Ca²⁺ spark amplitudes measured at 50 nM free cytosolic Ca²⁺ (Control, blue, n=3939), with early addition of CTT-A (red, n=3434) or CTT-B (green, n=2126). (P<0.0001 for ANOVA test; P<0.0001, CTT-A vs Control, CTT-A vs CTT-B, Tukey’s test) **P<0.001, ****P<0.0001

CTT-A and CTT-B dock to different motifs on RyR_2.

To gain insights into the interaction of CTT-A or CTT-B with the RyR₂ channel, we performed modeling and docking studies using our established molecular dynamic simulation method³⁷. We first compared the 3-D structure of CTT-A with CTT-B and found divergent structural patterns between the two peptides (Fig. 8A). Interestingly, the close similarity between CTT-A and the II-III loop peptide of the dihydropyridine receptor (DHPR) was observed (Fig. 8B).

Fig. 8. Docking of CTT-A and CTT-B with open and closed-state of the RyR₂ channel.

A. Structural comparison between CTT-A (cyan) and CTT-B (pink) in ribbon representation. The histidine-rich motif in CTT-A is colored red. B. Structural comparison of CTT-A, CTT-B and the II-III loop of DHPR. C. Four potential binding sites of CTT-A interaction with hRyR₂ were predicted from molecular docking. Both hRyR₂ (grey) and CTT-A are plotted in ribbon representation. CTT-A is colored differently at the four sites. D. CTT-A and CTT-B bind to hRyR₂ distinctively on Site 3 and Site 4. Left - binding to the closed structure of hRyR₂; right - binding to the open structure of hRyR₂. CTT-A (blue), CTT-B (pink) and hRyR₂ (grey) are plotted in ribbon representation.

Molecular docking of CTT-A and CTT-B onto the human RyR₂ (hRyR₂) showed four potential binding sites, which were numbered according to their rankings by the energy scoring function in MDockPP⁵⁶. Representative binding modes of CTT-A at these four sites on the closed structure of hRyR₂ are plotted in Fig. 8C. Among them, Site 1 and Site 2 exist in both open and closed structures of hRyR₂ for both CTT-A and CTT-B. Site 1, the best ranked site, is the hinge domain of hRyR₂³⁹, which has been shown to play a critical role in the activation of the RyR₂ channel. Site 2 consists of binding domain with FKBP12 (see PDB ID: 3J8H⁵⁷), which has demonstrated function in control of RyR channel opening. The zoom-in picture of CCT-A interacting with the closed structure of hRyR² is shown in Supplemental Online Fig. II.

Interestingly, CTT-A and CTT-B bind distinctively on Site 3 and Site 4 (Fig. 8D). Site 3 locates on the N-terminal domain (NTD), and Site 4 locates on the SPRY domain. CTT-A binds to Site 3 for both the open and closed state of hRyR₂, whereas CTT-B binds to Site 3 only for the closed state of hRyR₂. CTT-A binds to Site 4 only in the closed state, whereas CTT-B binds to Site 4 only in the open state. Previous studies by Tae et al showed that the SPRY domain of RyR potentially contributes to binding to DHPR for control of RyR channel activity⁵⁸. While our modeling study suggested conformation-dependent interaction of CTT-A with the SPRY domain, further biochemical and structure-function studies will be required to test if the modeled CTT-A/SPRY interaction or CTT-A’s binding to other domains of RyR₂ indeed play a role in the control of Ca²⁺ release from the SR.

DISCUSSION

The present study uncovered a novel function for TRIC-A as an accessory protein that interacts with RyR₂ to modulate intracellular Ca²⁺ signaling. Aside from its recognized role as a counter-ion channel that participates in excitation-contraction coupling of striated muscles, the physiological function of TRIC-A in heart physiology and disease has remained largely unexplored. Even with the recent resolution of the crystal structure of the TRIC channels, the interacting partners for TRIC have yet to be defined. We showed that CTT of TRIC-A, an important portion of the channel that is left out of the crystal structure determination, constitutes an active motif that interacts with RyR₂ to control SOICR function. We found that TRIC-B is ineffective in control of RyR₂-mediated SOICR in cultured cells, suggesting the specificity of TRIC-A interaction with the RyR₂ channel. Using reconstitution studies and Ca²⁺ spark measurements, we demonstrated that the synthetic CTT-A peptide could directly enhance the RyR₂ channel activity and stimulate intracellular Ca²⁺ release in isolated cardiomyocytes.

While many studies have suggested that altered function of SOICR from the SR in cardiomyocytes may contribute to the development of cardiac arrhythmias^{32, 47, 59–61}, searching for accessory proteins that modulate RyR₂ channel function and SR Ca²⁺ homeostasis should yield important clues to the function of SOICR in physiological and pathophysiological settings. We showed that TRIC-A ablation led to SR Ca²⁺ overload, yet the spontaneous Ca²⁺ sparks and store-overload induced Ca²⁺ waves were less frequent in cardiomyocytes derived from the TRIC-A^−/− mice compared with those from the WT mice. Moreover, such SR Ca²⁺ overload did not trigger SOICR in the TRIC-A^−/− cardiomyocytes under basal conditions. Through the determination of the ER Ca²⁺ content, we showed that overexpression of TRIC-A could increase Ca²⁺ leak across the ER in HEK293 cells. The TRIC-A mediated increase of ER Ca²⁺ leak reflects direct activation of the RyR₂ channel, for expression of TRIC-A alone in the absence of RyR₂ did not affect ER Ca²⁺ storage or passive Ca²⁺ leak across the ER. From these studies, we conclude that TRIC-A constitutes a physiological component of the SR Ca²⁺ release machinery, and TRIC-A deficiency could render RyR₂ channels less sensitive to physiological activation of Ca²⁺ signaling in cardiac muscle. The vulnerability of the TRIC-A knockout mouse heart to handling of isoproterenol stimulation further support the important physiological function of TRIC-A in heart physiology and disease.

It is known that changes in RyR₂ activity can impact the frequency, as well as the speed of stress-induced Ca²⁺ waves in cardiomyocytes. Studies from other investigators have shown that experimental manipulations or mutagenesis of the RyR₂ channel that caused reduction of the RyR₂ channel activity led to slower onset and prolonged decaying phase of voltage-induced Ca²⁺ transients and reduced speed of Ca²⁺ wave propagation^43–45; and conversely mutations of RyR₂ (e.g. R4496C) that enhance RyR₂ channel activity had the opposite effect on intracellular Ca²⁺ transients⁶². Moreover, Ca²⁺ wave velocity increased with treatment of caffeine and decreased with treatment of tetracaine or flecainide⁶². Ullrich et al⁶³ reported that cardiomyocytes expressing RyR₂-S2808A displayed slower Ca²⁺ wave velocity. Our data demonstrated that cardiomyocytes derived from the TRIC-A^−/− mice displayed reduced frequency of Ca²⁺ sparks and reduced speed of Ca²⁺ wave propagation, likely reflecting reduced RyR₂ activity. When cardiomyocytes are exposed to a high Ca²⁺ (10 mM) extracellular solution, less frequent store-overload induced Ca²⁺ waves were observed in the TRIC-A^−/− cardiomyocytes than those in the WT cardiomyocytes. Moreover, the frequency of stress-induced Ca²⁺ waves was less in the TRIC-A^−/− cardiomyocytes than that in the WT cardiomyocytes. Following electric pacing, there was a delayed onset of intracellular Ca²⁺ transients and prolonged decaying phase of Ca²⁺ transients in the TRIC-A^−/− cardiomyocytes compared with those in the WT cardiomyocytes. Similar observations were also made by other investigators who reported extended decay of Ca²⁺ transients in cardiomyocytes derived from the A4860G mice (with reduced RyR₂ activity)⁴⁵ and from the RyR₂-Ex3-del^+/− mice⁴⁴, indicating that alteration of RyR₂ activity alone would impact the decaying phase of intracellular Ca²⁺ transient. However, the apparent delayed phase of intracellular Ca²⁺ release from the TRIC-A^−/− cardiomyocytes may involve additional actions on other components of Ca²⁺ uptake and release across the SR, e.g. counter-ion movement or SERCA activity. While our data support the notion that alteration of RyR₂ activity associated with genetic ablation of TRIC-A could be a main contributing factor for the dysfunctional intracellular Ca²⁺ signaling in cardiomyocytes, we could not rule out the possibility that changes in SERCA activity could also contribute to the altered Ca²⁺ handling in the TRIC-A^−/− cardiomyocytes.

Smith and O’Neill⁶⁴ described the effect of low ATP and tetracaine on spontaneous Ca²⁺ waves in permeabilized cardiomyocytes. Lower ATP was shown to cause reduced frequency and slower Ca²⁺ wave propagation, and the changes in Ca²⁺ wave characteristics in 0.5 mM ATP were similar to those observed during perfusion with 50 μM tetracaine. Smith and O’Neill⁶⁴ also reported that the SR Ca²⁺ content increased as ATP was reduced or tetracaine was increased. Both tetracaine and lowering ATP are known to cause reduction of the RyR₂ channel activity. Many features of the changes in Ca²⁺ sparks and Ca²⁺ waves reported by Smith and O’Neill⁶⁴ are very similar to what we observed with the TRIC-A^−/− cardiomyocytes. This further support the notion that ablation of TRIC-A leads to reduced RyR₂ activity in cardiomyocytes.

A recent study from Sitsapesan and colleagues found that in skeletal muscle, the RyR₁ channel from the TRIC-A^−/− mice displays increased sensitivity to Mg²⁺ inhibition, defective response to protein kinase A phosphorylation, and physiological activator such as ATP is less effective in activating the individual RyR₁ channels reconstituted into the lipid bilayer membrane²². They also reported that the Ca²⁺-dependent control of RyR₁ channel was not altered in the absence of TRIC-A, suggesting that the modulatory effect of TRIC-A may involve other factors. These findings are consistent with a potential role of TRIC-A as an enhancer of RyR channels, such that the absence of TRIC-A leads to reduced RyR channel function under stress conditions.

Crystal structures of several TRIC channels from invertebrate and prokaryotic TRIC channels have been reported^23–26. All of these TRIC proteins share very similar folding and trimeric organizations. Interestingly, all of these crystal structures lack the highly flexible CTT domain, which investigators claim may decrease the stability of the crystal^23–26. Our findings that TRIC-A directly modulates RyR₂-mediated SOICR may have wide implications in cardiovascular research. Potential therapeutic interventions can be used to target the functional interaction between TRIC-A and RyR₂ to restore defective Ca²⁺ signaling in cardiovascular diseases.

In our studies testing CTT-A on Ca²⁺ sparks measured in permeabilized cardiomyocytes, we found an interesting phenomenon. The stimulatory effect of CTT-A on Ca²⁺ sparks was transient, and there appeared to be a delayed inhibitory effect on Ca²⁺ sparks with prolonged incubation of CTT-A. The transient increase in Ca²⁺ sparks induced by CTT-A may reduce the SR Ca²⁺ content that would reduce the frequency of Ca²⁺ sparks after the CTT-A stimulation, or alternatively the inhibitory effect of CTT-A likely reflected other components of Ca²⁺ movement across the SR membrane. Since the functional effect on intracellular Ca²⁺ release was only observed with CTT-A, not CTT-B, differential interaction between RyR₂ and CTT-A would be required. Using the available information about the 3-D structure of the RyR₂ channel³⁹, we performed modeling and docking studies and discovered potential sites of interaction between RyR₂ and CTT-A/CTT-B. Of the four sites plotted in Fig. 8C (see also Supplemental Online Fig. II), one of them is particularly interesting. The SPRY domain of RyR₂ constitutes a potential binding site (site 4) for interaction with CTT-A and CTT-B. CTT-A preferentially accesses the SPRY domain when the channel is in the closed state, whereas CTT-B appears to recognize the SPRY domain when the channel is in the open state. Several studies have demonstrated that the SPRY domain of RyR plays an important role in modulation of the Ca²⁺ release channel activity^{58, 65}. It should be emphasized that our modeling studies only provide a hint of possible domain-domain interaction between CTT-A/B and RyR. Further mutagenesis studies will be required to ascertain the functional impact of such interaction in control of intracellular Ca²⁺ release in muscle cells.

Previous studies from Chen and colleagues showed that reconstitution of RyR₂ gain-of-function mutants in HEK293 cells could lead to enhanced SOICR as reflected by increased Ca²⁺ oscillations and/or reduced threshold of luminal Ca²⁺ to trigger Ca²⁺ oscillations^{5, 6, 59, 66}. Interestingly, our study revealed that co-expression of TRIC-A and RyR₂ led to an apparent reduction of Ca²⁺ oscillation in HEK293 cells, which seems unexpected based on the role of TRIC-A as an enhancer of RyR₂ function. In fact, when TRIC-A was introduced into HEK293 cells, increased Ca²⁺ leak across the ER was observed that may reflect direct activation of the RyR₂ channel. It is possible that the TRIC-A/RyR₂-mediated ER Ca²⁺ leak may take place in the form of small Ca²⁺ release events which are un-synchronized and therefore do not appear as Ca²⁺ oscillations. One would ask why TRIC-A would lead to over activation of RyR₂ in HEK293 cells but not in native cardiomyocytes. It is known that coordinated activation and inactivation mechanisms must exist in cardiomyocytes for the proper function of Ca²⁺ signaling in the heart, and many of the stabilizers or inhibitory factors of RyR₂-mediated Ca²⁺ release are likely absent in HEK293 cells. One avenue of future study may take advantage of HEK293 cells to reconstitute the potential functional interaction of TRIC-A with other regulatory components of the intracellular Ca²⁺ release machinery. Our data support the dual function of TRIC-A as a counter-ion channel and an activator of RyR₂, and both functions would enhance RyR₂-mediated Ca²⁺ release in cardiac muscle. A direct interaction between TRIC-A and RyR₂ constitutes an important physiologic component of intracellular Ca²⁺ signaling in the heart. Studies from Fill and colleagues suggested that due to the non-selective nature of the Ca²⁺ release channels, RyR channels can provide a certain extent of counter currents by themselves^{28, 67, 68}. Quantitative simulations suggested that TRIC channels could contribute to the network of SR membrane potential to support Ca²⁺ release and reuptake⁶⁸. Thus, in addition to regulating the acute phase of Ca²⁺ release from the SR/ER store, TRIC-mediated movement of counter current flow could also play a role in balancing the electronegative influence of Ca²⁺ release on other aspects of Ca²⁺ homeostasis inside the SR/ER.

TRIC-A and TRIC-B have different functions in regulating SR and ER Ca²⁺ homeostasis in excitable and non-excitable cells, respectively^{11, 16–21}. TRIC-B is present in both excitable and non-excitable cells, whereas TR-A is predominantly expressed in excitable tissues such as striated muscles and brain tissues. We showed that the divergent CTT domains of TRIC-A and TRIC-B have different functions in modulation of RyR₂ channel activity. Previously we have demonstrated that epithelial cells derived from the TRIC-B knockout mice display abnormal function of IP₃ receptor (IP₃R) mediated Ca²⁺ release from the ER store²¹, and skeletal muscle derived from the TRIC-A knockout mice display abnormal function of RyR₁-mediated Ca²⁺ release from the SR store²⁰. The functional crosstalk between IP₃R and RyR-mediated Ca²⁺ signaling has been implicated in muscle and heart cells under physiologic and pathologic conditions^69–75. Dissecting the function of TRIC-A and TRIC-B in modulating RyR/IP₃R cross-talk for control of Ca²⁺ signaling in health and disease will be an important task of future research.

NOVELTY AND SIGNIFICANCE.

What Is Known?

• Dysfunction of ryanodine receptor (RyR)-mediated Ca²⁺ signaling contributes to arrhythmia and heart failure.

• TRIC channels provide counter-current movement during Ca²⁺ release from intracellular stores.

• Crystal structure of TRIC lacking the carboxyl-terminal tail (CTT) domain has been solved.

What New Information Does This Article Contribute?

• TRIC-A knockout mouse heart shows abnormal response to isoproterenol stimulation due to dysfunctional Ca²⁺ signaling.

• TRIC-A interacts with RyR to directly control Ca²⁺ release in cardiomyocytes.

Nonstandard Abbreviations and Acronyms:

SR

sarcoplasmic reticulum

CICR

Calcium induced Calcium release

SOICR

store overload induced calcium release

ECG

electrocardiogram

RyR

ryanodine receptor

TRIC

trimeric intracellular cation channels

CTT

carboxyl-terminal tail

CPA

cyclopiazonic acid