The MS4A (membrane-spanning 4-domain family, subfamily A) gene family contains 18 members in humans (Liang and Tedder, 2001; Zuccolo et al., 2010), and their encoded proteins show tissue-specific expression (Liang and Tedder, 2001; Zuccolo et al., 2010; 2013). CD20 (MS4A1) is specifically expressed only in B cells and is a highly selective cancer marker for B cell lymphoma (Herter et al., 2013; Reff et al., 1994). Ca²⁺ influx of CD20 is activated by the binding of anti-CD20 monoclonal antibody (mAb) (Beers et al., 2010; Bubien et al., 1993), which is unique as a successful surface target of mAbs (Walshe et al., 2008). These features have enabled the production of blockbuster antibodies, anti-CD20 antibodies, which have greatly contributed to the success of the biopharmaceutical industry (Rogers et al., 2014). Additionally, CD20 has been shown to have store-operated Ca²⁺ influx activity (Li et al., 2003; Vacher et al., 2015). As an analogue to CD20, the MS4A family member MS4A12, a colon-specific membrane protein, has also been found to exhibit tissue-specific expression and store-operated Ca²⁺ influx activity (Koslowski et al., 2008). However, characteristics of the Ca²⁺ channel activity of MS4A12 are unclear, and the potential of MS4A12 as a colon cancer marker is controversial.

Store-operated Ca²⁺ entry (SOCE) is a major Ca²⁺ entry pathway in the epithelia (Prakriya and Lewis, 2015; Sobradillo et al., 2014; Villalobos et al., 2017; Yuan et al., 2007). Among ion channels, SOCE channels have unique biophysical properties and regulation modes. The core uniqueness is regulation by STIM protein, the only known endoplasmic reticulum (ER) Ca²⁺ sensor until now (Prakriya and Lewis, 2015). Orai channels are representative SOCE channels regulated by STIM protein, the only known ER Ca²⁺ sensor identified to date (Prakriya and Lewis, 2015). STIM1 is located in the ER membrane and senses Ca²⁺ contents in the ER (Kim and Muallem, 2011; Prakriya and Lewis, 2015; Yuan et al., 2007). Once activated by Ca²⁺ store depletion, STIM1 proteins undergo conformational changes in their EF domain and then oligomerize and accumulate in puncta near the plasma membrane to gate the Orai1 channel (Feske et al., 2006; Wang et al., 2010). Other proteins also play a key role in SOCE channels that historically have SOC channel activity, such as TRPC1. Dynamic assembly of a TRPC1-STIM1-Orai1 ternary complex is involved in activation of Ca²⁺ entry, and this is a common molecular basis for SOC and CRAC channels (Ong et al., 2007; Sabourin et al., 2015). On the other hand, the pore component of L-type Ca²⁺ channel Ca_V1.2 is inhibited by over-expression of STIM1 (Park et al., 2010; Wang et al., 2010), and Orai1 increases the proximity of STIM1 to Ca_V1.2 (Wang et al., 2010). Reportedly, over-expression of MS4A12 increased Ca²⁺ permeability via thapsigargin, and the observed permeability decreased when MS4A12 expression was knock-downed by siRNA (Koslowski et al., 2008). However, MS4A12 has not been investigated as an SOCE channel regulated by Orai1 and STIM1.

Tumour cells exhibit distinct and acquired traits that are relevant to or influenced by changes in Ca²⁺ handling (Monteith et al., 2007; Parekh, 2010; Roderick and Cook, 2008). MS4A12 was first reported to show increased expression in patients with colon cancer (Koslowski et al., 2008). However, recent reports have indicated that MS4A12 is predominantly expressed in normal colon tissues and tends to show decreased expression when transformed into cancerous tissues (Dalerba et al., 2011; Drew et al., 2014; Roberts et al., 2015). Although the direction of changes in expression during cancer progression is unclear, MS4A12 is a promising colon cancer marker for predicting both cancer diagnosis and progression (Dalerba et al., 2011; Drew et al., 2014; Koslowski et al., 2008).

GCaMP fluorescence proteins are robust tools with which to monitor Ca²⁺ concentrations nearby ion channels (Akerboom et al., 2013; Ko et al., 2017; Nguyen et al., 2020). Indeed, multiple gating states and oscillation of Orai1 has been documented by genetically encoded Ca²⁺-indicators for optical imaging fused to the N-terminal of Orai1 protein (Dynes et al., 2016), and lysosomal Ca²⁺ efflux has been successfully measured by GCaMP3-tagged TRPML1 (Shen et al., 2012). In this study, we evaluated SOCE activity for MS4A12 based on Orai1 and STIM1 interactions. We used GCaMP6s fused with MS4A12 to measure Ca²⁺ influx near MS4A12 and to test hetero-complex formation with Orai1. Additionally, expression changes of MS4A12, as well as Orai1 and STIM1, were observed in normal and cancer tissues from the human colon.

Clones, cell lines, materials, and human colon samples

pCMV-SPORT6-MS4A12 clones (long isotype of hMS4A12) were purchased from Korea Human Gene Bank (hMU004204), and 3X flag-MS4A12 was sub-cloned into pCMV5 vectors using NotI/BamHI site. pLenti6.3-GCaMP6s-MS4A12 was produced using BamHI/SpeI sites for GCaMP6s and SpeI/XhoI sites for MS4A12 into pLenti6.3 vectors. C-term truncated MS4A12 was produced by insertion of a stop codon by mutagenesis at the 224th amino acid. Truncation mutants of STIM1 and STIM1^D76A were described in a previous study (Yuan et al., 2007). Flag-Orai1 was kindly provided by Shmuel Muallem. Thapsigargin (T9033), 2-APB (D9754), and BTP2 (YM-58483) were purchased from Sigma (USA). Anti-MS4A12 antibody (HPA057657) and anti-flag (9E10) antibody were purchased in Sigma. Anti-myc antibody (2276s), anti-GOK/STIM1 antibody (610954), anti-Orai1 antibody (ACC-062), and ZO-1 antibody (ab190085) were purchased from Cell Signaling Technology (USA), BD Bioscience (USA), Alomone (Israel), and Abcam (UK), respectively. Aldolase A (N-15, sc-12059; Santa Cruz Biotechnology, USA) and β-actin (sc-1616; Santa Cruz Biotechnology) were used, and anti-mouse, anti-rabbit, or anti-goat IgG conjugated with Alexa Fluor 488, 563 (Invitrogen, USA) was utilized as secondary antibody. Anti-mouse, anti-rabbit, or anti-goat IgG conjugated with HRP was purchased form Thermo Fisher Scientific (USA). Plasmids were transiently transfected into HEK293 cells using Lipofectamine 2000 (Invitrogen) and T84 cells using ExGen 500 (Thermo Fisher Scientific). siRNA was transfected with RNAimax (Invitrogen). HEK293, A549, and CaCo2 cells were maintained in DMEM-High Glucose (Invitrogen). T84 cells were maintained in a 1:1 mixture of Ham’s F-12 medium and DMEM (Invitrogen). H508 and HT29 cells were maintained in RPMI1640. All media were supplemented with 10% fetal bovine serum (Invitrogen), penicillin (50 IU/ml), and streptomycin (50 μg/ml). The experiments using human colon and cancer tissue samples were approved by the Institutional Review Board of Severance Hospital, Yonsei University College of Medicine, Seoul, Korea (IRB protocols No. 4-2014-0349). Normal tissues were obtained from tumor margins under inspection by an oncology surgeon. All subjects provided written informed consent, and their clinicopathologic information are outlined in Supplementary Table S1.

Plasma expression of GCaMP-MS4A12

HEK cells were transfected with GCaMP-MS4A12 alone and GCaMP-MS4A12 with untagged MS4A12 or Flag tagged Orai1 (plasmid 1:1 ratio) using Lipofectamine 2000, and were grown on glass coverslips. After growing for 36 h, the cells were stimulated with 1 μM thapsigargin in 5 mM Ca²⁺ regular solution for 5 min. Then, cells on coverslips were fixed with 4% paraformaldehyde in phosphate-buffered saline and mounted on slide glass. Images were obtained with a Zeiss LSM 780 confocal microscope (Zeiss, Germany).

Ca²⁺ influx measurement using GCaMP6s-MS4A12

No-tagged MS4A12 (nt-M12) and GCaMP-MS4A12 plasmids were mixed at a 1:1 ratio and then transfected into HEK293 cells grown on coverslips. In case of Orai-GCaMP-MS4A12 hetero complex, flag-tagged Orai1 was used instead of nt-M12. At 48 h after transfection, changes in GCaMP fluorescence were assessed using a Zeiss 780 confocal microscope (Zeiss). siRNA of Orai1 and STIM1 was also transfected 1 day before MS4A12 transfection. During fluorescence measurements, cells were continually perfused with a Ca²⁺ free regular solution (37°C) containing 150 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 10 mM glucose, and 10 mM HEPES at pH 7.4 with NaOH. Thapsigargin of 5 μM was used for stimulation and store depletion, after which 5 mM Ca²⁺ regular solution was perfused to visualize Ca²⁺ influx-induced fluorescence of GCaMP-MS4A12. To inhibit SOCE mediated Ca²⁺ influx, 5 μM 2-aminoethoxydiphenyl borate (2-APB) was treated during the measurement, and 500 nM BTP2 was pre-treated 5 min before starting the measurement. Ca²⁺ influx activity was calculated according to changes in GCaMP6s fluorescence (ΔF) at each time point over basal fluorescence (F₀; ΔF/F₀). Averages ΔF/F₀ of all measurements were obtained from more than three independent experiments (each comprising analyses of 4-20 cells).

Immunoprecipitation and Western blot analysis

For immunoprecipitation assay, HEK293 cells transfected with appropriate plasmid were lysed with sonication in lysis buffer (150 mM NaCl, 5 mM Na-EDTA, 10% glycerol, 20 mM Tris-HCl [pH 8.0], 0.5% Triton X-100, and proteinase inhibitors [Complete; Roche Applied Science, USA]). To capture the protein, appropriate antibody was incubated for more than 4 h, after which protein A/G beads were added to capture antibody bound protein complex. Beads were washed with 400 μl of lysis buffer four times, and 2× sample buffer was used to elute whole protein complexes. For Western blot analysis, cells were lysed in lysis buffer, and lysates were separated on 4% to 12% pre-made gradient sodium dodecyl sulfate-polyacrylamide electrophoresis gels (Koma Biotech, Korea) and transferred to membranes. For tissue experiments, human colon and cancer tissue were homogenized with lysis buffer after chopping them into small pieces. Appropriate antibodies were used to detect the proteins, and HRP-conjugated secondary antibody and ECL solution (Amersham Bioscience, UK) were used for detection by chemi-luminescence. Quantification of band intensity was performed using MultiGauge software (Fujifilm, Japan).

Reverse transcription polymerase chain reaction

Total RNA was extracted using TRIzol reagent (Invitrogen), and an equal amount of RNA from each sample was reverse-transcribed to cDNA using Superior Script II Reverse transcriptase (Enzynomics, Korea). The cDNA was amplified with specific primers and a Taq polymerase (2X TOPsimple Premix; Enzynomics). The polymerase chain reaction (PCR) primer sequences were as follows: hMS4A12 sense, ACA TTG CCC TGG GGG GTC TTC T; hMS4A12 antisense, CCA GAA ATG GCA GCA AAG AGG CT (205 bp); hSTIM1 sense, GGAAGACCTCAATTACCATGAC; hSTIM1 antisense, GCTCCTTAGAGTAACGG TTCTG (440 bp); hOrai1 sense, ACC TGC ATC CTG CCC AAC ATC; hOrai1 antisense, GCC CAG GCC AGC TCG ATG TG (107 bp); β actin-sense, GAC CCA GAT CAT GTT TGA GAC C; β-actin antisense, GGC CAT CTC CTG CTC GAA GTC; 18s ribosome sense, GTGGAGCGATTTGTCTGGTT; and 18s ribosome anti-sense, CGCTGAGCCAGTCAGTGTAG (199 bp).

Statistical analysis

Data are presented as the mean ± SEM. Statistical analysis was performed with Student’s t-tests, ANOVA, followed by Tukey’s multiple comparison, or one-way ANOVA using GraphPad Prism software (ver. 5.0; GraphPad Software, USA), as appropriate. P values ≤ 0.05 were considered statistically significant.

Characteristics of MS4A12 as a store-operated Ca²⁺ influx channel

To measure Ca²⁺ influx near MS4A12, GCaMP-MS4A12 (GCaMP-M12) was constructed by fusing MS4A12 with the GCaMP6s Ca²⁺ indicator, which fluoresces in response to a short-term Ca²⁺ concentration increase near MS4A12. However, GCaMP-M12 did not reach the cell membrane, likely due to the high solubility of GCaMP. HEK cells expressing GCaMP-M12 alone showed fluorescence throughout the cytosol when stimulated with thapsigargin in 5 mM Ca²⁺ solution (Fig. 1A). This problem was overcome by co-expressing GCaMP-M12 with nt-M12 at a 1:1 DNA transfecting ratio: GCaMP-M12 co-transfected with nt-M12 showed plasma membrane localization in the same thapsigargin condition (Fig. 1B). Successful plasma membrane localization proved that GCaMP-M12 complexed with nt-M12. Next, we evaluated the pharmacological characteristics of MS4A12 store-dependent Ca²⁺ influx using plasma membrane localized GCaMP-M12. First, we confirmed the spontaneous Ca²⁺ influx of MS4A12 (Fig. 1C). When 5 mM Ca²⁺ was perfused without stimulation, we observed no fluorescence increases in GCaMP-M12. However, when cells were stimulated with a store-depleting condition (5 μM thapsigargin for 5 min in the 5 mM Ca²⁺ extracellular solution), the fluorescence of GCaMP-M12 very rapidly and greatly increased (Supplementary Movie S1). Additionally, the increased fluorescence of GCaMP-M12 by store-depleting condition was inhibited by 2 mM 2-APB (Fig. 1E, Supplementary Movie S2). Furthermore, pre-treatment with 500 nM of BTP2 (Fig. 1F) completely blocked any florescence increases in the store-depleting condition.

Both the robust increase in GCaMP-M12 fluorescence under the thapsigargin and external 5 mM Ca²⁺ condition and the lack of fluorescence changes upon thapsigargin-induced internal Ca²⁺ efflux clearly demonstrated that the increased fluorescence of GCaMP-M12 occurred as a result of increased Ca²⁺ concentrations near GCaMP-M12 or increased Ca²⁺ influx through MS4A12 in plasma membrane. Accordingly, thapsigargin-dependent Ca²⁺ influx and its blockade by typical SOCE inhibitors revealed that the complex of GCaMP-M12 with nt-M12 monitors or functions as a store-operated Ca²⁺ influx pore.

Orai1 co-expression increases and Orai1 knockdown decreases thapsigargin-induced increases in GCaMP-M12 fluorescence

Measuring Ca²⁺ influx in Fura2-AM loaded HEK cells, we noted that over-expression of MS4A12 elicited increased thapsigargin-activated Ca²⁺ influx, which was further increased by STIM1 co-expression (Supplementary Fig. S1). However, since the magnitude thereof was much smaller than STIM1 effects on Orai1, we presumed that the reason why MS4A12 showed characteristics of SOCE might be that MS4A12 could hetero-complex with Orai1. Also, STIM1 could control Orai1 gating, resulting in Ca²⁺ influx through the hetero-complex of MS4A12 and Orai1. To investigate the role of Orai1 in the SOCE-like behaviour of MS4A12, we observed the plasma membrane expression of GCaMP-M12 co-transfected with Flag-tagged Orai1 in its N-terminal (Flag-Orai1). As when co-expressed with nt-M12 (Figs. 1A and 1B), the fluorescence of GCaMP-M12 was observed at the plasma membrane when co-expressed with Flag-Orai1, implying that GCaMP-M12 complexes with Flag-Orai1 (Fig. 2A). Next, we compared changes in fluorescence upon thapsigargin-induced Ca²⁺ influx between GCaMP-M12 with nt-M12 and GCaMP-M12 with Flag-Orai1. Interestingly, the relative fluorescence changes were significantly higher for GCaMP-M12 co-transfected with Flag-Orai1 than for GCaMP-M12 co-transfected with nt-M12 (Figs. 2B and 2C). Additionally, we examined thapsigargin-induced Ca²⁺ influx-mediated fluorescence changes for GCaMP-M12 co-expressed with nt-M12 after employing siRNA for Orai1 (siOrai1). In siOrai1 treatment, increases in GCaMP-M12 fluorescence were delayed, and the degree of fluorescence increase was reduced (Figs. 2D-2G). To prove protein interactions between MS4A12 and Orai1, immunoprecipitation experiments were performed, in doings so, we found that GCaMP-M12 indeed interacted with Flag-Orai1 and that their interaction was enhanced upon thapsigargin treatment. Moreover, GCaMP-M12 and Orai1 interactions were confirmed by live cell imaging using GCaMP-M12 and mCherry-Orai1 expressed HEK cells (Fig. 2H). Therein, both proteins co-localized at the plasma membrane, and their co-localization was enhanced by thapsigargin treatment (Fig. 2I). Altogether, these results demonstrated that MS4A12 complexes with Orai1 and that the Ca²⁺ influx activity of MS4A12 is dependent on Orai1.

STIM1 co-expression decreases and STIM1 knockdown increases thapsigargin-induced increases in GCaMP-M12 fluorescence

Next, we examined the effects of STIM1on the SOCE-like behaviour of MS4A12. To do so, GCaMP-M12 with nt-M12 expressing cells were co-expressed with STIM1 via pIRES2 DsRED vector (expresses STIM1 and DsRED fluorescent proteins separately) to distinguish between STIM1-expressing red fluorescent cells (yellow dotted line cells, Fig. 3A) and non-expressing non-fluorescent cells (white dotted line cells, Fig. 3A). Cells expressing STIM1 showed significantly reduced changes in GCaMP-M12 fluorescence(Figs. 3B and 3C), compared to cells not expressing STIM1 (white dotted line cells, Figs. 3D). Meanwhile, when STIM1 expression was decreased with siRNA against STIM1 (siSTIM1), the fluorescence of GCaMP-M12 complexed with nt-M12 was more intense (Figs. 3D-3G). This was an unexpected and very compelling result, and implied that Ca²⁺ influx through MS4A12 is inhibited by STIM1 co-expression. In immunoprecipitation experiments, Flag-MS4A12 with myc-STIM1 showed protein interaction in normal conditions, and this interaction was slightly enhanced by thapsigargin (Fig. 3H). In addition, immunoprecipitation using each truncated protein showed that the cytosolic C terminal region (amino acids 225 to 267) of MS4A12 interacts with the CCD2+3, C2+3 region of STIM1 (STIM1-Orai activating region [SOAR] of STIM1) (Figs. 3I and 3J, Supplementary Fig. S2). Next, to examine the effect of Orai1 on the binding of STIM1^D76A (constitutive active form) with MS4A12, Orai1 was over-expressed, and the binding strength of STIM1^D76A with MS4A12 was observed (Fig. 3K). Interestingly, MS4A12 binding to STIM1^D76A was reduced by Orai1 over-expression. These results indicated that STIM1 might face competitive in the presence of Orai1. Immunostaining for MS4A12 and STIM1 revealed that these two proteins are closely localised at some region and that, even in a normal state, thapsigargin treatment did not elicit big large changes in their co-localization (Fig. 3L). Overall, these data suggested that MS4A12 is not a canonical SOCE channel directly regulated by STIM1, because Ca²⁺ influx through MS4A12 is inhibited by STIM1 and because MS4A12 binding to STIM1^D76A is inhibited by Orai1 over-expression.

Apical expression of MS4A12 in human colonic crypts and polarised T84 cells decreases in colon cancer malignancy

The expression of MS4A12 in colon cancer is controversial (Dalerba et al., 2011; Drew et al., 2014; Koslowski et al., 2008). We compared the expression levels of MS4A12 and Orai1 in samples from patients with colon cancer and colon cancer cell lines using RT-PCR and quantitative PCR (Figs. 4A and 4B). H508, CaCO-2, and HT29 cells, human adenocarcinoma cell lines originating from the colon, were used to examine mRNA expression levels of MS4A12. A549 cells, a human adenocarcinoma originating from the lung, was used as a negative control of MS4A12 expression. All three colon cancer cell lines, as well as A549 cells, showed no MS4A12 mRNA expression, whereas Orai1 showed high mRNA expression. In addition, MS4A12 expression was significantly lower in the cancer tissue than normal tissue (margins of colon cancers) from surgically removed specimens from colon cancers patient (described in Supplementary Table S1), which indicated that the expression of MS4A12 decreases during cancerous transformation. Moreover, we examined protein integrity and expression levels in normal and cancer tissue lysates (Fig. 4C). MS4A12 protein was clearly observed in normal colon tissue lysates, but significantly decreased in cancer tissue lysates. Accordingly, we deemed that protein and mRNA expression of MS4A12 is decreased in cancer states of the human colon and that MS4A12 expression is normal in the human colon protein, not a colon cancer marker.

MS4A12 was reported as a store-operated Ca²⁺ channel (Koslowski et al., 2008). However, no studies have examined the specific molecular mechanisms of the store-operated Ca²⁺ channel function of MS4A12. Our study demonstrated that the store-operated Ca²⁺ influx activity of MS4A12 is driven by a hetero-complex formed by MS4A12 and Orai1. The plasma membrane localization of GCaMP-M12 by co-expression of nt-MS4A12 or flag-Orai1 indicates that the hydrophobicity required for the membrane insertion is achieved by homo (with nt-M12; Fig. 1A)- or hetero (with Flag-Orai1; Figs. 2A and 2B)-complex formation with GCaMP-M12. Immunoprecipitation and co-localization of MS4A12 and Orai1 confirmed that MS4A12 interacts with Orai1 and that their interaction is further enhanced by thapsigargin (Figs. 2H and 2I). Interestingly, STIM1 co-expression inhibited thapsigargin-activated increases in fluorescence for GCaMP-M12 complexed with nt-M12 (Figs. 3A-3G), and this was recovered by siSTIM1 treatment. In addition, Flag-Orai1 co-expression decreased MS4A12 protein interactions with STIM1 (Fig. 3K). Altogether, these data imply that MS4A12 contributes to store-operated Ca²⁺ influx via complex formation with Orai1. However, given the inhibited store-operated Ca²⁺ influx upon STIM1 over-expression, MS4A12 seems to be a non-canonical component protein for store-operated Ca²⁺ influx.

The characteristic of MS4A12 as a store-operated Ca²⁺ channel only when it forms a hetero-complex with Orai1 is similar to that of TRPC3, which store-operated Ca²⁺ influx activity dependent on TRPC1 (Yuan et al., 2007). However, TRPC3 does not bind to STIM1 directly. In this regard, the relationship between MS4A12 and STIM1 is different from that between TRPC3 and STIM1, in that MS4A12 can bind to STIM1, but their binding is competitively inhibited by Orai1. Additional observations are necessary to determine the physiological situation in which MS4A12 forms a hetero-complex with Orai1.

The locations of MS4A12, STIM1, and Orai1 appear to differ in cancer tissue and normal tissue (Supplementary Fig. S2). The apical major localization of MS4A12 in normal colon tissue (Supplementary Fig. S2A) was not observed in cancer tissues (Supplementary Fig. S2B). In addition, STIM1 and Orai1 localization changed in cancer tissue. Most of all, in cancer tissue, MS4A12 was nearly completely diminished, as seen in Western blot analysis (Fig. 4C). Considering the contribution of aberrant ion channel localization during cancer progression, such as Ca_V1.2 in colon cancer (Wang et al., 2000) and TRPM8 in prostate cancer (Monteith et al., 2007; Thebault et al., 2005), the altered location of Orai1 and STIM1, as well as the disappearance of MS4A12, may contribute to cancer progression.

Given that anti-CD20 monoclonal antibodies are effective immune-therapeutics for human malignancies, other MS4A family members have been identified as potential targets for therapeutic intervention (Koslowski et al., 2008; Liang and Tedder, 2001). However, MS4A12 is unlikely to be useful as a colon cancer-specific target for cancer immunotherapy because its expression is mostly eradicated during cancer transformation (Fig. 4). Whether MS4A12 can be used as a colon cancer immunotherapy target with Ca²⁺ influx function should be reconsidered. Rather, the decreased expression of MS4A12 may be a marker of poor prognosis (Dalerba et al., 2011; Drew et al., 2014). Based on our results showing reduced expression in cancer tissues, it is necessary to further investigate the role of MS4A12 under normal colon conditions and the effect of decreased MS4A12 expression on the malignant processes of colon cancer. In the meantime, we have identified another colon-expressed MS4A member, MS4A8, that shows increased expression upon cancer transition. However, this protein does not appear to be expressed in the plasma membrane according to immunocytochemistry data using MS4A8 over-expressing cells and live cell imaging of mCherry tagged-MS4A8 (Supplementary Fig. S3).

In conclusion, we suggest that MS4A12 is not a STIM1-regulated canonical store-operated Ca²⁺ influx channel and shows store-operated Ca²⁺ influx activity when it forms a hetero-complex with Orai1. Additionally, our results indicate that MS4A12 is a colon protein marker and that its expression is decreased during cancerous states.

This study was supported by a Korean government grant NFR-2019R1A2C1086348 and a faculty research grant from Yonsei University College of Medicine (4-2016-0600) for J.Y.K.

J.Y.K. and M.J. conceived of and designed the experiments. J.W.H. performed the most of the experiments. W.H., D.L., H.Y.K., and I.J. performed IP experiments. W.H. and C.K. performed live imaging experiment using GCaMP-MS4A12. H.H. and M.J. contributed the materials from patients. J.Y.K. and J.W.H. analysed the data. I.S. and M.G.L. participated in design experiments and contributed reagents, materials and instruments. J.Y.K. wrote the manuscripts. All authors read and approved the final manuscript.

The authors have no potential conflicts of interest to disclose.