Rad regulation of CaV1.2 channels controls cardiac fight-or-flight response

Fight-or-flight responses involve β-adrenergic-induced increases in heart rate and contractile force. In the present study, we uncover the primary mechanism underlying the heart’s innate contractile reserve. We show that four protein kinase A (PKA)-phosphorylated residues in Rad, a calcium channel inhibitor, are crucial for controlling basal calcium current and essential for β-adrenergic augmentation of calcium influx in cardiomyocytes. Even with intact PKA signaling to other proteins modulating calcium handling, preventing adrenergic activation of calcium channels in Rad-phosphosite-mutant mice (4SA-Rad) has profound physiological effects: reduced heart rate with increased pauses, reduced basal contractility, near-complete attenuation of β-adrenergic contractile response and diminished exercise capacity. Conversely, expression of mutant calcium-channel β-subunits that cannot bind 4SA-Rad is sufficient to enhance basal calcium influx and contractility to adrenergically augmented levels of wild-type mice, rescuing the failing heart phenotype of 4SA-Rad mice. Hence, disruption of interactions between Rad and calcium channels constitutes the foundation toward next-generation therapeutics specifically enhancing cardiac contractility.

direct phosphorylation of the Ca V 1.2 pore-forming α 1 -and/or auxiliary β-subunits is incorrect: β-adrenergic agonists still upregulate current through Ca V 1.2 in which all potential phosphorylation sites on the αand β-subunits have been removed [13][14][15] . We recently identified the small RGK G-protein Rad, an inhibitor of voltage-activated Ca 2+ channels 16,17 , as an alternative target 15 . Using an ascorbate peroxidase (APEX2)-catalyzed proximity labeling in transgenic mouse hearts, combined with quantitative proteomics, we observed that, under basal conditions, Rad was enriched in the neighborhood of Ca V 1.2; on exposure to a β-adrenergic agonist, however, Rad was depleted from around Ca V 1.2 (ref. 15 ). With the identification that PKA phosphorylation of the small RGK G-protein Rad relieves its inhibition of heterologously expressed, voltage-gated Ca 2+ channels 17 and the development of mice expressing mutant Rad and Ca V 1.2 channels, the mechanisms underlying the primordial fight-or-flight responses that boost heart rate and contractility can now be ascertained.

Adrenergic regulation of calcium influx
We developed knock-in mice (Extended Data Fig. 1a) in which the four evolutionarily conserved PKA-phosphorylated serine residues of the endogenous murine Rrad locus (Ser25, Ser38, Ser272 and Ser300) were replaced by alanine residues (4SA-Rad). The protein expression in cardiomyocytes of Rad (Extended Data Fig. 1b) or the principal Ca V 1.2 subunits, α 1C and β 2B , were not affected by introducing these mutations in Rad (Extended Data Fig. 1c). We interrogated the electrophysiological properties of Ca 2+ channels in ventricular cardiomyocytes isolated from WT and homozygous 4SA-Rad knockin mice using either traditional voltage-steps with Ca 2+ as the charge carrier (Fig. 1a,b) or a voltage ramp protocol applied every 6 s with Ba 2+ as the charge carrier (Fig. 1c). The ramp protocol enables the monitoring of agonist effects over time. The basal electrophysiological properties of Ca 2+ channels in the 4SA-Rad cardiomyocytes did not differ (Extended Data Fig. 1d,e), but for the amplitude of peak current that was reduced compared with wild-type (WT) cardiomyocytes (Fig. 1d), despite unchanged protein expression of Ca V 1.2 subunits. In sharp contrast to the Ca 2+ current of WT ventricular cardiomyocytes, neither was the Ca 2+ current of 4SA-Rad ventricular cardiomyocytes augmented nor was the membrane potential for half-maximum activation, V 50 , shifted by either the nonselective β-adrenoreceptor agonist isoproterenol or the adenylyl cyclase activator forskolin ( Fig.   1a-c,e,f and Extended Data Fig. 1f-h). Isoproterenol induced a rapid increase in current across all test potentials in WT but not 4SA-Rad ventricular cardiomyocytes (Fig. 1h,i). Cardiomyocytes isolated from heterozygous 4SA-Rad knock-in mice, which bear one WT allele and one 4SA-Rad allele, demonstrated an intermediate response to forskolin (Extended Data Fig. 1h).
Cardiac atrial contraction normally accounts for ~10% of left ventricular filling at rest and up to ~40% of ventricular filling at high heart rates, such as in fight-or-flight responses. The atria in the heart respond to isoproterenol with much larger increases in developed tension, contractility and relaxation rates than ventricular papillary muscles 18 . Consistent with these findings, we observed substantially greater isoproterenol-induced augmentation of the Ca 2+ current in atrial myocytes than in ventricular myocytes ( Fig. 1g-i). As in ventricular myocytes, isoproterenol failed to increase the current in 4SA-Rad atrial myocytes ( Fig.  1g-i). Thus, adrenergic augmentation of cardiac Ca 2+ currents in both atrial and ventricular myocytes is fully dependent on phosphorylation of Rad and not on phosphorylation of the principal channel α 1C -and β 2 -subunits [13][14][15] .

Calyculin A-induced changes in Ca V 1.2 neighbors
As the balance between kinase and phosphatase activity contributes to setting the basal Ca 2+ current in cardiomyocytes, which was reduced in 4SA-Rad cardiomyocytes (Fig. 1d), we set out to identify key components underlying basal Ca V 1.2 regulation. Application of various protein phosphatase inhibitors, such as okadaic acid, microcystin or calyculin A, to cardiomyocytes results in large increases in Ca 2+ current amplitude [19][20][21][22][23] via an unknown mechanism. In mice, the calyculin A-induced increase in current in WT cardiomyocytes is insensitive to the PKA inhibitor Rp-8-Br-cAMPS (Extended Data Fig. 1i) 20,23 .
Previously, we applied enzyme-catalyzed proximity labeling and multiplexed mass spectrometry (MS) in cardiomyocytes, which quantified isoproterenol-induced changes in the molecular environment of Ca V 1.2 channels of cardiomyocytes 15 . We now leverage this approach to identify the mechanism by which Ca 2+ influx can be modulated in cardiomyocytes at basal conditions. Using transgenic mice expressing Ca V 1.2 α 1C -APEX2 (ref. 15 ) and multiplexed tandem mass tag (TMT) MS, we assessed changes in the Ca V 1.2 neighborhood after adding calyculin A or isoproterenol. The remarkable 50% decrease in Rad level in proximity to Ca V 1.2 caused by calyculin A exposure is comparable to the isoproterenol-induced depletion of Rad (Fig. 2a,b and Supplementary Table 1), suggesting that Rad is also involved in nonadrenergic regulation of Ca V 1.2. In contrast to addition of isoproterenol (Fig. 2a), we did not observe recruitment of the PKA catalytic subunit to the channel neighborhood by calyculin A (Fig. 2b)  We hypothesized that the calyculin A-induced change in Rad localization may depend on Rad phosphorylation, probably on at least some of the four phosphorylated serine residues in Rad. We utilized flow cytometry Förster resonance energy transfer (FRET) two-hybrid assay 24 to probe potential calyculin-mediated changes in the macromolecular complex of Ca V 1.2. At baseline, robust binding is detected between Cerulean-tagged β 2B -subunit and Venus-tagged WT Rad expressed in human embryonic kidney (HEK) cells (Fig. 2c,d,f). Incubation of HEK cells with calyculin A markedly weakened this interaction (Fig. 2e,f). Calyculin A had no effect, in contrast, on the interaction between the β 2B -subunit and 4SA-Rad ( Fig. 2g-j). As definitive proof of the mechanism by which phosphatase inhibition increases ionic currents through Ca V 1.2 channels, we compared the effects of calyculin A in WT and 4SA-Rad ventricular myocytes. Compared with WT cardiomyocytes (Fig.  2k), calyculin A neither increased the Ca V 1.2 current amplitude (Fig. 2l,m) nor shifted the current-voltage relationship in a hyperpolarizing direction (Fig. 2n) in the 4SA-Rad ventricular cardiomyocytes. Thus, signaling pathways other than the β-adrenergic-PKA system are integrated by Rad phosphorylation and contribute to the setting of the basal Ca 2+ current in cardiomyocytes.

Adrenergic augmentation of Ca 2 transient
How relevant is the modulation of Ca 2+ channels in the broader context of changes in the Ca 2+ transient and contractility, processes that involve not only the Ca V 1.2 channel, but many other ion channels and transporters, all targets of adrenergic signaling? Excitationcontraction coupling in myocytes is initiated by myocyte membrane depolarization, leading to the opening of Ca V 1.2 channels and the influx of extracellular Ca 2+ that in turn activates RyR2, the SR Ca 2+ release channels (Fig. 3a). The transient increase in cytosolic Ca 2+ concentration induces contraction. Electrical field-stimulation-induced Ca 2+ transients of WT and 4SA-Rad ventricular cardiomyocytes, detected by the ratiometric Ca 2+ indicator Fura2-AM, were quantified before and after a 2-min superfusion of vehicle, isoproterenol or forskolin (Fig. 3b). Thereafter, the Ca 2+ content of the SR was assessed by rapid infusion of caffeine, which induces release of SR Ca 2+ .
As expected, isoproterenol and forskolin, but not vehicle, increased the Ca 2+ transient amplitude in WT ventricular cardiomyocytes (Fig. 3c-f). The effect of isoproterenol on the Ca 2+ transient amplitude was even more profound in WT atrial cardiomyocytes (Fig.  3g,h). The extent of isoproterenol-and forskolin-induced change in the Ca 2+ transient of WT ventricular cardiomyocytes was inversely related to the basal transient amplitude (Fig. 3d,e), similar to the inverse relationship between basal Ca 2+ current and the extent of the response to the β-adrenergic agonist 25 .
Compared with WT ventricular myocytes, the increase in the Ca 2+ transient by adrenergic agonists was markedly attenuated in 4SA-Rad ventricular myocytes by 70% (80% increase in 4SA-Rad versus 260% increase in WT) ( Fig. 3d-f). Compared with WT atrial myocytes, the isoproterenol-induced increase in Ca 2+ transient was attenuated by 93% (32% increase in 4SA-Rad versus 490% increase in WT) in 4SA-Rad atrial myocytes (Fig. 3g,h). The 4SA-Rad cardiomyocytes exhibited increased amplitude of the basal Ca 2+ transient compared with WT cardiomyocytes (Fig. 3i), despite the modest reduction in basal Ca 2+ conductance (Fig. 1d). This implies at least some downstream compensatory signaling to attenuate the loss of adrenergic augmentation of the Ca 2+ transient amplitude.
Assessed by the magnitude of caffeine-induced SR Ca 2+ release, isoproterenol and forskolin increased SR Ca 2+ load in WT but not 4SA-Rad cardiomyocytes (Fig. 4a), despite equivalent expression of SERCA and the Na + /Ca 2+ exchanger (NCX) (Fig. 4b,c). Although adrenergic augmentation of the Ca 2+ transient was markedly diminished in the 4SA-Rad cardiomyocytes, other fundamental adrenergic signaling pathways remained fully functional in these mice. First, both isoproterenol and forskolin accelerated the rate of Ca 2+ reuptake (Fig. 4d). Second, isoproterenol and forskolin induced phosphorylation of PLN (Fig. 4e), RyR2 (Fig. 4f) and troponin I (TnI) (Fig. 4g) in ventricular myocytes isolated from both WT and 4SA-Rad mice. Taken together, the bulk of the adrenergic Ca 2+ transient and contraction augmentation in isolated cardiomyocytes depends on Rad phosphorylation and the subsequent increase in Ca 2+ influx.

Rad phosphorylation is required for increased contractility
Mice expressing 4SA-Rad were born at normal Mendelian ratios (Extended Data Fig. 2a) and their survival was equivalent to WT animals up to age at least 6 months. Histological examination of 4SA-Rad hearts did not display increased fibrosis or changes in wall thickness (Extended Data Fig. 2b). Heart and lung weights in WT and 4SA-Rad mice did not differ (Extended Data Fig. 2c,d). Body weight of 4SA-Rad mice was slightly increased compared with littermate WT mice (Extended Data Fig. 2e).
Bulk RNA-seq demonstrated modest changes with 607 genes downregulated and 445 genes upregulated in homozygous 4SA-Rad hearts compared with WT hearts (Extended Data Fig. 3a,b). Upregulated Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways in the 4SA-Rad hearts included arrhythmogenic right ventricular cardiomyopathy, adrenergic signaling, regulation of heart contraction and metabolism (Extended Data Fig. 3b). Notably, the transcript levels for Ca V 1.2 α 1C -and β 2B -subunits, RyR2, SERCA2, PLN and Rad and other RGK GTPase family members were not substantially altered (Extended Data Fig. 3c,d and Supplementary Table 2). In contrast, transcripts levels for adrenergic receptors (Adra1a, Adra1b), HCN4, Ca V 3.1 (Cacn1g) and Ca V 3.2 (Cacn1h), Ca V α 2 δ 1 (Cacna2d1) and several K + channels were upregulated (Extended Data Fig. 3c,d). We surmise that these changes reflect compensation for the loss of adrenergic regulation of Ca 2+ channels.
To assess the role of Rad phosphorylation in cardiac contractility in vitro, we measured the changes in pacing-induced sarcomere length before and after exposure to forskolin. In WT cardiomyocytes, the sarcomere contraction increased from 3.4% to 11.3% (absolute difference 7.9%) in response to forskolin (Fig. 5a,c). In contrast, the sarcomere contraction increased only from 2.7% to 5.3% (absolute difference 2.6%) in forskolin-treated 4SA-Rad ventricular myocytes (Fig. 5b,c). The forskolin-induced acceleration in the relaxation speed after the pacing-induced contraction was equivalent in WT and 4SA-Rad cardiomyocytes (Fig. 5d), consistent with the normal adrenergic signaling to PLN and TnI in 4SA-Rad cardiomyocytes (Fig. 4e,g).
To assess the role of Rad phosphorylation in cardiac contractility in vivo, we performed echocardiography on isoflurane-anesthetized WT and 4SA-Rad mice. In the 4SA-Rad mice, we found ~25% reduction in baseline contractility, with expansion of both end-diastolic and end-systolic volumes of the left ventricle, and a reduction in both global circumferential strain (GCS) and global longitudinal strain (GLS; Fig. 5e-h and Extended Data Fig. 4ae). Next, we assessed the effects of intraperitoneal injection of isoproterenol on cardiac contractility (Fig. 5i). Isoproterenol increased the ejection fraction and fractional area of change by 81% and 86% (absolute difference 38% and 37%), respectively, in WT mice, but increased the ejection fraction and fractional area of change by only 9% and 14% (absolute difference 3% and 5%), respectively, in 4SA-Rad mice (Fig. 5j,k). Speckle tracking-based strain analysis confirmed the pronounced attenuation of the isoproterenol effect in 4SA-Rad mice (Fig. 5l). These data demonstrate that adrenergic augmentation of cardiac contractility in vivo strongly depends on Rad phosphorylation and the subsequent increased Ca 2+ influx.
We subjected WT and 4SA-Rad mice to treadmill exercise testing. After 2 d of treadmill acclimation and training, mice were subjected to a 20-min exercise session at an incline of either 0° or 15°. At constant speed (25 cm s −1 ) and no incline, substantially more 4SA-Rad mice than WT mice failed the 20-min testing session, with a shorter latency to failure in the 4SA-Rad mice (Fig. 5m). Although the number of mice failing to complete the 20-min testing session increased for both WT and 4SA-Rad groups at an incline of 15° and progressively increasing speed, the 4SA-Rad mice group was more prone to failure with reduced latency (Fig. 5m). Theoretically, we cannot exclude that phosphorylated Rad has roles in metabolism 26 or in other tissues relevant to exercise capacity, such as skeletal muscle 27 . Still, exercise intolerance is a hallmark of heart failure and/or inability to increase cardiac output, and these findings are consistent with the expected consequences of diminished fight-or-flight responses.

Adrenergic regulation of heart rate
In addition to increased myocyte contractility rate, the fight-or-flight response depends on increased heart rate. Cardiac automaticity is chiefly driven by membrane potential depolarization of sinoatrial nodal (SAN) cells during diastole, determined by the complex coupling of the 'membrane clock' and the 'Ca 2+ clock' 28 . The coupled clock model postulates that, along with the inward cation current through hyperpolarization-activated cyclic nucleotide-gated (HCN4) channels, spontaneous local Ca 2+ release events and NCX depolarize the membrane causing activation of T-type and L-type Ca 2+ channels (Ca V 1.2 and Ca V 1.3) and generation of the action potential 29 . Adrenergic agonists, via cAMP generation and PKA activation, accelerate pacemaker activity via integrated actions on multiple targets including HCN4, Ca 2+ channels, RyR2, PLN/SERCA, NCX and K + channels 29,30 . Ca V 1.3 channels are critical for the initiation of pacemaker activity in dormant mouse SAN cells by β-adrenergic stimulation 31 . Rad is expressed in SAN cells and Rad knockout mice demonstrate increased intrinsic and sleep-phase heart rates consistent with a role of Rad in modulating Ca 2+ current in SAN cells 32 . Using the 4SA-Rad mice generated in the present study, we assessed the role of adrenergic stimulation of Ca V 1.2 and Ca V 1.3 channels on sinus node function.
We implanted radio-telemeters in mice and recorded electrocardiograms (ECGs) under basal conditions to assess sinus node function. The 4SA-Rad mice demonstrated a reduction in minimum, mean and maximum heart rate over a 24-h period (Fig. 6a). Acute injection of isoproterenol induced an increase in maximum heart rate in both WT and 4SA-Rad mice (Fig. 6b). During the subsequent several hours after isoproterenol injection, however, we observed a greater slowing of the heart rate in 4SA-Rad mice, marked by prolonged episodes of irregularity and slow heart rates below 400 beats min −1 , which was rarely observed in WT mice (Fig. 6c,d). These data suggest that Rad phosphorylation has important protective effects in preventing slow heart rates and stabilizing sinus node pacemaker activity during periods of high stress. As heart rate increases with isoproterenol, the reduction in exercise capacity is not related to a lack of chronotropic response.

Rad-inhibited channels required for adrenergic regulation
Expression of Rad profoundly inhibits the open probability of Ca V 1.2 channels 15,33 . The presence of Ca 2+ current in cardiomyocytes under basal conditions implies that a substantial fraction of Ca 2+ channels is not bound to Rad, and that Rad-bound Ca 2+ channels form the functional reserve of Ca 2+ influx and cardiac contractility. Rad-null mice demonstrate high basal contractility and blunted adrenergic responses 34,35 . To test whether the Rad-bound Ca 2+ channel form the contractile reserve, we generated mice in which the interaction between the Ca 2+ channel β-subunit and Rad is reduced (Fig. 7a). Previous studies showed that substituting three aspartic acid β 2B residues, Asp244, Asp320 and Asp322, in the human β 2B -subunit, with alanine (3DA) attenuated Rad binding to the Ca 2+ channel β-subunit 36,37 , which we confirmed using a flow cytometry, Förster resonance energy transfer (FRET), two-hybrid assay in HEK293 cells 24 (Extended Data Fig. 5a). To facilitate generation of knock-in mice, because the aspartate-coding residues in exon 9 and exon 11 are separated by 9.3 kb, we determined that alanine substitutions of only the two aspartate residues in exon 11 (2DA) are sufficient to reduce Rad-β 2B -subunit interaction (Extended Data Fig. 5a). These mutations in the β-subunit did not affect the interaction between the I-II loop of α 1C and β, also assessed by a FRET assay (Extended Data Fig. 5b).
In ventricular cardiomyocytes isolated from mice with mutant Ca 2+ channel β subunits that cannot bind Rad, the basal conductance was increased (Fig. 7d) and the V 50 for activation was substantially hyperpolarized; neither V 50 nor conductance was substantially changed by adrenergic agonists (Fig. 7e,f and Extended Data Fig. 5d). Similarly, the amplitude of the Ca 2+ transient was increased under basal conditions to almost the levels of isoproterenoltreated WT cardiomyocytes (Fig. 7g) and did not substantially change after isoproterenol administration in 2DA-β 2B or 3DA-β 2B cardiomyocytes, in contrast to WT or transgenic WT β 2B cardiomyocytes (Fig. 7h,i). Basal cardiac contractility, assessed by echocardiography, was increased and the augmentation of contractility by isoproterenol was blunted in both lines of mice compared with either WT or WT β 2B -expressing transgenic mice controls (Fig.  7j,k). These findings reveal that a subpopulation of Rad-bound Ca 2+ channels is required for sympathetic nervous system regulation of Ca 2+ influx, transient and contractility.
To verify that reduced basal Ca 2+ influx and attenuated adrenergic agonist-induced augmentation of cardiac contractility in the 4SA-Rad mice are due to dysfunctional regulation of Ca 2+ channels, we crossed the 4SA-Rad mice with 3DA-β 2B transgenic mice. As expected, cardiomyocytes isolated from combined homozygous 4SA-Rad and transgenic 3DA-β 2B mice had activated Ca 2+ channels without exposure to adrenergic agonists, marked by increased conductance (Fig. 8a), hyperpolarized V 50 for activation ( Fig. 8b) and markedly attenuated adrenergic regulation (Fig. 8b,c). Furthermore, these myocytes displayed increased basal Ca 2+ transients (Fig. 8d) and diminished adrenergic agonist-induced augmentation of both Ca 2+ transients (Fig. 8e,f) and contractility (Fig.  8g). Isoproterenol accelerated the rate of Ca 2+ reuptake (Fig. 8h), revealing that other adrenergic signaling pathways were unperturbed in these mice. The homozygous 4SA-Rad mice crossed with transgenic 3DA-β 2B mice displayed increased basal left ventricular contractility and diminished adrenergic-induced augmentation of contractility (Fig. 8i). As prevention of 4SA-Rad from binding to Ca V β 2B increased the basal Ca 2+ transient and contractility compared with 4SA-Rad alone, we exclude confounding 'off-target' functions and conclude that the phenotypes imparted by 4SA-Rad are solely due to a direct effect on Ca 2+ channels. Moreover, that the Ca 2+ current, Ca 2+ transients and cardiac contractility can be augmented to near adrenergic-agonist levels independent of the β-adrenergic system and Rad phosphorylation suggests a therapeutic target for increasing cardiac contractility in patients with failing hearts.

Discussion
We show that PKA phosphorylation of Rad is essential for regulation by the sympathetic nervous system of Ca 2+ influx in atrial and ventricular myocytes and for augmentation of cardiac contractility. The adrenergic regulation of Ca V 1.2 channels can be fully abrogated by preventing β-subunit binding to α 1C 14 , introducing flexibility to the rigid linker 38 between the β-subunit-binding site in the I-II loop and the channel pore 33 , replacing WT Rad with a mutant Rad that cannot be phosphorylated or preventing β-subunit binding to Rad. That no acute adrenergic increase in Ca 2+ current is observed in any of these four distinct mouse lines suggests that plasma membrane insertion of additional Ca 2+ channels after adrenergic stimulation 39 is not a major contributor to augmentation of Ca 2+ influx.
In the absence of Rad phosphorylation, adrenergic agonist-induced enhancement of cardiac contraction is almost completely disabled. As Rad-bound channels have very low open probability and are essentially electrically silent, Rad-less Ca 2+ channels are the basis for the initiation of excitation-contraction coupling under basal conditions. The adrenergic reserve of Ca 2+ influx and the potential to boost the contractile output, in contrast, are fully dependent on the Rad-bound Ca 2+ channels in both the atrial and the ventricular chambers of the heart.
Is Rad phosphorylation and release of Ca 2+ channel inhibition sufficient to independently elevate contractility? To answer this question, we developed mice with an ablated Rad-Ca V β interaction through mutations of Ca V β. Solely releasing the subpopulation of Radbound Ca 2+ channels from inhibition without activating the β-adrenergic-PKA signaling pathways was sufficient to fully activate Ca 2+ current, and to substantially augment the Ca 2+ transient and contractility to the levels typically induced by adrenergic agonists. Our findings establish that the principal mechanism of adrenergic control of contractility is via enhancement of Ca 2+ influx via Ca V 1.2 channels. The several-fold increase in Ca 2+ current results in increased triggering of RyR2 channels and increased Ca 2+ release from the SR, leading to increased cardiac contractility. The small residual increase in the Ca 2+ transient and contractility by adrenergic agonists in 4SA-Rad, 2DA-β 2B and 3DA-β 2B mice is probably due to phosphorylation of RyR2 (ref. 8 ), PLN/SERCA 3,40 and perhaps other targets.
Patients hospitalized with severely decompensated heart function have limited therapeutic options and are typically treated with invasive implantation of mechanical pumps, or βadrenergic agonists or phosphodiesterase inhibitors, which increase PKA activity. Although the goals of these pharmacological interventions are to promote cardiac contractility 41 , their long-term use is limited by diminishing responses over time and substantial side effects in the cardiovascular and other organ systems 42,43 . We demonstrate that the disruption of the Rad-Ca V β interaction is an equally efficacious but substantially more specific downstream target than the currently available upstream cardiotherapeutic activators of the PKA signaling pathways. A therapeutic that targets this interaction would be the foundation for specific cardiac inotropes.

Generation of knock-in and transgenic mice
The present study conformed to the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (NIH) and protocols approved by the Institutional Animal Care and Use Committee of Columbia University. Animals were maintained under a standard 12-h light:12-h dark cycle and had free access to standard chow and water. We used male and female mice aged 6 weeks to 5 months. The investigators were blinded to group allocation during data acquisition and analysis.
The 4SA-Rad knock-in mouse line was generated by Genoway. The murine genomic region encompassing the targeted Rrad mouse gene from C57BL/6N mouse genomic (g)DNA was used for homologous recombination. Ser25 and Ser38 in exon 2 were mutated to alanine, and Ser272 and Ser300 in exon 5 were mutated to alanine. The LoxP/FRT Neo cassette in the intron between exons 4 and 5 was deleted through mating with C57BL/6N Cre-deleter mice (Extended Data Fig. 1a). The locus and a minimum of 1-kb downstream and upstream of each homology arm were sequenced. Mice were exclusively maintained in the C57BL/6N background by mating of heterozygous mice.
The WT and 3DA-β 2B transgenic mutant mouse constructs were created by ligating in-frame a 3× FLAG epitope to the amino-terminus of human CACNB2b cDNA (accession no. AAG01473) and mutating residues Asp244, Asp320 and Asp322 to alanine by site-directed mutagenesis. The WT and the 3DA-β 2B cDNAs were ligated into the pJG/α-myosin heavy chain (MHC) plasmid (a gift from J. Robbins, Addgene plasmid no. 55594) 44 , between the 5.5-kb murine α-MHC promoter and the human growth hormone polyadenylation sequence. The transgenic mice were created by the Genetically Modified Mouse Models Shared Resource at Columbia University. The WT and the 3DA-β 2B transgenic mice, on a B6CBA/F2 hybrid background, were bred with WT C57BL/6N mice or 4SA-Rad mice, which are on the C57BL/6N background.
The 2DA-β 2B knock-in mouse line, with alanine substitutions for the two aspartate residues in exon 11 (equivalent of human Asp320 and Asp322), was created using CRISPR-Cas9 gene editing. Validation of the single guide (sg)RNA and single-strand oligodeoxynucleotide (ssODN) was performed in the Genome Engineering and iPSC Center (GEic) at Washington University (Extended Data Fig. 5c). Zygotes isolated from C57BL/6N mice were electroporated with the sgRNA and ssODN at Mount Sinai School of Medicine Mouse Genetics and Gene Targeting Core. Identification of potential founders and germline transmission after crossing with WT C57BL/6N mice was performed at Washington University by deep sequencing of gDNA from tail biopsies. Heterozygous 2DA-β 2B offspring mice were crossed to obtain homozygotes. Genotypes were identified by PCR of gDNA and sequencing.

Histology
Total body weight and tibial length were measured for 2-to 9-month-old mice. Hearts and lungs were harvested and weighed. Hearts were fixed in 4% paraformaldehyde overnight and processed for routine paraffin histology. They were stained with hematoxylin and eosin and Masson's trichrome.

Cellular electrophysiology
Mice ventricular myocytes were isolated by enzymatic digestion using a Langendorff perfusion apparatus 13 To measure peak currents, we held the cell membrane potential at −60 mV and stepped it to +50 mV for 150 ms in 10-mV increments every 10 s. Cells without a stable baseline (possibly due to run-down or run-up) were not studied. Membrane currents were measured by the conventional whole-cell patch-clamp method using a MultiClamp 700B amplifier and pCLAMP 10.7 software (Molecular Devices). The acquisition sampling rate for this step protocol was 20 kHz. Capacitance transients and series resistance were compensated for (>85%). Voltage was corrected for liquid junction potential (−10 mV) during analysis. Leak currents were subtracted by a P/3 protocol. The conductance was normalized to cell size. The voltage-step protocol used in cardiomyocytes studies evaluated I peak = I peak (V), which was recalculated in CLAMPFIT to G = G(V) as G = I/(V − E rev ). The parameters of voltage-dependent activation were obtained using a Boltzmann approximation curve 15 .
In many experiments, we used a ramp protocol with a 200-ms voltage ramp from −60 mV to +60 mV (0.6 V s −1 ) applied every 6 s to monitor the current-voltage (I-V) relationship. Voltage was corrected for liquid junction potential (10 mV) during analysis. Leak currents were subtracted by a P/3 protocol. Capacitance transients and series resistance were compensated for (>85%). The acquisition sampling rate for the ramp protocol was 5 kHz. In these experiments the external solution contained 0.5 mM BaCl 2 instead of 1.8 mM CaCl 2 . Under these conditions, currents through Ca 2+ channels are small (<1 nA) and showed practically no inactivation. After establishing stable records (usually after 2-3 min), 10-15 traces were recorded for the control. Thereafter, isoproterenol or forskolin was superfused. After the response stabilized, typically within 2-3 min for isoproterenol and 3-6 min for forskolin, 10-15 additional traces were recorded. When no response was observed, we continued the experiments for 4-6 min.

RNA-seq
Mouse hearts were removed, the atria removed and the ventricles snap-frozen in liquid nitrogen before storing at −80 °C. RNA-seq was performed at the Genomics and High Throughput Screening Shared Resource at Columbia University. RNA was extracted from the samples with QIAGEN miRNeasy micro kit (catalog no. 217084) following the kit protocol, except 0.7 volume of 100% ethanol was used instead of 1.5 volume of 100% ethanol for binding of total RNA on to the column. We used poly(A) pull-down to enrich mRNAs from total RNA samples and then proceeded with library construction using Illumina TruSeq chemistry. Libraries were then sequenced using Illumina NovaSeq 6000. We multiplexed samples in each lane, which yielded targeted number of paired-end 100-bp reads for each sample. Real-time analysis (Illumina) was used for base calling and bcl2fastq2 (v.2.19) was used for converting BCL to fastq format, coupled with adapter trimming. We performed a pseudoalignment to a kallisto index created from transcriptomes (GRCm38) using kallisto (0.44.0). Differential gene expression analysis in the 4SA-RAD knock-in versus littermate control mice was performed using the R package DESeq2 (v.1.13.0) from unnormalized count matrix with a false discovery rate (FDR) cut-off of 0.05.

Fractional shortening of isolated cardiomyocytes
Freshly isolated myocytes were superfused with Tyrode's solution containing 1.2 mM CaCl 2 . Myocytes were field stimulated at 1 Hz. The percentage contraction of the sarcomere length was measured using the SarcLen module of Ionoptix and calculated as the difference of shortest sarcomere length during a contraction subtracted from the relaxed sarcomere length, divided by the relaxed sarcomere length, all averaged over at least eight contractions.

Calcium imaging
Cells were incubated at a final concentration of 2  (Teledyne Photometrics) and Nikon Elements (v.5-21-03). The emission due to excitation by 340 nm (F 340 ) was acquired for 20 ms and the emission due to excitation by 380 nm (F 380 ) was subsequently acquired for 20 ms. To minimize bleaching, however, fluorescence was acquired for only 10 s before superfusion of isoproterenol or forskolin, for 10 s after a 2-min superfusion of isoproterenol (100 nM) or forskolin (5 μM) and for 30 s during the superfusion of 10 mM caffeine at the conclusion of the experiment. The entire experiment (perfusion, recording and pacing) was automated using Nikon Elements, STV-2-4MX-1 valve (Takasago) and MyoPacer (IonOptix) to eliminate user variability during acquisition.
After the acquisition, cardiomyocytes were demarcated using Nikon Elements software. An area was selected without cardiomyocytes for background subtraction. Analysis was completed using a customized MATLAB (R2021b) script, which identified the basal diastolic fluorescence ratio (F 340 :F 380 ) and the pacing-induced Ca 2+ transient, ΔF (difference of F 340 :F 380 ratio of basal), before pacing (diastole) and at the peak after field stimulation (systole) for both pre-and post-infusion of isoproterenol or forskolin for the middle seven to eight of ten acquired transients. The time of fluorescence signal decay (t) to the levels of 37% peak for each transient was calculated in MATLAB and used to determine intracellular Ca 2+ decay kinetics. ΔF caffeine was the difference of peak F 340 :F 380 post-caffeine versus basal F 340 :F 380 pre-caffeine, which is proportional to SR Ca 2+ load.

Echocardiograms
Mice, aged 6-20 weeks, were anesthetized with 1-2% inhalational isoflurane and transthoracic echocardiography was performed using a 25-to 55-MHz linear-array transducer probe with a digital ultrasound system (Vevo 3100 Image System, VisualSonics). Vevo LAB 3.0 ultrasound analysis software (Fujifilm, VisualSonics) was used to measure and analyze images. Each echocardiographic parameter is the average of two measurements obtained from different cines. Vevo Strain speckle tracking and analysis software were used to calculate left ventricular strain values by the formula ((l 1 − l 0 )/l 0 ) × 100. Ejection fraction was calculated in Vevo Strain using Simpson's biplane method. GLS was calculated from an apical four-chamber view as follows: GLS = (L (t) − L 0 )/L 0 , where L 0 is the length of the heart in end-diastole. GCS was similarly calculated from a parasternal short-axis view using circumference rather than length. In a subset of mice, 2 mg kg −1 of isoproterenol via intraperitoneal injection was administered and echocardiograms were reacquired 2-5 min post-injection.

Telemetry and ECG analysis
Telemetry devices (Data Sciences International, model ETA-F10) were implanted in 6to 10-week-old mice. Recordings were begun 1 week after implantation. Intervals were measured using Ponemah software. Mice were housed in individual cages after telemeter implantation and for the entire experiment, on 12-h light:12-h dark cycle. Saline (control) or 2 mg kg −1 of isoproterenol dissolved in saline was administered via intraperitoneal injection.
ECGs were recorded both before and after injections. Heart rate values were averaged over 3-min time periods.

Exercise treadmill testing
Exercise testing was performed in the Mouse NeuroBehavior Core of the Institute of Genomic Medicine at Columbia University using established protocols. Male and female mice, aged 2-4 months, were acclimated to the treadmill (Panlab/Harvard Apparatus Treadmill) for 2 d (training). During training, the speed of the treadmill was gradually increased from 5 cm s −1 to 20 cm s −1 for up to 20 min, unless mice failed. Mice were considered to have failed the training if they spent 5 s consecutively in the 'fatigue zone', defined as one body length area toward the end of the belt, and/or if they received more than ten shocks. After training, the mice were challenged in a more difficult trial. On the challenge day, mice were run on the treadmill with no incline at 25 cm s −1 for 20 min.

Proximity proteomics and MS
Proximity labeling was performed 51 with minor modifications 15,52 . Isolated ventricular cardiomyocytes from mice expressing rabbit α 1C -APEX2 were incubated in labeling solution with 0.5 μM biotinphenol (Iris-biotech) for 30 min. During the final 10 min of labeling, 1 μM isoproterenol or 100 nM calyculin A was added. To initiate labeling, H 2 O 2 (Sigma-Aldrich, catalog no. H1009) was added to a final concentration of 1 mM. Exactly 1 min after H 2 O 2 treatment, the labeling solution was decanted and cells were washed 3× with cold quenching solution containing 10 mM sodium ascorbate (VWR 95035-692), 5 mM Trolox (Sigma-Aldrich, catalog no. 238813) and 10 mM sodium azide (Sigma-Aldrich, catalog no. S2002). After cells were harvested by centrifugation, the quenching solution was aspirated, and the pellet was flash-frozen and stored at −80 °C until streptavidin pull-down.
Subsequent protein processing procedures and MS analysis were performed as described 15,[53][54][55][56] . The digested peptides were labeled with TMTpro 16-plex (Thermo Fisher Scientific, catalog no. A44520) for 1 h. Data collection followed a MultiNotch MS 3 TMT method 57 using an Orbitrap Lumos mass spectrometer coupled to a Proxeon EASY-nLC 1200 liquid chromatography system (both Thermo Fisher Scientific) 15,55 . Peptides were searched with SEQUEST (v.28, rev. 12)-based software against a size-sorted forward and reverse database of the Mus musculus proteome (Uniprot 07/2014) with added common contaminant proteins and rabbit α1C sp|P15381|CAC1C_RABIT. For this, spectra were first converted to mzXML. Searches were performed using a mass tolerance of 20 p.p.m. for precursors and a fragment ion tolerance of 0.9 Da. For the searches maximally two missed cleavages per peptide were allowed. Carboxyamidomethylation on cysteine was set as a static modification (+57.0214 Da) and we searched dynamically for oxidized methionine residues (+15.9949 Da). We applied a target decoy database strategy and an FDR of 1% was set for peptide-spectrum matches after filtering by linear discriminant analysis 55,58 .
The FDR for final collapsed proteins was 1%. MS 1 data were calibrated post-search and searches performed again. Quantitative information on peptides was derived from MS 3 scans. Quantitative tables were generated requiring an MS 2 isolation specificity of >70% for each peptide and a sum of TMT (tandem mass tags) signal:noise ratio (s:n) of >200 over all channels for any given peptide, and then exported to Excel and further processed therein. Proteomics raw data and search results were deposited in the PRIDE archive 59,60 . The relative summed TMT s:n for proteins between two experimental conditions was calculated from the sum of TMT s:n for all peptides of a given protein quantified.

Statistics
The results are presented as the mean ± s.e.m. For multiple-group comparisons, a one-way analysis of variance (ANOVA) followed by multiple-comparison testing were performed. For comparisons between two groups, an unpaired, two-tailed Student's t-test was used.
Statistical analyses were performed using Prism 8 (Graphpad). Differences were considered statistically significant at values of P < 0.05.

Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Supplementary Material
Refer to Web version on PubMed Central for supplementary material.

Data availability
RNA-seq data have been uploaded to the Gene Expression Omnibus (accession no. GSE198903). Proteomics raw data and search results were deposited in the PRIDE archive and can be accessed via the ProteomeXchange under accession no. PXD033492. All other data are available in the main text and related files. Source data are provided with this paper.  anti-phospho-S2808 RyR2 antibody (upper) and anti-RyR2 antibody (middle) western blots. from left to right). e,i, Representative M-mode echocardiographic recordings before (e) and after (i) intraperitoneal injection of isoproterenol (ISO). f,g, Graphs of ejection fraction and fractional shortening from echocardiograms. Data are the mean ± s.e.m. ****P < 0.0001 by unpaired, two-tailed Student's t-test (n = 21 mice for each group). h, Graphs of GCS and GLS. Data are the mean ± s.e.m. **P < 0.01, ****P < 0.0001 by unpaired, two-tailed Student's t-test (n = 20, 23, 20 and 21 mice, from left to right). j, Graph of ejection fraction before and after isoproterenol. Data are the mean ± s.e.m. ****P < 0.0001 and P = 0.29 a, Graph of minimum, mean and maximum heart rate over a 24-h period. Data are the mean ± s.e.m. *P < 0.05 by unpaired, two-tailed Student's t-test (n = 4 WT mice; 7 4SA-Rad mice). b, Graph of heart rate before and 10 min after intraperitoneal injection of isoproterenol. Data are the mean ± s.e.m. *P < 0.05, ***P < 0.001 by unpaired, twotailed Student's t-test. c, Representative diary plots of heart rate during 4-h period after intraperitoneal injection of isoproterenol. The black trace is a WT mouse and the red trace a 4SA-Rad mouse. Representative ECGs are shown for the indicated timepoint. The arrows point to atrial activations (P waves). d, Graph of fraction of time in which heart rates were <400 beats min −1 (bpm), 380 bpm or 350 bpm during a 24-h period without isoproterenol and a 3-h period post-isoproterenol injection. Data are the mean ± s.e.m. *P < 0.05, **P < 0.01 by unpaired, two-tailed Student's t-test. Specific P values can be found in the associated Source data (see Supplementary Information) (n = 4 WT and 7 4SA-Rad mice). a, Schematic depicting mutation of the Ca V β subunit (DA-β) that prevents Rad binding. b, Anti-β subunit antibody (upper) and anti-β-actin (middle) antibody western blots. Lower: graph of densitometry, normalized to β-actin. Data are the mean ± s.e.m. *P < 0.05 two-tailed, unpaired Student's t-test (n = 3 mice in each group). c, Anti-α 1C , anti-β and anti-FLAG antibody western blots of anti-α 1C immunoprecipitation, representing three similar experiments. Ig, heavy chain of immunoglobulin. d, Graph of basal conductance density at −20 mV acquired using a ramp protocol. Data are the mean ± s.e.m. *P < 0.05; ****P < 0.0001 two-tailed, unpaired Student's t-test (n = 28, 24 and 24 cardiomyocytes from 3, 4 and 3 mice, respectively, from left to right). e, Graph of V 50 . P < 0.0001 by one-way ANOVA; ****P < 0.0001 by Sidak's multiple-comparison test. WT: 29, 29, 23, 23, 27 and 27 cells from 3, 3, 4, 4, 3 and 3 mice, respectively, from top to bottom. f, Ratio of Ba 2+ currents after isoproterenol to before isoproterenol versus voltage. Data are the mean ± s.e.m. Sample size same as e. g, Graph of basal Ca 2+ transient amplitude. The dashed blue line is the mean for WT + isoproterenol. The WT data are the same as in Fig. 3i. Data are the mean ± s.e.m. ****P < 0.0001 by unpaired, two-tailed Student's t-test (n = 462, 151, 245, 316 cardiomyocytes from 8, 3, 3 and 6 mice, respectively, from left to right). h, Scatter plot of isoproterenol-induced fold-change of Ca 2+ transient amplitude versus basal transient amplitude in WT (gray-data same as Fig. 3d) and 2DA-β 2B (green) myocytes. The dashed black line is a single-exponential fit of WT data (n = 225 WT cells, 4 mice; 151 2DA-β 2B cells, 3 mice). i, Scatter plot of isoproterenol-induced fold-change of Ca 2+ transient amplitude versus basal transient amplitude. The blue dashed line is a single-exponential fit of transgenic WT β 2B data (n = 106 WT-β 2B cells, 3 mice; 148 3DA-β 2B cells, 3 mice). j, Representative M-mode recordings. k, Graph of fractional area of change. The dashed lines are means for WT mice without and with isoproterenol. Data are the mean ± s.e.m. ****P < 0.0001 by unpaired, two-tailed Student's t-test (n = 5 mice from each group).