The ATP-binding site of brain phosphatidylinositol 4-kinase PI4K230 as revealed by 5'-p-fluorosulfonylbenzoyladenosine.

The ATP-binding site of purified bovine brain phosphatidylinositol 4-kinase 230 (PI4K230) was studied by its reaction with 5'-p-fluorosulfonylbenzoyladenosine (FSBA), an ATP-like alkylating reagent. Four hundred to eight hundred micromolar FSBA inactivated PI4K230 specifically with apparently first-order kinetics and resulted in 50% loss of enzyme activity in 36--130 min. The specificity of the reaction with FSBA was demonstrated by the lack of inactivation with 5'-p-fluorosulfonylbenzoyl chloride and by protection with ATP and ATP analogues against inactivation. Most ATP analogues competed with FSBA inactivation in order of their increasing hydrophobicity, parallel to their inhibitory potency in activity measurements. The specific binding of FSBA to PI4K230 was demonstrated also by Western-blot experiments. These results suggest that FSBA-reactive group(s) involved in the enzyme activity are located near to the ATP-binding site in a hydrophobic region of native PI4K230. Experiments with site-directed mutagenesis indicate that the conserved Lys-1792 plays essential role in the enzyme activity and serves as one target of affinity labelling by FSBA. Prevention of both Lys-1792-directed and Lys-1792-independent binding of FSBA by Cibacron Blue 3GA suggest that these sites are located spatially close to each other.

Human phosphatidylinositol 4-kinase isoform PI4K92. Expression of the recombinant enzyme and determination of multiple phosphorylation sites.

Human phosphatidylinositol 4-kinase, isoform PI4K92, was expressed as His6 tagged protein in Sf9 cells reaching a level of approximately 5% of cellular protein. The enzyme can be purified nearly to homogeneity in a single step by absorption/desorption on Ni/nitriloacetic acid agarose magnetic beads. High Km values in the millimolar range for ATP and PtdIns as well as only a moderate inhibition by adenosine and a sensitivity to Wortmannin (IC50 approximately 300 nM) characterize the enzyme as a type 3 PI4K. The enzyme produces PtdIns4P as product. The isolated enzyme is a phosphoprotein, additionally phosphate is incorporated by incubation with ATP/Mg or ATP/Mn. Phosphorylation sites were mapped by MALDI-MS and LC-MS/MS at the following positions: S258, T263, S266, S277, S294, T423, S496, T504. Accordingly, a stretch of 81 amino acids between the common and the C-terminal catalytic domain was designated phosphorylation domain.

Immunohistochemical localisation of two phosphatidylinositol 4-kinase isoforms, PI4K230 and PI4K92, in the central nervous system of rats.

The distribution and cellular localisation of the phosphatidylinositol 4-kinase isoforms, PI4K230 and PI4K92, that are believed to play important roles in the intracellular signalling mechanisms were studied in the rat brain (cortex, cerebellum, hippocampus and spinal cord) using immunocytochemistry with light and electron microscopy. PI4K230 was detected with a specific antibody purified by affinity chromatography from the egg yolk of chicken immunised with a 33-kDa fragment of bovine PI4K230, comprising amino acids 873-1175 of the native protein. PI4K92 was immunostained with a commercially available antibody raised in rabbit against amino acid residues 410-537 of human PI4K92. At the light microscopic level, the immunostaining of PI4K230 and PI4K92 showed a very similar distribution throughout the neurons and appeared as dense punctate labelling in the cytoplasm of perikarya and stem dendrites of various neurons. In addition to neurons, a strongly stained cell population was observed in the molecular layer of the cerebellar cortex that resembled Bergmann glia cells. Electron microscopy of neurons in the ventral horn of the spinal cord showed dense granular immunoprecipitates for both PI4K230 and PI4K92, mostly associated with the outer membrane of mitochondria and membranes of the rough endoplasmic reticulum. In addition, immunostaining of PI4K92 was also frequently found on the outer surface of cisterns and vesicles of Golgi complexes, whereas PI4K230 immunoreactivity was colocalised with some multivesicular bodies. Neither nuclear localisation nor a regular attachment to the cell membrane of these enzymes were observed. Our findings indicate that PI4K230 and PI4K92 are not involved directly in the ligand-stimulated turnover of phosphoinositides at the plasma membrane of neurons. However, they may provide regulatory phosphoinositides for intracellular vesicular traffic being associated with various organelles.

Muscle phosphorylase kinase is not a substrate of AMP-activated protein kinase.

AMP-activated protein kinase (AMPK) and cAMP-dependent protein kinase (cAMPK) have been reported to phosphorylate sites on phosphorylase kinase (PhK). Their target residues Ser 1018 and Ser 1020, respectively, are located in the so-called multi-phosphorylation domain in the PhK alpha subunit. In PhK preparations, only one of these serines is phosphorylated, but never both of them. The aim of this study was to determine whether phosphorylation by cAMPK or AMPK would influence subsequent phosphorylation by the other kinase. Surprisingly, employing four different PhK substrates, it could be demonstrated that, in contradiction to previous reports, PhK is not phosphorylated by AMPK.

Characterization of the cardiac holotroponin complex reconstituted from native cardiac troponin T and recombinant I and C.

Cardiac troponin I (cTnI), the inhibitory subunit of cardiac troponin (cTn), is phosphorylated by the cAMP-dependent protein kinase A at two adjacently located serine residues within the heart-specific N-terminal elongation. Four different phosphorylation states can be formed. To investigate each monophosphorylated form cTnI mutants, in which each of the two serine residues is replaced by an alanine, were generated. These mutants, as well as the wild-type cardiac troponin I (cTnI-WT) have been expressed in Escherichia coli, purified and characterized by isoelectric focusing, MS and CD-spectroscopy. Monophosphorylation induces conformational changes within cTnI that are different from those induced by bisphosphorylation. Functionality was assessed by measuring the calcium dependence of myosin S1 binding to thin filaments containing reconstituted native, wild-type and mutant cTn complexes. In all cases a functional holotroponin complex was obtained. Upon bisphosphorylation of cTnI-WT the pCa curve was shifted to the right to the same extent as that observed with bisphosphosphorylated native cTnI. However, the absolute values for the midpoints were higher when recombinant cTn subunits were used for reconstitution. Reconstitution itself changed the calcium affinity of cTnC: pCa50-values were higher than those obtained with the native cardiac holotroponin complex. Apparently only bisphosphorylation of cTnI influences the calcium sensitivity of the thin filament, thus monophosphorylation has a function different from that of bisphosphorylation; this function has not yet been identified.

Functional expression and characterisation of a new human phosphatidylinositol 4-kinase PI4K230.

By constructing DNA probes we have identified and cloned a human PtdIns 4-kinase, PI4K230, corresponding to a mRNA of 7.0 kb. The cDNA encodes a protein of 2044 amino acids. The C-terminal part of ca. 260 amino acids represents the catalytic domain which is highly conserved in all recently cloned PtdIns 4-kinases. N-terminal motifs indicate multiple heterologous protein interactions. Human PtdIns 4-kinase PI4K230 expressed in vitro exhibits a specific activity of 58 micromol mg-1min-1. The enzyme expressed in Sf9 cells is essentially not inhibited by adenosine, it shows a high Km for ATP of about 300 microM and it is half-maximally inactivated by approximately 200 nM wortmannin. These data classify this enzyme as type 3 PtdIns 4-kinase. Antibodies raised against the N-terminal part moderately activate and those raised against the C-terminal catalytic domain inhibit the enzymatic activity. The coexistence of two different type 3 PtdIns 4-kinases, PI4K92 and PI4K230, in several human tissues, including brain, suggests that these enzymes are involved in distinct basic cellular functions.

Stepwise subunit interaction changes by mono- and bisphosphorylation of cardiac troponin I.

Four phosphorylation degrees of cardiac troponin I (cTnI) have been characterized, namely, a dephospho, a bisphospho, and two monophospho states. Here we describe for the first time a role of the monophosphorylated forms. We have investigated the interaction between the cardiac troponin subunits dependent on the phosphorylation state of cTnI by surface plasmon resonance (SPR) spectroscopy. The monophosphorylated forms were generated by mutating each of the two serine residues, located in human cTnI at positions 22 and 23, to alanine. Association and dissociation rate constants of binary (cTnI-cTnT and cTnI-cTnC) and ternary (cTnI/cTnC complex-cTnT) complexes were determined. Mono- and consecutive bisphosphorylation of cTnI gradually reduces the affinity to cTnC and cTnT by lowering the association rate constants; the dissociation rate constants remain unchanged. Phosphorylation also affects formation of the ternary complexes; however, in this instance, association rate constants are constant, and dissociation rate constants are enhanced. A model of cardiac troponin is presented describing an induction of distinct conformational changes by mono- and bisphosphorylation of cTnI.

Phosphatidylinositol 4-kinases.

Polyphosphoinositides are involved in many signal transduction pathways in eukaryotic cells. The first committed step is catalysed by phosphatidylinositol 4-kinase leading to the formation of phosphatidylinositol 4-phosphate. In the last four years, ten cDNA molecules have been cloned which code isoforms of phosphatidylinositol 4-kinase; some of which are highly related. Characteristically, they contain a C-terminal catalytic domain which is similar to that of (poly)phosphoinositide 3-kinases and to that of more distantly related lipid/protein kinases. Alignment has characterised cDNAs from Chaenorabditis, Dictyostelium and Schizostaphyloccus pombe as those of phosphatidylinositol 4-kinases also. All these lipid kinases are related to the superfamily of protein kinases. Several amino acids are highly conserved in catalytic domains of lipid and protein kinases. Employing the catalytic subunit of the cAMP-dependent protein kinase as template, these residues can be assigned functionally. On the basis of the alignment, a phylogenetic tree of the superfamily of phosphatidylinositol kinases has been constructed. Three families, the phosphatidylinositol 4-kinases, phosphoinositide 3-kinases, and the phosphatidylinositol related lipid/protein kinases, can be recognised. Each family comprises two subfamilies. The involvement of the phosphatidylinositol 4-kinases in signal transduction processes is summarised and a new hypothesis for the function of their isoforms in polyphosphoinositide signalling is presented. The involvement of phosphatidylinositol 4-kinases in formation of lipid-protein interactions with cytoskeleton proteins and the metabolism of polyphosphoinositide in the nucleus is discussed.

Microanalysis and distribution of cardiac troponin I phospho species in heart areas.

Sequential phosphorylation and dephosphorylation of cTnI by the cAMP dependent protein kinase and by protein phosphatase 2A, respectively, produce the non-, mono- and bisphosphorylated species (Jaquet et al., 1995, Eur. J. Biochem. 231, 486-490). The aim of this study was to determine these forms even in small tissue samples, e.g. in biopsy probes of approximately 30 mg which would allow to define the phosphorylation state of cTnI in heart areas. In order to do so a micro isolation procedure for cTnI had to be established. cTnI is extracted from small bovine, rabbit and human heart tissue samples (30-100 mg) under special conditions avoiding dephosphorylation and is isolated by affinity chromatography on cTnC Sepharose. All three species, the bis-, mono- and dephospho cTnI, are precipitated quantitatively by acetone, then they are separated by non-equilibrium isoelectric focusing and quantified by scanning densitometry. The method presented here allows to quantify the three cTnI species reproducibly. No other phosphorylated species are detected. Truncated cTnI forms of each phospho species are found in human biopsy samples due to removal of a approximately 36 amino acid peptide from the C-terminus. In bovine, human and rabbit heart the pattern of the three cTnI phospho species is characteristic for left and right atrium, left and right ventricle and septum.

The identification of the phosphorylated 150/160-kDa proteins of sarcoplasmic reticulum, their kinase and their association with the ryanodine receptor.

In the present work we studied the relationship between the phosphorylated 150- and 160-kDa proteins and other SR proteins in the 150,000-170,000 range of molecular masses. on SDS-PAGE, the identification of their kinase, as well as the purification and structural interactions between these proteins and the rynodine receptor (RyR). The phosphorylated 150-kDa protein was identified as sarcalumenin based on: (a) its cross-reactivity with three different monoclonal antibodies specific for sarcalumenin. (b) its mobility in SDS-PAGE which was modified upon digestion with endoglycosidase H, (c) its elution from lentil-lectin column by alpha-methyl mannoside, (d) its resistance to trypsin, (e) its ability to bind Ca2+ and to stain blue with Stains-All. The phosphorylated 160-kDa protein was identified as the histidine-rich Ca2+ binding protein (HCP) based on: (a) its Ca(2+)-binding property and staining blue with Stains-All, (b) phosphorylation with the catalytic subunit of cAMP-dependent kinase. (c) its increased mobility in SDS-PAGE in the presence of Ca2+ (d) its heat stability and (e) its partial amino acid sequence. The endogenous kinase was identified as casein kinase II (CK II) based on the inhibition of the endogenous phosphorylation 160/150-kDa proteins by heparin, 5.6-dichlorobenzimidazole riboside, polyaspartyl peptide and hemin, and its ability to use [gamma-32P]GTP as the phosphate donor. The association of CK II with SR membranes, was demonstrated using specific polyclonal anti-CK II antibodies. The luminal location of CK II is suggested because CK II was extracted from the SR by l M NaCl only after their treatment with hypotonic medium, and CK II activity was inhibited with the charged inhibitors heparin and polyaspartyl peptide only after their incubation with the SR in the presence of NP-40. The 160- and 150-kDa proteins were purified on spermine-agarose column, and were phosphorylated by CK II. Like the endogenous phosphorylation of the 150/160-kDa proteins in SR. the phosphorylation of the purified proteins by CK II was inhibited by La3+ (Cl50 = 4 microM) and hemin. The results suggest the phosphorylation of the luminally located sarcalumenin and HCP with CK II.

VDAC/porin is present in sarcoplasmic reticulum from skeletal muscle.

In this study we demonstrate the existence of a protein with properties of the voltage-dependent anion channel (VDAC) in the sarcoplasmic reticulum (SR) using multiple approaches as summarized in the following: (a) 35 and 30 kDa proteins in different SR preparations, purified from other membranal systems by Ca2+/oxalate loading and sedimentation through 55% sucrose, cross-react with four different VDAC monoclonal antibodies. (b) Amino acid sequences of three peptides derived from the SR 35 kDa protein are identical to the sequences present in VDAC1 isoform. (c) Similar to the mitochondrial VDAC, the SR protein is specifically labeled by [14C]DCCD. (d) Using a new method, a 35 kDa protein has been purified from SR and mitochondria with a higher yield for the SR. (e) Upon reconstitution into a planar lipid bilayer, the purified SR protein shows voltage-dependent channel activity with properties similar to those of the purified mitochondrial VDAC or VDAC1/porin 31HL from human B lymphocytes, and its channel activity is completely inhibited by the anion transport inhibitor DIDS and about 80% by DCCD. We also demonstrate the translocation of ATP into the SR lumen and the phosphorylation of the luminal protein sarcalumenin by this ATP. Both ATP translocation and sarcalumenin phosphorylation are inhibited by DIDS, but not by atractyloside, a blocker of the ATP/ADP exchanger. These results indicate the existence of VDAC, thought to be located exclusively in mitochondria, in the SR of skeletal muscle, and its possible involvement in ATP transport. Together with recent studies on VDAC multicompartment location and its dynamic association with enzymes and channels, our findings suggest that VDAC deserves attention and consideration as a protein contributing to various cellular functions.

Bisphosphorylation of cardiac troponin I modulates the Ca(2+)-dependent binding of myosin subfragment S1 to reconstituted thin filaments.

We have reconstituted thin filaments comprising pyrene-labelled actin (pyr-actin), tropomyosin (Tm) and cardiac troponin (cTn). cTn was isolated in two defined phosphorylation states; completely dephosphorylated on all subunits and with only the cTnI subunit bisphosphorylated. The thin filament was saturated with cTn at a pyr-actin/Tm/cTn ratio of 7:1:1. The calcium-dependent binding of S1 to thin filaments was measured in a stopped-flow spectrophotometer and the dependence of the observed rate constant on [Ca2+] fitted to the Hill equation. The only significant difference between the two phosphorylation states of the filaments was a 0.36 decrease in the pCa50 on bisphosphorylation.

Identification of a 200 kDa polypeptide as type 3 phosphatidylinositol 4-kinase from bovine brain by partial protein and cDNA sequencing.

Two phosphatidylinositol 4-kinase isozymes, type 3 and type 2, have been separated on hydroxylapatite after solubilizing bovine brain microsomes with Triton X-114. Employing a newly developed renaturation procedure following SDS-PAGE, we demonstrate that a 200 kDa polypeptide carries the enzyme activity of this type 3 isoform. Chromatography on hydroxylapatite, Heparin-Sepharose, Superdex 200 and finally SDS-PAGE results in an approximately 30,000-fold purification. Tryptic peptides generated from the 200 kDa polypeptide after SDS-PAGE have been sequenced and the obtained data have been used for constructing and synthesizing degenerated oligonucleotides. Polymerase chain reaction as well as screening of cDNA libraries allowed several clones to be isolated from which a 4.7 kb contiguous sequence can be built up. The open reading frame covers 4.4 kb with a 0.3 kb untranslated 3' end which yields a deduced amino acid sequence of 1,467 amino acids. The C-terminal part of ca. 300 amino acids represents the catalytic domain. Sequence alignment of this domain with the mammalian counterpart, the human type 2 phosphatidylinositol 4-kinase, the yeast kinases STT4 and PIK1, as well as with the catalytic domains of bovine, human, mouse and yeast phosphatidylinositol 3-kinases reveals a high degree of identity: 26 of these approximately 300 amino acids are invariable in all of these eight catalytic domains. Five motifs indicate nuclear localization and DNA binding properties of the enzyme. Two leucine zipper motifs (amino acids 358-386, 862-882) are detectable. Furthermore, a helix loop helix motif (amino acids 716-729) as well as two nuclear localization signals (amino acids 838-854, 345-349) indicate the presence of the type 3 isoform in the nucleus.

Cloning of a phospholipase C-delta 1 of rabbit skeletal muscle.

The phospholipase C isoform responsible for the increase in the total myoplasmic inositol 1,4,5-trisphosphate concentration during tetanic contraction of isolated skeletal muscle and its mechanism of activation is not known. We have cloned and sequenced a phospholipase C cDNA of rabbit skeletal muscle coding for a protein of 745 amino acids with a molecular mass of 84,440 kDa. The deduced amino acid sequence exhibits the phospholipase C-specific domains X and Y which according to current knowledge very likely represent the catalytic centre of the enzyme. An overall sequence homology of 88% to the phospholipase C-delta 1 of rat brain suggests that the encoded protein represents a phospholipase C-delta 1 isoform of rabbit skeletal muscle. Northern blot analysis shows, that this phospholipase C-delta is dominantly expressed in skeletal muscle, less strongly in smooth muscle (uterus) and lung and weakly in heart, kidney and brain. In the N-terminal part of the primary structure a consensus sequence for a canonical EF-hand Ca2+ binding domain can be identified together with a short positively charged motif which recently has been suggested to be essential for the binding of phosphatidylinositol 4,5-bisphosphate. If these two domains which are unique for phospholipase C-delta are sufficient in establishing a mechanism for the activation of the enzyme, inositol 1,4,5-trisphosphate formation in skeletal muscle could be the consequence of an increase in myoplasmic Ca2+.

Pattern formation on cardiac troponin I by consecutive phosphorylation and dephosphorylation.

Two serine residues located adjacently in the heart-specific N-terminus of cardiac troponin I can be phosphorylated in vivo. Both residues are sequentially phosphorylated and dephosphorylated by cAMP-dependent protein kinase (PKA) and protein phosphatase 2A (PP2A). The concentration changes of the different troponin I species have been determined separately for the phosphorylation and dephosphorylation reaction and approximated by time courses predicted by a reaction model. Dependent on the concentration ratio of active protein kinase/protein phosphatase, four different troponin I species can be generated; one nonphosphorylated, two monophosphorylated and one bisphosphorylated. This pattern generation will be observed in proteins phosphorylated and dephosphorylated by a single protein kinase and phosphatase on more than one site and is a new principle inherent in signal cascades.

Does phosphorylase kinase control glycogen biosynthesis in skeletal muscle?

Immunoblotting as well as enzyme assays demonstrate the presence of the self-glucosylating protein, glycogenin, in the protein-glycogen complex, in the sarcoplasmic reticulum and in phosphorylase kinase. In all three compartments glycogenin occurs in different, albeit, defined glucosylated forms, which upon deglucosylation are converted into a 42 kDa form. We suggest that phosphorylase kinase might have a dual function in glycogen biogenesis: firstly, control of glycogen degradation in the protein-glycogen complex via phosphorylation of glycogen phosphorylase b; secondly, regulation of glycogen biosynthesis on the sarcoplasmic reticular membranes via phosphorylation and thereby inhibition of glycogen synthase.

Endogenous, Ca(2+)-dependent cysteine-protease cleaves specifically the ryanodine receptor/Ca2+ release channel in skeletal muscle.

The association of an endogenous, Ca(2+)-dependent cysteine-protease with the junctional sarcoplasmic reticulum (SR) is demonstrated. The activity of this protease is strongly stimulated by dithiothreitol (DTT), cysteine and beta-mercaptoethanol, and is inhibited by iodoacetamide, mercuric chloride and leupeptin, but not by PMSF. The activity of this thiol-protease is dependent on Ca2+ with half-maximal activity obtained at 0.1 microM and maximal activity at 10 microM. Mg2+ is also an activator of this enzyme (CI50 = 22 microM). These observations, together with the neutral pH optima and inhibition by the calpain I inhibitor, suggest that this enzyme is of calpain I type. This protease specifically cleaves the ryanodine receptor monomer (510 kD) at one site to produce two fragments with apparent molecular masses of 375 and 150 kD. The proteolytic fragments remain associated as shown by purification of the cleaved ryanodine receptor. The calpain binding site is identified as a PEST (proline, glutamic acid, serine, threonine-rich) region in the amino acid sequence GTPGGTPQPGVE, at positions 1356-1367 of the RyR and the cleavage site, the calmodulin binding site, at residues 1383-1400. The RyR cleavage by the Ca(2+)-dependent thiol-protease is prevented in the presence of ATP (1-5 mM) and by high NaCl concentrations. This cleavage of the RyR has no effect on ryanodine binding activity but stimulates Ca2+ efflux. A possible involvement of this specific cleavage of the RyR/Ca2+ release channel in the control of calpain activity is discussed.

Phosphoinositides in membranes that build up the triads of rabbit skeletal muscle.

The total membrane concentrations of PtdIns, PtdIns4P, and PtdIns(4,5)P2 contribute to the functional capacity of the Ins(1,4,5)P3 signalling system which is operating in skeletal muscle but the function of which is still unknown. Total amounts of these phosphoinositides have been determined in purified membranes of transverse tubules (TT) and terminal cisternae (TC) of the sarcoplasmic reticulum (SR) of rabbit skeletal muscle. PtdIns and PtdIns4P have been detected in both membrane systems whereas PtdIns(4,5)P2 (290 mumol/mol phospholipid) is confined only to TT. A much greater pool of PtdIns(4,5)P2 seems, however, to be located in the sarcolemma away from the triadic junction.

Characterization of the cardiac troponin I phosphorylation domain by 31P nuclear magnetic resonance spectroscopy.

Cardiac holotroponin can be phosphorylated at serine 23 and/or 24 in the heart-specific region of bovine troponin I. When isolated freshly it is composed of a mixture of non-, two mono-, and bisphosphorylated species. At neutral pH the monophosphorylated form carrying phosphate at serine 24 yields a resonance signal at 4.6 ppm and that carrying phosphate at serine 23 at 4.4 ppm; the two phosphate groups of the bisphosphorylated form yield only one 31P-NMR signal at 4.2 ppm. From the chemical shift dependence on pH, pKa values have been estimated to be 5.3 and 5.6 for the phosphate groups at serine 24 and serine 23, respectively. Both phosphates of the bisphosphorylated form exhibit very similar pKa values of approximately 5.8. Separation of bisphosphotroponin I from the complex results in a downfield shift and the appearance of two 31P-NMR signals at positions comparable to those of the two monophospho forms. Complex formation of cardiac troponin I with C or T does not alter the spectrum obtained with isolated troponin I; however, the original troponin spectrum is restored by reconstitution of the holocomplex from all three components T, I, and C. Two signals are also observed with a bisphosphorylated synthetic peptide [PVRRRS(P)S(P)ANYR] representing the phosphorylation domain. pKa values of about 5.3 and 5.6 have been determined for serine 7 (corresponding to serine 24 of troponin I) and serine 6 of the peptide (corresponding to serine 23 of troponin I).

Interaction sites on phosphorylase kinase for calmodulin.

Holophosphorylase kinase was digested with Glu-C specific protease; from the peptide mixture calmodulin binding peptides were isolated by affinity chromatography and identified by N-terminal sequence analysis. Two peptides originating from the alpha subunit, having a high tendency to form a positively charged amphiphilic helix and containing tryptophane, were synthesized. Additionally, a homologous region of the beta subunit and a peptide from the alpha subunit present in a region deleted in the alpha' isoform were also selected for synthesis. Binding stoichiometry and affinity were determined by following the enhancement in tryptophane fluorescence occurring upon 1:1 complex formation between these peptides and calmodulin. Finally, Ca2+ binding to calmodulin in presence of peptides was measured. By this way, the peptides alpha 542-566, alpha 547-571, alpha 660-677 and beta 597-614 have been found to bind specifically to calmodulin. Together with previously predicted and synthesized calmodulin binding peptides four calmodulin binding regions have been characterized on each the alpha and beta subunits. It can be concluded that endogenous calmodulin can bind to two calmodulin binding regions in gamma as well as to two regions in alpha and beta. Exogenous calmodulin can bind to two regions in alpha and in beta. A binding stoichiometry of 0.8 mol of calmodulin/alpha beta gamma delta promoter of phosphorylase kinase has been determined by inhibiting the ubiquitination of calmodulin with phosphorylase kinase. Phosphorylase kinase is half maximally activated by 23 nM calmodulin which is in the affinity range of calmodulin binding peptides from beta to calmodulin. Therefore, binding of exogenous calmodulin to beta activates the enzyme. A model for switching endogenous calmodulin between alpha, beta and gamma and modulation of ATP binding to alpha as well as Mg2+/ADP binding to beta by calmodulin is presented.