Inhibitor of Myogenic Family, a Novel Suppressor of Store-operated Currents through an Interaction with TRPC1*

Depletion of intracellular Ca2+ stores leads to the activation of Ca2+ inflow through store-operated Ca2+ channels. Although the identity of these channels is unknown, there is considerable evidence that the transient receptor potential channel 1 (TRPC1) participates in the formation of these channels. We show that TRPC1 physically interacts with the a-isoform of the inhibitor of the myogenic family (I-mfa), a known inhibitor of basic helix-loop-helix transcription factors, in vitro and in vivo. The interaction is mediated by the C-terminal cytoplasmic tail of TRPC1 and the C-terminal cysteine-rich domain of I-mfa. Using the whole cell configuration of the patch clamp technique, we show that ectopic expression of I-mfa in CHO-K1 cells reduces native store-activated Ca2+ currents, whereas knock-down of endogenous I-mfa in A431 cells by RNA interference enhances these currents. Pipette perfusion of purified recombinant I-mfa rescues the effect of I-mfa knock-down on store-operated conductance. Finally, cell dialysis with a monoclonal antibody specific to TRPC1 results in the suppression of store-activated conductance in cells lacking I-mfa, but not in I-mfa expressing cells. We propose that I-mfa functions as a molecular switch to suppress the store dependence of TRPC1.

G protein-coupled receptors and receptor tyrosine kinases comprise a very large group of cell surface receptors that elicit their physiological responses through the production of inositol (1,4,5)-trisphosphate (IP 3 ) 1 (1). Upon receptor activation, newly synthesized IP 3 acts on IP 3 receptors (IP 3 Rs) to trigger a rapid increase in the intracellular Ca 2ϩ concentration by releasing free Ca 2ϩ from intracellular stores (2). Intracellular Ca 2ϩ concentration returns to normal levels by extrusion of cytoplasmic Ca 2ϩ to the extracellular space by plasma membrane Ca 2ϩ -ATPase and Na ϩ -Ca 2ϩ exchangers, re-admission of Ca 2ϩ into the endoplasmic reticulum by the SERCA pump, and Ca 2ϩ entry via the store-operated Ca 2ϩ channels (3,4). Whereas all of these pathways contribute to the maintenance of normal Ca 2ϩ homeostasis, store-operated Ca 2ϩ entry is of particular interest because it is responsible for the regulation of diverse cellular functions (5) ranging from cell proliferation and gene expression in T lymphocytes (6) to endothelial cell function (7) and regulation of the acrosome reaction in germ cells (8).
Despite intensive investigation, the identity of the storeoperated channels and the cellular mechanisms underlying the coupling of store depletion and Ca 2ϩ entry remain elusive (9,10). Mammalian TRPCs have been proposed to function as store-operated channels (11,12). However, it is not clear whether they are store-, receptor-operated (13,14), and/or channels serving other functions.
TRPC1 was the first mammalian TRP channel cloned (15,16) and there has been considerable evidence to support its role in store-operated Ca 2ϩ entry. First, genetic elimination of TRPC1 by homologous recombination significantly reduces Ca 2ϩ release-activated Ca 2ϩ currents in DT40 cells (17). Second, TRPC1 knock-down by RNAi reduces store-operated currents in CHO-K1 cells (18). Third, heterologous expression of TRPC1 augments store-operated Ca 2ϩ entry in some cell types including CHO-K1 cells (19,20). Fourth, TRPC1 is responsible for the Ca 2ϩ /calmodulin-dependent feedback inhibition of store-operated Ca 2ϩ entry, a hallmark of native store-activated Ca 2ϩ conductance (21). Fifth, TRPC1 can physically interact with various isoforms of the IP 3 R (22-24) and polycystin-2 (25), both functioning as intracellular Ca 2ϩ release channels activated in response to IP 3 (26,27). This is consistent with the conformational coupling model of the store-operated Ca 2ϩ entry by which store-operated channels are gated by Ca 2ϩ release channels via protein-protein interactions (28).
In contrast, there are independent functional and biochemical studies showing that TRPC1 does not have channel function by itself, but it is required for the formation of receptor-but not store-activated channel complexes with TRPC4 and TRPC5 (29 -31). In addition, when TRPC1 is successfully expressed in CHO-K1 cells it does not reconstitute the properties of the prototypical Ca 2ϩ release-activated Ca 2ϩ current extensively studied in hematopoietic cells (19,32). Although these findings are difficult to reconcile with each other, most of the controversy appears to be because of results obtained from gainversus loss-of-function experiments. Thus, it has been suggested that as yet unidentified proteins may be required to successfully reconstitute a true store-operated channel containing one or more of the currently known TRPCs (6,33,34). Another possibility is that proteins with inhibitory activity may prevent activation of mammalian TRPCs by store-depletion, similar to the Calliphora vicina TRP homolog (35). It was recently shown that the PDZ containing domain protein, INAD (inactivation no afterpotential D) inhibited the activation of C. vicina TRP by store depletion (35).
I-mf is the founding member of a group of related proteins with an unusual cysteine-rich domain (36), first identified as an interacting protein with MyoD in a yeast two-hybrid screen. I-mf exists in three isoforms (a, b, and c) produced by alternative splicing (37), with each isoform containing a common Nterminal and a variable C-terminal region (a, b, or c) (37). All three transcripts encode small cytoplasmic proteins lacking any predicted transmembrane segments or nuclear localization signals (37). I-mfa, the a-isoform, inhibits the transcriptional activity of MyoD and other basic helix-loop-helix (bHLH) transcription factors such as myogenin, Myf5, Mash2, and Hand1 by preventing their nuclear localization (37,38). I-mfa has also been shown to function within the Wnt/␤-catenin pathway (39,40). Specifically, it binds and inhibits the transcriptional activity of the Xenopus HMG box transcription factor XTcf3 (39) and also participates in a multiprotein complex containing Axin (40). Consistent with its molecular role as an inhibitor of both the bHLH proteins and canonical Wnt/␤-catenin pathway, inactivation of the I-mfa gene by homologous recombination in mice resulted in skeletal patterning and placental defects (38). I-mfa mutant mice on a C57B1/6 background died around E10.5 because of a placental defect associated with a reduced number of trophoblast giant cells, whereas mutant mice on a 129/Sv background survived to adulthood and showed skeletal patterning defects associated with deregulated osteogenic differentiation (38). However, the exact mechanisms by which I-mfa mediates its effects during development still remain unknown.
In the present study, we have identified and characterized a physical interaction between I-mfa and TRPC1 in vitro and in vivo. Furthermore, gain-and loss-of-expression experiments show that I-mfa functions as an inhibitor of store-operated currents by negatively modulating TRPC1 activity.

EXPERIMENTAL PROCEDURES
Cell Culture-CHO-K1 and A431 cells were obtained from ATCC and maintained in F-12 (Ham's) supplemented with 0.05% sodium bicarbonate and 10% fetal bovine serum or Dulbecco's modified Eagle's medium plus 10% fetal bovine serum, respectively.
Plasmids-F-I-mfa was made by inserting the coding region of human I-mfa downstream of the FLAG epitope in the pCMV5-FLAG vector (Kodak). Mouse myogenin in pCMV-SPORT6 was obtained from Incyte Genetics. The "a-region" of human I-mfa was fused to pMAL-c2 (New England Biolabs) to generate maltose-binding protein (MBP)-Imfa(a). Portions of the C-terminal (Asp 639 -Ser 750 , accession number Z73903) and the N-terminal cytoplasmic region of human TRPC1 (Ser 27 -Ser 306 ) were cloned in pGEX-3X (Pharmacia) to generate glutathione S-transferase (GST)-TRPC1-C or GST-TRPC1-N, respectively. F-Imfa⌬5 and TRPC1⌬ were generated by the introduction of termination codons at positions Ser 170 and Arg 630 of human I-mfa (accession number NM_005586) and human TRPC1, respectively, by site-directed mutagenesis using QuikChange (Stratagene). All other constructs were described previously (25).
Production of Polyclonal Antibodies-Mouse ␣-TRPC1-N, rabbit ␣-TRPC1-C, or rabbit ␣-TRPC1-N was mouse polyclonal against the N-terminal cytoplasmic region, rabbit polyclonal against the C-terminal cytoplasmic tail, or rabbit polyclonal antibody made against the Nterminal region of human TRPC1, respectively. Mouse or chicken ␣-Imfa was raised against the entire I-mfa molecule. All antibodies were raised against fusions of TRPC1 or I-mfa with GST. Antibody titers were initially determined by enzyme-linked immunosorbent assay using purified TRPC1 or I-mfa fusions with MBP. Rabbit ␣-TRPC1-C and chicken ␣-I-mfa were further affinity purified using the AminoLink method (Pierce). All polyclonal antibodies recognized human, mouse, and rat TRPC1 or I-mfa protein.
Production of Monoclonal Antibody (1F1)-Monoclonal antibody was produced by standard methods (41). Briefly, female Balb/c mice were immunized to purified GST-TRPC1-N by three biweekly injections (intraperitoneally and subcutaneous) of 15 g each. Serum levels of specific antibody were determined by direct enzyme-linked immunosorbent assay. Four days following the final injection, the spleen of the animal determined to have the highest specific titer was removed and then fused to Sp/2O myeloma cells in 50% polyethylene glycol-500. Following hypoxanthine-aminopterine-thymidine selection, resulting clones were screened for specific antibody production by direct enzyme-linked immunosorbent assay using a purified fusion of the N-terminal of human TRPC1 with MBP (MBP-TRPC1-N). Positive clones were subcloned twice by limiting dilution. 1F1 showed the highest titers and was further used for ascites production. 1F1 was isotyped mouse IgG 1 and its specificity to TRPC1 was additionally characterized by immunoblotting and immunofluorescence staining in transfected cells. 1F1 recognized human, mouse, and rat TRPC1.
Yeast Two-hybrid-The L40 yeast strain containing both His and LacZ reporters under the control of LexA binding sites was sequentially transformed with a bait plasmid in pLexA containing a portion of the C-terminal cytoplasmic region of human TRPC1 (Asp 639 -Ser 750 , accession number Z73903) and a human fetal kidney library (Clontech) in pGAD10 or a whole mouse embryonic library in pVP16. Four million independent clones from each library were screened with pLexA-TRPC1. Two overlapping clones of human I-mfa were obtained from the human fetal kidney library and a single mouse clone was obtained from the mouse embryonic library.
In Vitro Binding Using Recombinant Proteins-MBP or GST fusions were expressed and purified from bacteria using amylose resin or glutathione-Sepharose 4B column, respectively. GST fusions were retained on the column, whereas MBP fusions were eluted by 10 mM maltose. The concentration and integrity of purified soluble or immobilized MBP or GST fusions, respectively, were determined by Coomassie Brilliant Blue staining of 12% SDS-PAGE gel containing known amounts of purified bovine serum albumin or MBP. In a typical in vitro binding assay, 2 g of immobilized GST fusions were incubated with 5 g of soluble MBP or MBP fusion in a total volume of 1 ml of immunoprecipitation solution (1% Triton X-100 buffer containing 150 mM NaCl, 10 mM Tris-HCl (pH 7.5), 1 mM EDTA, 1 mM EGTA, 0.2 mM sodium vanadate, 0.2 mM phenylmethylsulfonyl fluoride, 0.5% Nonidet P-40, aprotinin (1 g/ml), and pepstatin). Binding was done overnight at 4°C. The beads were washed five times in immunoprecipitation buffer, resuspended in 40 l of gel loading buffer, separated by 12% SDS-PAGE, and followed by Western blotting using ␣-MBP (1:10,000) (New England Biolabs).
Transient Transfections and Immunoprecipitations-Transient transfections and immunoprecipitations were done in HEK293T cells as described previously (42).
Stable Transfections in CHO-K1 Cells-CHO-K1 cells were transfected with pSV2-neo (CHO mock cells) or F-I-mfa, pSV2-neo, and CD4 (CHO I-mfa ) using LipofectAMINE PLUS (Invitrogen) and stable transfectants were selected in 500 g/ml G418 for 2 weeks. CD4 did not affect the interaction between TRPC1 and I-mfa, as determined by transient transfections followed by co-immunoprecipitations (data not shown). Individual clones were isolated, expanded, and sorted twice for CD4 ϩ cells using a phycoerythrin-conjugated monoclonal antibody against CD4 to obtain pure populations of F-I-mfa expressing cells. Individual clones were first tested for F-I-mfa expression by immunoprecipitation followed by immunoblotting using ␣-FLAG. Clone CHO I-mfa showed the highest levels of F-I-mfa.
Electrophysiology-The conventional whole cell voltage-clamp configuration was used to measure transmembrane currents in single cells. Patch clamp recordings were obtained from single cells at room temperature using a Warner PC-505B amplifier (Warner Instrument Corp., Hamden, CT) and pClamp 8 software (Axon Instrument, Foster City, CA). Glass pipettes (plain, Fisher Scientific) with resistances of 5-8 M⍀ were prepared with a pipette puller and polisher (PP-830 and MF-830, respectively, Narishige, Tokyo, Japan). After the whole cell configuration was achieved, cell capacitance and series resistance were compensated before each recording period. From a holding potential of 0 mV, voltage steps were applied from Ϫ100 to 80 mV in 20-mV increments with 100 ms duration at 5-s intervals. I-V relationships were measured by voltage ramping from Ϫ80 to 90 mV within 100 ms. Current traces were filtered at 1 kHz and analyzed off-line with pClamp 8. Statistical analysis was employed with the SigmaStat (Chicago, IL) software. Data were reported as mean Ϯ S.E. One-way repeated measures analysis of variance followed by Student-Newman-Keul's tests were used for comparisons among sequential treatments in the same group. Student's t test was used for comparisons between groups. Differences were considered significant at p Ͻ 0.05. The pipette solution contained (in mM): 135 cesium methane sulfonate, 8 NaCl, 1 MgCl 2 , 0.3 Mg-ATP, 0.03 GTP, 10 HEPES, and 10 EGTA (pH 7.2). The external (bath) solution contained (in mM): 110 NaCl, 5 CsCl, 1 MgCl 2 , 20 CaCl 2 , 10 HEPES, and 10 glucose (pH 7.4). Permeability ratio of Ca 2ϩ to Na ϩ was calculated from the modified GHK equation as described in Ref. 43, assuming a permeability ratio of Cs ϩ to Na ϩ equals to 1.
Ratiometric Ca 2ϩ Measurements-Cells were harvested in phosphate-buffered saline containing 0.5 mM EDTA, washed with normal extracellular solution (ECS) (140 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 10 mM glucose, 0.1% bovine serum albumin, 15 mM HEPES, pH 7.4), and loaded with 2 M Indo-1/AM in ECS containing 0.05% pluronic F-127 (Molecular Probes) for 40 min at room temperature. After the 40-min incubation, cells were washed three times with a nominally Ca 2ϩ -free solution (same as ECS but 2 mM CaCl 2 was replaced by 0.5 mM EGTA, ECS-EGTA). 2 ϫ 10 6 cells were resuspended in 2 ml of ECS-EGTA. Cells in ECS-EGTA were first incubated with 1 M thapsigargin (TG) for 1 h to deplete the internal stores and then Ca 2ϩ entry was determined by CaCl 2 re-addition (6.25 mM CaCl 2 ). Ca 2ϩ entry in response to Ca 2ϩ re-addition without prior store depletion was also determined. Ratiometric measurements representing free intracellular Ca 2ϩ concentration were obtained by a PTI QuantaMaster spectrofluorometer equipped with an excitation monochromator set at 350 nm and two emission monochromators set at 405 and 485 nm.
RNAi-To inactivate human I-mfa in A431 cells, we first constructed a vector that would allow the production of a small I-mfa-specific interfering RNA driven by the polymerase-III H1-RNA gene promoter. Specifically, a fragment of H1-RNA promoter was cloned into pUB/Bsd (Invitrogen) by EcoRI-BamHI digestion to generate pUB/Bsd/H1. This vector allowed blasticidin selection for stable integration. We then cloned a 64-mer I-mfa-specific oligo (sense strand: 5Ј-GATCCCCGAG-TAAGAGCAGCAGCAAATTCAAGAGATTTGCTGCTGCTCTTACTCT-TTTTGGAAA-3Ј) into pUB/Bsd/H1 digested with BamHI-HindIII to generate a construct that would allow the production of a short hairpin I-mfa RNA in mammalian cells. A431 cells were transfected in 10-cm plates using FuGENE (Qiagen). 48 h following transfection, cells were placed in selective media containing 10 g/ml blasticidin. After 2 weeks, individual clones were picked and expanded. Clone A431 I-mfa-RNAi showed the highest I-mfa reduction and was used in the described experiments.

I-mfa Is an Interacting Partner of TRPC1-A
yeast twohybrid screen using two independent libraries, a human fetal kidney or whole mouse embryonic (E9.5) library, resulted in the identification of I-mfa as an interacting partner with TRPC1. Two overlapping clones containing the entire I-mfa cDNA including a portion of the 5Ј-untranslated region or just the a-region were obtained from the human fetal kidney library, and a partial clone containing the a-region and a portion of the common region was obtained from the whole mouse embryonic library. DNA sequence analysis revealed that all interacting clones were in-frame with the transactivation domain of GAL4. Because the smaller interacting clone contained just the a-region of I-mfa, we inferred that TRPC1 might specifically interact with the a-but not the b-or c-isoforms of I-mf.
In Vitro Interaction and Co-localization of I-mfa and TRPC1 in Transfected Cells-To test whether the a-region of I-mfa specifically interacted with the C-terminal cytoplasmic region of TRPC1 (TRPC1-C), we employed purified recombinant proteins to examine these interactions. 2 g of immobilized GST, a fusion protein with GST and the C-terminal tail of human TRPC1 (GST-TRPC1-C), or the N-terminal region of human TRPC1 (GST-TRPC1-N) were incubated with 5 g of soluble, purified MBP or MBP fused to the a-region of human I-mfa (MBP-Imfa(a)). Fig. 1A shows that MBP-Imfa(a), but not MBP (lanes 3, 5, and 7), bound to GST-TRPC1-C (lane 6) but not GST (lane 4) or GST-TRPC1-N (lane 8). These results show that the C-terminal cytoplasmic tail of TRPC1 can mediate a direct, physical interaction with the a-region of I-mfa.
To determine whether full-length TRPC1 and I-mfa could interact in mammalian cells, we transfected FLAG-tagged Imfa (F-I-mfa), myc-tagged TRPC1 (M-TRPC1), or myc-tagged TRPC3 (M-TRPC3) in HEK293T cells and tested their interactions by co-immunoprecipitation experiments. F-I-mfa co-immunoprecipitated with M-TRPC1 (Fig. 1B, lane 3) but failed to interact with M-TRPC3 (Fig. 1B, lane 2) showing specificity of I-mfa for TRPC1 but not for a homologous channel protein. A TRPC1-truncation mutant lacking the domain used in the yeast two-hybrid screen to identify I-mfa (TRPC1⌬) failed to interact with I-mfa (Fig. 1C, lane 2). Conversely, a truncation mutant of I-mfa lacking almost the entire a-region (F-I-mfa⌬5) could not associate with TRPC1 (Fig. 1C, lane 4). HA-tagged I-mfb (HA-I-mfb) was also unable to associate with TRPC1 (Fig. 1D, lane 2) supporting the finding that the interaction between TRPC1 and F-I-mfa is mediated by the a-region of I-mf. Altogether, these results demonstrate not only the specificity of the interaction but also the lack of additional interaction domains in the two proteins.
Myogenin Competes with TRPC1 for Binding to I-mfa-Because the a-region of I-mfa was independently found to mediate an interaction with bHLH proteins (37), we tested whether the bHLH protein myogenin could compete with TRPC1 for binding to I-mfa. Fig. 1E shows that increasing amounts of transfected myogenin did inhibit the TRPC1-Imfa association in a concentration-dependent fashion. These results suggest that I-mfa may not bind to TRPC1 and myogenin at the same time, or to other bHLH-binding partners, thereby ruling out the possibility of a tripartite formation.
Endogenous I-mfa and TRPC1 Interact in Native Tissues-To determine whether endogenous TRPC1 associated with I-mfa, we examined whether they co-existed in a native complex in adult rat tissues. Fig. 2A shows that a mouse polyclonal antibody made against TRPC1 or I-mfa (mouse polyclonal ␣-TRPC1-N or ␣-I-mfa, respectively) immunoprecipitated native TRPC1 from rat liver (lanes 7 and 8), Nalm-6 cells (a human pre-B cell line) (lanes 11 and 12), and rat kidney (lanes 14 and 15) lysates. Endogenous TRPC1 was detected by a rabbit polyclonal antibody raised against the C-terminal cytoplasmic tail of human TRPC1 (␣-TRPC1-C) ( Fig. 2A, lanes 5-15) but not by rabbit preimmune serum ( Fig. 2A, lanes 1-4). The possibility that proteins recognized in lanes 7, 8, 11, 12, 14, and 15 represented mouse antibodies non-specifically recognized by rabbit ␣-TRPC1-C was ruled out by probing the immunocomplexes captured by preimmune, ␣-TRPC1-N, or ␣-Imfa antiserum from rat liver lysates with sheep anti-mouse IgG coupled to horseradish peroxidase. Fig. 2B shows that anti-mouse IgG identified just the heavy and light chains of the mouse antibodies used in immunoprecipitations but not any other protein with a molecular weight similar to transfected TRPC1. The specificity of the ␣-TRPC1-C was further confirmed by recognizing a single protein in native rat tissue lysates having a molecular size similar to transfected TRPC1 (Fig. 2C, upper panel). Transfected F-TRPC1 was used as a positive size control in Fig. 2A (lanes 5 and 9) and  9) are likely to represent nonspecific binding because of long exposure to identify endogenous TRPC1 in various tissues (Fig. 2C, lanes 2-5). It should be noted that endogenous TRPC1 is expressed at very low levels in the tissues tested and therefore long exposures were required for detection, in our hands. The relative expression of endogenous I-mfa in various rat tissues is shown in Fig. 2C (lower panel). The highest I-mfa expression was seen in lung followed by kidney, liver, and heart. This pattern of expression partially mirrored the expression of TRPC1, which showed the highest levels in liver and lung followed by heart and kidney. Overall, these results provided evidence for a physical association between TRPC1 and I-mfa in adult rat tissues. The interaction was also supported by the co-expression pattern of these proteins.
Ectopic Expression of I-mfa Inhibits Native Store-operated Currents in CHO-K1 Cells-To determine whether I-mfa was sufficient to regulate the function of TRPC1, we stably transfected I-mfa into CHO-K1 cells. We chose these cells because they express TRPC1 (Fig. 3A, lane 2) and lack endogenous I-mfa (Fig. 3B, lane 1). More importantly in CHO-K1 cells, TRPC1 knock-down by RNAi resulted in the elimination of store-operated currents suggesting that TRPC1 is an essential component of the native store-operated conductance in these cells (18). Mock-transfected (CHO mock ) or cells stably expressing I-mfa (CHO I-mfa ) were challenged with 1 M TG, an agent traditionally used to irreversibly deplete intracellular Ca 2ϩ stores, resulting in the activation of store-operated conductance without the interference of post-receptor activated events.
Whole cell currents were sequentially measured in CHO mock or CHO I-mfa cells using voltage step and ramp protocols before and after the addition of TG. Under basal conditions (Pre-TG), small currents were recorded in both cell types (Fig. 3, C, D, G, and H).
Depletion of the internal stores with 1 M TG enhanced the whole cell current density in CHO mock cells from Ϫ4.66 Ϯ 0.61 pA/pF to Ϫ12.69 Ϯ 2.1 pA/pF (mean Ϯ S.E.) at Ϫ100 mV (Fig. 3I). Similar results were obtained in a green fluorescent protein-expressing clone of CHO-K1 cells (data not shown) and in the parental CHO-K1 cells dialyzed with control mouse IgG 1 (see below). Examination of the I-V relation curve derived from the ramp protocol showed that TG induced a slightly inward and outward rectifying current, which reversed at positive potentials (Fig. 3G) characteristic of higher Ca 2ϩ over Na ϩ selectivity (P Ca :P Na ϭ 12.6:1). In contrast to CHO mock cells, CHO I-mfa cells did not show a significant response to TG (from Ϫ5.4 Ϯ 1.09 to Ϫ6.7 Ϯ 1.15 pA/pF at Ϫ100 mV) (Fig. 3I). Because La 3ϩ has been shown to inhibit store-operated channels at nano-to micromolar concentrations (44 -46), depending on the cell type, we examined the TG-evoked responses in the presence of 200 nM or 20 M La 3ϩ . . Cells were lysed and F-I-mfa was immunoprecipitated with ␣-FLAG. Immunocomplexes (two middle panels) and lysates (upper and lower panels) prior to immunoprecipitation were immunoblotted with ␣-mgn or ␣-myc, as indicated.
TG-induced, but not background currents were completely blocked by 20 M La 3ϩ in CHO mock cells (Fig. 3I), whereas they were only partially inhibited by 200 nM La 3ϩ (data not shown). In summary, these data showed that TG was able to induce native currents in CHO-K1 cells but not in cells expressing I-mfa, suggesting that I-mfa has a suppressive effect on native store-operated currents.
To determine whether the effect of I-mfa was direct and did not result from changes in transcription affecting the expression of proteins that may regulate TRPC1 activity, and to also determine the specificity of I-mfa on TG-induced currents, we tested whether pipette-perfused recombinant MBP-Imfa(a), MBP-Imfb(b), or MBP-Imfc at 200 ng/ml could affect TG-induced currents in CHO-K1 cells. To allow adequate time for the recombinant proteins to diffuse into the cytoplasm, we initiated recording 5 min following break-in. Consistent with previous results, only MBP-Imfa(a) suppressed these currents (Fig. 3J).
To determine whether the I-mfa-mediated reduction in storeoperated currents could also result in a reduction in storeoperated Ca 2ϩ entry, we measured TG-induced Ca 2ϩ influx in CHO mock and CHO I-mfa cells by ratiometric measurements of intracellular Ca 2ϩ concentration in cell populations using indo-1/AM. Ca 2ϩ entry in response to store depletion by 1 M TG was determined in the presence of a nominally Ca 2ϩ -free extracellular solution (0.5 mM EGTA) using a Ca 2ϩ re-addition protocol (Fig. 3K). Ca 2ϩ entry was significantly reduced in CHO I-mfa cells compared with CHO mock cells (Fig. 3K, n ϭ 4). Ca 2ϩ imaging experiments also verified the inhibitory effect of I-mfa on TG-induced Ca 2ϩ entry in individual cells (data not shown). Thus, ratiometric results supported previous current measurements providing evidence for an inhibitory role of transfected I-mfa in native store-operated Ca 2ϩ entry in CHO-K1 cells.
Endogenous TRPC1 Functionally Contributes to Store-operated Currents in CHO-K1 Cells-TRPC1 was shown to be an essential component of store-operated currents in CHO-K1 cells (18). To confirm these findings using an independent approach, we measured store-operated currents in the presence of 10 or 100 ng/ml 1F1, a mouse TRPC1-specific monoclonal antibody (47), or mouse IgG 1 in the intrapipette solution. The specificity of 1F1 for TRPC1 versus TRPC2, -3, and -4 was additionally tested by immunoblotting or flow cytometry in transfected HEK293T or permeabilized CHO-K1 cells, respectively (data not shown). Fig. 4 shows that TG failed to induce store-operated currents in 4 of 4 CHO-K1 cells dialyzed with 1F1 at 100 ng/ml (B, D, and E). The inclusion of 100 ng/ml mouse IgG 1 or 10 ng/ml 1F1 to the intrapipette solution did not affect store-operated currents (n ϭ 4) (Fig. 4, A, C, and E). These data showed that TRPC1 contributes greatly to native store-operated currents in these cells, thus confirming previous data (18) and validating the use of 1F1 as a TRPC1-specific neutralizing antibody.

Inactivation of Endogenous I-mfa Increases Store-operated
Currents in A431 Cells-Whereas transfected I-mfa suppressed native currents (Fig. 3), we wished to determine whether endogenous I-mfa had a similar effect on native store-activated conductance. Therefore, we knocked down I-mfa expression by RNAi using a short hairpin I-mfa-specific RNA (48) in A431 cells, a human epidermoid carcinoma cell line. The interfering sequence was designed to the common region of human I-mf. We chose the human A431 cell line because they showed the highest levels of I-mfa expression compared with all other cell lines tested, including Nalm-6 (data not shown). They also express detectable levels of TRPC1 (Fig. 5A, lanes 3 and 4). Immunoblotting of total cell lysates from A431 cells and cells stably transfected with our I-mfa-specific RNAi vector (A431 I-mfa-RNAi ) revealed that endogenous I-mfa was reduced by ϳ90% in the transfected cells (A431 I-mfa-RNAi ) (Fig. 5B). To confirm these results with two I-mfa-specific antibodies, we immunoprecipitated I-mfa using mouse ␣-I-mfa and probed the immunocomplexes with the chicken ␣-I-mfa (Fig. 5C).
Next, we compared whole cell currents in parental cells (A431) and cells lacking I-mfa (A431 I-mfa-RNAi ) using the same protocols, and extracellular bathing and intrapipette solutions as the ones used for CHO-K1 cells (Fig. 3). Under basal conditions only small currents were detected in both cell types (Fig.  5, D and E). Depletion of internal Ca 2ϩ pools by TG increased the membrane conductance in both cell types, indicating the activation of store-operated channels (Fig. 5, F and G). Notably, store-operated currents were larger in A431 I-mfa-RNAi cells compared with A431 cells at all tested potentials (Fig. 5, H and I). Specifically, at a holding potential of Ϫ100 mV, TG increased the current density from Ϫ3.98 Ϯ 0.4 pA/pF to Ϫ5.98 Ϯ 0.8 pA/pF in A431, whereas from Ϫ3.42 Ϯ 0.45 pA/pF to Ϫ9.28 Ϯ 1.25 pA/pF in A431 I-mfa-RNAi cells (Fig. 5J). Again, TG-induced responses but not background currents were completely inhibited by 20 M La 3ϩ (Fig. 5J), and partially inhibited by 200 nM La 3ϩ (data not shown). In our hands, 2 of 8 A431 and 2 of 10 of A431 I-mfa-RNAi cells were unresponsive to TG in the presence of EGTA in the pipette. We considered A431 responsive cells to be those in which TG enhanced currents by more than 10% above basal levels.
The I-V relation curve showed a significant shift in the reversal potential from 2.5 Ϯ 6.8 to 18.6 Ϯ 7.3 mV (n ϭ 6, paired t test, p Ͻ 0.01) (Fig. 5I) of TG-induced currents indicating the activation of channels with slightly higher selectivity for Ca 2ϩ over Na ϩ in A431 I-mfa-RNAi cells but not in A431 cells (Fig. 5H). Reversal potentials for background currents in A431 I-mfa-RNAi versus A431 and CHO mock versus CHO I-mfa cells were not significantly different (Fig. 5, I and H, and Fig. 3, G  and H, respectively). Overall, knock-down of endogenous I-mfa from A431 cells resulted in the enhancement of TG-induced currents probably because of the opening of relatively Ca 2ϩ selective store-operated channels. Taking into account that I-mfa is not a channel protein itself, its effects on TG-induced currents are of considerable magnitude suggesting that it may have a key regulatory role in native channel activity.
TRPC1 Functionally Contributes to Native Store-operated Conductance in A431 I-mfa-RNAi but Not in A431 Cells-Although down-regulation of I-mfa enhanced store-activated currents in A431 cells, the contribution of TRPC1 to this enhancement was unknown. Fig. 6 shows that dialysis of the cytoplasm of A431 cells with 1F1 did not result in a significant change in the whole cell store-operated currents compared with non-dialyzed A431 cells (Figs. 5J and 6C) suggesting that TRPC1 did not significantly contribute to TG-induced currents in these cells. In contrast, TG-induced currents were reduced by 85% in A431 I-mfa-RNAi cells dialyzed with 1F1 as compared with mouse IgG 1 (Fig. 6, D-G). The increase in current density at Ϫ100 mV in response to TG was reduced from Ϫ5.9 Ϯ 0.9 pA/pF in cells dialyzed with 100 ng/ml mouse IgG 1 to Ϫ0.9 Ϯ 0.4 pA/pF in cells dialyzed with 100 ng/ml 1F1 (Fig. 6H). These results indicated that TRPC1 was required for the function of native store-operated channels in A431 I-mfa-RNAi cells supporting the idea that removal of I-mfa from A431 cells enhanced the function of TRPC1 as a store-operated channel.

Recombinant MBP-Imfa(a) Rescues the Effect of I-mfa Knockdown in A431
Cells-To confirm that the effect of I-mfa inactivation on store-operated conductance in A431 cells was specific to I-mfa knock-down, we pipette-perfused recombinant MBP-Imfa(a) in A431 I-mfa-RNAi cells and tested whether it could reverse the increase in store-operated currents induced by Imfa knock-down. Fig. 6H shows that MBP-I-mfa(a) significantly reduced the increase in current density in response to TG from Ϫ5.9 Ϯ 0.9 pA/pF to Ϫ1.9 Ϯ 0.38 pA/pF in A431 I-mfa-RNA cells. DISCUSSION We provide three lines of evidence supporting a physical and functional interaction between I-mfa and TRPC1. First, TRPC1 and I-mfa associate in vitro and in vivo. Second, overexpression of I-mfa suppresses store-operated conductance, whereas down-regulation of I-mfa increases store-operated currents. Third, TRPC1 functionally contributes to store-operated conductance in cells lacking I-mfa but not in cells expressing I-mfa. These data provide evidence for a new function of I-mfa as a negative regulator of store-operated Ca 2ϩ entry. Using the yeast two-hybrid assay as an unbiased screen, we identified I-mfa as an interacting protein with TRPC1. The interaction was further verified by in vitro and in vivo coimmunoprecipitation experiments and in vitro binding assays using recombinant proteins. Furthermore, the interaction sites were mapped to the C terminus of both peptides. The a-region of I-mf has been shown to mediate interactions with a divergent group of proteins such as a subset of bHLH (37,38), XTcf3 (39), and axin (40). In vitro data showed that the I-mfa-TRPC1 interaction could be competed away by I-mfa interacting proteins, such as myogenin. We speculate that I-mfa exists in two separate but dynamically regulated pools; one pool of I-mfa associated with TRPC1 and another pool of I-mfa associated with bHLH/LEF1/TCF3 transcription factors. Although the functional and physiological significance of this competition is unknown at present, it provides an in vitro example of the dynamic nature of the interaction between TRPC1 and I-mfa and raises the intriguing possibility that TRPC1 function could be regulated by proteins that interact with I-mfa. Thus, the I-mfa-mediated inhibition of TRPC1 could be reversed under conditions limiting the amount of I-mfa available for interaction with TRPC1. Conversely, TRPC1-interacting proteins may displace inhibitory I-mfa from TRPC1. It is tempting to speculate that the latter may underlie a gating mechanism for storeoperated channels via TRPC1.
Several independent studies, including our own, show that endogenous TRPC1 functionally contributes to native storeoperated conductance in a variety of cells including hamster fibroblast CHO-K1, human submandibular gland HSG cells, and DT40 B lymphocytes (17,18,49). However, it cannot be concluded from these studies that store-operated currents are carried through TRPC1 and evidence from heterologous expression studies argues against this idea (19,32). Our data show that removal of I-mfa from A431 cells resulted in the activation of slightly Ca 2ϩ selective store-operated currents. In addition, TG activated a conductance with similar selectivity in CHO-K1 cells. The TG-induced currents are unlikely to be carried by TRPC1 itself, as TRPC1 forms a non-selective cation channel when successfully expressed alone in CHO-K1 cells (19). Furthermore, knock-down of either TRPC1 (18) or TRPC2 (50) in CHO-K1 cells resulted in more than 90% reduction in Ca 2ϩ entry in response to store depletion implying a functional interaction between TRPC1 and TRPC2. Thus, we favor the idea that native TRPC1 is an essential component of store- Step currents recorded 50 ms after establishing the initiation of the Ϫ100 mV voltage pulse were used in the analysis. *, p Ͻ 0.05, comparison among pre-, post-TG, and TG ϩ La 3ϩ treatments in the same group; #, p Ͻ 0.05, comparison of post-TG treatments between groups. n indicates the number of cells analyzed in each group. operated conductance, but that it may require other channel proteins to form a store-operated channel. Identification of these interacting proteins will be instrumental in the determination of the molecular entity of the elusive store-operated Ca 2ϩ channel complex.
In a previous study (18), TG-induced current densities were 4 -5 pA/pF in CHO-K1 cells, whereas in our case, TG-induced current densities were 8 pA/pF in the same cells. There are several explanations for these disparities. First, in the previous study, 200 nM TG was used to activate store depletion-sensitive currents instead of 1 M used in our experiments. It is unknown whether 200 nM TG had completely depleted internal Ca 2ϩ stores in these experiments. It has been reported that the magnitude of Ca 2ϩ entry through store-operated Ca 2ϩ channels is closely correlated with the extent of store depletion (51). The Ca 2ϩ concentration in the bathing solution was 2 mM in the experiments of Vaca and Sampieri (18), whereas 20 mM in ours. Because store-activated channel is permeable to Ca 2ϩ , the extracellular Ca 2ϩ concentration may have an impact on the amplitude of the whole cell currents. In addition, the Mg 2ϩ concentration in their pipette solution was dramatically different from ours. Specifically, the total intrapipette Mg 2ϩ concentration was 5 mM (2 mM MgCl 2 , 2 mM Mg-ATP, and 1 mM Mg-GTP) in the previous study, whereas it was 1.33 mM (1 mM MgCl 2 , 0.3 Mg-ATP, and 0.03 mM GTP) in our case. The effect of intracellular Mg 2ϩ on currents carried by CRAC, MIC, or TRP channels is summarized in a recent review (9). Despite differences in current amplitudes, two independent approaches, RNAi (18) and the use of a neutralizing TRPC1specific antibody, have demonstrated a necessary role of TRPC1 in native store-operated conductance in CHO-K1 cells. Because I-mfa expression suppressed store-activated currents as completely as 1F1 or the double stranded RNA used to inactivate TRPC1 in another study (18), we propose that I-mfa inhibited store-activated currents in CHO-K1 cells by a TRPC1dependent mechanism. This was further supported in A431 cells in which 1F1 reversed the increase in store-operated currents following I-mfa inactivation.
We found that knock-down of I-mfa by ϳ90% resulted in a 3-fold increase in the TG-induced currents in A431 cells. This increase was reduced by 85% using a TRPC1 neutralizing antibody or by 68% with recombinant I-mfa, both delivered via pipette-perfusion. These results support a direct functional interaction between I-mfa and TRPC1 and prompt us to conclude that native I-mfa negatively regulates the activity of TRPC1 as a store-activated channel in these cells. Store-operated channels in A431 have been studied extensively (52)(53)(54)(55)(56)(57). In a recent detailed study (58), it was shown that store-operated conductance in A431 cells is composed of two major currents, the Ca 2ϩ release-activated current (I CRAC ) and store- Step currents were recorded 50 ms after the delivery of the Ϫ100 mV voltage pulse. *, p Ͻ 0.05, comparison of 1F1 or MBP-Imfa(a) group to IgG 1 group. n indicates the number of cells analyzed in each group. operated current (I SOC ). Some of the differences between these two currents in relation to our studies were that I CRAC showed high Ca 2ϩ selectivity and lack of significant outward currents at positive potentials, whereas I SOC was less selective to Ca 2ϩ and showed outward currents. Under our experimental conditions, it is likely that both of these currents contribute to the TG-induced response. However, based on the shape of our I-V curve, we propose that I SOC might be the predominant contributor. Because 1F1 had a minimal effect on the TG-induced current in native cells, we speculate that TRPC1 may not significantly contribute to I SOC in these cells. In contrast, it may contribute up to 85% (Fig. 6H) of the same current in cells lacking I-mfa. A detailed analysis to determine the specific suppressive effect of I-mfa on I CRAC and I SOC will provide insight into the mechanism by which TRPC1 contributes to I CRAC and/or I SOC .
The interaction between TRPC1 and I-mfa was demonstrated in adult rat kidney and liver lysates, whereas significant co-expression of TRPC1 and I-mfa was seen in lung and heart. A widespread distribution of TRPC1 (for review see, Ref. 14) and I-mfa mRNAs (38) in adult tissues has also been reported. Based on our biochemical and functional studies, it is tempting to speculate that the suppression of store dependence of TRPC1 by I-mfa may have physiological consequences. However, it is also equally possible that one of the physiological roles of I-mfa may be the modulation of the ability of TRPC1 to function as a receptor-operated channel through the association with TRPC4 or TRPC5. Because, TRPC1 is likely to be co-expressed with TRPC4 and/or TRPC5 in native tissues, a more intriguing possibility is that I-mfa may function as a mode switch between store-operated and receptor-operated modes of TRPC1 activation. Experiments are underway to determine whether I-mfa can modulate TRPC1/TRPC4 function in transfected cells.
Overall, our study provides both biochemical and functional data supporting a role for I-mfa on Ca 2ϩ signaling via a protein-protein interaction with TRPC1. Given the role of I-mfa in the regulation of the activity of a diverse group of transcription factors (37-39), our results provide evidence for a novel mechanism by which a transcriptional regulator is involved in Ca 2ϩ signaling by direct interaction with a channel. Because of the latter, the interaction between TRPC1 and I-mfa could function as a ubiquitous switch influencing many developmental and physiological processes.