Search

KRP (telokin), an independently expressed C-terminal myosin-binding domain of smooth muscle myosin light chain kinase (MLCK), has been reported to have two related functions. First, KRP stabilizes myosin filaments (Shirinsky et al., 1993, J. Biol. Chem. 268, 16578-16583) in the presence of ATP. Secondly, KRP can modulate the level of myosin light chain phosphorylation. In this latter role, multiple mechanisms have been suggested. One hypothesis is that light chain phosphorylation is diminished by the direct competition of KRP and MLCK for myosin, resulting in a loss of contraction. Alternatively, KRP, through an unidentified mechanism, accelerates myosin light chain dephosphorylation in a manner possibly enhanced by KRP phosphorylation. Here, we demonstrate that KRP is a major phosphoprotein in smooth muscle, and use a comparative approach to investigate how its phosphorylation correlates with sustained contraction and forskolin-induced relaxation. Forskolin relaxation of precontracted artery strips caused little increase in KRP phosphorylation, while treatment with phorbol ester increased the level of KRP phosphorylation without a subsequent change in contractility. Although phorbol ester does not induce contraction of phasic tissues, the level of KRP phosphorylation is increased. Phosphopeptide maps of KRP from both tissues revealed multiple sites of phosphorylation within the N-terminal region of KRP. Phosphopeptide maps of KRP from gizzard were more complex than those for KRP from artery consistent with heterogeneity at the amino terminus and/or additional sites. We discovered through analysis of KRP phosphorylation in vitro that Ser12, Ser15 and Ser15 are phosphorylated by cAMP-dependent protein kinase, mitogen-activated protein (MAP) kinase and glycogen synthase kinase 3 (GSK3), respectively. Phosphorylation by GSK3 was dependent upon prephosphorylation by MAP kinase. This appears to be the first report of conditional or hierarchical phosphorylation of KRP. Peptides consistent with such multiple phosphorylations were found on the in vivo phosphopeptide maps of avian KRP. Collectively, the available data indicate that there is a complex relationship between the in vivo phosphorylation states of KRP and its effects on relaxation in smooth muscle.

Studies on transglutaminases usually focus on the polymerization of protein substrates by intermolecular N(epsilon)(gamma-glutamyl)lysine bridges, without considering the possibility that the monomeric protein units, themselves, could also become crosslinked internally. Both types of crosslinks are produced in the reaction of fibrinogen with red cell transglutaminase. We isolated the transglutaminase-modified, mostly monomeric form (92-96%) of fibrinogen with a N(epsilon)(gamma-glutamyl)lysine content of approximately 1.6 moles/mole of fibrinogen. The preparation was fully clottable by thrombin, but the rates of release of fibrinopeptides and clotting times were delayed compared with control. Hybrid Aalpha.gamma type of crosslinking, the hallmark of the reaction of the transglutaminase with fibrinogen, occurred by bridging the Aalpha(408-421) chain segment of the protein to that of gamma(392-406). Rotary shadowed electron microscope images showed many monomers to be bent, and the crosslinks seemed to bind the otherwise flexible alphaC domain closer to the backbone of fibrinogen.

Myosin light chain kinase (MLCK) is the key regulator of cell motility and smooth muscle contraction in higher vertebrates. We searched for the features of the high molecular weight MLCK (MLCK-210) associated with its unique N-terminal sequence not found in a more ubiquitous lower molecular weight MLCK (MLCK-108). MLCK-210 demonstrates stronger association with the Triton-insoluble cytoskeletons than MLCK-108, suggesting the role for this sequence in subcellular targeting. Indeed, the expressed unique domain of MLCK-210 binds and bundles F-actin in vitro and colocalises with the microfilaments in transfected cells reproducing endogenous MLCK-210 distribution. Thus, MLCK-210 features an extensive actin binding interface and, perhaps, acts as an actin cytoskeleton stabiliser.

A high-throughput screening (HTS) assay for inhibitors of nitric oxide (NO) production by activated microglia was developed and used to compare the relative activities of various anti-inflammatory compounds and cell-permeable protein kinase inhibitors. BV-2 cells, an immortalized line that retains phenotypic features of microglia and produces NO in response to lipopolysaccharide (LPS), were used in the activation paradigm for the HTS assay. A characteristic feature of the compounds that were the most potent dose-dependent inhibitors of NO production is their ability to modulate serine/threonine protein kinases. The anti-inflammatory compound K252a, an inhibitor of calmodulin (CaM)-regulated protein kinases, had one of the highest potencies in the assay. Other classes of kinase inhibitors, including the protein kinase A inhibitor H-89, the mitogen activated protein kinase inhibitors PD98059 and SB203580, and the tyrosine kinase inhibitor genistein, were less potent and efficacious than K252a or the general serine/threonine/tyrosine kinase inhibitor staurosporine. K252a suppresses production of the inducible nitric-oxide synthase (iNOS). The inhibitory effect of K252a is not due to cell toxicity and does not correlate with inhibition of NFkappaB nuclear translocation. The mechanism of action appears to involve inhibition of phosphorylation of the transcription factor CREB, a protein whose activity is modulated by phosphorylation by CaM-dependent protein kinases. These data suggest that signal transduction pathways mediated by CaM-dependent protein kinases warrant future study as potential drug discovery targets.

The 24 kDa flagellar calcium binding protein (FCaBP) of the protozoan Trypanosoma cruzi is a calcium-acyl switch protein. FCaBP is modified by the addition of myristate and palmitate at its amino terminal segment and both modifications are required for calcium-modulated flagellar membrane association. FCaBP has four sequence motifs for potential calcium binding, and comparison to other calcium-acyl switch proteins, such as recoverin, suggested that only two of these sites are functional. Because it is not possible to predict with certainty the calcium binding affinity or selectivity based on motif analysis alone, we determined the quantitative calcium binding activity of FCaBP by direct ligand binding using the flow dialysis method. The results demonstrated the presence of two calcium binding sites in the full length FCaBP and in a mutant (FCaBPdelta12) lacking the amino terminal pair of sites. FCaBPdelta12 retains its ability to localize to the flagellum. A mutant FCaBP lacking the two carboxyl-terminal sites (FCaBPdelta34), did not bind calcium with high affinity and selectivity under the conditions used. The calcium binding properties of FCaBP are therefore distinct from other myristoyl switch proteins such as recoverin. The results add to a growing body of knowledge about the correlation of sequence motifs with calcium binding activity. Moreover, they demonstrate the need to determine the apparently novel mechanism by which FCaBP undergoes calcium modulated flagellar membrane association and its relation to calcium signal transduction.

The enhancement of calmodulin's (CaM) calcium binding activity by an enzyme or a recognition site peptide and its diminution by key point mutations at the protein recognition interface (e.g., E84K-CaM), which is more than 20 A away from the nearest calcium ligation structure, can be described by an expanded version of the Adair-Klotz equation for multiligand binding. The expanded equation can accurately describe the calcium binding events and their variable linkage to protein recognition events can be extended to other CaM-regulated enzymes and can potentially be applied to a diverse array of ligand binding systems with allosteric regulation of ligand binding, whether by other ligands or protein interaction. The 1.9 A resolution X-ray crystallographic structure of the complex between E84K-CaM and RS20 peptide, the CaM recognition site peptide from vertebrate smooth muscle and nonmuscle forms of myosin light chain kinase, provides insight into the structural basis of the functional communication between CaM's calcium ligation structures and protein recognition surfaces. The structure reveals that the complex adapts to the effect of the functional mutation by discrete adjustments in the helix that contains E84. This helix is on the amino-terminal side of the helix-loop-helix structural motif that is the first to be occupied in CaM's calcium binding mechanism. The results reported here are consistent with a sequential and cooperative model of CaM's calcium binding activity in which the two globular and flexible central helix domains are functionally linked, and provide insight into how CaM's calcium binding activity and peptide recognition properties are functionally coupled.

A portion of the neurofibrillary tangles of Alzheimer's disease has the characteristics of cross-linked protein. Because the principal component of these lesions is the microtubule-associated protein tau, and because a major source of cross-linking activity within neurons is supplied by tissue transglutaminase (TGase), it has been postulated that isopeptide bond formation is a major posttranslational modification leading to the formation of insoluble neurofibrillary tangles. Here we have mapped the sites on two isoforms of human tau protein (tau23 and tau40) capable of participating in human TGase-mediated isopeptide bond formation. Using dansyl-labeled fluorescent probes, it was shown that eight Gln residues can function as amine acceptor residues, with two major sites being Gln351 and Gln424. In addition, 10 Lys residues were identified as amine donors, most of which are clustered adjacent to the microtubule-binding repeats of tau in regions known to be solvent accessible in filamentous tau. The distribution of amine donors correlated closely with that of Arg residues, suggesting a link between neighboring positive charge and the TGase selectivity for donor sites in the protein substrate. Apart from revealing the sites that can be cross-linked during the TGase-catalyzed assembly of tau filaments, the results suggest a topography for the tau monomers so assembled.

Gel filtration of the water-soluble extract from bovine lens yields a group of proteins, emerging between the peaks of beta H and beta L crystallins, which show a considerably greater sensitivity to heat-induced aggregation/precipitation than the far more abundant beta and gamma crystallins. However, the small heat shock protein: alpha crystallin was effective in protecting these trace constituents of the lens from precipitating out of solution at 55 degrees C (measured under the standard conditions in a pH 7.5 buffer containing 50 mM sodium phosphate, 100 mM NaCl, 1 mM EDTA and 0.05% NaN3). Prominent components of the precipitate, formed in the absence of a recombinant alpha B crystallin chaperone could be resolved by one- and two-dimensional electrophoresis. Identification by amino acid sequencing revealed that the heat-sensitive group of lens proteins comprised glyceraldehyde-3-phosphate dehydrogenase (M(r) approximately 39 kDa), enolase (approximately 48 kDa), leucine aminopeptidase (approximately 52 kDa) and aldehyde dehydrogenase (approximately 53 kDa). These findings indicate for the first time that the aggregation of such minor lens constituents could possibly contribute to initiating the process of opacification in the development of cataracts.

The apical and principal segments of the bovine acrosome contain a stable matrix complex that is bound to the outer acrosomal membrane and exhibits hydrolase-binding activity. The present study was undertaken to determine whether the outer acrosomal membrane-associated matrix complex (OMC) is composed of a unique set of acrosomal proteins and to define its fate during both capacitation and the acrosome reaction. A purified OMC fraction was isolated from ejaculated spermatozoa, and one polypeptide of 32 kDa (OMC32) was purified to homogeneity and used for N-terminal sequence analysis and preparation of monospecific antisera. Immunofluorescence staining of sperm with anti-OMC32 demonstrated that the polypeptide localized specifically to the apical and principal segments of the acrosome. Immunoelectron microscopy further revealed that OMC32 was restricted to the stable matrix assembly and was not associated with the inner acrosomal membrane or the equatorial segment. Immunoblot analyses of sperm lysates and of the purified OMC fraction revealed that anti-OMC32 recognized an antigenically related family of polypeptides between 38 and 19 kDa. These polypeptides exhibited no size processing during capacitation or the acrosome reaction, and they were not released during the acrosome reaction but remained in the particulate cell subfraction, associated with the hybrid membrane complex. N-terminal sequence analysis of OMC32 indicated a structural relationship to the SP-10 polypeptide family of human and baboon spermatozoa. The potential function of the OMC complex and differences in the intraacrosomal distribution of bovine OMC32-related polypeptides from that reported for acrosomal SP-10 polypeptides in other species are discussed.

The nonmuscle/smooth muscle myosin light chain kinase (MLCK) and the kinase related protein (KRP) that lacks protein kinase activity are myosin II binding proteins encoded in the vertebrate genome by a true gene within a gene relationship. The genomic organization and expression result in the same amino acid sequence in different molecular contexts from two different sizes of mRNA. We report here the identification and characterization of a third size class of gene products. The protein appears to be a higher molecular weight form of MLCK with additional amino terminal tail sequence which might provide differential subcellular targeting characteristics.

The interaction of avian smooth muscle caldesmon with calmodulin (CaM) was investigated by studying the ability of selected mutant calmodulins to induce fluorescence changes in caldesmon. Different types of CaM mutants were used including point charge mutants, cluster mutations, and mutations which alter the calcium binding of CaM. The caldesmon binding properties were only slightly affected by E84K-CaM or by the double mutation E84Q/E120Q-CaM. Affinity of calmodulin to caldesmon was decreased 2-4 times by point mutation G33V-CaM, double mutation E84K/E120K-CaM, deletion of residues 82-84, and by cluster mutations DEE118-120-->KKK or EEE82-84-->KKK. Mutations of the first (E31A-CaM) and the second (E67A-CaM) calcium binding sites reduced the affinity of calmodulin to caldesmon by at least 5-fold; in addition these calmodulin mutants exhibited smaller changes in the fluorescence spectra of caldesmon. Simultaneous mutation of the two negatively charged clusters of calmodulin EEE82-84-->KKK and DEE118-120-->KKK resulted in a more than 15-fold decrease in the affinity of calmodulin for caldesmon. The data indicate that charged and uncharged amino acids in both halves of CaM play an important role in the binding of calmodulin to caldesmon, and that Ca2+ binding must be maintained in the amino-terminal sites for maximal interaction with caldesmon.

Yeast calmodulin binds only three calcium ions in the presence of millimolar concentrations of magnesium due to a defective calcium-binding sequence in its carboxyl terminal domain. Yeast calmodulin's diminished calcium-binding activity can be restored to that of other calmodulins by the use of site-directed mutagenesis to substitute its fourth calcium-binding domain with that of a vertebrate calmodulin sequence. However, the repair of yeast calmodulin's calcium-binding activity is not sufficient to repair quantitatively yeast calmodulin's defective protein kinase activator activity. Yeast calmodulin's activator activity with smooth muscle and skeletal muscle myosin light chain kinases and brain calmodulin-dependent protein kinase II can be progressively repaired by additional substitutions of vertebrate calmodulin sequences, provided that the four calcium-binding sites remain intact. An unexpected result obtained during the course of these studies was the observation that myosin light chain kinases from smooth and skeletal muscle tissues can respond differently to mutations in calmodulin. These and previous results indicate that the binding of four calcium ions by calmodulin is necessary but not sufficient to bring about quantitative activation of protein kinases, and are consistent with the conformational selection/restriction model of the dynamic equilibrium among calcium, calmodulin and each calmodulin regulated enzyme.

An apparent paradox in smooth muscle biology is the ability of unphosphorylated myosin to maintain a filamentous structure in the presence of ATP in vivo, whereas unphosphorylated myosin filaments are depolymerized in vitro in the presence of ATP. This suggests that additional uncharacterized factors are required for the stabilization of myosin filaments in the presence of ATP. We report here that an abundant smooth muscle protein forms sedimentable complexes with unphosphorylated smooth muscle myosin, partially reverses the depolymerizing effect of ATP on unphosphorylated myosin, and promotes the assembly of minifilaments as revealed by electron microscopy. This protein is called kinase-related protein (KRP) because it is derived from a gene within the gene for myosin light chain kinase (MLCK) and has an amino acid sequence identical to the carboxyl-terminal domain of MLCK. Consistent with the results with purified KRP, deletion of the KRP domain within MLCK results in a diminished ability of MLCK to interact with unphosphorylated myosin. KRP binds to the heavy meromyosin fragment of myosin but not to myosin rod or fragments lacking the hinge region and light chains. Altogether, these results suggest that KRP may play a critical role in stabilizing unphosphorylated myosin filaments and that the KRP domain of MLCK may be important for subcellular targeting to filaments.

Calmodulin (CaM) is an integral subunit, called delta, of the phosphorylase kinase hexadecamer, and the activity of the isolated catalytic gamma-subunit of the kinase is stimulated by CaM. We report here the first analysis of functionally important features of CaM for activation of the gamma-subunit. A set of genetically engineered CaMs, in which acidic residues in each of the four E-helices of the "EF-hands" were changed to basic lysine residues, was used to probe the relative importance of charge features in each domain of CaM. The maximal activation of the isolated gamma-subunit was diminished by all of the charge reversal mutations. The gamma-subunit was especially sensitive to reversals in the second and third E-helix of CaM (residues 45-47 and 82-84), the latter being present in the central helix. The results suggest the functional importance of electrostatics in the interactions between the delta-subunit (CaM) and the catalytic gamma-subunit of phosphorylase kinase, which is similar to results obtained with CaM-dependent myosin light chain kinase (MLCK) from chicken gizzard and CaM-dependent protein kinase II (CaMPK-II). However, novel features of the interaction between CaM and the gamma-subunit of phosphorylase kinase are the significant contribution of electrostatics throughout the CaM molecule, including residues in both halves and on more than one face of CaM, and the lack of a major effect of the CaM mutations on substrate kinetic parameters, unlike the effects observed with MLCK and CaMPK-II. These results are consistent with a model in which the delta-subunit (CaM) of phosphorylase kinase interacts with an extended region or multiple regions of the gamma-subunit and suggest that the mechanism of CaM activation of the gamma-subunit may have features that are distinct from those of MLCK and CaMPK-II.

The heterodimer complex of calmodulin (CaM) and the protein kinase catalytic subunit of myosin light chain kinase from vertebrate smooth muscle and non-muscle tissues (sm/nmMLCK) is one of the most extensively characterized CaM-regulated enzyme complexes and it has an established in vivo role in the transduction of calcium signals into biological responses. We have used a combination of approaches to the study of CaM and sm/nmMLCK in order to derive initial insight into the key features of each protein and of the CaM-MLCK heterodimeric complex that are involved in protein-protein and calcium-protein recognition and regulation of enzyme activity. On-going studies are described here that include site-specific mutagenesis, fluorescence spectroscopy, enzymology and peptide analog analysis. These and previous results indicate that: (1), both electrostatic and hydrophobic features are important in the functionally correct interactions between CaM and MLCK; (2), even the interactions between CaM and peptide analogs of the CaM binding site of MLCK are heterogeneous and non-trivial in nature; (3), amino-acid residues that have been conserved in CaM across millions of years of evolution and that are conserved in CaMs with quantitative MLCK activator activity can be mutated without any detectable effect on activity and (4), structures different from the prototypical EF-hand domain of CaM can have similar calcium-binding activity in the presence of a CaM binding structure.

Peptide-induced conformational changes in five isofunctional mutants of calmodulin (CaM), each bearing a single tryptophan residue either at the seventh position of each of the four calcium-binding loops (i.e., amino acids 26, 62, 99, and 135) or in the central helix (amino acid 81) were studied by using fluorescence spectroscopy. The peptides RS20F and RS20CK correspond to CaM-binding amino acid sequence segments of either nonmuscle myosin light chain kinase (nmMLCK) or calmodulin-dependent protein kinase II (CaMPK-II), respectively. Both steady-state and time-resolved fluorescence data were collected from the various peptide-CaM complexes. Steady-state fluorescence intensity measurements indicated that, in the presence of an excess of calcium, both peptides bind to the calmodulin mutants with a 1:1 stoichiometry. The tryptophans located in loops I and IV exhibited red-shifted emission maxima (356 nm), high quantum yields (0.3), and long average lifetimes (6 ns). They responded in a similar manner to peptide binding, by only slight changes in their fluorescence features. In contrast, the fluorescence intensity of the tryptophans in loops II and III decreased markedly, and their fluorescence spectrum was blue-shifted upon peptide binding. Analysis of the tryptophan fluorescence decay of the last mentioned calmodulins supports a model in which the equilibrium between two (Trp-99) or three (Trp-62) states of these tryptophan residues, each characterized by a different lifetime, was altered toward the blue-shifted short lifetime component upon peptide binding. Taken together, these data provide new evidence that both lobes of calmodulin are involved in peptide binding. Both peptides induced similar changes in the fluorescence properties of the tryptophan residues located in the calcium-binding loops, with the exception of calmodulin with Trp-135. For this last mentioned calmodulin, slight differences were observed. Tryptophan in the central helix responded differently to RS20F and RS20CK binding. RS20F binding induced a red-shift in the emission maximum of Trp-81 while RS20CK induced a blue-shift. The quenching rate of Trp-81 by iodide was slightly reduced upon RS20CK binding, while RS20F induced a 2-fold increase. These results provide evidence that the environment of Trp-81 is different in each case and are, therefore, consistent with the hypothesis that the central helix can play a differential role in the recognition of, or response to, CaM-binding structures.

Trifluoperazine (TFP) binding by 14 calmodulins, including 12 produced by site-directed mutagenesis, was determined. While vertebrate calmodulin binds 4.2 +/- 0.2 equiv of TFP, Escherichia coli expressed but unmutated calmodulins bind about 5.0 +/- 0.5 equiv of TFP. The cause for this difference is not known. The E. coli expressed proteins consist of two different series expressed from different calmodulin genes, CaMI and SYNCAM. The wild-type genes code for proteins that differ by nine conservative amino acid substitutions. Both these calmodulins bind 5 equiv of TFP with similar affinities, thus none of these conservative substitutions has any additional effect on TFP binding. Some altered calmodulins (deletion of EE83-84 or SEEE81-84, changing DEE118-120----KKK, M124----I,E120----K, or E82----K) have no appreciable effect on TFP binding. Other mutations affect either the binding of one TFP (deletion of E84) or about two TFP (changing E84----K, EEE82-84----KKK, E67----A, DEQ6-8----KKK, or E11----K). The mutations that affect TFP binding are localized to three regions of calmodulin: The amino-terminal alpha-helix, the central helix between the two globular ends of calmodulin, and a calcium-binding site in the second calcium-binding domain. The results are consistent with each of these regions either directly participating in drug binding or involved structurally in maintaining or inducing the correct conformation for TFP binding in the amino-terminal half of calmodulin.

Presented in this paper is a knowledge-based experimental design system that incorporates the domain expertise used in nucleic acid engineering, thus automating the processing of error-prone, laborious low-level work, and many decision-making steps, and guiding the biologist toward a workable plan. This allows the biologist to work at a higher abstraction level, concentrating on more fundamental, difficult and challenging problems directly related to protein structure - function relationships. Cassette-based site-directed mutagenesis and synthetic gene designs are used as examples to illustrate the utility of the knowledge-based system approach to experimental design.

Calmodulin is a eukaryotic calcium binding protein which has several calcium-dependent in vitro activities. Presented in this report is a structural characterization of calmodulin from spinach leaves (Spinacia oleracea). Spinach calmodulin may be representative of higher plant calmodulins in general since calmodulin from the monocotyledon barley (Hordeum vulgare) is indistinguishable by a variety of physical, chemical, and functional criteria (Schleicher, Lukas, Watterson 1983 Plant Physiol 73: 666-670). Spinach calmodulin is homologous to bovine brain calmodulin with only 13 identified amino acid sequence differences, excluding a blocked NH(2)-terminal tripeptide whose sequence has not been elucidated. Two extended regions of sequence identity are in the NH(2)-terminal half of the molecule, while nine of the 13 identified differences are in the COOH-terminal half of the molecule. Two of the changes, a cysteine at residue 26 and a glutamine at residue 96, require a minimum of two base changes in the nucleotide codons. Both of these changes occur in the proposed calcium binding loops of the molecule. Five additional amino acid differences found in spinach calmodulin had not been observed previously in a calmodulin. As described in an accompanying report (Roberts, Burgess, Watterson 1984 Plant Physiol 75: 796-798), these limited number of amino acid sequence variations appear to result in differential effects on the activation of calmodulin-dependent enzymes by plant and vertebrate calmodulins.

The molecular mechanisms by which calcium regulates cellular processes such as metabolic and mechanochemical events probably involve interactions with a variety of molecules. A large body of evidence suggests that the targets of calcium's regulatory effects inside the cell are calcium-binding proteins. Our work attempts to correlate calcium-binding protein structure with activities. In this chapter we have presented some of our recent studies on these functional domains in calmodulin and calmodulin-binding proteins. Selected chemical modifications of known amino acid sequence positions have demonstrated the presence of multiple functional domains on calmodulin, have allowed the dissociation of calmodulin functions, and have provided the necessary tools for further investigations of the molecular basis of calmodulin action. One of these modifications, iodination of tyrosine-99, has allowed us to develop procedures to reproducibly detect calmodulin-binding proteins by a binding technique. This method is technically simple. It allows us to detect and study calmodulin-binding proteins (e.g., myosin heavy chain and membrane gap junction proteins) that would be difficult to study with immobilized calmodulin. We have developed a library of antibodies to calmodulin and related proteins such as troponin C and S100 beta. Some of these antisera appear to be site-specific gamma globulins. We have demonstrated that reactivity can be contained in an amino acid sequence as short as seven residues. We have demonstrated the feasibility of using an immunochemical mapping approach to study calmodulin and calmodulin-binding proteins. Comparative sequence analyses combined with functional analyses have allowed correlation of function and structure and have suggested logical candidates for functional domains on calmodulin and calmodulin-binding proteins. Although not discussed in detail in this chapter, these calmodulin-binding proteins appear to contain amino acid sequence homologies. This suggests, analogous to the approach used for calmodulin and related proteins, a logical starting point for domain analyses of calmodulin-binding proteins. Interestingly, structural homologies among calmodulin-binding proteins are reminiscent of the different phosphorylation sites found in many of the physiological substrates for protein kinases. Since a number of calmodulin-binding proteins are themselves substrates for protein kinases, these results suggest another possible point of interrelationships between calcium and cyclic nucleotide regulation.