The ATP,Mg-dependent protein phosphatase. Regulation by inhibitor-1 or modulator protein and stabilizing role of Mg2+ ions

The activation of the ATP,Mg-dependent protein phosphatase [Fc.M] has been shown to involve a transient phosphorylation of the modulator subunit (M) and consequent isomerization of the catalytic subunit (Fc) into its active conformation (Jurgensen, S., Shacter, E., Huang, C. Y., Chock, P. B., Yang, S. -D., Vandenheede, J. R., and Merlevede, W. (1984) J. Biol. Chem. 259, 5864-5870). The modulator subunit constitutes the inactivating force for the enzyme, but the slow intramolecular inactivation of the phosphatase can be prevented or blocked by the addition of either the phosphorylated inhibitor-1 or Mg2+ ions. Autodephosphorylation of the modulator subunit is not prevented by the phosphoinhibitor-1, suggesting that the ATP,Mg-dependent phosphatase binds the phosphomodulator subunit in a very specific manner, different from the way it binds exogenous phosphoprotein substrates. Alternatively, the autodephosphorylation of the modulator subunit is catalyzed at a separate active site on the enzyme, which is not influenced by the binding of phosphoinhibitor-1. The phosphoinhibitor-1 does not prevent the activation of the enzyme by kinase FA when added at concentrations that totally inhibit the potential phosphorylase phosphatase activity. These results, together with other already published information, suggest separate autonomic controls of the ATP,Mg-dependent phosphatase activity by inhibitor-1 and the modulator protein through the presence of specific regulatory subunits on the enzyme.     

The ATP, Mg-dependent phosphatase: role of Mg ions in the expression of the phosphorylase phosphatase activity

The activation of the ATP, Mg-dependent phosphatase [FCM] by kinase FA has been shown to involve the phosphorylation or thiophosphorylation of the modulator subunit [M] and the consequent isomerization of the catalytic subunit [FC] into the active conformation. The inactive catalytic subunit [free FC] exhibits substantial activity in the presence of non-physiological concentrations of Mn ions whereas the Mn2+-activation of the intact FCM-enzyme requires the proteolytic destruction of the modulator subunit. The present study points to the importance of Mg2+ in the activation of the phosphatase. The inactive catalytic unit can be activated by millimolar concentrations of Mg2+ and the thiophosphorylated FCM-enzyme only expresses its phosphorylase phosphatase activity after a subsequent trypsin treatment in the presence of Mg ions.     

Phosphorylation of the modulator protein of the ATP, Mg-dependent protein phosphatase by casein kinase TS. Reversal by PCS phosphatases and control by distinct phosphorylation site(s)

The phosphorylation by casein kinase TS (II) of the modulator protein of the ATP, Mg-dependent phosphatase increases after preincubation with the PCSH1 phosphatase or with the catalytic subunit of the ATP, Mg-dependent phosphatase. Dephosphorylation by the two phosphatases combined leads to the incorporation of 2 mol phosphate per mol modulator (at Ser residues). Occupancy of the ATP, Mg-dependent phosphatase phosphorylation site(s) is a negative determinant in the phosphorylation of the modulator by kinase TS. Among the PCS phosphatases PCSH1 shows the highest activity toward the 32P-Ser residues labeled by kinase TS in untreated or previously dephosphorylated modulator, while the ATP, Mg-dependent phosphatase is totally ineffective. Protamine stimulates all phosphatase activities, so that the catalytic subunit of the ATP, Mg-dependent phosphatase becomes almost as effective as the PCSC phosphatase in dephosphorylating the kinase TS sites.     

ATP x Mg-dependent protein phosphatase from rabbit skeletal muscle. I. Purification of the enzyme and its regulation by the interaction with an activating protein factor

An ATP x Mg-dependent protein phosphatase (FC) was purified to near homogeneity from rabbit muscle. The enzyme was completely devoid of any spontaneous activity but could be activated by a protein activator (FA) in the presence of ATP and Mg ions. The inactive phosphatase migrated as a single protein band on sodium dodecyl sulfate-gel electrophoresis, and in discontinuous gel electrophoresis, where the potential phosphatase activity was located in the main protein band. The molecular weight determined by sodium dodecyl sulfate electrophoresis or by sucrose density centrifugation was found to be 70,000. FC migrated on gel filtration as a 140,000 molecular weight species. The activation by FA was not paralleled by an incorporation of [32P]-phosphate into the ATP x Mg-dependent phosphatase, and from the kinetics of activation a protein-protein interaction with ATP x Mg as a necessary factor, can be inferred as the mechanism of activation. After activation by FA and ATP X Mg, the purified enzyme had a specific activity of 10,000 units/mg of protein, and a Km for rabbit muscle phosphorylase a of approximately 1.0 mg/ml. The activated enzyme did not release [32P]phosphate from 32[-labeled rabbit muscle synthase b, prepared from glucagon-treated dogs. It did, however, remove all the 32P label from phosphorylase b kinase, autophosphorylated to the level of 2.0 mol/mol of 1.3 X 10(6) molecular weight.     

The heat labile phosphatase modulator (inhibitor-2) complex from rabbit skeletal muscle

The heat stable phosphatase modulator protein (inhibitor-2) has been shown to play a crucial role in the reversible ATP, Mg-dependent activation of a multisubstrate protein phosphatase. The modulator activity is acid and heat stable and resides in a small asymmetrical protein which, after boiling migrates in sucrose density gradient centrifugation with a molecular weight of 17K. The present report shows that in unboiled rabbit skeletal muscle preparations all the modulator activity is found associated with a heat labile protein component, which imposes an important regulatory feature on the heat stable activity. The heat labile complex migrates in sucrose density gradient centrifugation as a M_r = 70K protein.     

Kinase FA-mediated regulation of rabbit skeletal muscle protein phosphatase. Reversible phosphorylation of the modulator subunit

A mechanism of activation of the ATP.Mg-dependent protein phosphatase (FC.M) has been proposed (Jurgensen, S., Shacter, E., Huang, C. Y., Chock, P. B., Yang, S.-D., Vandenheede, J. R., and Merlevede, W. (1984) J. Biol. Chem. 259, 5864-5870) in which a transient phosphorylation by the kinase FA of the modulator subunit (M) is the driving force for the transition of the inactive catalytic subunit (FC) into its active conformation. Incubation of FC.M with kinase FA and Mg2+ and adenosine 5'-(gamma-thio)triphosphate results in thiophosphorylation of M and also a conformational change in the phosphatase catalytic subunit; however, the enzyme remains inactive. Proteolysis of this inactive, thiophosphorylated complex causes proteolytic destruction of the modulator subunit and yields an active phosphorylase phosphatase species. Similar treatment of the native inactive enzyme does not yield active phosphatase. Evidence is presented, suggesting that a molecule of modulator is bound at an "inhibitory site" on the native enzyme. This modulator does not prevent the conformational change in the phosphatase catalytic subunit upon incubation with kinase FA and ATP.Mg but does partially inhibit the expression of the phosphorylase phosphatase activity.     

Role of the modulator protein in the interconversion of rabbit skeletal muscle protein phosphatase

The major active protein phosphatase present in a rabbit skeletal muscle extract is associated with the glycogen particle and migrates in sucrose density gradient centrifugation as a Mr = 70,000 protein and contains modulator activity. Addition of extra modulator protein causes a time- and concentration-dependent conversion of the enzyme to an inactive FA-ATP, Mg-dependent form. The intrinsic modulator in the active phosphatase is destroyed by limited proteolysis without an appreciable change in the phosphatase activity. The proteolyzed active enzyme has a lower molecular weight (Mr = 40,000) and it reassociates with the modulator producing a FA-ATP, Mg-dependent enzyme form (Mr = 60,000). The modulator protein is used stoichiometrically in the activation of the ATP, Mg-dependent phosphatase. This is in agreement with the presence of one unit of modulator activity per unit of native spontaneously active phosphatase.     

Dephosphorylation of the deinhibitor protein by the PCSH protein phosphatase

The deinhibitor protein, responsible for the decreased sensitivity of the ATP,Mg-dependent protein phosphatase to inhibitor-1 and the modulator protein, is inactivated by cyclic AMP-dependent protein kinase and reactivated by dephosphorylation. The specificity of this reaction was tested with the ATP,Mg-dependent phosphatase in its activated or spontaneously active form, four different forms of polycation-stimulated phosphatases (PCSH, PCSM, PCSL and PCSC) and calcineurin. Only the high -Mr polycation-stimulated protein phosphatase (PCSH), but not its catalytic subunit (PCSC), shows a high degree of specificity for the deinhibitor protein. Deinhibitor phosphatase activity of PCSH is affected neither by polycations nor by Mn ions.     

Selective activation of the two catalytic sites in the ATP.Mg-dependent phosphoprotein phosphatase by kinase Fa and Mn2+ ion

1. Although Mn2+ could mimic kinase FA/ATP.Mg to activate ATP.Mg-dependent protein phosphatase, strong indications have been obtained that the Mn2(+)-activated and FA/ATP.Mg-activated phosphatase forms are not identical in terms of their substrate specificities and catalytic properties. 2. Both Mn2(+)-activated and FA/ATP.Mg-activated phosphatase forms readily dephosphorylate 32P-labeled phosphorylase a and myelin basic protein (MBP), however the Mn2(+)-activated phosphatase displays activity preferentially against [32P]MBP and FA/ATP.Mg-activated phosphatase preferentially dephosphorylates [32P]phosphorylase a, representing a unique control mechanism to regulate the substrate specificity of multisubstrate protein phosphatase in mammalian tissues.     

Role of the deinhibitor protein in the interconversion of the ATP,Mg-dependent protein phosphatase

The small molecular weight (± 9,000) heat stable deinhibitor protein, isolated from dog liver, not only protects the multisubstrate protein phosphatase from inhibition by inhibitor-1 and the modulator protein. It prevents the conversion of the active enzyme to the ATP,Mg-dependent enzyme form brought about by the modulator protein, and also affects the activation of the ATP,Mg-dependent protein phosphatase, probably by stabilizing the enzyme in its active conformation during the reversible activation by protein kinase F_A. Therefore the deinhibitor protein could be an important factor in the process of glycogen synthesis, which requires glycogen synthase and phosphorylase as dephosphorylated enzymes.     

Purification, subunit composition and regulatory properties of the ATP X Mg2+-dependent form of type I phosphoprotein phosphatase from bovine heart

The ATP X Mg2+-dependent phosphoprotein phosphatase has been purified from bovine heart to near-homogeneity. It is a heterodimer (75 kDa) consisting of a catalytic (C) subunit (40 kDa) and a regulatory (R) subunit (35 kDa). The R subunit, which is identical to inhibitor-2, is transiently phosphorylated during activation of the enzyme catalyzed by phosphatase-1 kinase (FA). Maximal activation requires preincubation of the phosphatase with FA and ATP X Mg2+. However, relatively low yet definitively demonstrable basal activity can be expressed by Mg2+ alone (ranging from 3% to 10% of the FA X ATP X Mg activity, depending on the degree of endogenous proteolytic damage of the phosphatase during purification), but not by either FA or ATP alone. Limited trypsinization results in a rapid and total degradation of the R subunit and partial degradation of the 40-kDa C subunit to active proteins of 35-38 kDa. The resulting 'nicked' C subunit of 35-38 kDa is no longer dependent on FA for activation and can be fully activated by Mg2+ (or Mn2+) alone. Endogenous proteolytic damage of the R subunit also results in an increase of activity that can be expressed by M2+ alone with a concomitant decrease of the FA-dependent activation. Although Mn2+ is slightly more effective than Mg2+ in expressing the holoenzyme basal activity, the activation by Mn2+ is only about 60% of that of Mg2+ when FA and ATP are also present. In the activation by adenosine 5'-[gamma-thio]triphosphate (ATP[gamma S]), Co2+ is the most effective cofactor. The activation by ATP[gamma S] X Co2+ is more than 50% of that by ATP X Mg2+. The present studies indicate that Mg2+ is the natural divalent cation for the FA-catalyzed activation in which Mg2+ plays two distinctly different roles: it forms Mg2+ X ATP which serves as a substrate for the kinase; it acts as an essential cofactor for the catalytic function of the phosphatase. The discrepancies between the results obtained by this and other laboratories with respect to the effectiveness of Mg2+ and ATP[gamma S] in the activation of the phosphatase are discussed.     

Cyclic phosphorylation-dephosphorylation of rhodopsin in retina by protein kinase FA (the activator of ATP.Mg-dependent protein phosphatase).

The ATP.Mg-dependent protein phosphatase activating factor (protein kinase FA) was identified to exist in bovine retina. Furthermore, rhodopsin, the visual light pigment associated with rod outer segments in retina, could be well phosphorylated by kinase FA to about 0.9 mol of phosphates per mol of protein. Moreover, more than 90% of the phosphates in [32P]-rhodopsin could be completely removed by ATP.Mg-dependent protein phosphatase and the rhodopsin phosphatase activity was strictly kinase FA-dependent. Taken together, the results provide initial evidence that a cyclic phosphorylation-dephosphorylation of rhodopsin can be controlled by the retina-associated protein kinase FA, representing an efficient cyclic cascade mechanism possibly involved in the rapid regulation of rhodopsin function in retina.     

Insulin-induced increases in the activity of the spontaneously active and ATP.Mg-dependent forms of phosphatase-1 in alloxan-diabetic rat liver

Liver supernatant from normal and alloxan-diabetic rats was fractionated by DEAE-cellulose chromatography and the separated phosphoprotein phosphatase fractions were assayed with [32P]histone f2b, [32P]phosphorylase a and [32P]phosphorylase kinase as substrates. In diabetic rat liver, one of the phosphatase fractions found in the normal liver was significantly reduced. This fraction was identified as a mixture of the spontaneously active form and the ATP . Mg-dependent form of phosphoprotein phosphatase-1 (Fc) based on sensitivity to inhibitor-2, substrate specificity, and the fact that it could be activated 42-70% by glycogen synthase kinase-3 in the presence of ATP . Mg. Further analysis of this fraction showed that liver cytosol from diabetic rats contained 62-79% lower spontaneously active phosphatase-1 activity and 40-51% lower combined spontaneously active and ATP . Mg-dependent protein phosphatase-1 (Fc) activity. Insulin administration increased the spontaneously active and the ATP . Mg-dependent protein phosphatase-1 activities approximately 45% and 36%, respectively, in alloxan-diabetic rats. These data imply that the lower levels of spontaneously active phosphatase-1 activity in diabetic rat liver cannot be explained by presuming phosphatase-1 to have been present as Fc, the inactive form. Moreover, insulin restored the total activity of the spontaneously active and activatable forms of phosphatase-1 to those present in normal liver implying that both forms of phosphatase-1 activity are under hormonal control.     

Insulin induces activation of kinase FA in membranes and thereby promotes activation of ATP.Mg-dependent phosphatase in adipocytes

Exposure of rat adipocytes to physiological concentrations of insulin resulted in a time- and concentration-dependent activation-translocation of kinase FA (an activating factor of ATP.Mg-dependent protein phosphatase) in plasma membranes and the subsequent activation of ATP.Mg-dependent protein phosphatase in the cytosol. The insulin-induced activation of membrane-associated kinase FA and cytosolic ATP.Mg-dependent protein phosphatase occurred very rapidly, reaching the maximal activity levels within 3 min. Moreover, the insulin effect is transient; the insulin-stimulated FA activity in membranes and ATP.Mg-dependent phosphatase activity in the cytosol returned to control levels within 30 min. It is concluded that insulin may induce the activation of kinase FA in membranes and thereby promotes the activation of ATP.Mg-dependent multifunctional protein phosphatase in the cytosol of rat adipocytes in order to mediate some of its intracellular effects through the dephosphorylation reactions. The release of factor FA from plasma membranes may represent one of the transmembrane signalling mechanisms for insulin actions.     

Identification and characterization of an ATP.Mg-dependent protein phosphatase from pig brain

Substantial amounts of ATP.Mg-dependent phosphorylase phosphatase (Fc. M) and its activator (kinase FA) were identified and extensively purified from pig brain, in spite of the fact that glycogen metabolism in the brain is of little importance. The brain Fc.M was completely inactive and could only be activated by ATP.Mg and FA, isolated either from rabbit muscle or pig brain. Kinetical analysis of the dephosphorylation of endogenous brain protein indicates that Fc.M could dephosphorylate 32P-labeled myelin basic protein (MBP) and [32P]phosphorylase alpha at a comparable rate and moreover, this associated MBP phosphatase activity was also strictly kinase FA/ATP.Mg-dependent, demonstrating that MBP is a potential substrate for Fc.M in the brain. By manipulating MBP and inhibitor-2 as specific potent phosphorylase phosphatase inhibitors, we further demonstrate that 1) Fc.M contains two distinct catalytic sites to dephosphorylate different substrates, and 2) brain MBP may be a physiological trigger involved in the regulation of protein phosphatase substrate specificity in mammalian nervous tissues.     

Activation of bovine heart ATP-MG2+-dependent phosphoprotein phosphatase: isolation of a phosphoenzyme intermediate and its conversion to the active form via a Mg2+-dependent autodephosphorylation reaction

The ATP-Mg2+-dependent protein phosphatase, a holoenzyme form of type I protein phosphatase (phosphatase-1) requires the action of phosphatase-1 kinase (FA) for activation. The enzyme (75 kDa) purified from bovine heart consists of a catalytic (C) and a regulatory (R) subunit of 40 kDa and 34 kDa, respectively, and activation is associated with phosphorylation of the R-sub-unit. A procedure has been developed for isolation of [32P]phosphatase-1 ( [32P]E-P) in non-denatured form. In the absence of divalent cation, [32P]E-P is catalytically inactive and the phosphorylation is stable. Addition of Mg2+ triggers autodephosphorylation of [32P]E-P with concomitant generation of phosphorylase phosphatase activity. The autodephosphorylation/activation process is dependent on Mg2+ concentration. The KA value for Mg2+ is 0.6 mM. The phosphorylase phosphatase activity generated from the release of 1 mol. 32P is 1.1 X 10(12) units which is equivalent to 15,000 units per mg enzyme protein. The present findings provide direct evidence that the phosphorylated phosphatase-1 is not the active form (Ea). Instead, Ea is directly produced from the intermediate by a Mg2+-dependent autodephosphorylation reaction.     

Epidermal growth factor induces activation of protein kinase FA and ATP.Mg-dependent protein phosphatase in A431 cells

The cytosolic fractions from epidermal growth factor (EGF)-treated A431 cells exhibit a marked increase in activities of ATP.Mg-dependent protein phosphatase and its activating factor (protein kinase FA) when compared to controls in the absence of EGF. By contrast, the Triton X-100-solubilized membrane fractions from the same EGF-treated cells exhibit a corresponding decrease in protein kinase FA activity. The EGF-dependent activation of protein kinase FA and ATP.Mg-dependent protein phosphatase occurred within physiological concentrations of EGF (ED50 = 5 x 10(-10) M). The changes of kinase and phosphatase activities which were measured concomitantly exhibit very similar characteristics as to EGF sensitivity and time dependence. The EGF-induced kinase and phosphatase activation occurred very rapidly, reaching the maximal activity levels within 3 min. Moreover, the EGF effect is transient; both EGF-stimulated phosphatase and kinase activities returned to control levels within 30 min. Taken together, the results suggest that EGF may induce the activation of kinase FA in the membrane and thereby promotes the activation of ATP.Mg-dependent phosphatase in the cytosol. Exposure of A431 cells to exogenous phospholipase C also resulted in the activation of endogenous kinase FA and ATP.Mg-dependent phosphatase in a similar pattern produced by EGF. This further suggests that phospholipase C can mimic EGF to mediate the activation of kinase FA and ATP.Mg-dependent phosphatase in A431 cells. By its dual role as a multisubstrate protein kinase and as an activating factor of multisubstrate protein phosphatase, protein kinase FA may represent a transmembrane signal of EGF.     

Phosphorylation and dephosphorylation of human platelet surface proteins by an ecto-protein kinase/phosphatase system.

We have characterized a novel ecto-protein kinase activity and a novel ecto-protein phosphatase activity on the membrane surface of human platelets. Washed intact platelets, when incubated with [gamma-32P]ATP in Tyrode's buffer, showed the phosphorylation of a membrane surface protein migrating with an apparent molecular mass of 42 kDa on 5-15% SDS polyacrylamide gradient gels. The 42 kDa protein could be further resolved on 15% SDS gels into two proteins of 39 kDa and 42 kDa. In this gel system, it was found that the 39 kDa protein became rapidly phosphorylated and dephosphorylated, whereas the 42 kDa protein was phosphorylated and dephosphorylated at a much slower rate. NaF inhibited the dephosphorylation of these proteins indicating the involvement of an ecto-protein phosphatase. The platelet membrane ecto-protein kinase responsible for the phosphorylation of both of these proteins was identified as a serine kinase and showed dependency on divalent cations Mg2+ or Mn2+ ions. Ca2+ ions potentiated the Mg(2+)-dependent ecto-protein kinase activity. The ecto-protein kinase rapidly phosphorylated histone and casein added exogenously to the extracellular medium of intact platelets. Following activation of platelets by alpha-thrombin, the incorporation of [32P]phosphate from exogenously added [gamma-32P]ATP by endogenous protein substrates was reduced by 90%, suggesting a role of the ecto-protein kinase system in the regulation of platelet function. The results presented here demonstrate that both protein kinase and protein phosphatase activities reside on the membrane surface of human platelets. These activities are capable of rapidly phosphorylating and dephosphorylating specific surface platelet membrane proteins which may play important roles in early events of platelet activation and secretion.     

Protein tyrosine phosphorylation in the cyanobacterium Anabaena sp. strain PCC 7120.

Components of a protein tyrosine phosphorylation/dephosphorylation network were identified in the cyanobacterium Anabaena sp. strain PCC 7120. Three phosphotyrosine (P-Tyr) proteins of 27, 36, and 52 kDa were identified through their conspicuous immunoreactions with RC20H monoclonal antibodies specific for P-Tyr. These immunoreactions were outcompeted completely by free P-Tyr (5 mM) but not by phosphoserine or phosphothreonine. The P-Tyr content of the three major P-Tyr proteins and several minor proteins increased with their time of incubation in the presence of Mg-ATP and the protein phosphatase inhibitors sodium orthovanadate and sodium fluoride. Incubation of the same extracts with [gamma-32P]ATP but not [alpha-32P]ATP led to the phosphorylation of five polypeptides with molecular masses of 20, 27, 52, 85, and 100 kDa. Human placental protein tyrosine phosphatase 1B, with absolute specificity for P-Tyr, liberated significant quantities of 32Pi from four of the polypeptides, confirming that a portion of the protein-bound phosphate was present as 32P-Tyr. Alkaline phosphatase and the dual-specificity protein phosphatase IphP from the cyanobacterium Nostoc commune UTEX 584 also dephosphorylated these proteins and did so with greater apparent efficiency. Two of the polypeptides were partially purified, and phosphoamino analysis identified 32P-Tyr, [32P]phosphoserine, and [32P]phosphothreonine. Anabaena sp. strain PCC 7120 cell extracts contained a protein tyrosine phosphatase activity that was abolished in the presence of sodium orthovanadate and inhibited significantly by the sulfhydryl-modifying agents p-hydroxymercuriphenylsulfonic acid and p-hydroxymercuribenzoate as well as by heparin. In Anabaena sp. strain PCC 7120 the presence and/or phosphorylation status of P-Tyr proteins was influenced by incident photon flux density.     

Phosphorylation of tryptophanyl-tRNA-synthase in extracts of bovine pancreas

The polypeptide with a mobility of the tryptophanyl-tRNA-synthetase subunit can be labeled in bovine pancreas extracts from [gamma-32P]ATP. Immunoprecipitation analysis with monospecific polyclonal antibodies against the enzyme as well as identification of [32P]phosphoamino acids in the immunoprecipitate revealed that in bovine pancreas extracts tryptophanyl-tRNA-synthetase undergoes phosphorylation at serine residues. The level of phosphorylation does not change in the presence of activity modulators of cAMP-, cGMP- and Ca2(+)-dependent protein kinases, decreases after addition of phosphoseryl/phosphothreonyl-protein phosphatase inhibitors and increases in the presence of their activators. It was supposed that phosphorylation of tryptophanyl-tRNA-synthetase catalyzed by seryl/threonyl-specific protein kinase depends on the activity of phosphoseryl/phosphothreonyl-phosphatase.