Myristoylation-regulated Direct Interaction Between Calcium-bound Calmodulin and N-terminal Region of pp60v-src

doi:10.1016/j.jmb.2004.02.041

J. Mol. Biol. (2004) 338, 169–180

Myristoylation-regulated Direct Interaction Between Calcium-bound Calmodulin and N-terminal Region of pp60v-src Nobuhiro Hayashi1*, Chisako Nakagawa1, Yutaka Ito2 Akihiko Takasaki3, Yuji Jinbo4, Yoshinori Yamakawa5, Koiti Titani1 Keiichiro Hashimoto1, Yoshinobu Izumi4 and Norio Matsushima6 1 Division of Biomedical Polymer Science, Institute for Comprehensive Medical Science Fujita Health University Toyoake, Aichi 470-1192, Japan 2 Cellular and Molecular Biology Laboratory, RIKEN Yokohama 230-0045, Japan 3 School of Health Science Fujita Health University Toyoake, Aichi 470-1192, Japan 4

Graduate School of Engineering, Yamagata University, Yonezawa 992-8510, Japan 5 Equipment Center for Research and Education, Nagoya University School of Medicine Nagoya 466-8550, Japan 6

School of Health Sciences Sapporo Medical University S-1, W-17, Sapporo 060-8556 Japan *Corresponding author

pp60v-src tyrosine protein kinase was suggested to interact with Ca2þbound calmodulin (Ca2þ/CaM) through the N-terminal region based on its structural similarities to CAP-23/NAP-22, a myristoylated neuronspecific protein, whose myristoyl group is essential for interaction with Ca2þ/CaM; (1) the N terminus of pp60v-src is myristoylated like CAP-23/ NAP-22; (2) both lysine residues are required for the myristoylationdependent interaction and serine residues that are thought to regulate the interaction through the phosphorylations located in the N-terminal region of pp60v-src. To verify this possibility, we investigated the direct interaction between pp60v-src and Ca2þ/CaM using a myristoylated peptide corresponding to the N-terminal region of pp60v-src. The binding assay indicated that only the myristoylated peptide binds to Ca2þ/CaM, and the non-myristoylated peptide is not able to bind to Ca2þ/CaM. Analyses of the binding kinetics revealed two independent reactions with the dissociation constants (KD) of 2.07 £ 1029 M (KD1) and 3.93 £ 1026 M (KD2), respectively. Two serine residues near the myristoyl moiety of the peptide (Ser2, Ser11) were phosphorylated by protein kinase C in vitro, and the phosphorylation drastically reduced the interaction. NMR experiments indicated that two molecules of the myristoylated peptide were bound around the hydrophobic clefts of a Ca2þ/CaM molecule. The small-angle X-ray scattering analyses showed that the size of the peptide – Ca2þ/CaM ˚ smaller than that of the known Ca2þ/CaM – target molcomplex is 2 –3 A ecule complexes. These results demonstrate clearly the direct interaction between pp60v-src and Ca2þ/CaM in a novel manner different from that of known Ca2þ/CaM, the target molecules, interactions. q 2004 Elsevier Ltd. All rights reserved.

Keywords: myristoylation; calmodulin; pp60v-src; small-angle X-ray scattering; protein –protein interaction

Abbreviations used: CaM, calmodulin; Ca2þ/CaM, Ca2þ-bound CaM; vSrc, peptide, myristoylated peptides corresponding to the N-terminal region of pp60v-src, GSSKSKPKDPSQRRRSLE; myr-vSrc, peptide, myristoylated form of vSrc peptide; SAXS, small angle X-ray scattering; NMR, nuclear magnetic resonance spectroscopy; mC/N9, a myristoylated peptide corresponding to the N-terminal CaM-binding site of CAP-23/NAP-22; Na, a-amino; MAPK, mitogenactivated protein kinase; HSQC, heteronuclear single quantum coherence; MLCK, myosin light chain kinase; M13, a peptide based on the CaM-binding domain of MLCK; PGD, petunia glutamate decarboxylase. E-mail address of the corresponding author: [email protected]

Introduction The v-src is the Rous sarcoma virus oncogene, which is a homologue of the normal cellular gene of the pp60c-src tyrosine protein kinase (c-src), and encodes the pp60v-src tyrosine protein kinase. Although pp60c-src and pp60v-src have been the most extensively studied among tyrosine protein kinases, their physiological role still remains largely unknown. One of their common structural features is the Na-myristoylation. Both of them are the prototype members of the family of tyrosine protein kinases and are tightly associated with cellular membranes, and the main role of their

0022-2836/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.

170

N-terminal myristoyl groups has been so far assumed to be involved in protein –membrane interactions.1 – 4 However, as an anchor to cellular membranes, other functions have been presumed as the other roles of the myristoylation because the transforming activity of pp60v-src is dependent on its myristoylation.5,6 Nevertheless, it has been shown that without their myristoyl groups, pp60c-src and pp60v-src retain their activities of tyrosine protein kinase.7 Calmodulin (CaM)1 is a small calcium-binding protein (16.7 kDa) involved in a wide range of cellular Ca2þ-dependent signaling pathways through various enzymes, including protein kinases, protein phosphatases, nitric oxide synthase, inositol triphosphate kinase, nicotinamide adenine dinucleotide kinase, and cyclic nucleotide phosphodiesterase.8 We have recently found a novel mechanism of target recognition by CaM. It has been demonstrated that CAP-23/ NAP-22 isolated from rat brain is Na-myristoylated, and this modification is involved in its interaction with CaM in the presence of Ca2þ.9 The CaM-binding site was narrowed down to the myristoyl moiety together with the N-terminal basic domain of nine amino acid residues, GGKL SKKKK. The basic Na-myristoylated peptide corresponding to the N-terminal CaM-binding site of CAP-23/NAP-22 (mC/N9) has clearly different properties from known CaM-binding peptides. The most striking feature is that the binding of mC/N9 to CaM is dependent on the presence of the myristoyl moiety, whose general function so far has been assumed to be the membrane targeting of myristoylated proteins. Recently, a smallangle X-ray scattering (SAXS) study revealed that large structural changes of Ca2þ-bound calmodulin (Ca2þ/CaM) occur by the binding of mC/N9 to CaM.10 Moreover, the interaction between CAP23/NAP-22 and CaM is controlled by phosphorylation of serine residue at the N terminus. From these observations, we identified that mC/N9 is a new CaM-binding motif, differing from the sequence motifs of CaM recognition such as motifs termed 1-8-14 and 1-5-10 based on the position of conserved hydrophobic residues or IQ motifs seen in many proteins.11 Moreover, from the N-terminal amino acid sequence alignment of CAP-23/NAP-22 and other myristoylated proteins, we hypothesized that some myristoylated proteins, including pp60v-src, bind to CaM through the N-terminal domain.10 It has already been shown that one of the candidates, HIV nef, binds to CaM through its N-terminal myristoyl moiety.12 pp60v-src has a sequence of GSSKSKPKDPSQRRRSLE at the N terminus, which, in addition to the myristoylation, has some functionally important features similar to those of CAP-23/NAP-22 (GGKLSKKKK), the basic residues (lysine), and the target residue of protein kinase C (PKC) (serine).10,12 Thus, we hypothesized that there is a strong possibility that pp60v-src interacts directly with CaM, as observed in cases of

SAXS and NMR of CaM –pp60v-src Peptide Complex

CAP-23/NAP-22 and HIV nef. As noted, the SAXS study is a useful method to detect direct interaction between target peptides and CaM through measurable structural changes of CaM.13 Here, in order to investigate the interaction, we performed structural studies of Ca2þ/CaM in the presence of a myristoylated N-terminal peptide (myr-GSSKSKPKDPSQRRRSLE; vSrc peptide) of pp60v-src using nuclear magnetic resonance spectroscopy (NMR) and SAXS. The NMR and SAXS studies revealed that structural changes of Ca2þ/ CaM occurred in the presence of the vSrc peptide, indicating clearly that Ca2þ/CaM bound to the myristoylated peptide.

Results Myristoylation-dependent binding of the N-terminal peptides to CaM To study the interaction of pp60v-src tyrosine protein kinase with CaM, peptides were synthesized based on the N-terminal sequence of pp60v-src. Their binding to CaM was assessed using CaMagarose as described below. The longest peptide synthesized containing the N-terminal 18 residues (GSSKSKPKDPSQRRRSLE) and the shortest peptide of ten amino acid residues (GSSKSKPKDP) showed a similar Ca2þ-dependent binding to CaM. Because results for the shortest peptide were similar to those for the longest peptide and those of CAP-23/NAP-22 as described,9 results for the longest peptide, the vSrc peptide, are mainly shown here. Myristoylated vSrc peptide or non-myristoylated vSrc peptide was mixed with CaM beads in the presence or absence of Ca2þ, and the bound peptides were eluted from the beads with EGTA and Ca2þ, respectively. As shown in Figure 1(a), myristoylated vSrc peptide bound to the CaMagarose, and the bound peptides were eluted with the Ca2þ-free buffer. Therefore, the binding of the N-terminal region of pp60v-src to CaM is Ca2þdependent. In contrast, non-myristoylated vSrc peptide did not bind to the CaM beads under the same conditions. There seems to be an absolute requirement for N-terminal myristoylation in the N-terminal region of pp60v-src –CaM interaction. PKC-dependent phosphorylation of N-terminal domain of pp60v-src Since the N-terminal myristoylated domain has a serine residue (Ser3) with the motif of PKC-dependent phosphorylation ([Ser/Thr]-x-[Arg/Lys]),14 and, in the case of CAP-23/NAP-22, the phosphorylation significantly affects the myristoylation-dependent CAP-23/NAP-22–CaM interaction,9 phosphorylation of the N-terminal myristoylated domain of pp60v-src was examined. Myristoylated vSrc peptide was treated with PKC, desalted by passage through a pipette tip containing a small

171

SAXS and NMR of CaM –pp60v-src Peptide Complex

Figure 1. Effects of myristoylation and phosphorylation on the interaction with CaM. (a) Binding of the non-myristoylated and myristoylated N-terminal peptides, and (b) myristoylated, phosphorylated N-terminal peptides were assessed using CaM-agarose in the presence (þ Ca) or absence (2 Ca) of calcium ions. Total fraction (lane 1, þCa; lane 4, 2Ca), unbound fraction (lane 2, þ Ca; lane 5, 2 Ca), and bound fraction eluted by EGTA (lane 3, þ Ca; lane 6, 2Ca) are shown. MyrvSrc peptide phosphorylated at Ser2, at Ser11, and at both Ser2 and Ser11 are abbreviated as myr-Ser2P, myrSer11P, and myr-Ser2P-Ser11P, respectively.

C18 microcolumn (Ziptip), and characterized by matrix-assisted laser desorption ionization timeof-flight (MALDI-TOF) mass spectrometry (Figure 2). The presence of doubly phosphorylated species indicated that vSrc peptide contained two phosphorylation sites. To determine the phosphorylation sites in myristoylated vSrc peptide, the peptide was subjected to digestion with trypsin. The digestion yielded four peptides, GSSK (G1-K4), SKPK (S5-K8), DPSQR (D9-R13), SLE (S16-E18), and two arginine residues (R14, R15). The MALDI-TOF mass spectrometric analyses showed that G1-K4 and S5-K8 were found to be singly phosphorylated, suggesting that each peptide contained one phosphorylation site. Because S5-K8 has only one serine residue (Ser5), the residue is assumed to be phosphorylated. However, because G1-K4 has two serine residues, the precise site could not be identified by MALDI-TOF mass spectrometric analyses. To identify the phosphorylation site in the G1-K4 peptide, the peptide was subjected to tandem mass spectral (MS/MS) analysis (Figure 3). Although the region contained two phosphorylatable residues (Ser2, Ser3), the MS/MS spectra clearly identified the second Ser in the peptide (Ser2) as the phosphorylation site. Furthermore, the other phosphorylatable residues (Ser11) were identified as the other phosphorylation sites in the G1-E18 peptide using the same method.

Figure 2. Analysis of the phosphorylated and myristoylated vSrc peptide. Myristoylated vSrc peptide was treated with PKC, purified by the pipette tip containing a small C18 microcolumn (Ziptip), and characterized by MALDI-TOF mass spectrometry; myrvSrc peptide (up) and PKC phosphorylated myr-vSrc peptide (down).

Effects of phosphorylation on CaM binding We have examined whether the phosphorylation of the N-terminal myristoylated domain of pp60v-src affects the binding to CaM. Phosphorylated and myristoylated peptide, myr-Ser2P-vSrc peptide (phosphorylated at Ser2), myr-Ser11P-vSrc peptide (phosphorylated at Ser11), and myr-Ser2P-Ser11PvSrc peptide (phosphorylated at Ser2 and Ser11) were chemically synthesized. The binding assay revealed that the phosphorylation abolished the binding of the peptide to CaM significantly (Figure 1(b)). Interestingly, binding to CaM can be observed to some extent even in cases when the myristoylated peptide is phosphorylated. Phosphorylation of only one serine residue is not sufficient to prevent the myristoylated peptides from binding to CaM. On the other hand, addition of one more phospho group to the phosphorylated peptides almost restrained the interaction, and the importance of the number of the added phospho groups was confirmed by the result of the myr-Ser2P-Ser11PvSrc peptide.

172

SAXS and NMR of CaM –pp60v-src Peptide Complex

Figure 3. Tandem mass spectra of the myristoylated vSrc peptides. The singly charged ion of the G1-K4 peptide was subjected to the fragmentation in the ion trap mass analyzer. Fragment ions observed are indicated above the peptide sequence. The singly charged ion of the S5-K8, D9-R13 and S16-E18 peptides was also analyzed.

Analyses of binding kinetics and affinity A more quantitative analysis was carried out using a surface plasmon resonance (SPR) (Biacore) apparatus. While myristoylated vSrc peptide displayed reversible binding to CaM in the presence

of Ca2þ, no interaction was revealed with nonmyristoylated vSrc peptide. Analysis of the binding curves from several experimental series revealed two independent dissociation reactions with rate constants ðkd Þ of 1.14 £ 1025 s21 ðkd1 Þ and 6.19 £ 1021 s21 ðkd2 Þ; respectively. From the

Table 1. CaM binding to peptide variants

Non-myrg myrh myr-Ser2Pi myr-Ser11Pj myr-Ser2P-Ser11Pk

Ka1 (M21 s21)a

Kd1 (s21)b

KD1 (M)c

Ka2 (M21 s21)d

Kd2 (s21)e

KD2 (M)f

3.78 £ 102 5.48 £ 103 1.89 £ 104 1.67 £ 103 1.20 £ 102

1.04 £ 1022 1.14 £ 1025 1.11 £ 1021 1.98 £ 1022 5.52 £ 1022

2.76 £ 1025 2.07 £ 1029 5.88 £ 1026 1.18 £ 1025 4.59 £ 1024

ND 1.57 £ 105 ND ND ND

ND 6.19 £ 1021 ND ND ND

ND 3.93 £ 1026 ND ND ND

ND, not detected. a Binding rate constant of the first reaction. b Dissociation rate constant of the first reaction. c Dissociation constant of the first reaction. d Binding rate constant of the second reaction. e Dissociation rate constant of the second reaction. f Dissociation constant of the second reaction. g vSrc peptide. h myr-vSrc peptide. i myr-vSrc peptide phosphorylated at Ser2. j myr-vSrc peptide phosphorylated at Ser11. k myr-vSrc peptide phosphorylated at Ser2 and Ser11.

SAXS and NMR of CaM –pp60v-src Peptide Complex

173

Figure 4. Overlay of 1H – 15N HSQC spectra of Ca2þ/CaM alone (black) and the Ca2þ/CaM– vSrc peptide complex (red). The samples contained 0.5 mM CaM, 120 mM NaCl, 2.5 mM CaCl2, and 50 mM deuterated Tris – HCl (pH 7.5) in 90% H2O, 10% 2H2O. Cross-peaks showing larger shifts (Phe16, Ser17, Gly33, Ala57, Phe68, Ser101, Asp129, Phe141) are shown.

association rate constants ðka Þ (ka1 ¼ 5:48£ 103 M21 s21, ka2 ¼ 1:57 £ 105 M21 s21) and the dissociation rate constants, the dissociation constants ðKD Þ of 2.07 £ 1029 M (KD1) and 3.93 £ 1026 M (KD2) were calculated. These results suggest that the binding of first myristoylated peptide raises the association rate of the second binding. On the

other hand, non-myristoylated peptide showed no significant interaction (KD is 2.76 £ 1025 M). Furthermore, to elucidate the effects of the phosphorylation on the interaction, phosphorylated and myristoylated peptides were subjected to the analyses, and phosphorylations of the myristoylated peptides were found to reduce their affinities

Figure 5. The observed changes (Hz) in the 1H– 15N HSQC cross-peak position upon forming the Ca2þ/CaM– vSrc peptide complex. The changes in cross-peak positions were quantified by [(D15NHz)2 þ (D1HHz)2]1/2.

174

SAXS and NMR of CaM –pp60v-src Peptide Complex

Figure 6. Arrangements on the Connolly surface of the Ca2þ/CaM– M13 complex (Protein Data Bank accession number 1CDL) (both sides) of residues whose chemical shifts changed upon forming the Ca2þ/CaM– vSrc peptide complex. The white object is M13, which is shown to indicating the position of the bound target peptide. Red, orange, yellow, and green indicate the locations of amino acid residues whose [(D15NHz)2 þ (D1HHz)2]1/2 were over 65 (Phe16, Ser17, Gly33, Ala57, Phe68, Ser101, Asp129, and Phe141), between 65 and 55 (Ile27, Asp64, Glu67, Met109, Glu114, Lys115, Glu127, and Tyr138), between 55 and 45 (Lys21, Thr29, Val35, Glu83, Arg90, Val91, Gly113, Ala128, Ile130, and Lys148), and the others, respectively. In the lower row, the other residues which surround the contact surface of the Ca2þ/CaM– M13 complex are removed. The models were constructed and rendered on an IRIS Indigo 2 workstation (SGI) using Insight II (Molecular Simulations) and SYBYL/BASE software (Tripos, Inc.).

to Ca2þ/CaM drastically corresponding to the number of added phospho groups (Table 1). Nuclear magnetic resonance spectroscopy We have studied the interaction between the vSrc peptide and Ca2þ/CaM using two-dimensional

1

H – 15N heteronuclear single quantum coherence (HSQC) NMR spectroscopy. The NMR spectra for Ca2þ/CaM in the absence and presence of vSrc peptide are shown in Figure 4. When one molar equivalent of the vSrc peptide was added, some peaks were shifted slightly, and no significant shift was observed in the 1H – 15N HSQC NMR spectra

Figure 7. Guinier plots for the Ca2þ/CaM– vSrc peptide complex (Ca2þ/CaM– vSrc peptide ¼ 1:2): 1, 7.0 mg/ml; 2, 9.0 mg/ml; 3, 11.0 mg/ml; 4, 13.0 mg/ml; 5, 15.0 mg/ml.

175

SAXS and NMR of CaM –pp60v-src Peptide Complex

of Ca2þ/CaM. However, some drastic shifts of the peak were observed with the addition of two molar equivalents of the vSrc peptide (Figure 4). The residues that showed larger shifts contained Phe16, Ser17, Gly33, Ala57, Phe68, Ser101, Asp129, and Phe141 (Figures 4 and 5). As illustrated in Figure 6, these residues were primarily found in the hydrophobic clefts of the N and C-domain of CaM. Small-angle X-ray scattering

Figure 8. The radius of gyration, Rg ; for the Ca2þ/ CaM –vSrc peptide complex (Ca2þ/CaM– vSrc peptide ¼ 1:2).

Table 2. Radius of gyration Rg and maximum dimension dmax for Ca2þ/CaM– vSrc peptide complex ˚) Rg (A

˚) dmax (A

Reference

Ca /CaM–vSrc peptide Ca2þ/CaM–mC/N9a

20.0 ^ 0.3 9.8 ^ 0.3

60 50

This study 10

Ca2þ/CaMa Ca2þ/CaMa Ca2þ/CaM–M13a Ca2þ/CaM–W-7a

21.9 ^ 0.3 21.5 ^ 0.3 16.4 ^ 0.2 17.6 ^ 0.3

62 69 49 47

10 15 35 36

2þ

a

a Values at zero protein concentration obtained by SAXS experiment.

The Guinier plots for Ca2þ/CaM – vSrc peptide mixtures with the molar ratio of 1:2 at five protein concentrations are shown in Figure 7. Figure 8 shows Rg as a function of protein concentration. The Rg values of the Ca2þ/CaM – vSrc mixtures are given in Table 2. For comparison, Table 2 also contains Rg values for other Ca2þ/CaM complexes including Ca2þ/CaM –mC/N9 complex.10 The Rg ˚ is value for Ca2þ/CaM – vSrc complex 20.0(^ 0.3) A consistent with that of the Ca2þ/CaM – mC/N9 complex. Both Rg values are smaller than that for Ca2þ/CaM and larger than those Ca2þ/CaM complexes containing TFP, W-7, mastoparan, melittin, cyclosporin-A, substance P and synthetic peptides corresponding to the CaM-binding domains of myosin light chain kinase (MLCK) (M13), phosphorylase kinase (RhK5) and Ca2þ-pump (C24W). Figure 9 shows the pðrÞ functions for Ca2þ/CaM in the presence of the vSrc peptide at a molar ratio of 1:2 and Ca2þ/CaM alone.10 The pðrÞ function for ˚ the Ca2þ/CaM alone has a peak near 20 A (principally representing interatomic distances within each domain of Ca2þ/CaM) and a shoulder ˚ (mainly representing the interdomain at near 40 A distances).15 In contrast, in the CaM –vSrc peptide

Figure 9. Pair distance distribution function, pðrÞ; for the Ca2þ/CaM– vSrc peptide complex (Ca2þ/CaM– vSrc peptide ¼ 1:2). (W) the Ca2þ/CaM– vSrc peptide complex (Ca2þ/CaM –vSrc peptide ¼ 1:2); (A) Ca2þ-saturated CaM alone.10,13

176 ˚ but around complex, a peak appears not at 20 A ˚ . Moreover, the maximal pair distance ðdmax Þ 25 A of the Ca2þ/CaM – vSrc peptide complexes is ˚ smaller than that of Ca2þ/CaM in about 2 A isolation. Additionally, the pðrÞs of the Ca2þ/ CaM – vSrc peptide complexes are relatively symmetrical. These results of the Ca2þ/CaM – vSrc peptide complex correspond to those observed in other complexes including the Ca2þ/CaM – mC/ N9 complex. Thus, it can be concluded that the Ca2þ/CaM forms a complex with vSrc peptide. The Ca2þ/CaM complex with vSrc peptide adopts a unique and larger globular structure that is not similar to those of other known Ca2þ/CaM complexes.

Discussion Canonical CaM-binding motif A closer look at the N-terminal sequence shows that the domain contains a hydrophobic region (myristoyl moiety) in addition to the basic amino acid residues (Lys4, Lys6, Lys8). This is reminiscent of the canonical CaM-binding motif, a basic amphiphilic motif, in which basic hydrophilic residues and hydrophobic ones appear alternately, and this feature is also found in the N-terminal region of CAP-23/NAP-22 and HIV nef protein.9,10,12 From the results of this study and the previous results for CAP-23/NAP-22 and HIV nef protein, like the other myristoylated proteins, it is reasonable to assume that the two pp60v-src molecules interact with one CaM molecule. Direct protein –protein interaction between Ca21/CaM and vSrc peptide The present NMR and SAXS analyses demonstrated that the structure of Ca2þ/CaM changed drastically in the presence of the myristoylated N-terminal peptide of pp60v-src tyrosine protein kinase. Similarly, the binding of the myristoylated nonapeptide (mC/N9) to Ca2þ/CaM induced large structural changes in CaM.10 The radii of gyration ðRg Þ are consistent with each other (Table 2), and the pair distance distributions pðrÞ showed similar behaviors (Figure 9). The results indicate that the myristoylated vSrc peptides as well as mC/N9 can interact with Ca2þ/CaM. The involvement of the protein myristoylation in protein –protein interactions has been implied in various studies,16 – 18 but it has never been clearly demonstrated. Because protein myristoylation has been implicated in the regulation of various signal transduction proteins,19,20 and because, in addition to signal transduction proteins, there are many other potential myristoylated proteins whose myristoylations can be predicted from their homologous amino acid sequences (a database of myristoylated proteins is now under construction), there is a possibility that myristoylation-dependent

SAXS and NMR of CaM –pp60v-src Peptide Complex

protein– protein interaction plays important roles in some of these cases including pp60v-src tyrosine protein kinase.10 Myristoylated N-terminal peptide in pp60v-src tyrosine protein kinase as a novel Ca21/CaMbinding motif It is well known that there are two recognition motifs for calcium-dependent CaM interactions. The two motifs of CaM recognition are termed 1-8-14 and 1-5-10 based on the position of conserved hydrophobic residues.11 Among structures of Ca2þ/CaM, the target molecule complexes including these two motifs, those with the Ca2þ/ CaM-binding region of MLCK and CaMKII are available now. Also the structure of the Ca2þ/CaM complex with the Ca2þ/CaM-binding region of Ca2þ-CaM-dependent protein kinase kinase (CaMKK) has been solved.21 The Ca2þ/CaM-binding regions are characterized as a 1-15 motif. Their target peptides form 1:1 complexes with Ca2þ/ CaM and adopt an a-helix. Here, we indicate formation of the Ca2þ/CaM – vSrc peptide complex with the molar ratio of 1:2. Also Ca2þ/CaM –mC/N9 complex was suggested to form a 1:2 complex,10 and recombinant myristoylated NAP-22/CAP-23 has been found to form a Ca2þ/CaM – myr-NAP22/CAP-23 complex with the molar ratio of 1:2 (data will be published elsewhere). It is likely that mC/N9 adopts a nonhelical conformation even in the complex. As noted, since the mC/N9 and vSrc peptides are similar to each other, the vSrc peptide is considered to adopt a non-helical conformation in the complex. A 1:2 Ca2þ/CaM –peptide complex was observed in the C-terminal domain from petunia glutamate decarboxylase (PGD), and the peptide corresponding to the domain was shown to bind to Ca2þ/CaM, but unlike this study, the complex takes an a-helical conformation.22 The SPR analyses of this study and biophysical analyses of this and previous10,12 studies independently revealed two states of transitions following the interaction between the myr-vSrc peptide and Ca2þ/CaM. Together with the results of this study, it is revealed that when one mole of myr-vSrc peptide is bound to Ca2þ/CaM, the affinity is very high although the induced structural change of Ca2þ/ CaM is very little. On the other hand, when the second mole of the peptide binds to Ca2þ/CaM, Ca2þ/CaM gives rise to drastic structural change despite the lower affinity. As described below, phosphorylation of the peptide reduces the first step of the interaction. The radii of gyration ðRg Þ of the Ca2þ/CaM –vSrc peptide complex and the Ca2þ/CaM – mC/N9 complex are consistent with each other (Table 2), and their pair distance distributions pðrÞ showed similar behaviors (Figure 9). Thus, the final overall shape conformation appears to be similar to a “relaxed” globular structure proposed for the mC/N9–Ca2þ/CaM complex, presumably differing

177

SAXS and NMR of CaM –pp60v-src Peptide Complex

from the known compact globular structures of CaM induced by target peptides of MLCK23 and CaMKII.21 Interaction of vSrc peptide with the hydrophobic cleft in Ca21/CaM The chemical shifts of some residues located in the hydrophobic clefts of the N and C-domain of CaM changed upon forming the Ca2þ/CaM – vSrc peptide complex, suggesting that the binding site of the vSrc peptide is located in the hydrophobic clefts of Ca2þ/CaM (Figure 6). The residues whose chemical shifts changed drastically include Phe16, Ser17, Gly33, Ala57, Phe68, Ser101, Asp129, and Phe141 (Figure 4, colored red in Figure 6). Of these residues, Phe68 is located on the contact surface of the Ca2þ/CaM – MLCK peptide (M13) complex; this CaM residue makes contact directly with the target peptide (MLCK peptide).23 Ser17, ˚ from Asp129, and Phe141 are located within 5 A the target peptide. Phe16, Gly33, Ala57, and Ser101 do not make contact with the target on the Ca2þ/CaM –MLCK peptide complex. However, their neighboring residues are located on the contact surface of the MLCK complex, and they themselves are located near the contact surface. All residues whose chemical shifts were considerably changed (colored orange in Figure 6) are primarily located on and around the contact surface. These observations strongly suggest that the two vSrc peptide molecules bind to the hydrophobic clefts of CaM as seen in the M13 peptide. Other than at the binding site, however, the features of the binding are considered to be different from each other, because, among the residues whose chemical shifts considerably changed, the residue on the contact surface is only Phe68, and because other features (e.g. molar ratio and requirement of myristoylation) are distinctly different from each other. In addition to the case of these myristoylated proteins (CAP-23/NAP-22 and pp60v-src tyrosine protein kinase), an example in which not one but two peptides bind to Ca2þ/ CaM has been reported by Yuan & Vogel22 in relation to PGD. The role of protein Na-myristoylation Protein Na-myristoylation, one of the most common forms of protein fatty acylations, was first characterized in the catalytic subunit of cAMPdependent protein kinase24 and in the regulatory subunit of calcineurin.25 The target proteins are modified with myristate, a 14-carbon saturated fatty acid, and the enzymology of the myristoylation reaction is now well understood.26 Covalent attachment of fatty acids to proteins is now a widely recognized form of protein modification, and many fatty acylated proteins are found to play key roles in regulating cellular functions. It has so far been generally assumed that hydrophobic acyl groups including the myristoyl group

are involved in protein – membrane interactions. Thus, as to pp60v-src tyrosine protein kinase, one of the roles of the myristoylation was predicted to anchor the myristoylated proteins to the plasma membrane. However, interestingly, the presence of a myristoyl group is not sufficient for the stable membrane anchoring of the protein,27 and it is easy to imagine that the myristoylation has another role. We have demonstrated that CAP-23/ NAP-22 isolated from rat brain is Na-myristoylated, and this modification is involved in its interaction with CaM (CaM) in the presence of Ca2þ.9 Considering the results of this study, the other role of the myristoylation was found to make binding to CaM possible, and pp60v-src tyrosine protein kinase is also assumed to interact with CaM. Regulation of the myristoylation-dependent interaction by phosphorylation Like CAP-23/NAP-22,9 phosphorylation of the CaM-binding domain of pp60v-src reduced the interaction between the N-terminal domain of pp60v-src and CaM. In contrast, the myristoylation affects the phosphorylation of peptides by PKC (data not shown), and the protein myristoylation may also be involved in the protein –protein interaction between PKC and pp60v-src. Altogether, these results raise the possibility that the CaM binding of pp60v-src is regulated by the PKC phosphorylation, and that the domain is a cross-talk point between the intracellular signal transduction system regulated by CaM and that regulated by PKC. The result that the phosphorylation drastically reduced only the first step of the interaction is interesting from a biochemical point of view in terms of the regulation mechanism of intracellular signal transduction, and it also attracts the interest of structural biology in terms of the control mechanism of protein function by post-translational modifications. Furthermore, because the calmodulin-binding domains in some proteins are at the same time phosphorylation domains for PKC and because phosphorylation of the domains by PKC significantly reduces their ability to bind to membrane phospholipids and/or calmodulin,28 reversible translocation between membrane and cytosol in response to PKC-dependent phosphorylation of pp60v-src is strongly suggested. These regulatory functions are assumed to correlate with cell transformation induced by pp60v-src, but the exact mechanisms are not well understood. Functional implication of the interaction between pp60v-src tyrosine protein kinase and CaM Because, despite the inability of non-myristoylated pp60v-src to induce the transformation, pp60v-src retains high levels of kinase activity in vitro and in vivo, it was assumed that cellular substrates

178

involved in cell transformation are localized at or near the plasma membrane.29 Myristoylationdependent protein –protein interaction between pp60v-src and CaM also suggests a possibility that the transformation results from interference of pp60v-src in the intracellular signal transduction system regulated by CaM. We have demonstrated so far that phosphorylation of the CaM-binding domain of neuronal protein, GAP43, reduces its binding to CaM and to the plasma membrane.28 On the other hand, its binding to CaM is also assumed to reduce its binding to the plasma membrane. pp60v-src studied here also suggests the same regulation mechanism for association between the pp60v-src and plasma membrane. Although the interaction of pp60c-src, the homologue of pp60v-src, with CaM has been considered to be important in the mitogen-activated protein kinase (MAPK) cascade for the regulation of the intracellular signal transduction,30 – 32 the direct interaction between Src family tyrosine kinases and CaM has not so far been indicated. The findings here imply that, in the MAPK cascade, the calcium-mediated regulation of Src family tyrosine kinases proceeds through a CaMdependent mechanism. The interaction between CaM and pp60c-src may have no effect on the tyrosine kinase activity. CaM may regulate the function of pp60c-src not directly but indirectly through regulation of the translocation of pp60c-src to the plasma membrane.

SAXS and NMR of CaM –pp60v-src Peptide Complex

10 mg/ml of alpha-cyano-4-hydroxycinnamic acid (Sigma) in 0.1% (v/v) TFA, 50% (v/v) acetonitrile solution. The spectra were displayed and analyzed using the GRAMS-MS software, and calibration of the spectra was done using a calibration mixture 1 or 2 (PE Biosystems). MS/MS analysis was done with an electrospray ionization/ion trap quadrupole mass spectrometer (ThermoQuest ion trap LCQ). The trypsin digestion samples were directly analyzed by MS/MS. The spray voltage was set at 2.53 kV, and the sheath gas flow-rate was set at 0. The temperature of the heated capillary was set at 270 8C. To determine phosphorylation sites, the precursor ions were set at all trypsin-digested fragment masses, and all daughter ions generated from the precursor ions were analyzed regardless of the phosphorylation. Binding to CaM Bindings of the peptides to CaM were studied using CaM-agarose (Sigma). Peptides were mixed with CaMagarose in 50 mM Tris – HCl buffer (pH 6.8) containing 1 mM CaCl2 and 0.2 M NaCl. After a short centrifugation in a tabletop centrifuge (2500 g for one minute), the supernatants were removed to analyze the unbound fractions. After washing twice with the same buffer, the gel-bound peptides were eluted with 50 mM Tris – HCl buffer (pH 6.8) containing 0.2 M NaCl and 5 mM EGTA. Finally, the remaining peptides were eluted by heating at 95 8C for five minutes with 50 mM Tris – HCl buffer (pH 6.8) containing 0.1% (w/v) SDS. The fractions obtained were analyzed by SDS-PAGE. Biosensor analysis

Experimental Procedures

The synthetic peptides were phosphorylated by PKC purified from bovine brain as has been described.9 For the mass spectrometric studies, the phosphorylation reaction mixtures were desalted with a ZipTip apparatus (Millipore) for MALDI-TOF mass spectrometry, direct analyses by liquid chromatography/mass spectrometry, and trypsin digestion.

Binding experiments and kinetic analysis were performed by using a Biacore 1000 apparatus (Biacore K.K. Japan) at 25 8C with a constant flow-rate of 20 ml/minute. Rate constants were calculated using a global fitting method with a modified two-state model by the BIA 3.2 evaluation program (Biacore AB). A pre-coated streptavidin biosensor chip (SA) was used to immobilize biotinylated CaM (Calbiochem CA). Non-specific binding was evaluated in the same experiment using a biosensor chip saturated with biotin. vSrc peptide was diluted in running buffer (10 mM Hepes (pH 7.4), 150 mM NaCl, 0.005% (v/v) Surfactant P20), and to study the interaction, myristoylated vSrc peptide was injected in the presence (1 mM CaCl2) or in the absence (5 mM EGTA) of calcium ions. Regeneration was performed with buffer containing 5 mM EGTA or 20 mM NaOH. Evaluation and calculation of binding parameters were done according to the manual provided by Pharmacia Biosensor AB.

Digestion by trypsin

Nuclear magnetic resonance spectroscopy

For peptide digestion, trypsin (Sigma) with a ratio between 1:20 (w/w) to 1:100 (w/w) of enzyme to substrate was used. The peptide was dissolved in 100 mM ammonium bicarbonate (pH 8.5) and the reaction mixture was incubated for three hours at 37 8C.

All NMR experiments were carried out on a Bruker DMX-500 spectrometer using a 5 mm broadband, z-axis gradient-shielded probe at 298 K. Two-dimensional 1 H – 15N HSQC spectra were obtained by pulse field gradient selection.34 The sweep width was 12 ppm in the 1H dimension and 30 ppm in the 15N dimension, with the 1H carrier set at 500.1324 MHz and the 15N carrier at 50.6814 MHz. The size of the HSQC spectra was a 1024 £ 1024 real data matrix with eight scans for each experiment. Proton chemical shifts were referenced

Sample preparation Myristoylated vSrc peptide was designed according to the N-terminal amino acid sequence of pp60v-src,29 and purchased from Research Genetics Inc. (Huntsville, AL). Rat CaM was expressed in Escherichia coli and purified to homogeneity as described.33 Phosphorylation by PKC in vitro

Mass spectrometry MALDI-TOF mass spectrometry was done on a Voyager DE Pro (PE Biosystems) and, the matrix was

179

SAXS and NMR of CaM –pp60v-src Peptide Complex

to 2,2-dimethyl-2-silapentane-5-sulfonate as 0 ppm. Nitrogen-15 chemical shifts were referenced to liquid NH3. The sequential assignments for the backbone nuclei of the 15N/13C-labeled Ca2þ/CaM were mainly achieved by a set of experiments, CBCA(CO)NNH, CBCANNH, HNCO, and HN(CA)CO. Side-chain assignments were obtained using HBHA(CO)NNH, H(CCCO)NNH, CC(CO)NNH, and HCCH-TOCSY experiments. The details will be described elsewhere (N.H. et al., unpublished results). The assignments for the protein backbone nuclei of the 15N/13C-labeled Ca2þ/CaM –vSrc peptide complex were achieved by the two-dimensional 1H – 15N HSQC spectral changes induced by titrations of vSrc peptide to the Ca2þ/CaM. NMR spectra were processed on a Silicon Graphics Indigo2 workstation using Bruker XWIN-NMR and MSI Felix 95.0 software packages. Small-angle X-ray scattering The recombinant CaM was dissolved in Tris buffer (50 mM Tris – HCl, pH 7.6) containing 120 mM NaCl and a sufficient amount (five molar equivalents relative to CaM) of CaCl2. The protein concentration was determined by quantitative amino acid analysis. The solutions for the Ca2þ/CaM– vSrc peptide mixtures with a molar ratio of 1:2 were prepared at CaM concentrations of 6.0, 9.0, 11.0, 13.0 and 15.0 mg/ml. For the comparison, the solutions of Ca2þ/CaM were also prepared at a concentration of 9.0 mg/ml. The measurements and data analyses were performed as described.10

Acknowledgements This work was supported, in part, by grants-inaid from the Fujita Health University, from the Asahi Glass Foundation (to N.H.), by grants-in-aid for Scientific Research on Priority Areas “Genome Information Science” (14015231 and 15014230) (to N.H.) and Scientific Research (C) (to N.H.) from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and by a grant-in-aid for the Fujita Health University High-tech Research Center from the Ministry of Education, Science, Sports and Culture of Japan. SAXS measurements were performed with the approval of the Photon Factory Advisory Committee (proposal no. 00G147).

References 1. Cross, F. R., Garber, E. A., Pellman, D. & Hanafusa, H. (1984). A short sequence in the p60src N terminus is required for p60src myristylation and membrane association and for cell transformation. Mol. Cell. Biol. 4, 1834–1842. 2. Schultz, A. M., Henderson, L. E., Oroszlan, S., Garber, E. A. & Hanafusa, H. (1985). Amino terminal myristylation of the protein kinase p60src, a retroviral transforming protein. Science, 227, 427– 429. 3. Kamps, M. P., Buss, J. E. & Sefton, B. M. (1985). Mutation of NH2-terminal glycine of p60src prevents

4.

5. 6.

7.

8. 9.

10.

11. 12.

13.

14.

15.

16.

17.

18.

both myristoylation and morphological transformation. Proc. Natl Acad. Sci. USA, 82, 4625– 4628. Song, K. S., Sargiacomo, M., Galbiati, F., Parenti, M. & Lisanti, M. P. (1997). Targeting of a G alpha subunit (Gi1 alpha) and c-Src tyrosine kinase to caveolae membranes: clarifying the role of N-myristoylation. Cell. Mol. Biol. (Noisy-le-grand), 43, 293– 303. Garber, E. A. & Hanafusa, H. (1987). NH2-terminal sequences of two src proteins that cause aberrant transformation. Proc. Natl Acad. Sci. USA, 84, 80 – 84. David-Pfeuty, T., Bagrodia, S. & Shalloway, D. (1993). Differential localization patterns of myristoylated and nonmyristoylated c-Src proteins in interphase and mitotic c-Src overexpresser cells. J. Cell Sci. 105, 613 –628. Lydon, N. B., Gay, B., Mett, H., Murray, B., Liebetanz, J., Gutzwiller, A. et al. (1992). Purification and biochemical characterization of non-myristoylated recombinant pp60c-src kinase. Biochem. J. 287, 985 –993. Crivici, A. & Ikura, M. (1995). Molecular and structural basis of target recognition by calmodulin. Annu. Rev. Biophys. Biomol. Struct. 24, 85 – 116. Takasaki, A., Hayashi, N., Matsubara, M., Yamauchi, E. & Taniguchi, H. (1999). Identification of the calmodulin-binding domain of neuron-specific protein kinase C substrate protein CAP-22/NAP-22. Direct involvement of protein myristoylation in calmodulin-target protein interaction. J. Biol. Chem. 274, 11848 –11853. Hayashi, N., Izumi, Y., Titani, K. & Matsushima, N. (2000). The binding of myristoylated N-terminal nonapeptide from neuron-specific protein CAP-43/ NAP-22 to calmodulin does not induce the globular structure observed for the calmodulin-nonmyristoylated peptide complex. Protein Sci. 9, 1905– 1913. Rhoads, A. R. & Friedberg, F. (1997). Sequence motifs for calmodulin recognition. FASEB J. 11, 331– 340. Hayashi, N., Matsubara, M., Jinbo, Y., Titani, K., Izumi, Y. & Matsushima, N. (2002). Nef of HIV-1 interacts directly with calcium-bound calmodulin. Protein Sci. 11, 529– 537. Matsushima, N., Hayashi, N., Jinbo, Y. & Izumi, Y. (2000). Ca2þ-bound calmodulin forms a compact globular structure on binding four trifluoperazine molecules in solution. Biochem. J. 347, 211 – 215. Woodgett, J. R., Gould, K. L. & Hunter, T. (1986). Substrate specificity of protein kinase C. Use of synthetic peptides corresponding to physiological sites as probes for substrate recognition requirements. Eur. J. Biochem. 161, 177– 184. Matsushima, N., Izumi, Y., Matsuo, T., Yoshino, H., Ueki, T. & Miyake, Y. (1989). Binding of both Ca2þand mastoparan to calmodulin induces a large change in the tertiary structure. J. Biochem. (Tokyo), 105, 883– 887. Chow, M., Newman, J. F., Filman, D., Hogle, J. M., Rowlands, D. J. & Brown, F. (1987). Myristylation of picornavirus capsid protein VP4 and its structural significance. Nature, 327, 482– 486. Kawamura, S., Cox, J. A. & Nef, P. (1994). Inhibition of rhodopsin phosphorylation by non-myristoylated recombinant recoverin. Biochem. Biophys. Res. Commun. 203, 121–127. Senin, I. I., Zargarov, A. A., Alekseev, A. M., Gorodovikova, E. N., Lipkin, V. M. & Philippov, P. P. (1995). N-myristoylation of recoverin enhances its efficiency as an inhibitor of rhodopsin kinase. FEBS Letters, 376, 87 – 90.

180

SAXS and NMR of CaM –pp60v-src Peptide Complex

19. Resh, M. D. (1996). Regulation of cellular signalling by fatty acid acylation and prenylation of signal transduction proteins. Cell. Signal. 8, 403– 412. 20. Towler, D. A., Gordon, J. I., Adams, S. P. & Glaser, L. (1988). The biology and enzymology of eukaryotic protein acylation. Annu. Rev. Biochem. 57, 69 – 99. 21. Osawa, M., Tokumitsu, H., Swindells, M. B., Kurihara, H., Orita, M., Shibanuma, T. et al. (1999). A novel target recognition revealed by calmodulin in complex with Ca2þ- calmodulin-dependent kinase kinase. Nature Struct. Biol. 6, 819– 824. 22. Yuan, T. & Vogel, H. J. (1998). Calcium-calmodulininduced dimerization of the carboxyl-terminal domain from petunia glutamate decarboxylase. A novel calmodulin-peptide interaction motif. J. Biol. Chem. 273, 30328– 30335. 23. Ikura, M., Clore, G. M., Gronenborn, A. M., Zhu, G., Klee, C. B. & Bax, A. (1992). Solution structure of a calmodulin-target peptide complex by multidimensional NMR. Science, 256, 632– 638. 24. Carr, S. A., Biemann, K., Shoji, S., Parmelee, D. C. & Titani, K. (1982). n-Tetradecanoyl is the NH2terminal blocking group of the catalytic subunit of cyclic AMP-dependent protein kinase from bovine cardiac muscle. Proc. Natl Acad. Sci. USA, 79, 6128– 6131. 25. Aitken, A., Cohen, P., Santikarn, S., Williams, D. H., Calder, A. G., Smith, A. & Klee, C. B. (1982). Identification of the NH2-terminal blocking group of calcineurin B as myristic acid. FEBS Letters, 150, 314– 318. 26. Resh, M. D. (1999). Fatty acylation of proteins: new insights into membrane targeting of myristoylated and palmitoylated proteins. Biochim. Biophys. Acta, 1451, 1 – 16. 27. Resh, M. D. (1994). Myristylation and palmitylation of Src family members: the fats of the matter. Cell, 76, 411 – 413. 28. Hayashi, N., Matsubara, M., Titani, K. & Taniguchi, H. (1997). Circular dichroism and 1H nuclear magnetic resonance studies on the solution and

29. 30.

31.

32.

33.

34.

35.

36.

membrane structures of GAP-43 calmodulin-binding domain. J. Biol. Chem. 272, 7639– 7645. Jove, R. & Hanafusa, H. (1987). Cell transformation by the viral src oncogene. Annu. Rev. Cell. Biol. 3, 31 – 56. Dikic, I., Tokiwa, G., Lev, S., Courtneidge, S. A. & Schlessinger, J. (1996). A role for Pyk2 and Src in linking G-protein-coupled receptors with MAP kinase activation. Nature, 383, 547– 550. Eguchi, S., Matsumoto, T., Motley, E. D., Utsunomiya, H. & Inagami, T. (1996). Identification of an essential signaling cascade for mitogenactivated protein kinase activation by angiotensin II in cultured rat vascular smooth muscle cells. Possible requirement of Gq-mediated p21ras activation coupled to a Ca2þ/calmodulin-sensitive tyrosine kinase. J. Biol. Chem. 271, 14169 –14175. Della Rocca, G. J., van Biesen, T., Daaka, Y., Luttrell, D. K., Luttrell, L. M. & Lefkowitz, R. J. (1997). Rasdependent mitogen-activated protein kinase activation by G protein- coupled receptors. Convergence of Gi- and Gq-mediated pathways on calcium/calmodulin, Pyk2, and Src kinase. J. Biol. Chem. 272, 19125– 19132. Hayashi, N., Matsubara, M., Takasaki, A., Titani, K. & Taniguchi, H. (1998). An expression system of rat calmodulin using T7 phage promoter in Escherichia coli. Protein Expr. Purif. 12, 25 – 28. Grzesiek, S. & Bax, A. (1993). The importance of not saturating H2O in protein NMR. Application to sensitivity enhancement and NOE measurements. J. Am. Chem. Soc. 115, 12593– 12594. Heidorn, D. B., Seeger, P. A., Rokop, S. E., Blumenthal, D. K., Means, A. R., Crespi, H. & Trewhella, J. (1989). Changes in the structure of calmodulin induced by a peptide based on the calmodulin-binding domain of myosin light chain kinase. Biochemistry, 28, 6757– 6764. Osawa, M., Kuwamoto, S., Izumi, Y., Yap, K. L., Ikura, M., Shibanuma, T. et al. (1999). Evidence for calmodulin inter-domain compaction in solution induced by W-7 binding. FEBS Letters, 442, 173– 177.

Edited by G. von Heijne (Received 17 November 2003; received in revised form 12 February 2004; accepted 12 February 2004)