Orai channel C-terminal peptides are key modulators of STIM-Orai coupling and calcium signal generation

Abstract

Junctional coupling between endoplasmic reticulum (ER) Ca²⁺-sensor STIM proteins and plasma membrane (PM) Orai channels mediates Ca²⁺ signals in most cells. We reveal that PM-tethered, fluorescently tagged C-terminal M4x (fourth transmembrane helix contains a cytoplasmic C-terminal extension) peptides from Orai channels undergo a Leu-specific signature of direct interaction with the STIM1 Orai-activating region (SOAR), exactly mimicking STIM1 binding to gate Orai channels. The 20-amino-acid Orai3-M4x peptide associates avidly with STIM1 within ER-PM junctions, functions to competitively block native Ca²⁺ signals, and mediates a key modification of STIM-Orai coupling induced by 2-aminoethoxydiphenyl borate. By blocking STIM-Orai coupling, the Orai3-M4x peptide reveals the critical role of Orai channels in driving Ca²⁺ oscillatory signals and transcriptional control through NFAT. The M4x peptides interact independently with SOAR dimers consistent with unimolecular coupling between Orai subunits and STIM1 dimers. We reveal the critical role of M4x helices in defining the coupling interface between STIM and Orai proteins to mediate store-operated Ca²⁺ signals.

Figure 1.. PM-tethered Orai channel M4x peptides recruit cytosolic SOAR dimers and undergo Leu-specific profiles of FRET interaction

(A) Diagram of interacting constructs. LK-CFP-Orai-M4x comprises the Lyn kinase PM tether (LK), cyan fluorescent protein (CFP), and M4x peptide (controls were devoid of M4x); yellow fluorescent protein (YFP)-SOAR dimer is the concatenated dimer of SOAR tagged at the N terminus with YFP. (B) The M4x sequences of Orai channel subtypes highlighting key hydrophobic residues. (C) Cell localization images (left) and intensity plots (right) of control LK-CFP with YFP-SOAR dimer co-expressed in HEK-Orai1/2/3^TKO cells. (AUs, arbitrary units). (D–F) Cells as in (C) but co-expressing LK-CFP-O1M4x (containing the Orai1 267-301 M4x peptide) (D), LK-CFP-O2M4x (containing the Orai2 228-254 M4x peptide) (E), or LK-CFP-O3M4x (containing the Orai3 276-295 M4x peptide) (F), each with YFP-SOAR dimer. (G) E-FRET measurements for the interaction of YFP-SOAR dimer with LK-CFP-Orai1-M4x peptide, containing either wild-type Orai1-M4x (O1-WT); Orai1-M4x with the L273D, L276D, or combined L273D/L276D mutations; or the control (LK-CFP). (H) E-FRET between YFP-SOAR dimer and LK-CFP-Orai2-M4x peptide, with either WT Orai2-M4x (O2-WT), Orai2-M4x-I234D (O2-I234D), or the control (LK-CFP). (I) E-FRET between YFP-SOAR dimer and LK-CFP-Orai3-M4x peptide, with either WT Orai3-M4x (O3-WT), Orai3-M4x-L282D (O3-L282D), Orai3-M4x-L285D (O3-L285D), Orai3-M4x-L282D/L285D double mutant (O3-L282D/L285D), or the control (LK-CFP). (J) Cells as in (C) but co-expressing LK-CFP-O3M4x-L285D mutant with YFP-SOAR dimer. Images and intensity plots are representative of three independently repeated experiments. Scale bars represent 5 μm. One-way ANOVA analysis was performed on E-FRET results (**p < 0.01, ***p < 0.001, ****p < 0.0001). Results are means ± SEM, representative of at least three independent experiments. E-FRET analyses were on cells expressing a narrow range of LK-CFP-Orai-M4x fluorescence to assure accuracy of E-FRET values. The CFP levels are shown in Figure S3.

Figure 2.. Defining the Leu signature and mechanistic role of the Orai3 M4x peptide interaction with SOAR dimer

E-FRET analysis of transiently co-expressed LK-CFP-Orai3-M4x peptide with the YFP-SOAR dimer in HEK-Orai1/2/3^TKO cells. (A) E-FRET between YFP-SOAR and LK-CFP-Orai-M4x peptide, with either O3-WT, O3-L285D, Orai3-M4x-L285M (O3-L285M), Orai3-M4x-L285F (O3-L285F), Orai3-M4x-L285Y (O3-L285Y), or the control (LK-CFP) with no M4x peptide. (B) E-FRET between YFP-SOAR and LK-CFP-Orai-M4x peptide, with either O3-WT, O3-L282D, O3-L285D, Orai3-M4x-L288D (O3-L288D), or the control (LK-CFP). (C) Predicted structure of the hOrai3-M4x peptide to show positions of the four Leu residues (L282, L285, L288, and L292) mutated in (B). (D and E) Measurement of self-interactions between M4x peptides by E-FRET-analysis between CFP- and YFP-tagged M4x peptides in HEK-Orai1/2/3^TKO cells. E-FRET between LK-YFP-O1M4x and either LK-CFP-O1M4x WT (teal), LK-CFP-O1M4x-L273D/L276D double mutant (magenta), or LK-CFP control (black) (D). E-FRET measurements between LK-YFP-O3M4x and either LK-CFP-OM4x WT (green), LK-CFP-O3M4x-L282D/L285D double mutant (purple), or LK-CFP control (black) (E). E-FRET analyses for (E) and (D) were undertaken on cells expressing a narrow range of both LK-CFP-Orai-M4x and LK-YFP-Orai-M4x fluorescence levels to assure accuracy of E-FRET values. The CFP and YFP levels of cells are shown in Figure S4. (F) E-FRET interaction between YFP-SOAR dimer and full-length CFP-Orai3 WT (green), CFP-Orai3-L282D (red), CFP-Orai3-L285D (blue), or CFP-Orai3-L282D/L285D double mutation (purple). Orai constructs and YFP-SOAR dimer were transiently expressed in HEK-Orai1/2/3^TKO cells. One-way ANOVA analysis was performed on E-FRET results (****p < 0.0001). Results are means ± SEM of at least three independent experiments.

Figure 3.. Store-activated full-length STIM1 recruits the Orai3 M4x peptide into ER-PM junctions

(A) Time course of E-FRET interactions between transiently expressed STIM1-YFP with LK-CFP-O3M4x constructs (either WT M4x, single L282D or L285D M4x mutants, or the L282D/L285D double mutant) in HEK-Orai1/2/3^TKO cells. Store depletion was initiated by 2.5 μM ionomycin. (B) Summary data for experiments in (A) showing the change in E-FRET from baseline (0 s) to peak after ionomycin (140 s). (C and D) High-resolution fluorescence imaging of transiently expressed LK-CFP-O3M4x constructs with STIM1-YFP in ER-PM junctions measured at the PM layer adjacent to the coverslip in HEK-Orai1/2/3^TKO cells. Images were taken 5 min after 2.5 μM ionomycin addition to deplete Ca²⁺ stores. In (C), LK-CFP-O3M4x WT peptide was highly localized with STIM1-YFP, and image magnification reveals a complete overlap of fluorescence in the punctal areas. In contrast, LK-CFP-O3M4x bearing the single M4x L285D point mutation (D) showed no obvious overlap with STIM1-YFP fluorescence, and magnified STIM1 punctal areas reveal no co-localized M4x construct. Images are representative of at least three independent experiments with 5-μm scale bars. One-way ANOVA was performed on E-FRET results (****p < 0.0001). Results are means ± SEM of at least three independent experiments.

Figure 4.. The Orai3 M4x Leu profile reveals the peptide mediates STIM1-induced Orai3 channel activation

(A) Time course of E-FRET interactions between STIM1-YFP co-expressed with either full-length WT CFP-Orai3, CFP-Orai3 L282D or L285D single mutants, CFP-Orai3 L282D/L285D double mutant, or CFP control, before and after store depletion with 2.5 μM ionomycin. (B) Summary data for E-FRET analyses in (A), showing the change in E-FRET from baseline (0 s) to the peak after ionomycin addition (400 s). (C) Cytosolic Ca²⁺ signals measured by fura-2 ratiometric Ca²⁺ imaging in HEK-Orai1/2/3^TKO transiently expressing STIM1-mCherry together with either WT CFP-Orai3 (green) or L282D (red), L285D (blue), or L282D/L285D (purple) mutants of CFP-Orai3 (purple). Cells in Ca²⁺-free medium were treated with 2.5 μM ionomycin followed by 1 mM Ca²⁺ add back. (D) Summary statistics for the average peak of store-operated Ca²⁺ entry in (C). (E) I/V relationship of whole-cell I_CRAC measurements in HEK-Orai1/2/3^TKO cells transfected with the same CFP-Orai3 constructs shown in (C) together with STIM1-YFP. (F) Ca²⁺ levels in fura2-loaded HEK-Orai1/2/3^TKO cells transiently expressing either WT CFP-Orai3 (green); L282D (red), L285D (blue), or L282D/L285D (purple) mutants; or negative control LK-CFP (black). Constitutive Ca²⁺ entry and 2-APB responses were in Ca²⁺-free solution, replaced with 1 mM Ca²⁺ followed by 50 μM 2-APB, as shown. (G) Summary statistics for the 2-APB-induced Ca²⁺ entry peak in the experiments in (F). (H) I/V relationship of 2-APB-induced whole-cell current measurements for HEK-Orai1/2/3^TKO cells transfected with the same CFP-Orai3 constructs shown in (F). One-way ANOVA analyses were under taken on results in (B), (D), and (G). Ca²⁺ entry shown are means ± SEM (****p < 0.0001) and representative results from at least three independent experiments.

Figure 5.. The Orai3 M4x peptide is a powerful SOCE blocker and defines the physiological role of Orai channels in Ca²⁺ oscillation generation and downstream NFAT1 activation

(A) Endogenous SOCE in WT HEK293 cells transiently expressing LK-CFP-O3M4x (red), LK-CFP-O1M4x (green), or LK-CFP control (black). (B) SOCE as in (A) in cells transiently expressing LK-CFP-O3M4x (red), LK-CFP-O3M4x-L285D (blue), or LK-CFP control (black). (C) Summary statistics of endogenous SOCE responses shown in (A) and (B). (D) Receptor-activated Ca²⁺ responses induced by 100 μM carbachol (CCh) measured in fura-2-loaded WT HEK293 cells transiently expressing either LK-CFP-O3M4x (red), LK-CFP-O3M4x-L285D mutant (blue), or LK-CFP control (black). (E–G) Representative single-cell Ca²⁺ responses to CCh from each of the traces shown in (D) and the fraction of oscillating (OSC) and non-oscillating (NO) cells unde reach condition. Cells were transfected with LK-CFP(77.5% OSC, 22.5 NO; 44 cells) (E), LK-O3M4x-WT (19.7% OSC, 80.3% NO; 71 cells) (F), and LK-CFP-O3M4x-L285D (64.4% OSC, 35.6 NO; 59 cells) (G). (H–J) Time-dependent nuclear localization of NFAT1 in WT HEK293 cells transiently expressing NFAT1-GFP together with either LK-CFP control (H), LK-CFP-O3M4x (I), or LK-CFP-O3M4x-L285D mutant (J). Images of CFP, CFP/GFP, and GFP were taken before CCh (0 min) and at 3, 5, 10, and 15min after 100 μM CCh addition. Fluorescence of LK-CFP constructs and NFAT1-GFP are shown in red and green, respectively. (K) Summary of NFAT1-GFP translocation imaged in cells co-expressing LK-CFP (black), LK-CFP-O3M4x-WT (red), and LK-CFP-O3M4x-L285D (blue). Nuclear/cytosolic GFP-NFAT1 intensity ratios were measured in imaged cells treated as in (H)—(J) by using images taken before and after CCh addition, at the times shown. Oscillating cells were defined as having >4 peaks/10 min after initial Ca²⁺ release (400 s), with each peak at >0.03 F₃₄₀/F₃₈₀. One-way ANOVA analyses were undertaken on results in (C). Ca²⁺ entry results shown are means ± SEM (****p < 0.0001). All results shown are representative of at least three independent experiments. Scale bars represent 10 μm.

Figure 6.. The Orai3 M4x peptide overcomes a powerful STIM1 coupling mutant and mediates an important action of 2-APB

(A) Constitutive Ca²⁺ entry in fura-2-loaded HEK-STIM1/2^DKO cells transiently expressing CFP-Orai3 with either WT YFP-SOAR dimer (red), the YFP-SOAR-F394H homodimer mutant (blue), or YFP vector alone (black). A total of 1 mM Ca²⁺ was added to cells in Ca²⁺-free solution, as shown. (B) Summary statistics for the Ca²⁺ entry shown in (A). (C) E-FRET analysis of HEK293 cells transiently expressing either WT YFP-SOAR dimer (red) or the YFP-SOAR-F394H homodimer mutant (blue), together with either WT LK-CFP-O3M4x or its L285D mutant. (D) SOCE in fura-2-loaded HEK cells transiently expressing CFP-Orai3 with either WT STIMI-YFP (red), the STIM1-YFP-F394H homodimer mutant (blue), or YFP control (black). Store depletion was with 2.5 μM ionomycin, followed by 1 mM Ca²⁺ addback, and 50 μM 2-APB, as shown. (E) Enlarged detail for the 2-APB-mediated Ca²⁺ responses shown in (D) in cells expressing either STIM1-YFP-F394H or YFP control. (F) Summary statistics for Ca²⁺ entry experiments represented in (D). (G) Time course E-FRET measurements for HEK-Orai1/2/3^TKO cells transiently co-expressing CFP-Orai3 with either WT STIM1-YFP (red) or STIM1-YFP F394H mutant (blue). Additions of 2.5 μM ionomycin and 50 μM 2-APB were as shown. (H) Average change in E-FRET from baseline (before Ca²⁺) to store-depleted peak (iono), and additional E-FRET peak change after the 2-APB addition, from experiments in (G). (I) E-FRET time course measured in HEK-Orai1/2/3^TKO cells transiently co-expressing the Orai3-M4x peptide with either STIM1-YFP (red) or STIM1-YFP-F394H mutant (blue). Additions of 2.5 μM ionomycin and 50 μM 2-APB were as shown. (J) Summary statistics for E-FRET results shown in (I). Changes in FRET are shown from baseline (Ca²⁺-free) to store-depleted peak (iono) and additional FRET peak after 2-APB. One-way ANOVA analysis was performed in (B) and (F). Ca²⁺ entry results are means ± SEM (****p < 0.0001). Results are representative of at least three independent experiments

Figure 7.. Models depicting the role of Orai channel M4x in coupling with the SOAR domain of STIM proteins

(A) C-terminal sequences of Orai channels showing the highly conserved, flexible 5-aa “nexus” that attaches the M4 helix to the helical M4x extension of 35, 27, and 20 aa, in Orai1, Orai2, and Orai3, respectively. The M4x peptides contain four conserved hydrophobic residues (yellow highlighted); the first two have been thought to mediate antiparallel pairing between M4x peptides in adjoining Orai channel subunits. (B and C) Structural models of two adjacent subunits from the hexameric Orai channel. The orientation of the four transmembrane helices (M1–M4) are based on the dOrai crystal structure shown in Figure S1A. The central pore-forming M1 helices are surrounded by the M2, M3, and outermost M4 helices. In (B), the M4x peptides are shown in a paired or “latched” configuration predicted from the dOrai crystal structure (Hou et al., 2012) and suggested to form a STIM-binding pocket in Orai (Fahrner et al., 2014; Maus et al., 2015; Stathopulos et al., 2013; Yen and Lewis, 2018; Zhou et al., 2016). In (C), the two M4x peptides are shown as “unlatched” helices, with each able to undergo independent interactions with STIM as indicated by the current studies. (D–F) Predicted models of coupling between the active, dimeric SOAR units of STIM proteins, and M4x peptides in Orai channels. In (D), “unimolecular” coupling is shown between each of the two independent Orai-binding sites on the SOAR dimer (Zhou et al., 2015) and single M4x peptides. This model is consistent with the optimal 2:1 stoichiometry for STIM-Orai coupling (Hoover and Lewis, 2011). In (E), the unimolecular coupling between SOAR and M4x peptides allows for “intermolecular” coupling, consistent with the observed cross-linking and clustering of Orai channels by STIM1 (Zhou et al., 2018a). In (F), a “bimolecular” interaction between independent M4x peptides and SOAR dimers cannot be excluded, but this is distinct from formation of abinding pocket formed through M4x interactions and is not the stoichiometrically preferred coupling configuration.