Amino acid sequence of a region on the glycogen-binding subunit of protein phosphatase-1 phosphorylated by cyclic AMP-dependent protein kinase

Multiple interactions within the AKAP220 signaling complex contribute to protein phosphatase 1 regulation

The phosphorylation status of cellular proteins is controlled by the opposing actions of protein kinases and phosphatases. Compartmentalization of these enzymes is critical for spatial and temporal control of these phosphorylation/dephosphorylation events. We previously reported that a 220-kDa A-kinase anchoring protein (AKAP220) coordinates the location of the cAMP-dependent protein kinase (PKA) and the type 1 protein phosphatase catalytic subunit (PP1c) (Schillace, R. V., and Scott, J. D. (1999) Curr. Biol. 9, 321-324). We now demonstrate that an AKAP220 fragment is a competitive inhibitor of PP1c activity (K(i) = 2.9 +/- 0.7 micrometer). Mapping studies and activity measurements indicate that several protein-protein interactions act synergistically to inhibit PP1. A consensus targeting motif, between residues 1195 and 1198 (Lys-Val-Gln-Phe), binds but does not affect enzyme activity, whereas determinants between residues 1711 and 1901 inhibit the phosphatase. Analysis of truncated PP1c and chimeric PP1/2A catalytic subunits suggests that AKAP220 inhibits the phosphatase in a manner distinct from all known PP1 inhibitors and toxins. Intermolecular interactions within the AKAP220 signaling complex further contribute to PP1 inhibition as addition of the PKA regulatory subunit (RII) enhances phosphatase inhibition. These experiments indicate that regulation of PP1 activity by AKAP220 involves a complex network of intra- and intermolecular interactions.

Overexpression of the glycogen targeting (G(M)) subunit of protein phosphatase-1

The G(M) glycogen-targeting subunit of protein phosphatase-1 (PP1) is believed to be involved in dephosphorylation of the enzymes of glycogen metabolism. To assess the roles of G(M) on glycogen metabolism, we created site-directed G(M) mutants and overexpressed them in Chinese hamster ovary (CHO) cells expressing human insulin receptor. Overexpressed G(M) recruited glycogen synthase as well as PP1 to the glycogen pellet, and upregulated basal glycogen synthase activity. Overexpressed G(M)-67A (Ser-67 replaced with alanine) exhibited decreased sensitivity to suppression of glycogen synthase activity by forskolin, while overexpression of G(M)-48A (Ser-48 replaced with alanine) preserved glycogen synthase activation in response to insulin. These observations indicate that in CHO cells overexpressing G(M); (1) G(M) translocates glycogen synthase to the glycogen pellet and affected basal glycogen synthase, (2) Ser-67 might be involved in the suppression of glycogen synthase activity by glycogenolytic agents, and (3) Ser-48 might not commit to activation of glycogen synthase by insulin.

Introduction to signal transduction: a primer for untangling the web of intracellular messengers

The field of signal transduction developed from our need to understand the specific mechanisms involved in the transmission of extracellular signals to intracellular actions in target cells. As our comprehension of cellular function has grown, it has become clear that the need to understand signal transduction events has invaded all fields of biological science. Many scientists learned about adenylate cyclase and cyclic AMP when they were students but went on to focus on some other aspect of science; now they find the need to apply studies of signal transduction to their own work. However, the field of signal transduction has progressed so rapidly that from starting with a rudimentary knowledge of adenylate cyclase and cyclic AMP to understanding the field and applying the knowledge to one's own work seems insurmountable. The goal of this commentary is to provide a starting point for those who recognize the need to understand the mechanisms of signal transduction but do not know where or how to begin.

Identification and characterization of cAMP-dependent protein kinase and its possible direct interactions with protein phosphatase-1 in marine dinoflagellates

We have examined the nature of signal transduction involving reversible protein phosphorylation in marine Prorocentrale species. Of particular interest is the marine dinoflagellate Prorocentrum lima in which the tumour promoter okadaic acid is produced and may interfere with signal transduction. We have identified cAMP-dependent protein kinase (PKA) activity in P. lima, P. micans, and P. minimum. The P. lima enzyme was characterized biochemically and appears to consist of two different isoforms in the R2C2 configuration. Whole cell extracts of P. micans and P. minimum treated with the specific PKA inhibitor peptide PKI (5-24) or cAMP demonstrated altered intensities of phosphopeptide 32P labeling, most likely involving regulation of a protein phosphatase via PKA activity. A primary candidate for PKA regulation is protein phosphatase-1 (PP-1), which in P. lima possesses a classical PKA consensus phosphorylation site. We demonstrate that a peptide fragment of PP-1 from P. lima corresponding to this PKA phosphorylation site can be effectively phosphorylated by PKA and dephosphorylated by calcineurin. We speculate that PP-1 activity among several lower eukaryotes may be mediated directly by reversible phosphorylation. Higher eukaryotes may have developed inhibitor proteins to provide more complex regulation of protein phosphatase activity.

Stimulation of FA and casein kinase II by insulin in 3T3-L1 cells

Insulin stimulates protein phosphatase-1 and FA, assayed as phosphatase-1 activator, in 3T3-L1 cells. Since other kinases, such as casein kinase-II may also contribute to such FA activity, we assayed casein kinase-II and FA as peptide kinase on extracts from 3T3-L1 cells that had been exposed to insulin for various times. Under such conditions FA, assayed as phosphatase-1 activator, was stimulated 2-3-fold within 1-2 min. Casein kinase-II was stimulated about 2-fold but at a slightly later time (2-3 min) than FA, making it unlikely that casein kinase-II contributes to FA stimulation. Insulin slightly stimulated also the kinase activity of FA towards a synthetic peptide at 2 min, thus confirming the FA activation seen when FA was assayed as activator of phosphatase-1.

The molecular mechanism by which insulin stimulates glycogen synthesis in mammalian skeletal muscle

The ability of insulin to promote the phosphorylation of some proteins and the dephosphorylation of others is paradoxical. An insulin-stimulated protein kinase is shown to activate the type-1 protein phosphatase that controls glycogen metabolism, by phosphorylating its regulatory subunit at a specific serine. Furthermore, the phosphorylation of this residue is stimulated by insulin in vivo. Increased and decreased phosphorylation of proteins by insulin can therefore be explained through the same basic underlying mechanism.

Synthetic peptides as model substrates for the study of the specificity of the polycation-stimulated protein phosphatases

The substrate specificity of the different forms of the polycation-stimulated (PCS, type 2A) protein phosphatases and of the active catalytic subunit of the ATP, Mg-dependent (type 1) phosphatase (AMDC) was investigated, using synthetic peptides phosphorylated by either cyclic-AMP-dependent protein kinase or by casein kinase-2. The PCS phosphatases are very efficient toward the Thr(P) peptides RRAT(P)VA and RRREEET(P)EEE when compared with the Ser(P) analogues RRAS(P)VA and RRREEES(P)EEEAA. Despite their distinct sequence, both Thr(P) peptides are excellent substrates for the PCSM and PCSH1 phosphatases, being dephosphorylated faster than phosphorylase a. The slow dephosphorylation of RRAS(P)VA by the PCS phosphatases could be increased substantially by the insertion of N-terminal (Arg) basic residues. In contrast with the latter, the AMDC phosphatase shows very poor activity toward all the phosphopeptides tested, without preference for either Ser(P) or Thr(P) peptides. However, N-terminal basic residues also favor the dephosphorylation of otherwise almost inert substrates by the AMDC phosphatase. Hence, while the dephosphorylation of Thr(P) substrates by the PCS phosphatases is highly favored by the nature of the phosphorylated amino acid, phosphatase activity toward Ser(P)-containing peptides may require specific determinants in the primary structure of the phosphorylation site.

Identification of three in vivo phosphorylation sites on the glycogen-binding subunit of protein phosphatase 1 from rabbit skeletal muscle, and their response to adrenaline

The in vivo phosphorylation stoichiometries of 4 serines on the glycogen-binding (G)-subunit of protein phosphatase 1 (PP1) have been determined. In fed rabbits injected with propranolol stoichiometries (mol/mol) were: site 1 (0.67 +/- 0.09), site 2 (0.20 +/- 0.07), site 3a (0.23 +/- 0.01) and site 3b (0). After injection with adrenalin they became: site 1 (0.90 +/- 0.02), site 2 (0.72 +/- 0.01), site 3a (0.23 +/- 0.02) and site 3b (0). These results, together with other data, establish that site 2 phosphorylation by cyclic AMP-dependent protein kinase triggers dissociation of PP1 from the G-subunit in vivo. They also demonstrate that a residue phosphorylated in vitro by glycogen synthase kinase 3 (site 3a) is phosphorylated in vivo.

Regulation of protein phosphatase-1G from rabbit skeletal muscle. 1. Phosphorylation by cAMP-dependent protein kinase at site 2 releases catalytic subunit from the glycogen-bound holoenzyme

The glycogen-associated form of protein phosphatase-1 (PP-1G) is a heterodimer comprising a 37-kDa catalytic (C) subunit and a 161-kDa glycogen-binding (G) subunit, the latter being phosphorylated by cAMP-dependent protein kinase at two serine residues (site 1 and site 2). Here the amino acid sequence surrounding site 2 has been determined and this phosphoserine shown to lie 19 residues C-terminal to site 1 in the primary structure. The sequence in this region is: (sequence; see text) At physiological ionic strength, phosphorylation of glycogen-bound PP-1G was found to release all the phosphatase activity from glycogen. The released activity was free C subunit, and not PP-1G, while the phospho-G subunit remained bound to glycogen. Dissociation reflected a greater than or equal to 4000-fold decrease in affinity of C subunit for G subunit and was readily reversed by dephosphorylation. Phosphorylation and dephosphorylation of site 2 was rate-limiting for dissociation and reassociation of C subunit. Release of C subunit was also induced by the binding of anti-site-1 Fab fragments to glycogen-bound PP-1G. At near physiological ionic strength, PP-1G and glycogen concentration, site 2 was autodephosphorylated by PP-1G with a t0.5 of 2.6 min at 30 degrees C, approximately 100-fold slower than the t0.5 for dephosphorylation of glycogen phosphorylase under the same conditions. Site 2 was a good substrate for all three type-2 phosphatases (2A, 2B and 2C) with t0.5 values less than those toward the alpha subunit of phosphorylase kinase. At the levels present in skeletal muscle, the type-2A and type-2B phosphatases are potentially capable of dephosphorylating site 2 in vivo within seconds. Site 1 was at least 10-fold less effective than site 2 as a substrate for all four phosphatases. In conjunction with information presented in the following paper in this issue of this journal, the results substantiate the hypothesis that PP-1 activity towards the glycogen-metabolising enzymes is regulated in vivo by reversible phosphorylation of a targetting subunit (G) that directs the C subunit to glycogen--protein particles. The efficient dephosphorylation of site 2 by the Ca2+/calmodulin-stimulated protein phosphatase (2B) provides a potential mechanism for regulating PP-1 activity in response to Ca2+, and represents an example of a protein phosphatase cascade.

Glycogen-bound type-1 phosphatase: isolation and dissociation of a complex containing undegraded G-subunit

A high molecular mass type-1 phosphatase complex can be isolated from muscle glycogen particles by a fast procedure that preserves the glycogen-binding subunit of phosphatase called G from proteolysis. G can be dissociated from such complex by ion exchange chromatography on FPLC SI column, with recovery of unproteolyzed G completely separated from phosphatase catalytic subunit.

Multisite phosphorylation of the glycogen-binding subunit of protein phosphatase-1G by cyclic AMP-dependent protein kinase and glycogen synthase kinase-3

The glycogen-binding (G) subunit of protein phosphatase-1G is phosphorylated stoichiometrically by glycogen synthase kinase-3 (GSK3), and with a greater catalytic efficiency than glycogen synthase, but only after prior phosphorylation by cyclic AMP-dependent protein kinase (A-kinase) at site 1. The residues phosphorylated are the first two serines in the sequence AIFKPGFSPQPSRRGS-, while the C-terminal serine (site 1) is one of the two residues phosphorylated by A-kinase. These findings demonstrate that (i) the G subunit undergoes multisite phosphorylation in vitro; (ii) phosphorylation by GSK3 requires the presence of a C-terminal phosphoserine residue; (iii) GSK3 can synergise with protein kinases other than casein kinase-2.

The glycogen-binding subunit of protein phosphatase-1G from rabbit skeletal muscle. Further characterisation of its structure and glycogen-binding properties

The glycogen-bound form of protein phosphatase-1 (PP-1G) was previously purified as a heterodimer composed of a 37-kDa catalytic (C) subunit and a proteolytically sensitive 103-kDa glycogen-binding (G) subunit [Stråhlfors, P., Hiraga, A. & Cohen, P. (1985) Eur. J. Biochem. 149, 295-303]. In this paper we demonstrate by a variety of criteria that the intact G subunit is a 161-kDa protein, and that the 103-kDa species (now termed G') is itself a product of proteolysis. A second phosphorylation site for cAMP-dependent protein kinase (termed site 2) was identified on the G subunit. The site 2 serine was phosphorylated at a comparable rate to site 1, and near stoichiometric phosphorylation could be achieved in the presence and absence of glycogen. Site 2 was dephosphorylated by PP-1 at a slow rate, whereas site 1 was resistant to autodephosphorylation. PP-1G, as well as the proteolytic activity responsible for degradation of the G subunit, remained tightly associated with glycogen-protein particles during washing with a variety of solvents. The PP-1G holoenzyme was released from glycogen-protein particles by dilution, with a dissociation half point corresponding to about 10 nM PP-1G. Binding experiments with purified PP-1G and glycogen indicated a bimolecular process with Kapp values corresponding to about 8 nM glycogen and 4 nM PP-1G. Binding was not significantly affected by increasing ionic strength to 0.5 M or variation of pH from 6 to 8. The results are consistent with a high-affinity glycogen-binding domain on the G subunit, and indicate that a physiological concentrations of phosphatase and glycogen, PP-1G should be almost entirely bound to glycogen.

Phosphoserine as a recognition determinant for glycogen synthase kinase-3: phosphorylation of a synthetic peptide based on the G-component of protein phosphatase-1

Prior phosphorylation of its substrate has been shown to be important for substrate recognition by the protein kinase glycogen synthase kinase-3 (GSK-3). Phosphorylation of glycogen synthase by GSK-3 is known to be enhanced by the previous action of casein kinase II and the sequence -SXXXS(P)- was proposed as the minimal recognition determinant for GSK-3. The glycogen binding subunit of type 1 phosphoprotein phosphatase has been shown to be phosphorylated by cyclic AMP-dependent protein kinase at serine-13 in the sequence KPGFS(5)PQPS(9)RRGS(13)ESSEEVYV (F.B. Caudwell, A. Hiraga, and P. Cohen (1986) FEBS Lett. 194, 85-89). Inspection of the sequence revealed potential GSK-3 sites at residues 5 and 9. Using a synthetic peptide with the above sequence, we found that phosphorylation of serine-13 by cyclic AMP-dependent protein kinase permitted the recognition of serine-9 and serine-5 by GSK-3. The work provides another example of a substrate for GSK-3 and demonstrates that the action of GSK-3 is linked to the presence of phosphate in the substrate and not the action of any particular protein kinase. In the course of the analyses, a novel feature of trypsin cleavage of phosphopeptides was noted. In the sequence -SRRGS(P)- trypsin acted uniquely after the first arginine whereas in the sequence -S(P)RRGS(P)- it cleaved randomly at either arginine residue. The fact that GSK-3 could phosphorylate a peptide derived from a phosphatase subunit also raises the possibility that GSK-3 might be involved in controlling glycogen-associated type 1 phosphatase and, more generally, in mediating cyclic AMP control of protein phosphorylation in cells.

Identification of phosphorylation sites in peptides using a gas-phase sequencer

A simple procedure is described for determining the location of phosphorylation sites in phosphopeptides. The method employs measurement of 32P-labeled inorganic phosphate release during Edman degradation cycles using a gas-phase sequencer. The procedure is based on extracting peptides and inorganic phosphate from portions of the sample filter at strategic cycles in the sequence analysis followed by determination of the relative amounts of phosphate and phosphopeptide. One advantage of this technique is the very high recovery of the phosphate associated with the peptide, 80-97% in this study. In the course of this work, it was also found that phosphoserine residues themselves caused reduced efficiency of the Edman degradation as compared with unesterified serine residues. The present procedure has the merit of being simple and easy to apply.

Protein phosphorylation and hormone action

Many key regulatory proteins exist in cells as either a phosphorylated or a dephosphorylated form, their steady-state levels of phosphorylation reflecting the relative activities of the protein kinases and protein phosphatases that catalyse the interconversion process. Phosphorylation of seryl or threonyl (and occasionally tyrosyl) residues triggers small conformational changes in these proteins that alter their biological properties. Hormones and other extracellular signals transmit information to the interior of the cell by activating transmembrane signalling systems that control the production of a relatively small number of chemical mediators, termed 'second messengers'. These substances regulate the activities of protein kinases and phosphatases, and so alter the phosphorylation states of many intracellular proteins, accounting for the diversity of action of hormones. In this lecture I review recent work which demonstrates that a wide variety of cellular processes are controlled by relatively few protein kinases and protein phosphatases with pleiotropic actions. These enzymes provide the basis of an interlocking network that allows extracellular signals to coordinate biochemical functions.

Phosphorylation of the glycogen-binding subunit of protein phosphatase-1G in response to adrenalin

The glycogen-binding (G) subunit of protein phosphatase-1 is phosphorylated in vivo. In rabbits injected with propranolol the serine residue termed site-1 was phosphorylated in 56% of the molecules isolated, and phosphorylation increased to 82% after administration of adrenalin. It is concluded that the G-subunit is a physiological substrate for cyclic AMP-dependent protein kinase. The G-subunit remained largely bound to glycogen even after injection of adrenalin, whereas half of the protein phosphatase-1 activity associated with glycogen was released into the cytosol. The results indicate that adrenalin induces dissociation of the catalytic subunit from the G-subunit in vivo.

Observations on the quantitation of the phosphate content of peptides by fast-atom bombardment mass spectrometry

Equimolar mixtures of the phosphorylated and dephosphorylated forms of several peptides have been subjected to fast-atom bombardment mass spectrometry (FABMS), to investigate whether the stoichiometry of phosphorylation can be determined from the relative molecular-ion abundances of the phospho and dephospho derivatives. It is concluded that quantitation can be achieved for peptides with large positive or negative hydrophobicity/hydrophilicity indices (delta F values) where addition of a phosphate group does not alter the distribution of the peptide within the matrix significantly. For peptides with small positive or negative delta F values, phosphopeptides tend to be partially suppressed by their dephosphorylated counterparts. Suppression can be partially or totally overcome by conversion of the peptide to a hydrophobic derivative, and by the selection of an appropriate matrix. Alternatively, addition of a very strong acid, perchloric acid, can even reverse the original suppression effect. This last effect is believed to be due to the increased ionic strength in the matrix, which forces a relatively hydrophilic analyte to the matrix surface; and the ability of such a phosphorylated analyte to form a more stable gas-phase cation.

A new method for the selective isolation of phosphoserine-containing peptides

Meyer et al. [(1986) FEBS Lett 204, 61-66] have shown that phosphoserine can be converted to S-ethylcysteine by beta-elimination and addition of ethanethiol. I have utilised this modification to develop a rapid method for the selective purification of phosphoserine-containing peptides from complex mixtures. Changing phosphoserine to S-ethylcysteine increases the hydrophobicity of a peptide, altering its mobility during reverse-phase chromatography. The number of S-ethylcysteine residues in a peptide can be quantified at the picomolar level, following acid hydrolysis and conversion to the phenylthiocarbamyl derivative. The procedure may be particularly powerful for the analysis of peptides that are phosphorylated at multiple sites in vivo.

Further studies on the structure of the glycogen-bound form of protein phosphatase-1 from rabbit skeletal muscle

We have reported previously that the glycogen-bound form of protein phosphatase-1 (termed PP-1G) is a heterodimer, composed of the 37-kDa catalytic (C) subunit complexed to a 103-kDa G-subunit that anchors the enzyme to glycogen [Strålfors, P., Hiraga, A. and Cohen, P. (1985) Eur. J. Biochem. 149, 295-303]. An antibody raised against a synthetic peptide corresponding to the phosphorylation site on the G-subunit was found to immunoprecipitate PP-1G specifically. No precipitation occurred if the antibody was preincubated with the synthetic peptide, or if PP-1G was replaced by either the isolated C-subunit or protein phosphatase-2A. The results confirm by a new and independent method that the G-subunit is complexed to the C-subunit, and that it is not a contaminant. The G-subunit is remarkably sensitive to proteolysis. At the final stage of purification, PP-1G was eluted as a broad peak of activity. The leading fractions contained the 37-kDa C-subunit and 103-kDa G-subunit, while the central and trailing fractions comprised the 37-kDa C-subunit plus a number of bands with molecular masses ranging over 40-80 kDa. The 40-80-kDa bands were phosphorylated by cyclic-AMP-dependent protein kinase and tryptic digestion generated the identical phosphopeptides obtained by trypsinisation of the 103-kDa G-subunit. Furthermore, antibody to the G-subunit immunoprecipitated protein phosphatase activity quantitatively in the leading, central and trailing fractions. The results demonstrate that the 40-80-kDa polypeptides are fragments of the G-subunit, and that fragments as small as 40 kDa retain the sites of phosphorylation as well as the ability to interact with the C-subunit and with glycogen. Khatra [J. Biol. Chem. (1986) 261, 8944-8952] reported that the glycogen-bound form of protein phosphatase-1 did not contain a G-subunit and that it was a dimer composed of two identical C-subunits. The present work has shown that this proposal is incorrect, and that proteolysis of the G-subunit to fragments that stain very poorly with Coomassie blue can explain why this error was made.

Phosphorylation of the glycogen-binding subunit of protein phosphatase-1G by cyclic-AMP-dependent protein kinase promotes translocation of the phosphatase from glycogen to cytosol in rabbit skeletal muscle

The glycogen-bound form of protein phosphatase-1 (termed protein phosphatase-1G) is composed of the catalytic (C) subunit complexed to a glycogen-binding (G) subunit that anchors the enzyme to glycogen [Strålfors et al. (1985) Eur. J. Biochem. 149, 295-303]. Incubation of purified protein phosphatase-1G with cyclic-AMP-dependent protein kinase and MgATP, which leads to stoichiometric phosphorylation of the G-subunit [Caudwell et al. (1986) FEBS Lett. 194, 85-90], was found to promote the release of the phosphatase from glycogen; similar observations were made using glycogen-protein particle preparations. An intravenous injection of adrenaline decreased protein phosphatase-1 activity associated with the glycogen-protein particles by 50% with a corresponding increase in the amount present in the cytosol. By contrast, adrenaline did not affect the distribution of glycogen synthase or glycogen phosphorylase which remained entirely bound to glycogen in these experiments. The specific release of protein phosphatase-1 from glycogen may facilitate its inactivation by inhibitor-1 in the cytosol, thereby preventing dephosphorylation of the glycogen metabolising enzymes. Translocation of protein phosphatase-1 may represent a novel mechanism for the activation of glycogenolysis and inhibition of glycogen synthesis by adrenaline.