Lethal Arg9Cys phospholamban mutation hinders Ca²⁺-ATPase regulation and phosphorylation by protein kinase A

Abstract

The regulatory interaction of phospholamban (PLN) with Ca²⁺-ATPase controls the uptake of calcium into the sarcoplasmic reticulum, modulating heart muscle contractility. A missense mutation in PLN cytoplasmic domain (R9C) triggers dilated cardiomyopathy in humans, leading to premature death. Using a combination of biochemical and biophysical techniques both in vitro and in live cells, we show that the R9C mutation increases the stability of the PLN pentameric assembly via disulfide bridge formation, preventing its binding to Ca²⁺-ATPase as well as phosphorylation by protein kinase A. These effects are enhanced under oxidizing conditions, suggesting that oxidative stress may exacerbate the cardiotoxic effects of the PLN^R9C mutant. These results reveal a regulatory role of the PLN pentamer in calcium homeostasis, going beyond the previously hypothesized role of passive storage for active monomers.

Heart failure (HF) is the leading cause of morbidity and mortality worldwide (1, 2). The most prominent disorder leading to HF is dilated cardiomyopathy (DCM), a disease characterized by left ventricular dilatation and impaired systolic function (1, 2). DCM has both acquired and genetic etiologies (1, 2). Recent genome sequencing has revealed a high incidence of DCM-associated mutations in cytoskeletal, nuclear, as well as sarcomeric proteins (3). A number of mutations have been indentified in calcium handling proteins, which play a central role in the mechanics of heart muscle contractility (3–6).

Cardiac muscle contraction (systole) begins when an action potential causes membrane depolarization, activating the sarcolemmal L-type calcium (Ca²⁺) channels. Ca²⁺ flows through the L-type Ca²⁺-channels into the cytosol. This increase in Ca²⁺ concentration induces a large-scale release of Ca²⁺ into the cytosol from intracellular stores by the sarcoplasmic reticulum (SR) Ca²⁺-release channels (or ryanodine receptors). Ca²⁺ then moves toward the contractile apparatus, where it binds the troponin complex and initiates contraction. Muscle relaxation (diastole) occurs when Ca²⁺ is sequestered into the SR by the SR Ca²⁺-ATPase (SERCA) (7) a membrane-embedded Ca²⁺ pump (8). SERCA is regulated by phospholamban (PLN), which reduces its apparent Ca²⁺ affinity (9, 10). PLN’s inhibition is reversed by cAMP-dependent protein kinase A (PKA), which phosphorylates PLN at Ser16, enhancing cardiac contractility and reestablishing Ca²⁺ flux (11).

PLN is a single-pass membrane protein, which comprises three structural domains (12–14), further subdivided into four dynamic domains [cytoplasm: domain Ia (residues 1–16), loop (residues 17–22), domain Ib (residues 23–30); transmembrane: domain II (residues 31–52)] (15) (Fig. S1). In membranes, PLN forms homopentamers arranged in a pinwheel topology that are in equilibrium with monomers (16, 17) that bind SERCA with 1∶1 stoichiometry (6, 18–21). Also, it has been proposed that the PLN monomer-pentamer equilibrium plays a central role in SERCA regulation (6).

Several naturally occurring mutations in the PLN gene have been linked to hereditary DCM (5), including a substitution of Arg9 for Cys (PLN^R9C) located in the cytoplasmic domain Ia of PLN (Fig. S1), which has been identified in several cases of familial DCM (22). R9C cardiotoxic effects are correlated with inefficient Ca²⁺ handling (23) and show a dose-dependent inhibition of SERCA (24). Schmitt et al. hypothesized that PLN^R9C leads to DCM by binding irreversibly to the catalytic subunit of PKA (PKA-C) and preventing PLN^R9C and/or PLN^wt phosphorylation at Ser16 (22). To date, however, there are no firm conclusions on the molecular mechanisms that link PLN^R9C to DCM.

Here, we used an array of biochemical and biophysical techniques both in vitro and in live cells to establish the molecular determinants of the cardiotoxic effects of PLN^R9C. Specifically, we focused on the effects of this aberrant mutation on (i) the recognition and phosphorylation by PKA-C, (ii) the PLN monomer-pentamer equilibrium, and (iii) SERCA regulation. We found that the R9C mutation stabilizes the pentameric assembly, hindering PLN deoligomerization, phosphorylation by PKA-C, and SERCA regulation. Importantly, we discovered that these effects are exacerbated under oxidative environments, which are related to both physiological and pathophysiology conditions of cardiac myocytes resulting from myocardial ischemia (25, 26).

Results

Our immediate objectives were to determine the effects of R9C mutation on (a) the PKA-C recognition and phosphorylation, (b) the PLN monomer-pentamer equilibrium, and (c) SERCA regulation. Toward these goals, we utilized three different PLN constructs with and without the R9C mutation: (i) synthetic peptides spanning cytoplasmic residues of PLN ( or ), (ii) full-length recombinant pentamers (PLN^wt and PLN^R9C), and (iii) recombinant monomeric PLN (AFA-PLN), where the three transmembrane cysteines (Cys36, Cys41, Cys46) were mutated into Ala, Phe, Ala, respectively. This triple mutation abolishes PLN oligomerization without altering PLN’s inhibitory function (27). We carried out these experiments in the presence of dithiothreitol (DTT ranging from 1 to 20 mM) or hydrogen peroxide (H₂O₂ ranging from 1–100 μM), chemicals commonly used to mimic physiological redox conditions and oxidative stress (28, 29).

Effects of the R9C Mutations on the Phosphorylation Kinetics by PKA-C.

Under reducing conditions, phosphorylation kinetics of synthetic PLN peptides were monitored using a coupled enzyme assay (30), standardized with a synthetic peptide corresponding to the minimal recognition sequence for the kinase (Kemptide) (31). Under our experimental conditions, recombinant PKA-C shows a catalytic efficiency typical for Kemptide (k_cat/K_M ∼ 0.78) (30, 32). Interestingly, we found that PKA-C is able to phosphorylate both and peptides with similar catalytic efficiencies (Fig. 1A and Table S1). Moreover, we carried out competitive kinetic assays in the presence of products phosphorylated at Ser16 ( or ). Our measurements did not show any substantial product inhibition (Fig. 1A). Also, we performed the experiments under oxidative conditions, varying the concentration of H₂O₂ from 1 to 100 μM. Because the coupled enzyme assay is incompatible with oxidizing agents, we monitored peptide phosphorylation using electrophoretic mobility shift assay (EMSA) and identified the products with electrospray ionization mass spectrometry (ESI-MS). We found that the PLN peptides are phosphorylated under both oxidizing and reducing conditions (Fig. 1B), but under oxidizing conditions, forms dimers, which can still be fully phosphorylated by PKA-C. Phosphorylation reactions were repeated with full-length PLN^R9C and PLN^wt. Interestingly, we did not detect any phosphorylation for pentameric PLN^R9C under either reducing or oxidizing conditions (Fig. 1B). The absence of phosphorylation of pentameric PLN^R9C was confirmed by EMSA and MALDI-TOF mass spectrometry (Fig. S2). In contrast, we found complete phosphorylation for the PLN^wt and monomeric AFA-PLN^R9C. The observed gel shift (Fig. 1B) is typical of AFA-PLN phosphorylation at Ser16 (33). Based on these results, we conclude that phosphorylation by PKA-C is impaired only for the R9C pentamer.

Fig. 1.

PKA-C phosphorylation reaction and SERCA inhibition assays for wild-type and R9C constructs of PLN. (A) Steady state phosphorylation kinetics of and and competition assays. (Left) Plot of the initial rates as a function of substrate concentration. (Right) Plot of the apparent K_M as a function of phosphorylated products ( and ). (B) SDS-PAGE gels for the phosphorylation reactions of , , AFA-PLN, AFA-PLN^R9C, PLN^wt, and PLN^R9C carried out under oxidizing (100 μM H₂O₂) and reducing (10 mM DTT) conditions. The number of phosphates per peptide was detected by ESI mass spectrometry. (C) Histograms showing the change in apparent Ca²⁺ affinity (ΔpK_Ca) of SERCA in the presence of phosphorylated and unphosphorylated PLN monomers (AFA-PLN and AFA-PLN^R9C) and pentameric PLN^wt and PLN^R9C constructs.

SERCA Activity Assays.

To characterize the efficacy of PLN^R9C to bind and reduce SERCA’s apparent Ca²⁺ affinity (pK_Ca), we performed coupled enzyme activity assays in reconstituted lipids. In agreement with Schmitt et al. (24), we found that monomeric AFA-PLN^R9C is a loss-of-function (LOF) mutant, with a partial inhibitory effect on SERCA; i.e., slight reduction in pK_Ca (Fig. 1C, Left). Phosphorylation at Ser16 relieves the inhibitory effect for both AFA-PLN^wt and AFA-PLN^R9C (Fig. 1C, Center). Remarkably, the PLN^R9C pentameric species is a total LOF (Fig. 1C, Right). Taken together, these results suggest that if PLN^R9C were to deoligomerize, it would be able to reversibly inhibit SERCA.

Stability of PLN Pentamer and Cys Accessibility.

Based on the results above, we deduced that PLN^R9C could in principle regulate SERCA, although with a reduced degree of inhibition. However, EMSA and mass spectrometry data show that the pentameric species is not phosphorylated at Ser 16. Therefore, the cardiotoxic species is possibly the pentameric assembly, which prevents phosphorylation and hampers the monomer-pentamer equilibrium necessary for the regulation of SERCA. To further test this hypothesis, we measured the stability of the pentamers (PLN^wt and PLN^R9C) using thermal unfolding and gel electrophoresis. At 25 °C and in the presence of 10 mM DTT, the pentamer to monomer ratio detected by SDS-PAGE for PLN^wt is approximately 4∶1 (78% pentamer). Fitting of the densitometry data from the SDS gels gave a melting temperature (T_m) of 45 ± 1 °C for the wild-type pentamer. Under the same conditions, the pentamer/monomer ratio for PLN^R9C increases noticeably (92% pentamer) and its T_m is 51 ± 1 °C (Fig. 2). The thermostability of the mutant is even more pronounced under oxidative conditions (100 μM H₂O₂). We obtained T_m values of 52 ± 2 °C and 67 ± 6 °C for PLN^wt and PLN^R9C, respectively. The SDS gels of the oxidized pentamers (Fig. 2) show some important features: (i) PLN^R9C pentamers have slightly less mobility than the PLN^wt, which suggests a change in the protein structural topology (i.e., hydrodynamic radius), and (ii) the presence of heat-resistant dimers. The latter was also observed in the oxidation studies of PLN^wt carried out by Froehlich et al. (34) (Fig. 2). To test whether the formation of the dimers is due to the cytoplasmic (Cys 9) or transmembrane (Cys36, Cys41, Cys46) cysteines, we carried out the same experiments with AFA-PLN^R9C. Under reducing conditions, AFA-PLN^R9C runs as a monomer on an SDS-PAGE gel; whereas under oxidizing conditions it forms dimers (Fig. 3A). behaves in a similar manner, suggesting a marked tendency of Cys9 to form intermolecular disulfide bridges. To further support the formation of disulfide bridges, we probed the presence of free thiols for the PLN variants with Ellman’s reagent (5,5′-dithiobis-(2-nitrobenzoic) acid; DTNB). For , we detected the formation of disulfide bridges even in the presence of moderate amounts of reducing agent (DTT < 1 mM). Under stronger reducing conditions (DTT > 10 mM), however, the runs as a monomer (Fig. 3A) and forms a 1∶1 adduct (DTNB: ) after ∼80 min (Fig. 3B). Under oxidizing conditions (ranging from 1 to 100 μM H₂O₂), forms dimers, with the Cys residues becoming completely inaccessible to DTNB. We repeated these measurements with PLN^R9C and PLN^wt and found that under reducing conditions PLN^R9C is much more reactive with DTNB than PLN^wt. We monitored the DTNB reaction up to ∼90 min. To reach completion, however, the reaction requires more than 18 h (35). Because the membrane-embedded Cys residues of PLN^wt react with DTNB very slowly (> 300 min), we assigned the fast rise of the binding curve for PLN^R9C to the reactivity of Cys 9 (Fig. 3B), which is more exposed to the soluble DTNB. Under oxidizing conditions, both PLN^R9C and PLN^wt behave identically, with very sluggish reaction kinetics with DTNB. The latter is in quantitative agreement with previous cysteine accessibility measurements carried out by Karim et al. (35). Overall, these data suggest that oxidation of Cys 9 results in formation of stable dimers and confers greater thermostability to PLN^R9C oligomers.

Fig. 2.

Thermostability of PLN^wt and PLN^R9C pentamers using electromobility shift assay. (A) SDS-PAGE gels for PLN^wt and PLN^R9C incubated at different temperatures under reducing (Upper) and oxidizing (Lower) conditions. (B) Plots of the percent pentamers obtained from densitometry analysis versus temperature for both PLN^wt and PLN^R9C under reducing (Upper) and oxidizing (Lower) conditions.

Fig. 3.

Electromobility shift and DTNB cysteine accessibility assays for wild-type and R9C constructs of PLN. (A) SDS-PAGE gels in Tris buffer, 1 mM DTT, and 100 μM H₂O₂ of (Left), and AFA-PLN^R9C (Right). (B) Plots of cysteine reactivity over time. Cysteine reactivity is measured by moles TNB produced per peptide or full-length protomer from the reaction with DTNB under reducing and oxidizing conditions. (Upper) Black squares, in Tris buffer; red squares, in Tris buffer; open red squares, in 1 mM DTT; and red crosses, in 100 μM H₂O₂. (Lower) Black squares, PLN^wt in Tris buffer; red squares, PLN^R9C in Tris buffer; open black squares, PLN^wt in 100 μM H₂O₂; and open red circles, PLN^R9C in 100 μM H₂O₂.

Probing PLN Oligomerization and SERCA Binding Using Quantitative Fluorescence Resonance Energy Transfer (FRET).

We probed the oligomerization of the PLN^wt and PLN^R9C pentamers in live AAV-293 cells using quantitative FRET measurements between fluorescent protein tags fused to PLN’s cytoplasmic domains (36, 37). Specifically, to detect intrapentameric FRET (i.e., FRET between protomers in each pentamer), we engineered PLN^wt or PLN^R9C with cerulean fluorescent protein (Cer) and yellow fluorescent protein (YFP), respectively, and coexpressed them in AAV-293 cells (36, 37). For SERCA binding assays, we tagged SERCA with cyan fluorescent protein (CFP) and measured FRET with YFP-PLN constructs. To quantify the dependence of FRET on PLN expression levels, we carried out a cell-by-cell survey of quantitative FRET and fluorescence intensity (an index of protein concentration). For both PLN constructs, we found that intrapentameric FRET increased with protein concentration to a maximum value (FRET_max) (Fig. 4A). For PLN^R9C, FRET_max is slightly higher than PLN^wt (p < 0.05). This corresponds to a modest decrease in the average distances between the Cer-YFP probes within the PLN^R9C pentamer relative to PLN^wt, which may reflect the formation of disulfide bridges in the pentameric mutant (Table S2). The curve representing the concentration dependence for PLN^R9C is left-shifted with respect to that of PLN^wt (Fig. 4A). The calculated dissociation constant (K_d1) was approximately 53% lower than that of PLN^wt, suggesting the formation of more stable pentamers for the R9C mutant (Table S2). Interestingly, the left-shift of the intrapentamer FRET curve for PLN^R9C corresponds to a right-shift of SERCA/PLN^R9C binding curve (Fig. 4B), with the calculated apparent dissociation constant (K_d2) approximately 130% greater than that of PLN^wt. We conclude that the PLN^R9C has much lower affinity for SERCA than PLN^wt and SERCA is not able to deoligomerize PLN^R9C as efficiently as PLN^wt (18, 19). This explains the LOF character of PLN^R9C measured by ATPase activity (22). Notably, we did not observe a significant difference in FRET_max for SERCA complexes with either PLN^wt or PLN^R9C, which suggests that both species probably bind SERCA in a similar manner.

Fig. 4.

In cell FRET measurements of the pentamer stability and SERCA regulation for wild-type and R9C constructs of PLN. (A) Intrapentameric FRET versus protein expression level for Cer-PLN^wt/YFP-PLN^wt and Cer-PLN^R9C/YFP-PLN^R9C (B) Percent of FRET efficiency from CFP-SERCA to YFP-PLN^wt and YFP-PLN^R9C. (C) Percent of intrapentameric FRET efficiency for Cer-PLN^wt/YFP-PLN^wt homoligomers and Cer-PLN^R9C/YFP-PLN^wt heteroligomers. (D) Plot of FRET_max versus competitor to quantify nonspecific FRET. (E). Plot of Cer F/F₀ versus time for YFP-PLN^R9C and Cer-PLN^R9C. Arrows indicate the time of addition of 100 μM H₂O₂. (F) YFP/Cer FRET ratio versus time. Arrows indicate the time of addition of 100 μM H₂O₂. (G) Plots of K_d1 (arbitrary units) and FRET_max upon addition of 100 μM H₂O₂ (arrow). (H) Fluorescence microscopy images of cells expressing Cer-PLN^R9C and YFP-PLN^R9C. Scale bar = 20 μm. Ratio color scale = 0.4–2.6.

Because patients with DCM carry both the wild-type and the mutant alleles, with the latter showing a dominant inheritance pattern (22, 24), we also tested the stability of mixed PLN^wt/PLN^R9C pentamers. We coexpressed Cer-PLN^R9C and YFP-PLN^wt and measured FRET between them. Fig. 4C shows that FRET_max is reduced for the mixed pentamers compared to PLN^wt homopentamers. This is consistent with a clustering of Cer-PLN^R9C cytoplasmic domains away from YFP-PLN^wt oligomers. Notably, mixed pentamers showed a small but reproducible reduction of K_d1 compared to that PLN^wt, suggesting that the interactions between the R9C mutants within the mixed pentamers prevent deoligomerization. To determine FRET specificity, we carried out competition assays with increasing amounts of a competitor PLN that cannot serve as a FRET acceptor. We found that by increasing the competitor concentration we obtained a reduction in FRET_max to a minimum value of 4% (Fig. 4D). This value represents the amount of nonspecific FRET that is subtracted from total FRET for interprobe distance calculations (Table S2).

In vitro assays reported in the previous sections and work from other groups (26, 34) demonstrate that both PLN^R9C and PLN^wt are sensitive to oxidation. Therefore, we treated the AAV-293 cells with 100 μM H₂O₂. Upon addition of H₂O₂, Cer-PLN^R9C fluorescence decreased by approximately 4% over the course of 5 min (Fig. 4E), with approximately 5% increase in the emission of YFP-PLN^R9C. The FRET ratio YFP/Cer increased by 10% for PLN^R9C (Fig. 4F and Movie S1), whereas no increase was detected for PLN^wt and AFA-PLN (Fig. 4F). To quantify the relative contributions to the observed FRET ratio of protein oligomerization and changes in distance between the fluorescent probes, we measured FRET_max at regular intervals and estimated K_d1 after H₂O₂ treatment. Fig. 4G shows that both FRET_max and K_d1 changed after addition of H₂O₂, with a ∼40% reduction in K_d1 and a ∼10% increase in FRET_max. This suggests an increase in PLN oligomerization with a slightly more compact conformation of the pentameric assembly. Note that we did not detect any large-scale aggregation of PLN^R9C either before or after treatment with H₂O₂. Such aggregation would appear as fluorescent puncta, which would be visible by wide-field fluorescence microscopy (Fig. 4H) or total internal reflection fluorescence (Fig. S3).

Discussion

Based on coimmunoprecipitation experiments, Schmitt et al. (22) proposed that PLN^R9C binds PKA-C irreversibly, creating dead-end complexes that deplete the local reservoir of kinase. The latter would reduce phosphorylation levels of PLN, with concomitant dysregulation of SERCA, leading to DCM. This interpretation accounts for the observed weak adrenergic responsiveness and dominant effect of PLN^R9C in heterozygous patients (22). In the present study, we directly tested this hypothesis using both in vitro and in cell experiments. We found that PKA-C was able to phosphorylate both a truncated peptide and monomeric AFA-PLN^R9C, which is still able to reversibly inhibit SERCA, although with lower efficacy than PLN^wt or AFA-PLN. Most importantly, kinetic assays under reducing conditions show that PKA-C is able to quantitatively phosphorylate with the same catalytic efficiency of . Under oxidizing conditions, is able to be phosphorylated in a similar manner to . Thus, we did not find evidence that this single mutation at the P-7 site of the recognition sequence of PKA-C interferes with the phosphorylation reaction. However, we found that phosphorylation of pentameric PLN^R9C is significantly impaired, which is consistent with previous reports (24). Therefore, the stabilization of the pentamer by this Arg to Cys substitution prevents PLN phosphorylation. This finding emphasizes the role of monomer-pentamer equilibrium in the SERCA regulatory mechanism by PLN. The latter is supported by in vivo studies carried out by Kranias and coworkers in mice models, which demonstrate the importance of PLN^wt over the monomeric mutant PLN^C41F for the optimal relaxation of cardiomyocytes (38).

Disulfide bridges in the cytoplasmic domains of PLN^R9C stabilize the pentamer, making it practically inaccessible to PKA-C and unable to deoligomerize and regulate SERCA. An important finding is the presence of dimers in oxidized PLN^R9C. The latter has been previously observed by Froehlich et al. upon PLN oxidation by nitroxyl radicals, which promote the formation of disulfide bonds in the transmembrane region, generating noninhibitory oligomers that prevent SERCA regulation (34). Under oxidative conditions, PLN^R9C oligomerization is enhanced. This is important given that ischemic oxidative stress conditions are prevailing features of pathological states such as heart failure (39, 40). Additionally, oxidative stresses are also frequent under acute β-adrenergic stimulation and even in nonpathological conditions, where transient oxidative stress could cause cumulative damage (41). Therefore, it is possible that deteriorating redox conditions in PLN^R9C-induced heart failure would reinforce anomalous PLN oligomerization and exacerbate the mutation’s effects on calcium cycling. Furthermore, we found that for mixed PLN^wt/PLN^R9C pentamers enhanced oligomerization is a dominant effect. This is consistent with the observed R9C phenotypes for both heterozygous mice (24) and human patients (22).

Based on the above considerations, we propose a model for PLN-SERCA disruption by the R9C mutation (Fig. 5). The principal effect of R9C is the formation of interprotomer disulfide bonds in the cytoplasmic domains, which is transient under reducing conditions and increases upon oxidation, stabilizing the PLN pentamer and rendering the recognition site for the kinase inaccessible. This also prevents PLN dissociation into monomers and formation of the regulatory complex (PLN:SERCA). These effects (enhanced SERCA activity and diminished phosphorylation) are reminiscent of ablation of PLN observed by Kranias and coworkers (42). The formation of the disulfide bridges in the pentamer hinders PLN phosphorylation by PKA, and possibly induces conformational or topological changes in PLN. We propose that these combined effects are involved in the development of DCM. An important corollary of this study is the emerging role of the PLN pentamer and the monomer-pentamer equilibrium (43). This lethal mutation revealed that oligomerization and deoligomerization of the PLN pentamer within the membrane is directly involved in the SERCA regulatory process.

Fig. 5.

Proposed model of the effects of R9C mutation on the SERCA regulatory cycle. The cardiotoxic effects are due to the stability of the PLN^R9C pentameric assembly, which prevents deoligomerization, phosphorylation by PKA-C, and regulation of SERCA by the monomeric species. Oxidative stress pushes the equilibrium toward the pentamer, making the PLN^R9C pentamer more stable and causing the formation of dimeric species for both PLN^wt and PLN^R9C (not shown in the model) that are probably unable to regulate SERCA.

Experimental Procedures

Sample Preparation and Kinetic Assays.

Recombinant PKA-C was expressed, purified, and assayed as previously reported (32, 44). All peptides ( and ) were synthesized on a Liberty 12-channel Automated Microwave Synthesizer from CEM (Matthews, NC) (see SI Text). Phosphorylation reactions were performed at 25 °C and monitored by following coupled enzyme (12 units lactate dehydrogenase and 4 units pyruvate kinase) mediated consumption of NADH at 340 nm using a Spectromax microplate reader (Molecular Devices) (30) as previously reported (32, 44). Reaction solutions contained 50 mM 3-(N-morpholino) propanesulfonic acid (MOPS) (pH 7.0), 64 nM PKA-C, 5 mM ATP, and 10 mM MgCl₂, with substrate concentrations ranging between 20–300 μM (32, 44). Inhibitor studies were performed with the addition of phosphorylated or (0.5 or 1.0 mM) to the reaction solution. For phosphorylation reactions in the presence of H₂O₂, 300 μM of substrate was incubated for 10 min in the reaction solution before initiating phosphorylation with 64 nM PKA-C. The reactions were stopped after 10 min by the addition of 0.5% TFA, and analyzed by EMSA (25% SDS-PAGE gels) stained by Coomassie Blue and ESI-MS after desalting with a C8 Zip-Tip (Millipore).

Thermostability of the Pentamer.

PLN pentamer thermostability was monitored by SDS-PAGE gels. For the reducing conditions, PLN^wt and PLN^R9C samples contained 100 mM Tris buffer, pH 6.8, 3% sodium dodecyl sulfate, 8% (v/v) glycerol, using a range of 5 to 10 mM DTT; whereas for oxidizing conditions the samples were incubated with 100 μM H₂O₂ for 20 min at each temperature. Each sample (3 μg total mass) was loaded into 5% Next Gels (Amresco) or a 14% Tris-tricine SDS-PAGE gels. The Coomassie-stained gels were quantitated using ImageJ software (45).

SERCA Activity Assays.

PLN variants were coreconstituted with purified SERCA (46, 47) in lipid bilayer membranes (1,2-dioleoyl-sn-glycero-3-phosphocholine: 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine, DOPC:DOPE, 4∶1) at molar ratios of 10∶1 PLN:SERCA and 700∶1 lipids:SERCA. The Ca²⁺ dependence of the ATPase activity was measured using a coupled enzyme assay at 37 °C (48) and monitored as for the PKA-C assays. Initial rates of SERCA was measured as a function of calcium concentration (pCa), and data were fit to the Hill equation (48).

Cysteine Accessibility Measurements.

Free thiols were assayed via titrations with 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB) (49). Samples (typically 100 μM) were dissolved in 60 mM Tris buffer (pH 8.0) and 1 mM EDTA and added to a reaction vessel containing 100 mM Tris and 0.3 mM DTNB. The reactions were monitored at a wavelength of 412 nm (49). For oxidized and reduced conditions, we incubated PLN samples for 20 min in 100 μM H₂O₂ and 1 mM DTT, respectively. The excess of reducing agent was eliminated with NaAsO₂ (50).

Dynamic FRET.

Transfected cells were washed with phosphate buffered saline (PBS) and imaged by epifluorescence imaging at 1 min time intervals with excitation at 427/10 nm and emission at 472/30 nm (for Cer) or 542/27 nm (for YFP). After 5 min of acquisition, the buffer was replaced with 100 μM H₂O₂ in PBS and acquisition continued for 15 min. Mean F/F₀ of all cells (± SE) at each time point was calculated for each filter configuration. The FRET ratio was calculated as the quotient of (Cer F/F₀)/(YFP F/F₀), with a combined error of . Images of YFP fluorescence were divided by images of Cer fluorescence (both 427/10 nm excitation) using ImageJ software (45).

Quantitative FRET.

FRET_max and dissociation constants for the complexes were determined as described previously (36, 37). The observed FRET was calculated for each cell from the extent of donor fluorescence enhancement after acceptor photobleaching, according to E = 1 - (F_prebleach/F_postbleach). For repetitive, nondestructive measurements, FRET was quantified with a “3-cube” method (E-FRET) (36, 51). FRET efficiency of each cell was compared to that cell’s starting YFP fluorescence (an index of protein concentration). FRET concentration dependence was fit by a hyperbolic curve (36). Regulatory complex probe separation distance was calculated using R = (R₀)[(1/FRET_max) - 1)]^1/6 (52). The distance between fluorescent probes in PLN pentamers was calculated according to a ring-shaped oligomer model as previously described (36), with a Förster radius (R₀) of 49.2 Å for Cer-YFP energy transfer (53, 54). Non-specific FRET between unbound donors and acceptors was determined by measuring the reduction in FRET from YFP - PLN^wt to mCherry - PLN^wt in cells coexpressing increasing amounts of competing CFP - PLN^wt (Fig. S4).