Calmodulin binding domains: characterization of a phosphorylation and calmodulin binding site from myosin light chain kinase

Autoregulation of enzymes by pseudosubstrate prototopes: myosin light chain kinase

The myosin light chain kinase requires calmodulin for activation. Tryptic cleavage of the enzyme generates an inactive 64-kilodalton (kD) fragment that can be further cleaved to form a constitutively active, calmodulin-independent, 61-kD fragment. Microsequencing and amino acid analysis of purified peptides after proteolysis of the 61- and 64-kD fragments were used to determine the amino-terminal and carboxyl-terminal sequences of the 64-kD fragment. Cleavage within the calmodulin-binding region at Arg505 generates the catalytically inactive 64-kD fragment, which is incapable of binding calmodulin. Further digestion removes a carboxyl-terminal fragment, including the pseudosubstrate sequence Ser484-Lys-Asp-Arg-Met-Lys-Lys-Tyr-Met- Ala-Arg-Arg-Lys-Trp-Gln-Lys-Thr-Gly-His-Ala-Val-Arg505 and results in a calmodulin-independent 61-kD fragment. Both the 61- and 64-kD fragments have the same primary amino-terminal sequences. These results provide direct support for the concept that the pseudosubstrate structure binds the active site and that the role of calmodulin is to modulate this interaction. Pseudosubstrates may be utilized in analogous ways by other allosterically regulated enzymes.

p36, the major cytoplasmic substrate of src tyrosine protein kinase, binds to its p11 regulatory subunit via a short amino-terminal amphiphatic helix

Protein I is a hetero-tetramer which contains two copies each of p11 and p36. p36 (calpactin I, lipocortin II) is a major substrate of retrovirally encoded tyrosine protein kinases, while p11 modulates several Ca2+-induced properties also displayed by p36 alone. Here we have characterized the p11 binding site on p36 by fluorescence spectroscopy using porcine p36 labelled at cysteine 8 with the fluorophore Prodan (6-proprionyl-2-dimethylamino-naphthalene). We have used peptides of differing length from the amino-terminal domain of p36 to restrict the major binding site to the first 12 residues. Noticeable binding is still observed with a peptide containing only the first nine residues. Interestingly the N-terminal acetyl group of p36 forms a functional part of the p11 binding site. CD studies indicate that the binding region can form an alpha-helix, which seems to have amphiphatic properties when projected on a helical wheel. This structural element is also known for a calmodulin binding protein. Thus the question is raised whether other p11/calmodulin-related proteins interact with their target proteins via a similar mechanism. We also discuss how p11 could modulate p36 associated properties.

Mapping of calmodulin-binding domain of Ca2+/calmodulin-dependent protein kinase II from rat brain

Recent molecular cloning experiments have identified a 25 amino-acid region as the calmodulin-binding domain of the alpha-subunit of rat brain Ca2+/calmodulin-dependent multifunctional protein kinase II (CaM-K II). Synthetic peptides, derived from the deduced amino-acid sequence encompassing this region, were examined for their ability to bind calmodulin in a calcium dependent manner and to inhibit the Ca2+/calmodulin-dependent autophosphorylation of CaM-K II. Comparison of these structure-function relationships highlighted a region of 5 amino-acids, which was essential for calmodulin interaction and inhibition of kinase activity. This region demonstrated some homology with other calmodulin-binding peptides, and may represent a key site of interaction of the kinase with calmodulin. These analyses provide additional insight into the molecular mechanism underlying the Ca2+ regulation of CaM-K II.

Association of melittin with the isolated myosin light chains

Melittin is a 26-residue peptide which undergoes high-affinity calcium-dependent binding by calmodulin [Barnette, M.S., Daly, R., & Weiss, B. (1983) Biochem. Pharmacol. 32, 2929; Comte, M., Maulet, Y., & Cox, J.A. (1983) Biochem. J. 209, 269; Anderson, S.R., & Malencik, D.A. (1986) Calcium Cell Funct. 6, 1]. The results in this paper show that three different types of myosin light chain--the smooth muscle regulatory light chain, the smooth muscle essential light chain, and the skeletal muscle regulatory 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB) light chain--also associate with melittin. The resulting complexes have dissociation constants ranging from 1.1 to 2.5 microM in the presence of 0.10 M NaCl and from approximately 50 to approximately 130 nM in solutions of 20 mM 3-(N-morpholino)propanesulfonic acid alone. The regulatory smooth muscle myosin light chain exhibits two equivalent melittin binding sites while each of the others displays only one. The myosin light chains evidently contain elements of structure related to the macromolecular interaction sites present in calmodulin and troponin C but not in parvalbumin. The association of melittin and other peptides with the light chains requires consideration whenever assays of the calmodulin-dependent activity of myosin light chain kinase are used to determine peptide binding by calmodulin. The binding measurements performed on the DTNB light chain and melittin necessitated derivation of the equation relating complex formation to the observed fluorescence anisotropy of a solution containing three fluorescent components. This analysis is generally applicable to equilibria involving the association of two fluorescent molecules emitting in the same wavelength range.

Interaction of alpha-N-Acetyl-beta-endorphin and calmodulin

Acetylation at the alpha-amino terminal is a common post-translational modification of many peptides and proteins. In the case of the potent opiate peptide beta-endorphin, alpha-N-acetylation is a known physiological modification that abolishes opiate activity. Since there are no known receptors for alpha-N-acetyl-beta-endorphin, we have studied the association of this peptide with calmodulin, a calcium-dependent protein that binds a variety of peptides, phenothiazines, and enzymes, as a model system for studying acetylated endorphin-protein interactions. Association of the acetylated peptide with calmodulin was demonstrated by cross-linking with bis(sulfosuccinimidyl)suberate; like beta-endorphin, adducts containing 1 mol and 2 mol of acetylated peptide per mole calmodulin were formed. Some of the bound peptides are evidently in relatively close proximity to each other since, in the presence of amidated (i.e., lysine-blocked) calmodulin, cross-linking yielded peptide dimers. The acetylated peptide exhibited no appreciable helicity in aqueous solution, but in trifluoroethanol (TFE) considerable helicity was formed. Also, a mixture of acetylated peptide and calmodulin was characterized by a circular dichroic spectrum indicative of induced helicity. Empirical prediction rules, applied earlier to beta-endorphin, suggest that residues 14-24 exhibit alpha-helix potential. This segment has the potential of forming an amphipathic helix; this structural unit is believed to be important in calmodulin binding. The acetylated peptide was capable of inhibiting the calmodulin-mediated stimulation of cyclic nucleotide phosphodiesterase (EC 3.1.4.17) activity with an effective dose for 50% inhibition of about 3 microM; this inhibitory effect was demonstrated using both an enzyme-enriched preparation as well as highly purified enzyme. Thus, acetylation at the alpha-amino terminal of beta-endorphin, although abolishing opiate activity, does not interfere with the binding to calmodulin. Indeed, beta-endorphin and the alpha-N-acetylated peptide behave very similarly with respect to calmodulin association.

Stimulation of the erythrocyte Ca2+-ATPase and of bovine brain cyclic nucleotide phosphodiesterase by chemically modified calmodulin

Chemically modified calmodulins have been used to investigate structural features which are important for the interaction of the activator with targets. Carbamoylation of lysine residues had no influence on the ability of calmodulin to stimulate the plasma membrane Ca2+-ATPase whereas the stimulation of the bovine brain cyclic-nucleotide phosphodiesterase was reduced up to 50%. Different species of carbamoylated calmodulin have been isolated but no differences were detected in their interaction with the cyclic-nucleotide phosphodiesterase. Modification of arginine residues by 1,2-cyclohexanedione had no effect of the stimulation of the phosphodiesterase but reduced by 40% the stimulation of the erythrocyte Ca2+ ATPase. Mild oxidation of methionines by N-chlorosuccinimide produced a number of differently modified calmodulins. The different species have been purified and the modified residues have been identified. They affected the two different test enzymes to different extents indicating that methionines in the central helix of calmodulin are of greater importance for the interaction with the phosphodiesterase, whereas methionines located in the C-terminal half of calmodulin are more important for the interaction with the Ca2+-ATPase.

Amino acid sequence analysis of the neuronal type II calmodulin-dependent protein kinase by tandem mass spectrometry

The primary structure of the neuronal Type II calmodulin-dependent protein kinase has been examined by protein sequence analysis and compared to cDNA-derived sequence. Tandem mass spectroscopic analysis was used for the sequence determination. Comparison with published cDNA sequence data for the alpha subunit revealed that the difference between the alpha- and beta-subunits lay in two insertions into the sequence for the alpha-subunit and a short alpha-specific sequence. The N-terminal amino acid of the alpha subunit which is blocked to Edman degradation has been tentatively identified as N-acetyl-alanine.

Purification, characterization and substrate specificity of calmodulin-dependent myosin light-chain kinase from bovine brain

A substrate-specific calmodulin-dependent myosin light-chain kinase (MLCK) was purified 45,000-fold to near homogeneity from bovine brain in 12% yield. Bovine brain MLCK phosphorylates a serine residue in the isolated turkey gizzard myosin light chain (MLC), with a specific activity of 1.8 mumol/min per mg of enzyme. The regulatory MLC present in intact gizzard myosin is also phosphorylated by the enzyme. The Mr-19,000 rabbit skeletal-muscle MLC is a substrate; however, the rate of its phosphorylation is at best 30% of that obtained with turkey gizzard MLC. Phosphorylation of all other protein substrates tested is less than 1% of that observed with gizzard MLC as substrate. SDS/polyacrylamide-gel electrophoresis of purified MLCK reveals the presence of a major protein band with an apparent Mr of 152000, which is capable of binding 125I-calmodulin in a Ca2+-dependent manner. Phosphorylation of MLCK by the catalytic subunit of cyclic-AMP-dependent protein kinase results in the incorporation of phosphate into the Mr-152,000 protein band and a marked decrease in the affinity of MLCK for calmodulin. The presence of Ca2+ and calmodulin inhibits the phosphorylation of the enzyme. Bovine brain MLCK appears similar to MLCKs isolated from platelets and various forms of muscle.

Properties of a monoclonal antibody directed to the calmodulin-binding domain of rabbit skeletal muscle myosin light chain kinase

A synthetic peptide representing the calmodulin-binding domain of rabbit skeletal muscle myosin light chain kinase (K-R-R-W-K-K-N-F-I-A-V-S-A-A-N-R-F-K-K-I-S-S-S-G-A-L) was used as an antigen to produce a monoclonal antibody. The antibody (designated MAb RSkCBP1, of the IgM class) reacted with similar affinity (KD approximately 20 nM) by competitive enzyme-linked immunoassay (ELISA) with the antigen peptide and intact rabbit skeletal muscle myosin light chain kinase. MAb RSkCBP1 inhibited rabbit skeletal muscle myosin light chain kinase activity competitively with respect to calmodulin (Ki = 20 nM). The antibody also inhibited myosin light chain kinase activity in extracts of skeletal muscle from several mammalian species (rabbit, sheep, and bovine) and an avian species (chicken). The concentration of MAb RSKCBP1 required for 50% inhibition of enzyme activity was similar for the mammalian species (80 nM) but was significantly higher for the avian species (1.2 microM). A competitive ELISA protocol was used to analyze weak cross-reactivity to other calmodulin-binding peptides and proteins. This assay demonstrated no cross-reactivity with the venom peptides melittin or mastoparan; smooth muscle myosin light chain kinases from hog carotid, bovine trachea, or chicken gizzard; bovine brain calmodulin-dependent calcineurin; or rabbit skeletal muscle troponin I. These data support the contention that the synthetic peptide used as the antigen represents the calmodulin-binding domain of rabbit skeletal muscle myosin light chain kinase and that the calmodulin-binding domains of different calmodulin-regulated proteins may have distinct primary and/or higher order structures.

High-affinity formation of a 2:1 complex between gramicidin S and calmodulin

Two molecules of gramicidin S, a very rigid cyclic decapeptide rich in beta-sheet structure, can bind in a Ca2+-dependent way to a calmodulin molecule in the presence as well as in the absence of 4 M-urea. The flow-microcalorimetric titration of 25 microM-calmodulin with gramicidin S at 25 degrees C is endothermic for 21.3 kJ.mol-1; the enthalpy change is strictly linear up to a ratio of 2, indicating that the affinity constant for binding of the second gramicidin S is at least 10(7) M-1. In 4 M-urea the peptide quantitatively displaces seminalplasmin from calmodulin, as monitored by tryptophan fluorescence. An iterative data treatment of these competition experiments revealed strong positive co-operativity with K1 less than 5 X 10(5) M-1 and K1.K2 = 2.8 X 10(12) M-2. A competition assay with the use of immobilized melittin enabled us to monitor separately the binding of the second gramicidin S molecule: the K2 value is 1.9 X 10(7) M-1. By complementarity, the K1 value is 1.5 X 10(5) M-1. In the absence of urea the seminalplasmin displacement is incomplete: the data analysis shows optimal fitting with K1 less than 2 X 10(4) M-1 and K1.K2 = 3.2 X 10(11) M-2 and reveals that the mixed complex (calmodulin-seminalplasmin-gramicidin S) is quite stable and is even not fully displaced from calmodulin at high concentrations of gramicidin S. The activation of bovine brain phosphodiesterase by calmodulin is not impaired up to 0.2 microM-gramicidin S. According to our model the ternary complex enzyme-calmodulin-gramicidin is relatively important and displays the same activity as the binary complex enzyme-calmodulin. Gramicidin S also displaces melittin from calmodulin synergistically, as monitored by c.d. Our studies with gramicidin S reveal the importance of multipoint attachments in interactions involving calmodulin and confirm the heterotropic co-operativity in the binding of calmodulin antagonists first demonstrated by Johnson [(1983) Biochem. Biophys. Res. Commun. 112, 787-793].

Isolation and sequence analysis of a cDNA clone encoding the entire catalytic subunit of phosphorylase kinase

Synthetic oligonucleotides have been used to isolate a 1.85 kb clone containing the full length coding sequence for the catalytic subunit of rabbit skeletal muscle phosphorylase kinase from a cDNA library constructed in lambda gt10. Sequence analysis of the clone predicted an amino acid sequence in agreement with a published primary structure. Inspection of the codon usage revealed a strong preference for G or C nucleotides at the third codon position as found for several other skeletal muscle proteins. This cDNA clone should facilitate identification of functional domains, including the calmodulin-binding site, and investigation of the molecular basis of X-linked phosphorylase kinase deficiencies.

Molecular cloning of a brain-specific calcium/calmodulin-dependent protein kinase

A calcium/calmodulin-dependent protein kinase type II (CaM-K) alpha-subunit cDNA has been cloned from rat brain. This enzyme is encoded by a 5.1-kilobase mRNA expressed exclusively in the brain. Hybridization histochemistry reveals that the CaM-K mRNA expression corresponds to the distribution of the immunoreactive alpha-subunit protein, suggesting that the high enzyme levels in specific brain areas reflect regional differences in gene expression. The sequence of CaM-K alpha-subunit cDNA indicates a 478-amino acid (54-kDa) protein with three functional domains. The domain organization suggests a structural model for calcium/calmodulin-dependent and independent states that might subserve short- and long-term responses to transient stimuli.

Functional analysis of a complementary DNA for the 50-kilodalton subunit of calmodulin kinase II

The calcium-calmodulin-dependent protein kinase II is a major component of brain synaptic junctions and has been proposed to play a variety of important roles in brain function. A complementary DNA representing a portion of the smaller 50-kilodalton subunit of the rat brain enzyme has been cloned and sequenced. The calmodulin-binding region has been identified and a synthetic analog prepared that binds calmodulin with high affinity in the presence of calcium. Like the 50-kilodalton kinase polypeptide, the concentration of the messenger RNA varies both neuroanatomically and during postnatal development of the brain. The broad tissue and species cross-reactivity of the complementary DNA suggests that the 50-kilodalton subunit found in rat brain is evolutionarily conserved and is the product of a single gene.

Fluorescence properties of calmodulin-binding peptides reflect alpha-helical periodicity

A basic amphiphilic alpha-helix is a structural feature common to many calmodulin-binding peptides and proteins. A set of fluorescent analogues of a very tight binding inhibitor (dissociation constant of 200 picomolar) of calmodulin has been synthesized. The fluorescent amino acid tryptophan has been systematically moved throughout the sequence of this peptide. The fluorescence properties for the peptides repeat every three to four residues and are consistent with the periodicity observed for an alpha-helix.

Screening an expression library with a ligand probe: isolation and sequence of a cDNA corresponding to a brain calmodulin-binding protein

The use of cloning vectors that express inserted cDNA as fusion protein has led to the isolation of genes encoding a variety of eukaryotic proteins. In these instances antisera or monoclonal antibodies were used as probes to screen expression libraries. Since fusion proteins sometimes display biological activity reflective of the insert-specified portion, we tested the possibility that ligand-binding sites might exist in fusion proteins. Specifically we used 125I-labeled calmodulin as a probe to screen a mouse brain lambda gt11 library. One clone, lambda ICM-1 isolated using this approach, produces fusion protein that binds calmodulin with high affinity (Kd, 3-10 nM) in a Ca2+-dependent manner. Molecular genetic mapping experiments and deduction of the predicted higher-order structure from sequence data indicate the binding site is, or is within, a basic, amphiphilic alpha-helical domain composed of approximately 20 amino acids. lambda ICM-1 hybridizes with brain mRNA of 2.1 and 3.5 kb but not with mRNA from liver or kidney, suggesting possible restriction of the protein to brain. We discuss several observations that suggest lambda ICM-1 corresponds to Ca2+/calmodulin-dependent protein kinase II, an enzyme that phosphorylates several neuronal proteins, some of which apparently play a role in synaptic function. Our results suggest certain types of ligands may be useful probes to isolate genes encoding various receptor proteins, particularly when the protein is very rare or when it is difficult to obtain antibodies suitable for screening libraries.

Differentiation of the drug-binding sites of calmodulin

Calmodulin contains several binding sites for hydrophobic compounds. The apparent specificity of various 'calmodulin antagonists' for these sites was investigated. The Ki values for the inhibition of calmodulin-activated cyclic-nucleotide phosphodiesterase and myosin light-chain kinase was determined. In addition, the Kd values of the same compounds for binding to calmodulin were measured. The compounds could be separated into four groups. Group I and II compounds inhibited competitively the activation of the phosphodiesterase and myosin light-chain kinase by calmodulin. Group I compounds inhibited the activation of the phosphodiesterase and myosin light-chain kinase at identical concentrations. In contrast, group II compounds inhibited the activation of the phosphodiesterase at 5-10-fold lower concentrations than that of myosin light-chain kinase. Group III compounds inhibited the activation of these enzymes by an uncompetitive mechanism. Group IV compounds inhibited the activation of the phosphodiesterase with Ki values above 10 microM and did not affect the activation of myosin light-chain kinase. Binding of [3H]bepridil to calmodulin under equilibrium conditions yielded one high-affinity site (apparent Kd 0.4 microM) and four low affinity sites (apparent Kd 44 microM). Group I compounds interfered with the binding of bepridil to the high and low-affinity sites in a competitive manner. Group II compounds interfered in a non-competitive manner with the high-affinity site and apparently competed only with one of the low-affinity sites. Group III compounds did not compete with any of the bepridil-binding sites. Nimodipine, a group III compound, bound to one site on calmodulin with a Kd value of 1.1 microM. Other dihydropyridines competed with [3H]nimodipine for this site. The group I and II compounds, trifluoperazine and prenylamine, did not affect the binding of [3H]nimodipine. These data show that 'calmodulin antagonists' can be differentiated into at least three distinct groups. Kinetic and binding data suggest that the three groups bind to at least three different sites on calmodulin. Selective occupation of these sites may inhibit specifically the activation of distinct enzymes.

Deduced primary structure of the beta subunit of brain type II Ca2+/calmodulin-dependent protein kinase determined by molecular cloning

cDNA clones coding for the beta subunit of rat brain type II Ca2+/calmodulin-dependent protein kinase were isolated and sequenced. The clones, including one containing the entire coding region, hybridize at high stringency to a single band of poly(A)+ RNA of length 4.8 kilobases. The subunit coded for by the clones was identified by in vitro transcription of the cDNA followed by translation of the resulting RNA. The DNA sequence of the clones contains a single long open reading frame (1626 nucleotides) coding for a protein of 542 amino acids with a molecular weight of 60,333, the amino-terminal half of which is homologous to several other protein kinases. Potential ATP- and calmodulin-binding domains were identified. Two independent clones contain an identical 45-nucleotide deletion, relative to the clones described above, resulting in a shorter open reading frame coding for a protein of molecular weight 58,000. This suggests that the minor, 58-kDa beta' subunit of the type II Ca2+/calmodulin-dependent kinase may be synthesized on a separate message.

Predicted calmodulin-binding sequence in the gamma subunit of phosphorylase b kinase

A basic, amphiphilic alpha helix is a structural feature common to a variety of inhibitors of calmodulin and to the calmodulin-binding domains of myosin light chain kinases. To aid in recognizing this structural feature in sequences of peptides and proteins we have developed a computer algorithm which searches for sequences of appropriate length, hydrophobicity, helical hydrophobic moment, and charge to be considered as potential calmodulin-binding sequences. Such sequences occurred infrequently in proteins of known crystal structure. This algorithm was used to find the most likely site in the catalytic (gamma) subunit of phosphorylase b kinase for interaction with calmodulin (the delta subunit). A peptide corresponding to this site (residues 341-361 of the gamma subunit) was synthesized and found to bind calmodulin with approximately an 11 nM dissociation constant. A variant of this peptide in which an aspartic acid at position 7 in its sequence (347 of the gamma subunit) was replaced with an asparagine was found to bind calmodulin with approximately a 3 nM dissociation constant.

Affinity purification of seminalplasmin and characterization of its interaction with calmodulin

Bull seminalplasmin antagonizes with high potency and selectivity the activating effect of calmodulin on target enzymes [Gietzen & Galla (1985) Biochem. J. 230, 277-280]. In the present paper we establish that seminalplasmin forms a 1:1, Ca2+-dependent and urea-resistant complex with calmodulin. The dissociation constant equals 1.6 nM. In the absence of Ca2+ a low-affinity complex is formed that is disrupted by 4 M-urea. On the basis of these properties, a fast affinity purification of seminalplasmin was developed. The high specificity of seminalplasmin as a calmodulin antagonist was demonstrated for the multipathway-regulated adenylate cyclase of bovine cerebellum. Far-u.v. c.d. properties are consistent with a random form of seminalplasmin in aqueous solution; 23% alpha-helix is induced on interaction with calmodulin. The fluorescence properties of the single tryptophan residue of seminalplasmin are markedly changed on formation of the complex. These studies allowed us to locate tentatively the peptide segment that interacts with calmodulin, and to ascertain the structural homology between seminalplasmin and other calmodulin-binding peptides. Additional material, showing the inhibition of calmodulin-mediated activation of bovine brain phosphodiesterase by melittin and seminalplasmin and also the near-u.v. spectrum of affinity-purified seminalplasmin, has been deposited as supplement SUP 50135 (4 pages) at the British Library Lending Division, Boston Spa, Wetherby, West Yorkshire LS23 7BQ, U.K., from whom copies may be obtained on the terms indicated in Biochem. J. (1986) 233, 5.