Regulation of PP2AC Carboxylmethylation and Cellular Localisation by Inhibitory Class G-Protein Coupled Receptors in Cardiomyocytes

The enzymatic activity of the type 2A protein phosphatase (PP2A) holoenzyme, a major serine/threonine phosphatase in the heart, is conferred by its catalytic subunit (PP2AC). PP2AC activity and subcellular localisation can be regulated by reversible carboxylmethylation of its C-terminal leucine309 (leu309) residue. Previous studies have shown that the stimulation of adenosine type 1 receptors (A1.Rs) induces PP2AC carboxylmethylation and altered subcellular distribution in adult rat ventricular myocytes (ARVM). In the current study, we show that the enzymatic components that regulate the carboxylmethylation status of PP2AC, leucine carboxylmethyltransferase-1 (LCMT-1) and phosphatase methylesterase-1 (PME-1) are abundantly expressed in, and almost entirely localised in the cytoplasm of ARVM. The stimulation of Gi-coupled A1.Rs with N6-cyclopentyladenosine (CPA), and of other Gi-coupled receptors such as muscarinic M2 receptors (stimulated with carbachol) and angiotensin II AT2 receptors (stimulated with CGP42112) in ARVM, induced PP2AC carboxylmethylation at leu309 in a concentration-dependent manner. Exposure of ARVM to 10 µM CPA increased the cellular association between PP2AC and its methyltransferase LCMT-1, but not its esterase PME-1. Stimulation of A1.Rs with 10 µM CPA increased the phosphorylation of protein kinase B at ser473, which was abolished by the PI3K inhibitor LY294002 (20 µM), thereby confirming that PI3K activity is upregulated in response to A1.R stimulation by CPA in ARVM. A1.R-induced PP2AC translocation to the particulate fraction was abrogated by adenoviral expression of the alpha subunit (Gαt1) coupled to the transducin G-protein coupled receptor. A similar inhibitory effect on A1.R-induced PP2AC translocation was also seen with LY294002 (20 µM). These data suggest that in ARVM, A1.R-induced PP2AC translocation to the particulate fraction occurs through a GiPCR-Gβγ-PI3K mediated intracellular signalling pathway, which may involve elevated PP2AC carboxylmethylation at leu309.


Introduction
The type 2A protein phosphatase (PP2A) is a serine/threonine protein phosphatase that is ubiquitously expressed in all eukaryotic cells and may account for up to 1% of total cellular protein [1]. A significant proportion (,30%) of PP2A consists as a scaffold (A) subunit complexed with the catalytic (C) subunit to form PP2A A/C heterodimers [2]. The PP2A A/C heterodimer subunits provide a platform for the binding of a third component, the regulatory B subunit, which facilitates ''targeting'' of the heterotrimeric holoenzyme towards target substrates [3].
The PP2A catalytic subunit undergoes reversible post-translational phosphorylation and carboxylmethylation, both of which can alter catalytic activity and cellular distribution of PP2A. Phosphorylation of threonine304 and tyrosine307 residues in the carboxyl terminus of the protein is associated with inhibition of PP2A C activity [4,5]. Carboxylmethylation of PP2A C occurs at the C-terminal leucine309 (leu309) and is catalysed by leucine carboxylmethyltransferase-1 (LCMT-1), a member of a large methyltransferase family of enzymes that utilise S-adenosyl methionine (SAM/AdoMet) as a universal methyl donor [6]. The carboxylmethylation of leu309 increases the binding affinity of the PP2A A/C heterodimer for some, but not all, regulatory B subunits, which have been classified into four separate sub families and are encoded by 15 human genes: PPP2R2/B (A-D), PPP2R5/ B9 (A-E), PPP2R3/B0 (A-C) and the striatins/B90 (1, 3 and 4). Through alternative gene splicing, several of these genes can generate a number of splice variants, resulting in the expression of 20+ regulatory B subunits [7]. The importance of PP2A C leu309 carboxylmethylation by LCMT-1 for recruitment of regulatory B subunits to the PP2A A/C heterodimer can be considered a sliding scale, whereby it is an absolute prerequisite for PPP2R2/B subunit recruitment by the PP2A A/C heterodimer and progressively less important for the recruitment of PPP2R5/B9, PPP2R3/B0 and the striatin/B90 subunits to PP2A A/C [8].
The carboxylmethylation of PP2A C on leu309 is reversed by the protein phosphatase methyl esterase PME-1 [9], an enzyme found to be associated with an inactive and demethylated pool of PP2A C subunits [10]. Evidence suggests that PME-1 can displace the two metal ions from the active site that are required for PP2A C activity, thereby inhibiting PP2A C activity in a demethylation-independent manner (6). This inactive pool of PP2A C can also be reactivated by a phosphotyrosyl phosphatase activator (PTPA) in vitro [10]. The mechanism of PP2A C reactivation by PTPA remains undefined, however, evidence suggests that the PTPA-mediated peptidylprolyl cis/trans isomerase activity and consequent conformational change to the structure of PP2A C seems likely [11].
Structural X-ray crystallographic studies [12][13][14] have shown that the last 6 C-terminal amino acids (304TPDYFL309) of PP2A C interact with both the active site and a unique lid domain of LCMT-1, the latter of which may confer some degree of substrate specificity. These 6 C-terminal amino acids of PP2A C are highly conserved between species [15] and occupy the deep active site pocket of LCMT-1, which is facilitated by interaction between the catalytic sites of active PP2A C and LCMT-1, thereby suggesting that PP2A C can only be carboxylmethylated by LCMT-1 once in an active conformation [16]. This ensures that only active PP2A C subunits can be carboxylmethylated and is therefore thought to prevent uncontrolled PPP2R2/B-PP2Amediated dephosphorylation [16].
The importance of understanding the regulation of cellular PP2A C activity is confirmed by the number of human diseases such as diabetes [17], cancer [18], Alzheimer's disease [19] and heart failure [20,21] which appear to share (in part) a PP2A-based aetiology. Despite this, the role and regulation of PP2A C carboxylmethylation in cells is poorly understood. In the heart, several studies [22][23][24][25] demonstrate that stimulation of adenosine type 1 receptors can induce not only PP2A C carboxylmethylation but also PP2A C translocation to the membrane-rich particulate fraction of ventricular myocytes. Furthermore, previous work by ourselves [23] and others [22] has shown that the phosphorylation status of selected proteins that are expressed within the membranerich particulate fraction of ventricular myocytes is reduced, thereby altering the function of these proteins. Since type 2A protein phosphatases represent the only identifiable LCMT-1 substrates [26], it is surprising that this important LCMT-1/PME-1/PP2A C intracellular signalling axis has not been more extensively studied in cardiac cells.
Hence, the current study investigates the intracellular signalling mechanisms through which G i PCR stimulation regulates PP2A C carboxylmethylation and subcellular distribution in ARVM. We demonstrate that (i) the LCMT-1/PME-1/PP2A C /B55a intracellular signalling axis is mainly localised in the cytoplasm, (ii) PP2A C carboxylmethylation can be induced by multiple G i PCR agonists, (iii) A1.R stimulation increases the cellular association between PP2A C and LCMT-1 and (iv) A1.R-induced PP2A C translocation to a membrane-rich particulate fraction occurs through a Gbc -PI3K pathway.

Materials and Methods
Animal tissue used in this study was obtained in accordance with the UK Home Office Guidance on the Operation of the Animals (Scientific Procedures) Act 1986, published by Her Majesty's Stationery Office, London, and approved by the Institutional Animal Care and Use Committee (IACUC) of King's College London. Healthy animals were sacrificed by a schedule one procedure completed by a home office licensed individual such that animal suffering was categorised as minimal.
Cell culture of adult rat ventricular myocytes ARVM were isolated from the hearts of adult male Wistar rats (200-250 g, B & K Universal Ltd) by collagenase-based enzymatic digestion, as previously described [28]. The isolated myocytes were pelleted by brief centrifugation at 50 g and washed at room temperature with M199 culture medium (Invitrogen) containing 100 IU/ml penicillin/streptomycin and 10% FBS. Following further centrifugation at 50 g, ARVM were resuspended in modified M199 containing (in mM) creatine 2, carnitine 2, taurine 5, 100 IU/ml penicillin/streptomycin and 10% FBS. ARVM were then added to 6-well cell culture plates, that had been earlier coated with laminin as previously described [28] and allowed to adhere for 90 minutes in an incubator (37uC, 5% CO 2 ). The culture medium was then replaced with fresh modified M199 medium, prior to adenoviral infection.

Adenoviral infection of adult rat ventricular myocytes
The recombinant adenovirus which co-expresses enhanced green fluorescent protein (EGFP) with the bovine Galpha subunit of the transducin GPCR (AdV:Ga t1 ) was constructed using the AdEasy system [29] and a kind gift from Professor T Wieland, University of Heidelberg [27,30,31]. Adenovirus expressing EGFP alone (AdV:EGFP) was used as a control. The recombinant adenoviruses were amplified in HEK-293 cells and purified over CsCl 2 gradients as previously described [28], which produced high-titre viral stocks of 3-5610 9 plaque forming units (pfu)/ml, as determined by a serial end-point dilution assay [32]. ARVM were infected with AdV:EGFP (control) or AdV:Ga t1 at a multiplicity of infection (MOI) of 50 pfu/ARVM (which provided .90% transduction efficiency). The adenovirus-containing medium was then removed after 60 min and replaced with fresh modified M199 medium. ARVM were maintained in an incubator (37uC, 5%CO 2 ) and were used for experiments 18 h after infection.
Stimulation of adult rat ventricular myocytes with G i PCR agonists ARVM were exposed to the angiotensin II type 2 receptor agonist CGP42112 (0.01-100 nM, 10 min), the muscarinic receptor agonist carbachol (0.01-100 mM, 10 min) or adenosine type 1 receptor (A1.R) agonist CPA (0.01-100 mM, 10 min), to stimulate their cognate GPCRs. Control cells received vehicle (phosphate buffered saline or 0.1% DMSO in PBS) for an identical period. At the end of the exposure period, ARVM were lysed with either Laemmli SDS-PAGE sample buffer (for immunoblotting) or a lysis buffer for subcellular fractionation.
Immunoprecipitation of PP2A Ca/b from adult rat ventricular myocytes ARVM were exposed to vehicle control (DMSO, 0.1%) or CPA (10 mM) for 10 min and then lysed with ice-cold immunoprecipitation buffer (in mM); imidazole-HCl 20, EDTA 2, EGTA 2 and 0.1% Triton X-100. The culture plates were frozen on a volume of liquid N 2 and subsequently thawed at room temperature at which point cellular lysates were harvested. The insoluble cellular components were collected as a pellet following centrifugation at 14000 g for 10 min at 4uC. A 100 ml aliquot of the supernatant was then removed and to this 50 ml of 36 Laemmli sample buffer was added, this is referred to as the ''input''. To immunopreciptate PP2A Ca/b , 100 mg of protein from the remaining supernatant was incubated with 5 mg of immunoprecipitating antibody (mouse monoclonal anti-PP2A Ca/b ) or non-immune mouse IgG (Millipore, UK), overnight at 4uC. Then 50 ml of protein A sepharose beads (GE Healthcare) was added for a further 1.5 hours at 4uC. The beads were collected by centrifugation at 370 g for 1 min and at 4uC. The immunocomplex was then washed twice with ice-cold PBS and 50 ml of 26 Laemmli sample buffer was added to the immunocomplex. All samples were then heated to 95uC for 5 min prior to SDS-PAGE and Western immunoblotting. Equal volumes (20 ml) of each sample were loaded per well.

Subcellular fractionation of adult rat ventricular myocytes
ARVM were fractionated using a previously described protocol [33], with minor modifications. In brief, ARVM were lysed in icecold lysis buffer at pH 7.5 containing (in mM) Tris-HCl 50, EGTA 5, EDTA 2, DTT 5, as well as 0.05% digitonin and Mini-Complete protease inhibitor cocktail (Roche, Germany). The samples were then frozen by floating the culture plate on a volume of liquid N 2 and subsequently thawed at room temperature, at which point cellular lysates were harvested. Cell lysates were then centrifuged at 14000 g for 30 min at 4uC and the supernatant, which contained the cytoplasmic fraction, was removed. The pellet, which contained the membrane-rich particulate fraction, was then solubilised in an equal volume of the digitonin-based lysis buffer containing 1% Triton X-100. The determination of total PP2A c content in the cytosolic and particulate subcellular fractions was achieved by NaOH-mediated saponification of the methylated C-terminal leu309 residue of PP2A C . In brief, an equal volume of 200 mM NaOH (100 mM final [NaOH]) was added to each cytosolic and particulate fraction, followed by incubation for 30 min at 30uC and subsequent pH neutralisation by the addition of 1 M HCl, as previously described [23]. This ensured that all PP2A C protein in the sample was in a demethylated state, thereby allowing detection of total PP2A C content by the anti-PP2A C monoclonal antibody (4B7). Equal volumes of Laemmli SDS-PAGE sample buffer were added to both fractions and the fractionated proteins resolved by 12% SDS-PAGE followed by western immunoblotting.

Western immunoblotting
Western immunoblotting was carried out as previously described [34]. In brief, protein samples were separated by 12% SDS-PAGE, transferred to PVDF or nitrocellulose membranes where appropriate and probed with the primary antibodies as described earlier. Primary antibodies were detected by appropriate donkey anti-rabbit or sheep anti-mouse secondary antibodies linked to horseradish peroxidase (GE Healthcare, UK). Specific protein bands were detected by enhanced chemiluminescence (GE Healthcare, UK) and band intensity was quantified using a calibrated densitometer (GS-800, Bio-Rad) and Quantity OneH 1-D analysis software (v 4.6.2).

Statistical analysis
Data are presented as mean 6 SEM. Data were subjected to ANOVA (GraphPad Prism v6.0.1) and if a significant difference (p,0.05) was detected, further analysis by a Dunnett's modified ttest (for comparison of each group with a single control) was performed.

Results
Compartmentalisation of the PP2A C , LCMT-1 and PME-1 signalling axis in adult rat ventricular myocytes We initially aimed to determine the subcellular localisation of components comprising the LCMT-1/PME-1/PP2A signalling axis in ARVM. Previous studies have demonstrated that LCMT-1 protein expression in non-myocytes is restricted to the cytoplasm, whereas PME-1 has been shown to be predominantly expressed in the nucleus [35]. The subcellular fractionation of ARVM by digitonin/Triton X-100 based lysis is shown in Fig. 1 and confirms our own previous observation [23] that the PP2A C subunit is predominantly localised to the cytoplasmic fraction of ARVM (Fig. 1A). Furthermore, Fig. 1A also shows that B55a (PPP2R2A) regulatory subunit isoform that associates with the PP2A A/C heterodimer in a methylation dependent manner [8] is similarly localised to the cytoplasmic fraction of ARVM. Importantly, both PME-1 (Fig. 1B) and LCMT-1 (Fig. 1C) protein expression appears to be almost exclusively localised in the cytoplasm of ARVM, with negligible localisation in the soluble membrane and insoluble fractions. This suggests that all components thought to regulate the carboxylmethylation and methylation-dependent targeting of PP2A C to cellular substrates are predominantly localised in the cytoplasm of ARVM.

G i PCR-induced PP2A C carboxylmethylation in adult rat ventricular myocytes
Using an anti-methyl PP2A C antibody (2A10) to detect the methylation status of the C-terminal leu309 of PP2A C , data shown in Fig. 2A extends previously reported observations [22][23][24] that the stimulation of G i -protein coupled A1.Rs by N6-cylcopentyladenosine (CPA) induces carboxylmethylation of PP2A C , by establishing the concentration-dependence of the response. In addition, we provide evidence shown that the Angiotensin II type 2 receptor (AT 2 ) agonist CGP42112 (Fig. 2B) and the muscarinic receptor agonist carbachol (Fig. 2C) also induce PP2A C carboxylmethylation in a concentration-dependent manner. Equal protein loading was determined by using a polyclonal anti-PP2A C (FL-309) to detect both non-methylated and methylated forms of PP2A C . These data suggest that PP2A C carboxylmethylation can be induced not only by the stimulation of A1.Rs but also by the stimulation of other G i PCRs.
Using NaOH to remove the methyl moiety from the C-terminal leu309 of PP2A C by saponification, we were able to demonstrate that the anti-methyl PP2A C antibody (2A10) was no longer able to detect carboxylmethylated PP2A C . Methylation of leu309 within the TPDYFL C-terminal tail of PP2A C masks the 4B7 antibody epitope and sterically interferes with antibody recognition of PP2A C protein. We show in Fig. 3A that A1.R stimulation by CPA induces an increase in PP2A C carboxylmethylation, which is confirmed by two different antibodies raised against carboxylmethylated PP2A C (2A10) or demethylated PP2A C (4B7). Having used NaOH to remove the C-terminal (leu309) methyl moiety, the antibody raised against demethylated PP2A C (4B7) showed an equal amount of total PP2A C in all protein sample lanes. We believe that this constitutes an effective protocol for the determination of total PP2A C in any protein sample. Fig. 3B confirms that the increased level of PP2A C carboxylmethylation in response to either CGP42112 or carbachol is also sensitive to saponification by NaOH.

Effects of CPA on cellular PP2A C binding partners
Several studies have shown that PP2A C can associate with several regulatory proteins such as PME-1 [14], LCMT-1 [16] and PKB [36] in non-myocytes. In support of this data, figure 4 shows that PP2A C does indeed exist in association with both PME-1 and LCMT-1 in unstimulated adult rat ventricular myocytes. Interestingly the data also demonstrate that although the exposure of ARVM to CPA does not alter the association between PP2A C and PME-1, it increases the association between PP2A C and LCMT-1. The data also suggest that, unlike in other cell types [36], there is no apparent association between cellular PP2A C and PKB in adult rat ventricular myocytes.

Role of Gbc subunits in CPA-induced PP2A C translocation in adult rat ventricular myocytes
The classical effect of G i PCR stimulation is to induce Ga imediated inhibition of membrane bound adenylate cyclase [37], however, Gbc subunits are known to mediate a range of intracellular signalling events of their own [38]. In this study we have exposed ARVM to increasing concentrations of CPA (0-100 mM) for 10 minutes and determined the translocation of PP2A C to the membrane-rich particulate fraction. Figure 5A shows that CPA significantly (p,0.05) increased the total PP2A C content within the particulate fraction of ARVM in a concentration dependent manner. To test the hypothesis that Gbc subunits were mediators of the intracellular signalling cascade that led to G i PCR-induced PP2A C translocation to the particulate fraction of ARVM in response to G i PCR stimulation, we adenovirally expressed the Ga subunit of the ocular transducin GPCR, Ga t1 in these cells. Several studies have demonstrated that heterologous expression of the Ga t1 subunit in cells can sequester Gbc subunits and switch off Gbc subunit-dependent intracellular signalling [27,39,40]. Figure 5B illustrates the MOI-dependent levels of EGFP and Ga t1 protein co-expression and indicates that considerable heterologous expression of Ga t1 protein was achieved at an MOI of 50 pfu/ARVM. Hence, ARVM were infected with either a control EGFP-expressing or Ga t1 -expressing adenovirus (each at 50 MOI) and then exposed to 10 mM CPA for 10 minutes. Figure 5C shows that CPA-induced significant (p,0.05) PP2A C translocation to the particulate fraction in the presence of EGFP expression. However, this translocation was abrogated in ARVM infected to express Ga t1 in order to sequester Gbc subunits. These data suggest that CPA-induced PP2A C translocation to the particulate fraction of ARVM is mediated by Gbc subunits.

CPA-induced activation of phosphoinositide 3-kinase (PI3K) and redistribution of PP2A C in adult rat ventricular myocytes
Several studies have shown that G i PCR stimulation can activate the PI3K family of lipid/protein kinases [41][42][43]. To determine that this occurs under our experimental conditions, we used PKB (ser473) phosphorylation status as a surrogate index of PI3K activity. CPA was found to induce significant (p,0.05) phosphorylation of PKB at ser473 (Fig. 6A) and pretreatment of ARVM with the PI3K inhibitor LY294002 (20 mM) abrogated this response (Fig. 6B). These observations suggest that not only did the activation of G i -coupled A1.Rs induce the phosphorylation of PKB (ser473), but that this occurred in a PI3K-dependent manner in ARVM. Importantly, Fig. 6C shows that pretreatment of ARVM with LY294002 (20 mM) abrogated CPA-induced PP2A C translocation to the particulate fraction of ARVM, thereby suggesting that G i PCR-induced PP2A C translocation is dependent on PI3K activation in ARVM.

Discussion
Several studies [22][23][24] have previously shown that the subcellular distribution of PP2A C is predominantly restricted to the cytoplasm of ARVM, which we confirmed in the current study. Furthermore, the PP2A C regulatory B subunit (PPP2R2A/ B55a) which is known to associate exclusively with carboxylmethylated PP2A C [8] and form targetable PP2A heterotrimers, was also restricted to the cytoplasm of ARVM (Fig. 1A). This confirms that both PP2A C and B55a are localised to the same cellular compartment of ARVM. The overall cellular carboxylmethylation status of PP2A C is thought to reflect the balance in activity of LCMT-1 and PME-1 towards their common substrate PP2A C . X-ray crystallographic data suggest that the active sites of PP2A C , LCMT-1 [16] and PME-1 [14] can interact with each other to form tight PP2A C /LCMT-1 and PP2A C /PME-1 complexes, which regulate the methylation status of the Cterminal leu309 of PP2A C . These observations are supported by our own novel finding in ARVM, which suggest that cellular PP2A C does indeed basally form an association with both PME-1 and LCMT-1. Assuming that the activities of PME-1 and LCMT-1 remain constant, the increase in C-terminal leu309 PP2A C methylation as a result of G i PCR stimulation, may be explained either by a reduced PME-1 or an increased LCMT-1 association with PP2A C . Our data suggests that the latter may be true as G i PCR stimulation does not affect the association of PP2A C with PME-1 protein. This novel observation may explain the basis by which G i PCR stimulation induces an increase in C-terminal leu309 carboxylmethylation of PP2A C . Interestingly, immunoprecipitation of PP2A C removed virtually all PP2A C protein from the lysate (to undetectable levels) and a significant proportion of cellular LCMT-1 protein was immunoprecipitated along with it. Hence, the pool of cytoplasmic PP2A C in this complex may represent a reservoir of predominantly carboxylmethylated PP2A C subunits. Although PME-1 protein was shown to associate with PP2A C , a significant proportion of PME-1 protein was still present in the immunocomplex supernatant (post-immunoprecipitation) and it is likely that the PP2A C in this complex is predominantly in the demethylated state [9,10].
As PP2A C subunits are the only known LCMT-1/PME-1 substrates, it is possible that cytoplasmic PP2A C restricts the localisation of LCMT-1, and to a lesser extent PME-1 protein, in the cytoplasm. PME-1 protein in human HeLa cells contains a functional 270KRKK273 nuclear localisation sequence (NLS), which explains why in this cell type PME-1 protein is thought to be predominantly localised in the nucleus together with demethylated PP2A C [35]. The human and rat PME-1 proteins are highly (98%) homologous, however, much of the variance that exists between these two PME-1 proteins, does so in the residues adjacent to the NLS. Rat PME-1 protein contains the sequence 268VNKRKK273 which differs to the human PME-1 protein sequence 268ISKRKK273. It is possible that the NLS contained within rat PME-1 protein is in some way dysfunctional due to the preceding valine/asparagine residues, which may partly explain why in ARVM, PME-1 protein is predominantly found in the cytoplasm and not the nucleus.
The carboxylmethylation and subsequent translocation of PP2A C to the particulate fraction, in response to stimulated G i protein coupled A1.Rs, within ARVMs was reported several years ago by the Hofmann laboratory [22,24,44]. In these studies, the phosphorylation of proteins in response to isoprenaline, present in the particulate fraction (troponin I and phospholamban), was found to be decreased by the stimulation of A1.Rs in a phosphatase-dependent manner. Our previous studies not only Figure 3. Saponification of carboxylmethylated PP2A C by alkalinisation. ARVM were exposed to the A1.R agonist CPA (0-100 mM) for 10 minutes and PP2A C carboxylmethylation was detected by Western immunoblotting (IB) with either an anti-methyl PP2A C (2A10) or demethylated PP2A C (4B7) antibody before (2NaOH) or after (+NaOH) saponification of the C-terminal leu309 methylation by 100 mM NaOH. (B) Carboxylmethylation of PP2A C in response to CGP (10 nM), CPA (10 mM) or CCH (10 mM) as detected by Western immunoblotting with either an anti-methyl PP2A C (2A10) or demethylated PP2A C (4B7) antibody with or without treatment with 100 mM NaOH. Immunoblots are representative of 3 individual experiments. doi:10.1371/journal.pone.0086234.g003 confirmed these observations, but identified an additional membrane bound protein (Na + /H + exchanger isoform-1) as a novel PP2A C substrate and demonstrated that the translocation of PP2A C to the particulate fraction was sensitive to pertussis toxin [23]. Despite these studies suggesting that PP2A C can be regulated by G i PCR stimulation, very little has been reported since regarding the cellular mechanisms involved in the regulation of PP2A C by inhibitory class GPCRs. A possibility that the PP2A C translocation observed in ARVM may be A1.R-specific did exist. Hence, in this study we also chose to use other G i PCR agonists such as carbachol and CGP42112 to target M 2 .Rs and AT 2 .Rs, respectively. Our data show that the stimulation of M 2 .Rs and AT 2 .Rs also induced PP2A C carboxylmethylation in a concentration-dependent manner. The data suggests that PP2A C carboxylmethylation is not a unique consequence of A1.R stimulation but can occur downstream of other G i PCRs. To confirm that PP2A C carboxylmethylation in ARVM following exposure to CPA, carbachol or CGP42112 as detected by the anti-methyl PP2A C antibody was genuine, we used NaOH to remove the methyl moiety by saponification from the C-terminal leu309 of PP2A C [23,45,46]. Saponification abolished the signal detected by the anti-methyl PP2A C antibody and equalised the signal in all lanes, as detected by the anti-demethylated (4B7) PP2A C antibody. This confirmed that the methyl moiety on the C-terminal leu309 of PP2A C conferred antibody epitope masking as previously reported [23,45,46].
These novel observations led us to next investigate mechanisms downstream of G i PCR stimulation and we focused our attention on the role of Gbc subunits in G i PCR-induced cellular PP2A C redistribution. We chose to sequester Gbc subunits by adenovirally expressing the Ga t1 subunit of the ocular transducin GPCR, a protein not natively expressed in ARVM. Several studies [27,39,47] have successfully expressed Ga t1 subunits to implicate a role for Gbc subunits in intracellular signalling events. Our data demonstrate that heterologous expression of Ga t1 subunits abrogated G i PCR-induced PP2A C translocation to the particulate fraction of ARVM. This is supported by studies suggesting the existence of a G i PCR (A1.R)-Gbc signalling hub in cardiac tissue. The stimulation of A1.Rs by chlorocyclopentyl adenosine in murine hearts was shown to activate phospholipase C in a Gbcdependent manner [48]. Furthermore, the stimulation of the dual (G s /G i ) coupled b 2 -adrenoceptor (b 2 .AR) was reported to promote the survival of adult mouse cardiomyocytes following exposure to isoprenaline. In this study, the carboxyl terminus of b-adrenoreceptor kinase (bARK-ct) a commonly used Gbc inhibitor [49] abrogated isoprenaline-induced PKB activation and cellular survival, thereby implying that b 2 .AR/G i -induced cardiomyocyte survival involves Gbc subunit activation [41]. Our novel data suggests that a proximal signalling step linking G i PCRs to the regulation of PP2A C cellular localization is mediated by Gbc subunits.
Evidence [41,50] suggests that Gbc subunits initiate intracellular signalling cascades via the activation of PI3K [38]. It has been Figure 5. Role of Gbc subunits in CPA-induced PP2A C translocation. ARVM were lysed with a digitonin-based buffer to separate the cytoplasm and particulate fractions by centrifugation. Samples were then saponified with NaOH to abrogate any PP2A C carboxylmethylation. (A) Total PP2A C content in the particulate fraction of ARVM in response to increasing concentrations of CPA (0-100 mM) was indexed by Western immunoblotting (IB) with an anti-demethylated PP2A C antibody (4B7) following treatment with 100 mM NaOH.
Total PP2A C content in the particulate fraction was quantified by densitometry. (B) Multiplicity of infection (MOI)-dependent co-expression of EGFP and Ga t1 protein in ARVM infected with the AdV:Ga t1 . (C) ARVM were infected with either the control AdV:EGFP or AdV:Ga t1 for 18 hours and then exposed to10 mM CPA for 10 minutes. ARVM were then lysed and total PP2A C content in the particulate fraction was indexed by Western immunoblotting (IB) with an anti-demethylated PP2A C (4B7) antibody following treatment of samples with 100 mM NaOH. PP2A C content in the particulate fraction was quantified by densitometry. All columns represent mean values 6 SEM, n = 4 individual experiments, *p,0.05 vs 0 (control group). doi:10.1371/journal.pone.0086234.g005 considered for some time that class IB PI3Ks can be activated by Gbc dimers [51,52]. Class IB PI3Ks consist of a p110c catalytic subunit associated with a regulatory subunit referred to as p101 [51]. The regulatory subunit p101 is central to the Gbc-mediated activation of class IB PI3Ks [52]. Our observations with the PI3K inhibitor LY294002 suggest that A1.R stimulation induces PP2A C translocation to the particulate fraction via a Gbc/PI3K signalling axis. In support of our observations, Zhu et al [41] reported that the G i arm of the dual G s /G i -coupled b 2 -adrenoceptor could also activate PI3K via Gbc subunits in murine cardiomyocytes. Hence our data suggests that A1.Rs are linked to PP2A C by a Gbc/PI3K signalling axis. Figure 6. Role of PI3K in CPA-induced PP2A C translocation. (A) ARVM were exposed to the A1.R agonist CPA (0-100 mM) for 10 minutes and then lysed with SDS-PAGE Laemmli sample buffer for the determination of PKB phosphorylation (ser473) and total PKB by Western immunoblotting (IB). PKB (ser473) phosphorylation was quantified by densitometry. (B) ARVM were pretreated with either 0.1% DMSO or 20 mM LY294002 for 30 minutes and then exposed to 0.1% DMSO (CTR) or 10 mM CPA for 10 minutes. PKB phosphorylation (ser473) and total PKB were then determined by Western immunoblotting (IB). PKB (ser473) phosphorylation was quantified by densitometry. (C) ARVM were pretreated with either 0.1% DMSO or 20 mM LY294002 for 30 minutes and then exposed to 0.1% DMSO (CTR) or 10 mM CPA for 10 minutes. ARVM were then lysed with 0.05% digitonin and fractionated by centrifugation. Samples were then saponified with NaOH to remove any PP2A C carboxylmethylation. Total PP2A C content in the particulate fraction of ARVM was indexed by Western immunoblotting (IB) with anti-demethylated PP2A C (4B7) antibody following treatment with 100 mM NaOH. Total PP2A C content in the particulate fraction was quantified by densitometry. All columns represent mean values 6 SEM, n = 4 individual experiments, *p,0.05 vs 0 (control group). doi:10.1371/journal.pone.0086234.g006 Figure 7. Suggested intracellular signalling mechanism (s) through which A1.Rs induce PP2A C translocation. Our data suggests that the stimulation of G i protein-coupled adenosine A1 receptors by the agonist CPA increases the association of PP2A C with LCMT-1, thereby augmenting the leucine carboxylmethylation status of PP2A C . The stimulation of G i protein-coupled adenosine A1 receptors by the agonist CPA also elicits a cascade involving the release of Gbc subunits which activate PI3K. Both of these intracellular signalling events coordinate and facilitate the association of PP2A C with the membrane-rich particulate compartment of ARVM. doi:10.1371/journal.pone.0086234.g007 In conclusion, our data indicate that G i PCR agonists may induce the carboxylmethylation of PP2A C by increasing its association with the methyltransferase LCMT-1. G i PCR agonists also mediate the disassociation of Ga i bc heterotrimeric proteins and the released Gbc subunits in turn activate PI3K. The simultaneous activation of PI3K and induction of PP2A C carboxylmethylation appear to coordinate the translocation of PP2A C to the particulate fraction of ARVM in response to G i PCR stimulation (see figure 7).