Molecular basis of allosteric Orai1 channel activation by STIM1

Abstract

Store-operated Ca²⁺ entry through Orai1 channels is a primary mechanism for Ca²⁺ entry in many cells and mediates numerous cellular effector functions ranging from gene transcription to exocytosis. Orai1 channels are amongst the most Ca²⁺-selective channels known and are activated by direct physical interactions with the endoplasmic reticulum Ca²⁺ sensor stromal interaction molecule 1 (STIM1) in response to store depletion triggered by stimulation of a variety of cell surface G-protein coupled and tyrosine kinase receptors. Work in the last decade has revealed that the Orai1 gating process is highly cooperative and strongly allosteric, likely driven by a wave of interdependent conformational changes throughout the protein originating in the peripheral C-terminal ligand binding site and culminating in pore opening. In this review, we survey the structural and molecular features in Orai1 that contribute to channel gating and consider how they give rise to the unique biophysical fingerprint of Orai1 currents.

Graphical Abstract

Abstract figure legend Activation of store-operated Orai1 channels. Following the depletion of endoplasmic reticulum (ER) Ca²⁺ stores, STIM1 binds to Orai1 at ER–plasma membrane junctions. Inset: the channel activation process involves a conformational wave throughout the transmembrane domains of Orai1, starting from the peripheral C-terminal extensions and propagating towards the central pore.

Introduction

Store-operated Ca²⁺ entry

Rises and falls in cytosolic Ca²⁺ are at the crux of many cellular signalling pathways for transducing extracellular inputs received from cell-surface receptors into intracellular changes in gene expression and protein regulation (Clapham, 2007). One of the primary ways by which cellular Ca²⁺ elevations are mobilized, especially in non-excitable cells, is store-operated Ca²⁺ entry through Ca²⁺ release-activated Ca²⁺ (CRAC) channels (Prakriya & Lewis, 2015) (Fig. 1A). CRAC channels open in response to inositol trisphosphate-dependent release of endoplasmic reticulum (ER) Ca²⁺ stores following the engagement of many types of G-protein coupled receptors and receptor tyrosine kinases (Prakriya & Lewis, 2015) (Fig. 1A), and the ensuing Ca²⁺ entry mediates a wide range of essential cellular processes such as gene transcription, proliferation, migration and exocytosis (Prakriya & Lewis, 2015). Clinical studies have shown that patients bearing mutations in the genes encoding CRAC channel proteins suffer from devastating diseases including severe combined immunodeficiency, tubular aggregate myopathy, thrombocytopenia and ectodermal dysplasia (Lacruz & Feske, 2015; Prakriya & Lewis, 2015). Because of their causative roles in these diseases, CRAC channels have attracted attention in recent years as potential drug targets for these and a growing list of other pathologies (Gerasimenko et al. 2013; Prakriya & Lewis, 2015; Wen et al. 2015; Munoz & Hu, 2016; Ali et al. 2017; Groschner et al. 2017; Chalmers & Monteith, 2018; Michelucci et al. 2018; Secondo et al. 2018; Vaeth & Feske, 2018; Wegierski & Kuznicki, 2018).

Figure 1. Molecular machinery of store-operated calcium entry.

A, schematic diagram of CRAC channel activation. Stimulation of cell surface receptors including G-protein coupled receptors and receptor tyrosine kinases (GPCR/RTK) increases cytosolic inositol 1,4,5-trisphosphate (IP₃), which causes ER Ca²⁺ store depletion. STIM1 senses the decreased ER Ca²⁺ concentration and migrates to ER–plasma membrane junctions, where it gates Orai1 channels. The ensuing calcium entry mediates numerous downstream cellular effector functions. B, following STIM1 activation, the CAD domain is exposed enabling it to bind to the cytosolic surface of Orai1 channels to open the pore. NT, N terminus; CT, C terminus; CC1, coiled-coil domain 1; CAD, CRAC activation domain; PLC, Phospholipase C.

Overview of the channel activation process

The principal components of CRAC channels are the STIM and Orai proteins, which are both necessary and sufficient for reconstituting CRAC channel function (Zhou et al. 2010) (Fig. 1B). While there are three human homologues of Orai and two for STIM that are differentially expressed in various tissue types and which could potentially mediate different physiological functions (Prakriya & Lewis, 2015), this review will focus on the most widely studied canonical CRAC channels that are formed by Orai1 and STIM1. STIM1 is a single-pass ER membrane protein with luminal Ca²⁺-sensing EF-hand and sterile α-motif (SAM) oligomerization domains that link ER Ca²⁺ store depletion to the activation of Orai1, which forms the Ca²⁺-selective channel pore in the plasma membrane (Prakriya & Lewis, 2015) (Fig. 1B). When ER Ca²⁺ stores are depleted, dissociation of Ca²⁺ from the luminal EF-hand domain of STIM1 initiates conformational changes (Stathopulos et al. 2008) that result in the oligomerization and migration of STIM1 to ER–plasma membrane junctions, forming puncta of active STIM1–Orai1 channel complexes (Prakriya & Lewis, 2015) (Fig. 1B). A key event in this process is exposure of the Orai1-activating domain of STIM1, known as the CRAC activation domain (CAD) (Park et al. 2009), STIM–Orai activating region (SOAR) (Yuan et al. 2009; Yang et al. 2012) or CCb9 (Kawasaki et al. 2009), from a hidden configuration in the resting state, enabling it to directly bind to and activate Orai1 channels (Prakriya & Lewis, 2015) (Fig. 1B).

The crystal structure of Drosophila melanogaster Orai (dOrai; PDB ID: 4HKR), which is highly homologous to human Orai1 within the transmembrane domains (TMs), reveals a hexameric stoichiometry with sixfold symmetry in the pore and threefold symmetry at the outer C-terminal regions (Hou et al. 2012) (Fig. 2A). This stoichiometry is consistent with recent functional studies using engineered concatemers indicating that Orai1 channels are indeed composed of a trimer of dimers (Cai et al. 2016; Yen et al. 2016). Within the channel, the four TMs of each Orai1 subunit are arranged in concentric rings, with TMs 2–4 surrounding a central pore lined by TM1 (Hou et al. 2012) (Fig. 2A).

Figure 2. Global architecture of WT dOrai and H206A dOrai (H134A Orai1) channels.

A, top and side views of the crystal structure of dOrai (PDB ID: 4HKR; Hou et al. 2012) showing the arrangement of six subunits around a central pore with coiled-coil interactions between the three pairs of C-terminal extensions. TMs 1–4 and the C-terminal extension helices are shown in blue, red, green, yellow and orange, respectively. B, top and side views of the crystal structure of H206A dOrai (H134A Orai1) (PDB ID: 6BBF; Hou et al. 2018). Compared to the WT channel, the H206A mutant displays unlatching of the C-terminal extensions and dilatation of the N-terminal inner pore region.

Although a general framework of STIM1-mediated Orai1 activation has emerged, many crucial details surrounding the key events along this pathway remain poorly understood. These include the processes of STIM1 activation and CAD exposure following ER Ca²⁺ store depletion, the structural and molecular basis of STIM1–Orai1 coupling, and the mechanism by which STIM1 binding to Orai1 drives channel activation. This review focuses on the last question, in particular, how the conformational changes at the STIM1 binding site on Orai1 are allosterically communicated to the channel pore.

Biophysical properties of CRAC currents

CRAC channels exhibit a fascinating set of permeation and gating properties that render them unique among cation channels. Foremost among these properties is their exquisite Ca²⁺ selectivity: CRAC channels conduct Ca²⁺ >1000 times better than Na⁺ (P_Ca/Na >1000) under physiological conditions, making them amongst the most Ca²⁺-selective channels known (Hoth & Penner, 1993; Hoth, 1995; Prakriya et al. 2006). A second distinguishing feature is their low unitary conductance, which is estimated from noise analysis to be only 10–25 fS for Ca²⁺ and ~0.8 pS for Na⁺ (Zweifach & Lewis, 1993; Prakriya & Lewis, 2006). By comparison, conductance estimates of voltage-gated Ca²⁺ channels are at least 100 times larger(Hess et al.1986; Guia et al.2001). Incidentally, this feature has presented significant challenges towards understanding the gating properties of CRAC channels because the unitary currents are below the resolution of direct single-channel recordings. A third property is the extreme non-linearity between STIM1 binding and channel activation (Hoover & Lewis, 2011; Yen & Lewis, 2018; Yeung & Prakriya, 2018). Maximal channel activation requires at least two STIM1 molecules per Orai1 subunit (Li et al. 2007; Scrimgeour et al. 2009; Hoover & Lewis, 2011) and whole-cell current drops off sharply as the STIM1:Orai1 ratio is lowered (Hoover & Lewis, 2011). Related to this notable non-linearity between STIM1 binding and channel opening, noise analysis reveals an unusual gating mode wherein CRAC channels transit abruptly between a long-lived silent state and a highly active one with a high apparent open probability (P_o ≈ 0.7–0.8) (Prakriya et al. 2006; Yen & Lewis, 2018). Fourth, Orai1 channels exhibit striking coupling of gating and ion selectivity; unlike many other ligand-gated ion channels in which gating and ion permeation are independent channel properties, not only does STIM1 binding open the pore, it also renders the Orai1 channel Ca²⁺ selective (McNally et al. 2012). In this way, STIM1 is not only the ligand for the channel, but is essentially a key accessory subunit that confers the defining properties of CRAC channels that includes a narrow pore with high Ca²⁺ selectivity.

Molecular details and functional features of STIM1–Orai1 interaction

Where does STIM1 bind?

Each subunit of the hexameric Orai1 channel has four TMs, with two extracellular loops and one intracellular loop linking the TMs, as well as cytosolic N- and C-terminal tails (Fig. 1B). In addition to STIM1, these intracellular tails are suggested to interact with a growing list of channel modulators such as calmodulin, CRACR2A and septins (reviewed in Shim et al. 2015). Early studies using purified Orai1 fragments have suggested that STIM1 directly binds strongly to the Orai1 C-terminus and more weakly to the N-terminus, but not to the intracellular TM2–3 loop (Park et al. 2009; Zhou et al. 2010). In one widely accepted model, STIM1–Orai1 binding is postulated to occur through a two-step process in which STIM1 is first recruited to the channel via a high affinity site at the Orai1 C-terminus and then subsequently gates the channel by interacting with the N-terminus (Li et al. 2007; Muik et al. 2008; Zheng et al. 2013). More recently, however, an alternative model has emerged in which STIM1 binding to the C-terminus alone is sufficient for gating (Zhou et al. 2016). It is also possible that the N- and C-termini work synergistically as a single binding interface (McNally et al. 2013; Palty & Isacoff, 2016).

In all three schemes, the Orai1 C-terminus is noted to be essential for interaction with STIM1. This binding surface consists of an antiparallel coiled-coil motif formed by C-terminal extension helices between adjacent subunits (Hou et al. 2012; Stathopulos et al. 2013) (Fig. 2A). Because this coiled-coil interaction is driven by a hydrophobic core with residues L273 and L276 (Hou et al. 2012; Tirado-Lee et al. 2015), STIM1 binding is abolished when either residue is mutated to polar or charged amino acids or when this region is truncated, likely due to a decrease in the coiled-coil propensity of the C-termini (Li et al. 2007; Navarro-Borelly et al. 2008; Frischauf et al. 2009; McNally et al. 2013; Palty et al. 2015).

Disruptions in the putative N-terminal binding site also reduce STIM1–Orai1 binding and abrogate channel function (Li et al. 2007; Frischauf et al. 2009; Derler et al. 2013; McNally et al. 2013; Zheng et al. 2013; Palty et al. 2015). This evidence was interpreted in many early models to indicate that gating is predominantly controlled by the N-terminal interaction site (Li et al. 2007; Zheng et al. 2013). However, the idea that STIM1 binding to the N-terminus is essential for pore opening has been challenged by recent studies showing that mutations in the putative N-terminal STIM1-binding site (but not the C-terminal site) also abrogate currents in gain-of-function (GOF) Orai1 backgrounds independently of interaction with STIM1 (Zhou et al. 2016; Derler et al. 2018; Yeung et al. 2018). Moreover, in the context of the closed dOrai crystal structure (Hou et al. 2012), the N-terminus appears to be shielded by other TMs in the resting state (Fig. 2A), and it is difficult to picture how STIM1 can access this domain without significant steric clash. Further, new studies have indicated that the C-terminal STIM1-binding site plays a more active role in gating than previously appreciated (McNally et al. 2013; Palty et al. 2015; Zhou et al. 2016) and that channel activation involves a global conformational change of the entire protein rather than a localized motion at the N-terminus near TM1 (Frischauf et al. 2017; Yeung et al. 2018). Taken together, these results suggest that the N-terminus is required for an aspect of channel function independently of any role in STIM1 binding. Whether the N-terminus truly harbours a STIM1 binding site in the full-length protein remains unknown.

Non-linear dependence of channel gating on STIM1 binding

The stoichiometric requirements of STIM1–Orai1 binding has been a longstanding question of interest. Several early studies have revealed that activation of Orai1 channels is highly non-linear with respect to STIM1 binding, with at least two STIM1 molecules per Orai1 required to form fully active channels that recapitulate the properties of endogenous CRAC current (Li et al. 2007; Scrimgeour et al. 2009; Hoover & Lewis, 2011; McNally et al. 2012). Most recently, Yen and Lewis examined the contribution of each STIM1 binding site at the Orai1 C-terminus by introducing an L273D mutation that renders the subunit unable to interact with STIM1 (Li et al. 2007) into various positions within a hexameric concatemer (Yen & Lewis, 2018). Intriguingly, they determined that even when only one out of six subunits was mutated, the overall channel open probability diminished by 90% and the channel became less ion selective in divalent-free solutions (Yen & Lewis, 2018). This suggests that STIM1 binding to all six subunits is required for channel activation and that the gating process is highly cooperative. Because a large fraction of channels remain functionally silent during the 200 ms sampling windows of most steady-state noise analysis experiments (Prakriya & Lewis, 2006; Kilch et al. 2013; Yamashita & Prakriya, 2014; Mullins et al. 2016; Yen & Lewis, 2018), it appears reasonable to conclude that channels with one to five subunits bound have very low P_o (Yen & Lewis, 2018). The importance of the final transition from a nearly closed state of five subunit-bound channel to the fully active six subunit-bound one is consistent with previous work implying the stepwise recruitment of Orai channels from a silent gating mode to one with a high open probability during STIM1 activation (Prakriya & Lewis, 2006). Even more fascinating, the high P_o of the final open state of Orai channels appears to be independent of how they are activated, whether by STIM1 (Prakriya & Lewis, 2006; Kilch et al. 2013; Yamashita & Prakriya, 2014) or the Orai channel modulator 2-aminoethoxydiphenyl borate (2-APB) (Yamashita & Prakriya, 2014), and consistent across various mutant backgrounds (Mullins et al. 2016; Yen & Lewis, 2018). Therefore, the high P_o appears to be an intrinsic property of the channel (discussed further in Yeung & Prakriya, 2018). The molecular mechanisms by which the channel is stabilized in the high P_o state are unknown, but it is tempting to speculate that there may be specific motifs within the Orai1 protein that stabilize the open state of the channel.

Pore architecture and current gating models

Outer pore vestibule and selectivity filter

Broadly speaking, the channel pore can be divided into three regions: an acidic region in the outer pore, a hydrophobic central region, and a basic region in the inner pore (Fig. 3A). The wide outer vestibule of the pore has 18 aspartate residues (D110/D112/D114 on each subunit) and dramatically narrows near a ring of glutamate residues forming the selectivity filter (E106) at the pore entrance (Prakriya et al. 2006; Vig et al. 2006; Yeromin et al. 2006; McNally et al. 2009) (Fig. 3A). These negative charges in the outer vestibule and selectivity filter are important for attracting cations to the external mouth and enhancing permeation by funnelling ions into the pore (Frischauf et al. 2015) with E106 directly forming the high affinity Ca²⁺ binding site regulating Ca²⁺ selectivity. Interestingly, one study found that the binding affinity of Ca²⁺ at the Glu selectivity filter alone does not account for high Ca²⁺ selectivity, rather it is binding at the selectivity filter along with an unusually slow ion flow rate (likely due to its narrow pore dimensions) that confers high Ca²⁺ selectivity (Yamashita & Prakriya, 2014). Ion sizing experiments indicate that the STIM1-mediated boost in Ca²⁺ selectivity of the weakly Ca²⁺-selective constituently conducting V102C/A channels occurs with concomitant narrowing of the pore (McNally et al. 2012), which would be predicted to bring E106 residues closer together. Although the exact underlying biophysical mechanisms of how STIM1 enhances Ca²⁺ selectivity are unknown, one possibility is that the STIM1-dependent pore narrowing promotes the ability of Ca²⁺ to more effectively neutralize the negatively charged carboxylates than Na⁺ while also slowing ion flux by increasing the entry and exit barriers. Finally, in addition to regulating Ca²⁺ selectivity and permeation, the outer pore, and in particular, the selectivity filter E106 (Scrimgeour et al. 2012) and V107 (Bulla et al. 2019), are also implicated in external pH block on the channel. Whether this occurs through allosteric modulation of gating or simple pore block remains to be elucidated.

Figure 3. Pore helix rotation and inner pore dilatation models of Orai1 gating.

A, side view of the transmembrane domain 1 (TM1) pore helix (blue) and the N-terminal extension helix (cyan) from two diagonally facing subunits in the crystal structure of dOrai, showing the pore-lining residues (equivalent Orai1 residues in parentheses). B, schematic diagram of the pore helix rotation model. Positions of the hydrophobic gate F99 (yellow) and residue G98 (pink) in the pore are shown. Gating occurs through a slight rotation of TM1, which moves F99 away from the pore axis and decreases the free energy barrier in the V102/F99 hydrophobic stretch to allow ion permeation. C, schematic diagram of the inner pore dilatation gating model. The structures of the inner pore region in the closed WT and open H206A dOrai (H134A Orai1) are shown side by side (densities in the W148 region were not modelled in H206A dOrai). In this model, STIM1 binding to Orai1 causes the inner pore region to dilate, potentially relieving steric and electrostatic barriers formed by the basic residues to allow ion conduction (dOrai numbering shown with equivalent Orai1 residues in parentheses).

Hydrophobic stretch and pore helix rotation gating model

The central portion of the pore is lined by three rings of hydrophobic residues (V102, F99 and L95) (Fig. 3A), of which the outer two residues V102 and F99 function as a hydrophobic gate (McNally et al. 2012; Yamashita et al. 2017). Differences in accessibility of these pore residues in the presence and absence of STIM1 binding and molecular dynamics (MD) simulations indicate that pore opening is triggered by slight rotation of the pore helix to move the F99 residues away from the pore axis (Fig. 3B). This motion solvates the pore and lowers the overall energetic barrier in the hydrophobic stretch to allow ion conduction (Yamashita et al. 2017). Further, MD simulations of GOF and loss-of-function (LOF) mutants indicate that the orientation of F99 is tightly coupled to pore hydration (Yeung et al. 2018). When F99 is pore facing, the hydrophobic stretch repels water, and when F99 moves away, pore hydration increases. The reciprocal is also true; when the pore is filled with water (i.e. when the channel is open), it becomes an energetically unfavourable environment for F99, stabilizing the side-chain away from the pore axis. In this manner, F99 and V102 regulate the wetting–de-wetting transitions in the pore to control ion permeation. Consistent with their role as a hydrophobic gate, mutations of either V102 or F99 to more polar residues enhance pore hydration, leading to leaky channels (McNally et al. 2012; Dong et al. 2013; Yamashita et al. 2017), whereas increasing the collective energetic barrier by increasing the hydrophobicity at position F99 blocks gating even when STIM1 is bound (Yamashita et al. 2017).

Notably, unlike most other ion channels in which the channel gate and selectivity filter are located in discrete regions within the pore, the hydrophobic gate (V102/F99) of Orai1 is located only one turn below the selectivity filter (E106) (Hou et al. 2012) (Fig. 3A), which could account for the unique coupling between gating and ion selectivity of Orai1 (McNally et al. 2012). Thus, full channel activation from the closed state would not only move the F99 gate away from the pore axis and dilate the pore, but also reorient the E106 side chains as the helices rotate during gating to confer high Ca²⁺ selectivity in fully activated channels. By contrast, in partially open, non-selective Orai1 channels observed in some mutants or seen in wild-type (WT) channels under sub-liganded conditions (McNally et al. 2012; Yen & Lewis, 2018), only a partial conformational change may result (for example, slight pore dilatation from the closed state to move V102/F99 apart from each other and lower the hydrophobic energetic barrier without helix rotation). This latter idea awaits further experimental validation.

Inner pore basic residues and inner pore dilation gating model

Although the Orai1 N-terminus was originally thought to be disordered, the crystal structure of dOrai revealed that the sequence directly N-terminal to R91 (residues W76–S90) actually forms a contiguous helix with TM1 to create a 55 Å-long pore extending into the cytosol (Hou et al. 2012) (Fig. 3A). In the dOrai crystal structure, the inner pore is lined by three rings of basic residues K163/R159/R155 (Orai1 R91/K87/R83) bound to anions which block the cytosolic opening of the pore (Hou et al. 2012). Recently, the same group solved a 6.7 Å crystal structure of constitutively active mutant H206A dOrai (H134A Orai1; PDB ID: 6BBF), which revealed two prominent features that differentiate it from the closed WT dOrai structure, namely dilation of the inner pore and “unlatching” of the C-termini (discussed further below; Fig. 2B) (Hou et al. 2018). Although the structure is not of sufficient resolution to visualize individual side chains, the inner pore appears to be dilated by approximately 10 Å compared with WT channels (Hou et al. 2018) (Figs 2B and 3C). This feature led the authors of the study to conclude that inner pore dilatation is a defining part of the conformational change associated with H206A-driven channel activation (Hou et al. 2018) (Fig. 3C). Such a model could in principle be compatible with previous suggestions that the basic residues, and R91 in particular, could function as an inner channel gate, inhibiting ion conduction via electrostatic or steric blockade (Zhang et al. 2011; Frischauf et al. 2017).

Do the inner pore basic residues truly function as a channel gate?

Despite these data, there is also a considerable amount of evidence that raises questions about the idea that the inner pore residues, including R91, comprise a channel gate. For example, mutations of the inner pore residues do not cause channels to become leaky (Hou et al. 2018), implying that the hydrophobic region is enough to keep the channel in its closed state. Even the deletion of the entire inner pore region does not render Orai1 constitutively permeant, unless there is an additional mutation in the hydrophobic region (McNally et al. 2013; Gudlur et al. 2014; Palty et al. 2015), indicating that the primary gate is in the outer pore. In fact, most of the mutations or truncations of the inner pore lead to loss of channel activity (Li et al. 2007; Derler et al. 2009, 2013, 2018; McNally et al. 2013; Zheng et al. 2013; Hou et al. 2018), suggesting that these basic residues in the inner pore, rather than forming a barrier to ion conduction, may be essential for another aspect of channel function.

One potential scenario is that the inner pore basic residues are allosterically coupled to the hydrophobic gate and indirectly stabilize the opening of the outer gate. Another possibility is that this hub of positive charges in the inner pore is critical for ion permeation, perhaps by promoting pore hydration or coordinating anions. Indeed, based on MD simulations performed with strong electric fields applied across the membrane, Dong et al. proposed that anions such as Cl⁻ that are recruited to the inner pore can efflux through Orai to assist cation influx through the channel (Dong et al. 2014). This idea awaits experimental validation. In addition, parts of the N-terminal helical extension, specifically residues L74 and Y80, are implicated in modulation of Orai1 channels by cholesterol (Derler et al. 2016). Future studies will be required to illuminate the role of this enigmatic region. Nevertheless, whether the predominant channel gate is in the inner pore or the hydrophobic stretch or whether there is equal contribution from both gates, remains a point of contention (reviewed in Yeung et al. 2017).

How does STIM1 binding activate Orai1?

Essential role of the transmembrane domains in gating

The earliest evidence for the TM domains of Orai1 directly influencing channel opening came from studies showing that oxidation of the non-pore lining TM3 residue C195 or cross-linking of an exogenous cysteine (G158C) in the TM3 domain of 2-APB-gated Orai3 channels can inhibit or enhance CRAC channel function (Bogeski et al. 2010; Amcheslavsky et al. 2013). In a more direct line of evidence, there are now several dozens of mutations identified within the non-pore lining Orai1 TMs that inhibit channel function or cause constitutive channel activity (Srikanth et al. 2011; Palty et al. 2015; Zhou et al. 2016; Frischauf et al. 2017; Yeung et al. 2018; Bulla et al. 2019), including human mutations that cause tubular aggregate myopathy and other associated symptoms (Nesin et al. 2014; Endo et al. 2015; Garibaldi et al. 2017; Bohm et al. 2017) (Fig. 4, Table 1). Many of these mutations were discovered in a comprehensive GOF cysteine screen across all the TM residues (Yeung et al. 2018). The unforeseen high number of GOF mutations within Orai1 and the diffuse nature of the gating “hotspots” (Fig. 4, Table 1) point towards a metastable closed channel state that activates via a concerted motion across all four helices.

Figure 4. Mutations in Orai1 that cause constitutive channel activity and/or abrogate channel function.

A topology diagram of Orai1 is illustrated with the locations of point mutations causing gain-of-function (GOF), loss-of-function (LOF) or both (depending on the introduced amino acid) phenotypes shown in green, red, and blue, respectively. The effects of specific substitutions are listed in Table 1. The pervasiveness of GOF and LOF mutations throughout the channel suggests a highly allosteric gating mechanism involving the entire Orai1 protein. *These GOF mutants best recapitulate the hallmarks of Orai1 channel activity. Loci of human mutations are highlighted in bold italicized letters.

Table 1.

Gain-of-function and loss-of-function point mutations in Orai1

	Mutations
Domain	GOF	LOF	References
N-terminus		W76P	(Derler et al. 2013)
		81LSRAK85 → 81AARAE85	(Gudlur et al. 2014)
		K85A/C/E	(Lis etal. 2010; McNally et al. 2013)
TM1		R91F/I/L/V/W	(Feske et al. 2006; Derler et al. 2009)
		L96A	(Yeung etal. 2018)
	S97C/I/L/M/V	S97F/W/Y	(Garibaldi etal. 2017; Yeung etal. 2018)
	G98D/E/H/K/M/N/P/Q/S/T	G98A/C/G/F/I/K/L/V/W	(Zhang et al. 2011; McNally et al. 2012; Endo et al. 2015; Yamashita et al. 2017)
	F99C/G/M/S/T/W/Y	F99I/L/V	(Yamashita etal. 2017)
	A100C		(Yeung etal. 2018)
		M101A/C	(McNally etal. 2009; Yeung etal. 2018)
	V102A/C/G/S/T	V102D/E/F/H/K/H/P/R/T/W	(McNally etal. 2012)
		A103E	(McCarl etal. 2009)
		M104A	(Yeung etal. 2018)
		E106A/C/Q	(Prakriya etal. 2006; Vig etal. 2006; Gwack etal. 2007; Yamashita et al. 2007; McNally etal. 2009)
	V107M		(Bohm etal. 2017; Bulla etal. 2019)
TM1–2 loop		H113C	(McNally etal. 2009)
TM2	H134A*/C*/E/M/P/Q/S*/T*/V	H134K/W	(Yeung etal. 2018)
	A137C/L/M/V		(Frischauf et al. 2017; Yeung et al. 2018)
	L138F	L138A	(Endo et al. 2015; Frischauf et al. 2017)
	M139V		(Frischauf et al. 2017)
	S141C		(Yeung etal. 2018)
TM2–3 loop	S159L		(Frischauf et al. 2017)
TM3		L174D/K	(Zhou et al. 2016)
		A175K	(Zhou et al. 2016)
	W176C		(Srikanth et al. 2011; Yeung et al. 2018)
	A177D		(Frischauf et al. 2017)
	G183D		(Frischauf et al. 2017)
	T184M		(Bohm etal. 2017; Bulla etal. 2019)
	L185C		(Yeung etal. 2018)
		F187A	(Yeung etal. 2018)
	E190C		(Yeung etal. 2018)
		V191N	(Yeung etal. 2018)
		L194A/N/P	(McCarl etal. 2009; Yeung etal. 2018)
TM4	A235C		(Yeung etal. 2018)
	S239C		(Yeung etal. 2018)
	P245X (any amino acid, P245L)		(Nesin et al. 2014; Palty et al. 2015; Yeung etal. 2018)
	G247S		(Frischauf et al. 2017)
	F250C		(Yeung etal. 2018)
		L261D/K	(Zhou et al. 2016)
	261 LVSHK265→ANSGA*, ANSHK,		(Zhou et al. 2016)
	ANSGK, ANSHA, AKSGA,
	AKSHK
C-terminus		L273D/R/S	(Li etal. 2007; Muik etal. 2008; Frischauf etal. 2009; McNally etal. 2013; Palty etal. 2015)
		L276D	(Li et al. 2007; Navarro-Borelly et al. 2008; Frischauf etal. 2009; McNally etal. 2013; Palty etal. 2015)

^*These GOF mutants at H134 and at the base of TM4 best recapitulate the hallmarks of Orai1 channel activity. Loci of human mutations are marked in bold italicized letters.

An interesting feature of the constitutively active mutants is that they display varying degrees of current density and ion selectivity (Yeung et al. 2018). Intriguingly, when STIM1 is co-expressed, the extent of additional STIM1-mediated channel activation and change in Ca²⁺ selectivity are inversely correlated to the mutants’ baseline activity and selectivity (Yeung et al. 2018). This finding suggests that the mutations stabilize Orai1 channels in partially open, intermediate states along the STIM1 gating pathway (Palty et al. 2017; Yeung et al. 2018). Importantly, the presence of both GOF and LOF mutations within the TMs (Fig. 4, Table 1) demonstrates that alterations in inter-TM interactions are not only sufficient for pore opening, but also essential for STIM1-mediated gating (Zhou et al. 2016; Frischauf et al. 2017; Yeung et al. 2018).

Conformational wave through the transmembrane domains

How are the structural changes that arise from STIM1 binding propagated through the TMs to open the pore? A prevailing motif in biology is that form determines function. Thus, as a first step towards understanding this question, Yeung et al. mapped the packing densities of the interhelical surfaces of the crystal structure of dOrai in its resting state (Yeung et al. 2018). In general, residues exhibiting extensive interactions contribute more to the stability and coupling between structural elements than residues with less interaction surface. As a consequence, tightly packed protein regions are generally more rigid, whereas loosely packed regions more readily undergo conformational change. When complemented with hydrophobicity maps, this analysis revealed several distinct structural features in Orai that contribute to gating. The atomic packing density in Orai varies significantly across and within the TM1–TM4 helices, with the TM2/3 helix pair displaying substantially more contacts across its interfaces compared to TM1 or TM4 (Yeung et al. 2018) (Fig. 5A). Because they are wedged between TM1 and TM4, TMs 2–3 could function as an interwoven ring to relay the gating signal from the peripheral STIM1 binding site to the central TM1 pore helix bundle (Fig. 5A). The roles of specific gating loci along this pathway will be described next starting from the C-terminus and progressing towards the pore.

Figure 5. Atomic packing analysis and hydrophobicity maps at the TM1–TM2/3 ring interface reveal two topologically distinct regions.

A, top view of the dOrai crystal structure with interlocked TM2 and TM3 helices highlighted in green and red, respectively. These two helices form an interwoven ring that may help relay the gating signal to the pore. B, space-filling representation of the interface between TM1 and the TM2/3 ring coloured by the number of contacts per residue as assessed by atomic packing analysis. C, surface representation of TM1 residues facing the TM2/3 ring coloured according to amino acid hydrophobicity. The hydrophobic cluster and serine ridge regions are noted. B and C were adapted from Yeung et al. (2018).

Gating hinges at C-terminal region ²⁶³SHK²⁶⁵ and TM4 residue P245

At the intersection between the C-terminal extensions and TM4 is a short segment consisting of residues ²⁶³SHK²⁶⁵ that forms a dramatic kink, breaking the continuity between these two α-helices (Fig. 6A). This bend has been proposed to be crucial for setting up the correct conformation of the C-terminus for STIM1 binding. Specifically, one line of thought is that Orai1 C-termini undergo only modest conformational changes during STIM1 binding, sliding against each other to accommodate a proposed interaction with the STIM1 CC2 domain (Stathopulos et al. 2013; Tirado-Lee et al. 2015; Palty et al. 2017). This motion presumably necessitates flexibility of the ²⁶³SHK²⁶⁵ domain because mutations that are predicted to alter the orientation of the bend (e.g. mutation to Gly, Pro, Trp, or disulfide bond formation) have been found to impair STIM1–Orai1 association (Palty et al. 2015; Tirado-Lee et al. 2015).

Figure 6. Gating loci along the conformational wave from the C-terminus to the pore.

Locations of various gating loci in Orai1 implicated in controlling the open–closed transitions including the C-terminal gating hinges, hydrophobic cluster, serine ridge and H134 highlighted in side views of a monomer from the dOrai crystal structure (dOrai numbering shown with corresponding Orai1 residues in parentheses). A, mutation of C-terminal bend residues ²⁶¹LVSHK²⁶⁵ to ²⁶¹ANSGA²⁶⁵ and mutation of the TM4 bend at P245 to any other residue causes the channel to become constitutively active, indicating that these bends are crucial for keeping the channel in the closed state. B, a set of interdigitating hydrophobic residues between TM3 and TM1 (L96, M101, M104, F187, V191, L194) form a hydrophobic stack that is essential for Orai1 gating by STIM1. One dOrai monomer and an additional TM1 from a neighbouring subunit are shown. C, the serine ridge (S89, S90, S93, S97) likely interacts with the polar surfaces of the TM2/3 ring and endows conformational flexibility to the inner pore. D, TM2 residue H134 acts as a steric brake at the interface between TM1 and TM2, with mutations to small residues giving rise to GOF channels and large residues leading to LOF.

In contrast to the slight shifts of the C-termini predicted by the above model, a recently published structure of the dOrai open mutant H206A (H134A Orai1; Fig. 2B) shows that, unlike the closed state where the C-terminal extensions from adjacent subunits bend sharply to form coiled-coil interactions with each other (Hou et al. 2012), the C-termini in the activated channel protrude directly into the cytosol perpendicular to the plane of the plasma membrane, culminating in a conformational change termed “unlatching” (Hou et al. 2018). Importantly, this “unlatching” step was proposed to be essential for channel activation (Hou et al. 2018). One caveat to this interpretation is that the sequence contained in the C-terminal extension helices has a high propensity to form coiled-coil interactions. Therefore, in the resting state of dOrai, the C-termini kink significantly to form antiparallel coiled-coil interactions with neighbouring subunits (Hou et al. 2012); in the H206A background, however, the need for this interaction appears to be satisfied by crystal contacts between two different channels, one upright and one upside down, with the straightened C-termini from apposing channels binding to each other (Hou et al. 2018). It is not yet clear how a substantial conformational change such as unlatching could be favoured in a physiological context or in STIM1-free GOF H134 channels in the plasma membrane. However, this process could potentially be driven by an interaction with a specific domain within STIM1 or other binding partners.

In addition to its potential role in STIM1 binding, the ²⁶³SHK²⁶⁵ locus (Fig. 6A) is also part of a gating nexus consisting of five amino acids (²⁶¹LVSHK²⁶⁵) identified by Zhou et al. to be important for controlling the open–closed transition (Zhou et al. 2016) (Fig. 6A). Unlike the single point mutations in the ²⁶³SHK²⁶⁵ bend which abrogate STIM1 binding (Palty et al. 2015; Tirado-Lee et al. 2015), mutation of these five residues to ²⁶¹ANSGA²⁶⁵ causes the channel to become constitutively active, with many properties similar to that of the STIM1-activated channel including high Ca²⁺ selectivity, blockade by 2-APB, and sensitivity to mutations in the N-terminus (Zhou et al. 2016). These properties are different from leaky gate channels produced by polar substitutions of the hydrophobic gate (i.e. at V102 or F99) in the pore which show altered pore properties and are not impaired by N-terminal mutations (McNally et al. 2012, 2013; Yamashita et al. 2017). Based on these observations, Zhou et al. postulated that STIM1 gating is mediated solely through the Orai1 C-terminus and involves inter-TM coupling at the base of TM3/TM4, specifically through a hydrophobic interaction between L174 and L261 (Zhou et al. 2016).

In addition to the drastic bend between TM4 and the C-terminal extension, the TM4 helix also contains a shallower bend in the middle of the helix due to the presence of a proline residue at position P245 (Fig. 6A). Interestingly, a human mutation was identified at this locus (P245L) linked to Stormorken syndrome arising from constitutively active Orai1 channels (Nesin et al. 2014). Palty et al. examined the mechanism of this effect and determined that, in fact, substitutions of P245 to any other amino acid lead to GOF channels (Palty et al. 2015). They speculated that this phenotype may be mediated by straightening of TM4 (Palty et al. 2015). Consistent with this hypothesis, the kink at P288 (P245 Orai1) on TM4 appears to straighten in the crystal structure of the activated mutant H206A (Hou et al. 2018) (Fig. 2B). These conformational changes at the ²⁶³SHK²⁶⁵ and P245 hinges may be required to permit rigid body movements within TMs 1–3 that culminate in pore opening (Hou et al. 2018).

TM2/3 ring as an enforcer of cooperativity in STIM1-driven channel activation

The TM2/3 ring, wedged between TM4 and TM1, is perfectly positioned to relay gating signals from the peripheral STIM1 binding site to the channel gate (Fig. 5A). Because the TM2/3 ring is tightly packed (Yeung et al. 2018), it presumably rearranges as one unit, since decoupling the interlocked TM2 and TM3 helices would require a significant amount of enthalpic energy. Yen and Lewis recently showed that the effect of STIM1 binding at a single subunit in the hexamer is much stronger than would be expected if all six subunits acted independently (Yen & Lewis, 2018). This striking observation could be explained if the effects of STIM1 binding at each of the C-termini were diffused across the TM2/3 ring and were able to influence the conformation of all six pore helices. Thus, the tight coupling of TMs 2–3 as a single functional entity (Yeung et al. 2018) could serve as a logical “AND” gate that confers strong cooperativity between STIM1 binding at six individual subunits and channel gate opening (Yeung & Prakriya, 2018) (Fig. 5A).

Hydrophobic cluster at TM1–TM2/3 ring interface

The atomic packing analysis and hydrophobicity mapping of the interface between the TM2/3 ring and TM1 reveals two functionally distinct features (Yeung et al. 2018) (Fig. 5B and C). First, a prominent cluster of interdigitating bulky hydrophobic residues lining the external part of this interface (L96, M101, M104, F187, V191, L194) forms a tightly packed and therefore presumably rigid “hydrophobic cluster” (Yeung et al. 2018) that connects the TM1 pore helix to the surrounding TM2/3 ring (Fig. 5B and C). Because this hydrophobic cluster is located directly across from the pore-facing V102/F99 hydrophobic gate, it is ideally situated to control opening and closing of the channel gate (Fig. 6B). Mutation of these residues to small or polar residues abrogate channel gating despite normal levels of STIM1 binding (Yeung et al. 2018). This result suggests that although STIM1 is able to bind to these mutant channels, disruption of this hydrophobic stack prevents the gating signal from being effectively relayed from the TM2/3 ring to the pore and that this stack is an essential component of the STIM1-mediated conformational wave to activate Orai1 channels.

Serine ridge at TM1–TM2/3 ring interface

The second structural feature revealed by the atomic packing analysis is a “serine ridge” formed by strips of serine residues (S89, S90, S93, S97) lining the cytoplasmic surface of the pore helix (Yeung et al. 2018) (Figs 5B and C, and 6C). The serine ridges of the hexamer are partitioned by hydrophobic residues such as L96, causing the surface on the back of TM1 to be lined by alternating stripes of polar and hydrophobic residues (Fig. 5B and C). Correspondingly, these ridges are complemented by alternating stripes of polar–non-polar residues on the inner surface of the TM2/3 ring facing the pore helices (Yeung et al. 2018). This feature is particularly intriguing because a human mutation S97C, associated with tubular aggregate myopathy, exists in this serine ridge (Garibaldi et al. 2017). We postulate that S97C activates the channel by disrupting the polar interactions between TM1 and the TM2/3 ring that stabilize it in the closed state (Yeung et al. 2018). In addition, the relatively low packing density of the serine ridges could endow conformational flexibility to the inner pore required for optimal channel gating (Yeung et al. 2018).

TM2 gating locus H134

Perhaps the best-studied gating locus at the TM1–TM2/3 ring interface is at TM2 residue H134 which faces the non-pore-lining surface of TM1, approximately halfway into the membrane spanning region (Frischauf et al. 2017; Hou et al. 2018; Yeung et al. 2018) (Fig. 6D). Like the ²⁶¹ANSGA²⁶⁵ mutation at the C-terminal hinge, channels activated by GOF H134 mutations recapitulate several core features of STIM1-gated channels. Because H134A/C/S/T channels conduct large, Ca²⁺ selective currents that are not further enhanced by STIM1, they are thought to be one of the most intrinsically active open mutants described thus far (Frischauf et al. 2017; Yeung et al. 2018). Frischauf et al. first reported several substitutions at H134 that led to GOF channels and postulated that H134 forms a hydrogen bond with TM1 which stabilizes the closed conformation and whose disruption flips the channel into an open state (Frischauf et al. 2017). By contrast, Yeung et al. proposed that H134 acts as a steric brake, noting that channel activity was dependent on the side-chain surface area and independent of the introduced residue’s predicted ability to form hydrogen bonds with TM1 (Yeung et al. 2018). In the latter model, H134 is likely hydrogen bonded to the carbonyl backbone of TM2 residue L130 located one turn above, which stabilizes the H134 side chain towards the TM1–TM2/3 ring interface in the closed state. A conformational change mimicking the release of this intra-helical tether following STIM1 binding that significantly alters the H134 rotamer orientation could lead to pore opening, but the specific steps of this last step remain to be elucidated.

How do mutations at H134 activate the channel gate? The H206A dOrai (H134A Orai1) structure revealed conformational changes throughout the entire channel (Fig. 2B). In addition to the C-terminal unlatching and inner pore splaying described above, Hou et al. also noted a slight dilatation at the outer pore based on the presence of Ba²⁺ binding in the plane of E106 in the open state, instead of 4 Å above the plane of E106 in the closed state (Hou et al. 2018). Unfortunately, because the resolution of the structure was not sufficient to identify the rotameric conformations of individual side chains (Hou et al. 2018), while it shows global structural changes associated with the H134A mutant, it does not provide a definitive gate opening mechanism in the pore.

Using cysteine accessibility analysis, Yeung et al. found that the H134S mutation opens the hydrophobic gate via pore helix rotation (Fig. 3B) and slight dilatation of the outer pore, analogous to the structural change mediated by STIM1 gating (Yamashita et al. 2017; Yeung et al. 2018). This experimental finding was supported by the results of MD analysis which showed significantly increased pore helix rotation and displacement of F99 away from the pore, and increased water occupancy in the pore when a H206A mutation was introduced into the 4HKR structure (Yeung et al. 2018). On the other hand, Frischauf et al. used MD analysis of a humanized version of the 4HKR structure to report that the H134A mutation causes the R91 side-chain to be displaced away from the pore via hydrogen bonding with the S90 of neighbouring subunits (Frischauf et al. 2017). Whether this occurs during STIM1 gating remains unclear, since some mutants that should not be able to form R91–S90 hydrogen bonds, such as R91G and the S89G/S90G double mutant, remain store-operated (Derler et al. 2009). This result raises the possibility that the S90–R91 bond observed in H134A channels may not be required for STIM1-mediated channel activation.

The most prominent structural changes observed in the H206A structure, however, namely inner pore dilatation and C-terminal unlatching (Hou et al. 2018) (Fig. 2B), await further validation. Future experiments that could potentially validate these conformational changes include examining the accessibility of the pore to MTS reagents of varying sizes in the H134A channel to map its architecture and compare it with those of Orai1 V102A leaky gate channels as well as the WT Orai1 STIM1-gated channel (McNally et al. 2009), and determine whether the ability of the L273–L276 pair to interact with each other is altered in activated channels as would be expected in the unlatched conformation of the C-termini.

N-terminus and TM2–3 loop

Regardless of whether or not STIM1 binds to the Orai1 N-terminus, it is clear that this region is critical for several hallmarks of CRAC current, including STIM1-dependent gating, fast Ca²⁺-dependent inactivation, and increased permeation to Na⁺ in divalent-free solutions (Li et al. 2007; Derler et al. 2013, 2018; McNally et al. 2013; Zheng et al. 2013; Palty et al. 2015; Mullins et al. 2016). Because the activated states of channels typically have higher affinity for ligands than the resting states (Colquhoun, 1998), the apparent defect in STIM1 binding in N-terminally mutated channels could, in principle, arise from changes in gating that trap channels in states with lower STIM1-binding affinity. How then does the N-terminus contribute to gating?

One possibility is that the N-terminus may contribute to the transmission of the gating signal from the C-terminus to the pore by interacting with other regions on the cytosolic surface of Orai1 such as the TM2–3 loop. Fahrner et al. observed that the loss of current in Orai1 N-terminal truncation mutants (Δ1–78) can be partially restored by replacing the native Orai1 TM2–3 loop with that of Orai3 (Fahrner et al. 2018). They postulated that the inhibition of current seen in the N-terminally truncated Orai1 channels occurs because of an interaction between the truncated Orai1 N-terminus and the TM2–3 loop (Fahrner et al. 2018). Consistent with this idea, cysteine cross-linking between the N-terminus and the TM2–3 loop inhibited currents in STIM1-gated and P245L constitutively active mutant channels (Fahrner et al. 2018). This interesting finding suggests that an Orai1 N-terminus–TM2–3 loop interaction inhibits gating. Along the same vein, a recent study by Kim et al. postulated that in the context of Orai in Caenorhabditis elegans, the N- and C-terminal tails, which are bound to the TM2–3 loop at rest, are displaced during STIM1 binding (Kim et al. 2018). They proposed that this unbinding step between the TM2–3 loop and the N- and C-termini permits channel opening (Kim et al. 2018). Taken together, both studies imply that an interaction between the N-terminus and the TM2–3 loop favours the closed state whereas release of this interaction biases the channel towards an open configuration. However, it is still not known whether STIM1-driven Orai1 channel activation involves decoupling of these domains.

Molecular basis of the allosteric gating process

One of the most widely utilized frameworks for understanding allosteric protein regulation is the Monod–Wyman–Changeux (MWC) model, which posits two core ideas: (1) that allosteric proteins are symmetrical oligomers with the conformations of each protomer constrained by its interface with neighbouring subunits, and (2) that allosteric proteins can reversibly access multiple states with varying conformational constraints placed at the inter-protomer interfaces and that transitions between the different states involve cooperative interactions between subunits that preserve molecular symmetry across the protein (Changeux, 2012). Because ion channel pores are typically formed by multimers with an axis of symmetry along the pore, these tenets of the MWC model have served as a unifying theme exemplified across various families of ion channels and receptors (Changeux & Christopoulos, 2016).

The studies presented in earlier sections of this review have recently brought the concept of allosteric gating for Orai1 to the forefront. Orai1 channels have a unique overall architecture of three concentric rings of TMs with sixfold symmetry in the pore and threefold symmetry at the outermost ligand binding sites (Hou et al. 2012). Within the MWC framework of allostery, the overlapping C-terminal extensions and the densely packed inter-subunit interfaces along the TM2/3 ring serve as strategic areas for the Orai1 protomers to sense and respond collectively to STIM binding. Furthermore, the model predicts that the extent of cooperativity is determined by the free energy of the inter-protomer bonds between subunits, which can range from a linear graded response with little cooperativity to an all-or-none transition for highly cooperative protein complexes (Changeux, 2012). For Orai1, we propose that the TM2/3 ring, which is lined by interwoven large, hydrophobic residues and situated at the transition zone between the peripheral ligand binding sites and the pore (Yeung et al. 2018), serves as the enforcer of the extremely non-linear relationship between STIM1 binding and channel gating (Scrimgeour et al. 2009; Hoover & Lewis, 2011; Yen & Lewis, 2018). The TM2/3 ring, potentially along with other inter-subunit features that have yet to be described, likely stabilizes the closed and open states of the pore, accounting for the abrupt switching in Orai1 channels from a gating mode with an extremely low P_o to one with a high P_o (Prakriya & Lewis, 2006). Moreover, because STIM1-activated Orai1 channels may be at equilibrium between multiple states corresponding to different free energies, mutations within the TMs that disrupt the inter-subunit interfaces or affect the overall free energy landscape will inevitably lead to GOF or LOF phenotypes as observed through numerous recent studies (Palty et al. 2015; Zhou et al. 2016; Frischauf et al. 2017; Yeung et al. 2018; Bulla et al. 2019).

Conclusions and future directions

Like many ligand-gated ion channels, Orai1 channels are activated through a highly allosteric process. Recent studies have collectively pointed towards the propagation of a conformational wave originating from the distal STIM1 binding site at the Orai1 C-terminus through numerous gating hotspots within the non-pore-lining regions (e.g. C-terminal gating hinges, hydrophobic cluster, serine ridge, TM2 residue H134, TM2–3 loop) (Fig. 6) to ultimately open the channel pore (Palty et al. 2015; Zhou et al. 2016; Frischauf et al. 2017; Fahrner et al. 2018; Yeung et al. 2018). Mutation of various gating loci along this pathway, including several pathological human mutations (Nesin et al. 2014; Endo et al. 2015; Garibaldi et al. 2017; Bohm et al. 2017), lead to GOF or LOF channels (Palty et al. 2015; Zhou et al. 2016; Frischauf et al. 2017; Yeung et al. 2018; Bulla et al. 2019) (Fig. 4). Because some of these mutations may trap the channel in intermediate states, this growing library of Orai1 mutations is valuable in studying the channel activation mechanism as well as for drug development.

The tremendous potential of this emerging database could soon be harnessed by the ongoing breakthroughs in cryo-electron microscopy and other biophysical and computational methods. In particular, MD simulations of the different variants has become a powerful tool to study the dynamic positions and interactions of all atoms in Orai1 and determine free energy and hydration profiles in many channel states in a way that electrophysiology and traditional structural biology methods cannot (reviewed in Bonhenry et al. 2019). Although the specific structural rearrangements within the TMs that accompany Orai1 gating are currently unknown, there is reason to believe that the structure of the STIM1-gated channel as well as the dynamics of channel gating will be within reach in the coming years. These exciting milestones would mark significant steps forward in the pursuit of mechanism-based drug discovery targeting CRAC channels.

Biography

Priscilla Yeung is a MD/PhD student and Postdoctoral Fellow at Northwestern University who recently completed her PhD (Life Sciences) in the Prakriya laboratory. Prior to arriving at Northwestern, she obtained a BA in the Biological Basis of Behavior from the University of Pennsylvania. Her research focus is on the molecular mechanisms of Orai1 channel activation. Megumi Yamashita is a Research Assistant Professor at Northwestern University. Her work examines the gating mechanisms of store-operated Ca²⁺ channels. Murali Prakriya is a Professor at Northwestern University Feinberg School of Medicine. The Prakriya laboratory examines the molecular and cellular mechanisms by which store-operated Ca²⁺ channels are activated and their roles in the biology of immune cells, airway epithelial cells, and neurons and astrocytes.