If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
To whom correspondence should be addressed: Dept. of Chemistry, One Shields Ave., University of California, Davis, CA 95616. Tel.: 530-752-6358; Fax: 530-752-8995;.
* This work was supported by National Institutes of Health Grants EY012347 (to J. B. A.), EY11522 (to A. M. D.), and RR11973 (UC Davis NMR). This work was also supported by a Pennsylvania Department of Health CURE Formula grant (to A. M. D.) The authors declare that they have no conflicts of interest with the contents of this article. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
GCAP1, a member of the neuronal calcium sensor subclass of the calmodulin superfamily, confers Ca2+-sensitive activation of retinal guanylyl cyclase 1 (RetGC1). We present NMR resonance assignments, residual dipolar coupling data, functional analysis, and a structural model of GCAP1 mutant (GCAP1V77E) in the Ca2+-free/Mg2+-bound state. NMR chemical shifts and residual dipolar coupling data reveal Ca2+-dependent differences for residues 170–174. An NMR-derived model of GCAP1V77E contains Mg2+ bound at EF2 and looks similar to Ca2+ saturated GCAP1 (root mean square deviations = 2.0 Å). Ca2+-dependent structural differences occur in the fourth EF-hand (EF4) and adjacent helical region (residues 164–174 called the Ca2+ switch helix). Ca2+-induced shortening of the Ca2+ switch helix changes solvent accessibility of Thr-171 and Leu-174 that affects the domain interface. Although the Ca2+ switch helix is not part of the RetGC1 binding site, insertion of an extra Gly residue between Ser-173 and Leu-174 as well as deletion of Arg-172, Ser-173, or Leu-174 all caused a decrease in Ca2+ binding affinity and abolished RetGC1 activation. We conclude that Ca2+-dependent conformational changes in the Ca2+ switch helix are important for activating RetGC1 and provide further support for a Ca2+-myristoyl tug mechanism.
Activation and inhibition of photoreceptor guanylyl cyclase by guanylyl cyclase activating protein 1 (GCAP-1): the functional role of Mg2+/Ca2+ exchange in EF-hand domains.
Inactivation of EF-hands makes GCAP-2 (p24) a constitutive activator of photoreceptor guanylyl cyclase by preventing a Ca2+-induced “activator-to-inhibitor” transition.
Diversity of guanylate cyclase-activating proteins (GCAPs) in teleost fish: characterization of three novel GCAPs (GCAP4, GCAP5, GCAP7) from zebrafish (Danio rerio) and prediction of eight GCAPs (GCAP1–8) in pufferfish (Fugu rubripes).
) are all ∼200-amino acid residue proteins containing a covalently attached N-terminal myristoyl group and four EF-hand motifs (EF1 through EF4; Fig. 1). Mg2+ binds to GCAP1 in place of Ca2+ when cytosolic Ca2+ levels are below 50 nm in light-activated photoreceptor cells (
Activation and inhibition of photoreceptor guanylyl cyclase by guanylyl cyclase activating protein 1 (GCAP-1): the functional role of Mg2+/Ca2+ exchange in EF-hand domains.
Retinal guanylyl cyclase isozyme 1 is the preferential in vivo target for constitutively active GCAP1 mutants causing congenital degeneration of photoreceptors.
) showed that the four EF-hands form two semiglobular domains (EF1 and EF2 in the N-domain and EF3 and EF4 in the C-domain); Ca2+ is bound at EF2, EF3, and EF4, and the N-terminal myristoyl group in GCAP1 is buried inside the Ca2+-bound protein flanked by hydrophobic residues at the N and C termini (see the red residues in Fig. 1). The structure of the physiological activator form of GCAPs (Mg2+-bound/Ca2+-free state) is currently unknown.
FIGURE 1.Amino acid sequence alignment of bovine GCAP1, recoverin, and NCS-1. Secondary structural elements (α-helices and β-strands) were derived from NMR analyses (
The structure of the Ca2+-free/Mg2+-bound activator state of GCAP1 has remained elusive, in part because it tends to aggregate under conditions for NMR or x-ray crystallography (
). Here, we present a NMR structural analysis of Ca2+-free/Mg2+-bound GCAP1 mutant that has Val-77 replaced by Glu (called GCAP1V77E). The GCAP1V77E mutant retains functional Mg2+ and Ca2+ binding with intact tertiary structure. However, unlike the dimeric wild type GCAP1, GCAP1V77E is monomeric in solution and remains soluble under NMR conditions. Our NMR analysis indicates that Ca2+-free/Mg2+-bound GCAP1V77E is overall structurally similar to that of Ca2+-saturated GCAP1 (root mean square deviations of 2.0 Å), except that Mg2+ is bound at EF2, and the other EF-hands are unoccupied. The largest Ca2+-dependent structural differences in GCAP1 are seen for residues in EF4 and the adjacent helical region (residues 164–174, called Ca2+ switch helix). We propose that the Ca2+ switch helix may serve as a conduit that relays Ca2+-induced structural changes in EF4 to the RetGC binding site in the N-terminal domain, which provides further support of a Ca2+-myristoyl tug mechanism (
). Myristoylated GCAP1 and its mutants were produced in Escherichia coli strain harboring yeast N-myristoyl transferase and purified to homogeneity using a previously described method (
Activation and inhibition of photoreceptor guanylyl cyclase by guanylyl cyclase activating protein 1 (GCAP-1): the functional role of Mg2+/Ca2+ exchange in EF-hand domains.
). Uniformly 15N-labeled GCAP1, 15N-labeled GCAP1V77E, and triple-labeled 13C,2H,15N-labeled GCAP1V77E used in the NMR studies (0.5 mm) were dissolved in 5 mm Tris-d11 (pH 7.4), 5 mm CaCl2, 5 mm MgCl2, 5 mm dithiothreitol-d10, and 93%/7% H2O/D2O.
Mutagenesis
Mutations were introduced in bovine GCAP1 cDNA by “splicing by overlap extension” technique using PCR reactions catalyzed by high-fidelity Phusion Flash polymerase (Finnzymes/Thermo Scientific). The resultant products were ligated into the NcoI/BamHI sites of pET11d (Novagen/Calbiochem) vector, sequenced, and transformed into expressing cell lines as described previously in detail (
The intrinsic Trp fluorescence of GCAP1 and its mutants was recorded in the presence of variable-free Mg2+ and Ca2+ concentrations as previously described in detail (
The stoichiometry of Ca2+ binding to myristoylated and non-acylated GCAP1 and its mutants was determined using the fluorescent Ca2+ indicator dye method previously described in detail (
) using Fluo-4FF (Molecular Probes/Fisher). Free Ca2+ in the reaction mixture was calculated using the formula [Ca]free = Kd × (F − Fmin)/(100 − F), where F is the fluorescence intensity of the Ca2+ indicator in the assay mixture expressed as a percentage of the fluorescence of the Ca2+-saturated indicator (recorded at the end of each experiment in 1 mm [Ca]free), Fmin is the fluorescence intensity of the Ca2+ indicator in the absence of Ca2+ and also expressed as a percentage of the fluorescence of the Ca2+-saturated indicator, and Kd is a corrected constant of the indicator dye for Ca2+ (
). The fluorescence data were fitted by the equation ([Ca]bound/[GCAP]) = N × [Cafree]n([Cafree]n + Kdn), where [Ca]bound is the concentration of Ca2+ bound to GCAP1 calculated as [Ca]bound = [Ca]total − [Ca]free, where N is the number of Ca2+ ions bound per molecule of GCAP at saturation, Kd is the apparent affinity of GCAP1 for Ca2+, and n is the Hill coefficient. The data shown are representative from independent experiments producing virtually identical results.
Guanylyl Cyclase Assay
Recombinant human RetGC1 was expressed in HEK293 cells and assayed in vitro using [α-32P]GTP as a substrate as previously described in detail (
Activation and inhibition of photoreceptor guanylyl cyclase by guanylyl cyclase activating protein 1 (GCAP-1): the functional role of Mg2+/Ca2+ exchange in EF-hand domains.
Activation and inhibition of photoreceptor guanylyl cyclase by guanylyl cyclase activating protein 1 (GCAP-1): the functional role of Mg2+/Ca2+ exchange in EF-hand domains.
Samples for NMR analyses were prepared by dissolving unlabeled, 15N-labeled, 13C,15N-labeled, or 13C,2H,15N-labeled GCAP1 proteins in 0.5 ml of 90% H2O, 10% [2H]H2O containing 10 mm [2H11]Tris (pH 7.4) and either 5 mm MgCl2 (Mg2+-bound) or 5 mm CaCl2 (Ca2+-bound). All NMR experiments were performed at 37 °C on a Bruker Avance 800-MHz spectrometer equipped with a triple resonance cryoprobe and z axis gradient. NMR experiments and backbone assignments for Ca2+-saturated GCAP1 were described elsewhere (
). Backbone NMR resonance assignments of the Ca2+-free/Mg2+-bound GCAP1V77E activator state (Mg2+ bound at EF2) were obtained in this study by the analysis of three-dimensional NMR data as described previously (
A novel approach for sequential assignment of 1H, 13C, and 15N spectra of proteins: heteronuclear triple-resonance three-dimensional NMR spectroscopy; application to calmodulin.
). Two-dimensional 1H,15N HSQC with 2048 (1H) × 256 (15N) data points, three-dimensional HNCACB with 1024 (1H) × 64 (15N) × 120 (13C) data points, and three-dimensional HNCO with 2048 (1H) × 64 (15N) × 128 (13C) data points were all performed on a triple-labeled GCAP1V77E sample using a 800-MHz Bruker NMR spectrometer equipped with a triple resonance cryogenic probe. In addition, three-dimensional HNCOCACB with 1024 (1H) × 64 (15N) × 128 (13C) data points was performed on the triple-labeled sample using a 600-MHz Bruker NMR spectrometer equipped with a triple resonance cryogenic probe. 1H,15N residual dipolar coupling constants (DNH) were measured with a 15N-labeled GCAP1V77E (∼0.3 mm) containing 12 mg/ml Pf1 phage (Asla Biotech) and using a two-dimensional IPAP (inphase/antiphase) 1H,15N HSQC experiment (
An NMR-guided homology model structure of Mg2+-bound/Ca2+-free GCAP1V77E was generated based on NMR data (chemical shifts, residual dipolar couplings (RDCs) and NOEs) using the Xplor-NIH software suite (
). A template structure for the model calculation was first built using SWISS-MODEL based on the x-ray crystal structure of Ca2+-bound GCAP1 (PDB ID 2R2I). The N-terminal myristoyl group was attached to the template structure as it is in the crystal structure (PDB ID 2R2I). The three Ca2+ ions were deleted, and a single Mg2+ ion was added to the second EF-hand metal binding loop as described by Park et al. (
Determination of three-dimensional structures of proteins by simulated annealing with interproton distance restraints: application to crambin, potato carboxypeptidase inhibitor, and barley serine proteinase inhibitor 2.
). RDC and NMR restraints (NOEs and dihedral angles) were applied during the simulated annealing step. Dihedral angles were calculated by the Talos+ program (
).For under-assigned secondary structural motifs, theoretical restraints within the initial template structure were used to supplement the experimental restraints. Refinement of the final structure was initiated with high temperature annealing at 1000 K for 10 ps, and cooling from 1000 K to 25 K in 12.5 K steps. The duration of cooling dynamics run at each step was 0.2 ps. A total of 500 structures were obtained, and the 75 lowest energy structures were chosen to generate an energy-minimized average structure. The structural statistics are shown in Table 1.
TABLE 1NMR structural statistics for Mg2+-bound/Ca2+-free GCAP1V77E
ITC experiments were performed using a VP-ITC calorimeter (Micro-Cal) at 30 °C, and data were acquired and processed with MicroCal software as described previously (
). Metal-free GCAP1V77E samples were prepared by exchanging protein into buffer containing 15 mm Tris-HCl (pH 7.5), 100 mm NaCl, and 1 mm β-mercaptoethanol. The metal-free GCAP1V77E in the sample cell (50 μm, 1.5 ml) was titrated with either Ca2+ (2 mm) or Mg2+ (40 mm) using 40 injections of 5 μl each.
Results
GCAP1 Mutant (V77E) Binds Functionally to Mg2+ and Ca2+
ITC was used to monitor Mg2+ and Ca2+ binding to GCAP1V77E (Fig. 2). Titration of Mg2+ into apoGCAP1V77E produced an endothermic isotherm (Fig. 2A). Mg2+ binds to GCAP1V77E with an apparent dissociation constant (Kd) of 700 μm and enthalpy difference (ΔH) of +4.2 kcal/mol. The stoichiometry of Mg2+ binding was determined by analyzing 1H,15N HSQC NMR spectra of Mg2+-bound GCAP1V77E. NMR not only determined the number of Mg2+ ions bound per protein but also determined which particular EF-hands are bound to Mg2+ (
). The NMR spectrum of Ca2+-free/Mg2+-bound GCAP1V77E contains one downfield NMR peak at ∼10.5 ppm assigned to Gly-69 (conserved glycine in EF2 binding loop), indicating that 1 Mg2+ is bound to GCAP1V77E at EF2. (Fig. 2A, inset). Thus, GCAP1V77E binds to Mg2+ with the same affinity and stoichiometry as wild type GCAP1 (
), which demonstrates that Ca2+-free/Mg2+-bound GCAP1V77E is structurally and functionally intact.
FIGURE 2.GCAP1 mutant GCAP1V77E binds to Mg2+ and Ca2+ as measured by ITC and NMR (inset). ITC binding isotherms recorded at 30 °C are shown for Mg2+ binding (A) and Ca2+ binding (B) to GCAP1V77E. Binding isotherms were fit to a sequential model (solid line), and fitting parameters are given in “Results.” Downfield spectral regions of 1H,15N HSQC spectra of Mg2+-bound GCAP1V77E (inset, A) and Ca2+-bound GCAP1V77E (inset, B) are also shown. The downfield peak assigned to Gly-69 indicates Mg2+ is bound at EF2 (inset, A), whereas peaks assigned to Gly-69, Gly-105, and Gly-149 indicate Ca2+ is bound at EF2, EF3, and EF4.
Titration of CaCl2 into apoGCAP1V77E showed an ITC isotherm (Fig. 2B) with three separate phases that corresponded to Ca2+ binding at three EF-hand binding sites (EF2, EF3, and EF4). The Ca2+ binding isotherm of GCAP1V77E looked overall similar to that of wild type (
). The highest affinity site in GCAP1V77E (Kd = 0.08 μm and ΔH = −3 kcal/mol) exhibited binding energetics that are nearly identical to those of EF3 in wild type (
). Therefore, this highest affinity site in GCAP1V77E was assigned to EF3. The site with intermediate affinity (Kd = 0.2 μm and ΔH = +1.4 kcal/mol) was assigned to EF4 based on its similar binding affinity to EF4 in wild type under the conditions of microcalorimetry assay (
). By the process of elimination, the lowest affinity site in GCAP1V77E (Kd = 10 μm and ΔH = −0.5 kcal/mol) was assigned to EF2. The Ca2+ binding affinity for EF2 in GCAP1V77E is 10-fold weaker than that of wild type, consistent with the V77E mutation residing in EF2. The lower affinity Ca2+ binding to EF2 may also result from a loss of protein dimerization for GCAP1V77E compared with the dimeric wild type protein (
). For example, intermolecular dimer interactions involving EF2 could stabilize Ca2+ binding to EF2 in the dimeric wild type protein. Indeed, the ΔH for Ca2+ binding to both EF2 and EF4 in GCAP1V77E have opposite signs compared with that of wild type. The opposite sign of ΔH for both EF2 and EF4 in GCAP1V77E suggests that exposed residues in EF2 and EF4 might form Ca2+-dependent intermolecular contacts with each other in the dimeric wild type protein that get disabled (or otherwise altered) in the monomeric GCAP1V77E (
The stoichiometry of Ca2+ binding was determined by analysis of 1H,15N HSQC NMR spectra of Ca2+-bound GCAP1V77E (Fig. 2B, inset). Three downfield NMR peaks assigned to Gly-69 (EF2), Gly-105 (EF3), and Gly-149 (EF4) demonstrated that three Ca2+ ions are bound per protein at EF2, EF3, and EF4. Thus, GCAP1V77E binds to Ca2+ with the same stoichiometry and structure as wild type GCAP1 (
NMR Assignments for Ca2+-free/Mg2+-bound GCAP1V77E
NMR spectroscopy was used to demonstrate that GCAP1V77E adopts a native tertiary structure as determined by comparing 1H,15N HSQC spectra of wild type protein (Fig. 3A and Ref
) to that of GCAP1V77E (Fig. 3, B and C). The peaks in the 1H,15N HSQC NMR spectra represent main chain and side-chain amide groups that provide a residue-specific fingerprint of the overall protein conformation. The resonance assignments for Ca2+-saturated GCAP1WT (
) match quite well with the spectrum of Ca2+-saturated GCAP1V77E (Fig. 3C), indicating that GCAP1V77E is structurally intact. The NMR assignments for Ca2+-free/Mg2+-bound GCAP1V77E are shown by the labeled peaks in Fig. 3B. The Ca2+-free/Mg2+-bound GCAP1V77E forms a monomer in solution under NMR conditions (
). The relative peak positions and spectral patterns overall look similar when comparing spectra of wild type (Fig. 3A) to that of GCAP1V77E (Fig. 3B), indicating that Mg2+-bound/Ca2+-free GCAP1V77E retains the same main chain fold compared with wild type. The monomeric state of GCAP1V77E caused much sharper NMR line width s compared with that of dimeric wild type (Fig. 3, A versus B). The sharper NMR peaks observed for Ca2+-free/Mg2+-bound GCAP1V77E allowed ∼60% assignment of backbone resonances compared with <25% assignment for Ca2+-free wild type and ∼80% assignment for Ca2+-saturated GCAP1 (
). The unassigned residues for GCAP1V77E included the first 20 residues from the N terminus and unstructured regions (Lys-46–Trp-51, Met-74–Asp-108, Ile-119–Met-129, Asp-143–Leu-150, Arg-177–Gln-183). These unassigned residues have weak NMR intensities due to exchange broadening, suggesting that these residues undergo dynamical motions on the chemical shift time scale. The exchange broadening of residues at the domain interface in Ca2+-free/Mg2+-bound GCAP1 (residues Lys-91–Tyr-99) is similar to that seen in Ca2+-free recoverin (
FIGURE 3.NMR spectroscopy of GCAP1WT, GCAP1V77E, and mutants (SGL and ΔLeu-174). Two-dimensional (1H,15N HSQC) NMR spectra of 15N-labeled wild type GCAP1 in the Ca2+-free/Mg2+-bound state (A), GCAP1V77E in the Ca2+-free/Mg2+-bound state (B), GCAP1V77E in the Ca2+-bound state (C), and GCAP1 mutants (SGL, black; ΔLeu-174, red) in the Ca2+-bound state (D). Spectra were obtained at 37 °C. A downfield resonance at ∼10.5 ppm for Ca2+-free/Mg2+-bound GCAP1V77E is assigned to a conserved glycine residue (Gly-69) in EF2 and indicates Mg2+ is bound at EF2. Downfield resonances (at 10.45, 10.47, and 10.55 ppm) for Ca2+-bound GCAP1V77E are assigned to conserved glycine residues (Gly-105, Gly-69, and G149) and indicate that three Ca2+ are bound per protein at EF2, EF3, and EF4. Sequence-specific resonance assignments for Ca2+-free/Mg2+-bound V77E are indicated by the peak labels. Complete NMR assignments were deposited in the BMRB (accession no. 26688).
Downfield-shifted NMR peaks at ∼10.5 ppm for GCAP1V77E are assigned to conserved glycine residues in the EF-hand Ca2+ binding loops and are characteristic of Ca2+/Mg2+-bound EF-hands. In Ca2+-saturated GCAP1V77E (Fig. 3C), three downfield peaks assigned to Gly-69 (EF2), Gly-105 (EF3), and Gly-149 (EF4) confirm that Ca2+ is bound at EF2, EF3, and EF4 as seen in the crystal structure (
). In Ca2+-free/Mg2+-bound GCAP1V77E (Fig. 3B), one downfield peak assigned to Gly-69 (EF3) confirmed that Mg2+ is bound at EF2 as seen by fluorescence spectroscopy (
Activation and inhibition of photoreceptor guanylyl cyclase by guanylyl cyclase activating protein 1 (GCAP-1): the functional role of Mg2+/Ca2+ exchange in EF-hand domains.
NMR-derived Structural Model of Ca2+-free/Mg2+-bound GCAP1V77E
The sequence-specific NMR assignments above for Ca2+-free/Mg2+-bound GCAP1V77E and the secondary structure based on these assignments is summarized in Fig. 1. An NMR-derived structural model of Ca2+-free/Mg2+-bound GCAP1V77E was calculated using NOE-based distances, NMR-derived dihedral angle restraints, and RDC data (Fig. 4) that served as input for restrained molecular dynamics structure calculations (see “Experimental Procedures”). Initial residual dipolar coupling magnitude and rhombicity were calculated by fitting the measured residual dipolar couplings to the calculated structure using the PALES program (
). The RDC-refined structures have a quality Q-factor of 0.28 and an R-factor of 0.995 (Fig. 4C). The overall secondary structure and topology of Mg2+-bound GCAP1V77E is similar to that found in the crystal structure of Ca2+-bound GCAP1 (
). The NMR-derived structural model of Ca2+-free/Mg2+-bound GCAP1V77E (PDB ID 2NA0) was validated with PROCHECK: 86% of residues belonged to the most favorable region in the Ramachandran plot.
FIGURE 4.RDC structural analysis of GCAP1V77E.1H,15N IPAP (inphase/antiphase)-HSQC spectra of Ca2+-free/Mg2+-bound GCAP1V77E in the absence (A) and presence (B) of 12 mg/ml Pf1 phage. Spectral splittings for the isotropic condition (JNH) versus the anisotropic condition (JNH + DNH) are marked by vertical lines and were used to calculate RDCs as described under “Experimental Procedures.” C, RDCs calculated from the structure of Ca2+-free/Mg2+-bound GCAP1V77E in Fig. 5 are plotted versus the RDCs measured in Fig. 4B and show good agreement (Q-factor = 0.28 and an R-factor = 0.95 (
The NMR-derived structural model of Ca2+-free/Mg2+-bound GCAP1V77E (Fig. 5) contains a total of 11 α-helices and 4 β-strands: α1 (residues 7–14), α2 (18–28), α3 (36–43), α4 (51–63), α5 (73–83), α6 (90–99), α7 (109–119), α8 (132–143), α9 (153–161), α10 (164–174), α11 (176–183), β1 (33–35), β2 (70–72), β3 (106–108), and β4 (150–152). GCAP1V77E contains two separate domains comprising four EF hands: EF1 (green, residues 18–43) and EF2 (red, residues 51–83) form the N-terminal domain; EF3 (cyan, residues 90–119) and EF4 (yellow, residues 132–161) form the C-terminal domain. Two C-terminal helices are downstream of EF4 (α10 and α11 in Fig. 4). The helix immediately adjacent to EF4 (α10, highlighted in orange in Fig. 5B) is one-half turn longer in Ca2+-free/Mg2+-bound GCAP1V77E compared with that of Ca2+-bound GCAP1 (Fig. 6). The C-terminal helix (α11) has the same length in both Ca2+-free and Ca2+-bound GCAP1 and makes contacts with the N-terminal myristoyl group (Fig. 5).
FIGURE 5.NMR-derived structure of Ca2+-free/Mg2+-bound GCAP1V77E (PDB ID 2NA0). The main chain structure of Ca2+-free/Mg2+-bound GCAP1V77E (A) and the same view rotated by 180 degrees (B) show four EF-hands (colored as in Fig. 1) packed in a globular arrangement very similar to what is seen for Ca2+-bound GCAP1 (
). The secondary structural elements are labeled as defined in Fig. 1. The Ca2+ switch helix (α10) is highlighted in orange, bound Mg2+ is in blue, and the N-terminal myristoyl group is in magenta.
FIGURE 6.Ca2+-induced conformational changes in GCAP1. Shown in Ca2+-dependent amide chemical shift difference (Ca2+-free minus Ca2+-bound) plotted versus residue number (A) and chemical shift difference mapped onto the main chain structure (B). Residues Thr-171 and Leu-174 exhibited the largest Ca2+-induced chemical shift differences. Residues (Thr-62, Phe-140, Leu-151, Val-160, Leu-170, Thr-171, Leu-174) with a chemical shift difference higher than 0.8 are colored red. Residues (Ala-52, Tyr-55, Asp-68, Ile-115, Ala-118, Ser-141, Ser-152, Glu-158, Gln-161, Asp-168) with a chemical shift difference between 0.5 and 0.8 are colored magenta. Residues with a chemical shift difference <0.5 are colored light blue. C, close-up view of the Ca2+ switch helix (α10, orange) that is elongated by one turn in Ca2+-free/Mg2+-bound GCAP1V77E (left) compared with Ca2+-bound GCAP1WT (right). The angle between EF2 exiting helix (red) and EF3 entering helix (cyan) at the domain interface increased slightly (dotted line) due to Ca2+-dependent interactions with the Ca2+ switch helix.
GCAP1V77E contains Mg2+ bound at EF2 (blue sphere in Fig. 5) as evidenced by characteristic Mg2+-dependent amide chemical shift changes assigned to Gly-69 in EF2 (Fig. 2A, inset). The geometry of the coordinate covalent bonds formed between chelating amino acid residues in GCAP1V77E and the bound Mg2+ could not be observed directly in our NMR study. Instead, the stereochemical geometry and chelation of Mg2+ bound at EF2 was modeled with structural constraints derived from the x-ray crystal structure of Mg2+-bound CaM (
X-ray structures of magnesium and manganese complexes with the N-terminal domain of calmodulin: insights into the mechanism and specificity of metal ion binding to an EF-hand.
). GCAP1 residues at the 1, 3, and 5 positions of the EF-hand loop in EF2 were selected to chelate the bound Mg2+ (see D64 and D68 in Fig. 5A). Replacement of these two residues disrupted Mg2+ binding in our previous biochemical studies (
The four EF-hands of GCAP1V77E with one Mg2+ bound at EF2 (and no metal bound at EF1, EF3, and EF4) each adopt interhelical angles that are similar to those observed in the crystal structure of Ca2+-bound GCAP1 (Table 2). For the Ca2+-free/Mg2+-bound GCAP1V77E structure, the interhelical angles are 132° (EF1), 114° (EF2), 103° (EF3), and 106° (EF4). Therefore, the three functional EF-hands in GCAP1 (EF2, EF3, and EF4) each adopt a somewhat open conformation in the Ca2+-free state, and Ca2+ binding at these sites in GCAP1 cause only a slight change in interhelical angle. In essence, the three functional EF-hands in GCAP1 adopt a “pre-formed” open conformation in the Ca2+-free state akin to that of calbindin D9k (
). As a result, the Ca2+-binding free energy for GCAP1 is NOT coupled to an unfavorable conformational change, which may explain the very high nanomolar Ca2+ binding affinity for GCAP1 (
Chemical shift differences in GCAP1 versus Ca2+ were plotted as a function of residue number and reveal the location of Ca2+-dependent structural changes (Fig. 6A). Detectable chemical shift differences are seen for residues in EF1 (residues 26, 27, and 33) that are implicated in RetGC1 binding (
). Somewhat larger chemical shift changes are seen in EF2 (residues 55, 62, and 68), which represent residues at the domain interface and suggest Ca2+-dependent structural contacts between EF2 and EF3 like that seen previously for recoverin (
). The largest chemical shift differences (highlighted in red in Fig. 6B) are observed in EF4 (residues 140, 151–152, and 160) and helix α10 (residues 168, 170, and 171), called the Ca2+ switch helix. Thus, Ca2+-induced structural changes in EF4 appear coupled to structural changes in the adjacent Ca2+ switch helix. An overlay of GCAP1 structures in the Ca2+-free versus Ca2+-bound states shows overall main chain root mean square deviations of 2.0 Å. The most striking difference is seen in the Ca2+ switch helix (orange in Figs. 5B and 6C), which is one-half-turn longer in the Ca2+-free state. An expanded view of the Ca2+ switch helix (Fig. 6C) reveals two residues (Thr-171 and Leu-174) that are most affected by Ca2+. Thr-171 is exposed in the Ca2+-free structure, whereas it becomes buried and makes contact with Leu-92 in the Ca2+-bound structure. Conversely, Leu-174 is buried and makes contact with Leu-92 in the Ca2+-free structure, but it switches to a solvent-exposed environment in the Ca2+-bound structure. The Ca2+-dependent contacts formed by both Thr-171 and Leu-174 may be important for switching GCAP1 from the Ca2+-free activator to the Ca2+-bound inhibitor states.
Mutations in the Ca2+ Switch Helix Affect Metal Binding and Cyclase Activation
) proposes a structural link between the exiting helix of EF4 and residues that contact the myristoyl group in the N-domain. Ca2+-induced structural changes in EF4 exert a “tug” on downstream C-terminal residues that are in contact with the myristoyl group. These Ca2+-dependent contacts to the myristate are then relayed to the target binding site in the N-domain (
). We hypothesized that the Ca2+-induced shortening of the Ca2+ switch helix observed in our structure (Fig. 6) may transmit the tug action that connects Ca2+ binding at EF4 with target activation in the N-domain. Therefore, we tested if a change in the length of the Ca2+ switch helix by mutagenesis could affect the regulatory properties of GCAP1. We constructed a series of mutants in which the length of the helix was altered either by insertion or deletion of a single amino acid: a Gly was inserted between Ser-173 and Leu-174 (SGL mutant, Fig. 7), and single amino acids were deleted: Arg-172 (ΔArg-172), Ser-173 (ΔSer-173), or Leu-174 (ΔLeu-174) (Fig. 8). NMR HSQC spectra demonstrate that these mutations do not alter the conformation of α10 and these mutants are structurally intact (Fig. 3D). We found that each of these mutants (which changes the length of the Ca2+ switch helix) profoundly affected both the metal binding affinity of GCAP1 and its ability to activate RetGC. The Ca2+ binding stoichiometry in the SGL mutant remained three per GCAP1 molecule, but the apparent affinity decreased slightly with higher cooperativity compared with that of wild type (Fig. 6A). Metal-dependent changes in the intrinsic Trp fluorescence of the SGL mutant looked quite different from that of wild type (Fig. 7, B–D). The Ca2+-induced decrease in fluorescence (in the absence of Mg2+) occurred at higher Ca2+ levels in SGL compared with wild type (Fig. 7B). The presence of 10 mm Mg2+ (that normally causes a quite striking Ca2+-induced increase in Trp fluorescence for wild type) resulted in much smaller Ca2+-dependent change in Trp fluorescence for SGL (Fig. 7C). Thus, the transition from the Mg2+-bound activator state to the Ca2+-bound inhibitor state in the presence of 10 mm Mg2+ (
) was more difficult to monitor in SGL compared with wild type (Fig. 7C). In addition, the Trp fluorescence of the SGL mutant titrated as a function of Mg2+ revealed a nearly 10-fold lower Mg2+ binding affinity compared with that of wild type (Fig. 7D). Because Mg2+ binding to EF2 is essential for converting GCAP1 into its activated state in the absence of Ca2+ (
Activation and inhibition of photoreceptor guanylyl cyclase by guanylyl cyclase activating protein 1 (GCAP-1): the functional role of Mg2+/Ca2+ exchange in EF-hand domains.
), we tested if RetGC1 activation was also affected by this mutation. Indeed, the SGL mutant failed to activate RetGC1 at physiological concentrations of Mg2+ and when the Ca2+ concentration was reduced to levels typical of light-exposed photoreceptors (Fig. 7D).
FIGURE 7.Single amino acid residue insertion in the Ca2+ switch helix affects metal sensor properties of GCAP1.A, Ca2+ binding isotherm for wild type (black, ●) and SGL (red, ●) GCAP1. Ca2+ binding was assayed using titration of 20 μm GCAP1 in the presence of Fluo4FF as described under “Experimental Procedures.” B–D, tryptophan fluorescence titrations for monitoring metal-dependent conformational change in wild type (black, ●) and SGL (red, ●) GCAP1 caused by Ca2+ (B and C) or Mg2+ (D) binding. B and C, Ca2+ titration in the absence (B) or in the presence (C) of 10 mm Mg2+. AU, absorbance units. D, Mg2+ titration. E, RetGC1 activation in vitro by wild type (black, ●) and SGL (red, ●) GCAP1 in the presence of 1 mm Mg2+ was assayed as described under “Experimental Procedures.”
FIGURE 8.Effect of a single amino acid residue deletion in Ca2+ switch helix on metal sensor properties of GCAP1.A–C, tryptophan fluorescence titrationsfor monitoring Ca2+-dependent conformational change in ΔArg-172 (■) (A), ΔSer-173 (▴) (B), or ΔLeu-174 (♦) (C) GCAP1 in the absence (red symbols) or in the presence (blue symbols) of 10 mm Mg2+. AU, absorbance units. D and E, comparison of the Ca2+-dependent (D) or Mg2+-dependent (E) Trp fluorescence change in the wild type (black, ●), ΔArg-172 (□) (A), ΔSer-173 (▴) (B), or ΔLeu-174 (♢); no Mg2+ added in D. F, dose dependence of RetGC1 activation in vitro by wild type (black, ●) ΔArg-172 (red, □), ΔSer-173 (red, ▴), or ΔLeu-174 (red ♢) GCAP1 in the presence of 6 mm free Mg2+ and 2 mm EGTA; the data points were fitted using Synergy Kaleidagraph 4 software assuming a standard Michaelis hyperbolic function.
Making the Ca2+ switch helix shorter by a single residue produced a similar effect regardless of which amino acid was deleted, Arg-172, Ser-173, or Leu-174 (Fig. 8). In all three deletion mutants, the Ca2+-induced decrease in Trp fluorescence looked similar to that of wild type (Fig. 8, A–C, red symbols), except for a slight 2-fold shift toward higher Ca2+ concentrations (Fig. 8D). As seen above for SGL, the presence of 10 mm Mg2+ caused a much smaller change in Ca2+-dependent Trp fluorescence for each of the deletion mutants, thus making it difficult to monitor the transition from the Mg2+-bound activator state to the Ca2+-bound inhibitor state (Fig. 8, A–C, blue symbols) in contrast to the much larger Ca2+-induced fluorescence change seen for wild type (Fig. 7B). As seen above for SGL, the Trp fluorescence of the deletion mutants titrated as a function of Mg2+ revealed a nearly 10-fold lower Mg2+ binding affinity compared with that of wild type (Fig. 8E). Finally, the deletion mutants all failed to activate RetGC1 at physiological concentrations of Mg2+, consistent with their lack of Mg2+ binding (Fig. 8F).
Discussion
NCS proteins like recoverin and NCS-1 undergo large Ca2+-induced conformational changes that cause extrusion of the N-terminal myristoyl group, termed Ca2+-myristoyl switch (
Calcium binding, but not a calcium-myristoyl switch, controls the ability of guanylyl cyclase-activating protein GCAP-2 to regulate photoreceptor guanylyl cyclase.
), before this study little was known structurally about the Mg2+-bound/Ca2+-free activator state of GCAPs and how Ca2+-induced conformational changes in GCAPs control cyclase activation.
In this study we present NMR assignments (Fig. 3B), NOEs (Table 1), RDCs (Fig. 4), and mutagenesis data (Fig. 7) to probe Ca2+-dependent conformational changes between the Ca2+-saturated inhibitory state versus the Ca2+-free/Mg2+-bound GCAP1V77E (Fig. 6). Overall, the NMR chemical shift differences between Ca2+-saturated wild type GCAP1 and Ca2+-free/Mg2+-bound GCAP1V77E are relatively small (Fig. 6A), consistent with an overall similar main chain conformation for the two states (root mean square deviations = 2.0 Å). This contrasts with the large Ca2+-induced conformational changes seen for other EF-hand proteins, including recoverin (
). The large Ca2+-induced conformational changes in recoverin and CaM both drive the exposure of hydrophobic residues, which consumes Ca2+ binding free energy and accounts for their low binding affinity. The Ca2+-free EF-hands in GCAP1 adopt a preformed open conformation, like what is seen for calbindin D9k (
). The lack of any large scale Ca2+-induced protein conformational change in GCAP1, therefore, may explain why the GCAPs bind Ca2+ with at least 100-fold higher affinity compared with that of recoverin (
The most apparent Ca2+-induced structural changes in GCAP1 are localized in EF4 and the adjacent Ca2+ switch helix (α10). Ca2+ binding to EF4 is essential for switching GCAP1 between activator and inhibitor states (
Activation and inhibition of photoreceptor guanylyl cyclase by guanylyl cyclase activating protein 1 (GCAP-1): the functional role of Mg2+/Ca2+ exchange in EF-hand domains.
). Upon Ca2+ binding to EF4, its interhelical angle decreases by 2° (Table 2) along with a slight lengthening of the exiting helix (α9) that in turn exerts a force (tug) on the adjacent Ca2+ switch helix (α10). As a result, the Ca2+ switch helix is one half-turn longer in the Ca2+-free structure compared with the Ca2+-bound structure (Fig. 6, B and C). The Ca2+-induced shortening of the Ca2+ switch helix alters the disposition of key residues (Thr-171 and Leu-174) that form Ca2+-dependent contacts with EF3 (Leu-92 and Tyr-95). These Ca2+-dependent contacts alter the packing angle between helices α5 and α6 at the domain interface (Fig. 6C) that resembles the Ca2+-induced domain swiveling in recoverin (
). The Ca2+-dependent structural changes in the Ca2+ switch helix also exert a force on the C-terminal helix (α11) that alters its contact with the myristoyl group. We propose that this Ca2+-dependent perturbation of the myristoyl group could alter its contact with EF1 and EF2 and thereby affect the accessibility of GCAP1 residues (Met-26, Tyr-37, Val-77, and Ala-78) in the RetGC1 binding site (
). In essence, the Ca2+-induced structural changes in EF4 are relayed to the cyclase binding site via the Ca2+ switch helix. This relay scheme is consistent with a previously proposed mechanism called the Ca2+-myristoyl tug (
) and Mg2+ (FIGURE 7., FIGURE 8.). This effect on Mg2+ binding (FIGURE 7., FIGURE 8.) is especially important for the RetGC1 binding site in GCAP1, which includes the N-domain and requires Mg2+ binding in EF2 to interact with the cyclase at low Ca2+ levels that occur in light-adapted photoreceptors (
). Evidently, the length of the Ca2+ switch helix (α10) in wild type GCAP1 has evolved to optimize the proper tug action between the N- and C-domains, which tunes the Ca2+ and especially Mg2+ binding affinity into the physiological range.
GCAP1 residues in EF2 and EF3 located at the domain interface (Val-77–Trp-94, see FIGURE 1., FIGURE 6.) exhibit exchange-broadened NMR resonances and are conformationally dynamic. The dynamical nature of these residues is consistent with Ca2+-induced conformational changes in GCAP1 at the domain interface that were observed previously (
). The corresponding residues in recoverin (Fig. 1) also exhibited broad NMR resonances, and 15N NMR relaxation dispersion studies revealed that these residues at the domain interface exhibit backbone dynamics on the millisecond timescale (
). Our structural model of Mg2+-bound/Ca2+-free GCAP1V77E suggests a related but smaller Ca2+-induced rearrangement at the domain interface (Fig. 6). Residues Tyr-55, Thr-62, and Ala-118 at the domain interface exhibit large Ca2+-induced chemical shift differences (Fig. 6A). In addition, residues in EF3 (Leu-92 and Tyr-95) make Ca2+-dependent contacts with Thr-171 and Leu-174 in the Ca2+ switch helix (Fig. 6C). These Ca2+-dependent contacts to EF3 cause a change in packing angle between EF2 and EF3 at the domain interface (see the dotted line in Fig. 6C) that is somewhat reminiscent of the Ca2+-dependent domain swiveling observed for recoverin (
NMR relaxation data and size-exclusion chromatography analysis previously showed that GCAP1 is dimeric in solution at high micromolar protein concentration (
). A GCAP1 homolog, GCAP2, undergoes a Ca2+-sensitive dimerization at low micromolar protein concentrations, which originally suggested that reversible dimerization may control formation and activation of RetGC:GCAP in a 2:2 complex (
Instead of binding calcium, one of the EF-hand structures in guanylyl cyclase activating protein-2 is required for targeting photoreceptor guanylyl cyclase.
) does not appear to be Ca2+-sensitive, it is possible that GCAP1 dimerization might promote a functional interaction within a RetGC1 dimer on the disk membrane (
), consistent with a 2:2 complex that contains a RetGC1 dimer bound to a GCAP1 dimer. Previously we showed that the V77E mutation in GCAP1 eliminated GCAP1 dimerization and abolished its ability to bind RetGC but did not block its ability to undergo the functional transition between Ca2+-bound and Mg2+-bound states (
). If GCAP1 dimerization is important for RetGC activation, allosteric changes in the GCAP1 dimer quaternary structure may control cyclase activation. In this scenario a small Ca2+-induced change in tertiary structure could lead to a much larger change in quaternary structure akin to the O2-dependent conformational changes in hemoglobin (
). Hydrophobic residues at the GCAP1 domain interface (Val-77, Ala-78, Leu-82, Trp-94) are solvent-exposed and might mediate specific contacts that control RetGC1 binding and GCAP1 dimerization alike. However, it cannot be excluded that these exposed GCAP1 residues may create an artificial dimer at high concentrations and in the absence of the target enzyme. Future studies are needed to further probe the dimeric structure of GCAP1 and determine whether Ca2+-induced changes in quaternary structure might control its activation of RetGC1.
Author Contributions
J. B. A. directed the overall project and wrote the paper. S. L. performed NMR and ITC experiments, analyzed the NMR and ITC data, performed structure calculations, and helped write the manuscript. E. V. O. and A. M. D. constructed the mutants. I. V. P. assayed Trp fluorescence, Ca2+ binding stoichiometry, and RETGC activity. I. V. P. and A. M. D. analyzed the mutagenesis data and participated in writing the manuscript.
Acknowledgments
We thank Jeff Walton and Bennett Addison for technical support and help with NMR experiments.
Activation and inhibition of photoreceptor guanylyl cyclase by guanylyl cyclase activating protein 1 (GCAP-1): the functional role of Mg2+/Ca2+ exchange in EF-hand domains.
Inactivation of EF-hands makes GCAP-2 (p24) a constitutive activator of photoreceptor guanylyl cyclase by preventing a Ca2+-induced “activator-to-inhibitor” transition.
Diversity of guanylate cyclase-activating proteins (GCAPs) in teleost fish: characterization of three novel GCAPs (GCAP4, GCAP5, GCAP7) from zebrafish (Danio rerio) and prediction of eight GCAPs (GCAP1–8) in pufferfish (Fugu rubripes).
Retinal guanylyl cyclase isozyme 1 is the preferential in vivo target for constitutively active GCAP1 mutants causing congenital degeneration of photoreceptors.
A novel approach for sequential assignment of 1H, 13C, and 15N spectra of proteins: heteronuclear triple-resonance three-dimensional NMR spectroscopy; application to calmodulin.
Determination of three-dimensional structures of proteins by simulated annealing with interproton distance restraints: application to crambin, potato carboxypeptidase inhibitor, and barley serine proteinase inhibitor 2.
X-ray structures of magnesium and manganese complexes with the N-terminal domain of calmodulin: insights into the mechanism and specificity of metal ion binding to an EF-hand.
Calcium binding, but not a calcium-myristoyl switch, controls the ability of guanylyl cyclase-activating protein GCAP-2 to regulate photoreceptor guanylyl cyclase.
Instead of binding calcium, one of the EF-hand structures in guanylyl cyclase activating protein-2 is required for targeting photoreceptor guanylyl cyclase.
The atomic coordinates and structure factors (code 2NA0) have been deposited in the Protein Data Bank (http://wwpdb.org/).
A complete list of NMR assignments for Ca2+-free/Mg2+-bound GCAP1V77E has been deposited in the Biological Magnetic Resonance Bank (BMRB) (accession no. 26688).
43590-2&id=gr1.jpg){kind=link}
43590-2&id=gr2.jpg){kind=link}
43590-2&id=gr3.jpg){kind=link}
43590-2&id=gr4.jpg){kind=link}
43590-2&id=gr5.jpg){kind=link}
43590-2&id=gr6.jpg){kind=link}
43590-2&id=gr7.jpg){kind=link}
43590-2&id=gr8.jpg){kind=link}