Research ArticleSTRUCTURAL BIOLOGY

Cryo-EM structures of SERCA2b reveal the mechanism of regulation by the luminal extension tail

See allHide authors and affiliations

Science Advances  12 Aug 2020:
Vol. 6, no. 33, eabb0147
DOI: 10.1126/sciadv.abb0147

Abstract

Sarco/endoplasmic reticulum Ca2+ ATPase (SERCA) pumps Ca2+ from the cytosol into the ER and maintains the cellular calcium homeostasis. Herein, we present cryo–electron microscopy (cryo-EM) structures of human SERCA2b in E1∙2Ca2+–adenylyl methylenediphosphonate (AMPPCP) and E2-BeF3 states at 2.9- and 2.8-Å resolutions, respectively. The structures revealed that the luminal extension tail (LE) characteristic of SERCA2b runs parallel to the lipid-water boundary near the luminal ends of transmembrane (TM) helices TM10 and TM7 and approaches the luminal loop flanked by TM7 and TM8. While the LE served to stabilize the cytosolic and TM domain arrangement of SERCA2b, deletion of the LE rendered the overall conformation resemble that of SERCA1a and SERCA2a and allowed multiple conformations. Thus, the LE appears to play a critical role in conformational regulation in SERCA2b, which likely explains the different kinetic properties of SERCA2b from those of other isoforms lacking the LE.

INTRODUCTION

The endoplasmic reticulum (ER) is the organelle where secretory and membrane proteins are synthesized and acquire higher-order structure under the surveillance of cellular protein quality control systems (1, 2). Another important function of the ER is calcium ion (Ca2+) storage, and regulation of Ca2+ efflux from this organelle plays a critical role in controlling cellular growth, proliferation, differentiation, and death (3). Thus, the ER has developed the elaborate mechanisms for maintaining Ca2+ homeostasis that use multiple kinds of Ca2+ pumps and channels.

Sarco/endoplasmic reticulum calcium ion ATPase (adenosine triphosphatase) (SERCA) is a P-type ATPase family member that conducts Ca2+ uptake from the cytosol to the ER against an uphill gradient of >1000-fold difference in Ca2+ concentration (4). The SERCA family contains three isoforms, SERCA1 to SERCA3, among which SERCA2b serves as a ubiquitously expressed housekeeping enzyme involved in the cellular Ca2+ homeostasis. They share similar overall structure with three cytoplasmic domains comprising an actuator (A), a nucleotide-binding (N), and a phosphorylation (P) domain, and 10 transmembrane (TM) helices (Fig. 1A) that can be divided into three TM helix clusters: TM1–TM2, TM3–TM4, and TM5–TM10 (5, 6).

Fig. 1 Cryo-EM structures of SERCA2b WT in the E1∙2Ca2+-ATP and E2P states.

(A) Topology diagram of SERCA2b. Conserved A, N, and P domains and TM helices (TM1 to TM10) are colored yellow, magenta, cyan, and wheat, respectively. The C-terminal 11th TM helix (TM11) and the subsequent luminal extension tail (LE) are colored red, while L7/8 is colored blue. The scissors indicate the truncation site used to generate the T1032stop construct. (B) Catalytic cycle for SERCA to transport Ca2+ from the cytosol to the ER lumen through hydrolysis of ATP. The intermediate sates for which cryo-EM structures have been determined in this work are colored pink and cyan, respectively. (C and D) Overall cryo-EM structures of SERCA2b WT in the E1∙2Ca2+-AMPPCP (C) and E2-BeF3 (D) states. The color pattern is the same as in (A).

Crystal structures of a series of reaction intermediates and transition states stabilized with appropriate ligands or analogs (7, 8) have demonstrated the mechanism of actions of SERCA1a, an isoform specifically expressed in skeletal muscle, thereby revealing molecular details of the catalytic cycle of this Ca2+ pump. By contrast, structural and mechanistic information of SERCA2b remains scarce despite its physiological importance. SERCA2b shares 85% amino acid sequence identity with SERCA1a and shares residues 1 to 993 with SERCA2a, a splicing variant of SERCA2b. Unlike other isoforms, SERCA2b has a 49-residue C-terminal extension comprising a cytosolic loop (L10/11; Gly994-Asp1012), an additional (11th) TM helix (TM11; Gly1013-Tyr1030), and a luminal extension tail (LE; Ser1031-Ser1042). In this connection, SERCA2b exhibits a significantly lower maximal turnover rate than SERCA1a and SERCA2a, and almost twofold higher apparent calcium affinity (9).

While a previous study suggested physical contact of TM11 with TM7 and TM10 (10), our recent crystal structure of SERCA2b at 3.45-Å resolution revealed that TM11 is located adjacent to TM10 and engages in weak interactions with a part of the L8/9 loop and the N-terminal end of TM10 (11). Structural comparisons further demonstrated that because of the presence of TM11, SERCA2b has a different TM helix orientation relative to the cytosolic domains from those of SERCA2a and SERCA1a. However, electron density was missing for several critical segments including the LE. In addition, crystal structure of SERCA2b was solved only for the E1∙2Ca2+–adenylyl methylenediphosphonate (AMPPCP)–bound state. Thus, the mechanism underlying the SERCA2b regulation by the C-terminal extension remains poorly understood.

In the present work, we determined cryo–electron microscopy (cryo-EM) structures of human SERCA2b in the E1∙2Ca2+-AMPPCP and E2-BeF3 states at resolutions of 2.9 and 2.8 Å, respectively. The higher-resolution structures illuminated both the backbone and side-chain conformations over almost the entire part of SERCA2b. Furthermore, to define the location of the LE and gain deep insight into its regulatory roles, we determined cryo-EM structures of SERCA2b T1032stop, a truncated construct that spans the protein chain up to TM11 but lacks the LE. Structural comparison between wild-type (WT) SERCA2b and the T1032stop variant reveals the exact location of the LE and the mechanism of the LE-mediated structural regulation in SERCA2b. The present findings explain the regulated conformational transition between the E1∙2Ca2+-ATP and E2P states in SERCA2b.

RESULTS

Overall structures of SERCA2b in the E1∙2Ca2+-AMPPCP and E2-BeF3 states

We expressed and purified SERCA2b WT and T1032stop essentially as described previously (11), except that the buffer used for the final exchange comprised 50 mM Hepes (pH 7.0), 100 mM KCl, and 0.01% (w/v) lauryl maltose neopentyl glycol (LMNG), but contained no glycerol to allow acquisition of high-quality particle images. The cryo-EM structures of SERCA2b WT in the E1∙2Ca2+-AMPPCP and E2-BeF3 states were determined as mimics of the E1∙2Ca2+-ATP and E2P states (Fig. 1B) at a resolution of 2.9 and 2.8 Å, respectively (table S1). Repeated rounds of three-dimensional (3D) classification resulted in a single class of conformation of the E1∙2Ca2+-AMPPCP state (fig. S1). On the other hand, the E2-BeF3 state yielded another minor class (24.1%) of conformation after the same data processing (fig. S2). However, the density of TM11 and the LE was largely unclear in this minor class, suggesting that these additional C-terminal segments are highly disordered. In this regard, this minor class may not precisely reflect the structural features of SERCA2b, although we cannot exclude its physiological significance. Thus, we only focus on the major class of conformation with significant density of TM11 and the LE here.

In the cryo-EM structure of SERCA2b WT in the E1∙2Ca2+-AMPPCP state, all cytosolic domains and TM helices including TM11 are located at almost the same positions as observed in the previous crystal structure (Fig. 1C and fig. S3A) (11). The overall root mean square deviation (RMSD) of Cα atoms between these two structures is 1.10 Å. The higher-resolution cryo-EM structures provided more precise views of the modes of AMPPCP and Ca2+ binding (fig. S4, A and B). Moreover, no-alignment 3D classification with a mask covering TM11 resulted in clear density for the entire part of TM11 and its neighboring regions, even though density was still invisible for the major portion of L10/11, probably due to its intrinsically disordered structure. Thus, the cryo-EM structures of SERCA2b verified the close contacts between the side chains of W927 (L8/9) and Phe1018 (TM11) and between those of L967/L970/K971 (TM10) and V1029 (TM11) (fig. S3B), of which regulatory roles were proposed previously (11).

Notably, the LE of SERCA2b was visible in the cryo-EM maps of both the E1∙2Ca2+-AMPPCP and E2-BeF3 states (Fig. 1, C and D, and fig. S5, A and B). In sharp contrast, SERCA2b T1032stop did not show density around the corresponding regions (fig. S5, C and D), corroborating our identification of the LE. Consequently, the LE has been found to locate near the luminal ends of TM10 and TM7 and approach the short α helix (Phe866-Ser870) in L7/8 (Fig. 2, A and B), in partial agreement with the previous docking model, in which the LE was predicted to interact with L7/8 and modulate the apparent Ca2+ affinity of SERCA2b (10).

Fig. 2 Location of the LE in SERCA2b.

(A and B) Location of the LE in the E1∙2Ca2+-AMPPCP (A) and E2-BeF3 (B) states. The TM11-LE and L7/8 segments are colored red and blue, respectively. (C and D) Location of the LE relative to the first layer of phospholipids identified in crystal structures of SERCA1a. The cryo-EM structures of SERCA2b are superimposed to crystal structures of SERCA1a with the surrounding phospholipids in the E1∙2Ca2+-AMPPCP (C) and E2-BeF3 (D) states [Protein Data Bank (PDB) ID: 5XA8 and 5XA9, respectively] such that the RMSD of Cα atoms in TM7 to TM10 is minimized. Phosphorous atoms of the phospholipids are represented with cyan spheres. The color pattern for SERCA2b is the same as in (A), while the crystal structures of SERCA1a are colored gray.

In the preceding work (12), the entire first layer of phospholipids surrounding the TM helices of SERCA1a was visualized in the electron density maps obtained by x-ray solvent contrast modulation. By superimposing the present cryo-EM structures of SERCA2b to the crystal structures of SERCA1a with the surrounding phospholipids, we found that the LE runs parallel to the lipid-water boundary in the luminal side (Fig. 2, C and D). Accordingly, the LE contains polar and charged residues on the aqueous side, whereas the hydrophobic residues including F1036 and W1041 are directed toward the interior of the lipid bilayer. The LE contains no basic amino acids, which could, if any, anchor the nearby phospholipids through salt bridges with phosphate groups. It therefore seems unlikely that the LE serves to fix particular phospholipids or locally distort the lipid bilayer in accordance with the TM helix movements during the SERCA reaction cycle. Rather, the LE likely acts as an element that directly regulates the location and orientation of both the cytosolic and TM domains of SERCA2b, as described later.

Conformational transition from E1∙2Ca2+-ATP to E2P states in SERCA2b

The previous crystallographic studies of SERCA1a demonstrated that during the transition from the E1∙2Ca2+-ATP to E2P states, the nucleotide-binding site in the N domain and the Ca2+ exit channel constituted by TM4 to TM6 are opened to the cytosolic and luminal sides, respectively, to facilitate the release of adenosine 5′-diphosphate (ADP) and Ca2+ (5, 13). Similar conformational changes were observed for recently reported crystal structures of SERCA2a in E1∙2Ca2+-AMPPCP and cyclopiazonic acid (CPA)–stabilized E2-AlF4 states (11, 14). The present study revealed that, similarly to SERCA1a and SERCA2a, AMPPCP binds to the boundary of the P and N domains of SERCA2b, resulting in a compact headpiece cluster composed of the cytosolic A, N, and P domains (Fig. 1C and Fig. 3A, left). In the E2-BeF3 state, no Ca2+ is bound and a BeF3 molecule is covalently linked to the carboxylate side chain of Asp351 with assistance of a nearby Mg2+ ion (fig. S4, C and D). Because of the dissociation of the ADP moiety located at the interdomain interface, the cytoplasmic domains in the E2-BeF3 state are more separated from each other than those in the AMPPCP-bound state (Fig. 3A, right). Thus, the A, N, and P domains in the E2-BeF3 state are tilted down toward the membrane and rotated by 99°, 60°, and 20°, respectively, relative to the corresponding domains in the E1∙2Ca2+-AMPPCP state (Fig. 3, A and B). The movement of the A domain imposes a force on TM1 and TM2 in a lateral direction, which relocates TM3 and TM4 through van der Waals contacts between TM1 and TM3 and the luminal half of TM4 (Fig. 3C). Eventually, TM1 to TM4 are separated from TM5 and TM6 to open the calcium exit channel (Fig. 3A, right), whereas the other TM helices (TM7 to TM10) constitute a more rigid cluster. Thus, SERCA2b undergoes a significant rearrangement in both the cytosolic and TM domains upon ATP hydrolysis and subsequent ADP dissociation, leading to the facilitated release of Ca2+, as previously observed for SERCA1a.

Fig. 3 Conformational transition of SERCA2b from the E1∙2Ca2+-ATP to E2P states.

(A) Cryo-EM structures of SERCA2b WT in the E1∙2Ca2+-AMPPCP (left) and E2-BeF3 (right) states. The A, N, and P domains are colored yellow, magenta, and cyan, respectively. TM1 and TM2, TM3 to TM6, TM7 to TM10, and TM11 are colored blue, brown, wheat, and red, respectively. Ca2+ ions and Mg2+ ion are represented as purple and green spheres, respectively, AMPPCP and BeF3 are depicted by green and cyan sticks, respectively. (B) Top views of the cytosolic domains in the E1∙2Ca2+-AMPPCP and E2-BeF3 states of SERCA2b WT. The upper and lower panels illustrate the rearrangement of the A and N domains and the P domain, respectively, during the transition from the E1∙2Ca2+-AMPPCP to E2-BeF3 states. (C) Top view of the TM helix domain in the E1∙2Ca2+-AMPPCP and E2-BeF3 states of SERCA2b WT. The superposition illustrates the relocation of each TM helix during the transition from the E1∙2Ca2+-AMPPCP to E2-BeF3 states. For clarity, the cytosolic domains are not shown. All domains and TM helices in the E1∙2Ca2+-AMPPCP state are shown with transparency 0.5. Structures are superimposed such that the RMSD of Cα atoms in TM7 to TM10 is minimized.

Roles of the LE in the structural regulation of the E1∙2Ca2+-ATP form of SERCA2b

Detailed structural comparison of SERCA2b WT and T1032stop provided deep insight into the regulatory role of the LE. Unlike SERCA2b WT, the T1032stop mutant was classified into three different classes of conformations in a similar abundance (class 1 = 23.4%, class 2 = 26.7%, and class 3 = 26.7%) after the additional local search 3D classification with masks to remove micelle density (Fig. 4A and fig. S8), corroborating the LE-mediated conformational regulation in SERCA2b WT. All classes of T1032stop clearly showed density of bound AMPPCP, its neighboring Mg2+, and two bound Ca2+ in the cryo-EM maps (fig. S4, E and F), indicating that the preparation of the E1∙2Ca2+-AMPPCP form was homogeneous. Like WT, T1032stop retained the compact headpiece cluster of the A, N, and P domains in all classes. However, the cytosolic domains of T1032stop stood more upright relative to the TM domain than those of WT (Fig. 4B).

Fig. 4 Structural effects of the LE on SERCA2b in the E1∙2Ca2+-ATP state.

(A) Cryo-EM structures of three different classes (classes 1, 2, and 3) of SERCA2b T1032stop in the E1∙2Ca2+-AMPPCP state. The rotation angles of the cytosolic A, N, and P domains as compared with those of SERCA2b WT are indicated by red arrows showing the rotation direction. Classes 1, 2, and 3 are colored pale cyan, cyan, and dark cyan, respectively. (B) Cryo-EM structures of SERCA2b WT and class 1 of T1032stop in the E1∙2Ca2+-AMPPCP state are superimposed such that the RMSD of Cα atoms in TM7 to TM10 is minimized. Cylinder diagrams of SERCA2b WT and class 1 of T1032stop are colored red and cyan, respectively. The right inset presents a close-up view of the cytosolic parts of TM6 to TM11, L6/7, and L8/9. The left inset shows a side view of the TM helix domain, in which TM10 and TM11 are indicated by red (WT) and cyan (class 1 of T1032stop) cylinders. Residues critical for the positional shifts of the cytosolic domains and TM helices upon deletion of the LE are represented by sticks in the right inset. The movements of the cytosolic domains and TM helices induced by deletion of the LE are indicated by arrows in both insets. (C) The regions surrounded by violet, green, and red dotted squares in the right inset of (B) are highlighted in the left, middle, and right panels, respectively. The movements of residues in L6/7 and P domain induced by deletion of the LE are indicated by arrows. Red dotted lines in the left and right panels indicate a hydrogen bond between the R923 side chain and the G1013 main-chain carbonyl group and a salt bridge between the side chains of E340 and R821 formed in SERCA2b WT.

Structural comparison between WT and the three classes of T1032stop demonstrates that the cytosolic domains of classes 1, 2, and 3 get gradually farther apart from those of SERCA2b WT in this order. When the TM7 to TM10 helices are aligned in a conventional way (15, 16), the A, N, and P domains in class 1 rotate by 3.9°, 4.0°, and 4.0° from the corresponding domains in WT, respectively, to get more remote from the membrane (Fig. 4A). The rotation angles of the A, N, and P domains increase to 7.6°, 7.6°, and 7.5° in class 2 and to 10.4°, 10.0°, and 9.6° in class 3, respectively (Fig. 4A).

Superposition of SERCA2b WT and class 3 of T1032stop fitted with the TM7 to TM10 helices indicates that while L6/7 in class 3 approaches closely to L8/9, the P domain was situated even more upright relative to the membrane than that in WT (Fig. 4B, right inset), suggesting that the locations of L6/7 and L8/9 are linked to the orientations of the cytosolic domains. In more detail, deletion of the LE relocated L813 in L6/7 closer to R923 in L8/9 to make van der Waals contact between these two (Fig. 4C, left). The concomitant approach of L6/7 to L8/9 appears to influence the conformation of a triad constituted by R821 in L6/7, E340, and C344 in the P domain (Fig. 4C, right), accompanying the slight but significant repositioning of the intermediary residues (N754, D812, D814, K818, and Q919) (Fig. 4C, middle). Whereas this triad is highly conserved in P-type ATPases (17) and supposed to link the P domain to L6/7 and the cytoplasmic part of TM3 through hydrogen bonds (18), the side chain of E340 is significantly more apart from those of R821 and C344 in class 1 of T1032stop (Fig. 4C, right, and fig. S6A). In addition, P337 in the P domain and T247 in TM3 are markedly repositioned (Fig. 4C, right, and fig. S6A). These local conformational changes at the triad and its neighboring region allow the P domain to move away from the original position (Fig. 4B, right inset). Resultantly, the A and N domains undergo the upright rotation (Fig. 4A), as the P domain provides a flat surface for tight interactions with the A domain, and the bound AMPPCP molecule serves as a bridge that forces the N domain to move in accordance with the P domain. The triad thus altered likely allowed generation of different classes of conformations in T1032stop.

Notably, the overall domain arrangement of class 3 of T1032stop is highly superimposable to that of SERCA1a and SERCA2a (Fig. 4A and fig. S7, A and B); the RMSD values for Cα atoms between class 3 of T1032stop and SERCA1a/SERCA2a are 0.602 and 0.580 Å, significantly smaller than that between class 3 of T1032stop and SERCA2b WT (1.678 Å). Like class 3 of T1032stop, SERCA1a retains a hydrogen bond between R822 (R821 in SERCA2b) and Cys344, and its E340 residue is situated well far from these two residues to form a hydrogen bond with T247 (18), rendering the cytoplasmic domain cluster more upright relative to the membrane. In this context, a mutation of E340 in SERCA1a markedly slowed the Ca2+ binding transitions from E2 to E1∙2Ca2+ and the backward Ca2+ dissociation toward the cytoplasm (19), suggesting a correlation between the cytosolic domain arrangement and the kinetic behavior of Ca2+ binding/release in the SERCA family.

Deletion of the LE also altered the orientation of TM11 (Fig. 4B, left inset), which likely influenced the positions of the N-terminal end of TM11 and, hence, the locations of R923 and L8/9 (Fig. 4C, left), as described above. In this connection, the hydrogen bond formed between the R923 side chain and the G1013 main-chain carbonyl group in WT appears to be broken in T1032stop (Fig. 4C, left). Collectively, the LE regulates the location and orientation of both the cytosolic and TM domains in the E1∙2Ca2+-ATP state, serving to stabilize an overall conformation unique to SERCA2b.

Roles of the LE in the structural regulation of the E2P form of SERCA2b

To explore the regulatory role of the LE in the E2P form of SERCA2b, we also determined and compared the cryo-EM structures of both SERCA2b WT and T1032stop bound to BeF3. The LE in this form was located at a similar position to that in the E1∙2Ca2+-AMPPCP form (Fig. 2 and fig. S5). As observed with the E1∙2Ca2+-AMPPCP state, deletion of the LE in the E2-BeF3 state generated multiple classes (classes 1 and 2) with significantly different conformations present in almost equal abundance (class 1 = 43.4% and class 2 = 43.2%) (Fig. 5A and fig. S9). While neither class included Ca2+ at the Ca2+ binding sites, both showed significant density of bound BeF3 and its neighboring Mg2+ in the cryo-EM maps (fig. S4, G and H), indicating that generation of the two different classes cannot be ascribed to imperfect or heterogeneous preparation of the E2-BeF3 form.

Fig. 5 Structural effects of the LE on SERCA2b in the E2P state.

(A) Cryo-EM structures of two different classes (classes 1 and 2) of SERCA2b T1032stop in the E2-BeF3 state. The rotation angles of the cytosolic A, N, and P domains as compared with those of class 1 are indicated by green arrows showing the rotation direction. Classes 1 and 2 are colored pale cyan and yellow, respectively. The movements of TM4 (purple) and TM5 (blue) between these two classes are highlighted by cartoons on the right of ribbon diagrams. (B) Cryo-EM structures of two different classes (classes 1 and 2) of SERCA2b T1032stop in the E2-BeF3 state are superimposed such that the RMSD of Cα atoms in TM7 to TM10 is minimized. Classes 1 and 2 are colored as in (A). The left upper inset presents a close-up view of TM5 to TM11, in which TM helix movements during the conversion from class 1 to class 2 are indicated by arrows. The left lower inset highlights the switch residues involved in the interconversion between class 1 and class 2. The right upper and lower insets show close-up views of the movements of TM1 and TM2 and TM3 and TM4 during the conversion from class 1 to class 2. Note that the similar domain rearrangements take place during the transition from the E2P to E2 states in SERCA1a.

Class 1 of SERCA2b T1032stop retained almost the same arrangement for both the cytosolic and TM helix domains as those in SERCA2b WT (fig. S7D). On the other hand, the cytosolic A, N, and P domains in class 2 were tilted down toward the membrane by 9.7°, 10.6°, and 9.6° as compared with the corresponding domains in class 1, respectively, when the TM7 to TM10 helices were aligned to each other (Fig. 5, A and B). In addition to the cytosolic domain movement, class 2 underwent significant TM helix relocation; compared with those of class 1, TM7, TM10, and TM11 of class 2 were shifted laterally to be further apart from the other TM helices (Fig. 5B, left inset), indicating a role of the LE in fine-tuning the TM helix arrangement in the E2P state. Notably, TM4, which serves to control the calcium release gate (13), was significantly bent to open the gate in class 1. By contrast, the luminal half of TM4 in class 2 was positioned more linearly with its cytosolic half, resulting in the closed gate (Fig. 5, A and B), as seen in the E2-Pi and E2 states of SERCA1a (7, 13, 15).

It is noteworthy that Y836 in TM7 and R761 in TM5 adopt notably different side-chain rotamers between class 1 and class 2 (Fig. 5B, left lower inset and fig. S6B). Consequently, the angle of the cytosolic half of TM5 is altered (Fig. 5B, left upper inset), leading to significant inclination of the P domain toward the membrane. In this regard, the highly conserved GxxxG motif, an element previously proposed to trigger the bending of TM5 in SERCA1a (7, 20), is located at the middle of TM7 in SERCA2b (Fig. 5B, left lower inset). The inclination of the P domain induces a rotation of the A domain due to their close contact, resulting in relocation of TM1 and TM2 (Fig. 5B, right upper inset) and a subsequent change in orientation of TM3 and TM4 (Fig. 5B, right lower inset). Collectively, we propose that R761 and Y836 act as switch residues that mediate conformational interconversion between class 1 and class 2 via the rotamer changes.

A similar switching mechanism was reported to operate during the transition from E2P to E2 states in SERCA1a (15, 21). Hydrolysis of an aspartylphosphate at D351 allows the side chains of R762 and Y837 to adopt different rotamers (fig. S6C) and the subsequent rotation of the P domain in SERCA1a, which relocates TM5 closer to TM4 (fig. S6D). As a result, TM3 and TM4 undergo a significant positional shift to close the calcium exit channel (fig. S6D). Concomitantly, significant reorientations of TM1 and TM2 and rotation of the A domain are induced in SERCA1a due to the steric effects (fig. S6E), eventually leading to the transition to the E2 state.

DISCUSSION

Previously, TM11 of SERCA2b was reported to stabilize the E2/E2P states by interacting with TM7 and TM10 (10). The exogenous addition of a synthetic TM11 peptide to SERCA1a resulted in a lower Vmax value and higher apparent Ca2+ affinity, suggesting that TM11 acts as an intramolecular uncompetitive inhibitor for SERCA2b (22, 23). Kinetics analysis with various SERCA2b derivatives including T1032stop suggested that whereas TM11 plays a regulatory role in the E2P dephosphorylation and the transition from E2 to E1, the LE serves to retard the E1P-to-E2P transition by stabilizing the E1 state via interaction with L7/8 (23). Thus, distinct and independent regulatory roles of TM11 and the LE were proposed.

We herein demonstrated that the LE is located outside the luminal part of TM7 and approaches a short α helix segment of L7/8 (Figs. 1C and 2A), generating an overall domain arrangement that differs from SERCA1a and SERCA2a in the E1∙2Ca2+-AMPPCP state. In this context, deletion of the LE rendered the overall conformation of SERCA2b similar to those of SERCA1a and SERCA2a (fig. S7, A and B), while generating multiple conformations. It is thus conceivable that the LE of SERCA2b physically stabilizes the E1∙2Ca2+-ATP conformation unique to this isoform and thereby regulates the conformational transition to the E2P state (Fig. 6). This notion is consistent with the previous observation that the rate of the “E1∙2Ca2+-P–to–E2P” transition of SERCA2b T1032stop was significantly faster than that of SERCA2b WT (23). Although it remains to be fully elucidated how the LE remotely modulates the orientation and location of the cytosolic domains, the overall conformation characteristic of SERCA2b appears to be stabilized most likely with specific interactions formed between R821 (L6/7) and E340 (TM5) and between R923 (L8/9) and G1013 (TM11) (Fig. 4, B and C).

Fig. 6 Proposed mechanisms of SERCA2b regulation by the LE.

Structures of SERCA2b WT in the E1∙2Ca2+-ATP and E2P states are stabilized through interactions of the LE with part of L7/8 and other neighboring regions, leading to the regulated transition between these two states. Upon deletion of the LE, SERCA2b undergoes a significant relocation of both the cytosolic and TM domains accompanied by the generation of multiple overall conformations, allowing the unregulated transition from the E1∙2Ca2+-ATP to E2P states.

Multiple conformations in similar abundance were also observed for SERCA2b T1032stop in the E2-BeF3 state. Thus, the LE likely plays a significant role in stabilizing the Ca2+-unbound form of SERCA2b, as well as the Ca2+-bound form, and may thereby modulate the kinetic properties of SERCA2b. However, the rates of the dephosphorylation of E2P and the conformational transition of the E2 to E1 states of SERCA2b were hardly affected by deletion of the LE and considerably slower than those of SERCA1a and SERCA2a (23). It is thus suggested that the LE-mediated conformational regulation is not closely involved in the transitions beginning from the E2P state and that the late stages of the SERCA2b catalytic cycle depend critically on TM11, as previously proposed (22, 23). Nevertheless, the overall turnover rate of SERCA2b T1032stop is comparable to that of SERCA2a (24), which may be consistent with the present observations that deletion of the LE yielded significant portion of SERCA1a- or SERCA2a-like conformations in both E1∙2Ca2+-AMPPCP and E2-BeF3 states (Figs. 4 and 5, and fig. S7, A to E).

Although further studies are required to fully elucidate how the LE controls each step of conformational transitions in SERCA2b throughout the catalytic cycle, the present structural data in combination with previous biochemical and biophysical evidence suggest a possible link between the LE-mediated conformational stabilization in SERCA2b and the different kinetic properties of the enzyme from other isoforms lacking the LE. Time-resolved x-ray solution scattering experiments combined with molecular dynamic simulation recently accomplished for SERCA1a (25) will reveal yet unidentified mechanisms of action of SERCA2b. In conclusion, the regulatory role of the LE revealed in this study provides a framework for understanding the mechanism underlying the ER Ca2+ homeostasis ensured by the regulated cytosol-to-ER Ca2+ transport by SERCA2b.

MATERIALS AND METHODS

Expression of SERCA2b and SERCA2b T1032stop

To construct a plasmid for overexpression of SERCA2b T1032stop in human embryonic kidney (HEK) 293T cells, a stop codon was inserted after Thr1032 using a QuickChange site-directed mutagenesis kit with primers 5′-gatctgggtctatagctgagacactaactttag-3′and 5′- ctaaagttagtgtctcagctatagacccagatc-3′. Plasmid pcDNA3.1/PA-SERCA2bWT constructed in a previous study (11) served as the template.

Expression of WT SERCA2b and SERCA2b T1032stop was carried out as previously described (11). In brief, a PiggyBac Cumate Switch Inducible Vector harboring an open reading frame encoding human SERCA2b WT or T1032stop was introduced into HEK293T cells along with Super PiggyBac Transposase Expression Vector (System Bioscience, LLC, CA, USA) using polyethylenimine (Sigma-Aldrich) to generate a stable cell line for overexpression of each protein. Cells were grown in Dulbecco’s modified Eagle’s medium with 4% inactivated fetal calf serum and incubated in a humidified incubator with 5% CO2 at 37°C. After 2 days of incubation, expression of SERCA2b WT or T1032stop was induced with cumate (150 μg/ml) and phorbol 12-myristate 13-acetate (50 ng/ml). Cells were incubated at 37°C for another 48 hours and then harvested by centrifugation at 1000g for 15 min.

Purification of SERCA2b and SERCA2b T1032stop

All purification steps were performed at 4°C. Harvested cells were lysed using a Dounce homogenizer and solubilized in a buffer containing 50 mM Hepes-NaOH (pH 7.0), 100 mM NaCl, 20% glycerol, 1 mM CaCl2, 1 mM MgCl2, 1 mM dithiothreitol (DTT), 1 mM phenylmethylsulfonyl fluoride, and 1/100 Protease Inhibitor Cocktail (Nacalai). After homogenization, 1% (w/v) DDM (n-dodecyl-b-D-maltoside) was added to solubilize the membrane fraction. After 2 hours of gentle rotation at 4°C, the sample was centrifuged at 14,500 rpm for 1.5 hours to remove insoluble material. The supernatant was collected and incubated overnight with anti–PA Sepharose beads. Beads were washed with 10 column volumes of buffer containing 50 mM Hepes-NaOH (pH 7.0), 100 mM KCl, 20% glycerol, 1 mM CaCl2, 1 mM MgCl2, 1 mM DTT, and 0.05% (w/v) LMNG, followed by elution with the same buffer containing PA peptide (0.2 mg/ml). Eluted fractions were concentrated and then further purified by size exclusion chromatography (SEC) on a Superose 6 10/300 GL column (GE Healthcare) at 4°C with buffer containing 50 mM Hepes-NaOH (pH 7.0), 100 mM KCl, 20% glycerol, 1 mM CaCl2, 1 mM MgCl2, 1 mM DTT, and 0.05% (w/v) LMNG. Peak fractions were pooled and concentrated to 10 mg/ml with an Amicon filter device equipped with a 100-kDa cutoff membrane. Purification of SERCA2b T1032stop was performed as described for the WT protein. Representative results for SEC and SDS–polyacrylamide gel electrophoresis are shown in fig. S1.

Grid preparation for cryo-EM

To prepare the E1∙2Ca2+-AMPPCP form, 1 mM AMPPCP was added to purified SERCA2b WT or T1032stop. After overnight incubation, a second round of SEC was performed on the same column at 4°C with 50 mM Hepes-NaOH (pH 7.0), 100 mM KCl, 0% glycerol, 1 mM CaCl2, 1 mM MgCl2, 1 mM DTT, and 0.01% (w/v) LMNG to remove glycerol and decrease the detergent concentration. Preparation of the E2-BeF3 form was performed by treating the sample with 5 mM BeSO4, 10 mM NaF, and 1 mM EGTA overnight. All samples were concentrated to 4 to 8 mg/ml for cryo-EM measurement. Purified samples (2 to 3 μl) were applied to a freshly glow-discharged 300 mesh Quantifoil 1.2/1.3 carbon grid using a Vitrobot Mark IV instrument (FEI) with a blotting time of 3 to 4 s under 100% humidity at 4°C, and grids were plunge frozen in liquid ethane.

Cryo-EM data collection and image processing

Prepared grids were transferred to a Titan Krios G3i microscope (Thermo Fisher Scientific), which was operated at 300 kV and equipped with a Gatan Quantum-LS Energy Filter (GIF) and a Gatan K3 Summit direct electron detector in the electron counting mode. Imaging was performed at a nominal magnification of ×105,000, corresponding to a calibrated pixel size of 0.83 Å per pixel (University of Tokyo, Japan). Each movie was recorded for 3 s and subdivided into 60 frames. The electron flux was set to 14 e/pixel/s at the detector, resulting in an accumulated exposure of 60 e2 at the specimen. Data were automatically acquired by the image shift method using SerialEM software, with a defocus range of −0.8 to −1.8 μm. More than 4000 movies were acquired for each condition grid, and the total number of images is described in table S1. For all datasets, dose-fractionated movies were subjected to beam-induced motion correction using RELION-3, and contrast transfer function parameters were estimated using CTFFIND4.

All datasets were processed using the same workflow except for 3D classification after 3D refinement. Particles were extracted from motion-corrected micrographs with downsampling to a pixel size of 3.32 Å per pixel. These particles were subjected to one round of 2D classification and two rounds of 3D classification. The best particles were selected, reextracted with a pixel size of 1.245 Å per pixel, and subjected to auto-3D refinement. The resulting 3D maps and particle sets were subjected to per-particle defocus refinement, beam-tilt refinement, Bayesian polishing, second per-particle defocus refinement, and 3D refinement. The global resolution of SERCA2b WT in the E1∙2Ca2+-AMPPCP and BeF3 states was 2.9 and 2.8 Å, respectively. For the SERCA2b WT E1∙2Ca2+-AMPPCP state dataset, we performed additional no-align 3D classification using a mask covering TM11 and the LE. Particles from the best class were subjected to final 3D refinement. For SERCA2b T1032stop datasets, additional local search 3D classification was performed with masks to remove micelle density. Good classes resulting from this classification were individually subjected to 3D refinement. The global resolutions for SERCA2b T1032stop E1∙2Ca2+-AMPPCP state datasets were 2.8 Å (class 1), 2.9 Å (class 2), and 2.8 Å (class 3). The global resolutions for SERCA2b T1032stop BeF3 state datasets were 3.4 Å (class 1) and 3.1 Å (class 2). All resolutions were calculated according to the Fourier shell correlation (FSC) = 0.143 criterion. The local resolution was estimated using RELION. Detailed processing strategies are shown in figs. S1, S2, S8, and S9.

Model building, refinement, and validation

Cryo-EM structures of the E1∙2Ca2+-AMPPCP and E2-BeF3 forms of SERCA2b WT were modeled using crystal structures of human SERCA2b WT in the AMPPCP state [Protein Data Bank (PDB) ID: 5ZTF] and unpublished crystal structures of SERCA2b WT in the E2-BeF3 state as initial models, respectively. TM11 and the LE were built manually using Coot (26) based on the cryo-EM maps. After manual modification of the models, structure refinement was carried out using “phenix.real_space_refine” in PHENIX (27, 28). For the E1∙2Ca2+-AMPPCP form, linker regions composed of residues 81, 505 to 506, and 993 to 1011, which were located at L1/2, the N domain, and L10/11, respectively, could not be modeled due to missing density in the cryo-EM map. For the E2-BeF3 form, linker regions composed of residues 79 to 84, 502 to 507, 879 to 882, and 991 to 1014, located at L1/2, the N domain, L7/8, and L10/11, respectively, were not modeled due to missing density in the cryo-EM map.

Cryo-EM structures of the E1∙2Ca2+-AMPPCP and E2-BeF3 forms of SERCA2b T1032stop were built using the corresponding WT cryo-EM structures as initial models. After truncating the LE region and improving other regions of the models using Coot, structure refinement was carried out using “phenix real space refine” in PHENIX (27, 28). For the E1∙2Ca2+-AMPPCP form, linker regions composed of residues 993 to 1013 contained in L10/11 could not be modeled in any of the class 1, class 2, or class 3 structures due to missing electron density in the map. Similarly, for the E2-BeF3 form, linker regions composed of residues 78 to 85, 283 to 287, 373 to 377, 504 to 506, 876 to 877, and 991 to 1013, and residues 503 to 506, 876 to 881, and 991 to 1012, could not be modeled in either the class 1 or class 2 structures. The statistics for 3D reconstitution and model refinement are summarized in table S1.

SUPPLEMENTARY MATERIALS

Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/33/eabb0147/DC1

https://creativecommons.org/licenses/by-nc/4.0/

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.

REFERENCES AND NOTES

Acknowledgments: Funding: This work was supported by CREST, JST (grant number JPMJCR13M6 to K.I. and K.N., and grant number JPMJCR14M1 to M.K.), Grants-in-Aid for Scientific Research on Innovative Areas from MEXT to K.I. (18H03978), and the Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS) from the Japan Agency for Medical Research and Development (AMED) under grant number JP19am0101115 (support number 1025). Author contributions: Y.Z. performed almost all the experiments, structure modeling, and structure refinement. M.I. established a large-scale expression and purification system for recombinant human SERCA2b. A.T. performed acquisition and processing of cryo-EM image data. S.W. and K.I. assisted in structure modeling and refinement. T.N. assisted in data processing of cryo-EM images. K.I. supervised this work. Y.Z. and K.I. prepared figures and wrote the manuscript. Y.Z., A.T., S.W., T.N., K.N., M.K., and K.I. discussed the results and critically read the manuscript. All authors approved the manuscript for submission. Competing interests: The authors declare that they have no competing interests. Data availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data supporting the findings of this manuscript are available from the corresponding authors upon reasonable request. Atomic coordinates of human SERCA2b WT in the E1∙2Ca2+-AMPPCP and E2-BeF3 states have been deposited in the Protein Data Bank under accession codes 6LLE and 6LLY, respectively. Those of three classes of human SERCA2b T1032stop in the E1∙2Ca2+-AMPPCP state have been deposited in the Protein Data Bank under accession codes 6LN5 (class 1), 6LN6 (class 2), and 6LN7 (class 3). Those of two classes of human SERCA2b T1032stop in the E2-BeF3 state have been deposited in the Protein Data Bank under accession codes 6LN8 (class 1) and 6LN9 (class 2). Cryo-EM density maps were deposited in the Electron Microscopy Data Bank under accession codes EMD-0912 (SERCA2b WT in the E1∙2Ca2+-AMPPCP state), EMD-0915 (SERCA2b WT in the E2-BeF3 state), EMD-0924 (class 1 of SERCA2b T1032stop in the E1∙2Ca2+-AMPPCP state), EMD-0925 (class 2 of SERCA2b T1032stop in the E1∙2Ca2+-AMPPCP state), EMD-0926 (class 3 of SERCA2b T1032stop in the E1∙2Ca2+-AMPPCP state), EMD-0927 (class 1 of SERCA2b T1032stop in the E2-BeF3 state), and EMD-0928 (class 2 of SERCA2b T1032stop in the E2-BeF3 state).
View Abstract

Stay Connected to Science Advances

Navigate This Article