Nanomolar affinity of EF-hands in neuronal calcium sensor 1 for bivalent cations Pb²⁺, Mn²⁺, and Hg²⁺

Abstract

Abiogenic metals Pb and Hg are highly toxic since chronic and/or acute exposure often leads to severe neuropathologies. Mn²⁺ is an essential metal ion but in excess can impair neuronal function. In this study, we address in vitro the interactions between neuronal calcium sensor 1 (NCS1) and divalent cations. Results showed that non-physiological ions (Pb²⁺ and Mn²⁺) bind to EF-hands in NCS1 with nanomolar affinity and lower equilibrium dissociation constant than the physiological Ca²⁺ ion. (K_{d, Pb²⁺} = 7.0 ± 1.0 nM; K_{d, Mn²⁺} = 34.0 ± 6.0 nM; K). Native ultra-high resolution mass spectrometry (FT-ICR MS) and trapped ion mobility spectrometry—mass spectrometry (nESI-TIMS-MS) studies provided the NCS1-metal complex compositions—up to four Ca²⁺ or Mn²⁺ ions and three Pb²⁺ ions (M⋅Pb_1-3Ca_1-3, M⋅Mn_1-4Ca_1-2, and M⋅Ca_1-4) were observed in complex—and similarity across the mobility profiles suggests that the overall native structure is preserved regardless of the number and type of cations. However, the non-physiological metal ions (Pb²⁺, Mn²⁺, and Hg²⁺) binding to NCS1 leads to more efficient quenching of Trp emission and a decrease in W30 and W103 solvent exposure compared to the apo and Ca²⁺ bound form, although the secondary structural rearrangement and exposure of hydrophobic sites are analogous to those for Ca²⁺ bound protein. Only Pb²⁺ and Hg²⁺ binding to EF-hands leads to the NCS1 dimerization whereas Mn²⁺ bound NCS1 remains in the monomeric form, suggesting that other factors in addition to metal ion coordination, are required for protein dimerization.

Graphical Abstract

Graphical Abstract.

Association of Pb²⁺ and Hg²⁺ to EF hands stabilizes a NCS1 dimer whereas Mn²⁺ bound NCS1 favors a monomeric form.

Abbreviation

neuronal calcium sensor 1

(NCS1)

1-anilinonaphthalene-8-sulfonic acid

(1,8- ANS)

equilibrium dissociation constant

(K_d)

Introduction

Calcium ion (Ca²⁺) is considered to be a universal second messenger ion as a change in its intracellular concentration regulates a wide variety of biological processes ranging from muscle contraction,¹ to neuronal signaling,² fertilization at the beginning of the life cycle, cell cycle progression, to cell division.^3,4. The majority of Ca²⁺ sensor proteins contain Ca²⁺ binding sites, so called EF-hands that bind Ca²⁺ with high specificity and affinity. Ca²⁺ association to the EF hands triggers a conformation transition and ultimately modulates affinity of Ca²⁺ sensors for downstream effector proteins.⁵ EF-hands were shown to bind abiogenic ions such as Pb²⁺ and Cd²⁺ and biogenic Mn²⁺ with nano to micro molar affinity and association of these metals often promotes structural changes that are analogous to those triggered by Ca²⁺.^6–8 Here, we focused on three metal ions (Pb²⁺, Mn²⁺, and Hg²⁺) as potential electron acceptors for neuronal calcium sensor 1 (NCS1). Pb²⁺ and Hg²⁺ have ionic radii of 1.20 Å and 1.02 Å, and a charge to radius ratio of 1.68 eÅ^–1 and 1.96 eÅ^–1, respectively,⁸ that are comparable to the physical properties of Ca²⁺ ion; ionic radius and charge to radius ratio of 0.99 Å and 2.02 eÅ^–1, respectively.⁹ Although Mn²⁺ has the smaller ionic radius and the larger charge to radius ratio (0.66 Å and 2.98 eÅ^–1, respectively) compared to the Ca²⁺ ion, this cation exhibits similar preferences towards the negatively charged amino acid residues and a coordination geometry analogous to that seen with Ca²⁺ and Mg^{2+  8} and thus can potentially interact with EF-hands in neuronal calcium sensors (NCS).

NCS1 is a 200 amino acid residue protein (i.e. the most ancient member of the NCS protein family) and has orthologues ranging from Saccharomyces cerevisiae to humans.¹⁰ In the human genome NCS1 and other NCS family members are encoded by 14 genes.^11,12 NCS1 contains four EF-hands. A non-functional EF-1 and Mg²⁺/Ca²⁺ binding EF-2 are in the N-terminal domain whereas the C-terminal domain carries Ca²⁺ binding EF-3 and EF-4 (Fig. 1).^13,14 The 12-residue canonical EF-hand loops bind Ca²⁺ ions through oxygen atoms in the sidechain carboxyl group or carbonyl group of the residues located in the position 1, 3, 5, 7, and 9. The sixth coordinating residue, Glu, is located in the exiting helix and serves as a bidentate ligand (Fig. 1).^13,15

Fig. 1.

Left: Cartoon presentation of the NCS1 structure (PDB entry 2LCP). Non-functional EF-hand 1 is colored in green, Mg²⁺/Ca²⁺ binding EF-hand 2 is shown in red, and Ca²⁺ binding EF-hand 3 and 4 are shown in yellow and cyan, respectively. Ca²⁺ ions are shown as green spheres and the sidechains of W30 and W103 are colored in green and yellow, respectively. Right top: Sequence alignment for NCS1 EF-hand 2, 3, and 4. Positions of amino acid residues that coordinate Ca²⁺ ions are labeled using arrows. Right bottom: Structure of Ca²⁺ bound EF-hand 4 in NCS1 with the Ca²⁺ coordinating residues shown as sticks (PDB entry 2LCP).

NCS1 structure shows 2 × 2 structural arrangements of four EF-hands and a hinge loop that connects the N-terminal EF-2 with the C-terminal EF-3.¹⁶ Ca²⁺ binding to NCS1 causes a positional shift in the C-terminal EF-hands leading to exposure of the hydrophobic crevice on protein surface, modulating affinity for target molecules.^13,17 The pairing of the EF-hands plays a significant role in modulating Ca²⁺ binding affinity and a subtle change in Ca²⁺ concentration can be detected^15,18 which points to the evolution of NCS1 as a Ca²⁺ sensor protein rather than merely as a Ca²⁺ buffer.¹⁹ The NCS1 sequence contain two Trp residues that can serve as intrinsic fluorescence probes for metal binding as their emission quantum yield is modulated by Ca²⁺ association to the EF-hands.²⁰ W103 is located in the entrance helix of EF-3 and its indole ring is partially shielded from solvent as it is positioned between the N- and C-terminal domain (Fig. 1). This residue is highly conserved among NCS proteins, pointing towards its potential structural and/or functional role. The second Trp residue, W30 is located in the entering helix of EF hand 1 and its sidechain is solvent exposed.

NCS1 regulates a wide number of cellular functions including exocytosis,²¹ neurite outgrowth,²² neuroprotection,²³ axonal regeneration,²⁴ and nuclear regulation.²⁵ Several NCS1 intracellular partners have been identified including ADP-ribosylation factor 1,^26,27 α1 subunit of voltage-gated Ca²⁺ channel 2.1,^28,29 dopamine D2 receptor,^30–32 and inositol-1,4,5-triphosphate receptor (InsP3R).^33–35 NCS1 also interacts with interleukin receptor accessory protein like-1 (IL1RAPL1), particularly found in the hippocampus, and this interaction may mediate the downregulation of N-type calcium channels.^36–38 NCS1 has been implicated in several neuro-pathologies and non-neurological diseases such as aggressive form of breast cancer.³⁹ A point mutation in NCS1 sequence, R102Q, has been found in patients with autism and associated with structural deficit in this protein.⁴⁰ NCS1 has also been linked to schizophrenia, bipolar disorder, and Parkinson's disease.^41,42 Recently NCS1 was also implicated in the susceptibility to cocaine addiction and the effectiveness of nicotine replacement therapy through the interaction with dopamine D2 receptor.^43–45

Recent studies from our group have shown that another member of the NCS1 family downstream regulatory element antagonist modulator (DREAM) can bind divalent metals Pb²⁺ and Cd²⁺ with an affinity superior to that for Ca²⁺.⁶,⁷ Considering a high sequence and structural similarity between NCS1 and DREAM; here, we address the potential role of NCS1 as a target protein for non-physiological metals Pb²⁺, Hg²⁺ , and Mn²⁺. Importantly, NCS1 was implicated in physiological and pathological processes that were associated with Mn²⁺ and Pb²⁺ neurotoxicity such as learning difficulties, decrease in IQ, and cognitive deficiencies,^46,47 suggesting that non-physiological metals interactions with NCS1 may represent a molecular mechanism of acute and chronic Pb²⁺, Mn²⁺, and Hg²⁺ poisoning.

Material and methods

Isolation and purification of NCS1 protein.

The human NCS1, variant 1, expression vector was purchased from GeneCopoeia (product number EX-V0686-B31) and the gene sequence is provided in the Supplementary information (Table S1). Expression of recombinant NCS1 with a C-terminally attached His-tag was accomplished in BL21 (DE3) Escherichia coli (E. coli) competent cells. The recombinant protein was purified following the published protocol with minor modifications.⁴⁸ Briefly, the NCS1 expression vector was transformed into BL21 (DE3) E. coli competent cells and cells were grown in Terrific Broth medium at 37°C in the presence of 100 μg mL^–1 ampicillin until reaching an optical density of 0.6 at 600 nm. The recombinant protein expression was induced by addition of 0.5 mM IPTG to the cell culture and the cells were incubated with shaking (250 rpm) for an additional 12 h at room temperature and then harvested through 20 min centrifugation at 5000 rpm. The cell pellets were resuspended in Lysis buffer (100 mM Tris buffer, pH 8.0, 0.5 M NaCl, 1 mM CaCl₂, 5 mM MgCl₂, 5 mM imidazole, 1.0% Triton-X100, 0.5% Tween 20, and 10% glycerol), supplemented with 20 μg ml^–1 DNase and 200 μg ml^–1 lysozyme, and sonicated on an ice bath for 30 min at 50% duty cycle. The supernatant obtained upon centrifugation of the cell lysate at 5000 rpm for 5 h was loaded to a Ni-NTA column equilibrated with buffer A (20 mM Tris-HCl, pH 7.40, 300 mM NaCl, 0.5 mM CaCl₂) and the column was washed extensively with buffer A containing increasing concentration of imidazole (5 to 25 mM). Protein was eluted upon passing buffer A with 30 mM imidazole through the column. Isolated protein was then dialyzed twice against 20 mM Tris, pH 7.40 for 12 h and stored in same buffer. The protein purity was assessed using SDS-PAGE electrophoresis.

Sample Preparation. All protein samples were prepared in 20 mM Tris, pH 7.40 and the concentrations were determined by single beam UV-vis spectrophotometer (Cary 50, Varian) using ε_280nm = 22 000 M^–1cm^–1.⁴⁹ The 1-anilinonaphthalene-8-sulfonic acid (1,8-ANS) concentration was determined using ε_350nm = 4999 M^–1cm^–1.⁵⁰ For native MS and TIMS-MS measurements, 50 μM protein was dialyzed against 50 mM ammonium acetate buffer (pH 6.9) overnight and appropriate salt (20 μM CaCO₃, 100 μM Pb (AcO)₂, or 300 μM MnCO₃) was added to the protein sample before MS analysis. All solvents for MS analysis were HPLC grade from Thermo Fisher Scientific Inc. (Waltham, MA).

Steady State Fluorescence Emission. Steady state emission spectra were measured at room temperature (20°C) using a 0.2 × 1 cm path length quartz cuvette by a custom Chronos^FD spectrofluorimeter (ISS, Champaign IL). Protein samples were excited along 0.2 cm path length and emitted light was detected through a 1.0 cm path length. For Trp emission measurements, samples were excited at 295 nm while 350 nm excitation was used to characterize the emission of the 1,8-ANS: NCS1 complex. To determine the equilibrium dissociation constants for Pb²⁺, Mn²⁺, or Hg²⁺ binding to NCS1, small aliquots of metal ion stocks (6 mM Pb (AcO)₂, 100 mM Mn CO₃, 50 mM HgCl₂ dissolved in 20 mM Tris, pH 7.40) were titrated into 10 μM NCS1 protein samples in the presence/absence of 1 mM Ca²⁺, 20 mM Tris buffer, pH 7.40. Due to the precipitation of NCS1 samples at increased Pb²⁺ concentration, the maximum concentration of Pb²⁺ used in titration studies was 160 μM. To prepare apoNCS1 samples, freshly prepared protein was dialyzed against 20 mM Tris, 200 μM EDTA, pH = 7.40 for 12 h. NCS1 protein sample was further dialyzed for 12 h against 20 mM Tris buffer, pH 7.40 for two cycles. The fraction of Pb²⁺, Mn²⁺, and Hg²⁺ bound protein in the absence or presence of Ca²⁺ (f_b) was determined from the decrease in fluorescence emission intensity at 330 nm or 335 nm, using Equation 1.

(1)

where I_i is the fluorescence emission intensity at various metal ions concentrations; I₀ is the intensity in the absence of metal; and I_max is the emission intensity of the saturated solution of metal bound NCS1.

The equilibrium dissociation constants for Pb²⁺ and Mn²⁺ binding to NCS1 were determined using a quadratic equation (Equation 2).⁵¹  

(2)

where P_t is the total protein concentration, L_t is the total ligand concentration, and n denotes the number of binding sites. In case of Ca²⁺ displacement measurements, K_d in Equation 2 represents apparent dissociation constant, K_app, for Pb²⁺ or Mn²⁺ binding to Ca²⁺NCS1 or Pb²⁺, Mn²⁺, or Hg²⁺ binding to 1,8-ANS: Ca²⁺NCS1. The titration curve for Hg²⁺ association to apo and Ca²⁺NCS1 were analysed using Equation 3.

(3)

where K_d indicates the apparent dissociation constant, and c¹ and c² are the scaling factors.

The K_app and K_d values and associated errors reported in Table 1 were obtained as the mean values and standard deviations, respectively, from three independent measurements. Based on the apparent dissociation constant for metals binding to Ca²⁺NCS1, the dissociation constants for metal binding to apoNCS1 (K_d) was calculated using Equation 4.⁵²  

(4)

where K_{d, Ca2+} is the equilibrium dissociation constant for Ca²⁺ binding to apoNCS1, which was found to be 96±48 nM ⁵³ and [Ca²⁺]_T is the total calcium concentration, which is 1 mM.

As the emission intensity of 1,8-ANS: NCS1 complexes increases considerably in the presence of metals, we used the change in 1,8-ANS: NCS1 emission to determine equilibrium dissociation constant for metal binding to NCS1. In those measurements, small aliquots of metal stocks were added to 1,8-ANS: NCS1 complex (20 μM NCS1, 40 μM 1,8-ANS, 20 mM Tris buffer, pH 7.40), and the emission spectra were collected. The increase in emission intensity at the emission maximum of 466 nm was used to calculate the fraction of metal bound to 1,8-ANS: protein complex by Equation 1, and data were fitted to Equation 2 or Hill equation (Equation 5) to determine the dissociation constant for metal binding to 1,8-ANS: Ca²⁺ NCS1 complexes.

(5)

Circular Dichroism. Far-UV circular dichroism spectra of 10 μM NCS1 in 20 mM Tris buffer, pH 7.40 in presence of 1 mM EDTA, 1 mM Ca²⁺, or 120 μM Pb²⁺/1 mM Mn²⁺/1 mM Hg²⁺ were recorded using a quartz cuvette (1 mm × 10 mm) by J-810 Jasco CD spectrometer and data were collected through a 1 mm path from 200 to 250 nm at room temperature (20°C).

Quenching Study. Trp quenching studies of apo, and metal bound NCS1 were performed with acrylamide, a polar organic compound that is frequently used in Trp quenching studies. Steady state emission spectra of intrinsic Trp residues were detected from 300 nm to 450 nm using the excitation wavelength of 295 nm. NCS1 protein was titrated with the freshly prepared acrylamide solution in the presence or absence of metals. Data were analysed using the Stern–Volmer equation:

(6)

where I₀ and I denote the emission intensity in absence and presence of the quencher acrylamide, respectively, K_SV is the Stern–Volmer quenching constant, and [Q] stands for quencher concentration.

Frequency Domain Lifetime and Time Resolved Anisotropy. Frequency domain lifetime and time resolved anisotropy measurements were carried out for both intrinsic Trp and 1,8-ANS: NCS1 complexes. Samples for fluorescence lifetime and time resolved anisotropy study contained 10 μM NCS1, 20 mM Tris buffer, pH 7.40, 1 mM EDTA, 1 mM Ca²⁺, and/or 120 μM Pb²⁺, 1 mM Mn²⁺ or 1 mM Hg²⁺. 40 μM 1,8-ANS was added to samples for time resolved measurements of 1,8-ANS: NCS1 complex. The measurements were carried out at room temperature (20°C) using ChronosFD spectrofluorometer (ISS, Champaign IL). For Trp study, apo and metal bound NCS1 samples were excited using a 280 nm laser diode and the emissions were collected through 305 nm long pass filter. A 300 nm laser diode was used for 1,8-ANS: NCS1 complex excitation and emission was collected through 400 nm long pass and 600 nm short pass filter. 1,4-bis(5-phenyloxazol-2-yl) benzene (POPOP) in ethanol was used as a reference in time resolved measurements (τ = 1.35 ns). Trp and 1,8-ANS:NCS1 data were analysed using a sum of three discrete exponential decay model according to Equation 7.⁵⁴ We attempted to analyse fluorescence decay data using other models such as a single exponential decay model, a sum of two discreet exponential decay model, or a Gaussian distribution model. Those models provided less satisfactory fits than the three discrete exponential decay model based on the χ² parameter and residuals. The data were analysed using Vinci software (ISS Inc., Champaign, IL).

(7)

where I(t) is the emission intensity, α_i is the pre-exponential factor, t is the time, and τ_i is the lifetime. The time resolved anisotropy data were analysed using a double-exponential decay model according to Equation 8⁵⁵ using Vinci software.

(8)

where Θ₁ and Θ₂ represent the rotational correlation times, and r₁ and r₂ are anisotropy amplitudes.

Nano electrospray Ionization Fourier transform ion cyclotron resonance mass spectrometry (nESI-FT-ICR MS). Ultrahigh resolution MS experiments were performed in a Solarix 7T FT-ICR MS spectrometer with an infinity ICR cell (Bruker Daltonics Inc., MA) and a custom-built nanoESI source operated in positive ion mode. Small aliquots (7-10 μL) of NCS1 samples were loaded into a glass capillary tip located at the instrument inlet. Typical operating conditions were 1600 V capillary voltage, skimmer 15 V, funnel RF amplitude 230 Vpp, transfer line RF 400 Vpp, and collision cell RF 1000 Vpp. Broadband MS spectra (500–600 scans) in the mass range of 600–3500 Da were collected at 1 MW data acquisition size. External mass calibration was performed using Agilent Tuning Mix calibration standard.

Native Trapped Ion Mobility Spectrometry − Mass Spectrometry (nESI-TIMS−MS). In TIMS, ions are separated in terms of their mobility where TIMS holds the ions stationary using an electric field against a moving column of gas. The separation in a TIMS device works by the center of the mass frame using the same principles as in a conventional IMS drift tube. The TIMS analyser was coupled to a maXis Impact Q-UHRToF instrument (Bruker Daltonics Inc., Billerica, MA). Data acquisition was controlled using in-house software, written in National Instruments Lab VIEW (2012, version 12.0f3) and synchronized with the maXis Impact acquisition program. TIMS separation was performed using nitrogen as a bath gas at ∼300 K, and typical P1 and P2 values are 1.8 and 0.6 mbar, respectively. The same RF (880 kHz and 200 − 350 Vpp) was applied to all electrodes, including the entrance funnel, the mobility separating section, and the exit funnel. A custom-built, nano electrospray ionization source was coupled to the TIMS − MS analyzer and was used for all analyses. A typical source voltage of 600−1200 V was used, and analyses were performed in positive ion mode. Mobility calibration was performed using the Tuning Mix calibration standard (G24221A, Agilent Technologies, Santa Clara, CA) in positive ion mode (e.g. m/z = 322, K₀ = 1.376 cm² V^–1 s^–1 and m/z = 622, K₀ = 1.013 cm² V^–1 s^–1).⁵⁶ The TIMS operation was controlled using in-house software, written in National Instruments Lab VIEW, and synchronized with the maXis Impact Q-ToF acquisition program. Theoretical CCS were calculated using the IMoS (v1.04b) package with nitrogen as a bath gas at ca. 300 K.^57,58

Results

Impact of non-physiological metals binding to NCS1 on Trp emission properties

Tryptophan emission is a powerful tool to probe structural transitions in proteins. We took advantage of the presence of two Trp residues (W30 and W103) in the sequence of NCS1 to (i) monitor metal binding to the protein and its impact on the protein's tertiary structure and (ii) determine the affinity of NCS1 for Pb²⁺, Mn²⁺ , and Hg²⁺. The fluorescence emission spectra of apoNCS1, NCS1 in the presence of physiological metals (Mg²⁺ or Ca²⁺), and non-physiological ions (Pb²⁺, Mn²⁺, or Hg²⁺) are shown in Fig. 2.

Fig. 2.

Intrinsic fluorescence emission of 10 μM NCS1 in the apo form (1 mM EDTA) and metal bound form (1 mM Mg²⁺and 1 mM EGTA, 1 mM Ca²⁺ and/or 120 μM Pb²⁺, 1 mM Mn²⁺, 1 mM Hg²⁺). Conditions: 10 μM NCS1 solubilized in 20 mM Tris buffer, pH 7.40, λ_exc = 295 nm and temperature of 20°C.

ApoNCS1 exhibits an emission maximum at 335 nm and the emission intensity increases by 18% upon Ca²⁺ addition and a further increase is observed upon addition of Mg²⁺ in the absence of Ca²⁺ due to distinct surrounding of Trp residues in apo-, Mg²⁺ , and Ca²⁺ bound NCS1 protein, in agreement with previously published data.²⁰ Interestingly, NCS1 emission spectra measured in the presence of non-physiological metal ions are distinct from those measured for apo- and Mg²⁺ or Ca²⁺ bound NCS1. Specifically, addition of non-physiological metals to NCS1 results in quenching of Trp emission and in case of Pb²⁺ to an additional hypsochromic shift of 5 nm with respect to the spectrum for the apo protein. Such shift in the emission maximum points towards a more hydrophobic surrounding of Trp residues upon complexation with Pb²⁺. Interestingly, the emission spectrum recorded upon Pb²⁺, Mn²⁺, or Hg²⁺ addition to Ca²⁺NCS1 exhibits nearly identical emission intensity as the spectra of Pb²⁺NCS1, Mn²⁺NCS1, or Hg²⁺NCS1 in the absence of Ca²⁺. This result indicates that non-physiological metal ions may either displace Ca²⁺ from the EF hands or they bind to a distinct site on the protein surface and trigger analogous changes in the protein structure in the absence and presence of Ca²⁺. Unlike Trp emission intensity, the Trp absorbance at 278 nm is not sensitive to the presence of Ca²⁺ and other ions studied here.

Equilibrium dissociation constants for non-physiological metals binding to NCS1

The equilibrium dissociation constants for Pb²⁺, Hg²⁺, Mn²⁺ association to NCS1 were determined either by direct titration, i.e. adding small aliquots of metal ion solution to apoNCS1, or using the displacement approach where small aliquots of metal ion solution were added to NCS1 in the presence of 1 mM Ca²⁺. Fig. S1(a–b) shows emission spectra for Ca²⁺NCS1 as a function of increasing concentration of Pb²⁺ and Mn²⁺. The direct and displacement titration curves were determined by monitoring changes in the intrinsic Trp emission intensity and are presented in Fig. 3 and Fig. S2.

Fig. 3.

Titration curves for Pb²⁺ left, Mn²⁺ (middle), Hg²⁺ (right) binding to apoNCS1 (in red), and Ca²⁺NCS1 (in black). Conditions: 10 μM apoNCS1 or 10 μM NCS1 and 1 mM Ca²⁺ in 20 mM Tris buffer, pH 7.40. The solid line represents the fit of the experimental data according to Equation 2 for Pb²⁺ and Mn²⁺ titration curve and Equation 3 for Hg²⁺ association to NCS1.

The titration curves for Pb²⁺ or Mn²⁺binding to apoNCS1 or Ca²⁺NCS1 were analysed using Equation 2 to determine the equilibrium dissociation constant for Pb²⁺ and Mn²⁺ binding to the apo protein and the apparent equilibrium dissociation constant for Pb²⁺ and Mn²⁺ binding to Ca²⁺NCS1, respectively. The measured equilibrium dissociation constants are listed in Table 1. Both approaches, direct titration and displacement approach, provide nearly identical equilibrium dissociation constants for Pb²⁺ and Mn²⁺ binding to apoNCS1 with K_{d, Pb²⁺} = 7.0 ± 1.0 nM and K_{d, Mn²⁺} = 34.0 ± 6.0 nM. The very good agreement between the K_d values obtained by both approaches supports the binding mechanism where Pb²⁺ and Mn²⁺ displaces Ca²⁺ from the EF-hands. The number of binding sites (n parameter in Equation 2) obtained by analysing the titration curves were approximately 1 for Pb²⁺ and Mn²⁺ binding to apoNCS1 and Ca²⁺NCS1. This value is lower than number of binding sites identified using MS spectrometry. The discrepancy is likely due to the fact that the fluorescence change due to the ion binding to individual binding sites may not be additive or binding to certain EF-hands may not lead to a measurable fluorescence signal. The titration curves for Hg²⁺ binding to both apoNCS1 and Ca²⁺NCS1 were fitted using a two binding site model, Equation 3, as the data analysis using one binding site model did not provide a satisfactory fit, suggesting the presence of at least two Hg²⁺ binding sites with distinct affinities. The analysis of titration curves for Hg²⁺ binding to apoNCS1 and Ca²⁺NCS1 provided nearly identical equilibrium dissociation constants, suggesting that Hg²⁺ does not displaces Ca²⁺ from EF-hands.

Table 1.

Emission λ_max and metal ions binding properties of metal: NCS1 complexes.

Metal Ion		Kd,appCa2+NCS1 (μM)	KdapoNCS1 (μM)	KdapoNCS1 (μM)	Kd,appCa2+NCS1:ANS (μM)
			Ca2+ displacement	direct titration
Ca2+a	332	n. a.	n. a.	0.096 ± 0.048	n. d.
Pb2+	330	68.0 ± 2.0 (1.0 ± 0.06)	0.006 ± 0.002	0.007 ± 0.004 (0.95 ± 0.05)	103 ± 23
Mn2+	335	340 ± 67 (1.1 ± 0.03)	0.034 ± 0.006	0.037 ± 0.003 (1.05 ± 0.05)	924 ± 185
Hg2+	335	5.1 ± 1.3 (0.6) 272 ± 126 (0.4)	n.a.	5.1 ± 2.6 (0.7) 315 ± 129 (0.3)	18.8 ± 3.0

^aValue from Ref.⁵³ Values in parenthesis represent the n value in Equation 2 for Pb²⁺ and Mn²⁺ titration and scaling factors in Equation 3 for Hg²⁺ titration.

Impact of non-physiological metals binding to NCS1 on the protein secondary structure

To determine whether the metal binding alters the secondary structure of the NCS1 protein, far-UV circular dichroism experiments were carried out for the apo and metal bound NCS1 at room temperature (Fig. 4).

Fig. 4.

Far-UV CD spectra of 10 μM NCS1 protein in apo and metal bound form. Experimental conditions as in the Fig. 2.

The CD spectra recorded between 200 and 250 nm show that binding of Ca²⁺ leads to a decrease in the CD signal at 220 and 208 nm, in agreement with previous data.⁵⁹ Addition of non-physiological metals to NCS1 causes a more negative CD signal at 208 and 220 nm, compared to apo- and Ca²⁺ bound protein, due to an increase of the protein α-helical content or rearrangement of α- helices. Notably, the CD spectra measured upon Pb²⁺, Mn²⁺, or Hg²⁺ addition to Ca²⁺NCS1 are nearly identical to that of Pb²⁺, Mn²⁺, Hg²⁺ bound NCS1 in the absence of Ca²⁺.

Solvent accessibility of Trp residues

To further investigate the changes in the protein structure due to the non-physiological metal binding, we have characterized the solvent accessibility of W30 and W103, by probing Trp emission quenching using a neutral quencher, acrylamide. The emission spectra of Pb²⁺NCS1 as a function of increasing concentration of acrylamide are shown in Fig. S1(d). The ratio of the fluorescence intensities in absence and presence of acrylamide was plotted as a function of the quencher concentration and the Stern–Volmer constant was determined according to Stern–Volmer equation (Equation 5). The Stern–Volmer plot and the calculated results are presented in Fig. 5 and Table 2, respectively.

Fig. 5.

Stern–Volmer plot for acrylamide quenching of Trp emission in apo and metal bound NCS1 protein. Conditions: 10 μM NCS1, 1 mM EDTA, 1 mM Mg²⁺, 1 mM Ca²⁺ and/or 120 μM Pb²⁺, 1 mM Mn²⁺, or 1 mM Hg² in 20 mM Tris buffer, pH 7.40. Solid lines correspond to the linear fit of the experimental data, and the error bar represents the standard deviations from three independent measurements.

Table 2.

K_sv and k_q values for NCS1 emission quenching by acrylamide.

	KSV (M–1)	<τ0> (ns)	<τq> (ns)	kq (M–1s–1)
ApoNCS1	2.82 ± 0.34	4.06	1.74	(6.9 ± 0.8) ×1010
Mg2+NCS1	3.36 ± 0.06	4.6	1.82	(7.3 ± 0.1) ×1010
Ca2+NCS1	3.11 ± 0.05	3.59	1.78	(8.7 ± 0.2) ×1010
Pb2+NCS1	1.66 ± 0.20	2.64	2.31	(6.3 ± 0.7) ×1010
Mn2+NCS1	2.54 ± 0.08	3.67	1.56	(6.9 ± 0.2) ×1010
Hg2+NCS1	2.51 ± 0.15	3.06	2.01	(8.2 ± 0.5) ×1010

Acrylamide quenching of Mg²⁺NCS1 exhibits the Stern–Volmer quenching constant (Ksv = 3.36 ± 0.06 M^–1) that is larger than that determined for Ca²⁺NCS1 (Ksv = 3.11 ± 0.05 M^–1) and for the apo protein (K_SV = 2.82 ± 0.34 M^–1). The increase in the K_SV value is consistent with the enhanced Trp accessibility in the presence of Mg²⁺ ion. The comparable K_SV values determined for apoNCS1 and Ca²⁺NCS1 indicate that even through Ca²⁺ triggers changes in the NCS1 secondary structure and an increase in Trp emission intensity, the solvent accessibility of W30 and W103 sidechains is not affected by the Ca²⁺ binding to EF-hands. This is consistent with the identical Trp emission maximum in apo and Ca²⁺NCS1, indicating similar polarity of Trp microenvironment in the metal free and Ca²⁺ bound protein.

On the other hand, binding of Pb²⁺ and to similar extent Hg²⁺ and Mn²⁺ binding, result in a decreased accessibility of Trp residues as evident from smaller K_SV values (Ksv = 1.66 ± 0.20 M^–1 for Pb²⁺NCS1). The linear Stern–Volmer plots and the ability of acrylamide to equally quench the emission intensity and Trp lifetime, suggest that the acrylamide acts as a collisional quencher for the Trp fluorophores. This is further supported by the values for bimolecular quenching constant, k_q ∼10¹⁰ M^–1s^–1 i.e. consistent with the diffusion-limited mechanism for acrylamide quenching of Trp fluorescence. The acrylamide quenching data were also analysed using a modified Stern–Volmer equation (Equation 9)⁶⁰:

(9)

where I₀ and I are the emission in the absence and presence of the quencher, respectively, and f_a is the fraction of fluorophore accessible to the quencher. Analysis of the quenching data for apo, Ca²⁺ , and Pb²⁺ bound NCS1 using the modified Stern–Volmer equation provided f_a values of ∼1 (Fig. S3 and Table S2) suggesting that W30 and W103 are fully quenched by acrylamide in the apoNCS1 and Ca²⁺ or Pb²⁺ bound protein. Interestingly, in case of Hg²⁺NCS1, at least one Trp residue is only partially accessible to the quencher as the f_a value for acrylamide quenching is 0.78. As the modified Stern–Vomer plot for Mn²⁺NCS1 is not linear, the data at low and high concentration of acrylamide were analysed separately using Equation 9. The f_a value obtained at high concentration of quencher is close to unity, f_a = 0.96, in agreement with the complete quenching of both Trp residues. The f_a value determined at the quencher concentrations below 250 mM is ∼0.57 suggesting limited accessibility of either one or both Trp residue in Mn²⁺NCS1.

Time resolved study of Trp emission

Trp lifetime is highly sensitive to interactions between Trp indole ring and surrounding residues and provides important insight into conformational dynamics of W30 and W103 surrounding in apo and metal bound forms of NCS1. The fluorescence decay data (Fig. S4) were analysed using a sum of three exponential decay model and determined parameters are summarized in Table 3. The binding of Mg²⁺ to NCS1 leads to an increase in the average lifetime with respect to the lifetime for the apo protein, <τ> = 4.06 ns for the apo protein and <τ> = 4.60 ns for the Mg²⁺ bound protein. Binding of other metals, including Ca²⁺, results in shorter average fluorescence lifetimes compared to apoNCS1. Namely, association of Pb²⁺ to NCS1 and Ca²⁺NCS1 leads to a decrease in the fluorescence lifetime with <τ> = 2.64 ns and 2.84 ns, respectively, mainly due to the decrease of τ₁ and increase of the pre-exponential factor α₁. Similar decrease in the average value of fluorescence lifetime was observed for other non-physiological metals (<τ> = 3.67 ns for Mn²⁺NCS1 and <τ> = 3.06 ns for Hg²⁺NCS1).

Table 3.

Fluorescence decay parameters for NCS1 in the presence/absence of metal ions.

	τ1 (ns)	α1 (%)	f 1 (%)	τ2 (ns)	α2 (%)	f 2 (%)	τ3 (ns)	α3 (%)	f 3 (%)	<τ> (ns)	χ2
ApoNCS1	0.64	49.65	12.70	3.36	39.33	60.00	6.73	11.00	26.30	4.06	1.60
Mg2+NCS1	0.43	51.55	10.00	3.79	44.51	49.67	7.16	11.50	33.00	4.60	1.10
Ca2+NCS1	0.60	34.45	12.95	2.82	47.10	50.20	5.38	18.50	36.90	3.59	1.10
Pb2+NCS1	0.03	96.67	23.77	1.79	1.98	28.83	4.34	1.35	47.50	2.64	1.20
Ca2+Pb2+NCS1	0.04	90.24	13.00	1.39	4.93	24.00	3.98	23.50	44.00	2.84	1.13
Mn2+NCS1	0.21	62.91	10.00	2.80	23.82	33.33	5.84	13.20	56.70	3.67	1.18
Ca2+Mn2+NCS1	0.26	51.85	10.00	2.84	40.74	27.00	6.55	7.40	63.00	3.61	1.04
Hg2+NCS1	0.17	71.50	12.00	1.79	15.78	31.00	4.30	12.40	56.30	3.06	1.15
Ca2+Hg2+NCS1	0.15	60.00	4.50	1.47	22.08	28.50	4.10	18.40	66.50	3.18	0.90

α_i denotes the amplitude factors and f_i  corresponds to fractional intensity

To determine the impact of non-physiological metals binding on NCS1 oligomerization, we measured the anisotropy decay of Trp residues in apo, and metal bound NCS1 using frequency domain approach (Fig. 6). The data were analysed using a sum of two rotational correlation times: Θ_int corresponds to the rotational correlation time of the fluorophore and Θ_prot reflects the rotational correlation time of the protein. The theoretical Θ value was calculated based on the Stokes—Einstein equation (Θ_calc = ηV/RT)⁵⁰ where η is the viscosity of water (η = 1.02 cP),⁵⁰ V is the hydrodynamic molecular volume, R is the ideal gas constant, and T is the temperature (293 K). Based on molar weight of NCS1 of 23 kDa and considering the shape of NCS1 to be spherical, we estimated the rotational correlation time for the NCS1 monomer to be 12 ns and the NCS1 dimer to be 24 ns, respectively. The measured rotational correlation times for the NCS1 samples in the presence and absence of metals ions are listed in Table 4. For all samples, a sub nanosecond rotational correlation time, Θ_int, was observed and can be attributed to a local motion of Trp side chains around C_α-C_β or C_β-C_γ bond as its values is comparable to the reported values for Trp sidechain mobility in troponin C and its mutants (Θ_int = 0.33 ns to 1.46 ns).⁶¹ The binding of metals to NCS1 has relatively small impact on the Trp side chain mobility as Θ_int increases from 0.11 ns for the apo protein to 0.39 ns for the Ca²⁺ bound protein. The values measured for NCS1 in complex with non-physiological metals are somewhat smaller than the value determined for Ca²⁺NCS1 and range from 0.22 ns to 0. 33 ns. These value indicate a decreased mobility of Trp sidechains in NCS1 in the presence of Ca²⁺ and non-physiological metals possibly due to a more compact environment of Trp residues upon metal binding to EF-hands. The longer rotational correlation time, Θ_prot, can be attributed to a tumbling motion of the entire protein and reflects the changes in the protein's oligomerization state. Shorter Θ_prot was observed for apoNCS1 (Θ_prot = 14.3 ns) and Mn²⁺NCS1 (Θ_prot = 13.0 ns) and this value is comparable to Θ_calc calculated for the NCS1 monomer suggesting monomeric form of NCS1 in the apo form and in the presence of Mn²⁺. The value of, Θ_prot doubles for NCS1 in complex with Ca²⁺, Pb²⁺ , and Hg²⁺, Θ_prot = 24.0, 24.6 and 25.0 ns, respectively, indicating that Ca²⁺, Pb²⁺ , and Hg²⁺ binding to the protein triggers the dimerization of NCS1.

Fig. 6.

Time resolved anisotropy data for apo and metal bound NCS1. Solid and open symbols represent Δ phase shift and amplitude ratio, respectively, measured as a function of modulation frequency. Solid line represents the non-linear fit of the experimental data using a two exponential decay model. Conditions: 10 μM NCS1, 20 mM Tris buffer, pH 7.40, 1 mM EDTA, 1 mM Ca²⁺, 120 μM Pb²⁺, 1 mM Mn²⁺ or 1 mM Hg²⁺. Temperature 20 °C.

Table 4.

Anisotropy decay parameters for NCS1 and 1,8-ANS: NCS1 complex in the presence/absence of metal ions.

	NCS1					1,8-ANS: NCS1
Metal Ion	Θint (ns)	f 1 (%)	Θprot (ns)	f 2 (%)	χ2	ΘANS (ns)	f 1 (%)	Θprot (ns)	f 2 (%)	χ2
Apo	0.11 ± 0.01	88	14.3 ± 1.6	12	0.5	0.11	33	14 ± 2.7	67	0.3
Ca2+	0.39 ± 0.02	75	24.0 ± 0.6	25	1.0	0.11	23	18 ± 5.0	77	0.6
Pb2+	0.21 ± 0.04	86	24.6 ± 1.0	14	0.7	0.11	17	23 ± 6.2	82	0.4
Mn2+	0.23 ± 0.07	80	13.0 ± 1.3	20	1.1	0.11	45	10 ± 2.8	55	0.6
Hg2+	0.33 ± 0.05	88	25.0 ± 1.6	12	0.8	0.11	29	37.0 ± 3.5	71	0.6

Non-physiological metal induced changes in NCS1 surface hydrophobicity

1,8-ANS is a hydrophobic fluorescence probe that is nearly non-fluorescent in the aqueous solution and it's quantum yield increases in the non-polar environment. This probe is often used to monitor conformational changes in proteins as it binding to hydrophobic cavities on protein surface results in fluorescence increase and hypsochromic shift in the emission maximum.⁶² It was shown previously by Aravind et al.²⁰ that Ca²⁺ and Mg²⁺ association to NCS1 promotes binding of 1,8-ANS molecules to hydrophobic cavities on NCS1 surface and a similar impact of Ca²⁺ binding on protein hydrophobic surfaces was reported for Ca²⁺ binding to DREAM.⁶ Here, we investigated whether binding of non-physiological metals impacts NCS1 affinity for 1,8-ANS. The emission spectrum of 1,8-ANS: apoNCS1 complex exhibits a maximum at 466 nm and the spectra for both apo and metal bound complexes are shown in Fig. 7. Addition of Ca²⁺ or Mg²⁺, has a distinct impact on the emission spectrum of 1,8-ANS: NCS1 complex, as emission intensity decreases for 1,8-ANS: Mg²⁺NCS1 complex and increases for 1,8-ANS: Ca²⁺NCS1 complex, in agreement with previously published results.²⁰ Addition of non-physiological metals leads to an increase in the emission intensity of the 1,8-ANS: NCS1 complex without concomitant change in the emission maximum. These results demonstrate that association of non-physiological metals leads to an increase in 1,8-ANS binding that can be associated with the enhanced solvent exposed hydrophobic surfaces. The emission spectra for Pb²⁺, Mn²⁺, Hg²⁺ bound 1,8-ANS: NCS1 complexes show almost identical intensity in presence and absence of Ca²⁺ ion which is consistent with the Trp emission and far-UV CD results.

Fig. 7.

Emission spectra of NCS1:1,8-ANS complexes in absence and presence of metal ions. Samples were excited at 350 nm. The experimental conditions as in Fig. 2.

We took advantage of the distinct emission intensity of metal free, and metal bound 1,8-ANS: NCS1 complex, to characterize the equilibrium dissociation constants for non-physiological metal binding to this protein. To determine the affinity of the non-physiological metals, small aliquots of metal solution were added to 1,8-ANS: Ca²⁺NCS1 complex and the changes in 1,8-ANS emission were measured (1,8-ANS: Ca²⁺NCS1 emission traces as a function of Pb²⁺ concentration are shown in Fig. S1(c)). The titration curves for non-physiological metals binding to NCS1 in presence of 1,8-ANS are shown in Fig. S5 and the determined apparent dissociation constants are summarized in Table 2. The titration curve for Hg²⁺ and Mn²⁺ binding to 1,8-ANS: Ca²⁺NCS1 complex are hyperbolic and provide an apparent dissociation constant of 18.8 ± 3.0 μM and 924.0 ± 185.0 μM, respectively. The recovered apparent dissociation constants are larger than that determined by monitoring the emission of intrinsic Trp residues, suggesting the 1,8-ANS association to the hydrophobic sites on the protein surface modulates the affinity for metals. Unlike Mn²⁺ and Hg²⁺, the titration curve for Pb²⁺ binding to Ca²⁺NCS1 in the presence of 1,8-ANS is cooperative with = 103.0 ± 23.0 μM and n = 1.57 ± 0.06, suggesting allosteric interactions between 1,8-ANS binding site(s) and Pb²⁺ binding sites.

Identification of ANS binding sites

To better understand the impact of non-physiological metals binding to NCS1 on the protein's hydrophobic surface, we have identified 1,8-ANS binding sites through ligand docking simulations using Auto dock software version 4.2.⁶³ The NMR structure of Ca²⁺ bound NCS1 (PDB entry 2LCP) was used as the static macromolecule and the 1,8-ANS ligands as flexible molecules. The binding sites were identified based on the free energy values and the presence of favorable hydrophobic and electrostatic interactions between the fluorescent probe and the protein. Non-specific binding sites were excluded from this study due to unfavorable binding energy and hydrophobic/polar interactions. Two potential 1,8-ANS binding sites were found (Fig. S6) with the first binding site located at the interface between EF-hand 3 and EF-hand 4. In this site, 1,8-ANS molecule is surrounded by mostly hydrophobic residues (L97, L101, A104, L107, Y108, F169, I179, V180, A182, L183). Site-1 also contains three polar residues (K100, Q181, S184) with positively charged K100 sidechain stabilizing an 1,8-ANS molecule through ionic interactions with the sulfate group. The second binding site is found in the interface between EF- 1 and EF-2 with 1,8-ANS molecule surrounding being more polar as three polar residues (Q29, K36, D37, Q54) and seven hydrophobic residues (W30, G33, F34, L43, I51, F55, F85) surround the probe. In addition to hydrophobic interactions, the fluorescent probe forms an ionic bond with the sidechain of the nearby K36 residue.

Determination of lifetime for 1,8-ANS bound to NCS1: metal complexes

To further characterize the properties of the 1,8-ANS binding sites, we have determined the fluorescence lifetime of 1,8-ANS bound to the protein in the absence and presence of metal ions (Fig. S7). Analysis of lifetime data using sum of three discrete exponential decay model provide three lifetimes (Table S3). The shortest lifetime corresponds to an unbound 1,8-ANS molecule (τ₁ = 0.28 ns) whereas additional two lifetimes reflect the fluorescence decay of 1,8-ANS bound to at least two binding sites on the protein surface. We attribute the longest lifetime, τ₃ = 13.2 ns to 17.2 ns to 1,8-ANS associated to the binding site between EF-hand 3 and EF-hand 4, as the probe in this site is surrounded by hydrophobic residues, resulting in the stabilization of the 1,8-ANS excited state. The modulation of τ₃ and α₃ values by metal binding to NCS1 indicates the polarity of this site is influence by the metal binding to EF-hands. Analogously, the τ₂ reflects the fluorescent probe that is located in the binding site between EF-hand 2 and EF-hand 1 as its surrounding is formed by more polar residues compared to the binding site located between EF-hand 3 and 4. The value of both lifetimes, τ₂ and τ₃, and corresponding α₂ and α₃ values decrease upon Mg²⁺ binding to NCS1, which agrees with steady state data that shown the lowest emission intensity for 1,8-ANS bound to Mg²⁺NCS1. On the other hand, binding of Ca²⁺ and non-physiological metals leads to the increase in both lifetimes and pre-exponential factors, suggesting that binding of non-physiological metals has analogous impact on exposure of hydrophobic surfaces on NCS1 surface as Ca²⁺ ion.

Determination of rotational correlation times for 1,8-ANS: NCS1 complexes in time resolved anisotropy measurements provides additional way to determine oligomerization state of NCS1 in the presence or absence of metals. The time resolved anisotropy data (Fig. S8) were analysed using a sum of two rotational correlation times with Θ_ANS, corresponding to the unbound probe, and Θ_prot reflecting the tumbling motion of the entire protein in the presence of 1,8-ANS probe and recovered parameters are listed in Table 4. The Θ_ANS value was fixed to be 0.11 ns whereas other fitting parameters were allowed to vary during the analysis. For the apoNCS1, Mn²⁺ NCS1 and Pb²⁺NCS1, the Θ_prot values is nearly identical to those determined by monitoring time resolved anisotropy of intrinsic Trp probe, in agreement with the fact that these values reflect the size of the protein and thus are independent of the fluorescent probe. The small value of Θ_prot 14 ns, observed for apo protein and Mn²⁺NCS1 indicate that the protein is predominantly monomeric, whereas Θ_prot = 23 ns determined for Pb²⁺NCS complex agrees with the dimeric form of Pb²⁺ NCS1. Interestingly, the Θ_prot value measured for Ca²⁺NCS1, Θ_prot = 18 ns, is between the values expected for the monomeric and dimeric proteins. Therefore, the time resolved anisotropy curves were fitted using the sum of three exponential decay model. This approach provided Θ_prot1 = 10 ns and Θ_prot2 = 20 ns, suggesting the presence of the monomeric and dimeric forms of 1,8-ANS: Ca²⁺NCS1. The result of two exponential decay analysis for 1,8-ANS: Hg²⁺NCS1 complex provide Θ_prot = 37 ns. This value is significantly larger than expected for a dimeric protein and may indicate distinct conformation of 1,8-ANS: NCS1 in the presence of Hg²⁺.

Native FT-ICR MS and TIMS-MS analysis of Pb²⁺, Mn²⁺, and Hg²⁺ binding to NCS1

Newly developed mass spectrometry techniques became excellent tools in the field of structural biology due to their capability to investigate protein structure and structural changes including metal ions interactions with metalloproteins.⁶⁴ Specifically, nESI-FT-ICR MS allows to quantify the binding stoichiometry and identify presence of distinct proteoforms based on the charge state distribution.⁶⁴ TIMS-MS is well suited for monitoring impact of metal ions on protein structure in terms of changes in collisional cross section (CSS) values that reflect protein three-dimensional shape. The nESI-FT-ICR MS analysis of the apo form of NCS1 resulted in a neutral mass of 22,976 Da, in good agreement with the theoretical average neutral mass (22 975.6 Da) calculated based on the protein sequence (Fig. S9a bottom). The native MS spectra of NCS1 in the presence of Ca²⁺ or Mn²⁺ (Fig. S9a–c), and Pb²⁺ (Fig. 8) show a bimodal distribution of charge states with a narrow distribution centered at the 8 + charge state. Additional broad envelopes of low abundant species centered at 19 + for NCS1 in the presence of Ca²⁺ and 16 + for NCS1 in the presence of Pb²⁺, Mn²⁺ are also observed. The narrow distribution of low charges (7+ to 9+) are characteristic of native like conformations whereas the higher charges can be attributed to partially destabilized structures. Interestingly, NCS1 remains in the metal, predominantly Ca²⁺ bound form in all charge states observed including higher charge states (16+ to 20+) pointing towards enhanced structural stability and affinity for bivalent metals of this protein (Fig. 8 and Fig. S9a–c).

Fig. 8.

Expanded view of the 8 + charge state with annotated adduct species and the isotopic pattern of the ion [M•2Pb•2Ca]⁸⁺ shown in the inset (A). Typical broadband nESI-FT-ICR MS spectrum of NCS1 in the presence of Pb²⁺ (B). Deconvoluted spectrum showing the neutral species. The isotopic pattern shown in the inset confirms the theoretical isotopic average mass of the NCS1 protein 22.976 kDa (C).

A closer inspection of the 8 + charge state for NCS1 in the presence of Ca²⁺ (Fig. S9a) reveals the presence of several adducts including apo protein of a low abundance, and one to four Ca²⁺ bound NCS1 adducts with the NCS1: (Ca²⁺)₃ complex being the most abundant, in agreement with the Ca²⁺ binding to three EF-hands (2, 3, and 4). The presence of NCS1: (Ca²⁺)₄ adduct indicates that fourth Ca²⁺ ion can either bind to the non-functional EF-1 as observed previously for Ca²⁺ binding to EF-hand 1 in DREAM ⁶⁵ or to a distinct binding site on the protein surface in the gas phase. The analysis of the charge state 8 + for NCS1 in the presence of Pb²⁺ is more complex (Fig. 8), as it contains NCS1: (Ca²⁺)_1-4 adducts and NCS1: (Pb²⁺)_1-3: (Ca²⁺)_1-3 adducts. As observed for NCS1 samples in the presence of Pb²⁺, the MS analysis of NCS1 in the presence of Mn²⁺ (Fig. S9b) reveals several adducts with NCS1: (Mn²⁺)₃ being the most abundant. Notably, the total metal binding stoichiometry of 4:1 is consistent with four EF-hands being the metal binding sites. Also, the absence of adducts with 4 Ca²⁺ and additional Pb²⁺ or Mn²⁺ ions bound to the protein agrees with the mechanism when Pb²⁺ or Mn²⁺ displaces Ca²⁺ from the EF-hands. Interestingly, we were unable to detect peaks corresponding to the NCS1 dimer in the MS spectra for Ca²⁺NCS1 and Pb²⁺NCS1. Such absence of the dimeric form is consistent with a low stability of the NCS1 dimer in the gas phase.

The native nESI-TIMS—MS data provided additional insight into the protein's conformation in the presence of various metals. The CCS values of ∼ 2145 Ų to 2164 Ų determined for the charge state 8 + of NCS1 in the presence of Ca²⁺, Pb²⁺, and or Mn²⁺ (Fig. S10) are nearly identical to the CCS values observed for 8 + charged state of apo protein 2153 Ų and 2158 Ų (Table S4) and the CCS value of 2428 Ų calculated based on the Ca²⁺ bound NCS1 structure (PDB entry 2LCP) suggesting small impact of metal binding to EF-hands on the overall conformation of NCS1. Such negligible influence of metal binding on the overall protein structure may be a common feature for NCS proteins. These observations are in good agreement with reported NMR structures of one and two Ca²⁺ ion bound recoverin (PDB entry 4M2O and 4YI8) in the non-myristoylated form.

Discussion

The results reported here show NCS1’s capability of binding various abiogenic metals including neurotoxic Pb²⁺ and Hg²⁺ as well as biogenic Mn²⁺ with sub-nanomolar affinity. Interestingly, the association of non-physiological metals triggers conformational changes that are distinct from those triggered by physiological metal ions Ca²⁺ or Mg²⁺ pointing towards a high conformational flexibility of NCS1. EF-hand protein CaM has been frequently used as a model protein to investigate protein interactions with ions as its EF-hands bind various biogenic and abiogenic metals including divalent ions Pb²⁺, Cd²⁺, Zn²⁺, Hg²⁺, Sr²⁺, Mn²⁺ and trivalent ions La³⁺, Tb³⁺, and Sm³⁺.⁹ The interactions of non-physiological metal ions with CaM seem to be complex as they either directly compete with Ca²⁺ to bind to EF-hands and/or bind to so called auxiliary sites.⁶⁶ Interactions between CaM and Pb²⁺ attracted significant attention as CaM was proposed to be a putative target for Pb²⁺ and Pb²⁺ neurotoxicity.⁶⁷ As a soft-to-borderline metal ion, Pb²⁺ binds to oxygen ligands from sidechains of Asp and Glu residues and can adopt either hemi-directed geometry as observed in the X-ray structure of Pb²⁺ bound CaM (PDB entry 1N0Y) or holo-directed geometry as proposed for Pb²⁺ binding to CaM using computational approach.^67,68

The nearly identical equilibrium dissociation constants determined by direct titration and Ca²⁺ displacement and the stoichiometry data showing maximum total number of metals binding to NCS1 to be four demonstrate that Pb²⁺ replaces native Ca²⁺ in EF-hands of NCS1 and apoNCS1 has 15 times higher affinity for Pb²⁺ compared to its native ion Ca²⁺. Both Pb²⁺NCS1, and 1,8-ANS: Pb²⁺NCS1 complex exhibit distinct fluorescence properties from Ca²⁺NCS1 and 1,8-ANS: Ca²⁺NCS1 complex, suggesting that Pb²⁺ binding does not trigger identical conformational changes in NCS1 as determined for Ca²⁺. This may be due to the distinct coordination geometry, as Pb²⁺ prefers six ligands in hexagonal geometry whereas Ca²⁺ favors a pentagonal bipyramidal configuration with seven-oxygen ligands.⁶⁷ Notably, unlike Pb²⁺ association to NCS1, Pb²⁺ binding to 1,8-ANS: NCS1 complex is cooperative, n = 1.57 ± 0.06, and Pb²⁺ binds to 1,8-ANS: NCS1 with an affinity constant i.e. two times smaller than that for Pb²⁺ binding to NCS1 in the absence of 1,8-ANS pointing towards the presence of allosteric interactions between 1,8-ANS binding site and Pb²⁺ bound EF-hands. This is reminiscent of a non-cooperative Ca²⁺ association to NCS1 and its cooperative binding to myristoylated NCS1.⁶⁹ Although the Trp residues in Pb²⁺NCS1 show distinct photo-physical properties compare to Trp residues in Ca²⁺NCS1 complex, Pb²⁺ association to EF-hands triggers NCS1 dimerization as observed previously for Ca²⁺ suggesting, that long-range structural changes required for protein dimerization are analogous in Ca²⁺NCS1 and Pb²⁺NCS1.

Analogous to Pb²⁺, Mn²⁺ also prefers six oxygen ligands in its coordination geometry.⁸ To predict Mn²⁺ binding sites in NCS1 we have used the Metal Ion-Binding Site Prediction and Docking Server (MIB).⁷⁰ The docking results are summarized in Fig. S11. Potential residues for Mn²⁺ coordination overlap with the residues from EF-hand 2, 3, and 4 that coordinate Ca²⁺ in the NCS1 X-ray structures and together with the MS data, these results suggest that Mn²⁺ ions occupy three EF-hands in NCS1 in a similar position as Ca²⁺ ions. This is consistent with the X-ray structure of CaM with Mn²⁺ bound to the N-terminal EF-hands shows that Mn²⁺ binds to a position i.e. similar to that for Ca²⁺ but unlike Ca²⁺ coordination, the Glu residue in the 12^th position serves as a monodentate ligand.⁸ The MS data indicate binding up to four Mn²⁺ ions to EF-hands in NCS1. In the gas phase, Mn²⁺ ions can bind to EF-hand 1 that is inactive in the solution or additional Mn²⁺ binding may be due to non-specific Mn²⁺ interactions with NCS1. Also, we cannot exclude that more than one Mn²⁺ ion binds to a single EF-hand as observed in the Mn²⁺ bound CaM structure (PDB entry 3UCT). Interestingly, although photo-physical properties of Trp residues in Mn²⁺NCS1 complex and properties of Mn²⁺NCS1:1,8-ANS complex are similar to those determined for Pb²⁺NCS1, i.e. increased quenching of Trp emission and increased binding of the 1,8-ANS probe in the presence of Pb²⁺ or Mn²⁺, the overall conformational changes triggered by Mn²⁺ and Pb²⁺ binding to EF-hands are distinct as Mn²⁺ binding to EF-hands in NCS1 does not undergo dimerization and the Mn²⁺ binding to NCS1:1,8ANS complex does not exhibit cooperativity.

The relatively high affinity of Hg²⁺ for NCS1 is surprising as this metal favors sulfur atoms in linear arrangement as the coordination ligands, although other coordination geometries such as tetrahedral and pseudo-octahedral with sulfur, nitrogen, and/or oxygen atoms as ligands were observed in structures of metalloproteins deposited in PDB.⁷¹ The prediction of Hg²⁺ binding sites in NCS1 using MIB was inconclusive as the potential coordination residues vary significantly based on different PDB structures used for the prediction. Unlike well documented Pb²⁺ affinity for EF-hands in CaM and troponin C, significantly less is known about Hg²⁺ interactions with EF-hand proteins. Using NMR spectroscopy, Ouyang and Vogel reported that Hg²⁺ binds to EF-hand 4 and EF-hand 1 in CaM with EF-hand 4 having larger affinity for Hg²⁺ compared to EF-hand 1.⁷² Nearly overlapping titration curves for Hg²⁺ binding to apoNCS1 and Ca²⁺ NCS1 provide comparable equilibrium K_d values, suggesting that Hg²⁺ binding sites on NCS1 surface are distinct from these for Ca²⁺. It is also possible that Hg²⁺ can bind to EF-hand(s) in the presence Ca²⁺. Considering that Hg²⁺ binding to NCS1 leads to an increase in the protein secondary structure, enhances 1,8-ANS binding to NCS1 and most importantly, triggers protein dimerization in a similar way as Ca²⁺ does, it seems to be unlikely that Hg²⁺ binding site is distinct from that for Ca²⁺. We propose that Hg²⁺ occupies EF-hands in the apo protein and in Ca²⁺ bound NCS1 shares EF hand with Ca²⁺. Although Hg²⁺ ionic radius of 1.02 Å is comparable to the Ca²⁺ ionic radius of 0.99 Å, suggesting sterically restrictions for two metals binding to a single EF hand, Hg²⁺ smaller coordination number and a plasticity of EF hands can promote binding of both metals. The binding of two ions to EF hands in NCS1 was reported previously by Tsvetkov et al.⁷³ who proposed that Zn²⁺ binds to EF-hand 2 and 4 in NCS1 in the presence of Ca²⁺. In such case, Hg²⁺ binding to Ca²⁺ NCS1 partially reverses the conformational changes triggered by Ca²⁺, as the emission and CD spectra of Hg²⁺NCS1 are nearly identical in the presence and absence of Ca²⁺.

Importantly, the distinct emission properties of NCS1 in apo, Mg²⁺, Ca²⁺, or non-physiological metal ions bound form indicate that NCS1 carry enhanced structural flexibility as it can adopt several conformations based on the various coordination geometry of the metals and/or distinct number of EF-hands being in the metal bound form. Such conformational plasticity is likely to be crucial for functional properties of NCS1 as it may facilitate its distinct response to temporary changes in intracellular Ca²⁺ concentration as well as enable NCS1 to efficiently interact with various binding surfaces on effector proteins and provide molecular basis for discrimination among various intracellular targets. Indeed, analogous structural diversity was previously reported for CaM ⁷⁴ and our data indicate that this feature may be also common among NCS proteins.

Also, the comparison of metal selectivity between CaM and two members of the NCS1 family, NCS1 and DREAM (Table S5), show that although all three proteins carry canonical Ca²⁺ binding EF-hands and all bind bivalent Pb²⁺ and trivalent Tb³⁺ with sub-millimolar affinity, EF-hands in NCS1 and DREAM demonstrate increased selectivity as NCS 1 does not bind Cd²⁺ and DREAM does not exhibit sub-millimolar affinity for divalent ions such as Mn²⁺ and Hg²⁺ that bind to EF-hands in CaM.⁹ Previously it was proposed that several factors impact the selectivity of EF-hands for various ions including (i) flexibility of metal binding loop; (ii) coordinating side chains in the binding loop and their hydrophobic contact (iii) hydrogen bonding network within the coordination loop (iv) long range electrostatic interactions in the proteins.⁷⁵

The average net charge of the Ca²⁺ binding loops in EF-hands in CaM is more negative, with the individual net charges of −2, −4, −2, −5, compared to NCS proteins (Fig. S10). However, NCS1 and DREAM Ca²⁺ binding loops show comparable net charges (−2 and −3, depending on the EF-hand), so the net negative EF-hand charge cannot explain the different metal selectivity observed between DREAM and NCS1. According to the ‘side chain interaction model’ packing and multiple contacts between sidechain of coordinating residues contribute to ionic size selectivity.⁷⁶ To estimate the packing of the individual EF hands we have determined the sum of the van der Waals volumes (vdW) for Ca²⁺ coordinating residues in each EF hand (Fig. S12). The total vdW volume of coordinating residues in DREAM and NCS1 is larger than that for coordinating residues in CaM EF-hands, suggesting an increased rigidity of the EF hand sidechains in DREAM and NCS1 that can potentially contribute to the increase in the metal binding energy and decreased affinity for certain ions. Also an additional Gly residue in the fourth position in the sequence of Ca²⁺ binding loops in CaM (Fig. S12) that is absent in the coordinating loops in DREAM and NCS1 EF hands, can provide additional loop flexibility, potentially allowing accommodation of various metals with different ionic radii, though Drake et al. reported relatively modest impact of Gly in the sixth position on ion affinity and selectivity of the EF-hand in E. coli galactose binding protein.⁷⁷ The introduction of a negatively charged sidechain in the ninth position of EF-hand loop, so called ‘gate away’ position, was shown to lead to a decreased affinity for divalent ions and size selectivity.⁷⁸ Only EF-2 in CaM and NCS1 carry negatively charge residue in the ninth position, so it is unlikely that the residue in this position enhances ion selectivity in DREAM and NCS1. Importantly, the metal binding selectivity is not only modulated by coordinating residues in individual EF-hands as the protein matrix influences the structure and rigidity of the metal binding site⁷⁹ and long range electrostatic interactions and protein electric field were proposed to impact cation selectivity of EF hands in proteins.⁷⁴

Summary

The data presented here indicate that in addition to physiological metals Ca²⁺ and Mg²⁺, NCS1 is capable of binding other biogenic and abiogenic metals with an affinity i.e. superior to that for Ca²⁺. Although, the impact of Pb²⁺ and Hg²⁺ on NCS1 structure in terms of Trp emission and secondary structure stabilization is distinct from Ca²⁺, both non-physiological metals promote NCS1 dimerization and thus their binding may modulate NCS1 interactions with intracellular partners in a similar way as Ca²⁺ ion. There results indicate that in addition to CaM, NCS1 may serve as a molecular target for toxic Pb²⁺ and Hg²⁺ as well as for biogenic metal Mn²⁺ and interactions between non-physiological metals and NCS1 are likely to contribute to the neurotoxic effects of those metals under acute and or chronical exposure.

Conflicts of interest

There are no conflicts to declare.

Data availability statement

The data underlying this article are available in the article and in its online supplementary material.