FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4200006038> ?p ?o ?g. }

• W4200006038 abstract "Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract The dimeric ER Ca2+ sensor STIM1 controls store-operated Ca2+ entry (SOCE) through the regulated binding of its CRAC activation domain (CAD) to Orai channels in the plasma membrane. In resting cells, the STIM1 CC1 domain interacts with CAD to suppress SOCE, but the structural basis of this interaction is unclear. Using single-molecule Förster resonance energy transfer (smFRET) and protein crosslinking approaches, we show that CC1 interacts dynamically with CAD in a domain-swapped configuration with an orientation predicted to sequester its Orai-binding region adjacent to the ER membrane. Following ER Ca2+ depletion and release from CAD, cysteine crosslinking indicates that the two CC1 domains become closely paired along their entire length in the active Orai-bound state. These findings provide a structural basis for the dual roles of CC1: sequestering CAD to suppress SOCE in resting cells and propelling it toward the plasma membrane to activate Orai and SOCE after store depletion. Editor's evaluation This study uses complementary approaches to advance our mechanistic understanding of STIM1 activation, with elegant single molecule methods providing new details on STIM1 structure and dynamics. The data clarifies some of the controversy between domain packing in two differing X-ray and NMR structures and substantially contributes to a mechanistic and structural understanding of the STIM1 activation process. https://doi.org/10.7554/eLife.66194.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Store-operated Ca2+ entry (SOCE) is a nearly ubiquitous signaling pathway activated by extracellular stimuli that deplete Ca2+ from the endoplasmic reticulum (ER). SOCE is essential for diverse physiological functions of excitable and non-excitable cells, including gene expression, secretion, and motility (Prakriya and Lewis, 2015). Accordingly, loss-of-function mutations in the core SOCE components lead to serious human pathologies such as severe combined immunodeficiency, autoimmunity, myopathy, ectodermal dysplasia, and anhidrosis (Lacruz and Feske, 2015), whereas gain-of-function mutations cause Stormorken syndrome, miosis, myopathy, thrombocytopenia, and excessive bleeding (Böhm and Laporte, 2018). These clinical manifestations underscore the need for precise regulation of SOCE to ensure that it is silent when ER Ca2+ stores are full yet reliably activated by stimuli that drive store depletion. The dimeric ER membrane protein STIM1 (Figure 1A, top) controls SOCE through the regulated interaction of its CRAC activation domain (CAD [Park et al., 2009], also known as SOAR [Yuan et al., 2009] or CCb9 [Kawasaki et al., 2009]) with Orai1 Ca2+ channels in the plasma membrane. In resting cells, Ca2+ bound to the luminal EF hands of the STIM1 dimer suppresses STIM1 activity by promoting intramolecular association of the coiled-coil 1 (CC1) domains with CAD, referred to as the ‘inhibitory clamp’ (Korzeniowski et al., 2010; Yang et al., 2012; Yu et al., 2013; Fahrner et al., 2014; Ma et al., 2015). Upon ER Ca2+ depletion, dissociation of Ca2+ from the luminal EF hands triggers conformational changes that release CC1 from CAD and promote STIM1 accumulation at ER-plasma membrane (ER-PM) junctions where the CAD domain binds, traps, and activates Orai1 channels diffusing in the PM (Wu et al., 2006; Wu et al., 2014; Liou et al., 2007; Park et al., 2009; Yuan et al., 2009). Thus, to elucidate the regulatory mechanism of SOCE, it is essential to understand how CC1 interacts with CAD to control STIM1 activity. Figure 1 with 3 supplements see all Download asset Open asset Parallel orientation of CC2 domains in ctSTIM1 is consistent with the CAD crystal structure. (A) (top) Schematic overview of domains in ctSTIM1, comprising the membrane-proximal coiled-coil 1 (CC1) domain with helical regions α1 (aa 238–271), α2 (aa 278–304), and α3 (aa 308–337), and the CC2 (345–391) and CC3 (aa 408–437) domains which are part of CAD. (bottom) The orientations of the CC2 domains distinguish the NMR structure of CC1α3-CC2 (bottom left, anti-parallel CC2 domains; 2MAJ.pdb) from the crystal structure of CAD (bottom right, parallel CC2 domains; 3TEQ.pdb). Sites tested by inter-subunit smFRET are indicated. Although aa 337 is not part of the crystal structure (aa 344–443), it is sufficiently close to permit comparison to the NMR structure. (B) Two methods for immobilizing single ctSTIM1 dimers for smFRET. ctSTIM1 peptides were encapsulated in liposomes and immobilized on a glass slide by a biotin-neutravidin interaction or were attached directly via a C-terminal avitag. (C) Example recordings of donor (green) and acceptor (red) fluorescence evoked by donor excitation for sites 337:337´, 363:363´, and 378:378´. Single-step bleaching events indicate a single molecule, while the anti-correlated response of the donor to the acceptor bleach event is a hallmark of smFRET. Calculated FRET ratios (black) are shown below. (D) Ensemble smFRET histograms for sites 337:337´, 363:363´, and 378:378´ displayed predominantly high FRET, indicating close parallel apposition. (E) A comparison of smFRET-derived distances (black) with simulated inter-fluorophore distances from the crystal (white) and NMR structures (gray). The measured distances closely match the crystal structure but deviate strongly from the NMR structure, particularly for 337:337´ and 378:378´. smFRET error bars indicate the expected distance deviation corresponding to an uncertainty of ±0.05 in FRET measurement. 'Crystal' and 'NMR' error bars indicate the s.e.m. of inter-fluorophore distance from 100 simulations of dye position (see Materials and methods). The CC1 domain comprises three helical segments, CC1α1–3 (Soboloff et al., 2012; Figure 1A). Current evidence supports a model in which CC1α1, the most ER-proximal segment, acts as a bimodal switch to stabilize both the quiescent and active states of STIM1. According to this model, in cells with full ER Ca2+ stores, CC1α1 binds to the CAD CC3 helix to sequester CAD (Fahrner et al., 2014; Ma et al., 2015); upon store depletion, the luminal EF-SAM domains dimerize, CC1α1 unbinds from CAD, and the transmembrane and proximal CC1α1 regions form a coiled-coil (Stathopulos et al., 2006; Hirve et al., 2018; Schober et al., 2019), projecting CAD toward the plasma membrane. Although several residues in CC1α1 and CC3 have been identified as essential for forming the inhibitory clamp (Muik et al., 2011; Zhou et al., 2013; Fahrner et al., 2014; Ma et al., 2015), the helical arrangement of the CC1α1:CAD binding interface is completely unknown. The CC1α3 domain is also implicated in stabilizing the inactive state (Korzeniowski et al., 2010; Yang et al., 2012), but it is unclear whether this involves binding to CAD or another mechanism (Zhou et al., 2013). Attempts to determine the structure of full-length STIM1 in the inactive or active state have been unsuccessful. Instead, structures of several fragments from the STIM1 cytosolic region have offered potential clues, although it is not yet clear how they might relate to physiological states of the full-length protein. The crystal structure of human STIM1 CAD (aa 345–444, with mutations L374M, V419A, and C437T) depicts a parallel V-shaped dimer comprising the CC2 and CC3 domains (Figure 1A, right; Yang et al., 2012). It is unknown whether this structure represents an active or inactive conformation, and the absence of CC1 precludes inferences about its interaction with CAD. A solution NMR structure of a CC1α3-CC2 fragment depicts a dimer of wedged antiparallel CC2 helices (Figure 1A, bottom left; Stathopulos et al., 2013). This NMR structure can bind Orai1 C-terminal fragments in vitro and has been proposed to be an intermediate in the activation of STIM1 (Stathopulos et al., 2013), but it has never been shown to bind intact Orai1 channels in situ and lacks the CC3 helix needed to evoke SOCE (Covington et al., 2010). Finally, two contrasting structures have been reported for the conformation of the CC1 region. In a crystal structure of CC1 (aa 237–340, with mutations M244L and L321M), two elongated CC1 helices form an antiparallel coiled-coil dimer within the CC1α2 and CC1α3 domains (Cui et al., 2013; see Figure 5C ), but it is unclear how this elongated structure with its widely separated C-termini could attach to the closely paired N-termini of CAD. In contrast, a recent solution NMR study of CC1 (aa 234–343) depicts a compactly folded monomeric structure in which CC1α1-α3 form a tight three-helix bundle (Rathner et al., 2021). Although the sequences of these various structures partially overlap, their conformations are incompatible with each other, raising fundamental questions about their potential relation to the inactive and active forms of STIM1. What is the conformation of CAD in the context of the entire cytosolic domain? How is CC1 arranged relative to CAD in the inactive state, and how does this change during activation? Which regions of STIM1 are stable, and which are flexible? To address these questions, we examined the structure and dynamics of the full cytosolic domain of STIM1 (ctSTIM1, aa 233–685) using single-molecule Förster resonance energy transfer (smFRET) measurements at sites throughout CC1 and CAD. smFRET is commonly applied to estimate intramolecular distances and reveal asynchronous molecular transitions that are typically obscured in population measurements. Our results show that CAD in ctSTIM1 resembles the CAD crystal structure, but with unexpected structure and flexibility in an apical region that is critical for Orai binding and activation. CC1α1 interacts with CAD in a domain-swapped configuration to form the inhibitory clamp, while the Stormorken syndrome mutation R304W releases CC1α1 from CAD to activate ctSTIM1. In studies of full-length STIM1 in intact cells, cysteine crosslinking after store depletion suggests that after unbinding from CAD, the two CC1 domains pack together along their entire length, in contrast to the CC1 crystal and CC1α3-CC2 NMR structures. Together, these findings offer the first structural view of the STIM1 inhibitory clamp and reveal the massive conformational changes evoked by store depletion, which serve to reorient and translocate CAD towards the PM where it can bind Orai1 and activate SOCE. Results The CC2 domains in ctSTIM1 are oriented in a parallel configuration We initially applied smFRET to test whether the conformation of CAD within the cytosolic domain (ctSTIM1, aa 233–685) resembles that of the CC2-CC3 crystal structure (Yang et al., 2012) or the CC1α3-CC2 NMR structure (Stathopulos et al., 2013), referred to below as the ‘crystal structure’ or ‘NMR structure’ respectively. A major distinction between the two structures is that the CC2 domains (aa 345–391) in CAD are parallel in the crystal structure but antiparallel in the NMR structure (Figure 1A). After removing the single native cysteine in ctSTIM1 with a C437S mutation, we created three mutants with cysteines at positions 337, 363, and 378, for which the two structures predicted very different inter-subunit distances, particularly for 337:337´ and 378:378´ (hereafter ‘:’ is used to denote a pair of sites in the dimer and ‘´’ indicates a site on the adjacent subunit) (Figure 1A). The ctSTIM1 mutants were expressed in E. coli and purified, yielding full-length protein of >90% purity (see Materials and methods). After randomly labeling the cysteine-mutant ctSTIM1 molecules with FRET donor (Alexa Fluor 555) and acceptor (Alexa Fluor 647) dyes, single ctSTIM1 dimers were surface-tethered to a glass coverslip for imaging by TIRF microscopy, either encapsulated in liposomes or attached directly through a tag on the C-terminus (Figure 1B). In several instances we tested, both attachment methods yielded similar results (Figure 1—figure supplements 1–3); liposome encapsulation was used for all measurements except intramolecular FRET measurements in Figure 2 and 4 (see Materials and methods, ‘Sample preparation’ for more details). Upon excitation of the donor dye, the donor and acceptor emissions of individual molecules were ratioed to calculate FRET efficiencies, which were high and stable at all three label sites (Figure 1C and D). These results are consistent with a CAD conformation in which the CC2 domains are closely apposed and parallel, as in the crystal structure. Inter-fluorophore distances calculated from simulations of the donor and acceptor dye positions on the crystal structure closely matched those derived from smFRET, while simulations based on the NMR structure deviated substantially (Figure 1E). We conclude that in the context of the complete ctSTIM1 in solution, the CC2 domains in CAD adopt a parallel conformation consistent with the CAD crystal structure and not the NMR structure. Figure 2 with 1 supplement see all Download asset Open asset The apex of CAD in ctSTIM1 deviates from the CAD crystal structure. (A) Representative smFRET recording at the base of CAD (431:431´) showing stable, high FRET. (B) Ensemble density plots of the initial 2 s of inter-subunit smFRET recordings for sites throughout CAD. Predominant FRET levels are shown by arrowheads to the right of each plot. (C–D) Comparison of inter-fluorophore distances from smFRET and the CAD crystal structure. Simulated fluorophores are represented by the average position of a central atom, indicated here as ball-on-stick models (see Materials and methods). For sites near the dimer interface (the 'base' of CAD, C), the smFRET-derived distances (black bars) agree closely with the crystal structure (white bars). Measurements in the apex (D) deviate from the crystal structure. (E) Modified structure (red) generated by a smFRET-constrained optimization of the crystal structure (gray). Outward rotation of the distal CC2 region (aa 379–391) around G379 as a pivot point restored close correspondence between smFRET-derived distances and the crystal structure (bottom). (F) Representative example of smFRET fluctuations in the CAD apex (388:388´). The apex of CAD deviates from the crystal structure We made additional smFRET measurements throughout the entire CAD to allow a more complete comparison with the crystal structure. Near the C-terminal end of the CC3 domain (aa 408–437), the 431:431´ label pair produced a stable, narrowly defined high-FRET state (Figure 2A), and inter-subunit smFRET throughout the CAD exceeded 0.5, consistent with a compact parallel dimer (Figure 2B). Near the dimer interface (hereafter termed the ‘base’ of CAD), smFRET-derived distances quantitatively agreed with those from the crystal structure, and an intra-subunit measurement 431:363 directly confirmed the compact folding of CC3 onto CC2 within each subunit (Figure 2C). In contrast to the base of CAD, smFRET measurements of the distal CC2 and CC2-CC3 linker regions (the CAD ‘apex’) deviated from those indicated in the crystal structure (Figure 2D). This is particularly interesting as mutagenesis studies have identified this region as critical for Orai1 binding and activation (Calloway et al., 2010; Korzeniowski et al., 2010; Wang et al., 2014; Thompson et al., 2018; Butorac et al., 2019). To provide a rough indication of the extent of the deviation implied by smFRET, we generated an alternative model of the CAD apex based on the crystal structure by allowing rotation around G379 to bring the simulated FRET efficiencies in line with our experimental results. In the resulting model the distal CC2 regions were rotated outward, while the regions around aa 400 were pulled closer together, effectively compacting the apical ‘V-shape’ (Figure 2E and see Materials and methods). Whereas inter-subunit smFRET at aa 399–401 appeared as a well-defined peak, FRET levels at the distal end of CC2 were broadly distributed and appeared multimodal, particularly at aa 388 (Figure 2B). Accordingly, smFRET measurements in this region displayed fluctuations (Figure 2F), suggesting conformational flexibility that was not expected from the uninterrupted rigid helix of the crystal structure. These differences presumably arose from stabilization of an alternate apical conformation in the crystal by hydrogen bonds between the apex and adjacent CAD subunits (Figure 2—figure supplement 1B). The smFRET-derived model in Figure 2E thus likely represents just one of a number of potential apical conformations sampled by spontaneous fluctuations of the distal CC2 domain. As a complementary structural test, we analyzed inter-subunit cysteine-cysteine (cys-cys) crosslinking upon oxidation by copper phenanthroline (CuP, see Materials and methods). Cysteines introduced throughout the CAD apex formed inter-subunit crosslinks, detected as ctSTIM1 dimers on non-reducing SDS-PAGE gels (Figure 2—figure supplement 1A). This result shows that opposing residues in the apical region at least transiently come within the ~6 Å range for cys-cys crosslinking (Qin et al., 2015), again inconsistent with the CAD crystal structure. Moreover, crosslinking efficiencies were similar for three adjacent apical cysteine pairs (399:399´, 400:400´, and 401:401´), suggesting significant rotational flexibility beyond the 100 ms temporal resolution of our smFRET recordings. Taken together, these data show that flexibility and spontaneous fluctuations cause the CAD apical region to deviate significantly from the crystal structure, unlike the stable basal region. CC1α1 docks parallel to CC3 in a domain-swapped configuration The CC1 domain controls the activation of STIM1 in vivo by releasing CAD for interaction with Orai1 (Zhou et al., 2013; Ma et al., 2015). Although it is clear that the membrane-proximal CC1α1 domain can interact with CAD to inhibit STIM1 (Ma et al., 2015), their relative orientations remain undefined. We detected high inter-subunit FRET efficiencies between labels at the N-termini of CC1α1 and CC3 (red histogram in Figure 3A, top), as well as between labels at their C-termini (blue histogram in Figure 3A, bottom), indicating that the CC1α1 domain (a continuous α-helix [Cui et al., 2013; Rathner et al., 2021]) is oriented parallel to CC3 in the predominant ctSTIM1 conformation. Figure 3 with 1 supplement see all Download asset Open asset The CC1α1 domain is near and parallel to the CC3 domain of the opposing CAD subunit. (A) Asymmetric inter-subunit smFRET measurements between CC1α1 and CC3 reveal a parallel orientation, with high FRET levels at 242:400´ (top, red) and 274:431´ (bottom, blue) indicating close apposition. The crystal structure of CAD is shown with schematic representation of CC1α1 domains. (B) The predominant smFRET level between CC1α1 and CC3 was consistently higher for inter-subunit measurements (red) than intra-subunit measurements (blue). This pattern indicates a predominant domain-swapped configuration in which CC1α1 docks close to the CC3 domain of its opposing subunit. (C) Inter- and intra-subunit recordings (242:431´ and 242:431, respectively) displayed brief smFRET transitions between two states indicated by red and black bars. In both cases, fluctuations occurred between the same two FRET levels, reflected in the main and sub cluster peaks of the ensemble FRET histogram below, suggesting the CC1α1 domains switched sides on CAD (see also Figure 3—figure supplement 1). (D) Schematic illustration of spontaneous conformational transitions of the CC1α1 domain. In the predominant conformation, CC1α1 is closely apposed and parallel to CC3 on its opposing subunit in a domain-swapped configuration ('trans'). Occasionally, CC1α1 switches to interact briefly with CC3 on its own subunit ('cis'). Rate constants were derived from dwell time analysis of smFRET traces (Figure 3—figure supplement 1E,F). To determine whether CC1α1 associates with the CAD of the same subunit or the adjacent one within the dimer, we compared inter-subunit and intra-subunit FRET efficiencies between labels at the N-terminus of CC1α1 and three sites in CAD (Figure 3B). The predominant inter-subunit FRET efficiencies were consistently higher than intra-subunit FRET efficiencies, demonstrating that in the predominant ctSTIM1 conformation, CC1α1 engages CAD in a domain-swapped configuration. We also observed spontaneous and reversible transitions from the predominant FRET state, in which FRET efficiencies fluctuated between the same two levels for both inter- and intra-subunit measurements (Figure 3C). Similar fluctuation behavior and transition kinetics were observed at all three CAD sites (Figure 3—figure supplement 1), suggesting a conformational transition in which the two CC1α1 domains briefly switch sides on CAD, probably using the same binding interface (Figure 3D). The CC1α3 domains are compactly folded and directed away from CAD In vivo, mutations or deletions of the CC1α3 domain activate STIM1 (Korzeniowski et al., 2010; Yang et al., 2012) but the underlying mechanism, in particular the proposal that CC1α3 interacts directly with CAD (Yang et al., 2012), has been questioned (Zhou et al., 2013). We measured predominantly high inter-subunit FRET efficiencies of ~0.9 between labels at the N- or C-terminal ends of the CC1α3 domains (aa 312 and 337, respectively), indicating that the two domains are closely apposed and parallel (Figure 4A). To determine their orientation relative to CAD, we measured intra-subunit FRET efficiencies between labels at either end of the CC1α3 domain to three sites in CC2. While FRET efficiency from the CC1α3 C-terminus to the proximal CC2 domain was high as expected (red histogram in Figure 4B, top), low FRET efficiencies were measured from the CC1α3 N-terminus, indicating that the N-terminus is directed away from CAD (Figure 4B, bottom). Figure 4 with 1 supplement see all Download asset Open asset The CC1α2 and CC1α3 domains are closely apposed and directed away from CAD. (A) High inter-subunit smFRET at the N- and C-termini of the CC1α3 domain (aa 312 and 337, respectively) indicates close parallel apposition of the two helices. The crystal structure of CAD is shown with a schematic depiction of CC1α2 and CC1α3 domains. (B) (top) High intra-subunit smFRET from aa 337 to sites on CC2 declined progressively with distance, while intra-subunit smFRET from aa 298 (near the CC1α2/α3 linker) to sites in CC2 was low (bottom), indicating that the CC1α2 and α3 domains are directed away from CAD. (C) Inter-subunit recordings near the CC1α3 N-terminus displayed brief transitions from the predominant high-FRET state to a low-FRET state. Ensemble histograms show similar main and sub cluster FRET states for neighboring sites aa 309 and 312. (D) Schematic depiction of spontaneous conformational transitions of the CC1α3 domains. CC1α3 domains are closely apposed (left) but occasionally transition to a open configuration (right) at rates derived from dwell time analysis of smFRET traces (Figure 4—figure supplement 1E,F). Stable high inter-subunit FRET efficiencies at the CC1α3 C-terminus (aa 337) were consistent with its proximity to the compact, stable CAD base. However, label sites near the N-terminus of CC1α3 displayed brief (0.8 s average lifetime), large transitions to a low FRET state, suggesting that the N-termini of CC1α3 intermittently splay out from a stable pivot point at the base of CAD (Figure 4C and D). While these large conformational transitions were most prominent, further analysis revealed additional substates in the smFRET trajectories (Figure 4—figure supplement 1). A model for the CC1-CAD complex Through the integration of many inter-molecular distance measurements, smFRET can be applied to build models of tertiary protein structure (Brunger et al., 2011). We used 36 unique inter- and intra-subunit smFRET measurements (Supplementary file 1) to develop a structural model of the resting conformation of CC1 subdomains relative to the optimized model of CAD depicted in Figure 2E. CC1 in the ctSTIM1 dimer was assumed to comprise three rigid α-helical sections (CC1α1-α3, comprising residues aa 246–271, 275–305, and 310–337, respectively) connected by flexible linkers, consistent with bioinformatic analysis (Soboloff et al., 2012), NMR measurements of the CC1 fragment (Rathner et al., 2021), and distances estimated from intra-subunit smFRET efficiencies (239:274, 274:307, and 307:337; Figure 1—figure supplement 2). We selected randomly generated models that met the criteria of satisfying smFRET-derived distance constraints while avoiding steric clashes with each other or with CAD (see Materials and methods and Figure 5—figure supplement 1A for details of the modeling process). Two classes of models emerged that were highly similar but distinguishable by the organization of the CC1α2/α3 domains. In ~70 % of solutions, the CC1α2/α3 domains were stacked parallel (average model in Figure 5A from an ensemble of models in Figure 5—figure supplement 1B), while the remaining solutions had an alternative, wedged CC1α2/α3 topology (Figure 5A and Figure 5—figure supplement 1C). Importantly, both classes of solutions displayed the defining features described above, including closely apposed CC1α3 domains directed away from CAD and CC1α1 domains oriented parallel to CC3 in a domain-swapped configuration. As an independent test of the smFRET-derived models, we applied the lysine crosslinking agent BS3 combined with mass spectrometry (XLMS) to identify proximal pairs of residues (Rappsilber, 2011). BS3 XLMS confirmed key structural arrangements of both models of the CC1-CAD complex, including the compact packing of CC1α2/α3 domains and the domain-swapped CC1α1-CAD interaction (Figure 5—figure supplement 2). Figure 5 with 3 supplements see all Download asset Open asset smFRET-derived models of the configuration of CC1 and CAD in ctSTIM1. (A) 36 smFRET-derived distances were used to reconstruct the orientations of CC1 domains relative to CAD, yielding two classes of solutions with the CC1α2/α3 domains in a ’stacked' (left) or 'wedged' (right) configuration. In both classes, CC1α1 domains are in close parallel apposition to the CC3 domains on the adjacent subunit in a domain-swapped configuration, with the CC1α2 and CC1α3 domains forming a compact parallel structure directed away from CAD. The average of 50 model solutions is shown for each class (see Figure 5—figure supplement 1 for individual solutions). The complete list of smFRET values and distance constraints used to generate the models is shown in Supplementary file 1. (B) smFRET-derived models suggest hydrophobic stabilization of the CC1α2/α3 complex by antiparallel apposition of CC1α2 and CC1α3´. (C) The crystal structure of CC1 peptides depicts an antiparallel interaction of CC1α2 and CC1α3´ domains (left, dashed box), in which hydrophobic residues form a tightly packed dimer interface (top right, adapted from 4O9B.pdb). (Bottom right) Tight antiparallel packing of CC1α2/α3´ in ctSTIM1 was confirmed by inter-subunit crosslinks with EDC (red lines) (see also Figure 5—figure supplement 3). (D) Parallel apposition of hydrophobic residues on CC1α1 and CC3´. Many of these were previously identified by mutagenesis to stabilize the inactive state of STIM1, including L248, L251, L258, L261, L416, V419, and L423. (E) A model of the hydrophobic CC1α1:CC3´ interface obtained by computational peptide docking (see Materials and methods). Although the models do not specify the orientation of side chains in the CC1 helices, they highlight two regions for potential inter-subunit coiled-coil interactions in the CC1-CAD complex. The antiparallel apposition of hydrophobic residues in CC1α2 and CC1α3´ (Figure 5B) is reminiscent of the tight, antiparallel interaction of CC1α2 and CC1α3´ domains in the CC1 dimer crystal structure (Cui et al., 2013), an arrangement that was further supported by the formation of amine-carboxy crosslinks in ctSTIM1 by 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC), a zero-length crosslinker (Figure 5C and Figure 5—figure supplement 3). Interestingly, the CC1α1:CC3´ interface in the smFRET-derived models juxtaposes multiple hydrophobic residues that have been previously implicated through mutagenesis in stabilizing the CC1-CAD complex: L248, L251, L258, and L261 in CC1α1 and L416, V419, and L423 in CC3 (Muik et al., 2011; Zhou et al., 2013; Fahrner et al., 2014; Ma et al., 2015; Figure 5D). To refine the CC1α1:CC3´ interface, we aligned the crystal structure of the CC1α1 domain (aa 246–271) to the smFRET-derived model and performed a computational docking simulation using Rosetta software (see Materials and methods). The resulting model (Figure 5E) indicates an extensive pairing of hydrophobic residues predicted to stabilize the CC1α1:CC3´ interface, providing a plausible explanation for the activating effects of disruptive mutations at these sites. The Stormorken mutation R304W activates ctSTIM1 by releasing CC1α1 from CAD The close packing of CC1α1 and CAD in our model suggests that ctSTIM1 is primarily inactive. Accordingly, when expressed with Orai1 in HEK293 cells, mCh-ctSTIM1 only slightly increased [Ca2+]i compared to mCh-CAD (Figure 6—figure supplement 1), in agreement with previous evidence that ctSTIM1 is a relatively weak activator of Orai1 (Muik et al., 2009; Park et al., 2009; Yuan et al., 2009; Korzeniowski et al., 2010). Deletion of the CC1 domain (mCh-ctSTIM1-ΔCC1) restored ctSTIM1 activity to the level seen with CAD, as did the introduction of R304W, a naturally occurring mutation that causes" @default.
• W4200006038 created "2021-12-31" @default.
• W4200006038 creator A5005991308 @default.
• W4200006038 creator A5008400031 @default.
• W4200006038 creator A5018168832 @default.
• W4200006038 creator A5019402916 @default.
• W4200006038 creator A5023990361 @default.
• W4200006038 creator A5051508940 @default.
• W4200006038 creator A5056884743 @default.
• W4200006038 creator A5079895950 @default.
• W4200006038 date "2021-10-12" @default.
• W4200006038 modified "2023-09-27" @default.
• W4200006038 title "Author response: Conformational dynamics of auto-inhibition in the ER calcium sensor STIM1" @default.
• W4200006038 doi "https://doi.org/10.7554/elife.66194.sa2" @default.
• W4200006038 hasPublicationYear "2021" @default.
• W4200006038 type Work @default.
• W4200006038 citedByCount "0" @default.
• W4200006038 crossrefType "peer-review" @default.
• W4200006038 hasAuthorship W4200006038A5005991308 @default.
• W4200006038 hasAuthorship W4200006038A5008400031 @default.
• W4200006038 hasAuthorship W4200006038A5018168832 @default.
• W4200006038 hasAuthorship W4200006038A5019402916 @default.
• W4200006038 hasAuthorship W4200006038A5023990361 @default.
• W4200006038 hasAuthorship W4200006038A5051508940 @default.
• W4200006038 hasAuthorship W4200006038A5056884743 @default.
• W4200006038 hasAuthorship W4200006038A5079895950 @default.
• W4200006038 hasBestOaLocation W42000060381 @default.
• W4200006038 hasConcept C121332964 @default.
• W4200006038 hasConcept C12554922 @default.
• W4200006038 hasConcept C145912823 @default.
• W4200006038 hasConcept C178790620 @default.
• W4200006038 hasConcept C185592680 @default.
• W4200006038 hasConcept C24890656 @default.
• W4200006038 hasConcept C41008148 @default.
• W4200006038 hasConcept C519063684 @default.
• W4200006038 hasConcept C86803240 @default.
• W4200006038 hasConceptScore W4200006038C121332964 @default.
• W4200006038 hasConceptScore W4200006038C12554922 @default.
• W4200006038 hasConceptScore W4200006038C145912823 @default.
• W4200006038 hasConceptScore W4200006038C178790620 @default.
• W4200006038 hasConceptScore W4200006038C185592680 @default.
• W4200006038 hasConceptScore W4200006038C24890656 @default.
• W4200006038 hasConceptScore W4200006038C41008148 @default.
• W4200006038 hasConceptScore W4200006038C519063684 @default.
• W4200006038 hasConceptScore W4200006038C86803240 @default.
• W4200006038 hasLocation W42000060381 @default.
• W4200006038 hasOpenAccess W4200006038 @default.
• W4200006038 hasPrimaryLocation W42000060381 @default.
• W4200006038 hasRelatedWork W1990058038 @default.
• W4200006038 hasRelatedWork W1993472122 @default.
• W4200006038 hasRelatedWork W2029549495 @default.
• W4200006038 hasRelatedWork W2036875053 @default.
• W4200006038 hasRelatedWork W2057121960 @default.
• W4200006038 hasRelatedWork W2062979101 @default.
• W4200006038 hasRelatedWork W2070228801 @default.
• W4200006038 hasRelatedWork W2748952813 @default.
• W4200006038 hasRelatedWork W2899084033 @default.
• W4200006038 hasRelatedWork W949016938 @default.
• W4200006038 isParatext "false" @default.
• W4200006038 isRetracted "false" @default.
• W4200006038 workType "peer-review" @default.