Dimerization of SERCA2a Enhances Transport Rate and Improves Energetic Efficiency in Living Cells

Abstract

The type 2a sarco/endoplasmic reticulum (ER) Ca²⁺-ATPase (SERCA2a) plays a key role in intracellular Ca²⁺ regulation in the heart. We have previously shown evidence of stable homodimers of SERCA2a in heterologous cells and cardiomyocytes. However, the functional significance of the pump dimerization remains unclear. Here, we analyzed how SERCA2a dimerization affects ER Ca²⁺ transport. Fluorescence resonance energy transfer experiments in HEK293 cells transfected with fluorescently labeled SERCA2a revealed increasing dimerization of Ca²⁺ pumps with increasing expression level. This concentration-dependent dimerization provided means of comparison of the functional characteristics of monomeric and dimeric pumps. SERCA-mediated Ca²⁺ uptake was measured with the ER-targeted Ca²⁺ sensor R-CEPIA1er in cells cotransfected with SERCA2a and ryanodine receptor. For each individual cell, the maximal ER Ca²⁺ uptake rate and the maximal Ca²⁺ load, together with the pump expression level, were analyzed. This analysis revealed that the ER Ca²⁺ uptake rate increased as a function of SERCA2a expression, with a particularly steep, nonlinear increase at high expression levels. Interestingly, the maximal ER Ca²⁺ load also increased with an increase in the pump expression level, suggesting improved catalytic efficiency of the dimeric species. Reciprocally, thapsigargin inhibition of a fraction of the population of SERCA2a reduced not only the maximal ER Ca²⁺ uptake rate but also the maximal Ca²⁺ load. These data suggest that SERCA2a dimerization regulates Ca²⁺ transport by improving both the SERCA2a turnover rate and catalytic efficacy. Analysis of ER Ca²⁺ uptake in cells cotransfected with human wild-type SERCA2a (SERCA2a^WT) and SERCA2a mutants with different catalytic activity revealed that an intact catalytic cycle in both protomers is required for enhancing the efficacy of Ca²⁺ transport by a dimer. The data are consistent with the hypothesis of functional coupling of two SERCA2a protomers in a dimer that reduces the energy barrier of rate-limiting steps of the catalytic cycle of Ca²⁺ transport.

Significance

The type 2a sarco/endoplasmic reticulum Ca2+-ATPase (SERCA2a) Ca²⁺ pump plays a central role in heart function. The speed at which SERCA2a removes Ca²⁺ from the cytosol is an important determinant of the rate of heart relaxation and the strength of heart contraction. It is not surprising that impaired SERCA2a function has been reported in several pathological conditions, including heart failure. Thus, understanding mechanisms of SERCA2a regulation is of great clinical importance. This study revealed a noteworthy aspect of transporter regulation. Increasing SERCA2a expression level enhances the pump’s dimerization, which in turn improves SERCA2a kinetics and catalytic efficacy. These findings may provide new insight into the mechanisms of abnormal Ca²⁺ regulation under conditions of decreased SERCA2a level, as seen in heart failure.

Introduction

Regular heart contraction heavily relies on well-controlled intracellular Ca²⁺ cycling. Playing a particularly important role in this process is the type 2a sarco/endoplasmic reticulum (ER) Ca²⁺-ATPase (SERCA2a). The speed at which SERCA2a removes Ca²⁺ from the cytosol determines the rate of heart relaxation during diastole. SERCA2a activity also sets the total amount of Ca²⁺ in the sarcoplasmic reticulum (SR; SR Ca²⁺ load), which determines the amplitude of SR Ca²⁺ release and the strength of systolic contraction. Impaired SERCA2a function has been reported in a number of pathological conditions, including heart failure (HF) (1, 2, 3). Consequently, SERCA2a has attracted attention as a potential target for therapeutic approaches to improve heart function in HF patients (4). So far, recent clinical trials that aimed to restore HF function by overexpressing SERCA2a have yielded disappointing outcomes (5). Therefore, there is a critical need to determine the mechanisms of SERCA2a regulation, mechanisms that may be the basis of alternative new therapies to enhance the function of endogenous pumps in the diseased heart.

During an action potential, global SR Ca²⁺ release via ryanodine receptors (RyRs) initiates heart contraction. For relaxation to occur, RyRs must close, allowing SERCA2a to pump cytosolic Ca²⁺ back into the SR and the Na⁺-Ca²⁺ exchanger to extrude Ca²⁺ from the cell (6,7). Competition between these two Ca²⁺ transporters determines the SR Ca²⁺ load and the strength of contraction. Because of its central role in intracellular Ca²⁺ homeostasis and the cardiac cycle, SERCA2a activity must be tightly regulated. Through regulation of SERCA2a Ca²⁺ uptake rate, the heart can swiftly adjust cardiac output during a transition from rest to exercise and vice versa. The principal mechanism of SERCA2a regulation occurs through an interaction with the small transmembrane peptide phospholamban (PLB). In resting conditions, PLB binding to SERCA2a causes an inhibition of SR Ca²⁺ transport by decreasing pump affinity for cytosolic Ca²⁺. A decreased Ca²⁺ transport rate leads to a decrease in SR Ca²⁺ load, decreased SR Ca²⁺ release, and weaker myocyte contraction. Conversely, during adrenergic stimulation, PLB phosphorylation by protein kinase A relieves SERCA2a inhibition and increases SR Ca²⁺ uptake (8,9). Stimulation of SR Ca²⁺ uptake by PLB phosphorylation enhances cardiac muscle relaxation (improving filling of the ventricles) and increases contraction strength (improving cardiac output) (10). Besides PLB regulation, SERCA2a also interacts with a wide array of small proteins, including histidine-rich protein, calreticulin, S100A, and newly discovered micropeptides (11,12). Additionally, several groups have shown evidence of a physical interaction of SERCA protomers in functionally coupled oligomeric complexes (13, 14, 15, 16). Recently, we provided evidence that these oligomeric complexes are SERCA2a homodimers and this dimerization is constitutive, irrespective of binding of PLB, ATP, thapsigargin, or Ca²⁺ to the pump (17). Although the functional significance of SERCA dimerization remains unclear (15,18,19), previous kinetics studies generated the intriguing hypothesis that the pumps in a dimer cycle asynchronously and their physical motions are coupled together. In this model of “conformational coupling,” the favorable conformational changes of one pump are physically coupled to the unfavorable transitions of the other member of the dimer. This proposed mechanism enhances cycling kinetics by reducing the activation energy barrier of the rate-limiting step (20).

Here, we take advantage of a newly developed approach that allows direct measurement of SERCA-mediated Ca²⁺ transport in living cells (21). Using this assay, together with the SERCA-SERCA fluorescence resonance energy transfer (FRET) analysis, (17), we correlated the SERCA2a transport kinetics and thermodynamics with pump dimerization to explore the functional role of dimerization. We found that the pump Ca²⁺ uptake rate and the maximal ER Ca²⁺ load are enhanced with increasing SERCA2a dimerization. We also found that the pump’s ATPase activity on both protomers is essential for enhancing kinetic and catalytic efficacy of a SERCA2a dimer. These findings revealed new insight into the role of SERCA2a dimerization in modulating the pump’s transport kinetics and catalytic efficacy.

Materials and Methods

Vectors production

pCMV R-CEPIA1er was a gift from Dr. Masamitsu Iino (Addgene, Watertown, MA). The vector encoding the human RyR2 cDNA fused to GFP at the N-terminus domain was kindly provided by Dr. Christopher George (University of Cardiff, Cardiff, UK). The vector encoding human wild-type (WT) SERCA2a (SERCA2a^WT) cDNA was kindly provided by Dr. David Thomas (University of Minnesota, Minneapolis, MN). The SERCA2a cDNA was cloned into the mCerluean-M1 (mCer)-modified plasmid (Addgene) using KpnI and NotI restriction enzymes, yielding SERCA2a fused to a modified Cerulean fluorescent protein (mCer) at the N-terminus. The mCyRFP1 (22) vector and the pcDNA3.1-mMaroon1 (23) were obtained from Addgene. Human SERCA2a was fused to the C-terminus of mCyRFP1-C1. mMaroon1 was amplified from the pcDNA3.1 vector using the polymerase chain reaction and inserted into YFP-C1 (24) in place of the YFP fluorescent tag. The sequence for human PLB was synthesized by GenScript (Piscataway, NJ) and inserted at the 3′ end of mMaroon1 to create a mMaroon1-PLB fusion construct. SERCA2a cDNA was also cloned into the inducible expression vector pcDNA5/FRT/TO for the generation of a SERCA2a stable cell line. The sequences were all verified by single-pass primer extension analysis (ACGT, Wheeling, IL). SERCA2a Y762G and E309Q mutants were generated using a site-directed mutagenesis kit (Q5 Site-Directed Mutagenesis Kit; New England BioLabs, Ipswich, MA). Specific primers containing the mutations were used to amplify the whole plasmid containing the Cerulean-tagged SERCA2a gene. After verification of the mutagenesis by single-pass analysis, the plasmids were amplified and used for experimentation.

Generation of SERCA2a stable line

A stable inducible Flp-In T-Rex-293 cell line expressing SERCA2a was generated using the Flp-In T-REx Core Kit (Invitrogen, Carlsbad, CA), as described before (21). Flp-In T-REx-293 cells were cotransfected with the pOG44 vector encoding the Flp recombinase and the expression vector pcDNA5/FRT/TO containing the SERCA2a cDNA (without a fluorescent tag). 48 h after transfection, the growth medium was replaced with a selection medium containing 100 μg/mL hygromycin. The hygromycin-resistant cell foci were selected and expended. Stable line cells were cultured in high-glucose Dulbecco’s Modified Eagle Medium supplemented with 100 units/mL penicillin, 100 mg/mL streptomycin, and 10% fetal bovine serum at 5% CO₂ and 37°C. Expression of SERCA2a in the stable cell line was verified by Western blot analysis 48 h after induction of recombinant pump expression with 1 μg/mL tetracycline.

Western blot analysis

The SERCA2a stable cell line was used to collect three different protein samples for Western blot analysis. In the first and second samples, SERCA2a expression was induced 24 and 48 h, respectively, before protein collection. In the third sample, SERCA2a stable Flp-In T-Rex-293 cells were transfected with mCer-SERCA2a plasmid. Protein collection occurred 48 h after transfection and induction of SERCA2a expression. Cell samples were lysed in a buffer containing 1% Triton and protease inhibitors. Protein quantification was performed using the Lowry-based approach, DC Protein Assay (Bio-Rad Laboratories, Hercules, CA). Equal amounts of protein were run on an SDS-PAGE and blotted on nitrocellulose using the Turbo Transfer System (Bio-Rad Laboratories). Membranes were incubated with primary antibody anti-SERCA2 (IID8; Santa Cruz Biotechnology, Dallas, TX) and developed with the HRP-conjugated secondary antibody. Western blots were imaged in a ChemiDoc (Bio-Rad Laboratories) apparatus after incubation with the HRP Chemiluminescent Substrate (Millipore, Burlington MA).

Experimental protocol

A similar experimental protocol, as described in (21), was used to study SERCA2a function. Human embryonic kidney (HEK)293 cells at 60–80% confluency were transiently cotransfected with plasmids containing the cDNA of RyR2, SERCA2a (WT or mutant), and R-CEPIA1er. Experiments were conducted 48 h after transfection to obtain the optimal level of recombinant proteins expression. Flp-In T-Rex-293 stably expressing SERCA2a were cotransfected with plasmids containing the cDNA of RyR2 and R-CEPIA1er, using the same conditions for the HEK293 cells. The surface membrane was permeabilized with saponin (0.005%). Experiments were conducted after washout of saponin with an experimental solution: 100 mM K-aspartate; 15 mM KCl; 5 mM KH₂PO₄; 5 mM MgATP; 0.35 mM EGTA; 0.22 mM CaCl₂; 0.75 mM MgCl₂; 10 mM HEPES; and dextran (MW: 40,000) 2% (pH 7.2). Free [Ca²⁺] and [Mg²⁺] of this solution were 200 nM and 1 mM, respectively.

Acceptor sensitization FRET and high-throughput cell scoring

Acceptor sensitization FRET was quantified, as previously described (17). HEK293 cells were transiently transfected with mCer-SERCA2a and YFP-SERCA2a in a 1:5 M plasmid ratio, and FRET was measured by automated fluorescence microscopy before and after permeabilization with 0.005% w/v saponin, as described above. For each condition, four sets of 240 images were collected with a 40×0.75 NA objective with 100-ms exposure for each channel: Cer, YFP, and FRET (Cer excitation/YFP emission). Cells that had a diameter of 35–75 μm and expressed both mCer and YFP at an average intensity of 100 counts above the background were automatically scored for fluorescence intensity using a multiwavelength cell scoring application in MetaMorph. FRET efficiency was calculated according to E = G/(G + 4.742 × F_Cer), where G = F_FRET − a × F_YFP − d × F_Cer and F_FRET, F_YFP, and F_Cer are the fluorescence intensities from FRET, YFP, and Cer images, respectively, and G represents FRET intensity corrected for the bleedthrough of the channels. The parameters a and d are bleedthrough constants, calculated as a = F_FRET/F_YFP for a control sample transfected only with YFP-SERCA and d = F_FRET/F_Cer for a control sample transfected only with mCer-SERCA. For our experimental setup, a and d were 0.075 and 0.877, respectively. FRET efficiency for each scored cell was plotted as a function of expressed protein concentration, as determined from the fluorescence intensity of the yellow fluorescent protein channel. The maximal saturating FRET efficiency (FRET_max) was derived from a hyperbolic fit of the data. The value was used to estimate the donor-acceptor distance, using a simulation in MATLAB (The MathWorks, Natick, MA) that assumes that the mCer- and YFP-labeled SERCA2a protomers randomly sort into donor-acceptor, donor-donor, and acceptor-acceptor dimers (25). The model used a Förster radius (R₀) of 54.1 Å for the mCer/YFP pair.

Time-correlated single-photon counting

Time-correlated single-photon counting (TCSPC) data acquisition was performed, as previously described (17). TCSPC histograms were obtained using HEK AAV-293 cells expressing mCyRFP1-SERCA2a alone or coexpressing mCyRFP1-SERCA2a with mMaroon1-PLB. Fluorescently labeled proteins were manually selected using 500-mm focal length planoconvex lens in a flip mount to defocus the excitation supercontinuum laser beam (Fianium, Southampton, U.K.) to excite the whole cell. To excite mCyRFP1, the excitation laser beam was filtered through a 464–500-nm bandpass filter, and emitted fluorescence was detected using 607–683-nm bandpass filter. After selection of a cell for spectroscopy, the defocusing lens was removed from the light path, the laser intensity was attenuated with a 1.0 neutral-density filter, and the laser focus was positioned on a region of the cell that yielded a fluorescence intensity of 100,000 photons/s. Under these excitation conditions, we did not see significant photobleaching during the acquisition period. Fluorescence was detected through a 60 × 1.2 NA water-immersion objective with a photon counting detector (Micro Photon Devices, Bolzano, Italy) and TCSPC module (PicoHarp300; PicoQuant, Berlin, Germany) using a time channel width of 16 ps. 60 s of acquisition was performed for each cell, excluding cells that showed significant fluorescence intensity changes due to movement. Under these conditions, the free mCyRFP1 showed single-exponential decay kinetics. After fusion of mCyRFP1 to SERCA2a, the decay showed an additional decay component, but this was a minor fraction (<4% of total amplitude). The decay was still well described by a single-exponential fit (average χ² = 1.065) with a fluorescence lifetime of 3.52 ns, and this lifetime was recovered in subsequent global fitting of donor-plus-acceptor FRET samples. Fluorescence decay histograms obtained from cells expressing mCyRFP1-SERCA and mMaroon1-PLB were analyzed using global analysis in SymPhoTime 64 software with shared fluorescence lifetimes and freely variable amplitudes. The data were poorly described by single- and double-exponential models, as evidenced by an uneven distribution of residuals and a poor reduced χ². Fitting was significantly improved by a triple-exponential model. There was only modest additional improvement with a quadruple-exponential fit that yielded implausible parameter values, including negative amplitudes. The FRET efficiency was determined for each of the lifetime components, according to the relationship E = 1 − (τ_DA/τ_D), where τ_D is the lifetime of the donor alone and τ_DA is the shorter lifetime due to FRET. Probe separation distances were calculated from FRET efficiency using the relationship R = R₀(E⁻¹ − 1)^1/6, where R₀ is the Förster distance for mCyRFP1-mMaroon1 pair (63.3 Å) (22) and E is the average FRET efficiency from lifetime measurements.

Confocal microscopy

Expression of recombinant proteins and changes in the luminal ER [Ca²⁺] ([Ca]_ER) were measured with laser scanning confocal microscopy (Radiance2000 MP; Bio-Rad Laboratories, Watford, UK) equipped with a ×40 oil-immersion objective lens (NA = 1.3). To measure the expression of SERCA2a, mCer was excited with the 458-nm line of the argon laser, and the signal was collected at >485 nm. To verify and quantify expression of RyR2, the GFP was excited with the 488-nm line of the argon laser, and the signal was collected at >515 nm. Two-dimensional images were collected at a speed of 6 ms/line. [Ca²⁺]_ER was recorded as changes in fluorescence intensity of the genetically encoded ER-targeted Ca²⁺ sensor R-CEPIA1er (21,26). R-CEPIA1er was excited with a 514-nm line of the argon laser, and the signal was collected at >560 nm. Line-scan images were collected at a speed of 20 ms/line. The R-CEPIA1er signal (F) was converted to [Ca²⁺]_ER by the following formula: [Ca²⁺]_SE = K_d × [(F − F_min)/(F_max − F)]. F_max was recorded in 5 mM Ca²⁺ and 5 μM ionomycin, and F_min was recorded after ER Ca²⁺ depletion with 5 mM caffeine. The K_d (Ca²⁺ dissociation constant) was 564 μM (27). SERCA-mediated Ca²⁺ uptake was calculated as the first derivative of [Ca²⁺]_ER refilling (d[Ca²⁺]_ER/dt) after RyR2 inhibition with ruthenium red (15 μM) and tetracaine (1 mM). The RyR2-independent Ca²⁺ leak was analyzed as the first derivative of [Ca²⁺]_ER decline (d[Ca²⁺]_ER/dt) after simultaneous inhibition of RyR2 and SERCA (with thapsigargin, 1 μM). ER Ca²⁺ uptake and Ca²⁺ leak rates were plotted as a function of [Ca²⁺]_ER to analyze the maximal ER Ca²⁺ uptake rate and maximal ER Ca²⁺ load. All two-dimensional images and line-scan measurements for [Ca²⁺]_SR were analyzed with ImageJ software (National Institutes of Health, Bethesda, MD).

Statistics

Data are presented as mean ± standard error of the mean (SEM) of n measurements. Statistical comparisons between groups were performed with the Student’s t-test for unpaired data sets. Differences were considered statistically significant at p < 0.05. Statistical analysis and graphical representation of averaged data were carried out on OriginPro7.5 software (OriginLab).

Results

SERCA2a dimerizes in permeabilized cells

We have previously detected dimerization of SERCA in cardiac myocytes using photoaffinity cross-linking (17). To evaluate SERCA-SERCA interactions under conditions suitable for quantitative Ca²⁺ uptake measurements, we performed Western blot analysis of SERCA2a-expressing Flp-In T-Rex-293 cells without addition of any cross-linking reagents. After 24 h of expression, SERCA dimers were detected as a band with a molecular weight of 200 kDa (Fig. 1 A, Lane 1), and the degree of dimerization was further increased with increasing protein expression at 48 h (Fig. 1 A, Lane 2). As expected, coexpression of SERCA2a fused to mCer (MW 100 + 27 kDa) resulted in an additional band above 100 kDa and widened the indistinct dimer band above 200 kDa (Fig. 1 A, Lane 3). Similar results were obtained in four other experiments.

Figure 1.

Effect of SERCA2a expression on the pump dimerization in HEK293 cells. (A) Western blot analysis of SERCA2a expression in HEK cells is shown. (B) FRET from donor-labeled mCer-SERCA2a to acceptor-labeled YFP-SERCA2a increased with increasing protein expression, suggesting SERCA dimerizes in permeabilized HEK cells. Data represent 640 cells from four independent experiments. The inset shows a SERCA dimer labeled with a donor (D) and acceptor (A) separated by 58 Å. (C) The fluorescence decay of mCyRFP-SERCA alone (black) and mCyRFP-SERCA interacting with mMaroon1-PLB (red) is shown. The inset shows multiexponential decay kinetics, suggesting a complex of SERCA-SERCA-PLB (blue) with two FRET distances; D1-A = 90 Å, D2-A = 56 Å. (D) The relative amplitude is given for each decay component for 11 different cells from three independent experiments, plotted as a function of each cell’s amplitude-weighted average lifetime. As FRET increased, the two shorter lifetimes increased together at the expense of the non-FRET (donor-only) lifetime (3.53 ns). The data are consistent with two FRET distances in a single complex.

Presumably, SERCA dimers that survive the denaturing conditions of SDS-PAGE represent a small fraction of the native SERCA-SERCA interactions. To quantify SERCA2a dimerization in cells, we performed FRET microscopy, as previously described (17). Acceptor sensitization FRET measurements of HEK293 cells transiently transfected with mCer-SERCA2a and YFP-SERCA2a showed that FRET was low for cells expressing low levels of YFP-SERCA2a, and FRET increased with increased protein expression (Fig. 1 B). This SERCA-SERCA binding curve was well described by a hyperbolic fit with a maximal value of 32% (FRET_max). At this maximum, all the protein-protein interactions are saturated—that is, all mCer-SERCA2a molecules are interacting with YFP-SERCA2a. Thus, the value FRET_max represents the intrinsic FRET of the SERCA2a dimer (Fig. 1 B, inset), corresponding to a donor-acceptor distance of 58.3 Å, according to a MATLAB model of intraoligomeric FRET (25). Saponin permeabilization for the purpose of Ca²⁺ transport analysis resulted in a modest decrease in FRET_max from 32 to 28%. The decreased FRET_max is likely due to sensitivity of the fluorescent protein tags to the replacement of the cellular cytoplasm with permeabilization buffer.

We previously inferred the presence of SERCA dimers from fluorescence lifetime measurements of FRET from SERCA to its regulatory partner PLB. In that study, TCSPC experiments resolved two populations of FRET donors. At maximal expression levels, at which FRET was saturated, only 50% of the donors (GFP-SERCA) participated in FRET with an acceptor (mCherry-PLB) at a donor-acceptor separation distance of 54 Å. The other 50% of donors showed an unchanged fluorescence lifetime, suggesting the mCherry acceptor on PLB was too far away for detectable FRET. Such a “donor-donor-acceptor” assembly was compatible with a regulatory complex consisting of two SERCAs and a single PLB (17). In this study, we measured the longer donor-acceptor distance by engineering new fusion constructs with the FRET pair mCyRFP1 and mMaroon1. This pair has a Förster distance of 63Å, 9 Å longer than the GFP-mCherry pair. Fig. 1 C shows the fluorescence decay of a cell expressing mCyRFP1-SERCA (Fig. 1 C, black). This donor-only fluorescence decay was mostly monoexponential, with a fluorescence lifetime of 3.51 ns. Coexpression of mMaroon1-PLB yielded a shorter lifetime, consistent with FRET (Fig. 1 C, red). Global fitting revealed three decay components: 3.53 ns (donor alone) and two lifetimes attributable to FRET, 3.15 ns (long distance) and 1.14 ns (short distance). These lifetimes correspond to distances of 90 and 56 Å. The latter value compares favorably with the previous estimate of 54 Å obtained with a different FRET pair (17). The 90-Å distance measured with the mCyRFP1-mMaroon1 pair would yield only 4% FRET with the GFP-mCherry pair, too low to resolve against the donor-only decay component in our previous study.

Transiently transfected cells expressed fluorescently labeled proteins at different concentrations. Cell-by-cell comparison showed that the fluorescence lifetime decreased (FRET increased) with increasing protein expression, as previously observed (17). Fig. 1 D shows the results of global fitting that revealed the relationship between the relative amplitude of each decay component (on the ordinate) and the average lifetime measured from each cell (on the abscissa). The contribution of the population with a 3.53-ns lifetime (donor alone) decreased toward zero as average lifetime decreased toward a minimum of ∼2.25 ns (on the abscissa). Over this range of average lifetimes, the 3.15-ns component and the 1.14-ns component increased together, maintaining consistently similar amplitudes. This is consistent with the hypothesis that both FRET distances arise from the same ternary complex, comprising a single PLB bound to a SERCA-SERCA dimer.

ER Ca²⁺ transport analysis in permeabilized cells

In the next experiments, we used a newly developed approach to measure SERCA-mediated Ca²⁺ transport in living cells (21). Changes in [Ca²⁺]_ER were directly recorded with the Ca²⁺ sensor R-CEPIA1er in HEK293 cells expressing mCer-SERCA2a. The cells were also transfected with the Ca²⁺ release channel RyR2, which enables pharmacological manipulation of ER Ca²⁺ load. The plasma membrane was permeabilized with saponin to control the cytosolic environment, including [Ca²⁺] and [ATP]. The SERCA2a function was quantified from the rate of [Ca²⁺]_ER reuptake after full ER Ca²⁺ depletion by caffeine (Fig. 2 A) followed by RyR2 inhibition with ruthenium red and tetracaine (RR + Tetr). For each individual cell, the SERCA2a expression level was estimated from the mCer signal. The contribution of the endogenous SERCA2b (Endo; Fig. 2 A) to the total Ca²⁺ uptake in these cells were relatively constant and accounted for less than 10% (21). Thus, changes in ER Ca²⁺ transport between different cells were mainly due to differences in the SERCA2a expression level.

Figure 2.

Effect of SERCA2a expression level on ER Ca²⁺ uptake and ER Ca²⁺ load. (A) Experimental protocol to measure SERCA-mediated ER Ca²⁺ uptake is shown. Changes in [Ca²⁺]_ER in control conditions during RyR activation with caffeine (Caf) and RyR inhibition with ruthenium red and tetracaine (RR+Tetr) are shown. (B) For each individual cell, ER Ca²⁺ uptake was analyzed as the first derivative (d[Ca²⁺]_ER/dt) and plotted as a function of [Ca²⁺]_ER. At the end of experiment, the RyR-independent Ca²⁺ leak was measured (data not shown) and subtracted from ER Ca²⁺ uptake to determine the maximal Ca²⁺ uptake rate and the maximal Ca²⁺ load. (C) The relationships between the maximal ER Ca²⁺ uptake rate and [SERCA2a] are shown. (D) The relationships between the maximal ER Ca²⁺ load and [SERCA2a] are shown. Experimental data from 39 different cells were grouped according to the mCerSERCA2a signal (bin size was 50 AU). Data are mean ± SE.

R-CEPIA1er was calibrated to convert the fluorescent signal to [Ca²⁺]_ER (for details, see Material and Methods). To quantify SERCA Ca²⁺ transport, the first derivative of ER Ca²⁺ uptake (d[Ca²⁺]_ER/dt) was plotted against the corresponding [Ca²⁺]_ER to estimate the maximal ER Ca²⁺ uptake rate and the maximal ER Ca²⁺ load (Fig. 2 B). The maximal uptake rate is determined by the kinetics of the SERCA catalytic cycle. The maximal ER Ca²⁺ load is determined by the catalytic efficiency of the pump and the thermodynamic limit set by the ATP/ADP ratio. Importantly, maximal ER Ca²⁺ load should be distinguished from physiological cardiac SR Ca²⁺ load. In cardiomyocytes, the peak diastolic Ca²⁺ load in the SR is not at its maximal potential because it is reduced from the theoretical limit by Ca²⁺ leak pathways (mostly through RyR2 (28,29)). Here, we avoid confounding effects of ER Ca²⁺ leak by inhibition of RyR2 with RR and Tetr during the Ca²⁺ refilling phase (Fig. 2 A). Moreover, the time allowed for ER Ca²⁺ uptake is sufficient to achieve steady state, revealing the maximal ER load that SERCA2a can support. At the end of each experiment, the RyR2-independent Ca²⁺ leak was measured (data not shown) and subtracted from the ER Ca²⁺ uptake to determine the true value of the maximal Ca²⁺ load. This value provides insight into the efficiency of coupling of ATP hydrolysis to Ca²⁺ transport.

Effect of SERCA2a expression on ER Ca²⁺ release and Ca²⁺ uptake

Coexpression of SERCA2a and RyR2 produced periodic Ca²⁺ waves due to spontaneous activation of RyR2 followed by SERCA Ca²⁺ reuptake (Fig. 2 A, cell #1). Ca²⁺ waves did not occur in cells with low expression of SERCA2a (cell #2) and untransfected cells (Endo), indicating that a certain threshold Ca²⁺ uptake rate is required to bring [Ca²⁺]_ER to a critical level that can trigger Ca²⁺ waves. As expected, the Ca²⁺ uptake load analysis revealed that ER Ca²⁺ uptake is faster in cells with higher SERCA2a expression level (Fig. 2 B; cell #1). The maximal ER Ca²⁺ load was also higher in those cells, suggesting that the catalytic efficacy of SERCA2a-mediated Ca²⁺ transport is also dependent on the pump’s expression level.

For each individual cell studied under similar conditions described in Fig. 2, the maximal ER Ca²⁺ uptake rate and the maximal ER Ca²⁺ load were plotted against SERCA2a expression level. The experimental points were then binned and averaged within five different groups to analyze the effect of the pump’s expression level on SERCA2a function. These analyses revealed that the maximal ER Ca²⁺ uptake rate increased as a function of SERCA2a expression, with a particularly steep nonlinear increase at higher pump expression levels (Fig. 2 C). Thus, increased SERCA2a expression increased both the pump’s dimerization (Fig. 1, A and B) and the maximal uptake rate (Fig. 2 C), consistent with the hypothesis that SERCA2a dimerization and conformational coupling enhance the pump Ca²⁺ transport rate. Remarkably, a higher level of SERCA2a expression also caused an increase in the maximal ER Ca²⁺ load (Fig. 2 D). Because these experiments quantify and correct for ER Ca²⁺ leak (21), one would expect the maximal ER Ca²⁺ load to be thermodynamically limited by the [ATP]/[ADP] ratio and the pump’s catalytic efficiency. Therefore, maximal load is expected to be insensitive to the SERCA2a expression level and should yield a flat ER Ca²⁺ load versus SERCA2a expression relationship. Instead, we observed a clear dependence of maximal load on SERCA2a expression (Fig. 2 D), suggesting the pump’s dimerization can also improve the pump’s catalytic efficacy.

Effect of SERCA2a inhibition on ER Ca²⁺ transport

In the following experiments, ER Ca²⁺ uptake was studied in the inducible SERCA2a stable line (T-Rex-293 cells) to obtain a homogenous pump expression and thus to control the degree of the pump dimerization between cells. Two days after SERCA2a gene induction with tetracycline, SERCA2a expression increased >10 times over endogenous SERCA2b (21). ER Ca²⁺ uptake was sensitive to the cytosolic [ATP]/[ADP] ratio and thapsigargin (TG) (Fig. 3 A). Decreasing [ATP] from 5 to 2.5 mM and increasing [ADP] to 2.5 mM (ADP) slowed ER Ca²⁺ uptake by 69% (n = 8; Fig. 3 B). As expected, the decrease in the [ATP]/[ADP] ratio also decreased the catalytic efficacy (or thermodynamic limit) of SERCA Ca²⁺ transport. The maximal ER Ca²⁺ load decreased by 58% (n = 8) in the presence of the low [ATP]/[ADP] ratio (Fig. 3 C). Surprisingly, a partial inhibition of SERCA2a activity with TG also decreased the catalytic efficacy of ER Ca²⁺ transport. On average, inhibition of ER Ca²⁺ uptake by 52% (n = 7; Fig. 3 B) with 50 nM TG decreased the maximal ER Ca²⁺ load by 48% (n = 7; Fig. 3 C). We found that a dose-dependent inhibition of SERCA2a Ca²⁺ uptake positively correlates with a decrease in ER Ca load. Because TG does not affect the SERCA2a dimerization (17), these results suggest that an irreversible inhibition of a fraction of Ca²⁺ pumps decreases ER Ca²⁺ uptake by reducing the number of functionally coupled SERCA2a dimers.

Figure 3.

Effects of [ATP]/[ADP] ratio and thapsigargin on ER Ca²⁺ uptake. (A) Representative traces of ER Ca²⁺ uptake mediated by endogenous and recombinant SERCA2a expressed in T-Rex-293 cells are shown. (B) The effect of the low [ATP]/[ADP] ratio (ADP) and thapsigargin (50 nM TG) on ER Ca²⁺ uptake is shown. (C) The maximal ER Ca²⁺ uptake rate and the maximal ER Ca²⁺ load in control conditions (Ctrl; n = 20 cells), in the presence of the low [ATP]/[ADP] ratio (n = 8 cells) or TG (25 nM, n = 5 cells; 50 nM, n = 7 cells; and 75 nM, n = 4 cells) is shown. ^∗p < 0.05 vs. control. Data are mean ± SE.

Effect of SERCA2a mutants with different catalytic activity on ER Ca²⁺ transport

The following experiments were aimed to define whether the functional coupling within a dimer is mediated by the pump’s conformational changes during its catalytic cycle. Here, we studied the effect of overexpression of two SERCA2a mutants with different catalytic activity on ER Ca²⁺ transport by SERCA2a^WT. The Y762G mutation (SERCA2a^Y762G) inhibits Ca²⁺ transport without a significant effect on the pump’s catalytic activity (i.e., causing uncoupling between ATP hydrolysis and Ca²⁺ uptake (30)). Another E309Q mutation (SERCA2a^E309Q) inhibits both the pump’s catalytic activity and Ca²⁺ transport (31). To keep the SERCA2a^WT expression at similar level in both experimental groups, the experiments were conducted in the inducible SERCA2a^WT stable line. The mutants were predominantly expressed in the ER (estimated from the overlap between the mCer-SERCA2a and the R-CEPIA1er signals; Fig. 4 A), with a distribution pattern similar to SERCA2a^WT (21). FRET analysis of HEK293 cells transfected with mCer-SERCA2a mutant (Y762G or E309Q) and YFP-SERCA2a^WT confirmed that the corresponding mutations do not affect SERCA2a dimerization (Fig. 4 B). Analysis of ER Ca²⁺ dynamics revealed that overexpression of the “catalytically dead” SERCA2a^E309Q significantly reduced the maximal ER Ca²⁺ uptake rate by 29% and the maximal ER Ca²⁺ load by 22% (n = 8) compared with SERCA2a^Y762G (n = 10; Fig. 4, B and C). The effect of SERCA2a^E309Q overexpression on ER Ca²⁺ uptake was similar to the TG action (Fig. 3). Interestingly, expression of SERCA2a^Y762G did not significantly change Ca²⁺ transport by SERCA2a^WT. The data suggest that it is the coupling of ATP hydrolysis activities of the two protomers that results in enhanced catalytic function. We conclude that the coupling of energy transduction in the SERCA dimer enhances the kinetic and catalytic efficacy of transport. By this mechanism, the transport activity of the SERCA2a dimer is greater than the sum of its parts.

Figure 4.

Effect of SERCA2a mutants (Y762G and E309Q) with different catalytic activity on ER Ca²⁺ uptake. (A) Images of T-Rex-293 cells expressing mCer-SERCA2a^Y762G or mCer-SERCA2a ^E309Q together with R-CEPIA1er are shown. Merged images show colocalization of the mCer and the R-CEPIA1er signals, indicating that SERCA2a mutants were expressed in the ER. (B) FRET from mCer-SERCA2a^{WT,Y762G,E309Q} to YFP-SERCA2a^WT increased with increasing protein expression, confirming SERCA dimerization. Data are from three independent experiments. (C) Representative traces of ER Ca²⁺ uptake mediated by SERCA2a^WT together with SERCA2a mutants (Y762G or E309Q) are shown. (D) The maximal ER Ca²⁺ uptake rate and the maximal ER Ca²⁺ load mediated by SERCA2a^WT cotransfected with Y762G (n = 10 cells) or E309Q (n = 8 cells) are shown. ^∗p < 0.05 vs. control. Data are mean ± SE.

Discussion

Oligomerization is thought to be an important functional mechanism for many P-type ion pumps, including the Na⁺/K⁺-ATPase (32, 33, 34, 35), the plasma membrane Ca²⁺-ATPase (36,37), the H⁺/K⁺-ATPase (38), the Cu²⁺-ATPase (39), and the subject of this study, the SERCA pump (13, 14, 15, 16). For some ATPases, oligomerization has been shown to increase transport activity (40,41). In the case of the Na⁺/K⁺-ATPase, oligomerization may facilitate enzymatic coupling and enhanced activity. In this model, catalytic cycling of the second protomer of an Na⁺/K⁺-ATPase dimer is engaged when the concentration of ATP is high enough to occupy nucleotide binding sites on both protomers, increasing overall transport rate (40). There has not been a consensus in the literature about the functional significance of the SERCA-SERCA interaction. SERCA oligomerization has been proposed as a regulatory mechanism, and different groups have correlated dimerization to pump inhibition (16,19,42) or, conversely, relief of inhibitory regulation (18). Alternatively, SERCA dimerization has been proposed to be an intrinsic kinetic mechanism that enhances SERCA turnover. Support for this concept comes from studies that showed partial inhibition of SERCA results in a larger-than-expected decrease in transport function (43,44). Moreover, kinetic studies suggested that the SERCA protomers in the dimer operate asynchronously, with energetically favorable conformational changes of one SERCA molecule assisting unfavorable conformational changes of the second SERCA (20). One may consider the analogy of pedaling a bicycle using only the right leg; when the leg is fully extended, the crank gets stuck at the bottom of the circle of rotation, impairing turnover. Adding the action of the left leg couples two rotational cycles together. Thus, one leg is always in a favorable position, allowing faster and more efficient cycling.

Our data presented here are compatible with this model of functional dimerization of the SERCA2a pump. Biochemical and spectroscopic measurements are consistent with SERCA2a-SERCA2a binding (Fig. 1, A and B) and with PLB interaction with a SERCA2a dimer (Fig. 1, C and D). In support of the mechanism of SERCA conformational coupling, we observed increased ER Ca²⁺ uptake at protein concentrations that are sufficiently high to support SERCA-SERCA dimerization. Specifically, in the absence of a coupling effect, one would predict a linear increase in ER Ca²⁺ uptake with increasing expression of SERCA2a. Instead, we found the maximal Ca²⁺ uptake rate increased supralinearly with an increasing SERCA2a expression level (Fig. 2 C). Interestingly, we also found that maximal ER Ca²⁺ load unexpectedly increased with an increased SERCA2a expression level (Fig. 2 D), suggesting that dimerization improves the thermodynamics of Ca²⁺ transport. The data indicate that dimerization increases SERCA2a catalytic efficiency through decreased Ca²⁺ transport cost—that is, more efficient coupling between ATP hydrolysis and ion translocation (Fig. 5). Consistent with this hypothesis was the observation that partial inhibition of the pump with a low dose of TG, inhibiting ∼50% of the population of pumps, also decreased maximal ER Ca²⁺ load. This was unexpected; a reduced population of pumps should be able to achieve the same thermodynamically limited ER Ca²⁺ load (Fig. 3 A; red dashed line). Instead, the data suggest a threshold density of active transporters is needed for maximal catalytic efficiency. We have previously shown that TG binding to SERCA2a does not affect SERCA-SERCA physical dimerization (17); thus, the reduced pump catalytic efficiency must be due to coupling of active protomers with TG-inhibited “catalytically dead” partners (Fig. 5). Similarly, coupling of functional WT pump with the transport-inactive, ATP hydrolysis-inactive mutant E309Q decreased the maximal ER Ca²⁺ load. On the other hand, an overexpression of the transport-dead but ATP hydrolysis-active Y762G mutant did not significantly affect the efficacy of Ca²⁺ transport by SERCA^WT (Fig. 5). Because this mutant can still perform ATP hydrolysis (30), it can still conformationally couple with a SERCA2a^WT partner in mixed dimers, improving the transport kinetics of that active WT protomer. All together, these data suggest that it is the ATPase activity that is the critical function, transmitting energy by conformational coupling from one protomer to the other, conferring thermodynamic and kinetic advantages to the dimer complex. Other precedents for protein-protein interactions altering SERCA Ca²⁺ transport cost have been described. PLB phosphorylation has been shown to increase SERCA activity, possibly because of decreased “slip-mode” cycling of the transporter (45). Conversely, sarcolipin decreases the energetic efficiency of SERCA, resulting in thermogenesis (46). Our study suggests another SERCA2a regulatory interaction, dimerization, can improve the pump’s energetics.

Figure 5.

A model of SERCA2a regulation by dimerization. Monomeric SERCA2a is capable of ATP hydrolysis and Ca²⁺ transport by itself, but when conformationally coupled with another protomer in a dimeric complex, the kinetics of transport are increased, and transport cost is reduced. Conformational coupling can still occur when WT-SERCA is bound to SERCA2a^Y762G, which cannot transport Ca²⁺ itself but still hydrolyzes ATP. Conformational coupling cannot occur when SERCA2a^WT is bound to TG-inactivated SERCA or to SERCA2a^E309Q, which cannot hydrolyze ATP.

Collectively, the results of this study indicate that SERCA2a dimerization enhances ER Ca²⁺ uptake by improving kinetics (maximal Ca²⁺ uptake rate) and catalytic efficacy (maximal Ca²⁺ load) of the pump. The observations may provide insight into the mechanisms of abnormal Ca²⁺ regulation in HF, which is characterized by decreased SERCA2a expression and poor Ca²⁺ handling (1, 2, 3,47,48). These results suggest that decreased SERCA2a expression may result in decreased dimerization and conformational coupling, exacerbating the Ca²⁺ uptake deficit in diseased cardiac myocytes. In addition, we showed that pharmacological inhibition of a fraction of the population of pumps impairs conformational coupling and decreases catalytic efficiency (Fig. 3). Pathological correlates of this condition include oxidative modification of SERCA (49), which is associated with aging; SERCA carbonylation (50,51), which is increased in diabetes; and proteolytic cleavage of SERCA, which increases after ischemia/reperfusion (52). Such modifications impair SERCA function, and the data presented here suggest the damaged SERCA pumps may negatively impact unmodified SERCA protomers through conformational coupling in a dimer. Moreover, decreased catalytic efficiency could worsen the already poor energy balance of the failing heart (28,53). Conversely, improving SERCA2a conformational coupling may be a worthwhile goal in the design of future therapies.

Author Contributions

E.B., S.R.C., J.S., S.L.R., and A.V.Z. contributed to the conception and design of the study. E.B., R.N., S.R.C., J.S., and D.K. performed the experimental work and analysis of results. E.B., S.R.C., J.S., S.L.R., and A.V.Z. contributed to writing of the manuscript. All authors have approved the version to be published.

Editor: Heping Cheng.