MxA, a Member of the Dynamin Superfamily, Interacts with the Ankyrin-like Repeat Domain of TRPC*

Mammalian transient receptor potential canonical channels have been proposed as the molecular entities associated with calcium entry activity in nonexcitable cells. Amino acid sequence analyses of TRPCs revealed the presence of ankyrin-like repeat domains, one of the most common protein-protein interaction motifs. Using a yeast two-hybrid interaction assay, we found that the second ankyrin-like repeat domain of TRPC6 interacted with MxA, a member of the dynamin superfamily. Using a GST pull-down and co-immunoprecipitation assay, we showed that MxA interacted with TRPC1, -3, -4, -5, -6, and -7. Overexpression of MxA in HEK293T cells slightly increased endogenous calcium entry subsequent to stimulation of Gq protein-coupled receptors or store depletion by thapsigargin. Co-expression of MxA with TRPC6 enhanced agonist-induced or OAG-induced calcium entry activity. GTP binding-defective MxA mutants had only a minor potentiating effect on OAG-induced TRPC6 activity. However, a MxA mutant that could bind GTP but that lacked GTPase activity produced the same effect as MxA on OAG-induced TRPC6 activity. These results indicated that MxA interacted specifically with the second ankyrin-like repeat domain of TRPCs and suggested that monomeric MxA regulated the activity of TRPC6 by a mechanism requiring GTP binding. Additional results showed that an increase in the endogenous expression of MxA, induced by a treatment with interferon α, regulated the activity of TRPC6. The study clearly identified MxA as a new regulatory protein involved in Ca2+ signaling.

[Ca 2ϩ ] i regulates many cellular functions, including cell growth, differentiation, contraction, and secretion. In most nonexcitable cells, [Ca 2ϩ ] i elevations are initiated in response to hormones and other stimuli that activate phospholipase C ␤ or ␥. This activation causes the hydrolysis of phosphatidylinositol 4,5-bisphosphate into two second messengers, inositol 1,4,5-trisphosphate (IP 3 ) 1 and diacylglycerol. IP 3 activates its receptor/channel on the endoplasmic reticulum and causes the first phase of [Ca 2ϩ ] i elevation by releasing Ca 2ϩ from the intracellular pool. The second phase of [Ca 2ϩ ] i elevation involves Ca 2ϩ entry from the extracellular compartment. This second phase allows the maintenance of [Ca 2ϩ ] i at a higher than basal level, which is essential for maintaining cell functions (1,2).
Mammalian members of the TRPC (transient receptor potential canonical) subfamily (TRPC1 to -7) are calcium-permeable cation channels involved in the mechanism of [Ca 2ϩ ] i elevation in cells stimulated with G q -coupled receptors or tyrosine kinase receptor agonists (3,4). Functional TRPC complexes are presumed to be homo-as well as heterotetramers, and each TRPC subunit contains typical domains that probably interact with other proteins. There are growing efforts to identify proteins that interact with TRPCs and to clarify the role of these interactions in the regulation of the Ca 2ϩ entry. The IP 3 receptor, calmodulin, HOMER, INAD, NHERF, the a-isoform of the inhibitor of the myogenic family (I-mfa), and stathmin 2 (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17) all interact with TRPCs. A typical TRPC subunit is composed of six membrane-spanning domains and a hydrophobic segment between the fifth and sixth transmembrane domains that forms the putative pore region loop. Cytosolic N terminus and C terminus tails flank the six transmembrane domains. Each TRPC subunit possesses two coiled-coil regions (one in the N terminus and the other in the C terminus) and a hyperconserved region composed of three to four ankyrin-like repeat domains (ARDs) in the N terminus (18).
ARDs are found in more than 4,000 different proteins, making one of the most common motifs involved in protein-protein interactions. The 33 amino acids of ARDs form two ␣-helices linked by a ␤-hairpin/loop (19,20). When ARDs are involved in protein-protein interactions, most of the binding occurs at the ␤-hairpin/loop region (21), but certain complexes can also form at the surface of the inner ␣-helices (22). The ARDs of proteins like GA binding protein, p53, CDK6, and Notch (23) play a role in the regulation of the activity of these proteins.
To determine which proteins can interact specifically with the ARD of TRPC6, we performed a yeast two-hybrid screening of a cDNA library from human fetal brain. MxA, an interferoninduced 76-kDa GTPase that inhibits the multiplication of several RNA viruses, was one of the proteins that interacted with the ARD. In the study presented here, we showed that the C terminus of MxA bound to the second ankyrin-like repeat of TRPC6. In vitro and in vivo assays showed that MxA also interacted with TRPC1, -3, -4, -5, and -7 and that the overexpression of the wild-type form of MxA enhanced the channel activity of TRPC6.
To obtain the full-length coding region of MxA (GenBank TM accession number M30817), overlapping partial cDNA fragments were amplified from HEK293 total RNA by reverse transcription-PCR, subcloned into pCR2.1 (Invitrogen), and sequenced to confirm the identity of the inserts. MxA cDNA (nucleotides Ϫ84 to 2441) was assembled and subcloned into pcDNA-3.1 using a standard molecular biology approach. Mutations were introduced into the MxA cDNA by PCR-based sitedirected mutagenesis using Expand high fidelity polymerase and the following corresponding sense primers: K83A, 5Ј-GACCA-GAGCTCGGGCGCTAGCTCC-3Ј; T103A, 5Ј-AGCGGGATCGTGGCCA-GATGCCCGCTG-3Ј; L612K, 5Ј-ACGTACGGCCAGCAGAAACAGAAG-3Ј, and their complementary antisense primers. The resulting PCR products were sequenced to confirm the mutations and the integrity of the amplified fragments. The resulting PCR products and pcDNA3.1-MxA were digested with appropriate restriction endonucleases and purified, and suitable fragments were ligated together. The same strategy was employed to introduce the FLAG epitope (DYKD-DDDK) at the N terminus of MxA.
To determine the IFN-␣-induced MxA expression in HEK293 stably expressing TRPC6, cell monolayers were treated with 1,000 units/ml IFN-␣ for 16 h. RNA was isolated by using TRIzol according to the manufacturer's directions. One g of total RNA was reverse transcribed in a total volume of 30 l, with mouse leukemia virus reverse transcriptase according to the manufacturer's instructions by using 2.5 g/l hexamer as primer. This reverse cDNA was then used as template for the subsequent PCR. Primers used were as follows: for MxA, 5Ј-AACAGCTCTGTGATACCATTTAACTTGTTG-3Ј and 5Ј-TTCCTCCAG-CAGATCCCTGAAATATGGGTG-3Ј; for ␤-actin, 5Ј-TCAAGATCATT-GCTCCTCCTGAGC-3Ј and 5Ј-TACTCCTGCTTGCTGATCCACATC-3Ј. PCRs were carried out for the indicated number of cycles by using an annealing temperature of 53°C and a 120-s elongation at 72°C. Products were electrophoresed on 1.2% agarose gels and visualized by ethidium bromide staining.
Cell Culture and Transfection-HEK293T cells and HEK293 stably expressing TRPC6 cells were maintained at subconfluence in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, penicillin (50 units/ml), and streptomycin (50 g/ml) at 37°C in a humidified atmosphere containing 5% CO 2 . Cells were transiently transfected using Lipofectamine 2000 transfection reagent. Briefly, 6-well plates were treated with poly-L-lysine for 30 min, rinsed with phosphate-buffered saline, and dried. Plasmid DNA (1 g) diluted in 250 l of Opti-MEM I was added to each well before adding 2 l of Lipofectamine 2000 diluted in 250 l of Opti-MEM I. The mixture was incubated for 20 min at room temperature. HEK293T cells (7.5 ϫ 10 5 ) diluted in 1.5 ml of culture medium without antibiotic were then added to the DNA-Lipofectamine 2000 complex and incubated for 16 h at 37°C in a humidified atmosphere containing 5% CO 2 . The medium was then replaced by complete culture medium, and the cells were incubated for a further 24 h. For [Ca 2ϩ ] i measurements, the content of a single well was trypsinized and plated on four poly-L-lysine-treated coverslips. To determine the transfection efficacy, HEK293T cells were transfected with 1 g of cDNA coding for the M5 muscarinic receptor, and the proportion of cells showing a CCh-induced Ca 2ϩ mobilization was evaluated 48 h after transfection. Typically, the transfection efficacy varied between 55 and 65%.
Measurement of [Ca 2ϩ ] i -We used the method described by Zhu et al. (28) to measure [Ca 2ϩ ] i . Briefly, cells attached to coverslips were washed twice with HBSS (120 mM NaCl, 5.3 mM KCl, 0.8 mM MgSO 4 , 10 mM glucose, 20 mM Hepes, pH 7.4, 1.8 mM CaCl 2 ) and loaded with fura-2/AM (0.05 M in HBSS) for 20 min at room temperature in the dark. After washing and a 30-min de-esterification step in fresh HBSS at room temperature, the coverslips were inserted into a circular openbottom chamber and placed on the stage of a Zeiss Axiovert microscope fitted with an Attofluor Digital Imaging and Photometry System (Attofluor Inc., Rockville, MD). Isolated fura-2-loaded cells were selected and the [Ca 2ϩ ] i in these cells was measured by fluorescence video microscopy at room temperature using alternating excitation wavelengths of 334 and 380 nm and monitoring emitted fluorescence at 520 nm. Free [Ca 2ϩ ] i was calculated from the 334/380 fluorescence ratio according to the method of Grynkiewicz et al. (29). All reagents were diluted to their final concentration in HBSS and applied to the cells by surface perfusion. For Ba 2ϩ entry measurements, the composition of HBSS was modified (120 mM NaCl, 5.3 mM KCl, 0.8 mM MgCl 2 , 10 mM glucose, 20 mM Hepes, pH 7.4, 1.0 mM BaCl 2 ). To select transfected cells in the OAG-induced Ca 2ϩ entry experiments, the medium was replaced by Ca 2ϩ -free HBSS (plus 0.5 mM EGTA) 3 min after OAG stimulation, and 100 nM AngII was added to the medium. Only cells that showed AngIIinduced Ca 2ϩ release were considered as transfected. The results were represented as average data obtained from 8 -12 coverslips, each containing 20 -30 transfected cells. The S.E. bars were usually smaller than the size of the symbols.
Immunoprecipitation Assay-Transfected cells were rinsed twice with phosphate-buffered saline without Ca 2ϩ /Mg 2ϩ (137 mM NaCl, 3.5 mM KCl, 10 mM sodium phosphate buffer, pH 7.4) and lysed with 600 l of lysis buffer (150 mM NaCl, 1% Triton X-100, 0.5% deoxycholate, 5 mM EDTA, 1 g/ml soybean trypsin inhibitor, 0.5 g/ml leupeptin, 100 M phenylmethylsulfonyl fluoride, and 20 mM Tris-HCl, pH 8.0). The cells were incubated for 30 min at 4°C with gentle agitation followed by 10 passages through a 20-gauge needle and five passages through a 25gauge needle. The solubilized material was cleared by centrifugation for 15 min at 13,000 rpm at 4°C. The supernatant was mixed with 2 l of anti-HA antibody and 100 l of a 50% slurry of protein A-Sepharose CL-4B and incubated overnight at 4°C. Before use, the protein A-Sepharose CL-4B was precoated for 1 h at 4°C, with 0.1% bovine serum albumin in lysis buffer. Samples were centrifuged at room temperature for 2 min at 5,000 rpm and then washed three times with 1 ml of ice-cold lysis buffer. Immunoprecipitated proteins were dissolved in 50 l of 2ϫ Laemmli buffer and incubated for 30 min at 60°C before separation on a SDS-polyacrylamide gel and transfer to nitrocellulose for immunoblotting.
Immunoblots-For immunoblots, cell lysates, and immunoprecipitated proteins were separated by SDS-PAGE and transferred to a 0.2-m nitrocellulose membrane (Bio-Rad) in 150 mM glycine, 20 mM Tris-base, and 20% methanol (350 mA, 3 h, 4°C). The blots were stained with Ponceau S to visualize marker proteins, destained with TBST (20 mM Tris-HCl, pH 7.5, 137 mM NaCl, 0.3% Tween 20), and blocked for either 3 h at room temperature or overnight at 4°C in TBST containing 5% (w/v) nonfat skim milk. The blots were then incubated for 3 h at room temperature or overnight at 4°C with rabbit anti-HA (dilution 1:1,000), mouse anti-FLAG M2 (dilution 1:1,000), or mouse anti-c-Myc antibodies (dilution 1:1,000). After three washes with TBST, the blots were incubated with a peroxidase-conjugated donkey anti-rabbit-IgG (1:20,000) or peroxidase-conjugated sheep anti-mouse IgG antibody (1: 10,000) for 2 h at room temperature in TBST. The blots were washed three times with TBST, and the immune complexes were visualized with the ECL Plus detection system.
GST Fusion Proteins and GST Pull-down Assays-Amino acids 572-662 of MxA were cloned in frame into the pGEX-4T-1 plasmid to express GST fusion proteins in Escherichia coli BL21. Expression of GST fusion proteins was induced with 0.2 mM isopropyl-D-thiogalactoside for 2 h at 30°C. The cells were subsequently collected by centrifugation at 2,500 ϫ g for 15 min, sonicated in lysis buffer (20 mM Tris-HCl, pH 7.5, 1.0% Triton X-100, 100 mM NaCl, 5 mM EDTA, 1 mM dithiothreitol, 1 g/ml soybean trypsin inhibitor, 100 M phenylmethylsulfonyl fluoride) on ice, and clarified by centrifugation for 30 min at 2,500 ϫ g at 4°C. The clarified lysate was incubated with glutathione-Sepharose beads for 1 h at room temperature and then washed (five times) with ice-cold lysis buffer.
Yeast Two-hybrid Screening-cDNA encoding the amino acids 3-259 of TRPC6 was used as bait and cloned in-frame with the GAL4 DNAbinding domain in pGBKT7-BD and transformed into the yeast strain AH109. AH109 expressing the bait was then mated with the Y187 yeast strain pretransformed with a cDNA library from human fetal brain cloned into the pACT2-AD vector according to the manufacturer's instructions. Based on the mating efficiency and the titer of the library, 6.4 ϫ 10 6 clones were screened. Mated yeast cells were first grown on low stringency selection plates (ϪLeu, ϪTrp, ϪHis) and then on high stringency selection plates (ϪLeu, ϪTrp, ϪHis, ϪAde, ϩX-␣-gal). Colonies were tested for ␤-galactosidase activity and positive colonies were further segregated on (ϪLeu, ϪTrp, ϩX-␣-gal) media. pACT2-AD plasmids containing the library inserts from positive colonies were isolated and transformed into the DH5␣ bacterial strain. The selected plasmids were then sequenced and analyzed with the Blast alignment tool from NCBI (30). To eliminate false positive clones, cDNAs in frame with the GAL4-activating domain were used to transform the AH109 yeast strain that was mated with the Y187 yeast strain containing either a cDNA encoding amino acids 3-96 of TRPC6 (which is missing the ARD of TRPC6) cloned into pGBKT7-BD, or a cDNA encoding amino acids 66 -230 of human lamin C cloned into pGBKT7-BD, or the plasmid pGBKT7-BD alone. Mated yeast cells were plated on (ϪLeu, ϪTrp, ϪHis, ϪAde ϩX-␣-gal) and (ϪLeu, ϪTrp) media. Only those colonies that grew only on (ϪLeu, ϪTrp) medium were considered as expressing a protein that could specifically interact with the ARD of TRPC6.
Yeast Protein Extraction-Five ml of selective medium (ϪLeu/ϪTrp) was inoculated with a single yeast colony, vigorously mixed, and incubated overnight at 30°C on a rotating plate at 250 rpm. An aliquot (1.5 ml) of overnight culture was centrifuged for 5 min at 13,000 ϫ g at 4°C. The pellet was dissolved in 1 ml of 0.25 N NaOH containing 1% ␤-mercaptoethanol, agitated well, and incubated for 10 min on ice. Trichloroacetic acid (160 l of 50%) was added to the mixture, which was mixed well and reincubated for a further 10 min on ice. After a 10-min centrifugation at 13,000 ϫ g at 4°C, the pellet was washed with 1 ml of ice-cold acetone before being centrifuged for 10 min at 13,000 ϫ g at 4°C. The pellet was dried by evaporation and dissolved in 500 l of 2ϫ Laemmli buffer containing 10% ␤-mercaptoethanol. Samples were boiled for 5 min, resolved on a 12% SDS-polyacrylamide gel, and transferred to a nitrocellulose membrane for immunoblotting.

RESULTS
Based on homology with other proteins, several specific domains or motifs have been identified in the intracellular N and C termini of TRPC proteins. One of them, ARD, is known to play a major role in protein-protein interactions and is found in many types of proteins, including those involved in signal transduction. We performed an amino acid sequence alignment of the ARDs of 15,401 different proteins in the Pfam data base and calculated the frequency of each amino acid and group of amino acids at precise positions. Amino acids present at a high frequency (in at least 10% of all ARDs) at precise positions determined the "common consensus motif " shown in Fig. 1A. According to this consensus motif, amino acid insertions can be found between positions 15 and 16 and between positions 25 and 26 of some ARDs. Also, amino acids after position 26 or further are deleted in some ARDs. Amino acid sequence alignments of TRPC1, TRPC4, and TRPC6 with the consensus motif  (68). B, structure-based consensus of ARDs proposed by Sedgwick and Smerdon (23). Secondary structure elements in the structure are depicted schematically with ␣-helices, ␤-strands, and loops denoted by cylinders, arrows, and thin lines, respectively. C, amino acid sequence alignments of ankyrin repeats from human TRPC1 (amino acids 44 -183, GenBank TM accession number P48995), human TRPC4 (amino acids 29 -166, GenBank TM accession number Q9UBN4), and human TRPC6 (amino acids 95-243, GenBank TM accession number Q9Y210) were obtained using Clustal alignment analysis. Residues matching the amino acids with a frequency of Ͼ10% in the consensus regions of the ankyrin repeats shown in A are shaded in black, and residues matching the amino acids with a frequency of 7.5-9.9% in the consensus regions of the ankyrin repeats shown in A are shaded in gray.
revealed that each TRPC contained four ARDs. For the ARDs of TRPC1, TRPC4, and TRPC6, 44.2, 50.8, and 49.2%, respectively, of the amino acids were identical to those of the consensus motif (Fig. 1C).
To identify proteins that interact with TRPCs, we performed a yeast two-hybrid screen of a human brain cDNA library with the amino acids 3-259 of TRPC6 as bait. From this initial screening, we obtained 32 potential positive clones that grew on a Trp/Leu/ His/Ade-deficient medium, that were positive for ␣and ␤-galactosidase, and that were in frame with the GAL4-activating domain. To determine the specificity of their interactions with the ARD, each potential clone was transformed into the AH109 yeast strain and mated with the Y187 yeast strain expressing either lamin C, the N terminus of TRPC6 without the ARD (TRPC6 3-93), or the N terminus of TRPC6 with the ARD (TRPC6 3-259). Mated yeast were plated on a high stringency medium (ϪTrp/ ϪLeu/ϪHis/ϪAde ϩX-␣-gal) and a Trp/Leu-deficient medium. Despite the fact that all of the clones were highly expressed in the diploid yeast (Fig. 2B), only four were able to grow with TRPC6 3-259 but not with TRPC6 3-93 or lamin C ( Fig. 2A). One of the four clones contained a 0.8-kb cDNA fragment that encoded the C terminus of MxA, a member of the dynamin superfamily, which has antiviral properties.
To identify the exact region of the ARD required for the interaction with MxA, we made progressive deletions from the N terminus and C terminus of the ARD of TRPC6 (Fig. 3A) and tested for interactions using the yeast two-hybrid assay. All of the deletion mutants of the ARD of TRPC6, as well as the empty bait plasmid pGBKT7 and lamin C, were transformed individually into the Y187 yeast strain and mated with the AH109 yeast strain carrying the prey plasmid pACT2-MxA 547-662. Fig. 3B shows that all of the mutants of the ARD of TRPC6 were expressed in diploid yeast cells plated on a Trp/ Leu-deficient medium. However, only diploid yeast cells that expressed the TRPC6 3-159, TRPC6 3-196, and TRPC6 122-259 sequences were able to grow on a high stringency medium. In contrast, yeast cells containing the empty plasmid pGBKT7, lamin C, or the ARD of TRPC6 that is missing the sequence 131-157 failed to grow on a high stringency medium. These results pointed to the importance of the second ankyrin-like repeat of TRPC6 for the interaction with MxA.
The ability of MxA to interact with the ARDs of TRPCs was confirmed by in vitro and in vivo binding assays. In GST pulldown experiments, the in vitro translated c-Myc-tagged ARDs of TRPC3, -4, -5, -6, and -7 were incubated with GST-MxA 572-662 or GST immobilized on glutathione-Sepharose 4B affinity beads. As shown in Fig. 4A, whereas no significant binding was observed with GST alone, the c-Myc-tagged ARDs of all of the TRPCs tested were efficiently retained by GST-MxA 572-662. These results were confirmed in vivo with a co-immunoprecipitation approach. FLAG-tagged MxA was transiently co-overexpressed in HEK293T cells with HA-tagged TRPC1, -3, -4, or -6. As shown in Fig. 4B, TRPCs immunoprecipitated with an anti-HA antibody and immunoblotting with an anti-FLAG antibody revealed the presence of MxA in the immunoprecipitates. This typical band was found in the immunoprecipitates from TRPC1, -3, -4, and -6. It is important to note that lysates from HEK293T cells contained an endogenous protein of molecular weight similar to that of transfected FLAG-MxA (78 -80 kDa), and this protein was nonspecifically recognized by the anti-FLAG antibody. This endogenous protein was absent in the HA immunoprecipitates from cells transfected with TRPCs alone. The interaction between TRPC6 and MxA was also seen after immunoprecipitating FLAG-MxA and immunoblotting for the presence of TRPC6 (data not shown). These results showed that MxA interacted with all TRPCs tested.  2. Amino acids 547-662 of MxA interact specifically with the ARD of TRPC6. A, the S. cerevisiae Y187 strain pretransformed with the GAL4-binding domain (GAL4-BD) constructs pGBKT7 (Vector), pGBKT7-human lamin C, pGBKT7-TRPC6 3-93, or pGBKT7-TRPC6 3-259 were mated with S. cerevisiae AH109 expressing the GAL4-activating domain (GAL4-AD) containing amino acids 547-662 of MxA. A high stringency selective medium (ϪLeu, ϪTrp, ϪHis, ϪAde, ϩX-␣-gal) was used to grow mated clones. B, the mated clones expressing the constructions described in A were grown overnight at 30°C in selective liquid medium (ϪTrp). Proteins were extracted, trichloroacetic acid-precipitated, resuspended in 2ϫ Laemmli buffer, and analyzed by SDS-PAGE, followed by specific immunoblotting directed against the N-terminal c-Myc epitope.

FIG. 3. Amino acids 547-662 of MxA interact specifically with the second ankyrin-like repeat of TRPC6 in a yeast two-hybrid assay.
A, representation of the various GAL4-BD-TRPC6-fusion fragments used in the yeast two-hybrid assay for the identification of the TRPC6 region that interacts with MxA. B, S. cerevisiae Y187 strains pretransformed with the GAL4-BD-TRPC6-fusion fragments shown in A were mated with S. cerevisiae AH109 pretransformed with the GAL4-AD fused to amino acids 547-662 of MxA. A high stringency selective medium (ϪLeu, ϪTrp, ϪHis, ϪAde, ϩX-␣-gal) was used to grow the interacting clones. A selective medium (ϪLeu, ϪTrp) was also used to monitor that the mating step was correctly achieved. Mated clones were grown overnight at 30°C in a selective liquid medium (ϪLeu, ϪTrp). Proteins were extracted, trichloroacetic acid-precipitated, resuspended in 2ϫ Laemmli buffer, and analyzed by SDS-PAGE, followed by specific immunoblotting directed against the N-terminal c-Myc epitope.
To investigate the functional effect of MxA on TRPC6 activity, HEK293T cells were co-transfected with different combinations of cDNAs coding for the AT 1 receptor, MxA, and TRPC6. TRPC6 is directly activated by OAG, whereas the endogenous Ca 2ϩ entry channels expressed in HEK293T cells were not activated by OAG. As shown in Fig. 5A, basal levels of [Ca 2ϩ ] i in all fura-2-loaded transfected HEK293T cells varied between 50 and 60 nM. When 50 M of OAG was applied to control or MxA-transfected cells, the [Ca 2ϩ ] i reached a value of 140 nM, indicating that the expression of MxA did not modify endogenous OAG-induced Ca 2ϩ entry. When TRPC6 was transfected into HEK293T cells, OAG caused a rapid increase in Ca 2ϩ , which reached a plateau level of 180 nM. However, the co-transfection of MxA with TRPC6 enhanced OAG-induced Ca 2ϩ entry, which reached a high plateau level of 225 nM. These results indicated that MxA potentiated the activity of TRPC6 in HEK293T cells.
As was shown of TRPC3, TRPC6 is permeable to Ba 2ϩ , whereas endogenous Ca 2ϩ entry channels expressed in HEK293T cells are poorly permeable to Ba 2ϩ (31). Fig. 5B shows that CCh-induced Ba 2ϩ entry was very low (0.51 fluorescence units) in control HEK293T cells. In MxA-transfected cells, CCh-induced Ba 2ϩ entry was similar to that of control cells and increased by only 0.57 fluorescence units. In TRPC6transfected cells, CCh-induced Ba 2ϩ entry was very pronounced (2.73 fluorescence units) and, interestingly, was potentiated in TRPC6-and MxA-co-transfected cells (3.68 fluorescence units). These results further suggested that MxA potentiated the activity of TRPC6 but not the activity of endogenous Ca 2ϩ entry channels.
We also investigated the functional effect of MxA on Ca 2ϩ entry in HEK293T cells co-transfected with different combinations of cDNAs coding for the AT 1 receptor, MxA, and TRPC6. To discriminate between AngII-induced Ca 2ϩ release and AngII-induced Ca 2ϩ entry, we used a Ca 2ϩ depletion/Ca 2ϩ readdition protocol. Fura-2-loaded transfected HEK293T cells were incubated for 30 s in a Ca 2ϩ -free medium before their intracellular Ca 2ϩ stores were depleted with 100 nM AngII. Once the [Ca 2ϩ ] i had returned to the basal level, extracellular Ca 2ϩ was restored to 1.8 mM. As shown in Fig. 6A, the basal [Ca 2ϩ ] i was ϳ60 nM in control HEK293T cells. In the absence of extracellular Ca 2ϩ , 100 nM AngII produced a transient elevation in [Ca 2ϩ ] i that corresponded to the depletion of the intracellular Ca 2ϩ store. Upon the addition of 1.8 mM CaCl 2 to the external medium, a sustained entry of Ca 2ϩ raised the [Ca 2ϩ ] i to a plateau level of 110 nM. In MxA-transfected cells, the basal level of Ca 2ϩ and the AngII-induced Ca 2ϩ transient were similar to those observed in control cells, but Ca 2ϩ entry upon the addition of external Ca 2ϩ slightly increased compared with control cells. In TRPC6-transfected HEK293T cells, the basal level of Ca 2ϩ and the AngII-induced Ca 2ϩ transient were also similar to those observed in control cells, but Ca 2ϩ entry upon the addition of external Ca 2ϩ increased significantly compared with control cells and reached a high plateau level of 160 nM. When MxA was transiently co-transfected with TRPC6, Ca 2ϩ entry reached a relatively low plateau level of 135 nM. These results suggested that the combination of MxA and TRPC6 affects endogenous Ca 2ϩ entry activity. To verify this hypothesis, we measured store-operated Ca 2ϩ entry with thapsigargin (TG), which releases Ca 2ϩ from intracellular stores by specific inhibition of microsomal Ca 2ϩ -ATPase activity. Depletion of Ca 2ϩ stores activates store-operated Ca 2ϩ entry in a variety of cell types (32). As shown in Fig. 6B, in a Ca 2ϩ -free medium, the basal level of Ca 2ϩ in HEK293T cells was ϳ100 nM. The addition of TG produced a transient elevation in [Ca 2ϩ ] i that slowly returned to the basal level. Upon the addition of 1.8 mM CaCl 2 to the external medium, the [Ca 2ϩ ] i reached a high plateau level of 200 nM. In cells transfected with TRPC6, store-operated Ca 2ϩ entry was the same as that observed in control cells. These results demonstrate that TRPC6 was not activated by depletion of the internal Ca 2ϩ stores, as previously reported in other studies (33,34). In MxA-transfected HEK293T cells, TGinduced Ca 2ϩ entry reached a plateau level of 225 nM, which was significantly higher than that in control cells. However, in cells co-transfected with MxA and TRPC6, TG-induced Ca 2ϩ entry reached a plateau level of 190 nM, which was significantly lower than that in control cells. These results demonstrate that co-expressed MxA and TRPC6 decreased endogenous storeoperated calcium entry activity of HEK293T cells.
As reported in previous studies, binding of GTP is required for complete activation of MxA. GTP binding-deficient mutants (MxA K83A and MxA T103A) (35,36) were used to determine the role of GTP in the modulation of TRPC6 activity by MxA. HEK293T cells were transfected with TRPC6 and with or without MxA, MxA K83A, or MxA T103A. In TRPC6-transfected HEK293T cells, the basal level of Ca 2ϩ varied between 80 and 95 nM, and OAG caused a substantial increase in [Ca 2ϩ ] i , which reached a plateau level of 145 nM within about 2 min (Fig. 7B). Co-expression of MxA increased that plateau level to 180 nM, whereas co-expression of GTP binding-deficient mutants MxA K83A or MxA T103A, at levels similar to that of MxA (Fig. 7A), did not modify OAG-induced Ca 2ϩ entry through TRPC6 (Fig.  7B). However, co-transfection with the GTP hydrolysis-deficient mutant MxA L612K, which retained its capacity to bind GTP (37,38), enhanced OAG-induced Ca 2ϩ entry through TRPC6 to the same extent as wild-type MxA. In co-immunoprecipitation experiments, GTP binding-deficient mutants retained the capacity to interact with TRPC6 (data not shown). These results suggest that the potentiating effect of MxA on TRPC6 activity was highly dependent on the binding of GTP to MxA.
To determine whether endogenous MxA could regulate the activity of TRPC6, cells were treated with 1000 units/ml of IFN-␣ for 24 h. As shown in Fig. 8A, treatment with IFN-␣ significantly increased the endogenous expression of MxA in cells stably expressing TRPC6. The treatment did not affect the CCh-induced Ca 2ϩ transient, indicating that the activities of the muscarinic receptor, the IP 3 receptor, the phospholipase C, and the Ca 2ϩ -ATPase had not been affected (Fig. 8B). In untreated cells, restoration of extracellular Ca 2ϩ to 1.8 mM caused a rapid increase in [Ca 2ϩ ] i , which reached a high level of 245 nM. Similarly to results obtained with cells transiently transfected with MxA and TRPC6, treatment of TRPC6 cells treated with 1000 units/ml IFN-␣ showed a decreased CCh-induced Ca 2ϩ entry (Fig. 8B). Similar results were obtained with a TG-induced Ca 2ϩ entry protocol (data not shown). Additionally, as also observed with MxA-transfected cells (Fig. 5A), treatment with 1000 units/ml IFN-␣ enhanced OAG-induced Ca 2ϩ entry (Fig. 8C). These results strongly suggest that the endogenously expressed MxA can also regulate the activity of TRPC6.

DISCUSSION
In the study presented here, we demonstrated that the C terminus of MxA interacted with the second ankyrin-like repeat of all TRPCs and enhanced the activity of TRPC6. ARDs are 33-residue sequence motifs found in over 4,000 proteins and are defined as a ␤-hairpin-helix-loop-helix (␤ 2 ␣ 2 ) structures responsible for protein-protein interactions (19,20). An evaluation of the ARD sequences of all of these proteins revealed a common consensus motif. The ARDs of TRPCs had some sequence deviations compared with the common consensus motif. The most noticeable deviations were amino acid insertions after the second helix of the first and third ankyrin repeats as well as amino acid deletions after the second helix of the second ankyrin repeat. Other proteins also exhibit some distinctive variations in their ARDs. For example, the NMR and x-ray analyses of IB␣ and of the yeast protein Swi6 structures showed that these proteins have insertions in the connecting linkers of their ARDs that form an additional helix (39 -41). Swi6 also has a shorter ␤-hairpin section between two of its helix-loop-helix motifs. These variations, however, do not affect the functionality of the ARDs, which are important for the biological functions of IB␣ and Swi6. It is thus likely that the insertions and deletions in the ARDs of TRPCs do not affect their basic structures and that these ARDs can play a significant functional role.
Since MxA is a member of the large dynamin GTPase superfamily and enhances the activity of TRPC6, we considered the possibility that MxA could be involved in the trafficking of TRPC6. We and others have shown that some TRPCs undergo exocytosis insertion after cell stimulation (42)(43)(44). Interestingly, the deletion of the first 131 amino acids of TRPC6, a sequence that contains the first predicted ARD, blocks the routing of TRPC6 to the plasma membrane (45). Also, Wedel et al. (46) reported that the deletion of the first 199 amino acids of TRPC3, a sequence that contains the entire ARD, causes intracellular retention with no detectable labeling of the plasma membrane and a loss of function of TRPC3 in response to methacholine or OAG. A shorter deletion of the first 27 amino acids of TRPC3, a sequence just upstream from the ARD, produced a functional channel that was still able to enhance OAG-induced Ca 2ϩ entry. However, co-expression with MxA did not significantly increase the amount of TRPC6 at the plasma membrane and did not enhance its maturation process (data not shown). Nevertheless, we do not rule out the possibility that MxA could be involved in the intracellular trafficking of TRPCs, and this phenomenon would not be observed in overexpression conditions. Further studies are needed to provide a clear answer to this important question.
Until now, the unique cellular role attributed to MxA is the protection of cells against viral infections (for a review, see Ref. 47). Whereas the exact mechanism of action of MxA is unclear and seems to vary depending on the nature of the infecting virus, we identified some similarities in the antiviral activity of MxA and its TRPC6 regulatory activity. Like the interaction with the ARDs of TRPCs, the last stretch of amino acids at the C terminus of MxA, which is predicted to be a leucine zipper (48), also interacts with viral particles and is important for the specificity of the antiviral activity (36,49,50). This feature is conserved in classic dynamin, which contains, in its C terminus, a proline-rich domain responsible for its interaction with many proteins involved in trafficking, including amphiphysin and endophilin (51,52), or in signal transduction, including calcineurin, Gbr2, and phospholipase C ␥ (53)(54)(55). Another similarity between the antiviral activity of MxA and its TRPC6 regulatory activity is its requirement for GTP binding but not for GTPase activity (35)(36)(37)(38). These results strongly suggest that MxA uses an identical mechanism of action to regulate Ca 2ϩ signaling and to protect cells against viral infections.
MxA may act by interfering the formation of vesicles responsible for the intracellular trafficking of TRPC6. Although MxA is generally considered to be cytosolic, in some cell types it is associated with the smooth endoplasmic reticulum or the cytoskeleton (47,56,57). In addition, it was demonstrated that MxA can bind to lipids and modulate the shape of membranes in vitro (57). As with classic dynamin (58 -62), MxA can selfassemble into large oligomeric complexes, forming rod or ringlike structures (61), and the leucine zipper in the C terminus of MxA is responsible for this oligomerization step (37,66). MxA interacts with the nucleocapsid proteins of some viruses and mislocates them into a membrane-associated, large perinuclear complex (63)(64)(65). However, the L612K MxA mutant, which enhances that activity of TRPC6 as efficiently as wild-type MxA, is unable to form oligomers (38,66). These results suggest that monomeric MxA modulates the activity of TRPC6, and, most likely, this does not require the formation of vesicles.
Ca 2ϩ entry through endogenously expressed TRPCs in HEK293T cells was slightly enhanced by MxA. This effect was stronger when Ca 2ϩ entry was caused by emptying the endoplasmic reticulum with TG. In cells expressing TRPC6, Ca 2ϩ entry was significantly modified by the co-expression of MxA. On the one hand, MxA clearly increased the activity of TRPC6 when protocols that measured direct activation with OAG or CChinduced Ba 2ϩ entry were used. On the other, MxA decreased (by about 50%) the net entry of Ca 2ϩ into cells when protocols that measured AngII-induced Ca 2ϩ entry or TG-induced Ca 2ϩ entry were used. Similar results were obtained with cells treated with IFN-␣, a cytokine known to induce an antiviral activity in HEK293 cells (67). In wild-type HEK293T cells, agonist-induced and TG-induced Ca 2ϩ entry is dependent on the content of the intracellular Ca 2ϩ store. This process is known as store-operated Ca 2ϩ entry (SOCE). Interestingly, HEK293T cells overexpressing TRPC6 display a Ca 2ϩ entry process that is highly dependent on the activation of a G q protein-coupled receptor. This process is known as receptor-operated Ca 2ϩ entry (33,34). In our system, the co-expression of MxA and TRPC6 probably caused a reduction in the SOCE but an increase in the receptor-operated Ca 2ϩ entry, which resulted in an agonist-induced Ca 2ϩ entry that was similar to that of control cells. A reasonable explanation for the reduction in the SOCE could be that MxA increases or stabilizes the interaction of TRPC6 with an essential but limited protein or factor necessary for the SOCE, making it unavailable for the activation of endogenous SOCE channels. Another possibility could be that MxA is involved in the process of a post-transcriptional modification such as a sumoylation. Previous yeast twohybrid screenings have shown an interaction of Mx1 (a mouse ortholog of MxA) with proteins, such as SUMO (small ubiquitinrelated modifier)-1, Uba2, Ubc9, and PIAS1, that are implicated in the sumoylation process. These studies also revealed interactions between Mx1 and PML and SP100, proteins that can be sumoylated (69). Sumoylation is a post-transcriptional modification that influences the activity of many proteins (70). For exam-ple, sumoylation of GLUT1 and GLUT4, two insulin-sensitive glucose transporters, have opposite effects on their activities (71). In this context, MxA could be implicated in sumoylation or in other post-transcriptional modification processes that could influence TRPC activity. Further studies are needed to clarify the exact mechanism by which MxA regulates the activity of TRPCs, whether it be through a modulatory interaction or through a post-transcriptional modification.
In conclusion, this study provided evidence that MxA is a binding partner that interacts with the ARDs of TRPCs and that can modulate, directly or indirectly, the activity of TRPC6. The binding of GTP to MxA appears to be important for its ability to modulate TRPC6 activity. These results further demonstrate the importance of ARDs for the correct activation and localization of TRPCs.