Hydrolysis of Phosphatidylinositol 4,5-Bisphosphate Mediates Calcium-induced Inactivation of TRPV6 Channels*

TRPV6 is a member of the transient receptor potential superfamily of ion channels that facilitates Ca2+ absorption in the intestines. These channels display high selectivity for Ca2+, but in the absence of divalent cations they also conduct monovalent ions. TRPV6 channels have been shown to be inactivated by increased cytoplasmic Ca2+ concentrations. Here we studied the mechanism of this Ca2+-induced inactivation. Monovalent currents through TRPV6 substantially decreased after a 40-s application of Ca2+, but not Ba2+. We also show that Ca2+, but not Ba2+, influx via TRPV6 induces depletion of phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2 or PIP2) and the formation of inositol 1,4,5-trisphosphate. Dialysis of DiC8 PI(4,5)P2 through the patch pipette inhibited Ca2+-dependent inactivation of TRPV6 currents in whole-cell patch clamp experiments. PI(4,5)P2 also activated TRPV6 currents in excised patches. PI(4)P, the precursor of PI(4,5)P2, neither activated TRPV6 in excised patches nor had any effect on Ca2+-induced inactivation in whole-cell experiments. Conversion of PI(4,5)P2 to PI(4)P by a rapamycin-inducible PI(4,5)P2 5-phosphatase inhibited TRPV6 currents in whole-cell experiments. Inhibiting phosphatidylinositol 4 kinases with wortmannin decreased TRPV6 currents and Ca2+ entry into TRPV6-expressing cells. We propose that Ca2+ influx through TRPV6 activates phospholipase C and the resulting depletion of PI(4,5)P2 contributes to the inactivation of TRPV6.

TRPV6 is a member of the transient receptor potential superfamily of ion channels that facilitates Ca 2؉ absorption in the intestines. These channels display high selectivity for Ca 2؉ , but in the absence of divalent cations they also conduct monovalent ions. TRPV6 channels have been shown to be inactivated by increased cytoplasmic Ca 2؉ concentrations. Here we studied the mechanism of this Ca 2؉ -induced inactivation. Monovalent currents through TRPV6 substantially decreased after a 40-s application of Ca 2؉ , but not Ba 2؉ . We also show that Ca 2؉ , but not Ba 2؉ , influx via TRPV6 induces depletion of phosphatidylinositol 4,5-bisphosphate (PI(4,5)P 2 or PIP 2 ) and the formation of inositol 1,4,5-trisphosphate. Dialysis of DiC 8 PI(4,5)P 2 through the patch pipette inhibited Ca 2؉ -dependent inactivation of TRPV6 currents in whole-cell patch clamp experiments. PI(4,5)P 2 also activated TRPV6 currents in excised patches. PI(4)P, the precursor of PI(4,5)P 2 , neither activated TRPV6 in excised patches nor had any effect on Ca 2؉ -induced inactivation in whole-cell experiments. Conversion of PI(4,5)P 2 to PI(4)P by a rapamycin-inducible PI(4,5)P 2 5-phosphatase inhibited TRPV6 currents in whole-cell experiments. Inhibiting phosphatidylinositol 4 kinases with wortmannin decreased TRPV6 currents and Ca 2؉ entry into TRPV6-expressing cells. We propose that Ca 2؉ influx through TRPV6 activates phospholipase C and the resulting depletion of PI(4,5)P 2 contributes to the inactivation of TRPV6.
Calcium signaling orchestrates a myriad of physiological functions such as muscle contraction, neurotransmitter release, bone formation, and fertilization. Ca 2ϩ entry through plasma membrane ion channels regulates numerous physiological and pathophysiological processes. The essential role of transient receptor potential (TRP) 2 channel proteins in the reg-ulation of cellular Ca 2ϩ signaling has begun to be appreciated in the recent past (1)(2)(3). The mammalian TRP superfamily comprises six main subfamilies named the TRPC (canonical), TRPM (melastatin), TRPV (vanilloid), TRPP (polycystin), TRPML (mucolipin), and TRPA (ankyrin) groups. Among all TRP channels, TRPV5 and TRPV6, the members of the vanilloid subfamily, are the only ones that exhibit high Ca 2ϩ selectivity (4).
TRPV6 is expressed in Ca 2ϩ -transporting epithelial cells, and it plays an important role in the active Ca 2ϩ absorption by the intestines (5,6). TRPV6 channels have been reported to be expressed in a variety of other tissues such as kidney, prostate, stomach, brain, and lung (7) and have also been shown to be expressed aberrantly in human malignancies (8). Until now no electrophysiological study has been performed characterizing these channels in native cells that express them endogenously. The physiological importance of these channels is understood from studies with genetically modified mice. Mice lacking TRPV6 have been shown to exhibit reduced intestinal Ca 2ϩ reabsorption, increased urinary Ca 2ϩ excretion, decreased bone density, reduced fertility, and skin abnormalities (9).
Electrophysiological characterization of TRPV6 channels in heterologous expression systems reveals that they exhibit strong inward rectification and reverse at positive potentials (10). They exhibit high Ca 2ϩ selectivity and conduct monovalent cations in the absence of divalent cations. These monovalent currents through TRPV6 channels are much larger than those carried by Ca 2ϩ at physiological Ca 2ϩ concentrations (11,12). Ca 2ϩ that enters through TRPV6 or an increase in intracellular Ca 2ϩ has been reported to cause inactivation of these channels (12)(13)(14). This also inactivates monovalent currents upon subsequent removal of extracellular Ca 2ϩ (12). TRPV6 is also permeable to Ba 2ϩ , but Ba 2ϩ influx induces less inactivation than Ca 2ϩ or no inactivation at all depending on the conditions (12,14). Recovery from Ca 2ϩ -induced inactivation is quite slow (12), and it was shown for the closely related TRPV5 that this recovery lags significantly behind restoration of cytoplasmic Ca 2ϩ levels (15). TRPV6 channels were proposed to function as Ca 2ϩ sensors, i.e. at low cytoplasmic [Ca 2ϩ ] they open and let more Ca 2ϩ in, and at high [Ca 2ϩ ] they close and reduce further Ca 2ϩ entry (10).
In this study we examined the role of phosphoinositides in the Ca 2ϩ -induced inactivation of TRPV6. We demonstrate that the activity of TRPV6 depends on PI(4,5)P 2 , using different approaches including direct activation by PI(4,5)P 2 in excised patches, dialyzing DiC 8 PI(4,5)P 2 via the patch pipette in wholecell recordings, and dephosphorylating plasma membrane PI(4,5)P 2 using a rapamycin-inducible PI(4,5)P 2 5-phosphatase. We also show that Ca 2ϩ flowing through TRPV6 activates phospholipase C (PLC), which leads to the depletion of PI(4,5)P 2 . Taken together, we provide evidence for a model that envisages the activation of PLC by Ca 2ϩ , which results in the hydrolysis of PI(4,5)P 2 , causing inactivation of TRPV6 channels.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-HEK293 cells were maintained in minimal essential medium supplemented with 10% fetal bovine serum and penicillin/streptomycin. The human TRPV6 tagged with the Myc epitope on the N terminus, subcloned into the expression vector pCMV-Tag3A (Stratagene), was used for the experiments (31), and cells were transfected using the Effectene reagent. For the intracellular Ca 2ϩ imaging and electrophysiology experiments, transfection was confirmed by measuring the fluorescence of co-transfected GFP. For experiments with rapamycin, the cells were transfected with the Myctagged TRPV6, the plasma membrane-targeted, CFP-tagged FRB, and the RFP-tagged FKBP12 linked to the phosphatase domain of PIP 2 5-phosphatase (32). For control experiments, RFP-FKBP12phosphatase domain was replaced with the RFP-tagged FKBP12 without the 5-phosphatase domain.
Mammalian Electrophysiology-Whole-cell patch clamp measurements were performed using a continuous holding protocol at Ϫ60 mV. Recordings were performed 36 -72 h posttransfection in HEK293 cells using a bath solution containing, in mM, 137 NaCl, 5 KCl, 10 glucose, 10 HEPES, pH adjusted to 7.4 (designated as nominally divalent free, NDF). The same solution was used for the fluorescence measurements and Ca 2ϩ imaging. Borosilicate glass pipettes (World Precision Instruments) of 2-4-megaohm resistance were filled with a solution containing, in mM, 135 potassium-gluconate, 5 KCl, 5 EGTA, 1 MgCl 2 , 2 ATP disodium, 10 HEPES, pH adjusted to 7.2. The cells were kept in NDF solution for 20 min before measurements. After formation of gigaohm resistance seals, whole-cell configuration was established and currents were measured using an Axopatch 200B amplifier (Axon Instruments). Data were collected and analyzed with pCLAMP 9.0 software. All measurements were performed at room temperature, 20 -25°C. For experiments with wortmannin (WMN), cells expressing TRPV6 were preincubated with NDF containing 10 M WMN for 30 min at room temperature. Control cells were treated with vehicle (DMSO) in NDF. All conditions and solutions remain the same for patch clamp, Ca 2ϩ imaging, and fluorescence resonance energy transfer (FRET) experiments unless otherwise specified.
Excised Patch Measurements in Xenopus Oocytes-Macropatch experiments were performed with borosilicate glass pipettes (World Precision Instruments) of 0.8 -1.7-megaohm resistance. After establishing gigaohm resistance seals on devitellinized surfaces of Xenopus oocytes, inside-out configuration was established, and currents were measured using an Axopatch 200B amplifier (Axon Instruments). The pipette solution contained, in mM, 96 LiCl, 1 EGTA, and 5 HEPES, pH 7.4. For the measurements in Fig. 3, the perfusion solution contained, in mM, 96 KCl, 5 EGTA, 10 HEPES, pH adjusted to 7.4. For the measurements shown in Fig. 7, the perfusion solution contained, in mM, 93 potassium gluconate, 5 HEDTA, 5 HEPES, pH 7.4. The free Ca 2ϩ and Mg 2ϩ concentrations were calculated using the MaxChelator program. To result in 10 M free calcium, 3.85 mM Ca 2ϩ was added (calcium-gluconate) and to result in 43 M free Mg 2ϩ , we added 2.35 mM Mg 2ϩ (magnesium-gluconate). For these measurements the bath was connected with the ground electrode through an agar bridge. Data were analyzed with pCLAMP 9.0 software (Axon Instruments) and plotted using Microcal Origin.
FRET Measurements-HEK293 cells were transfected with the CFP-and YFP-tagged PH domains of PLC␦1 and TRPV6. Measurements were performed using a Photon Technology International (PTI) (Birmingham, NJ) photomultiplier-based system mounted on an Olympus IX71 microscope, equipped with a DeltaRAM excitation light source. For the FRET measurements, excitation wavelength was 425 nm and emission was detected parallel at 480 and 535 nm using two interference filters and a dichroic mirror to separate the two emission wavelengths. Data were collected using the Felix software (PTI), and the ratio of traces obtained at the two different wavelengths correlating with FRET were plotted (33). Measurements were performed at room temperature (20 -25°C).
Inositol Phosphate Turnover-Measurements were performed similarly to that described in Ref. 34. HEK293 cells were transfected with either TRPV6-Myc and GFP or with GFP alone (controls) and incubated with 20 Ci of [ 3 H]myo-inositol overnight in growth medium. Before the experiments the cells were kept in NDF for 20 min and for an additional 10 min in NDF containing 10 mM LiCl. Then the cells were treated with NDF containing no Ca 2ϩ or 2 mM Ca 2ϩ or 2 mM Ba 2ϩ for 25-30 min in the continued presence of LiCl. The cells were scraped, treated with 4% perchloric acid, and centrifuged at 12000 rpm for 2-4 min. The supernatants (1.2 ml) were transferred into glass tubes containing 180 l of 10 mM EDTA, pH 7.0, and to each tube 1.3 ml of a freshly prepared mixture of trioctylamine/ Freon was added, vortexed, and centrifuged at 12000 rpm for 4 min. The top aqueous layer (ϳ1.2 ml) was transferred into plastic vials and 3.6 ml of sodium bicarbonate was added. This solution was then added to Dowex columns filled with AG1 ϫ 8 resin (formate form). The columns were washed four times each with 5 ml of distilled water. Fractions 1-4 (5 ml each) 0.4 M ammonium formate/0.1 M formic acid, pH 4.75 (IP 2 fraction); fractions 5-8 (5 ml each) 0.7 M ammonium formate/0.1 M for-mic acid, pH 4.75 (IP 3 fraction); and fractions 9 -12 (5 ml each) 1.2 M ammonium formate/0.1 M formic acid, pH 4.75 (IP 4 , IP 5 , and IP 6 fraction) were collected. One ml of each of the collected fractions was transferred into a scintillation vial, 10 ml of scintillation mixture was added, and 3 H activity was determined in a scintillation counter.
Ca 2ϩ Imaging-Cells transfected with TRPV6 and GFP (marker for cell selection) were grown on 25-mm circular coverslips and loaded with fura-2 AM (2 M) for 30 -40 min at room temperature in NDF. The coverslips were then washed in NDF, placed in a stainless steel holder (bath volume, ϳ0.8 ml; Molecular Probes), and viewed in a Zeiss Axiovert 100 microscope coupled to an Attofluor digital imaging system. Cells expressing GFP were selected and monitored simultaneously on each coverslip. Results are presented as the ratio (R) of fluorescence intensities at excitation wavelengths of 334 and 380 nm. Cells were continuously superfused with NDF, and Ca 2ϩ entry was initiated by addition of a solution containing 2 mM CaCl 2 . All experiments were performed at room temperature.
Confocal Microscopy-HEK293 cells were transfected with TRPV6 and the GFP-tagged PLC␦1 PH domain (supplemental Fig. S1) or the components of the PI(4,5)P 2 phosphatase recruitment system (supplemental Fig. S3). Experiments were performed 2 days after transfection. The cells were observed with a Zeiss LSM-510 confocal microscope in the Confocal Imaging Facility of the New Jersey Medical School. Images were saved as TIFF files and were analyzed with IMAGE J software.
Materials-Fura-2 AM was obtained from TefLabs (Austin, Texas). Effectene was obtained from Qiagen. Cell culture media, antibiotics, and sera were obtained from Invitrogen. DiC 8 phosphoinositides were obtained from Cayman and Echelon. All other chemicals were purchased from Sigma.
Data Analysis-Data for all figures were expressed as mean Ϯ S.E. Statistical significance was evaluated by Student's t test. * stands for p Ͻ 0.05, and ** for p Ͻ 0.001.

RESULTS
Ca 2ϩ , but Not Ba 2ϩ , Inactivates Na ϩ Currents through TRPV6-We studied the mechanism of Ca 2ϩ -induced inactivation of TRPV6 by measuring monovalent currents before and after exposing the TRPV6-expressing cells to Ca 2ϩ (or Ba 2ϩ ) -containing solutions in whole-cell configuration at a constant holding potential of Ϫ60 mV (Fig. 1). Monovalent currents were initiated by the application of a solution containing 2 mM EGTA and no added divalent ions. After a 40-s application of Ca 2ϩ , but not Ba 2ϩ , monovalent currents were markedly decreased despite the substantial entry of Ba 2ϩ observed in fluorescence measurements (see Fig. 2D). The average current amplitude during the first pulse of 0 Ca 2ϩ (EGTA) was 2.27 Ϯ 0.65 nA, which was significantly higher than the second (0.86 Ϯ 0.38 nA) and third (0.69 Ϯ 0.27 nA) pulses. When Ba 2ϩ was used instead of Ca 2ϩ , the average current amplitude for the first pulse was 1.87 Ϯ 0.48 nA. The average current amplitudes for the second and third pulses were 1.83 Ϯ 0.34 and 1.95 Ϯ 0.32 nA, respectively. The same protocol did not induce any current in cells transfected with GFP ( Fig. 1, C and F) or non-transfected HEK cells (data not shown). Monovalent currents through TRPV6 are also blocked by extracellular Mg 2ϩ (11); thus we omitted Mg 2ϩ from the extracellular NDF solution throughout the experiments.
Ca 2ϩ , but Not Ba 2ϩ , Influx through TRPV6 Channels Reduces PI(4,5)P 2 Levels-Ca 2ϩ influx through TRPM8 channels has been suggested to induce PI(4,5)P 2 depletion via PLC activation (23). To explore the mechanism of the differential behavior of Ca 2ϩ and Ba 2ϩ on the inactivation of TRPV6 channels, we used a FRET-based technique (23,33) to show that Ca 2ϩ , but not Ba 2ϩ , induces depletion of PI(4,5)P 2 (Fig. 2, A and B). This technique is based on the translocation of the CFP/YFP-tagged PLC␦1 PH domains from the plasma membrane to the cytoplasm upon PIP(4,5) 2 depletion, which is shown in the figure as downward deflection of the FRET ratio traces. This technique has been shown to display good correlation with translocation of the GFP-tagged PLC␦1 PH domain as measured with confocal microscopy (33). We also show that Ca 2ϩ induces translocation of the GFP-tagged PLC␦1 PH domain using confocal microscopy (supplemental Fig. S1). Fig. 2C summarizes the percentage of change in FRET ratio caused by addition of 2 mM Ca 2ϩ or Ba 2ϩ . Fig. 2D shows that addition of 2 mM Ca 2ϩ or Ba 2ϩ resulted in similar change in the fluorescence ratio of fura-2- loaded TRPV6 cells, indicating that both Ca 2ϩ and Ba 2ϩ enter through these channels.
We have also measured IP 3 production in HEK cells in response to Ca 2ϩ and Ba 2ϩ influx through TRPV6. Fig. 2E shows that IP 3 production is increased in HEK cells expressing TRPV6 in response to the application of Ca 2ϩ , but not Ba 2ϩ . HEK cells not expressing TRPV6 did not respond with increased formation of IP 3 to the application of Ca 2ϩ (Fig. 2F), but they responded to extracellular ATP, which activates PLC in these cells via P2Y cell surface receptors (35). Supplemental Fig. S2 shows that application of extracellular Ca 2ϩ in HEK293 cells not expressing TRPV6 did not significantly increase cytoplasmic Ca 2ϩ levels and did not induce any changes in FRET, demonstrating that Ca 2ϩ entry and Ca 2ϩ -induced PI(4,5)P 2 hydrolysis depend on the presence of TRPV6 channels.
In whole-cell patch clamp experiments, dialysis of DiC 8 PI(4,5)P 2 via the patch pipette relieved TRPV6 currents from Ca 2ϩ -induced inactivation (Fig. 4C), whereas DiC 8 PI(4)P had no effect (Fig. 4B). Control cells (Fig. 4, A and D)   . PI(4,5)P 2 activates TRPV6 in excised patches. Currents were measured in large membrane patches excised from Xenopus oocytes expressing TRPV6 using the ramp protocol from Ϫ100 to ϩ100 mV applied every second (0.25 mV/ms). Representative traces show currents at ϩ100 and Ϫ100 mV, upper and lower traces, respectively. Phosphoinositides were applied directly to the intracellular surface of the patch, as indicated by the horizontal bars. A, representative trace describing the effects of various phosphoinositides. To facilitate rundown, 30 g/ml poly-lysine was added before the application of the phosphoinositides in two of six experiments; in the remaining experiments phosphoinositides were applied after spontaneous run down of the currents. The effects of phosphoinositides were indistinguishable in polylysine-treated and untreated patches, and thus the results were pooled. B illustrates statistics for Ϫ100 mV (n ϭ 6) of the effects of different phosphoinositides compared with PI(4,5)P 2 . The average current Ϯ S.E. for various phosphoinositides was normalized to PI(4,5)P 2 responses. All phosphoinositides were short acyl chain (DiC 8 ) applied at a concentration of 50 M. monovalent currents in cells dialyzed with DiC 8 PI(4)P (Fig. 4, B and E) were 1.65 Ϯ 0.35, 0.48 Ϯ 0.16, and 0.34 Ϯ 0.1 nA for the first, second, and third pulses, respectively. In cells dialyzed with DiC 8 PI(4,5)P 2 (Fig. 4, C and F), the current amplitude for the first pulse was 1.1 Ϯ 0.14, for the second pulse it was 1.18 Ϯ 0.17, and it was 1.16 Ϯ 0.18 nA for the third pulse.
WMN Inhibits TRPV6 Channels-WMN at high concentrations inhibits some isoforms of phosphatidylinositol 4-kinase (38). WMN was reported to deplete PI(4,5)P 2 in intact cells and inhibit PI(4,5)P 2 -sensitive ion channels (39,40). We found that preincubation with 10 M WMN for 30 min significantly inhibited TRPV6 currents compared with untreated controls (Fig. 6, A-C). We measured the currents at two different time points to compare the peak and the sustained TRPV6 currents in WMN-treated cells and vehicle-treated controls. In control cells (Fig. 6A) To further confirm the effect of WMN, we measured intracellular Ca 2ϩ in cells expressing TRPV6. Treatment of cells with 10 M WMN for 30 min inhibited Ca 2ϩ entry significantly compared with untreated controls (Fig. 6D). Fig. 6E summarizes the change in fluorescence ratio in control and WMNtreated cells measured at two different time points of 30 and 150 s after the addition of Ca 2ϩ .
Protein Kinase C Activation Does Not Affect TRPV6 Channel Activity-Protein kinase C is also generally activated upon PLC activation, and Ca 2ϩ -dependent protein kinase C activation has  been suggested to mediate menthol-induced desensitization of TRPM8 (41,42). To test the effect of protein kinase C, we examined the effect of 1-oleoyl-2-acetyl-sn-glycerol, a cell-permeable diacylglycerol analogue. 1-Oleoyl-2-acetyl-sn-glycerol (100 M) failed to inhibit monovalent currents through TRPV6 measured at Ϫ60 mV, and it also failed to affect Ca 2ϩ signals in TRPV6-expressing cells (data not shown). Direct Application of Ca 2ϩ to Excised Patches Has Only a Negligible Effect on TRPV6-Finally, we tested the effects of direct application of Ca 2ϩ (10 M) on TRPV6 in excised patches. These measurements were performed in Xenopus oocytes, which endogenously express a Ca 2ϩ -activated Cl Ϫ current. We could not fully inhibit these channels with niflumic acid or flufenamic acid (43) at 300 M. Higher concentrations of these agents inhibited TRPV6 currents (data not shown).
Thus, we circumvented this problem by detecting TRPV6 currents at the reversal potential of the chloride current the following way. For the bath solution (cytoplasmic) we used a gluconate-based Cl Ϫ -free solution, and the pipette solution (extracellular) contained Cl Ϫ as the main anion (see "Experimental Procedures" for details). The Ca 2ϩ -activated Cl Ϫ channels have a small but detectable permeability to gluconate (44); thus, we found that the reversal potential of the Cl Ϫ (gluconate) currents under these conditions was Ϫ103 mV. At this potential we could measure TRPV6 currents, whereas we could monitor the chloride currents at ϩ100 mV (Fig. 7). We applied 10 M free Ca 2ϩ (buffered with HEDTA) shortly after excision, where it induced only a negligible inhibition of TRPV6 currents (Fig. 7A). As the currents under these conditions exhibited a variable level of rundown (see also Fig. 3), we also applied Ca 2ϩ after the channels were re-activated with 50 M diC 8 PI(4,5)P 2 (Fig. 7). Under these conditions 10 M Ca 2ϩ slightly potentiated TRPV6 currents, but this effect was variable and not statistically significant (p ϭ 0.094, n ϭ 8).
To induce uniform rundown of TRPV6 currents in all experiments, we applied Mg 2ϩ (43 M free Mg 2ϩ ) to the excised patches before reactivating TRPV6 with PI(4,5)P 2 . Mg 2ϩ serves as a cofactor for lipid phosphatases and thus promotes depletion of PI(4,5)P 2 (45) that leads to current rundown. Mg 2ϩ also has a direct inhibitory effect on TRPV6 (11), which is mainly prevalent at positive voltages at this concentration; note the fast inhibition of the outward currents in Fig. 7. After the washout of diC 8 PI(4,5)P 2 when TRPV6 currents completely disappeared, we applied a third pulse of Ca 2ϩ in each measurement to confirm the absence of Cl Ϫ current at Ϫ103 mV. We conclude that direct binding of Ca 2ϩ to the cytosolic surface of TRPV6 is unlikely to significantly contribute to the marked Ca 2ϩ -induced inactivation we observe in whole-cell patch clamp measurements.

DISCUSSION
Ca 2ϩ -induced Inactivation of TRPV6-TRPV6 is a constitutively active Ca 2ϩ -selective channel that mediates Ca 2ϩ uptake through the apical membrane of epithelial cells (4). When Ca 2ϩ enters the cells through these channels, they inactivate, which is mediated by an increase in cytoplasmic [Ca 2ϩ ]. This Ca 2ϩ -induced inactivation may play a role as a feedback loop to regulate cytoplasmic Ca 2ϩ levels, preventing Ca 2ϩ overload through these channels (10). This Ca 2ϩ -induced inactivation has been shown to consist of a fast and a slower component when studied by activating the channels with fast voltage steps to negative membrane potentials (12)(13)(14). Recovery from this Ca 2ϩ -induced inactivation is quite slow (12), suggesting either a process with a very slow off rate or a need for the resynthesis of a cofactor that is lost during the inactivation process.
Most earlier studies examined Ca 2ϩ -induced inactivation at relatively high extracellular Ca 2ϩ concentrations (Ն10 mM) and on a relatively short time scale (1-2 s), and Ca 2ϩ entry was initiated by a short voltage pulse to negative membrane potentials (12,13). Our study focused on the effects of steady-state Ca 2ϩ entry at a constant holding potential on a longer time scale (min), as this presumably resembles the native conditions of these channels in epithelial cells. In patch clamp experiments we measured monovalent currents through TRPV6 because these are much easier to detect than the much smaller Ca 2ϩ currents in physiological extracellular Ca 2ϩ concentrations. We show that under these conditions TRPV6 channels undergo inactivation when conducting Ca 2ϩ , but not Ba 2ϩ , which is consistent with earlier findings using different protocols (14). We show that PI(4,5)P 2 is depleted in response to application of extracellular Ca 2ϩ , but not Ba 2ϩ , in cells expressing TRPV6. We also show that Ca 2ϩ , but not Ba 2ϩ , stimulates the formation of IP 3 in TRPV6-expressing cells; thus, it is likely that the mechanism of PI(4,5)P 2 depletion is the Ca 2ϩ -induced activation of PLC. The lack of effect of Ba 2ϩ on PI(4,5)P 2 depletion and IP 3 formation is consistent with earlier findings showing that PLC activation, as measured by the sustained phase of IP 3 production, was significantly reduced in angiotensin-2-stimulated adrenal glomerulosa cells when Ca 2ϩ was replaced with Ba 2ϩ in the extracellular medium (46).
The Role of Phosphoinositides in the Regulation of TRPV6-The activation of PLC leads to a multitude of events, such as formation of IP 3 and other soluble inositol phosphates, Ca 2ϩ release from intracellular stores, the activation of protein kinase C, and reduction of plasma membrane PI(4,5)P 2 levels. Many TRP channels need PI(4,5)P 2 for activity; thus, PI(4,5)P 2 depletion is an attractive candidate to mediate the inactivation of TRPV6. We have shown here that dialyzing PI(4,5)P 2 through the patch pipette essentially eliminated the Ca 2ϩ -induced inactivation of TRPV6. This is compatible with the inactivation being mediated by PI(4,5)P 2 depletion, and it is incompatible with the role of all other candidates, because supplying more substrate for PLC would presumably increase the formation of all the other messengers. Our negative control was PI(4)P, which, unlike PI(4,5)P 2 , did not activate TRPV6 in excised patches and did not inhibit Ca 2ϩ -induced inactivation of TRPV6. We have also shown that activation of protein kinase C with the diacylglycerol analogue 1-oleoyl-2-acetyl-sn-glycerol did not inhibit TRPV6, confirming that Ca 2ϩ -induced inactivation is not mediated by protein kinase C. We also did not detect substantial inhibition by Ca 2ϩ in excised patches; thus, it is unlikely that direct binding of Ca 2ϩ to cytoplasmic parts of the channel significantly contributes to Ca 2ϩ -induced inactivation.
We have shown that PI(4,5)P 2 depletion is necessary for Ca 2ϩ -induced inactivation of TRPV6, but is it sufficient to inhibit these channels? To test this, we have utilized two different tools to decrease membrane PI(4,5)P 2 levels without activating PLC and thus not forming IP 3 and diacylglycerol. First we used the recently described rapamycin-inducible PI(4,5)P 2 5-phosphatase recruitment system to selectively deplete PI(4,5)P 2 by converting it to PI(4)P (32). Rapamycin-induced PI(4,5)P 2 depletion inhibited TRPV6 currents, which is compatible with the role of PI(4,5)P 2 keeping this channel open and the lack of ability of PI(4)P to activate it. To confirm these data, we have also utilized wortmannin at concentrations where it inhibits phosphoinositol 4 kinases thus inhibiting the supply of the precursor of PI(4,5)P 2 , leading to slow depletion of PI(4,5)P 2 . Wortmannin inhibited both TRPV6 currents and Ca 2ϩ signals in TRPV6-expressing cells. These data together demonstrate that depletion of PI(4,5)P 2 is sufficient to inhibit TRPV6.
PI(3,4)P 2 and PI(3,4,5)P 3 , the products of phosphatidyl inositol 3 kinase, also activated TRPV6 in excised patches even though they were less effective than PI(4,5)P 2 . These lipids are thought to be at lower concentrations in the plasma membrane than PI(4,5)P 2 (47), and thus their effects are probably overridden by the latter.
In summary, our data demonstrate that TRPV6 channels require PI(4,5)P 2 for activity and that the hydrolysis of this lipid by Ca 2ϩ -induced activation of PLC contributes to inactivation of this channel. This mechanism may serve as a feedback loop for the regulation of TRPV6, allowing this channel to function as a Ca 2ϩ sensor and thus regulate cytoplasmic Ca 2ϩ levels.
Acknowledgments-We thank Dr. J. P. Reeves for insightful comments on the manuscript and for kind permission to use the Ca 2ϩ -imaging setup in the Reeves laboratory. We also thank Dr. A. P. Thomas   Holding potential was 0 mV, and then a 150-ms long step to Ϫ103 mV, followed by a step to ϩ100 mV, was applied once every second. The changes at ϩ100 mV mainly correspond to the Ca 2ϩactivated Cl Ϫ currents (I Cl Ϫ), whereas the changes at Ϫ103 mV correspond exclusively to TRPV6 currents. Note the lack of any current in response to Ca 2ϩ at Ϫ103 mV after TRPV6 current rundown at the end of the experiment (third pulse of Ca 2ϩ ). B, normalized current changes induced by Ca 2ϩ at Ϫ103 mV (mean Ϯ S.E., n ϭ 8).