The 2β Splice Variation Alters the Structure and Function of the Stromal Interaction Molecule Coiled-Coil Domains

Abstract

Stromal interaction molecule (STIM)-1 and -2 regulate agonist-induced and basal cytosolic calcium (Ca²⁺) levels after oligomerization and translocation to endoplasmic reticulum (ER)-plasma membrane (PM) junctions. At these junctions, the STIM cytosolic coiled-coil (CC) domains couple to PM Orai1 proteins and gate these Ca²⁺ release-activated Ca²⁺ (CRAC) channels, which facilitate store-operated Ca²⁺ entry (SOCE). Unlike STIM1 and STIM2, which are SOCE activators, the STIM2β splice variant contains an 8-residue insert located within the conserved CCs which inhibits SOCE. It remains unclear if the 2β insert further depotentiates weak STIM2 coupling to Orai1 or independently causes structural perturbations which prevent SOCE. Here, we use far-UV circular dichroism, light scattering, exposed hydrophobicity analysis, solution small angle X-ray scattering, and a chimeric STIM1/STIM2β functional assessment to provide insights into the molecular mechanism by which the 2β insert precludes SOCE activation. We find that the 2β insert reduces the overall α-helicity and enhances the exposed hydrophobicity of the STIM2 CC domains in the absence of a global conformational change. Remarkably, incorporation of the 2β insert into the STIM1 context not only affects the secondary structure and hydrophobicity as observed for STIM2, but also eliminates the more robust SOCE response mediated by STIM1. Collectively, our data show that the 2β insert directly precludes Orai1 channel activation by inducing structural perturbations in the STIM CC region.

Keywords: Fura-2; alterative splicing; calcium signaling; coiled-coil; store-operated calcium entry (SOCE); stromal interaction molecule-1 (STIM1); stromal interaction molecule-2 (STIM2); structure.

Figure 1

Human stromal interaction molecule (STIM1) and (STIM2) domain architecture. The residue ranges encompassing each conserved domain are indicated. The Orai-activating STIM fragment (OASF) constructs used in the present work are shown above the full-length architectures. The relative locations of the CRAC activating domain (CAD)/STIM-Orai activating region (SOAR) and the CC1-CC2 NMR structure fragment are indicated below the STIM1 architecture. The location of the 2β insert in the CC2 region of STIM2 is highlighted. The acidic residues found in the ID sequences are shown in red text. S, signal peptide; EF1, canonical EF-hand; EF2, non-canonical EF-hand; SAM, sterile α-motif; TM, transmembrane; CC1, coiled-coil-1; CC2, coiled-coil-2; CC3, coiled-coil-3; ID, inhibitory domain; N, amino terminus; C, carboxyl terminus.

Figure 2

Secondary structure of STIM OASF proteins. (A) Far-UV CD spectra of STIM1-OASF and STIM1-2β-OASF. (B) Comparison of STIM1-OASF and STIM1-2β-OASF mean residue ellipticity (MRE) at 222 nm. (C) Far-UV CD spectra of STIM2-OASF and STIM2-2β-OASF. (D) Comparison of STIM2-OASF and STIM2-2β-OASF MRE at 222 nm. Spectra were acquired using 0.2 mg·mL⁻¹ protein at 20 °C. Data are means ± SEM of n = 3 separate protein purifications. In (B,D), * p < 0.05 using Student’s t-test.

Figure 3

Thermal stability of STIM OASF proteins. (A) Thermal unfolding profile of STIM1-OASF and STIM1-2β-OASF. (B) Comparison of the STIM1-OASF and STIM1-2β-OASF T_m extracted from the thermal melts. (C) Thermal unfolding profile of STIM2-OASF and STIM2-2β-OASF. (D) Comparison of the STIM2-OASF and STIM2-2β-OASF T_m extracted from the thermal melts. Thermal melts were acquired using 0.2 mg·mL⁻¹ protein. Data are means ± SEM of n = 3 separate protein purifications. In (B), ** p < 0.01 using Student’s t-test.

Figure 4

Extrinsic hydrophobicity of STIM OASF proteins. ANS fluorescence emission spectra of STIM1-OASF (A), STIM1-2β-OASF (B), STIM2-OASF (C), and STIM2-2β-OASF (D) as a function of increasing CaCl₂ concentrations. In (A–D), the black spectra represent the buffer control scans. Spectra were acquired using 0.14 mg·mL⁻¹ protein, 50 μM ANS and an excitation wavelength of 372 nm at 35 °C. (E) Change in the maximum ANS fluorescence intensity for STIM1-OASF (blue), STIM1-2β-OASF (red), STIM2-OASF (green) and STIM2-2β-OASF (purple) plotted as a function of CaCl₂ concentration. Data are means ± SEM of n = 3 separate protein purifications.

Figure 5

Hydrodynamic size distributions of STIM OASF proteins. (A) Distribution of hydrodynamic radii for STIM1-OASF. (B) Distribution of hydrodynamic radii for STIM1-2β-OASF. (C) Distribution of hydrodynamic radii for STIM2-OASF. (D) Distribution of hydrodynamic radii for STIM2-2β-OASF. In (A–D), the sizes are derived from the regularization deconvolution of the autocorrelation functions acquired using 0.5 mg·mL⁻¹ protein at 35 °C in the absence (colored circles) and presence of 25 mM CaCl₂ (open circles). (E) Direct comparison of the STIM1-OASF and STIM1-2β-OASF size distributions in the absence of CaCl₂. (F) Direct comparison of the STIM2-OASF and STIM2-2β-OASF size distributions in the absence of CaCl₂.

Figure 6

SAXS-derived bead models of STIM OASF proteins. DAMMIF-derived bead models of STIM1-OASF (conformation 1) (A), STIM1-OASF (conformation 2) (B), STIM1-2β-OASF (C), STIM2-OASF (D), and STIM2-2β-OASF (E). The reconstructed scattering profiles based on the STIM1-OASF, STIM1-2β-OASF, STIM2-OASF and STIM2-2β-OASF DAMMIF models are plotted in Figures S6A,D and S7A,D, respectively.

Figure 7

Fura-2 cytosolic Ca²⁺ assessment of full-length STIM1 and STIM1-2β function in SOCE. (A) Representative Fura-2 ratiometric fluorescence traces reporting relative changes in cytosolic Ca²⁺ of mCh-STIM1 (blue), mCh-STIM1-2β insert (red) and empty mCh-vector (grey) expressing HEK293 cells also stably expressing yellow fluorescence protein (YFP)-Orai1. Cells were initially bathed in Ca²⁺-free buffer. The relative change in Fura-2 fluorescence was monitored after 2 μM TG and subsequently 2 mM CaCl₂ additions to the external medium. (B) Maximal ER Ca²⁺ release (i.e., F/F₀ taken after TG addition in 0 mM external CaCl₂. (C) Maximal SOCE taken after addition of 2 mM CaCl₂ to the external medium. (D) Maximum mCh fluorescence emission of cell suspensions using an excitation wavelength of 565 nm. (E) Maximum YFP fluorescence emission of cell suspensions using an excitation wavelength of 490 nm. Data in (B–E) are means ± SEM of n = 4 separate transfections. In (C) * p < 0.05 and (D) ** p < 0.01 using One-way ANOVA and Tukey’s multiple comparisons test.

Figure 8

STIM CC1-CC2 structures identifying the location of the 2β variations. (A) Cartoon view of the backbone atoms making up the CC1 and CC2 helices in the dimeric human STIM1 CC1-CC2 solution structure. The location of the two human Orai1 C-terminal helices (yellow ribbons) are indicated (top view). The hydrophobic (brown sticks) and positively charged (blue sticks) residues within the CC2 helices making up the Orai1 C-terminal helix binding pocket are indicated and labeled (bottom view). (B) Cartoon view of the backbone atoms making the CC1 and CC2 helices in the dimeric human STIM2 homology model. The hydrophobic and positively charged residues making up the putative Orai1 C-terminal helix binding sites are indicated at top and bottom views. In (A,B), the relative locations of the 2β variations within CC2 are shown (red arrows). The STIM2 homology model was created from the STIM1 CC1-CC2 solution structural coordinates (2MAK.pdb) in Modeller [54]. Structure images were rendered using PyMOL.