The unfolding and activation of STIM1 in store-operated calcium signal generation

Store-operated Ca²⁺ signals are generated by STIM proteins which are ER luminal Ca²⁺ sensors that couple across ER-PM junctions to activate Ca²⁺ entry through PM Orai channels [1]. These Ca²⁺ signals control a spectrum of cellular responses including transcription, secretion and cell growth. Intriguing has been the extraordinarily dynamic operation of STIM proteins which undergo a series of elaborate unfolding, translocation, and channel-coupling events. Recent single-molecule Förster resonance energy transfer (smFRET) and biochemical cross-linking approaches from van Dorp et al. [2] reveal some critical new information on the molecular architecture of STIM1 and its unfolding to span across ER-PM junctions and trigger Orai channel gating [3].

Much has been learned about STIM1 structure and function, but uncertainty has remained on how the C-terminal cytoplasmic domain is able to unfold and extend across the approximately 15 nm gap in ER-PM junctions in order to tether and activate Orai channels [3]. STIM1 has a tightly coiled ER luminal N-terminus including EF-hands which sense luminal Ca²⁺ changes (Fig. 1A,B). A single transmembrane helix joins this domain to a large cytoplasmic region comprising several helical domains (CC1, CC2, and CC3) followed by a long flexible tail ending in a lysine-rich (KR) sequence (Fig. 1A). STIM1 is always dimeric. In its resting state (ER stores replete with Ca²⁺), the two N-terminal domains within the dimer are unattached giving distance between the two transmembrane helices (Fig. 1B). The resting C-terminal domain is folded such that there is complete occlusion of the critical, active domain known as SOAR (STIM-Orai activating region) or CAD (channel activating domain) [1]. SOAR comprises a tight intertwined cluster of the CC2 and CC3 helices from each of the two STIM1 monomers. Cytoplasmic expression of just the 100-amino acid SOAR sequence is sufficient to give full constitutive activation of Orai1 channels [1, 3].

Fig. 1.

The structure and function of STIM1. A. Sequence domains of STIM1: SP, signal peptide; cEF, canonical EF-hand; hEF, hidden EF-hand; SAM, sterile alpha motif; TM, transmembrane helix; CC1-CC3 coiled-coil regions; Cα1–3, the CC1α1/ CC1α2/ CC1α3 helices within CC1; SOAR, STIM-Orai Activating Region; CAD, Channel Activating Domain; Sα1–4, the four helices within SOAR; K, lysine-rich C-terminal domain. B. The resting state of the STIM1 dimer. SOAR is occluded within a “clamped” state through “domain-swapping” inter-dimer interactions between the CC1α1 helices and the CC3 coil of SOAR. TM, transmembrane domain; PL, poly-lysine C-terminus. C. After Ca²⁺ is depleted in the ER, the distant luminal EF-SAM domains associate, causing the TM and proximal CC1α1 helices to move together resulting in SOAR being “squeezed” and hence expelled and to begin to flip-out to the rear of the STIM1 protein (shown flipped by 60°). D. The CC1α1 and CC1α2 helices continue to bind together extending the length of STIM1, SOAR continues to flip-out (now at 120°), and the flexible C-termini are able to reach toward the PM, where their poly-lysine tails can associate with acidic phospholipids in the PM, tethering STIM1 to the PM. E. The three CC1 helices (CC1α1, CC12α2, and CC1α3) have fully bound and “zipped” together, giving STIM1 its fully extended conformation, which allows SOAR to bind to and activate Orai channels in the PM, permitting Ca²⁺ to enter and generate “store-operated” Ca²⁺ signals.

Through clever fluorophore-labeling of discrete amino acids within the STIM1 cytoplasmic domain, van Dorp et al. [2] were able to measure distances between key residues in the purified STIM1 dimer. Distance measurements within the SOAR domain are consistent with an earlier crystal structure [4] showing tightly-associated CC2 helices in a dimeric, parallel configuration. The results militate against an NMR-derived antiparallel configuration derived from isolated nonfunctional CC2-containing fragments [5]. Interestingly, the smFRET measurements reveal the “apical” region of SOAR has quite some flexibility not revealed in the static crystal structure [2], in keeping with the crucial ability of this region to conformationally couple with Orai [3].

The studies provide important new understanding on how the folding-up of resting STIM1 occludes SOAR [2]. They support findings [6, 7] that the CC1α1 helix is closely associated with the CC3 helix within SOAR. Unexpectedly, this interaction occurs in a “domain-s-wapping” configuration, wherein the CC1α1 helix from one monomer interacts with the CC3 helix from the other monomer (Fig. 1B). This intriguing intertwining of helices within STIM1 explains much about the “clamping” of the SOAR entity and its subsequent release during activation. The CC1α1/CC3 smFRET results also reveal that SOAR is inverted relative to the configuration needed for Orai-activation, with the SOAR apices close to the ER membrane (Fig. 1B). Further smFRET distance measurements revealed that the two parallel CC1α3 helices are closely apposed with their N-termini oriented away from SOAR. The CC1α2 and CC1α3 helices are folded antiparallel above SOAR (Fig. 1B). Interestingly, molecular modeling of the SOAR/CC1α1 helices suggests the hydrophobic residues shown earlier to hold SOAR in the clamped configuration [6, 7] are indeed mediating SOAR/CC1α1 docking. A series of elegant cysteine cross-linking studies [2] now reveal that all three CC1 helices undergo homologous inter-dimer parallel pairing to mediate the full extension of STIM1 in its activated state, to span the ER-PM junction (Fig. 1E).

The interesting results from van Dorp et al. [2] provide a critical new understanding of the extraordinary activation and extension of STIM1. The unfolding and interactions build upon those suggested earlier [8, 9] with the SOAR domain occluded within the CC1α1 helices. It contrasts with a recent model suggesting CC1α1/CC1α2/CCα3 helices form a compact domain, but this was based on NMR of detergent-induced monomeric CC1 [10].

How does the actual unfolding and exposure of SOAR proceed? The earlier NMR-derived model that SOAR itself might unfold in order to interact with Orai [5], seems intrinsically unlikely given the fundamental stability of SOAR in the new studies [2] and earlier studies [3]. Instead, a new SOAR “flip-out” model appears to be the most likely (Fig. 1C–E). In this model, we envisage the initial driving force for STIM1 activation as pairing of the luminal N-termini driven by Ca²⁺ dissociation from EF-hands. The resulting close-approach of TMs and CC1α1 helices may have the effect of “squeezing” SOAR from its resting niche between the CC1α1 coils, causing it to start flipping out. Thereafter, the STIM1 CC1 helices begin to zip together from the ER surface. This allows the flexible C-termini to reach and tether STIM1 to the membrane, assisting SOAR to dock with and activate the Orai channel. Thus, STIM1 is a multipurpose signal transducer – it creates SOAR as a highly specific key to unlock Orai channels, but keeps the key safely hidden until Ca²⁺ store-depletion triggers STIM1 to release it and to assure its safe delivery to the PM. Moreover, it has the capacity to reverse – to rapidly retrieve and sequester SOAR back toward the ER surface as the luminal Ca²⁺ tide returns and Ca²⁺ stores become replete.

Glossary

STIM

stromal interaction molecule