FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2043935934> ?p ?o ?g. }

• W2043935934 endingPage "6750" @default.
• W2043935934 startingPage "6741" @default.
• W2043935934 abstract "Two neuronal protein kinase C substrates, RC3/neurogranin and GAP-43/neuromodulin, preferentially bind to calmodulin (CaM) when Ca2+ is absent. We examine RC3•CaM and GAP-43•CaM interactions by circular dichroism spectroscopy using purified, recombinant RC3 and GAP-43, sequence variants of RC3 displaying qualitative and quantitative differences in CaM binding affinities, and overlapping peptides that cumulatively span the entire amino acid sequence of RC3. We conclude that CaM stabilizes a basic, amphiphilic α-helix within RC3 and GAP-43 under physiological salt concentrations only when Ca2+ is absent. This provides structural confirmation for two binding modes and suggests that CaM regulates the biological activities of RC3 and GAP-43 through an allosteric, Ca2+-sensitive mechanism that can be uncoupled by protein kinase C-mediated phosphorylation. More generally, our observations imply an alternative allosteric regulatory role for the Ca2+-free form of CaM. Two neuronal protein kinase C substrates, RC3/neurogranin and GAP-43/neuromodulin, preferentially bind to calmodulin (CaM) when Ca2+ is absent. We examine RC3•CaM and GAP-43•CaM interactions by circular dichroism spectroscopy using purified, recombinant RC3 and GAP-43, sequence variants of RC3 displaying qualitative and quantitative differences in CaM binding affinities, and overlapping peptides that cumulatively span the entire amino acid sequence of RC3. We conclude that CaM stabilizes a basic, amphiphilic α-helix within RC3 and GAP-43 under physiological salt concentrations only when Ca2+ is absent. This provides structural confirmation for two binding modes and suggests that CaM regulates the biological activities of RC3 and GAP-43 through an allosteric, Ca2+-sensitive mechanism that can be uncoupled by protein kinase C-mediated phosphorylation. More generally, our observations imply an alternative allosteric regulatory role for the Ca2+-free form of CaM. Calmodulin (CaM) 1The abbreviations used are: CaMcalmodulinPKCprotein kinase C9AC9-anthroylcholineMOPS4-morpholinepropanesulfonic acidTFEtrifluoroethanol. activates numerous proteins in response to Ca2+ fluxes. Its N- and C-terminal globular domains, which are connected by a flexible, α-helical “tether,” each contain a pair of helix-loop-helix-binding sites with slightly different Ca2+ affinities (for reviews, see Forsén et al., 1986; Means and Conn, 1987; Cohen and Klee, 1988; Means, 1988; Strynadka and James, 1989; McPhalan et al., 1991; Kretsinger, 1992; Weinstein and Mehler, 1994). Occupation of at least the high affinity pair by Ca2+ causes the flexible tether to bend, bringing the globular domains closer together to form a more compact structure. At the same time, more subtle rearrangements within each globular domain cause increased exposure of hydrophobic side chains (LaPort et al., 1980; Tanaka and Hidaka, 1980). Collectively, these conformational changes enable CaM to bind and activate many classes of proteins, including protein kinases, inositol triphosphate kinase, nicotinamide adenine dinucleotide kinase, calcineurin, calcium pumps, cyclic nucleotide phosphodiesterase, cyclases, nitric oxide synthase, and cytoskeletal proteins (Kincaid and Vaughan, 1986; Tallant and Cheung, 1986; Ryu et al., 1987; Cohen and Klee, 1988; Means, 1988; Bredt and Snyder, 1990; Klee, 1991; Carafoli et al., 1992). Transient activations of some or all of these proteins are essential for the development, growth, and environmental adaptation of virtually all cell types, as well as for plasticity within the mammalian central nervous system. calmodulin protein kinase C 9-anthroylcholine 4-morpholinepropanesulfonic acid trifluoroethanol. Two neuronal proteins are unusual because they bind CaM with a greater affinity in the absence of Ca2+ (Andreasen et al., 1983; Baudier et al., 1991). One is GAP-43, also known as neuromodulin or b-50, which is associated with prenatal axonal growth and can be induced in adults by axonal injury (for reviews see Skene, 1989; Liu and Storm, 1990; Benowitz and Perrone-Bizzozero, 1991; Coggins and Zweirs, 1991; Gispen et al., 1991; Strittmatter et al., 1992). GAP-43 associates with the cytoplasmic face of axonal growth-cone membranes and has been implicated in presynaptic events that contribute to synaptic development and such neuroplastic phenomena as neurite extension (Doster et al., 1991; Fitzgerald et al., 1991; Meiri et al., 1991; Sommervaille et al., 1991; Lin et al., 1992), modulation of neurotransmitter release (De Graan et al., 1991; Dekker et al., 1991) and long term potentiation (Lovinger et al., 1985; Gianotti et al., 1992). The second protein, known as RC3, neurogranin, or BICKS, is a neuron-specific, postnatal-onset protein which accumulates in forebrain dendritic spines, loosely associated with postsynaptic structures (Watson et al., 1990, 1992; Represa et al., 1990; Baudier et al., 1991; Coggins et al., 1991). It is probably involved in neonatal synaptogenesis and has been implicated in the post-synaptic second messenger cascade of long term potentiation (Klann et al., 1992; Chen et al., 1993a). Recently one other protein, the neuron-specific Drosophila protein igloo (Neel and Young, 1994) has been reported to interact preferentially with CaM when Ca2+ is absent. It contains three regions that are homologous with the overlapping PKC recognition and CaM-binding domains of RC3 and GAP-43. Other proteins have been described that bind to CaM regardless of ambient Ca2+, but these interactions are either Ca2+-independent or stronger when Ca2+ is present (Sharma and Wang, 1986; Ladant, 1988; Dasgupta et al., 1989). We will use the term “Ca2+ sensitive” to refer to interactions that require the absence of Ca2+. The CaM-binding regions of RC3 and GAP-43 share primary sequence similarities with those of proteins that interact with CaM only when Ca2+ concentrations are high: many basic amino acids, no acidic amino acids, and an abundance of hydrophobic residues with a periodicity suggestive of an amphiphilic α-helix (Anderson and Malencik, 1986; Blumenthal and Krebs, 1988; O'Neil and Degrado, 1990). In the presence of Ca2+, CaM binding stabilizes a basic, amphiphilic α-helix within target proteins that can be detected by circular dichroism (CD) and nuclear magnetic resonance (NMR) spectroscopies (Cox et al., 1985; DeGrado et al., 1985; McDowell et al., 1985; Klevit et al., 1985; Seeholzer et al. 1986; MacLachlan et al., 1990; Ikura et al., 1992a, 1992b; Precheur et al., 1991, 1992; Vorherr et al., 1990; Munier et al., 1993; Zhang et al., 1993; Zhang and Vogel, 1994a, 1994b). The overlapping CaM-binding and PKC recognition domains of both RC3 and GAP-43 are predicted to form an α-helix when analyzed by the method of Garnier et al.(1978) or with the Gascuel and Golmard (1988) basic statistical method. A pair of strongly hydrophobic amino acids separated by 8 or 12 residues (Ile33 or Phe37, and Ile46 in RC3, Ile30 or Phe43, and Leu52 in GAP-43) may correspond to residues in Ca2+/CaM-activated proteins that interact with one or both of the pliant hydrophobic patches exposed by CaM in response to Ca2+ (Ikura et al., 1992a; Meador et al., 1993). RC3 and GAP-43 belong to a family of proteins that bind CaM or CaM-like molecules and share an IQ motif with a consensus sequence of IQxxxRGxxxR (Mercer et al., 1991; Cheney and Mooseker, 1992). The isoleucine in this consensus sequence corresponds to Ile33 in RC3 and Ile30 in GAP-43. This motif is found in p68 RNA helicase (Buelt et al., 1994) and in heavy chains of many myosins in regions that interact with light chains (light chains are structurally and functionally similar to CaM). Xie et al.(1994) solved the structure of scallop myosin and observed that the two light chains stabilize a long α-helix within the heavy chain containing its IQ motifs. We recently characterized the RC3•CaM interaction by fluorescence emission spectroscopy and proposed that RC3 acted as a capacitor for CaM availability within dendritic spines (Gerendasy et al., 1994). That study, which utilized purified recombinant RC3 and sequence variants containing amino acid substitutions within their CaM-binding domains (Gerendasy et al., 1995), suggested that there were two CaM-affinity forms of RC3, but did not address directly the structural basis of these binding modes. To evaluate the hypothesis that CaM binding stabilizes a helix within RC3 and GAP-43, we investigated the binding interactions by CD. We conclude that CaM mediates the Ca2+-sensitive stabilization of basic, amphiphilic α-helices within RC3 (residues 26-47) and GAP-43 (residues 31-52) under physiological salt concentrations. When Ca2+ is present, RC3 and GAP-43 associate with CaM with little or no detectable stabilization of structure. Our findings provide structural support for the notion of two CaM affinity forms of RC3 and, by extension, GAP-43, and further suggest a mode by which their activities might be regulated. More generally, they suggest that CaM induces a helical conformation within certain target proteins only when Ca2+ is low or absent, implying that the Ca2+-free form of CaM may also be an active regulatory molecule. Salt-free calmodulin from bovine brain (>98% purity) was obtained from Sigma and 9-anthroylcholine from Molecular Probes Inc. (Eugene, OR). All solutions, proteins, and peptides used in these experiments were passed through Bio-Rex sample preparation discs containing Chelex Chelating Ion Exchange membranes (Bio-Rad) to strip remaining Ca2+ from CaM and to remove Ca2+ and other trace metals from the other reagents. Sequences corresponding to the protein coding region of RC3, its sequence variants, and GAP-43 were amplified by polymerase chain reaction, cloned into pRK172, expressed, and purified as described, (Gerendasy et al., 1995). The identities of all proteins were verified and their purity and concentrations determined in duplicate by quantitative amino acid analyses using a Beckman model 6300 high performance analyzer modified with a 25-cm long cation exchange column for improved resolution. Nine peptides derived from the amino acid sequence of different regions of RC3 that, together, span the entire protein (Table 1), and the Ser-to-Ala variant of residues 25-47 were synthesized on an Applied Biosystems peptide synthesizer (model 430A) using solid-phase t-boc chemistry and purified by reverse-phase high performance liquid chromatography using a VYDAC C-18 column. Absence of contaminating Ca2+ and other potentially confounding cations was verified by fast ion bombardment. Quantitative amino acid content analysis, performed in duplicate, was used to determine peptide concentrations. Because we were particularly interested in comparing the mean residue ellipticity of RC325-47 with S36A25-37, amino acid content analysis of each peptide was performed in quintuplicate, yielding concentrations with standard errors of 4.6 and 6.8%, respectively.Tabled 1 All CD spectra were recorded on an AVIV model 61DS spectropolarimeter calibrated with a standard solution of 10-camphorsulfonic acid. Spectra were measured in either a quartz cuvette with a path length of 0.2 cm or in tandem quartz cells with path lengths of 0.437 cm. In mixing experiments using the tandem cuvette, 3.5 μM RC3 or 3.5 μM peptide was placed in one compartment, and 3.5 μM CaM was placed in the other. Mixing experiment involving GAP-43 required lower concentrations (1.25 μM) of each component to achieve an optimum signal. In all other experiments, protein concentrations ranged from 5 to 10 μM, while peptide concentrations ranged from 67 to 198 μM. All buffers contained 100 μM Tris, pH 7.5, and 20 μM EGTA or 5 mM CaCl2 unless otherwise indicated. Mean residue ellipticity was determined from the relationship [θ]MRE = (θ•100•Mr)/(c•d•Na) where θ is the measured ellipticity in degrees, c is the protein concentration in mg/ml, d is the path length in cm, Mr is the peptide molecular weight, and Na is the number of amino acid residues/peptide. When a tandem cuvette was used Mr, c, and Na of each component were averaged. Spectra were recorded with a 0.50-nm bandwidth, 0.5-nm step size, and a time constant of 4.0 s. Emission fluorescence spectroscopy of 9-anthroylcholine (9AC) was performed as described (LaPort et al., 1980) with an SPF-500C spectrofluorimeter at 24°C using a path length of 1 cm. The excitation beam (361 nm) was passed through a 2-nm slit, and the emitted spectrum (375-600 nm) was collected in 1-nm increments through a 7.5-nm slit. Each data point was sampled 32 times and averaged. Emission spectra were generated by 9AC (5 μM) dissolved in 50 mM MOPS, pH 8.0, 200 mM KCl containing 2 μM CaM and 1 mM EGTA or 2 mM CaCl2 as indicated. In two cases 2 μM RC3 was also included. Interactions between CaM and purified, recombinant RC3 or GAP-43, or a synthetic peptide corresponding to RC3 residues 25-47, were monitored by CD spectroscopy using a tandem cuvette. RC325-47 contains the CaM-binding site, PKC recognition domain, and region of homology of RC3 with GAP-43. In each case RC3, GAP-43 or RC325-47 was placed in one compartment of the cuvette while an equimolar concentration of CaM was placed in the other. A spectrum was obtained before and after mixing the contents of the two chambers, allowing detection of changes in spectra that resulted from molecular interactions between the components in the two chambers (Fig. 1, A and B). These experiments were performed in the presence and absence of Ca2+. Likewise, the effects of Ca2+ or KCl on the conformation of individual proteins was examined by collecting spectra before and after mixing a solution of the protein with buffer containing KCl or Ca2+ (Fig. 2, A and B).Figure 2:Effect of Ca2+, KCl, and intermolecular interactions on the conformations of RC3, GAP-43, and CaM. A and B, RC3 or CaM (indicated), in the presence of EGTA, was placed in one compartment of a tandem cuvette and buffer containing 5 mM Ca2+ (A) or 150 mM KCl (B) was placed in the other. CD spectra were collected before (solid lines) or after (dotted lines) mixing the components of each compartment. C and D, mixing experiments were performed between CaM and RC3 (C) or GAP-43 (D), as described in Fig. 1 and under “Experimental Procedures,” except that increasing concentrations of KCl (0, 50, 100, 150, 200, 250, 300, and 400 mM) were included in both compartments of the tandem cuvette. CD spectra were collected before (open circles) or after (closed triangles) mixing the components of each compartment. Pre- and post-mix spectra generated in the absence of KCl are indicated. With increasing concentration of KCl, the pre-mix spectra tend to descend and the post-mix spectra tend to ascend when viewed between 205 and 240 nm. E, summary of mixing experiments displayed in panels C and D. Mean residue ellipticities, measured at 222 nm ([θ]MRE222) are plotted against concentration of KCl. Data derived from interactions between CaM and GAP-43 (ovals) or RC3 (rectangles) are indicated. Light stippling is used within symbols that represent pre-mix data points while dark stippling is used in symbols representing post-mix data points. F, fluorescence emission spectroscopy of 9AC in the presence of Ca2+/CaM, Ca2+/CaM+RC3, EGTA/CaM, or EGTA/CaM+RC3 was performed as described under “Experimental Procedures.”View Large Image Figure ViewerDownload (PPT) The CD spectra generated individually by RC3, GAP-43, and RC325-47 showed marginal negative ellipticities at 222 nm (Fig. 1C; top spectrum in each panel) suggesting a general lack of structure for each species. The spectra generated by RC3 was not altered appreciably by KCl (Fig. 2B), the presence of reducing agents, changes in pH, or by changes in protein concentration (data not shown). Mixing RC3 (Fig. 2A) or GAP-43 (data not shown) with Ca2+ to a final concentration of 2 mM also had no appreciable effect. Examination of CaM by itself revealed conspicuous minima at approximately 222 and 205 nm, characteristic of an α-helical conformation, which increased significantly at 2 mM Ca2+ (Fig. 2A). When RC3, GAP-43, or RC3 25-47 was placed in one chamber and equimolar CaM in the other, each spectrum corresponded to the sum of the spectra of the individual components (Fig. 1,A and B). Because RC3 and RC325-47 are small and relatively unstructured compared to CaM, their respective pre-mix spectra were dominated by that of CaM. On the other hand, GAP-43 which is larger and proportionately less structured makes a greater contribution to the pre-mix spectra. After mixing, no differences in spectra were observed in the presence of Ca2+ (Fig. 1B); however, markedly different spectra were seen when Ca2+ was absent (Fig. 1A). The differences involved significant enhancement of negative ellipticities between wavelengths 205-240 nm, with minima at approximately 205 and 222 nm, characteristic of an increase in α-helical content. These enhanced spectra resembled those obtained when Ca2+ was present (Fig. 1B), which are dominated by the contribution from the Ca2+-CaM interaction (Fig. 2A). Calcium had little effect on the spectra generated by RC3 or GAP-43 alone, raising the disturbing possibility that the RC3, GAP-43, and RC325-47 preparations contained contaminating Ca2+ or the intriguing possibility that these proteins caused conformational changes in CaM by mimicking Ca2+. The first possibility is unlikely. During purification, RC3 and GAP-43 were precipitated with trichloroacetic acid, resuspended in buffers containing EGTA, and never again exposed to Ca2+. Furthermore, RC325-47 and other synthetic peptides were screened for ionic impurities by ion mass spectroscopy and found to be uncontaminated. Thus, the proteins themselves appear to be responsible for changes in conformation. The second possibility, that RC3 or GAP-43 induce conformational changes in CaM analogous to those caused by Ca2+, is inconsistent with the following experimental data: calcium induces both gross and fine conformational changes within CaM. At the gross level these changes can be imitated by salt, a phenomenon described previously by Hennessey et al.(1987). As salt-induced CaM α-helicity reaches saturation, the ability of Ca2+ to promote additional α-helicity disappears and, at physiological salt concentrations, Ca2+ fails to elicit a response that can be detected by CD spectroscopy. Whereas increasing ionic strength had no effect on RC3 α-helical content, it caused a detectable increase in that of CaM (Fig. 2B). To evaluate the effect of KCl on the structural consequences of complex formation, we performed a series of mixing experiments with RC3 or GAP-43 in one chamber of a tandem cuvette and CaM in the other and varied the concentrations of KCl equally in both chambers. Due to the strong contribution of CaM to the pre-mix spectra, the minima at 205 and 222 prior to mixing increased with salt concentration until saturation was reached (Fig. 2, C-E). The change in ellipticity that was caused by mixing CaM and GAP-43 diminished with increasing salt concentrations until a KCl concentration of 200 mM, at which point no difference could be discerned between pre- and post-mix spectra (Fig. 2E). On the other hand, the interaction between RC3 and CaM continued to produce a change in ellipticity, even when 400 mM KCl was present (Fig. 2E). At this concentration of salt, CaM does not exhibit increased negative ellipticity in response to addition of Ca2+. When similar mixing experiments were performed in the presence of Ca2+, little if any difference could be discerned between pre- and post-mix spectra (data not shown). Assuming KCl and Ca2+ have similar effects, the observed increase must have been caused either by the stabilization of an α-helical conformation within RC3 or structural changes within CaM that are fundamentally different than those caused by Ca2+. KCl does not cause CaM to activate Ca2+/CaM-dependent enzymes, so more subtle conformational changes must also be mediated by Ca2+ and these have been detected using various fluorescent probes (Forsén et al., 1986). For example, at physiological salt concentrations, Ca2+-induced conformational changes within CaM cause exposure of hydrophobic clefts within each globular domain that results in an increased quantum yield of the fluorescence probe (9AC) (LaPort et al., 1980). Since the exposure of a pliant hydrophobic face appears to be a prerequisite for the activation of Ca2+/CaM-dependent enzymes, we tested RC3 for ability to induce such structural transformations by monitoring RC3-induced changes in emitted fluorescence from a mixture of CaM and 9AC (Fig. 2F). When Ca2+ was added to the mixture, the quantum yield increased as expected. Inclusion of RC3 in either the presence or absence of Ca2+ quenched the signals slightly, perhaps due to interactions between RC3 and CaM or to environmental effects. Nevertheless, RC3 did not increase fluorescence emission in the absence of Ca2+. Thus, interactions between CaM and RC3, GAP-43, or RC325-47 lead to an increased α-helicity within one or both molecules only when Ca2+ is absent, and the effect is different than that which occurs when Ca2+ binds to CaM. Many targets of CaM are highly fluid and fail to adopt a detectable helical conformation unless stabilized through interactions with Ca2+/CaM or helix-stabilizing agents such as trifluoroethanol (TFE) (Precheur et al., 1991; Vorherr et al., 1990, 1992; Munier et al., 1993; Zhang et al., 1993; Zhang and Vogel, 1994a, 1994b). Although random coiled conformations predominate within RC3, GAP-43, and RC325-47 (Fig. 1C, Table 1 and Table 2), addition of increasing concentrations of TFE causes dramatic increases in negative ellipticity indicating a strong predilection for an α-helical conformation in all three cases.Tabled 1 To map the region(s) of RC3 responsible for inducing global increases in negative ellipticity, we compared nine peptides, with amino acid sequences that, cumulatively, span the entire sequence of RC3, for their capacities to increase α-helicity when mixed with CaM (Fig. 3, 5A, and Table 1). In the absence of Ca2+ (Fig. 3A), RC324-37 showed an 85% increase in negative ellipticity when mixed with CaM, while peptides RC334-51, RC340-55, and RC344-54 showed changes on the order of 40%. No appreciable changes were observed with the other five peptides when mixed with CaM in the absence of Ca2+, and none of the nine provoked responses greater than 10% when Ca2+ was present (Fig. 3B). CD spectra of each of the peptides were also obtained individually and in the presence of increasing concentrations of TFE (Fig. 3C, 5A, and Table 1). Prior to the addition of TFE, peptides RC31-20, RC312-24, and RC316-42 exhibited mean residue ellipticities (MRE[θ]222 expressed in deg cm2 dmol-1) of approximately −1.0 × 104, while peptides derived from more C-terminal sequences displayed mean residue ellipticities of approximately −0.4 × 104 to 0.4 × 104. The CD spectra of these peptides suggest a random coiled conformation (data not shown). When increasing amounts of TFE were added, RC316-42, RC325-47, RC334-51, RC340-55, and RC344-54 responded with increasing negative ellipticities, indicating a propensity for an α-helical conformation. The remaining four peptides showed no appreciable sensitivity to TFE. In the presence of 80% TFE, peptide RC316-42 exhibited the greatest helical content followed closely by that of RC325-47, while those of RC334-51, RC340-55, and RC344-54 were relatively small. Nevertheless, all five peptides showed large proportional responses in that 80% TFE induced negative ellipticities manyfold higher than that displayed in a completely aqueous environment. Thus, with the exception of RC316-42, which lacks crucial CaM-binding residues, the helical propensity and degree of fluidity displayed by individual peptides correlated closely with their ability to evoke a change in α-helicity when interacting with CaM in the absence of Ca2+. Control experiments performed in the presence of dithiothreitol ruled out the possibility that spectral changes resulted from the formation of cystine dimers (data not shown). We previously demonstrated that replacing RC3 Phe37 with Trp (F37W) caused an increased affinity for CaM in the presence or absence of Ca2+, while substituting Ala for Ser36 (S36A) only caused a large increase in affinity when Ca2+ was absent (Gerendasy et al., 1994, 1995). Replacing Ser36 with Lys (S36K) had little effect under low ionic conditions while replacing it with Asp (S36D) mimicked the phosphorylation of RC3 by eliminating all detectable affinity for CaM. The biochemical phenotype of F37W suggested that Ca2+-independent binding was augmented by increasing the hydrophobic bulk of residue 37. We were, therefore, curious to learn whether the interaction between F37W and CaM would result in a Ca2+-sensitive, Ca2+-dependent or Ca2+-independent increase in helicity. Comparison of RC3, its variants F37W, S36A, S36K and S36D, and GAP-43 revealed that each of the proteins except for S36D evoked large increases in negative ellipticity when mixed with CaM in the absence but not presence of Ca2+ (Fig. 4, Table 2). If CaM stabilizes an α-helix within the CaM-binding domains of RC3 and GAP-43, mutations that enhance Ca2+-sensitive binding, such as the substitution of Ala for Ser36 in RC3 might be expected to do so by augmenting helical tendencies within the binding domain. To test this hypothesis we examined the negative ellipticities of RC3 and sequence variants F37W, S36A, S36K, and S36D with and without TFE. (Table 2). We also compared the ellipticities of peptide RC325-47 and one corresponding to the same region of S36A (S36A25-47) (Table 1). Although S36A and S36A25-47 appeared to exhibit greater negative ellipticities than RC3 and RC325-47, respectively, the differences fell within the standard errors of the amino acid content analysis used to determine protein and peptide concentrations. RC3 and GAP-43 participate in conformational changes involving the stabilization of an α-helix when they interact with CaM. These changes are only prominent when Ca2+ is absent. Although the observed shifts in CD spectra are virtually indistinguishable from those caused by Ca2+-induced conformational changes within CaM, fluorescence spectroscopy in the presence of 9AC and salt saturation experiments indicates that the spectral shifts caused by mixing RC3 or GAP-43 with CaM were caused by different phenomena. Additionally, the salt saturation experiments suggest that CaM is not the source of the spectral shift in the case of mixing experiments involving RC3. Spectroscopy of RC3 or GAP-43 alone indicates that their individual conformations are not significantly affected by Ca2+ or salt, despite their high degree of structural fluidity and strong predilection for the α-helical conformation, as demonstrated by TFE studies. Mapping experiments with peptides demonstrated, with one exception, a correlation between those regions exhibiting structurally fluid α-helical tendencies and those that elicit Ca2+-sensitive increases in negative ellipticity when interacting with CaM. The exception was RC316-42, which exhibited a high degree of fluidity and strong α-helical tendencies but failed to interact with CaM, almost certainly because of the absence of residues 43 through 47 (RKKIK). These residues are conserved in GAP-43 and required for its binding to CaM (Alexander et al., 1988; Baudier et al., 1991; Chapman et al., 1991). This highly basic stretch of amino acids is probably required for initial anchoring of RC3 or GAP-43 to CaM as discussed below. The data summarized above constitutes strong circumstantial evidence that an α-helix is stabilized within the CaM-binding domains of RC3 and GAP-43 upon binding to CaM when Ca2+ is absent and is consistent with the helical structure observed in the heavy chain of scallop myosin (Xie et al., 1994), another protein that contains an IQ motif. The conclusion is further supported by its ability to explain the dissimilar CaM binding affinities of sequence variants F37W and S36A in spite of the close proximity and similarly hydrophobic nature of their respective sequence substitutions (the former exhibits a greater Ca2+-independent affinity while the latter displays an increased Ca2+-sensitive affinity). One would expect that a Ser-to-Ala substitution would be helix stabilizing while a Phe-to-Trp substitution would not (Marqusee, et al., 1989; Padmanabhan et al., 1990). Increasing the hydrophobic bulk of residue 37 could enhance Ca2+-independent binding while replacing Ser36 with a helix-promoting residue would increase Ca2+-sensitive binding. Unfortunately, the methods used here lack the necessary precision to demonstrate increased α-helicity on the part of S36A. Interactions between F37W or GAP-43 and CaM generated only a small or undetectable increase in α-helicity, respectively, when Ca2+ was present, even though a significant portion of each would be expected to form a complex with CaM, independent of Ca2+ levels (both have micromolar dissociation constants in Ca2+). In low ionic environments such as those used here, RC3, S36K, and S36A also interact, although to lesser extents, in Ca2+ and induce small changes in ellipticity when they associate with CaM (S36D has the lowest affinity and elicits the smallest change in ellipticity). Alexander et al.(1987) have demonstrated that GAP-43-CaM affinity is Ca2+-independent when salt concentrations approach physiological levels (150 mM KCl). The salt sensitivity displayed by GAP-43 has been interpreted to imply a significant ionic component to the interaction, although the affinity appears to be inversely proportional to the induction by salt of conformational changes within CaM, implying a possible causal relationship between affinity and KCl-induced conformational changes. (Fig. 2E). We demonstrated that changes in ellipticity occur when CaM and GAP-43 interact in 150 mM KCl when Ca2+ is absent but not when present. As salt increased from 0 to 200 mM, the change in mean residue ellipticity decreased along with affinity, probably because fewer molecules formed complexes. Similar mixing experiments revealed negligible change in the presence of Ca2+ and salt (data not shown), indicating that the stabilization of helical structure within GAP-43 by CaM is Ca2+ sensitive even under circumstances that render its affinity Ca2+-independent. RC3•CaM interactions are less sensitive to salt, indicating either a smaller ionic component or less sensitivity to KCl-induced gross conformational changes within CaM. The data imply that an α-helix with a positively charged face and a hydrophobic face is stabilized within RC3 between residues 25 and 47 by CaM when Ca2+ is absent (Fig. 6). In the presence of Ca2+ or the absence of CaM, a randomly coiled state dominates, although the significant negative ellipticity elicited by TFE within RC316-42 and RC335-47 suggests that residues 25 through 35 occasionally sample the helical conformation, even when CaM is not present (Fig. 5E). This region appears to stabilize and be stabilized by interactions with CaM involving the previously determined CaM-binding domain C-terminal to Ser36. Coincident stabilization of an α-helix on both the N-terminal and C-terminal sides of Ser36 may be influenced by the ease with which an α-helix can be propagated through residue 36. Hence, the fraction of molecules sampling the helical conformation in the unbound state may increase when Ser36 is replaced with alanine and this, in turn, could translate into a greater Ca2+-sensitive affinity for CaM.Figure 5:Graphic representation of peptide data. A-C, nine overlapping peptides that collectively span the entire 78-residue sequence of RC3 are represented by rectangles. The data are extracted from Table 2. A, [θ]MRE222 of each peptide prior to the addition of TFE. B, [θ]MRE222 of each peptide in the presence of 80% TFE. C, percent change in [θ]MRE222 that occurred when each peptide was mixed with CaM in the presence of EGTA. The PKC recognition and minimal CaM-binding domains (Apel et al., 1990; Alexander et al., 1988; Baudier et al., 1991; Chapman et al., 1991; Houbre et al., 1991) are indicated at the top of each panel. D, primary structure of RC3 and homologous region of GAP-43. The overlapping PKC recognition and CaM-binding domains are indicated over the primary structure of RC3. The minimal CaM-binding domain as determined by Alexander et al.(1998) and Chapman et al.(1991) is indicated by a solid line while additional residues deemed to be important for Ca2+-sensitive interactions with CaM based on the data presented here are indicated by a dotted line. E, hypothetical secondary structure of the that region. The helical propensity exhibited by each peptide, inferred from its behavior in TFE, and each peptide's individual ability to elicit a Ca2+-sensitive increase in helicity when mixed with CaM, suggests that sequences N-terminal to Ser36, which occupy a predominantly random coiled conformation (left), are able to transiently sample a helical conformation (center). Direct binding of CaM, in the absence of Ca2+, to residues C-terminal to Ser36 are hypothesized to stabilize a helix on the C-terminal side. This propagates and is further stabilized by the formation of a helix N-terminal to Ser36 (right).View Large Image Figure ViewerDownload (PPT) Negative ellipticities exhibited by RC325-47, RC334-51 but not RC340-55 when exposed to TFE or CaM suggest that amino acids N-terminal to Ser36 (ANAAAAKIQAS) contribute substantially to the formation of an α-helix. There are three chemically non-conservative amino acid differences between RC3 and GAP-43 in this region (Fig. 5D), which could account for their different affinities for CaM and the sensitivity of those affinities to salt. RC3 and GAP-43 differ from other known CaM-binding proteins in that their CaM-binding domains only assume a stable amphiphilic, α-helical conformation when interacting with CaM in the absence of Ca2+. Sequences N-terminal to Ser36 may be responsible for this unconventional behavior. Interactions between RC3 or GAP-43 and CaM, PKC, and possibly membrane phospholipids may be mediated through positively charged and hydrophobic faces of the α-helix (Fig. 6). This would be consistent with the hypothesis that ACTH antagonizes GAP-43•CaM and GAP-43•PKC interactions through its presumed ability to form a competing amphiphilic α-helix (Zweirs and Coggins, 1991). It is also consistent with the observation that RC3 and de-myristalted GAP-43 associate with liposomes composed of acidic phospholipids (Houbre et al., 1991). In this model, Lys32, His40, Arg43, Lys44, Lys45, and Lys47 constitute one positively charged surface that winds around one side of the helix, unbroken except for Ser36, the target of PKC. Phosphorylation of Ser36 or its substitution with aspartate would place a negative charge in the center of this positively charged face, possibly disrupting interactions with CaM. Residue 36 appears to be pivotal for the propagation of an α-helix, thus explaining how its phosphorylation or substitution with aspartate abrogates both low and high affinity binding. Chen et al.(1993) have demonstrated that Lys32 is important for PKC specificity, implying that the positively charged face is important for recognition by PKC. Ser36 also lies immediately adjacent to another continuous surface composed of hydrophobic residues Ala28-Ala31, Ile33, Phe37, Met41, Ala142, and Ile46. The amino acid differences in GAP-43 do not alter the nature of these surfaces. Mutagenesis and fluorescence binding studies with peptides and recombinant proteins indicate that at least one of the residues within the hydrophobic face, Phe37 in RC3 or Phe43 in GAP-43, interacts with CaM and with PKC (Alexander et al., 1988; Chapman et al., 1991, Chen et al., 1993; Gerendasy, et al., 1994). Masure et al.(1986) have concluded that GAP-43 is rod shaped based on its sedimentation rate during velocity centrifugation and that its CD spectrum indicated a predominantly random coiled configuration and very little α-helicity (1%). In partial agreement, Coggins et al.(1989) reported that they were unable to detect any α-helical structure in GAP-43 by proton nuclear magnetic-resonance spectroscopy. Gel filtration chromatography suggests that RC3 is also rod shaped (Huang et al., 1993). Based on CD spectra of peptides corresponding to the sequence of RC3, its N terminus (residues 1-30) appears to be somewhat more structured than its C terminus (residues 50-78), which is similar in sequence to collagen and forms a random coil conformation. The distinctive CD spectrum generated by the triple helix conformation of collagen is not evident in any of the spectra generated here. The central domain (residues 30-50) is unstructured but has a strong helical propensity that is stabilized by CaM. Thus, one might infer that both RC3 and GAP-43 are rod shaped but quite fluid within their CaM-binding domains. Stabilization of a helix within this region could decrease the freedom of motion at the ends as well as within the immediate binding domain resulting in a stiffer, more pronounced rod-like structure when Ca2+ levels are low. This structure and the less organized form may correspond to the two affinity states of RC3 which were previously inferred by fluorescence emission spectroscopy (Gerendasy et al., 1994). Alexander et al.(1987, 1988) and Liu and Storm (1990) proposed that GAP-43 sequesters CaM next to the inner leaf of the axonal growth cone membrane, releasing it with increasing Ca2+ and/or phosphorylation by PKC. We recently proposed that RC3 acts like a capacitor in dendritic spines, releasing CaM rapidly in response to large increases in Ca2+ and much more slowly in response to smaller increases. Neither of these functions precludes the possibility that RC3 or GAP-43 are themselves biologically active as has been suggested by Gispen(1986), Van Hoof et al.(1988), Strittmatter et al.(1990, 1992, 1993), and Cohen et al.(1993). The stabilization of a helix within RC3 and GAP-43 only by the Ca2+-free form of CaM suggests that CaM regulates their activities in a Ca2+-sensitive manner. Phosphorylation of either substrate by PKC would uncouple this regulation leading to constitutive Ca2+-independent activity. This hypothesis is consistent with a number of observations. Cohen et al.(1993) demonstrated that phosphorylated RC3 stimulates inositol triphosphate-mediated mobilization of Ca2+ in response to activation of G protein-coupled receptors in Xenopus oocytes. They also recently observed that S36D but not S36A behaves the same as phosphorylated RC3. 2J. B. Watson, unpublished results. Strittmatter et al.(1990) demonstrated that the N terminus of GAP-43, which lacks a CaM-binding domain, is able to stimulate G0 in vitro and that GAP-43 had the same effect as RC3 in Xenopus oocytes (Strittmatter et al., 1993). We report here that CaM can stabilize a conformational state within RC3, S36A, and GAP-43 but not S36D when Ca2+ is absent. Together, these data argue that RC3 and GAP-43 are subject to Ca2+-sensitive, allosteric regulation by CaM that can be uncoupled by phosphorylation. Perhaps, more generally, calmodulin's ability to stabilize a basic, amphiphilic, α-helix within a specific class of target proteins, solely when Ca2+ is absent, implies that the Ca2+-free form of CaM has the potential to regulate the actions of other proteins through an allosteric mechanism similar to that employed by the Ca2+-containing form. Since RC3 may also regulate CaM availability, the model proposed here suggests that CaM and RC3/neurogranin regulate each other's activities. We gratefully acknowledge Dr. Peter Wright and Dr. Reza Ghadiri for the use their circular dichroism spectrophotometers, Dr. Michael Buchmeier for peptide synthesis and purification, Kerry Guinn and the Scripps Research Institute Protein Service for quantitative amino acid analysis of peptides and purified proteins, Dr. Gary Siuzdak and the mass spectrometry laboratory for analysis of peptides and purified proteins by fast ion bombardment, Dr. Michael Pique and the Molecular Graphics Laboratory for the creation of Fig. 5E and Fig. 6." @default.
• W2043935934 created "2016-06-24" @default.
• W2043935934 creator A5046657525 @default.
• W2043935934 creator A5053680151 @default.
• W2043935934 creator A5058691210 @default.
• W2043935934 creator A5063683312 @default.
• W2043935934 date "1995-03-01" @default.
• W2043935934 modified "2023-10-18" @default.
• W2043935934 title "Calmodulin Stabilizes an Amphiphilic α-Helix within RC3/Neurogranin and GAP-43/Neuromodulin Only When Ca2+ Is Absent" @default.
• W2043935934 cites W1484977141 @default.
• W2043935934 cites W1485888367 @default.
• W2043935934 cites W1496223818 @default.
• W2043935934 cites W1498962226 @default.
• W2043935934 cites W1503320832 @default.
• W2043935934 cites W1506302012 @default.
• W2043935934 cites W1511885651 @default.
• W2043935934 cites W1512693490 @default.
• W2043935934 cites W1518396254 @default.
• W2043935934 cites W1526773019 @default.
• W2043935934 cites W1529053562 @default.
• W2043935934 cites W1545633418 @default.
• W2043935934 cites W1556906420 @default.
• W2043935934 cites W1575693343 @default.
• W2043935934 cites W1586520658 @default.
• W2043935934 cites W1592556850 @default.
• W2043935934 cites W1634735131 @default.
• W2043935934 cites W1838152396 @default.
• W2043935934 cites W1965611215 @default.
• W2043935934 cites W1967230963 @default.
• W2043935934 cites W1968901680 @default.
• W2043935934 cites W1969953781 @default.
• W2043935934 cites W1977134775 @default.
• W2043935934 cites W1985587671 @default.
• W2043935934 cites W1991767840 @default.
• W2043935934 cites W1993626595 @default.
• W2043935934 cites W2000478781 @default.
• W2043935934 cites W2000521810 @default.
• W2043935934 cites W2003697928 @default.
• W2043935934 cites W2008924730 @default.
• W2043935934 cites W2009064225 @default.
• W2043935934 cites W2012704789 @default.
• W2043935934 cites W2018217893 @default.
• W2043935934 cites W2023147195 @default.
• W2043935934 cites W2024142216 @default.
• W2043935934 cites W2024782877 @default.
• W2043935934 cites W2026660096 @default.
• W2043935934 cites W2030669640 @default.
• W2043935934 cites W2030998945 @default.
• W2043935934 cites W2031811477 @default.
• W2043935934 cites W2038585972 @default.
• W2043935934 cites W2042748946 @default.
• W2043935934 cites W2050217213 @default.
• W2043935934 cites W2050889339 @default.
• W2043935934 cites W2052829813 @default.
• W2043935934 cites W2057553906 @default.
• W2043935934 cites W2058835113 @default.
• W2043935934 cites W2058925305 @default.
• W2043935934 cites W2059379142 @default.
• W2043935934 cites W2063627391 @default.
• W2043935934 cites W2071380535 @default.
• W2043935934 cites W2076203826 @default.
• W2043935934 cites W2076664539 @default.
• W2043935934 cites W2077446863 @default.
• W2043935934 cites W2077642580 @default.
• W2043935934 cites W2079076786 @default.
• W2043935934 cites W2082554893 @default.
• W2043935934 cites W2088002244 @default.
• W2043935934 cites W2088716547 @default.
• W2043935934 cites W2089831707 @default.
• W2043935934 cites W2095450147 @default.
• W2043935934 cites W2096716204 @default.
• W2043935934 cites W2104274727 @default.
• W2043935934 cites W2117005920 @default.
• W2043935934 cites W2120142493 @default.
• W2043935934 cites W2120633292 @default.
• W2043935934 cites W2141888986 @default.
• W2043935934 cites W2145961910 @default.
• W2043935934 cites W2164555150 @default.
• W2043935934 cites W2173578956 @default.
• W2043935934 cites W2182766924 @default.
• W2043935934 cites W2183120559 @default.
• W2043935934 cites W2344277549 @default.
• W2043935934 cites W2473867347 @default.
• W2043935934 cites W2744186677 @default.
• W2043935934 cites W276226909 @default.
• W2043935934 cites W4229913877 @default.
• W2043935934 doi "https://doi.org/10.1074/jbc.270.12.6741" @default.
• W2043935934 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/7896819" @default.
• W2043935934 hasPublicationYear "1995" @default.
• W2043935934 type Work @default.
• W2043935934 sameAs 2043935934 @default.
• W2043935934 citedByCount "84" @default.
• W2043935934 countsByYear W20439359342013 @default.
• W2043935934 countsByYear W20439359342014 @default.
• W2043935934 countsByYear W20439359342015 @default.
• W2043935934 countsByYear W20439359342016 @default.
• W2043935934 countsByYear W20439359342017 @default.
• W2043935934 countsByYear W20439359342018 @default.

• next