Introduction

Calcium (Ca²⁺) is a ubiquitous second messenger with central roles in signal transduction (1). One of the predominant Ca²⁺ influx pathways into the cytosol of non-excitable cells is the store-operated Ca²⁺ entry (SOCE) pathway (2–6). SOCE is mediated by Ca²⁺ release–activated Ca²⁺ (CRAC) channels (7, 8), which are formed by the pore-forming subunits Orai1/2/3 (9), and activated by the endoplasmic reticulum (ER)-resident Ca²⁺ sensors STIM1 and STIM2 (10, 11) when ER Ca²⁺ stores are depleted. Each of the three Orai and the two STIM proteins is encoded by a distinct gene. Further, Orai1 exists as two alternatively translated variants, a longer, mammalian-specific isoform (Orai1α) and shorter isoform (Orai1β) lacking the first N-terminal 63 amino acids (12, 13). Upon ER Ca²⁺ store depletion, STIM proteins oligomerize, move to plasma membrane-ER contact sites and gate Orai channels, enabling Ca²⁺ influx from the extracellular space to replenish ER stores and to activate cellular signaling pathways that couple to transcriptional and metabolic activities (14). The Ca²⁺ microdomains generated at the cytosolic mouth of CRAC channels are decoded by a network of Ca²⁺-responsive signaling pathways, culminating in the activation of multiple transcription factors (15, 16). Among these, nuclear factor of activated T cells (NFAT) is well-characterized as a canonical downstream target for SOCE (17–19). In resting cells, NFAT is retained in the cytoplasm through phosphorylation that masks its nuclear localization sequence. Ca²⁺ influx through Orai channels activates the phosphatase calcineurin, which dephosphorylates NFAT, enabling its nuclear translocation and thereby linking CRAC-mediated Ca²⁺ influx to gene programs that orchestrate diverse biological processes, including proliferation, immune activation, cytokine production, and metabolic regulation (14, 20–23). The NFAT family comprises five members, four of which (NFAT1–4) are activated in a Ca²⁺-dependent manner and display distinct activation kinetics: NFAT4 nuclear translocation occurs rapidly in response to modest cytosolic Ca²⁺ rise, whereas NFAT1 requires stronger and more prolonged cytosolic Ca²⁺ signals (24–26).

Orai homologues share a common activation mechanism, but they diverge in tissue expression, regulation and biophysical properties, including their susceptibility to fast Ca²⁺-dependent inactivation (CDI), from the high fast CDI of Orai3 to the minimal fast CDI of Orai1β (12, 27, 28). These differences, together with their ability to assemble in varying stoichiometries, potentially enable native CRAC channels to fine-tune Ca²⁺ signal amplitude, duration and dynamics, thereby broadening the bandwidth of Ca²⁺ signaling events (20, 29–31). For instance, knockdown of Orai2 or Orai3 enhances SOCE, suggesting that these isoforms act as regulatory brakes on CRAC channel activity (30, 32, 33). Yet, how the unique Ca²⁺ signatures of individual Orai isoforms shape the transcriptional landscape remains largely unknown. Further, whether Orai1α and Orai1β isoforms are each capable of supporting SOCE-dependent transcriptional and functional responses in a physiological context such as in primary CD4⁺ T cells has not been examined. The longer Orai1α isoform contains several unique motifs and residues shown to affect SOCE itself and SOCE-dependent downstream signaling. Substitution of serines 27 and 30, putative protein kinase C phosphorylation sites unique to Orai1α, resulted in increased SOCE suggesting that Orai1α phosphorylation on S27/S30 regulates CRAC channel function (34). Moreover, phosphorylation of serine 34 following Ca²⁺ influx through Orai1α was shown to mediate CRAC channel inactivation in a cyclic AMP and protein kinase A-dependent manner (35). Orai1α contains a caveolin consensus site between residues 52-60 that have been shown to mediates its internalization from the plasma membrane during oocyte meiosis (36). Previous studies also proposed that residues 39–59 of Orai1α can bind the scaffolding protein AKAP79, thus coupling SOCE to calcineurin and NFAT activation (37, 38).

Here we set out to analyze side by side all four Orai isoforms (Orai1α/1β/2/3) at their native levels of expression in HEK293 and primary CD4⁺ T cells for their ability to mediate SOCE, activate NFAT isoforms, and regulate gene expression. Using HEK293 cells natively expressing only one Orai homologue, we found that Orai1 is the principal driver of NFAT1/4 activation whereas Orai2 and Orai3 less efficiently supported NFAT1/4 nuclear translocation. Comparing Orai1α and Orai1β isoforms, we found that both isoforms can promote NFAT nuclear translocation in HEK293 cells across diverse activation conditions with Orai1β-mediating significantly bigger SOCE and NFAT1/4 nuclear translocation compared to Orai1α. Expression of Orai1α and Orai1β at near native levels in primary Orai1^−/− CD4⁺ T cells showed that both restored SOCE, NFAT translocation, cytokine production and gene expression without appreciable differences. An analysis of genetic and clinical data showed that human individuals with homozygous null mutations unique to Orai1α who only express Orai1β protein are healthy in contrast to patients with null mutations in residues shared by both Orai1α and Orai1β(39). Moreover, the analysis of T cells from patients heterozygous or homozygous for an Orai1α null mutation showed increased SOCE and NFAT1 activation compared to healthy controls expressing both isoforms. Collectively, our findings demonstrate that all three Orai homologues and both Orai1 isoforms can mediate NFAT activation and cell functions, albeit with different efficiencies that correlate with the magnitude of SOCE through these channels.

Results

Native Orai1 promotes more efficient NFAT1/4 nuclear translocation upon maximal store depletion compared to native Orai2 and Orai3

Ca²⁺ influx through Orai channels generates spatially restricted cytosolic Ca²⁺ microdomains that couple SOCE to transcriptional regulation, controlling multiple gene expression programs (15, 22, 40–43). To dissect Orai homologue-specific contributions to NFAT activation, and ultimately gene expression, we used CRISPR/Cas9-engineered HEK293 clones lacking two Orai isoforms (double knockout, Orai-DKO), thereby retaining expression of the third Orai isoform at endogenous expression levels. These knockout models were previously validated at the genomic and transcript levels, with no compensatory changes in expression of STIM1, STIM2 or IP₃ receptors (30, 35). NFAT1 and NFAT4 were shown previously to require different types of Ca²⁺ signals with regard to magnitude and source of Ca²⁺ (30, 35, 44). We therefore expressed NFAT1-GFP and NFAT4-GFP reporters in each DKO line, and tracked their localization by live imaging following stimulation with 2 μM thapsigargin in the presence of extracellular Ca²⁺ to induce strong, sustained SOCE. Nuclear NFAT accumulation was quantified as the nuclear-to-cytoplasmic GFP ratio over time on a cell-by-cell basis. Native Orai1 induced the most robust NFAT nuclear translocation compared to Orai2 or Orai3 with regard to both the maximum level and kinetics of translocation (Fig. 1A-H). Orai1 was more efficient than Orai2 and Orai3 in inducing the translocation of both NFAT1 and NFAT4. Of note, under thapsigargin stimulation Orai2 was more efficient at inducing NFAT4 nuclear translocation compared to Orai3. These findings demonstrate that under conditions of maximal SOCE induction, Orai1 is the main driver of both NFAT1 and NFAT4 nuclear translocation.

Figure 1.

Orai1 supports robust NFAT4 nuclear localization upon physiological receptor stimulation

To determine the differential roles of Orai homologues in NFAT nuclear translocation under more physiological conditions of activation, we next stimulated cells with the G-protein coupled receptor (GPCR) agonist carbachol (CCh) at 10 µM, a relatively low dose that typically generates robust oscillations in HEK293 cells (29, 30, 45). Although these oscillations are initiated by IP₃R-mediated release, CRAC channel activity is essential for sustaining them and for driving downstream transcriptional programs (29, 30). Because NFAT1 nuclear translocation requires robust and sustained CRAC channel activity, low concentrations of GPCR agonists do not cause detectable activation of NFAT1. Consistent with this, we previously showed that 10 µM CCh fails to induce NFAT1 nuclear translocation in HEK293 cells (35). Therefore, we focused on NFAT4. Using the homologue-specific Orai-DKO cell lines, we expressed NFAT4-GFP and monitored its localization by live imaging following stimulation with 10 µM CCh. Native Orai1 mediated robust NFAT4 nuclear accumulation compared to Orai2 and Orai3 (Fig. 1I–L). These findings demonstrate that Orai1 is the principal Orai family member coupling SOCE to NFAT4 nuclear localization after receptor-mediated stimulation of SOCE.

Orai1α and Orai1β both promote NFAT1/4 nuclear localization

We next asked whether the two alternatively translated Orai1 isoforms, Orai1α and Orai1β, differ in their ability to activate NFAT1/4 nuclear translocation. Orai1β arises from an alternative translation start site (TSS) at methionine 64, producing a shorter isoform lacking the distal N-terminal 63 amino acids (Fig. 2A). To directly compare the effects of both Orai1 isoforms, we used HEK293 triple Orai knockout (Orai-TKO) cells lacking all three Orai homologues and reintroduced either Orai1α or Orai1β cDNA plasmid. To achieve near-native levels of expression, Orai1α and Orai1β cDNA constructs were under the control of the relatively “weak” chicken thymidine kinase (TK) promoter. This system offers several advantages over conventional plasmid overexpression driven by the robust cytomegalovirus (CMV) promoter. Using both immunofluorescence and Western blotting, we show by comparison to CMV-driven plasmids, TK-driven plasmids achieve near-endogenous levels of Orai isoform expression (Fig. S1). TK-driven expression achieves more physiologically relevant expression levels and better preserves Orai-STIM stoichiometry, thereby eliminating the need for STIM1 co-expression and avoiding artifacts associated with excessive cDNA loading (35). This is important because overexpression of Orai/STIM genes has been shown to obscure Orai isoform-specific differences and alter the effects of endogenous SOCE regulators (12, 46, 47).

Figure 2.

We studied side by side only cells that show similar expression levels of Orai1α, Orai1β, Orai2 and Orai3 (Fig. S1D). We compared the ability of these four channel proteins to promote the activation of NFAT1 and NFAT4 in response to weak and strong induction of SOCE. NFAT1-GFP and NFAT4-GFP reporter assays revealed that both Orai1α and Orai1β efficiently supported NFAT1 and NFAT4 nuclear translocation upon stimulation with 2 μM thapsigargin (Fig. 2B–I). A quantitative analysis confirmed significant increases in both the maximal nuclear/cytoplasmic ratio and the slope of translocation kinetics for both Orai1α and Orai1β isoforms compared to Orai2, Orai3, or Orai-TKO controls. Of note, Orai1β was moderately but significantly more efficient at mediating NFAT translocation compared to Orai1α (Fig. 2C, D, G, H), which is consistent with the enhanced fast Ca²⁺-dependent inactivation (CDI) of Orai1α compared to Orai1β (12). Using more physiological activation conditions by stimulation with 10 µM CCh yielded similar results. Orai1α and Orai1β both effectively promoted NFAT4 nuclear localization with Orai1β showing significantly faster and more complete NFAT4 nuclear translocation (Fig. 2J–M). Furthermore, ectopic Orai2 expression in Orai-TKO cells generally supported higher NFAT nuclear translocation compared to Orai3, although this difference varied depending on the NFAT isoform and the condition of stimulation considered (Fig. 2B-M). These results demonstrate that both Orai1α and Orai1β effectively couple SOCE to NFAT1/4 activation under conditions of maximal store depletion and physiological receptor stimulation.

The magnitude of NFAT nuclear translocation is commensurate with SOCE levels and is not Orai isoform specific

To determine whether the magnitude and kinetics of NFAT activation reflect differences in Ca²⁺ entry mediated by Orai1α, Orai1β, Orai2, and Orai3, we directly measured SOCE in Orai-TKO cells ectopically expressing individual Orai homologues and isoforms using TK promoter-driven constructs. Equivalent expression levels of YFP-tagged proteins in individual cells were validated (Fig. S1B-D). SOCE, which in this situation is supported by native STIM proteins, was measured following maximal store depletion with thapsigargin. These recordings showed that Orai1β expression resulted in the largest SOCE magnitude, followed by Orai1α, Orai2, and Orai3 (Fig. 3A–B). This rank order tracked closely with NFAT nuclear translocation (Fig. 2), suggesting that NFAT responses reflect the relative Ca²⁺ influx mediated by each Orai isoform.

Figure 3.

If NFAT activation is mainly governed by the magnitude of SOCE, we reasoned that the homologue mediating the lowest SOCE, Orai3, might still support robust NFAT activation if it is highly expressed under native conditions. In MCF7 breast cancer cells SOCE is mediated predominantly by Orai3, which is highly expressed and contributes to breast cancer cell proliferation and survival (48–53). MCF7 cells offered a model with high expression of native Orai3 that are functional as SOCE channels while preserving endogenous STIM expression and avoiding potential overexpression artifacts. To ensure that the bulk of SOCE in MCF7 cells is mediated by Orai3 and to rule out contributions from Orai1 to SOCE, we deleted Orai1 using CRISPR/Cas9 gene editing (MCF7-O1KO)(27) (Fig. 3C). Consistent with our previous findings, SOCE was only slightly reduced in the absence of Orai1 (Fig. 3D-E), confirming that Orai3 is the major contributor to CRAC channel activity in MCF7 cells (48, 51, 52, 54). We then monitored NFAT1 nuclear translocation and found that following thapsigargin stimulation MCF7-O1KO cells sustained robust NFAT1 activation comparable to that of wildtype MCF7 cells (Fig. 3F–G). As a negative control, wildtype MCF7 cells pretreated with 20 μM BAPTA-AM to chelate cytosolic Ca²⁺ showed no NFAT1 nuclear translocation (Fig. 3F-G). These data demonstrate that abundant, native Orai3 is able to promote NFAT1 nuclear translocation, supporting the conclusion that NFAT1 activation scales with the capacity of individual Orai homologues to mediate SOCE.

Orai1α- and Orai1β-dependent NFAT activation is governed by Ca²⁺ microdomains at the mouth of the channel

Previous studies have shown that Ca²⁺ signals can be decoded by Ca²⁺-binding proteins in the vicinity of Ca²⁺ channels and relayed to the nucleus to achieve distinct transcriptional patterns (55, 56). Given the dominant role of Orai1 in NFAT activation in most cell types, we asked whether the SOCE-dependent nuclear translocation of NFAT1 is regulated by the magnitude of the global Ca²⁺ signal or localized Ca²⁺ signals generated at the mouth of Orai channels. To this end, we studied the impact of slow and fast Ca²⁺-binding chelators, namely membrane-permeant forms of EGTA-AM and BAPTA-AM respectively, on Orai isoform dependent SOCE and subsequent NFAT1 nuclear translocation. Both chelators have similar affinities for Ca²⁺ but differ in their binding rates. EGTA’s slower binding kinetics allows a region of relatively unbuffered intracellular Ca²⁺ concentration at the vicinity of individual Ca²⁺ channels. On the other hand, BAPTA’s faster rate results in a sharp decrease in Ca²⁺ concentrations within a small distance away from a Ca²⁺ channel (57, 58). Therefore, BAPTA is likely to inhibit both local and global Ca²⁺ signals, while EGTA suppresses only the global Ca²⁺ signal (Fig. 4A). These differential effects have been extensively employed to investigate local channel-effector coupling in diverse contexts (55, 56, 59–61). We pre-loaded Orai-TKO HEK293 cells expressing either Orai1α or Orai1β with EGTA-AM or BAPTA-AM and we then analyzed SOCE and NFAT1-GFP nuclear translocation. Both EGTA and BAPTA completely suppressed the global Ca²⁺ increase after thapsigargin stimulation of cells expressing either Orai1α or Orai1β (Fig. 4B-C). By contrast, only BAPTA, but not EGTA, strongly interfered with the nuclear translocation of NFAT1 in response to thapsigargin stimulation (Fig. 4D-I). These findings, which are consistent with earlier reports (26, 37, 62, 63), indicate that NFAT1 activation is regulated by localized Ca²⁺ signals in the vicinity of the CRAC channel pore. Of note, the inhibitory effect of BAPTA on NFAT1-GFP translocation was observed in cells expressing either Orai1α or Orai1β indicating that it is not isoform specific.

Figure 4.

The mechanisms underlying regulation of NFAT activation by local Ca²⁺ signals mediated by Orai channels are not well understood. In neurons, the binding of the Ca²⁺ sensor calmodulin to voltage-gated Ca²⁺ channels is a known mechanism to recruit calmodulin kinases and activate the transcription factor CREB near the channel pore (55, 64). A recent study showed that Orai1 may use a similar mechanism that includes binding of the adaptor AKAP79 to a 21 amino acid region (39–59) in the N-terminus of Orai1α, thus facilitating AKAP79-dependent recruitment of calcineurin to Orai1 channels and NFAT1 activation (38). Because the AKAP79-binding region is missing in Orai1β, we hypothesized that NFAT1 activation would be reduced in cells expressing Orai1β compared to Orai1α. Contrary to this hypothesis, nuclear translocation of NFAT1 and NFAT4 was comparable in Orai1β and Orai1α expressing cells (Fig. 2). Collectively, our findings indicate that while local Ca²⁺ signals near the CRAC channel pore mediate the activation of NFAT, they are unlikely to be decoded by the N-terminal region unique to Orai1α.

Orai1α and Orai1β effectively promote NFAT activation and cytokine production in primary CD4⁺ T cells

We next sought to determine if Orai1α and Orai1β show differences in their ability to promote NFAT nuclear translocation and NFAT-dependent gene expression in cells in which Orai1 is essential for function. Orai1 and NFAT are indispensable for T cell activation, cytokine production, and effector functions (16, 65). Native Orai1α and Orai1β proteins can be resolved using Western blotting of deglycosylated proteins samples as described previously(12, 13). Primary naïve CD4⁺ T cells as well as CD4⁺ T cells differentiated into either Th1, Th2, regulatory (Treg), conventional Th17 (cTh17) or pathogenic Th17 (pTh17) express both Orai1α and Orai1β isoforms with varying ratios (Fig. S2), making T cells an excellent model to probe the functional significance of Orai1α and Orai1β signaling. To this end, we isolated primary murine CD4⁺ T cells from the spleens of Orai1^fl/fl Cd4Cre mice (Orai1^CD4) with T cell specific deletion of both Orai1α and Orai1β (66). T cells were retrovirally transduced with either Orai1α or Orai1β in plasmids containing the fluorescent reporter Ametrine and cultured under T helper (Th) 0 or Th1 polarizing conditions (Fig. S3A). Ametrine (Amt) fluorescence correlated well with Orai1α and Orai1β expression levels detected by antibody staining (Fig. S3B). This allowed us to use Amt reporter fluorescence as a proxy for Orai1 levels, which is important to ensure similar expression levels of Orai1α and Orai1β. Because Orai1 overexpression suppresses SOCE – likely due to an increased Orai1:STIM1 ratio (67, 68) We gated on Amt^high and Amt^low cells to control for Orai1 expression levels (Fig. S3C-D). Measurements of intracellular Ca²⁺ concentrations using Indo-1 showed that Orai1^CD4 T cells expressing high levels of either Orai1α or Orai1β (Amt^high) had low thapsigargin-induced SOCE similar to cells expressing EV control (Fig. S3F-G). By contrast, T cells expressing low levels of Orai1 (Amt^low) showed increased SOCE compared to mock-transduced cells (Fig. 5A).

Figure 5.

We next examined the ability of Orai1α and Orai1β expressed in Orai1^CD4T cells to induce nuclear translocation of native NFAT1. Amt^low CD4⁺ T cells were enriched by flow cytometric cell sorting and stimulated with 1 μM ionomycin for 30 min or left untreated. Permeabilized cells were stained with anti-NFAT1 antibody (clone D43B1; Cell Signaling Technology, Cat. #14324) and analyzed by ImageStream. Unstimulated T cells had predominantly cytosolic NFAT1, whereas stimulation induced robust nuclear translocation in Orai1^CD4 T cells transduced with Orai1variants but not EV (Fig. 5B-C). Importantly, no differences regarding nuclear to cytoplasmic NFAT1 ratios were observed between Orai1α- and Orai1β-transduced CD4⁺ T cells. Because NFAT is essential for the transcriptional regulation of cytokine production in CD4⁺ T cells, we analyzed the levels of IL-2, TNF and IFN-γ in Orai1^CD4T cells reconstituted with either Orai1α or Orai1β. CD4⁺ T cells were stimulated with phorbol 12-myristate 13-acetate (PMA) and ionomycin for 6h followed by intracellular cytokine staining and analysis of Amt^low CD4⁺ T cells by flow cytometry. Orai1 reconstitution in Orai1^CD4 T cells significantly increased the expression levels of IL-2, TNF, and IFN-γ compared to mock-transduced cells and the frequencies of cytokine expressing cells (Fig. 5D-E). Importantly, no significant differences were observed between the ability of Orai1α or Orai1β to restore cytokine production in Orai1^CD4 T cells. Together these findings indicate that both Orai1 isoforms are able to effectively promote NFAT1 activation and cytokine gene expression in CD4⁺ T cells.

Orai1α and Orai1β regulate similar transcriptional programs in primary CD4⁺ T cells

SOCE and NFAT regulate many other transcriptional programs in T cells besides cytokine expression including metabolism (14, 69, 70), CD4⁺ T cell differentiation (16, 70–74), anergy and exhaustion (75–77). To determine the broader transcriptional consequences of Orai1α and Orai1β signaling in CD4 T cells, we conducted RNA-sequencing of primary murine CD4⁺ T cells from Orai1^CD4 mice that were reconstituted with Orai1α, Orai1β, or EV. Cells were either left untreated or stimulated by anti-CD3/CD28 stimulation for 6h. We first confirmed isoform-specific expression of Orai1 by aligning sequencing reads to exons 1 and 2 of the Orai1 gene (Fig. 6A). Principal component analysis (PCA) showed a clear separation between unstimulated and stimulated T cells (Fig. 6B). Within the stimulated samples, Orai1α- and Orai1β-expressing Orai1^CD4 CD4⁺ T cells clustered closely together and away from EV-transduced Orai1^CD4 T cells and WT controls. The analysis of differentially expressed genes (DEG) identified hundreds of genes that were either up- or downregulated in their expression in Orai1α- and Orai1β transduced Orai1^CD4 CD4⁺ T cells compared to EV controls (Fig. 6C). A direct comparison of DEGs in cells expressing either Orai1α or Orai1β showed almost no differences between the two isoforms. The total number of DEGs either up- or downregulated in Orai1α transduced cells was slightly larger than in cells expressing Orai1β (Fig. 6C-D). A correlation analysis of DEGs in T cells expressing either Orai1α or Orai1β, however, showed that log₂ fold-changes (FC) of gene expression were nearly identical between both isoforms (R ≈ 0.92; Fig. 6E), indicating that the observed differences in DEG counts are primarily an artifact of statistical thresholding rather than evidence of major biological divergence between the isoforms. Pathway enrichment analyses confirmed that both Orai1 isoforms regulated the same gene expression programs, including interferon responses, cytokine signaling, and other immune activation pathways (Fig. 6F). Moreover, the analysis of the most upregulated genes revealed that expression levels of DEGs in Orai1α and Orai1β transduced CD4⁺ T cells were strongly induced compared to EV controls but highly similar to each other (Fig. 6G). These include known NFAT-regulated cytokine genes such as IL-3, IL-10 and IFN-γ. Taken together, these results show that Orai1α and Orai1β drive nearly indistinguishable global transcriptional signatures in T cells.

Figure 6.

Homozygosity for a common frameshift mutation selectively abolishing Orai1α expression is not associated with the classical CRAC channelopathy phenotype

Homozygous null mutations in the ORAI1 gene that either abolish protein expression or impair CRAC channel function cause immunodeficiency associated with recurrent severe, often lethal infections early in life (39, 78). Patients successfully treated by hematopoietic stem cell transplantation develop a characteristic set of non-immunological symptoms including amelogenesis imperfecta, anhidrosis and muscular hypotonia. Notably, all known ORAI1 null mutations causing CRAC channelopathy are located in residues shared by Orai1α and Orai1β and thus affect both isoforms (Fig. 7A). To gain information about the physiological role of Orai1α in humans, we used population genetic databanks to identify and analyze genetic variants in the N terminus of Orai1 that exclusively affect Orai1α. We identified two frameshift mutations confined to this region (p.P43TfsX45 and p.P46SfsX42) that are predicted to selectively disrupt Orai1α without affecting Orai1β (Fig. 7A). Using deglycosylation and Western blotting to separate Orai1α and Orai1β protein bands (see methods), we confirmed that overexpression of the Orai1 p.P43TfsX45 variant in HEK293 cells indeed resulted in the loss of Orai1α protein while preserving the expression of Orai1β, which is translated from M64 downstream of the mutation (Fig. 7B). Analysis of population genetic datasets including the Genome Aggregation Database (gnomAD), UK Biobank (UKBB), Qatar Biobank (QBB), and the All of Us (AoU) Research Program revealed that both N-terminal truncating variants are present at higher allele frequencies compared to mutations affecting regions shared by both isoforms (Fig. 7C). Homozygous carriers of the Orai1α-specific truncating variants p.P43TfsX45 and p.P46SfsX42 were identified across multiple population cohorts, including 43 and 41 individuals in gnomAD, 2 and 0 individuals in QBB, and 8 and 8 individuals in AoU, respectively, whereas no homozygotes were identified for ORAI1 loss-of-function affecting regions shared by both Orai1 isoforms that are associated with CRAC channelopathy. (39). These results suggest tolerance of mutations that selectively disrupt Orai1α while preserving Orai1β. To determine whether Orai1 p.P43TfsX45 and p.P46SfsX42 variants are associated with clinical phenotypes, we examined diagnostic codes associated with CRAC channelopathy due to ORAI1 null mutations in UK Biobank (UKBB) (39). Individuals homozygous for Orai1α-specific truncation variants did not exhibit an enrichment of disease codes for infectious diseases, ectodermal dysplasia or myopathy characteristic of CRAC channelopathy (Fig. 7D–E). Together, these findings indicate that selective disruption of Orai1α does not phenocopy ORAI1 mutations affecting both Orai1α and Orai1β and that preservation of Orai1β is sufficient to prevent disease.

Fig 7.

Further, we identified individuals carrying the Orai1 p.P43TfsX45 variant in the Qatar Genome Project healthy population study (79). All individuals in this population study were recruited as healthy with no apparent clinical phenotype. We chose three individuals of the same gender but different age who are homozygous, heterozygous and wild-type for the Orai1 p.P43TfsX45 variant. We used immunofluorescence to assess NFAT1 translocation using anti-NFAT1 antibody (clone D43B1; Cell Signaling Technology, Cat. #14324) both at baseline before stimulation and following store depletion with thapsigargin to activate SOCE. Stimulation induced significant nuclear translocation of NFAT1 that was more pronounced in both the heterozygous and homozygous donors compared to cells from the wild-type donor (Fig. 8A-B). This is consistent with increased NFAT1 translocation in HEK293 cells expressing Orai1β compared to Orai1α (Fig. 2). Interestingly we noted that even before stimulation the levels of NFAT1 in the nucleus were elevated in T cells from the homozygous and heterozygous individuals compared to the wild-type control (Fig. 8B). This suggests that even in the absence of maximal store depletion with thapsigargin, the loss of a single or both alleles of Orai1α induces higher levels of NFAT1 translocation at basal physiological conditions. Our studies in HEK293 cells and primary CD4⁺ T cells argue that the extent of NFAT1 translocation depends on the magnitude of SOCE. We therefore measured SOCE in T cells from the three individuals who are homozygous, heterozygous and wild-type for the Orai1 p.P43TfsX45 variant. T cells from the individuals homozygous and heterozygous for Orai1 p.P43TfsX45 displayed elevated SOCE after thapsigargin stimulation compared to the wildtype control (Fig. 8C-D), consistent with results shown in Fig. 3A and the enhanced fast CDI of Orai1α compared to Orai1β reported earlier (12). To test whether increased SOCE in the homozygous and heterozygous individuals is associated with altered STIM1 recruitment into puncta, a hallmark of CRAC channel activation (2), we immunostained endogenous STIM1 following store depletion and quantified the volume of STIM1 clusters after 3D reconstruction from a z stack of images at the single cell level. The results show no significant difference in the volume of STIM1 clusters in the three individuals arguing against differential STIM1 recruitment to Orai1α and Orai1β (Fig. S4A-B). Together, these data show that an Orai1 mutation resulting in the expression of Orai1β alone does not impair, but rather enhances, SOCE and NFAT1 activation in primary human T cells.

Fig 8.

Discussion

Here we provide a comprehensive profiling of the coupling between native Orai channel homologues, SOCE and NFAT activation. While prior studies established that Orai1 is significantly more efficient than Orai3 at coupling to NFAT activation, they used overexpression models of STIM1, Orai1 and Orai3 and relied on heterologous expression of receptors for physiological stimulation (37, 38). Further, previous attempts to address the differences Orai1α and Orai1β coupling to SOCE and transcription in a native system were performed in Orai1 knockout HEK293 cells that retained native expression of Orai2 and Orai3 isoforms (35). Here we used multiple CRISPR/Cas9-engineered HEK293 models that isolate a single Orai homologue on a clean background under native levels of expression and conditions of activation that mimic the range of physiological Ca²⁺ signaling events. This strategy allowed us to resolve quantitative differences in NFAT nuclear localization across Orai isoforms in response to relatively low and maximal levels of store depletion. Our data reveal that NFAT activation positively correlates with the amount of Ca²⁺ influx mediated by each Orai homologue, and inversely correlates with the extent of Ca²⁺-dependent fast inactivation reported for each homologue (12, 28, 35). We also compared the abilities of both Orai1 translational variants to mediate SOCE and activate NFAT. We found that Orai1β more potently activated both NFAT1 and NFAT4 compared to Orai1α or the other Orai homologues, Orai2 and Orai3.

We extended our analysis of Orai coupling to transcription in primary immune cells. We chose murine CD4⁺ T cells where Orai1 is indispensable for T cell activation, metabolic reprogramming, secretion of cytokines, and effector functions (14). In this physiological setting, we focused on the alternatively translated Orai1α and Orai1β isoforms and asked whether either could compensate for the loss of endogenous Orai1. Reconstitution with either isoform restored NFAT nuclear translocation, and cytokine production. Moreover, a global transcriptomic analysis of CD4⁺ T cells expressing either Orai1α or Orai1β revealed highly overlapping gene expression signatures that were enriched for canonical immune activation pathways, closely resembling wild-type CD4⁺ T cells. These results demonstrate that Orai1α and Orai1β are equally competent in driving the Ca²⁺-dependent transcriptional networks that underpin murine T cell effector function. This conclusion is further supported by human population genetic data and the analysis of human T cells from individuals carrying mutations that selectively abolish Orai1α protein expression. Individuals homozygous for variants that selectively disrupt Orai1α without affecting Orai1β do not recapitulate the severe CRAC channelopathy phenotype associated with LOF mutations in regions shared by Orai1α and Orai1β(39). Furthermore, both SOCE and NFAT activation levels in primary T cells from individuals homozygous or heterozygous for the Orai1α null variant are not impaired but moderately increased consistent with similar observations in HEK293 cells.

While overexpressed Orai1α or Orai1β were equally able to promote SOCE, NFAT activation and transcriptional responses in CD4⁺ T cells from Orai1^CD4 mice, we observed a moderately greater ability of Orai1β compared to Orai1α to mediate SOCE and NFAT activation in HEK293 cells and T cells from the Orai1α null patient. The reason for this discrepancy is not clear. A likely explanation is that the ectopic expression levels of Orai1α and Orai1β in murine CD4⁺ T cells were sufficiently high to mask the subtle differences between both Orai1 isoforms as shown previously (12, 13). Increased SOCE and consequently NFAT activation in HEK293 cells expressing low levels of Orai1β driven by a TK promoter construct and native Orai1β in CD4⁺ T cells of the patient with Orai1α null mutation are consistent with the enhanced fast CDI of Orai1α compared to Orai1β we have reported earlier (12). Moreover, Orai1β lacks two serine residues, S27 and S30, whose PKCβ mediated phosphorylation we had previously shown to inhibit SOCE and CRAC currents (34). Similarly, we showed that Orai1α-mediated SOCE is inhibited through phosphorylation by PKA of S34, which is lacking in Orai1β (35). These inhibitory mechanisms unique to Orai1α provide a potential explanation why Orai1β is more efficient than Orai1α at promoting SOCE and NFAT activation in some cells. Collectively, our findings from engineered HEK293 cells and primary murine CD4⁺ T cells expressing either Orai1α or Orai1β as well as human individuals with Orai1α specific mutations demonstrate that Orai1α and Orai1β are each capable of coupling SOCE to NFAT activation and promoting Ca²⁺-dependent transcriptional programs.

Our data using the Ca²⁺ chelators BAPTA and EGTA indicate that NFAT1 activation is mediated by localized Ca²⁺ signals near the CRAC channel pore, which is consistent with conclusions from previous studies of Orai1 channels (38) and more generally other Ca²⁺ channels such as voltage-gated Ca²⁺ channels (55, 64). How these localized Orai1-derived Ca²⁺ signals are converted into the activation of Ca²⁺-regulated transcription factors such as NFAT, however, is not well understood. A recent study showed that amino acids 39-59 of Orai1α form a domain that allows binding of AKAP79, the recruitment of calcineurin and thus, activation of NFAT near the channel pore (38). This is an attractive model to explain the coupling of localized Ca²⁺ signals to NFAT activation and gene transcription. This study proposed that Orai1β cannot effectively couple SOCE to NFAT activation, which is however not what we observed. Instead, our findings show that Orai1β supports robust NFAT activation and transcription. The reasons for this discrepancy are currently not clear. We cannot exclude that the N terminus unique to Orai1α plays subtle roles in feedback inhibition of CRAC channels by phosphorylation of serines S27/S30 and S34 or the activities of PKC, PKA or AKAP79 anchored near the channel (34, 35). Moreover, Orai1α and Orai1β were reported to have different biophysical properties (for instance sensitivity to CDI, pH dependence, protein stability, membrane organization, and structural or heteromeric assembly capabilities (4, 12, 34, 35, 80–82), which may allow each isoform to function more effectively under distinct signaling or environmental conditions. This likely broadens the operational range of Orai1-mediated Ca²⁺ influx.

The conclusion that Orai1β is fully competent at coupling SOCE to NFAT activation and transcription is further supported by its evolutionary conservation from invertebrates such as Drosophila Melanogaster and Caenorhabditis elegans through vertebrates including zebrafish, frog, and chicken and all the way to mammals, including mouse, rat and human (Fig. S5). By contrast, Orai1α is found only in mammals despite the fact that Ca²⁺-calcineurin-NFAT dependent transcription is an ancient signaling module that predates mammals (83–85). For instance, Drosophila Melanogaster expresses a single Orai homologue (dOrai) and SOCE mediated by dOrai effectively couples to NFAT induction despite the fact that the N terminus of dOrai does not share sequence homology with the N terminus of the exclusively mammalian Orai1α (9, 86). This phylogenetic evidence argues that Orai1β represents the ancestral isoform of Orai1 channels and is sufficient to effectively couple SOCE to transcription in most cells.

Future work will need to extend the physiological and pathophysiological roles of Orai1α and Orai1β in different cells types and organs in vivo. In immune cells, it will be important to define how Orai1α and Orai1β shape other immune functions such as T and B cell differentiation, effector functions, immunological memory and tolerance. Beyond immunity, clarifying whether Orai isoforms are differentially expressed or regulated in disease states will reveal whether changes in the relative proportions of Orai1α and Orai1β could bias transcriptional programs. The discovery of unique physiological or pathological functions of each Orai isoform could spur the development of pharmacological tools capable of selectively modulating each Orai isoform for therapeutic interventions in disease.

Methods

Cell culture and transfection

HEK293 cells obtained from ATCC (Catalog # CRL-1573) were grown in Dulbecco’s modified Eagle’s medium (DMEM) containing high glucose (4.5 g/L), supplemented with 1% Antibiotic–Antimycotic solution and 10% heat-inactivated fetal bovine serum. The cells were incubated in a humidified CO2 incubator at 37 °C under standard cell culture conditions (5% CO2, 95% air).

Transfections were conducted using lipofectamine 2000 with 1:3 Lipofectamine: cDNA ratio. 75-100K cells were seeded to 25mm glass coverslips and incubated for 24 h prior to experimentation. During transfection, the Antibiotic–Antimycotic solution was taken off from the media. For experiments performed in Orai double knockout HEK293 cells (DKO), 1.0 μg of NFAT-GFP plasmids were transfected. For experiments performed in Orai triple knockout HEK293 cells (TKO), 1.0 μg of NFAT-GFP plasmids together with either 0.3-0.5 μg of TK-ORAI1α, TK-ORAI11β, TK-ORAI2 or TK-ORAI3 plasmids were used. Transfections were confirmed by recording the corresponding fluorescence, calcium imaging and immunoblotting or qPCRs.

Genetic knockout of ORAI in HEK293 cells using CRISPR/Cas9

To create ORAI-KO HEK293 cell lines, two steps were used. The first step involved using a single ORAI specific gRNA to generate insertions/deletions in the ORAI gene through transfection of corresponding ORAI1 or ORAI3 targeting vector followed by puromycin selection, colony screening by Guide-it Mutation Detection Kit, and Sanger sequencing for genetic knockout confirmation. This step was used to generate single knockouts (SKO), ORAI1-SKO and ORAI3-SKO clones. The second step was to achieve double knockout (DKO) and subsequently triple knockout (TKO) in the SKO clones. This involved subcloning of two ORAI2 or ORAI3 specific gRNAs flanking the entire gene into fluorescent vectors (pSpCas9(BB)-2A-GFP and pU6-(BbsI)-CBh-Cas9-T2AmCherry: Addgene) and sorting single cells with high GFP and mCherry expression into 96-well plates. Positive clones were identified through screening with specific PCR primers distinguishing the WT and Knockouts and confirmed by Sanger sequencing. All gRNA sequences used can be found in our previously published study(30).

Generation of Orai1-knockout MCF-7 cells (CRISPR/Cas9)

MCF-7 cells (ATCC, HTB-22) were engineered to delete ORAI1 using CRISPR/Cas9(27). A gRNA targeting human ORAI1 (GTTGCTCACCGCCTCGATGT) was cloned into lentiCRISPR v2 (Addgene) using annealed oligos PX11 (5′-CACCGGTTGCTCACCGCCTCGATGT-3′) and PX12 (5′-AAACACATCGAGGCGGTGAGCAACC-3′). Cells were transfected with the gRNA–Cas9 construct, puromycin-selected for 5 days, and plated at single-cell density in 96-well plates. Expanded clones were genotyped by PCR across the cut site (primers PX39 and PX11) and re-screened with flanking primers (PX39 and PX3). Amplicons were cloned and sequenced to confirm indels in all ORAI1 alleles leading to frameshift and premature stop codons. Validated Orai1-KO clones were further confirmed by loss of Orai1-dependent SOCE and used for Ca²⁺ imaging and NFAT1 activation assays.

Constructs and transfections

In order to produce a vector that exclusively expresses long ORAI1 isoform, a wild-type ORAI1-YFP cDNA clone was utilized. This involved replacing the weak native ORAI1 Kozak sequence located upstream of the first ATG start site (TGCTCCATG) with a stronger Kozak sequence (GCCACCATG). This resulted in the translational machinery being forced to start almost exclusively at the first ATG. Conversely, to create a vector that exclusively expressed ORAI1β, the first ATG start site of ORAI1-YFP cDNA was mutated to GCG (alanine) or ATC (Isoleucine), which left only the start site at methionine-64. Further details of how Orai1α and Orai1β vectors were constructed have been described previously(35). Constructs were validated by sequencing, and expression in cells was validated by fluorescence imaging and Western blotting (Fig. S1A-E) and functionally by their ability to rescue SOCE in HEK293 TKO cells. The florescence tags in both constructs were swapped to either YFP, m-Cherry, or Ametrine instead of CFP based on the experimental needs. TK-driven ORAI1/2/3 constructs, and gRNA were generated using primers compatible with the in-fusion cloning system. The resulting plasmids were confirmed by sequencing and by PCR using selective primers to ensure proper cDNA insertion or mutation. NFAT1/4-GFP plasmids are from Addgene (Catalog #11107/#21664).

Assessment of expression of YFP/CFP-tagged Orai isoforms in ORAI TKO HEK293 cells

The expression of Orai1-YFP in HEK293 cells was confirmed by capturing cell fluorescence using a Leica DMI8 confocal microscope with a Plan-Apochromat 40×/1.4 oilDIC M27 objective at 1× zoom. The 488-nm laser line was used to excite YFP, and fluorescence emission was detected between 491 nm − 695 nm with a 9-nm bandwidth, using digital photon counting mode and a pinhole of 180 μm with 8-line averaging.

Western blot analysis

The cells were rinsed with ice-cold phosphate-buffered saline and then lysed in RIPA buffer supplemented with protease and phosphatase inhibitors at 4 °C for 30 min. The lysates were centrifuged at 14,000 RPM for 15 min at 4 °C, and the clarified supernatant was collected for protein concentration measurement using the Pierce BCA assay. To perform deglycosylation either Protein Deglycosylation Mix II O-Glycosidase kit (NEB #P0733) or N-Glycosidase F (Millipore Sigma) were used. For the Protein Deglycosylation Mix II the manufacturer instructions were followed, and for N-Glycosidase F (Millipore Sigma) 5 μl of N-Glycosidase were added to 50 μg of protein lysate and incubated in a 37°C water bath overnight. Afterward, 30-60 μg of protein lysate was loaded into a 4-12% Bis-Tris gel. The deglycosalted protein extracts were loaded into NuPAGE 4–12% Bis-Tris precast protein gels and subjected to electrophoresis. The proteins were then transferred to polyvinylidene difluoride membranes overnight at 4 °C. Subsequently, the membranes were blocked with LI-COR Tris-buffered saline (TBS) buffer and incubated overnight at 4 °C with primary antibody. The probed membranes were then washed with TBST and incubated with secondary antibodies for 1 h at room temperature with shaking. Finally, after washing with TBST the blots were imaged using the LI-COR Odyssey imaging system and quantified by Image Studio software (LI-COR).

Calcium measurements of HEK293 cells

Fluorescence measurements were performed in HEK293 cells loaded with the radiometric calcium-sensitive dye, Fura-2AM. Briefly, collagen was used to coat 25 mm glass coverslips at a concentration of 30% in a 30% ethanol solution for 30 minutes prior to cell seeding. Day 0: 48 hours before imaging, the cells were seeded on the coated coverslips at a density of 100*10³ cells per coverslip and allowed to settle for 24 hours in the incubator. Day 1, the transfection was performed as mentioned above. Finally, on day 2, coverslips mounted into an Attofluor cell chamber (Thermo Fisher). The cells were then incubated with 2 μM Fura-2-AM (Thermo Fisher) in HEPES-buffered salt solution (HBSS) containing 140 mM NaCl, 4.7 mM KCl, 1.13 mM MgCl2, 2.0 mM CaCl2, 10 mM glucose, and 10 mM HEPES (pH 7.4) for 30 minutes at room temperature, washed four times with HBSS, and imaged using a Leica DMi8 fluorescence microscope with a 20X Fluor objective. Fura-2 was excited at 340 and 380 nm using a fast shutter wheel (Sutter Instruments), and emissions were captured at a wavelength of 510 nm with a Hamamatsu Flash 4 camera. The ratio of fluorescence emission upon alternate excitation at 340 nm and 380 nm (F340/F380) was analyzed using regions of interest drawn around the perimeters of 10-20 cells per coverslip with the Leica Application Suite X (LAS X) software. The mean ± SEM traces resulting from this analysis were represented.

NFAT-GFP nuclear translocation assays in HEK293 cells

After HEK293 cell transfections with NFAT1/4-GFP as described above, the GFP fluorescence in the cytoplasm and nucleus was collected on a Leica DiM8 microscope with a 20x objective and a Hamamatsu Flash 4 camera, with Leica Application Suite X software used for processing. NFAT1/4-GFP was excited with a 488 nm fast wheel filter and its emission spectra were selectively captured through a 510 nm GFP filter cube. The equation NFAT translocation = Nuclear F510/cytoplasmic F510 was used to quantify NFAT nuclear translocation as a function of time in response to 10 μM Cch, or 2 μM Tg after 1 min of baseline recording as previously described(29, 30, 35). Background fluorescence was subtracted from each region of interest and the fluorescence of each cell was normalized to the starting fluorescence values from the same cell to eliminate the cell-to-cell variability in expression levels.

Chelator loading experiments (EGTA-AM vs BAPTA-AM)

To distinguish between local and global Ca²⁺ signaling, HEK293 ORAI-TKO cells transfected with either Orai1α or Orai1β were pre-incubated with either 20 μM EGTA-AM or 20 μM BAPTA- AM (Thermo Fisher) for 25 min at 37 °C prior to imaging. Following chelator loading, cells were washed and either (i) loaded with 2 μM Fura-2-AM as described above for SOCE measurements, or (ii) co-transfected with NFAT1-GFP to monitor nuclear translocation in response to 2 μM thapsigargin stimulation. SOCE was quantified with Fura-2 340/380 ratio measurements, while NFAT activation was measured as the nuclear/cytoplasmic fluorescence ratio as described above.

Mice

Orai1^fl/flCd4^Cre mice were generated by crossing Orai1^fl/fl mice (62) to Cd4^Cremice (RRID:IMSR_JAX:017336). Sex-matched male and female mice were used between 10 and 12 weeks of age. Mice were maintained under specific pathogen–free conditions. All experiments were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee of at New York University School of Medicine.

Murine T cell culture

CD4⁺ T cells from Orai1^fl/flCd4^Cre mice were isolated from spleens using EasySep™ Mouse CD4+ T Cell Isolation Kit (Stemcell Technologies, Cat# 19852A) following the vendor’s manual instruction. CD4⁺ T Cells were cultured in complete RPMI1640 medium (Corning, 10-040-CV) containing 10% FBS, 2 mM L-glutamine, 50 U/ml penicillin-streptomycin, and 5.5μM β-mercaptoethanol. For preparing retrovirally transduced T cells, CD4⁺ T cells were first purified from spleen from Orai1^fl/flCd4^Cre mice by EasySep Mouse CD4⁺ T Cell Isolation Kit (StemCell, Cat. 19852) following the vendor’s instruction. CD4⁺ T cells were activated with anti-CD3ε (1ug/mL) anti-CD28 (1 μg/mL) on 20 μg/mL rabbit-anti-hamster IgG (Thermo Scientific, Cat. A18891) coated 12-well plate with seeding density 1×10⁶/mL in the presence of 5 μg/mL anti-IFN-γ (BioXcell, Cat. BE0055, Clone. XMG1.2) and anti-IL-4 (BioXcell, Cat. BE0045, Clone. 11B11) antibodies. Twenty-four hours post activation, T cells were transduced by spin-transduction (2.500 rpm, 90 min, 32 °C) in the presence of retroviral supernatant and 8 μg/ml Polybrene (Sigma-Aldrich, Cat. 107689). Retroviral supernatant from T cells was removed 30 min after spin infection and replaced with fresh complete RPMI-1640 media. T cells were detached 24 h later and split into 6-well plate with 5mL of complete RPMI-1640 media containing recombinant human interleukin (IL)-2 (20 U/mL) (PeproTech, Cat. 200-02) and IL-12 (20ng/mL, PeproTech, Cat. 210-12) to expand for 2 more days before analysis. Retroviral supernatant was produced in the Platinum-E retroviral packaging cell line (87). Platinum-E cells were transfected by GeneJet (Fisher, Cat. FERK0481) with retroviral expression plasmids pMSCV-Orai1a-IRES-Ametrine, pMSCV-Orai1b-IRES-Ametrine and pMSCV-IRES-Ametrine empty vector with the ecotropic packaging vector pCL-Eco. Retroviral supernatant was collected 36 and 60h after transfection.

Murine CD4⁺ T differentiation assays

Naive CD4⁺CD25⁻CD44^low T cells were purified from spleens and lymph nodes (cervical, inguinal, brachial, and axillary) of wild type C57BL/6 mice using a CD4⁺ T cell isolation kit (Miltenyi Biotech) according to the manufacturer’s protocol. Purified cells were activated with plate-coated anti-CD3 and soluble anti-CD28 (anti-CD3: 4 μg/ml, Clone 145-2C11; BD Biosciences; anti-CD28: 2 μg/ml, Clone 37.51; BDBiosciences) on flat-bottom plates (1 × 10⁵/well) for Th17 cells or coated with anti-CD3 and anti-CD28 (both 2 μg/ml; BD Biosciences) on flat-bottom plates (1 × 105/well) for Th1, Th2, or Treg cells. Skewing conditions were as follows: cTh17: 2.5 ng/ml rhTGF-β1 (BioLegend) plus 40 ng/ml rmIL-6 (BioLegend); pTh17: 40 ng/ml rmIL-6, 40 ng/ml rmIL-23 plus 40 ng/ml rmIL-1β; Th1: 20 ng/ml rmIL-12 plus 20 ng/ml rmIL-2; Th2: 20 ng/ml rmIL4, 20 ng/ml rmIL-2 plus 10 μg/ml anti-IFNγ; and iTreg polarization: 3 ng/ml plus 20 ng/ml rmIL-2 rhTGF-β1 (BioLegend).

Intracellular Ca²⁺ measurements of murine T cells

The flow cytometric analysis of intracellular Ca²⁺ levels in mouse CD4⁺ T cells was described previously (88). Briefly, retroviral transduced CD4⁺ T cells from Orai1^fl/flCd4^Cre mice were first stained with either PE-Cyanine7 conjugated anti-CD4 (Biolegend, Cat. 100528) or APC-Cyanine7 conjugated anti-CD4 (Biolegend, Cat. 100414). Cells were then washed, combined, and loaded with 2 μM Indo-1 in fresh complete RPMI-1640 medium for 15 min. After Indo-1 loading, cells were washed and resuspended in 510 μL of Ca²⁺-free Ringer buffer. Intracellular Ca²⁺ flux was measured using an LSR Fortessa flow cytometer on medium speed. Baseline fluorescence was recorded first, and at 100 s, 40μL of 10 μM Thapsigargin (Sigma-Aldrich, Cat. No. 586005) was added to the cell suspension to induce ER Ca²⁺ store depletion. At 400 s, 240 μL of 2 mM Ca²⁺ solution was added to the suspension to induce Ca²⁺ influx. Data acquisition continued for a total of 700 s. Fluorescence signals were collected in BUV396 and BUV496 channels. Intracellular Ca²⁺ levels were quantified as the fluorescence intensity ratio of BUV396 to BUV496.

ImageStream analysis of NFAT translocation

Nuclear translocation of NFAT1 in primary murine CD4⁺ T cells was assessed by imaging flow cytometry. CD4⁺ T cells from Orai1^fl/flCd4^Crewere retrovirally transduced with empty vector (EV), Orai1α or Orai1β, and enriched for Ametrine^dim cells by fluorescence-activated cell sorting as described above. 3 × 10⁶ cells per condition were either left unstimulated or stimulated with 1 µM ionomycin for 30 minutes at room temperature. Cells were fixed and permeabilized then incubated with an anti-NFAT1 rabbit monoclonal antibody conjugated to Alexa Fluor 488 (clone D43B1; Cell Signaling Technology, Cat. #14324) at a 1:100 dilution in 100 µL of Flow Staining Buffer for 30 minutes at room temperature in the dark. Cells were washed twice with 1× Flow Cytometry Perm Buffer and once with Flow Staining Buffer, then resuspended in 40 µL of 1× PBS. DAPI (10 µL of 1.25 µg/mL in 1× PBS) was added immediately before acquisition to a final concentration of 0.25 µg/mL to stain nuclei.

Images were acquired on an Amnis ImageStream X Mk II imaging cytometer. Brightfield images were obtained with channels 1 and 9 (430–480 nm); DAPI nuclear signal was collected in channel 7 (excitation 405 nm, emission 430–505 nm); and NFAT1-AF488 signal was collected in channel 2 (excitation 488 nm, emission 528–565 nm). Excitation laser power was adjusted prior to acquisition to prevent pixel saturation. Data were analyzed using IDEAS software version 6.2 (Amnis Corporation) using the built-in transcription factor translocation guided analysis workflow. Gating was applied to select single, in-focus cells that were double-positive for DAPI and NFAT1 staining. Nuclear translocation was quantified using the similarity score, a log-transformed Pearson correlation coefficient between the NFAT1 signal intensity and the DAPI-defined nuclear mask. Raw similarity scores were exported to CSV files for statistical analysis.

Cytokine production by murine T cells

Intracellular cytokine staining of murine CD4⁺ T cells was performed as previously described (72). Briefly, CD4⁺ T cells were stimulated with 1μM ionomycin (SigmaAldrich, Cat. 407952) and 20 nM phorbol 12-myristate 13-acetate (PMA; Calbiochem, Cat. 524400) for 4 hours in the presence of 5 μM Brefeldin A (eBioscience, 00-4506-51). Surface staining was conducted first, cells were then fixed with an intracellular fixation kit (Invitrogen, 88-8824-00) for 20 min, permeabilized with Permeabilization Buffer (eBioscience, 00-8333-56) and incubated with anti-cytokine antibody cocktails (listed in Supplementary Table S1) for 30 min. Data was analyzed using FlowJo software (FlowJo v.10.8.1). Slope, area under curve (AUC) and peak was calculated by Prism Software (GraphPad).

Bulk RNA-sequencing

Retroviral transduced CD4⁺ T cells from Orai1^fl/flCd4^Cre mice were stimulated with anti-CD3ε (1ug/mL) anti-CD28 (1 μg/mL) on 20 μg/mL rabbit-anti-hamster IgG (Thermo Scientific, Cat. A18891) coated 12-well plate for 6 hours, then sorted on Ametrine^dim (lower 25%) by flow cytometric cell sorting using a FACSAriaII (BD Biosciences). Cell purity was enriched to at least 93.3% for RNA isolation. RNA was purified using the RNeasy Micro RNA Isolation Kit (Qiagen, Cat no. 74104) and RNA-Seq libraries were prepared using the Clonetech SMART-Seq HT Kit (Takara, Cat no. 634792). The amplified libraries were purified using AMPure beads (Beckman Coulter, Cat no. A63882), quantified using a Qubit 2.0 fluorometer (Life Technologies), and visualized on an Agilent Tapestation 2200. The libraries were pooled equimolarly, loaded onto on a NovaSeq6000 system (Illumina) using a SP 100 Cycle Flow Cell v1.5 to generate 100 base pair paired-end reads. FASTQ files were generated using the bcl2fastq2 Conversion software for each sample (v2.20). Raw fastq files were assembled using Seq-N-Slide pipeline (10.5281/zenodo.5550459). For gene expression analysis, reads were trimmed using trimmomatic (v20.36) and then aligned to the mouse reference genome (GRCm38.85 / mm10) using STAR aligner (v.2.5.0c). Feature count matrices were generated with the featureCounts function of the subread package (v2.0.3). Differential expression was performed in an R statistical environment with the DESeq2 (v1.50.2) package. Gene set enrichment analysis was performed using clusterprofileR(v4.18.4) using gene sets acquired from the molecular signature database. The RNA-seq has been deposited in the GEO database under accession number: GSE334106.

Population genetic analysis of ORAI1 variants

Allele frequencies and homozygous carrier counts for ORAI1 variants were obtained from the UK Biobank (UKBB), Genome Aggregation Database (gnomAD), Qatar Biobank (QBB), and the All of Us (AoU) Research Program(89). Variants were classified according to whether they selectively affected the Orai1α-specific N-terminus or regions shared by both Orai1α and Orai1β. Allele frequencies and numbers of homozygous carriers are summarized in Figure 7C.

Human study population

QBB is a prospective, population-based cohort study established in 2012 in Qatar (90). Participants are adults who are either Qatari nationals or have resided in Qatar for at least 15 years. All participants provided written informed consent before inclusion. The study was approved by the Hamad Medical Corporation Ethics Committee and the QBB institutional review board. Associations between homozygous protein-changing variants and extreme protein and metabolite levels were analyzed in the first 14,669 participants enrolled in QBB.

Whole-genome sequencing

Genomic DNA was extracted from peripheral blood using the automated QIASymphony SP instrument according to the manufacturer’s instructions (Qiagen, Germany). DNA concentrations were quantified using the Quant-iT dsDNA Assay (Invitrogen, USA) on a FlexStation 3 (Molecular Devices, USA). Whole-genome sequencing was carried out using 2 × 150 bp paired-end reads on a HiSeq X Ten sequencer (Illumina, USA) at the Sidra Clinical Genomics Laboratory Sequencing Facility, achieving a mean coverage depth of 30×. Read-level quality control was assessed with FastQC (v0.11.2). Reads were aligned to the GRCh37 (hs37d5) reference genome using bwa.kit (v0.7.12). Post-alignment quality control of mapped reads was performed with Picard (v1.117, CollectWgsMetrics).

Variant calling followed GATK 3.4 best practices. Indel realignment and base quality score recalibration were applied to the initial BAM files, after which HaplotypeCaller was run on each sample to generate per-sample genomic variant call files (gVCFs). Joint variant calling was then performed across all samples: GenomicsDB was used to combine samples by genomic region, and GenotypeGVCFs was subsequently applied to each region. Variant quality score recalibration (VQSR) was performed separately for SNVs and indels. In total, 134 million variants passed the VQSR filter across the 14,669 QBB individuals (90). No individuals were excluded for high missingness (>0.1), as assessed using Plink (v2.0) (90). Approximately five million variants were removed due to a low call rate (<0.05).

Isolation of human T cells

Peripheral blood mononuclear cells (PBMCs) were provided by Qatar Biobank (QBB) and experimentation was performed following approved QBB IRB protocol number E-2023-QF-QBB-RES-ACC-00153-0245. To activate and expand human T cells, thawed PBMCs were resuspended in complete Advanced RPMI 1640 medium supplemented with GlutaMAX (Gibco) and 10% fetal bovine serum (FBS). Cells were stimulated with 1 μg/ml purified mouse anti-human CD3 (clone HIT3a; BD Pharmingen) and anti-human CD28 (clone CD28.2; BD Pharmingen) antibodies and incubated at 37 °C in a humidified incubator with 5% CO₂ for 48 hours. Following activation, cells were maintained in culture in the presence of 30 U/ml recombinant human IL-2 (hIL-2; PeproTech). Fresh IL-2-containing medium was added every two days starting on day 3 post-activation, and T cells were expanded for up to 21 days.

Intracellular Ca²⁺ imaging of human T cells

Ca²⁺ imaging was performed using a PTI Easy Ratio Pro system (software version 1.6.1.0.101; Horiba Scientific) composed of a DeltaRAMX monochromator and a CoolSnapHQ² camera attached to an Olympus IX71 inverted microscope fitted with a 20x/0.75 lens. The standard saline contained (in mM) 145 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 10 Glucose, 10 HEPES, pH 7.2, for Ca²⁺-free experiments, the Ca²⁺ was exchanged equimolarly with Mg²⁺. For Ca²⁺ imaging the cells were first plated for 30 minutes on CellTak (Corning) coated glass coverslips and were then loaded for 30 min with 2 µM Fura2-AM in a Ca²⁺-containing media at room temperature and washed three times in Ca²⁺-free media prior to the experiments. Cells were excited at 340 nm and 380 nm for 100 msec at a 0.1 Hz frame rate and the ratio of the fluorescence intensity at 340/380 measured. Tg (1 µM) was applied for 20 minutes (enough for the Ca²⁺ baseline to recover) in a Ca²⁺-free media. The extracellular Ca²⁺ concentration was then raised to 2 mM through direct injection of Ca²⁺ into the dish. The Ca²⁺ traces were analyzed using Clampfit (11.4.3, Molecular Devices).

Immunocytochemistry of human T cells

T cells were plated as indicated for Ca²⁺ imaging and treated with thapsigargin (Tg, 1 µM) for 10 min before fixation for 10 min with 4% paraformaldehyde. Cells were then permeabilized with 0.1 % Triton X100 for 10 min and saturated for 1h using a mix of 10% goat serum and 1% bovine serum albumin. Incubation in the primary antibody for NFAT1 (5861S, Cell Signaling) was performed overnight at 4°C, secondary antibody was anti-rabbit coupled Ax488 (Molecular Probes) and used at a 1:2000 dilution at room temperature for 2 hours, staining of the nuclei was performed using Hoechst (8 µM). Confocal images were acquired using a Zeiss LSM880 confocal microscope (63x/1.4, 1 Airy Unit) and the following parameters (λ_ex=405, λ_em=410-495 for Hoechst, λ_ex=488, λ_em=495-630 for Ax488, 1AU). Z-stacks were recorded at 1 µm intervals. Analysis was performed using FIJI ((91). Fixed T cells (see main manuscript methods) were incubated with a STIM1 primary antibody overnight at 4°C (4916S, Cell signaling,) and for 2 hours at room temperature with an Ax555 coupled secondary antibody (Molecular Probes) and counterstained with Hoechst to stain the nuclei. High resolution images were acquired using the Airyscan detector of a Zeiss LSM880 confocal microscope using the super-resolution mode (SR) and default image processing parameters. The 405 and 561 laser lines were used to acquire alternating z-stacks typically 0.18 µm intervals. Clusters identification was performed using the 3D Spot Segmentation plugin of FIJI using the watersheding option and default parameters. Spot seeding threshold was defined as the intensity in the middle of the weakest well defined cluster. Cluster volume measurement was performed using the built-in 3D object counter of FIJI. Outliers were removed from the data using Graphpad Prism “identify outliers” function.

Statistics

GraphPad Prism 8 software was used to conduct statistical analyses. The Student’s t-test was used to compare two groups, while One-way analysis of variance was used to compare more than two groups with Dunnett’s test for multiple comparisons. Prism 8.0 was used to determine data normality. Parametric statistical tests were used for normally distributed data, and non-parametric statistical tests were used for data that were not normally distributed. Results are expressed as the mean and error reported as ±SEM. Throughout the manuscript, asterisks were used to indicate p-values are shown on the graphs.

Supplementary Figures

Figure S1.

Figure S2.

Figure S3.

Figure S4.

Figure S5.

Supplementary Tables

Table species1.

Data availability

The RNA-seq has been deposited in the GEO database under accession number: GSE334106. All data generated or analyzed during this study are included in the manuscript and supporting files; source data files have been provided for all figures. Allele frequencies and homozygous carrier counts for ORAI1 variants were obtained from the UK Biobank (UKBB), Genome Aggregation Database (gnomAD), Qatar Biobank (QBB), and the All of Us (AoU) Research Program. Variants were classified according to whether they selectively affected the Orai1α-specific N-terminus or regions shared by both Orai1α and Orai1β. Allele frequencies and numbers of homozygous carriers are summarized in Figure 7C.

Acknowledgements

This research has been conducted using the UK Biobank Resource under Application Number 105564.

Additional information

Funding

This work was supported by the National Institutes of Health (NIH) grants: NIH/ National Heart, Lung, and Blood Institute R35HL150778 (M.T.), R35HL161177 (A.C.S), and AI175276, AI180128 (S.F.); and by the Qatar Research, Development and Innovation (QRDI) Pathways to Precision Medicine grant PPM 06-0522-230038 (A.B.) and by the Biomedical Research Program at Weill Cornell Medical College in Qatar (BMRP), a program funded by Qatar Foundation. The imaging core at WCMQ is supported by the BMRP. The statements made herein are solely the responsibility of the authors.

Author Contributions

A.E.A., K.M, S.F., and M.T. designed research; A.E.A., Y.W., J.C.B., I.P., M.S., M.M., M.J., A.T.E., S.C.-R., R.C., A.B., F.Y., D.S.P., S.Y., and P.X. performed research; A.C.S., F.F.S., N.H., W.F.H., D.B.B., K.M., S.F., and M.T. contributed new reagents, analytic tools, or clinical/genetic resources; A.E.A., Y.W., J.C.B., I.P., M.S., M.M., M.J., R.C., A.B., F.Y., S.Y., D.B.B., K.M., S.F., and M.T. analyzed data; and A.E.A., K.M., S.F., and M.T. wrote the paper with input from all authors.

Funding

HHS | NIH | National Heart, Lung, and Blood Institute (NHLBI) (HL150778)

• Mohamed Trebak