|Home | About | Journals | Submit | Contact Us | Français|
The second of three SPRY domains (SPRY2, S1085-V1208) located in the skeletal muscle ryanodine receptor (RyR1) is contained within regions of RyR1 that influence EC coupling and bind to imperatoxin A, a toxin probe of RyR1 channel gating. We examined the binding of the F loop (P1107-A1121) in SPRY2 to the ASI/basic region in RyR1 (T3471-G3500, containing both alternatively spliced (ASI) residues and neighboring basic amino acids). We then investigated the possible influence of this interaction on excitation contraction (EC) coupling. A peptide with the F loop sequence and an antibody to the SPRY2 domain each enhanced RyR1 activity at low concentrations and inhibited at higher concentrations. A peptide containing the ASI/basic sequence bound to SPRY2 and binding decreased ~10-fold following mutation or structural disruption of the basic residues. Binding was abolished by mutation of three critical acidic F loop residues. Together these results suggest that the ASI/basic and SPRY2 domains interact in an F loop regulatory module. Although a region that includes the SPRY2 domain influences EC coupling, as does the ASI/basic region, Ca2+ release during ligand- and depolarization-induced RyR1 activation were not altered by mutation of the three critical F loop residues following expression of mutant RyR1 in RyR1-null myotubes. Therefore the electrostatic regulatory interaction between the SPRY2 F loop residues (that bind to imperatoxin A) and the ASI/basic residues of RyR1 does not influence bi-directional DHPR-RyR1 signaling during skeletal EC coupling, possibly because the interaction is interrupted by the influence of factors present in intact muscle cells.
The Ca2+ release channel in the sarcoplasmic reticulum (SR) of skeletal muscle is the type 1 ryanodine receptor (RyR1). The in vivo function of RyR1 in skeletal muscle excitation-contraction (EC) coupling requires the physical presence of the skeletal muscle isoform of the dihydropyridine receptor (DHPR), and more specifically, the II-III loop of its α1S subunit.1 RyR1 is composed of numerous structural domains, including three “SPRY” domains in the N-terminal 1/3 of each of its four subunits. These domains were named SPRY because they were identified in both Dictyostelium discoideum tyrosine kinase spore lysis A (SplA) and the mammalian RyR.2 The domains generally function as protein-protein interaction motifs and contain two antiparallel β-sheets with β-strands separated by unstructured or α-helical linkers.3 The three SPRY domains in the RyR are conserved across species from insects to mammals, but their function is unknown. The second of the three domains in RyR1 (SPRY2, S1085-V1208), may function in EC coupling because it overlaps with larger sequences that support skeletal muscle EC coupling4 and is an in vitro binding partner for both N-terminal residues in the II-III loop of the skeletal muscle DHPR, as well as for the scorpion toxins imperatoxin A (IpTxA) and maurocalcine.5,6 IpTxA, maurocalcine and the N-terminal part of the II-III loop are used as in vitro probes of RyR1 function.7–12 All or part of the SPRY2 domain is common to each of the sequences implicated in II-III loop and toxin binding and, since we previously found that the II-III loop and IpTxA interact with the isolated SPRY2 domain,3,13,14 it is likely that their interaction with SPRY2 underlies the effects of the peptides on RyR channel gating.
Both the conservation of the SPRY2 sequence across species and the likelihood that both IpTxA and the N-terminus of the II-III loop produce significant changes in RyR1 gating by binding to the SPRY2 domain suggest that the SPRY2 domain regulates in vitro RyR1 channel gating. Since the binding of both compounds to the SPRY2 domain increases channel activity, it is possible that the SPRY2 domain contributes to an inhibitory regulatory module within RyR1 such that channel activity increases when this module is disrupted by either toxin or II-III loop peptide binding. Curiously, although IpTxA strongly influences the gating of isolated RyR1 channels, it does not alter RyR1-mediated Ca2+ release during skeletal EC coupling.7 This suggests that the RyR1 channel becomes insensitive to IpTxA in muscle cells when it is under the control of the DHPR.7 If the inhibitory regulatory module including the SPRY2 domain is interrupted by binding of IpTxA, then its interruption may similarly have a minimal influence on RyR1-mediated Ca2+ release during skeletal EC coupling. However, the impact of the SPRY2 domain regulatory module on RyR1-mediated Ca2+ release during EC coupling has not been determined and is examined here.
The impact of a regulatory module containing the SPRY2 domain on Ca2+ release during EC coupling was also of interest because SPRY2 is included in a larger N-terminal region of RyR1 that influences conformational DHPR-RyR1 coupling.4 One loop of the SPRY2 domain (P1107-A1121, the F loop) binds to the positively charged α-helical region of the N-terminal part of the DHPR II-III loop (between residues E666-L690).3,13,14 These basic II-III loop residues do not overlap with the “critical” amino acids (720–765) in the central region of the II-III loop15 that are required for the conformational coupling between the DHPR and RyR1 that defines skeletal muscle EC coupling. This suggests that SPRY2 domain binding to the II-III loop does not contribute to the essential role of the II-III loop in EC coupling. However, it is possible that the SPRY2 domain influences EC coupling through interaction with another site, but this possibility has not been specifically examined. Here we directly assessed the specific role of the F loop of the SPRY2 domain on RyR1 function and investigated the possible role of three specific residues in the F loop in regulating in vitro RyR1 channel activity and bi-directional DHPR-RyR1 signaling16 during EC coupling in myotubes.
We also addressed the likely binding partner for the SPRY2 domain in the inhibitory regulatory module within RyR1. RyR1 contains a region that exhibits close structural and functional similarity to the N-terminal residues of the DHPR II-III loop17–19 and structural similarity to IpTxA and maurocalcine. This 30 amino acid region in RyR1 contains alternatively spliced (ASI) residues plus a neighboring polybasic sequence and is designated here as the ASI/basic region (T3471-G3500). Although its binding partners are unknown, several lines of evidence indicate that the small ASI/basic region is a functionally very important region in RyR1. Firstly, the variably spliced residues D3483-G3487 are present in the adult RyR1 isoform (ASI(+)RyR1), but excluded from the juvenile isoform (ASI(−)RyR1). Secondly, ASI(−)RyR1 is upregulated in adults with myotonic dystrophy.20 Thirdly, the ASI residues contribute to a regulatory module in the RyR1 complex.21 Finally, mutations of the ASI residues or adjacent basic residues produce significant effects on skeletal muscle EC coupling.22,23 Specifically, omission of the 5 variably spliced residues leads to a significant increase in the gain of EC coupling22 and the cluster of adjacent basic residues (L3495-R3499) contributes to a binding site for the DHPR β1a subunit and its deletion leads to a substantial reduction in voltage-dependent Ca2+ release during skeletal EC coupling.23
The cluster of positively charged residues in the ASI/basic region and in the N-terminal region of the DHPR II-III loop are aligned along one surface of an α-helix. As a result, there is competition between the positively charged residues in the II-III loop and positively charged residues in the ASI/basic region for in vitro binding to a negatively charged region of RyR1.22 Therefore, since the negatively charged F loop of the SPRY2 domain binds to the DHPR II-III loop in vitro,14 we predicted that by homology it may also bind to the ASI/basic region. Moreover, there is also evidence that the SPRY2 domain and the ASI/basic region are located within 4–6 nm of each other in the three dimensional RyR1 structure, and thus, could interact (Professor M. Samso, personal communication). The SPRY2 domain has been localized to the clamp region of RyR1 using 3 different anti-SPRY2 antibodies (Perálvarez-Marín A, Tae HS, Board PG, Casarotto MG, Dulhunty AF and Samsó M—in preparation for publication), including the antibody used in the present work (Results, Figs. 2 and and33). The ASI/basic region has not been localized on the RyR 3D structure. However the residues are within the IPTxA binding sequence (3,210–3,661) and the IPTxA binding site is located 4 nm from the SPRY2 domain.5,24 Furthermore, in the linear RyR1 sequence, ASI/basic residues are separated by only 100–150 residues from the calmodulin binding site which is located 6 nm from the SPRY2 domain.25 Thus, depending on the degree of folding, it is possible that the SPRY2 domain and the ASI/basic regions are close enough to interact to form the inhibitory module that is disrupted by IpTxA, the N-terminal residues of the II-III loop or a peptide corresponding to the RyR1 ASI/basic region.22 Here, we set out to validate this proposed electrostatic binding interaction between the negatively charged residues in the SPRY2 F loop and the positively charged ASI residues and to investigate the possible functional impact of this interaction on RyR1 channel activity in planar lipid bilayers and during physiological ligand and depolarization-induced activation following expression in RyR1-null myotubes.
The role of the SPRY2 domain and its F loop in influencing RyR1 channel activity was explored using a peptide with the F loop sequence (F peptide, Methods) and an antibody to the SPRY2 domain.
One F peptide concentration was tested in each experiment to evaluate any time-dependent effects on channel activity. The peptide was applied to the cytoplasmic solution and channel activity followed for ~20 min before the peptide was removed by perfusion. The peptide at 1 to 10 µM (Fig. 1A) activated channels, while higher concentrations (100 µM) inhibited. No time-dependent effects were seen; channel activity changed within 30 s of peptide application and was maintained until the peptide was removed. Bilayers were maintained through to washout in at least one or two experiments at each peptide concentration. In each case, activity returned to control levels within 1–2 min as shown in Figure 1A and B.
The average values in Figures 1, ,33 and and44 combine data at +40 and −40 mV for simplicity of presentation as there was no consistent difference between parameters obtained at the two potentials (Table 1A). Data for each channel was normalized to its internal control to avoid anomalies introduced by the usual variability in baseline gating between RyR1 channels,26,27 which is reflected in the range of control values in Table 1B. The results in Figure 1A–C are consistent with a higher affinity activation by 1–10 µM F peptide and a lower affinity inhibition with ≥50 µM peptide. The fit to the data in Figure 1C with standard Hill equations (Methods) show an AC50 of 0.55 µM and Hill coefficient of 4.8 for activation, and an IC50 of 17.4 µM and Hill coefficient of 2.4 for inhibition. The coefficient of ~4 for activation suggests that the F peptide binds to four high affinity sites on the RyR1 complex, perhaps to one binding site per subunit. The coefficient of 2.4 for inhibition suggests that low affinity binding to only two subunits may be required for inhibition.
The effects of the F peptide on channel gating are shown as the average difference between log10 of each control and test parameter (Fig. 1D–F). The increase in channel open probability with 1 and 10 µM F peptide and the decline in activity at 50 and 100 µM F peptide are significant and are accompanied by small changes in mean open time and larger changes in mean closed times. The open times with 1 µM peptide were significantly longer and closed durations were significantly shorter than control with 0.5, 1 and 10 µM peptide. Open times with the F peptide at 50 and 100 µM were significantly shorter than with 0.5, 1 and 10 µM, while closed times were significantly longer. When compared to control, the open times with 100 µM peptide were significantly shorter, and mean closed time significantly longer. Control experiments with unrelated peptides were not conducted here as we previously demonstrated the absence of effects on RyR1 activity of several unrelated peptides.22,28,29 Moreover, the lack of stimulation of the SPRY2 RyR1 mutant by the F loop peptide in Figure 4 provides an additional important negative control for F loop peptide specificity.
One interpretation of the increase in RyR1 activity with low F loop peptide concentrations is that the peptide disrupts a closed state stabilization module in RyR1 involving an interaction between the endogenous SPRY2 F loop and another part of the RyR1 protein complex. As suggested for the actions of other domain peptides,30 the F loop peptide could bind and displace the endogenous F loop binding partner. Support for this proposed mechanism can be inferred from experiments in which an antibody to the domain peptide exerts a similar action, presumably by binding to the endogenous domain to interrupt the same inhibitory module.31 Thus we examined the actions of a polyclonal antibody raised against the SPRY2 domain. Since the homology model of SPRY2 shows that the F loop is on the outer surface of the domain,3,14 it was likely that the F loop region provided an epitope for antibody recognition in the polyclonal mixture.
Western blots show that the anti-SPRY2 antibody recognized the recombinant SPRY2 domain (Fig. 2A) and native RyR1 channels in SR vesicles (Fig. 2B). Moreover, the antibody immunoprecipitated purified RyR1, but with a lower efficacy than the widely used anti-RyR 34C antibody (Fig. 2C and D, for details see Methods).
Addition of the anti-SPRY2 antibody to the cytoplasmic solution caused changes in open probability similar to those with the F loop peptide (Fig. 3A). Specifically, 1 µg/ml anti-SPRY2 antibody substantially increased RyR1 activity, while 5 µg/ml nearly abolished activity. Both the increase and the decrease in activity were seen within 30 s of antibody addition and maintained for 10–30 min. The antibody concentration was increased further to 10 µg/ml in three experiments, and to 20 µg/ml then 40 µg/ml in another, but no further decrease in activity was detected below that recorded with 5 µg/ml of the antibody. The purified anti-γ-GCT antibody had no effect on channel activity in six experiments (Fig. 3B).
On average, 1 µg/ml of the anti-SPRY2 antibody caused a significant increase in open probability and 5 µg/ml produced a significant decrease to values less than that of control (Fig. 3C). In marked contrast, relative channel open probability (Po) was unaltered by the control anti-γ-GCT antibody (1 or 5 µg/ml, or even 10 µg/ml in one experiment). The average basal Po was 0.14 ± 0.04 for the anti-SPRY2 antibody experiments and 0.19 ± 0.08 for anti-γ-GCT experiments. The changes in open probability with the anti-SPRY2 antibody were due to opposing effects on channel open (Fig. 3D) and closed (Fig. 3E) times. With 1 µg/ml anti-SPRY2 antibody, the relative open time increased significantly (from 3.2 ± 0.6 ms), while the relative closed time fell significantly (from 82.0 ± 31.5 ms). When anti-SPRY2 antibody was increased to 5 µg/ml, the relative open time fell significantly below control while the relative closed time increased significantly. The basal open and closed times before antibody addition were 3.2 ± 0.6 ms and 82.0 ± 31.5 ms, respectively. Neither 1 µg/ml nor 5 µg/ml anti-γ-GCT antibody altered either open (control, 4.35 ± 1.35 ms) or closed times (control, 73.5 ± 33.2 ms). Thus, the anti-SPRY2 antibody regulates RyR1 channels in a qualitatively similar biphasic manner as that of the SPRY2 F loop peptide.
Overall, the results in Figures 1–3 suggest that the F loop of the SPRY2 domain is a functionally significant modulator of RyR1 channel gating. We next examined the F loop contribution to a protein-protein interaction(s) within the RyR1 complex, and whether this interaction significantly impacts channel activity and EC coupling. The activating effect of the F loop and SPRY2 antibody on RyR1 is similar to previously reported effects of the ASI/basic peptide.21 The ASI/basic region contains a stretch of positively charged residues that could interact with acidic residues in the SPRY2 F loop to form a regulatory module within RyR1. The involvement of the SPRY2 F loop in an inter-domain interaction with the ASI region was first explored in biochemical experiments to examine the ability of the two regions to bind to each other.
The binding of the recombinant SPRY2 domain to two peptides corresponding to the ASI/basic region was examined. These peptides (see Methods) were; (1) ASI(+) containing the variably spliced residues, (2) ASI(−) lacking the residues and (3) two ASI(−) mutants. The ASI(−) peptide bound to the SPRY2 domain with a Kd of 1.5 ± 0.1 µM (Table 2), which was >4-fold lower than Kd for the ASI(+) peptide (6.6 ± 0.2 µM). The higher affinity for the ASI(−) peptide is reminiscent of its 4- to 6-fold higher affinity for RyR1 in [3H]ryanodine binding experiments and its greater ability to increase SR Ca2+ release and enhance RyR1 channel activity, with affinities also in the 1–10 µM range.21
The influence of the ASI peptides on RyR1 activity depends on the peptides having an α-helical structure with positively charged residues aligned on one side of the helix.22 Therefore, we determined the impact of this positive charge density and α-helical structure on ASI peptide binding to the SPRY2 domain using two mutant ASI(−) peptides. In the 1st mutant (ASI(−)mutant), three positively charged residues were substituted with alanine or leucine (K3495A, K3496A and R3498L). This peptide retains an α-helical structure, but has limited ability to activate RyR1.22 Its affinity for binding to the SPRY2 domain was reduced >8-fold (Table 2A). The second mutant, ASIshort (R3493-G3500), retains the positively charged residues, but is too short to sustain an α-helical structure. ASIshort was relatively ineffective in activating RyR1,22 and its affinity for the SPRY2 domain was reduced >10-fold (Table 2A). Thus, the charge and structure of the ASI/basic region dictates its interaction with the SPRY2 domain and its regulation of RyR1 activity in a very similar way, consistent with the hypothesis that the ASI region may regulate RyR1 through binding to the SPRY2 domain.
One acidic residue E1108 in a negatively charged region of the SPRY2 F loop influences its in vitro binding to the basic N-terminal region of the DHPR II-III loop.22 Thus, we tested if the same acidic region also interacts with basic residues in the α-helix of the structurally analogous ASI/basic region. The E1108A F loop substitution reduced the affinity for ASI(−) ~7-fold (mutant-1, Table 2B). The D1112A substitution (Mutant-2) caused a 9-fold decline in affinity. Mutating both residues (mutant-6) produced a 19-fold fall in affinity, while an additional E1114A (mutant-7) substitution abolished binding. Curiously, however, neither the single E1114A substitution (mutant-3), D1118A (mutant-4) nor E1119A (mutant-5) alone significantly altered binding. Thus, E1108, D1112 and E1114 are essential for in vitro binding of the SPRY2 domain to the ASI/basic residues. Mutation of the three acidic residues abolished SRRY2 domain binding to the ASI/basic region, in contrast to only a partial reduction in SPRY2 binding to the N-terminal residues of the II-III loop with E1108A.14 This observation indicates that the binding site for the N-terminal residues of the II-III loop and for the ASI/basic region overlap but is not identical. However, the overlap may be sufficient for binding of the N-terminal II-III loop residues to disrupt the interaction between the F-loop and the ASI/basic region.
The results thus far demonstrate that the isolated ASI/basic region binds to the SPRY2 domain in RyR1. We reported that the ASI/basic residues of RyR1 strongly impact RyR1 activity and EC coupling.22 Specifically, the ASI(−) peptide activates RyR to a greater extent than the ASI(+) peptide and full-length juvenile ASI(−) RyR1 exhibits enhanced depolarization-induced Ca2+ release compared to the adult ASI(+)RyR1.22 This gain in function in the full-length ASI(−)RyR1 may reflect a tighter interaction between the SPRY2 domain and ASI(−)/basic region. Thus, we tested the hypothesis that the strength of EC coupling depends on SPRY2/basic interaction by assessing the effects on RyR1 channel activity and EC coupling of mutations in the SPRY2 domain that disrupt binding of the domain to the ASI(−) peptide. For these experiments, the SPRY2 domain residues E1108, D1112 and E1114 were substituted with alanines in the full-length juvenile ASI(−) RyR1 since these three mutations abolished in vitro binding of the SPRY2 domain to the ASI region (mutant-7, Table 2B). We predicted that the SPRY2 domain mutations would disrupt the SPRY2-ASI/basic inhibitory interaction, increase isolated RyR1 channel activity and enhance Ca2+ release during EC coupling.
We first determined the effect of the three mutations on regulation of ASI(−)RyR1 single channel activity by the F loop peptide (Fig. 4). Wild-type ASI(−)(ASI(−)RyR1 WT) and the ASI(−)RyR1 triple mutant (E1108A, D1112A and E1114A, ASI(−) RyR1 EDE) were expressed in HEK293 cells and the RyR1 protein was solubilized and incorporated into lipid bilayers (Fig. 4A and B). As predicted, the average basal channel open probability of the ASI(−)RyR1 EDE was significantly greater than that of ASI(−)RyR1 WT (Fig. 4C). The F loop peptide was added to cytoplasmic solution bathing the channels firstly at a concentration of 1 µM, and then the concentration was increased after 5–8 min to 100 µM. Similar to ASI(+) RyR1 (Fig. 1), ASI(−) RyR1 WT channels responded to 1 µM F loop peptide with an increase in activity (Po increased from 0.13 ± 0.03 to 0.34 ± 0.07) and then activity decreased below control levels after increasing the peptide concentration to 100 µM (to 0.083 ± 0.02, Fig. 4A and D). In marked contrast, the open probability of ASI(−) RyR1 EDE channels did not change significantly with 1 µM F loop peptide (Po: control 0.25 ± 0.02; F-loop, 0.25 ± 0.03) and decreased substantially after addition of 100 µM of the peptide (Po reduced to 0.16 ± 0.02; Fig. 4B and D). Thus, the SPRY2 mutations selectively abolished channel activation by low concentrations of F loop peptide.
We next examined the possible impact of the SPRY2 mutations in ASI(−)RyR1 on RyR1 function and bi-directional DHPR-RyR1 signaling during EC coupling following expression in RyR1-null myotubes and HEK-293 cells (Figs. 5 and and66). Figure 5A shows representative electrically evoked and ligand-induced (30 mM caffeine and 500 µM 4-CMC) responses from indo-1 am-loaded RyR1-null myotubes expressing either wild-type ASI(−) (ASI(−)RyR1 WT) or the ASI(−)RyR1 triple mutant (ASI(−)RyR1 EDE). Average peak responses to each stimulus are summarized in Figure 5B and were not significantly different between RyR1-null myotubes expressing either ASI(−) RyR1 WT or ASI(−)RyR1 EDE. As shown previously in reference 32, peak electrically-evoked Ca2+ transient amplitude is smaller than that produced by exposure to caffeine and 4-CMC due to the much longer RyR1 activation during ligand activation (30 s) compared to that produced by a single action potential stimulus (~10 ms). In addition, following expression in HEK293 cells, the magnitude and sensitivity of 4-CMC-induced Ca2+ release was also similar for ASI(−)RyR1 WT (Fig. 5C filled circle; ΔRmax = 1.3, EC50 = 97 µM and Hill coefficient = 1.4) and ASI(−)RyR1 EDE (Fig. 5C open circle; ΔRmax = 1.4, EC50 = 113 µM and Hill coefficient = 1.3).
The effects of the SPRY2 domain mutations that disrupt binding to the ASI(−) region on bi-directional DHPRRyR1 coupling are summarized in Figure 6. The whole-cell voltage clamp technique was used to simultaneously record depolarization-induced L-type Ca2+ currents and intracellular Ca2+ transients in RyR-null myotubes expressing either ASI(−) RyR1 WT, the E1108A, D1112A ASI(−) double mutant (ASI(−) RyR1 ED) or ASI(−)RyR1 EDE. The voltage dependence and peak magnitude of both L-type Ca2+ currents (Fig. 6A and C and Table 3) and intracellular Ca2+ transients (Fig. 6B and D and Table 3) were similar for all three constructs. A small (~6 mV) depolarizing shift in the voltage dependence L-type Ca2+ channel activation was the only parameter significantly altered by both the ED and EDE SPRY2 domain mutations (Fig. 6C and Table 3). The SPRY2 domain ED double mutant introduced into full-length adult ASI(+)RyR1 also did not significantly alter the magnitude or voltage dependence of either L-type Ca2+ currents or intracellular Ca2+ transients (data not shown). Together, the results in Figures 5 and and66 demonstrate that the bi-directional DHPR-RyR1 signaling is not dramatically altered by F loop mutations that destabilize an interaction between the SPRY2 and ASI domains in vitro. It is possible that the influence of the SPRY2-ASI/basic interaction on RyR1 activity observed in vitro is masked by interactions of the channel with other proteins present within cells.
The data presented in this study provides several novel observations and raises a number of new questions. Our results are the first to demonstrate that the SPRY2 domain binds to the ASI/basic region in RyR1 (Table 2) and regulates RyR1 channel activity (Fig. 1), but that this interaction is not an essential determinant of in situ RyR1 function or bi-directional DHPR-RyR1 coupling in myotubes (Figs. 5 and and6;6; Table 3). The SPRY2-ASI/basic regulatory interaction is indicated by similar stimulatory effects on RyR1 activity by low concentrations of both a peptide corresponding to the SPRY2 domain F loop and an antibody to the SPRY2 domain. Moreover, we demonstrated direct binding between the SPRY2 F loop and ASI/basic peptides and that this interaction is strongly influenced by three specific acidic residues in the SPRY2 F loop. Disruption of the putative ASI/SPRY2 interaction by mutating these acidic residues in the full-length protein increased the RyR1 channel open probability, abolished channel activation by low concentrations of F loop peptide, but did not significantly alter either ligand- or depolarization-induced RyR1 Ca2+ release following expression in RyR1-null myotubes and HEK293 cells. These results indicate that while SPRY2 contributes to an in vitro RyR1 regulatory domain, the influence of this domain is substantially reduced in intact cells, under the conditions examined here. Thus, although the SPRY2 domain (aa 1,085–1,208) is contained within regions of RyR1 shown to influence skeletal muscle EC coupling (aa 1–1,681),4 our results suggest that these effects do not involve a regulatory module including the three acidic F loop residues required for intramolecular binding to the ASI/basic region.
Both the F peptide (Fig. 1) and the anti SPRY2 antibody (Fig. 3) activate RyR1 channels at low concentrations, the F peptide with a Hill coefficient of ~4. This suggests that the SPRY2 domain may contribute to an intramolecular inhibitory module in each subunit of the RyR1 tetramer and that competing F loop peptide and anti-SPRY2 antibody activates the channel by disrupting the four inhibitory modules within the RyR1 tetramer.
The dual effects of the peptide and antibody on RyR1 channel activity (activation at low and inhibition at high concentrations) could potentially be due to the peptide and the antibody binding to two sites on RyR1 with the lower affinity interaction being non-specific. Alternatively, the biphasic effect could arise from the binding of peptide or antibody to a common site in which activation and inhibition reflect binding to a different number of RyR1 subunits. However, our finding that channel activation by low concentrations of F loop peptide was selectively abolished by mutations in the SPRY2 domain that disrupt its interaction with the ASI region, but not the inhibition observed at high concentrations of F peptide suggests that the biphasic effect most likely reflects binding to two distinct sites.
Parallel effects of antibodies and peptides on release channel function have previously been interpreted in terms of interruption of important intramolecular regulatory domains that involve interactions between different regions of RyR1.30,31 Within this framework, our results are consistent with the SPRY2 F loop being involved in a functionally important inhibitory module of the channel. Although we found that the F loop peptide binds with relatively high affinity to the ASI/basic region, the myotube RyR1 reconstitution studies suggest that the ASI/basic region may not serve as the physiologically relevant binding site in intact muscle cells.
We defined essential structural parameters required for the ASI/basic region to interact with the SPRY2 domain. The interaction site on the ASI peptide involves a series of basic residues aligned along one side of an α-helix. Both neutralization of the basic residues and disruption of the α-helix reduce the affinity of the ASI/basic region for the SPRY2 domain, although the remaining lower affinity binding is consistent with other residues in the ASI/basic region contributing to binding (Table 2). Conversely, the substitution of three acidic residues, E1108, D1112 and E1114 on the SPRY2 F loop with non-charged alanine residues abolished binding, indicating a tightly circumscribed binding site on the F loop (Table 2). In addition, the SPRY2 domain binds to the ASI/basic region with higher affinity than to the N-terminal II-III loop residues, indicating a tighter and more constrained binding interaction between ASI/basic residues and the SPRY2 domain.
A surprising finding was that mutations that abolish SPRY2 domain binding to the ASI/basic region did not significantly impact either ligand-induced RyR1 Ca2+ release or bi-directional DHPR-RyR1 signaling (Figs. 5 and and66 and Table 3). Thus, although the ASI/basic region strongly influences depolarization-induced Ca2+ release during skeletal muscle EC coupling,22,23 this regulation appears not to depend on an interaction with the SPRY2 domain. Alternatively, the SPRY2-ASI/basic interaction may influence in situ RyR1 function and EC coupling and the absence of a detectable functional effect in myotube experiments could reflect failure of the three mutations to disrupt the interaction in the full-length protein. However, the fact that the mutations selectively abolished the stimulatory effect of the F loop peptide on recombinant RyR1 single channel activity suggests that the mutations indeed disrupts the interaction even in the full-length protein. Another possibility is that the effect of ASI/basic region on EC coupling reported previously,22 reflects modifications of its interaction with the DHPR β1a subunit23 in a manner that enhances opening of the RyR1 pore to release SR Ca2+ during membrane depolarization.
A final possibility is that the influence of the ASI/basic regulatory module on native RyR1 activity is negated by conditions that exist in the cellular environment. For example, the DHPR-RyR1 physical interaction that defines skeletal muscle EC coupling33 may override or interrupt the SPRY2-ASI regulatory interaction, as it interrupts the IpTxA interaction with RyR1.7 However the observation that the SPRY2/ASI/basic interaction does not alter 4-CMC induced Ca2+ release in HEK293 cells suggests that other factors are involved in interrupting the functional effects of the interaction in cells. These factors might include inhibition by interactions between adjacent RyR1 tetramers assembled in orthogonal arrays, oxidation state or by the regulatory interaction being overwhelmed by the strong positive feedback of Ca2+-induced Ca2+ release, where free Ca2+ is not held constant as it is in the bilayer studies.
Thus, the in vivo function of the SPRY2 domain and its interaction with the ASI-basic region remains elusive. The fact that it is interrupted by cellular factors present under normal conditions in myotubes and HEK293 cells suggests that any in vivo functional importance may emerge only when the normal cellular conditions are altered as in disease (e.g., myotonic dystrophy) or in aging. The overall effect of interrupting such an inhibitory interaction in muscle during disease or aging would be to allow enhanced activation of the RyR1 during excitation-contraction coupling. These possibilities remain to be explored.
The following peptides were synthesized by the JCSMR Biomolecular Resource Facility using an Applied Biosystems 430A Peptide Synthesiser, with purification to 98–100% using HPLC and mass spectroscopy. Stock peptide (5 mM) in H2O was frozen in 20 µl aliquots.
|SPRY2 F loop:|
The cDNA encoding the entire SPRY2 domain was amplified by PCR and cloned in-frame down stream of a poly-histidine-tagged ubiquitin sequence in the plasmid pHUE.13,14 The wild-type and mutant F loop sequences in the SPRY2 domain were:
|Wild-type||P1107ELR PDV ELG ADE LA1121|
|Mutant-1||P1107ALR PDV ELG ADE LA1121|
|Mutant-2||P1107ELR PAV ELG ADE LA1121|
|Mutant-3||P1107ELR PDV ALG ADE LA1121|
|Mutant-4||P1107ELR PDV ELG AAE LA1121|
|Mutant-5||P1107ELR PDV ELG ADA LA1121|
|Mutant-6||P1107ALR PAV ELG ADE LA1121|
|Mutant-7||P1107ALR PAV ALG ADE LA1121|
|Mutant-8||P1107ALR PAV ALG AAA LA1121|
The antibody was produced in rabbits using standard procedures with Freund's adjuvant and three weekly immunizations with purified recombinant SPRY2 domain. The antibody was purified by specific binding to purified SPRY2 domain immobilised on nitrocellulose membranes.34 The anti-γ-glutamycyclotransferase (γ-GCT) antibody control was similarly purified against recombinant γ-GCT. The use of an unrelated antibody as a control allows the inclusion of non-crossreacting immunoglobulins in the experiments. This is not possible if pre immune serum is used as a control. Purified antibodies were stored at −20°C.
Protein A sepharose beads were equilibrated overnight with the anti-SPRY2 antibody (2 µg/ml) or with the commercially available 34C antibody (0.05 µg/ml) (Abcam Cambridge, MA USA) or with no antibody. All procedures were performed at 4°C. The beads were washed three times with 500 µl of 20 mM NaPIPES pH 7.4 and incubated with 40 ng of purified RyR1 (80 µl of 25 µg/ml, in 20 mM NaPIPES pH 7.4 plus a protease inhibitor cocktail (Roche) including Calpain I, Calpain II (1.25 µl per 5 ml buffer); Benz, Pepstatin A, Leupeptin, AEBSF, A protinin (1 µl per 1 ml buffer)) for 4 h. The beads were then spun down for 10 s and washed three times with 500 µl of 20 mM NaPIPES pH 7.4 + protease inhibitors. For electrophoresis, the beads were equilibrated with 20 µl of sample buffer for 10 min at 60°C. The supernatant following centrifugation was run on a 4–15% 15-well BioRad precast gel at 200 V for ~40 min under denaturing conditions. The samples were blotted onto nitrocellulose transfer membrane at 125V for ~1½ hours in 37 mM Tris, 140 mM glycine, 20% methanol, blocked for 1 h at room temperature in 5% skim milk in PBS, washed and incubated with primary antibody (34C anti-RyR1 antibody) 1:5,000 in TPBS at 4°C, then washed and incubated with secondary anti-mouse antibody conjugated to HRP 1:5,000 in TPBS for 2 h at room temperature.
The efficacy of the anti-SPRY2 antibody in immunoprecipitating RyR1 was compared with the efficacy of a widely used and commercially available anti-RyR antibody (34C; Abcam Cambridge, MA USA) (Fig. 2C). In the experiment shown in Figure 2D, the anti-SPRY antibody was incubated with Sepharaose A beads at a concentration of 2 µg/ml immunoprecipitated the purified RyR1 with 36% of the efficiency (Fig. 2D) of the 34C antibody incubated with beads at a concentration of 0.05 µg/ml. Therefore, the 34C antibody was ~118 times more efficient at pulling down purified RyR than the anti-SPRY antibody. It is worth noting that while a similar increased efficacy in RyR1 labeling for 34C was also observed in western Blot experiments, the two antibodies exhibited a similar efficacy in dot blot experiments (Professor Montserrat Samso-personal communication). These results indicate that SPRY antibody binding is strongly dependent on the functionally competent native conformation of the protein, which is maintained in lipid bilayer experiments.
Equilibrium dissociation constants for SPRY2 binding to the ASI peptides were determined with a Perkin-Elmer LS 50B6 Spectrofluorimeter using changes in the intrinsic fluorescence in the 203 residue recombinant SPRY2 domain (S1085-V1208, containing two tryptophan residues) in 50 mM sodium phosphate (pH 8) and 300 mM NaCl buffer at 25°C, when aliquots of peptide were added. Excitation was at 280 nm and fluorescence emission monitored at 340 nm. Data, was corrected for dilution effects and for intrinsic peptide fluorescence, and fitted with nonlinear regression using GraphPad Prism software. For analysis, fluorescence (Y) at quencher concentrations (X) were fitted with equation 1 where Fo is the initial fluorescence of the SPRY2 domain without quencher, Fs is the final SPRY2 domain fluorescence with quencher, Eo is the concentration of SPRY2 and Kd is the dissociation constant.35
HEK293 cells were grown in MEM (Minimal Essential Medium) with 10% FCS, at 37°C, 5% CO2 in T175 culture flasks. The day before transfection, cells were plated out at 50% confluence and the medium was changed 3 hours before transfection. Cells were transfected with a mammalian expression vector (pCIneo) encoding either ASI(−)RyR1 WT or ASI(−)RyR1 EDE (80 µg DNA per flask) using the calcium phosphate precipitation method. Cells were incubated at 37°C for 24 h before the medium was changed and for another 24 hours before they were harvested. For homogenization and solubilization of the recombinant RyR1 constructs, the cell pellet was resuspended in 3 ml of ice-cold Hypotonic Buffer (1 mM EDTA 10 mM Hepes, pH 7.4) containing protease inhibitor cocktail (Roche). The resuspended cells were homogenized once with a French Press (Thermo Scientific) at 1,000 psi. The homogenate was then centrifuged at 3,000 g for 10 min at 4°C. Solubilization buffer (3.75 ml of 33 mM Pipes pH 7.2, 266 µM EDTA, 1.33 M NaCl, 287 µM DTT, 3.33%/1.67% CHAPS/PC) was added to the supernatant and carefully mixed in a tissue grinder with 15 strokes and incubated on ice for 30 min. The mixture was then centrifuged at 163,200 g for 45 min. Then the supernatant was aliquoted and stored at −80°C until further use.
HEK293 cells were grown in DMEM/10% FBS at 37°C, 5% CO2 and transfected with full-length ASI(−) WT RyR1 or EDE mutant cDNA using lipofectamine at 70% cell confluence and a DNA:lipofectamine ratio of 1:2.5 (12 µg DNA/10 cm culture dish). This transfection protocol resulted in an ~70% transfection efficiency. Ca2+ measurements were carried out 36–48 h after transfection. Following loading with 5 µM Fura-2 in Ringer solution for 45 min at 37°C, cells were alternately excited by 340 and 380 nm (510 nm emission) every 1 s (20 ms exposure per wavelength and 1 × 1 binning) using a monochrometer-based illumination system and captured using a high speed, digital QE CCD camera (TILL Photonics, Pleasanton, CA USA). Thirty seconds of different concentrations of 4-chloro-m-cresol (4-CMC) were applied using a local profusion system, with a 50 s of washout with control Ringer solution between applications. Only 2–3 concentrations of 4-CMC were applied to any given cell in order to minimize 4-CMC toxicity. Fura-2 ratios (R = F340/F380) were generated offline and used to determine peak 4-CMC responses (ΔR = Rpeak − Rbase). Concentration-response curves were generated by combining peak responses to 1, 5, 10, 30, 50, 100, 300, 1,000 µM 4-CMC and then the results fitted using a standard Hill equation.
Primary cultures and nuclear cDNA injections of myotubes from RyR1-null mice were performed as described previously.20 Briefly, cDNAs encoding either wild-type ASI(−)RyR1 (ASI(−)RyR1 WT), E1108A/D1112A double mutant (ASI(−) RyR1 ED), or E1108A/D1112A/E1114A triple mutant (ASI(−) RyR1 EDE) were microinjected into nuclei of individual myotubes at 0.5 µg/µl. Co-injection of CD8 cDNA (0.1 µg/µl) was used to identify expressing myotubes by incubation with CD8 antibody-coated beads.
Myotubes expressing wild-type or ASI mutant RyR1s were loaded for 1 h at 37°C with 6 µM Indo-1 AM (TEFLABS, Austin, TX USA), in a Ringer's solution containing: 146 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM HEPES, pH 7.4 Intracellular Ca2+ transients were measured as described previously.32 Indo-1 loaded myotubes were excited at 350 nm and fluorescence emission measured at 405 nm and 485 nm. A series of stimuli were sequentially applied to myotubes including electrical stimulation (80 V, 5 ms) at a frequency of 0.1 Hz, followed by 30 s of 30 mM caffeine and 500 µM 4-CMC, with a 30 s washout in between. All stimulations were delivered locally either by glass pipette electrode or by local perfusion. Results were presented as the ratio of fluorescence at 405 nm and 485 nm (F405/F485).
Voltage-gated L-type Ca2+ currents (L-currents) and intracellular Ca2+ transients were simultaneously recorded from WT and mutant ASI(−)RyR1-expressing myotubes using the whole-cell patch clamp technique.36 For these experiments, the external solution contained: 145 mM TEA-Cl, 10 mM CaCl2, 10 mM HEPES, pH 7.4 and the pipette internal solution contained: 145 mM Cs-aspartate, 0.1 mM EGTA, 1.2 mM MgCl2, 0.2 mM K5-Fluo-4, 5 mM Mg-ATP, 10 mM HEPES, pH 7.4. Cell capacitance was determined by integration of the capacity transient resulting from a +10 mV pulse applied from the holding potential (−80 mV) and was used to normalize Ca2+ currents (pA/pF) from different myotubes. Peak L-current density (pA/pF) was plotted as a function of membrane potential and fitted according to:
where Gmax is the maximum L-channel conductance, Vm is the test potential, Vrev is the reversal potential, VG1/2 is the voltage for half-maximal activation of Gmax and kG is a slope factor. Relative changes in intracellular Ca2+ in patch clamp experiments were simultaneously recorded during each test pulse and are reported as ΔF/F, where F is the baseline fluorescence immediately prior to depolarization and ΔF is the fluorescence change from baseline. Fluorescence amplitudes at the end of each test pulse were plotted as a function of membrane potential and fitted according to:
where (ΔF/F)max is the maximal fluorescence change, Vm is the test potential, VF1/2 is the voltage for half activation of (ΔF/F)max and kF is a slope factor.
Rabbit skeletal SR vesicles were isolated and incorporated in lipid bilayers37–39 using a cis (cytoplasmic) solution containing: 230 mM CsCH3O3S, 20 mM CsCl, 1.0 mM CaCl2, 10 mM TES and 500 mM mannitol (pH 7.4, with CsOH) and a trans (luminal) solution containing 30 mM CsCH3O3S, 20 mM CsCl, 1 mM CaCl2 and 10 mM TES (pH 7.4).38,39 Following RyR incorporation, the cis solution was replaced with a similar solution, lacking mannitol and having 10 µM free Ca2+ (buffered with BAPTA, adjusted using a Ca2+ electrode). 200 mM CsCH3O3S was added to the trans solution to achieve symmetrical 250 mM Cs+. Bilayer potential, Vcis − Vtrans, was switched between −40 and +40 mV every 30 s. Channel activity under each condition was analyzed over 60 to 120 s using the program Channel 2 (developed by PW Gage and M Smith, JCSMR). Threshold levels for channel opening were set to exclude baseline noise at ~20% of the maximum single channel conductance. A curve describing the peptide concentration - dependence of relative Po (RPo) was constructed using the sum of Hill equations for activation and inhibition:40
RPomax and RPomin are the relative Po of the minimally and maximally activated channel, AC50 and IC50 are the F [peptide] for half-maximal activation and inhibition respectively, and Ha and Hi are the corresponding Hill coefficients. Curves were fitted to the data using the least squares method.
Average data are given as mean ±SEM. Statistical significance was evaluated using paired or unpaired Student's t-test as appropriate or ANOVA. Numbers of observations (n) are given in tables and figure legends. For data shown in Figures 1 and and22, either one peptide concentration was tested in each experiment (Fig. 1) or antibodies not always added sequentially at 1 µg/ml then 5 µg/ml (Fig. 2: in five cases 1 µg/ml was followed by 5 µg/ml, in a further five cases the bilayer broke after exposure to 1 µg/ml, before exposure to 5 µg/ml antibody). Thus there were separate sets of control data for each concentration of F peptide or antibody, necessitating the use of ANOVA and post-hoc Mahalanobis tests. To reduce effects of variability in control parameters (PoC, ToC, TcC), and to evaluate parameters after F peptide (PoF, ToF, TcF) or antibody (PoA, ToA, TcA) addition, data were expressed as the difference between log10 XC and log10XF (or log10XA) for each channel (e.g., log10PoF − log10PoC). The difference from control was assessed with a paired t-test applied to log10PoC and log-10PoF. The difference between each concentration was assessed using ANOVA on log10PoC − log10PoF at each concentration with the post-hoc multidimensional Mahalanobis test. A p value of <0.05 was considered significant.
We demonstrate that the SPRY2 domain of RyR1, and specifically the F loop domain, contributes to RyR1 ion channel gating by the F loop interacting with a regulatory site on the RyR tetramer. We also show that the SPRY2 domain and its F loop bind to the ASI/basic region of the RyR via an electrostatic interaction mediated by specific acidic residues in the F loop with a series of basic residues in the ASI/basic region, as well as additional structural constraints within the ASI/basic α-helix. The results indicate that the functional consequence of the F loop interdomain interaction with the ASI/basic region may be interrupted by the influence of conditions specific to the intact cells on RyR1, in a similar manner to how DHPR interactions with RyR1 disrupt imperatoxin A sensitivity. The results also indicate that the previously identified impact of the ASI/basic region on EC coupling is not mediated by an interaction with the SPRY2 domain F loop.
The authors are grateful to S. Pace and J. Stivala for SR vesicle preparation, to M. Coggan for assistance with SPRY2 domain expression and purification, and L. Groom for generating the full-length RyR1 SPRY2 F loop mutants. We also thank Dr. P.D. Allen for providing access to the RyR1-null mice used in this study. The work was supported by the NHMRC (#471462 to Angela F. Dulhunty), the NIH (AR44657 to Robert T. Dirksen) and the Academia Dei Lincei Fund (Lan Wei).