Store-operated Ca2+ Current and TRPV6 Channels in Lymph Node Prostate Cancer Cells*

The contribution of endogenous and recombinant transient receptor potential vanilloid type 6 (TRPV6) channels to Ca2+ entry across the plasma membrane was studied in the human lymph node prostate cancer cell line (LNCaP). LNCaP cells do express the TRPV6 gene, and Ca2+ entry currents in these cells were detected after active and passive Ca2+ store depletion by intracellular application of inositol 1,4,5-trisphosphate, Ca2+ chelators, and the sarcoplasmic/endoplasmic reticulum Ca2+-ATPase inhibitor thapsigargin. This store-operated Ca2+ current (ISOC) had biophysical properties similar to those of the Ca2+ release-activated Ca2+ current (ICRAC) in rat basophilic leukemia cells such as the activation mechanism, inward rectification, and Ca2+ selectivity. These properties are also shared by the Ca2+-sensing Ca2+ current (ITRPV6) recorded after heterologous expression of TRPV6 cDNA in human embryonic kidney and rat basophilic leukemia cells (Bödding, M., Wissenbach, U., Flockerzi, V. (2002) J. Biol. Chem. 277, 36656-36664). TRPV6 cDNA transfection of LNCaP cells restored recombinant ITRPV6, which can be distinguished from ISOC by the mechanism of activation, the voltage dependence of monovalent currents in the absence of external divalent cations, and the changes in Ca2+ current densities due to different membrane potentials. In addition, ISOC was not affected by antiandrogen or 1,25-dihydroxyvitamin D3 treatment of LNCaP cells, which up-regulates TRPV6 gene expression, or by androgen treatment, which has the opposite effect. Therefore, native channels responsible for ISOC are different from those for recombinant ITRPV6 and do not appear to be affected if one of their assumed subunits, TRPV6, is up- or down-regulated, suggesting a rather rigid subunit composition in vivo.

during receptor stimulation is regulated by both Ca 2ϩ release from intracellular inositol 1,4,5-trisphosphate (IP 3 )-sensitive stores and Ca 2ϩ entry across the plasma membrane. In many cases, the underlying plasma membrane ion channels require phospholipase C for activity, and stimulation of phospholipase C could be linked to channel gating via production of IP 3 and/or diacylglycerol. According to one mechanism referred to as store-operated Ca 2ϩ entry, transient release of Ca 2ϩ from internal stores induces sustained Ca 2ϩ influx. Of particular interest is a highly Ca 2ϩ -selective store-operated channel referred to as the Ca 2ϩ release-activated Ca 2ϩ (CRAC) channel, which has been implicated to be crucial in T-cell activation and mast cell functions (3,4). The molecular structure of this channel is still elusive. However, the mammalian transient receptor potential proteins (5), structurally related to the transient receptor potential cation channel underlying transduction in Drosophila melanogaster photoreceptors, have been proposed to participate in phospholipase C-and/or phosphoinositide-mediated Ca 2ϩ entry, including store-operated Ca 2ϩ influx (6,7).
Recently, a member of the vanilloid type transient receptor potential subfamily, TRPV6, was reported to display electrophysiological properties resembling those of the CRAC channel, such as high Ca 2ϩ selectivity and the current-voltage (I-V) signature under physiological conditions (8). However, there were also differences between both channels, including the protocols necessary to activate I CRAC in rat basophilic leukemia (RBL) cells and to activate recombinant I TRPV6 recorded after overexpressing its cDNA in RBL cells (1) or human embryonic kidney (HEK) cells (1,9). Both currents, I CRAC and I TRPV6 , were activated under conditions of high intracellular Ca 2ϩ buffering in TRPV6-transfected RBL cells, whereas only I CRAC was activated at low intracellular Ca 2ϩ buffering after Ca 2ϩ store depletion (1). It remains possible that TRPV6 channels may also be activated by procedures known to deplete cellular Ca 2ϩ stores and that differences between I TRPV6 and I CRAC are due to aberrant TRPV6 expression levels in the heterologous expression systems or may be a consequences of heteromultimerization between TRPV6 and other channel subunits.
The TRPV6 gene is primarily expressed in pancreatic acinar cells and placental syncytiotrophoblasts and trophoblasts (10) and in various intestinal epithelial cells (11). However, to our knowledge, no data are available demonstrating activation of endogenously expressed TRPV6 channels using the various protocols sufficient to activate recombinant I TRPV6 . RBL cells, which we used in our previous study (1), and the human Jurkat T-lymphocytes are common model systems to record I CRAC and therefore should contain the essential components necessary for channel activation by store-operated mechanisms. However, TRPV6 transcripts could not be detected in RBL cells by Northern blot analysis (1,8) or in Jurkat T-cells (12). We therefore performed this study on the androgen-sensitive human lymph node prostate cancer cell line LNCaP, which, in contrast to RBL cells, endogenously expresses the TRPV6 gene (see Fig. 1). Moreover, store-operated currents have been suggested to be present in these cells (13)(14)(15)(16)(17). Therefore, LNCaP cells are an appropriate system to study Ca 2ϩ entry currents with respect to store-operated and TRPV6 channels.
In this study, we describe a store-operated Ca 2ϩ current (I SOC ) in LNCaP cells that closely resembles I CRAC in RBL cells with respect to its activation by protocols known to deplete intracellular Ca 2ϩ stores, inward rectification, and high Ca 2ϩ selectivity. LNCaP cells endogenously express the TRPV6 gene; but unlike recombinant I TRPV6 in HEK or RBL cells, I SOC is not increased under elevated Ca 2ϩ buffering conditions, and the activation kinetics correlate with the expected time course of active or passive store depletion, which is not the case for recombinant I TRPV6 . TRPV6 cDNA transfection of LNCaP cells restores I TRPV6 with its typical responses due to changes in the membrane potential and [Ca 2ϩ ] i and in the absence of external divalent cations. Endogenous TRPV6 mRNA expression in LN-CaP cells is suppressed by androgen treatment (18) and is induced by a specific androgen receptor antagonist (18) and by 1,25-dihydroxyvitamin D 3 (19,20). However, I SOC was not affected by any of these treatments, indicating that TRPV6 alone at the low expression levels found in non-transfected LNCaP cells is not sufficient for SOC channel formation, whereas at high expression levels in transfected cells, it underlies I TRPV6 .

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-LNCaP cells (CRL-10995, lot 1216242) were from American Type Culture Collection (Manassas, VA) and from a second, independent source (ACC256, Deutsche Sammlung von Mikroorganismen und Zellkulturen, Braunschweig, Germany). Cells were cultured in RPMI 1640 medium with L-glutamine, 10% fetal calf serum, 50,000 IU/liter penicillin G (sodium), and 50 mg/liter streptomycin sulfate (Invitrogen, Karlsruhe, Germany). Cells were grown in 250-and 750-ml flasks (BD Biosciences) under 5% CO 2 at 37°C and plated onto glass coverslips in 35-mm diameter Petri dishes at least 24 h prior to electrophysiological experiments. One day after trypsinization, cells were transiently transfected with 4 g of DNA in 5 ml of PolyFect® reagent (QIAGEN, Hilden, Germany) as described (1). The bicistronic expression plasmid contained the entire protein-coding regions of human TRPV6b (DDBJ/GenBank TM /EBI accession number CAC20417), followed by an internal ribosomal entry site and the green fluorescent protein DNA (10). Patch-clamp experiments were performed 24 -36 h after transient transfection. Coverslips with cells were used for periods of up to 1 h to minimize osmotic stress.
Electrophysiological Recordings and Solutions-For experiments, coverslips with cells were transferred to the recording chamber and kept in a standard modified Ringer's solution of the following composition: 145 mM tetraethylammonium chloride, 10 mM CaCl 2 , 10 mM CsCl, 2.8 mM KCl, 2 mM MgCl 2 , and 10 mM HEPES adjusted to pH 7.2 with NaOH. In one set of recordings, the above solution was temporarily replaced with an otherwise identical external solution in which CaCl 2 was substituted with MgCl 2 . The divalent-free solution contained 135 mM NaCl, 20 mM tetraethylammonium chloride, 10 mM CsCl, 2.8 mM KCl, 5 mM EDTA, 5 mM EGTA, and 10 mM HEPES adjusted to pH 7.2 with NaOH. If necessary, the osmolarity was kept constant between internal and external solutions by the appropriate addition of mannitol to the bath. Solution changes were performed by pressure ejection from a wide tipped pipette. Patch-clamp experiments were conducted in the tight-seal whole-cell configuration (21) using an EPC-9 amplifier (HEKA Electronics, Lambrecht, Germany). Pipettes were pulled from borosilicate glass (Kimax®) and coated with Sigmacote®. Their resistances were between 2 and 3 megaohms when filled with the standard internal solution containing 145 mM cesium glutamate, 10 mM HEPES, 8 mM NaCl, 1 mM MgCl 2 and 2 mM MgATP adjusted to pH 7.2 with CsOH. Series resistances were typically in the range of 5-10 megaohms and were not electronically compensated. Currents were filtered using an 8-pole Bessel filter at 2.9 kHz and digitized at 100 s. Immediately following establishment of the whole-cell configuration, voltage ramps of 50-ms duration from Ϫ110 to 90 mV were delivered from a holding potential of Ϫ10 mV at a rate of 0.5 Hz using PULSE software (HEKA Electronics). For analysis, the current elicited by the first voltage ramp prior to current activation was subtracted for all subsequent current records unless stated otherwise. The temporal development of the in-ward currents was extracted from the leak-corrected individual ramp current traces by measuring the amplitude at Ϫ80 mV. I-V relationships were measured by the application of voltage steps from a holding potential of Ϫ10 mV to test potentials from Ϫ110 to 110 mV in 10-mV increments of 200-ms duration at 1 Hz. Currents were measured at the end of each pulse during the last 40 ms. Several parameters (capacitance, series resistance, and holding current) were displayed simultaneously at a slower rate (2 Hz) using the X-Chart display (HEKA Electronics). Membrane potentials were corrected for 10-mV liquid junction potential. All experiments were carried out at room temperature (20 -23°C); internal solutions were kept on ice to minimize ATP hydrolysis. Analysis was performed with PulseFit and programs written in the IGOR macro language (Wave Metrics, Lake Oswego, OR). Where applicable, data are given as means Ϯ S.E. for n number of cells.
Microfluorometry-Ester loading of intact cells was performed by incubation of cells in the culture medium at 37°C with 2 M fura-2 as an acetoxymethyl ester (fura-2/AM) for 30 min. Cells were then rinsed several times with a nominally Ca 2ϩ -free bath solution containing 135 mM NaCl, 20 mM tetraethylammonium chloride, 10 mM CsCl, 2.8 mM KCl, 2 mM MgCl 2 , and 10 mM HEPES adjusted to pH 7.2 with NaOH. The setup was an inverted microscope (Axiovert S100, Zeiss, Oberkochen, Germany) with an oil immersion objective (Fluar, ϫ40/1.3, Zeiss). A monochromatic light source (Polychrom II, TILL Phototonics, Planegg, Germany) was used for fluorescence excitation, and a CCD camera (SensioCam, PCO, Kelheim, Germany) was used to acquire fluorescence ratio images from up to 20 regions of interest simultaneously. The Ca 2ϩ indicator dye was excited alternately at 340 and 380 nm (10-ms exposures), and images were taken once every second at an emission wavelength of 510 nm. Calcium mobilization was induced by bath application of 2.5 M ionomycin and 5 mM Ca 2ϩ . Paired experiments were repeated three times by alternating Ca 2ϩ measurements in control and 1,25-dihydroxvitamin D 3 -treated cells.
Northern Blot Analysis-For Northern blot experiments, LNCaP cells were cultured in the presence of dihydrotestosterone (1 M) for 24 h, bicalutamide (1 M) for 48 h, 1,25-dihydroxyvitamin D 3 (1 M) for 24 h, and Me 2 SO (0.01%) for 48 h. LNCaP cells from three 750-ml (175-cm 2 ) flasks for each treatment were harvested by trypsinization. Cell extracts were prepared with a Potter-Elvehjem-type homogenizer, and RNA was purified using GOLD RNAPure TM (peqlab Biotechnologie GmbH, Erlangen, Germany). Total RNA was extracted with chloroform, precipitated with isopropyl alcohol, washed with 70% ethanol, and solved in diethyl pyrocarbonate-treated water. Isolation of mRNA was carried out using the poly(A) Spin TM mRNA isolation kit (New England Biolabs GmbH, Frankfurt, Germany). Poly(A) ϩ RNA (10 g/lane) was separated by electrophoresis on denaturing formamide/formaldehydecontaining 0.8% agarose gels and transferred to Hybond N membranes (Amersham Pharmacia Biotech Europe GmbH, Freiburg, Germany) by capillary transfer overnight. The RNA was fixed on the nylon membranes by UV light, followed by heating at 80°C for 2 h. The blots were prehybridized at 65°C for 1 h and a second time with fresh solutions at 42°C for 1 h in 50% (v/v) formamide, 50 mM Tris-HCl (pH 7.5), 0.1% (w/v) Na 4 P 2 O 7 , 1% SDS, 0.2% (w/v) polyvinylpyrrolidone 400,000, 0.2% (w/v) Ficoll 400, 5 mM EDTA (pH 8.0), 750 mM NaCl, 75 mM sodium citrate, and 0.15 mg/ml freshly denatured salmon sperm DNA. Hybridizations were performed at 42°C for 18 -20 h in the presence of the prehybridization solution including the radioactive probe at 3-20 ϫ 10 6 cpm/ml. The probe was the 571-bp PstI cDNA fragment of human TRPV6 encoding amino acid residues 40 -228 labeled by random priming with [␣-32 P]dCTP (Rediprime TM II random prime labeling system, Amersham Biosciences) and isolated from non-incorporated nucleotides by gel filtration (Sephadex G-50 Nick TM columns, Amersham Biosciences). After hybridization, the filters were repeatedly washed at increasing temperature (50 -60°C) using a solution of 0.1% SDS and 0.5ϫ SSC for the final wash step. Autoradiographs were exposed to Fujifilm Type BAS-MP imaging plates for 1-2 days and developed on a Fujifilm BAS-2500 phosphor imager (Fuji Photo Film, Kanagawa, Japan). Thereafter, filters were exposed for 7-20 days at Ϫ80°C to x-ray films using intensifying screens (see Fig. 1A). TRPV6 mRNA expression in the presence of 0.01% (v/v) Me 2 SO (control), dihydrotestosterone, bicalutamide, and 1,25-dihydroxyvitamin D 3 was normalized to ␤-actin expression using the Advanced Image Data Analyzer software (Raytest Isotopenme␤gerä te, Straubenhardt, Germany). Individual background subtraction was performed for each lane. Glyceraldehyde-3-phosphate dehydrogenase was not used as a positive control because its mRNA level is influenced by androgens and antiandrogens (22).
Immunoprecipitation and Immunoblotting-Microsomal proteins and cell lysates were prepared from HEK-293 cells (CRL-1573, American Type Culture Collection) and LNCaP cells, and immunoprecipita-tion was performed using 7.5 mg of microsomal membrane proteins prepared from LNCaP cells using antibody 429 (23). Immunoprecipitated proteins and microsomal proteins were separated by electrophoresis on an 8% SDS-polyacrylamide gel and transferred to nitrocellulose membranes (0.45 m) by tank-blotting (Bio-Rad, Mü nchen, Germany) in the presence of 20% methanol at 350 mA for 1.5 h. Blots were stained with Ponceau red and incubated in 5% nonfat dry milk in Tris-buffered saline at 21°C for 1 h before the primary antibody, monoclonal antibody 20C6 (23), was added (4°C for 12 h). Blots were washed three times with Tris-buffered saline and incubated for 1 h in the presence of horseradish peroxidase-labeled secondary antibody (anti-mouse IgG at 1:30,000; Dianova, Hamburg, Germany). Again the blot was washed three times and developed with the Renaissance Western blot chemiluminescence reagent (PerkinElmer Life Sciences, Rodgau-Jü gesheim, Germany).
Drugs-All chemicals for electrophysiological solutions were from Sigma, except BAPTA, which was purchased from Molecular Probes, Inc. Bicalutamide (Casodex®) was a gift from AstraZeneca (Cheshire, United Kingdom). Bicalutamide, dihydrotestosterone, and 1,25-dihydroxyvitamin D 3 stock solutions were prepared in Me 2 SO and 70% absolute ethanol, respectively.

RESULTS
LNCaP Cells Endogenously Express TRPV6 mRNA and Protein-The androgen-sensitive human lymph node prostate cancer cell line LNCaP expressed detectable levels of TRPV6 transcripts of ϳ3 kb (Fig. 1A), which have also be shown to be present in poly(A) ϩ RNA isolated from TRPV6-expressing human tissues such as placenta, pancreas, and prostate cancer tissue (10). The calculated relative molecular mass of human TRPV6 is 83,210 Da, which corresponds to the proteins recognized by antibodies in HEK-293 cells expressing TRPV6 (Fig.  1B, lane 2) and in LNCaP cells (lanes 3 and 4). The antibodies recognized TRPV6 proteins with molecular masses ranging from 69 to 83-98 kDa in TRPV6-expressing HEK-293 cells (Fig.  1B, lane 2) and of 87 and 98 kDa in LNCaP cells (lanes 3 and  4). These proteins were not detected in non-transfected HEK-293 cells (Fig. 1B, lane 1), which do not express TRPV6 (24). The various proteins recognized by the antibodies may represent differentially glycosylated TRPV6 proteins, as suggested previously (23,25).
Store Depletion Protocols Activate an Inward Current (I SOC ) in LNCaP Cells-In the following, we first investigated the effects of Ca 2ϩ store depletion protocols on membrane currents in LNCaP cells. Patch-clamp recordings were performed using experimental conditions that were originally described to measure I CRAC in RBL and Jurkat cells (3,4,26). Briefly, the driving force for Ca 2ϩ was magnified by increasing the external Ca 2ϩ concentration to 10 mM. The intracellular solution usually contained 10 mM EGTA to decrease the [Ca 2ϩ ] i and thereby prevent the activation of Ca 2ϩ -dependent channels. Contamination by K ϩ currents was eliminated by internal and external Cs ϩ and external tetraethylammonium. Osmolarities of the pipette and bath solutions were carefully adjusted and kept within the physiological range to minimize contributions of volume-sensitive channels.
Dialyzing IP 3 (30 M) and the sarcoplasmic/endoplasmic reticulum Ca 2ϩ -ATPase inhibitor thapsigargin (2 M) into the cell has been shown to deplete intracellular Ca 2ϩ stores rapidly (3). Shortly after establishing the whole-cell configuration, a small inward current activated ( ϭ 27 Ϯ 3 s, n ϭ 4; I max ϭ Ϫ1.8 Ϯ 0.4 pA/pF, n ϭ 5) (Fig. 2A). Similar results were obtained with either IP 3 ( ϭ 22 Ϯ 1 s, n ϭ 15; I max ϭ Ϫ2.4 Ϯ 0.1 pA/pF, n ϭ 18) (Fig. 2B) or thapsigargin ( ϭ 24 Ϯ 3 s, n ϭ 5; I max ϭ Ϫ2.0 Ϯ 0.1 pA/pF, n ϭ 10) (Fig. 2C) alone in the pipette solution. In the absence of IP 3 and thapsigargin, the same small inward current was activated, but with a much slower time course. After a delay of ϳ1 min, the current slowly developed and reached a similar peak current amplitude as in the recordings with IP 3 and/or thapsigargin (I max ϭ Ϫ1.8 Ϯ 0.2 pA/pF, n ϭ 3) (Fig. 2D). Because the cell was dialyzed by a solution containing 10 mM EGTA, it is likely that this chelator prevented re-uptake of the Ca 2ϩ that had leaked out back into the stores. Such store depletion is known to be considerably slower than depletion caused by inhibition of Ca 2ϩ re-uptake by thapsigargin or stimulation of Ca 2ϩ release by IP 3 and might explain the different activation kinetics of the inward current under the various conditions. Similar results were obtained with the Ca 2ϩ chelator BAPTA (10 mM; n ϭ 5) (data not shown) instead of EGTA.
Current-Voltage Relationship of I SOC -The current-voltage relationship of I SOC was the same in all experiments (Fig. 3): the current showed a strong inward rectification and reversed at positive potentials (Ͼ30 mV). During voltage steps, there was a biexponential current inactivation at potentials more negative than Ϫ50 mV, similar to the inactivation behavior of I CRAC in RBL cells (data not shown).
Ca 2ϩ -permeable Store-operated Channel in LNCaP-The Ca 2ϩ permeability of the store-operated channel in LNCaP cells was tested by replacing Ca 2ϩ with Mg 2ϩ in the standard bath solution. During application of this Ca 2ϩ -free solution, the inward current was abruptly abolished (Fig. 4). The block was reversible because the current could be restored in the presence of the Ca 2ϩ -containing bath solution. The loss of inward current in the absence of extracellular Ca 2ϩ and the positive reversal potentials in the presence of extracellular Ca 2ϩ indicate that the store-operated inward current is carried by Ca 2ϩ .
Activation of I SOC by Store Depletion at Low Intracellular Ca 2ϩ Buffering-It is possible to activate I CRAC under conditions of low intracellular Ca 2ϩ buffering if sarcoplasmic/endoplasmic reticulum Ca 2ϩ -ATPases are inhibited; under these conditions, I TRPV6 is not activated (1). In the presence of 2 M thapsigargin, 40 M IP 3 , and 0.5 mM EGTA (instead of the 10 mM EGTA used before), an inward current was activated in LNCaP cells ( ϭ 26 Ϯ 4 s, n ϭ 5; I max ϭ Ϫ3.1 Ϯ 0.3 pA/pF, n ϭ 6) (Fig. 5), which was identical to I SOC measured under conditions of high intracellular Ca 2ϩ before in terms of its I-V relation and Ca 2ϩ selectivity. The same current developed if the identical internal solution was used, but with only 0.1 mM EGTA instead of 0.5 mM (n ϭ 3). However, no current developed if the pipette solution contained only low concentrations of EGTA, but no store-depleting agents such as thapsigargin and IP 3 (n ϭ 3 for 0.1 mM and n ϭ 4 for 0.5 mM EGTA).

Recombinant I TRPV6 in LNCaP Cells Resembles, but Is Not
Identical to, I SOC -The data presented so far argue in favor of a Ca 2ϩ -selective store-operated current in LNCaP cells similar to I CRAC in RBL and Jurkat cells. However, unlike recombinant I TRPV6 in HEK and RBL cells, I SOC in LNCaP cells was not increased by raising the intracellular EGTA concentration to 30 mM (I max ϭ Ϫ2.1 Ϯ 0.3 pA/pF, n ϭ 6; compare with Fig. 2B). These data suggest that I SOC was already maximally activated if Ca 2ϩ stores were depleted under conditions of lower Ca 2ϩ buffering (Figs. 2 and 5). In addition, the activation kinetics of I SOC were slow when cells were dialyzed with EGTA or BAPTA at high concentrations in the absence of IP 3 and thapsigargin, which contrasts with the rapid activation of recombinant I TRPV6 . To examine whether TRPV6 might be regulated differentially in LNCaP cells or whether current activation depends on TRPV6 expression levels, we transfected LNCaP cells with the TRPV6 cDNA. Dialyzing these cells with 10 mM EGTA resulted in rapid activation of a large inward current (Fig. 6A) very similar to the recombinant I TRPV6 previously recorded in TRPV6-transfected HEK and RBL cells (1). Peak amplitudes were reached within 1 min of whole-cell recording; and subsequently, current densities decreased until a steady-state level was reached (Fig. 6A). Like I SOC , I TRPV6 is an inward rectifying current with a positive reversal potential (Fig. 6B). However, its activation precedes the depletion of intracellular Ca 2ϩ stores caused by intracellular perfusion with high concentra- tions of EGTA. In contrast, the activation of endogenous I SOC in LNCaP cells (Fig. 2D) and of I CRAC in RBL cells strongly depends on the kinetics of Ca 2ϩ release from intracellular stores.
Ca 2ϩ -selective channels like voltage-gated Ca 2ϩ channels (27) and CRAC channels (3) become permeable for monovalent ions if divalent cations are removed from the extracellular solution. Similar properties of recombinant I TRPV6 in HEK cells have been described (1,9,10,28). Under these conditions, I TRPV6 showed a typical voltage-dependent Mg 2ϩ block (9, 1), resulting in a negative slope at strong hyperpolarization, which was also observed in TRPV6-transfected LNCaP cells (Fig. 6D), but not in non-transfected LNCaP cells (n ϭ 5) (data not shown).
Modulation of I SOC and I TRPV6 by the Membrane Potential-Previously, it has been shown that increasing the holding potential augments recombinant I TRPV6 , most probably by reducing the driving force for Ca 2ϩ to enter the cell in between pulses and subsequently the Ca 2ϩ -induced negative feedback (10). However, these typical properties of recombinant I TRPV6 were not detected in non-transfected LNCaP cells using various activation protocols (10 mM EGTA, 30 M IP 3 , and 2 M thapsigargin (n ϭ 5); 0.5 mM EGTA, 30 M IP 3 , and 2 M thapsigargin, (n ϭ 6); and 10 mM EGTA and 30 M IP 3 (n ϭ 4)), whereas Ca 2ϩ currents in TRPV6-transfected LNCaP cells reversibly increased during membrane depolarization (Fig. 7). I SOC in Androgen-, Antiandrogen-, and 1,25-Dihydroxyvitamin D 3 -treated LNCaP Cells-So far, the data demonstrate that LNCaP cells contain store-operated channels similar to CRAC channels in RBL and Jurkat cells, whereas recombinant I TRPV6 can be recorded only if the TRPV6 cDNA is overexpressed in these cells. Apparently, TRPV6 at low expression levels might constitute a subunit of the store-operated channel intrinsic to LNCaP and contribute to I SOC . At high expression levels as obtained after TRPV6 cDNA transfection, TRPV6 protein levels may not be matched by protein levels of the other channel subunits and/or of molecules of the signal transduction mechanism that sense store depletion, and the resulting homomeric TRPV6 channel gives rise to I TRPV6 . Interestingly, it has been previously shown that androgens and antiandrogens (18) and 1,25-dihydroxyvitamin D 3 (19,20) 6). B, representative I-V relationship from a TRPV6transfected LNCaP cell shown with the corresponding voltage protocol. C, I-V signature of TRPV6-transfected LNCaP cells under divalent-free conditions. Voltage steps of 200-ms duration were applied every second from Ϫ110 to 110 mV from a holding potential of Ϫ10. Current amplitudes were measured as means during the last 20% of each pulse. Averaged data with double-sided S.E. are shown (n ϭ 12). D, representative high resolution Ca 2ϩ currents evoked by the same voltage protocol used in C. No background current subtraction was performed for the large monovalent currents in C and D. The dashed lines represent zero current and, in C, also a membrane potential of 0 mV. Norm., normalized.

DISCUSSION
In this study, we have demonstrated the endogenous expression of the TRPV6 gene in the human lymph node prostate cancer cell line LNCaP. We have characterized a store-operated inward current (I SOC ) in these cells, which closely resembles I CRAC with respect to its activation mechanism, inward rectification, and Ca 2ϩ selectivity. These biophysical properties are also typical for recombinant I TRPV6 . It was, however, possible to discriminate between I SOC and I TRPV6 in three independent ways; the activation mechanism, the behavior in divalent-free external solution, and the response due to changes in the membrane potential were different. In addition, cell culturing in the presence of androgen, antiandrogen, or 1,25-dihydroxyvitamin D 3 , each of which is known to change TRPV6 mRNA expression, did not affect I SOC with respect to its peak current amplitude and activation and inactivation kinetics.
I SOC in LNCaP cells could be activated by three independent protocols that lead to depletion of intracellular Ca 2ϩ stores. 1) Ca 2ϩ release from endogenous stores can be elicited using IP 3 in the pipette solution.
2) The sarcoplasmic/endoplasmic retic-ulum Ca 2ϩ -ATPase inhibitor thapsigargin leads to store depletion because Ca 2ϩ pumps are blocked and no longer able to replenish the pools. 3) Dialyzing high Ca 2ϩ chelator concentrations into the cell is believed to prevent the re-uptake of Ca 2ϩ that has leaked out from the stores. Subsequently, the Ca 2ϩ pools become empty, and store-operated Ca 2ϩ influx is activated. Depletion of Ca 2ϩ stores is thought to be rapid with IP 3 and thapsigargin, but slow with EGTA or BAPTA alone. The kinetics of I SOC activation in LNCaP cells correlate well with the presumed time courses of store depletion due to the various protocols.
It has previously been reported that the I-V relationship of recombinant I TRPV6 is similar to the prototype of the storeoperated current I CRAC (8,9,1). Despite these similar properties, it is possible to discriminate between both Ca 2ϩ currents. Whereas I CRAC is activated by store depletion, TRPV6 functions as a Ca 2ϩ -sensing Ca 2ϩ channel, and its current amplitude increases if the Ca 2ϩ concentration close to the intracellular mouth of the channel decreases (1). In contrast, I SOC in LNCaP cells is activated by applying store depletion protocols as described above, and its amplitude is not changed when store depletion occurs at low intracellular Ca 2ϩ buffering (0.5 or 0.1 mM EGTA in the presence of IP 3 and thapsigargin) or at high intracellular Ca 2ϩ buffering.
Besides the different activation mechanisms of I SOC and I TRPV6 , it is also possible to separate both currents by their I-V relationship in the absence of extracellular divalent cations. Under these conditions, monovalent currents through recombinant TRPV6 channels are characterized by a positive slope at very negative potentials in TRPV6-overexpressing cells. This behavior is not seen if intracellular Mg 2ϩ is omitted, indicating a voltage-dependent block and unblock of the TRPV6 channel by Mg 2ϩ , as shown previously by Voets et al. (9,28). However, no such voltage-dependent Mg 2ϩ block of I SOC was detected in non-transfected LNCaP cells.
Another typical property of TRPV6 channels is the modulation of Ca 2ϩ current amplitudes by the membrane potential (10). Depolarization resulted in I TRPV6 augmentation, whereas hyperpolarization induced a reduction of I TRPV6 . In contrast, changes in the holding potential did not change the amplitude of I SOC .
Previously, it has been shown that the levels of endogenously expressed TRPV6 can be changed by treating cells with androgens, antiandrogens, and 1,25-dihydroxyvitamin D 3 (18 -20). Changing expression levels of endogenous TRPV6 by hormones might be a more physiological procedure than overexpression of its cDNA. Accordingly, we found that TRPV6 transcript expression levels increased by ϳ1.3and 2-fold after treating LNCaP cells with bicalutamide and 1,25-dihydroxyvitamin D 3 , respectively, but were reduced by ϳ62-79% after incubation with dihydrotestosterone. However, no corresponding changes in I SOC were detectable. These data, together with the recordings demonstrating differences in the activation mechanism in the behavior under divalent-free conditions as well as the diverse properties due to changes in the membrane potential, suggest that the TRPV6 protein alone cannot form store-operated Ca 2ϩ channels in LNCaP cells. These results were confirmed using LNCaP cells from a second, independent source. As in RBL cells (1), overexpression of the TRPV6 cDNA resulted in the appearance of I TRPV6 rather than an augmentation of endogenous I SOC , indicating that under conditions of overexpressed protein, TRPV6 is a Ca 2ϩ -sensing Mg 2ϩ -gated channel.
It is, however, also possible to speculate that TRPV6 is one potential component of the protein complex forming store-operated channels. One could consider that TRPV6 is tightly coupled to its signal transduction machinery. Moreover, this FIG. 9. Store-dependent monovalent current in androgen-, antiandrogen-, and 1,25-dihydroxyvitamin D 3 -treated LNCaP cells. Cells were incubated, and patch-clamp experiments were performed as described for Fig. 8. The bath solution was divalent-free. Mean data with two-sided S.E. are shown for dihydrotestosteronetreated (n ϭ 8) (A), bicalutamide-treated (n ϭ 5) (B), and 1,25-dihydroxyvitamin D 3 -treated (n ϭ 7) (C) cells and control cells (n ϭ 9). Norm., normalized. coupling cannot be disturbed by overexpression of TRPV6 or by manipulation of its endogenous expression levels as shown in this study. In fact, TRPV6 has been suggested to be a component of store-operated channels in RBL cells (8,29), Jurkat T-lymphocytes (30), and LNCaP cells (16,17), although TRPV6 transcripts are not detectable by Northern blotting in RBL cells (1,8) and Jurkat cells (12). Apparently, the TRPV6-related functions elicited by physiological stimuli have to be identified in TRPV6-expressing primary cells like pancreatic acinar cells, syncytiotrophoblasts, and trophoblasts (10) and in various epithelia of the intestine (11).