|Home | About | Journals | Submit | Contact Us | Français|
Protein phosphatase 2A (PP2A) is a family of multifunctional serine/threonine phosphatases consisting of a catalytic C, a structural A, and a regulatory B subunit. The substrate and therefore the functional specificity of PP2A are determined by the assembly of the enzyme complex with the appropriate regulatory B subunit families, namely B55, B56, PR72, or PR93/PR110. It has been suggested that additional levels of regulating PP2A function may result from the phosphorylation of B56 isoforms. In this study, we identified a novel phosphorylation site at Ser41 of B56α. This phosphoamino acid residue was efficiently phosphorylated in vitro by PKCα. We detected a 7-fold higher phosphorylation of B56α in failing human hearts compared with nonfailing hearts. Purified PP2A dimeric holoenzyme (subunits C and A) was able to dephosphorylate PKCα-phosphorylated B56α. The potency of B56α for PP2A inhibition was markedly increased by PKCα phosphorylation. PP2A activity was also reduced in HEK293 cells transfected with a B56α mutant, where serine 41 was replaced by aspartic acid, which mimics phosphorylation. More evidence for a functional role of PKCα-dependent phosphorylation of B56α was derived from Fluo-4 fluorescence measurements in phenylephrine-stimulated Flp293 cells. The endoplasmic reticulum Ca2+ release was increased by 23% by expression of the pseudophosphorylated form compared with wild-type B56α. Taken together, our results suggest that PKCα can modify PP2A activity by phosphorylation of B56α at Ser41. This interplay between PKCα and PP2A represents a new mechanism to regulate important cellular functions like cellular Ca2+ homeostasis.
Protein phosphatase 2A (PP2A)2 is one of the major classes of serine/threonine protein phosphatases affecting the phosphorylation status of many phosphoproteins in different cell types. It represents nearly half of the total cellular serine/threonine phosphatase activity and has been linked to the regulation of cellular signaling and (patho)physiology. To its multiple functions belong the regulation of the cell cycle, apoptosis, signal transduction, DNA replication, and myocardial contractility (for reviews see Refs. 1–3). During the last decade, PP2A emerged as an important regulator of oncogenesis and Alzheimer disease (4, 5). The backbone of this functional importance is formed by the broad diversity of PP2A subunit combinations. The PP2A core enzyme is composed of a heterodimer, including a scaffolding A and a catalytic C subunit (PP2ACA). The association of regulatory B subunits with the core dimer confers substrate specificity and intracellular targeting of the heterotrimeric PP2A holoenzymes. Up to now, four distinct families of regulatory B subunits have been described, B55 (B), B56 (B′), PR72 (B″), and PR93/PR110 (B). The B56 subunits are the most diverse, consisting of α (PPP2R5A), β (PPP2R5B), γ (PPP2R5C), δ (PPP2R5D), and ϵ (PPP2R5E) isoforms, which are differentially expressed in many tissues and cell types (6, 7).
B56 subunits are crucial to regulate the subcellular targeting of the PP2A heterodimeric core enzyme to specific substrates and subcellular domains (6, 8). The targeting is mediated by specific adaptor proteins allowing a differential regulation of PP2A activity. For example, ankyrin-B co-localizes with B56α in cardiomyocytes leading to tethering of PP2ACA to cellular ion pumps and channels like the sodium/calcium exchanger, Na/K-ATPase, and inositol 1,4,5-trisphosphate receptor (9). Besides this critical role in PP2A targeting, B56 subunits can also act as receptors of second messengers as shown for the lipid ceramide (10). Finally, PP2A activity and subcellular localization are regulated by post-translational modifications of B56 subunits. It has been demonstrated that most of the B56 family members are phosphoproteins (6). Several protein kinases (e.g. PKA and PKR) have been reported to phosphorylate B56 subunits (11, 12). In detail, the phosphorylation of B56δ at Ser566 by PKA increases the PP2A activity that catalyzes dephosphorylation of DARPP-32, thereby coordinating the efficacy of dopaminergic neurotransmission in striatal neurons (12). Moreover, PKA-dependent phosphorylation of B56δ, which is anchored to PDE4D3 by muscle A kinase-anchoring protein, promotes the dephosphorylation of this cAMP-specific phosphodiesterase (13). This inhibits PDE4D3 activity and thereby mediates a cAMP-induced positive feedback mechanism after activation of adenylyl cyclase and B56δ phosphorylation.
Previous work has shown the phosphorylation of PP2A by PKCα at one of its regulatory B subunits (14). These authors detected a 55-kDa band that became phosphorylated in the presence of PKCα but were not able to identify the isoform of this B subunit. The classical PKC isotypes (e.g. PKCα) display a physiological requirement for Ca2+ and diacylglycerol (15). The cPKC family members are known to play an important (patho)physiological role in regulating cellular functions, including proliferation, differentiation, apoptosis, oncogenesis, and myocardial/vascular smooth muscle contraction (16), indicating that cPKC isotypes and PP2A are acting on the same signaling pathways and molecular targets. Indeed, the activation of PKCα by the phorbol ester PMA was followed by the occurrence of a membrane-associated PP2A heterotrimeric complex resulting in the dephosphorylation and desensitization of the kinase (17). Thus, the aim of this study was the identification and characterization of the missing link between PKCα and PP2A as several studies raised the possibility that B56α might mediate the kinase-phosphatase interaction. Here, we report that PKCα inhibits PP2A via phosphorylation of B56α at Ser41, leading to an altered ER Ca2+ release.
[γ-32P]ATP was obtained from Hartmann Analytic GmbH. Sf21 insect cells were supplied by Invitrogen. HEK293 cells were obtained from ATCC-LGC Standards. PMA was used to activate PKC (Sigma). All other chemicals were of reagent grade. A polyclonal antibody for phospho-Ser41 B56α was generated in rabbit and affinity-purified by use of a peptide pair of 12 amino acids, comprising residues 37–48 of human B56α (Perbio Science). The phospho-specific peptide was synthesized with a phosphoserine residue at position 41 of B56α.
Left ventricular (LV) tissue was received from patients undergoing heart transplantation due to end-stage heart failure resulting from ischemic (ICM) or dilated cardiomyopathy (DCM) and from nonfailing (NF) hearts that could not be transplanted due to medical reasons or blood group incompatibility (18). The study was approved by the local Ethic Committee of the University Hospital of Münster and the St. Vincent's Hospital Human Research Ethics Committee in Sydney, Australia (file number H03/118; Title, Molecular Analysis of Human Heart Failure). The investigation conformed to the principles outlined in the Declaration of Helsinki.
cDNA from human left ventricle (BioChain Institute Inc.) was amplified using Pfu DNA polymerase (Promega) and B56α primers as follows. The forward primer included 1 bp of the 5′-UTR, and the reverse primer extended to bp 31 of the 3′-UTR downstream of the translational stop codon. The amplified cDNA fragment was inserted into SalI of the pJET1 cloning vector (Thermo Fisher Scientific). This construct was used as template for subsequent cloning experiments. PCR was utilized for generation of a B56α fragment containing a 5′-NdeI and a 3′-engineered PstI restriction enzyme site. The amplified fragment was subcloned in-frame into the corresponding cloning sites of pAcHLTA (BD Biosciences). This vector contains a His6 tag nucleotide sequence upstream of the multiple cloning site. Alternatively, a TaqDNA polymerase-amplified PCR product, using primers that generate a His6 tag nucleotide sequence upstream of the 5′ end of B56α, was directly inserted into the pCR2.1-TOPO (Invitrogen). The His6-B56α cDNA was then excised from the pCR2.1-TOPO vector with BamHI/NotI and subcloned into corresponding coning sites of pVL1393 (BD Biosciences). In parallel, we also constructed a pCR2.1-TOPO vector that contained B56α without the His6 tag nucleotide sequence as well as the stop codon. After digestion with XhoI/BamHI, the resulting B56α fragment was subcloned into the multiple cloning site of the pAcGFP1-N1 expression vector (Clontech). Thus, B56α was fused in-frame to the N terminus of AcGFP1 (green fluorescent protein from Aequorea coerulescens). Finally, the originally constructed B56α PCR fragment was inserted into the SalI site of the pIRES2-DsRed-Express vector (Clontech). This vector allows the control of an efficient transfection and expression of the gene of interest by detection of red fluorescence.
Mutagenesis of wild-type B56α was performed according to the protocol of Braman et al. (19). This strategy allows the mutation of B56α at single residues by direct use of the harboring cloning vector. In detail, 200 ng of DNA of either pAcGFP1-B56α, pIRES2-DsRed-Express-B56α, or pAcHLTA-B56α were amplified by DNA polymerase with proofreading activity (Expand Long Template PCR System, Roche Applied Science) and 27-mer complementary primer pairs, including mismatched bases (S41A sense primer, 5′-cgctcccagggcgcgtcgcagtttcgc-3′; S41D sense primer, 5′-cgctcccagggcgactcgcagtttcgc-3′). The sample was digested by DpnI for 2 h at 37 °C. After preparative agarose gel electrophoresis, the product bands were excised, and nucleic acids were extracted and purified with the Invisorb Spin DNA extraction kit (Invitek). Escherichia coli DH10B were transformed and plated on selective medium (pAcHLTA, ampicillin; pAcGFP1 and pIRES2-DsRed-Express, kanamycin). Colonies grown overnight were inoculated to liquid LB culture (including antibiotic agent), and plasmid DNA was extracted. The insertion of the mutation was verified by sequencing of the B56α coding region.
Sf21 insect cells were cultured at 27 °C in Grace's medium containing 10% FBS (v/v), 50 μg/ml gentamicin, and 2.5 μg/ml amphotericin B. Sf21 insect cells were co-transfected with a transfer plasmid (pAcHLTA or pVL1393) carrying cDNA of wild-type or mutated B56α and linearized wild-type baculovirus cDNA using the BaculoGoldTM kit (BD Biosciences). Recombinant baculoviruses were enriched by the plaque purification method (20) and used for infection of insect cells for recombinant expression of wild-type and mutated B56α. For this purpose, 9 × 106 cells were seeded into monolayer culture and then infected with recombinant baculoviruses encoding wild-type B56α or B56α mutants. 72 h after infection, the cells were centrifuged, and the corresponding pellets were stored at −20 °C.
HEK293 cells or Flp293 cells (derived from HEK293 cells that stably express the human α1A-adrenoreceptor (21)) were cultured (37 °C and 5% CO2) in Dulbecco's modified Eagle's medium (Invitrogen) with 10% FBS (v/v). Cells were transfected with expression vectors pAcGFP1-N1 or pIRES2-DsRed-Express (100–200 ng) carrying wild-type or mutated B56α using Lipofectamine 2000 (Invitrogen) following the manufacturer's protocol for transient transfection of adherent cells. 72 h after transfection, the degree of expression of the B56α-GFP fusion protein was tested by detection of the fluorescence signal using confocal microscopy (Zeiss LSM 710), and cells were washed in phosphate-buffered saline and harvested. Expression of PP2A subunits in cell pellets was also controlled by SDS-gel electrophoresis and immunoblotting. For stimulation experiments, cells were incubated with 10 μm phenylephrine (PE, agonist on α-adrenoceptors), 1 μm propranolol (blocks β-adrenoceptors), and 100 nm PMA (dissolved in 0.0001% dimethyl sulfoxide, DMSO) for the final 4 h before lysis.
For detection of the IP3-related fluorescence, Flp293 cells were seeded on round 18-mm diameter glass slides in 12-well plates (Nunc) and transfected using TurboFectTM in vitro transfection reagent (Fermentas) according to the manufacturer's protocol. Cells were transfected with 1 μg of pIRES2-DsRed-Express-B56α (wild-type or mutants S41A/S41D). 48 h post-transfection, cells were incubated with 4 μm of the calcium indicator Fluo-4 AM (Invitrogen) in Ca2+-free Tyrode's solution (140 mm NaCl, 4 mm KCl, 1 mm MgCl2, 5 mm HEPES, 10 mm glucose, pH 7.4) for 10 min in the incubation chamber. Then the dye was washed out using Ca2+-free Tyrode's solution. Detection of Fluo-4 fluorescence was recorded using a Zeiss LSM 710 confocal fluorescence microscope and ZEN software (excitation at 488 nm and Fluo-4 emission at 493–552 nm). Cells transfected with wild-type or mutant B56α pIRES-dsRed-Express constructs were identified by detection of dsRed fluorescence (excitation at 543 nm and dsRed emission at 552–747 nm). A possible spectral cross-talk between Fluo-4 and dsRed emissions was excluded by linear unmixing of the signals. 60 s after the start of the experiment, 100 μm PE in 2 mm Ca2+ Tyrode's solution was directly applied into the incubation chamber resulting in a final concentration of 1 mm Ca2+ and 50 μm PE.
Cultured HEK293 cells (see above) were maintained on 24-well plates (Corning Glass) and transfected using TurboFectTM in vitro transfection reagent (Fermentas) according to manufacturer's protocol. Cells were co-transfected with 200 ng of pAcGFP1-B56α (wild type or mutants), 200 ng of pAcGFP1-N1 as a negative control, 500 ng of the Photinus pyralis luciferase Icer promoter construct pP2Luc (22), and 50 ng of the control Renilla reniformis luciferase construct hRluc/TK (Promega) per well. 24 h post-transfection, cells were lysed and assayed using the Dual-Luciferase Reporter Assay System (Promega) as described in the manufacturer's protocol on a Mithras LB 940 microplate analyzer (Berthold Technologies). All luciferase results were normalized to the Renilla luciferase activity as internal control.
Sf21 insect cell pellets were resuspended in a buffer (50 mm Tris-HCl, 150 mm NaCl, 10 mm imidazole, pH 8.0) containing lysozyme, benzonase, and a proteinase inhibitor mixture (Roche Applied Science). Fragmentation of cells was achieved by sonification on ice. Soluble and insoluble fractions were separated by centrifugation at 20,000 × g for 30 min (4 °C). The purification of His6-tagged wild-type or mutated B56α was performed by use of Ni-NTA affinity column chromatography according to the manufacturer's instructions (Qiagen). The purification procedure was controlled by loading small samples of all fractions on 8% polyacrylamide gels that were stained with Coomassie dye. All elution fractions were pooled, desalted, and concentrated by AmiconTM ultracentrifugal filter units (Millipore). The recombinant proteins were found to be ~95% pure by use of this method.
Recombinant wild-type or mutated B56α was phosphorylated by PKCα in a Ca2+-dependent manner. The standard reaction mixture contained 25 mm Tris-HCl, pH 7.5, 5 mm MgCl2, B56α (0.5–2 μg per reaction), 5 mm NaF, 100 μm ATP, [γ-32P]ATP, 0.5 mm CaCl2, 100 μg/ml phosphatidylserine, and 20 μg/ml diolein. The reaction was initiated by addition of 0.2 μl of recombinant PKCα (Sigma), incubated up to 60 min at 30 °C, and terminated by addition of ice-cold EDTA (200 mm, pH 8.0). For determination of PKC activity, an aliquot was spotted on P81 phosphocellulose papers (Whatman), which were washed four times with 75 mm phosphoric acid. Filters were dried, and radioactivity was determined in a liquid scintillation counter. For visualization of PKCα-dependent phosphorylation of wild-type and mutated B56α, an aliquot of the reaction was added to 2.5% SDS buffer for the following separation of solubilized protein on 8% polyacrylamide gels. Gels were Coomassie-stained after electrophoresis, dried on a gel dryer (Bio-Rad), exposed for 24 h, and then analyzed by use of a STORM 860 (GE Healthcare). Moreover, aliquots of PKCα-phosphorylated B56α reactions, where NaF was omitted from the mixture, were dephosphorylated subsequently by either recombinant catalytic subunit of PP1 (Sigma) or purified heterodimeric PP2ACA (Millipore), consisting of the catalytic C and scaffolding subunit A. Where indicated, dephosphorylation was performed in the absence or presence of 500 μm inhibitor-2 of PP1 or 3 nm okadaic acid (inhibiting PP2A without affecting PP1 activity). Samples were analyzed by subjection to SDS-PAGE (see above) or detection of protein phosphatase activity.
Mass spectrometric (MS) analysis was performed using nanoAcquity UPLC (column BEH130 C18 inner diameter 100 μm, length 100 mm, pore size 130 Å, particle size 1.7 μm, 30 min water/acetonitrile gradient) coupled to Q-Tof Premier (Waters Corp.). For best sensitivity, the protein content of one individual gel band (after tryptic in-gel digestion) was used for the detection of a putative phosphorylation site by tandem MS on the selected m/z value.
Protein phosphatase activity was assayed using [32P]phosphorylase a as substrate as described previously (23). Protein phosphatase activity was determined on purified heterodimeric PP2ACA in the absence or presence of PKCα-(non)phosphorylated recombinant wild-type or mutated B56α. In addition, protein phosphatase activity was also measured in HEK293 cells transfected with cDNA of wild-type or mutated B56α. Briefly, cell pellets were sonicated at 4 °C six times for 10 s in buffer containing 4 mm EDTA, pH 7.4, and 15 mm 2-mercaptoethanol. Lysed cells were centrifuged at 14,000 × g for 20 min at 4 °C, and the supernatants were used for determination of phosphorylase phosphatase activity. The dephosphorylation reactions were initiated by adding [32P]phosphorylase a to a final concentration of 0.5 mg/ml (40,000 cpm/nmol) and carried out at 30 °C for 20 min, except that time-dependent measurements were performed up to 40 min. Where appropriate, 3 nm okadaic acid was added before the initiation of the reaction. No more than 18% of the substrate was utilized in the assay to ensure linearity of the reaction.
The amount of [32P]phosphorylase a binding to B56α was determined by incubation of nonphosphorylated or PKCα-phosphorylated His6-tagged wild-type B56α (4 μg each) with purified heterodimeric PP2ACA under conditions described above. After addition of [32P]phosphorylase a, the reaction mixture was incubated for 40 min at 30 °C and then added to 25 μl of Ni-NTA-agarose beads. Further incubation was performed for 2 h at 4 °C on a rocking platform. After centrifugation at 14,000 × g for 2 min, beads were washed three times with PBS, pH 7.4. Finally, bound His6-tagged B56α was eluted with 400 mm imidazole. SDS-solubilized eluted proteins were subjected to 10% polyacrylamide gels. The upper part of the gel was stained with Coomassie dye, dried, and then exposed for 24 h for further analysis of B56α-bound 32P-labeled phosphorylase a. The expression level of utilized B56α was tested by SDS-gel electrophoreses and subsequent immunoblotting.
Mouse cardiomyocytes were isolated as described previously (24). After incubation with 100 μm phenylephrine, 1 μm propranolol, and 100 nm PMA for 30 min at 37 °C, cells were shock-frozen in liquid nitrogen. Cardiomyocytes were resuspended in 4 °C cold PBS supplemented with 0.5% Triton X-100 and a protease inhibitor/phospho-STOP mixture (Roche Applied Science). Cell lysis was achieved by sonification six times for 10 s on ice. Thereafter, the homogenate was held on ice for an additional 30 min. Soluble and insoluble fractions were separated by centrifugation (14,000 × g, 1 h, 4 °C). For preclearing, the supernatant (0.5–1 μg/μl) was incubated with a mixture of protein A/G beads (Millipore) and 2 μg of normal rabbit IgG (Millipore) for 1 h at 4 °C. After centrifugation at 14,000 × g for 5 min, the supernatant was incubated overnight with a polyclonal antibody against B56α (Bethyl, 2 μg/mg supernatant protein) at 4 °C on a rocking platform. Protein A/G beads were added for an incubation period of 2 h at 4 °C. After centrifugation, beads were washed consecutively (three times) with PBS. The SDS-solubilized supernatant above the pelleted beads was processed for immunoblotting and analysis of the phosphorylation status using the ProQ Diamond phosphoprotein stain (18).
Protein expression analysis was performed on the following probes: 1) homogenates of wild-type and mutated B56α expressed in Sf21 insect (as His6-tagged proteins) or HEK293 cells (as GFP fusion proteins); 2) purified wild-type and mutated recombinant B56α (by Ni-NTA affinity column chromatography) and heterodimeric PP2ACA (Millipore); 3) immunoprecipitated mouse B56α; 4) mouse heart lysates; and 5) ventricular homogenates of nonfailing and failing human hearts. To this end, Sf21 and HEK cell pellets were resuspended in a buffer containing 30 mm histidine, pH 7.4, 250 mm sucrose, and a protease inhibitor mixture (Roche Applied Science). Lysed homogenates were mixed with equal volumes of reducing 5% SDS sample buffer and boiled for 5 min before loading. Purified or immunoprecipitated B56α as well as PP2ACA were also solubilized in SDS sample buffer for further processing by gel electrophoresis. In addition, 50 mg of mouse heart tissue were homogenized at 4 °C for 1 min in 0.5 ml of a medium containing 5% SDS and 10 mm NaHCO3 using a Polytron PT-10 (Kinematica). Homogenates were centrifuged at 14,000 × g for 20 min, and supernatant lysates were subjected to SDS-gel electrophoresis. Finally, human heart homogenates were prepared from frozen LV tissue as described previously (18). For immunoblot analysis of all proteins, 100–200 μg of individual samples were electrophoretically separated on 8 or 10% SDS-polyacrylamide gels (25, 26). After transfer of proteins to nitrocellulose, the blots were incubated with different antibodies raised against the following proteins: human B56α of PP2A (Bethyl, amino acids 25–75), human B56α of PP2A (Acris, amino acids 417–446), human phosphoserine 41 B56α (custom-made), human B56β of PP2A (Thermo Scientific, amino acids 434–497), human B56γ of PP2A (Bethyl, amino acids 490–524), human B56δ of PP2A (Bethyl, amino acids 1–50), mouse Aα of PP2A (Santa Cruz Biotechnology), human Cα of PP2A (PTG), canine calsequestrin (24), rabbit GAPDH (Ambion), and Tetra-His epitopes (Qiagen). The amounts of bound antibodies were detected by use of secondary antibodies (ECL rabbit/goat IgG, HRP-linked whole antibody, GE Healthcare). Signals were visualized and quantified with the ECL Plus detection system (ECL Plus, GE Healthcare) and the STORM blot imaging system, respectively.
Total RNA was extracted from transfected HEK293 cells with the use of TRIzol® (Invitrogen). Total RNA (1 μg) was randomly reversely transcribed to cDNA using Transcriptor First Strand cDNA synthesis kit (Roche Applied Science). The real time RT-PCR was carried out using a LightCycler 480 System (Roche Applied Science), and detection was performed by the use of carboxyfluorescein-labeled universal ProbeLibrary probes (UPL, Roche Applied Science). Primers (Invitrogen) were designed using ProbeFinder Assay Design Center (Roche Applied Science). The PCRs were set up in a 96-well plate in a volume of 20 μl. The reaction components were 1 μl of undiluted cDNA, 10 μl of LightCycler® 480 Probes Master, 4 μl of H2O, and 0.8 μl for each primer and 0.4 μl for the UPL probe (all 10 μm). Reactions were incubated at 95 °C for 10 min followed by 45 cycles at 95 °C for 10 s, 60 °C for 30 s, and 72 °C for 1 s. Relative levels of particular cDNAs were determined with the help of LightCycler® 480 software with appropriate calibration curves obtained with different amounts of control cDNAs. Crossing points were determined by using the second derivative method. Relative quantification was performed by calculating relative expression ratios using the ΔΔCT method and the relative expression software tool (REST© Version 2.013; see Ref. 27, 28). Random statistical analysis was performed with 10,000 iterations, and hypoxanthine-guanine phosphoribosyltransferase (Hprt) was used as a reference gene.
Statistical differences between groups were calculated by analysis of variance or Kruskal-Wallis one-way analysis of variance on ranks followed by Bonferroni's or Dunn's post hoc tests, respectively. p < 0.05 was considered significant. Statistical analysis of real time PCR data were performed using REST© software (REST© Version 2.0.13; see Refs. 27, 28). Random statistical analysis was performed with 10,000 iterations.
To study whether the regulatory subunit B is phosphorylated by PKC in vivo, isolated cardiomyocytes were treated with phenylephrine, propranolol, and PMA to achieve a maximum activation of PKCα (representing the main cardiac isoform). B56α was enriched in supernatants of lysed cells by immunoprecipitation, eluted from protein A/G beads, and separated by SDS-gel electrophoresis. The prominent phosphorylated mobility form detected by ProQ Diamond phosphoprotein stainings at a molecular mass of 56 kDa corresponds to mouse B56α (Fig. 1A). This was the first indication of a PKC-dependent phosphorylation of B56α. In the next step, we tested whether B56α is also phosphorylated in vitro. For this purpose, we generated recombinant B56α by an insect cell expression system. The N-terminal His6-tagged B56α was purified by Ni-NTA affinity column chromatography. An example of a Coomassie-stained gel of the purification procedure is given for wild-type B56α (Fig. 1B). After elution, a single band at the expected height of the fusion protein was present in the desalted and concentrated fraction. This fraction was used for further expression analysis by immunoblotting using specific antibodies directed against the N-terminal region of B56α or the His6 tag (Fig. 1C).
The purified recombinant B56α was subjected to in vitro phosphorylation by PKCα in the presence of [γ-32P]ATP. The phosphorylation of B56α by PKCα was time-dependent (Fig. 2A) reaching its maximum after 30 min. The maximal phosphorylation stoichiometry obtained after 30 min of incubation with PKCα was ~2.5 mol of phosphate incorporated per mol of B56α either in the absence or presence of purified heterodimeric PP2ACA. The 32P-labeled wild-type B56α was detected by SDS-PAGE in an aliquot of the 60-min fraction (Fig. 2A). To identify the site of 32P incorporation into B56α, we subjected the PKCα-phosphorylated protein to MS analysis specifically targeting peptides of interest. For peptide 38SQGSSQFR45, phosphorylation was indicated at its fourth amino acid, Ser41, by the gas phase fragmentation experiment (m/z 488.7 in Fig. 2B). This amino acid residue was predicted as a potential PKCα phosphorylation site by GPS 2.1, a group-based prediction system for kinase-specific phosphorylation sites (Fig. 2C) (29). Of note, this residue as well as adjacent amino acids are conserved between mammalian species, birds, fish, and amphibians (Fig. 2C).
The newly identified PKCα phosphorylation site was confirmed by mutagenesis of the Ser41 residue to alanine (S41A). After a 30-min incubation period, the incorporation of 32P into the B56α mutant S41A was reduced by 80% compared with the wild-type form (Fig. 3A). The remaining 32P signal may result from unspecific phosphorylation of non-PKC sites. Moreover, we cannot exclude the presence of additional PKCα phosphorylation sites, which so far has possibly eluded detection due to the low abundance of phosphorylated peptides in the protein digest. Nevertheless, this has no impact on the inhibitory potency of B56α on PP2A. The PKCα-dependent phosphorylation of B56α was also confirmed by an antibody specific for the phosphorylated Ser41 site of human B56α. The B56α phosphoserine 41 antibody recognized recombinant wild-type B56α phosphorylated with PKCα (Fig. 3B). The nonphosphorylated wild-type B56α was also detected by this antibody. However, this was to a much lesser extent, suggesting a nonexclusive specificity for phosphoserine 41 of B56α. Antibodies against the C-terminal region (Fig. 3B) or N-terminal region (data not shown) of B56α detected a comparable loading of nonphosphorylated and PKCα-phosphorylated protein samples. The phospho-specific antibody was also studied on homogenates of human LV tissue of nonfailing and failing hearts (Fig. 3C). We detected a higher phosphorylation level of B56α at Ser41 in failing hearts suffering from ischemic or dilated cardiomyopathy compared with nonfailing hearts (Fig. 3D). In contrast, the expression of total B56α was reduced by 61% in DCM hearts (Fig. 3E). Thus, the ratio of phosphorylated to total B56α was increased by 220 and 600% in ICM and DCM hearts, respectively, compared with NF hearts (Fig. 3F). The expression of GAPDH, which served as a control, was unchanged between all groups (data not shown).
To investigate whether the phosphorylation of B56α can be reversed, PKCα-phosphorylated wild-type B56α was incubated with different protein phosphatases. Dephosphorylation of B56α was reversed by PP2A but not PP1, another major serine/threonine protein phosphatase (Fig. 4A). The specificity of a PP2A-dependent dephosphorylation was further confirmed by preincubation with 3 nm okadaic acid, which totally blocks the enzyme activity. In addition, the inhibition of PP1 by one of its potent endogenous inhibitors, inhibitor 2, had no influence on the phosphorylation level of wild-type B56α (Fig. 4A). PP2A subunits of the dephosphorylation reaction were detected by immunoblotting using specific antibodies (Fig. 4B). The ratio of all PP2a subunits, namely the purified heterodimeric PP2ACA and the recombinant wild-type B56α, is denoted. Mouse heart homogenate was loaded as a control demonstrating that the tagged recombinant B56α fusion protein runs higher on SDS gels than endogenous native B56α (Fig. 4B).
To test whether B56α can influence the PP2A activity per se, we incubated different amounts of recombinant wild-type B56α with purified heterodimeric PP2ACA. B56α inhibited PP2A activity on phosphorylase a in a concentration-dependent manner (Fig. 5A).
Wild-type B56α displayed high inhibitory activity on PP2ACA with an IC50 of 2 nm. PP2A activity was completely inhibited by wild-type B56α at concentrations of ~100 nm. In the next step, we studied whether phosphorylation by PKCα can modulate the inhibitory effect of B56α on PP2A activity. The PKCα-phosphorylated form shifted the dose-response curve to the left resulting in an IC50 of 0.5 nm (Fig. 5A). To evaluate the mechanism(s) of PP2A inhibition, we measured the interaction between [32P]phosphorylase a and B56α in the presence of purified PP2ACA. The binding of 32P-labeled phosphorylase a was lower in PKCα-phosphorylated B56α compared with nonphosphorylated B56α (Fig. 5B). The assay was performed under conditions where the difference in PP2A inhibition was maximal between nonphosphorylated and phosphorylated B56α. The B56α mutant S41D reduced PP2A activity on phosphorylase a to a comparable left shift as seen under application of PKCα-phosphorylated wild-type B56α (Fig. 5C). Therefore, S41D mimics phosphorylated wild-type B56α. S41D exhibited a high inhibitory activity on PP2ACA with an IC50 of 0.5 nm, which is the same potency as measured for phosphorylated wild-type B56α. These data suggest that the activity of the heterodimer PP2ACA depends directly on the PKCα-mediated phosphorylation of B56α. The mutant S41A, lacking the PKCα phosphorylation site, had an inhibitory potency similar to that of wild-type B56α, which under the same condition was ~1.5 nm (Fig. 5D). When Ser41 of B56α was mutated to alanine, the inhibitory effect of PKCα phosphorylation on PP2A activity was abolished (Fig. 5D), indicating that this phosphorylation site is crucial for regulation of PP2A by PKCα-dependent phosphorylation of B56α. In addition, recombinant (non)phosphorylated wild-type B56α had no effect on PP1 activity (data not shown).
The effects of PKCα-dependent B56α phosphorylation at Ser41 on type-2A phosphatase activity were further confirmed in vivo by measuring total protein phosphatase activity in HEK293 cells transfected with the expression vector pAcGFP1 carrying either wild-type or mutated B56α (S41A or S41D). The construction of the expression vector allows the visual control of the transfection efficiency by GFP fluorescence (Fig. 6A). By use of this vector, we achieved abundant expression of recombinant B56α-GFP fusion proteins (Fig. 6B). The endogenous B56α expression in HEK293 cells was not influenced by transfection of the empty vector, and it was not detectable in cells exhibiting expression of recombinant B56α. The expression of exogenous B56α (Table 1) and of endogenous scaffolding Aα and catalytic Cα subunits was not different between wild-type and mutated B56α (Fig. 6C). Moreover, exogenous B56α did not affect the protein expression of other B56 subunits (Fig. 6D and Table 1). The mRNA expression of single B56 subunits was only slightly reduced in transfected HEK293 cells expressing wild-type B56α or S41A (Table 1). Confocal microscopy revealed that recombinant B56α is located mainly in the cytosol of transfected HEK293 cells (Fig. 6E). To determine whether the depression in protein phosphatase activity associated with phosphorylation of B56α at Ser41 corresponds to similar changes in an in vivo system, total phosphatase activity was measured in homogenates of transfected HEK293 cells (Fig. 6F). Stimulation of PKC by phenylephrine and PMA (30) resulted in a reduced total phosphatase activity in cells transfected with wild-type B56α. This decrease was abolished in homogenates of HEK293 cells expressing S41A. When cells were transfected with pAcGFP1 carrying the B56α mutant, S41D, total phosphatase activity was already depressed under basal conditions to the same level as observed in HEK293 cells transfected with wild-type B56α after stimulation. Pharmacological stimulation of S41D-transfected HEK293 cells was not able to further augment the inhibition of total phosphatase activity (Fig. 6F).
Protein phosphatase activity in transfected HEK293 cell homogenates was also assayed in the presence of okadaic acid, which allows the discrimination between PP1 and PP2A activity. Purified PP2ACA was completely inhibited at a concentration of 3 nm okadaic acid, which selectively inhibits PP2A (Fig. 7A). Thus, the remaining protein phosphatase activity in HEK293 cells, transfected with the empty pAcGFP1 vector, represents PP1 activity. By use of 3 nm okadaic acid, we found decreased PP2A activity in wild-type B56α-transfected HEK293 cells after stimulation of PKC with both phenylephrine and PMA (Fig. 7B). This effect on PP2A activity was abolished in cells expressing S41A. Consistently, expression of S41D, the pseudophosphorylated form of B56α, was associated with a depressed PP2A activity under basal conditions compared with wild-type B56α (Fig. 7B). Under conditions of maximal PKC stimulation, no additional effect on PP2A activity was measurable. Taken together, these findings indicate that phosphorylation of B56α at Ser41 and resulting regulation of PP2A occurs not only in vitro but also in living cells.
Finally, the functional relevance of PKCα-dependent phosphorylation of B56α was examined in regard to downstream cellular signaling. For this purpose, α1A-adrenoreceptor-expressing Flp293 cells were transfected with wild-type or mutated B56α expression constructs using the pIRES2-DsRed-Express vector that allows a discrimination between transfected (red fluorescence) and nontransfected control cells (Fig. 8A). The IP3-induced Fluo-4 fluorescence was detected in transfected and nontransfected cells under application of 50 μm phenylephrine (Fig. 8B). Transfection of Flp293 cells with wild-type B56α resulted in a 22% lower IP3-mediated Ca2+ release measured as the F1/F0 peak amplitude of individual Ca2+ spikes (Fig. 8C). Lack of the PKCα phosphorylation site by mutation of Ser41 to alanine resulted in an ER Ca2+ release amplitude that was comparable with the wild-type form of B56α. This is consistent with the similar basal inhibitory activity of both nonphosphorylated B56α isoforms (wild type and S41A) on PP2ACA (Fig. 5, A and D) and the comparable PP2A activity under basal (nonstimulated) conditions in HEK cells transfected with wild-type and mutated (S41A) B56α (Fig. 7B). We co-transfected Flp293 cells also with a pIRES2-DsRed-Express construct expressing the B56α mutant S41D, which imitates a constitutive phosphorylation by PKCα at this site. This mutation enhanced the ER Ca2+ release amplitude by 23% compared with wild-type B56α resulting in similar Fluo-4 fluorescence F1/F0 peak amplitudes in transfected and nontransfected cells (Fig. 8C). Ca2+ decay kinetics were unchanged between all groups (Fig. 8D). In addition, application of 1 μm isoprenaline was not sufficient to elicit Ca2+ spikes in α1A-adrenoreceptor-expressing Flp293 cells (data not shown). Thus, our data suggest that PKCα can modulate PP2A activity and thereby IP3R activity in vivo by phosphorylation of B56α.
Furthermore, we investigated the relevance of the Ser41 phosphosite for transcriptional gene regulation by expression of either wild-type or mutated B56α together with a luciferase reporter gene construct containing the inducible cAMP early repressor (Icer) promoter in HEK293 cells. Lack of the PKCα phosphorylation site by mutation of Ser41 to alanine, S41A, resulted in a luciferase expression that was comparable with the wild-type form of B56α (data not shown). This is consistent with the similar inhibitory activity of both nonphosphorylated B56α isoforms (wild-type and S41A) on PP2ACA (Fig. 5, A and D). In contrast, transcriptional activity of the Icer promoter was decreased by 20 ± 2% in S41D-transfected HEK293 cells (p < 0.05, n = 20 from three independent transfections).
The findings of this study provide in vitro and in vivo evidence for the functional relevance of a novel phosphorylation site in B56α at Ser41. The regulatory function of B56α phosphorylation on PP2A activity was confirmed by several independent lines of evidence using a mutant B56α subunit (S41A) that cannot be phosphorylated by PKCα and by the generation of a specific antibody against Ser41-phosphorylated B56α. Ser41 phosphorylation was increased in LV tissue of failing human hearts. PP2A activity was inhibited by PKCα-dependent phosphorylation at this B56α phosphosite. The decrease in PP2A activity was observed both in vitro on recombinant B56α and in cultured HEK293 cells by use of a B56α mutant (S41D) mimicking phosphorylated B56α. The transient transfection of wild-type B56α decreased IP3-mediated ER Ca2+ release in cultured Flp293 cells, although this decrease was not observed after transfection of the S41D mutant. Thus, the PKCα-dependent phosphorylation of B56α regulates the ER Ca2+ release.
Previous work has demonstrated that most of the B56 subunits are phosphoproteins, with phosphorylation on serine (6). This is in contrast to the regulation of the PP2A catalytic subunit C, occurring upon both serine/threonine and tyrosine phosphorylation. The phosphorylation of B56 subunits by different protein kinases can affect either the assembly of the PP2A heterotrimer or its enzymatic activity depending on the localization of the phosphosite. Letourneux et al. (31) found that ERK-phosphorylated B56γ1 exhibited a decreased binding to A and C subunits. This is in contrast to a previous report showing that the formation of the PP2A trimeric holoenzyme occurred independently of its phosphorylation state (6). ERK-dependent phosphorylation of B56γ1 was associated with a lower PP2A activity on ERK and was localized to phospho-Ser327 (31). The phosphorylation of B56γ3 by checkpoint kinase Chk2 resulted in an increase in PP2A activity against myelin basic protein as well as on autophosphorylated Chk2 (32). An increase of PP2A activity was also observed for phosphorylation of B56α by PKR (11) and of B56δ by PKA or PKC at Ser566 (12, 33). In this study, we found a depressed PP2A activity after a PKCα-dependent phosphorylation of B56α at Ser41. A lower PP2A activity was also observed after phosphorylation of recombinant wild-type B56α by PKA, whereas CaM kinase II-phosphorylated B56α enhanced PP2A activity (data not shown). This mechanism may explain, at least in part, the increased PP2A activity in CaM kinase II-overexpressing hearts of transgenic mice (34). Thus, besides the diversity of B56 subunits, the phosphorylation of B56 isoforms may contribute to the substrate specificity of PP2A trimeric holoenzymes. Moreover, the phosphorylation studies raise the question on how the phosphorylation of B56 isoforms at multiple phosphoacceptor sites may influence the catalytic activity of PP2A by structural changes.
The B56 subunits exhibit a superhelical structure comprising 18 α-helices, which are organized into eight HEAT-like repeats (35). There are a number of highly conserved amino acid residues to interact with the A and C subunits. The interaction to the A subunit is mediated by two binding domains located on the convex side of B56 subunit C-terminal HEAT-like repeats. The interface between both subunits is relatively weak but is enhanced by binding of the methylated C-terminal tail of the C subunit to form a stable heterotrimeric structure. The interaction with the C subunit is mediated by three interfaces, including the intra-repeat loop 2, HEAT-like repeats 4–6, and intra-repeat loops of HEAT repeats 6–8 of B56γ1 (35). However, there are no direct contacts between B56γ and residues near the active site of the catalytic subunit C, suggesting that regulatory B56 subunits cannot directly modify the catalytic site or activity. Thus, the major role of B56 isoforms is rather to tether the PP2A holoenzyme complex to its substrates. This is caused by formation of the PP2A heterotrimer bringing the concave surface of B56 subunits to the active site pocket of the catalytic subunit C. The acidic concave may shape a surface that recruits different phosphoproteins through electrostatic interactions (36). Therefore, the phosphorylation of B56 subunits at different sites at the concave surface, giving an additional negative charge to the modified protein, may favor charge-charge repulsion or attraction between B56 isoforms and the phosphorylated substrates leading to a limited or improved access of PP2A substrates to the catalytic site, respectively (Fig. 9A). This hypothesis is supported by data from this study showing that phosphorylase a exhibited a lower binding to PKCα-phosphorylated B56α compared with nonphosphorylated B56α, which might explain the reduced PP2A activity when extra-charged B56α (= phosphorylated by PKCα) was applied (Figs. 5B and and99A). Thus, different protein kinases can compel B56α-associated PP2A to a lower (by PKCα) or higher (by CaM kinase II) activity depending on the phosphorylation at different specific phosphosites of the regulatory subunit. Alternatively, the assembly of the PP2A trimeric holoenzyme may be disturbed due to the charge repulsion between the PKCα-phosphorylated B56α and a negatively charged surface on the A subunit binding area. The phosphorylation of B56α at Ser41 occurs at the N-terminal tail of the protein, a region that is unlikely to be involved in the formation of the PP2A heterotrimer. However, it may affect transient interactions between the B56α core domain and the AC dimer because B56 isoforms, truncated at the N-terminal sequence, did not interact with A and C subunits in vivo (37–39).
The functional relevance of the newly identified phosphosite in B56a was examined by transfection of cultured Flp293 cells with a pseudophosphorylated B56α mutant (S41D), which resulted in a 23% increase in the F1/F0 amplitude of Fluo-4 fluorescence compared with wild-type B56α. This increase was paralleled by a reduced PP2A activity under basal conditions in Flp293 cell homogenates (Fig. 7B) suggesting that the IP3R function is modulated by the PKCα-dependent phosphorylation of B56α. PP2A as well as PP1 are components of the IP3R macromolecular signaling complex (40). It was demonstrated that PP1 facilitates the dephosphorylation of the PKA-phosphorylated IP3R, reversing the increase of the channel sensitivity to activation by IP3 (41). The physiological significance of PP2A in regulating the IP3R function remained unclear because both basal and dopamine-induced PKA phosphorylation of the IP3R were not affected by 10 nm okadaic acid (41). However, our data rather suggest a more sophisticated model of IP3R regulation by the phospho-B56α/PP2A heterotrimer than a simple on-off of the catalytic activity of the PP2A heterodimer. This focuses on a potential role of protein kinases (e.g. PKCα) in influencing PP2A activity. Normally, PKCα and PP2ACA heterodimers are co-localized in the cytosol of resting cells (17, 42). Activation of PKCα by (auto)phosphorylation or PMA induces its translocation along with PP2A because of their physical association from the cytosol to the membrane (17, 42, 43). From these studies, it can be suggested that a PKCα-dependent phosphorylation of B56α, which is mimicked by the B56α mutant S41D, results in a lower membrane-associated PP2A activity. This is followed by a higher PKA-mediated phosphorylation and activation of the IP3R, i.e. an increased ER Ca2+ release (Fig. 9B). The presence of PP2A also correlated with a PKCα phosphatase activity in membrane fractions suggesting the inhibition and desensitization of PKCα through dephosphorylation by PP2A (14). In addition, not only PKCα but also PKCα-phosphorylated B56α is a target for PP2A leading to dephosphorylation of the regulatory subunit as shown in our studies. This suggests a multitude of autoregulatory mechanisms to restore normal PP2A activity as demonstrated for PKR-dependent phosphorylation of B56α (11). It is also conceivable that the increased ER Ca2+ release in S41D-expressing Flp293 cells leads to activation of the CaM kinase II, which phosphorylates B56α (data not shown) and reverses the inhibition of the PP2A activity. Alternatively, the activation of PP2B by the released Ca2+ may dephosphorylate DARPP-32, which relieves the inhibition of PP1 and closes the feedback loop by dephosphorylating the IP3R (41).
Previous findings showed that the PP2A is able to dephosphorylate the transcription factor cAMP-response element-binding protein (CREB) at Ser133, which inhibits its transcriptional activity (44). A prominent target gene promoter of CREB is the inducible cAMP early repressor (Icer) (45). Accordingly, an inhibition of PP2A by the B56α mutant S41D should increase Icer promoter activity. However, a luciferase reporter gene assay in HEK293 cells revealed that expression of S41D is followed by decreased transcriptional activity of the Icer gene promoter compared with wild-type B56α. Shanware et al. (46) reported that B56γ-associated PP2A inhibits CREB hyperphosphorylation at Ser121 resulting in a decreased binding of CREB to the CREB-binding protein. The histone acetyltransferase activity of CREB-binding protein is able to enhance the ability of CREB to activate the transcription of its target genes (47). Hence, an inhibition of PP2A activity by the B56α mutant S41D, which is associated with a hyperphosphorylation of CREB at Ser121 and a decreased interaction with CREB-binding protein, may explain the observed decreased Icer promoter activity. Although the exact mechanism underlying the inhibition of cAMP-response element-mediated gene transcription by S41D is not fully understood, the altered promoter activity suggests the possible relevance of the Ser41 phosphosite of B56α for the regulation of gene transcription.
The higher phosphorylation level of B56α at Ser41 in failing human hearts suggests an important role of the PKCα-dependent phosphorylation of B56a not only under physiological conditions (e.g. modulation of IP3R function) but also in disease states. There is evidence that PKCα activity regulates cardiac contractility. The phosphorylation of PKCα substrates (e.g. inhibitor-1 of PP1, G-protein-coupled receptor kinase 2, cardiac troponin I) is thereby associated with a reduced sarcoplasmic reticulum Ca2+ load, an uncoupling of β-adrenergic receptors, a lower myofilament Ca2+ sensitivity, and a decreased contractility (48). Consistently, the increased PKCα expression and activity contributes to myofilament dysfunction in failing human hearts and in experimental congestive heart failure (49, 50). Moreover, heart-directed overexpression or adenovirus-mediated gene transfer of PKCα resulted in a reduced ventricular performance (51). Thus, it is conceivable that a depressed PP2A activity due to a higher PKCα-dependent phosphorylation of B56α may contribute to the effects of increased PKCα on cardiac troponin I phosphorylation in hypertrophy and end-stage heart failure. A model of protein phosphatase regulation was demonstrated for the PKCα-dependent phosphorylation of PP1 inhibitor-1 leading to inhibition of sarcoplasmic reticulum Ca2+ uptake and diminished contractile response (51). These authors did not detect changes in PP2A activity in homogenates of PKCα-overexpressing hearts. However, this does not rule out the possibility that PP2A activity is altered in certain subcellular compartments relevant for cardiac contractility and/or progression of heart failure. The stimulation of PKCα, which is physically associated with PP2A, initiated its translocation to membrane fractions in general (17, 42, 43). In cardiomyocytes, the activation of PKCα by phorbol esters was followed by the translocation of the enzyme to the T-tubular membrane network (51) where phosphosubstrates of PP2A (e.g. ryanodine receptor type 2, L-type Ca2+ channel, phospholamban) are localized. Consistently, PKCα overexpression in REH cells was associated with a depressed PP2A activity in mitochondrial membrane fractions (52). The lower PP2A activity was not paralleled by a decreased expression of the catalytic subunit Cα but a reduction of the B56α protein levels indicating the redistribution of PP2A subunits. Thus, taking into account the close physical association of PKCα and PP2A as well as the wide ranging effects of this protein kinase-phosphatase complex in the heart, the investigation of the exact mechanism of how PKCα regulates PP2A and vice versa under (patho)physiological conditions is of considerable interest in future studies.
In summary, here we present data showing for the first time that PP2A activity is regulated by a newly identified phosphosite on B56α. The phosphorylation at Ser41 by PKCα converts B56α into a more potent inhibitor of PP2A. The inhibition of PP2A activity was associated with an increase in the IP3-mediated ER Ca2+ release and an altered cAMP-response element-mediated transcriptional activity in vivo. Moreover, the PKCα-dependent Ser41 phosphorylation of B56a may play an important role in the pathophysiology of human heart failure.
We thank N. Hinsenhofen and N. Nordsiek-Goda for excellent technical assistance.
*This work was supported by the Innovative Medizinische Forschung Münster Grant KI 2 1 10 10 and Deutsche Forschungsgemeinschaft Grant Mu1376/11-1.
2The abbreviations used are: