A perinuclear calcium compartment regulates cardiac myocyte hypertrophy

Abstract

The pleiotropic Ca²⁺/calmodulin-dependent phosphatase calcineurin is a key regulator of pathological cardiac myocyte hypertrophy. The selective activation of hypertrophic calcineurin signaling under stress conditions has been attributed to compartmentation of Ca²⁺ signaling in cardiac myocytes. Here, perinuclear signalosomes organized by the scaffold protein muscle A-Kinase Anchoring Protein β (mAKAPβ/AKAP6β) are shown to orchestrate local Ca²⁺ transients, inducing calcineurin-dependent NFATc nuclear localization and myocyte hypertrophy in response to β-adrenergic receptor activation. Fluorescent biosensors for Ca²⁺ and calcineurin and protein kinase A (PKA) activity, both diffusely expressed and localized by nesprin-1α to the nuclear envelope, are used to define an autonomous mAKAPβ signaling compartment in adult and neonatal rat ventricular myocytes. Notably, β-adrenergic-stimulated perinuclear Ca²⁺ and PKA and CaN activity transients depended upon mAKAPβ expression, while Ca²⁺ elevation and PKA and CaN activity in the cytosol were mAKAPβ independent. Buffering perinuclear cAMP and Ca²⁺ prevented calcineurin-dependent NFATc nuclear translocation and myocyte hypertrophy, without affecting cardiac myocyte contractility. Additional findings suggest that the perinuclear Ca²⁺ transients were mediated by signalosome-associated ryanodine receptors regulated by local PKA phosphorylation. These results demonstrate the existence of a functionally independent Ca²⁺ signaling compartment in the cardiac myocyte regulating hypertrophy and provide a premise for targeting mAKAPβ signalosomes to prevent selectively cardiac hypertrophy in disease.

1. Introduction

The second messenger Ca²⁺ is central to the function of the cardiac myocyte, not only for excitation-contraction coupling, but also for regulation of metabolism, cell death and survival, and gene expression [1,2]. In particular, Ca²⁺ activates the Ca²⁺/calmodulin-dependent serine-threonine phosphatase calcineurin (CaN, also known as protein phosphatase 2B and protein phosphatase 3), a signaling enzyme required for the induction of pathological cardiac hypertrophy [3]. Cytosolic Ca²⁺ cycles between 100 nmol/L and 1–2 μmol/L with each cardiac myocyte contraction [4]. However, CaN is not usually active in the normal myocyte, despite its potential to being activated by systolic Ca²⁺ levels [3]. Multiple mechanisms have been proposed to explain this paradox, including a requirement for persistently elevated time-averaged Ca²⁺ levels sufficient for CaN activation, as might be present in diastolic dysfunction [1,3]. However, consistent with in vivo models in which elevated cytosolic Ca²⁺ is not associated with activated CaN [5], cardiac myocyte CaN activation is likely restricted to spatially segregated, insulated Ca²⁺ microdomains. Evidence supporting the existence of these microdomains includes the identification of scaffold proteins that can co-localize CaN and relevant CaN substrates in different cellular compartments and the identification of ion channels that specifically activate CaN signaling without apparently serving an important role in excitation-contraction coupling [3,6]. Defining the molecular architecture and function of these putative microdomains should inspire novel therapeutic strategies for CaN inhibition that selectively inhibit pathological cardiac hypertrophy, while not adversely affecting cardiac contractility and other essential CaN-dependent cellular processes.

mAKAPβ is a 230 kDa scaffold protein localized in the cardiac myocyte to the outer nuclear membrane by direct binding to the integral membrane protein nesprin-1α [7,8]. There, mAKAPβ both organizes signalosomes that modulate the activity of transcription factors and class IIa histone deacetylases regulating gene expression and contributes to the regulation of the microtubule cytoskeleton [9,10]. Originally identified by its binding to PKA, mAKAPβ binds in addition type 5 adenylyl cyclase (AC5), type 4D3 cAMP-specific phosphodiesterase (PDE4D3), and the cAMP effector Epac1, thereby orchestrating a signalosome capable of autonomous, local cAMP signaling. As a large multivalent scaffold, mAKAPβ also organizes signaling modules for mitogen-activated protein kinase, phosphatidylinositol, and Ca²⁺ signaling [11]. In particular, mAKAPβ binds CaN, including the CaN Aβ catalytic subunit required for pathological cardiac hypertrophy [12,13]. CaN binding to mAKAPβ is distinguished from that to other scaffolds as CaN-mAKAPβ binding is Ca²⁺/calmodulin-dependent, increased by adrenergic stimulation, and does not result in inhibition of the phosphatase [14]. In myocytes mAKAPβ facilitates the association of CaN with nuclear factor of activated T-cells (NFAT) and myocyte enhancer factor 2 (MEF2), both of which are required for pathological cardiac remodeling [2]. In addition, mAKAPβ is required for the dephosphorylation and activation of these transcription factors by CaN [14–16]. Accordingly, deletion of the discrete binding site within mAKAPβ for CaN, as well as inhibition of CaN-mAKAPβ binding using competitive binding peptides, inhibited the adrenergic receptor-induced hypertrophy of neonatal rat ventricular myocytes in vitro [14,17].

In vivo evidence for mAKAPβ regulation of pathological cardiac remodeling has been obtained in mice by inducible mAKAPβ gene deletion and by treatment with a mAKAP small hairpin RNA (shRNA) adeno-associated virus gene therapy vector [18,19]. Cardiac myocyte-specific mAKAPβ targeting in the adult inhibited cardiac hypertrophy induced by chronic catecholamine infusion, pressure overload and myocardial infarction. In addition, scaffold targeting also prevented the development of heart failure and improved the survival of mice subjected to long term pressure overload. With regards to CaN signaling, NFAT activation induced by pressure overload was attenuated by mAKAPβ knock-out [18]. However, despite biochemical evidence obtained in vitro and analysis of mAKAPβ-directed mouse models, it remains unclear whether mAKAPβ signalosomes are located within a discrete Ca²⁺ compartment and how CaN at mAKAPβ is activated. This issue is compounded by evidence suggesting that nesprin-1α localizes mAKAPβ, mainly if not exclusively, to the outer nuclear membrane [7,8], where mAKAPβ signalosomes might be exposed to cytosolic Ca²⁺ oscillations. Like CaN, PKA binding to mAKAPβ was required for the induction of neonatal myocyte hypertrophy [12]. As mAKAPβ signalosomes include the PKA-stimulated, Ca²⁺ release channel ryanodine receptor (RyR2) [20,21], we have proposed that at mAKAPβ signalosomes, PKA-dependent phosphorylation of RyR2 will result in elevated Ca²⁺, activating local CaN and inducing myocyte hypertrophy via activation of relevant CaN-dependent transcription factors. We now provide evidence for this model obtained by live cell imaging using genetically encoded fluorescent reporters and tools designed to specifically alter signaling at perinuclear mAKAPβ signalosomes. These results suggest that CaN activity responsible for NFAT activation is maintained in pathologic states by local, cAMP-dependent perinuclear Ca²⁺ transients restricted to a compartment where Ca²⁺-CaN signaling may be targeted without adversely affecting myocyte contractility.

2. Materials and methods

2.1. Animal studies

Sprague Dawley Rats were purchased from Charles River. All research described herein was approved by the Institutional Animal Care and Use Committee at the University of Connecticut Health Center and conformed to the NIH Guide for the Care and Use of Laboratory Animals.

2.2. Adenoviruses

All adenoviruses were generated by transfection of HEK293 cells with Adeno-X plasmids (Clontech Adeno-X Tet-Off Expression System 1) into which genes of interest were subcloned using the I-Ceu I and PI-Sce I restriction sites. Adenovirus were purified using Vivapure^® Adeno-PACK^™ 20 kits (Sartorius), and titers were determined by end-point dilution method for HEK293 cell viability. Adenovirus for mAKAP and control shRNA were as previously described [12]. The adenovirus shuttle vectors pS-mCherry-Parv-Nesprin, pS-mCherry-Nesprin, pSCameleon, pS-AKAR4, and pS-AKAR4-nesprin, that direct cDNA expression under control of the cytomegalovirus immediate early promoter, were as previously described [22–24]. New plasmids were constructed by Genewiz using the method of the company’s choice. pS-mCherry-PDE4D_C(−)(ERK-)-nesprin expressing an inactive mutant PDE4D catalytic domain was the same as the previously described pS-mCherry-PDE4D_C (ERK-)-nesprin expression plasmid that contains a cDNA for PDE4D3 aa 225–673 (NP_006194.2) with missense mutations K455A/K456A/S579A/F597A/Q598A/F599A [23], but now with an additional T349A mutation. pTRE-GcAMP6S (containing the conditional TRE promoter and requiring co-infection with Adeno-tTA virus, Clontech Adeno-X Tet-Off Expression System 1) and pS-CaNAR2 contain cDNAs for Ca²⁺ and CaN activity sensors derived from pGP-CMVGCaMP6s (RRID:Addgene_40753, a gift from Douglas Kim & GENIE Project [25]) and pCDNA3-CaNAR2 (RRID:Addgene_64728, a gift from Jin Zhang [26]), respectively. pS-Cameleon-nesprin is the same as pSCameleon except that a human nesprin-1α cDNA (AAN60442.1; aa 7799–8797) is included in frame 3′ to the Cameleon cDNA. pS-CaNAR2-nesprin and pS-GCaMP6s-nesprin are similar to pS-Cameleon-nesprin except that the Cameleon cDNA has been replaced with a CaNAR2 [26] or GCaMP6s [25] cDNA [25]. Adenovirus for NFATc1-GFP was acquired from Seven Hills Bioreagents (Catalog no. JMAd-98). Plasmid sequences and maps are available upon request.

2.3. Antibodies

The following were used for immunocytochemistry: mouse anti-α-actinin (monoclonal EA-53, Sigma), rabbit anti-Atrial Natriuretic Peptide (polyclonal AB5490, Sigma), Goat anti-mouse IgG (H + L) Alex Fluor 488 (A-11001, Thermofisher), Goat anti-rabbit IgG (H + L) Alex Fluor 488 (A32731, Thermofisher), and Donkey anti-rabbit IgG (H + L) Alexa Fluor 555 (A31572, Thermofisher). All antibodies used for immunocytochemistry were used at 1:1000 dilution of liquid stocks delivered from supplier in phosphate-buffered saline (PBS) containing 0.2% BSA and 1% horse serum. SlowFade Diamond Antifade Mountant with DAPI (S36964, ThermoFisher) was used one drop per slide. For Western blotting, RyR2 (Badrilla, A010–35AP), RyR2 pSer2808 (Badrilla, A010–30), RyR2 pSer2030 (Badrilla, A010–32), IP₃ receptor (Abcam, ab108517), and polyclonal mCherry (Invitrogen, PA5–34974) antibodies were used at 1:1000 dilutions of the stock solution as supplied or following reconstitution in buffer. The previously described mAKAP FL100 and nesprin OR009 antibodies were used at 1:10,000 dilution [7,18].

2.4. Neonatal rat ventricular myocyte isolation and culture

Neonatal rat ventricular myocytes were isolated as previously described from 2 to 3-day old Sprague-Dawley pups [12,14,27]. Euthanasia for neonatal rats was by decapitation. After dissociation of heart tissue through several cycles of trypsin wash and serum neutralization, cells were collected by centrifugation and strained with a 70 μm mesh cell strainer. Cells were pre-plated for 2 h to remove fibroblasts. Unattached cells were collected by centrifugation and plated on 1% gelatin-coated plates (500,000 myocytes per 35 mm plate) in Dulbecco’s Modified Eagle Medium: Nutrient Mixture F-12 (DMEM/F12) supplemented with 1% penicillin/streptomycin (Gibco-BRL), 10% horse serum (HS), and 5% fetal bovine serum (FBS). After 24 h, cells were washed and cultured in Maintenance Medium (serum-free DMEM/F12 with penicillin/streptomycin). Cells were either infected with adenovirus or treated with drugs as indicated.

2.5. Adult rat ventricular myocyte isolation and culture

Adult rat ventricular myocytes were isolated and cultured as previously described using 2–3-month old male and female Sprague-Dawley rats [28]. Prior to tissue collection, rats were anti-coagulated with 300 U heparin intraperitoneal injection. 30 min later, rats were anesthetized with ketamine (80 mg/kg) and xylazine (8 mg/kg) for heart excision. Hearts were placed into chilled perfusion buffer (mmol/L: NaCl 120, KCl 5.4, Na₂HPO₄ 1.2, NaHCO₃ 20, MgCl₂ 1.6, Taurine 5, Glucose 5.6, 2,3-Butanedione monoxime 10), equilibrated with 95% O₂ and 5% CO₂). The heart was cannulated via the aorta to the condenser of a Harvard Langendorff apparatus and perfused with buffer at 37 °C at a constant rate of 2.2 mL/min for 5 min, followed by perfusion for 45 min with 50 mL digestion buffer (perfusion buffer with 120 mg type II collagenase (Worthington, 315 U/mg), 5 mg protease (Sigma type XIV), and 55 mg BSA), recycling enzyme once reduced to 30 mL remaining. Once the hearts became soft, the atria were removed and the ventricles cut into pieces before suspension in 5 mL digestion buffer. Myocytes were dissociated by trituration using a large bore pipette and filtered through 150–200 μm nylon mesh before collection by centrifugation. The supernatant was discarded, and the cells subjected to gradual Ca²⁺ stepwise addition (0.25, 0.5, and 1 mmol/L Ca²⁺). The remaining myocytes were plated (100,000 myocytes per 35 mm plate) on laminin (Corning, 10 μg laminin per dish) coated plates for 1.5 h followed by washing and maintenance in ACCT medium [Medium 199, 5 mmol/L Creatine (Sigma), 2 mmol/L l-carnitine (Sigma), 5 mmol/L Taurine (Sigma), 25 mmol/L Hepes (Sigma), 10 mmol/L 2, 3-Butanedione monoxime (ACROS Organics), 1% penicillin/streptomycin, 0.2% BSA (fatty acid free, Sigma), 0.1% ITS]. Adult myocytes were infected with adenovirus the same day as preparation, with drug treatments occurring the next day.

2.6. Live cell imaging

For imaging of both parent and nesprin-1α-targeted Cameleon, CaNAR2 and AKAR4 Forster resonance energy transfer (FRET) sensors, adult and neonatal myocytes were infected with adenovirus (multiplicity of infection [MOI] 5–50) the day after plating and were imaged within the subsequent two days. Cells were washed once with and then placed in Hanks’ Balanced Salt Solution (Gibco, mmol/L: KH₂PO₄ 0.44; Na₂HPO₄ 0.34; NaHCO₃ 4.2; NaCl 138, KCl 5.3; d-glucose 5.6; CaCl₂ 1.3; MgCl₂ 0.49; MgSO₄ 0.41). As Ca²⁺ transients were highly variable in the presence of extracellular Ca²⁺, myocytes expressing Cameleon or Cameleon-nesprin were imaged in Ca²⁺- and Mg²⁺-free Hanks’ Balanced Salt Solution (Gibco, as above without CaCl₂, MgCl₂, and MgSO₄). Cells were imaged on a Zeiss Pascal confocal microscope using a 40×/1.2 numerical aperture objective, a 440 nm laser (Toptica Photonics), and HQ535/50 M and HQ480/40 M emission and 510DCLP dichroic filters (Chroma Technology). Images were acquired at 15 s intervals, with the exception of CaNAR2 and CaNAR2-nesprin imaging for which images were acquired at 1-min intervals. FRET for regions of interest was quantified using background-subtracted images and Image J, with statistical analyses performed using Graphpad Prism 8. FRET ratio “R” was defined as net FRET ÷ donor signal and normalized to R₀ (ratio for time = 0). For each experiment, traces were obtained for cells obtained from at least 3 different myocyte preparations.

Adult myocytes expressing the intensiometric GCaMP6s and GcaMP6s-nesprin sensors were imaged in Ca²⁺- and Mg²⁺-free Hanks’ Balanced Salt Solution on an inverted Zeiss LSM800 confocal microscope with a 63×/1.4 numerical aperture oil immersion objective and a 488 nm laser, with images acquired at 5 s intervals. Fluorescent signal “F” reflecting Ca²⁺ binding to the intensiometric GCaMP6s sensor was calculated for regions of interest using background-subtracted images and Image J and analyzed using Graphpad Prism 8, with normalization to fluorescence signal at time = 0 (F₀).

2.7. Immunoprecipitation

Confluent 60 mM plates of approximately 1 million neonatal myocytes were washed twice with PBS and lysed with 1 ml HSE buffer (20 mmol/L HEPES, pH 7.4, 150 mmol/L NaCl, 5 mmol/L EDTA, 1% Triton X-100) supplemented with protease inhibitors (AEBSF, benzamidine, leupeptin/pepstatin) and okadaic acid (500 nmol/L). Following centrifugation at 13.2 krpm at 4 °C, soluble lysates were incubated with 5 μL mAKAP or nesprin antibodies in the presence of 25 μLTrueBlot Anti-Rabbit Ig agarose Beads (Rockland) overnight at 4 °C. Beads were pelleted and washed three times with HSE buffer and boiled in 25 μL 2× SDS loading buffer. Samples were size-fractionated by 5% (for RyR2) or 7.5% SDS-PAGE and transferred to nitrocellulose membranes. Blots were blocked in 5% milk for one hour, followed by incubation in primary antibody overnight at 4 °C. Following washes, membranes were incubated with secondary antibodies (Rabbit Trueblot Anti-Rabbit IgG HRP, Rockland, 1:5000) for one hour. Signals were visualized with an enhanced chemiluminescence reagent (Abcam) and exposed to X-ray film. Quantification was determined by ImageJ analysis.

2.8. Immunocytochemistry

Myocytes on glass coverslips were fixed with 3.7% formaldehyde in PBS, 10 min for neonatal, 1 h for adult myocytes. Cell were permeabilized with 0.3% Triton X-100 in PBS, and blocked with PBS containing 0.2% BSA and 1% horse serum for 30 min. Slides were incubated for 1 h with primary antibodies, followed by 1-h incubation with Alexa fluorescent dye-conjugated specific-secondary antibodies. Blocking buffer was used for antibody incubations and subsequent washes. Slide were mounted with SlowFade Diamond Mountant with DAPI. Widefield fluorescent images were acquired by a Zeiss Observer Z1 fluorescent microscope with an Axiocam camera.

NFATc1 Localization and ANF expression assays:

Neonatal myocytes were infected the day after plating with adenovirus as indicated (MOI 5–50). The next day, cells were washed with maintenance medium and drugs added as indicated, with overall culture time as indicated. For each slide, at least 25 cells were examined for NFATc1 localization or for perinuclear ANF staining. NFATc1-GFP expressing cells were scored for those with fluorescent intensity greater within the nucleus or the cytosol.

Myocyte Hypertrophy Assay:

Adult myocytes were infected adenovirus as indicated for 24 h (MOI 5–50), followed by washing with ACCT media, and treatment with drug as indicated. For each slide, at least 22 cells were measured for maximum length and width.

2.9. Cell shortening

Adult myocytes were perfused with Tyrode solution (137 mmol/L NaCl, 5.4 mmol/L KCl, 1.8 mmol/L CaCl₂, 0.5 mmol/L MgCl₂, 10 mmol/L Hepes, 10 mmol/L glucose, pH 7.4), with 1 μmol/L isoproterenol added as indicated. Cell shortening was elicited by field stimulation at 1 Hz at room temperature and measured by video edge detection using a sequential-scanning video camera attached to a fluorescent microscope (Crescent Electronics, Sandy, UT). The camera was rotated during imaging to align the parallel video detector raster lines with the myocyte long axis. Each cell was measured for maximum cell width and length, with calibration provided by 10 μm beads under the same magnification, and monitored for changes in myocyte length with each contractile cycle. Data were analyzed using AxoScope9.0 (Axon) and represent the average of 6 traces from 3 separate myocyte preparations for each condition.

2.10. Statistical analysis

Statistics were calculated using Graphpad Prism 8. All data are expressed as mean ± s.e.m. Unpaired, two-tailed t-tests or two-way ANOVA (four groups) followed by Tukey’s multiple comparison post-hoc testing was performed as appropriate. Repeated symbols used as follows: * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001; n.s. - not significant.

3. Results

3.1. mAKAPβ is located within an independent perinuclear Ca²⁺ compartment

We have previously shown that in cultured neonatal rat ventricular myocytes CaN is recruited to the perinuclear mAKAPβ scaffold over the first hour following adrenergic stimulation [14]. To assay Ca²⁺ levels and CaN activity at mAKAPβ signalosomes in real time in live myocytes using FRET biosensors, we fused the Ca²⁺ sensor Cameleon (D3cpv, K’_d = 0.6 μmol/L) and the CaN activity reporter CaNAR2 to the N-terminus of nesprin-1α (Fig. 1A) [26,29]. In contrast to the parent sensors, Cameleon- and CaNAR2-nesprin were restricted to the nuclear envelope when expressed in myocytes using adenovirus (Fig. 1B–C).

Fig. 1.

Identification of an mAKAPβ-dependent perinuclear Ca²⁺ compartment regulating local calcineurin activity. A, Compartment-specific biosensors: D3cpv Cameleon contains mutant calmodulin (mCam) and Ca²⁺/calmodulin-binding domain of myosin light chain kinase (mMLCK) inserted between ECFP and cpVenus fluorescent proteins, such that Ca²⁺-dependent mCaM binding of the mMLCK peptide increases FRET signal [29]. CaNAR2 contains between Cerulean 3 and YPet fluorescent proteins the N-terminal regulatory domain of NFATc2, that is highly phosphorylated in unstimulated cells and that upon CaN dephosphorylation changes conformation increasing FRET signal [26]. The intensiometric Ca²⁺ sensor GCaMP6s contains a calmodulin-binding M13 peptide, circularly permuted green fluorescent protein (cpGFP), and a variant mCam, that upon Ca²⁺-dependent M13-mCam binding increases cpGFP fluorescence [25]. Nesprin-1α (blue bar) has spectrin-like repeat (SR) domains, that mediate mAKAPβ binding, and a C-terminal transmembrane Klarsicht, ANC-1, Syne Homology (KASH) domain, that results in nuclear envelope localization [7]. B, Neonatal rat ventricular myocytes infected with adenovirus expressing Cameleon, Cameleon-nesprin, or CaNAR2-nesprin were imaged by confocal microscopy. Cyan images are shown in grayscale. The outline of cells in this and other figures are hand-drawn in orange when the outline is not readily apparent on the presented fluorescent channel. Bar - 10 μm. C, Adult rat ventricular myocytes infected with adenovirus expressing GCaMP6s-nesprin, CaNAR2-nesprin, and parent CaNAR2 sensor as in B (GCaMP6s-nesprin – GFP channel). Bar – 10 μm. D, Pseudocolored images showing FRET ratio R (net FRET ÷ background-subtracted donor signal) for representative isoproterenol (Iso, 1 μmol/L) stimulated control shRNA tracings in D, E, and G. E-H, FRET imaging of Iso-treated neonatal myocytes co-infected with adenovirus for Cameleon (D), Cameleon-nesprin (E), or CaNAR2-nesprin (G,H) and either control (Ctrl) or mAKAP shRNA and in the presence or absence of the CaN inhibitor cyclosporin A (CsA, 1 μmol/L) as indicated. Representative tracings and peak amplitudes for FRET ratio (R normalized to baseline R₀) are shown. I–K, Imaging of Iso-treated adult myocytes co-infected with adenovirus for GCaMP6s-nesprin (I), CaNAR2-nesprin (J), or CaNAR2 (K) and either control (Ctrl shRNA) or mAKAP shRNA. For GCaMP6s-nesprin, intensiometric fluorescent signals (F) are provided normalized to baseline measurement (F₀). Each data point in peak amplitude plots represents separate tracings (n ≥ 6) obtained using cells from 3 independent myocyte preparations. Peak amplitude data were compared by unpaired, two-tailed t-test. ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001.

Stimulation of neonatal myocytes with the β-adrenergic agonist isoproterenol (Iso, 1 μmol/L) induced a generalized Ca²⁺ transient detected by the parent Cameleon sensor (Fig. 1D–E). Note that FRET imaging with Cameleon is too slow to detect action potentials associated with individual myocyte contractions, but rather is an assay for average Ca²⁺ concentration. As expected, an Iso-induced Ca²⁺ transient was also detected by the nuclear envelope-localized Cameleon-nesprin sensor (Fig. 1F). Adenovirus-mediated expression of a previously characterized small hairpin RNA (shRNA) was used to inhibit mAKAPβ expression [12]. Surprisingly, the perinuclear transient, but not the cytosolic transient, was lost upon mAKAPβ shRNA expression (Fig. 1E–F). These results demonstrated the existence of a β-adrenergic-regulated perinuclear Ca²⁺ compartment detectable by the nesprin-fusion sensor that is insulated from cytosolic Ca²⁺ and presumably dependent upon mAKAPβ signalosome formation.

Iso induced a CaNAR2-nesprin transient in neonatal myocytes that was blocked by the CaN inhibitor cyclosporin A (CsA, Fig. 1G). The increase in perinuclear CaN activity was relatively slow in onset consistent with the relatively slow recruitment of CaN to the mAKAPβ scaffold in stimulated myocytes [14]. Like the perinuclear Ca²⁺ transient, CaNAR2-nesprin signal was dependent upon mAKAPβ expression in neonatal myocytes (Fig. 1H). Together, these results demonstrated the existence of a perinuclear compartment for Ca²⁺ and CaN signaling at mAKAPβ signalosomes in neonatal myocytes.

While neonatal myocytes have proven useful for the study of CaN signaling in hypertrophy, there are significant differences in ultra-structure between neonatal and adult myocytes, including in extent of the transverse tubule system and in Ca²⁺ handling [3]. To determine if a comparable perinuclear Ca²⁺ compartment was present in adult rat ventricular myocytes, the higher sensitivity, intensiometric Ca²⁺ sensor GCaMP6s (K_d = 144 nmol/L) [25] was fused to nesprin-1α (Fig. 1A–B). Like in neonatal myocytes, Iso induced in adult myocytes a perinuclear Ca²⁺ transient that was dependent upon mAKAPβ expression (Fig. 1I). Accordingly, Iso also induced in adult myocytes a CaNAR2-nesprin that was mAKAPβ-dependent (Fig. 1J). Although we were unable to detect an Iso induced transient with the diffusely localized parent GCaMP6s sensor in these un-paced, non-beating adult myocytes (Supplementary Figure), Iso induced CaN activity detected by the parent CaNAR2 in these cells was, as expected, not dependent upon mAKAPβ expression (Fig. 1K). These results confirm that there is a Ca²⁺-CaN signaling compartment at the nuclear envelope in cardiac myocytes organized by the mAKAPβ scaffold protein.

3.2. mAKAPβ-associated calcineurin is activated by local Ca²⁺

The existence of a perinuclear Ca²⁺ compartment distinct from cytosolic Ca²⁺ suggested that once bound to mAKAPβ, CaN activity would be maintained by Ca²⁺ localized to that compartment. Carp parvalbumin-β is a high-affinity (K_a = 29 nmol/L) EF-hand domain Ca²⁺ binding protein [30]. In recent investigations of neuronal axon growth, we inhibited depolarization-induced perinuclear Ca²⁺ elevation by expression of a mCherry-tagged parvalbumin β-nesprin fusion protein (mCherry-Parv-nesprin) designed to buffer Ca²⁺ in that intracellular compartment (Fig. 2A) [22]. Here we show that in neonatal myocytes mCherry-Parv-nesprin, but not a control mCherry-nesprin fusion protein blocked Iso-induced perinuclear Ca²⁺ transients (Fig. 2B). Like results obtained for mAKAP shRNA (Fig. 1D), perinuclear Ca²⁺ buffering with mCherry-Parv-nesprin had no effect on cytosolic Ca²⁺ transients detected by the parent Cameleon sensor (Fig. 2C). Notably, mCherry-Parvnesprin expression not only blocked perinuclear Ca²⁺ transients, but also completely inhibited Iso-induced CaNAR2-nesprin signals in both neonatal and adult myocytes (Fig. 2D–E). Importantly, mCherry-parv-nesprin had no effect on cytosolic CaN activity, as shown in adult myocytes (Fig. 2F). These results demonstrate that CaN activity at mAKAPβ signalosomes was dependent upon elevated local Ca²⁺ levels.

Fig. 2.

Local Ca²⁺ ion and ryanodine receptor activity are required for Ca²⁺ −CaN signaling in the perinuclear mAKAPβ – nesprin-1α compartment. A, Recombinant mCherry-tagged nesprin-1α fusion proteins for the study of perinuclear Ca²⁺ signaling. Carp parvalbumin-β contains 2 EF-hand domains that bind Ca²⁺ with high affinity (K_d ≈ 10⁻⁸ d mol/L) [30]. B–D, FRET imaging of Iso-stimulated (1 μmol/L) neonatal rat ventricular myocytes co-infected with adenovirus expressing Cameleon-nesprin (B), Cameleon (C), or CaNAR2-nesprin (D) and either mCherry-parv-nesprin or mCherry-nesprin control. E–F, FRET imaging of Iso-stimulated (1 μmol/L) adult rat ventricular myocytes co-infected with adenovirus expressing CaNAR2-nesprin (E) or CaNAR2 (F) and either mCherry-parv-nesprin or mCherry-nesprin control. G-H, Protein complexes were immunoprecipitated from neonatal myocyte extracts using antibodies for nesprin-1α and mAKAPβ and detected by Western Blot. n = 3. I–K, Imaging of neonatal myocytes expressing Cameleon-nesprin and treated with Iso (1 μmol/L), ryanodine (5 nmol/L or 5 μmol/L) or xestospongin C (3 μmol/L) as indicated. Representative tracings and peak amplitudes for FRET ratio (R normalized to baseline R₀) are shown for imaging experiments. Each data point in peak amplitude plots represents separate tracings (n ≥ 6) obtained using cells from 3 independent myocyte preparations. Peak amplitude data were compared by unpaired, two-tailed t-test. ** p ≤ 0.01; **** p ≤ 0.0001.

RyR2, but not IP₃-receptor Ca²⁺ release channels have been detected by co-immunoprecipitation assay to be associated with mAKAPβ-nesprin-1α signalosomes in cardiac myocytes (Fig. 2G–H) [7,20]. By locking open or occluding the channel pore, low dose (5 nmol/L) and high dose (5 μmol/L) ryanodine will promote and inhibit RyR2 channel currents, respectively [31,32]. To test whether RyR2 might be responsible for Ca²⁺ influx into the mAKAPβ compartment, neonatal myocytes expressing Cameleon-nesprin were stimulated with low dose ryanodine, inducing a readily detectable perinuclear Ca²⁺ transient that was not detected in cells stimulated with a high dose of the ligand (Fig. 2I). High dose ryanodine did, however, block Iso-induced Cameleon-nesprin transients (Fig. 2J). Consistent with the differential co-immunoprecipitation of the two Ca²⁺ release channels, the IP₃-receptor inhibitor xestospongin C (3 μmol/L) did not affect the Iso-induced transients (Fig. 2K). These results suggest that RyR2 is responsible for the release of Ca²⁺ into the perinuclear mAKAPβ compartment.

3.3. mAKAPβ-associated Ca²⁺ transients regulate myocyte hypertrophy, but not contractility

mAKAPβ signalosomes bind NFATc family members, facilitating NFATc dephosphorylation by CaN and nuclear translocation [12,14,18]. To determine whether the local Ca²⁺ transients at mAKAPβ signalosomes were required for signaling downstream of CaN, the nuclear translocation of adenovirus-expressed green fluorescent protein (GFP)-tagged NFATc1 was assayed in neonatal myocytes. Iso (1 μmol/L) stimulation for 24 h resulted in a near complete translocation of GFP-NFATc1 to the nucleus that was blocked by the mCherry-Parv-nesprin fusion protein (Fig. 3A–B). Accordingly, Iso-induced expression of the hypertrophy marker atrial natriuretic factor (ANF) was similarly inhibited by buffering of perinuclear Ca²⁺ (Fig. 3C–D).

Fig. 3.

Perinuclear Ca²⁺ in the mAKAPβ – nesprin-1α compartment ion is required for NFAT nuclear translocation and myocyte hypertrophy, but not myocyte contractility. A-B, Neonatal rat ventricular myocytes, infected with adenovirus expressing GFP-tagged NFATc1 and either mCherry-nesprin or mCherry-Parv-nesprin, were stimulated with Iso (1 μmol/L) for 24 h, before assay of NFAT localization by fluorescence microscopy. Fraction of cells with majority NFAT nuclear localization was determined. Bar - 10 μm. C—D, Neonatal myocytes expressing either mCherry-nesprin or mCherry-Parv-nesprin were stimulated for 48 h with Iso (1 μmol/L) before staining with atrial natriuretic factor antibody (ANF, red). Fraction of cells with prominent ANF staining was determined. E–F, Adult rat ventricular myocytes were infected with adenovirus expressing either mCherry-nesprin (Ctrl) or mCherry-Parv-nesprin (PDE) and stimulated for 48 h with Iso (1 μmol/L) before staining for α-actinin (green) and nuclei (DAPI, blue) and fluorescent imaging. Myocyte length and width were measured. Bar - 25 μm. In A-F each data point represents the mean value of >22 cells that were measured for each individual biological replicate; n = 3–6 independent myocyte preparations. G-H, Adult rat ventricular myocytes were infected with adenovirus expressing either mCherry-nesprin or mCherry-Parv-nesprin for 24 h prior to study of contractility. Cell shortening was elicited by field stimulation at 1 Hz and measured through video edge detection. Iso (100 nmol/L) stimulation was as indicated. n = 6 individual tracings using cells from 3 independent myocyte preparations. All data were analyzed by matched 2-way ANOVA and Tukey post-hoc testing. * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001.

Iso treatment of adult ventricular myocytes induces myocyte growth in both length and width (Fig. 3E–F) [24,28]. Like inhibition of perinuclear CaN activity (Fig. 2E), expression of the mCherry-Parv-nesprin completely inhibited Iso-induced adult myocyte hypertrophy. In contrast, mCherry-Parv-nesprin expression did not affect myocyte contraction (Fig. 3G–H). mCherry-Parv-nesprin had no effect on the shortening of paced (1 Hz) myocytes, either in the absence or presence of Iso (100 nmol/L) stimulation. Together, these results demonstrate that the mAKAPβ-dependent perinuclear Ca²⁺ compartment modulates CaN-NFAT signaling that selectively regulates cardiac myocyte hypertrophy.

3.4. Perinuclear cAMP-PKA signaling regulates CaN-NFAT signaling and hypertrophy

RyR2 Ca²⁺-release channels associated with mAKAPβ are PKA phosphorylated in response to β-adrenergic stimulation [20,21]. To assay in live myocytes mAKAPβ-bound PKA activity that might regulate local Ca²⁺-CaN signaling, we used adenovirus to express the Cerulean-cpVenus PKA activity reporter AKAR4 (AKAR-nesprin, Fig. 4A) in fusion to nesprin-1α [23,33]. Like Cameleon and CaNAR2, the parent AKAR4 sensor was expressed diffusely, while AKAR4-nesprin was localized to the nuclear envelope in neonatal myocytes (Fig. 4B). While adrenergic stimulation with norepinephrine (NE, 10 nmol/L) induced PKA transients detected by both the localized and parent AKAR4 reporters, only the AKAR4-nesprin transient was mAKAPβ-dependent (Fig. 4C–E), consistent with the direct binding and intracellular localization of mAKAPβ by nesprin-1α [7].

Fig. 4.

mAKAPβ organizes a perinuclear cAMP-PKA signaling compartment. A, AKAR4 consists of a forkhead-associated domain (FHA1) and PKA substrate (LRRATLVD) flanked by Cerulean and cpVenus; PKA substrate phosphorylation induces FHA1 binding, increasing FRET signal [33]. mCherry-PDE-nesprin contains a mutated PDE4D catalytic domain (T349A, K455/456A, S579A, FQF597–599AAA). B, Neonatal rat ventricular myocytes were infected with adenovirus for AKAR4 or AKAR4-nesprin and imaged by confocal microscopy. Bar - 10 μm. Cyan image is shown in grayscale. C, Pseudocolored images showing FRET ratio R (net FRET ÷ background-subtracted donor signal) for representative control shRNA tracings shown in D-G. D-E, FRET measurements for adenovirus-infected neonatal myocytes expressing AKAR4-nesprin (D) or AKAR (E) and either control (Ctrl) or mAKAP shRNA. F-G, FRET measurements for adenovirus-infected neonatal myocytes expressing AKAR4-nesprin (F) or AKAR (G) and either mCherry-PDE-nesprin or mCherry-nesprin control. D-G, Myocytes were stimulated with norepinephrine (NE, 10 nmol/L). Each data point in peak amplitude plots for FRET ratios (R/R₀) represents separate tracings (n ≥ 6) obtained using cells from 3 independent myocyte preparations. Peak amplitude data were compared by unpaired, two-tailed t-test. *** p ≤ 0.001; **** p ≤ 0.0001.

To inhibit perinuclear cAMP signaling in myocytes, we designed an mCherry-tagged nesprin-1α fusion protein containing a mutant phosphodiesterase catalytic domain (Fig. 4A). The PDE4D3-derived domain was mutated to prevent extracellular signal-regulated kinase (ERK) phosphorylation and to decrease catalysis (mutant residues: T349A, K455/456A, S579A, FQF597–599AAA; Fig. 4A) [34]. Expression of mCherry-PDE-nesprin, but not a control mCherry-nesprin fusion protein inhibited NE stimulated AKAR4-nesprin transients (Fig. 4F). Cytosolic PKA transients were not affected by the nuclear envelope-localized cAMP buffer (Fig. 4G), demonstrating the specificity of the mCherry-PDE-nesprin construct for the mAKAPβ compartment in myocytes.

Having found that RyR2 currents were involved in perinuclear Ca²⁺ signaling (Fig. 2G–I), we assayed whether PKA activity, that might potentiate RyR2 channel opening [35,36], was also required for perinuclear Ca²⁺ signaling. Pretreatment of neonatal myocytes with the PKA inhibitor H89 blocked both Iso-induced Ca²⁺ and CaN perinuclear transients (Fig. 5A–B). In addition, expression of the perinuclear cAMP buffer mCherry-PDE-nesprin, that inhibited local PKA activity (Fig. 4F), blocked β-adrenergic-induced CaNAR2-nesprin signals (Fig. 5C).

Fig. 5.

Perinuclear cAMP is required for mAKAPβ-associated calcineurin activity. A-B, FRET imaging of neonatal rat ventricular myocytes expressing Cameleon-nesprin (A) or CaNAR2-nesprin (B) were stimulated with isoproterenol (1 μmol/L) and inhibited with H89 PKA inhibitor (1 μmol/L) as indicated. C, FRET imaging of neonatal myocytes expressing CaNAR2-nesprin and either mCherry-PDE-nesprin or mCherry-nesprin control and stimulated with norepinephrine (NE, 10 nmol/L). For each panel, representative tracings for FRET ratio (R/R₀) are shown. Each data point in peak amplitude plots for FRET ratios (R/R₀) represents separate tracings (n ≥ 6) obtained using cells from 3 independent myocyte preparations. Peak amplitude data were compared by unpaired, two-tailed t-test. *** p ≤ 0.001; **** p ≤ 0.0001.

RyR2 serine residues 2808 and 2030 have been shown to be PKA phosphorylated and have been implicated in RyR2 regulation [37,38]. To determine if PKA-dependent regulation of perinuclear Ca²⁺ and CaN transients correlated with perinuclear RyR2 phosphorylation, Ser-2808 and −2030 phosphorylation of RyR2 in nesprin-1 antibody immunoprecipitates was assayed following β-adrenergic stimulation (Fig. 6). mAKAPβ expression was irrelevant to the phosphorylation of the bulk of cellular RyR2 detected in total neonatal myocyte extracts, that presumably represent RyR2 at the sarcoplasmic reticulum (Fig. 6A–B). Remarkably, mAKAPβ expression was required for the Iso-induced phosphorylation of nesprin-1-associated RyR2 at both PKA sites, while notably not required for RyR2 association with nesprin-1α. In addition, inhibition of perinuclear cAMP signaling using mCherry-PDE-nesprin inhibited Ser-2808 and −2030 RyR2 phosphorylation only for perinuclear RyR2 (in this case immunoprecipitated with mCherry antibody), but not for RyR2 in total extracts (Fig. 6C–D). These results imply that perinuclear RyR2-mediated Ca²⁺ transients associated with mAKAPβ signalosomes are both spatial segregated and differentially regulated from RyR2-mediated transients elsewhere in the cardiac myocyte, including presumably those essential for excitation-contraction coupling.

Fig. 6.

The phosphorylation of perinuclear ryanodine receptors by PKA is independently regulated by mAKAPβ and local cAMP signaling. Neonatal myocytes infected with adenovirus expressing mAKAP or control shRNA (A,B) or mcherry-PDE-nesprin or control mCherry-nesprin (C,D) were stimulated with isoproterenol (1 μmol/L) as indicated (hr – hours). Protein complexes immunoprecipitated with nesprin-1 (A,B) or mCherry antibody (C,D) were assayed for RyR2 Ser-2808 phosphorylation (A,C), RyR2 Ser-2030 phosphorylation (B,D) and total RyR2 protein by western blotting. Iso induced both Ser-2808 and −2030 RyR2 phosphorylation in both immunoprecipitates and in total extracts (p ≤ 0.0001), although only the phosphorylation of immunoprecipitated RyR2 was dependent upon mAKAP and perinuclear cAMP. n = 4, each using myocytes from an independent preparation. Data were analyzed by matched 2-way ANOVA and Tukey post-hoc testing. **** p ≤ 0.0001.

3.5. Perinuclear cAMP induced by the mAKAPβ – nesprin-1α signalosome is required for NFAT nuclear translocation and myocyte hypertrophy

To compare localized cAMP signaling at mAKAPβ signalosomes with downstream hypertrophic CaN signaling, the effect of mCherry-PDE-nesprin in neonatal myocytes on GFP-NFATc1 nuclear translocation (Fig. 7A–B) and ANF expression (Fig. 7C–D) was determined. Norepinephrine (NE) stimulation for 24 h resulted in a near complete translocation of GFP-NFATc1 to the nucleus, correlating with an increased expression of the hypertrophy marker atrial natriuretic factor (ANF). Both were blocked by the co-expression of the mCherry-PDE-nesprin fusion protein. Accordingly, inhibition of perinuclear cAMP signaling using mCherry-PDE-nesprin blocked the growth and width of adult rat ventricular myocytes induced by norepinephrine (NE) stimulation (Fig. 7 E–G). Together, these results suggest that perinuclear cAMP-PKA signaling is required for activation of the CaN-NFAT pathway that regulates cardiac myocyte hypertrophy.

Fig. 7.

Perinuclear cAMP at mAKAPβ – nesprin-1α signalosomes is required for NFAT nuclear translocation and myocyte hypertrophy. A-B, Neonatal myocytes were infected with adenovirus expressing GFP-tagged NFATc1 and either mCherry-PDE-nesprin or mCherry-nesprin control followed by stimulation with NE (10 nmol/L) for 24 h before staining with DAPI (blue) and imaging to determine NFAT localization. Representative images of myocytes with predominantly cytosolic and nuclear NFATc1-GFP are shown in A. Bar - 10 μm. C—D, Neonatal myocytes were infected with adenovirus expressing either mCherry-PDE-nesprin or mCherry-nesprin control followed by stimulation with NE (10 nmol/L) for 48 h before staining with ANF antibody and DAPI (blue). Representative images of ANF staining are shown in C. E-G, Adult rat ventricular myocytes were infected with adenovirus expressing either mCherry-nesprin or mCherry-PDE-nesprin and stimulated for 48 h with NE (10 nmol/L) before staining with α-actinin antibodies (green) and DAPI (blue) and fluorescent imaging. Representative images of NE-stimulated myocytes are shown in E. Bar - 25 μm. For all panels, each data point represents the mean value of >50 cells that were measured for each individual biological replicate; n = 3–7 independent myocyte preparations. Data were analyzed by matched 2-way ANOVA and Tukey post-hoc testing. * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001.

4. Discussion

Biochemical and physiological studies have implicated the perinuclear mAKAPβ signalosome in the regulation of gene expression responsible for cardiac myocyte hypertrophy [9]. Live cell imaging experiments using novel tools for compartment-specific signaling modulation now provide evidence that mAKAPβ organizes an independent Ca²⁺ signaling compartment that regulates CaN and NFAT transcription factor required for hypertrophy (Fig. 8). Depletion of mAKAPβ by RNA interference in neonatal myocytes resulted in a loss of β-adrenergic-stimulated Ca²⁺ transients at the nuclear envelope, but not in the cytosol, showing that elevated cytosolic Ca²⁺ could not activate a nesprin-targeted reporter (Fig. 1). Besides an elevation in local Ca²⁺, cAMP signaling at mAKAPβ was found to be required for CaN activation in that compartment. These results provide a mechanism for crosstalk between cAMP and Ca²⁺ second messenger pathways that together regulate stress-induced gene expression independently of myocyte contractility.

Fig. 8.

Model for mAKAPβ signalosomes forming an independent perinuclear Ca²⁺ and cAMP signaling compartment regulating local calcineurin activity, NFAT nuclear translocation, and myocyte hypertrophy. See text for discussion. Drawing created with BioRender.com.

Results obtained with the nesprin sensors differ from results we previously obtained using Ca²⁺ and CaN reporters fused to the membrane-associated protein Cdc42-interacting protein 4 (CIP4/TRIP10), that is expressed in a punctate pattern throughout the myocyte [24]. At CIP4, both Ca²⁺ and CaN FRET reporter signals were maximal within 1–6 min. In contrast, signals detected with the CaNAR2-nesprin sensor increased slowly over 15–40 min after addition of agonist. The main difference contributing to the delay in CaN dephosphorylation of the CaNAR2-nesprin sensor may be the relatively low basal stoichiometry of mAKAPβ-CaN binding and the slow recruitment of CaN to mAKAPβ following adrenergic stimulation [14]. Although CIP4-CaNAβ signalosomes also regulate pathological myocyte hypertrophy, surprisingly, CaNAβ signaling at CIP4 did not appear to be associated with direct NFAT activation [24]. In contrast, mAKAPβ signalosome targeting, both as performed here and by prior genetic deletion of the scaffold in vivo, robustly inhibited NFAT activation [18]. Consistent with the known translocation of CaN to the nucleus following activation [6], perinuclear sites like mAKAPβ signalosomes may be primarily responsible for later CaN signaling required to maintain or amplify hypertrophic signaling via NFAT-dependent signaling. In this regard, a requirement for CaN activity at mAKAPβ signalosomes would serve as a checkpoint for potentially deleterious hypertrophic signaling, such that only prolonged CaN activation would result in an overt pathophysiological phenotype.

While intracellular Ca²⁺ is often discussed in terms of average cytosolic concentration, Ca²⁺ is present at different concentrations in different regions of the myocyte due to Ca²⁺ buffering and the action of differentially localized Ca²⁺ pumps, transporters, and ion channels [1,39]. For example, nucleoplasmic Ca²⁺ oscillations, that are generally slower and lower in amplitude than those in the cytosol, are coupled by the passive permeability of nuclear pore complexes to cytosolic Ca²⁺ oscillations regulated by extra-nuclear RyR2 and sarco-endoplasmic reticulum Ca²⁺-ATPase 2a (SERCA2a) [4]. Nucleoplasmic transients, however, can be modulated in response to Gq-protein coupled receptor signaling by IP₃-receptors enriched on both the inner and outer cardiac myocyte nuclear membranes [4,40]. The association of IP₃-receptors with hypertrophic signaling in the nucleoplasm by CaN, as well as by Ca²⁺/calmodulin-dependent protein kinase, is well established [41,42]. In contrast, mAKAPβ signalosomes are located at the outer nuclear membrane both in neonatal and adult myocytes by binding nesprin-1α [7,8,10,18]. Remarkably, IP₃-receptors were not found to associate with mAKAPβ signalosomes, while RyR2 and perinuclear Ca²⁺ and cAMP appeared to co-regulate mAKAPβ signalosome signaling independently of any potential perinuclear IP₃-receptor activity. As RyR2 are not thought to be inserted into the nuclear envelope, RyR2 that associate with mAKAPβ and nesprin-1α are presumably constituents of perinuclear sarcoplasmic reticulum-transverse tubule dyads that the Franzini-Armstrong group showed by electron microscopy to be 200 nm or less from the outer nuclear membrane in adult myocytes [40]. Although it has been suggested that mAKAPβ can bind RyR2 directly [43], recent structural data do not support the direct binding of mAKAPβ to RyR2 via leucine zipper motifs [44], and data presented here show that nesprin-1α associates with RyR2 in the absence of mAKAPβ (Fig. 6). How RyR2 is recruited to the perinuclear compartment remains unclear, but may involve other nesprin-1α binding partners such as kinesin [8,45].

The inclusion of RyR2 in mAKAPβ signalosomes suggests a mechanism for cross-talk between Ca²⁺ and PKA signaling. Although the relevance of PKA-catalyzed RyR2 phosphorylation to excitation-contraction coupling has been controversial [35,36], PKA-catalyzed RyR2 phosphorylation at serine residues 2808 and/or 2030 is poised to play a role at mAKAPβ signalosomes where the stoichiometry and compartmentation might favor PKA-potentiated RyR2 opening. By back-phosphorylation assay, we found that ~2/3 of the available PKA phosphorylation sites on mAKAPβ-associated RyR2 were phosphorylated in response to β-adrenergic stimulation [12]. As RyR2 associates with nesprin-1α independently of mAKAPβ (Fig. 6), it is an absence of PKA, as opposed to an absence of RyR2 in the relevant compartment, that can explain the lack of β-adrenergic-induced perinuclear Ca²⁺ transients in mAKAP shRNA-expressing adult and neonatal myocytes (Fig. 1). Taken together, PKA-dependent RyR2 phosphorylation at mAKAPβ and nesprin-1α complexes appears crucial for the regulation of CaN-NFAT signaling that induces pathological cardiac hypertrophy. Moreover, the binding of phosphatases by mAKAPβ [11], including CaN and protein phosphatase 2A that can oppose RyR2 phosphorylation [46], allows tight control of hypertrophic RyR2 signaling in this critical compartment, explaining in part the digital nature of many of the results presented herein.

A major challenge in the development of new therapeutics for the treatment of pathological cardiac remodeling and heart failure has been the pleiotropy of many potential targets for intervention, including prominently CaN [47]. The identification of intracellular signaling compartments dedicated to specific cellular functions may provide opportunities to target selectively individual pools of signaling enzyme required for adverse cardiac remodeling. The identification here of an independent perinuclear Ca²⁺ compartment responsible for activating the CaN-NFAT pathway in hypertrophy provides compelling evidence supporting the development of a translational pipeline targeting the mAKAPβ signalosome in cardiovascular disease.

Abbreviations:

AKAR4

PKA activity reporter

CaN

calcineurin

CaNAR

calcineurin activity reporter

FRET

Forster resonance energy transfer

mAKAPβ

muscle A-Kinase Anchoring Protein

MEF2

myocyte enhancer factor 2

NFAT

nuclear factor of activated T-cells

Parv

parvalbumin

PDE

phosphodiesterase

PKA

protein kinase A

Data availability statement

The data that support the findings of this study are available from the corresponding author upon request.