The N Terminus of Sarcolipin Plays an Important Role in Uncoupling Sarco-endoplasmic Reticulum Ca2+-ATPase (SERCA) ATP Hydrolysis from Ca2+ Transport*

Background: Both phospholamban (PLB) and sarcolipin (SLN) regulate SERCA activity, however, only SLN uncouples SERCA. Results: The N and C termini of SLN, or the N terminus and transmembrane region of PLB, confer protein-specific function. Conclusion: SLN N terminus plays a role in dynamic interaction and uncoupling of SERCA. Significance: SERCA uncoupling by SLN increases heat production implicating SLN-SERCA interaction in muscle thermogenesis. The sarcoendoplasmic reticulum Ca2+-ATPase (SERCA) is responsible for intracellular Ca2+ homeostasis. SERCA activity in muscle can be regulated by phospholamban (PLB), an affinity modulator, and sarcolipin (SLN), an uncoupler. Although PLB gets dislodged from Ca2+-bound SERCA, SLN continues to bind SERCA throughout its kinetic cycle and promotes uncoupling of Ca2+ transport from ATP hydrolysis. To determine the structural regions of SLN that mediate uncoupling of SERCA, we employed mutagenesis and generated chimeras of PLB and SLN. In this study we demonstrate that deletion of SLN N-terminal residues 2ERSTQ leads to loss of the uncoupling function even though the truncated peptide can target and constitutively bind SERCA. Furthermore, molecular dynamics simulations of SLN and SERCA interaction showed a rearrangement of SERCA residues that is altered when the SLN N terminus is deleted. Interestingly, transfer of the PLB cytosolic domain to the SLN transmembrane (TM) and luminal tail causes the chimeric protein to lose SLN-like function. Further introduction of the PLB TM region into this chimera resulted in conversion to full PLB-like function. We also found that swapping PLB N and C termini with those from SLN caused the resulting chimera to acquire SLN-like function. Swapping the C terminus alone was not sufficient for this conversion. These results suggest that domains can be switched between SLN and PLB without losing the ability to regulate SERCA activity; however, the resulting chimeras acquire functions different from the parent molecules. Importantly, our studies highlight that the N termini of SLN and PLB influence their respective unique functions.

The sarcoendoplasmic reticulum Ca 2ϩ -ATPase (SERCA) 6 pump is a P-type ATPase that catalyzes the transport of two Ca 2ϩ ions into the lumen of the sarcoplasmic reticulum (SR), at the expense of energy from the hydrolysis of one ATP molecule (1). SERCA pump activity in cardiac and skeletal muscle is regulated by the Ca 2ϩ ion gradient across the SR and by its regulatory proteins phospholamban (PLB) and sarcolipin (SLN), which physically interact with SERCA in a Ca 2ϩ -sensitive manner (2)(3)(4)(5). PLB is a 52-amino acid protein, consisting of a single transmembrane (TM) 22-residue ␣-helix connected to a 30-residue ␣-helical cytoplasmic domain. The role of PLB as an important regulator of cardiac contractility has been extensively studied (6,7). PLB modulates SERCA pump affinity for Ca 2ϩ and binds to SERCA in the Ca 2ϩ -free state (8 -10). PLB phosphorylation has been shown to relieve inhibition of SERCA and accelerate Ca 2ϩ uptake causing increased muscle relaxation (7). Compared with PLB, SLN is a much shorter 31-amino acid protein with a TM ␣-helix, an unstructured 7-residue cytosolic N terminus, as well as a unique C-terminal luminal tail (RSYQY) that is highly conserved from mouse to human (2,3,11). Unlike PLB, SLN is expressed abundantly in skeletal muscle tissues but its expression in cardiac muscle is restricted to the atria, although it is reported to be up-regulated in the hypertrophied and failing ventricle (12).
Initial studies on SLN suggested that SLN is an uncoupler of the SERCA pump and could enhance heat production at the expense of ATP hydrolysis (13,14). Using SLN Ϫ/Ϫ and SLN overexpression mouse models, we recently showed that SLN is an important player in muscle thermogenesis and metabolism (4,(15)(16)(17). Our studies revealed that mice lacking SLN were unable to maintain body temperature when exposed to cold but could be rescued by re-expression of SLN. Additionally the SLN Ϫ/Ϫ mice developed obesity when fed a high fat diet suggesting SLN regulation of SERCA activity contributes to whole body energy metabolism (4). However, the mechanistic details of how SLN binding promotes uncoupling of SERCA resulting in heat production are unclear. Thus a major interest of our current research is to elucidate the molecular mechanism of the SLN-SERCA interaction.
Toward this goal we have developed a SLN and SERCA expression system in HEK cells and employed chemical crosslinking, SERCA ATP hydrolysis, and Ca 2ϩ uptake assays to investigate how SLN interacts with SERCA during Ca 2ϩ transport. Using this approach we recently showed that monomeric SLN binds to SERCA directly and the nature of SLN interaction with SERCA differs significantly from PLB (3,4). A novel finding of this study was that although increasing the Ca 2ϩ concentration reduced the interaction between SERCA and SLN, it did not abolish the interaction. Interestingly, SLN was also able to bind to SERCA in the Ca 2ϩ -bound, phosphoenzyme conformation of the pump, E1P. In contrast, PLB binds to SERCA only in the Ca 2ϩ -free E2 state of the pump (3,8,10,18). These studies highlighted a key difference between SLN and PLB: that only SLN can bind to SERCA during Ca 2ϩ transport and this unique property of SLN leads to uncoupling of SERCA. Despite these findings, the detailed mechanism of how SLN binding to SERCA promotes uncoupling of the pump and causes release of Ca 2ϩ back to the cytosol is yet to be understood.
To date, attention has been focused on the C-terminal tail of SLN as the regulator of SERCA activity (19 -21). Furthermore, recent x-ray crystallographic studies on co-crystals of SERCA-SLN and SERCA-PLB have added valuable information about the interaction of monomeric PLB or SLN with SERCA (22)(23)(24). However, these crystal structures were unable to identify where the N terminus of SLN or PLB interact with SERCA, due to the dynamic nature of the N terminus. Previous studies have found that the N-terminal domain of PLB is critical for the modulation of SERCA pump function (25)(26)(27). The most conspicuous structural differences between SLN and PLB lie in their N-and C-terminal regions, which may dictate their functional differences in SERCA regulation. Therefore, a major objective of this study was to investigate the structural regions of SLN necessary for interaction and uncoupling of SERCA. We especially investigated whether the N terminus regulates the uncoupling function of SLN. In the present study we generated chimeras by swapping SLN and PLB domains to test the specific function of each. Our studies for the first time demonstrate that the N terminus of SLN plays an important role in its uncoupling action on SERCA.

Experimental Procedures
Materials-Lipofectamine, DMEM, and other cell culture reagents were obtained from Invitrogen. The cross-linking reagent 1,6-bismaleimidohexane was purchased from Pierce (Thermo Scientific). Thapsigargin, sodium orthovanadate, AlCl 3 , and KOH were purchased from Sigma. 45 CaCl 2 was obtained from PerkinElmer Life Sciences.
Generation of SERCA, SLN, and PLB Constructs-The SLN, PLB, and SERCA constructs were generated by PCR amplification and cloning into pcDNA3.1(ϩ) expression vector as reported previously (3). The rat SERCA1 cDNA sequence that has 99% homology with mouse SERCA1 was cloned into the pcDNA3.1(ϩ) vector. Mouse SLN and PLB cDNAs were PCR amplified and cloned into pcDNA3.1(ϩ) vector. The desired mutagenesis of rat SERCA1, mouse SLN, and mouse PLB were done using the QuikChange TM site-directed mutagenesis kit (Agilent Technologies). All cDNA clones and mutated constructs were confirmed by direct sequencing. We have previously shown that E7C SLN and N30C PLB are similar in function to WT SLN and WT PLB (3). The mutants E7C SLN and N30C PLB were used in this study as native SLN and PLB, and also to generate all chimeras to enable cross-linking with SERCA at Cys 318 (hence, E7C SLN and N30C PLB will be denoted as SLN and PLB throughout this text for simplicity, as shown in Fig. 1, A and B). Residues 2-6 of SLN were deleted to make the SLN NTdel mutant. The SLN N terminus (SLN residues 1-7) was replaced with the PLB N-terminal domain (PLB residues 1-30) to make Chimera-1. The last three residues (MLL) FIGURE 1. Design of SLN and PLB chimeras. A, mouse E7C SLN and N30C PLB cDNAs were used for generation of the constructs. SLN (31 aa) is represented as: cytosolic N terminus (black) consisting of 7 residues, TM segment (light green) formed by residues 8 -26 spanning the SR membrane and the luminal C terminus (dark green), which contains residues 27-31. Phospholamban (52 aa) is represented as: cytosolic N terminus (red) containing residues 1-30 and TM (blue) containing residues 31-52. SLN NTdel construct was made by deleting SLN-N-terminal residues 2-6 (ERSTQ). Chimera-1 contained PLB residues 1-30 and SLN residues 8 -31. Chimera-2 had PLB residues 1-49, followed by RSYQY, the C terminus of SLN. Chimera-3 contained SLN, N-terminal residues 1-7, residues 31-49 of the PLB TM domain, and SLN residues RSYQY at the C terminus. B, sequence alignment of SLN, PLB, and their chimeras showing the conserved amino acids in the transmembrane domain (red). Highlighted blue residues were mutated to cysteine for cross-linking with Cys 318 on SERCA.
in the C-terminal of PLB were replaced with SLN C-terminal tail residues, RSYQY to obtain Chimera-2. The length of PLB TM was kept intact to prevent possible mis-localization of PLB (28), as well as to ensure ideal interaction of RSYQY with SERCA. The N-terminal residues 1-30 of Chimera-2 were replaced with residues 1-7 of SLN to get Chimera-3 ( Fig. 1, A  and B). We further mutagenized residues 3 RS to AA, 5 TQ to AA, as well as 3 RSTQ to AAAA to obtain N-terminal alanine mutants of SLN. Protein expression of the chimeras was determined using the SLN antibody, which detects the C-terminal RSYQY sequence of SLN. PLB expression was determined by PLB antibody, which recognizes residues 2-13 of PLB N-terminal domain.
Expression of Proteins and Microsome Preparation-HEK293 cells were co-transfected with SERCA and SLN or PLB cDNAs at a 1:2 ratio, using Lipofectamine 2000. The cells were harvested in PBS 48 h post-transfection and the pellet was flash frozen in liquid nitrogen and stored at Ϫ80°C. Microsomes from transfected cells were prepared as described previously (3).
SERCA-mediated Ca 2ϩ Uptake and ATP Hydrolysis Assays-Oxalate-supported Ca 2ϩ uptake assay was performed to determine the effect of SLN, PLB, or their mutants on SERCA-mediated Ca 2ϩ transport. The proteins were expressed in HEK cells by transfection of DNA using Lipofectamine (3). HEK homogenates were incubated in Ca 2ϩ uptake buffer with the desired free Ca 2ϩ level as determined by Maxchelator webware. Aliquots of the reaction were collected and filtered at 30, 60, and 90 s. The filters were washed and bound radioactive 45 Ca 2ϩ was measured (29). ATP hydrolysis activity was measured using the Biomol green phosphate assay reagent (3,30). The reaction was initiated by addition of 5 g of microsomes in the Ca 2ϩ uptake buffer and samples were collected at the end of 30 min. The amount of P i release was calculated as nanomole of P i /mg/ min. Values of reactions carried out in the presence of thapsigargin (TG), a SERCA inhibitor, were subtracted from all samples to obtain SERCA specific activity (3).
All ATP hydrolysis and Ca 2ϩ uptake assays were carried out with a SERCA control for each sample. The SERCA protein level in the samples was normalized to that in the control to obtain the final absolute activity values. The final values were expressed as a percentage of the maximum control SERCA activity. Activity curves with n ϭ at least 3 measurements were generated using sigmoidal dose-response curve fitting in the GraphPad Prism 6 software to get the best-fit values Ϯ S.E. Analysis of data were done using ANOVA followed by Dunnett's multiple comparisons test.
Chemical Cross-linking-Chemical cross-linking of proteins was performed using the 10-Å homo-bifunctional sulfhydryl cross-linker, 1,6-bismaleimidohexane, as previously described (3). Briefly, 15 g of microsomes were mixed with 3 mM ATP in cross-linking buffer followed by addition of 0.1 mM crosslinker. The reaction was stopped after incubation for 1 h at 25°C, by the addition of SDS-PAGE sample-loading buffer containing 100 mM dithiothreitol. Specific cross-linking of SLN, SLN NTdel , Chimera-1, and Chimera-3 to SERCA showed a band above 110 kDa probed with anti-SLN antibody, whereas PLB or the Chimera-2 interaction with SERCA showed a 116-kDa band probed with anti-PLB antibody. The effect of TG was studied by adding 0 -10 M TG to the reaction mixture without Ca 2ϩ , in the presence of ATP, before the addition of crosslinker (3).
Cross-linking to Different Kinetic States of SERCA-Stable analogs of the major intermediates of SERCA kinetic steps (E2, E1, E1PCa 2 ⅐ADP, and E2P) were obtained by incubating microsomes in a cross-linking reaction buffer for 45 min at 25°C with different chemicals, before addition of 1,6-bismaleimidohexane (3,31,32). Incubating the microsomes without ATP induced the E2 state. 100 M free Ca 2ϩ with 3 mM ATP in the buffer resulted in a prevalence of the E1⅐PCa 2 state. Incubation with 50 M AlCl 3 and 3 mM KF and 3 mM ADP resulted in formation of the E1⅐AlF x ⅐ADP complex, an E1PCa 2 ⅐ADP analog. Last, the analog of the E2P to E2 transition state (E2⅐AlF 4 Ϫ ) was produced by addition of 50 M AlCl 3 and 3 mM KF to the buffer.
MD Simulations-The simulations were performed with the GROMACS 5.0 program (33). For modeling the system we used the crystal structure of SERCA and SLN in the Mg 2ϩ -bound E1 state (Protein Data Bank code 3W5A) (24). Here the ATP analog TNP-AMP was deleted, whereas the bound Mg 2ϩ -ion and crystal waters were retained. For the truncated version we deleted residues 2-6 of SLN after equilibration. We used the same protonation scheme as Espinoza-Fonseca et al. (34), namely that, Ca 2ϩ -binding residues Glu 771 , Asp 800 , and Glu 309 were kept unprotonated, whereas Glu 908 was protonated. The system consisted of SERCA, SLN, 500 POPC molecules, and ϳ55,000 TIPS3P water molecules. The protein was inserted into a pre-equilibrated bilayer using the "membed" option of GROMACS 5.0 (35). Furthermore, Na ϩ and Cl Ϫ ions were added to ensure a salt concentration of 150 mM. Virtual sites were used for all hydrogens in the system (36) allowing for a 5-fs time step in conjunction with the CHARMM36 force field (37). Both systems were run for 200 ns of which 50 ns were discarded as equilibration. Periodic boundary conditions were used, electrostatic forces were calculated using the Particle Mesh Ewald (38) method with a real space cut off of 1.2 nm, and the nonbonded interactions were calculated using a force switch in the range 1.0 to 1.2 nm. The temperature was kept at 310 K using the Nose-Hoover thermostat (39, 40) with a 10-ps time constant, and the pressure was coupled semi-isotropically using the Parrinello-Rahman barostat (41) as required by the virtual sites method, a reference value of 1.0 bar and a time constant of 50 ps. All bonds were constrained with LINCS (42).

Deletion of SLN Cytosolic Portion (N-terminal 2-6 Residues)
Leads to Loss of SLN Uncoupling Effect on SERCA-SLN has a unique unstructured N terminus consisting of 7 aa, whereas PLB has a 30-aa long ␣-helical N-terminal domain. X-ray crystallographic studies on SLN-SERCA co-crystals were unable to map the location of the N terminus and the interaction of N-terminal residues with SERCA remains a mystery so far (22,24). We therefore wanted to determine whether the unique N-terminal residues are responsible for the uncoupling action of SLN on SERCA. Toward this goal, we first deleted the N-terminal 2-6 residues from SLN (SLN NTdel ). The deleted con-struct was co-expressed with SERCA in HEK cells. The mutant protein was expressed at high levels and localized to the ER. Microsomes containing SERCA and SLN NTdel were then crosslinked at increasing Ca 2ϩ concentrations. As shown in Fig. 2A, SLN interacted with SERCA at all Ca 2ϩ concentrations tested, however, cross-linking intensity was reduced with increasing Ca 2ϩ concentration (3,4). SLN NTdel was also able to cross-link with SERCA; however, the intensity of the cross-linking remained unchanged with increasing Ca 2ϩ concentration ( Fig.  2A). This data suggested that SLN NTdel lost its ability to interact with SERCA in a Ca 2ϩ -sensitive manner but it remained bound to SERCA constitutively. We further tested if SLN NTdel affected SERCA activity by performing SERCA ATP hydrolysis and Ca 2ϩ uptake assays. Our results showed that both SLN and SLN NTdel had no significant effect on ATP hydrolysis activity (Fig. 2B). As expected, SLN decreased the V max of Ca 2ϩ uptake by 37% (Fig. 2C), which is in agreement with our previous study (3). Interestingly SLN NTdel had no effect on SERCA Ca 2ϩ uptake (Fig. 2C). These data suggested that although SLN NTdel was able to bind to SERCA constitutively, the deletion of the SLN N terminus made the molecule functionally inactive and it no longer affected SERCA ATP hydrolysis activity or Ca 2ϩ uptake.

Replacing SLN N Terminus (1-7 aa) with PLB N-terminal Domain (1-30 aa) Confers PLB-like Characteristics-
The above studies pointed out that deletion of the SLN N terminus resulted in a loss of function and clearly suggested that the N terminus is an important functional region. We then studied if the SLN function could be restored by addition of a non-homologous sequence to the N terminus. For this, we added the FLAG tag (DYKDDDDK) onto the N terminus of SLN NTdel . Similar to SLN NTdel , the resulting peptide was able to bind to SERCA at all Ca 2ϩ concentrations tested but had no effect on SERCA function (data not shown) and behaved similar to SLN NTdel . We next wanted to determine whether adding the PLB cytosolic domain (1-30 aa) (that has been shown to be important for PLB regulation of SERCA) to the N-terminaltruncated SLN would confer PLB-like function to the resulting Chimera-1 (Fig. 1B). We therefore compared the SERCA-Chimera-1 interaction with that of SERCA-SLN as well as SERCA-PLB at increasing Ca 2ϩ concentrations. As shown in Fig. 3A, SLN and PLB behaved as we have earlier reported (3); i.e. SLN was able to interact with SERCA at all Ca 2ϩ concentrations tested but PLB interaction with SERCA was abolished at higher Ca 2ϩ concentrations. Interestingly, Chimera-1 interacted with SERCA up to 1 M and the cross-linking was abolished at 2 M Ca 2ϩ (Fig. 3A), thus Chimera-1 behaved like PLB. The ATP hydrolysis assay showed that like PLB, Chimera-1 decreased the K 0.5 of SERCA activity (K 0.5 SERCA ϭ pCa 6.787 Ϯ 0.039; K 0.5 SERCA ϩ Chimera-1 ϭ pCa 6.638 Ϯ 0.014, p Ͻ 0.05; K 0.5 SERCA ϩ PLB ϭ pCa 6.248 Ϯ 0.063, p Ͻ 0.0001, mean Ϯ S.E.) (Fig. 3B). The Ca 2ϩ uptake assay showed that Chimera-1 did not inhibit the V max or affinity of SERCA for Ca 2ϩ (K 0.5 SERCA ϭ pCa 6.745 Ϯ 0.02; K 0.5 SERCA ϩ Chimera-1 ϭ pCa 6.535 Ϯ 0.134; K 0.5 SERCA ϩ PLB ϭ pCa 6.358 Ϯ 0.078, p Ͻ 0.01, mean Ϯ S.E.) (Fig. 3C). These results collectively suggest that replacing the SLN N terminus with the PLB N terminus leads to a loss of SLN-like function in the resulting Chimera-1, despite the fact that it has both the TM and C-terminal residues of SLN.
SLN N Terminus Plays a Role in Uncoupling the SERCA Pump-We additionally generated Chimeras-2 and -3 to determine the functional relevance of the SLN C and N terminus. In Chimera-2, the C-terminal residues of PLB (MLL) were replaced with the C-terminal residues of SLN (RSYQY). To test the role of the SLN N terminus we created Chimera-3, using Chimera-2 from which the N terminus of PLB was replaced with that of SLN (MERSTQ) (although this chimera contained the TM domain of PLB, no PLB antibody specific for this region is available. We therefore had to replace the C-terminal residues of PLB (MLL) with those of SLN (RSYQY) for immunodetection.) SLN, PLB, and the chimeras were each expressed with SERCA in HEK cells and cross-linking and functional assays were carried out as described above. Cross-linking studies showed that the PLB interaction with SERCA was highly sensitive to Ca 2ϩ and was abolished between 1 and 2 M Ca 2ϩ . On the other hand, Chimera-2 interaction with SERCA was stronger and persisted even at 2 M Ca 2ϩ but was abolished at 5 M Ca 2ϩ (Fig. 4A), suggesting that the addition of the SLN C terminus (RSYQY) residues onto PLB increased its affinity to SERCA. Interestingly, Chimera-3 cross-linked to SERCA at all Ca 2ϩ concentrations tested, in a Ca 2ϩ -sensitive manner (Fig.  4A) as observed for SLN ( Fig. 2A).
To further determine how Chimeras-2 and -3 affect SERCA activity, we measured SERCA ATP hydrolysis and Ca 2ϩ uptake. The ATP hydrolysis assay showed that similar to PLB, the presence of Chimera-2 decreased the affinity of the SERCA pump for Ca 2ϩ , without a significant effect on V max of the pump (K 0.5 SERCA ϭ pCa 6.595 Ϯ 0.044; K 0.5 SERCA ϩ Chimera-2 ϭ pCa 6.378 Ϯ 0.05, p Ͻ 0.01, mean Ϯ S.E.). The Ca 2ϩ uptake assay also showed that Chimera-2 decreased the apparent affinity of SERCA for Ca 2ϩ uptake, without affecting V max (K 0.5 SERCA ϭ pCa 6.785 Ϯ 0.023; K 0.5 SERCA ϩ Chimera-2 ϭ pCa 6.398 Ϯ 0.065, p Ͻ 0.0001; mean Ϯ S.E.), suggesting that Chimera-2 retains PLB-like function. These results revealed that although the C terminus of SLN increases the affinity of Chimera-2 to SERCA, it functions like PLB. When we swapped the N terminus of Chimera-2 with that of SLN, the resulting Chimera-3 did not affect ATP hydrolysis activity (K 0.5 SERCA ϭ pCa 6.595 Ϯ 0.044; K 0.5 SERCA ϩ Chimera-3 ϭ pCa 6.609 Ϯ 0.044, mean Ϯ S.E.) (Fig. 4B) and behaved similar to SLN. The Ca 2ϩ uptake assay showed that Chimera-3 did not affect the apparent affinity of SERCA for Ca 2ϩ uptake (K 0.5 SERCA ϭ pCa 6.785 Ϯ 0.023; K 0.5 SERCA ϩ Chimera-3 ϭ pCa 6.706 Ϯ 0.058, mean Ϯ S.E.) but showed a 25% decrease in the V max of SERCA pump (Fig. 4C). These data using Chimera-3 suggest that replacing the PLB N terminus of Chimera-2 with that of SLN converts it to a SLN-like molecule.
The N Terminus of SLN Enables Its Dynamic Interaction with SERCA-The SERCA pump undergoes four major kinetic steps (E2, E1P, E2P, and E2) during Ca 2ϩ transport (1). We have  previously shown that SLN interacts with SERCA through various kinetic steps of the Ca 2ϩ transport cycle, whereas PLB cannot bind to Ca 2ϩ bound SERCA (3). Therefore we investigated if SLN binding to SERCA in its various kinetic states is dependent on the N terminus of SLN, by the use of these chimeras. We assessed the ability of each chimera to bind to the five different kinetics states of SERCA as described previously (3). We measured protein cross-linking in the E2 state with and without ATP; the Ca 2ϩ bound phospho-intermediate and transition state, as well as the E2P state. Our cross-linking results showed that only SLN and Chimera-3 having the SLN N terminus were able to interact with SERCA across all tested kinetic states. However, PLB, Chimera-1, and Chimera-2 (with the PLB N terminus) bound only to the Ca 2ϩ -free E2 conformation and this interaction was enhanced in the presence of ATP (Fig. 5A).
It has been previously shown that PLB cannot interact with Ca 2ϩ -bound SERCA, whereas SLN can (3,4). To further determine how Ca 2ϩ binding to both sites in SERCA affect its interaction with the different chimeras, we performed cross-linking with the D351A-SERCA mutant. The D351A-SERCA mutant can bind two Ca 2ϩ ions but cannot hydrolyze ATP; therefore in the presence of Ca 2ϩ and ATP at pH 7.0, all the pumps remain in a Ca 2ϩ -bound state. Our results showed that SLN and Chimera-3 (having SLN N terminus) were able to bind to D351A-SERCA but Chimeras-1 and -2 dissociated from SERCA below 250 nM Ca 2ϩ (Fig. 5B). Additionally, cross-linking of SLN NTdel and chimeras to SERCA in the presence of increasing concentrations of TG (inhibitor that blocks SERCA irreversibly) showed that except for SLN NTdel , the cross-linking of all other chimeras with SERCA was abolished at 10 M TG. This suggested that the chimeras bound to active SERCA pumps and the interaction was dynamic, whereas SLN NTdel constitutively bound to SERCA (Fig. 5C).
Alanine Mutagenesis of SLN Cytosolic Residues Alters Ca 2ϩdependent Interaction with SERCA-The N terminus of SLN (MERSTQ) is relatively short and the ability of SLN to remain bound to SERCA may involve specific interactions of the N terminus with SERCA. To test this idea we mutated the N-terminal residues to alanines and analyzed the ability of these mutants to bind to SERCA in a Ca 2ϩ -dependent manner. We generated double alanine mutants at positions 3, 4 ( 3 RS-AA), and 5, 6 ( 5 TQ-AA) at the N terminus of SLN. In addition we mutated all 4 residues together to alanine ( 3 RSTQ-AAAA) to determine the effect of these residues on the function of SLN. Our chemical cross-linking results showed that the 3 RS-AA mutant had a minimal effect on SLN cross-linking with SERCA at a high Ca 2ϩ concentration. The 5 TQ-AA mutant showed decreased interaction with SERCA at a high Ca 2ϩ concentration, whereas mutation of all 4 residues resulted in further reduction of the interaction, which was abolished at 2 M Ca 2ϩ (Fig. 6). We also performed SERCA functional assays in the presence of these SLN mutants to study the role of the N-terminal residues (Table 1). Although the V max values were low in the presence of the mutant SLNs, we found no significant difference in the ATP hydrolysis activity of SERCA when expressed with the mutants, as compared with SERCA alone. Additionally we observed that only the 3 RSTQ-AAAA SLN mutant significantly reduced the V max of SERCA Ca 2ϩ uptake, similar to native SLN. Although the SLN 3 RS-AA and 5 TQ-AA lowered the V max of SERCA Ca 2ϩ uptake, this reduction was not significantly different from the SERCA control or even the SERCA ϩ SLN V max . The simulations also showed that residue Glu 2 of SLN forms a salt bridge with Arg 324 of SERCA, which is located on the cytosolic half of the M4 helix (M4C) and is oriented toward the groove in SERCA where SLN is bound. This helix M4C has been previously implicated in inter-domain communication in the SERCA Ca 2ϩ transport cycle (43,44). Fig. 7C shows representative snapshots from the two systems. In SERCA ϩ SLN NTdel , the loss of this salt bridge results in changes in interactions with Asn 111 , Asn 114 , and Ala 118 . For Asn 114 and Ala 118 the interaction becomes stronger (distance decreases) and for Asn 111 the interaction becomes weaker when removing the N terminus as can be seen in Fig. 7D.

ATP hydrolysis activity
Calcium uptake

Discussion
A major objective of this study was to determine the structural regions that enable SLN interaction with SERCA and promote uncoupling of the pump from Ca 2ϩ transport. We recently showed that monomeric SLN binds to the TM groove on SERCA consisting of TMs M2, M6, and M9, a region previously characterized to bind PLB (45). The binding of SLN to this groove, however, involves a different set of SERCA residues (3,45). Most importantly, we showed that SLN remains bound to SERCA in the presence of high Ca 2ϩ , whereas PLB is displaced from the groove. We also showed that SLN interacts with SERCA throughout the Ca 2ϩ transport cycle and proposed that its ability to interact with SERCA during the Ca 2ϩ transport cycle promotes uncoupling of the SERCA pump. These and other published studies suggest that PLB is an affinity modulator of SERCA (inhibits SERCA at submaximal Ca 2ϩ concentrations), whereas SLN is an uncoupler of the pump and decreases V max of Ca 2ϩ uptake (3,7,10,13,46,47). It is, however, unclear how SLN is able to uncouple the SERCA pump.
To date, the structural components that enable SLN to uncouple SERCA remain to be characterized. SLN is a short 31-amino acid protein consisting of an unstructured 7-residue (MERSTQE) N terminus, a TM ␣-helix made of 19 residues, and a unique 5-residue (RSYQY) C-terminal luminal tail (48 -51). Previous functional studies on SLN suggested that the C-terminal RSYQY sequence is important for SERCA regulation (19 -21). The published x-ray crystallographic studies by Toyoshima et al. and Winther, Bublitz, and co-workers (22,24) are an important milestone, which has shown that SLN indeed binds to the M2, M6, and M9 TM groove and stabilizes the E1 state of SERCA. The crystal structures showed interaction in the TM region but were unable to identify the N terminus locations of SLN on SERCA; therefore the nature of the SLN N terminus interaction with SERCA and its role in regulating SERCA function have remained elusive. We knew that SLN and PLB bind to the same SERCA groove but produce different functional outcomes (3,(13)(14)45). Hence, we chose to generate chimeras between PLB and SLN to dissect their unique functional domains and determine if the domains could be exchanged without losing function. Results from our study show that we were successful in creating functional chimeras. Moreover, these chimeras inherited the function of their respective N terminus.
The TM and C Terminus Are Sufficient for SLN Localization and Interaction with SERCA-Despite recent progress in localizing SLN to the SERCA TM groove, it remained unclear if the uncoupling function could be localized to a particular portion or a set of residues in SLN. In the present study, we focused our attention on investigating the role of the SLN N terminus in the uncoupling of SERCA. We found that deletion of the N terminus (ERSTQ) of SLN disabled its function (Fig. 2). The mutant peptide (SLN NTdel ) constitutively bound to SERCA, failed to show Ca 2ϩ sensitivity, and had no effect on SERCA Ca 2ϩ transport. We have earlier shown that native SLN or PLB are displaced from SERCA by TG, a SERCA inhibitor, which locks the pump irreversibly in the E2 conformation (3). Interestingly, the SLN NTdel peptide could not be dislodged from SERCA in the presence of TG (Fig. 5C). Thus SLN without its N terminus remained in the TM groove and lost its ability to dynamically interact with SERCA. These studies suggest that the occupation of the SERCA groove by the truncated SLN peptide is not sufficient to cause uncoupling. At the same time this mutant could successfully localize to SERCA, which suggests that the TM and C terminus are vital for SLN localization and interaction with SERCA in the TM groove (21), which is also in agreement with recent x-ray crystallographic studies (22,24).
The N Termini of SLN and PLB Are Essential for Their Unique Function-Because the truncated SLN peptide failed to show any effect on SERCA Ca 2ϩ transport, we added a heterologous N-terminal FLAG tag and also the N-terminal domain (1-30 aa) of PLB (Chimera-1) to observe an effect on SERCA. Only the addition of the PLB N terminus showed an effect on SERCA function and caused Chimera-1 to decrease the apparent affinity of SERCA but not V max (Fig. 3). This is in agreement with previous findings showing that the inhibitory function of PLB can be localized to the N terminus (25,26,52).
Similarly, the PLB N terminus, not the SLN C terminus, determined the function of Chimera-2. Although the SLN C terminus increased the binding affinity of Chimera-2 to SERCA, which is in agreement with recent studies by Gorski et al. (20), we, however, found that it was unable to promote SERCA uncoupling in the presence of the PLB N-terminal domain (Fig. 4). An interesting observation was that Chimera-2 behaved more like PLB than Chimera-1. The only variable factor between these two chimeras is the TM segment, which suggests that the TM of PLB plays a significant role in defining PLB characteristics. We additionally investigated if the swapping of the PLB N terminus of Chimera-2 with the SLN N terminus would promote SERCA uncoupling. The resulting Chimera-3 behaved much like SLN, with no effect on the apparent affinity of SERCA but a decreased V max of Ca 2ϩ uptake (Fig. 4). Although a direct comparison between SLN and Chimera-3 was not made, Chimera-3 also highlighted that the TM domain can be switched easily without affecting SLN properties. These studies employing chimeras allowed us to conclude that the N termini of SLN and PLB influence their respective identity as an uncoupler or affinity modulator. However, the TM and C terminus are important for SLN localization within the SERCA groove and without them SLN will not be able to interact with SERCA. In this study we additionally show that the TMs of PLB and SLN could be switched without compromising SLN function.
Although the above studies identified the SLN N terminus as an important uncoupling portion, it remained unclear how this region interacts with SERCA in a Ca 2ϩ -sensitive manner. Therefore we mutagenized the N-terminal RSTQ residues into alanines and found that mutagenesis from 2 to 4 residues increasingly weakened the mutants' ability to bind to SERCA (Fig. 6) as compared with native SLN. This could be a result of unfavorable interaction between SERCA residues and the N-terminal alanines on SLN, causing a change in the orientation of the SLN TM and C terminus within the SERCA binding groove. On the other hand, compared with alanine mutagenesis, deletion of the N terminus caused the truncated SLN peptide to constitutively bind to SERCA even at a high 5 M Ca 2ϩ concentration and unlike native SLN, it lost its SERCA regulating function and remained bound in the presence of TG. The SLN N terminus is thus important for its binding and dynamic interaction with SERCA during the kinetic cycle of the pump. The N-terminal alanine mutations, however, did not significantly alter the functional effect of SLN on SERCA (Table 1). It should be noted that these mutagenesis studies are only preliminary and further investigation with exhaustive mutagenesis using a range of amino acids is necessary to provide a definitive picture of how the side chain interactions in this region affect SLN function.
Evolution and Functional Divergence of SLN and PLB-Our studies using native SLN, PLB, and their chimeras, show that domain function can be separated and the domains can function independently of each other. The very fact that they both bind to the same SERCA groove, and that the two proteins have retained considerable homology in the TM region, provide support to their common origin. Recent studies in Drosophila have identified an ancestral protein, Sarcolamban, as a regulator of SERCA, which shares most homology in the TM region with PLB and SLN. This finding and studies on the SLN and PLB structure suggest that these proteins must have evolved from a common ancestral gene involved in Ca 2ϩ homeostasis (53). However, somewhere during evolution, the ancestral gene duplicated and diverged to code for PLB and SLN with very distinct functions even before the dawn of vertebrates (53,54). Thus PLB and SLN represent a unique set of proteins that bind to and interact with the same SERCA pump but have entirely different functional outcomes.
Mechanism of SLN Uncoupling of SERCA-Although the mechanism of SLN uncoupling remains to be understood, our study represents an important first step toward this goal. During Ca 2ϩ transport, the SERCA pump transports two Ca 2ϩ ions per molecule of ATP hydrolyzed (1,55,56). However, when SLN is bound to SERCA, the energy from ATP hydrolysis does not lead to accumulation of Ca 2ϩ in the SR (3,4), but is released as heat, thus contributing to thermogenesis in muscle (3,4). Crystallography data of the various SERCA kinetic states show that transferring Ca 2ϩ to the SR lumen requires rotation of the A domain of SERCA, which results in the sliding of TM helices that releases Ca 2ϩ ions into the SR lumen (1,(57)(58)(59). We propose that the interaction of SLN within the SERCA TM groove during Ca 2ϩ transport may interfere with TM sliding, resulting in slippage of Ca 2ϩ back into the cytosol (60). The MD simulations suggest the possibility that the short and flexible N terminus of SLN is capable of interacting with residues in the SERCA cytosolic domain. Deletion of the SLN N terminus affected spatial orientations of several SERCA residues, which are known to have an influence on SERCA function. Previous studies have shown that mutations N111A and N114A lead to a reduction in the transport of Ca 2ϩ (61), R324A has the same effect and also inhibits the E1P-E2P transition of SERCA (the step at which Ca 2ϩ is internalized into the SR lumen) (43,44). Furthermore, SERCA Glu 45 lies near the cytosolic portion of the Ca 2ϩ ion binding pathway in SERCA. Although the alanine mutagenesis results for the role of SLN residue Arg 3 are inconclusive, its interaction with SERCA Glu 45 may play a role in SLN-induced Ca 2ϩ slippage from the pump.
In summary, for the first time we report that the cytosolic N terminus in SLN plays an important role in uncoupling SERCA. The N terminus of SLN may be responsible for ideal positioning of the SLN TM and C terminus in the SERCA groove, to enable uncoupling of the pump. We also show that localization to the SR membrane and interaction in the SERCA groove depends on the SLN TM and C terminus. Our studies highlight the modular nature of the domains in SLN and PLB, which can be interchanged without losing their individual function. Furthermore, we show that when exchanged, the N terminus confers identity to the new protein. We have only begun to understand the role of this region in SLN. Future studies should be aimed at understanding in further detail the interaction of the N-terminal residues of SLN with SERCA.