Lanthanides Report Calcium Sensor in the Vestibule of Ryanodine Receptor

Article

Lanthanides Report Calcium Sensor in the Vestibule of Ryanodine Receptor Sa´ndor Sa´rko¨zi,1 Istva´n Koma´romi,2 Istva´n Jo´na,1 and Ja´nos Alma´ssy1,* 1 Department of Physiology, Faculty of Medicine and 2Division of Clinical Laboratory Science, Department of Laboratory Medicine, Faculty of Medicine, University of Debrecen, Debrecen, Hungary

ABSTRACT Ca2þ regulates ryanodine receptor’s (RyR) activity through an activating and an inhibiting Ca2þ-binding site located on the cytoplasmic side of the RyR channel. Their altered sensitivity plays an important role in the pathology of malignant hyperthermia and heart failure. We used lanthanide ions (Ln3þ) as probes to investigate the Ca2þ sensors of RyR, because they specifically bind to Ca2þ-binding proteins and they are impermeable to the channel. Eu3þ’s and Sm3þ’s action was tested on single RyR1 channels reconstituted into planar lipid bilayers. When the activating binding site was saturated by 50 mM Ca2þ, Ln3þ potently inhibited RyR’s open probability (Kd Eu3þ ¼ 167 5 5 nM and Kd Sm3þ ¼ 63 5 3 nM), but in nominally 0 [Ca2þ], low [Eu3þ] activated the channel. These results suggest that Ln3þ acts as an agonist of both Ca2þ-binding sites. More importantly, the voltage-dependent characteristics of Ln3þ’s action led to the conclusion that the activating Ca2þ binding site is located within the electrical field of the channel (in the vestibule). This idea was tested by applying the pore blocker toxin maurocalcine on the cytoplasmic side of RyR. These experiments showed that RyR lost reactivity to changing cytosolic [Ca2þ] from 50 mM to 100 nM when the toxin occupied the vestibule. These results suggest that maurocalcine mechanically prevented Ca2þ from dissociating from its binding site and support our vestibular Ca2þ sensor-model further.

INTRODUCTION Ryanodine receptor (RyR) is a Ca2þ-gated Ca2þ release channel of the sarco/endoplasmic reticulum. Ca2þ-dependent regenerative activation of RyR channels is the underlying mechanism for the process termed ‘‘Ca2þ-induced Ca2þ release’’, which is the essential step of excitation-contraction coupling (ECC) in the heart. It plays an amplifying role in skeletal muscle ECC and in Ca2þ wave propagation (for example, in pancreatic acinar cells (1–3)). In single channel experiments, skeletal muscle RyR (RyR1) is closed at nanomolar, activated by micromolar, and inhibited by higher than 100 mM cytosolic [Ca2þ]. This biphasic response to Ca2þ is mediated by two regulatory Ca2þ binding sites (Ca2þ sensors) located on the cytosolic region of the channel: a high-affinity activating site (Kd ¼ 9.4 mM) and a low-affinity inactivating site (Kd ¼ 298 mM), both characterized by a Hill coefficient of 1 (4–6). Many other ligands (ATP, caffeine, Mg2þ) modulate RyR activity, but Ca2þ is the most important one, because Ca2þ is the only ligand whose changing concentration plays a crucial role

Submitted September 27, 2016, and accepted for publication March 23, 2017. *Correspondence: [email protected] Editor: Eric Sobie. http://dx.doi.org/10.1016/j.bpj.2017.03.023

in the physiology of ECC by triggering Ca2þ-induced Ca2þ release. Due to the central role of Ca2þ activation in RyR function and to understand RyR regulation in health and disease, defining the molecular identity and the location of Ca2þ sensors in RyR sequence is essential. Defective Ca2þ regulation of RyR has been demonstrated in various muscle diseases, including malignant hyperthermia and heart failure (7–15). Recently, the three-dimensional (3D) structure of RyR has been solved at near-atomic resolution using cryo-electron microscopy (16–18) (see Fig. 1). These high-resolution structures show that EF-hand Ca2þ binding domains (potential Ca2þ sensors) are located at the C-terminal end of the central pillar in the cytoplasmic region of RyR (named ‘‘core-solenoid’’). The core solenoid lies very close to the S2-S3 linker helix and interacts with the C-terminal domain, which is linked to the S6 transmembrane helix. This helix forms the channel’s gate and the central pore. Apparently, this architecture provides a tight allosteric coupling model between the core solenoid and the gate and suggests a mechanism for Ca2þ-induced gating of the channel. In the presence of Ca2þ, the conformation of the EF-hand domain changes, which coincides with the rotation of the last two helices in the core solenoid that leads to the rearrangement of the C-terminal domain and the S2-S3 linker. These

Ó 2017 Biophysical Society.

Biophysical Journal 112, 2127–2137, May 23, 2017 2127

Sa´rko¨zi et al.

FIGURE 1 3D model of RyR1 structure (PDB:3J8E). RyR1 modeling was performed in the software Jmol. Side view of RyR is given. For better visibility, two subunits were removed from the structure. E4032 is labeled by blue dots, the EF-hand domain is colored in red, and the putative Ca2þ binding site is in green (16–18,32). To see this figure in color, go online.

conformational changes are transmitted to the S6 transmembrane helix, which twists to dilate the pore (16–18). This EF-hand pair (EF1 and EF2) was originally identified between residues 4081 and 4127 (19,20). Disrupting the spatial geometry of the EF2 abolished [3H] ryanodine binding of the channel, while EF1 mutant showed significantly increased Ca2þ sensitivity for activation, but lower for inactivation (21). These results suggest that EF1 and EF2 form the activating and the inhibiting Ca2þ sensors respectively. However, this calmodulin-like domain was also proposed to form intramolecular interactions with calmodulin binding sites in the RyR, implying that the calmodulin-like domain would regulate RyR activity by acting as a competitor for calmodulin (22). In contrast, studies using RyR1-RyR2 chimeras have shown that by transferring EF1 and EF2 sequences from RyR2 to RyR1, RyR2’s Ca2þ-inactivation properties could have been fully conferred, so these domains are involved in RyR inactivation (23,24). More recently, it was demonstrated that the EF-hand domain is not required for cytosolic Ca2þ activation but plays an important role in the regulation of cardiac RyR by luminal Ca2þ (25). Unfortunately, the lack of solid functional evidence to support the idea that the EF-hand domain is the activating Ca2þ binding site, and that Ca ions bound to the EF-hands are not visualized in the structure, raises concerns about the role of the EF-hand domain in Ca2þ-gating of RyR. Because negatively charged residues often form Ca2þ binding sites (for example, in the SERCA pump and Ca2þ channels (26–28)), hunting for the Ca2þ sensors in RyR was also pursued by exploring the glutamate-rich regions of conserved sequences in the C-terminal quarter of RyR (19). As a result, mutational analysis revealed that certain homolog glutamates (E4032 in RyR1, E3987 in RyR2, and E3885 in RyR3) are essential in the Ca2þ-activation mechanism of the channels (29–31). According to the 3D structure of RyR1, this E4032 residue is one of a cluster

2128 Biophysical Journal 112, 2127–2137, May 23, 2017

of negatively charged residues that may contribute to the formation of a Ca2þ binding pocket (25), which is located in the core solenoid next to the C-terminal domain and may have direct access to the vestibule of the channel (Fig. 1). These results suggest a different mechanistic model for Ca2þ gating. However, just recently, the model of EF-hand mediated Ca2þ-activation of RyR1 was revised by comparison of difference maps calculated from 3D reconstructions of RyR1 in different states. These calculations show that the putative Ca2þ-binding site is located at C-terminal domain-C-solenoid interface, which is part of the inner lining of the vestibule. The amino acid residues involved in the coordination of Ca2þ are E3893, E3967, and T5001 (32) (Fig. 1). Despite the fact that neither the primary sequence nor the 3D structure predicted functional voltage sensors within RyR, the channel’s open probability (Po) was found to be voltage sensitive (33,34). Channel activity was demonstrated to be enhanced by increased luminal to cytoplasmic Ca2þ gradients and by voltages that favor ion movements from the luminal to the cytoplasmic (trans to cis) sides of channels incorporated into lipid bilayers. Therefore, it was concluded that luminal Ca2þ regulates channel activity by binding to the cytosolic-activating Ca2þ binding site after flowing through the pore (feed-through activation). However, in experiments using the same [Ca2þ] on both sides (50 mM cis and 50 mM trans), channels showed higher Po at voltages that favor ion movements from the cis to the trans side (see Fig. 2) (33). Because this phenomenon cannot be explained by the feed-through activation mechanism, the basis of the voltage-dependent behavior of RyR remains to be elucidated. Our aim was to reveal the underlying mechanism of the voltage-sensitive Ca2þ activation of RyR1 by using the lanthanide (Ln3þ), europium (Eu3þ), and samarium (Sm3þ) ions as tools. Ln3þ has very similar chemical properties but slightly different ionic radii to Ca2þ, and they are commonly used as probes of Ca2þ binding proteins (35–44). We began the study with demonstrating that Eu3þ and Sm3þ are appropriate probes to investigate Ca2þ sensors in RyR, because 1) they specifically bind to Ca2þ binding sites of RyR and 2) RyR is impermeable to Ln3þ. Therefore, these ions allow sideselective investigation of the cytoplasmic binding sites without feed-through effects. Afterwards, we utilized their high charge (3þ) to examine the voltage dependence of Ca2þ activation. MATERIALS AND METHODS Ethical Approval All experiments complied with the Hungarian Animal Welfare Act, the 2010/63/EU guideline of the European Union, and had the approval of the Animal Welfare Committee of the University of Debrecen (22/2012/ ´ B). Male rabbits (n ¼ 5) were killed by using a guillotine. DEMA

Ln3þ Report Ca2þ Sensor in RyR Vestibule PIPES (pH 7.0), and pelleted by centrifugation at 124,000  g for 60 min in a Ti45 rotor. The pellet was resuspended in 300 mM sucrose, 10 mM K-PIPES (pH 7.0). Vesicles were aliquoted, snap-frozen in liquid nitrogen, and stored in a deep freezer until further use in [3H] ryanodine binding assay or were immediately used for RyR purification.

RyR Purification HSR vesicles were solubilized for 2 h in 1% CHAPS, 1 M NaCl, 100 mM EGTA, 150 mM CaCl2, 5 mM AMP, 0.45% phosphatidylcholine, 20 mM Na-PIPES (pH 7.2) at 4 C. The sample was loaded onto a 10–28% linear sucrose gradient and centrifuged overnight at 90,000  g in a SW27 rotor. RyR-containing fractions of the gradient were identified, snap-frozen in liquid nitrogen, and stored at 70 C in small aliquots (45,46).

RyR Reconstitution and Single-channel Recording

FIGURE 2 Voltage-dependent activity of the RyR1 Ca2þ release channel. (A and B) Average open probabilities 5 SEM together with individual data points of single RyR1 channels are plotted as the function of membrane potential. Single channel currents were recorded in a symmetric 250 mM KCl, 50 mM Ca2þ solution. Differences between data of similar symbols are statistically significant (*, #, $, x; p < 0.05). Representative records are shown. The closed state of the channel is marked by c. (B) Schematic diagram of our working hypothesis constructed to guide our further investigation, which were designed to explain the voltage-dependent activity of RyR.

Materials The chemicals, if not specified, were purchased from Sigma-Aldrich (St. Louis, MO).

Heavy Sarcoplasmic Reticulum Vesicle Isolation Heavy sarcoplasmic reticulum (HSR) vesicles were isolated from rabbit skeletal muscle by differential centrifugation as described in the literature (45,46). All the steps were performed on ice or at 4 C in the presence of protease inhibitors. After homogenization in 100 mM NaCl, 20 mM EGTA, 20 mM Na-HEPES (pH 7.5), cell debris was pelleted at 3500  g, for 35 min using a tabletop centrifuge. Crude microsomes were collected from the supernatant by centrifugation in a Ti45 rotor at 40,000  g, for 30 min. To dissolve the actomyosin content the pellet was resuspended in 600 mM KCl, 10 mM K-PIPES, 250 mM sucrose, 1 mM EGTA, 0.9 mM CaCl2 (pH 7.0). After incubation for 1 h at 4 C, the microsomes were centrifuged at 109,000  g for 30 min, then the pellet was resuspended and loaded onto a 20–45% linear sucrose gradient (105 mM NaCl, 10 mM PIPES, 0.1 mM EGTA, 0.09 mM CaCl2 (pH 7.0)). After spinning overnight at 90,000  g in a SW27 rotor, HSR vesicles were collected from the 36–38% regions of the sucrose gradient. The microsomes were washed with a buffer, containing 475 mM sucrose, 1 mM NaCl, 10 mM

Purified RyRs were incorporated into planar lipid bilayers. Bilayers were formed across a 200-mm-diameter aperture of a Delrin cap (Warner Instruments, Hamden, CT), which had two separated chambers. Both chambers contained 250 mM KCl, 50 mM CaCl2, 20 mM PIPES-Tris (pH 7.2). Some experiments were performed in a Ca2þ-free cytosolic recording medium, which did not contain added Ca2þ and was made using Chelextreated distilled water (Bio-Rad, Hercules, CA) to remove traces of Ca2þ. In these experiments, luminal [Ca2þ] was set to 50 mM. Most RyRs incorporated into the bilayer so that their cytosolic foot region faced toward the ground electrode. Transmembrane voltages are referred to this ground (trans to cis). The lipid mixture contained phosphatidylethanolamine, phosphatidylserine, and phosphatidylcholine (Avanti Polar Lipids, Alabaster, AL) in a ratio of 5:4:1 dissolved in n-decane in the final lipid concentration of 20 mg/mL. The bilayer currents were recorded using an AxoPatch 200 amplifier and pCLAMP 6.03 software (Axon Instruments, Sunnyvale, CA). Currents were filtered at 1 kHz through an eight-pole low-pass Bessel filter and digitized at 3 kHz (45,46).

[3H] Ryanodine Binding Assay Ryanodine-binding assay was carried out using [3H] ryanodine in a medium composed of 1 M NaCl, 25 mM Na/PIPES (pH 7.1), 1 mM Pefabloc (SigmaAldrich), and Chelex-treated (Bio-Rad) water. Aliquots (32 mL) containing 16 mg protein were incubated at 37 C for 120 min, with various concentrations of the radioligand, or [Ln3þ] as indicated in the figures. The reaction was terminated by filtering and washing the samples using a BIO-DOT 96-well filter apparatus (Bio-Rad) and 0.45 mm nitrocellulose membrane (Millipore, Billerica, MA). Nonspecific binding was determined in the presence of 1000 excess of ryanodine, which had been added to the incubation mixture before the radioligand. The membrane-bound radioactivity was determined using a liquid scintillation counter (Beckman Coulter, Brea, CA) (46). Bound [3H] ryanodine values were calculated at different Ln3þ concentrations and normalized to data measured in the absence of Ln3þ.

Visualization of RyR Structure Graphic visualization of RyR1 in Fig. 1 was made using PDB:3J8E in the software tool Jmol (an open-source JAVA viewer (Oracle, Redwood City, CA) for chemical structures in three dimensions; www.jmol.org/).

In Silico Studies The cryo-electron microscopy structure (32) deposited into the Protein Data Bank (PDB; www.rcsb.org, PDB:5TAL) was used as a starting geometry for

Biophysical Journal 112, 2127–2137, May 23, 2017 2129

Sa´rko¨zi et al. molecular modeling; however, simplifications had to be made to have tractable sizes for the modeling software we used. The electrostatic potential values at the solvent-excluded surface of the C-terminal domains of the RYR1 proteins (residues 3639–5037) were calculated by means of the APBS software (47). The surface was colored according to the electrostatic potential values and represented by the PyMol modeling package (PyMOL Molecular Graphics System, Ver. 1.8; Schro¨dinger, New York). Ca2þ and Eu3þ metal ion binding sites for the RYR1 domain were predicted by the integral-equation-based reference interaction site model (RISM) theory (48,49) using the rism1d and rism3d software implemented in the AMBER software package (ambermd.org) (50) and the PLACEVENT program (51) downloaded from the Placevent website (http://dansindhikara.com/ Software/Entries/2012/6/22_Placevent_New.html). The missing Eu3þ parameters for RISM calculations were adopted from the article published by Li et al. (52). The RISM and PLACEVENT calculations were carried out only on the cytosolic part of the C-terminal (from residue 3747) protein. Only a monomeric RYR1 protein chain was considered in these calculations.

Statistics Open probabilities (Po) were determined using the pClamp software suite (Molecular Devices, Sunnyvale, CA). Statistical analysis was made in the software Origin 7.0 (OriginLab, Northampton, MA) and in the software Excel (Microsoft, Redmond, WA). Averages were expressed as mean 5 SE. Relative Po and [3H] ryanodine binding data were calculated by normalizing each set of datapoints to their own control. F-test was performed to compare the distribution of the groups. The number of experiments is denoted in the graphs.

RESULTS In our initial experiments we reproduced the results published by Tripathy and Meissner (33). Purified RyR1 channels were incorporated into artificial lipid bilayers and single channel currents were recorded in symmetric 250 mM KCl and 50 mM Ca2þ-containing solution. Representative records in Fig. 2 demonstrate that RyR’s open probability (Po) at 60 mV was significantly higher compared to þ60 mV. (Current directions through the channel are indicated at negative and positive potentials by the cartoons in Fig. 2 B.) These results agree with those published by Tripathy and Meissner (33). In addition, we observed similar preference of channel activity to negative voltages over a wide voltage range (80 to þ80 mV) (Fig. 2 A). These results suggest that the Ca2þ binding site’s occupancy is higher at negative potentials, which raises the possibility that the activating Ca2þ binding site is located in the electrical field (i.e., in the pore or vestibule of the channel), and the local Ca2þ concentration in the vestibule is affected by voltage. To test the first possibility, we used the toxin maurocalcine (MCA) as a tool. MCA is a 33-mer peptide that was previously shown to lock RyR in a long-lasting subconductive state (LLSS) in a highly voltage-dependent manner (with preference to negative membrane potentials, when MCA tends to reside in the vestibule) (53–57). The interaction was found to be electrostatic, because decreasing the electrostatic potential of the toxin by substituting positively charged amino acids with alanine in the 19–24 position

2130 Biophysical Journal 112, 2127–2137, May 23, 2017

attenuated the interaction and reduced the fractional time of LLSS (57). In Fig. S1 we show that LLSS fractional time also depends on the concentration of the charge carrier (i.e., current), which—along with the voltage dependence of LLSS—strongly suggests that MCA binds to the cytoplasmic vestibule of RyR1, where voltage drops through the funnel (56,57). If the Ca2þ sensor is really located in the vestibule, MCA should influence the Ca2þ sensitivity of RyR1. In Fig. 3 A, control single channel currents were recorded at 100 nM cis Ca2þ, which allows low Po. Then, we tested the channel’s retained Ca2þ sensitivity by raising cis [Ca2þ] to 50 mM. Po raised as expected from a functionally intact channel (Fig. 3 B). In this state of the channel, MCA was added to the cis chamber and after the characteristic LLSS developed (this is the sign that MCA occupied the vestibule), Ca2þ was lowered to 100 nM by the addition of EGTA. The LLSSs were occasionally interrupted by normal gating bursts, when MCA left the vestibule. During these bursting events at 100 nM Ca2þ, Po was supposed to be as low as in Fig. 6 A. In contrast, RyR1 displayed high Po, similar to that of measured at 50 mM Ca2þ (Fig. 3 C). This finding suggests that in the presence of MCA, RyR1 became resistant to lowering [Ca2þ]. Furthermore, when the LLSS was cancelled by switching the polarity to positive membrane potential, RyR1 immediately recovered from MCAblock and returned to high Po bursting mode instead of low Po mode (Fig. 3 D). Further on, under the same voltage, Po gradually decreased with time and slowly approached the Po normal for RyR1 at 100 nM Ca2þ.

FIGURE 3 Modified Ca2þ sensitivity of RyR in the presence of MCA. Shown here is the representative current record under control conditions in (A) 100 nM Ca2þ and (B) 50 mM Ca2þ; (C) during MCA (20 nM) treatment at 60 mV; (D) then immediately after being switched to þ60 mV and (E) 10 s later. Closed states are labeled by C and the subconductive states are labeled with S. The cartoons describe the proposed model for the voltage-dependent behavior of MCA and Ca2þ in the vestibule.

Ln3þ Report Ca2þ Sensor in RyR Vestibule

Further testing the location of Ca2þ binding sites required probes that specifically bind to the Ca2þ sensors of RyR1, but do not cross the conducting pore to allow cytoplasmicside-selective examination of the Ca2þ binding sites. Our candidates for this job were the lanthanides, Eu3þ and Sm3þ. Nevertheless, Tb3þ was shown previously to specifically bind Ca2þ binding sites of RyR1 (44) and a number of other Ca2þ binding proteins were shown to bind other Ln3þs (35–43) with high affinity; we started the study with testing whether Eu3þ and Sm3þ were adequate probes of our Ca2þ sensors. For this end, we used a [3H] ryanodine (rya) binding assay, because the voltage across RyR in these experiments is 0 mV and no ionic gradients develop across the membrane. Therefore, we did not have to take into account any voltage-dependent mechanisms and ionic currents through RyRs, making data interpretation easier. The assay utilizes the feature of ryanodine that it binds only to open channels and the binding is irreversible. Consequently, the amount of bound ryanodine relates to the average Po of RyRs in the sample. HSR vesicles were incubated with different concentrations of Eu3þ or Sm3þ in Ca2þ-free medium in the presence of 12 nm [3H] rya. The amount of membrane-bound ryanodine was determined and plotted as the function of [Ln3þ] (Fig. 4, A and B). These graphs show that the action of Ln3þ was biphasic. In relatively low concentrations, Ln3þ increased [3H] rya binding with a Kd of 4.13 5 1.47 mM for Eu3þ and 0.56 5 0.26 mM for Sm3þ. At higher concentrations, Ln3þ decreased [3H] rya binding with a Kd of 15.6 5 0.3 mM for Eu3þ and 11.7 5 0.3 mM for Sm3þ. This bell-shaped dose-response relationship resembles that of Ca2þ’s (11), which suggests that low concentrations of

Ln3þ activate the channel by binding to the activating (high affinity) Ca2þ binding site, while at higher concentrations Ln3þ binds to the inhibiting (low-affinity) Ca2þ binding sites, deactivating the channel (Fig. 4, A and B). The peak activities were at 7.5 mM Eu3þ and 4 mM Sm3þ, suggesting that at these Ln3þ concentrations the activating binding sites were saturated, but the inhibiting Ca2þ binding sites’ occupancy was low(er). Higher than 20-mM Ln3þ concentrations saturated the inhibiting Ca2þ binding sites too, and caused full inhibition. Rya binding at various radioactive rya concentrations was also determined under control conditions (in Ca2þ-free medium), and at maximally activating and inhibiting Ln3þ concentrations. Fig. 4, Ai and Bi, shows pooled data of these experiments. Quantities of 7.5 mM Eu3þ and 4 mM Sm3þ enhanced [3H] rya binding, while 18 and 20 mM Eu3þ and 11 and 45 mM Sm3þ diminished the amount of membrane-bound [3H] rya. Bmax of the control experiment was 3.73 5 0.64 pmol rya/mg protein. The effect of these Ln3þ concentrations was also tested at different Ca2þ concentrations. Under control conditions, [3H] rya binding showed a classic bell-shaped curve with a peak at 50 mM Ca2þ (11) (Fig. 4 C). Between 0 and 2000 mM Ca2þ, 7.5 mM Eu3þ activated the channel, except for 11 mM, when Eu3þ lost effectiveness. As it is reasonable to expect stronger electrostatic interactions from an ion with higher charge, we used a conceptual framework to explain our data. For this concept, we suggest that Eu3þ is a more potent agonist of the Ca2þ binding sites than Ca2þ itself (in other words, Ca2þ is a partial agonist) and that the activating site is saturated at 7.5 mM Eu3þ, while the inhibiting Ca2þ binding site is partially occupied (RyR1’s activating

FIGURE 4 The effect of Ln3þ on the [3H] ryanodine binding properties of HSR vesicles. (A) Relative ryanodine binding of HSR vesicles as a function of [Eu3þ] (open squares) is shown. (Inset) Shown here is relative ryanodine binding in control (solid squares), and then at 7.5 mM (shaded diamonds), at 18 mM (open triangles), and at 20 mM (open spheres) Eu3þ. (B) Relative ryanodine binding of HSR vesicles as a function of [Sm3þ] (open squares) is given. Inset shows the relative ryanodine binding in control (solid squares), and then at 4 mM (shaded diamonds), at 11 mM (open triangles), and at 45 mM (open spheres) Sm3þ. (C and D) Shown here is the relative ryanodine binding as a function of [Ca2þ] in the absence (solid squares) and in the presence of different Eu3þ (C) and Sm3þ (D) concentrations. Eu3þ was applied at 7.5 mM (shaded triangles) and 20 mM (open spheres). Sm3þ was applied at 4 mM (shaded triangles) and 45 mM (open spheres).

Biophysical Journal 112, 2127–2137, May 23, 2017 2131

Sa´rko¨zi et al.

and inhibiting Ca2þ curves overlap). The fact that at <11 mM Ca2þ, Ca2þ binds to the activating but not to the inhibiting binding site, and that in this [Ca2þ] range the difference between Eu3þ’s effect and control gradually decreased, suggests that Ca2þ competes with Eu3þ for the activating site. In the 50–2000 mM Ca2þ range—which saturates the inhibiting binding site too—[3H] rya binding exceeded the highest value observed in the absence of Eu3þ. This infers that at >11 mM, Ca2þ competes with Eu3þ on the inhibiting site and the activating action of Ca2þ on the activating site overcomes the inhibition until the inhibiting Ca2þ binding site gets saturated at 10 mM Ca2þ to inhibit channel activity. At 11 mM Ca2þ, the two impacts were in balance. Here, 20 mM Eu3þ inhibited [3H] rya binding activity in the presence of Ca2þ, but the shape of the curve remained similar to control. This result suggests that at 20 mM Eu3þ, the inhibiting site is highly occupied (see also Fig. 4 A) and that Ca2þ competes for this site with Eu3þ. We obtained qualitatively similar results for 4 mM Sm3þ (Fig. 4 D). However, activation by 4 mM Sm3þ was lower between 2 and 50 mM Ca2þ compared to Eu3þ-treated vesicles. In addition, the inhibiting effect of 45 mM Sm3þ could not be reversed by any Ca2þ concentrations applied. Altogether, these data suggest that Eu3þ and Sm3þ are strong agonists of both cytoplasmic Ca2þ sensors. These data and the conclusion are in agreement with those published earlier about the action of Tb3þ by Hadad et al. (44). Next, to describe the effects of Ln3þ on RyR1 in detail, single channel current measurements were performed in symmetric 250 mM KCl and 50 mM Ca2þ. The representative current record shown in Fig. 5 A demonstrates that 1 mM Eu3þ added to the cytosolic (cis) side of the channel exerted two distinct effects. First, it induced long-lasting closed events at 80 mV. These closed events were not detected at lower Eu3þ concentrations and were highly voltage dependent, as they were absent at þ80 mV (Fig. 5 B). This finding implies that Eu3þ (at R1 mM) acts as a classic pore blocker at negative voltages, when cytosolic cations are driven into the deep pore (see illustration in the middle). As a matter of fact, this pore blocking action of trivalents is very similar to that observed in L-type Ca2þ channels (58–61). Second, at þ80 mV, Eu3þ’s action was qualitatively different: it did not cause long closed events, but significantly reduced the Po by decreasing the number of openings. Sm3þ acted similarly: it caused a voltage-dependent pore-occlusion preferably at negative membrane potentials, but the closed states also appeared at þ80 mV (in contrast to Eu3þ). In addition, Sm3þ also significantly decreased the number of openings in the bursts separated by the long closed events (Fig. 5, C and D). The block could be reversed by the addition of equimolar EGTA (Fig. 5 D). In summary, these results imply that cis Ln3þs occlude the pore and that they reduce Po in a voltage-dependent manner. Consequently, Ln3þ can enter the pore, but cannot

2132 Biophysical Journal 112, 2127–2137, May 23, 2017

FIGURE 5 RyR is impermeable to Ln3þ. (A) Representative current records under control conditions and in the presence of Eu3þare given. Long-lasting closed events of RyR evoked by cis Eu3þ at 80 mV are shown in the bottom trace. Closed state is labeled by c. (B) RyR activity at 1 mM Eu3þ cis, þ80 mV with no long-lasting closures, is shown. (C) Long-lasting closed events of RyR evoked by 90 nM Sm3þ cis at 80 mV are shown. (D) Long-lasting closed events of RyR at 250 and 500 nM Sm3þ at þ80 mV are given. The closed events were reversible by the addition of EGTA. At the end of the experiment, ryanodine was added to the cis chamber to verify RyR. Cartoons of RyR in the middle show the direction of current at the corresponding voltages.

reach the luminal side from the cytoplasmic side. This feature makes Ln3þ an excellent side-selective probe to examine the voltage-dependent cytosolic Ca2þ regulation of RyR. The dose-response relationship and voltage dependence of Eu3þ was further investigated in the virtual absence of Ca2þ. 100–500 nM cis Eu3þ was found to activate RyR1, while higher concentrations inhibited the channel (Fig. 6 A). These results confirm our [3H] rya binding data that Ln3þ is an agonist of both the activating and inhibiting Ca2þ binding sites. Analysis of the voltage dependence of Eu3þ’s action revealed that the activation was more prominent at 60 mV, compared to þ60 mV (Fig. 6 B). Because Eu3þ is driven into the pore by negative voltages, this result also raises the possibility that the activating binding site is located in the cytoplasmic vestibule of the channel. The dose-response curves of Ln3þ’s effect were also determined by using 50 mM Ca2þ in the recording solution. In Fig. 7, A and B single channel current records are shown under control conditions and during Ln3þ treatment at þ60 mV. Dose-response curves demonstrate strong inhibition with dissociation constants of 167 5 5 nM for Eu3þ and 63 5 3 nM for Sm3þ with the Hill coefficient of 2. Under these conditions (50 mM Ca2þ), the activating Ca2þ binding site is saturated and therefore, Ln3þ inhibits RyR1 by binding to the inhibiting Ca2þ binding site. In Fig. 7 C we show the voltage dependence of Eu3þ’s action. Po values in the presence of 0.5–1.5 mM cis Eu3þ were normalized to

Ln3þ Report Ca2þ Sensor in RyR Vestibule

FIGURE 6 Dose-response relationship and voltage sensitivity of Eu3þ action on RyR at nominally 0 mM Ca2þ. (A) Biphasic concentration-dependent effect of cis [Eu3þ] on the open probability (Po) of RyR1 is given. Inset shows representative current records in control and at 250 nM Eu3þ. (B) Voltage dependence of the activating action of 250 nM Eu3þ is shown. Relative effect of Eu3þ in five individual experiments were plotted at 60 and þ60 mV membrane potentials. Cartoons of RyR on the sides show the direction of currents at the negative and positive voltages, respectively.

the control Po values (of the same voltage and same channel) and were plotted as the function of membrane potentials. High concentrations of Eu3þ were shown in Fig. 5 A to occlude the pore at negative voltages. Because these poreblockage-related long closed events would have distorted the analysis of Eu3þ’s effect on the Ca2þ sensors, these events were not included in the Po analysis. The analysis revealed that Eu3þ was significantly less potent to reduce channel Po at negative voltages. In contrast, inhibiting potency was higher at positive potentials. This behavior is unexpected for a small cationic pore blocker, which is driven into the pore by negative membrane potentials. The voltage-dependent action of Sm3þ was also tested at 560 mV with similar results. Based on the results shown in Figs. 6 and 7, Eu3þ’s activating action is Ca2þ dependent. In nominally Ca2þ free solution, Eu3þ activated, whereas in 50 mM Ca2þ, it inhibited RyR in the <500 mM range in a voltage-dependent manner. Based on these results, we hypothesize that the inactivation

FIGURE 7 Dose-response relationship and voltage sensitivity of Ln3þ action on RyR in 50 mM Ca2þ is shown. Relative RyR1 open probabilities as a function of cis [Eu3þ] (A) and [Sm3þ] (B) are given. Hill fit revealed dissociation constants of 167 5 5 nM for Eu3þ and 63 5 3 nM for Sm3þ. Hill coefficients were ‘‘2’’ in both cases. (C). The voltage dependence of Ln3þ action is given. Po values in the presence of 0.5–1.5 mM cis Eu3þ were normalized to their own control Po values (solid squares). Sm3þ data at 560 mV are given (open spheres). (Insets) Schematic illustrations of the RyR channel show the current direction through the pore.

site is always (at least partially) occupied by Ln3þ and that the binding of Eu3þ to the activation site overcomes this inhibited state when Eu3þ is driven toward the activation site at negative voltages but that inhibition is more enhanced when Eu3þ is driven away from the site at positive voltages. In silico studies were performed to examine the Ca2þ binding site found in 2016 by des Georges et al. (32). In Fig. 8 we show that this binding site is accessible from

Biophysical Journal 112, 2127–2137, May 23, 2017 2133

Sa´rko¨zi et al.

FIGURE 8 In silico studies of the activating Ca2þ binding sites. Shown here is the solvent-excluded surface of the C-terminal end of RyR1. The surface color represents the electrostatic potential at the surface points calculated by the software APBS (http://www.poissonboltzmann.org/). The negative to positive electrostatic potential values are shown by red to blue colors, respectively. The tetramer structure was cut in half by a plane, which includes the main axis of the pore. The Ca2þ binding site determined by cryoelectron microscopy is circled with green. To see this figure in color, go online.

the pore. It is also immediately apparent in this figure that the inner surface of the ion-channel vestibule can be featured by highly negative electrostatic potential values, similar to the cavity where the Ca2þ binding site is located. From RISM and subsequent PLACEVENT calculations the proposed primary ion binding site for both Ca2þ and Eu3þ ions was the same, and it corresponded to the Ca2þ binding site, which was shown in the cryo-electron microscopy structure (PDB:5TAL) (32). At <20 mM Eu3þ, the model predicted one Eu3þ binding site, whereas at 50 mM ion concentrations it predicted seven Eu3þ and one Ca2þ bound to the truncated RyR subunit. These calculations indicate that in our concentration range, Eu3þ binds only to the Ca2þ binding site; however, our results should be interpreted with caution, because there are examples in the literature showing that Ln3þ also binds to nonspecific metal binding sites (62,63). DISCUSSION The voltage-dependent rectification of RyR1 raised the possibility that the voltage-dependent activity is due to the voltage-dependent Ca2þ gating of the channel. In this study, the side-selective Ca2þ analogs Eu3þ and Sm3þ were used as probes to discover the underlying mechanism for this behavior and to learn more about the Ca2þ sensors of RyR1. We show that Ln3þs function as agonists of both Ca2þ sensors. Nevertheless, our [3H] rya binding data should be interpreted with caution, because nonspecific

2134 Biophysical Journal 112, 2127–2137, May 23, 2017

binding of Ln3þ to calcium binding proteins was reported in the literature (62,63). These putative nonspecific binding sites may be responsible for the different dissociation constants in single-channel versus [3H] rya-binding experiments. However, the lack of transmembrane potential in [3H] rya experiments can account for the differences, too. Next, based on the finding that RyR is blocked (occluded) by Ln3þ, we concluded that RyR is impermeable to these ions. By utilizing these characteristics, we examined the voltage-dependent action of Ln3þ on RyR in single channel experiments. We show that at Ca2þ concentrations that saturate the activating Ca2þ binding site (50 mM), Ln3þ inhibited the channel but this inhibition was stronger at positive potentials. In contrast, in Ca2þ-free recording solution, Eu3þ activated the channel, which was stronger at negative membrane potential. To explain this anomalous voltagedependent behavior of Ln3þ’s action and the voltage-dependent Ca2þ-gating in symmetric 50 mM Ca2þ (Fig. 2), we constructed a model (see cartoon in Fig. 6 B) that locates the activating but not the inhibiting Ca2þ binding site in the electrical field of the channel (into the cytoplasmic vestibule). This model relies on the result that suggests Ln3þ acts as an agonist of both the activating and inhibiting Ca2þ-binding sites. According to this model, weak inhibiting action developed at negative voltages because negative potential drives Ln3þ into the vestibule, toward the activating Ca2þ binding site. Therefore, the activating action of Ln3þ overcomes the inhibition, while at positive voltages activation is weaker than inhibition. The model is further supported by the result that activation by Eu3þ in Ca2þfree solution is stronger at negative membrane potentials, when Eu3þ is driven into the cytoplasmic vestibule. The structure of the vestibule also accounts for the voltage dependence of Ca2þ gating. The channel’s pore has an asymmetric structure with a funnel-like geometry. The wide, conical opening of the funnel creates the cytoplasmic vestibule, which tapers toward the inner pore and is coupled to a narrow stem facing toward the SR lumen (16–18). The big cis-to-trans conductance of RyR is permitted by a wide cytoplasmic mouth (or vestibule) of the channel, which functions as a large ion-capture zone. Electrostatic models of channels with a narrow pore (ion conduction pathway) between two mouths suggest that the voltage drop extends beyond the constricted pore out into the regions with large diameters. The extension of the applied voltage is determined by pore geometry. It depends on the ratio of the length and the diameter of the conduction pathway (64–70). For example, in the case of a channel with a relatively short, but wide pore (like RyR’s), the electrical field will expand significantly out into the vestibule. Therefore, the movement of ions in the water-filled vestibule is not governed by free diffusion, but is also under the influence of trans-membrane voltage (electrodiffusion) (70). To enhance ion entry and conductance, the vestibule is lined by negatively charged amino acids (71). According

Ln3þ Report Ca2þ Sensor in RyR Vestibule

to our model, the structural basis for the voltage-dependent activity of RyR is the negatively charged funnel, which focuses Ca2þ near the pore and to the Ca2þ sensor (Fig. 8). Therefore, negative voltages, which favor ionic flow from cis-to-trans side of the channel, accumulate Ca2þ around the putative Ca2þ sensor, while positive potentials (which guide ions out from the cytoplasmic vestibule) deplete Ca2þ from the funnel and the Ca2þ binding pocket. Based on the recently solved 3D structure of RyR1, it has been proposed that the EF-hand Ca2þ binding domain mediates the Ca2þ gating in RyR through a conformational switch. EF-hands are located in the C-terminal end (amino acids 4071–4131) of the core solenoid and their Ca2þdependent conformational change is probably transmitted through the C-terminal domain toward the S6 transmembrane helix (the gate), which twists to open the pore (16–18). However, more recently, Guo et al. (25) found that mutating the amino acid residues that are critical for Ca2þ binding in the EF1 or EF2 motif or deleting the entire EF-hand Ca2þ binding domain of RyR2 did not affect the Ca2þ sensitivity of the RyR2 channel to cytosolic activation. Previously, E4032 was reported to be critical for Ca2þ sensing in RyR1 (29), but in lacking a structural model it remained obscure whether E4032 was part of the Ca2þ binding site or only an essential transducer on the gating pathway. New structural data show that E4032 only forms an interface with the C-terminal domain and by doing so it stabilizes the conformation that allows Ca2þ binding, but it does not contribute to the formation of the Ca2þ binding site directly. Ca2þ was proposed to be bound by E3893 and E3967 of the core solenoid, and T5001 from the C-terminal domain (32). In this RyR1 model (see Fig. 1) the EF-hand does not look like to be accessible from the pore, whereas E3893, E3967, and T5001 line the cytosolic vestibule and are part of a negatively charged pocket (Figs. 1 and 8). This structure is highly conserved in RyR and the other intracellular Ca2þ-dependent Ca2þ release channel, the inositol trisphosphate receptor (IP3R) (while analogous EF-hand motifs are missing from the IP3R) (72–74). In addition, the new binding pocket is located in the core solenoid C-terminal domain interface, which is directly connected to the S6 helix (Fig. 1), therefore it could relay similar conformational changes as EF-hand was suggested to. Consequently, this indirect evidence also supports our vestibular Ca2þ sensor model and argues against the hypothesis that the EF-hand domain is the activating Ca2þ binding site. Our previous results proposed that the orientation of the toxin MCA is likely to be the most important factor in its binding to the site, the one that is responsible for the initiation of the subconductive state—and that the electric field determines the orientation of the toxin. This suggests that the MCA binding site is near the opening of the channel pore (in the mouth), where the electric gradient is extremely large, and thus the electric momentum dominates the orien-

tation of the MCA molecule (57). To explain the result that RyR lost Ca2þ sensitivity with MCA in the vestibule (Fig. 3), quantitative analysis of the voltage dependence of MCA-block was performed according to Woodhull’s single-site blocking scheme (Eq. 1) (68–70,75), as follows:  G ¼

1þ

 ½MCa  expðzd  FV=RTÞ ; Kd 0

(1)

where G is the difference of relative conductances of MCAinduced subconductive states measured at two different negative voltages (slope of voltage dependence of conductance); V is the membrane potential difference; and R, T, and F have their usual meanings. The Kd of MCA at 0 mV (Kd 0) was determined using a [3H] ryanodine binding assay, and was 5 mM. The calculation yielded the effective valence of block (zd), which is the product of the valence of the blocking ion (z MCA ¼ þ4) and the distance into the electrical field at which the blocker binds. As a result, we got 0.11 for d, which means that MCA’s binding site is located at 11% of the whole voltage drop measured from the cytosolic end of the channel. It means that this large, 33-amino-acid toxin’s charged surface falls to the inner end of the vestibule, while the rest of the peptide covers the outer regions of the mouth, where the Ca2þ sensor is possibly located. Based on the result of this analysis, we concluded that MCA worked as the lid of the vestibule and closed Ca2þ in a restricted vestibular space near the Ca2þ sensor. Therefore, with the lid on (at substate), RyR was resistant to the modification of [Ca2þ] in the bulk solution. Releasing MCA from the channel (cancelling LLSS) opened the lid and allowed Ca2þ to diffuse out the vestibule, and Po slowly adapted to the new local [Ca2þ] (see cartoons in Fig. 3). We need to consider the alternative possibility that voltage-sensitive Ca2þ gating of RyR would develop because Ca2þ sensor’s conformation was directly or indirectly affected by voltage. If this was the case, the process should not depend on the charge of the ion. However, Ln3þ’s voltage-dependent action was stronger compared to the voltage-dependent action of Ca2þ (Fig. 2), which most probably arises from the fact that 3þ charged ions are more susceptible to the electrostatic effects of the electrical field. Our other argument against the voltage-dependent Ca2þ sensor option is that the transition between high Po and low Po mode in the MCA experiment at 100 nM Ca2þ was very slow. This indicates that the adaptation of Po to low [Ca2þ] was due to a relatively slow diffusion of Ca2þ out from the vestibular binding pocket after MCA was released from the pore. Therefore, the voltage-dependent Ca2þ gating cannot be attributed to the conformational change of the Ca2þ sensor domains. Because RyR is a large allosteric protein, a direct approach should be used to identify the Ca2þ sensors. Because electron scattering increases with the square of

Biophysical Journal 112, 2127–2137, May 23, 2017 2135

Sa´rko¨zi et al.

atomic number, the high atomic number of Ln3þ (63 for Eu3þ and 62 for Sm3þ versus 20 for Ca2þ) may offer a chance for imaging Ln3þ binding to the Ca2þ binding site by using high-resolution cryo-electron microscopy. SUPPORTING MATERIAL One figure is available at http://www.biophysj.org/biophysj/supplemental/ S0006-3495(17)30342-9.

AUTHOR CONTRIBUTIONS S.S. performed research, and analyzed and interpreted data. I.K. performed in silico experiments, interpreted data and wrote the manuscript. I.J. conceived the idea, designed research, and interpreted data. J.A. conceived the idea, designed research, performed experiments, analyzed and interpreted data, and wrote the manuscript.

ACKNOWLEDGMENTS This work was supported by grants provided to J.A. and I.J. from the Hungarian Scientific Research Fund (Nos. OTKA PD 112199 and OTKA 81923). J.A. is supported by the Janos Bolyai Research Scholarship of the Hungarian Academy of Sciences and the Lajos Szodoray Scholarship of the University of Debrecen.

REFERENCES 1. Bers, D. M. 2002. Cardiac excitation-contraction coupling. Nature. 415:198–205. 2. Stern, M. D., G. Pizarro, and E. Rı´os. 1997. Local control model of excitation-contraction coupling in skeletal muscle. J. Gen. Physiol. 110:415–440. 3. Straub, S. V., D. R. Giovannucci, and D. I. Yule. 2000. Calcium wave propagation in pancreatic acinar cells: functional interaction of inositol 1,4,5-trisphosphate receptors, ryanodine receptors, and mitochondria. J. Gen. Physiol. 116:547–560. 4. Smith, J. S., R. Coronado, and G. Meissner. 1986. Single channel measurements of the calcium release channel from skeletal muscle sarcoplasmic reticulum. Activation by Ca2þ and ATP and modulation by Mg2þ. J. Gen. Physiol. 88:573–588.

12. Tung, C. C., P. A. Lobo, ., F. Van Petegem. 2010. The amino-terminal disease hotspot of ryanodine receptors forms a cytoplasmic vestibule. Nature. 468:585–588. 13. Szigeti, G. P., J. Alma´ssy, ., I. Jo´na. 2007. Alterations in the calcium homeostasis of skeletal muscle from postmyocardial infarcted rats. Pflugers Arch. 455:541–553. 14. Marx, S. O., S. Reiken, ., A. R. Marks. 2000. PKA phosphorylation dissociates FKBP12.6 from the calcium release channel (ryanodine receptor): defective regulation in failing hearts. Cell. 101:365–376. 15. Jo´na, I., and P. P. Na´na´si. 2006. Cardiomyopathies and sudden cardiac death caused by RyR2 mutations: are the channels the beginning and the end? Cardiovasc. Res. 71:416–418. 16. Efremov, R. G., A. Leitner, ., S. Raunser. 2015. Architecture and conformational switch mechanism of the ryanodine receptor. Nature. 517:39–43. 17. Yan, Z., X. C. Bai, ., N. Yan. 2015. Structure of the rabbit ryanodine receptor RyR1 at near-atomic resolution. Nature. 517:50–55. 18. Zalk, R., O. B. Clarke, ., A. R. Marks. 2015. Structure of a mammalian ryanodine receptor. Nature. 517:44–49. 19. Takeshima, H., S. Nishimura, ., S. Numa. 1989. Primary structure and expression from complementary DNA of skeletal muscle ryanodine receptor. Nature. 339:439–445. 20. Xiong, H., X. Feng, ., G. Meissner. 1998. Identification of a two EF-hand Ca2þ binding domain in lobster skeletal muscle ryanodine receptor/Ca2þ release channel. Biochemistry. 37:4804–4814. 21. Fessenden, J. D., W. Feng, ., P. D. Allen. 2004. Mutational analysis of putative calcium binding motifs within the skeletal ryanodine receptor isoform, RyR1. J. Biol. Chem. 279:53028–53035. 22. Xiong, L., J. Z. Zhang, ., S. L. Hamilton. 2006. A Ca2þ-binding domain in RyR1 that interacts with the calmodulin binding site and modulates channel activity. Biophys. J. 90:173–182. 23. Gomez, A. C., and N. Yamaguchi. 2014. Two regions of the ryanodine receptor calcium channel are involved in Ca2þ-dependent inactivation. Biochemistry. 53:1373–1379. 24. Du, G. G., V. K. Khanna, and D. H. MacLennan. 2000. Mutation of divergent region 1 alters caffeine and Ca2þ sensitivity of the skeletal muscle Ca2þ release channel (ryanodine receptor). J. Biol. Chem. 275:11778–11783. 25. Guo, W., B. Sun, ., S. R. Chen. 2016. The EF-hand Ca2þ binding domain is not required for cytosolic Ca2þ activation of the cardiac ryanodine receptor. J. Biol. Chem. 291:2150–2160. 26. Clarke, D. M., T. W. Loo, ., D. H. MacLennan. 1989. Location of high affinity Ca2þ-binding sites within the predicted transmembrane domain of the sarcoplasmic reticulum Ca2þ-ATPase. Nature. 339:476–478.

5. Meissner, G. 1994. Ryanodine receptor/Ca2þ release channels and their regulation by endogenous effectors. Annu. Rev. Physiol. 56:485–508.

27. Heinemann, S. H., H. Terlau, ., S. Numa. 1992. Calcium channel characteristics conferred on the sodium channel by single mutations. Nature. 356:441–443.

6. Sa´rko¨zi, S., C. Szegedi, ., I. Jo´na. 2000. Regulation of the rat sarcoplasmic reticulum calcium release channel by calcium. J. Muscle Res. Cell Motil. 21:131–138.

28. Yang, J., P. T. Ellinor, ., R. W. Tsien. 1993. Molecular determinants of Ca2þ selectivity and ion permeation in L-type Ca2þ channels. Nature. 366:158–161.

7. Mickelson, J. R., and C. F. Louis. 1996. Malignant hyperthermia: excitation-contraction coupling, Ca2þ release channel, and cell Ca2þ regulation defects. Physiol. Rev. 76:537–592.

29. Du, G. G., and D. H. MacLennan. 1998. Functional consequences of mutations of conserved, polar amino acids in transmembrane sequences of the Ca2þ release channel (ryanodine receptor) of rabbit skeletal muscle sarcoplasmic reticulum. J. Biol. Chem. 273:31867– 31872.

8. Fill, M., R. Coronado, ., C. F. Louis. 1990. Abnormal ryanodine receptor channels in malignant hyperthermia. Biophys. J. 57:471–475. 9. Shomer, N. H., C. F. Louis, ., J. R. Mickelson. 1993. Reconstitution of abnormalities in the malignant hyperthermia-susceptible pig ryanodine receptor. Am. J. Physiol. 264:C125–C135. 10. Loke, J., and D. H. MacLennan. 1998. Malignant hyperthermia and central core disease: disorders of Ca2þ release channels. Am. J. Med. 104:470–486. 11. Mickelson, J. R., L. A. Litterer, ., C. F. Louis. 1990. Stimulation and inhibition of [3H] ryanodine binding to sarcoplasmic reticulum from malignant hyperthermia susceptible pigs. Arch. Biochem. Biophys. 278:251–257.

2136 Biophysical Journal 112, 2127–2137, May 23, 2017

30. Li, P., and S. R. Chen. 2001. Molecular basis of Ca2þ activation of the mouse cardiac Ca2þ release channel (ryanodine receptor). J. Gen. Physiol. 118:33–44. 31. Chen, S. R., K. Ebisawa, ., L. Zhang. 1998. Molecular identification of the ryanodine receptor Ca2þ sensor. J. Biol. Chem. 273:14675–14678. 32. des Georges, A., O. B. Clarke, ., J. Frank. 2016. Structural basis for gating and activation of RyR1. Cell. 167:145–157.e17. 33. Tripathy, A., and G. Meissner. 1996. Sarcoplasmic reticulum lumenal Ca2þ has access to cytosolic activation and inactivation sites of skeletal muscle Ca2þ release channel. Biophys. J. 70:2600–2615.

Ln3þ Report Ca2þ Sensor in RyR Vestibule 34. Herrmann-Frank, A., and F. Lehmann-Horn. 1996. Regulation of the purified Ca2þ release channel/ryanodine receptor complex of skeletal muscle sarcoplasmic reticulum by luminal calcium. Pflugers Arch. 432:155–157.

55. Altafaj, X., W. Cheng, ., M. Ronjat. 2005. Maurocalcine and domain A of the II-III loop of the dihydropyridine receptor Cav 1.1 subunit share common binding sites on the skeletal ryanodine receptor. J. Biol. Chem. 280:4013–4016.

35. Brown, P. H., R. A. H. Rathjen, ., D. E. Tribe. 1990. Rare earth elements in biological systems. In Handbook on the Physics and Chemistry of Rare Earths, Vol. 13. K. A. Gschneidner, and L. Eyring, editors.. Elsevier Science, Amsterdam, the Netherlands, pp. 423–452.

56. Altafaj, X., J. France, ., M. Ronjat. 2007. Maurocalcine interacts with the cardiac ryanodine receptor without inducing channel modification. Biochem. J. 406:309–315. 57. Luka´cs, B., M. Sztretye, ., L. Csernoch. 2008. Charged surface area of maurocalcine determines its interaction with the skeletal ryanodine receptor. Biophys. J. 95:3497–3509.

36. Jona, I., and A. Martonosi. 1986. The effects of membrane potential and lanthanides on the conformation of the Ca2þ-transport ATPase in sarcoplasmic reticulum. Biochem. J. 234:363–371. 37. Dux, L., K. A. Taylor, ., A. Martonosi. 1985. Crystallization of the Ca2þ-ATPase of sarcoplasmic reticulum by calcium and lanthanide ions. J. Biol. Chem. 260:11730–11743. 38. Remedios, C. G. 1981. Lanthanide ion probes of calcium-binding sites on cellular membranes. Cell Calcium. 2:29–51. 39. Evans, H. C. 1990. Biochemistry of the Lanthanides. Springer Science & Business Media, New York. 40. Switzer, M. E. 1978. The lanthanide ions as probes of calcium ion binding sites in biological systems. Sci. Prog. 65:19–30. 41. Pidcock, E., and G. R. Moore. 2001. Structural characteristics of protein binding sites for calcium and lanthanide ions. J. Biol. Inorg. Chem. 6:479–489. 42. Permyakov, E., and H. R. Kretsinger. 2011. Calcium Binding Proteins. John Wiley, Hoboken, NJ. 43. Brittain, H. G., F. S. Richardson, and R. B. Martin. 1976. Terbium(III) emission as a probe of calcium(II) binding sites in proteins. J. Am. Chem. Soc. 98:8255–8260. 44. Hadad, N., A. C. Zable, ., V. Shoshan-Barmatz. 1994. Ca2þ binding sites of the ryanodine receptor/Ca2þ release channel of sarcoplasmic reticulum. Low affinity binding site(s) as probed by terbium fluorescence. J. Biol. Chem. 269:24864–24869. 45. Geyer, N., G. Diszha´zi, ., J. Alma´ssy. 2015. Bile acids activate ryanodine receptors in pancreatic acinar cells via a direct allosteric mechanism. Cell Calcium. 58:160–170. 46. Sa´rko¨zi, S., C. Szegedi, ., I. Jo´na. 2005. Effect of gadolinium on the ryanodine receptor/sarcoplasmic reticulum calcium release channel of skeletal muscle. FEBS J. 272:464–471.

58. Lansman, J. B. 1990. Blockade of current through single calcium channels by trivalent lanthanide cations. Effect of ionic radius on the rates of ion entry and exit. J. Gen. Physiol. 95:679–696. 59. Obejero-Paz, C. A., I. P. Gray, and S. W. Jones. 2004. Y3þ block demonstrates an intracellular activation gate for the a1G T-type Ca2þ channel. J. Gen. Physiol. 124:631–640. 60. Babich, O., J. Reeves, and R. Shirokov. 2007. Block of CaV1.2 channels by Gd3þ reveals preopening transitions in the selectivity filter. J. Gen. Physiol. 129:461–475. 61. Malasics, A., D. Boda, ., D. Gillespie. 2010. Simulations of calcium channel block by trivalent cations: Gd3þ competes with permeant ions for the selectivity filter. Biochim. Biophys. Acta. 1798:2013–2021. 62. Henao, F., S. Orlowski, ., P. Champeil. 1992. The metal sites on sarcoplasmic reticulum membranes that bind lanthanide ions with the highest affinity are not the ATPase Ca2þ transport sites. J. Biol. Chem. 267:10302–10312. 63. Szebenyi, D. M., and K. Moffat. 1986. The refined structure of vitamin D-dependent calcium-binding protein from bovine intestine. Molecular details, ion binding, and implications for the structure of other calciumbinding proteins. J. Biol. Chem. 261:8761–8777. 64. Latorre, R., and C. Miller. 1983. Conduction and selectivity in potassium channels. J. Membr. Biol. 71:11–30. 65. Jordan, P. C. 1984. Effect of pore structure on energy barriers and applied voltage profiles. I. Symmetrical channels. Biophys. J. 45:1091–1100. 66. Jordan, P. C. 1984. Effect of pore structure on energy barriers and applied voltage profiles. II. Unsymmetrical channels. Biophys. J. 45:1101–1107.

47. Baker, N. A., D. Sept, ., J. A. McCammon. 2001. Electrostatics of nanosystems: application to microtubules and the ribosome. Proc. Natl. Acad. Sci. USA. 98:10037–10041. 48. Kiyota, Y., and M. Takeda-Shitaka. 2014. Molecular recognition study on the binding of calcium to calbindin D9k based on 3D reference interaction site model theory. J. Phys. Chem. B. 118:11496–11503.

67. Yellen, G. 1987. Permeation in potassium channels: implications for channel structure. Annu. Rev. Biophys. Biophys. Chem. 16:227–246.

49. Ratkova, E. L., D. S. Palmer, and M. V. Fedorov. 2015. Solvation thermodynamics of organic molecules by the molecular integral equation theory: approaching chemical accuracy. Chem. Rev. 115:6312–6356.

69. Tinker, A., and A. J. Williams. 1993. Charged local anesthetics block ionic conduction in the sheep cardiac sarcoplasmic reticulum calcium release channel. Biophys. J. 65:852–864.

50. Case, D. A., R. M. B. W. Botello-Smith, ., P. A. Kollman. 2016. AMBER 2016. University of California, San Francisco, CA.

70. Williams, A. J., D. J. West, and R. Sitsapesan. 2001. Light at the end of the Ca2þ-release channel tunnel: structures and mechanisms involved in ion translocation in ryanodine receptor channels. Q. Rev. Biophys. 34:61–104. 71. Xu, L., Y. Wang, ., G. Meissner. 2006. Two rings of negative charges in the cytosolic vestibule of type-1 ryanodine receptor modulate ion fluxes. Biophys. J. 90:443–453.

51. Sindhikara, D. J., N. Yoshida, and F. Hirata. 2012. PLACEVENT: an algorithm for prediction of explicit solvent atom distribution-application to HIV-1 protease and F-ATP synthase. J. Comput. Chem. 33:1536–1543. 52. Li, P., L. F. Song, and K. M. Merz, Jr. 2015. Parameterization of highly charged metal ions using the 12-6-4 LJ-type nonbonded model in explicit water. J. Phys. Chem. B. 119:883–895. 53. Chen, L., E. Este`ve, ., I. N. Pessah. 2003. Maurocalcine and peptide A stabilize distinct subconductance states of ryanodine receptor type 1, revealing a proportional gating mechanism. J. Biol. Chem. 278:16095–16106. 54. Este`ve, E., S. Smida-Rezgui, ., M. Ronjat. 2003. Critical amino acid residues determine the binding affinity and the Ca2þ release efficacy of maurocalcine in skeletal muscle cells. J. Biol. Chem. 278:37822– 37831.

68. Tinker, A., A. R. G. Lindsay, and A. J. Williams. 1992. Large tetraalkyl ammonium cations produce a reduced conductance state in the sheep cardiac sarcoplasmic reticulum Ca2þ-release channel. Biophys. J. 61:1122–1132.

72. Fan, G., M. L. Baker, ., I. I. Serysheva. 2015. Gating machinery of InsP3R channels revealed by electron cryomicroscopy. Nature. 527:336–341. 73. Miyakawa, T., A. Mizushima, ., M. Iino. 2001. Ca2þ-sensor region of IP3 receptor controls intracellular Ca2þ signaling. EMBO J. 20:1674–1680. 74. Tu, H., E. Nosyreva, ., I. Bezprozvanny. 2003. Functional and biochemical analysis of the type 1 inositol (1,4,5)-trisphosphate receptor calcium sensor. Biophys. J. 85:290–299. 75. Woodhull, A. M. 1973. Ionic blockage of sodium channels in nerve. J. Gen. Physiol. 61:687–708.

Biophysical Journal 112, 2127–2137, May 23, 2017 2137