Conformational memory in the association of the transmembrane protein phospholamban with the sarcoplasmic reticulum calcium pump SERCA

The sarcoplasmic reticulum Ca2+-ATPase SERCA promotes muscle relaxation by pumping calcium ions from the cytoplasm into the sarcoplasmic reticulum. SERCA activity is regulated by a variety of small transmembrane peptides, most notably by phospholamban in cardiac muscle and sarcolipin in skeletal muscle. However, how phospholamban and sarcolipin regulate SERCA is not fully understood. In the present study, we evaluated the effects of phospholamban and sarcolipin on calcium translocation and ATP hydrolysis by SERCA under conditions that mimic environments in sarcoplasmic reticulum membranes. For pre-steady-state current measurements, proteoliposomes containing SERCA and phospholamban or sarcolipin were adsorbed to a solid-supported membrane and activated by substrate concentration jumps. We observed that phospholamban altered ATP-dependent calcium translocation by SERCA within the first transport cycle, whereas sarcolipin did not. Using pre-steady-state charge (calcium) translocation and steady-state ATPase activity under substrate conditions (various calcium and/or ATP concentrations) promoting particular conformational states of SERCA, we found that the effect of phospholamban on SERCA depends on substrate preincubation conditions. Our results also indicated that phospholamban can establish an inhibitory interaction with multiple SERCA conformational states with distinct effects on SERCA's kinetic properties. Moreover, we noted multiple modes of interaction between SERCA and phospholamban and observed that once a particular mode of association is engaged it persists throughout the SERCA transport cycle and multiple turnover events. These observations are consistent with conformational memory in the interaction between SERCA and phospholamban, thus providing insights into the physiological role of phospholamban and its regulatory effect on SERCA transport activity.

The sarcoplasmic reticulum (SR) 4 Ca 2ϩ -ATPase (SERCA) plays a central role in muscle relaxation by removing calcium from the cytosol and transporting it into the lumen of the SR. The mechanism of active calcium transport by SERCA has been well described (1)(2)(3). According to the original E1-E2 scheme (4), SERCA activation requires the binding of two calcium ions (E1-Ca 2 ) followed by formation of a phosphoenzyme intermediate from ATP (E1P). The free energy from ATP, captured in phosphoenzyme formation, is used for a conformational transition (E1P to E2P) that favors the release of calcium into the SR in exchange for luminal protons. Hydrolytic cleavage of the phosphoenzyme intermediate (dephosphorylation) is the final reaction step, which returns the enzyme to the ground state and allows the initiation of a new transport cycle. Crystal structures of most of the intermediates in the reaction cycle have been determined, providing a thorough description of SERCA at atomic detail (for reviews, see Refs. [1][2][3]. SERCA transport activity in muscle cells is regulated by the homologous membrane proteins phospholamban (PLN; 52 amino acids) and sarcolipin (SLN; 31 amino acids). PLN is primarily expressed in cardiac muscle where it regulates SERCA2a, whereas SLN expression predominates in skeletal muscle where it regulates SERCA1a. PLN and SLN interact with SERCA in a groove formed by transmembrane helices M2, M6, and M9 as characterized by cross-linking studies (5), molecular modeling (6), NMR (7,8), and X-ray crystallography (9 -11). PLN lowers the apparent calcium affinity of SERCA, and this inhibitory effect is reversed upon PLN phosphorylation by protein kinase A (12), calcium-calmodulin-dependent protein kinase II (13), or Akt (14). It has been hypothesized that SLN alters the apparent calcium affinity of SERCA in a manner similar to PLN. However, the molecular mechanism (15,16) and physiological consequences (17) of SLN inhibition are distinct from PLN.
In addition to cytosolic calcium levels and phosphorylation status, the oligomeric state of PLN plays a critical role in SERCA regulation. PLN exists as an equilibrium mixture of monomers and pentamers, both of which are required for normal cardiac function (18). The main inhibitor is the PLN monomer, which physically interacts with SERCA and alters its apparent affinity for calcium. The PLN pentamer has been described as an inactive storage form; however, experimental evidence suggests that the PLN pentamer also interacts with SERCA (19), and it forms a cation-selective pore (20 -22). Although the consensus model for PLN regulation of SERCA involves the reversible association of a monomer under calcium-free conditions, there is a growing body of evidence that PLN remains associated with SERCA and adopts multiple conformational states in the SERCA-PLN complex (19,(23)(24)(25).
To address this, we evaluated the effects of PLN and SLN on ATP-dependent calcium translocation by SERCA using presteady-state current measurements on a solid-supported membrane (SSM; Fig. 1). The SSM method has been used to investigate the ion translocation mechanism of P-type ATPases (26,27). We performed current measurements on co-reconstituted proteoliposomes containing SERCA and PLN or SLN. Such proteoliposomes have been used extensively to characterize PLN structure and function (7, 19, 28 -33) and more recently to investigate SERCA inhibition by SLN (15). The proteoliposomes were adsorbed to the SSM and activated by a calcium and/or ATP concentration jump. Following SERCA activation, an electrical current was detected related to the displacement of calcium ions within the protein during the first catalytic cycle (27,34,35). These results from pre-steady-state charge translocation during the first SERCA turnover were compared with steady-state measurements of ATPase activity during continuous turnover (32,33). The choice of substrate conditions promoted particular conformational states of SERCA from which catalytic turnover could be initiated. We found that the effect of PLN on SERCA activity depended on the starting conditions, and our observations were consistent when comparing presteady-state charge translocation with steady-state ATPase activity. In contradiction to previous hypotheses, we conclude that PLN can interact with at least two conformational states of SERCA with distinct effects on SERCA's maximal activity, apparent calcium affinity, and cooperative calcium binding. Moreover, once a particular SERCA-PLN inhibitory interaction was established, it persisted during continuous catalytic turnover of SERCA. This "conformational memory" (36) in the SERCA-PLN regulatory complex provides insights into the functional relevance of the published crystal structures of SERCA bound to PLN (11) and SLN (9,10).

Results
In the present study, we focused on SERCA1a in the presence of PLN and SLN. The skeletal muscle SERCA1a isoform is a validated surrogate for the cardiac SERCA2a isoform (e.g. see Figs. 3 and 9 in Ref. 37). Perhaps most importantly, SERCA1a was shown to functionally substitute for SERCA2a in the heart, including regulation by endogenous PLN (38,39). Herein, we use the term "SERCA" to designate the SERCA1a isoform.

Pre-steady-state charge movement
ATP jumps-SERCA proteoliposomes were characterized by pre-steady-state current measurements on an SSM following ATP jumps ( Fig. 2; preincubation with calcium and reaction initiation with ATP; condition ii in Table 1). The charge obtained by integration of the ATP-induced current transient is A schematic diagram of a proteoliposome containing SERCA and PLN adsorbed to a SSM and subjected to an ATP concentration jump is shown. If the ATP jump induces net charge displacement across the protein, a current transient is measured (the red circles are electrons) when the potential difference applied across the whole system is kept constant.

Substrate dependence of SERCA-phospholamban interactions
attributable to the translocation of calcium into the proteoliposomes during the first SERCA transport cycle (27,34,35). To confirm that the measured current is due to SERCA, we performed an ATP jump in the presence of thapsigargin (40), which caused nearly complete suppression of the current transient (not shown). ATP jumps were then carried out on SERCA-PLN proteoliposomes. The ATP-induced current signals at 3 and 0.3 M free calcium were reduced compared with proteoliposomes containing SERCA alone (Fig. 3A). As expected, PLN affected ATP-dependent calcium translocation by SERCA. ATP jumps were performed over a range of calcium concentrations, and the resulting displaced charges were analyzed using the Hill equation (Fig. 4 and Table 1; calcium concentration at halfmaximal activity (K Ca ) of 0.70 M). It is interesting to compare this value with that determined for proteoliposomes containing SERCA alone (K Ca of 0.38 M). Overall, these pre-steady-state values are in good agreement with previous measurements of SERCA ATPase activity in the absence and presence of PLN (32,33).
We next performed current measurements on SERCA-SLN proteoliposomes. In this case, the displaced charges measured after ATP jumps at 3 and 0.3 M free calcium were indistinguishable from those obtained for proteoliposomes containing SERCA alone (Figs. 3A and 4; K Ca of 0.43 M and Hill coefficient (n H ) of 1.6 in the presence of SLN). There have been conflicting reports on the regulatory effect of SLN on SERCA. One study reported that SLN decreased the apparent calcium affinity of SERCA and stimulated the maximal calcium uptake rate (41). Another recent study reported that SLN decreases the maximal activity of SERCA, but it does not affect the apparent calcium affinity (16). Based on our recent studies using the SERCA-SLN proteoliposomes, we believe that SLN decreases both the apparent calcium affinity and maximal activity of SERCA (15). Nonetheless, these perplexing observations may reflect the nature of the experimental systems (cells versus purified components and calcium transport versus ATP hydrolysis).
Finally, we found that a high magnesium concentration (5 mM) eliminated the effect of PLN on the ATP-induced, SERCAdependent charge movement in proteoliposomes regardless of the calcium concentration (Fig. 3B). Unexpectedly, 5 mM magnesium decreased charge translocation by SERCA proteoliposomes, which became indistinguishable from SERCA-PLN proteoliposomes in the presence of either 5 or 1 mM magnesium. In other words, 5 mM magnesium promotes a SERCA response that is comparable with PLN inhibition, but this response cannot be further modulated by PLN (Fig. 3, A and B, compare bars for SERCA and SERCA ϩ PLN). The decrease in charge translocation may reflect competitive binding of calcium and magnesium to SERCA where the magnesium affinity is low (ϳmM) but significant at low calcium concentrations (2). As seen in recent crystal structures (9,10), magnesium may stabilize the calcium-binding sites of an E1-like SERCA conformation, which is not catalytically active for charge transport until the magnesium has been displaced.
Simultaneous calcium and ATP jumps-To gain further insights into the effect of PLN on the SERCA transport cycle, we performed simultaneous ATP (100 M) and calcium (varying Table 1 Kinetic parameters for SERCA and SERCA-PLN proteoliposomes derived from pre-steady-state charge movement and steady-state ATPase activity measurements Statistical significance was evaluated based on a threshold p value (p Ͻ 0.05). See Table S1 for a full comparison of all kinetic parameters and the associated p values.
a Statistically significant compared with SERCA in the absence of PLN under the same conditions (see Table S1 for p values). b Statistically significant compared with SERCA in the presence of PLN for condition i (see Table S1 for p values).

Substrate dependence of SERCA-phospholamban interactions
concentrations) jumps with proteoliposomes containing SERCA and PLN (starting condition i in Table 1). As shown in Fig. 5 and Table 1, we observed a decrease in the calcium affinity of SERCA (K Ca of 0.79 M) and comparable cooperativity for calcium binding (n H of 1.8) with respect to the values obtained following simultaneous ATP and calcium jumps for proteoliposomes containing SERCA alone (K Ca of 0.49 M and n H of 1.6). The simultaneous addition of calcium and ATP mimics conditions used for steady-state measurements of SERCA ATPase activity. Thus, the kinetic parameters agree well with previous observations (32,33). Simultaneous ATP and calcium jumps with proteoliposomes containing SERCA and SLN were again indistinguishable from those obtained for proteoliposomes containing SERCA alone (not shown). We conclude that, under the conditions used to measure charge translocation during the first SERCA transport cycle, SLN does not alter SERCA function.
Calcium jumps-The differences between ATP jumps and simultaneous calcium-ATP jumps prompted us to investigate calcium as the initiating event for SERCA-dependent charge displacement (starting condition iii in Table 1). Calcium jumps were carried out on proteoliposomes containing SERCA and PLN preincubated with ATP (100 M). These conditions resulted in a much lower apparent calcium affinity of SERCA (K Ca of 1.15 M) and reduced cooperativity (n H of 1.1) for calcium binding ( Fig. 5 and Table 1) as compared with the results obtained following calcium jumps for proteoliposomes containing SERCA alone preincubated with ATP (K Ca of 0.40 M and n H of 1.4). Thus, preincubation of SERCA-PLN with saturating ATP followed by calcium jumps led to a reduction in the apparent calcium affinity and cooperativity for calcium binding by SERCA. The results support the view that calcium and ATP promote slightly different SERCA conformational states from which the transport cycle can be initiated. Although this latter point may seem obvious, PLN inhibition still occurred under the different initiating conditions, although the kinetic effects and nature of PLN inhibition depend on how SERCA was poised before initiation of the transport cycle.

Steady-state ATPase activity
Measurement of ATP hydrolysis rates of the SERCA proteoliposomes in the absence and presence of PLN (31-33) provided an important benchmark for the pre-steady-state measurements. We determined the ATPase activity of SERCA alone and in the presence of PLN with the preincubation and starting conditions described above. Proteoliposomes containing SERCA alone or SERCA and PLN (Fig. 6) were subjected to the following conditions (Table 1): condition i, reaction initiation by the simultaneous addition of calcium and ATP, condi-   Table 1). Error bars represent S.E. of at least four independent measurements.

Substrate dependence of SERCA-phospholamban interactions
tion ii, reaction initiation by the addition of ATP, and condition iii, reaction initiation by the addition of calcium (0.1-10 M). The outcome of these experiments prompted us to include condition iv with more physiological substrate concentrations. It should be noted that the measurement of steady-state ATPase activity was performed with 4 mM ATP and 5 mM magnesium. These conditions differ from those used in pre-steady-state measurements.
For SERCA proteoliposomes, the traditional method of initiating the reaction, simultaneous calcium and ATP addition, provided baseline values for the apparent calcium affinity, maximal activity, and cooperativity of SERCA. The values obtained (Table 1) were similar to those reported previously (31)(32)(33). Varying the starting conditions impacted the ATPase activity (Fig. 6) and the kinetic parameters determined from fitting of the Hill equation. The apparent calcium affinity varied slightly (K Ca ranged from 0.26 to 0.39 M; p Ͻ 0.01) with the highest affinity occurring when SERCA was preincubated with calcium (ATP jump). The cooperativity for calcium binding also varied (average n H of 1.6) with reduced cooperativity occurring when SERCA was preincubated with calcium (ATP jump; n H of 1.2). Finally, the maximal activity of SERCA varied the most depending on starting conditions, and this paralleled the effect on the apparent calcium affinity of SERCA. The simultaneous calcium-ATP jump had the lowest affinity and highest maximal activity, whereas the ATP jump had the highest affinity and lowest maximal activity. In fact, there was a linear correlation between K Ca and maximal activity (V max ) for the different preincubation and starting conditions (not shown).
Our usual conditions are to initiate ATPase activity by SERCA-PLN proteoliposomes with the simultaneous addition of calcium and ATP (31)(32)(33) such that we observe effects of PLN on the apparent calcium affinity (ϳ2-fold decrease), maximal activity (ϳ1.4-fold increase), and cooperativity (ϳ1.3-fold increase) of SERCA (condition i in Table 1). Varying these starting conditions had a minor impact on the ability of PLN to alter the apparent calcium affinity of SERCA (K Ca varied from 0.82 to  Table 1). Error bars represent S.E. of at least four independent measurements.  Table 1). Each data point is the mean Ϯ S.E. (n Ն 4). Error bars represent S.E. 0.89 M; p Ͻ 0.01). Although the changes were small, there was a similar trend toward higher affinity when SERCA was preincubated with calcium (ATP jump; K Ca of 0.82 M). As was seen for SERCA proteoliposomes, the maximal activity varied significantly depending on starting conditions. Again, the simultaneous calcium-ATP starting condition had the lowest affinity and highest maximal activity, the ATP starting condition had the highest affinity and lowest maximal activity, and a linear correlation was observed (not shown). Finally, the largest change occurred in the cooperativity for calcium binding (n H varied from 1.3 to 2.0; p Ͻ 0.04). The loss of cooperativity occurred when SERCA was preincubated with either calcium or ATP, and this same trend was observed for pre-steady-state charge displacement (single SERCA turnover) and steady-state ATPase activity (continuous SERCA turnover). In other words, PLN inhibition increased the cooperativity for calcium binding only when SERCA was preincubated in the absence of both substrates.

Mimicking physiological conditions
The dependence of SERCA inhibition by PLN on starting conditions prompted us to explore more physiological substrate concentrations (starting condition iv in Table 1). At rest, the substrate concentrations that SERCA experiences are low cytosolic calcium (Ͻ0.1 M) and high ATP (1-10 mM). We preincubated SERCA and SERCA-PLN proteoliposomes with 60 nM calcium and 4 mM ATP followed by calcium jump experiments in which the addition of calcium was used to achieve various concentrations (0.1-10 M; Fig.  7). Overall, the SERCA proteoliposomes behaved in a manner similar to the other starting conditions with comparable kinetic parameters (Table 1). However, the SERCA-PLN proteoliposomes exhibited a striking increase in cooperativity for calcium binding to SERCA. Indeed, this is the largest change in cooperativity (n H value of 3.0) observed with our SERCA-PLN proteoliposome preparation (31-33). Thus, PLN inhibition increased the cooperativity for calcium binding when SERCA was preincubated in the presence or absence of both substrates but not in the presence of individual substrates (calcium or ATP).

Pre-steady-state charge movement versus steady-state ATPase activity
The SERCA-dependent current transients described above reflect pre-steady-state calcium translocation across the proteoliposome membrane during the first SERCA transport cycle. The conditions tested poise SERCA in distinct conformations prior to the initiation of the SERCA transport cycle. These same conditions were tested in steady-state measurements of SERCA ATPase activity where calcium-dependent ATP hydrolysis by SERCA remained constant over time. Overall, the kinetic parameters for pre-steady-state charge movement and steadystate ATPase activity agreed well (Table 1). However, there was one notable difference: the K Ca values for SERCA in the presence of PLN under calcium jump conditions. This difference may reflect the nature of the experimental measurements and incomplete coupling between calcium transport and ATP hy-   Table 1). Each data point is the mean Ϯ S.E. (n Ն 4). Error bars represent S.E.

Conformational memory in the SERCA-PLN regulatory complex
The choice of substrate conditions promotes particular conformational states of SERCA from which catalytic turnover and calcium transport can be initiated (Fig. 8). Preincubation of SERCA in the absence of substrates promotes the E2 state (42), preincubation of SERCA with calcium promotes the E1 calcium-bound state (43), and preincubation of SERCA with ATP promotes an E2 ATP-bound state (44). The SERCA conformational state that is traditionally associated with PLN inhibition is the calcium-free E2 state (6,7). However, crystal structures of SERCA-SLN complex (9, 10) revealed a new calcium-free, magnesium-bound E1-like state. The structure of the SERCA-PLN complex is remarkably similar to this E1-like state (11), although magnesium is absent. Nonetheless, molecular models of the SERCA-PLN complex based on the calcium-free E2 state are instructive for considering the mechanism of PLN inhibition. In the absence of calcium, PLN binds in a groove formed by transmembrane helices M2, M6, and M9 of SERCA. Because this groove closes upon calcium binding, PLN alters the apparent calcium affinity of SERCA by impeding groove closure and the E2-E1 transition. This model is consistent with kinetic arguments that PLN alters the activation energy for a slow conformational change triggered by calcium binding to SERCA (45), which impacts both the apparent calcium affinity and cooperativity for calcium binding.
SERCA inhibition by PLN occurred under the different starting conditions (Figs. 4 -7), although the kinetic effects and nature of PLN inhibition depended on how SERCA was poised before the initiation of transport. The most significant change occurred in the cooperativity for calcium binding, reflected by the large change in Hill coefficients across the different preincubation conditions (Table 1). In the absence of substrates or under physiological substrate conditions (low calcium and high ATP), PLN strongly increased the cooperativity for calcium binding. The theoretical limit for the Hill coefficient is 2 (two Ca 2ϩ ions transported per SERCA molecule); however, we observed a value of 3 (n H of 3.0). The simplest explanation for this is substrate-and PLN-dependent coupling between SERCA molecules (46). Nonetheless, these conditions promote an "on-pathway" reaction trajectory from the E2 calcium-free state to the E1-ATP-Ca 2ϩ state of SERCA such that PLN induces a strongly cooperative transition between the calciumfree and calcium-bound states (Fig. 8). The substrate-free condition promotes the E2 state of SERCA, which requires a large multidomain conformational change to form the catalytically active E1-ATP-Ca 2ϩ state (42). Thus, PLN stabilization of the E2 state of SERCA offers a plausible explanation for the increase in cooperativity: PLN lowers the probability of binding the first calcium ion, but once binding occurs there is a higher probability for binding the second calcium ion. The increased cooperativity was also observed under more physiological substrate conditions (low calcium and high ATP), suggesting a common structural basis. However, the conformational state of SERCA under these latter conditions is unknown.
In an opposite trend, there was reduced cooperativity for calcium binding when SERCA was preincubated with either calcium or ATP. Again, this same observation occurred in presteady-state charge displacement (condition i versus condition iii; n H decreased from 1.8 to 1.1; p Ͻ 0.025) and steady-state ATPase activity (condition i versus condition iii; n H decreased from 2.0 to 1.3; p Ͻ 0.04). The following question then arises: what causes the reduced cooperativity when SERCA is preincubated with either ATP or calcium in the presence of PLN? Because only a single substrate is present, these conditions are described as an "off-pathway" reaction trajectory from the E2 calcium-free state to the catalytically active E1-ATP-Ca 2ϩ state of SERCA. Calcium promotes the E1 calcium-bound state in a concentration-dependent manner. Although this offers an explanation for the reduced cooperativity, PLN can still alter the apparent calcium affinity of SERCA. In contrast, ATP alone promotes an E2 ATP-bound state; however, the reduced cooperativity suggests that SERCA adopts a conformation more compatible with calcium binding. These characteristics are reminiscent of the novel E1-like state identified in crystal structures of the SERCA-SLN (9, 10) and SERCA-PLN (11) complexes. In this state, the calcium-binding sites are formed and exposed to the cytoplasm. PLN stabilization of this E1-like state offers a plausible explanation for the reduced cooperativity for calcium binding to SERCA: the conformational transition that establishes cooperativity and forms the calcium-binding sites has already occurred. In the crystal structures of the SERCA-SLN complex (9,10), magnesium ions stabilize the calcium sites in the E1-like state. This offers an explanation for PLN inhibition and the decrease in the apparent calcium affinity of SERCA: magnesium ions must be displaced before calcium can bind. This latter point is consistent with the effects of the higher magnesium concentration (5 mM) on pre-steady-state charge displacement by SERCA-PLN proteoliposomes (Fig. 3).
Overall, our observations support the concept that PLN can establish an inhibitory interaction with multiple SERCA conformational states, minimally a calcium-free E2 state and a more E1-like state. We hypothesize that the calcium-free inhibitory state is the canonical complex as depicted by molecular modeling (6,7) and that the E1-like inhibitory state is represented by the more recent crystals structures (9 -11). Furthermore, once these modes of PLN interaction are established, they persist through multiple cycles of SERCA transport. If SERCA transport activity is initiated in the absence of substrates, PLN establishes and retains the ability to interact with the E2 state of SERCA through multiple turnover events (effect on both K Ca and n H ). Instead, if SERCA is preincubated with ATP or calcium, PLN establishes and retains the ability to interact with an E1-like state of SERCA through multiple turnover events (effect only on K Ca ). More importantly, PLN does not appear to access the E2 state of SERCA under these conditions. Recall that resting conditions in cardiac muscle are typically associated with low calcium (Ͻ0.1 M) and high ATP (1-10 mM) and magnesium (ϳ2 mM) concentrations.
SERCA is a molecular machine that carries out the complex functions of ATP hydrolysis and active calcium transport. To achieve this, SERCA undergoes plastic deformations and large domain movements enabled by flexibility and spontaneous Substrate dependence of SERCA-phospholamban interactions conformational fluctuations (47). However, the transition between different conformational states is not random but instead is driven by the availability of substrates and the suppression of backward transitions. PLN regulates SERCA by altering the transition between the E2 and E1 conformational states, which manifests as a decrease in SERCA's calcium affinity. Herein, our data are consistent with conformational memory (36) in the SERCA-PLN regulatory complex. The SERCA-PLN complex attains distinct structural states depending on substrate conditions (Fig. 8), and these states are retained during SERCA catalytic turnover and conformational cycling. In other words, there are multiple modes of interaction between SERCA and PLN, and once a particular mode of interaction is engaged it remains throughout the SERCA calcium transport cycle and multiple turnover events.

Co-reconstitution of SERCA and recombinant PLN or SLN
Human PLN and SLN were expressed and purified as described (48). Peptides were confirmed by DNA sequencing (TAGC Sequencing, University of Alberta) and MALDI-TOF mass spectrometry (Alberta Proteomics and Mass Spectrometry Facility, University of Alberta). SERCA1a was purified from rabbit skeletal muscle SR (49,50) and co-reconstituted with PLN and SLN as described (15,31). The purified proteoliposomes yielded a final molar ratio of 1 SERCA:4.5 PLN (or SLN): ϳ120 lipids. For the calculation of specific activity, the SERCA, PLN, and SLN concentrations were determined by quantitative SDS-PAGE (29).

Electrical measurements
All current measurements were performed with the SURFE 2 R One device (Nanion Technologies, Munich, Germany). The temperature was maintained at ϳ23°C. Charge movement was measured by adsorbing proteoliposomes containing (i) SERCA, (ii) SERCA and PLN, or (iii) SERCA and SLN onto a hybrid alkane thiol/phospholipid bilayer anchored to a gold electrode (SSM; Fig. 1) (26,51). Once the proteoliposomes were adsorbed, SERCA was activated by a concentration jump of a suitable substrate (e.g. ATP). If the substrate concentration jump induces charge displacement across the membrane, a current transient can be recorded due to capacitive coupling between the proteoliposome and SSM (27,34). The numerically integrated current transient is related to the net charge movement in the proteoliposomes, which depends upon the electrogenic calcium translocation by SERCA. Our premise is that the SSM technique detects pre-steady-state current transients within the first catalytic cycle of SERCA and that it is not sensitive to stationary currents following the first cycle (27,34,52).
For the ATP concentration jump experiments, "non-activating" and "activating" buffered solutions were used. The nonactivating solution contained 100 mM KCl, 0.25 mM EGTA, 0.2 mM DTT, 1 mM MgCl 2 , 25 mM MOPS (pH 7.0), and various concentrations of CaCl 2 (0.1-10 M free Ca 2ϩ ; as calculated by the WinMAXC program (53); e.g. control measurements utilized 0.25 mM CaCl 2 to achieve 10 M free Ca 2ϩ ). The activating solution was identical except for addition of 100 M ATP. To prevent calcium accumulation into proteoliposomes, 1 M cal-cium ionophore (A23187; calcimycin) was used. In inhibition experiments, SERCA was preincubated with 100 nM thapsigargin for 10 min. The ATP-induced current signal observed in the absence and presence of thapsigargin was then compared.
To verify the reproducibility of the current transients on the same SSM, each single measurement was repeated five times and then averaged to improve the signal-to-noise ratio. Standard deviations did not exceed 5%. At the beginning and end of a concentration jump experiment, initial and final control measurements (i.e. 100 M ATP jumps in the presence of 10 M free Ca 2ϩ ) were carried out. This allowed comparison of maximum current transients at the start and end of each experiment and the detection of signal deterioration during the time course of each experiment (e.g. due to protein degradation or loss of proteoliposome adsorption). Each set of measurements was reproduced using four to six different gold sensors.

ATPase activity assays
Calcium-dependent ATPase activities of the co-reconstituted proteoliposomes were measured by a coupled enzyme assay (54). The coupled enzyme assay reagents were of the highest purity available (Sigma-Aldrich). The proteoliposomes contained SERCA alone, SERCA in the presence of PLN, or SERCA in the presence of SLN. A minimum of three independent reconstitutions and activity assays were performed for each set of proteoliposomes, and the calcium-dependent ATPase activity was measured over a wide range of calcium concentrations (0.1-10 M). The assay was adapted to a 96-well format utilizing a BioTek Synergy 4 hybrid microplate reader (340-nm wavelength; 150-l volume; 10 -20 nM SERCA/well; 30°C; data points collected every 39 s for 1 h). The reactions were initiated by (i) the simultaneous addition of calcium and ATP (preincubation of the proteoliposomes without calcium or ATP), (ii) the addition of ATP (preincubation of the proteoliposomes with calcium), or (iii) the addition of calcium (preincubation of the proteoliposomes with ATP). The K Ca , V max , and n H were calculated based on non-linear least-squares fitting of the activity data to the Hill equation using Sigma Plot software (SPSS Inc., Chicago, IL). Errors were calculated as the standard error of the mean for a minimum of three independent measurements. Comparison of K Ca , V max , and n H parameters was carried out using analysis of variance (between-subjects, one-way) followed by the Holm-Sidak test for pairwise comparisons (Sigma Plot).