Chamber-specific regulation of myofilament Ca²⁺ sensitivity by nNOSβ via troponin I phosphorylation: Predominance in the right ventricle

Nitric oxide (NO) is a well-established signaling molecule with a wide range of physiological effects, including a critical role in regulating cardiovascular function. Among the three NO synthase (NOS) isoforms (neuronal NOS [nNOS or NOS1], cytokine-inducible NOS [NOS2], and endothelial NOS [eNOS or NOS3]), nNOS plays a pivotal role in excitation–contraction (E-C) coupling and calcium handling in cardiomyocytes (Shah and MacCarthy, 2000; Sears et al., 2003).

In cardiomyocytes, nNOS is spatially localized near the SR and the plasma membrane. This compartmentalization enables precise modulation of calcium dynamics via posttranslational modification of key enzymes involved in Ca²⁺ release and uptake (Xu et al., 1999; Carnicer et al., 2017). Specifically, nNOS-derived NO modulates ryanodine receptor (RyR2) function; while it can enhance SR Ca²⁺ leak under pathological conditions, it also provides essential physiological regulation for normal Ca²⁺ signaling and contraction-relaxation coupling (Gonzalez et al., 2007; Vielma et al., 2016). Furthermore, nNOS influences cardiomyocyte contractility and relaxation (lusitropy) by affecting L-type calcium channel (LTCCs) current (I_Ca,L) and phospholamban activity (Zhang et al., 2014).

Mechanistically, nNOS-derived NO stimulates cGMP production via stimulating guanylate cyclase, leading to the cGMP-dependent PKG pathways, promoting phosphorylation of phospholamban and subsequent enhancement of SR calcium reuptake via SR/ER Ca²⁺-ATPase (SERCA) 2a (Massion et al., 2003; Carnicer et al., 2013; Zhang et al., 2014). Additionally, NO can directly modify signaling molecules via S-nitrosylation of cysteine residues, altering protein functions responsible for cardiac E-C coupling and contractility (Lima et al., 2010).

Previous studies using both genetic deletion of nNOS (nNOS^−/−) and pharmacological inhibition have demonstrated that loss of nNOS activity leads to prolonged relaxation and a slower rate of cytosolic Ca²⁺ decay due to reduced phospholamban phosphorylation, suggesting that nNOS is a key regulator of cardiac lusitropy (Massion et al., 2003; Carnicer et al., 2013). In addition, nNOS inhibition has been reported to increase the amplitude of I_Ca,L, indicating a negative inotropic effect under physiological conditions (Sears et al., 2003). Importantly, nNOS is also associated with myofilaments, where it exerts compartmentalized signaling to facilitate myofilament Ca²⁺ desensitization and relaxation, a mechanism notably observed in hypertensive rats (Kojda and Kottenberg, 1999; Massion et al., 2003; Sears et al., 2003; Zhang et al., 2014).

Despite these extensive studies, our understanding of nNOS in cardiac physiology and pathophysiology has been largely limited to left ventricular cardiomyocytes (LVCMs), with comparable investigations in right ventricular cardiomyocytes (RVCMs) notably lacking. The right ventricle (RV), characterized by a thinner wall and crescent shape, is adapted to pump blood into the low-resistance pulmonary circulation. Although the RV and left ventricle (LV) eject the same stroke volume, they operate under markedly different mechanical conditions. The normal RV exhibits a triangular pressure–volume (P–V) relationship, with ejection beginning during pressure development and continuing during pressure decline, whereas the LV has more clearly defined isovolumic phases in the P–V relation (P–V loop) of the cardiac cycle (Redington et al., 1988). In addition, RV free-wall contraction has a greater longitudinal or meridional component, in contrast to the more prominent circumferential shortening of the LV lateral wall (Naito et al., 1995). Thus, cardiomyocytes in the two chambers experience distinct temporal and directional patterns of mechanical loading. Recent findings in isolated cardiomyocytes, including our own, suggest that normal RVCMs differ significantly from LVCMs in several aspects, such as action potential duration, E-C coupling efficiency, troponin expression patterns, and intracellular calcium buffering capacity (Kondo et al., 2006; Molina et al., 2016; Jeon et al., 2022).

In cardiac myocytes, intracellular Ca²⁺ mediates E-C coupling by binding to cardiac troponin (cTn) C (cTnC), the Ca²⁺-sensing subunit of the cTn complex. The cTn complex also includes cTnI, which inhibits actin–myosin interaction, and cTnT, which anchors the complex to tropomyosin. Ca²⁺-binding to cTnC relieves cTnI-mediated inhibition, permitting crossbridge formation and contraction. Beyond initiating contraction, cTnC functions as a major cytosolic Ca²⁺ buffer. This myofilament Ca²⁺ buffering critically influences the magnitude and decay kinetics of Ca²⁺ transients, thereby shaping contractile performance. In our previous work, we demonstrated that RVCMs exhibit a lower myofilament Ca²⁺ buffering capacity than LVCMs, primarily due to reduced expression of the cTn complex (Jeon et al., 2022). However, how this buffering capacity is modulated by physiological stimuli remains unclear.

The affinity of cTnC for Ca²⁺ is modulated by posttranslational modifications of cTnI; specifically, cTnI phosphorylation (p-cTnI) decreases myofilament Ca²⁺ sensitivity and facilitates Ca²⁺ dissociation from cTnC (Kooij et al., 2010; Salhi et al., 2014). Myofilament-localized nNOS has been implicated in regulating p-cTnI, positioning it as a key modulator of myofilament Ca²⁺ sensitivity (Kaye et al., 1999; Zhang et al., 2014). While the role of nNOS in myofilament sensitivity is recognized, the quantitative relationship between sensitivity and buffering capacity, which are related but distinct properties, has not been directly established (Eisner et al., 2017). Moreover, the chamber-specific effects of nNOS on functional myofilament Ca²⁺ sensitivity and endogenous Ca²⁺ buffering, potentially through regulation of the troponin complex, remain unexplored. Although we previously identified basal differences in Ca²⁺ buffering between RVCMs and LVCMs (Jeon et al., 2022), the specific regulation of this buffering by nNOS signaling in the RV versus the LV has yet to be elucidated.

In the present study, we investigated the role of nNOS in freshly isolated cardiomyocytes by combining pharmacological inhibition using S-methyl-L-thiocitrulline (SMTC) with simultaneous analysis of sarcomere length (SL) dynamics and E-C coupling. We also compared I_Ca,L via whole-cell patch-clamp recordings and assessed protein phosphorylation by immunoblot analysis. Intriguingly, SMTC treatment revealed a distinct phenotype in RVCMs compared with LVCMs: nNOS inhibition led to a more pronounced increase in contractility in RVCMs, despite a blunted increase in I_Ca,L and intracellular Ca²⁺ transients. This discrepancy suggests that nNOS inhibition primarily affects myofilament sensitivity rather than Ca²⁺ influx in the RVCMs. Consistent with this hypothesis, we observed a significant nNOS-dependent modulation of Ca²⁺ sensitivity and buffering capacity specifically in RVCMs. Furthermore, RVCMs exhibited higher levels of myofilament-associated nNOSβ and a more prominent nNOS-dependent regulation of p-cTnI. The physiological implications of these chamber-specific mechanisms are discussed.

Ventricular cardiomyocytes were isolated from male Sprague–Dawley rats (RRID:RGD_1566457). All experimental procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No.: 85-23, revised 1996) and were approved by the Institutional Animal Care and Use Committee of Seoul National University (IACUC approval No.: SNU-160128-1). A total of 55 rats were used. At the time of experimentation, rats were 12 wk old with a mean body weight of 322.7 ± 8.95 g.

Cardiomyocytes were isolated using a standard enzymatic dissociation protocol as previously described (Jin et al., 2012), with minor modifications. Rats were anesthetized by intraperitoneal injection of ketamine (90 mg·kg⁻¹) and xylazine (10 mg·kg⁻¹). Adequate depth of anesthesia was confirmed by absence of a toe-pinch reflex. Hearts were rapidly excised and mounted on a Langendorff perfusion system maintained at 37°C. Hearts were perfused at a constant flow rate (8–10 ml·min⁻¹) according to the following sequence: (1) 8 min with Ca²⁺-free Tyrode solution; (2) 8 min with Ca²⁺-free Tyrode solution containing collagenase (1 mg·ml⁻¹; Worthington Biochemical Co.), protease (0.133 mg·ml⁻¹), BSA (1.65 mg·ml⁻¹), and Ca²⁺ (0.05 mM); and (3) an additional 10 min digestion with collagenase-containing Tyrode solution. The interventricular septum was excluded from analysis. To ensure the homogeneity of the RVCMs and LVCMs, only tissue samples from the free walls of the RVs and LVs were collected, excluding the interventricular septum. Isolated cardiomyocytes were resuspended either in cardioplegic storage solution for contractility measurements or in Kraft–Brühe (K-B) solution for electrophysiological experiments (Table 1). The yield of rod-shaped, calcium-tolerant cardiomyocytes ranged from 60% to 80%. Cardiomyocytes designated for physiological experiments were used immediately after isolation, whereas those intended for biochemical analyses were centrifuged and stored at −80°C until further use. For SMTC treatment, cardiomyocytes were preincubated with 100 nM SMTC for 20 min prior to the start of experiments, and the same concentration of SMTC was maintained in all perfusion solutions throughout data acquisition.

NO production was indirectly assessed by quantifying nitrite levels in cardiomyocytes using a colorimetric NO assay kit based on the Griess Reagent System (Promega). After homogenization, lysates were centrifuged to collect the supernatant, and total protein concentration was determined using the Bradford assay (Bio-Rad). For nitrite measurement, 50 μl of each supernatant was transferred to a 96-well microplate and mixed with 50 μl of sulfanilamide solution, followed by incubation for 5–10 min at room temperature in the dark. Subsequently, 50 μl of N-(1-naphthyl)ethylenediamine dihydrochloride (Sigma-Aldrich) was added to each well, and the reaction was allowed to proceed under the same conditions. The resulting azo dye, proportional to nitrite concentration, was quantified by measuring absorbance at 540 nm using a microplate spectrophotometer (PowerWave XS; BioTek Instruments). A standard curve was generated using serial dilutions of sodium nitrite (Sigma-Aldrich), and sample nitrite concentrations were calculated by interpolation from this curve.

Fractionation of cardiomyocytes for immunoblotting was performed using two distinct protocols, as previously described (Jang et al., 2015). This fractionation procedure was previously shown to yield a myofilament-enriched fraction with no detectable GAPDH contamination, as demonstrated in Fig. 5 A of Jang et al. (2015). To obtain membrane and cytosolic fractions, cells were lysed in buffer containing 25 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.5 mM EGTA, and 1% Triton X-100 supplemented with protease and phosphatase inhibitor cocktails (Roche). The lysates were centrifuged at 15,000 rpm for 30 min, and the supernatant containing soluble membrane and cytosolic proteins was collected for further analysis. For isolation of the myofilament fraction, cardiomyocytes were rinsed with ice-cold phosphate-buffered saline and transferred to screw-cap microtubes containing 2.5-mm zirconia/silica beads and cold extraction buffer (60 mM KCl, 30 mM imidazole, 2 mM MgCl₂, and 1% Triton X-100, supplemented with protease and phosphatase inhibitors). Cells were homogenized using a Mini-Beadbeater-8 (BioSpec Products) for 20 s, followed by centrifugation at 8,000 rpm for 10 min at 4°C. The supernatant was discarded, and the homogenization-centrifugation cycle was repeated at least six times to ensure efficient removal of soluble proteins. The remaining pellet, enriched in myofilament proteins, was resuspended in high-salt buffer (Bio-Rad) and used for subsequent biochemical analyses.

Protein concentrations were determined using the Bradford assay (Bio-Rad). Equal amounts of protein were denatured at 95°C for 5 min in sample buffer, separated by SDS-PAGE, and transferred to polyvinylidene difluoride membranes (Bio-Rad) using transfer buffer containing 25 mM Tris, 192 mM glycine, 0.01% SDS, and 20% methanol. Membranes were blocked for 1 h at room temperature in TBS with 0.1% Tween-20 containing 5% BSA. Membranes were incubated overnight at 4°C with primary antibodies against nNOS (RRID:AB_304965, Abcam), phospho-nNOS (RRID:AB_304964, Abcam), eNOS (RRID:AB_397690, BD Biosciences), phospho-eNOS^Thr495 (RRID:AB_2153176, Cell Signaling Technology), phospho-eNOS^Ser1177 (RRID:AB_823493, Cell Signaling Technology), TnC (RRID:AB_306435, Abcam), TnI (RRID:AB_2206278, Cell Signaling), phospho-TnI (RRID:AB_2206275, Cell Signaling Technology), TnT (RRID:AB_261723, Sigma-Aldrich), and α-actin (RRID:AB_476695, Sigma-Aldrich). After washing, membranes were incubated with appropriate HRP-conjugated secondary antibodies. Immunoreactive bands were detected using enhanced chemiluminescence (ECL Plus; Merck) and quantified by densitometry. Protein levels were normalized to α-actin. Each sample represented an independent biological replicate from a different animal. Samples were excluded if α-actin levels varied by >10% between left and right ventricular lysates from the same animal. The phosphorylation site is referred to as Ser¹⁴¹⁷ according to the antibody nomenclature; the homologous site is commonly designated Ser¹⁴¹² in rodent nNOS literature.

The physiological properties of isolated cardiomyocytes were assessed using the IonOptix system (IonOptix). Contractile function was evaluated by measuring SL shortening in response to electrical field stimulation (EFS) at a frequency of 2 Hz. The SL was calculated in real time using the SarcLen module, which analyzes periodic striation patterns within the cell via fast Fourier transform of the optical density profile across the cell. Freshly isolated cardiomyocytes were transferred into a bath mounted on the stage of an inverted microscope (Diaphot 300; Nikon). The bath (0.15 ml) was continuously perfused with normal Tyrode solution (NT, Table 1) at a constant rate (5 ml·min⁻¹). All experiments were performed at 37 ± 1°C. For ratiometric fluorescence measurements using fura-2, cardiomyocytes were incubated with 2 μM fura-2 acetoxymethyl ester for 15 min at room temperature. Following gravity sedimentation, the supernatant was carefully removed, and the cells were resuspended in a perfusion buffer containing 500 μM Ca²⁺ for 10 min to allow for intracellular de-esterification of the dye. The fura-2–loaded cardiomyocytes were then transferred to a temperature-controlled imaging chamber (37 ± 1°C) and alternately illuminated with excitation wavelengths of 360 and 380 nm using a dual-wavelength excitation system. The corresponding emission signals were detected in real time using photomultiplier tubes (PTIs). The ratio of fluorescence intensities (F₃₆₀/F₃₈₀) was continuously monitored as a surrogate index of [Ca²⁺]_c dynamics. For each condition, F₃₆₀/F₃₈₀ was averaged over a minimum of 20 steady-state contraction cycles per cardiomyocyte.

Freshly isolated cardiomyocytes were transferred to a recording chamber mounted on the stage of an inverted microscope (Ti, Nikon). The chamber (volume: 0.15 ml) was continuously perfused at a constant rate of 5 ml·min⁻¹, and voltage-clamp experiments were performed at room temperature (22–25°C). Conventional and Nystatin-perforated whole-cell patch-clamp recordings were obtained using borosilicate glass pipettes with a tip resistance of ∼2.5 MΩ. Series resistance, estimated from the time constants of the capacitive transients, was kept below 10 MΩ in the whole-cell configuration (calcium imaging) and 16 MΩ in the perforated whole-cell configuration. LTCC current was recorded using the nystatin-perforated whole-cell patch-clamp technique with an I_Ca,L pipette solution (Table 1). To normalize for cell size, current amplitudes were divided by cell capacitance and expressed as pA·pF⁻¹. Patch pipettes were connected to the CV-203BU headstage of an Axopatch 200B patch-clamp amplifier (Molecular Devices). Data acquisition and command pulse generation were performed using pCLAMP software (version 10.6.2) in conjunction with a Digidata 1440A interface (Molecular Devices). Currents were sampled at 10 kHz and low-pass filtered at 5 kHz using a Bessel filter. Offline data analysis was performed using Clampfit software (version 11.0.1; Molecular Devices).

As in our previous research (Jeon et al., 2022), Ca²⁺ buffering capacity was assessed using simultaneous whole-cell patch-clamp recording and fura-2 ratiometric fluorescence imaging. Briefly, cardiomyocytes were dialyzed with a Cs⁺-based pipette solution (Cs-pipette; Table 1) supplemented with 50 µM fura-2 pentapotassium salt, and fluorescence signals were monitored using alternating 340/380 nm excitation with emission collected at 510 nm via a PTI.

Accurate calibration of intracellular Ca²⁺ concentration in cardiomyocytes is technically challenging because cells undergo hypercontraction at $[Ca2+]c$ above ∼1 µM, which alters optical path length and can compromise membrane integrity (Cheung et al., 1989). To circumvent these limitations, we determined the calibration parameters (R_min, R_max, and β) in separate experiments using the lanthanum (La³⁺) method, which allows estimation of calibration constants without exposing intact cells to high Ca²⁺ levels (Borzak et al., 1990), as described in our previous work (Jeon et al., 2022). Fluorescence conversion factors were obtained using a cuvette-based spectrophotometer with an identical excitation source and were calculated using the same values as in our previous study.

To estimate the effective dissociation constant of fura-2 in the cytosolic environment, we performed an in vitro calibration following Berlin et al. (1994) and our previous work. Among standard, viscosity-corrected (sucrose), and protein-containing calibration solutions, the protein-containing condition most closely reflected the intracellular environment. Accordingly, and in line with our previous work (Jeon et al., 2022), we adopted the previously reported calculated value of $Kd,fura−2$ = 381.8 nM for subsequent analyses.

Total cellular volume (⁠$voltc$⁠) was estimated from membrane capacitance (⁠$Cm$⁠), assuming a $Cm/voltc$ ratio of 6.76 pF·pl⁻¹ for 12-wk-old rats (Satoh et al., 1996). Because mitochondria and other organelles occupy ∼35% of cardiomyocyte volume, the fura-2–accessible cytosolic volume (⁠$volac$⁠) was taken as 0.65·$voltc$ (Berlin et al., 1994). To estimate the effective cytosolic concentration of fura-2 ([fura-2]_tot), we measured fluorescence at the isosbestic wavelength at the end of each experiment. Cellular fluorescence (⁠$Fcell$⁠) was normalized to $volac$ and compared with a standard fluorescence (⁠$Fstandard$⁠) obtained from a glass capillary (20-µm inner diameter) filled with 50 µM fura-2 positioned in the same region of interest. [fura-2]_tot was calculated as described in our previous work (Jeon et al., 2022).

Solutions for electrophysiology, Ca²⁺ measurement, and the IonOptix system are listed in Table 1. We used four external solutions: K-B, NT, 0Na/0Ca, and 0Na/2Ca. Three types of pipette solutions were used depending on the experimental conditions (Cs-pipette and I_Ca,L-pipette). Unless otherwise specified, all drugs and chemicals were purchased from Sigma-Aldrich.

Animals were randomly assigned to experimental groups where applicable. Blinding was not feasible for experiments requiring drug treatment during data acquisition, as treatment conditions were evident to the experimenter. In contrast, immunoblotting and NO measurements were performed with investigators blinded to group identity. Data processing and quantitative analyses were also conducted in a blinded manner whenever feasible.

Data are presented as mean ± standard deviation (SD), with n indicating the number of cells and N indicating the number of animals. Statistical comparisons were performed using two-tailed t tests or two-tailed nested t tests, where appropriate, to account for the hierarchical structure of the data, with cells nested within animals (n cells within N animals) (Sikkel et al., 2017; Eisner, 2021). Differences in I_Ca,L current amplitudes were evaluated using multivariate ANOVA (MANOVA). To directly compare treatment effects between chambers, cell-level measurements were averaged within each animal and condition, and data were analyzed using two-way repeated-measures ANOVA with chamber and treatment. Differences in treatment responses between chambers were determined from the chamber-by-treatment interaction. Statistical significance was accepted at P < 0.05. Exact P values are reported for statistically significant comparisons in the figure or figure legend, except for values below 0.0001, which are reported as P < 0.0001. Curve fitting was performed using a least-squares method in Origin (Microcal Software) or custom-written R scripts.

Fig. S1 presents immunohistochemical analyses showing the expression levels and relative ratios of eNOS and phospho-eNOS in the cytosolic fraction, as well as the expression patterns of nNOS in RVCMs and LVCMs. Fig. S2 shows the sarcomere shortening analysis for each individual animal, and Fig. S3 shows the Ca²⁺ transient analysis for each individual animal. The box-and-whisker plots shown in Figs. S2 and S3 display the five-number summary of the data distribution, including the minimum, first quartile, median, third quartile, and maximum. Fig. S4 provides a representative example of the experimental traces used for buffer analysis, illustrating how the data presented in Fig. 6 were obtained.

Immunoblot analysis of the membrane/cytosol fraction revealed two distinct immunoreactive bands corresponding to nNOS isoforms: a lower band below 140 kDa representing the β-variant (nNOSβ) and a higher band above 165 kDa corresponding to α/μ isoforms (nNOSα/μ), consistent with previously reported isoform-specific migration patterns (Jang et al., 2015). In this study, we quantified cytosolic nNOS using the band corresponding to the nNOSα form, whereas nNOS in the myofilament fraction was assessed using the nNOSβ band (Jang et al., 2015).

Quantitative densitometric analysis revealed no significant difference in total nNOS protein expression between LVCMs and RVCMs (Fig. 1, A and B). The subcellular distribution of nNOS was also comparable between LVCMs and RVCMs, exhibiting a diffuse cytosolic pattern with relatively dense nuclear signal in both populations (Fig. S1 A). We further assessed the phosphorylation status of nNOS at serine 1417 (p-Ser¹⁴¹⁷), a well-characterized regulatory site known to enhance electron transfer from the reductase to the oxygenase domain and thereby increase catalytic efficiency (Andrabi et al., 2023) (Fig. 1, C and D). Neither p-Ser¹⁴¹⁷ levels nor the ratio of phosphorylated nNOS (p-nNOS) to total nNOS differed significantly between the two groups.

While nNOS plays a more direct role in regulating contractile function, eNOS is also known to participate in cardiomyocyte NO signaling. Therefore, we examined eNOS levels to distinguish the contributions of these isoforms (Tang et al., 2014). Accordingly, we analyzed both total eNOS and its phosphorylated forms (Fig. S1, B–E). In the RVCM cytosolic fraction, total eNOS expression was modestly lower, whereas phosphorylation at Ser¹¹⁷⁷, one of the two major activating sites, was slightly higher than that of LVCMs (Tang et al., 2014). This reciprocal change suggests that the elevated phosphorylation likely compensates for the lower protein abundance, maintaining overall eNOS activity at levels comparable with LVCMs - a conclusion further supported by our pharmacological inhibition experiments (Fig. 1 E).

To directly quantify and compare intracellular NO levels between LVCMs and RVCMs, we performed the Griess assay (Fig. 1 E). Under basal conditions, NO production was comparable between the two groups. To dissect the relative contributions of NOS isoforms, we employed pharmacological inhibitors: the nonselective NOS inhibitor N-omega-Nitro-L-arginine methyl ester hydrochloride (L-NAME, 100 μM), and the nNOS-selective inhibitor SMTC (100 nM). Treatment with L-NAME markedly reduced NO levels to a similar extent in both LVCMs and RVCMs. Similarly, SMTC caused a significant and comparable reduction in NO production in both populations. These findings suggest that global nNOS-derived NO generation is functionally similar between LVCMs and RVCMs. Furthermore, the observation that nNOS inhibition largely recapitulated the effects of nonselective blockade implies that eNOS contributes minimally to the total cytosolic NO pool, likely exerting its effects in a compartmentalized manner. Consequently, SMTC was selected as the specific nNOS inhibitor for subsequent experiments.

Using the IonOptix system (1 kHz sampling), we assessed SL dynamics in cardiomyocytes paced at 2 Hz via EFS. Fig. 2, A and B, display the averaged SL traces synchronized to EFS and their normalized counterparts, respectively. We analyzed the extent of sarcomere shortening (ΔSL; diastolic SL minus minimal SL) and post-contraction relaxation kinetics (Fig. 2, C–H; individual-animal data are shown in Fig. S2).

While SMTC treatment significantly increased ΔSL in both LVCMs and RVCMs (Fig. 2, D and E), the SMTC-induced increase in mean ΔSL was significantly greater in RVCMs (0.030 µm) than in LVCMs (0.015 µm; Fig. 2 D). To directly compare the magnitude of the SMTC response between chambers, the resulting animal-level means were analyzed using two-way repeated-measures ANOVA with chamber and treatment. The significant chamber-by-treatment interaction (F_1,9 = 24.13, P = 0.0008) demonstrated that the SMTC-induced increase in ΔSL was greater in RVCMs. Interestingly, nNOS inhibition significantly reduced diastolic SL in RVCMs, whereas this reduction was not statistically significant in LVCMs (Fig. 2, A and C). After SMTC treatment, RVCMs exhibited a significantly shorter diastolic SL (1.74 ± 0.035 µm; Fig. 2 C), while the corresponding value in LVCMs (1.76 ± 0.021 µm) did not differ significantly. A significant chamber-by-treatment interaction was also observed for diastolic SL (F_1,9 = 6.15, P = 0.0349), indicating a greater reduction in RVCMs. As a result, the increase in fractional shortening (expressed as percentage of diastolic SL) was markedly higher in RVCMs than in LVCMs, indicating a preferential inotropic response in the RV (Fig. 2 E).

Given the established lusitropic role of nNOS in cardiomyocytes (Zhang et al., 2014), inhibition of nNOS was expected to slow the relaxation following EFS-induced contraction. To compare the relaxation kinetics, SL traces were normalized (scale 0–1). These normalized averaged traces clearly demonstrated a markedly greater slowing of relaxation in SMTC-treated RVCMs compared with LVCMs (Fig. 2 B). We quantified the time to 10%, 50%, and 90% relaxation, reflecting the early, mid, and late phases, respectively (Fig. 2, F–H). SMTC treatment prolonged 90% relaxation time in both cell types, while the time to 10% relaxation remained unaltered. The lack of effect on the early phase likely reflects its dependence on cytosolic Ca²⁺ decay, whereas mid-to-late phases are more influenced by myofilament crossbridge kinetics (Biesiadecki et al., 2014). Consistent with this, SMTC treatment revealed a more robust lusitropic effect of nNOS in the mid-to-late phases in RVCMs. Although the time to 90% relaxation was increased in both groups, RVCMs exhibited a markedly larger delay (mean increase: 9.27 ms) than LVCMs (mean increase: 3.28 ms; Fig. 2 H). Although the numerical increase was greater in RVCMs than in LVCMs, the chamber-by-treatment interaction was not statistically significant (F_1,9 = 2.80, P = 0.1284). These findings indicate that nNOS inhibition impairs relaxation in both ventricles.

To explore the mechanism underlying the chamber-specific contraction–relaxation responses, we examined intracellular Ca²⁺ transients in LVCMs and RVCMs paced at 2 Hz, using fura-2 fluorescence ratiometry (F₃₆₀/F₃₈₀) with an IonOptix system at a sampling frequency of 250 Hz (Fig. 3, A and B; individual-animal data are shown in Fig. S3). Basal diastolic Ca²⁺ levels (resting F₃₆₀/F₃₈₀) were comparable between LVCMs and RVCMs and remained unaltered by SMTC treatment (Fig. 3 C). However, nNOS inhibition significantly increased both the peak fluorescence ratio (peak F₃₆₀/F₃₈₀) and the amplitude of Ca²⁺ transients (ΔF₃₆₀/F₃₈₀) in LVCMs (F_1,9 = 6.48, P = 0.0314). In contrast, these effects were modest and did not reach statistical significance in RVCMs (Fig. 3, D and E). Notably, this pattern was strikingly discordant with the SL data, where nNOS inhibition elicited a much stronger contractile enhancement in RVCMs.

A similar dissociation was evident in the decay kinetics of Ca²⁺ transients. We quantified the decay times to 10%, 50%, and 90% recovery (Fig. 3, F–H). While SMTC tended to prolong Ca²⁺ decay in both cell types, the effect was statistically significant only in LVCMs during the mid-phase (time to 50% decay), whereas RVCMs showed no significant prolongation (Fig. 3 G). Considering that the mid-phase of Ca²⁺ decay temporally overlaps with the late phase of relaxation, the observation that SMTC markedly slowed mechanical relaxation in RVCMs without a parallel slowing of Ca²⁺ decay supports the existence of chamber-specific differences in myofilament regulatory properties.

Previous studies reported that nNOS-derived NO attenuates LTCCs through S-nitrosylation- or cGMP-dependent mechanisms (Sears et al., 2003). To determine if the differential Ca²⁺ transients resulted from altered Ca²⁺ entry, we recorded L-type Ca²⁺ currents (I_Ca,L) using the nystatin-perforated patch-clamp technique. Step-like depolarizing voltage commands were applied from a holding potential of −40 mV. Voltage-dependent activation of inward I_Ca,L was observed (Fig. 4, A and C). Peak current amplitudes, normalized to membrane capacitance (C_m), were plotted against test potentials to generate (I-V relationships, which did not differ between LVCMs and RVCMs under basal conditions (Fig. 4, B and D). Upon nNOS inhibition, I_Ca,L density increased in both populations; however, the magnitude of augmentation was substantially greater in LVCMs (ΔI_Ca,L = 3.4 ± 0.39 pA/pF @ 0 mV) compared with RVCMs (ΔI_Ca,L = 1.6 ± 0.49 pA/pF @ 0 mV, Fig. 4 C). Biophysical properties, including steady-state inactivation, remained unchanged (data not shown). Collectively, these findings indicate that while nNOS inhibition enhances I_Ca,L-mediated Ca²⁺ influx in both chambers, the effect is blunted in RVCMs. This result provides an electrophysiological basis for the smaller augmentation of Ca²⁺ transient amplitude observed in RVCMs following nNOS inhibition, further reinforcing the paradox of their exaggerated contractile response to SMTC treatment. This mismatch between Ca²⁺ signaling and mechanical output strongly suggests an additional regulatory mechanism downstream of Ca²⁺ influx in the RVCMs.

To compare the relative magnitudes of the electrical, Ca²⁺, and mechanical responses, we calculated the percentage change relative to the corresponding control value in each chamber. In LVCMs, SMTC treatment increased peak I_Ca,L at 0 mV, Ca²⁺ transient peak amplitude, and sarcomere shortening by 25.11%, 8.31%, and 13.36%, respectively. In RVCMs, the corresponding increases were 12.30%, 4.55%, and 29.27%, respectively. Thus, although the increases in Ca²⁺ influx and Ca²⁺ transient were greater in LVCMs, the increase in sarcomere shortening was markedly greater in RVCMs. This relative mismatch between the Ca²⁺-handling and mechanical responses indicates that the enhanced contractility in RVCMs cannot be explained solely by changes in Ca²⁺ influx, suggesting an involvement of downstream myofilament regulation.

To elucidate the mechanism underlying the dissociation between the exaggerated sarcomere shortening and the blunted Ca²⁺ transients in SMTC-treated RVCMs, we performed simultaneous measurements of SL and calcium signal (F₃₆₀/F₃₈₀) (Fig. 5). Plotting SL against F₃₆₀/F₃₈₀ over a single beat revealed characteristic hysteresis loops. In these phase-plane plots, the upper right portion of the loop corresponds to the rapid rise in intracellular Ca²⁺ during contraction, whereas the lower left portion represents the relaxation phase. The rapid rising phase is prone to distortion because force development lags behind the surge in Ca²⁺, rendering conventional steady-state measures (e.g., force-pCa) insufficient for describing dynamic behavior (Yagi et al., 1988; Chung et al., 2016). In contrast, the relaxation phase represented by the descending limb of the loop reflects a slower, quasi-equilibrium state where crossbridge detachment and myofilament properties dominate. Therefore, the slope and the horizontal position of the relaxation trajectory of the Ca²⁺–SL hysteresis loop provide a more reliable index of effective myofilament Ca²⁺ sensitivity. Accordingly, we focused our analysis on this relaxation phase.

In LVCMs, the contractile enhancement by SMTC was accompanied by a clear elevation in intracellular Ca²⁺ levels. While a modest leftward shift of the loop (indicating increased sensitivity) was observed, it did not reach statistical significance during the relaxation phase. This suggests that in LVCMs, nNOS inhibition augments contraction primarily by increasing Ca²⁺ availability rather than by altering myofilament Ca²⁺ sensitivity. Conversely, in RVCMs, nNOS inhibition resulted in a pronounced leftward shift of the relaxation trajectory (Fig. 5 B). Notably, RVCMs exhibited substantially greater shortening at equivalent Ca²⁺ levels compared with controls, and a marked reduction in SL was evident even at basal Ca²⁺ levels. Collectively, these data indicate that the augmented contraction in RVCMs is primarily attributable to a significant increase in myofilament Ca²⁺ sensitivity, with only a modest contribution from elevated intracellular Ca²⁺ levels. In contrast, in LVCMs, the enhanced contraction is driven mainly by increased intracellular Ca²⁺ concentration, with a comparatively limited contribution from changes in myofilament Ca²⁺ sensitivity.

An increase in functional myofilament Ca²⁺ sensitivity may be accompanied by altered Ca²⁺-binding within the myofilament compartment, including changes involving cTnC, corresponding to a reduced dissociation constant (⁠$Kd$⁠) and, consequently, an enhanced myofilament Ca²⁺ buffering capacity as suggested by Eisner and colleagues (Eisner, 2014; Eisner et al., 2017; Eisner et al., 2023). While Ca²⁺ sensitivity cannot be attributed solely to cTnC-Ca²⁺ binding, such augmented myofilament buffering is predicted to modulate Ca²⁺ transient kinetics and exert a disproportionate influence on mechanical relaxation, even in the absence of marked changes in total intracellular Ca²⁺ levels. Because other Ca²⁺ buffers such as calmodulin and ATP contribute minimally within the physiological range of [Ca²⁺]_c (Smith and Eisner, 2019), we quantified the endogenous cytosolic Ca²⁺ buffering coefficient (⁠$κS$⁠) in nNOS-inhibited RVCMs and LVCMs after functionally eliminating Ca²⁺ fluxes mediated by Ca²⁺ release from SR as well as the removal by SERCA and NCX, using caffeine (20 mM), thapsigargin (2.5 µM), and a Na⁺-free bath solution, respectively (Eq. 1, Fig. 6). A detailed schematic of the experimental protocol is provided in Fig. S4.

To quantify myofilament Ca²⁺ buffering capacity and kinetics, we applied depolarizing pulse protocols in a Na⁺-free bath solution with caffeine and thapsigargin to measure cumulative Ca²⁺ entry via voltage-operated LTCCs. Changes in intracellular Ca²⁺ concentration and buffering parameters (⁠$κB$ and $[B]tot$⁠, Eq. 1) were calculated using the added-buffer approach (Neher, 1995), as detailed in our previous study (Jeon et al., 2022). To estimate the endogenous Ca²⁺ buffer pool (⁠$[S]tot$⁠), the contribution of the exogenous indicator fura-2 ([fura-2]_tot) was subtracted (Fig. 6 C). Since the dissociation constant of fura-2 (K_d,fura-2) was determined via in vitro calibration and the effective intracellular concentration of fura-2 was established in our previous work, the relationship between the Ca²⁺-binding ratio of fura-2 (κ_fura-2) and the median [Ca²⁺]_c was fitted using the differential form of the Michaelis–Menten equation (triangular symbols in Fig. 6 A). Subsequently, the endogenous buffering coefficient (⁠$κS$⁠) was derived by subtracting κ_fura-2 from the total buffering coefficient (⁠$κB$⁠).

Fitting $κS$ as a function of [Ca²⁺] to Eq. 1 yielded two parameters describing the apparent properties of the endogenous Ca²⁺ buffer system: the total concentration of the endogenous buffer (⁠$[S]tot$⁠) and its apparent dissociation constant (⁠$Kd,S$⁠). Consistent with our previous findings demonstrating a lower myofilament content in RVCMs (Jeon et al., 2022), $[S]tot$ was significantly lower in RVCMs than in LVCMs, and this parameter (⁠$[S]tot$⁠) was unaffected by SMTC treatment in either group. Intracellular fura-2 concentrations were comparable across all four groups, indicating consistent fura-2 loading (Fig. 6 D).

Although $Kd,S$ represents the combined apparent behavior of endogenous intracellular buffers and could not be interpreted as a direct measurement of cTnC affinity, the calculated apparent dissociation constant for the endogenous buffer (⁠$Kd,S$⁠) values (LVCMs: 0.59 ± 0.049 µM; RVCMs: 0.68 ± 0.059 µM, Fig. 6 E) were similar to the reported dissociation constant of troponin C (⁠$Kd,cTnC$ <0.6 µM; Smith and Eisner [2019]). Consistent with the increase in myofilament Ca²⁺ sensitivity observed in the loop analysis, nNOS inhibition reduced $Kd,S$ in both ventricular cardiomyocyte types. Notably, the magnitude of this reduction was substantially greater in RVCMs (⁠$ΔKd,S$ = −0.265 µM) than in LVCMs (⁠$ΔKd,S$ = -0.077 µM). This indicates that nNOS inhibition produces a larger change in the apparent Ca²⁺ affinity of the endogenous buffering system in RVCMs, which may partly reflect altered Ca²⁺ binding within the cTn complex (Fig. 6 E).

Using these parameters (⁠$κB$⁠, $κS$⁠, κ_fura-2, $Kd,B$⁠, $Kd,S$⁠, $[B]tot$⁠, [fura-s]_tot, and $[S]tot$⁠), we calculated the amount of Ca²⁺ bound to intracellular buffers as a function of cytosolic Ca²⁺ concentrations ([Ca²⁺]_c) using the Michaelis–Menten equation (Fig. 6 B). In the Ca²⁺-bound Ca²⁺ relationship, solid lines represent untreated cardiomyocytes, whereas dashed lines denote SMTC-treated cells. Considering that the changes in the dissociation constant of intrinsic buffer (⁠$Kd,S$⁠) occur within the physiological range of [Ca²⁺]_c (<1.0 µM), the impact of nNOS inhibition on Ca²⁺ binding became more apparent within this physiological Ca²⁺ range, rather than at supraphysiological levels. Thus, nNOS inhibition produced a clearly distinguishable upward shift in bound Ca²⁺ within the physiological range in both cell types, with the effect being markedly more prominent in RVCMs.

Since the Ca²⁺–SL loop analysis and buffering quantification pointed toward a preferential increase in myofilament Ca²⁺ sensitivity and endogenous Ca²⁺-binding properties in RVCMs, we next investigated the molecular mechanism by quantifying the phosphorylation status of cardiac cTnI, a canonical regulator of myofilament Ca²⁺ sensitivity (Fig. 7). It has been reported that p-cTnI mediated by PKA or nNOS/PKG signaling could decrease myofilament Ca²⁺ sensitivity (Zhang et al., 1995; Kooij et al., 2010; Jin et al., 2012).

In LVCMs, SMTC treatment resulted in a trend toward decreased p-cTnI, but this did not reach statistical significance. In contrast, nNOS inhibition in RVCMs elicited a marked and statistically significant reduction in p-cTnI levels (Fig. 7 C). As noted in our previous study (Jeon et al., 2022), basal total cTnI expression is lower in RVCMs than in LVCMs (Fig. 7 B). Therefore, to better assess relative differences in phosphorylation, we compared the normalized ratio of p-cTnI with total cTnI (Fig. 7 D). While SMTC significantly reduced the p-cTnI/total cTnI ratio in both groups, the magnitude of this reduction was substantially greater in RVCMs. These molecular data align closely with the functional data showing a more robust increase in Ca²⁺ sensitivity in the RV. Collectively, these findings provide a molecular explanation for the chamber-specific phenotype: nNOS inhibition preferentially reduces p-cTnI in RVCMs, providing a potential molecular basis for the greater increase in functional myofilament Ca²⁺ sensitivity and the associated change in apparent endogenous buffering properties.

Cardiac nNOS is known to be spatially confined to distinct subcellular compartments, including the cytosol, sarcolemma, SR, and myofilaments (Jang et al., 2015; Carnicer et al., 2017). In this study, we initially confirmed that global cellular levels of total and p-nNOS did not differ between LVCMs and RVCMs (Fig. 1). However, whole-cell lysates can mask spatial differences between functional microdomains such as myofilaments. To address this, we isolated the myofilament fraction to specifically examine the pool of nNOS (Fig. 8). While total nNOS protein levels were comparable between groups, the level of Ser¹⁴¹⁷ p-nNOS was markedly higher in the myofilament fraction of RVCMs, with a consequently higher p-nNOS/nNOS ratio (Fig. 8, C and D). These findings reveal a larger pool of active p-nNOS enriched within the myofilaments of RVCMs, which may help explain the greater effect of nNOS inhibition on RVCM contractility.

The present results indicate that nNOS signaling exerts negative inotropic and positive lusitropic effects in both LVCMs and RVCMs. However, these regulatory influences are substantially more pronounced in RVCMs, suggesting a chamber-specific functional specialization of nNOS in contraction–relaxation dynamics. Our data suggest that the myofilament Ca²⁺ sensitivity is critically regulated by the myofilament-associated pSer¹⁴¹⁷ nNOSβ isoform via Ser^23/24 phosphorylation of cTnI in RVCMs.

NO signaling in cardiomyocytes is increasingly recognized as a highly compartmentalized process rather than a diffusely acting pathway. nNOS localizes to distinct subcellular microdomains, including the SR and sarcolemma, where it spatially regulates targets involved in E-C coupling and β-adrenergic signaling (Carnicer et al., 2017). Importantly, previous studies have also demonstrated the presence of an nNOS pool within the myofilament fraction, suggesting that sarcomere proteins are subject to regulation by locally generated NO (Jang et al., 2015). Such spatial restriction provides a mechanistic framework whereby nNOS exerts selective functional effects without necessarily altering whole-cell NO levels, as downstream signaling is shaped by microdomain-specific protein interactions and local redox environments.

In the present study, although total cellular nNOS expression was comparable between LVCMs and RVCMs, we identified significantly higher levels of phosphorylated nNOSβ at Ser¹⁴¹⁷ within the myofilament-enriched fraction of RVCMs. Since phosphorylation at Ser¹⁴¹⁷ is a well-established marker of enhanced nNOS enzymatic activity (Kar et al., 2015; Andrabi et al., 2023), this finding indicates that nNOSβ activity diverges between the two chambers at the posttranslational level and is preferentially compartmentalized within the myofilament region of the RV.

Phosphorylation of the stimulatory C-terminal site of nNOS is regulated by several upstream pathways. In the intact heart, oxidative stress, mechanical cues, and β-adrenergic receptor signaling can influence nNOS phosphorylation through AMP-activated protein kinase- or Akt-dependent processes (Watts et al., 2013; Garbincius and Michele, 2015; Kar et al., 2015). Because the RV and LV experience distinct loading patterns and exhibit differences in adrenergic signaling, these inputs may contribute in vivo to the establishment of chamber-specific nNOSβ regulation (Redington et al., 1988; Naito et al., 1995). However, the present experiments were performed in freshly isolated, unloaded cardiomyocytes under common experimental conditions, and cells used for biochemical analyses were not maintained in culture medium. Thus, acute mechanical load or ligand-dependent receptor signaling cannot by itself account for the difference in myofilament-associated p-nNOSβ at the time of measurement. Instead, the observed difference may reflect a molecular state established in vivo and retained after isolation, such as a persistent local kinase–phosphatase balance, stable nNOSβ microdomain or scaffold interactions, or a pre-existing posttranslational state. The present study does not distinguish among these possibilities or establish whether the retained phenotype is developmentally programmed or a lasting consequence of the in vivo mechanical and neurohumoral environment.

Functionally, the enrichment of activated nNOSβ in the myofilament fraction offers a potential mechanistic explanation for the greater sensitivity of RVCMs to nNOS inhibition. Pharmacological inhibition of nNOS resulted in a more pronounced reduction in p-cTnI at Ser^23/24 and a larger shift in the Ca²⁺-contraction relationship in RVCMs compared with LVCMs (Figs. 5 and 7). p-cTnI at Ser^23/24 is a central determinant of myofilament Ca²⁺ sensitivity, classically mediated by PKA but also modulated by NO-dependent signaling pathways (Zhang et al., 1995; Kooij et al., 2010; Jin et al., 2012). Physiologically, the higher myofilament-associated nNOSβ in RVCMs likely promotes myofilament Ca²⁺ desensitization via local regulation of p-cTnI. This microdomain-specific organization provides a molecular basis for the chamber-dependent regulation by nNOS activity in ventricular contractile behavior.

In addition to the chamber-dependent differences in myofilament-associated nNOS signaling, the response of LTCCs to nNOS inhibition also differed markedly between ventricular myocytes. In ventricular cardiomyocytes, nNOS has been shown to negatively regulate I_Ca,L through multiple mechanisms, including activation of the cGMP/PKG pathway and direct S-nitrosylation of LTCCs, thereby limiting Ca²⁺ influx under basal conditions (Sumii and Sperelakis, 1995; Sears et al., 2003). Consistent with this established inhibitory role of nNOS, pharmacological inhibition with SMTC resulted in a robust increase in I_Ca,L in LVCMs. In contrast, RVCMs exhibited only a modest and statistically nonsignificant increase in I_Ca,L following SMTC treatment (Fig. 4), despite comparable total nNOS expression and similar NO suppression. This suggests that nNOS-dependent regulation of LTCCs is intrinsically less prominent in RVCMs.

LTCC function is governed by an integrated network involving PKA, CaMKII, and phosphodiesterase within membrane-associated macromolecular complexes (Kamp and Hell, 2000; Bers and Grandi, 2009; Leroy et al., 2011). A previous study reported that the magnitude of I_Ca,L modulation by NOS inhibitors showed minimal changes in I_Ca,L despite detectable effects on intracellular Ca²⁺ transients (Ebner et al., 2020). Thus, variations in the composition or hierarchy of these signaling complexes could attenuate the functional impact of nNOS inhibition on I_Ca,L in the RV. While the precise molecular mechanisms remain to be elucidated, these data imply differential coupling between nNOS and LTCC regulatory pathways between LVCMs and RVCMs.

In theories of E-C coupling in cardiomyocytes, “calcium sensitivity” and “calcium buffering” are often discussed separately or interchangeably; they represent fundamentally distinct but interconnected concepts (Bers, 2002; Eisner et al., 2017; Eisner et al., 2023). Calcium sensitivity refers to the magnitude of contractile response at a given intracellular Ca²⁺ concentration, typically described by the pCa–force relationship (or, in the present study, the Ca²⁺–SL relationship). In contrast, calcium buffering describes the kinetic properties of Ca²⁺ binding and unbinding within the cell. Among the variety of binding sites, cTnC occupies a unique position as both the sensor/regulator for contraction and a major buffer modulating free Ca²⁺ availability.

When Ca²⁺ buffer affinity increases, a larger fraction of Ca²⁺ becomes bound, particularly to myofilament sites such as cTnC, resulting in a lower free cytosolic Ca²⁺ concentration ([Ca²⁺]_c) for a given amount of Ca²⁺ entering the cytosol. This redistribution can shift the relationship between Ca²⁺ and sarcomere shortening (Ca²⁺–SL), giving the appearance of increased Ca²⁺ sensitivity in the Ca²⁺–SL relationship. Thus, increased intrinsic buffer affinity may contribute to the observed enhancement of Ca²⁺ sensitivity. We observed that pharmacological inhibition of nNOS reduced the p-cTnI significantly more in RVCMs. Since p-cTnI reduces the Ca²⁺-binding affinity of cTnC (Kentish et al., 2001; Solaro and Kobayashi, 2011), we predicted a greater increase in cTnC Ca²⁺ affinity in RVCMs following nNOS inhibition. By employing an “added-buffer approach” under conditions that minimized SERCA and NCX fluxes (Fig. 6), we could estimate the apparent properties of endogenous buffer, and observed a significantly decreased dissociation constant (⁠$Kd,S$⁠) of the endogenous Ca²⁺ buffer (binding species “S”) in both cell types (Mathias et al., 1990). The reduced $Kd,S$ is indicative of increased binding affinity; Ca²⁺ binds more readily at lower cytosolic concentrations. Notably, this reduction in apparent $Kd,S$ was quantitatively greater in RVCMs (⁠$ΔKd,S$ = −0.265 µM vs. −0.077 µM in LVCMs), paralleling the greater reduction in p-cTnI. These findings suggest that enhanced Ca²⁺ binding to troponin may have contributed, at least in part, to the observed change in K_d,S. Mg²⁺ occupancy should also be considered when interpreting $Kd,S$ because the C-terminal structural sites of cTnC can bind both Ca²⁺ and Mg²⁺. All groups were studied using the same MgATP- and MgCl₂-containing pipette solution, but free intracellular Mg²⁺ was not directly measured. Thus, the absolute $Kd,S$ values are conditional on the Mg²⁺ environment of the whole-cell recording, and Mg²⁺-dependent effects cannot be separated from the combined apparent behavior of the endogenous buffering system.

Moreover, this effect was most pronounced within the physiological range of free Ca²⁺ concentrations (<1 µM), where differences in buffer affinity exert the greatest influence on Ca²⁺ partitioning between bound and free pools (Fig. 6 B). These results indicate that nNOS inhibition alters apparent endogenous Ca²⁺ buffering, with a greater effect in RVCMs. The parallel reduction in p-cTnI is consistent with a contribution from altered myofilament Ca²⁺ regulation (Fig. 7). This chamber-specific response is consistent with a greater contribution of myofilament-localized nNOSβ in RV (Fig. 8).

The majority of intracellular Ca²⁺ in cardiomyocytes resides in a buffer-bound state (Bers, 2002), and alterations in buffer affinity are expected to influence Ca²⁺ transient kinetics (Smith and Eisner, 2019). From this perspective, the slowing of Ca²⁺ decay observed following nNOS inhibition is consistent with an increase in Ca²⁺ buffer affinity. Notably, the effect on mechanical relaxation was substantially greater than the change in Ca²⁺ decay in RVCMs (Fig. 2 vs. Fig. 3), suggesting that altered endogenous Ca²⁺ buffering, potentially including changes at myofilament binding sites such as TnC, may contribute to the delayed mechanical relaxation. This interpretation is supported by the concept that Ca²⁺ transient kinetics reflect coordinated interactions between Ca²⁺ removal systems and intracellular buffering processes (Briston et al., 2014). Thus, altered endogenous Ca²⁺ buffering, potentially including a myofilament component, may contribute to the greater slowing of relaxation in RVCMs. Accordingly, alterations in myofilament Ca²⁺ buffer properties may act not merely as secondary modifiers but as primary physiological determinants of contractile kinetics.

From this perspective, the effects of nNOS inhibition on Ca²⁺ transients and myofilament shortening in both LVCMs and RVCMs can be interpreted primarily in terms of a change in apparent endogenous Ca²⁺ buffering. Following SMTC treatment, the decay of the Ca²⁺ transient was modestly but significantly slowed in both cell types (Fig. 3), as reflected by a statistically significant prolongation of the time to 90% relaxation. The reduction in the apparent dissociation constant (⁠$Kd,S$⁠) of the endogenous Ca²⁺ binding species in both LVCMs and RVCMs suggests that an increase in Ca²⁺ buffer affinity contributes, at least in part, to the observed slowing of Ca²⁺ decay. Previous studies have reported variable effects of nNOS inhibition on SERCA function (Wang et al., 2008; Carnicer et al., 2013; Vielma et al., 2016). Thus, it is plausible that altered apparent endogenous Ca²⁺ buffering, potentially including a myofilament component, and a modest decrease in SERCA-mediated Ca²⁺ reuptake both contribute to the slowing of Ca²⁺ transient decay observed under nNOS inhibition.

Notably, the magnitude of the change in Ca²⁺ transient decay observed in the present study was smaller than that reported in several previous studies (Sears et al., 2003; Wang et al., 2008). This likely reflects our use of a relatively low concentration of SMTC (100 nM), chosen deliberately in light of reports indicating off-target effects of SMTC on other NOS isoforms at higher concentrations (Narayanan and Griffith, 1994; Awasthi et al., 2026). Importantly, even at this low concentration, clear and reproducible signaling effects were detected (Fig. 7).

Our quantitative analysis revealed a specific RVCM phenotype following nNOS inhibition: (1) a modest increase in Ca²⁺ transient peak, but (2) a larger increase in sarcomere shortening and (3) a significant reduction in diastolic SL. When integrated, these findings indicate a shift in the Ca²⁺-SL relationship toward higher Ca²⁺ sensitivity.

It is important to acknowledge that our real-time [Ca²⁺]_c–SL loop analysis differs from conventional steady-state pCa–force measurements. The dynamic nature of the contraction cycle involves hysteresis, and we deliberately excluded the systolic phase, during which nonlinear and highly dynamic interactions dominate. Thus, our analysis focused on the relaxation phase and diastolic levels where conditions approach quasi-equilibrium. Within these limitations, the data clearly demonstrate that in RVCMs, a greater contractile response is elicited at equivalent or even lower [Ca²⁺]_c following nNOS inhibition. This pattern is consistent with an increase in calcium sensitivity in the classical sense. However, our buffering data suggest this should be reinterpreted: the RV phenotype is driven by a change in the efficiency of Ca²⁺-contraction coupling, mediated by altered buffering capacity (increased bound fraction), rather than simply by the magnitude of the Ca²⁺ transient. This highlights the value of simultaneously measuring sensitivity and buffering to provide a unified interpretation of cardiac mechanics. The reduction in diastolic SL without a substantial increase in diastolic free Ca²⁺ also suggests persistent mechanical activation involving altered thin-filament regulation or crossbridge kinetics. The decrease in apparent $Kd,S$ is directionally consistent with greater endogenous Ca²⁺ binding at a given free Ca²⁺ concentration and may contribute to this phenotype. However, the present measurements do not establish whether the magnitude of the buffering shift is quantitatively sufficient to account for the observed reduction in diastolic SL, because endogenous Ca²⁺ binding could not be solely related to troponin occupancy.

The buffering analysis provides complementary but distinct information. The reduction in apparent K_d,S indicates altered apparent Ca²⁺-binding behavior of the combined endogenous buffering system, but it does not directly demonstrate an increase in the biochemical Ca²⁺ affinity of cTnC. Together with the reduction in p-cTnI, the changes in the Ca²⁺-SL relationship and apparent endogenous buffering properties are consistent with altered myofilament Ca²⁺ regulation in the SMTC-treated RVCMs. Thus, the Ca²⁺–SL relationship provides a functional index of Ca²⁺ sensitivity, whereas the added-buffer analysis describes the apparent behavior of the endogenous buffering system. It has to be noted that neither measurement directly determines the biochemical Ca²⁺ affinity of isolated cTnC.

The RV is structurally more compliant and operates under lower afterload conditions than the LV. The higher basal nNOSβ activity and consequent lower Ca²⁺ sensitivity (and lower buffering affinity) in RVCMs may serve a physiological role to preserve lusitropy and limit energetic cost under basal conditions. Higher myofilament-associated nNOSβ activity in RVCMs may provide a cellular mechanism that favors relaxation under basal conditions. Such regulation could contribute to RV diastolic function with markedly lower afterload than LV, although its significance in the mechanically loaded intact ventricle remains to be determined.

Alterations in this pathway may have implications for pathophysiology. In our study, nNOS inhibition produced cellular features that may be relevant to pathological RV remodeling, including reduced diastolic SL and impaired relaxation. In conditions like pulmonary hypertension or RV hypertrophy, if nNOS activity or cGMP-PKG signaling is impaired, the resulting loss of TnI phosphorylation could lead to “hypersensitivity” to calcium, contributing to diastolic dysfunction and energetic inefficiency. Whether disruption of this pathway contributes to RV dysfunction during pressure overload, and whether its restoration improves RV performance, should be tested in intact-heart and disease models.

While this study provides novel insights into the compartmentalized role of nNOSβ in RVCMs, certain limitations warrant mention. Firstly, our experiments were conducted under basal conditions; the dynamic regulation of this pathway during adrenergic stimulation, exercise, or pathological stress remains to be elucidated. Secondly, while SMTC is widely used as a selective nNOS inhibitor, potential off-target effects or incomplete specificity cannot be entirely excluded. Thirdly, the SL measured in this study represents unloaded isolated cardiomyocytes and should not be considered equivalent to the end-diastolic SL of the mechanically loaded myocardium.

With regard to the diastolic SL, Bub et al. (2010) reported a modal SL of 1.95 µm in cardioplegically arrested, Langendorff-perfused rat hearts, compared with 1.71 µm in acutely isolated ventricular cardiomyocytes. However, this intact-heart value was obtained during cardioplegic arrest and was influenced by progressive edema, and therefore it should not be regarded as a direct measurement of physiological end-diastolic SL. Studies of intact rat LV myocardium have also shown that diastolic SL varies with ventricular filling pressure, transmural location, and residual wall stress (Grimm et al., 1983; Rodriguez et al., 1993). Moreover, direct side-by-side measurements of physiological end-diastolic SL in the normal rat RV and LV remain limited. The absolute SL and magnitude of the responses observed here therefore cannot be directly extrapolated to the mechanically loaded myocardium.

Despite the limitations mentioned above, isolated cardiomyocytes provide a useful model for comparing cellular phenotypes of LVCMs and RVCMs under identical experimental conditions. Previous studies have demonstrated that chamber-specific differences in action-potential configuration, Ca²⁺ handling, contraction, troponin expression, and intracellular Ca²⁺ buffering are retained after isolation (Kondo et al., 2006; Jeon et al., 2022). In the present study, cells were collected separately from the free walls of the RV and LV, excluding the interventricular septum, thereby allowing direct comparison of chamber-specific cellular properties. However, isolated cells do not reproduce physiological preload and afterload, the extracellular matrix, myocardial architecture, ventricular interaction, or neurohumoral regulation. Thus, our findings identify a chamber-specific cardiomyocyte phenotype that remains detectable after isolation, although its quantitative contribution to intact ventricular function remains to be established. Importantly, its detection after isolation does not establish that the phenotype is independent of prior in vivo mechanical loading or neurohumoral signaling.

In conclusion, our findings reveal a more pronounced role for myofilament-associated nNOSβ in the RV. By acting locally to phosphorylate cTnI Ser^23/24, nNOSβ acts as a crucial “brake” on myofilament Ca²⁺ sensitivity, thereby facilitating relaxation and maintaining diastolic compliance. This chamber-specific mechanism emphasizes that cardiac NO signaling is not uniform but is finely tuned by subcellular localization to meet the distinct hemodynamic demands of the RV.

The numerical data underlying Figs. 1–8 and S1–S4, including individual animal- and cell-level values, are available from the corresponding author upon request.

David A. Eisner served as editor.

This work was supported by the National Research Foundation of Korea funded by the Ministry of Science and ICT of the Republic of Korea (No.: RS-2024-00456454 and RS-2025-02303718 to S.J. Kim; No.: RS-2024-00360549 to Y.K. Jeon). This work was also supported by the Phase III (Postdoctoral fellowship) grant of the SPST (Seoul National University [SNU]-SNU Hospital [SNUH] Physician Scientist Training to Y.K. Jeon). Open Access funding provided by Seoul National University.

Author contributions: Jae Won Kwon: conceptualization, data curation, formal analysis, investigation, resources, validation, and visualization. Hyerim Park: writing—review and editing. Ryeon Heo: writing—review and editing. Jihyun Jang: conceptualization and writing—review and editing. Jintae Kim: investigation. Seung Beom Oh: writing—review and editing. Yin Hua Zhang: conceptualization, methodology, supervision, validation, and writing—review and editing. Sung Joon Kim: conceptualization, resources, supervision, and writing—review and editing. Young Keul Jeon: conceptualization, data curation, formal analysis, funding acquisition, investigation, project administration, resources, supervision, validation, visualization, and writing—original draft, review, and editing.