Interaction of TRPC2 and TRPC6 in erythropoietin modulation of calcium influx.

Erythropoietin (Epo) modulates calcium influx through voltage-independent calcium-permeable channel(s). Here, we characterized the expression of transient receptor potential channels (TRPCs) in primary erythroid cells and examined their regulation. Erythroblasts were isolated from the spleens of phenylhydrazine-treated mice, and Epo stimulation resulted in a significant and dose-dependent increase in [Ca](i). Among the classical TRPC channels, expression of three N-terminal splice variants of TRPC2 (clones 14, 17, and alpha) and of TRPC6 were demonstrated in these erythroblasts by both reverse transcriptase-PCR and Western blotting. Confocal microscopy confirmed localization to the plasma membrane. To determine the function of individual TRPC channels in erythropoietin modulation of calcium influx, digital video imaging was used to measure calcium influx through these TRPCs in a Chinese hamster ovary (CHO) cell model. Single CHO-S cells, expressing transfected Epo-R, were identified by detection of green fluorescent protein. Cells that express transfected TRPCs were identified by detection of blue fluorescent protein. [Ca](i) was monitored with Fura Red. Epo stimulation of CHO-S cells transfected with single TRPC2 isoforms (clone 14, 17, or alpha) and Epo-R resulted in a significant increase in [Ca](i). This was not observed in cells transfected with Epo-R and TRPC6. In addition, coexpression of TRPC6 with TRPC2 and Epo-R inhibited the increase in [Ca](i) observed after Epo stimulation. Immunoprecipitation experiments demonstrated that TRPC2 associates with TRPC6, indicating that these TRPCs can form multimeric channels. These data demonstrate that specific TRPCs are expressed in primary erythroid cells and that two of these channels, TRPC2 and TRPC6, can interact to modulate calcium influx stimulated by erythropoietin.

Erythropoietin (Epo) 1 stimulation of its receptor results in an increase in intracellular calcium in primary erythroid cells, which is mediated through a voltage-independent calcium channel (1,2). Understanding the regulation of calcium channel gating by erythropoietin as well as other hematopoietic growth factors is of fundamental biological importance. Regulation of intracellular calcium by erythropoietin is an important signaling mechanism controlling the proliferation and differentiation of erythroid progenitors and precursors (3)(4)(5)(6)(7)(8)(9)(10). Epo-induced murine erythroid colony growth is enhanced by treatment with the ionophore A23187 and inhibited by treatment with EGTA, a nonspecific calcium chelator (6). Erythropoietin has been shown to activate calcium influx and influence cell proliferation and viability via stimulation of its receptor in nonerythroid cells as well (11)(12)(13)(14)(15)(16)(17)(18). Epo stimulates myoblast proliferation to expand the progenitor population during differentiation and increases cytoplasmic calcium in these cells (11). Epo also has an important neuroprotective and neurotrophic effect on brain tissue, enhancing viability of neuronal cells through a mechanism dependent on calcium influx (12)(13)(14)(15)(16)19). However, the signaling pathways through which erythropoietin activation of its receptor modulates calcium influx are largely unknown.
A transient receptor potential (TRP) protein superfamily, related to the archetypal Drosophila TRP, has been identified. It consists of a diverse group of calcium-permeable cation channels that are expressed on nonexcitable mammalian cells (20 -22). The TRP superfamily has been divided into six subfamilies: C, V, M, N, P, and ML (21). We studied the classical (C) or short TRPCs because they are voltage-independent calciumpermeable channels that can be agonist-regulated and because erythropoietin modulates calcium influx through one of these channels, TRPC2 (23). All mammalian TRPCs share six putative transmembrane domains similar to the core structure of many pore-forming subunits of voltage-gated channels, except they lack charged residues necessary for the voltage sensor. Individual TRPC channels can be activated both by calcium store depletion-dependent and receptor-operated mechanisms. The physical associations between TRPCs and with other proteins are complex (21,24,25). TRPC channels are proposed to assemble as homo-or heterotetrameric ion channels, and the subunit composition influences their biophysical and regulatory properties (21, 26 -30). For many cell types, including primary hematopoietic cells, information about the physiological composition, regulation, and function of the expressed TRPCs is nearly completely lacking.
Here, we utilized a murine erythroblast model to study erythropoietin regulation of calcium influx in primary hematopoietic cells (31)(32)(33)(34). Erythropoietin stimulates a significant and dose-dependent increase in [Ca] i in erythroid cells harvested from the spleens of phenylhydrazine-treated mice, and we show expression of three N-terminal splice variants of TRPC2 (clones 14, 17, and ␣) and TRPC6 in these primary cells with RT-PCR, Western blotting, and confocal microscopy. Using digital video imaging and transfected CHO-S cells in which the TRPC composition can be controlled, we studied the ability of erythropoietin to modulate calcium influx through N-terminal splice variants of TRPC2. Clones 17 and ␣ lack the first 111 and 287 N-terminal amino acids of clone 14, respectively. Erythropoietin stimulated a significant increase in [Ca] i in cells transfected with all three TRPC2 isoforms but not in cells transfected with TRPC6. When coexpressed, TRPC2 and TRPC6 associated and calcium influx was inhibited, suggesting that TRPC2 and TRPC6 can form a multimeric channel with different regulatory properties than the individual TRPC. This conclusion is given further support by the observation that TRPC2 and TRPC6 coassociate in primary erythroid cells.

EXPERIMENTAL PROCEDURES
Tissue and Cell Lines-Brain tissue was obtained from C57Bl/6 mice, frozen in liquid nitrogen, and kept at Ϫ80°C until use. CHO-S cells were cultured in Dulbecco's modified Eagle's medium with 10% fetal calf serum and 0.1 mM nonessential amino acids.
For crude membrane preparations, cell pellets were suspended in Buffer I (10 mM Tris-HCl, pH 7.4, 1ϫ protease inhibitor mixture; Roche Applied Science) and sonicated. Brain was homogenized with a Dounce homogenizer on ice and centrifuged, and Buffer I was then added to the pellet. An equal volume of Buffer II (10 mM Tris-HCl, pH 7.4, 300 mM KCl, 20% sucrose, 1ϫ protease inhibitor mixture) was added. The suspension was then centrifuged at 10,000 ϫ g for 10 min at 4°C, and the supernatant was spun at 100,000 ϫ g for 1 h at 4°C. Crude membranes were solubilized in buffer containing 62 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol. Protein was quantified, and Western blotting was performed as described above.
Immunolocalization of TRPC2 Channels in Primary Erythroid Cells-Ter-119 Ϫ splenic erythroblasts (3 ϫ 10 5 cells/chamber) were placed in each well of Lab-Tek Permanox Chamber Slides precoated with fibronectin. After 45 min, cells were washed twice with PBS, fixed in methanol at Ϫ20°C for 10 min, and permeabilized in 0.5% Triton X-100 in PBS for 5 min. Incubation for 10 min in 20% normal goat serum preceded staining with primary antibodies (anti-TRPC2 clone 14, anti-TRPC2, anti-TRPC6, anti-Epo-R; 1:50) for 30 min at room temperature followed by secondary antibody (goat anti-rabbit Alexa 488, Molecular Probes, Inc., Eugene, OR) for 30 min in the dark. Slides were stained with propidium iodide in Vectashield mounting medium (Vector Laboratories, Burlingame, CA). Images were acquired with the Leica TCS SP2 confocal microscope.
Immunoprecipitation-To determine whether mTRPC2 isoforms interact with mTRPC6, CHO-S cells were transfected with mTRPC2 clone 14 or mTRPC2 ␣ (in pcDNA3.1/V5-His TOPO), mTRPC6 (in pCMV-Tag 2A), or combinations of these vectors. Lysates were prepared from cell pellets with lysis buffer containing 50 mM Tris, pH 7.5, 150 mM NaCl, 1% Triton X, 1 mM EDTA, and 1ϫ protease inhibitor mixture. Protein lysates were incubated with anti-V5 antibody (4 g; Invitrogen) for 2-4 h at 4°C. Protein G-Sepharose was added for 2-4 h 4°C with mixing, and immunoprecipitates washed three times. Sample buffer (2ϫ) was then added to the pellet, and the samples were boiled. For immunoprecipitation of FLAG-TRPC6, cell lysates were preabsorbed with Protein A-Sepharose CL-4B (Amersham Biosciences) and immunoprecipitated with 50 l of anti-FLAG M2 affinity gel (Sigma) for 2 h at 4°C. Samples were washed three times, and peptide elution was performed by adding 40 l of FLAG peptide (at 0.5 mg/ml; Sigma) twice. Western blotting was performed as described above, and blots were probed with anti-V5-HRP (1:1000; Invitrogen), or anti-TRPC6 (1:200) antibodies.
For immunoprecipitation studies with Ter-119 Ϫ erythroblasts, protein lysates were preabsorbed with Protein A-Sepharose CL-4B beads and then incubated with anti-TRPC6 antibody (10 g) for 3 h at 4°C. Protein A-Sepharose CL-4B beads were then added for 2 h at 4°C, and immunoprecipitates were washed. Sample buffer was added, and the samples were boiled.

Measurement of [Ca] i with Digital Video
Imaging-The fluorescence microscopy-coupled digital video imaging system used to measure changes in [Ca] i has been described previously (1-4, 23, 40). Following isolation and separation with the VarioMACS, Ter-119 Ϫ and Ter-119 ϩ splenic erythroblasts were incubated in Iscove's modified Dulbecco's medium containing 2% fetal calf serum and 50 M ␤-mercaptoethanol for 2-4 h without growth factor. Splenic erythroblasts were then adhered to fibronectin-coated glass coverslips and loaded for 20 min with 0.1 M Fura-2/AM (Molecular Probes). Fura-2-loaded cells were visualized with digital video imaging, and fluorescence was quantitated using the fluorescence intensity ratio (R) of the emission (510 nm) measured following excitation at 350 nm divided by the emission following excitation at 380 nm. Base-line fluorescence intensity ratio (R) and the changes in R of individual cells after stimulation with 0 -40 units/ml of recombinant erythropoietin (2000 units/ml; Amgen, Thousand Oaks, CA) were measured. Mature erythrocytes were excluded from analysis of [Ca] i in Ter-119 ϩ cells; most failed to adhere to coverslips or appeared significantly crenated, unlike nucleated cells, which had round cytoplasmic borders and appeared healthy.
To study changes in [Ca] i in transfected cells, we were not able to use Fura-2 as the detection fluorophore, because its excitation and emission wavelengths overlap with those of green fluorescent protein. Instead, we used the fluorescent indicator Fura Red (excitation 460 and 490 nm, emission 600-nm, long pass), a dual wavelength excitation probe (41,42). We initially calibrated the constants S f2 , S b2 , and the K D Ј of Fura Red utilizing a KCl-NaCl-Hepes-MgCl 2 buffer (100 mM KCl, 25 mM Hepes, 10 mM NaCl, 1 mM MgCl 2 , pH 7). To measure R min , 4 mM EGTA was added to the buffer, and to measure R max , 4 mM CaCl 2 was added to the buffer. CHO-S cells were loaded for 25 min at 37°C with 5 M Fura Red in the presence of Pluronic F-127 to enhance loading. CHO-S cells were then exposed to the R min or R max buffers, ionomycin (2 M) was added, and fluorescent measurements were obtained over the next 15 min. We determined R min (minimum fluorescence intensity ratio R of the emission following excitation at 460 nm divided by the emission following excitation at 490 nm), R max (maximum R), and the constants S f2 /S b2 so that [Ca] i could be calculated using the formula . Transfected CHO-S cells grown on glass coverslips were loaded at 48 h with 5 M Fura Red for 25 min at 37°C in the presence of Pluronic F-127. Cells were treated with 0 or 40 units/ml of Epo.
[Ca] i was measured at base line and over a 20-min interval following treatment. Statistical significance of results was analyzed with one-way analysis of variance.

Erythropoietin Modulation of Calcium Influx in Primary
Murine Erythroblasts-Erythropoietin regulation of calcium influx in primary erythroblasts was studied here using a well described murine erythroblast model (31,32). C57Bl/6 mice were injected intraperitoneally with phenylhydrazine on days 1 and 2, and spleens were removed on day 5. Morphological examination of cells following Wright's staining confirmed that greater than 90% of cells were erythroid precursors, consistent with previous reports (31,32). To study cells at different stages of erythroid differentiation, splenic erythroblasts separated into Ter-119 Ϫ and Ter-119 ϩ populations (35) were examined by Wright's staining. The Ter-119 Ϫ population consisted predom-inantly of large, less mature erythroid precursors (proerythroblasts and basophilic erythroblasts), whereas the Ter-119 ϩ population consisted predominantly of smaller, more mature erythroid precursors and erythrocytes (35). There was some overlap in the distribution of erythroblasts at different stages of maturation between cells that sorted as Ter-119 Ϫ and Ter-119 ϩ . Cells were also stained with immunofluorescent antibodies to confirm morphological examination. In contrast to normal spleens in which T cells, B cells, and macrophages represented greater than 90% of cells, in both the Ter-119 Ϫ and Ter-119 ϩ cell populations isolated from the spleens of phenylhydrazine-treated mice, less than 10% of cells stained with antibodies to these other cell types.
To examine erythropoietin regulation of calcium influx in Ter-119 Ϫ and Ter-119 ϩ erythroblasts, [Ca] i was measured with our digital video imaging system in single Fura-2-loaded cells at base line and over 20 min following stimulation with 0 -40 units/ml recombinant erythropoietin. The percentage of Ter-119 Ϫ and Ter-119 ϩ cells that responded to Epo with an increase in intracellular calcium, compared with control cells treated with vehicle (PBS), is shown in Fig. 1A. A greater percentage of Ter-119 Ϫ than Ter-119 ϩ cells responded at all Epo concentrations studied. A significant (p Ͻ 0.002) and dosedependent increase in [Ca] i , measured as F 350 /F 380 , was observed in both Ter-119 Ϫ and Ter-119 ϩ cells in response to Epo stimulation, compared with cells treated with PBS (Fig. 1B). For Ter-119 Ϫ cells, the peak increase of F 350 /F 380 was 85 Ϯ 6% above base line when cells were stimulated with 10 units/ml of Epo, compared with 9 Ϯ 2% above base line for cells treated with 0 units/ml Epo (Fig. 1B). The peak increase in F 350 /F 380 in individual cells characteristically occurred between 10 and 20 min after Epo stimulation, a time course similar to that observed in human BFU-E-derived erythroblasts (4). These results demonstrate that Epo stimulates calcium influx in both Ter-119 Ϫ and Ter-119 ϩ primary erythroblasts.
Expression of TRPC in Primary Murine Erythroid Cells-To identify the classical TRPC (1-7) that are expressed in primary erythroid cells, RT-PCR was performed using RNA isolated from Ter-119 Ϫ and Ter-119 ϩ splenic erythroblasts. RNA was also isolated from murine brain, used as a positive control. As shown in Fig. 2A, RNA for only TRPC2 and TRPC6 was detected in primary erythroid cells using our PCR conditions, whereas RNA for all seven TRPCs was detected in brain. No bands were observed when PCR was performed without the reverse transcriptase step, demonstrating that these products did not result from contaminating DNA. The identity of RT-PCR bands was confirmed by sequencing.
Four splice variants of murine TRPC2 have been identified: clones 14 (36) (GenBank TM accession number AF111108), 17 (36) (accession number AF111107), ␣ (37) (accession number AF230802), and ␤ (37) (accession number AF230803). To determine which isoforms are expressed in primary erythroblasts, RT-PCR was performed using primers designed to distinguish them based on sequence differences or size of the PCR product. A schema comparing cDNAs of the four mTRPC2 isoforms and differences in translation initiation sites is shown in Fig. 2B. Primer nucleotide sequences are provided under "Experimental Procedures." Relative locations of the different primer sets used to distinguish TRPC2 cDNAs are also shown in Fig. 2B. RNA encoding mTRPC2 clones 14, 17, and ␣, but not ␤, were found in Ter-119 Ϫ and Ter-119 ϩ primary erythroid cells using RT-PCR (Fig. 2C).
Generation of Antibodies Specific to mTRPC2 Isoforms-To study the expression and localization of mTRPC2 isoforms, affinity-purified antibodies were generated to an epitope unique to the N terminus of clone 14 (anti-TRPC2 c14), an epitope unique to the N terminus of clone 17 (anti-TRPC2 c17), and an epitope present in the N terminus of all four mTRPC2 isoforms (anti-TRPC2). To characterize antibody specificity, in vitro transcription/translation was performed using cDNAs for mTRPC2 c14, mTRPC2 c17, mTRPC2 ␣, and mTRPC6 cloned into pcDNA3. Western blotting was performed with each of these in vitro translation products and with anti-TRPC2 c14, anti-TRPC2 c17, or anti-TRPC2 antibodies (Fig. 3). Anti-TRPC2 c14 recognized only TRPC2 c14 at 135 kDa. A crossreacting band was observed with this antibody at ϳ110 kDa. Anti-TRPC2 c17 recognized only TRPC2 c17 at 120 kDa. Anti-TRPC2 recognized TRPC2 c14, c17, and ␣ (100 kDa). None of these antibodies recognized TRPC6, which was recognized by anti-TRPC6 antibody. TRPC6 prepared with in vitro translation ran at a slightly lower molecular weight than endogenous TRPC6 or TRPC6 produced in transfected cells.
Expression and Subcellular Localization of mTRPC2 Isoforms-To examine the subcellular localization of mTRPC2 and mTRPC6 in primary erythroid cells, crude membranes were prepared from Ter-119 Ϫ and Ter-119 ϩ erythroblasts (23). Western blotting was performed with protein isolated in the 100,000 ϫ g pellet or supernatant using anti-TRPC2 antibodies. Representative results of three experiments demonstrating membrane expression of TRPC2 clones 14 (135 kDa), 17 (120 kDa), and ␣ (100 kDa) are shown in Fig. 4A. Minimal or absent TRPC2 expression was observed in the supernatants (results not shown). Western blotting was also performed on membrane preparations from Ter-119 Ϫ and Ter-119 ϩ cells with anti-TRPC6 antibody. Brain was used as the positive control. Representative results of three experiments, shown on Fig. 4B, demonstrate expression of TRPC6 in membrane preparations from brain, Ter-119 Ϫ , and Ter-119 ϩ cells. Indistinct higher molecular weight bands were observed in Western blots of TRPC6 and are consistent with reports that TRPC6 is highly glycosylated in certain cell types (43)(44)(45). The quality of membrane preparations was demonstrated by reprobing blots with antibodies to murine Epo-R. Epo-R was predominantly located in the membrane fraction of Ter-119 Ϫ and Ter-119 ϩ cells (Fig.  4, A and B).
To confirm that TRPC are expressed in the plasma membrane of primary erythroid cells, immunolocalization studies were performed with Ter-119 Ϫ erythroblasts using anti-TRPC2 clone 14, anti-TRPC2, anti-TRPC6, and anti-Epo-R antibodies and confocal microscopy. Representative results, shown in Fig.  5, demonstrate that endogenous TRPC2 isoforms are localized at or near the plasma membrane in primary erythroid cells. TRPC6 also localized on or near the plasma membrane. Nonhomogeneous punctate cytoplasmic staining was occasionally observed, which may represent intracellular membrane structures. Epo-R localized strongly to the plasma membrane.
Erythropoietin Modulates Calcium Influx through TRPC2 Isoforms but Not through TRPC6 -We established a system to examine regulation of calcium influx through individual TRPCs. Single CHO-S cells were transfected with wild type Epo-R in pTracer-CMV and with TRPC in pQBI50. The pTracer-CMV vector (Invitrogen) contains an SV40 promoter driving expression of the green fluorescent protein gene and a CMV promoter driving expression of mEpo-R (23,40). The pQBI50 vector contains a CMV promoter, which drives expression of SuperGlo BFP fused through a flexible linker to TRPC2 or TRPC6. For digital video imaging, successful transfection of CHO-S cells with Epo-R was verified by detection of green fluorescent protein (excitation, 478 nm; emission, 535 nm) and successful transfection of TRPC with BFP (excitation, 380 nm; emission, 435 nm) (1-4, 23, 40). [Ca] i was measured simultaneously in successfully transfected cells (23). CHO-S cells are a good model system, because they lack endogenous Epo-R, express undetectable amounts of mTRPC2 or mTRPC6 orthologs, and have the necessary transducers required (23,40,46). Cotransfection of Epo-R and TRPC into CHO-S cells permits the study of the interaction of Epo-R with specific TRPC. This system was utilized to determine whether erythropoietin modulates calcium influx through each of the three TRPC2 splice variants expressed in primary erythroid cells, which have significant differences in their N terminus.
CHO-S cells were cotransfected with Epo-R in pTracer-CMV and TRPC2 clone 14, 17, or ␣ or TRPC6 in pQBI50. To confirm that all TRPC2 isoforms and TRPC6 in transfected cells have a similar subcellular distribution, crude membrane fractions were prepared from CHO-S cells cotransfected with TRPC2 c14, c17, ␣, or TRPC6 in pQBI50. Western blotting was performed with protein isolated in the 100,000 ϫ g pellet or the supernatant. Representative results of three experiments demonstrating membrane expression of all three TRPC2 N-terminal splice variants and TRPC6 are shown in Fig. 6A. Western blotting was also performed with the supernatants from the 100,000 ϫ g centrifugation step, but little or no TRPC2 or TRPC6 was observed in this fraction (not shown). Hence, failure to localize to the membrane does not explain the lack of response to Epo of cells transfected with TRPC6. The higher molecular weights of proteins shown here compared with reticulocyte lysates or cells transfected with TRPC in pcDNA3 is the result of linkage to BFP. In addition, indistinct higher molecular weight bands were observed in TRPC6-transfected CHO-S cells. These bands are consistent with previous reports that TRPC6 is highly glycosylated in certain cell types (43)(44)(45) and are more intense than observed with endogenous TRPC6.
TRPC6 Inhibits Calcium Influx through TRPC2-TRPC are proposed to function as heterotetrameric calcium-permeable channels, and three TRPC2 isoforms and TRPC6 are expressed on primary erythroid cells. The influence of TRPC subunit expression on calcium influx in response to Epo was examined in transfected CHO-S cells in which the composition of TRPC can be controlled. CHO-S cells were cotransfected with Epo-R in pTracer-CMV and with combinations of TRPC2 c14, TRPC2 ␣, and TRPC6 in pQBI50. The amount of DNA transfected for each plasmid was equivalent. At 48 h, cells were loaded with Fura Red, and digital video imaging was performed to determine the effect of Epo stimulation on [Ca] i in successfully transfected cells (Table II). Cotransfection of CHO-S cells with Epo-R, TRPC2 c14, and TRPC2 ␣ demonstrated a significant increase in [Ca] i of 299 Ϯ 72% above base line for cells stimulated with Epo, which was not significantly different from cells transfected with Epo-R and TRPC2 c14 or ␣ individually (Table  I) 14) or in vitro translation products (TRPC2 clones 17 and ␣). Western blotting was performed with anti-TRPC2 c14, anti-TRPC2 c17, or anti-TRPC2 antibodies. Anti-Epo-R antibody was used to confirm the quality of the membrane preparations. B, membrane localization of TRPC6 in primary murine erythroid cells. Western blotting was performed on membranes prepared from Ter-119 Ϫ and Ter-119 ϩ erythroblasts (150 g/lane). Brain was used as a positive control. Western blotting was performed with anti-TRPC6 antibody, and anti-Epo-R antibody was used to confirm the quality of the membrane preparations. and TRPC6 (44 Ϯ 10%) was significantly less than that of cells transfected with Epo-R/TRPC2 c14 (226 Ϯ 21%, p Ͻ 0.0001) or Epo-R/TRPC2 c14/TRPC2 ␣ (299 Ϯ 72%, p Ͻ 0.0001). Similarly, the increase in [Ca] i in response to Epo of CHO-S cells cotransfected with Epo-R, TRPC2 ␣, and TRPC6 (64 Ϯ 27%) was significantly less than that of cells transfected with Epo-R/ TRPC2␣ (201 Ϯ 28%, p Ͻ 0.02) or Epo-R/TRPC2 c14/TRPC2 ␣ (p Ͻ 0.03). The increase in [Ca] i in response to Epo in cells cotransfected with Epo-R, TRPC2 c14, TRPC2 ␣ and TRPC6 (67 Ϯ 14%) was also significantly less than that for cells transfected with Epo-R, TRPC2 c14, and TRPC2 ␣ (299 Ϯ 72%, p Ͻ 0.001). These data demonstrate the ability of TRPC6 to inhibit calcium influx through TRPC2 isoforms.
Western blotting was performed to establish that the inhibi-tion of calcium influx in response to Epo in cells expressing TRPC6 did not result from decreased expression of functional TRPC2 isoforms in TRPC6 cotransfected cells. Crude membrane fractions were prepared from CHO-S cells cotransfected with TRPC2 c14 and ␣, TRPC2 c14 and TRPC6, and TRPC2 c14, ␣, and TRPC6 in pQBI50, under conditions used in digital video imaging experiments. Representative results of two experiments, shown in Fig. 6B, demonstrate that the amount of TRPC2 clone 14 and ␣ expressed in doubly transfected cells, which responded to Epo with a significant increase in [Ca] i , is similar to that expressed in cells that were also cotransfected with TRPC6, which had a significantly reduced response to Epo. Direct Protein Interaction between TRPC2 Isoforms and TRPC6 -To determine whether TRPC6 might inhibit calcium FIG . 5. Immunolocalization of TRPC2, TRPC6, and Epo-R in Ter-119 ؊ erythroblasts. Ter-119 Ϫ erythroblasts fixed to glass coverslips were stained with anti-TRPC2 clone 14 (A and F), anti-TRPC2 (B and G), anti-TRPC6 (C and H), or anti-Epo-R (D and I) antibodies or nonimmune rabbit serum (J). Goat anti-rabbit Alexa 488 was used as the secondary antibody. Propidium iodide was used to stain DNA, and merged images of Alexa 488 and propidium iodide are shown in F-J. Sections in the midplane of the cell were obtained with confocal microscopy. E shows a bright light image of a representative Ter-119 Ϫ cell. a Percentage above base line ϭ peak ͓Ca͔ i divided by base-line ͓Ca͔ i ϫ 100% Ϫ 100% (base line). b n ϭ number of cells. c Significant increase above PBS control, p Ͻ 0.001.
influx by a direct interaction with TRPC2, TRPC2 isoforms were tagged with V5 (c14 and ␣), and TRPC6 was tagged with FLAG. Western blotting of each construct, expressed in transfected CHO-S cells, with anti-TRPC2 c14, anti-TRPC2, anti-V5-HRP, anti-TRPC6, and anti-FLAG antibodies, demonstrates the ability of each antibody to specifically recognize the appropriate tagged channel (Fig. 7). TRPC2 c14 tagged with V5 (pcDNA3.1/V5-His TOPO) and TRPC6 fused to the FLAG tag (pCMV-Tag 2A) were then expressed individually or together in CHO-S cells and lysates prepared. Immunoprecipitation was performed with anti-FLAG affinity gel or anti-V5 antibody. Western blotting demonstrated that anti-FLAG immunoprecipitated TRPC2 c14 in the presence of FLAG-TRPC6 and that anti-V5 antibody immunoprecipitated TRPC6 in the presence of V5-TRPC2 c14, confirming their ability to coassociate (Fig. 8). Immunoprecipitation with anti-FLAG effectively precipitated the higher molecular weight TRPC6 bands, which previous reports demonstrated represent glycosylated protein (43)(44)(45), as well as lower molecular weight bands. Similarly, anti-FLAG immunoprecipitated TRPC2 ␣ in the presence of FLAG-TRPC6, and anti-V5 antibody immunoprecipitated TRPC6 in the presence of V5-␣, demonstrating that TRPC2 ␣ and TRPC6 can also coassociate (Fig.  9). Examination of the supernatant fractions from each immunoprecipitation, not shown here, demonstrated that whereas immunoprecipitation of the target TRPC was very efficient, only a portion of the coassociated channel present in the cell immunoprecipitated with the target TRPC. TRPC6 was not immunoprecipitated by anti-V5 in cells expressing FLAG-TRPC6 alone; nor was V5-TRPC2 c14 or V5-TRPC2 ␣ immunoprecipitated by anti-FLAG in cells expressing each of these constructs alone. In addition, each channel was expressed to a similar extent in lysates from cells expressing V5-TRPC2 and FLAG-TRPC6 or each construct alone (Figs. 8 and 9; lysates not shown), demonstrating that the ability to detect each channel after immunoprecipitation was not a secondary consequence of increased channel stability in cells expressing both proteins.
To confirm the physiological significance of these findings, the ability of TRPC2 and TRPC6 to physically interact was studied in primary erythroid cells. Immunoprecipitation with anti-TRPC6 was performed with lysates of Ter-119 Ϫ erythroblasts. Western blots were probed first with anti-TRPC2 c14 antibody, and a band was observed at the expected molecular weight of TRPC2 c14 (Fig. 10). The effectiveness of immunoprecipitation was demonstrated by reprobing blots with anti-TRPC6. Immunoprecipitation with preimmune serum did not precipitate TRPC2 or TRPC6 (Fig. 10). These results provide evidence that TRPC2 and TRPC6 coassociate in primary erythroid cells. Reciprocal experiments with anti-TRPC2 c14 or anti-TRPC2 antibodies could not be performed, because these antibodies did not effectively immunoprecipitate their targets in transfected cells or primary erythroid cells. This may because they are not high titer antibodies or because the epitopes recognized by these antibodies are not highly accessible in the native protein.  8. Coimmunoprecipitation of TRPC2 c14 and TRPC6. TRPC2 c14 C-terminally tagged with V5 and TRPC6 N-terminally tagged with FLAG were expressed individually or together in CHO-S cells. Immunoprecipitation (IP) was performed on lysates prepared with 1% Triton X-100 from transfected cells with anti-FLAG M2 affinity gel or anti-V5 antibody. Western blotting (WB) of eluates of bound proteins (shown here) and supernatants (not shown) was performed on each blot with anti-TRPC6 and anti-V5 HRP antibodies.

DISCUSSION
Erythropoietin has been shown to increase [Ca] i in human and murine erythroid cells through voltage-independent ion channels (2, 7). A major impediment in determining the mechanisms through which erythropoietin modulates calcium entry and the functional role of calcium influx in cell proliferation, differentiation, and survival has been the difficulty in identifying and cloning the calcium-permeable channel(s) involved. Recently, mammalian isoforms of Drosophila calcium-permeable TRPCs have been described that are involved in the sustained phase of calcium entry in nonexcitable mammalian cells. Here, the expression of two classical TRPCs, TRPC2 and TRPC6, was demonstrated in primary erythroid cells. A CHO cell model system was used to examine regulation of individual TRPCs, which are expressed in erythroid cells. Erythropoietin modulated calcium influx through all three TRPC2 N-terminal splice variants but not TRPC6. Since erythroid cells express both TRPC2 isoforms and TRPC6, the possibility that heteromultimeric interactions contribute to receptor-modulated channel activity was explored. TRPC6 suppressed the Epoinduced increase in [Ca] i through TRPC2, demonstrating an important functional consequence of TRPC2 and TRPC6 interaction.
The first important finding of this report is the identification of two classical TRPCs, TRPC2 and TRPC6, which are expressed in primary murine erythroid cells, by RT-PCR, Western blotting, and confocal microscopy. For detection of TRPC protein in primary erythroid cells by Western blotting, large quantities of membrane protein were loaded (200 g/lane), and long exposure times were required. Immunofluorescence with anti-TRPC2 or anti-TRPC6 antibodies was not intense compared with fluorescence with other antibodies including anti-Epo-R. Together, these data suggest that the quantity of endogenous TRPC protein expressed in erythroblasts is low. This is consistent with observations made with other ion channels, since relatively few channels are required to significantly influence [Ca] i . Because precise quantitation of TRPC2 and TRPC6 protein levels is difficult, we could not definitively determine whether the TRPC2/TRPC6 ratio in the plasma membrane changes in Ter-119 Ϫ cells, compared with Ter-119 ϩ cells, during erythroid differentiation. However, our studies confirm that erythropoietin modulates calcium influx through TRPC2 (23) and provide support for the hypothesis that TRPC2 may have an important role in hematopoiesis, in addition to its reported roles in sensory activation of vomeronasal organ neurons and in regulation of sustained Ca 2ϩ entry in sperm (47,48). The ability of erythropoietin to modulate calcium influx through three splice variants of TRPC2 expressed on primary erythroid cells was explored. These N-terminal splice variants differ in that clone 17 is missing the first 111 amino acids of clone 14, and its first 11 amino acids are unique (Fig. 3) (36). Domains that are present on clone 14 but absent from clone 17 include glycosylation, myristoylation, and protein kinase C phosphorylation sites. TRPC2 ␣ is missing the first 287 amino acids of clone 14 (37). The ␣ isoform is lacking a putative nuclear localization signal, arginine-rich regions, and a phospholipase C␥ Src homology 3 binding site present on clones 14 and 17. Although controversy exists about the expression of TRPC2 ␣ in brain and testis (37,49), we were able to detect RNA and protein for TRPC2 ␣ in murine erythroid cells. Despite deletions in their N terminus, all TRPC2 isoforms predominantly localized to the plasma membrane. The second major finding of this report is that erythropoietin modulation of calcium influx through clone 14, 17, or ␣ is not significantly different; this observation suggests that the domains absent from the ␣ isoform are not critical for regulation of TRPC2 channel opening by erythropoietin.
TRPC proteins have been proposed to function as unitary subunits of tetrameric ion channels, whose characteristics differ with the type of TRPC that coassemble to form the channel (21,50). Several examples of TRPC interactions have been reported. First, N-terminal splice variants of Drosophila (26,27) and mammalian (51) TRPC can bind to and alter the activity of their full-length proteins. For example, transient receptor potential channel subfamily melastatin 1 (MLSN1) is expressed in melanocytes, and its expression level correlates inversely with melanoma aggressiveness and the potential for melanoma metastasis (52,53). A short N-terminal splice variant of MLSN1, devoid of any transmembrane segment, inhibits membrane localization and calcium influx through association with full-length MLSN1 (54). In addition, the ability of different TRPCs to form heteromultimeric channels is a critical determinant of their biophysical and functional properties, although the cellular principles regulating homo-and heteromeric channel assembly have not been elucidated. Multimeric channel formation has been reported for TRPC1 and TRPC3 (25); TRPC1 and TRPC5 (28); TRPC4 and TRPC5 (30, 55); TRPC3, TRPC6, and TRPC7 (30, 55); and TRPV5 and TRPV6 (29). Characteristic of heteromultimeric compared with homo- FIG. 9. Coimmunoprecipitation of TRPC2 ␣ and TRPC6. TRPC2 ␣ C-terminally tagged with V5 and TRPC6 N-terminally tagged with FLAG were expressed individually or together in CHO-S cells. Immunoprecipitation (IP) was performed on lysates from transfected cells with anti-FLAG M2 affinity gel or anti-V5 antibody. Western blotting (WB) of eluates of bound proteins (shown here) and supernatants (not shown) was performed on each blot with anti-TRPC6 or anti-V5 HRP antibodies.
FIG. 10. Immunoprecipitation of TRPC6 in Ter-119 ؊ erythroblasts. Immunoprecipitation (IP) was performed on lysates from Ter-119 Ϫ erythroblasts with anti-TRPC6 antibody or preimmune rabbit serum (NS). Western blotting (WB) of eluates of bound protein was performed with anti-TRPC2 c14 or anti-TRPC6 antibodies. Representative results of two experiments are shown. meric channels, coexpression of TRPC1 and TRPC3 in HEK293 cells resulted in a cation channel with properties substantially different from those of the individual TRPC; coexpression of TRPC1 and TRPC3 suppressed substantial carbachol-induced Ca 2ϩ entry and abolished 1-oleoyl-2-acetyl-sn-glycerol-induced Sr 2ϩ entry signals observed with TRPC3 (25). Our data demonstrating that TRPC6 can inhibit erythropoietin-induced calcium influx through TRPC2 are consistent with the observation that subunit interaction is important in the physiological regulation of TRPCs. Previously, TRPC2 was not found to interact with any other TRPC (30), but in our studies, an association between TRPC2 and TRPC6 was clearly demonstrated. This is not completely unexpected because of the relatively close phylogenetic relationship of these two TRPCs (30). The third major conclusion of this report is that the stoichiometry of expression of TRPC2 and TRPC6 may be an important mechanism for regulating calcium influx in erythroid cells in response to erythropoietin stimulation. TRPC6 has been reported to be down-regulated in tumor cells (56), and although TRPC6 is expressed in human neutrophils, its expression is not detectable in HL-60 cells (57). We speculate that the ability of TRPC6 to modulate calcium influx through other TRPCs may be part of the mechanism through which TRPC6 influences the proliferative capacity these cells.
Important questions remain regarding the mechanisms through which erythropoietin regulates TRPC2 and through which TRPC6 inhibits TRPC2 channel activity. The observation that rat TRPC2 localizes to the microvillar plasma membrane of VNO receptor neurons, which do not have calcium stores (47), supports our conclusion that receptor-activated, second messenger pathways can regulate TRPC2 opening. These pathways need to be defined for erythropoietin, as does the functional role of TRPC2 activation in erythroid proliferation, differentiation, and cell survival. Future studies will be required to determine how the stoichiometry of TRPC2 and TRPC6 expression and heteromultimeric channel formation is regulated in hematopoietic cells.