Abstract B56α is a protein phosphatase 2 A (PP2A) regulatory subunit which modulates the heart's inotropic response to acute β-adrenergic receptor (β-AR) stimulation, although knowledge of the underlying molecular mechanisms is limited. In this study, mice deficient for B56α and wildtype controls received an intraperitoneal injection of isoproterenol (0.1 mg/kg) to activate β-AR signalling in vivo, and their hearts examined two minutes post-injection by quantitative phosphoproteomics to identify mechanisms of acute β-adrenergic signalling. We identified site- and genotype-specific phosphorylation changes on >200 proteins, including 25 hyperphosphorylated proteins harbouring a B56 binding motif as putative substrates. Functional enrichment analysis pointed to cardiac Ca^2+ release and contractility as key processes impacted by B56α deficiency, as well as cardiac muscle hypertrophy as a potential disease mechanism. In vitro, loss of B56α in cardiomyocytes blunted acute isoproterenol-induced increases in intracellular calcium transient amplitude, confirming that B56α plays a key role in calcium handling. In vivo, loss of B56α protected mice from developing systolic dysfunction in response to sustained isoproterenol infusion (60 mg/kg/day for 14 days), despite comparable increases in heart mass. These findings reaffirm a key role for B56α as a mediator of physiologically important cardiac responses to β-AR stimulation and reveal potential new molecular mechanisms for this regulatory function, including putative cardiac B56α substrates. Keywords: Phosphatases, Phosphoproteomics, Calcium handling, SPEG, CaMKII Graphical abstract [37]Unlabelled Image [38]Open in a new tab Highlights * • Loss of the PP2A regulatory subunit B56α alters the cardiac phosphoproteome * • We identified 25 putative substrates of B56α downstream of β-adrenergic receptors * • Ca^2+ handling proteins were hyperphosphorylated in B56α-deficient hearts with ISO * • Loss of B56α blunts isoproterenol-induced cardiomyocyte Ca^2+ transient amplitudes * • Loss of B56α protects against isoproterenol-induced systolic dysfunction 1. Introduction Phosphorylation is a key post-translational modification that modulates protein function by altering protein activity, conformation, stability or localisation. In the heart, dynamic phosphorylation and dephosphorylation of proteins – by kinases and phosphatases, respectively – allows precise control of key processes, including calcium handling, myofilament contraction, gene expression and metabolism. Numerous studies have demonstrated the importance of reversible protein phosphorylation in maintaining cardiac function and structure, where disturbed phospho-regulation of even a single protein can be detrimental [[39][1], [40][2], [41][3]]. Beta-adrenergic receptor (β-AR) stimulation by catecholamines represents a powerful physiological mechanism for regulation of cardiac function, with protein kinase A (PKA) and Ca^2+/calmodulin-dependent protein kinase (CaMKII) contributing to the heart's inotropic and lusitropic response to noradrenergic stimulation via the phosphorylation of calcium handling and myofilament proteins [[42]4,[43]5]. We recently reported a broad spectrum quantitative profiling of the cardiac phosphoproteome in wildtype mice following in vivo β-AR stimulation, which identified hundreds of phospho-sites that were significantly regulated [[44]6]. Interestingly, phospho-sites that displayed reduced phosphorylation in response to acute β-AR stimulation were comparable in number to those that displayed increased phosphorylation, implicating phosphatases in the physiological responses elicited by β-AR stimulation [[45]6]. In support of such a role for phosphatases in β-adrenergic regulation of protein function, we demonstrated that acute β-AR stimulation leads to PP2A-mediated dephosphorylation and nuclear accumulation of histone deacetylase 5 (HDAC5), where HDAC5 then inhibits the pro-hypertrophic transcription factor MEF2 [[46]7]. PP2A is a heterotrimeric serine/threonine phosphatase which is composed of scaffolding (A), catalytic (C) and regulatory (B) subunits and whose dysregulated activity compromises cardiac structure and function [[47]8]. Targeting of the PP2A holoenzyme to its substrates is directed by one of >20 regulatory B subunit isoforms that associate with the AC dimer. We and others have identified B56α as an abundant cardiac B subunit isoform that modulates cardiac functional responses to acute β-AR stimulation [[48][9], [49][10], [50][11]]. A proteomic analysis of skinned cardiomyocytes treated with isoproterenol (ISO, a non-selective β-AR agonist) revealed that β-AR stimulation leads to translocation of B56α from the myofilaments to the cytosol [[51]12], suggesting a role for PP2A-B56α in the neurohormonal regulation of myofilament contraction. Subsequent studies showed that the in vivo inotropic response to acute β-AR stimulation was blunted in a mouse model with targeted disrupted of the gene encoding B56α [[52]9]. Immunoblotting for phosphorylated moieties of proteins with established roles in regulation of cardiac function revealed altered phosphorylation of the ryanodine receptor (RyR[2]) and voltage-gated Na^+ channel (Na[v]1.5) in hearts of mice deficient for B56α [[53]9,[54]11,[55]13]. However, it remains unclear if these proteins are direct or indirect targets of PP2A-B56α, and whether PP2A-B56α has other substrates that modulate the heart's functional response to β-AR stimulation. In our previous study, the phosphorylation of functional residues in other key Ca^2+ handling and myofilament proteins (cardiac troponin I (cTnI), cardiac myosin binding protein C (cMyBPC) and phospholamban) was not altered by B56α deletion [[56]9], providing impetus for us to investigate possible mechanisms using a non-targeted approach (i.e. phosphoproteomics). In this study, we utilised Super-SILAC mouse technology to perform a quantitative, comparative analysis of the cardiac phosphoproteome in mice with wildtype (WT) or homozygous (HOM) expression of a hypomorphic B56α allele, acutely treated in vivo with either saline or ISO. Using a customised mass spectrometry-based workflow ([57]Fig. 1A), our study identified hundreds of phosphosites that displayed altered phosphorylation in the hearts of WT and HOM mice that received ISO, including those with roles in cardiomyocyte calcium handling, contractility and cardiac hypertrophy. In light of our findings, we assessed and identified differences in the cardiomyocyte calcium response to acute β-AR stimulation in the absence and presence of B56α. We also examined how hearts of WT and HOM mice remodel and function in response to sustained β-AR stimulation in vivo. Fig. 1. [58]Fig. 1 [59]Open in a new tab B56α regulates phosphorylation responses to acute β-adrenergic stimulation in vivo. A) 10-week-old male mice expressing wildtype (WT) B56α or homozygous (HOM) for a hypomorphic B56α allele received an intraperitoneal (i.p.) injection of isoproterenol (ISO, 0.1 mg/kg in 0.9% NaCl) or saline. Hearts were explanted after 2 min and ventricular lysates mixed with lysate from SILAC-labelled mouse hearts (2:1 ratio) prior to processing for proteomics and phosphoproteomics using liquid chromatology tandem mass spectrometry (LC-MS/MS). B) Volcano plot showing relative phosphosite abundance in saline-treated HOM vs WT hearts, FDR 0.05, S0 0.1, n = 3/group. C) Volcano plot showing relative phosphosite abundance in ISO-treated HOM vs WT hearts, FDR 0.05, n = 3/group. D) Gene Ontology enrichment analysis of proteins that displayed altered (Molecular Function) or increased (Cellular Component, Biological Process) phosphorylation in HOM vs WT mice with ISO. 2. Materials and methods Mice on a C57BL6/N background were maintained at St Thomas' Hospital (London, UK). Animal experiments were conducted in accordance with the Home Office Guidance on the Operation of Animals (Scientific Procedures) Act 1986, published by Her Majesty's Stationary Office, London. 2.1. Preparation of mouse ventricular tissue for phosphoproteomics Ventricular tissue from 10-week-old male mice homozygous (HOM) for a mutated Ppp2r5a allele [[60]9] and wildtype (WT) littermate controls was collected following acute stimulation with ISO or saline. In brief, mice were anaesthetised with 1.5% isoflurane, then injected with ISO (0.1 mg/kg body weight in vehicle, i.p.) or an equivalent volume of vehicle (0.9% NaCl). Two minutes post-injection, hearts were explanted and washed quickly in ice-cold saline. The atria and the base of the heart were removed, and the remaining ventricular tissue snap-frozen in liquid nitrogen. Protein was extracted, combined with protein from a ^13C[6]-Lys-SILAC mouse heart standard at a ratio of 2:1, and subjected to Lys-C digestion and TiO[2] phosphopeptide enrichment prior to high-pH reverse phase fractionation and mass spectrometric analysis, as described [[61]6]. Raw data files were processed and analysed using MaxQuant software, as described [[62]6]. The cardiac phosphoproteome of WT mice subjected to this experimental protocol has been previously reported [[63]6]. In the current study, we provide the first phosphoproteomic characterisation of B56α-deficient hearts, and compare the ISO response in hearts from B56α HOM and WT mice. Gene Ontology enrichment analyses were performed using GOrilla ([64]https://cbl-gorilla.cs.technion.ac.il) [[65]14] and motifs with similarity to the B56 short linear motif LxxIxE identified using ProViz ([66]https://slim.icr.ac.uk/proviz/) [[67]15]. Simple Enrichment Analysis ([68]https://meme-suite.org/meme/tools/sea) was performed as described [[69]16]. The filtering of significantly regulated phosphosites and proteins was analysed using two-sample t-test with a Permutation-based FDR <0.05 and S0 (background variability) of 0.1. 2.2. Preparation of adult mouse ventricular myocytes (AMVM) Adult mouse ventricular myocytes (AMVM) were isolated from male 8–10-week-old mouse hearts by collagenase-based enzymatic digestion. Briefly, hearts were consecutively perfused with Ca^2+-free HEPES–Tyrode solution containing 100 μM EGTA for 7 min and HEPES–Tyrode solution containing 100 μM CaCl[2] and 1 mg/mL Type II collagenase (Worthington Biochemical Corp.) for 8 min. Any remaining ventricular tissue was trimmed using fine iris scissors and subjected to further digestion for 5 min using the same collagenase buffer at 37°C in a Petri dish. Isolated myocytes were separated from undigested ventricular tissue by filtering through 200 μm nylon gauze, and the cells allowed to settle by gravity for 8 min. Cells were transferred to HEPES–Tyrode solution containing 1% BSA and 250 μM CaCl[2]. Calcium was re-introduced in increments of 250 μM and finally pooled and resuspended in HEPES–Tyrode solution containing 1 mM CaCl[2]. 2.3. Measurement of calcium transients in AMVM Intracellular calcium transients were assessed in isolated AMVM loaded with 1 μM Fura-2 AM (Abcam, ab120873) diluted in pluronic acid (Sigma, P2443) using IonOptix equipment paired with a Nikon Eclipse TE300 microscope, as previously described [[70]17]. The dual excitation beam wavelengths used were 340 nm and 380 nm, at a transmission intensity of 50%. The cells were only exposed to the fluorescent light source during measurements. Cells were paced at 1 Hz throughout the experiment, and measurements recorded for at least 30 s prior to and following 2 min stimulation with 10 nM ISO or Tyrode's solution. I[340]/I[380] ratios were normalised to allow comparison across different cells and conditions by dividing the change in I[340]/I[380] ratio by the baseline I[340]/I[380] ratio. 2.4. Chronic β-adrenergic stress model Osmotic mini pumps (Model 1002, Alzet) were implanted in male mice at 8–10 weeks of age, as described previously [[71]9]. Anaesthesia was maintained via nose-cone inhalation of 3–4% isoflurane throughout the procedure. Vetergesic (buprenorphine hydrochloride; 0.03 mg/mL) was administered via single intraperitoneal injection during mini-pump implantation as a post-operative analgesic. Mice were randomly allocated to receive either a pump containing vehicle (0.1% ascorbic acid in 0.9% NaCl) or isoprenaline hydrochloride (Sigma I5627) dissolved in vehicle and delivered at a dose of 60 mg/kg/day for 14 days. Mice were imaged via echocardiography after 14 days of infusion. 2.5. Echocardiography Echocardiography was performed using the Visual Sonics Vevo® 770 imaging system (Scanhead: RMV707B, 15–45 MHz, cardiac mouse) as described previously [[72]9]. Anaesthesia was maintained via nose-cone inhalation of 1.5–2.0% isoflurane throughout the procedure. Body temperature was monitored and maintained at 37 ± 1.5°C to sustain a heart rate of approximately 400–600 bpm. 2.6. Statistical analyses of calcium transient and echocardiography data Data are presented as scatterplots with lines indicating the mean ± SEM. Data sets were analysed using two-sided unpaired/paired t-tests or two-way ANOVA as appropriate, using Graph-Pad Prism 9.0 and 10.0 software. Tukey's post-hoc tests were performed if the two-way ANOVA revealed a significant effect of both factors or if there was a significant interaction between factors. Sidak's post-hoc tests were performed if the two-way ANOVA revealed a significant effect of one factor only. Differences were considered significant when P < 0.05. 2.7. Data availability The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE [[73]18] partner repository with the dataset identifier PXD058709. 3. Results 3.1. Loss of B56α has minimal impact on the cardiac proteome but alters the phosphoproteome in unstimulated conditions We quantified 5685 proteins and 9669 phosphosites in ventricular homogenates from WT and HOM mice that had received an intraperitoneal injection of saline or ISO 2 min prior to tissue collection (see [74]Fig. 1A). The ventricular proteome was largely unaffected by B56α deficiency, with only 5 proteins differing in abundance, including a marked reduction in B56α itself ([75]Supp Fig. 1). Apart from the expected significant reduction in the abundance of B56α in HOM hearts, we observed modest (∼30–50%) but statistically significant reductions in the abundance of PP2A scaffold subunit Aα (encoded by Ppp2r1a), thiosulfate sulfurtransferase (encoded by Tst) and Ppm1k (a phosphatase of the PPM family that localises to the mitochondria and regulates the mitochondrial permeability transition pore). In addition, we observed an ∼85% reduction in immunoglobulin heavy constant gamma 2B (encoded by Igh3). To evaluate the impact of B56α deficiency on cardiac protein phosphorylation in unstimulated conditions, we analysed the phosphoproteome of ventricular tissue obtained from WT and HOM mice that received an injection of saline. We identified 23 phosphorylation sites on 20 proteins that were differentially phosphorylated between genotypes ([76]Fig. 1B). These included phosphorylation sites on kinases (SNRK, CKMT2, MAST2), phosphatases (B56α, B56ε, DUSP27) as well as important calcium handling proteins (RyR[2], LTCC, CASQ2; see [77]Table 1). Table 1. Loss of B56α alters the cardiac phosphoproteome. Differentially phosphorylated proteins in ventricular tissue from male 10-week-old WT and HOM mice under non-stimulated conditions. HOM/WT (fold difference) ↑ or ↓ Gene name Protein name Amino acid P-value 7.516 ↑ Sdpr Serum deprivation-response protein S247 1.1E-05 6.681 ↑ Ednrb Endothelin B receptor S435 1.9E-03 5.502 ↑ Trappc12 Trafficking protein particle complex subunit 12 S232 6.2E-04 5.315 ↑ Ppp2r5a Serine/threonine-protein phosphatase 2A 56 kDa regulatory subunit alpha isoform S5 2.2E-04 5.063 ↑ Snrk SNF-related serine/threonine-protein kinase S518 1.2E-06 3.531 ↑ Snrk SNF-related serine/threonine-protein kinase S162 3.0E-07 3.204 ↑ Lig1 DNA ligase 1 S67 1.5E-04 2.713 ↑ Snrk SNF-related serine/threonine-protein kinase S351 2.6E-09 1.647 ↑ Ppp2r5e Serine/threonine-protein phosphatase 2A 56 kDa regulatory subunit epsilon isoform T7 2.6E-05 0.582 ↓ Cacna1c Voltage-dependent L-type calcium channel subunit alpha-1C S870 6.2E-05 0.578 ↓ Sptbn1 Spectrin beta chain, non-erythrocytic 1 S2340 9.1E-07 0.540 ↓ Ryr2 Ryanodine receptor 2 S2915 2.5E-05 0.503 ↓ Dbt Lipoamide acyltransferase component of branched-chain alpha-keto acid dehydrogenase complex, mitochondrial S220 2.5E-04 0.415 ↓ Ccdc85a Coiled-coil domain-containing protein 85 A S308 3.5E-04 0.366 ↓ Casq2 Calsequestrin-2 S36 1.3E-04 0.240 ↓ Ckmt2 Creatine kinase S-type, mitochondrial Y208 2.6E-04 0.238 ↓ Spg20 Spartin S376 5.6E-04 0.171 ↓ Cmya5 Cardiomyopathy-associated protein 5 (Myospryn) S1993 1.1E-05 0.148 ↓ Dusp27 Inactive dual specificity phosphatase 27 S597 8.3E-04 0.141 ↓ Fyco1 FYVE and coiled-coil domain-containing protein 1 S329 7.2E-04 0.127 ↓ Mast2 Microtubule-associated serine/threonine-protein kinase 2 S1710 1.1E-04 0.087 ↓ Perm1 PGC-1 and ERR-induced regulator in muscle protein 1 S198 1.1E-04 0.087 ↓ Perm1 PGC-1 and ERR-induced regulator in muscle protein 1 S208 1.1E-04 0.042 ↓ Nckap1 Nck-associated protein 1 S2 7.2E-04 [78]Open in a new tab 3.2. Impact of B56α deficiency on the cardiac phosphoproteome following acute β-AR stimulation We recently reported that acute β-AR stimulation altered the phosphorylation of 197 unique phosphosites in wildtype mice in vivo [[79]6]. In conducting the reported study, we subjected HOM mice to acute β-AR stimulation in parallel, which identified 287 phosphosites that were responsive to β-AR stimulation with ISO ([80]Supp Table 1). To determine the impact of B56α deficiency on the cardiac protein phosphorylation response to acute β-AR stimulation, we compared the phosphoproteome of hearts from WT and HOM mice that received ISO. We identified 292 phospho-sites in 202 different proteins that displayed a statistically significant increase or decrease in phosphorylation between ISO-treated WT and HOM hearts ([81]Fig. 1C, [82]Supp Table 1). Of these, 191 phospho-sites (65%) displayed increased phosphorylation and 101 phospho-sites (35%) displayed reduced phosphorylation. To explore whether the phosphopeptides that displayed altered phosphorylation between the two genotypes resided in networks of proteins that are associated with certain cellular functions, we performed pathway enrichment analyses using the Gene Ontology (GO) database ([83]Supp Table 2). Functional analysis of enriched GO terms highlighted important molecular functions that are likely to be regulated by B56α, including phosphatase regulator activity, kinase binding, and ankyrin binding ([84]Fig. 1Di). The observed enrichment of phosphatase regulatory activity and kinase binding terms reaffirm the role for B56α in regulating the activity of PP2A and suggest a potential role for its regulation of other phosphatases and kinases via direct and indirect interactions. Moreover, the previously published interaction between B56α and ankyrin-B [[85]19] appeared to be impacted by B56α deficiency in HOM mice, giving rise to enrichment of the ankyrin binding term. We also analysed the enriched GO terms representing biological processes, which highlighted ‘cardiac muscle contraction’, ‘regulation of release of sequestered Ca^2+ into cytosol by SR’ and ‘cardiac muscle hypertrophy’ amongst the most prominent processes ([86]Fig. 1Dii). Analysis of GO terms representing cellular components revealed an enrichment for cardiomyocyte contractile machinery proteins and interactors ([87]Fig. 1Diii), consistent with our previous report showing that B56α localises to the Z disc within the I band of myofilaments [[88]12]. 3.3. Identification of putative B56α substrates downstream of acute β-AR stimulation To identify putative substrates of PP2A-B56α, we performed motif analysis to search for a known B56 docking motif in proteins that were hyperphosphorylated in HOM mice following ISO treatment. Several structural and biochemical studies have shown that PP2A, similar to kinases, binds conserved short linear motifs (SLiMs) found in its substrates and regulators [[89]20,[90]21]. A SLiM which binds to a highly conserved, hydrophobic binding pocket on all B56 regulatory subunit isoforms has been identified (LxxIxE) [[91]22,[92]23]. We scanned protein sequences for regions that display similarity to this experimentally determined motif using the ProViz interface built on the UniProt protein search engine. We identified 25 proteins that were 1) hyperphosphorylated in HOM hearts with ISO, and 2) harboured the B56 docking motif (see [93]Table 2), suggesting potential direct interactions between PP2A-B56α and these proteins upon cardiac β-AR stimulation. Using Simple Enrichment Analysis in the MEME Suite of motif analysis tools, we confirmed that there was a significant enrichment of motifs with consensus similarity to the B56 binding motif in proteins that were hyperphosphorylated in HOM-ISO vs WT-ISO hearts compared with those that were not differentially phosphorylated (enrichment ratio 3.44, q = 0.04). In contrast, there was not a significant enrichment of the motif in proteins that were hypophosphorylated in HOM-ISO vs WT-ISO hearts (enrichment ratio 1.69, q = 0.56). Table 2. Putative substrates of cardiac B56α following acute β-adrenergic stimulation. Proteins containing a B56 docking motif (LxxIxE) that were hyperphosphorylated in HOM vs WT hearts with ISO. Normalised S^pep = Position Specific Scoring Matrix (PSSM) score of the peptide / maximum possible peptide score for a PSSM. Consensus similarity to B56 docking motif sequence: *P < 0.0001, **P < 0.00001, ***P < 0.000001. Disorder score (computed from AlphaFold2 structures): 0 most likely globular, 1 most likely disordered. Gene Protein name Motif sequence(s) Start Stop Normalised S^pep Consensus similarity Disorder score Vcan Versican core protein MPDISEIKE ISEIKEEEL 453 456 461 464 0.60 0.47 * * 0.69 0.68 Pcm1 Pericentriolar material 1 protein LSYIEEKEQ 987 995 0.52 * 0.35 Dync1i2 Cytoplasmic dynein 1 intermediate chain 2 LAQIREEKK 18 26 0.49 * 0.76 Arhgef7 Rho guanine nucleotide exchange factor 7 LQNILETEH 255 263 0.60 * 0.27 Ppip5k2 Inositol hexakisphosphate and diphosphoinositol-pentakisphosphate kinase 2 LRKVPEMSS 1044 1052 0.47 * 0.70 Snrk SNF-related serine/threonine-protein kinase LNQIFEEGE 486 494 0.48 * 0.70 Ap4e1 AP-4 complex subunit epsilon-1 LRMIKENAS 229 237 0.47 * 0.05 Nes Nestin LRSLDENQE 839 847 0.49 * 0.62 Sorbs2 Sorbin and SH3 domain-containing protein 2 LPQIPERNS 844 852 0.60 * 0.54 Myo9b Unconventional myosin-IXb IQSIKEEKE LSPLPEAAA 1927 1991 1935 1999 0.58 0.57 * * 0.47 0.79 Afap1l1 Actin filament-associated protein 1-like 1 LTVIKEEQL 240 248 0.52 * 0.21 Dennd4c DENN domain-containing protein 4C LRPITEAPS 580 588 0.57 * 0.33 Cobll1 Cordon-bleu protein-like 1 LEEIDEKEE 927 935 0.54 * 0.81 Speg Striated muscle-specific serine/threonine-protein kinase MPSIPEEPE LRPIPELLR LTSVHEDDS 1175 1895 1033 1183 1903 1041 0.85 0.60 0.47 *** * * 0.75 0.31 0.15 Sptbn1 Spectrin beta chain, non-erythrocytic 1 MQLISEKPE LEPLSERKH 1338 1471 1346 1479 0.56 0.48 * * 0.48 0.36 Perm1 PGC-1 and ERR-induced regulator in muscle protein 1 CDTIEEEDE 602 610 0.69 ** 0.53 Xirp2 Xin actin-binding repeat-containing protein 2 LEKIKEESG LEDIREDKK LDSINELDE LDSIHESED 588 693 1270 1131 596 701 1278 1139 0.49 0.55 0.50 0.53 * * * * 0.41 0.39 0.54 0.30 Limch1 LIM and calponin homology domains-containing protein 1 MPKILERSH 462 470 0.58 * 0.65 Synj2 Synaptojanin-2 LYPLQEEAK 104 112 0.58 * 0.32 Cdyl Chromodomain Y-like protein LCTLPEKAE 15 23 0.53 * 0.80 Myh9 Myosin-9 LKKIRELET IMGIPEDEQ 1103 328 1111 336 0.47 0.49 * * 0.52 0.24 Flnb Filamin-B CQEIPEGYK 2431 2439 0.59 * 0.39 Cavin4 Caveolae-associated protein 4 LSSVTEDED LGPIHEFHS 18 301 26 309 0.48 0.57 * * 0.53 0.53 Fhod3 FH1/FH2 domain-containing protein 3 LSSISELSA 1193 1201 0.54 * 0.06 Son Protein SON ILPISETKE 1595 1603 0.47 * 0.51 [94]Open in a new tab We have also explored whether there is overlap between phosphosites that are less abundant in WT-ISO versus WT-VEH hearts and those that are more abundant in HOM-ISO versus WT-ISO hearts, which might further support the identification of protein that are dephosphorylated through a PP2A-B56α-mediated mechanism in response to β-AR stimulation. This revealed 11 overlapping phosphosites from 10 proteins (Ser7 in Adssl1, Ser154 in Cdyl, Ser438 in Clcc1, Ser368 in Cttn, Ser223 in Kdm5a, Ser757 in Limch1, Ser 1366 in Mast2, Ser126 in Mepce, Ser1805 & Ser1808 in Pcm1 and Ser94 in Son). Interestingly, 4 of these proteins (Cdyl, Limch1, Pcm1 & Son) harboured putative B56α binding motifs (see [95]Table 2), supporting the possibility that the affected phosphosites within them might be direct substrates for PP2A-B56α. 3.4. Loss of B56α reduces cardiomyocyte calcium transient concentrations following acute β-AR stimulation Amongst the list of proteins that displayed altered phosphorylation between the hearts of WT and HOM mice that received ISO were several important calcium handling proteins, including the LTCC and RyR[2] ([96]Supp Table 1). Moreover, our pathway enrichment analysis identified cardiac muscle contraction and calcium regulatory processes as the top pathways that were impacted by the reduced abundance of B56α in the hearts of HOM mice. In addition, we recently reported that HOM mice display a blunted cardiac inotropic response to acute β-AR stimulation in vivo, indicating potential alterations in calcium handling at the cardiomyocyte level [[97]9]. In light of these observations, we investigated calcium transient responses to acute β-AR stimulation in cardiomyocytes isolated from WT and HOM mice, in the absence and presence of ISO. [98]Fig. 2A shows representative traces of calcium transients from WT and HOM cardiomyocytes before and during ISO exposure, obtained by averaging measurements over 30 contractions during each phase. Each data point in [99]Fig. 2B represents the average intracellular calcium concentration [Ca^2+][i] transient amplitude from 5 to 8 cardiomyocytes per animal. Our analyses revealed that, as expected, acute β-AR stimulation with ISO led to a significant increase in calcium transient amplitude in cardiomyocytes from WT mice ([100]Fig. 2B). However, this response to ISO was markedly blunted in cardiomyocytes from HOM mice ([101]Fig. 2B). Interestingly, calcium transient amplitudes in unstimulated conditions (i.e. pre-ISO) were significantly lower in HOM cardiomyocytes compared with WT ([102]Fig. 2B), suggesting a role for B56α in regulating calcium handling during basal excitation-contraction coupling also. To further investigate the impact of B56α deficiency on cardiomyocyte calcium transient amplitudes, we plotted the data from all cardiomyocytes in each group and performed paired statistical analysis between the values obtained pre-ISO and during ISO treatment ([103]Fig. 2C). We also calculated the percentage change in calcium transient amplitude for each cell ([104]Fig. 2D). Our analysis revealed a significant increase in calcium transient amplitude in both genotypes ([105]Fig. 2C), however the magnitude of the increase was considerably smaller in HOM cardiomyocytes compared with WT cardiomyocytes (∼65% increase vs ∼25% increase, P < 0.0001, [106]Fig. 2D). In contrast, the ISO effect on calcium transient kinetics was unaffected by loss of B56α. Isoproterenol reduced the decay time constant, τ ([107]Fig. 2E, F), and the time to 50% [Ca^2+][i] transient decay ([108]Fig. 2H, I), however the magnitude of the ISO-induced reduction in these parameters was comparable between genotypes ([109]Fig. 2G, J). These data demonstrate that B56α is a key regulator of the cardiomyocyte intracellular calcium response to acute β-AR stimulation. Fig. 2. [110]Fig. 2 [111]Open in a new tab Loss of B56α blunts isoproterenol-induced Ca^2+transient amplitudes in adult cardiomyocytes. Representative recordings of intracellular Ca^2+ concentration ([Ca^2+][i]) transients in cardiomyocytes isolated from adult mice expressing wildtype (WT) B56α or homozygous (HOM) for a hypomorphic B56α allele (averaged from 30 beats). Pre-ISO: cardiomyocytes under superfusion with Tyrode solution. ISO: cardiomyocytes after 2 min of superfusion with Tyrode solution containing 10 nM ISO. B, C) Quantitation of [Ca^2+][i] transient amplitudes. D) ISO-induced change in [Ca^2+][i] amplitude. E, F) Quantitation of the decay time constant, τ. G) ISO-induced change in τ. H, I) Quantitation of time to 50% [Ca^2+][i] transient decay. For B, E & H: Individual data points represent the average of all measurements from 5 to 8 cells per heart. Two-way ANOVA followed by Tukey's post-hoc tests. *P < 0.05. For C, F & I: Individual data points represent all measurements (5–8 cells per heart, 5 hearts per genotype). Multiple paired t-tests. *Adjusted P < 0.05. For D, G, J: Unpaired t-tests. *P < 0.05. 3.5. Impact of B56α deficiency on cardiac structural and functional responses to sustained β-AR stimulation in vivo We previously reported that sustained β-AR stimulation with 30 mg/kg/day of ISO over 14 days induces cardiomyocyte hypertrophy in WT mice, which was attenuated in HOM counterparts [[112]9]. The hypertrophic growth we observed in our previous study was modest and not associated with cardiac dysfunction [[113]9]. In the current study we identified Gene Ontology terms related to cardiac muscle hypertrophy, calcium handling and cardiac muscle contraction as some of the most prominent processes that were impacted by B56α deficiency in HOM mice. Moreover, we found that HOM cardiomyocytes displayed an attenuated calcium transient response to acute β-AR stimulation in vitro ([114]Fig. 2). These findings prompted us to investigate the potential impact of B56α deficiency on cardiac responses to more pronounced β-adrenergic stress (60 mg/kg/day ISO for 14 days), which has previously been reported to induce hypertrophy and impair contractile function (i.e. reduce ejection fraction) in mice [[115]24]. We assessed ISO-induced cardiac hypertrophy by echocardiography and gravimetric analysis 2 weeks post-implantation. ISO infusion for 2 weeks led to a significant increase in heart weight normalised to tibia length in both WT and HOM mice ([116]Fig. 3A). Regional differences in left ventricular wall thicknesses were apparent by echocardiography, with ISO leading to more pronounced thickening of the interventricular septum (IVS) and anterior wall (LVAW) in WT mice ([117]Fig. 3B, [118]Table 3) and more pronounced thickening of the posterior wall (LVPW) in HOM mice ([119]Fig. 3C). Left ventricular volumes measured at diastole were larger in HOM mice vs WT and increased with ISO treatment ([120]Fig. 3D), however the magnitude of this increase was unaffected by genotype (22% increase in WT ISO vs VEH, 23% increase in HOM ISO vs VEH). Treatment with ISO gave rise to an expected increase in heart rate of ∼60 bpm in both genotypes ([121]Table 3). Ejection fraction was lower in WT mice treated with ISO vs saline ([122]Fig. 3E; P < 0.05 by unpaired t-test) but tended to be higher in HOM mice treated with ISO vs saline ([123]Fig. 3E). In WT mice treated with ISO, stroke volume was maintained despite the fall in ejection fraction ([124]Fig. 3F), likely because of increased LV volumes, and cardiac output tended to increase because of the increase in heart rate ([125]Fig. 3G). In HOM mice treated with ISO, both stroke volume and cardiac output were significantly increased ([126]Fig. 3F, G). Collectively, these data indicate that loss of B56α protected the heart from systolic dysfunction induced by sustained ISO infusion. Fig. 3. [127]Fig. 3 [128]Open in a new tab Loss of B56α alters the remodelling response and protects against systolic dysfunction induced by sustained ISO infusion in vivo. Male mice expressing wildtype (WT) B56α or homozygous (HOM) for a B56α mutant allele were administered isoproterenol (ISO, 60 mg/kg/day) or vehicle via subcutaneously implanted osmotic minipumps for 2 weeks. A) Heart weight (HW) normalised to tibia length (TL). B) Left ventricular anterior wall thickness at diastole (LVAW;d), C) left ventricular posterior wall thickness at diastole (LVPW;d), D) left ventricular volume at diastole (LV Vol;d), E) ejection fraction (EF), F) stroke volume, and G) cardiac output, as assessed by echocardiography 2 weeks after implantation of osmotic minipumps. Two-way-ANOVA followed by Tukey's or Sidak's post-hoc test, as appropriate. n = 9–11/group. For D & E: P-values from the ANOVA are shown (no significant differences in post-hoc analyses). Note: HW was not recorded at the time of dissection for one mouse in the HOM ISO group. Table 3. Echocardiography data from male mice expressing wildtype B56α or homozygous for a B56α mutant allele following 2 weeks of isoproterenol (60 mg/kg/day) or vehicle infusion delivered via subcutaneously implanted osmotic minipump. __________________________________________________________________ Wildtype __________________________________________________________________ Homozygous __________________________________________________________________ 2-way ANOVA (P-value) __________________________________________________________________ Vehicle Isoproterenol Vehicle Isoproterenol Treatment Genotype Interaction n 10 10 9 9 Heart rate (bpm) 544 ± 16 604 ± 9 545 ± 17 589 ± 16 0.001 0.62 0.60 IVS;d (mm) 0.78 ± 0.07 0.98 ± 0.08 0.74 ± 0.02 0.81 ± 0.02 0.02 0.06 0.21 LVPW;d (mm) 0.70 ± 0.04 0.81 ± 0.04 0.63 ± 0.04 0.88 ± 0.03 <0.001 0.93 0.07 LVAW;d (mm) 0.83 ± 0.04 0.93 ± 0.03 0.79 ± 0.02 0.82 ± 0.03 0.06 0.03 0.23 LVID;d (mm) 3.88 ± 0.12 4.22 ± 0.15 4.18 ± 0.07 4.55 ± 0.13 0.007 0.02 0.89 LVID;s (mm) 2.71 ± 0.12 3.18 ± 0.16 3.15 ± 0.08 3.32 ± 0.19 0.03 0.04 0.29 LVvol;d (μL) 66.1 ± 4.8 80.9 ± 7.4 77.7 ± 2.8 95.8 ± 6.9 0.008 0.03 0.78 LVvol;s (μL) 28.0 ± 3.3 41.7 ± 5.6 39.8 ± 2.4 46.6 ± 7.0 0.04 0.10 0.48 Stroke volume (μL) 38.1 ± 3.0 39.2 ± 2.6 37.9 ± 3.2 49.2 ± 2.6 0.04 0.10 0.08 Cardiac output (mL/min) 20.6 ± 1.6 23.6 ± 1.5 20.6 ± 1.8 28.7 ± 1.3 0.001 0.11 0.11 Fractional shortening (%) 30 ± 2 25 ± 1 24 ± 2 27 ± 2 0.54 0.35 0.03 Ejection fraction (%) 58 ± 3 50 ± 2 48 ± 3 53 ± 3 0.51 0.30 0.04 [129]Open in a new tab 4. Discussion B56α is an abundant cardiac PP2A regulatory subunit isoform that localises to the Z disc of cardiomyocytes and translocates to the cytosol upon β-AR stimulation [[130]12]. We previously showed that loss of B56α blunted the inotropic response to acute β-AR stimulation in vivo, however the pertinent molecular mechanisms remained unclear [[131]9]. In this study, we used quantitative phosphoproteomics to compare cardiac phosphorylation responses to acute β-AR stimulation in wildtype and B56α-deficient mice. Our analysis identified 25 putative substrates of B56α and numerous biological processes and molecular functions that were impacted by B56α deficiency. These included calcium handling, cardiac muscle contraction and cardiac muscle hypertrophy, amongst others. Using cardiomyocytes isolated from adult wildtype and B56α-deficient hearts, we demonstrate that B56α is required for isoproterenol-induced augmentation of intracellular calcium transient concentrations. We also show that loss of B56α protects hearts from systolic dysfunction induced by sustained isoproterenol infusion. These findings provide new insights into cardiac B56α signalling and a foundation for the experimental validation and further investigation of B56α substrates and associated molecular pathways in the heart. Historically, the identification of phosphatase substrates has been challenging due to the transient nature of phosphatase-substrate interactions. Various methods have been developed to improve the sensitivity of affinity purification mass spectrometry approaches, such as proximity-dependent biotinylation and substrate-trapping mutagenesis [[132]25,[133]26]. Phosphoproteomics, and the computational discovery and experimental validation of short linear motifs (SLiMs) that function as binding motifs for specific phosphatases, has expedited the identification of phosphatase substrates and generated novel insights into phosphatase functions [[134]27]. Binding motifs have been identified for several phosphatases, including protein phosphatase 2B (calcineurin) [[135]28], protein phosphatase 4 (PP4) [[136]29], CDC14 [[137]30], and PTP1B [[138]31]. For PP2A, conserved binding motifs have been identified for the B56 family of regulatory subunits [[139]22] as well as the B55 family [[140]32,[141]33]. In this study, we searched for proteins harbouring the LxxIxE B56 binding motif amongst those proteins that were hyper-phosphorylated with isoproterenol in hearts of B56α-deficient mice compared with wildtype. This allowed us to distinguish proteins that are likely to be directly dephosphorylated by PP2A-B56α from those that are indirectly phospho-regulated by PP2A-B56α (i.e. by downstream kinases or phosphatases). Our approach identified 25 putative substrates of PP2A-B56α. It will be important to determine experimentally which of these proteins are bona fide substrates. The only putative substrate with a known link to PP2A was SNRK, which has been shown to interact with the PP2A catalytic subunit and B56δ (a family member of B56α) in adipocytes [[142]34]. Interestingly, this interaction facilitated the phosphorylation of B56δ by SNRK, which amplified insulin signalling by preventing PP2A-mediated dephosphorylation and inactivation of Akt [[143]34]. Whether SNRK is also a substrate of PP2A-B56 enzymes is yet to be investigated. Our in vitro investigations showed that calcium transient amplitudes were lower in cardiomyocytes from B56α-deficient mice vs wildtype in unstimulated conditions, and that isoproterenol-induced increases in calcium transient amplitude were blunted in B56α-deficient cardiomyocytes. These data suggest that B56α is required for facilitating the rise in intracellular calcium concentration following an action potential. Interestingly, overexpression of B56α in cardiomyocyte-specific transgenic mice was previously reported to blunt isoproterenol-induced calcium transient amplitudes [[144]10], similar to what we observed in our loss-of-function mouse model. The reasons for this apparent discrepancy are unclear but may relate to a potential loss of localised function and specificity with overexpression. Additionally, the difference in the concentration of isoproterenol used (1 μM in the previous study vs 10 nM in the current study) is also a likely contributor. In this context, supraphysiologic and physiologic concentrations of isoproterenol had different effects on the peak L-type Ca^2+ current in cardiomyocytes from Rad-deficient mice [[145]35], and signalling enzymes, including the Ca^2+-sensitive phosphatase calcineurin, display concentration-dependent activation by isoproterenol in cardiomyocytes [[146]36]. Thus, it is possible that differing concentrations of isoproterenol differentially affect B56α localisation and substrate targeting in cardiomyocytes. During excitation-contraction coupling, the LTCC allows calcium to enter the cell triggering calcium-induced calcium release from the sarcoplasmic reticulum via the RyR[2]. Activation of PKA or CaMKII downstream of β-AR augments calcium transients via phosphorylation of the LTCC and RyR[2] [[147]4,[148]37]. Phosphopeptides containing the canonical PKA and CaMKII phosphorylation sites from these two proteins (Ser1928 in LTCC, and Ser2808 & Ser2814 in RyR[2]) were not detected in our mass spectrometric analysis of heart homogenates. However, with acute isoproterenol stimulation we detected reduced phosphorylation of LTCC at Ser870 and reduced phosphorylation of RyR[2] at Ser1317 and Ser2915. The functional implications of phosphorylation and dephosphorylation at these amino acid residues has not been investigated. However, these data support our previous Western blot findings which suggest that PP2A-B56α does not directly target RyR[2] for dephosphorylation, but regulates its phosphorylation via an indirect mechanism (if RyR[2] was a direct substrate of PP2A-B56α, it would be hyper-phosphorylated in a setting of B56α deficiency) [[149]9]. The same appears to be the case for LTCC, as it was also hypo-phosphorylated in hearts of B56α-deficient mice. Phospholamban is another key protein involved in excitation-contraction coupling that may influence calcium transient amplitudes by modulating SERCA2a activity and, consequently, sarcoplasmic reticulum Ca^2+ load. Phosphorylation of Ser16 by PKA or Ser17 by CaMKII reduces phospholamban inhibition of SERCA2a, allowing faster and more efficient Ca^2+ reuptake into the sarcoplasmic reticulum. We previously showed that ISO-stimulated phosphorylation of phospholamban at Ser16 is unaffected by loss of B56α [[150]9]. In the current study, we were unable to explore the effects of B56α deficiency on phospholamban phosphorylation at other sites, as no phosphopeptides derived from phospholamban were detected in our analysis. Amongst the putative B56α substrates identified in this study was striated muscle-specific serine/threonine-protein kinase (SPEG). This protein was of particular interest given its previously reported roles in cardiomyocyte calcium handling. SPEG stabilises the interaction between excitation-contraction coupling proteins in cardiac dyads [[151]38], and phosphorylates RyR[2] at Ser2367, which reduces calcium leak from the sarcoplasmic reticulum [[152]39]. In addition, SPEG is phospho-regulated by Akt and has been shown to play critical roles in calcium handling downstream of insulin receptor signalling in cardiomyocytes [[153]40]. Specifically, phosphorylation of SERCA2a by activated SPEG accelerated calcium uptake into the sarcoplasmic reticulum by promoting SERCA2a oligomerisation, and ablation of the Akt phosphorylation sites in SPEG reduced systolic function in mice [[154]40]. Our analysis revealed that SPEG contains three motifs with amino acid similarity to the B56 binding motif ([155]Table 2) and was hyper-phosphorylated in HOM ISO vs WT ISO hearts at Ser554, Ser2288 and Ser2327 ([156]Supp Table 1). These phosphosites are distinct from the Akt phosphorylation sites (which were not detected in our dataset) and have not been functionally characterised. Given the established role of SPEG in calcium handling, our observation that the magnitude of intracellular calcium transients is dependent on B56α, and our identification of SPEG as a potential B56α substrate, it is tempting to speculate that SPEG is dephosphorylated by B56α in response to β-AR stimulation and that this has a functional impact on calcium handling within the cardiomyocyte. However, our observation that calcium transient kinetics were unaffected by B56α deletion suggests any contribution of SPEG in this context is independent of its actions on SERCA2a. It would have been valuable to measure sarcoplasmic reticulum Ca^2+ content via caffeine stimulation to further interrogate possible mechanisms contributing to blunted [Ca^2+][i] levels following ISO stimulation. This is a limitation of the current study. In addition to terms associated with calcium handling, our pathway enrichment analysis of phosphosites that were differentially impacted by isoproterenol in wildtype and B56α-deficient hearts revealed an over-representation of proteins involved in cardiac muscle hypertrophy and contraction. We previously reported that loss of B56α attenuated cardiomyocyte hypertrophy in response to sustained isoproterenol infusion in vivo [[157]9]. However, the degree of hypertrophy in this model was mild and not accompanied by a decline in cardiac contractile performance. This prompted us to explore if loss of B56α would also impact the cardiac response to a more pronounced form of sustained β-AR stimulation, mimicking the setting of chronic sympathetic stimulation that leads to pathological cardiac remodelling and heart failure (60 mg/kg/day isoproterenol infusion). We found that B56α-deficient mice were protected from developing systolic dysfunction in this setting, despite comparable increases in heart mass. Investigation of the mechanisms contributing to this phenotype were beyond the scope of this study. However, considering our previously reported observation that HOM mice display an attenuated inotropic response to acute β-AR stimulation, it seems likely that the altered response to sustained isoproterenol stimulation in HOM mice might partially arise from a diminished cardiac workload. Reduced signalling via calcium-dependent pathways may also have contributed, given our observation of reduced intracellular calcium transients in B56α-deficient cardiomyocytes both basally and in response to acute isoproterenol stimulation. Acute treatment with isoproterenol increases activity of the calcium-dependent phosphatase calcineurin in isolated cardiomyocytes in vitro and in hearts in vivo [[158]36]. Chronic calcineurin activation, as occurs in settings of sustained β-AR stimulation, activates signalling cascades involved in pathological cardiac remodelling and dysfunction and leads to failure [[159]41]. Similarly, CaMKII is a calcium-dependent kinase that is activated in response to sustained isoproterenol infusion [[160]37] and which has been identified as a key contributor to pathological cardiac hypertrophy and dysfunction in cardiac stress settings [[161]42]. Thus, reductions in intracellular calcium concentrations would be anticipated to blunt calcineurin and CaMKII signalling, which may have prevented the development of systolic dysfunction in B56α-deficient mice. In our phosphoproteomic dataset from mice receiving acute treatment with isoproterenol, no phosphopeptides derived from calcineurin were detected. However, our analysis revealed modest (40–65%) decreases in the abundance of CaMKIIδ phosphopeptides corresponding to amino acids Thr330, Ser333, Ser336, Ser347 and Thr351 in HOM ISO hearts vs WT ISO hearts. While the Thr287 autophosphorylation site (which leads to CaMKII activation) was not detected, these data reveal altered phosphorylation of CaMKIIδ in B56α-deficient hearts at several novel sites, which may have physiological significance. In this context, a quantitative phosphoproteomics analysis of mouse brain tissue revealed that CaMKII isoforms are phosphorylated on at least 6–7 amino acid residues, and that distinct subcellular pools of CaMKII are differentially phosphorylated at specific sites and interact with different networks of proteins [[162]43]. Our identification of novel cardiac CaMKIIδ phosphosites that are differentially impacted by isoproterenol in wildtype and B56α-deficient hearts opens new avenues of investigation into CaMKII biology. Taken together, our findings validate and extend previous work identifying B56α as a regulator of excitation-contraction coupling proteins, kinase/phosphatase binding, and protein localisation and compartmentation. Amongst the proteins that were differentially phosphorylated in WT and HOM hearts following ISO stimulation were key proteins with established roles in cardiac physiology and pathophysiology (e.g. cMyBP-C, RyR[2], myosin regulatory light chain 2, voltage-dependent L-type calcium channel, calsequestrin, CaMKII, PKARIα, GSK3β, SNRK, β-AR kinase 1, PP1 and PP4, Ankyrin B, AKAPs), positioning B56α as an important regulator of numerous cardiac signalling networks and processes. In addition, our analysis identified several novel putative substrates of PP2A-B56α. The unique phosphoproteomics datasets generated in this study and our earlier work in wildtype hearts treated with ISO [[163]6] will facilitate future studies on the molecular mechanisms underlying β-AR-mediated regulation of cardiac function in health and disease. The following are the supplementary data related to this article. Supplementary Fig. 1 Loss of B56α on the cardiac proteome. 10-week-old male mice expressing wildtype (WT) B56α or homozygous (HOM) for a hypomorphic B56α allele received an intraperitoneal (i.p.) injection of isoproterenol (ISO, 0.1 mg/kg in 0.9 % NaCl) or saline. Hearts were explanted after 2 min and ventricular tissue processed for phosphoproteomics using liquid chromatology tandem mass spectrometry (LC-MS/MS). Volcano plot showing relative peptide abundance in saline-treated HOM vs WT hearts, FDR 0.05, n = 3/group. [164]mmc1.docx^ (223.3KB, docx) Supplementary Table 1 Outputs from phosphoproteomic analysis of mice with homozygous (HOM) expression of a hypomorphic B56alpha allele and wildtype (WT) littermates [165]mmc2.xlsx^ (7.8MB, xlsx) Supplementary Table 2 Outputs from Gene Ontology enrichment analyses of phosphoproteomic data from mice with homozygous (HOM) expression of a hypomorphic B56alpha allele or wildtype (WT) littermates. [166]mmc3.xlsx^ (76.9KB, xlsx) Disclosures None. CRediT authorship contribution statement Alican Güran: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Yanlong Ji: Visualization, Investigation, Formal analysis, Data curation. Pan Fang: Data curation. Kuan-Ting Pan: Data curation. Henning Urlaub: Resources, Funding acquisition, Conceptualization. Metin Avkiran: Writing – review & editing, Supervision, Resources, Project administration, Methodology, Funding acquisition, Conceptualization. Christof Lenz: Writing – review & editing, Supervision, Project administration, Methodology, Conceptualization. Kate L. Weeks: Writing – review & editing, Writing – original draft, Visualization, Supervision, Investigation, Formal analysis. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements