The Cell Cycle Checkpoint System MAST(L)-ENSA/ARPP19-PP2A is Targeted by cAMP/PKA and cGMP/PKG in Anucleate Human Platelets

The cell cycle is controlled by microtubule-associated serine/threonine kinase-like (MASTL), which phosphorylates the cAMP-regulated phosphoproteins 19 (ARPP19) at S62 and 19e/α-endosulfine (ENSA) at S67and converts them into protein phosphatase 2A (PP2A) inhibitors. Based on initial proteomic data, we hypothesized that the MASTL-ENSA/ARPP19-PP2A pathway, unknown until now in platelets, is regulated and functional in these anucleate cells. We detected ENSA, ARPP19 and various PP2A subunits (including seven different PP2A B-subunits) in proteomic studies of human platelets. ENSA-S109/ARPP19–S104 were efficiently phosphorylated in platelets treated with cAMP- (iloprost) and cGMP-elevating (NO donors/riociguat) agents. ENSA-S67/ARPP19-S62 phosphorylations increased following PP2A inhibition by okadaic acid (OA) in intact and lysed platelets indicating the presence of MASTL or a related protein kinase in human platelets. These data were validated with recombinant ENSA/ARPP19 and phospho-mutants using recombinant MASTL, protein kinase A and G. Both ARPP19 phosphorylation sites S62/S104 were dephosphorylated by platelet PP2A, but only S62-phosphorylated ARPP19 acted as PP2A inhibitor. Low-dose OA treatment of platelets caused PP2A inhibition, diminished thrombin-stimulated platelet aggregation and increased phosphorylation of distinct sites of VASP, Akt, p38 and ERK1/2 MAP kinases. In summary, our data establish the entire MASTL(like)–ENSA/ARPP19–PP2A pathway in human platelets and important interactions with the PKA, MAPK and PI3K/Akt systems.

. Amino Acid sequence alignment of ENSA isoform 1 and ARPP19. Clustal Omega (Version 1.2.1) was used for sequence alignment. S62/67 is marked in red, S104/109 in yellow. Stars demonstrate identical amino acids. Empty space shows that there is no amino acid sequence similarity between ARPP19 and ENSA.

Recombinant Wildtype and Mutant ENSA Protein Expression and Purification
E. coli BL21 were transfected with pET28 vectors including the DNA for wildtype or mutant (S67A/S109A/S109D) HisENSA. Protein expression was induced with isopropyl β-d-1thiogalactopyranoside (IPTG) (0.1 mM) and proteins were isolated 20 h after induction. Therefore, E. coli were pelleted at 4225× g for 10 min at 4 °C. The pellets were resuspended in ice-cold lysis buffer (50 mM NaH2PO4, 300 mM NaCl, 1 mM MgCl2, 15 mM imidazole, pH 8.0, cOmplete TM protease inhibitor cocktail) and four times sonicated with 30 s of resting on ice in between. The lysate was centrifuged for 45 min at 16,900× g and 4 °C. The recombinant wildtype and mutant HisENSA proteins were purified using metal ion affinity chromatography (1 mL HisTrap HP (Ni 2+ -ions) columns and an ÄKTA prime, GE Healthcare). The proteins were eluted with elution buffer (20 mM Na3PO4, 500 mM NaCl, 500 mM imidazole, pH 7.4), afterwards, the buffer was exchanged to 1× TBS buffer (1.37 M NaCl, 0.2 M tris, pH 7.4) with PD-10 desalting columns (GE Healthcare).

Culture, Treatment, and Sample Generation of HEK293 Cells
HEK293 cells were grown in DMEM at 37 °C and 5% (v/v) CO2 on 6-well plates until they reached 80% confluence. They were transfected with pCMV-3Tag-1A vector including FLAG-ENSA sequence using PolyJet TM transfection reagent (1 µg of vector/well). One day after transfection, cells were incubated in the presence or absence of 10 µM forskolin or vehicle control (0.1% (v/v) DMSO) or iloprost (1 µM) for 15 min or in the presence or absence of thrombin (0.1 U/mL) for 10 min at 37 °C and 5% (v/v) CO2. As control, non-transfected cells were used. Cells were lysed with 2× lysis buffer (0.1 M tris, 0.3 M NaCl, 10 mM MgCl2, 2% (v/v) triton, pH 7.5) and western blot samples were generated in 3× Laemmli buffer, heated for 10 min at 95 °C. Amino Acid sequence alignment of ENSA isoform 1 and ARPP19. Clustal Omega (Version 1.2.1) was used for sequence alignment. S62/67 is marked in red, S104/109 in yellow. Stars demonstrate identical amino acids. Empty space shows that there is no amino acid sequence similarity between ARPP19 and ENSA.

Recombinant Wildtype and Mutant ENSA Protein Expression and Purification
E. coli BL21 were transfected with pET28 vectors including the DNA for wildtype or mutant (S67A/S109A/S109D) HisENSA. Protein expression was induced with isopropyl β-d-1-thiogalactopyranoside (IPTG) (0.1 mM) and proteins were isolated 20 h after induction. Therefore, E. coli were pelleted at 4225× g for 10 min at 4 • C. The pellets were resuspended in ice-cold lysis buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, 1 mM MgCl 2 , 15 mM imidazole, pH 8.0, cOmplete TM protease inhibitor cocktail) and four times sonicated with 30 s of resting on ice in between. The lysate was centrifuged for 45 min at 16,900× g and 4 • C. The recombinant wildtype and mutant HisENSA proteins were purified using metal ion affinity chromatography (1 mL HisTrap HP (Ni 2+ -ions) columns and an ÄKTA prime, GE Healthcare). The proteins were eluted with elution buffer (20 mM Na 3 PO 4 , 500 mM NaCl, 500 mM imidazole, pH 7.4), afterwards, the buffer was exchanged to 1× TBS buffer (1.37 M NaCl, 0.2 M tris, pH 7.4) with PD-10 desalting columns (GE Healthcare).

Culture, Treatment, and Sample Generation of HEK293 Cells
HEK293 cells were grown in DMEM at 37 • C and 5% (v/v) CO 2 on 6-well plates until they reached 80% confluence. They were transfected with pCMV-3Tag-1A vector including FLAG-ENSA sequence using PolyJet TM transfection reagent (1 µg of vector/well). One day after transfection, cells were incubated in the presence or absence of 10 µM forskolin or vehicle control (0.1% (v/v) DMSO) or iloprost (1 µM) for 15 min or in the presence or absence of thrombin (0.1 U/mL) for 10 min at 37 • C and 5% (v/v) CO 2 . As control, non-transfected cells were used. Cells were lysed with 2× lysis buffer (0.1 M tris, 0.3 M NaCl, 10 mM MgCl 2 , 2% (v/v) triton, pH 7.5) and western blot samples were generated in 3× Laemmli buffer, heated for 10 min at 95 • C.

Preparation of Washed Human Platelets
Washed human platelets were isolated from citrate-anticoagulated (3.2% (v/v) tri-sodium-citrate) whole blood, taken from healthy volunteers not on platelet-affecting drugs for at least 10 days before blood collection. All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the local Ethics Committee of the University Medical Center Mainz (Study No. 837.302.12; 25.07.12; FF109/2015). Platelet-rich plasma (PRP) was generated by centrifugation of citrated blood, supplemented with 2 mM EGTA, at 200× g for 10 min at room temperature (RT). PRP was diluted 1:1 with CGS buffer (120 mM NaCl, 12.9 mM trisodium citrate dihydrate, 30 mM d-glucose, pH 6.5) and centrifuged at 69× g for 10 min at RT, to pellet the leukocytes. The supernatant was centrifuged at 400× g for 10 min at RT. The platelet pellet was resuspended in 3 mL CGS buffer and centrifuged again at 400× g for 10 min at RT. Finally, platelets were resuspended in HEPES buffer (150 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 10 mM d-glucose, 10 mM HEPES, pH 7.4). The isolated platelets were adjusted to 5 × 10 8 or 1 × 10 9 or 2 × 10 9 platelets/mL and kept at 37 • C for 15 min.

Generation of Platelet Lysates from Washed Human Platelets
Washed human platelets were incubated with 1 mM Ca 2+ in the presence or absence of 2 µM okadaic acid (OA) for 15 min at 37 • C. Platelets were centrifuged at 16,900× g for 1 min at 4 • C. The pellet was resuspended in 50% (v/v) of the original volume of lysis buffer (50 mM tris, 150 mM NaCl, 1 mM EDTA, 0.1 mM EGTA, 0.25% (v/v) NP-40, pH 7.4; cOmplete TM protease inhibitor cocktail) and vortexed.

Phosphorylation of GST-ARPP19 or HisENSA with Recombinant Kinases for Dephosphorylation Experiments in Platelet Lysates and for PP2A Phosphatase Activity Assay
Recombinant GST-ARPP19 or HisENSA were phosphorylated as described in 2.9. For western blot analysis, samples were taken before (0 min) and after 20 min of incubation. The phosphorylation reaction was stopped with 6 mM EDTA. For the PP2A activity assay, samples were desalted and buffer was exchanged with Zeba TM spin desalting columns (7K MWCO, 0.5 mL) to 5× PP2A reaction buffer (250 mM imidazole, 1 mM EGTA, 0.1% (v/v) 2-mercaptoethanol, 0.5 mg/mL BSA, pH 7.2).

Phosphorylation of Recombinant HisENSA and GST-ARPP19 in Platelet Lysates
Kinase phosphorylation buffer (10×: 100 mM HEPES, 50 mM MgCl 2 , 10 mM DTT, 0.2 mM EDTA, pH 7.2) and 1 mM ATP was added 1:10 to platelet lysates. As controls, lysates were incubated in the absence or presence of OA (2 µM or 2 nM). The reaction was started with addition of recombinant HisENSA (250 nM) or GST-ARPP19 (87 nM). Western blot samples were taken directly after protein addition (0 min), after 3, 10 and 30 min or after 10, 20 and 40 min of protein incubation at 30 • C, in 3× Laemmli buffer and heated for 5 min at 95 • C.

Ser/Thr-Protein Phosphatase Inhibition by OA in Intact Human Platelets
Washed human platelets were incubated in the presence of 1 mM Ca 2+ with OA (50 nM, 200 nM, 2 µM or 10 µM) or vehicle control (0.1% (v/v) or 0.25% (v/v) EtOH) at 37 • C for up to 40 min. Western blot samples were taken directly after OA/vehicle addition (0 min) and after 10, 20, 30 and 40 min of OA/vehicle addition into 3× Laemmli buffer and heated for 5 min at 95 • C.

Colorimetric Ser/Thr/PP2A Phosphatase Activity Assay
To assess PP2A activity in platelet lysates, a colorimetric molybdate-based Ser/Thr phosphatase assay was used according to the manufacturer's protocol (Promega Serine/Threonine Phosphatase Assay System V2460). Platelet lysate was generated as described in 2.6. and free phosphate was removed using the spin columns provided by the assay system. Platelet count was adjusted to 1 × 10 8 platelets/mL with lysis buffer (50 mM tris, 150 mM NaCl, 1 mM EDTA, 0.1 mM EGTA, 0.25% (v/v) NP-40, pH 7.4; cOmplete TM protease inhibitor cocktail). 35 µL platelet lysate were added per well to start the reaction and incubated with phosphopeptide in PP2A reaction buffer for 60 min at 37 • C. 0 min control samples were stopped immediately after lysate addition with a mixture of molybdate dye/additive (50 µL/well). After 60 min samples were also stopped, the reaction was incubated additional 15 min at RT in the dark. For inhibitory experiments, OA, fostriecin, tautomycetin or recombinant proteins were added to the platelet lysate. As recombinant proteins were in 5× PP2A reaction buffer (250 mM imidazole, 1 mM EGTA, 0.1% (v/v) 2-mercaptoethanol, 0.5 mg/mL BSA, pH 7.2), the amount of additional 5× PP2A reaction buffer was adjusted, according to the amount of added recombinant protein. P2A activity was measured by absorbance of 630 nm using a 96-well plate reader (Dynex Opsys MR TM , Tarporley, CW6 9BL UK).

Statistical Analysis
Experiments were performed at least three times with at least three different healthy donors when platelet samples were involved. Data are presented as means ± standard deviation (SD). Statistical analysis was performed using GraphPad Prism 8 for Windows (GraphPad Software, San Diego, CA, USA). One-way ANOVA multiple comparisons test was used for comparison of more than two groups. p < 0.05 was considered as significant.

Results and Discussion
3.1. The PP2A Inhibitors ENSA and ARPP19 are Present in Human Platelets and Phosphorylated by Both PKA and PKG In our earlier proteomic studies, we detected significant expression of both ENSA and ARPP19 in the small anucleate human platelets [34]. Subsequently, in our analysis of the iloprost/cAMP-stimulated phosphoproteome of human platelets, more than 130 cAMP/PKA regulated phosphoproteins were detected including ENSA (phosphorylated at S109) and ARPP19 (at S104) [35]. Platelets are strongly inhibited by both cAMP-and cGMP-elevating agents mediated by the corresponding cAMP-and cGMP-dependent protein kinases (PKA/PKG), which have overlapping specificity [18,19]. Therefore, we investigated the effect of the cGMP/PKG-pathway on ENSA/ARPP19 phosphorylation using various NO-donors [S-nitrosocysteine (SNC), sodium-nitroprusside (SNP), diethylamine NONOate (DEA-NO)] and the soluble guanylyl cyclase (sGC)-stimulator riociguat in comparison to the iloprost (cAMP) pathway, using vasodilator-stimulated phosphoprotein (VASP) as established PKA/PKG substrate and marker. Both ENSA (at S109) and ARPP19 (at S104) were strongly phosphorylated (several-fold stimulation over basal level, similar to VASP S239) in response to different cGMP-elevating agents as well as in response to the cAMP-elevating iloprost (Table 1). We also noted that both cAMP-and cGMP-elevating conditions reduced the basal phosphorylation of a second ENSA phospho-site, S67, ( Table 1). Washed human platelets were incubated without (control) or with the following test substances : iloprost (5 nM, 2 min); diethylamine NONOate (DEA-NO 5 µM, 2 min.); riociguat (10 µM, 5 min); sodium S-nitrosocysteine (SNC, 5 µM, 2 min) or sodium nitroprusside (SNP, 5 µM, 2 min). The reaction was stopped with 4× lysis buffer, samples were shock frozen in liquid nitrogen and analysed by quantitative phosphoproteomics as described [40]. The change of various phosphosites in response to the test substance compared to control (increase or decrease) is shown as average ratio (av. ratio). The copy number indicates the number of molecules of a given protein per platelet obtained in earlier studies [34].
Previously, we demonstrated that under these conditions riociguat had no detectable effect on cAMP/PKA in human platelets [41]. Therefore, we performed a second, comprehensive phosphoproteomic analysis of iloprost-and riociguat-treated human platelets with three biological replicates. These phosphoproteomic data demonstrated that both ENSA (at S109) and ARPP19 (at S104) are phosphorylated in response to cAMP/PKA as well as to cGMP/PKG stimulation in intact human platelets, similar to VASP S239 (Table 2). Washed human platelets from three healthy donors (three biological replicates) were incubated with buffer (control), iloprost (5 nM, 2 min at 37 • C) or riociguat (10 µM, 5 min at 37 • C). After incubation, samples were stopped by the addition of 4× lysis buffer, shock frozen in liquid nitrogen and analysed by quantitative phosphoproteomics as described [40]. The fold increase of phosphorylation of various phosphosites compared to control is shown as average ratio of all quantified peptides bearing the corresponding site. As a measure of reliability, the 'p-value fraction' (for each site: peptides p < 0.05/ site-bearing peptides) and # of averaged peptides is presented. The change of various phosphosites in response to the test substance compared to control (increase or decrease) is shown as average ratio (av. ratio). The p-value fraction of all sites shown here is 100%, which represents an excellent reliability of these phosphosite measurements. * All trypsin digests except VASP S157 (subtilisin digest).
Phosphorylation of the PKA-preferred site of VASP, S157, was also detected but required a special protease treatment (subtilisin) for detection by phosphoproteomics. It is of special interest for this study that phosphorylation of the serine/threonine protein phosphatase 2A (PP2A) at its regulatory B56δ subunit (S573) was also detected for the iloprost/cAMP pathway and as well but less phosphorylated for the riociguat/cGMP pathway. Others previously showed that the PKA phosphorylation of PP2A B56δ at S573 activates this special PP2A heterotrimer, leading to a decrease in protein phosphorylation [42]. We also confirmed this important B56δ S573 phosphorylation in our experiments with iloprost-/riociguat-treated human platelets, using the specific phospho-antibody as described [42], see Figure S1.
Our phosphoproteomic data show that both ENSA and ARPP19 are phosphorylation targets in response to the platelet inhibitory pathways cAMP and cGMP. It is indeed surprising to find the important phosphoproteins and cell cycle regulators ENSA and ARPP19 in non-dividing human platelets. We therefore confirmed our proteomic data via western blot analysis of washed human platelets and HEK293 cells using a well-established anti-ENSA antibody. Both human platelets and human HEK293 cells contained one ENSA species, (apparent 17 kDa in SDS-PAGE), independent of the experimental conditions ( Figure 2a). We also cloned, expressed and purified human ENSA of the known canonical sequence to study the direct PKA-/PKG-specific phosphorylation of human recombinant HisENSA (including S109A and S109D phosphosite mutants) and, in comparison, recombinant GST-ARPP19 using pure PKA (C-subunit) and PKG (Figure 2b,c).
HisENSA and GST-ARPP19 were strongly phosphorylated by both PKA and PKG, which was abolished by ENSA S109A or S109D mutations. Other data showed that the PKA-and PKG-induced phosphorylation of ARPP19 S104 resulted in a complete shift of GST-ARPP19 in phostag-gels, indicating stoichiometric phosphorylation of ARPP19 at S104.
All data obtained with intact human platelets and with recombinant proteins proved that ENSA and ARPP19 are not only PKA but also excellent PKG targets. In fact, the extent of S109 ENSA phosphorylation in response to various cGMP-elevating platelet inhibitors (Tables 1 and 2) compared to VASP 239 places especially ENSA among the best PKG targets studied. In contrast, the iloprost/cAMP pathway was more effective than the riociguat/cGMP pathway with respect to the phosphorylation of PP2A-B56δ (PPP2R5D) at S573, an established PKA substrate [42]. It is of special interest that both cGMP-and cAMP-elevating platelet inhibitors increased the phosphorylation of S109 ENSA while they decreased S67 ENSA phosphorylation (Table 1). Therefore, the properties of S67 ENSA and S62 ARPP19 phosphorylation in platelets were addressed next.

Phosphorylation of ENSA S67/ARPP19 S62 by a MASTL-Related Protein Kinase in Human Platelets
As introduced earlier, ENSA and ARPP19, when phosphorylated at S67/S62, strongly inhibit certain holoenzymes of PP2A. Considering this important functional role of S67 ENSA/S62 ARPP19 phosphorylation in other cell systems such as Xenopus oocytes and our preliminary detection of these phosphosites in human platelets by phosphoproteomics (Table 1), we investigated the possible S67 ENSA/S62 ARPP19 phosphorylation in human platelets and their lysates. With the Cell Signaling ® anti-pS67 ENSA/anti-pS62 ARPP19 phospho-antibody, which recognizes the conserved pS67 ENSA site as well as pS62 ARPP19 (sequence 100% similar), positive signals were initially detectable only in HEK293 cell samples ( Figure S2, used as positive control). However, it has been reported for other systems (Xenopus oocytes, HeLa cells) that Gwl/MASTL-phosphorylated ENSA/ARPP19 are rapidly dephosphorylated by PP2A, which could be completely prevented by the PP2A inhibitor okadaic acid (OA) [43,44]. In our experiments pre-incubation of intact platelets with OA at low (200 nM) and high (10 µM) concentrations induced a similar time-dependent phosphorylation of endogenous ENSA S67 (Figure 3a). This ENSA pS67 band may include some ARPP19 pS62 since the phospho-antibody recognizes both ENSA pS67 and ARPP19 pS62, and platelets contain about 3-fold more ENSA than ARPP19 (Table 2). regulatory B56δ subunit (S573) was also detected for the iloprost/cAMP pathway and as well but less phosphorylated for the riociguat/cGMP pathway. Others previously showed that the PKA phosphorylation of PP2A B56δ at S573 activates this special PP2A heterotrimer, leading to a decrease in protein phosphorylation [42]. We also confirmed this important B56δ S573 phosphorylation in our experiments with iloprost-/riociguat-treated human platelets, using the specific phospho-antibody as described [42], see Figure S1.  The molecular weight of endogenous ENSA in human platelets and HEK293 cells is similar and is not affected by different treatments (thrombin (thr), iloprost (ilo) or sodium nitroprusside (SNP)). (b) Recombinant HisENSA wildtype and phosphosite mutants S109A and S109D were phosphorylated in the presence of the C-subunit of PKA or of PKGIβ (with 5 µM cGMP for PKG) and 1 mM ATP. Samples were taken in Laemmli buffer before (0 min) and after 0.5, 2 and 10 min of ATP addition and analysed by immunoblotting. HisENSA S109A and S109D mutants were not phosphorylated confirming S109 is the phosphorylated amino acid. (c) Recombinant GST-ARPP19 was phosphorylated in the presence of PKA C-subunit or PKGIβ (and 5 µM cGMP) and 1 mM ATP. Samples were taken in Laemmli buffer before (0 min) and after 5, 10 and 20 min of ATP addition. The western blots showing HisENSA and GST-ARPP19 phosphorylation are representative for n = 3 independent experiments. part). This experiment additionally validated the phosphorylated site and the phospho-antibody studied. A consistent phosphorylation of endogenous ENSA at S67 already in the beginning of the lysate incubation can also be noted, probably due to the intracellular OA pre-incubation effects before lysis. OA added to lysates of untreated platelets also induced the phosphorylation of S67 HisENSA, and as well of S62 GST-ARPP19 (Figure 3c). HisENSA phosphorylation was only detectable when OA was added to intact platelets or platelet lysates. Intact platelets were pre-incubated without or with 2 µM OA and lysed afterwards. HisENSA and 1 mM ATP were added to the platelet lysate. Samples were taken in Laemmli buffer before (0 (b) S67 HisENSA phosphorylation was only detectable when OA was added to intact platelets or platelet lysates. Intact platelets were pre-incubated without or with 2 µM OA and lysed afterwards. HisENSA and 1 mM ATP were added to the platelet lysate. Samples were taken in Laemmli buffer before (0 min) and after 3, 10 and 30 min of 1 mM ATP addition. In contrast to endogenous ENSA and added HisENSA(wt), the added S67A HisENSA mutant was not phosphorylated, serving as important negative control. (c) HisENSA or GST-ARPP19 as well as 1 mM ATP were added to platelet lysate without (only with HisENSA) or with 2 µM OA in the lysate. Samples were taken before (0 min) and after 10, 20, 40 and 60 min of 1mM ATP addition. The western blots are representative for n = 3 independent experiments.
We then performed phosphorylation experiments with lysates from human platelets, pre-incubated with 2 µM of OA or vehicle control. Only lysates from OA pre-treated platelets (but not controls) supported a time-dependent S67-phosphorylation of added recombinant HisENSA wildtype (Figure 3b, upper part) whereas a S67A HisENSA mutant was negative (Figure 3b, lower part). This experiment additionally validated the phosphorylated site and the phospho-antibody studied. A consistent phosphorylation of endogenous ENSA at S67 already in the beginning of the lysate incubation can also be noted, probably due to the intracellular OA pre-incubation effects before lysis. OA added to lysates of untreated platelets also induced the phosphorylation of S67 HisENSA, and as well of S62 GST-ARPP19 (Figure 3c).
A control without OA showed no S67 HisENSA signal after 40 min of incubation. These results clearly indicate the presence of a protein kinase in human platelets that is able to phosphorylate ENSA and ARPP19 at S67 or S62, respectively. The protein kinase seems to be active in intact platelets as well as in platelet lysates but its activity can only be detected, when an ENSA/ARPP19-specific serine/threonine protein phosphatase (most likely PP2A) is inhibited.
These experiments with intact human platelets and their lysates establish for the first time a significant activity of a protein kinase, which phosphorylates endogenous and/or added HisENSA/GST-ARPP19 at the sites S67/S62 known in other systems to be responsible for PP2A inhibition. The absent S67 ENSA phosphorylation under basal conditions but clear detection when PP2A is inhibited ( Figure 3) suggests that the protein kinase responsible for S67 ENSA phosphorylation in the platelets is inhibited by PP2A-mediated dephosphorylation. In addition/or alternatively a very fast dephosphorylation of pS67 ENSA under basal and other conditions in human platelets is mediated by PP2A and inhibited by OA.
In mammalian cells, the only two protein kinases with the capacity to phosphorylate S67 ENSA/S62 ARPP19 (at the identical phosphosite motif KGQKYFDSGDYNMAK) described so far are MASTL (microtubule-associated serine/threonine kinase-like, the human orthologue of the Greatwall kinase, Gwl) [45,46] and the related MAST3 kinase [47,48]. However, in our proteomic studies of human platelets [34] and related studies of murine platelets [49] neither MASTL nor MAST3 were detected. Also, a comprehensive analysis of the human and platelet transcriptome did not detect significant expression levels of MASTL or MAST3, in contrast to significant expression levels of ENSA and ARPP19 [50]. At present, there are two possible explanations for our data on MASTL kinase protein (negative) and MASTL activity (positive) in human platelets. Despite the negative protein expression data so far, there are perhaps low, not yet detectable MASTL/MAST3 levels in platelets, which mediate the observed S67 ENSA/S62 ARPP19 phosphorylation. Alternatively, platelets may contain other MASTL-related protein kinases, which catalyse ENSA/ARPP19 phosphorylation, since human platelets express, at the protein level, about 150 different protein kinases from all classes [34,51]. Clearly, it will be important to elucidate the identity of the S67 ENSA/S62 ARPP19 protein kinase and its regulation within platelets (see also our limitation paragraph, end of the discussion).
It is of considerable interest that the discovery of MASTL as human Gwl orthologue [45] also defined a MASTL E167D mutation in humans, which is associated with thrombocytopenia [45,46,52]. Very recently, a megakaryocyte specific MASTL mutation (E167D knock-in) and a complete MASTL KO in mice was reported with a reduced platelet count and decreased half-life of platelets in MASTL KO and mutant knock-in mice. In addition, increased annexinV-levels, probably associated with platelet apoptosis or platelet activation, increased bleeding times and defective clot retraction were observed [53]. Overall, this MASTL mutation (E167D) was considered a gain of MASTL kinase function, leading to a stronger inhibition of PP2A. However, the study did not address the presence of MASTL protein or activity in platelets.
3.3. Phosphorylation of HisENSA and GST-ARPP19 by Human MASTL (S67/S62) and by PKA C-Subunit/ PKGIβ (S109/S104) or Combinations An important goal of this project was to investigate the possible inhibitory effect of ENSA and ARPP19 on serine/threonine protein phosphatases in human platelets. As prerequisite, the phosphorylation of the recombinant HisENSA and GST-ARPP19 proteins was studied in more detail (Figure 4). As already shown, PKA and PKG strongly phosphorylated HisENSA S109 and GST-ARPP19 S104. The ENSA S109 phosphorylation was abolished with ENSA S109A/D mutants but not with the S67A ENSA mutant ( Figure S3). Importantly, PKA and PKG phosphorylation of GST-ARPP19 caused a complete time-dependent mobility shift of GST-ARPP19 in phostag-gels (shown for PKA C-subunit in Figure 4c; data for PKG similar shown in Figure S4b) demonstrating stoichiometric (complete) phosphorylation of the PKA/PKG phosphosite under these conditions, as phostag-gels are separating the proteins by their number of phosphorylated sites and not by their molecular weight.  As already shown, PKA and PKG strongly phosphorylated HisENSA S109 and GST-ARPP19 S104. The ENSA S109 phosphorylation was abolished with ENSA S109A/D mutants but not with the S67A ENSA mutant ( Figure S3). Importantly, PKA and PKG phosphorylation of GST-ARPP19 caused a complete time-dependent mobility shift of GST-ARPP19 in phostag-gels (shown for PKA C-subunit in Figure 4c; data for PKG similar shown in Figure S4b) demonstrating stoichiometric (complete) phosphorylation of the PKA/PKG phosphosite under these conditions, as phostag-gels are separating the proteins by their number of phosphorylated sites and not by their molecular weight.
This phostag-gel analysis worked very well with GST ARPP19 but did not work with HisENSA, maybe due to interference of the His-Tag with Zn 2+ -ions of the phostag-gel. Recombinant human MASTL kinase phosphorylated HisENSA and GST-ARPP19 strongly and time-dependently at a site recognized by the pS67 ENSA/pS62 ARPP19 antibody. With GST-ARPP19, MASTL phosphorylation caused a time-dependent complete mobility shift of GST-ARPP19 in phostag-gels, demonstrating stoichiometric phosphorylation of the MASTL-site S62. Next, consecutive phosphorylation of HisENSA or GST-ARPP19 by kinase combinations (MASTL+PKA, Figure 4b,c; MASTL+PKG, Figure S4a) were investigated. The two important phosphorylation sites (MASTL: S67/S62; PKA/PKG: S109/S104) were effectively phosphorylated by the responsible kinases without interference by the other kinase. Specifically, MASTL had no major effect on PKA/PKG-mediated ENSA/ARPP19 phosphorylation, and PKA/PKG had no major effect on the MASTL-mediated ENSA/ARPP19 phosphorylation. This was clearly demonstrated by phostag-analysis of ARPP19 phosphorylation and its two independent shifts by S62 and S104 phosphorylation (Figure 4b,c, Figure S4b). At present, minor interactions between both kinases cannot be excluded since detailed kinetics analyses were not performed. Also, we cannot rule out interferences of the protein tags used in our experiments with ENSA/ARPP19 proteins. However, we have no evidence for this, and such tagged ENSA/ARPP19 proteins have been successfully used by many investigations in this field. Finally, the effects may be different in intact cells and may be cell type-/system-specific, as discussed below.
Mochida reported in 2014 the different phosphorylation sites of ENSA in Xenopus oocytes and their influence on PP2A inhibition [31]. Whereas human ENSA (and ARPP19) apparently only have two major phosphorylation sites (S67/S62 for MASTL, S109/S104 for PKA), Xenopus ENSA and ARPP19 have a third site, T28 for ENSA and S28 for ARPP19. It was concluded that the phosphorylation of all three sites influences PP2A inhibition. However, pS67 ENSA was shown to be the most potent PP2A inhibitor interacting with both B55-and C-subunits and blocking the catalytic centre of PP2A. Mochida calculated the IC 50 values for pENSA and PP2A B55 inhibition for his experiments using Xenopus oocytes and obtained an IC 50 of 0.47 nM for pS67 ENSA, an IC 50 of 0.52 nM for pS67+pS109 ENSA and much higher values for pS109 ENSA alone [31]. PKA phosphorylation of Xenopus ENSA S109 had little effect in these systems.
Also in Xenopus oocytes, PKA did not affect Gwl-induced ARPP19 phosphorylation nor the ability of Gwl-phosphorylated ARPP19 to inhibit PP2A B55δ. The authors concluded that the effect of S67 phosphorylation was dominant over the negative-function of S109-phosphorylation [54], in agreement with earlier studies by others [31].
Andrade et al. showed in 2017 that brain ARPP16 phosphorylation sites (the shorter isoform of ARPP19) were reciprocally phosphorylated when forskolin (activator of adenylate cyclase and thus of PKA) was added to striatal slices [47,48]. Forskolin induced a reduced phosphorylation of the MASTL/MAST site S46 (similar to S62 ARPP19/S67 ENSA) and stimulated S88 phosphorylation via PKA (similar to PKA sites S104 ARPP19/S109 ENSA). In that paper, several mechanisms were discussed to explain how S46 ARPP16 phosphorylation could be blocked resulting in reduced PP2A inhibition: • PKA phosphorylation of ARPP16 could interfere with the extent of PP2A inhibition by pS46 ARPP16 • PKA could phosphorylate and inhibit the ARPP16 S46 kinase MAST3 • Forskolin-stimulated PKA could activate PP2A by phosphorylation of PP2A B56δ [42] resulting in reduced S46 ARPP16 phosphorylation Some of these effects may perhaps be specific for brain, as was recently reviewed [55]. However, in our phosphoproteomic analysis, we clearly detected PP2A B56δ S573 phosphorylation in human platelets (Table 2) in response to iloprost (cAMP system) and riociguat (cGMP system), which we were able to confirm using phospho-specific antibodies ( Figure S4).
In our studies, PKA-phosphorylation of ENSA/ARPP19 did not affect their properties as MASTL substrates or in other in-vitro functions tested so far, but further work is required here. In intact platelets (Table 1) and transfected HEK293 cells ( Figure S2) ENSA S109 was strongly phosphorylated (transfected ENSA protein stronger than endogenous) when PKA was activated and, at the same time, down-regulation of S67 ENSA phosphorylation was observed (stronger signal with transfected ENSA, weaker with endogenous ENSA) ( Figure S2). Although our data with human platelets and the published results with murine striatum [47,48] are very similar (PKA reduced ENSA S67 or ARPP16 S62 phosphorylation, respectively), it will be necessary to identify the ENSA S67/ARPP19 S62 protein kinase in platelets for further mechanistic studies (see also the paragraph limitation).

Serine/Threonine Protein Phosphatases in Human Platelets
The availability of recombinant ENSA and ARPP19 in various phospho-variants allowed the analysis of their effects on platelet serine/threonine protein phosphatases. Although the general importance of serine/threonine protein phosphatase including PP2A as signalling opponent of serine/ threonine protein kinases has been recognized [25,26,56], the diversity and distinct regulatory mechanisms of serine/threonine protein phosphatases were only recently discovered, also supported by the complete description of the human protein kinome [57] and protein phosphatome [58]. With this background and our quantitative human platelet proteome data [34] it was possible to compile an overview of most serine/threonine protein phosphatases present in human platelets which include three catalytic subunits of PP1 and two catalytic subunits (α,β) of PP2A (Table S1). Table S1 also shows the intracellular concentrations of the catalytic subunits of serine/threonine protein phosphatases present in human platelets. It is remarkable that the major serine/threonine protein phosphatases such as PPM1, PP1, PP2A are present in human platelets at µM concentrations. This certainly is important for their catalytic capacity since most potential substrates have a similar intracellular concentration.
Whereas serine/threonine phosphatases were considered for a long time as unregulated components which 'simply' remove phosphates from serine/threonine phosphoproteins, it is now clear that serine/threonine protein phosphatases, and in particular PP1 and PP2A, are multimeric enzymes which are tightly controlled by regulatory subunits and additional activators and inhibitors [26].
Considering the enormous heterogeneity of PP2A composition and function, it was important to establish quantitative information on the expression level of PP2A subunits and ENSA/ARPP19 in human platelets, which is based on our proteomic databases (Table S2). We also estimated the intraplatelet concentration of these proteins, which is important for ENSA/ARPP19-PP2A interactions (see below). It is of interest that the PP2A A subunit concentration (1.67 µM) is higher than that of the two C subunits (1.07 µM) and similar to the concentration of all B subunits (1.73 µM) (Table S2). This indicates that there is no free PP2A C subunit in platelets, which agrees with observations in other cells and systems [26].
Human platelets contain the two catalytic subunits (α,β) of PP2A, only one of the two scaffolding A (α), two B55 subunits (α,δ), five B56 subunits (α,β,γ,δ,ε) and the PP2A activator, which is known to activate PP2A [59]. Interestingly, the distribution of PP2A isoforms in human platelets is similar to murine platelets, although some quantitative differences exist. For the scaffolding subunit A, only the α-isoform was found in human platelets, similar to other human cells [60]. The proteomic data suggest that 14 different PP2A trimeric holoenzymes and two different dimeric forms may exist in human platelets, perhaps even more due to splice variants of the regulatory subunits. The content of serine/threonine protein phosphatases in platelets and their subunit composition (especially of PP2A) is important to evaluate the following phosphatase assays.

Effects of PP2A Inhibitors and MASTL-Phosphorylated HisENSA/GST-ARPP19 on Human Platelet PP2A Activity
To evaluate the effects of S67/S62-phosphorylated ENSA/ARPP19 on PP2A activity of human platelets, we established a PP2A assay, which measures the release of free phosphate by the absorbance of a molybdate:phosphate complex [61]. This Ser/Thr-phosphatase assay was optimized for PP2A by selective assay conditions including a specific buffer, selective peptide substrate and appropriately desalted recombinant proteins and cell lysates ( Figure 5).

Effects of PP2A Inhibitors and MASTL-Phosphorylated HisENSA/GST-ARPP19 on Human Platelet PP2A Activity
To evaluate the effects of S67/S62-phosphorylated ENSA/ARPP19 on PP2A activity of human platelets, we established a PP2A assay, which measures the release of free phosphate by the absorbance of a molybdate:phosphate complex [61]. This Ser/Thr-phosphatase assay was optimized for PP2A by selective assay conditions including a specific buffer, selective peptide substrate and appropriately desalted recombinant proteins and cell lysates ( Figure 5).  This measured phosphatase activity was potently inhibited by OA (concentrations of 1 nM and even below) and by fostriecin (0.3-3.0 nM) (Figure 5a), in agreement with the reported IC 50 of 0.1-0.3 nM OA or of 1-3 nM fostriecin for PP2A inhibition, respectively [56,62]. In contrast, the PP1 inhibitor tautomycetin (reported IC 50 for PP1~0.5 nM, IC50 50 for PP2A~62 nM) [56,63] had inhibitory effects on the measured phosphatase activity only at concentrations of 50 nM and higher. These data established the conditions to study and compare the effects of known (low dose OA) and putative inhibitors of PP2A such as S67/S62-phosphorylated ENSA/ARPP19. Addition of both, HisENSA or GST-ARPP19 phosphorylated by MASTL at S67/S62 in vitro, resulted in a significant reduction of platelet PP2A activity (Figure 5b). In contrast, non-phosphorylated ENSA/ARPP19 proteins had no effect. As control, 1 nM OA resulted in complete inhibition of phosphatase activity. Although these results established for the first time a role for phosphorylated ENSA and ARPP19 in the inhibition of PP2A in human platelets, the effects observed were moderate (~20% inhibition). A likely explanation is that ENSA and ARPP19 target only certain B subunit containing PP2A heterotrimers present in human platelet lysates (see Tables S1 and S2). In contrast, OA binds directly to and inhibits all PP2A catalytic subunits, independent of the regulatory B subunits, and also inhibits structurally closely related serine/threonine protein phosphatases such as PP4/PP6 [62]. S67-/S62-phosphorylated ENSA/ARPP19 are thought to achieve their inhibitory effects via direct binding to a limited number of PP2A regulatory B-subunits (B55/B56 subfamily) [31,33,48,55]. However, there are very few studies available on the regulation of mammalian/human PP2A by the ENSA/ARPP family, and none for platelets and other blood cells.

Ser-Phosphorylated ARPP19 is Both a Substrate for and Inhibitor of Platelet PP2A
Xenopus ENSA, when phosphorylated at S67 by Gwl kinase, is both a substrate (but poor one) and inhibitor of Xenopus PP2A [31,43]. This led to the concept that S67-phosphorylated ENSA is a PP2A 'inhibitor by unfair competition' and that PP2A could also inactivate its own inhibitor. Therefore, we tested whether S62-or S104-phosphorylated GST-ARPP19 are substrates and inhibitors of protein phosphatases in platelet lysates. We performed these experiments using GST-ARPP19 since we could prepare and monitor stoichiometric MASTL-or PKA/PKG-phosphorylated pS62 GST-ARPP19 or pS104 GST-ARPP19, respectively, (Figure 4, Figure S4), which could then be tested as PP2A substrates or inhibitors.
Western blot analysis and quantification showed a fast and complete dephosphorylation of PKA-phosphorylated pS104 GST-ARPP19 (50% or more dephosphorylated at~3 min, Figure 6a, upper panel, first four lanes and Figure 6b) and a slower, but also complete dephosphorylation (50% dephosphorylated at 10 min) of MASTL-phosphorylated pS62 GST-ARPP19 (Figure 6a, lower panels, first four lanes and Figure 6c) constitutively in lysates of washed human platelets. The addition of 2 nM OA to platelet lysates completely inhibited the dephosphorylation of both sites, indicating that this dephosphorylation is mediated by one of the PP2A protein phosphatases present in human platelets (Table S1).  Then we tested whether S62-phosphorylated GST-ARPP19 can serve as PP2A inhibitor. For this, we used thio-phosphorylated S62 GST-ARPP19, since this prevents dephosphorylation and inactivation of ENSA/ARPP proteins as reported for Xenopus oocytes, HeLa cells and brain cells [31,43,48]. In platelet lysates, thio-pS62 GST-ARPP19 potently diminished the dephosphorylation of pS104 GST-ARPP19, whereas unphosphorylated GST-ARPP19 had no effects (Figure 6a,b). On the other hand, thio-phosphorylated S104 GST-ARPP19 and unphosphorylated GST-ARPP19 did not influence the dephosphorylation of pS62 GST-ARPP19 (Figure 6a,c). These data indicate that MASTL-as well as PKA-phosphorylated pS62 GST-ARPP19/pS104 GST-ARPP19 are substrates of a platelet PP2A, sensitive to low doses of OA. Furthermore, pS62 GST-ARPP19 is a potent inhibitor as well as a substrate of these PP2A forms. The data fully agree with the concept of 'inhibition by unfair competition' [43] of platelet PP2A by S62/S67 phosphorylated ARPP19 and ENSA.
The original concept of 'inhibition by unfair competition' also established an important requirement for this mechanism to work in intact cells: S67-phosphorylated ENSA should be present in molar excess over PP2A-B55 in mitosis to account for the total absence of PP2A activity during M phase [43]. This actually was shown for Xenopus oocytes with an estimated intracellular concentration of 100-250 nM for PP2A-B55 and 500 nM-1 µM for ENSA, which could be phosphorylated at S67 up to 50%. It is of considerable interest that similar or even somewhat higher concentrations of these components are found in human platelets (1.73 µM all PP2A B subunits, 1.7 µM ENSA-ARPP19, Table S2). There is important evidence that the family of regulatory B subunits defines special properties of PP2A protein phosphatases, including ENSA/ARPP19 inhibitory effects, PP2A substrate specificity, interaction with additional proteins and subcellular localization [26,[64][65][66][67]. Interestingly, mammalian cells have a much higher number of B subunit genes than lower eukaryotes [68]. Furthermore, phylogenetic research suggests that the expansion of the B subunit gene family was driven by the functional diversification of PP2A [69]. It is important to notice that human platelets contain the B55α, B55δ, B56α, B56β, B56γ, B56δ and B56ε PP2A subunits (Table S2), but only some of them (underlined) were shown so far to interact with Xenopus ENSA [31] or murine ARPP16 [48]. For human platelets, it will be very important now to establish which of the multiple platelet PP2A B subunits directly interacts with phosphorylated ENSA, ARPP19 or both.
It is therefore of importance to define both, inhibitors and substrates of PP2A in platelets. In contrast to protein kinases and their substrates/inhibitors, no reliable consensus sequence for serine/threonine protein phosphatases including PP2A have been established to date [26] in spite of considerable research efforts [65,70]. Nevertheless, multiple PP2A substrates have been studied and established within various processes such as intracellular signalling, cell cycle regulation, cell morphology and development as well as in brain function [23,55,71,72]. Because of the possible contribution of PP2A substrates to cell growth, cell survival and tumour suppression, phosphoproteins of the MAP kinase pathway and PI3K/Akt signalling have been particularly well studied [72][73][74]. For human platelets, very few studies have addressed PP2A substrates, even though, the PKA/PKG substrate and cytoskeletal protein VASP [75,76] and the p38 MAPK [36] were described as PP2A targets using OA and fostriecin as inhibitors of platelet PP2A. We therefore next studied the effect of OA on platelet activation/aggregation and on the state of phosphorylation of several known or putative PP2A substrates. We used 200 nM OA because it increased the phosphorylation of S67 ENSA already after 10 min incubation time reaching near-maximal effects at 20-40 min (Figure 3a). A 10 min incubation with 200 nM OA inhibited thrombin-induced aggregation of washed human platelets (Figure 7a; Figure S5a), which was associated with stable S67 phosphorylation of endogenous ENSA (Figure 7b). A 30 min incubation of intact human platelets with 200 nM OA strongly increased the phosphorylation of VASP (at S157 and S239), Akt (at T308 and S473), p38 (at T180/Y182) and p44/42 MAPK (Erk 1/2) (at T202/Y204) with initial effects observed within 10 min (Figure 7; Figure S6) after OA addition. Here, about 40% of VASP shifted to the upper band at 30 min OA treatment, indicating that 40% of VASP is S157 phosphorylated. This phosphorylation of VASP is shown here as ratio of the upper band compared to the total VASP signal in relation to each loading control. Data are presented as means +/-SD of three independent experiments. It needs to be considered that 200 nM OA used here with intact human platelets is relatively high compared to the PP2A inhibition by OA in cell free lysates (Ki 0.032 nM) and may affect PP1 (inhibition by OA in cell free lysates, Ki 150 nM) [56]. However, the conditions with intact cells (here intact platelets) are determined by additional factors. Although OA is membrane-permeant, it often requires concentrations of more than 50 nM to achieve significant effects within 30 min of incubation time [77]. A very important factor for OA effects in intact cells is also the concentration of the primary OA targets. In human platelets, we estimated here for the first time (see Tables S1 and S2) the concentration of PP2A catalytic subunits (α + β) as~1 µM, that of PP1 catalytic subunits (α,β,γ) 7.1 µM. Considering these biochemical parameters, it appears reasonable that biochemical and functional OA-/PP2A-mediated effects with human platelets require incubation times of about 10-30 min and concentrations of 50-200 nM as used in our experiments with intact cells. Furthermore, it appears unlikely that these conditions affect PP1 since initially all OA entering platelets will be bound by the OA target with the highest OA affinity (PP2A) which acts like a sink (PP2A has a~4.700 fold higher OA affinity compared to PP1). Only with higher concentrations or longer incubation times, other targets are hit when PP2A sites are saturated. We have also obtained supporting experimental data for this important topic. A 30 min incubation of only 50 nM OA with intact human platelets showed comparable effects for all phosphoproteins mentioned above ( Figure S6). To rule out some participation of PP1 in these OA effects, we also compared the effects of both OA and tautomycetin (a strong PP1 inhibitor) on thrombin-induced shape change and aggregation of intact human platelets. Platelet shape change is mediated by myosin light chain phosphorylation, which is controlled by myosin light chain kinase (MLCK) and, as MLCK opponent, by myosin light chain phosphatase (MLCP), a specialized PP1 [13,78]. Inhibition of MLCP, for example by cGMP/PKG, strongly inhibits platelet shape change. In our experiments here, thrombin-induced activation and aggregation of human platelets was completely inhibited after 30 min of incubation with 30 µM tautomycetin, and shape change was inhibited as well ( Figure S5b). The absence of a shape change clearly indicates inhibition of the MLCP (PP1) in human platelets [13]. In comparison, high amounts (1 µM) of OA completely inhibited platelet aggregation, but did not inhibit the PP1-mediated shape change ( Figure S5a). Based on these results and on our other information on platelet PP2A and PP1, we can exclude a strong inhibition of PP1 via OA under our conditions. The observed effects upon OA treatment are therefore very likely due to inhibition of PP2A or closely related protein phosphatases in intact human platelets.
Furthermore, OA inhibited aggregation of washed human platelets even with much lower OA concentrations (50 nM, 200 nM, 10 min preincubation) in a dose dependent manner in response to thrombin, collagen and TXA 2 analog. Interestingly, 1 µM OA completely prevented platelet aggregation but not the shape change. We also report that the low OA concentrations (50 nM, 200 nM) increased phosphorylation of several proteins within distinct signaling pathways (Figure 7). Although these functional data clearly support the concept that global inhibition of PP2A by OA results in inhibition of platelet activation it remains to be elucidated, which of the various PP2A subtypes is involved and whether such effects are perhaps PP2A subtype specific. These are important topics for future investigations.
These results show that inhibition of PP2A regulates the phosphorylation and activation of several pathways such as cytoskeletal adaptor proteins (VASP), proteins related to MAPK signalling (MEK, Erk, p38) and PI3K/Akt signalling components. While phosphorylation of VASP, a PKA/PKG substrate, is closely associated with the inhibition of platelet activation [80,81], phosphorylation and activation of MEK, Erk and p38 in platelets is especially important for initiation of platelet activation [51,76,[82][83][84] and Akt phosphorylation/activation for amplifying of platelet activation responses [85]. The observed increased phosphorylation of p38 with 200 nM of OA is of considerable interest, since it is well known that there is an extensive cross-talk between p38, other MAP kinases and PP2A [74]. Others showed that PP2A regulates PLA 2 phosphorylation/activity in human platelets upon platelet activation with thrombin [86] and that p38 is dephosphorylated and therefore inactivated by PP2A [36,85].
The observed OA inhibition of thrombin-stimulated platelet aggregation and OA-induced phosphorylation of MAP kinase components and Akt (which all lead to platelet activation) is at first sight a paradox. However, the OA targets PP2A, the MAP kinases and also VASP are components of large modular networks, which may have different functional endpoints. It is well established that PP2A can positively and negatively affect the Erk-MAP kinase pathway in other systems [74].

Limitations
A S67 ENSA/S62 ARPP19 protein kinase activity has been firmly established for human platelets by our data. However, the precise identity and regulation of this kinase (presumably a MASTL-related protein kinase) remains to be established, which presently is a limitation of our study. Also, while inhibition of platelet PP2A by MASTL-phosphorylated ENSA/ARPP19 has been clearly demonstrated, the precise spectrum of involved PP2A subfamilies is not known yet. Our data also suggest that there are multiple PP2A substrates of different signalling pathways in human platelets. The elucidation of the physiological and pathophysiological function of PP2A in human platelets will require to characterize properties and regulation of the distinct PP2A subtypes and their substrates. However, the similar expression of PP2A subunits in human and murine platelets, as demonstrated here, suggests that murine genetic models with defects or KOs for ENSA/ARPP19 or different PP2A subunits will be important for future studies.

Conclusions
This work shows that the well-established MASTL-ENSA/ARPP19-PP2A pathway, which is essential for the cell cycle control of most dividing eukaryotic cells, is present and regulated in non-dividing, anucleate human platelets (Figure 8).
In our phosphoproteomic studies with intact human platelets, both cAMP-/cGMP-elevating platelet inhibitors strongly induced ENSA/ARPP19 (S109/S104) phosphorylation with concomitant reduced S67/S62 phosphorylation. This suggests an inhibition of MASTL-induced ENSA/ARPP19 phosphorylation by PKA/PKG, similar to an effect of PKA on MAST3 observed in brain [47]. Our data also established that both MASTL-and PKA/PKG-phosphorylated ENSA/ARPP19 are PP2A substrates, whereas only MASTL-phosphorylated ENSA/ARPP19 are PP2A inhibitors. The elucidation of the PKA/PKG phosphorylation effects on ENSA/ARPP19-PP2A function in human platelets may require more detailed kinetic analysis and/or analysis of individual PP2A complexes rather than a mixture of all 14 PP2A forms.
Generally, it has been shown that dephosphorylation by protein phosphatases such as PP2A could inhibit, stimulate or modify the activity/function of substrate proteins resulting in various cellular effects [26]. In human platelets, we could show that detection of proteins being phosphorylated upon PP2A inhibition is possible. This shifts the equlibrium towards kinase activity/phosphorylation as shown in Figures 7 and 8. It will be of considerable interest to establish a larger spectrum of PP2A substrates in platelets, which is now possible by the methods we have developed. Figure 8. Model of PP2A regulation by S67/S62-phosphorylated ENSA/ARPP19 in human platelets and the impact of PKA/PKG on it. Human platelets contain the two closely related, but distinct proteins ENSA and ARPP19, which are phosphorylated (at S67/62) by MASTL (or a MASTL-like kinase) and then potently inhibit certain PP2A variants, leading to an increase in protein phosphorylation. PP2A enzymes are heterotrimers and may have more than 14 different isozymes in human platelets due to the heterogeneity of the B-subunit. One of them, B56δ, is phosphorylated by PKA, which activates the corresponding heterotrimer, leading to decreased protein phosphorylation. Both platelet PKA and PKG strongly phosphorylate ENSA/ARPP19 at their C-terminus (S109/S104), but the functional consequence of this phosphorylation remains to be elucidated. Platelet PKA/PKG also antagonize MASTL-mediated phosphorylation of ENSA/ARPP19 (S67/S62), but the mechanism needs to be elucidated. PP2A can reduce the phosphorylation of the major PKA subtrate VASP (at S157, S239) and thereby antagonize the PKA pathway. The PP2A family in platelets, which can be potently inhibited by low-dose okadaic acid, dephosphorylates many platelet phosphoproteins as partially shown here. PP2A-mediated dephosphorylation antagonizes the effects of many serine/threonine protein kinases including MAPKs, PKA, PKG and their substrates.
Due to the enormous PP2A diversity and possibilities of redundant functions, individual PP2A family members and their regulators/inhibitors need to be studied independently. This is now possible using murine and human genetics [72,[89][90][91], novel selective PP2A inhibitors [62] and cellular models such as human platelets.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Table S1. Diversity, copy number and intracellular concentration of serine/threonine protein phosphatases (catalytic subunits) in human platelets. Table S2. Diversity, copy numbers and intracellular concentration of PP2A subunits (catalytic C, scaffolding A, regulatory B) and of ENSA, ARPP19 and VASP in human and murine platelets. Figure S1. Human platelet PP2A B56δ is phosphorylated at S573 by PKA as well as PKG in intact human platelets. Figure  S2. Forskolin-effects on the S67-/S109-phosphorylation of endogenous ENSA or transfected FLAG-ENSA in HEK293 cells. Figure S3. Site-specific phosphorylation of HisENSA (wildtype and S67A mutant) by MASTL and PKGIβ in combination. Figure S4. Site-specific phosphorylation of GST-ARPP19 by MASTL kinase, PKGIβ and in combination. Figure S5. Inhibition of the PP2A family in human platelets affects platelet aggregation but not shape change. Figure S6. Inhibition of the PP2A family with 50 nM and 200 nM OA in human platelets affects multiple signaling pathways. Figure 8. Model of PP2A regulation by S67/S62-phosphorylated ENSA/ARPP19 in human platelets and the impact of PKA/PKG on it. Human platelets contain the two closely related, but distinct proteins ENSA and ARPP19, which are phosphorylated (at S67/62) by MASTL (or a MASTL-like kinase) and then potently inhibit certain PP2A variants, leading to an increase in protein phosphorylation. PP2A enzymes are heterotrimers and may have more than 14 different isozymes in human platelets due to the heterogeneity of the B-subunit. One of them, B56δ, is phosphorylated by PKA, which activates the corresponding heterotrimer, leading to decreased protein phosphorylation. Both platelet PKA and PKG strongly phosphorylate ENSA/ARPP19 at their C-terminus (S109/S104), but the functional consequence of this phosphorylation remains to be elucidated. Platelet PKA/PKG also antagonize MASTL-mediated phosphorylation of ENSA/ARPP19 (S67/S62), but the mechanism needs to be elucidated. PP2A can reduce the phosphorylation of the major PKA subtrate VASP (at S157, S239) and thereby antagonize the PKA pathway. The PP2A family in platelets, which can be potently inhibited by low-dose okadaic acid, dephosphorylates many platelet phosphoproteins as partially shown here. PP2A-mediated dephosphorylation antagonizes the effects of many serine/threonine protein kinases including MAPKs, PKA, PKG and their substrates.
Finally, what is the role of a cell cycle checkpoint in the anucleate platelets? We suggest that PP2A checkpoint systems not only control the cell cycle, but also other essential cell functions, especially in platelets. Research efforts with yeast presenting a well-established Greatwall-ENSA-PP2A pathway, very recently emphasized the need to search for novel PP2A substrates and their functions [87,88]. The various PP2A holoenzymes present in human platelets could have overlapping, distinct and also opposing effects on important disease-relevant platelet responses, such as integrin activation, secretion and thrombin generation. Therefore, it will be important to address the substrate and signalling specificity of the various PP2A forms and which of them is inhibited by S67-phosphorylated ENSA, S62-phosphoryated ARPP19 or both.
Due to the enormous PP2A diversity and possibilities of redundant functions, individual PP2A family members and their regulators/inhibitors need to be studied independently. This is now possible using murine and human genetics [72,[89][90][91], novel selective PP2A inhibitors [62] and cellular models such as human platelets.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4409/9/2/472/s1, Table S1. Diversity, copy number and intracellular concentration of serine/threonine protein phosphatases (catalytic subunits) in human platelets. Table S2. Diversity, copy numbers and intracellular concentration of PP2A subunits (catalytic C, scaffolding A, regulatory B) and of ENSA, ARPP19 and VASP in human and murine platelets. Figure  S1. Human platelet PP2A B56δ is phosphorylated at S573 by PKA as well as PKG in intact human platelets. Figure S2. Forskolin-effects on the S67-/S109-phosphorylation of endogenous ENSA or transfected FLAG-ENSA in HEK293 cells. Figure S3. Site-specific phosphorylation of HisENSA (wildtype and S67A mutant) by MASTL and PKGIβ in combination. Figure S4. Site-specific phosphorylation of GST-ARPP19 by MASTL kinase, PKGIβ and in combination. Figure S5. Inhibition of the PP2A family in human platelets affects platelet aggregation but not shape change. Figure S6