Structure of a Ca2+-Myristoyl Switch Protein That Controls Activation of a Phosphatidylinositol 4-Kinase in Fission Yeast*

Neuronal calcium sensor (NCS) proteins transduce Ca2+ signals and are highly conserved from yeast to humans. We determined NMR structures of the NCS-1 homolog from fission yeast (Ncs1), which activates a phosphatidylinositol 4-kinase. Ncs1 contains an α-NH2-linked myristoyl group on a long N-terminal arm and four EF-hand motifs, three of which bind Ca2+, assembled into a compact structure. In Ca2+-free Ncs1, the N-terminal arm positions the fatty acyl chain inside a cavity near the C terminus. The C14 end of the myristate is surrounded by residues in the protein core, whereas its amide-linked (C1) end is flanked by residues at the protein surface. In Ca2+-bound Ncs1, the myristoyl group is extruded (Ca2+-myristoyl switch), exposing a prominent patch of hydrophobic residues that specifically contact phosphatidylinositol 4-kinase. The location of the buried myristate and structure of Ca2+-free Ncs1 are quite different from those in other NCS proteins. Thus, a unique remodeling of each NCS protein by its myristoyl group, and Ca2+-dependent unmasking of different residues, may explain how each family member recognizes distinct target proteins.

The high degree of sequence identity among NCS proteins suggests that their three-dimensional structures should be quite similar. Indeed, the overall structures of the Ca 2ϩ -bound state of the unmyristoylated forms of several NCS family members look rather similar, as determined by x-ray crystallography (28 -32) or nuclear magnetic resonance (NMR) spectroscopy (33)(34)(35). The four EF-hands are packed in a tandem array, in contrast to the dumbbell-shaped arrangement seen in CaM (36) and troponin C (37). However, much less is known about the structure of the Ca 2ϩ -free and/or myristoylated forms of NCS proteins. The fact that NCS proteins, like recoverin, GCAP1, and NCS-1, all recognize different physiological target proteins suggests that they must have some distinguishing structural characteristic that may be conferred by the interaction of the N-myristoyl moiety with the rest of the protein. Consistent with this notion, the structures of myristoylated recoverin (PDB code 1iku) and myristoylated GCAP1 (PDB code 2r2i) are quite different from each other. Moreover, although there is only one NCS protein (recoverin) whose structure has been determined for its myristoylated form in both the absence (38) and presence of Ca 2ϩ (39), the structure of Ca 2ϩ -free recoverin is quite different from the structure of Ca 2ϩ -bound recoverin. In its Ca 2ϩ -free state, the N-terminal myristoyl group of recoverin is sequestered inside a deep hydrophobic cavity in the N-domain. In its Ca 2ϩ -bound state, the N-myristoyl group is extruded, permitting the now solventexposed fatty acyl chain to interact with membranes (40,41), allowing recoverin (42) and other NCS proteins (30,43,44) to bind to target membranes when the Ca 2ϩ level is high. The Ca 2ϩ -induced conformational change that exposes the myristoyl group has been dubbed a Ca 2ϩ -myristoyl switch.
Frq1 in budding yeast both activates the PtdIns 4-kinase Pik1 (21,23) and its N-terminal myristoyl group enhances its membrane binding. When bound to the enzyme, Frq1 occupies residues 121-174 of Pik1, which forms a U-shaped structure that lies upstream of the catalytic domain (residues 792-1066) (26). Mammalian NCS-1 can interact with yeast Pik1 (45) and reportedly regulates PtdIns 4-kinase activity in animal cells (46,47). Ca 2ϩ -dependent activation of PtdIns 4-kinase by NCS-1 may be especially important in neurons because modulation of phosphoinositide synthesis by intracellular Ca 2ϩ controls synaptic vesicle exocytosis (48) and is involved in synaptic plasticity (49). However, previously, we were unable to obtain structural information for myristoylated Frq1.
Here, we report the NMR structures of Ca 2ϩ -free myristoylated Ncs1 in solution and the Ca 2ϩ -loaded form of the same protein bound to its target site (residues 111-159) in fission yeast Pik1, hereafter referred to as Pik1(111-159). Strikingly, the location of the myristoyl-binding site in Ca 2ϩ -free Ncs1 and the resulting structure of Ncs1 are quite different from that of either recoverin (38) or GCAP1 (50). Our data support the conclusion that myristoylation shapes each NCS protein into a distinct structure. Moreover, we find that Ncs1 undergoes large Ca 2ϩ -induced conformational changes that lead to extrusion of the myristoyl group, causing hydrophobic residues in the C-domain that sequester the fatty acyl chain in Ca 2ϩ -free Ncs1 to become solvent-exposed in Ca 2ϩ -bound Ncs1 and thus available to interact with Pik1(111-159). In vivo, this Ca 2ϩ -myristoyl switch presumably promotes membrane localization of Ncs1 and its association with Pik1. Furthermore, based on the structure of Pik1-bound Ncs1, we propose a mechanism for simultaneous Ca 2ϩ -induced membrane localization and activation of the enzyme. Finally, given the profound structural differences between the Ca 2ϩ -free states of myristoylated Ncs1 (this study), myristoylated recoverin (38), and myristoylated FIGURE 1. A, amino acid sequence alignment of S. pombe Ncs1 with other NCS proteins (sequence numbering is for S. pombe Ncs1). Secondary structure elements (helices and strands), EF-hand motifs (EF1, green; EF2, red; EF3, cyan; and EF4, yellow), and residues that interact with the myristoyl group (highlighted magenta) are mapped onto the amino acid sequence of Ncs1. Swiss Protein Data base accession numbers are Q09711 (S. pombe Ncs1), Q06389 (S. cerevisiae Frq1), P21457 (bovine recoverin), and P43080 (human GCAP1). Secondary structural elements indicated schematically were derived from the analysis of NMR data ( 3 J HNH␣ , chemical shift index (54), and sequential NOE patterns). B, amino acid sequence alignment (one-letter code) of the N-terminal Frq1-binding region of S. cerevisiae Pik1 (ScPik1) and the N-terminal Ncs1-binding region of S. pombe Pik1 (ScPik1). Identities are indicated in boldface with a colon; conservative substitutions are indicated with a period. Residues in ScPik1 implicated in Frq1 binding are indicated in boldface red with an asterisk (26); residues in SpPik1 implicated in Ncs1 binding are indicated in boldface purple with an asterisk (this study). Residues that are ␣-helical in the ScPik1-Frq1 complex (125-135 and 156 -169) (26) are depicted as blue cylinders; residues that are ␣-helical in the SpPik1-Ncs1 complex (114 -127 and 143-156) (this study) are depicted as green cylinders. GCAP1 (50), we propose the general idea that N-terminal myristoylation is critical for shaping each NCS family member into a unique structure, which upon Ca 2ϩ -induced extrusion of the myristoyl group exposes a unique set of previously masked residues, thereby accounting for the target specificity of each NCS protein.

EXPERIMENTAL PROCEDURES
Protein Expression and Purification-Expression and purification of recombinant myristoylated S. pombe Ncs1 (and S. cerevisiae Frq1) was described previously (24). Briefly, Ncs1 and Frq1 (without any affinity tags) were expressed using pET11 vector harboring the ncs1 (or FRQ1) coding sequence and was co-expressed with yeast N-myristoyl-CoA transferase using pBB131 vector harboring the N-myristoyl-CoA transferase coding sequence in Escherichia coli strain BL21(DE3). Ncs1 was labeled with 15 N or 15 N/ 13 C isotopes by growing cells at 37°C in M9 minimal media supplemented with 15 N-labeled ammonium chloride and 13 C-labeled glucose as described previously (26). Unlabeled or 13 C 14 -labeled myristic acid was added to the culture 20 min prior to induction that is needed for N-terminal myristoylation. Cells were harvested, lysed, and spun down to collect supernatant. Protein was then purified from the supernatant using hydrophobic interaction (butyl-Sepharose), anion exchange (DEAE-Sepharose), and size exclusion (Superdex 200) columns. The final protein purity is ϳ95% judged by SDS-PAGE, and more than 90% of the protein was myristoylated determined by mass spectrometry analysis.
A functional polypeptide fragment of S. pombe Pik1 (residues 111-159, named Pik1(111-159)) uniformly labeled with nitrogen-15 and/or carbon-13 and tagged with a C-terminal His 6 tract was expressed in E. coli strain BL21(DE3)-RIL (Stratagene) carrying the pET23d vector (Novagen) harboring the PIK1(111-159) coding sequence grown in M9 medium containing [ 15 N]NH 4 Cl and [ 13 C]glucose (26). Labeled Pik1(111-159) was isolated from the insoluble fraction of bacterial cell lysates dissolved in 8 M urea buffer and purified using Ni 2ϩchelate affinity chromatography on a nitrilotriacetate resin (Qiagen), according to the manufacturer's instructions. Peak fractions were then dialyzed extensively against 4 liters of 25 mM sodium acetate (pH 5.0) to remove urea. After dialysis, the Pik1(111-159) polypeptide remained soluble at pH 5.0 and was concentrated about 10-fold to a final concentration of 1 mM used in NMR experiments.
Lipid Kinase Assay-PtdIns-4-OH kinase activity of S. pombe Pik1 was assayed using a procedure described previously (27) with minor modifications. The reaction product, [ 32 P]PtdIns 4-phosphate, was visualized by autoradiography of TLC plates. The band intensity on the TLC plates was measured and quantified using a PhosphorImager (GE Healthcare). The activity of Pik1 in cell-free extracts prepared from cells completely lacking Ncs1 was studied. For this purpose, a fission yeast strain that lacked Ncs1 expression (h ϩ his3-D1 ade6-M216 leu1-32 ura4-D18 ncs1⌬::his3 (24)) with overexpression of Pik1 (using plasmid pREP1-PIK1 (a kind gift from S. Hemmingsen) was grown at 30°C to midexponential phase in minimal medium lacking His and Ura to ensure maintenance of the plasmid, and extracts were prepared by glass bead breakage and clarified by brief centrifugation at 500 ϫ g to remove beads, unbroken cells, and large debris. Samples were assayed under conditions specific for Pik1 activity (51).
NMR Spectroscopy-NMR samples of Ca 2ϩ -free myristoylated or unmyristoylated Ncs1 (ϳ0.7 mM) or Frq1 (0.  15 N-or 13 C/ 15 Nlabeled Pik1(111-174) bound to 1 eq of unlabeled Ncs1) were also prepared for some of the NMR experiments. All NMR experiments were performed at 37°C on Bruker Avance 600 or 800 spectrometers with an Ultrashield Bruker magnet equipped with a four-channel interface, triple resonance probe, and cryo-probe with z axis pulsed field gradients. 15 N-1 H HSQC spectra were recorded on samples of 15 N-labeled Ncs1 in the presence or absence of unlabeled Pik1(111-159) in 95% H 2 O, 5% 2 H 2 O. The number of complex points and acquisition times were as follows: 256, 180 ms ( 15 N (F 1 )) and 512, 64 ms ( 1 H (F 2 )). The 13 C(F1)-edited and 13 C(F3)-filtered NOESY-HSQC spectra (see Figs. 4A and 5A) were recorded on a sample of unlabeled Ca 2ϩ -free Ncs1 protein attached to a 13 C-labeled myristoyl group (Fig. 4A) or unlabeled Ncs1 bound to 13 C-labeled Pik1(111-159) (Fig. 5A) as well as 13 C-labeled Ncs1 bound to unlabeled Pik1(111-159) (data not shown). Intermolecular NOESY experiments were performed as described previously (52). Stereospecific assignments of chiral methyl groups of valine and leucine were obtained by analyzing 1 H-13 C HSQC experiments performed on a sample that contained 10% 13 C labeling in either Ncs1 or Pik1(111-159). All triple resonance experiments were performed, processed, and analyzed as described previously (53)  The triple resonance and NOESY spectra measured above were analyzed to determine secondary and tertiary structure in Ca 2ϩ -free myristoylated Ncs1 and Ca 2ϩ -bound unmyristoylated Ncs1-Pik1(111-159) complex. The chemical shift index (see Ref. 54) for detailed description), 3 J NH␣ coupling constants, and NOE connectivity patterns for each residue were analyzed and provided a measure of the overall secondary structure. Small 3 J NH␣ coupling constants (Ͻ5 Hz), strong NOE connectivities (NN(i,i ϩ 1) and ␣N(i,i ϩ 3)), and positive chemical shift index are characteristic of residues in an ␣-helix. Conversely, large 3 J NH␣ coupling constants (Ͼ8 Hz), strong ␣N(i,i ϩ 1), and weak NN(i,i ϩ 1) NOE connectivities and negative chemical shift index are characteristic of residues in a ␤-strand. The results of the secondary structure analysis of Ncs1 and Pik1(111-159) are summarized schematically in Fig. 1.
Structure Calculation-Three-dimensional 15 N-NOESY-HSQC and 13 C-NOESY-HSQC and two-dimensional homonuclear NOESY spectra of Ca 2ϩ -free myristoylated Ncs1 and Ca 2ϩ -bound Ncs1 (bound to Pik1(111-159)) were analyzed to obtain 1553 and 1225 NOE distance restraints used in the structure calculations, respectively. In addition to the NOE-derived distance constraints, the following additional constraints were included in the structure calculation: 18 distance constraints involving Ca 2ϩ bound to loop residues 1, 3, 5, 7, and 12 in each EF-hand motif (EF-2, EF-3, and EF-4); 124 distance constraints for 80 hydrogen bonds; and 200 dihedral angle constraints ( and ) derived from TALOS (55). Fifty independent structures were calculated by XPLOR-NIH software (56) implemented with YASAP protocol (57), and the 15 structures of lowest energy were selected and overlaid with r.m.s.d. of 0.69 Å (Ca 2ϩfree) and 0.9 Å (Ca 2ϩ -bound). Structures of Ca 2ϩ -free unmyristoylated Ncs1 were also calculated in a similar fashion. Figures of NMR structures in this paper were prepared with PyMOL or VMD (University of Illinois at Urbana-Champaign).

RESULTS
Structure of Ca 2ϩ -free Myristoylated Ncs1-15 N-1 H HSQC NMR spectra of Ca 2ϩ -free Ncs1 (both myristoylated and unmyristoylated forms) exhibited the expected number of highly dispersed peaks with uniform intensities, indicating that Ca 2ϩ -free Ncs1 adopts a stable three-dimensional fold. Analysis of 15 N relaxation parameters (T 1 and T 2 ) indicates an average rotational correlation time of 9.65 Ϯ 0.5 ns, suggesting that Ca 2ϩ -free Ncs1 is monomeric in solution under NMR conditions. The spectral similarity for both myristoylated and unmyristoylated forms of Ca 2ϩ -free Ncs1 suggests their overall protein structures are similar, and thus we were able to determine the NMR structures of both myristoylated and unmyristoylated forms of Ca 2ϩ -free Ncs1. The sequence-specific NMR assignments of Ca 2ϩ -free myristoylated Ncs1 were analyzed and described previously (BMRB 16446) (58). The assigned resonances in the HSQC spectrum represent main chain and side chain amide groups that serve as fingerprints of the overall conformation. Three-dimensional protein structures derived from the NMR assignments were calculated on the basis of NOE data, chemical shift analysis, and 3 J NH␣ spin-spin coupling constants (see "Experimental Procedures"). The final NMR-derived structures of Ca 2ϩ -free myristoylated Ncs1 are illustrated in Fig. 2, A and B (atomic coordinates have been deposited in the RCSB Protein Data bank, code 2l2e). Table 1 summarizes the structural statistics calculated for the 15 lowest energy conformers.
Structure of Ca 2ϩ -free Unmyristoylated Ncs1-The NMR chemical shift assignments and structure for Ca 2ϩ -free myristoylated Ncs1 are very similar to those of Ca 2ϩ -free unmyristoylated Ncs1 (see overlaid structures in Fig. 3A). The root mean square deviation between the two structures is 1.18 Å. The main structural differences are detected in the N-terminal arm (residues Gly 2 -Arg 21 ) and C-terminal helix, both of which are somewhat destabilized and shortened by the absence of the myristoyl group (e.g. helix ␣1, residues 10 -18, and helix ␣10, residues 178 -186). The 15 N chemical shift differences between myristoylated and unmyristoylated Ncs1 are plotted as a function of residue number in Fig. 3B. The largest chemical shift differences are observed for the residues that contact the myr-istoyl group (Fig. 3C). The overall main chain structures for myristoylated and unmyristoylated Ncs1 are nearly identical in the EF-hand regions (r.m.s.d. ϭ 1.04 Å). Thus, the myristoyl group does not alter the main chain structure of the EF-hands in Ca 2ϩ -free Ncs1, but instead the myristoyl chain penetrates inside the protein by displacing hydrophobic side chains. Indeed, the Ca 2ϩ -free myristoylated Ncs1 has a slightly higher melting temperature compared with unmyristoylated Ncs1 (measured by differential scanning calorimetry, data not shown), consistent with stabilization of the hydrophobic core by the sequestered myristoyl group.
Myristoyl-binding Site in Ncs1-The structural environment around the covalently attached myristoyl group in Ncs1 was determined by analyzing NMR experiments (three-dimensional ( 13 C/F 1 ) edited and ( 13 C/F 3 ) filtered NOESY-HSQC) performed on unlabeled Ca 2ϩ -free Ncs1 samples that contained a 13 C-labeled myristoyl group (Fig. 4A). These NMR spectra selectively probed atoms in Ncs1 that lie within 5 Å of the 13 C-labeled fatty acyl chain. We analyzed nuclear Overhauser effect (NOE) dipolar interactions between the C14 methyl of the myristoyl group ( 13 C 14 , F 2 ϭ 16.62 ppm) and the protein (Fig. 4A, upper panel), and between the C2 methylene of the myristoyl chain ( 13 C 2 , F 2 ϭ 38.05 ppm) and the protein (Fig. 4A, lower panel). The spectra probing the C 14 methyl group (Fig. 4A, upper panel) exhibit many off-diagonal peaks, which could be assigned to residues with aromatic ring protons  Table 1). EF-hands and myristoyl group (magenta) are colored as defined in Fig. 1 (Tyr 129 and Phe 169 ) and to protons in aliphatic side chains (Leu 101 , Ile 124 , Val 125 , Met 121 , Ile 179 , and Leu 183 ). Thus, the C 14 methyl group is surrounded by hydrophobic side chains from residues in EF3 and EF4 and the C-terminal helix. The spectra probing the C 2 position of the myristoyl moiety (Fig. 4A, lower panel) exhibit off-diagonal peaks assigned to residues in the loop between EF3 and EF4 (Val 132 , Val 136 , and Pro 139 ) and the C-terminal helix (Thr 178 , Ala 182 , and Leu 185 ). On the basis of these NMR data, the N-terminal myristoyl group in Ncs1 resides inside a protein cavity located in the C-terminal domain; in marked contrast, the myristoyl group in recoverin (38) and in GCAP1 (50) is housed in an N-terminal cavity in both proteins. The myristoyl group attached to Ncs1 adopts an extended conformation (Fig. 4B) that is about 75% buried inside the protein (Fig. 4C). The C 14 methyl group of the myristate makes close contacts with hydrophobic side chains from Val 125 , Phe 169 , and Ile 179 located inside the hydrophobic core (Fig. 4, B and C). Thus, the C 14 methyl end of the myristoyl group protrudes deep inside the protein.  (21). Both Frq1 and Pik1 are highly conserved in fission yeast, and therefore we set out to verify whether S. pombe Pik1 is similarly activated by Ncs1. The effects of Ca 2ϩ and Ncs1 on the catalytic activity of PtdIns 4-kinase in fission yeast were monitored using an in vitro enzyme assay described previously (see supplemental Fig. 1) (27). Although S. pombe Pik1 displayed detectable basal activity in the absence of exogenously added Ncs1 and Ca 2ϩ , the enzyme was stimulated close to 5-fold in the presence of a saturating concentration of Ca 2ϩ and myristoylated Ncs1 (supplemental Fig. 1). Thus, Ca 2ϩbound Ncs1 activates S. pombe Pik1 similar to Frq1 activation of S. cerevisiae Pik1 (21). In the absence of added Ca 2ϩ , addition of myristoylated Ncs1 had a negligible stimulatory effect (supplemental Fig. 1). The most likely explanation for this observation is that sequestration of the myristoyl group in Ca 2ϩ -free Ncs1 (Fig. 2) blocks the residues in Ncs1 that are necessary to contact its binding site on Pik1. In agreement with this view, unmyristoylated Ncs1 yielded modest, but detectable, stimulation of the lipid kinase even in the absence of added Ca 2ϩ (supplemental Fig. 1). Furthermore, as we demonstrate below, Ncs1 undergoes a Ca 2ϩ -myristoyl switch in which Ca 2ϩ -induced extrusion of the myristoyl group exposes critical residues involved in interacting with the PtdIns 4-kinase (see under "Discussion").
Next, we set out to determine the NMR structure of Ca 2ϩbound Ncs1 bound to S. pombe Pik(111-159), as was done previously for Frq1 bound to Pik1(121-174) (26). The NMR spectra and assignments of Ca 2ϩ -bound Ncs1-Pik1(111-159) complex are shown in supplemental Fig. 5. More than 85% of the backbone assignments for Ncs1 and Pik1(111-159) were obtained as described under "Experimental Procedures." The unassigned residues were located in unstructured regions as follows: loop residues 134 -139 for Pik1(111-159) and the last eight residues at the C terminus for Ncs1. The NMR assignments in supplemental Fig. 5 then served as the basis for analyzing both intramolecular and intermolecular NOESY spectra as described for the Frq1-Pik1(121-174) complex (26). This analysis of the NOESY spectra provides distance constraints for determining the overall protein fold and probing contacts from key residues at the protein interface (Fig. 5A). The NMR-derived structure of Ca 2ϩ -bound Ncs1-Pik1(111-159) complex is shown in Fig. 5, B-D, and structural statistics are given in Table 2.
The four EF-hands in Ncs1 are arranged in a tandem array and, overall, form a globular structure with a concave solventexposed groove lined by two separate hydrophobic patches (highlighted yellow in Fig. 5C). These two hydrophobic surfaces represent bipartite binding sites on Ncs1 that interact with Pik1 (111-159). The structure of Pik1(111-159) in the complex adopts a conformation that contains two ␣-helices (residues  . 5D). Interestingly, these same hydrophobic residues in Ca 2ϩ -free Ncs1 make close contacts with the myr-  istoyl group. Therefore, Ca 2ϩ -induced extrusion of the myristoyl group causes exposure of hydrophobic residues in Ncs1 that forms part of the Pik1-binding site. This hydrophobic interaction was further confirmed using site-specific mutagenesis in which the L119A mutant of Pik1(111-159) binds to Ncs1 with ϳ3-fold lower affinity (supplemental Fig. 2). The opposite face of the Pik1 N-terminal helix contains polar and positively charged residues (Lys 119 , Arg 120 , Asn 123 , and Arg 124 ) that point outward toward the solvent. The C-terminal helix of Pik1(111-159) contains many hydrophobic residues (Val 145 , Ala 148 , Ile 150 , and Ile 154 ) that contact the exposed N-terminal hydrophobic groove of Ncs1 (Trp 30 , Phe 34 , Phe 48 , Ile 51 , Tyr 52 , Phe 55 , Phe 85 , and Leu 89 ). The two helices of Pik1(111-159) do not interact with one another or with the unstructured connecting loop and are highly stabilized by interactions with Ncs1. Ca 2ϩ -induced Conformational Changes in Ncs1-Comparing the structures of Ca 2ϩ -free (Fig. 2B) and Ca 2ϩ -bound (Fig.  5B) Ncs1 reveals large Ca 2ϩ -induced protein conformational changes (see supplemental Movie 1) analogous to the Ca 2ϩmyristoyl switch in recoverin (39,60). The structures of Ca 2ϩfree and Ca 2ϩ -bound Ncs1 have an overall r.m.s.d. of 7 Å when comparing all heavy atoms. The topology of Ca 2ϩ -free Ncs1 in the EF-hand regions is somewhat similar to that of Ca 2ϩ -free recoverin (r.m.s.d. ϭ 1.8 Å) except that the N-terminal myristoyl group is buried in a cavity formed by EF3 and EF4 (Fig. 2E) rather than the cavity in recoverin formed by EF1 and EF2 (Fig.  2F). Ca 2ϩ -binding at EF2, EF3, and EF4 in Ncs1 causes the familiar closed-to-open transition in the EF-hands that promotes a 45°swiveling about Gly 95 in the domain linker and results in a repacking of the domain interface (supplemental Fig. 3). The swiveling of the two domains then pulls the N-terminal myristoyl group out of the protein cavity at the C terminus, resulting in Ca 2ϩ -induced exposure of the myristoyl group and a concomitant exposure of hydrophobic residues (Leu 101 , Val 125 , Val 128 , Leu 138 , Ile 152 , Leu 155 , and Phe 169 ) that contact the myristoyl group in the Ca 2ϩ -free protein. Ca 2ϩ -induced conformational changes in the N-domain occur simultaneously and result in the exposure of many additional hydrophobic res-idues (Trp 30 , Phe 34 , Phe 48 , Ile 51 , Tyr 52 , Phe 55 , Phe 85 , and Leu 89 ), forming an exposed hydrophobic crevice (Fig. 5C), similar to those seen in all other Ca 2ϩ -bound NCS proteins examined to date (28 -32). In essence, Ca 2ϩ binding to Ncs1 leads to domain swiveling that causes extrusion of the myristoyl group. As a result of conformational changes in both domains, two separate hydrophobic patches are formed on the surface of Ca 2ϩ -bound Ncs1 that can now accommodate the hydrophobic faces of the two amphipathic helices from Pik1 (Fig. 5, B-D).

DISCUSSION
In this study, we determined NMR structures for Ca 2ϩ -free myristoylated Ncs1 and Ca 2ϩ -bound Ncs1 complexed to a fragment (residues 111-159) of a fission yeast PtdIns 4-kinase. Ca 2ϩ -free Ncs1 adopts a novel structure where the N-terminal myristoyl group is sequestered inside a protein cavity located near the C terminus (Figs. 2B and 4C). The structural location and environment around the myristoyl group in Ncs1 is very different from that in either recoverin (38) or GCAP1 (Fig. 2) (50). We suggest that each NCS protein adopts a distinct structure because its N-terminal myristoyl group associates with patches of hydrophobic residues that are unique to that protein; thus, upon Ca 2ϩ -evoked extrusion of the myristoyl group, a distinctive ensemble of hydrophobic residues is unmasked, exposing surface residues that allow each class of NCS protein to associate specifically with a particular physiological target. This scenario explains how a Ca 2ϩ signal can cause each NCS family member to engage a different physiological target, despite the high degree of sequence similarity among NCS family members (1).
As we have documented here, Ca 2ϩ binding to Ncs1 causes very large protein structural changes that causes extrusion of the myristoyl group, quite analogous to the Ca 2ϩ -myristoyl switch described previously for recoverin (39). The Ca 2ϩ -induced exposure of the myristoyl group for Ncs1 explains why myristoylated (but not unmyristoylated) Ncs1 binds to S. pombe cell membranes only at high Ca 2ϩ levels (24). The Ca 2ϩinduced extrusion of the myristoyl moiety also exposes two hydrophobic patches on a concave surface of the Ca 2ϩ -bound protein that provide sites for making important contacts with PtdIns 4-kinase (Fig. 5). Therefore, the Ca 2ϩ -myristoyl switch promotes both the capacity of Ncs1 to bind and activate the lipid kinase and controls the delivery of the Ncs1-Pik1 complex to the membrane where the substrate for this enzyme resides (Fig. 6).
Previous studies have shown that various frequenins (mammalian NCS-1 (30) and yeast Frq1 (61)) appear localized to membranes even at low Ca 2ϩ levels, suggesting that NCS-1 and Frq1 may not possess a functional myristoyl switch (62). Indeed, NMR structural studies on Frq1 suggested the Ca 2ϩfree myristoylated Frq1 protein is in a partially unfolded molten-globule state, and the myristoyl group remains solvent-exposed regardless of Ca 2ϩ level (63). These observations of a constitutively exposed myristoyl group are in stark contrast with the Ca 2ϩ -induced extrusion of the myristoyl group in recoverin (39) and Ncs1 (this study). Mutagenesis studies have suggested that particular residues in NCS-1 (62) might be responsible for preventing a Ca 2ϩ -myristoyl switch. However, these residues are somewhat conserved in both S. pombe Ncs1 and mammalian NCS-1 and do NOT prevent the Ca 2ϩ -myristoyl switch in this case. The very high sequence identity (Ն60%) among NCS-1, Frq1, and Ncs1, would imply that the threedimensional structures of Ncs1, NCS-1, and Frq1 must all be very similar. We considered the possibility that the persistent exposure of the myristoyl group in Frq1 and NCS-1 observed in prior work might be an artifact caused by protein misfolding due to the tags (His 6 or GFP) attached to the C terminus in those previous studies because our structure of Ncs1 shows that the myristoyl group makes important contacts with residues close to the C terminus (Ile 179 , Leu 183 , and Leu 185 ). Indeed, NMR spectra of Ca 2ϩ -free myristoylated Frq1 prepared without a C-terminal His 6 tag exhibit methyl resonances below 0 ppm (due to aromatic ring currents in the hydrophobic core), which are similar to those observed for Ncs1 and characteristic of a folded protein (supplemental Fig. 4). By stark contrast, NMR spectra of Ca 2ϩ -free myristoylated Frq1 that contains a C-terminal His 6 tag completely lack any resonances below 0 and above 9 ppm, indicative of a molten-globule (unfolded) state. Based on these findings and our structure of untagged myristoylated Ncs1 in the absence and presence of Ca 2ϩ , we feel that it is highly likely that all frequenins (from mammalian NCS-1 to yeast Frq1) will have structures very similar to that of Ncs1 (Figs. 2 and 5) and will undergo a Ca 2ϩ -myristoyl switch that is critical for its function. Thus, it seems clear from our findings that some aspects of studies of the subcellular localization and cellular roles of NCS family members might have been compromised by the use of tagged derivatives that may have caused significant structural perturbations. Nonconserved residues of NCS proteins at the C terminus (␣10) and immediately following EF3 (Fig. 1A) interact closely with the N-terminal myristoyl group in Ncs1 and thus help stabilize the novel structure of Ca 2ϩ -free Ncs1 (Fig. 2B). The corresponding residues in recoverin and GCAP1 do not contact the myristoyl group. Instead, both recoverin and GCAP1 have nonconserved residues near the N terminus (called an N-terminal arm highlighted purple in Fig. 2) that make specific contacts with the myristoyl moiety. GCAP1 also contains an extra helix at the C terminus that contacts the N-terminal arm and myristoyl group (Fig. 2D). Thus, nonconserved residues at the N and C termini and loop between EF3 and EF4 all play a role in creating a unique environment around the myristoyl group. In Ncs1, the long N-terminal arm and particular hydrophobic residues in the C-terminal helix are crucial for placing the C14 fatty acyl chain in a cavity between EF3 and EF4 (Fig. 2E). By contrast, the much shorter N-terminal arm in both recoverin and GCAP1 prevents the myristoyl group from reaching the C-terminal cavity and instead places the fatty acyl chain between EF1 and EF2 (Fig. 2F). Likewise, nonconserved residues at the N and C termini and/or loop between EF3 and EF4 may play a role in forming unique myristoyl binding environments in other NCS proteins, such as visinin-like proteins, neurocalcins, and hippocalcins that may explain their capacity to associate with functionally diverse targets once they have undergone a Ca 2ϩ -induced conformational change. In Ca 2ϩ -free Ncs1, an N-terminal arm (purple) places the myristoyl group (red) in a hydrophobic cavity (yellow) flanked by a C-terminal helix (␣ 10 ). Right panel shows Ca 2ϩ -induced conformational changes in Ncs1 that cause exposure of myristate and two hydrophobic patches (yellow), followed by structural rearrangement in Pik1 induced by its binding to Ca 2ϩ -bound Ncs1. Ncs1 binding to Pik1 imposes a U-turn in the main chain of Pik1 that is necessary to allow the LKU domain (gray) to interact with the catalytic domain (orange). In addition, N-myristoylation of Ncs1 (red) helps deliver the complex to membranes where it can bind to substrate (PtdIns). Inset shows schematic diagram of a somewhat different Ca 2ϩ -myristoyl switch mechanism for recoverin.
Aside from NCS family members, N-terminal myristoylation confers important structural effects in many other classes of proteins. The structures of myristoylated forms of ARF (64), Bcr-Abl (65), c-Abl (66), and the HIV-1 matrix protein (67) all reveal intimate contacts between the protein and the fatty acyl chain that help mold these proteins into biologically active structures. Interaction of the myristoyl group with oncogenic Bcr-Abl protein is critical for activating tyrosine kinase activity, and drugs (e.g. imatinib) important for treating human leukemias prevent access to the active conformation by occluding the myristoyl binding pocket (65). In the ARF protein, the myristoyl moiety interacts with important switch residues that control GTPase activity involved in regulating vesicular trafficking. Finally, interactions of the myristoyl group with other parts of the HIV-1 matrix protein control its oligomerization, an important step in how this virus targets a host cell. These highly diverse structural interactions demonstrate that N-terminal myristoylation is an important tool for shaping protein structures into distinct and physiologically active conformations.
The structures of Ca 2ϩ -free Ncs1 and Ncs1-Pik1 complex (Figs. 2 and 5) suggest how a Ca 2ϩ -myristoyl switch might promote activation of PtdIns 4-kinase (Fig. 6). Under resting basal conditions, cytosolic Ca 2ϩ levels are presumably maintained below 100 nM and Ncs1 exists in its Ca 2ϩ -free state with a sequestered myristoyl group buried in the C-domain that covers part of its binding site for PtdIns 4-kinase (highlighted yellow in Fig. 6) and prevents binding of Ncs1 to Pik1. The fatty acyl chain has the same molecular dimensions (length and width) as the N-terminal helix of Pik1(111-159), which explains why the myristoyl group and Pik1 helix can effectively compete for the same binding site in Ncs1. A rise in cytosolic Ca 2ϩ will cause Ca 2ϩ -induced conformational changes in Ncs1, resulting in extrusion of the N-terminal myristoyl group (see supplemental Movie 1). Ca 2ϩ -induced extrusion of the myristoyl group exposes a hydrophobic crevice in the C-terminal domain of Ncs1, and concomitantly, Ca 2ϩ -induced structural changes in its N-domain result in formation of a second exposed hydrophobic crevice, also seen in all other Ca 2ϩbound NCS proteins examined to date (26, 28 -31). These two separate hydrophobic sites on the surface of Ca 2ϩ -bound Ncs1 are different from Ca 2ϩ -bound recoverin that contains only one exposed hydrophobic patch (Fig. 6, inset) that interacts with a single target helix in rhodopsin kinase (68). The two exposed hydrophobic sites on Ncs1 bind to the hydrophobic faces of the two antiparallel amphipathic ␣-helices in Pik1(111-159) (colored magenta in Figs. 5B and 6), akin to the mechanism proposed previously for S. cerevisiae Frq1 and Pik1(121-174) (26). The Ca 2ϩ -induced binding of Ncs1 to PtdIns 4-kinase may promote a long range structural interaction between the LKU and catalytic domains, which are conserved in both S. cerevisiae and S. pombe Pik1 (26), leading to acquisition of the conformation optimal for lipid kinase activity. Simultaneously, Ncs1 binding to PtdIns 4-kinase will also promote membrane localization of the lipid kinase, because Ca 2ϩ -bound Ncs1 contains an extruded myristoyl group that serves as a membrane anchor. Thus, Ncs1 controls both delivery of PtdIns 4-kinase to the membrane where its substrates are located and formation of the optimally active state of the enzyme. We propose that a corre-sponding Ca 2ϩ -induced membrane localization and activation of PtdIns 4-kinase-␤ by NCS-1 may take place in neurons and may serve to couple phosphoinositide cascades with calcium signaling pathways, which is thought to be important in synaptic plasticity (49).