Cavβ1 regulates T cell expansion and apoptosis independently of voltage-gated Ca2+ channel function

TCR stimulation triggers Ca2+ signals that are critical for T cell function and immunity. Several pore-forming α and auxiliary β subunits of voltage-gated Ca2+ channels (VGCC) were reported in T cells, but their mechanism of activation remains elusive and their contribution to Ca2+ signaling in T cells is controversial. We here identify CaVβ1, encoded by Cacnb1, as a regulator of T cell function. Cacnb1 deletion enhances apoptosis and impairs the clonal expansion of T cells after lymphocytic choriomeningitis virus (LCMV) infection. By contrast, Cacnb1 is dispensable for T cell proliferation, cytokine production and Ca2+ signaling. Using patch clamp electrophysiology and Ca2+ recordings, we are unable to detect voltage-gated Ca2+ currents or Ca2+ influx in human and mouse T cells upon depolarization with or without prior TCR stimulation. mRNAs of several VGCC α1 subunits are detectable in human (CaV3.3, CaV3.2) and mouse (CaV2.1) T cells, but they lack transcription of many 5’ exons, likely resulting in N-terminally truncated and non-functional proteins. Our findings demonstrate that although CaVβ1 regulates T cell function, these effects are independent of VGCC channel activity.

C hanges in intracellular Ca 2+ concentration ([Ca 2+ ] i ) are essential for signal transduction in all eukaryotic cells including T lymphocytes 1,2 . The best characterized Ca 2+ influx pathway in T cells is store-operated Ca 2+ entry (SOCE) mediated by Ca 2+ release-activated Ca 2+ (CRAC) channels encoded by ORAI1 and its homologue ORAI2. TCR stimulation results in the generation of the second messenger inositol 1,4,5trisphosphate (IP 3 ), the opening of IP 3 receptor channels and Ca 2+ release from the endoplasmic reticulum (ER). Ca 2+ efflux from the ER causes the activation of STIM1 resulting in its binding to ORAI1 and opening of CRAC channels. CRAC channels are critical for T cell function and immunity to infection as evidenced by the immunodeficiency of CRAC-deficient human patients and mice 2 .
Other Ca 2+ channels that have been proposed to mediate Ca 2+ influx in T cells are voltage-gated Ca 2+ channels (VGCCs). They are critical for Ca 2+ signaling in excitable cells such as neurons, cardiomyocytes, skeletal muscle and secretory cells 3 3,4 . VGCCs are composed of a Ca 2+ conducting, pore-forming α1 subunit and several auxiliary β, α 2 δand γ-subunits. The α 1 subunits are composed of four domains (I-IV), each consisting of 6 α-helical transmembrane domains (S1-S6). S1-S4 form the voltage sensing domain (VSD) with S4 containing positively charged amino acids that sense changes in membrane potential, while the S5-S6 subunits constitute the ion conduction pore and selectivity filter 3,5 . Each VGCC has one β-subunit, which binds via its α 1binding pocket (ABP) to the cytosolic α 1 −interacting domain (AID) in the linker region between domains I and II of the α 1 subunit. The four β subunit homologues (β1-β4) increase the plasma membrane expression of α 1 subunits, enhance Ca 2+ currents and modulate the voltage-dependence and kinetics of activation and inactivation of VGCCs 6,7 . The α1 subunit furthermore binds to the extracellular α 2 δ-subunit and the transmembrane γ-subunit consisting of four transmembrane domains. Mutations in VGCCs and altered Ca 2+ influx are associated with a plethora of human diseases including cardiac arrhythmias and psychiatric diseases (Cav1.2), autism spectrum disorder and primary aldosteronism (Cav1.3), various X-linked retinal disorders (Cav1.4), familial hemiplegic migraine (Cav2.1), epilepsy (Cav1.3, Cav2.1, Cav3.2) and several forms of ataxia (Cav2.1, Cav3.1) [8][9][10][11] . Similarly, deletion or mutation of α 1 and auxiliary VGCC subunits in mice has been reported to cause a large spectrum of neurological, cardiovascular, musculoskeletal and endocrinological phenotypes 7,8,12,13 . Ca 2+ channel blockers targeting VGCCs such as nimodipine, verapamil and diltiazem are in wide clinical use for the treatment of arterial hypertension 5 . A common denominator of human diseases associated with VGCC mutations, phenotypes of genetically altered mice and the effects of Ca 2+ channel blockers is that they originate from the altered function of excitable cell types such as neurons, cardiomyocytes or secretory cells including pancreatic β cells or adrenal chromaffin cells 14 .
Several studies have reported that VGCCs are modulating immune responses and Ca 2+ signaling in T cells by using Ca 2+ channel blockers, RNA interference (RNAi) and genetic deletion of various α 1 and β subunits in mice 1 . RNAi-mediated deletion of Cav1.2 and Cav1.3, or β subunits with antisense oligonucleotides, showed reduced TCR-induced Ca 2+ influx, cytokine production and experimental asthma in CD4 + T cells polarized into T helper 2 (Th2) cells 15,16 . T cells from Cacna1f −/− mice lacking Cav1.4, which is highly expressed in the retina, had reduced Ca 2+ influx and Ba 2+ currents in T cells and showed a defect in the function, development and survival of naïve T cells and in T cell responses to intracellular pathogens in vivo 17,18 . Genetic deletion of the T-type VGCC Cav3.1 in mice had no effect on TCR-induced Ca 2+ influx in T cells despite reduced low-voltage activated Ca 2+ currents 19 . However, Cav3.1-deficient mice were protected from experimental autoimmune encephalomyelitis (EAE), which was associated with reduced numbers of IFN-γ and GM-CSF producing T cells in vivo and defects in Th17 cell function in vitro including Ca 2+ influx, NFAT activation, and the expression of RORγt and IL-17A 19 . In addition, several studies have implicated β subunits of VGCCs in T cell function. For example, T cellspecific deletion of Cavβ2 resulted in a severe defect in T cell development due to impaired thymocyte proliferation and survival 20 . T cells from Cacnb2 −/− (Cavβ2 knockout) mice had moderately reduced TCR-induced Ca 2+ influx, which was associated with loss of Cav1.2 and Cav1.3 protein expression 20 . Similarly, T cells of lethargic mice with a spontaneously occurring mutation in Cacnb4 (encoding Cavβ4) exhibited splenic and thymic involution and lymphocytopenia 21,22 . TCR-induced Ca 2+ influx was moderately reduced in T cells of Cavβ4 mutant mice and those from mice with targeted deletion of Cavβ3 23 . Although T cells of Cacnb3 −/− mice had moderately reduced TCR-induced Ca 2+ influx, the survival of naive CD8 + T cells was profoundly impaired due to altered expression of pro-and antiapoptotic genes 24 . Cavβ3 deficiency was associated with loss of Cav1.4 protein expression, suggesting that Cavβ3 may regulate Ca 2+ influx in T cells through Cav1.4 24 . Collectively, these studies suggest that VGCCs contribute to T cell development and function, potentially by regulating TCR-stimulated Ca 2+ signaling. VGCC as Ca 2+ channels in T cells, however, are not universally accepted 25 , and biophysical evidence of VGCC currents in T cells is limited 17,19 . Given the non-excitable nature of T cells it also remains unclear how VGCCs are activated in the context of T cell activation.
In this study, we identify Cacnb1 (Cavβ1) as a regulator of T cell function. Although Cavβ1 is well-studied in skeletal muscle, where it modulates excitation/contraction coupling 26 , its function in T cells has not been reported. Using a pooled shRNA screen to identify ion channels that regulate T cell responses to viral infection in vivo, we find that deletion of Cacnb1 impairs the clonal expansion of antigen specific T cells after viral infection in vivo by enhancing T cell apoptosis. Cacnb1 deletion does not affect TCR-induced Ca 2+ signaling and production of Ca 2+ regulated cytokines, suggesting that its function in T cells differs from its canonical one in excitable cells modulating the function of VGCCs. Indeed, a detailed search for voltage-gated Ca 2+ currents and Ca 2+ signals in human and mouse T cells fails to provide evidence for the existence of functional VGCCs in T cells. While mRNAs of several VGCC α1 subunits are detectable in T cells by RNA-Seq (Cav3.3, Cav3.2 and Cav2.1), these transcripts are incomplete, and lack expression of multiple 5′ exons that encode the first two (of four) Ca V domains. We conclude that full-length transcripts of α 1 subunits of VGCCs are not expressed in T cells, providing an explanation for the absence of VGCC currents and Ca 2+ influx upon depolarization in T cells.

Results
shRNA screen in vivo identifies Cacnb1 as a VGCC subunit required for clonal expansion of T cells during LCMV infection. To identify ion channels and transporters (ICTs) that regulate T cell function and T cell-mediated immunity during viral infection in vivo, we generated a library of 658 ICTs and regulatory factors, of which 602 ICTs were annotated in both mouse and human genomes. These ICTs were analyzed for their mRNA expression levels in immune cells using the Immunological Genome Project (ImmGen) 27 and Fantom5 28,29 databases, respectively ( Supplementary Fig. 1A). We identified 154 ICTs that are expressed at least twofold above the population average in both mouse and human CD4 + T cells ( Supplementary Fig. 1B, C). Similar analyses were conducted in 11 other immune cell populations, resulting in a total of 223 ICTs with >2-fold above average expression across all cell types ( Supplementary Fig. 1B). We used this information to generate a customized, pooled shRNA library targeting 223 mouse ICT genes. To delete ICTs, CD4 + CD45.1 + T cells were isolated from SMARTA mice that express a transgenic TCR specific for the LCMV GP [61][62][63][64][65][66][67][68][69][70][71][72][73][74][75][76][77][78][79][80] epitope 30 , and transduced with the shRNA library. shRNAtransduced (Ametrine + ) T cells were sorted and injected into CD45.2 + congenic WT host mice, which were next infected with the Armstrong strain of LCMV (LCMV ARM , Fig. 1A). LCMV ARM causes an acute viral infection and a well-characterized CD4 + and CD8 + T cell response 31 . 7 days later, donor T cells were isolated from the spleens of host mice, enriched for CD4 + CD45.1 + Amt + T cells by cell sorting and analyzed by next generation sequencing (NGS) for the enrichment or depletion of ICTs targeted by specific shRNAs (Fig. 1A). Among the positive controls whose suppression resulted in the significant depletion (P < 0.05, Log2-FC < 0.5) of T cells 7 days after LCMV infection were genes that are critical for T     Fig. 1 shRNA screen identifies Cacnb1 as a regulator of antiviral T cell responses. A In vivo ion channels and transporters (ICT) screen. CD4 + T cells from SMARTA mice were transduced with a pooled shRNA library targeting 223 ICTs (1342 shRNAs including control shRNAs), enriched by cell sorting for transduced (Ametrine + ) cells and injected into host mice. 7 days after infection with LCMV ARM , CD4 + CD45 + Amt + donor T cells were isolated from host spleens and analyzed by next generation sequencing (NGS) for the depletion or enrichment of shRNAs. B Scores of depleted shRNAs and their p values calculated based on the negative-binomial model using the MAGeCK software package 88 . Cacnb1 (blue dot), other ICTs (gray) and positive controls (red) are indicated. Shown are the pooled data from three independent screens. C, D mRNA expression of Cavβ subunits in mouse (C) and human (D) T cells compared to other immune cells based on data from ImmGen 27 and Fantom5 28,29 databases. Each column represents a different type of immune cell. Heatmaps represent % raw min-max expression for each gene. NK natural killer, NKT natural killer T, Mo monocyte, B B cell, DC dendritic cell, E eosinophil, M macrophage, N neutrophil, LC Langerhans cell. E, F Absolute mRNA expression of auxiliary β, α2δ and γ subunits of VGCCs in mouse (E) and human (F) T cells and reference tissues. Mouse CD4 + and CD8 + T cells were left unstimulated (−) or stimulated for 12 or 24 h with anti-CD3 + CD28 antibodies. Human CD4 + T cells were left unstimulated (−) or stimulated for 6 h with anti-CD3 + CD28 antibodies. mRNA expression was analyzed by RNA sequencing. HD represents the averaged data from three individual healthy donors (HD) and a patient with a STIM1 null mutation (STIM1 null ). RNA-Seq for mouse T cells and for human and mouse heart, skeletal muscle, brain, frontal cortex, and retina were extracted from GEO datasets (Supplementary Table 2 cell signaling, survival and function (CD3e, Cd4, Lck, Rpa3, Zap70, Bcl2l1). By contrast, suppression of Prdm1, which encodes BLIMP1 and inhibits the differentiation of follicular T helper (T FH ) cells after viral infection, resulted in significant enrichment (P < 0.05, Log2-FC > 2) of T cells (Fig. 1B, and Supplementary Fig. 1D).
Among the ICTs whose shRNA-mediated knockdown significantly depleted T cells in vivo was Cacnb1, which encodes the auxiliary Cavβ1 subunit of VGCCs (Fig. 1B, and Supplementary  Fig. 1D). Given the importance of β subunits for VGCC function in excitable cells and the reported function of β2, β3 and β4 in T cells 20,23,24 , we first analyzed the expression levels of all Cavβ subunits in mouse and human T cells using the ImmGen 27 and Fantom 5 28,29 gene expression databases, respectively. Compared to other immune cells, mouse T cells express higher levels of Cavβ1 (Cacnb1), whereas human T cells express both β1 and β3 (encoded by CACNB1 and CACNB3, Fig. 1C, D). We next analyzed absolute mRNA expression levels of β subunits in mouse and human T cells compared to reference tissues with known VGCC function using our own and published RNA-Seq data. In T cells of wildtype (WT) mice and healthy human donors (HDs), β1 and β3 were the only Cavβ subunits showing robust mRNA levels ( Fig. 1E, F). Whereas β1 and β3 mRNA levels in human T cells remained high after TCR stimulation, their expression decreased in mouse CD4 + and CD8 + T cells at 12-24 h after stimulation. Analysis of other auxiliary subunits of VGCCs showed that although γ subunits are generally expressed at very low levels, α 2 δ 4 is robustly expressed in both human and mouse T cells (Fig. 1E, F). It is noteworthy that expression levels of all four β subunits and α subunits did not markedly differ between human CD4 + and CD8 + T cells ( Supplementary Fig. 1E, F). The shRNA screen and expression data suggest that Cavβ1 may have a non-redundant function in T cells during LCMV infection.
Cavβ1 is required for T cell expansion and survival in vitro and in vivo. Because Cavβ1 is highly and selectively expressed in mouse and human T cells compared to other β subunits, and because β2, β3 and β4 were reported to regulate T cell development and function 20,23,24 , we further investigated the function of Cavβ1 in T cells. To delete Cacnb1 expression in mouse T cells we used two approaches: (1) CRISPR/Cas9 gene editing by retrovirally transducing CD4 + T cells of Cas9 LSL GFP Cd4Cre knock-in mice with small guide (sg) RNAs targeting Cacnb1, and (2) shRNA-mediated knockdown by transducing mouse CD4 + T cells with individual shRNAs targeting Cacnb1. Both approaches achieved~50-70% reduction of Cavβ1 mRNA and protein levels ( Fig. 2A, B and Supplementary Fig. 2A, B). Deletion of Cavβ1 reduced the numbers of transduced (Amt + ) CD4 + T cells after TCR stimulation in vitro relative to T cells transduced with control sgRNAs or shRNAs under co-culture conditions ( Fig. 2C and Supplementary Fig. 2C). No significant defects in CD4 + T cell proliferation were detectable after Cacnb1 deletion by either sgRNAs or shRNAs ( Fig. 2D and Supplementary Fig. 2D). By contrast, we observed a significant increase in apoptosis in Cavβ1deficient CD4 + T cells compared to control T cells ( Fig. 2E and Supplementary Fig. 2E) suggesting that Cavβ1 in T cells contributes to T cell survival.
We next investigated the ability of Cacnb1-deficient T cells to expand after infection of mice with LCMV ARM . CD4 + T cells from Cas9 LSL GFP Cd4Cre knock-in mice that had been crossed to SMARTA mice were transduced with sgRNAs against Cacnb1 (Ametrine + ) and mixed at a 1:1 ratio with CD4 + T cells transduced with control sgRNAs (GFP + ), allowing us to investigate the effects of Cavβ1 deletion in CD4 + T cells compared to mock-transduced T cells in the same host mice. As an additional control, CD4 + T cells were transduced with control sgRNAs encoded by vectors expressing either Ametrine or GFP reporters that were also mixed at a 1:1 ratio (sgCtrl Amt / sgCtrl GFP ). Following T cell transfer WT host mice were infected with LCMV ARM (Fig. 2F). 7 days post-infection, we observed a significant~2.4-fold decrease of CD4 + T cells transduced with sgCacnb1 (Amt + ) compared to sgCtrl (Amt + ) when normalized to sgCtrl (GFP + ) transduced cells (Fig. 2F). Similar observations were made using an orthogonal approach by transducing CD4 + T cells from SMARTA mice with shCacnb1 (Amt + ) and shCtrl (GFP + ) followed by transfer of T cells at a 1:1 ratio and infection with LCMV ARM (Supplementary Fig. 2F). Whereas the ratio of shCtrl (Amt + ) to shCtrl (GFP + ) transduced CD4 + T cells remained unchanged 7 days after infection, we found a strong, sixfold reduction in the ratio of shCacnb1 (Amt + ) to shCtrl (GFP + )-transduced T cells ( Supplementary Fig. 2F). Together, these data indicate that Cavβ1 is required for the clonal expansion of CD4 + T cells in vitro and in vivo after viral infection by regulating T cell survival.
Cavβ1 expression in T cells is dispensable for TCR-mediated Ca 2+ influx and cytokine production. The major canonical function of Cavβ subunits is to regulate the function of VGCCs and thereby Ca 2+ signaling in excitable cells 7 . Ca 2+ signals are critical for many T cell functions including cell proliferation, survival and cytokine production 1,2 . Because previous studies had shown that Cavβ3 and β4 subunits regulate Ca 2+ influx in T cells and T cell function 23,24 , we next investigated the effects of Cavβ1 deletion on Ca 2+ signaling and T cell function. Following deletion of Cavβ1 by transducing CD4 + T cells with sgCacnb1 or shCacnb1, T cells were stimulated by CD3 crosslinking and analyzed for cytosolic Ca 2+ concentrations. Cavβ1 deletion in T cells by either sgRNA or shRNA did not impair TCR-induced Ca 2+ influx (Fig. 3A, B and Supplementary Fig. 3A, B). Likewise, SOCE induced by 1 μM ionomycin (to deplete ER Ca 2+ stores) was not affected by suppression of Cacnb1 expression. By contrast, deletion of Stim1 to suppress CRAC channel activation strongly suppressed TCR-mediated Ca 2+ influx. We next analyzed if the expression of cytokines that are known to be regulated by Ca 2+ is impaired in Cacnb1-deficient T cells. Deletion of Cavβ1 in CD4 + T cells by transduction with sgRNAs and shRNAs had no effect on the production of IL-2, TNF and IFN-γ in response to PMA/ionomycin stimulation compared to T cells transduced with control sgRNAs and shRNAs (Fig. 3C, D and Supplementary Fig. 3C, D). By contrast, deletion of STIM1strongly suppressed the production of all three cytokines. Collectively, these data demonstrate that Cavβ1 is dispensable for TCR-induced Ca 2+ signaling and cytokine production in T cells.
Depolarization does not evoke Ca 2+ influx or Ca 2+ currents in T cells. Whether VGCCs are functional as Ca 2+ channels in T cells and regulate Ca 2+ signaling has remained controversial. The normal Ca 2+ signals in Cavβ1-deficient T cells despite altered T cell function prompted us to investigate whether VGCC function is detectable in human and mouse T cells. To this end, we measured Ca 2+ signals in T cells following exposure to high extracellular concentrations of K + ([K + ] o ) to depolarize the membrane potential (V m ) as was first demonstrated in lymphocytes by Deutsch et al. 32 and shown to activate VGCCs in excitable cells 33 (Fig. 4B). This reduction is expected because depolarization of V m collapses the electrical gradient required for Ca 2+ influx through store-operated CRAC channels. To demonstrate that depolarization of V m by application of high extracellular K + is able to activate VGCCs, we transfected HEK293 cells, which are not excitable, with the α1 subunit of Cav1.2 together with β, γ and α2δ subunits and subjected these cells to the same depolarization protocol. Addition of 150 mM extracellular K + evoked Ca 2+ influx in Cav1.2 transfected cells, but not in untransfected cells ( Supplementary  Fig. 4A). As expected, Ca 2+ influx in Cav1.2 expressing cells could be blocked by the Ca 2+ channel blocker nimodipine (8 μM) ( Supplementary Fig. 4B). Moreover, we tested if depolarization of V m by high extracellular K + induces voltage-dependent Ca 2+ influx in PC12 cells which endogenously express VGCCs 34 . Exposure of PC12 cells to 150 mM [K + ] o induced a robust, transient increase in intracellular Ca 2+ levels ( Supplementary  Fig. 4C).
To more precisely and dynamically control the membrane potential in T cells and to measure voltage-dependent Ca 2+ influx and Ca 2+ currents in T cells, we investigated VGCC channel function by patch-clamp electrophysiology. Human T cells from healthy donors (HD) were loaded with the Ca 2+ sensitive dye Indo-1 to measure [Ca 2+ ] i concentrations and patch-clamped to record VGCC currents. The perforated-patch configuration was chosen to minimize run-down of VGCC currents that commonly occurs during whole-cell recordings. Simultaneous measurements of [Ca 2+ ] i and VGCC currents provided two independent ways to measure depolarization-evoked Ca 2+ signals. In separate experiments, we measured VGCC currents in the presence of 110 mM Ba 2+ as the charge carrier, which conducts through the channel about twofold better than Ca 2+ and confers the added advantage that Ba 2+ currents, unlike Ca 2+ currents, do not inactivate, thereby optimizing detection of even small VGCC currents in T cells. To activate VGCCs, human T cells of HD were depolarized stepwise from −60 to +60 mV from a holding potential of −80 mV. No inward currents were detectable with this protocol, either in isotonic Ba 2+ (Fig. 4C) or in 20 mM Ca 2+ (Fig. 4D). Furthermore, steps to various voltages between −60 and +60 mV failed to evoke increases in [Ca 2+ ] i in human T cells (Fig. 4D). Although this result strongly suggests that human T cells do not express functional VGCCs, this conclusion may not be valid if T cells only expressed a few functional VGCCs per cell and the resulting small current amplitudes and Ca 2+ signals were hard to detect. To circumvent this potential limitation, we repeatedly activated VGCCs by delivering depolarizing stimuli every second with the goal of eliciting a more pronounced rise in [Ca 2+ ] i . We applied 40-50 depolarizing steps (each lasting 200 ms) to +10 mV from the holding potential every second. This protocol also failed to elicit a rise in [Ca 2+ ] i in human T cells (Fig. 4E). To exclude the possibility that the sensitivity of our recording system is too low to record small Ca 2+ currents, we measured Ca 2+ influx resulting from the activation of CRAC channels on the same experimental set-up. After treatment with thapsigargin to deplete Ca 2+ stores and activate CRAC channels, human T cells were held at a positive membrane potential of +60 mV and repeatedly stimulated with voltage steps to    Fig. 6A, B). The fact that membrane depolarization decreases, rather than increases, Ca 2+ signals in T cells is consistent with the requirement for a negative V m to provide the electrical gradient for Ca 2+ influx through CRAC and other Ca 2+ channels. To directly evaluate whether functional VGCC channels are present in mouse T cells, we carried out measurements of wholecell Ca 2+ currents using patch-clamp electrophysiology in response to depolarizing voltage steps. CD4 + T cells isolated from the spleen of C57BL/6 mice were kept overnight in culture medium supplemented with IL-7 or stimulated for 2 days with anti-CD3/CD28 antibodies before analysis of channel function.
To activate VGCCs, T cells were depolarized stepwise from −80 to +60 mV from a holding potential of −70 mV. These recordings failed to induce any inward currents in naïve unstimulated or stimulated mouse T cells (Fig. 5C, D). As a control for our recording conditions, we measured K + currents resulting from the voltage dependent activation of Kv1.3, which is well known to regulate the membrane potential of T cells and T   35 . Stepwise depolarization of unstimulated and stimulated CD4 + T cells from −100 to +100 mV evoked robust, easily detectable voltage-activated outwardly rectifying K + currents whose voltage dependence and kinetics were similar to well-described Kv1.3 channels (Fig. 5E, F) 36 . We next attempted to detect VGCC currents in human T cells after acute TCR stimulation with OKT3 (Fig. 5G, H). Examination of voltage-gated currents in T cells from HDs is complicated by the fact that TCR stimulation triggers the activation of CRAC currents and a rise in [Ca 2+ ] i . The presence of CRAC currents precluded us from employing leak subtraction using the P/N method 37 and VGCC currents were therefore measured without leak subtraction. Instead, we tested whether currents observed during T cell depolarization were affected by nimodipine, a potent blocker of L-type VGCCs. Stepwise depolarization of OKT3-treated human T cells failed to evoke inward Ba 2+ currents resembling VGCC currents. Furthermore, application of 8 μM nimodipine failed to affect the observed membrane currents (Fig. 5G, H), suggesting that dihydropyridine (DHP)sensitive currents are not present in stimulated human T cells. Together, our data indicate that VGCCs do not contribute to Ca 2+ signals in T cells in response to TCR stimulation.
In neurons, the function of L-type VGCCs is modulated by the phosphorylation of α 1 subunits mediated by protein kinase A (PKA) 38 . Activation of PKA by cAMP downstream of βadrenergic receptor stimulation results in increased L-type Ca 2+ currents 3 . Protein kinase C (PKC) also regulates L-type Ca 2+ channel function and has both stimulatory and inhibitory effects on Ca 2+ currents 39 . PKA and PKC also mediate T cell signaling and control T cell function 40,41 . To test the hypothesis that PKA or PKC may be required for VGCC function in T cells, we preincubated human T cells with the PKC agonist phorbol 12-myristate 13-acetate (PMA) or the PKA agonist 8-Bromoadenosine 3′,5′-cyclic monophosphate (8-Br-cAMP) for 10 min before depolarizing cells with 60 mM [K + ] o . Neither treatment with PMA nor 8-Br-cAMP induced Ca 2+ influx upon depolarization ( Supplementary Fig. 7A-D). Collectively, these data demonstrate that neither voltage-gated Ca 2+ influx nor Ca 2+ currents are detectable in human or mouse T cells, regardless if T cells were left unstimulated or restimulated by TCR activation or direct activation of PKC or PKA signaling.
Lack of ORAI1, STIM1 and CRAC channel function does not induce compensatory VGCC currents and Ca 2+ influx in T cells. The contribution of VGCCs to TCR-induced Ca 2+ influx may be difficult to assess in WT T cells because TCR stimulation results in the activation of CRAC channels and SOCE, which could mask smaller Ca 2+ signals emanating from VGCCs. To circumvent this problem, we investigated whether VGCC currents are unmasked in the absence of functional CRAC channels. To this end, we used T cells from a patient with a loss-of-function (LOF) mutation in ORAI1 (p.R91W), which abolishes CRAC channel function and SOCE 42 . We activated VGCCs in ORAI1-deficient T cells by depolarizing the membrane stepwise from −60 to +60 mV from a holding potential of −80 mV using the perforated patch configuration and simultaneously measured [Ca 2+ ] i with Indo-1 (Fig. 6A). No inward Ca 2+ currents or depolarization-evoked Ca 2+ signals could be detected with this protocol using either external solutions containing 20 mM Ca 2+ (Fig. 6B). To again exclude the possibility that activation of VGCCs in T cells requires TCR signaling in addition to depolarization, ORAI1-deficient T cells were first stimulated with OKT3 followed by stepwise depolarization. TCR crosslinking in the absence of CRAC channels failed to unmask depolarization-evoked membrane currents and Ca 2+ signals (Fig. 6C). We reasoned again that we might not be able to detect small VGCC currents if T cells expressed only a few functional VGCCs per cell. We therefore measured [Ca 2+ ] i in Indo-1-loaded T cells that were repeatedly depolarized every 1 s to produce a buildup of cytosolic Ca 2+ . Neither ORAI1-deficient T cells that were left untreated nor cells stimulated with OKT3 showed increases in [Ca 2+ ] i upon repeated depolarization (Fig. 6D, E). By contrast, stimulation of ORAI1-deficient T cells with a high dose of ionomycin to bypass Ca 2+ influx through CRAC channel resulted in the robust elevation of [Ca 2+ ] i (Fig. 6F). Collectively, these data demonstrate that the absence of functional CRAC channels does not result in compensatory activity of VGCCs.
Previous reports have shown that activation of STIM1, which is essential for the activation of CRAC channels by binding to ORAI1 43 , inhibits Ca 2+ influx through L-type Cav1.2 channels in response to depolarization 44,45 . These studies suggested that STIM1 reciprocally activates CRAC channels while suppressing Cav1.2. To test the hypothesis that STIM1 suppresses VGCC function in T cells, we first analyzed Ca 2+ influx in T cells from a patient with a null mutation in STIM1 (c.497 + 776 A > G) that abolishes STIM1 protein expression. Application of 60 mM [K + ] o to depolarize V m cells did not evoke a rise in [Ca 2+ ] i in T cells from either a HD or the STIM1-deficient patient (Fig. 6G). Whereas ionomycin induced SOCE in HD T cells, no increase in [Ca 2+ ] i was observed in the absence of STIM1 (Fig. 6G). We next analyzed Ca 2+ signals in T cells from WT and Stim1 fl/fl Cd4Cre mice with conditional deletion of STIM1 in T cells. Depolarization of Stim1-deficient mouse T cells by application of 60 mM or 150 mM [K + ] o failed to induce a rise in [Ca 2+ ] i (Fig. 6H, I). In WT T cells, depolarization with high [K + ] o suppressed SOCE induced by ionomycin stimulation (Fig. 6H, I). To exclude the possibility that VGCC function in Fig. 4 Depolarization of T cells fails to evoke Ca 2+ influx and Ca 2+ currents. A Cytosolic Ca 2+ levels in mouse CD4 + T cells. T cells were stimulated with anti-CD3 + CD28, cultured for 3 days and exposed to 60 mM (top) and 150 mM (bottom) KCl followed by stimulation with ionomycin (Iono). B Cytosolic Ca 2+ levels in human T cells from a healthy donor (HD) cultured for 10 days in vitro, exposed to 60 mM KCl and stimulated with ionomycin. Averaged Ca 2+ traces (left) and quantification (right) of the mean F340/F380 ratio during the time periods indicated by dotted lines. Data shown are the mean ± SEM of n = 3-4 (in A) and n = 7 (in B) independent experiments conducted in duplicates. C-H No voltage-gated Ca 2+ currents and signals in human T cells. C Membrane currents in HD T cells were recorded in 110 mM Ba 2+ in response to voltage steps from −60 to +60 mV for 200 ms from a holding potential of −80 mV. Current traces were leak-subtracted using the P/8 method with steps from −100 mV. D Current-voltage (I-V) plots (top) and [Ca 2+ ] i concentrations (bottom) measured using Indo-1 in the same cell stimulated in the presence of 20 mM extracellular Ca 2+ using the voltage protocol shown in (C). E [Ca 2+ ] i was measured in Indo-1 loaded HD T cells held at −80 mV for 20-30 s to establish the baseline [Ca 2+ ] i followed by application of 40-50 depolarizing steps to +10 mV every 1 s. F, G Simultaneous measurements of [Ca 2+ ] i and I CRAC . F T cells pretreated with TG were exposed to a step-ramp voltage protocol comprising a −100 mV step (50 ms) followed by a ramp from −100 to +100 mV (50 ms) every 1 s. The holding potential was +60 mV to prevent Ca 2+ influx during the interpulse interval.  Fig. 8C, D). In WT T cells, depolarization suppressed Ca 2+ influx following TCR and ionomycin stimulation as expected. Collectively, these data demonstrate that STIM1 deletion in either human or mouse T cells fails to induce voltage-gated Ca 2+ channel activity.   Given the limited specificity of most antibodies for all but some VGCC α 1 subunits, we also investigated their expression by analyzing RNA-Seq data of human and mouse T cells from a HD or WT mice, respectively, that were left untreated or stimulated by TCR crosslinking. We focused on VGCCs whose TPM normalized mRNA expression values exceeded 1 in T cells. These included, in descending order of total mRNA expression, mouse Cacna1a (Cav2.1), Cacna1i (Cav3.3) and Cacna1h (Cav3.2), as well as human CACNA1I (Cav3.3), CACNA1H (Cav3.2) and CACNA1F (Cav1.4) (Fig. 7A, B). Except for human Cav3.3, mRNA levels of VGCCs in T cells were markedly lower than those in reference tissues known to express VGCCs (brain, heart, skeletal muscle, retina). A similar mRNA expression pattern of α 1 subunits was observed in mouse CD4 + T cells that had been polarized in vitro into Th1, Th2, Th17 or iTreg cells as well as naturally occurring Treg cells 47 (Supplementary Fig. 10A). Again, Cacna1a (Cav2.1) was the most highly expressed α 1 subunit across all CD4 + T cell subsets. Compared to other T cell subsets, relatively higher mRNA levels were observed for Cacna1c (Cav1.2) in Th2 cells and Cacna1g (Cav3.1) in Treg cells ( Supplementary Fig. 10B). Because the α1 channel subunits Cav1.2, Cav1.4 and Cav3.1 have been implicated in the function of Th2 and Th17 cells, respectively 15,19,46,48 , we tested whether the differentiation of naïve CD4 + T cells into these T cell subsets is associated with the occurrence of voltage activated Ca 2+ influx. The polarization of mouse CD4 + T cells into Th2, Th17 and induced Treg (iTreg) cells was associated with the expected upregulation of lineage-specific transcription factors including GATA3 (Th2), Foxp3 (iTreg) and RORγt (Th17) ( Supplementary  Fig. 10C, D). Depolarization of Th2, Th17 and iTreg cells with 60 mM or 150 mM K + in the extracellular buffer, however, failed to evoke detectable Ca 2+ influx ( Supplementary Fig. 10E, F). By contrast, ionomycin induced robust SOCE in all T cell subsets, which was suppressed by high extracellular K + as expected.
Given the apparent expression of several α 1 subunits of VGCCs, we hypothesized that T cells may express alternatively spliced and non-functional isoforms of VGCCs. Indeed, splice variants of Cav1.1 and Cav1.4 have been reported in T cells 49,50 . We conducted an exon-level alignment of RNA-seq data for those VGCCs that we had found to be expressed in human and mouse T cells. mRNA for Cav3.3 is the most abundant of all VGCCs in human T cells. However, only transcripts of exons 12-37 were detectable in human T cells, which was in contrast to brain (frontal cortex) where all 37 exons of Cav3.3 (encoded by CACNA1I) are expressed (Fig. 7C). We found two putative transcription start sites (TSS) 5′ of exon 12 in human T cells by searching the refTSS dataset 51 (Fig. 7D). We made similar observations for Cav3.2 (CACNA1H), which is the second highest expressed VGCC in human T cells. Transcript levels of exons 1-12 were undetectable or very low in human T cells, whereas all exons (1-35) were expressed in brain tissue (Fig. 7E). We detected a putative TSS in exon 13, which may initiate mRNA expression in T cells (Fig. 7F). Cav2.1 (encoded by Cacna1a) is the most highly expressed α 1 subunit in mouse T cells at the transcriptional level. Exon usage analysis demonstrated that exons 1-33 (of 49) were not or weakly expressed in T cells (Fig. 7G). We detected 3 putative TSS in exon 34 of mouse Cacna1a, which may initiate transcription of a truncated mRNA (Fig. 7H). Collectively, these data demonstrate that although mRNAs for several VGCCs can be detected in mouse and human T cells, the transcripts are incomplete and result in N-terminally truncated proteins. For example, the non-transcribed exons 1-11 of CACNA1I encode amino acids (aa) 1-715 of the Cav3.3 protein, which form the N terminus of Cav3.3, its first channel domain (I), the I-II linker including the α-interaction domain (AID) and TM1-3 of the second channel domain (II) (Fig. 7H). Our data predict similar N-terminally truncated proteins for human Cav3.2 and mouse Cav2.1. Even if these proteins were stable and properly located in the plasma membrane, they would very likely not be functional Ca 2+ channels, providing an explanation for the absence of VGCC currents and Ca 2+ influx upon depolarization in T cells.

Discussion
We here identified Cavβ1 as a regulator of clonal expansion of T cells. By using an shRNA screening approach to identify ion channels that control T cell-mediated immunity, we found that Cavβ1 was required for the clonal expansion of CD4 + T cells after LCMV infection in vivo. Whereas deletion of Cavβ1 did not affect the proliferation of T cells, it was required to prevent T cell apoptosis following TCR stimulation in vitro. Three other Cavβ Fig. 6 Lack of ORAI1 or STIM1 does not induce voltage-gated Ca 2+ current or Ca 2+ influx in T cells. A-C Perforated patch recordings of human T cells from a patient with a loss-of-function (LOF) mutation (p.R91W) in ORAI1 (ORAI1 LOF ). T cells were left untreated (B) or stimulated with OKT3 (C) for 5-25 min prior to measurements. To record Ca 2+ current and cytosolic Ca 2+ levels, T cells were stepped from −60 to +60 mV for 200 ms from a holding potential of −80 mV. Displayed are membrane currents (top), I-V plots (middle) and [Ca 2+ ] i traces (bottom) measured simultaneously in the same cells in 20 mM Ca 2+ solution. Currents were leak-subtracted using the P/8 method. Data shown in (B, C) are representative of n = 7 and n = 3 cells, respectively. D, E T cells of the ORAI1 LOF patient were loaded with Indo-1 and either left unstimulated (D) or stimulated with OKT3 (E). T cells were stepped to +10 mV for 200 ms every second from a holding potential of −80 mV. Ca 2+ traces in (D, E) are representative of n = 5 and n = 4 cells, respectively. F Quantification of [Ca 2+ ] i at +10 mV in unstimulated and OKT3-stimulated T cells and after treatment with 5 μM ionomycin (Iono). Δ[Ca 2+ ] i was measured as the difference between the [Ca 2+ ] i prior to and at the end of 30 depolarization pulses. Data shown are the mean ± SEM from n = 4-5 cells. G Cytosolic Ca 2+ levels in human T cells from a patient with a STIM1 c.497 + 776 A > G null mutation (STIM1 null ). T cells were cultured for 10 days in vitro, loaded with Fura-2 and depolarized with Ringer's solution containing 60 mM KCl. Averaged Ca 2+ traces (left) and quantification (right) of the peak F340/F380 ratios during the time periods indicated by dotted lines. Data are the mean ± SEM from n = 5 independent experiments. (Note that Ca 2+ traces of the HD T cells are the same as those shown in Fig. 4B subunits have also been implicated in T cell function. Lack of functional Cavβ2 and Cavβ4 was associated with a severe defect in T cell development 20 , thymic and splenic involution and lymphocytopenia 22 . Cavβ3 deficiency resulted in impaired survival of naive CD8 + T cells 24 . Together, these studies and our data indicate that Cavβ subunits regulate T cell function, especially T cell survival. In contrast to the Cavβ1 deficiency phenotype described here, deletion or mutation of Cavβ2, Cavβ3 or Cavβ4 subunits resulted in moderately reduced TCR-induced Ca 2+ influx in T cells 20,23,24 , which was explained as arising from reduced Cav1.2/Cav1.3 and Cav1.4 protein expression in Cavβ2 and Cavβ3 deficient T cells, respectively 20,24 . It is noteworthy that we only observed robust expression of Cavβ1 and Cavβ3 in mouse and human T cells by RNA-Seq, but not that of Cavβ2 and  Cavβ4. Because we analyzed T cells from secondary lymphoid organs and blood of mice and humans, respectively, it is possible that Cavβ1 and Cavβ3 regulate the function of mature T cells such as survival, whereas Cavβ2 and Cavβ4 are critical in immature T cells during their development.
The mechanisms by which Cavβ1 controls apoptosis remain unclear. In excitable cells Cavβ subunits regulate VGCC function by promoting the cell surface expression of α 1 subunits and controlling channel activation and inactivation 7 . Reduced surface expression or activation of α1 subunits in the absence of Cavβ1 might therefore result in impaired Ca 2+ influx in T cells and explain impaired T cell survival. However, Cavβ1 deletion in T cells did not impair TCR-induced Ca 2+ influx, suggesting that Cavβ1 function is independent of regulating VGCC channels. Ca 2+ signaling is an important regulator of apoptosis in T cells with both pro-and anti-apoptotic effects observed 52 . However, our observation that Ca 2+ influx in T cells is unaffected by the loss of Cavβ1 excludes the possibility that increased apoptosis is caused by effects of Cavβ1 on Ca 2+ signals. Cavβ1-deficient T cells also showed normal production of cytokines such as IFNγ, TNF and IL-2 whose transcription is dependent on Ca 2+ signaling 2,53 , further demonstrating that Cavβ1 is not required for Ca 2+ influx in mouse T cells.
Although Cavβ proteins have mostly been thought of in terms of auxiliary subunits of VGCCs, a significant body of evidence demonstrates that they interact with many other proteins and have many VGCC-independent functions 7,26 . For instance, Cavβ proteins interact with other ion channels including ryanodine receptors, membrane receptors, Ras-related monomeric small GTP-binding (RGK) proteins, dynamin, actin and the scaffolding protein AHNAK1 26 . One of the most intriguing functions of Cavβ proteins is their function in controlling gene expression in the nucleus. A Cavβ4 splice variant was shown to interact with heterochromatin protein 1 γ (HP1γ), which mediates gene silencing 54 . Full-length Cavβ3 interacts with Pax6(S), an isoform of the transcription factor Pax6, to repress its transcriptional activity 55 . Moreover, overexpression of Cavβ4 in HEK293 cells was shown to modulate gene expression 56 . The Cavβ1a isoform was shown to localize to the nucleus of muscle progenitor cells (MPC) and bind to the myogenin promoter 57 . Deletion of Cavβ1a altered MPC expansion in vitro and in vivo, and changed global gene expression 57 . It is possible that Cavβ1 also controls gene expression in T cells, which would be distinct from its canonical purpose of regulating VGCC function in excitable cells 58 .
Because we identified Cavβ1 as a regulator of T cell function and previous reports had implicated other α 1 and β subunits in Ca 2+ influx and the function of T cells, we investigated the contribution of VGCCs to Ca 2+ signaling in T cells. Evidence supporting a function of VGCCs in T cells comes from the use of DHP Ca 2+ channel blockers, RNAi mediated knockdown of VGCC expression and knockout mice 15,17,19,20,23,24,[59][60][61] . Using a variety of measurement protocols in both human and mouse T cells, we were unable, however, to detect any evidence of functional VGCCs in T cells. Depolarization of T cells failed to evoke Ca 2+ influx and VGCC currents even under electrophysiological recording conditions greatly optimized to detect small currents. Similar results were obtained in T cells simultaneously activated by TCR stimulation based on the hypothesis that depolarization may not be sufficient to activate VGCCs in T cells and that additional stimuli may be required for their activation. Moreover, VGCC currents were undetectable in T cells lacking CRAC channel function. Because CRAC channels are the dominant Ca 2+ influx pathway in T cells after TCR stimulation, we reasoned that lack of CRAC channel function may result in a compensatory upregulation of VGCC currents, which was not the case. Lastly, deletion of STIM1, which was reported to inhibit Cav1.2 ectopically expressed in HEK293 or Jurkat cells 44,45 , did not evoke Ca 2+ influx upon depolarization of mouse or human T cells. Together, our studies fail to provide evidence for the existence of functional VGCCs in T cells.
How can these findings be reconciled with reports of VGCC function in T cells? Initial evidence supporting VGCC function in T cells came from the use of VGCC blockers including DHPs (amlodipine, nicardipine, nimodipine) and non-DHPs (verapamil, diltiazem), which were reported to inhibit Ca 2+ influx in T cells [60][61][62][63] . Micromolar concentrations of VGCC blockers, however, also inhibit several K + channels, notably Kv1.3 in T cells, thereby reducing V m and the electrical driving force for Ca 2+ influx through CRAC channels 35,64 . In fact, selective blockade of Kv1.3 and KCa3.1 channels in T cells inhibits Ca 2+ influx [65][66][67] . It is noteworthy that there is no compelling evidence showing that patients treated with VGCC blockers have impaired immune responses despite the widespread clinical use of these drugs for the treatment of many cardiovascular conditions 68 . More specific evidence supporting VGCC function in immunity comes from studies in mice. Targeted disruption of Cacna1f encoding Cav1.4 resulted in reduced CD4 + and CD8 + T cell numbers and function, most strikingly impaired T cell responses after infection with L. monocytogenes and murine gammaherpesvirus 68 (MHV-68) 17,18 . T cells of Cacna1f -/mice had reduced TCR stimulation-induced Ca 2+ influx and voltagedependent Ba 2+ currents 17 . We were not able to observe Ba 2+ currents upon depolarization of human peripheral blood T cells. Although this discrepancy could be due to a specific function of Cav1.4 in mouse but not human T cells, we failed to observe depolarization-induced Ca 2+ influx in both human and mouse T cells. Another study showed that deletion of Cacna1g encoding Cav3.1 impairs the production of cytokines by Th17 cells in vitro Fig. 7 Lack of full-length transcripts of VGCC α 1 subunits in T cells. A, B Absolute mRNA expression levels of α 1 subunits of VGCCs in human (A) and mouse (B) T cells and reference tissues. A CD4 + T cells were left unstimulated (−) or stimulated for 6 h with anti-CD3 + CD28 antibodies. HD represents the averaged data from three individual healthy donors (HD) and a patient with a STIM1 null mutation (c.497 + 776 A > G; STIM1 null ). B Mouse CD4 + and CD8 + T cells were left unstimulated (−) or stimulated for 12 or 24 h with anti-CD3 + CD28 antibodies. VGCC expression in T cells in (A, B) was analyzed by RNA sequencing and compared to that in human and mouse heart, skeletal muscle, frontal cortex, whole brain and retina using published datasets (Supplementary Table 2). C-F Exon usage of human CACNA1I (Cav3.3) (in C, D) and CACNA1H (Cav3.2) (in E, F) in CD4 + T cells and frontal cortex. C, E Normalized mRNA expression in unstimulated T cells from n = 3 HDs (averaged; red) and frontal cortex (blue) per exon. D, F Transcript levels (as RPKM, reads per kb of transcript, per million mapped reads) in frontal cortex and CD4 + T cells from an individual HD superimposed on exons 12-15 (for CACNA1I in D) and exons 12-14 (for CACNA1H in F). In T cells, exon 12 and exon 13 are the first transcribed exons of CACNA1I and CACNA1H, respectively. Green boxes and nucleotide sequences indicate predicted alternative transcriptional start sites (TSS). G Exon usage of Cacna1a (Cav2.1) in mouse CD4 + T cells (red) and brain (blue). Normalized Cacna1a mRNA expression per exon and exon-intron structure. and renders mice resistant to EAE, a mouse model of multiple sclerosis 19 . CD4 + T cells of Cacna1g -/mice had reduced voltageactivated Ca 2+ currents and Ca 2+ influx when exposed to extracellular Ca 2+ . This defect was more prominent in Th17 cells, which expressed higher levels of Cacna1g mRNA than Th1 or Th2 cells. By contrast, no defect in Ca 2+ influx was observed in Cacna1g -/-T cells after TCR stimulation. This study suggested that Cav3.1 channels mediate Ca 2+ influx at the resting membrane potential in T cells, especially in Th17 cells.
These findings regarding Cav1.4 and Cav3.1 functions in mouse T cells 17,19 are in apparent contrast to our failure to detect VGCC currents in human peripheral blood T cells. A potential explanation for this discrepancy could be that mouse T cells have functional VGCCs whereas human T cells do not. Arguing against this explanation is the fact that we failed to detect voltage-dependent Ca 2+ influx in mouse T cells under any of the conditions tested. To activate VGCCs, we depolarized human and mouse T cells to− 24 mV and~0 mV (with 60 mM and 150 mM [K + ] o , respectively), which would be sufficient to activate L-type (Cav1.4) and T-type (Cav3.1) channels that open at membrane potentials between −40 to −10 mV and −60 to −70 mV, respectively. Depolarization should, therefore, have induced voltage-dependent Ca 2+ currents and Ca 2+ influx in our experiments. An intriguing feature of T-type VGCCs such Cav3.1 that may explain their function in T cells is their activation at low voltages in the −60 to −70 mV range, which coincides with the resting V m of T cells (−53 to −59 mV) 69 . Window currents at these membrane potentials result from the overlap of activation and inactivation of T-type channels, and in the CNS were found to be important for the regulation of neuronal arousal 70,71 . In T cells, they may contribute to Ca 2+ influx and T cell function at resting V m and explain the effects of Cav3.1 deletion on T cell function in Cacna1g -/mice. Notwithstanding the potential importance of window currents for T cell function, if functional T-type VGCCs were expressed in T cells, we should have been able to detect Ca 2+ currents and Ca 2+ influx upon depolarization. Our inability to detect T-type VGCC currents in mouse T cells is consistent with the lack of Cav3.1 mRNA expression in mouse or human T cells.
Our RNA-Seq data demonstrated transcription of several genes encoding α 1 pore subunits in human and mouse T cells, but the α 1 pore subunits we detected were different from those reported before, namely Cav1.2, Cav1.3, Cav1.4 and Cav3.1 15,17,19,46 . The most abundant α 1 subunits we detected were Cav3.3 and Cav2.1 in human and mouse T cells, respectively, whereas transcripts encoding Cav1.2, Cav1.3, Cav1.4 and Cav3.1 were not or barely detectable. Moreover, we were unable to detect Cav1.2 or Cav1.4 proteins in either mouse or human T cells. The analysis of protein expression of other α1 subunits was limited by the quality of available antibodies, which included the validation of Cav3.3 protein expression. A potential explanation for our inability to detect VGCC function and mRNA of previously reported Cav α1 subunits is that their expression is restricted to particular T cell subsets. Cav1.2 levels were reported to be increased in human Th2 cells compared to Th1 or Th0 cells 46 , and Cav3.1 mRNA levels were transiently increased in Th17 cells 19 . Our analysis of published RNA-Seq data of mouse CD4 + T cells 47 confirms relatively higher mRNA levels for Cacna1c (Cav1.2) and Cacna1g (Cav3.1) in Th2 and Treg cells, respectively, compared to naïve T cells, Th1 and Th17 cells. However, electrophysiological evidence of functional VGCCs in Th2 or Treg cells is currently lacking.
Intriguingly, neither Cav3.3 and Cav2.1 have been implicated in T cell function before. To understand why their robust mRNA expression was not associated with VGCC function, we took advantage of the fact that RNA-sequencing reads can be aligned to individual exons to determine exon usage and potential splice variants in T cells. Whereas mRNA corresponding to all 37 exons of human CACNA1I was detected in the brain, where Cav3.3 function has been reported 72,73 , only mRNA corresponding to exons 12-37 of CACNA1I was detectable in human T cells. We observed a similar lack of transcription of 5′ exons for human CACNA1H (Cav3.2) and mouse Cacna1a (Cav2.1). We detected two TSS preceding exon 12 of CACNA1I and a TSS in exon 13 of CACNA1H, which coincide with the 5′ end of the mRNA transcripts in T cells. These 5′ truncated mRNAs are predicted to encode proteins that lack the NH 2 terminus of the α 1 subunits they encode including the entire channel domain I, the cytosolic I-II linker and part of domain II. It is unlikely, however, that the truncated proteins are stable or properly localize in the plasma membrane. While it is theoretically conceivable that a protein comprising domains III and IV of Cav3.3 is expressed and may assemble into a homomeric or heteromeric Ca 2+ channel that is gated by a voltage-independent mechanism, we consider this possibility to be remote.
Alternative splicing of VGCC α 1 subunits is well documented and an important mechanism to produce channels with distinct functional properties 5 . However, alternative splicing typically occurs in the C-terminus, and involves the alternative usage of individual exons. For instance, alternative splicing of Cav3.3 was shown to involve exon 9 encoding the I-II linker domain and exons 33 and 34 encoding part of the C-terminus of the channel, thereby giving rise to variants with distinct biophysical properties 74,75 . Alternative splicing of VGCCs has also been reported in T cells 49,50 . These include two variants of Cav1.4 in human T cells that lack expression of exons 31-34 & 37, and 32 & 37, respectively, predicted to delete the VSD of domain IV and its DHP binding site, which may render the channel insensitive to depolarization 49 . The alternative exon usage of human CACNA1I and CACNA1H as well as mouse Cacna1f in T cells we demonstrate here has not been reported before. Of note, Man et al. recently reported that human and mouse cardiomyocytes do not express full-length transcripts of the neuronal voltage-gated sodium channel Nav1.8 (Scn10a), but instead express a short transcript (Scn10a-short) comprising only the last 7 exons 76 . Transcription of this short variant occurred from an intronic enhancer-promoter complex. Overexpression of Nav1.8-short protein was shown to modulate the function of the main cardiac sodium channel Nav1.5 and heart rhythm. Whether truncated Cav3.3 protein is expressed in human T cells and suppresses VGCC function remains to be elucidated.
While our study does not support the existence of functional VGCCs in T cells, it suggests that Cavβ1 has alternative, VGCCindependent functions in T cells. These findings have implications for the assessment of the safety of VGCC channel blockers that are in wide clinical use for cardiovascular diseases, which based on our data are not expected to result in suppression of immune function. This interpretation is consistent with a lack of clinical evidence for immunosuppression in patients treated with Ca 2+ channel blockers.

Methods
Mice. All experiments were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee of at New York University Grossman School of Medicine. Stim1 fl/fl Cd4 Cre77 , Cacna1f-mutant 78 , congenic CD45.1 + SMARTA 79 have been described previously. Congenic CD45.2 + (strain 000664), Cd4Cre (strain 017336) and Rosa26-LSL-Cas9 knock-in (strain 024857) mice were purchased from the Jackson laboratory (Bar Harbor, ME). SMARTA; LSL-Cas9; Cd4Cre mice were generated by crossing SMARTA, Cd4Cre and LSL-Cas9 mice. All animals were on a pure C57BL/6 genetic background. Male and Female mice were used between 8 and 16 weeks of age. Mice were maintained under specific pathogen-free conditions with a 12 h dark/light cycle, at 22-25°C and 50-60% humidity with water and food provided ad libitum.
Human cells. Experiments using human cells were conducted in accordance with protocols approved by the Institutional Review Board of the New York University Grossman School of Medicine. Informed consent for the studies was obtained from the patients' parents and HDs in accordance with the Declaration of Helsinki. T cells from HDs and a patient homozygous for an autosomal recessive ORAI1 p.R91W LOF mutation (referred to as ORAI1 LOF ) were isolated from whole blood and stimulated as described 42 . Peripheral blood mononuclear cells were isolated from a 7yr-old male patient and his father, who were homozygous and heterozygous, respectively, for a cryptic splice acceptor site mutation in an intronic region between exons 4 and 5 of STIM1 (c.497 + 776 A > G). The mutation (referred to as STIM1 null ) abolished STIM1 protein expression and store-operated Ca 2+ influx (SOCE), and caused CRAC channelopathy syndrome in the patient 80 . A detailed case report describing the phenotype and molecular characterization of this patient will be published in a forthcoming manuscript.
Generation of shRNA library targeting ICTs. The shRNA library targeting 223 ion channels, transporters and regulatory proteins (ICTs) was custom-generated based on the analysis of mRNA expression levels of 658 ICTs in human and mouse immune cells using the Immunological Genome Project (ImmGen) 27 and Fantom5 28 databases. 223 ICTs were determined to be expressed at least twofold above the population average across all immune cell types and included in the shRNA library. Each ICT was targeted by five shRNAs. Also included in the pooled shRNA library were 34 positive controls (genes known to regulate T cell expansion and survival) and 13 negative controls (genes not expressed in mammalian cells nor in T cells). shRNAs were designed as described in 84 . Briefly, siRNAs targeting 223 ICTs were designed using the DSIR algorithm 85 and filtered further to select shRNAs with effective shRNAmir processing and potent knockdown 86 . For de novo generation of shRNA plasmids, 97-mer oligonucleotides (IDT Ultramers) coding for the respective shRNAs were synthesized on 55k arrays (Agilent) and cloned into the pLMPd recipient vector, which is based on miR-E vector and encodes Ametrine (Amt) as fluorescent reporter 81 . The pooled shRNA library containing 1342 shRNAs was sequenced by HiSeq 2500 (Illumina) to confirm equal representation of shRNAs and subsequently used to transfect Plat E cells for the production of pseudotyped retroviruses.
In vivo shRNA screen. LCMV-specific CD4 + T cells isolated from the spleens of CD45.1 + SMARTA mice were purified using the MagniSort Mouse CD4 + T cell Enrichment Kit and stimulated with plate-bound anti-CD3 and anti-CD28 antibodies. T cells were transduced 24 h after stimulation with the shRNA library packaged in pseudotyped retroviral particles at~0.3 multiplicity of infection in the presence of 8 μg/ml polybrene. 30 min after spin infection, the retroviral supernatant was diluted 1:2 with complete RPMI containing 20 IU/ml rh-IL-2 and 2.5 ng/ml IL-7. 16 h later, the T cell supernatant was replaced with fresh complete RPMI containing IL-2 and IL-7 16 h later. 3 days after transduction, transduced T cells (Amt + ) were enriched by cell sorting using a SY3200 (HAPS1) cell sorter (Sony) under sterile conditions. 1 × 10 6 Amt + SMARTA T cells were retro-orbitally injected into congenic CD45.2 + host mice. 1 × 10 6 Amt + SMARTA T cells were kept as input control. 2 days after adoptive T cell transfer, host mice were infected i.p. with 2 × 10 5 PFU of LCMV Armstrong (LCMV ARM ). 7 days later, CD45.1 + donor T cells were isolated from the spleen of host mice and CD4 + CD45.1 + Amt + T cells were purified by cell sorting. Genomic DNA (gDNA) was isolated from sorted cells using the FlexiGene DNA Kit (Qiagen, 51204) following the manufacturer's instructions. Next generation sequencing (NGS) libraries were generated from gDNA by PCR amplification of shRNA guide strands using primers that tag the product with Illumina HiSeq 2500 adaptors (primerA + mirE, AATGA-TACGGCGACCACCGAGAATTCTAGCCCCTTGAAGTC; primerB + BarCode + mirE-Loop, CAAGCAGAAGACGGCATACGAXXXXXTAGTGAAGCCACA-GATGTA). shRNAs were amplified in at least three individual 50 μl PCR reactions containing 500 ng of gDNA, 1U Platinun pfx DNA polymerase (Thermo Scientific, 11708013), 1× Platinum pfx buffer, 0.3 μM primers and 0.3 mM dNTPs using the following conditions: 94°C for 5 min; 35 cycles of 94°C for 15 s, 55°C for 30 s and 68°C for 20 s; 72°C for 7 min. The amplified shRNA library (~122 bp) was purified by phenol extraction and using the QIAEX II Gel Extraction Kit (Qiagen, 20021). The quality of the amplified library was assessed using the Agilent 2200 TapeStation Bioanalyzer (Agilent Technologies). The libraries were sequenced using the custom mirE-EcoR1 sequencing primer TAGCCCCTTGAAGTCCGAGGCAG-TAGGCA and the single read 50 bp (SR50) rapid-mode of the HiSeq 2500 DNA Sequencer (Illumina). Sequencing results were analyzed using the Model-based Analysis of Genome-wide CRISPR-Cas9 Knockout (MAGeCK) package 87 to identify shRNAs that were depleted or enriched in donor T cells after LCMV infection compared to input T cell samples.
Design and cloning of shRNAs and sgRNAs. shRNA target sequences against mouse Cacnb1 were extracted from the shRNA library. sgRNA target sequences were designed using Benchling software (https://www.benchling.com/crispr/). In addition, shRNA or sgRNA targeting a non-mammalian gene (Renilla luciferase) or human VEGF were used as controls. shRNA and sgRNA were cloned into the pLMPd 81 and pMRI 82 retroviral expression vectors, respectively, that encode Ametrine or GFP reporters. shRNA and shRNA sequences are provided in Supplementary Table 1.
RNA-sequencing. RNA-Seq data derived from human and mouse tissues were extracted from the Gene Expression Omnibus (GEO) database and are listed in Supplementary Table 2. FASTQ files were trimmed with Trimmomatic v0.36 88 and aligned with STAR/2.6.1 89 to the corresponding human GRCh38/hg38 or mouse GRCh38/mm10 genome assembly. BigWig files were generated using deeptools v3.1.0 90 and normalized using Reads Per Kilobase of transcript per Million reads mapped (RPKM). Human transcripts were counted using Salmon v0.14.1 91 with annotation from Gencode v30 Gene Transfer Format (GTF) file, and mouse transcripts were aligned and counted using the featureCounts function in the subread package v1.6.3 92 with annotation from Gencode Gene Transfer M21 (GTF GRCm38.p6) format file. The final data to generate heatmaps of gene expression for human and mouse samples was TPM normalized to allow comparison of expression levels between different genes. The final heatmap visualization was done in python using the mwaskom/seaborn: v0.8.1 package 93 .
Additional RNA-Seq data were generated using human CD4 + T cells from a STIM1 null patient (homozygous for the STIM1 c.497 + 776 A > G mutation) as well as his healthy brother and father (both heterozygous for the STIM1 mutation) and an unrelated healthy donor. CD4 + T cells were left untreated or stimulated with 5 μg/ml plate-bound anti-CD3 (clone OKT3, 14-0037-82) and 10 μg/ml soluble anti-CD28 (clone CD28.2, 14-0289-82, both eBioscience) antibodies for 6 h in RPMI1640 medium supplemented with 10% FBS, 1% L-glutamine, and 1% penicillin-streptomycin. Total RNA was extracted using the RNeasy Micro Kit (Qiagen, 74004) and RNA quality was analyzed using a Bioanalyzer 2100 (Agilent) using PICO chips. RNA-Seq libraries were prepared using the TruSeq RNA sample prep v2 kit (Illumina TruSeq Stranded mRNA, RS-122-2001) and 100 ng total RNA following the manufacturer's instructions. The amplified libraries were purified using AMPure beads (Beckman Coulter, A63881), quantified by Qubit 2.0 fluorometer (Life Technologies), and visualized using an Agilent Tapestation 2200. The libraries were pooled equimolarly, loaded on the HiSeq 2500 DNA Sequencer and run as single 50 nucleotide reads. For analysis of VGCC expression, data from the healthy donor, brother and father of the STIM1 null patient were averaged and are referred to as HD (healthy donor).
ICT expression and correlation analysis. ICT mRNA expression in human CD4 + and CD8 + T cells was analyzed using publicly available RNA-Seq datasets GSE87508 94 and GSE133822 95 as well as data from Haemopedia (https://www. haemosphere.org/) 96 . Raw data from Haemopedia and GSE133822 were normalized using the transcripts per million (TPM) method, and GSE87508 data was normalized using the Relative Log Expression method. Heatmaps, scatterplots and statistics were generated using MATLAB (R2019, Mathworks).
Exon usage analysis and transcriptional start site analysis. The exon-level quantification of VGCC mRNA expression in human and mouse tissues and cells was performed using the HTSeq package (v0.11.2) 97 with annotations from the human (GRCh38.p12) Gencode v30 GTF file and mouse (GRCm38.p6) Gencode M25 GTF file, respectively. The exon-level RNA expression data was normalized using the median of ratios method from DESeq2 98 Bioconductor package in R (version 3.6.1). This method was used to allow comparison between RNA expression per exon in T cells and reference tissues. Bars graphs showing quantification of DESeq2 normalized mRNA expression for each exon were generated using MATLAB R2019. The visualization of bigWig files generated from aligned reads after performing RPKM normalization was done using pyGenomeTracks 99 . Potential TSS near the 5′ end of mRNA transcripts in T cells were obtained from refTSS 51 .
Flow cytometry. Cells from tissue culture or isolated from mouse spleens were washed in cold PBS containing 3% FBS and 2 mM EDTA ("FACS buffer"). Staining of cell surface molecules with fluorescently labelled antibodies was performed at RT for 10 min in the dark. Intracellular (IC) cytokine staining was performed using the IC staining buffer kit (Invitrogen, 88-8824-00) according to the manufacturer's instructions. All antibodies were used at 1:200 dilution and the complete list of antibodies used in this study can be found in Supplementary Table 3. Briefly, T cells transduced with sgRNAs or shRNAs were stimulated with 1 μM ionomycin (Invitrogen, I-24222) and 20 nM phorbol myristate acetate (PMA, Calbiochem, 524400) for 6 h in the presence of brefeldin A (eBioscience, 00-4506-51). For apoptosis measurements, cells were incubated for 10 min at RT with Annexin V-AF647 in Annexin V Binding buffer (BioLegend, 422201). Samples were acquired using FACSDiva on a LSR Fortessa cell analyzer (BD Biosciences) and analyzed using FlowJo software (FlowJo 10.5.3). Cell sorting of transduced, Ametrine + T cells was performed using a MoFlo XDP cell sorter (Beckman Coulter).
Intracellular Ca 2+ measurements. T cells were loaded with the Ca 2+ sensitive dye Fura-2 (Invitrogen, F1221) for 30 min in complete RPMI1640 medium, and analyzed either using single-cell time-lapse microscopy or a microplate reader. For time-lapse microscopy, human T cells were loaded with 1 μM Fura-2-AM for 30 min and co-stained with anti-CD4-PE (Biolegend, 300550) and anti-CD8-FITC (Biolegend, 344704) at 1:200 dilution in the last 10 min of incubation, plated onto UV-sterilized coverslips pre-coated with 0.01% poly-L-lysine (Sigma-Aldrich, P8920) and perfused with buffer solutions in a RC-20 flow chamber (Warner Instruments). The chamber was mounted on an IX81 inverted epifluorescence microscope (Olympus). Some time-lapse imaging experiments were performed simultaneously with patch-clamping electrophysiology using a Nikon Diaphot TMD microscope equipped with a Nikon Fluor objective (NA 1.3). Fura-2 emission at 510 nm were measured following excitation at 340 nm and 380 nm using a Lambda DG-4 Plus light source and wavelength switcher (Sutter Instruments). F340/F380 emission ratios were calculated for each time point, and region of interests were used to discriminate human CD4 and CD8 T cells using the red (RFP) and green (GFP) filters, respectively, using SlideBook imaging software v4.2 (Olympus). Baseline Ca 2+ levels were determined after washing cells with 2 mM Ca 2+ Ringer's solution containing (in mM) 155 NaCl, 4.5 KCl, 2 CaCl 2 , 1 MgCl 2 , 10 D-glucose, and 5 HEPES (pH 7.4). T cells were depolarized by perfusion with Ringer's solution containing (in mM) either 60 KCl, 95 NaCl, 2 CaCl 2 , 1 MgCl 2 , 5 HEPES and 10 D-glucose or 150 KCl, 2 CaCl 2 , 1 MgCl 2 , 5 HEPES and 10 D-glucose. Ca 2+ levels in T cells were analyzed by measuring either the peak of F340/380 ratios or the area under the curve (AUC) for a specific time period as indicated in each experiment.
Patch-clamp electrophysiology. Patch-clamp experiments were conducted in the standard whole-cell recording configuration at room temperature using an Axopatch 200 amplifier (Axon Instruments) interfaced to an ITC-16 input/output board (Instrutech) and a Macintosh G3 computer. Recording electrodes were pulled from 100 μl pipettes coated with Sylgard and fire-polished to a final resistance of 2-5 MΩ. Stimulation and data acquisition and analysis were performed using in-house routines (R. Lewis, Stanford University) developed on the Igor Pro platform (Wavemetrics). Currents were filtered at 1 or 2 kHz with a 4-pole Bessel filter and sampled at 5 kHz. Data are corrected for the liquid junction potential of the pipette solution relative to Ringer's in the bath (−10 mV). Unless noted otherwise, all data were corrected for leak currents collected in Ca 2+ -free Ringer's solution (for CRAC currents) or using the P/8 method 37 for VGCC currents, and averaged results are presented as the mean ± SEM. Curve fitting was done by leastsquares methods using built-in functions in Igor Pro 4.0. For simultaneous measurements of [Ca 2+ ] i , human T cells were loaded with 1 µM Indo-1-AM in culture medium at 37°C for 15 min, washed and plated onto coverslips on the stage of a Nikon Diaphot TMD microscope. Cells were illuminated for 50-100 ms at 360 nm (360/25 filter; Chroma Technology) through a 40X Nikon Fluor objective (NA 1.3), and the fluorescence emissions at 405 nm and 485 nm (405/25 and 485/25 filters; Chroma Technology) were collected simultaneously with two photomultipliers (Hamamatsu). [Ca 2+ ] i was estimated from the relation [Ca 2+ ] i = K* (R-R min )/ (R max -R). K*, R min , and R max were measured in control human T cells in situ as previously described 100 . Ca 2+ -free Ringer's solution was prepared by substituting 1 mM EGTA + 2 mM MgCl 2 for CaCl 2 . For measurements of I CRAC and VGCC Ca 2+ currents using the perforated-patch configuration, 20 mM CaCl 2 was used. The isotonic Ba 2+ solution contained 110 BaCl 2 , 10 HEPES, and 10 D-glucose. The Δ[Ca 2+ ] i in cells stimulated with a train of depolarizing pulses was determined from the difference in [Ca 2+ ] i immediately prior to the first +10 mV pulse and following termination of the pulse train. For the Δ[Ca 2+ ] i rises evoked by TGactivated CRAC currents in response to −100 mV hyperpolarizing pulses, Δ[Ca 2+ ] i was determined as the difference in [Ca 2+ ] i following re-addition of extracellular Ca 2+ (20 mM) for 30 s. Internal patch-clamp solutions varied with the identity of the current being studied and had the following compositions (in mM): I CRAC : 150 Cs-methanesulfonate, 8 MgCl 2 , 10 BAPTA, and 10 Cs-HEPES (pH 7.2). L-type Ca 2+ current: 115 Cs-aspartate, 1 CaCl 2 , 5 MgCl 2 , 10 HEPES, and 10 NaCl. All internal solutions had a pH of 7.2. Thapsigargin (TG, LC Biochemicals) was diluted from a 1 mM stock in DMSO to a final concentration of 1 µM. The extracellular solution contained charybdotoxin (Sigma) at a final concentration of 2 nM. Solutions were applied to the cells using a multi-barrel local perfusion pipette with a common delivery port. The time for 90% solution exchange was measured to be <1 s, based on the rate at which the K + current reversal potential changed when the external Immunoblotting. Total cell lysates were prepared on ice using lysis buffer containing (in mM) 150 NaCl, 10 EDTA, 10 EGTA, 50 Tris, 1% Triton X-100, and 1% protease inhibitor cocktail (Sigma, P8340). For some protein extractions, lysis buffer was supplemented with 1% Nonidet P-40 (BioVision, 2111-100) and defined proteinase inhibitors: 10 μg/ml pepstatin A (Sigma-Aldrich 516481), 1 μg/ml leupeptin (Sigma-Aldrich L2884), 2 μg/ml aprotinin (Sigma-Aldrich A1153), 200 nM phenylmethanesulfonyl fluorid (PMSF; Millipore Sigma 10837091001). Cell debris