Circular Dichroism and 1H Nuclear Magnetic Resonance Studies on the Solution and Membrane Structures of GAP-43 Calmodulin-binding Domain*

(Received for publication, August 13, 1996, and in revised form, January 13, 1997)

Nobuhiro Hayashi , Mamoru Matsubara , Koiti Titani and Hisaaki Taniguchi §

From the Division of Biomedical Polymer Science, Institute for Comprehensive Medical Science, Fujita Health University, Toyoake, Aichi 470-11, Japan

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
Acknowledgment
REFERENCES


ABSTRACT

Growth-associated protein-43 (GAP-43) is believed to be palmitoylated near the N terminus and the modification is assumed to be involved in the membrane anchoring of the protein. However, GAP-43 isolated from bovine brain is not palmitoylated as shown by mass spectrometric analysis, but still retains the ability to bind phospholipids, suggesting that other parts of the molecule are involved in the interaction. Upon addition of acidic phospholipids, purified GAP-43 showed a conformational change from random coil to alpha -helix as indicated by a change in CD spectra. A synthetic peptide corresponding to the calmodulin-binding domain showed a similar conformational change from random coil to alpha -helix in the presence of various acidic phospholipids. These results suggest that the calmodulin-binding domain of GAP-43 is directly involved in the GAP-43-membrane interaction and undergoes a conformational change upon binding to phospholipid membranes. After phosphorylation by protein kinase C, the phospholipid-induced conformational changes were no longer observed. Structural characteristics of the calmodulin-binding domain peptide in aqueous and hydrophobic solvents were further studied in detail by two-dimensional 1H nuclear magnetic resonance. The results obtained suggest that the domain assumes a nascent alpha -helical structure in aqueous solution, which is stabilized under hydrophobic environments.


INTRODUCTION

GAP-43 (growth-associated protein-43, also known as B50, F1, P56, or neuromodulin)1 has been characterized as one of the major phosphoproteins in neuronal growth cone as well as in adult nerve terminals, and is thought to be involved in neurite extension as well as in neuronal plasticity and regulation of neurotransmitter release (for reviews, see Refs. 1 and 2). Although the protein has been believed to be a specific in vivo substrate of protein kinase C (PKC), recent studies on the in vivo phosphorylation sites revealed that the protein is a substrate of so-called proline-directed protein kinases as well (3, 4). Since casein kinase II has also been shown to phosphorylate GAP-43 in vitro (5, 6), physiological functions of GAP-43 such as sequestering of calmodulin (7) and signal transduction through the binding to GO (8) may be regulated through various protein kinases in a very complex manner.

GAP-43 is found associated with membrane fractions or with so-called membrane-cytoskeleton fractions (9). Since the protein is very hydrophilic and lacks any apparent membrane-binding domain (10), the interaction of the protein with membranes is thought to be effected by palmitoylation at two cysteine residues near the N terminus (11, 12). However, our recent mass spectrometric study on the in vivo post-translational modifications of GAP-43 revealed that the protein isolated from the membrane fractions of adult brain lacks the palmitoyl group (3), suggesting that part(s) of the molecule other than the palmitoyl moiety is involved in the interaction of the protein with the membrane. In fact, the involvement of the calmodulin-binding domain, which is at the same time the phosphorylation domain by PKC, in the membrane association has been suggested (13, 14).

We have previously shown that the PKC phosphorylation and calmodulin-binding domain of myristoylated alanine-rich protein kinase C substrate, another in vivo major substrate of PKC belonging to the same family of acidic hydrophilic membrane-associated proteins (15), is directly involved in the interaction of the protein with membrane phospholipids (16). Furthermore, the phosphorylation of the domain regulates the reversible membrane association of myristoylated alanine-rich protein kinase C substrate (14, 16). Thus, the calmodulin-binding domain of basic amphiphilic alpha -helical nature seems to serve as a phosphorylation- and calmodulin-dependent membrane binding motif as well (16, 17). A similar polybasic motif found in the Src family proteins has been implicated in the membrane association of these proteins (18, 19).

In the present study, structural characteristics of the calmodulin-binding domain of GAP-43 were studied by circular dichroism (CD) spectroscopy and two-dimensional 1H nuclear magnetic resonance (NMR). Purified non-palmitoylated GAP-43 and a synthetic peptide corresponding to the domain showed similar conformational changes from random coil to alpha -helix upon binding to acidic phospholipids. The calmodulin-binding domain of GAP-43 seems to assume a "nascent" alpha -helical structure in aqueous solution, which is further stabilized under hydrophobic environments.


EXPERIMENTAL PROCEDURES

Materials

GAP-43 was purified from membrane fractions of bovine brain after detergent extraction as described (3). PKC was purified from bovine brain as described previously (20). A peptide (QASFRGHITRKKLKGEK) corresponding to the calmodulin-binding domain of GAP-43, named GAP peptide, was synthesized using conventional t-butoxycarbonyl chemistry using an ABI 430A peptide synthesizer (Applied Biosystems), and purified over a C18 reversed-phase column (Vydac 218TP1010, The Separations Group, Hesperia, CA) using H2O-acetonitrile gradient in the presence of 0.1% trifluoroacetic acid. Phospholipids (Avanti Polar Lipids) were suspended in 5 mM phosphate buffer (pH 7.5) and sonicated in a sonicator (Branson Sonifier 250) for 30 min. After centrifugation in a tabletop centrifuge for 20 min, the supernatant was used as unilamellar liposomes.

Preparation of Phosphorylated GAP-43 and GAP Peptide

Phosphorylation of the intact GAP-43 and GAP peptide by PKC was carried out in 25 mM Tris-HCl buffer (pH 7.5) containing 10 mM MgCl2, 100 µM CaCl2, 80 µg/ml phosphatidylserine (PS), 8 µg/ml dioleoyl glycerol, and 1 mM ATP at 35 °C for 90 min. The reaction was stopped by adding 0.1% final concentration of trifluoroacetic acid. The extent of the phosphorylation was analyzed by mass spectrometry as described previously (3, 21). The phosphorylated GAP peptide was purified by reversed-phase HPLC using a Vydac C18 column (218TP52, 0.46 × 25 cm), while the phosphorylated GAP-43 protein was purified by chromatography on a Mono Q column (HR 5/5) using a linear gradient of NaCl (0-0.5 M) in 20 mM Tris-HCl buffer (pH 7.5) containing 1 mM EDTA and 1 mM dithiothreitol. The purified phosphorylated protein was concentrated by ultrafiltration using a Centricon 10 cartridge (Amicon).

Mass Spectrometric Analysis

GAP-43 purified from the membrane fractions of bovine brain by the detergent extraction method (3, 20) was digested with L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin (Worthington), and the peptides were separated over a reversed-phase HPLC column as described previously (3, 21). Fractions containing the N-terminal peptide were combined and dried with a Speed-Vac concentrator. The peptide was dissolved in 5 ml of 50% acetonitrile containing 0.1% trifluoroacetic acid. Electrospray tandem mass spectra were recorded in a PE Sciex API-III mass spectrometer in tandem mass mode. A singly-charged precursor ion was selected with Q1 for collision-induced decomposition with argon in Q2. Fragment ion spectra were obtained by scanning Q3 of the triple quadrupole system.

Circular Dichroism (CD) Spectrometry

CD spectra were recorded at 25 °C in a JASCO J-720 CD spectropolarimeter using a 0.1-cm cell. The concentration of the peptide was 20 µM in 5 mM phosphate buffer (pH 7.3) unless otherwise indicated. The contents of alpha -helix, beta -sheet, and beta -turn structures were estimated from the CD spectra on an IRIS Indigo XS24 workstation (Silicon Graphics, Inc.) using a CONTIN program (22) modified by Dr. F. Arisaka, Tokyo Institute of Technology.

NMR Spectrometric Analysis

1H NMR spectra were recorded on a Bruker DMX-500 spectrometer operating at 500 MHz. Chemical shifts were measured relative to the methyl resonance of 4,4-dimethyl-4-silapentane-1-sulfonate, used as an internal reference. GAP peptide (5 mM) was dissolved in H2O/D2O (9/1,v/v), D2O (99.98%), H2O/D2O/trifluoroethanol (TFE)-d3 (5/1/4), or D2O/TFE-d3 (6/4). The pH of the samples was 4.0 (direct meter reading). The sequence-specific assignment of resonances was obtained from two-dimensional TOCSY (23), NOESY (24, 25), DQF-COSY with phase cycling (26) or with pulsed field gradient (27, 28), and TQF-COSY with pulsed field gradient (27, 28) spectra, which were acquired at 25 °C in the phase-sensitive mode using the time proportional phase increment technique. Various mixing times were used for NOESY (200, 100, 50, 25, and 10 ms) and TOCSY (75, 50, and 25 ms).

Water suppression was accomplished using WATERGATE (29, 30) or presaturation. A total of 512 measurements with increasing t1 values were made, and 64 transients were accumulated for each measurement. For t2, 2048 data points were taken, and the spectral widths along f2 were 5000 Hz. The data were zero filled once in the f1 dimension, and a cosine window function was used in f1 dimension before Fourier transformation, and a Gaussian function was used in f2 dimension, except for the NOESY spectra whose time domain data were multiplied by Gaussian functions in both dimensions. All spectra were processed using Bruker XWIN-NMR or MSI Felix 95.0 software packages.

Sedimentation Assay

The direct binding of GAP-43 and that of GAP peptide to phospholipid membranes were assessed by cosedimentation analysis. GAP-43 (1 µM) or GAP peptide (20 µM) was mixed with indicated amounts of multilamellar liposomes containing 20% PS and 80% phosphatidylcholine (PC) in 5 mM potassium phosphate buffer (pH 7.5) containing 150 mM KCl, and incubated at 25 °C for 30 min. After centrifugation at 200,000 × g for 60 min, protein or peptide remained in supernatants were analyzed by SDS-gel electrophoresis and Coomassie-stained bands were quantified by densitometry (Molecular Dynamics, PDSI densitometer).

Titration Calorimetry

Heat production by peptide binding to unilamellar vesicles of PC containing 20% PS was measured with a MicroCal MC-2 high-sensitivity titration calorimeter (MicroCal, Northampton, MA) as described previously (31, 32). Five microliters of the lipid vesicles (13.9 mM) were injected into the calorimeter cell containing 1.3 ml of 30 µM GAP peptide in 5 mM phosphate buffer containing 150 mM KCl at 25 °C.

Other Analytical Methods

Protein concentration was determined by quantitative amino acid analysis. Purity and authenticity of the peptide during synthesis and purification, and those of the proteins were routinely monitored by mass spectrometry as described previously (3, 21).


RESULTS

Characterization of GAP-43 Isolated from Bovine Brain

GAP-43 purified from membrane fractions of bovine brain was digested with trypsin and the resulting peptide mixture was analyzed by the liquid chromatography-electrospray mass spectrometry (3, 21). One peptide with a mass of 796.3 Da has been tentatively assigned as the N-terminal peptide in the previous study (3). The N-terminal peptide should give a theoretical mass of 798.3 Da, when the N-terminal acetylation is assumed. The reversed-phase column fractions containing the peptide were combined, and the sample was analyzed by the electrospray tandem mass spectrometry (3). In this technique, a precursor ion corresponding to the peptide is selected with the first quadrupole mass analyzer, and the fragments obtained by the collision with argon are analyzed using the second mass analyzer. Most of the cleavage occur randomly in the middle of the peptide bonds, resulting in the formation of "ladder" mixtures (33). As shown in Fig. 1, the tandem mass spectrum of the peptide clearly indicated that the N terminus is acetylated, and that the two Cys residues, which have been assumed to be palmitoylated (11), are not palmitoylated at all, but form an intramolecular disulfide bond. GAP-43 has only two Cys residues, and no other modification has been detected by mass spectrometric analysis (3). Since the mass spectrometric analysis of the whole protein showed a single peak with a minor peak corresponding to a phosphorylated species (3), we concluded that the GAP-43 protein as isolated does not contain palmitoylated species in significant amounts.


Fig. 1. Tandem mass spectrum of GAP-43 N-terminal peptide. Product (fragment) ion spectrum of singly-charged precursor ion of N-terminal peptide (m/z = 796.3) was obtained as described under "Experimental Procedures." bn and yn ions, formed by the cleavage of the peptide bond between the nth and n+1th amino acids from the N terminus and the C terminus, respectively, refer to the N-terminal and the C-terminal peptide fragments (50). yn -17 ions are formed by the loss of NH3 from yn ions.
[View Larger Version of this Image (25K GIF file)]


To confirm that the non-palmitoylated GAP-43 still has the ability to bind to phospholipid membranes, the binding was studied by sedimentation analysis. When GAP-43 was mixed with multilamellar liposomes made of a mixture of PC (80%) and PS (20%), which mimics the compositions of biomembranes, at a physiological ionic strength, most of GAP-43 (89%) was found associated with the lipids (Fig. 2a). No binding was observed with pure PC liposomes (data not shown). The non-palmitoylated GAP-43, therefore, still retains the ability to bind to phospholipid membranes containing physiologically relevant combinations of lipids. A synthetic peptide corresponding to the calmodulin-binding domain of GAP-43 showed a similar binding to the phospholipid membranes. With increasing concentrations of vesicles containing PC (80%) and PS (20%), the amounts of peptide remained in the supernatant decreased (Fig. 2, b and c). On the other hand, no significant binding was observed when vesicles containing pure PC was used (Fig. 2c). A quantitative analysis of the peptide-phospholipid interaction was carried out further by high-sensitivity titration calorimetry (31, 32). As shown in Fig. 3a, each injection of lipid vesicles into a peptide solution in the measuring cell produced a marked heat production. With increasing numbers of injections the magnitude of the reaction enthalpy decreased, since less and less peptide is available for binding. From the cumulative heat of reaction plotted against the injection numbers which is correlated to the lipid concentration, one can calculate thermodynamic parameters (32). The partition constant thus obtained, Kp = 3.35 × 103 M-1, suggests that GAP peptide has an affinity comparable to other membrane-binding peptides (14, 32). A similar study with the whole GAP-43 protein was not feasible due to the amounts of samples needed and its tendency to aggregate at high concentrations.


Fig. 2. Binding of GAP-43 and GAP peptide to phospholipid membranes. a, binding of GAP-43 (1 µM) to PC liposomes (1.2 mg/ml) containing 20% PS was analyzed by co-sedimentation assay as described under "Experimental Procedures." GAP-43 remained in the supernatant after centrifugation was analyzed by SDS-gel electrophoresis. Lane 1, protein only; lane 2, protein plus liposomes. b, GAP peptide (20 µM) was incubated with increasing amounts of the same liposomes, and the peptide remained in the supernatant was analyzed by SDS-gel electrophoresis. Concentrations of liposomes were 0, 0.1, 0.5, 1.0, 3.0, 5.0, 7.0, and 9.6 mg/ml. c, the amounts of GAP peptide remained in the supernatant were determined by densitometry and plotted against lipid concentration (bullet ). A similar analysis was done with pure PC liposomes (black-triangle).
[View Larger Version of this Image (23K GIF file)]



Fig. 3. Titration calorimetry of GAP peptide binding to lipid vesicles. a, the raw data were obtained by titrating a 30 µM GAP peptide solution with 13.9 mM unilamellar PC liposomes containing 20% PS in 5 mM phosphate buffer (pH 7.3) containing 150 mM KCl at 25 °C as described under "Experimental Procedures." b, the cumulative reaction enthalpy obtained by integrating the peak areas in a. The solid line corresponds to the best theoretical fit calculated with Delta H = -10.7 kcal/mol, and a molar partition coefficient Kp = 3350 M-1, and an effective charge <z> = 0.99. The Gouy-Chapman model was applied to the theoretical calculation for correction of electrostatic effects (32, 34).
[View Larger Version of this Image (22K GIF file)]


Conformational Change of GAP-43 Induced by Phospholipid Binding

To get more insights into the mode of the GAP-43-phospholipid interactions, effects of phospholipids on the conformation of GAP-43 were studied by CD spectroscopy. Purified GAP-43 in aqueous buffer showed a CD spectrum with a single large negative peak at around 200 nm (Fig. 4a), suggesting that the protein has a random structure. The addition of an acidic phospholipid, phosphatidylglycerol (PG), caused a small but significant change in the CD spectrum, which was dependent on the concentration of PG added. There was a decrease in the intensity and a shift of the largest negative peak around 200 nm with a concomitant increase in the negative ellipticity between 210 and 230 nm. The characteristics of the change can be more clearly observed in the difference spectrum (Fig. 4a, inset). A broad negative peak between 210 and 230 nm with a positive peak below 200 nm suggests that a part of the protein molecule underwent a conformational change from random coil to alpha -helix.


Fig. 4. Effects of phosphatidylglycerol on CD spectra of GAP-43 and those of GAP peptide. CD spectra of GAP-43 were obtained in the presence of varying concentrations of PG (a). Concentration of GAP-43 was 3 µM. bullet , GAP-43 alone; open circle , 0.10 mM PG; ×, 0.16 mM PG; black-square, 0.25 mM PG. Difference spectrum between CD spectra of GAP-43 observed in the presence and in the absence of 0.25 mM PG is shown (a, inset). CD spectra of GAP peptide were obtained in the presence of varying concentrations of PG (b). Concentration of GAP peptide was 20 µM. bullet , GAP peptide alone; open circle , 0.06 mM PG; ×, 0.40 mM PG; black-square, 1.58 mM PG.
[View Larger Version of this Image (28K GIF file)]


Interaction of Calmodulin-binding Domain Peptide with Phospholipids

The CD spectrum of the GAP peptide showed again a typical CD spectrum of random coil with a single large negative peak at around 198 nm (Fig. 4b). When increasing amounts of PG were added to the peptide, a broad negative peak between 210 and 230 nm due to alpha -helix increased accordingly (Fig. 4b). The maximal extent of the change in the CD spectra per mole of molecule was comparable for the intact protein (0.98 × 105 degree·cm2·dmol-1) and the peptide (1.22 × 105 degree·cm2·dmol-1), suggesting that only the calmodulin-binding domain interacts with the lipids and undergoes the conformational change.

Phospholipid Specificity in the Binding of GAP Peptide

Various phospholipids at a fixed concentration were mixed with GAP peptide, and CD spectra were recorded to study the phospholipid specificity. Although neutral phospholipids such as PC did not affect the spectrum significantly, all acidic phospholipids examined caused similar changes in the CD spectra (Fig. 5). Although the final spectral change observed at saturating concentrations were very similar regardless of the phospholipid used (data not shown), the affinities of various acidic phospholipids varied appreciably, suggesting that the binding of the peptide to the phospholipids involves not only ionic but also specific structural interactions. Among various acidic phospholipids, phosphatidic acid showed the highest affinity to GAP peptide. When ionic strength of the buffer was increased, the apparent affinity between the peptide and the phospholipids decreased, suggesting that the electrostatic interaction plays an important role in the binding (14, 34).


Fig. 5. Phospholipid specificity of GAP peptide-phospholipid interaction. CD spectra of GAP peptide (20 µM) were measured in the presence of various phospholipids at a fixed concentration of phospholipids (0.25 mM). bullet , GAP peptide alone; open circle , PC; ×, PS; black-square, phosphatidylinositol; square , phosphatidylglycerol; black-triangle, phosphatidic acid. Spectral changes at saturating concentrations were similar regardless of the acidic phospholipids used.
[View Larger Version of this Image (20K GIF file)]


Effects of TFE on the Conformation of GAP Peptide

To study the structural characteristics of the calmodulin-binding domain of GAP-43 under hydrophobic environments in detail, effects of TFE, a membrane mimicking reagent, on the conformation of GAP peptide were studied. As shown in Fig. 6, the addition of TFE caused a concentration dependent induction of a CD spectrum typical for alpha -helix with two negative peaks at 222 and 208 nm, a cross-over near 200 nm, and a maximum near 192 nm (35). The alpha -helical content reached almost 100% in 40% TFE (Fig. 6, inset). Therefore the secondary structure of GAP peptide is converted from random coli to almost pure alpha -helix in the presence of TFE. Structural characteristics of GAP peptide were further analyzed by two-dimensional 1H NMR. To reduce the amide exchange rate and simultaneously increase the correlation time of the peptide, most of the spectra were obtained at acidic pH (Fig. 7a). The assignment of the NMR resonances was obtained by following standard procedures for two-dimensional 1H NMR of proteins (36). Fig. 8a summarizes the short and medium range NOE cross-peaks observed with GAP peptide. A substantial number of alpha beta (i, i+3) and alpha N (i, i+3) NOE cross-peaks were observed with GAP peptide in 50% H2O, 10% D2O, 40% TFE-d3 solution, suggesting that GAP peptide assumes an alpha -helical structure in the presence of TFE. Furthermore, comparison of the chemical shifts of the alpha  protons observed with GAP peptide to those obtained with peptides in random coil (37) shows upfield shifts characteristic for alpha -helix (38, 39) as shown in Fig. 8b. Altogether, these results established that GAP peptide, especially about two-thirds of the N-terminal side of the peptide, assumes a regular alpha -helix in the presence of TFE.


Fig. 6. Effects of TFE on CD spectra of GAP peptide. CD spectra of GAP peptide (20 µM) were obtained in the presence of varying concentrations of TFE (v/v). bullet , 0% TFE; open circle , 10% TFE; ×, 20% TFE; black-square, 30% TFE; square , 40% TFE; black-triangle, 50% TFE. CD spectral change at 222 nm and content of alpha  helix calculated as described under "Experimental Procedures" were plotted against TFE concentration (inset). bullet , change at 222 nm; open circle , alpha -helix content.
[View Larger Version of this Image (29K GIF file)]



Fig. 7. alpha H(F1)/NH(F2) region of 500 MHz DQF-COSY spectrum of GAP peptide in TFE and in aqueous solution. NMR spectra of GAP peptide were obtained in 50% H2O, 10% D2O, 40% TFE-d3 solution (a) or in 90% H2O, 10% D2O solution (b). Concentrations of GAP peptide were 5 mM. Sequence-specific assignments of the alpha H/NH cross-peaks are indicated. Regions where Lys, Arg, Gln, Ser, His, Phe, and Gly in random coil are observed are indicated.
[View Larger Version of this Image (32K GIF file)]



Fig. 8. Summary of the NMR studies obtained with GAP peptide in TFE. NOE connectivities of specified proton pairs that were observed are marked with bars (a). The thick and thin bars indicate the intensity (medium and weak, respectively) of the NOEs observed in the NOESY spectra. Deviation of the chemical shifts of alpha  protons from those observed with typical random coil (36, 37) are plotted (b). Negative values correspond to higher magnetic field shifts.
[View Larger Version of this Image (41K GIF file)]


Conformation of GAP Peptide in Aqueous Solution

In the absence of TFE, many cross-peaks in the NMR spectra could not be assigned due to resonance overlaps. However, it was possible to assign several residues from their unique spin systems. Since Ala2, Ile8, Thr9, and Leu13 occurs only once in the peptide, their well resolved methyl group signals in higher magnetic field region were used to assign peaks belonging to these residues. Interestingly, the chemical sifts of all the assigned residues showed intermediate values between those typical for random coil (37) and those for alpha -helix. Furthermore, the overlapped signals of cross-peaks between alpha  protons and amide protons as a whole showed a tendency of the upfield shift (Fig. 7b). Again, most of the signals showed intermediate chemical shifts between those observed with alpha -helical and random coil peptides. These results suggest that the GAP peptide in aqueous solution is not in a "pure" random coil, but assumes an intermediate state between the random coil and the alpha -helix.

Effects of Phosphorylation on GAP-43-membrane Interaction

To examine effects of PKC-dependent phosphorylation of GAP-43 on the GAP-43-membrane interaction, GAP-43 and GAP peptide were phosphorylated by PKC and purified as described under "Experimental Procedures." The basic domain of GAP-43 contains only one phosphorylatable residue (Ser41) and the phosphorylation of the protein prevents its binding to calmodulin (7). Phosphorylated GAP-43 and phosphorylated GAP peptide showed CD spectra similar to those of non-phosphorylated species. When acidic phospholipids such as PG was added, no significant change in the CD spectra was observed with both the phosphorylated protein and the peptide (data not shown), suggesting that the phosphorylation by PKC reduced their affinities to the phospholipids. Phosphorylation of GAP peptide has already been shown to affect the interaction (14). It is easily conceivable that the incorporation of negative charges reduces the net positive charge of the basic domain, which should affect the electrostatic interaction involved in the binding.

Although the CD spectrum of the phosphorylated peptide was very similar to that of the non-phosphorylated peptide in aqueous solution, those obtained in the presence of TFE differed significantly. Even in the presence of 50% TFE, the phosphorylated peptide showed an intermediate state between the random and alpha -helical structures (Fig. 9), similar to that observed with the non-phosphorylated peptide in 20% TFE (Fig. 6). These results suggest that phosphorylation of the Ser41, which is located near the N terminus of the peptide, not only changes the net charge of the peptide, but also impairs significantly the ability of the peptide to form the alpha -helical structure.


Fig. 9. Effects of TFE on CD spectra of phosphorylated GAP peptide. GAP peptide was phosphorylated by PKC, and purified over an HPLC reversed-phase column as described under "Experimental Procedures." CD spectra of phosphorylated GAP peptide (20 µM) were measured in the absence (bullet ) or presence (open circle ) of 50% TFE.
[View Larger Version of this Image (18K GIF file)]


Effects of Calmodulin on GAP-43-membrane Interaction

Since the same domain shows the ability to bind both acidic phospholipids and calmodulin, it is of interest to examine the effects of calmodulin on the GAP-43-membrane interactions. As shown in Fig. 10, the addition of calmodulin to the membrane-bound GAP-43 reversed the binding, and a part of the protein was found in the supernatant. In the presence of 0.15 M KCl, the amounts of GAP-43 remained in the supernatants were similar regardless of Ca2+, reflecting the fact that GAP-43 has similar binding constants to both Ca2+-free apo-calmodulin and Ca2+-calmodulin at this ionic strength (40). However, at a low ionic strength, where GAP-43 shows higher affinity to apo-calmodulin than to Ca2+-calmodulin (40), the amounts of GAP-43 remained in supernatant were higher in the absence of Ca2+ than in the presence of Ca2+ (Fig. 10b). Therefore, the cellular concentration of calmodulin available for binding can regulate the GAP-43-membrane interactions. It should be noted that the addition of Ca2+ also affected the GAP-43-phospholipid interaction, although the effects seems to be less significant compared with those observed with calmodulin under the experimental conditions employed. This is probably due to the direct binding of Ca2+ to PS, which has been well documented. These results suggest that the GAP-43-membrane interaction is regulated in a very complex manner including the PKC-dependent and the calmodulin-dependent pathways.


Fig. 10. Effects of calmodulin on GAP-43-phospholipid interaction. GAP-43 (1 µM) was mixed with PS liposomes (1 mM), and incubated at 25 °C for 30 min in the presence of 0.5 mM Ca2+ or 2 mM EGTA. The solutions were divided, and calmodulin (CaM) (35 µM) was added to samples indicated. The protein remained in supernatant after centrifugation was analyzed as described in the legend to Fig. 2. The experiments were carried out in 5 mM phosphate buffer (pH 7.3) containing 150 mM KCl (a) or in 20 mM phosphate buffer containing no additional salt (b). The relative amounts of GAP-43 remained in supernatants were determined by densitometry to be 4.7% (+PS), 11.4% (+PS +Ca2+), 60.9% (+PS +calmodulin), and 61.1% (+PS +calmodulin + Ca2+) at 150 mM KCl. The corresponding values obtained at low ionic strength were 3.9% (+PS), 8.9% (+PS +Ca2+), 35.1% (+PS +calmodulin), and 19.1% (+PS +calmodulin + Ca2+).
[View Larger Version of this Image (39K GIF file)]



DISCUSSION

Although GAP-43 has been reported to be palmitoylated and the acylation has been believed to be responsible for the membrane localization of the protein, our previous (3) and present studies established that GAP-43 purified from membrane fractions is not palmitoylated but still retains the ability to bind to phospholipid membranes. Since GAP-43 is a very hydrophilic protein without any clear membrane-binding domain, this seems to be puzzling. However, we have already shown that myristoylated alanine-rich protein kinase C substrate, which itself is a major PKC substrate protein and shows stimulation-dependent reversible membrane binding, interacts with phospholipid membranes through the PKC phosphorylation domain of basic amphiphilic nature (16). A similar observation has been reported with the GAP-43 calmodulin-binding domain (14). The calmodulin-binding domain of nitric oxide synthase, too, seems to serve as a membrane-binding domain (41).2 Although the calmodulin-binding motif has no clear conserved amino acid sequence, basic hydrophilic and hydrophobic amino acids appear alternately at certain intervals. When it assumes an alpha -helical structure, the two groups of amino acids segregate on opposite sides of the helices (42, 43). The calmodulin-binding domain of GAP-43, which is at the same time the phosphorylation domain by PKC, belongs to the same phosphorylation-dependent membrane binding motif.

As shown in the present study, both the intact protein and the phosphorylation domain of GAP-43 bind to phospholipid membranes and adopt alpha  helical structures under hydrophobic environments. The degree of change in the CD spectra of the whole protein induced by the phospholipid binding is similar to that observed with GAP peptide, suggesting that upon phospholipid binding only the phosphorylation domain undergoes a conformational change while the conformation of the rest of the molecule remains unchanged. It seems reasonable to assume that only the phosphorylation domain of GAP-43 interacts with phospholipid liposomes, since the whole protein except for the domain is very hydrophilic and acidic.

Compared with conventional CD measurements, two-dimensional 1H NMR studies gave more accurate and residue-specific information on the conformation. Most parts of the peptide, especially the N-terminal two-thirds, formed a regular alpha -helix in the presence of TFE, as was evidenced by the consecutive NOE connectivities and the characteristic upfield shift of alpha  protons. The C-terminal side of the peptide, on the other hand, showed less tendency to form an alpha -helix. The advantage to use the NMR technique was more pronounced in the analysis of the conformation of the peptide in aqueous solution. The CD spectrum of GAP peptide in aqueous solution showed no indication of the presence of any regular structure. On the other hand, chemical shifts of most of the alpha  protons obtained in aqueous solution showed intermediate values between those observable with alpha -helix and those seen with "true" random coil. So-called random coil can be considered to be a set of many different conformations, within which rapid exchanges occur. While the CD spectroscopy does not indicate the nature of the conformations included in such sets of the conformations, the NMR studies suggest that a significant portion of the peptide molecules adopts an alpha -helical conformation. Such a nascent helical structure may deviate from ideal geometry, and/or the ends of the alpha -helix can fray (44, 45). The interaction of GAP peptide with phospholipids seems to stabilize the conformation to induce an alpha -helical structure. Several calmodulin-binding peptides have been reported to form such a nascent alpha -helical structure in aqueous solution. Such a structure is usually further stabilized either by addition of TFE (46, 47) or by binding to calmodulin (48, 49). It should be noted that the shape of the CD spectrum obtained in the presence of phospholipids was not the same as that observed in TFE, suggesting that the conformation of GAP peptide bound to phospholipids is not a simple alpha -helix. Since the extents of the change observed upon phospholipid binding were also smaller than those observed in the presence of TFE, it is possible that only a part of the domain undergoes such a conformational change. A more detailed study of the interaction between phospholipid liposomes and the GAP-43 phosphorylation domain by NMR is now in progress.

The calmodulin-binding domain of GAP-43 interacts with membrane phospholipids, and the phosphorylation by PKC reduces the interaction. The phosphorylation seems to affect the interaction not only by reducing the positive charge of the domain but also by impairing the ability of the domain to form an alpha -helical structure (Fig. 9). It is reasonable to assume that the same mechanism is operational in the regulation of the binding of the domain to calmodulin by PKC-dependent phosphorylation (7). Since the binding of calmodulin, in turn, affects the interaction of the domain with PKC and phospholipids, physiological functions of GAP-43, whatever it is, are regulated in a very complex manner which involves various components of the signal transduction pathways (Fig. 11). The domain of basic amphiphilic alpha -helical nature may function as one of the cross-talk points in the calcium-dependent signal transduction.


Fig. 11. Schematic model of GAP-43-membrane interaction and its regulation by PKC and calmodulin. CaM, calmodulin.
[View Larger Version of this Image (11K GIF file)]



FOOTNOTES

*   This work was supported in part by Grants-in-Aid from the Fujita Health University, a Science Research Promotion Fund from the Japan Private School Promotion Foundation, a Research Grant from the Naito Foundation for Medical Research, a Grant-in-Aid for Scientific Research (C) (06680773) and Grants-in-Aid for Scientific Research on Priority Areas (06253218, 06276218, 07268221, 07279242, 08249240, and 08260220) from the Ministry of Education, Science, Sports and Culture, Japan.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
   Research Fellow of the Japan Society of the Promotion of Science.
§   To whom all correspondence should be addressed. Tel.: 81-562-93-9381; Fax.: 81-562-93-8832; E-mail: htanigut{at}fujita-hu.ac.jp.
1   The abbreviations used are: GAP-43, growth-associated protein-43; GAP peptide, peptide corresponding to the phosphorylation and calmodulin-binding domain of GAP-43; TFE, trifluoroethanol; PKC, protein kinase C; CD, circular dichroism; PG, phosphatidylglycerol; PC, phosphatidylcholine; PS, phosphatidylserine; NOE, nuclear Overhauser enhancement.
2   M. Matsubara, K. Titani, and H. Taniguchi, manuscript in preparation.

Acknowledgment

We thank M. Suzuki for technical assistance.


REFERENCES

  1. Coggins, P. J., and Zwiers, H. (1991) J. Neurochem. 56, 1095-1106 [Medline] [Order article via Infotrieve]
  2. Skene, J. H. P. (1989) Annu. Rev. Neurosci. 12, 127-156 [CrossRef][Medline] [Order article via Infotrieve]
  3. Taniguchi, H., Suzuki, M., Manenti, S., and Titani, K. (1994) J. Biol. Chem. 269, 22481-22484 [Abstract/Free Full Text]
  4. Spencer, S. A., Schuh, S. M., Liu, W.-S., and Willard, M. B. (1992) J. Biol. Chem. 267, 9059-9064 [Abstract/Free Full Text]
  5. Pisano, M. R., Hegazy, M. G., Reimann, E. M., and Dokas, L. A. (1988) Biochem. Biophys. Res. Commun. 155, 1207-1212 [Medline] [Order article via Infotrieve]
  6. Apel, E. D., Litchfield, D. W., Clark, R. H., Krebs, E. G., and Storm, D. R. (1991) J. Biol. Chem. 266, 10544-10551 [Abstract/Free Full Text]
  7. Liu, Y., and Storm, D. R. (1990) Trends Pharmacol. Sci. 11, 107-111 [CrossRef][Medline] [Order article via Infotrieve]
  8. Strittmatter, S. M., Valenzuela, D., Kennedy, T. E., Neer, E. J., and Fishman, M. C. (1990) Nature 344, 836-841 [CrossRef][Medline] [Order article via Infotrieve]
  9. Meiri, K. F., and Gordon-Weeks, P. R. (1990) J. Neurosci. 10, 256-266 [Abstract]
  10. LaBate, M. E., and Skene, J. H. P. (1989) Neuron 3, 299-310 [Medline] [Order article via Infotrieve]
  11. Skene, J. H. P., and Virág, I. (1989) J. Cell Biol. 108, 613-624 [Abstract]
  12. Zuber, M. X., Strittmatter, S. M., and Fishman, M. C. (1989) Nature 341, 345-348 [CrossRef][Medline] [Order article via Infotrieve]
  13. Houbre, D., Duportail, G., Deloulme, J.-C., and Baudier, J. (1991) J. Biol. Chem. 266, 7121-7131 [Abstract/Free Full Text]
  14. Kim, J., Blackshear, P. J., Johnson, J. D., and McLaughlin, S. (1994) Biophys. J. 67, 227-237 [Abstract]
  15. Blackshear, P. J. (1993) J. Biol. Chem. 268, 1501-1504 [Free Full Text]
  16. Taniguchi, H., and Manenti, S. (1993) J. Biol. Chem. 268, 9960-9963 [Abstract/Free Full Text]
  17. Kim, J., Shishido, T., Jiang, X., Aderem, A., and McLaughlin, S. (1994) J. Biol. Chem. 269, 28214-28219 [Abstract/Free Full Text]
  18. Sigal, C. T., Zhou, W., Buser, C. A., McLaughlin, S., and Resh, M. D. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 12253-12257 [Abstract/Free Full Text]
  19. Resh, M. D. (1994) Cell 6, 411-413
  20. Manenti, S., Sorokine, O., Van Dorsselaer, A., and Taniguchi, H. (1992) J. Biol. Chem. 267, 22310-22315 [Abstract/Free Full Text]
  21. Taniguchi, H., Manenti, S., Suzuki, M., and Titani, K. (1994) J. Biol. Chem. 269, 18299-18302 [Abstract/Free Full Text]
  22. Provencher, S. W., and Glockner, J. (1981) Biochemistry 20, 33-37 [Medline] [Order article via Infotrieve]
  23. Bax, A., and Davies, D. G. (1985) J. Magn. Reson. 65, 393-402
  24. Jeener, J., Meier, B. H., Bachman, P., and Ernst, R. R. (1979) J. Chem. Phys. 71, 4546-4553 [CrossRef]
  25. Macura, S., Hyang, Y., Suter, D., and Ernst, R. R. (1981) J. Magn. Reson. 43, 259-281
  26. Rance, M., Sorensen, O. W., Bodenhausen, G., Wagner, G., Ernst, R. R., and Wüthrich, K. (1983) Biochem. Biophys. Res. Commun. 177, 479-485
  27. Baker, P., and Freeman, R. (1985) J. Magn. Reson. 64, 334-340
  28. Hurd, R. E. (1990) J. Magn. Reson. 87, 422-428
  29. Piotto, M., Saudek, V., and Sklenar, V. (1992) J. Biomol. NMR 2, 661-665 [Medline] [Order article via Infotrieve]
  30. Sklenar, V., Piotto, M., Leppik, R., and Saudek, V. (1993) J. Magn. Reson. 102, 241-245 [CrossRef]
  31. Wiseman, T., Williston, S., Brandts, J. F., and Lin, L. (1989) Anal. Biochem. 179, 131-137 [Medline] [Order article via Infotrieve]
  32. Beschiaschvili, G., and Seelig, J. (1992) Biochemistry 31, 10044-10053 [Medline] [Order article via Infotrieve]
  33. Biemann, K. (1990) Methods Enzymol. 193, 455-479 [Medline] [Order article via Infotrieve]
  34. McLaughlin, S. (1989) Annu. Rev. Biophys. Biophys. Chem. 18, 113-136 [CrossRef][Medline] [Order article via Infotrieve]
  35. Greenfield, N., and Fasman, G. D. (1969) Biochemistry 8, 4108-4116 [Medline] [Order article via Infotrieve]
  36. Wüthrich, K. (1986) NMR of Proteins and Nucleic Acids, J. Wiley, New York
  37. Bundi, A., and Wüthrich, K. (1979) Biopolymers 18, 285-298
  38. Pastore, A., and Saude, V. (1990) J. Magn. Reson. 90, 165-176
  39. Wishart, D., Sykes, B., and Richards, F. (1991) J. Mol. Biol. 222, 311-333 [Medline] [Order article via Infotrieve]
  40. Alexander, K. A., Cimler, B. M., Meier, K. E., and Storm, D. R. (1987) J. Biol. Chem. 262, 6108-6113 [Abstract/Free Full Text]
  41. Venema, R. C., Sayegh, H. S., Arnal, J.-F., and Harrison, D. G. (1995) J. Biol. Chem. 270, 14705-14711 [Abstract/Free Full Text]
  42. O'Neil, K. T., and DeGrado, W. F. (1990) Trends Biochem. Sci. 15, 59-64 [CrossRef][Medline] [Order article via Infotrieve]
  43. Blumenthal, D. K., Takio, K., Edelman, A. M., Charbonneau, H., Titani, K., Walsh, K., and Krebs, E. G. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 3147-3149
  44. Dyson, H. J., Merutka, J., Waltho, J. P., Lerner, R. A., and Wright, P. E. (1992) J. Mol. Biol. 226, 795-817 [Medline] [Order article via Infotrieve]
  45. Manning, M. C., Illangasekare, M., and Woody, R. W. (1988) Biophys. Chem. 31, 77-86 [CrossRef][Medline] [Order article via Infotrieve]
  46. Munier, H., Blanco, F. J., Precheur, B., Diesis, E., Nieto, J. L., Craescu, C. T., and Barzu, O. (1993) J. Biol. Chem. 268, 1695-1701 [Abstract/Free Full Text]
  47. Zhang, M., and Vogel, H. J. (1994) J. Biol. Chem. 269, 981-985 [Abstract/Free Full Text]
  48. Ikura, M., Clore, G. M., Gronenborn, A. M., Zhu, G., Klee, C. B., and Bax, A. (1992) Science 256, 632-638 [Medline] [Order article via Infotrieve]
  49. Gerendasy, D. D., Herron, S. R., Jennings, P. A., and Sutcliffe, J. G. (1995) J. Biol. Chem. 270, 6741-6750 [Abstract/Free Full Text]
  50. Biemann, K. (1990) Methods Enzymol. 193, 886-887 [Medline] [Order article via Infotrieve]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.