Voltage-dependent Changes of TRPV6-mediated Ca2+ Currents*

The physiological role and activation mechanism for most proteins of the transient receptor potential (TRP) family are unknown. This is also the case for the highly Ca2+ selective transient receptor potential vanilloid type 6 (TRPV6) channel. Patch clamp experiments were performed on transiently transfected human embryonic kidney (HEK) cells to address this issue. Currents were recorded under various conditions of intracellular Ca2+ buffering and monitored at the same voltage throughout. No TRPV6-mediated Ca2+ entry was detected under in vivo Ca2+ buffering conditions at a slightly negative holding potential; however, moderate depolarization resulted in current activation. Very similar results were obtained with different Ca2+ chelators, either EGTA or BAPTA dialyzing the cell. TRPV6 channel activity showed a negative correlation with the intracellular free Ca2+ concentration ([Ca2+]i) and was modulated by the membrane potential: Hyperpolarization decreases and depolarization increases TRPV6-mediated currents. Monovalent ions permeated TRPV6 channels in the absence of extracellular divalent cations. These currents were resistant to changes in the holding potential while the negative correlation to the [Ca2+]i was conserved, indicating that the voltage-dependent current changes depend on blocking and unblocking the charge carrier Ca2+ within the pore. In summary, these results suggest that the voltage dependence of TRPV6-mediated Ca2+ influx is of physiological importance since it occurs at cytosolic Ca2+ buffering and takes place within a physiologically relevant membrane potential range.

exclusive Ca 2ϩ selectivity and are thought to be responsible for Ca 2ϩ uptake in the kidney and intestine, respectively (7,8). The TRPV6 protein functions as a Ca 2ϩ -sensing Ca 2ϩ pore in HEK and RBL cells and its current amplitude is inversely correlated with the [Ca 2ϩ ] i (9). However, the calculated [Ca 2ϩ ] i necessary for channel activity are several magnitudes below physiological values (9). Furthermore, even high concentrations of fast exogenous Ca 2ϩ chelators are unable to control the microdomain in the close proximity of the channel (9). Thus, the physiological activation mechanism of TRPV6 is not clear because no current was detected under in vivo Ca 2ϩ buffering conditions (10).
In the present study the role of the membrane potential in modulating TRPV6-mediated Ca 2ϩ currents was studied in TRPV6-transfected HEK cells. Moderate changes in the holding potential within the physiological range induced current augmentation at depolarization and current inhibition at hyperpolarization under in vivo conditions of intracellular Ca 2ϩ buffering. This effect is caused by Ca 2ϩ itself, which most likely binds to the TRPV6 protein in a voltage-dependent manner.

EXPERIMENTAL PROCEDURES
Cell Culture, Transfected cDNA, and Transfection-HEK-293 (ATCC, 1573-CRL) cells were from the American Type Culture Collection (Manassas, VA). Cell culture was done as described previously (9). Cells were transiently transfected with 4 g of DNA in 5 ml of the PolyFect® reagents (Qiagen, Hilden, Germany). The bicistronic expression plasmid pdiCaT-Lb was constructed as described (11) and contained the entire protein-coding regions of the b-variant of human TRPV6 (formerly CaT-Lb, DDBJ/EMBL/GenBank TM , accession number CAC20417) followed by an internal ribosomal entry side and the green fluorescence protein DNA.
Site-directed Mutagenesis-Mutagenesis of single amino acids in the presumable core region were carried out using the QuikChange TM sitedirected mutagenesis kit (Stratagene, La Jolla, CA) as reported previously for the TRPV6 D542A point mutant (10). For the double mutant two complementary oligonucleotides introduced alanine residues at positions 542 and 550 (sense primer, 5Ј-C CAT CAT CGC TGG CCC AGC CAA CTA CAA CGT GGC CCT GCC CTT C-3Ј and antisense primer, 5Ј-G AAG GGC AGG GCC ACG TTG TAG TTG GCT GGG CCA GCG ATG ATG G-3Ј). The mutated TRPV6 PCR products were excised from the pcDNA3 vector (Invitrogen, Karlsruhe, Germany) and subcloned into the pCAGGS-IRES-GFP vector. The nucleotide sequence of the complete inserts including the Kozak sequence and the cloning sites were verified by sequencing the corresponding DNAs on both strands. The truncated construct TRPV6⌬ 693-725 and the TRPV6 T702A mutant have been described previously (12).
Electrophysiological Recordings and Solutions-Patch clamp experiments were performed in the whole cell configuration (13) using an EPC-9 amplifier (HEKA Elektronik, Lambrecht, Germany). Cells were measured 24 -32 h after transfection in a modified Ringer's solution containing (in mM): 145 NaCl, 10 CaCl 2 , 10 CsCl, 2.8 KCl, 2 MgCl 2 , 11 glucose, 10 HEPES, adjusted to pH 7.2 with NaOH. The divalent-free (DVF) solution contained (in mM): 145 NaCl, 2.8 KCl, 10 CsCl, 11 glucose, 10 EGTA, 10 HEPES, adjusted to pH 7.2 with NaOH. The external solution for the optical recordings contained either 2 mM Ca 2ϩ or 1 mM EGTA. Extracellular solution changes were made by pressure ejection (ϳ5 cm H 2 O) from a wide-tipped pipette positioned about 10 m from the cell. Patch pipettes pulled from borosilicate glass (Kimax®) had resistances between 2 and 3 M⍀ when filled with the standard internal solution. This solution contained (in mM): 145 Cs-glutamate, 10 HEPES, 8 NaCl, 1 MgCl 2 , 2 Mg-ATP adjusted to pH 7.2 with CsOH. The free Ca 2ϩ concentration was clamped by the addition of 10 mM EGTA and 3.64 mM CaCl 2 to the Cs-glutamate-based pipette solution. The nominally Mg 2ϩ -free solution contained (in mM): 145 Cs-glutamate, 10 HEPES, 10 EGTA, 6 Na 2 -ATP, 8 NaCl adjusted to pH 7.2 with CsOH. The intracellular Mg-ATP concentration was increased to 6 mM for experiments in the absence of external divalent cations. Perforated patch recordings were performed with the standard intracellular solution supplemented with 100 g ml Ϫ1 nystatin (14). The tips of the patch pipettes were filled with nystatin-free internal solution by capillary force. When challenged with step depolarizations, perforation of the membrane patch was indicated by characteristic changes of the capacitance transient and by a continuous decline of the input resistance. To standardize experiments, recordings were started when the input resistance fell to a value of 20 M⍀ or less; this point was reached usually 5-10 min after gigaseal formation. The series resistances in the tight seal whole cell experiments were typically in the range of 5-10 M⍀. Currents were filtered using an 8-pole Bessel filter at 2.9 kHz and digitized at 100 s. Voltage clamp recordings were performed using ramps (Ϫ110 mV to ϩ90 mV in 50 ms) applied every 2 s using PULSE software (HEKA Electronics) on a personal computer. Cells were held at Ϫ80, Ϫ10, or 50 mV between ramps. Several parameters (capacitance, series resistance, holding current) were displayed simultaneously at a slower rate (2 Hz) using the X-Chart display (HEKA Electronics). The membrane potential values were corrected for 10 mV liquid junction potential. No additional voltage correction was performed for the experiments under divalent-free conditions. Effects of changes in surface charge screening were ignored. All experiments were carried out at room temperature (20 -23°C). Internal solutions were kept on ice to minimize hydrolysis of the nucleotides and IP 3 .
Data Analysis-Analysis was performed with PulseFit and programs written in the IGOR macro language (Wave Metrics, Lake Oswego, OR). Currents were elicited by voltage ramps and measured at Ϫ80 mV. Current densities are shown after background current subtraction in Fig. 4, A-C and Fig. 5. This was done by subtracting either the first ramp in TRPV6 transfected cells or by averaging the first 2-4 ramps after establishing the whole cell mode in non-transfected cells and then subtracting the mean from all subsequent traces. No correction for the initial leak current was performed for the data from TRPV6-expressing HEK cells in all other figures. Peak Ca 2ϩ current amplitudes were determined during 120 s of membrane depolarization to 50 mV. Both values were compared with the estimate of the depolarization-induced current increase for each cell. Beside these relative values, the absolute current increase was analyzed, too. Throughout, average data are given as means Ϯ S.E. for n number of cells. Student's unpaired t test was used for comparison of the means.
Drugs-All chemicals were purchased from Sigma. Nystatin was dissolved in Me 2 SO to yield a final concentration of 50 mg ml Ϫ1 , and a new stock solution was used for each experiment.

Time Course of Currents and Current-Voltage (I-V) Relationships from TRPV6-expressing HEK Cells at Various Holding
Potentials-Patch clamp experiments were carried out to investigate the effects of different holding potentials on TRPV6mediated Ca 2ϩ currents in transfected HEK cells. After establishing the tight seal whole cell configuration the cell was dialyzed with 10 mM EGTA from the internal solution. Inward currents were monitored at Ϫ80 mV during voltage ramps, which were applied immediately after breakin at a frequency of 0.5 Hz.
No current developed if the plasma membrane was clamped at Ϫ80 mV between voltage ramps (Fig. 1A, top). The I-V relationship showed a negative reversal potential, most likely caused by a basal Cl Ϫ conductance present in these cells (15). The inward current at potentials around 30 mV may arise through the activation of voltage-dependent channels, which are expressed in this cell line (16) and are not deactivated at such negative potentials (Fig. 1B, top).
At a holding potential of Ϫ10 mV the initial inward current measured by the first voltage ramp was larger than at more hyperpolarized membrane potentials (Ϫ2.3 Ϯ 0.8 pA/pF, n ϭ 10 at Ϫ80 mV and Ϫ8.2 Ϯ 2.4 pA/pF, n ϭ 7 at Ϫ10 mV). The current increased in size, peaked after 10 -20 s of whole cell recording and subsequently decayed until a steady-state level was reached (Fig. 1A, middle). It was inwardly rectifying and reversed at positive potentials indicating a high Ca 2ϩ selectivity (Fig. 1B, middle). No such current was recorded in the absence of external Ca 2ϩ from TRPV6-transfected HEK cells (data not shown, n ϭ 5). Likewise, this current was not detected in either mock or non-transfected HEK cells (data not shown, n ϭ 5 each).
Depolarization of the membrane resulted in a further increase of the TRPV6-mediated Ca 2ϩ entry (Ϫ26.4 Ϯ 3.3 pA/pF, n ϭ 7 at Ϫ10 mV and Ϫ47.8 Ϯ 7.8 pA/pF, n ϭ 12 at 50 mV). The time course and voltage dependence are shown at the bottom of Fig. 1. Taken together, these results indicate that the size of TRPV6-mediated Ca 2ϩ currents are strongly dependent on the holding potential, as suggested previously (10,11,17,18).
Current Changes during Voltage Steps of TRPV6-expressing HEK Cells with High Intracellular Ca 2ϩ Buffering-To test whether the voltage-dependent modulation of TRPV6 channel activity depends on the preceding state of channel activation or inactivation, the holding potential was changed during the recording from the same cell. With 10 mM EGTA in the pipette solution and a holding potential of Ϫ10 mV the Ca 2ϩ current activated rapidly. After the peak amplitude was reached at about 30 s, the current inactivated until current densities were stable at about 200 and 300 s. After this time, the cell is expected to be almost completely dialyzed with the EGTA containing intracellular solution (19). Next, the cell membrane was clamped at 50 mV between voltage ramps. As a consequence the current amplitude slowly increased with an exponential time course ( ϭ 58 s). After stepping back to Ϫ10 mV, the Ca 2ϩ entry rapidly decayed in a complex manner, before current densities were reconstituted to values similar to the ones recorded before depolarization ( Fig. 2A). This voltage-de- Note the different scaling of the current densities. Representative examples are shown (n ϭ 10, 7, 12 for Ϫ80, Ϫ10, and 50 mV, respectively). pendent modulation of Ca 2ϩ entry was reversible because it was recorded several times from the same cell (data not shown, n ϭ 20). Intracellular diffusible factors do not seem to be responsible for this effect because there was no correlation with the duration of whole cell perfusion. It was possible to elicit these changes in current densities with respect to the holding potential at physiological potentials. Depolarization increased the TRPV6-mediated Ca 2ϩ currents and hyperpolarization had the opposite effect (data not shown, holding potentials of Ϫ110 to 90 mV tested).
High resolution Ca 2ϩ currents evoked by voltage ramps after maximum activation at Ϫ10 and 50 mV holding potential are shown in Fig. 2B. On the basis of the similar I-V characteristics it is most likely that only TRPV6-mediated Ca 2ϩ influx was measured without major contamination of other currents. This voltage protocol was used in the following experiments to study the effect of the holding potential on the TRPV6 channel in more detail.
Other members of the TRP family, namely TRPV1 and TRPM8 have recently been shown to behave in a voltage-dependent manner. Both cation channels can be activated at different temperatures and thereby current activation at depolarized potentials precedes that at more negative potentials (20). However, such a behavior was not detected for the strongly inward rectifying TRPV6-mediated Ca 2ϩ current when activated with EGTA dialysis or membrane depolarization ( Fig. 2A). Likewise, no obvious changes in the voltage dependence of TRPV6-mediated Ca 2ϩ influx was recorded at the two holding potentials with a tail current protocol (Fig. 3).
Current Changes during Voltage Steps of TRPV6-expressing HEK Cells at Low Intracellular Ca 2ϩ Buffering-It has been previously shown, that TRPV6 channel activity correlates with the intracellular Ca 2ϩ chelator concentration (9). However, the calculated global [Ca 2ϩ ] i at which TRPV6 channels start to activate is extremely low and several magnitudes below physiological values (9). Therefore, it is crucial to find out whether TRPV6 channels open because of changes in membrane potential under more physiological conditions. At moderate cytoplasmic Ca 2ϩ buffering obtained by dialyzing cells with low concentrations of EGTA (0.1 mM), no current developed at a holding potential of Ϫ10 mV (Ϫ1.7 Ϯ 0.3 pA/pF, n ϭ 12; Fig. 4A). In contrast, large Ca 2ϩ currents were recorded at higher intracellular EGTA concentrations (Figs. 1 and 2) indicating that TRPV6 channels are strongly dependent on [Ca 2ϩ ] i and are significantly augmented by decreasing [Ca 2ϩ ] i . Switching the holding potential from Ϫ10 to 50 mV led to a gradual current increase until steady state was reached (Ϫ8.9 Ϯ 0.8 pA/pF, n ϭ 3; Fig. 4A). This current is likely to be TRPV6-mediated since it was strongly inward rectifying and reversed at positive potentials. Repolarization from 50 to Ϫ10 mV resulted in current decay and a very similar level was reached compared with the one before depolarization. Therefore, it is indeed possible to demask TRPV6-mediated Ca 2ϩ entry under quite physiological conditions of intracellular Ca 2ϩ

FIG. 2. Changes in current density during voltage steps in a TRPV6-expressing HEK cells with high intracellular Ca 2؉ buffering.
A, development of currents in a typical TRPV6transfected HEK following dialysis with the Cs-glutamate-based internal solution containing 10 mM EGTA. The membrane potential was changed from Ϫ10 mV to 50 mV as indicated. The correlating changes in current density at Ϫ80 mV and 80 mV are plotted versus time. B, I-V relationships were recorded with the ramp protocol shown at the top. At the time indicated in A current traces were measured. Note the different scaling of the current density in 2. A typical example is shown for five similar recordings. buffering and membrane potentials. In non-transfected HEK cells no such current and voltage-dependent behavior was seen (Fig. 4A).
The intracellular Ca 2ϩ dependence of TRPV6 channels with respect to its voltage-dependent behavior was further investigated with an internal solution containing 100 nM free Ca 2ϩ using the appropriate concentration of EGTA (10 mM) and CaCl 2 (3.64 mM; Fig. 4B). Again, no current activation was detected like in the experiments using 0.1 mM EGTA. In contrast to the experiments with 0.1 mM EGTA, the very same depolarization of the plasma membrane was now unable to activate TRPV6-mediated Ca 2ϩ entry (Ϫ0.7 Ϯ 0.3 pA/pF, n ϭ 3; Fig. 4B). Thus, the TRPV6-mediated Ca 2ϩ current density depends on both the intracellular EGTA concentration and the membrane potential (Fig. 5A).
Perforated patch clamp recordings do not alter the endogenous Ca 2ϩ buffering because there is no washout of cytosolic components (14). Therefore, nystatin patches were used to find out whether changes in the membrane potential can modulate TRPV6-mediated Ca 2ϩ entry in a situation with physiological Ca 2ϩ buffering. This was indeed the case as shown in Fig. 4, C-E. The voltage step from Ϫ10 to 50 mV induced the slow activation of an inward current with the typical voltage dependence of the TRPV6 protein (Ϫ1.9 Ϯ 0.5 pA/pF, n ϭ 5; Fig.  4C). No such current was recorded either before or after depolarization at the holding potential of Ϫ10 mV (Fig. 4, C-E). Likewise, no inward rectifying Ca 2ϩ current was measured in non-transfected HEK cells if voltage ramps were applied from either Ϫ10 or 50 mV (Ϫ0.6 Ϯ 0.1 pA/pF, n ϭ 4; Fig. 4C). In summary, these results favor the idea that TRPV6-mediated Ca 2ϩ currents are regulated in a voltage-dependent manner under in vivo Ca 2ϩ buffering conditions.
Voltage-dependent Modulation of TRPV6-mediated Ca 2ϩ Currents Is Dependent on Intracellular Ca 2ϩ -The preceding data demonstrate that the amplitude of TRPV6-mediated Ca 2ϩ currents is reduced as the holding potential between ramps becomes negative and is increased at positive voltages. It might be possible that cations bind closely to or even within the pore and thereby plug the channel. During membrane depolarization the blocking particle is electrostatically removed, which favors Ca 2ϩ entry. However, voltage-dependent current changes were relatively slow in comparison to the rapid changes in the holding potential. The following experiments should clarify the underlying mechanism for this voltage-dependent modulation of TRPV6 channel gating.
Negative membrane potentials provide a favorable driving force for Ca 2ϩ entry and it is conceivable that the enhanced Ca 2ϩ influx could inactivate TRPV6 channels in a Ca 2ϩ -dependent manner. This Ca 2ϩ -dependent inactivation arises from the build-up of a microdomain of Ca 2ϩ in the vicinity of each channel (9). This local Ca 2ϩ gradient might be controlled if high concentration of a fast Ca 2ϩ chelator (60 mM BAPTA) were dialyzed into the cell (Fig. 6). TRPV6 channel activation was monitored by sporadic application of short voltage ramps (25 ms duration, every 10 s). Under these conditions Ca 2ϩ currents were significantly larger (Ϫ169 Ϯ 21 pA/pF, n ϭ 10 versus Ϫ38 Ϯ 11 pA/pF, n ϭ 6 after 200 s whole-cell recording) and activated faster ( ϭ 25 Ϯ 2 s, n ϭ 11 versus ϭ 36 Ϯ 5 s, n ϭ 6) than with the standard voltage protocol (50 ms duration, every 2 s). Steadystate levels were reached within 100 s and no rundown was detected (Fig. 6). After about 200 s of whole cell perfusion the intracellular BAPTA concentration is thought to be roughly the same as in the pipette solution (19). Now the repetitive voltage ramps were no longer applied to minimize Ca 2ϩ entry to allow an efficient Ca 2ϩ buffering close to the TRPV6 pore. After a pause of 1 min, Ca 2ϩ current amplitude did not increase, suggesting that BAPTA was able to achieve an equilibrium within the presumable Ca 2ϩ microdomain responsible for inactivation during 10 s or

FIG. 4. Changes in current density during voltage steps in TRPV6-expressing HEK cells with low intracellular Ca 2؉ buffering.
Time course of currents in TRPV6-transfected HEK cells after dialysis with internal solutions containing EGTA (0.1 mM) in A or 100 nM free Ca 2ϩ in B. Perforated patch clamp recordings were carried out in C-E. The holding potential was switched from Ϫ10 mV to 50 mV at 300 s for 120 s. Currents were subtracted from the initial inward current as described under "Experimental Procedures" in A-C but not in D and E. Mean data with double-sided S.E. are shown in A-C (A, n ϭ 3 for HEK-TRPV6 and HEK-wt; B, n ϭ 3 for HEK-TRPV6 and n ϭ 3 for HEK-TRPV6 and HEK-wt; C, n ϭ 5 for HEK-TRPV6 and n ϭ 4 for HEK-wt) and a representative example in E. less (Fig. 6B). Thus, Ca 2ϩ currents seemed to be maximally activated, but switching the holding potential from Ϫ10 to 50 mV resulted in an augmentation of Ca 2ϩ influx without changes in the voltage dependence (Fig. 6B). These data suggest that the presumable Ca 2ϩ binding site within the TRPV6 protein responsible for the voltage-dependent modulation of the Ca 2ϩ currents may not be accessible from the cytosolic side. Another possibility is that it binds Ca 2ϩ with such a high affinity that the chelator cannot interfere.
Voltage-dependent Modulation of TRPV6-mediated Ca 2ϩ Currents with Varying Intracellular Mg 2ϩ -Intracellular Mg 2ϩ can block K ϩ channels in a voltage-dependent manner thereby inducing an inward rectification (21). In analogy, Mg 2ϩ ions might plug TRPV6 channels from the inside at hyperpolarized potentials. At positive potentials this block then might be removed leading to enhanced Ca 2ϩ entry. This idea was tested by examining the effects of Mg 2ϩ -free conditions on the voltagedependent modulation of Ca 2ϩ currents (Fig. 7). TRPV6-expressing HEK cells were dialyzed with a nominally Mg 2ϩ -free internal solution. Switching the holding potential from Ϫ10 to 50 mV after 300 s of whole cell perfusion resulted in a 5.4-fold current increase (Ϫ4.4 Ϯ 1.7 pA/pF to Ϫ23.9 Ϯ 7.1 pA/pF, n ϭ 4; after initial conductance correction). Thus, dramatically reducing the intracellular Mg 2ϩ concentration did not abolish the inward rectification and voltage-dependent modulation of Ca 2ϩ currents in TRPV6-expressing HEK cells. This was also the case when neither Mg 2ϩ nor ATP was added to the internal solution excluding a protein kinase-mediated effect (data not shown, n ϭ 4).
Voltage-dependent Modulation of TRPV6-mediated Ba 2ϩ Currents-As shown above the voltage-dependent behavior of TRPV6 channels was not abolished when either the intracellular Mg 2ϩ or Ca 2ϩ concentration was decreased dramatically. Therefore, Ba 2ϩ was used next as the charge carrier instead of Ca 2ϩ because Ba 2ϩ permeates Ca 2ϩ channels but substitutes poorly at Ca 2ϩ -binding proteins. Replacing Ca 2ϩ by Ba 2ϩ in the bath solution resulted in similar inward currents compared with the experiments with 10 mM intracellular EGTA (peak current densities: Ϫ27.9 Ϯ 4.7 pA/pF, n ϭ 7 for I Ba 2ϩ and Ϫ23.8 Ϯ 4.8 pA/pF, n ϭ 6 for I Ca 2ϩ). Likewise, the effect of holding potential was conserved switching from Ca 2ϩ to Ba 2ϩ currents (Fig. 8). Thus, Ba 2ϩ is able to substitute for the physiological charge carrier without abolishing the voltage dependence on current amplitude.
Voltage-dependent Modulation of TRPV6-mediated Monovalent Currents-Given the presumable importance of the charge carrier it was decided to record monovalent currents in the absence of extracellular divalent cations (Fig. 9). TRPV6-expressing HEK cells were dialyzed with 10 mM EGTA to activate prominent Ca 2ϩ influx. The holding potential was changed after 300 s of whole cell recording for 120 s and induced the typical augmentation of Ca 2ϩ entry as previously shown. Under DVF conditions the inward current was initially depressed, as one would expect for anomalous mole fraction behavior. Subsequently, large monovalent currents developed with an exponential time course, reached their peak amplitudes, which stayed stable with time. Now the holding potential was changed again, but without any effect on current density. After the application of the DVF solution was stopped, the inward current decayed rapidly. The very same depolarization of the membrane potential was now able to increase the TRPV6mediated Ca 2ϩ current. I-V relationships allowed the identification of the typical TRPV6-mediated currents in the presence or absence of external divalent cations (Fig. 9B). Ca 2ϩ and monovalent currents were elicited by standard voltage ramps from a holding potential of Ϫ10 mV immediately before depolarization to 50 mV, at the end of the 120-s interval at 50 mV and after repolarization to Ϫ10 mV. In all cases the inward current reversed at slightly positive potentials and showed the characteristic negative slope below Ϫ80 mV, which is due to time-dependent removal of an intracellular Mg 2ϩ block (9,22). Strikingly, no increase in monovalent current amplitude was detected due to the changes in membrane potential (Fig. 9). These data suggest that the charge carrier is crucial in generating the voltage-dependent conductance changes of the TRPV6 channel. In addition, channel activity is still dependent on [Ca 2ϩ ] i even if monovalent ions instead of Ca 2ϩ permeate the TRPV6 pore (Fig. 5B).
Voltage-dependent Modulation of TRPV6 Channels Is Independent of Calmodulin Binding at the C Terminus and Ca 2ϩ Binding to the Aspartate Residue at Position 542 within the Presumable Pore Region-The results shown so far indicate a divalent cation-dependent mechanism responsible for TRPV6 channel activity. Previously it was shown that Ca 2ϩ -dependent CaM binding at the C terminus of TRPV6 facilitates channel inactivation, which can be counteracted by PKC-mediated phosphorylation of a threonine residue at position 702 within this CaM binding site (12). Therefore, the following experiments were designed to test whether CaM and PKC contribute to the voltage-dependent modulation of TRPV6 channels.
The CaM binding site covering amino acid residues 694 -716 of TRPV6 is absent in the truncated TRPV6⌬ 693-725 variant, which lacks the C-terminal 33 amino acid residues including the CaM binding site (12). However, the typical Ca 2ϩ current increase was elicited by changing the holding potential from Ϫ10 to 50 mV (Fig. 10A). Similar results were obtained at lower intracellular Ca 2ϩ buffering (0.1 mM EGTA instead of 10 mM, data not shown, n ϭ 3). Replacing the threonine residue at position 702 by an alanine residue (TRPV6 T702A ), does not affect CaM binding but the TRPV6 protein can no longer be phosphorylated at this site (12). However, the effect of the membrane potential on Ca 2ϩ current amplitude was also present when the TRPV6 T702A variant was expressed (data not shown, n ϭ 6). Similar recordings were performed on wild-type TRPV6 expressing HEK cells incubated with the broad spectrum kinase blocker staurosporine (2 M for at least 30 min). In addition, ATP was omitted from the EGTA (10 mM) containing pipette solution to further reduce kinase activity. Under these conditions PKC activity is expected to be extremely low; however, Ca 2ϩ current changes in response to the membrane potential remained (data not shown, n ϭ 5). Thus, neither CaM binding at position 693-725 nor PKC-mediated phosphorylation at Thr 702 was responsible for the voltage-dependent conductance changes.
Previously it has been shown that the aspartate residue at position 541 of the mouse TRPV6 protein determines Ca 2ϩ permeation and Mg 2ϩ block (23). The mutation of the corresponding aspartate residue to alanine in the human TRPV6 yields proteins inserted in the plasma membrane, which do not function as ion conducting channels (10). Also switching the holding potential from Ϫ10 mV to 50 mV did not allow the detection of Ca 2ϩ currents in HEK cells expressing the TRPV6 D542A protein (Fig.  10B). Likewise, no differences to non-transfected cells were detected from cells expressing the double mutant where the aspartate residue at position 542 and the glutamate residue at position 550 were mutated into alanine residues (data not shown, n ϭ 4).

DISCUSSION
The present study shows that the dose response curve for TRPV6-mediated Ca 2ϩ currents in dependence of the [Ca 2ϩ ] i (9) is shifted by the membrane potential: Depolarization induced an augmentation of Ca 2ϩ entry whereas hyperpolarization reduced current densities. No prominent Ca 2ϩ current was detected if TRPV6-expressing HEK cells were clamped at Ϫ10 mV between voltage ramps and whole cell recordings were performed with 0.1 mM intracellular EGTA. Increasing either the membrane potential or the chelator concentration resulted in the activation of TRPV6 channels with their characteristic I-V signature. Likewise, no TRPV6-mediated current was detected with 10 mM EGTA during hyperpolarization whereas the typical TRPV6 current developed if the holding potential became positive. Thus, the TRPV6 protein functions as a Ca 2ϩ channel and its current amplitude is strongly dependent on both [Ca 2ϩ ] i and membrane potential.
The voltage-dependent modulation of TRPV6 channels is an unexpected finding since the driving force for Ca 2ϩ was only changed between pulses and currents were measured at the same potential throughout by applying voltage ramps repetitively. One explanation for these data is that due to membrane depolarization a blocking particle is removed from the pore. Unlike inward rectifying K ϩ channels and NMDA receptors, it does not seem to involve voltage-dependent block by intracellular Mg 2ϩ . This result is of particular interest since it has been previously shown that Mg 2ϩ can block TRPV6 channels in a voltage-dependent manner (9,22). The role of Ca 2ϩ ions, which may plug the TRPV6 channel from the cytosolic side was examined in experiments with 60 mM BAPTA in the pipette solution. As previously reported, current densities did not reach steady state but continuously increased in comparison to current amplitudes recorded with 30 mM BAPTA (9). Changing the holding potential affected the TRPV6-mediated Ca 2ϩ current suggesting that intracellular Ca 2ϩ is not important for this effect. Another possibility is that even these high concentrations of a fast chelator are unable to access the putative Ca 2ϩ binding site at the pore and thereby relieve channels from partial Ca 2ϩ -dependent inactivation. Since the changes in current density could be repeatedly evoked during prolonged whole cell recordings, it is unlikely that a small diffusible factor could be involved. Such a molecule would be effectively washed out of the cell into the recording pipette.
The mechanism by which TRPV6 channels operate in a voltage-dependent manner involves the charge carrier. Like Ca 2ϩ currents, Ba 2ϩ currents were augmented during depolarization and reduced at hyperpolarization. However, in the absence of divalent external cations monovalent ions permeate the TRPV6 pore. These currents, mainly carried by Na ϩ ions, were independent to changes in the holding potential. Therefore, one might imagine a scenario in which permeating Ca 2ϩ ions block the pore under physiological conditions. Depolarization of the cell membrane could result in electrostatic repulsion of these Ca 2ϩ ions, thus facilitating Ca 2ϩ flux and vice versa at hyperpolarization. Stepping to 50 mV, the Ca 2ϩ current usually developed slowly and decayed rapidly when returning back to Ϫ10 mV. The different kinetics probably reflect the regulation of TRPV6 by Ca 2ϩ that may dissolve slowly but rebind rapidly. With this idea in mind, the slower kinetics of Ba 2ϩ current activation and inactivation could be explained as resulting from the different binding affinities of Ba 2ϩ and Ca 2ϩ to the pore. Both, TRPV6-mediated Ca 2ϩ and monovalent currents show a negative correlation of amplitude and [Ca 2ϩ ] i . However, only Ca 2ϩ influx, but not monovalent currents can be modulated by the membrane potential. Thus, both processes are regulated independently.
Voltage-operated Ca 2ϩ channels compromise four glutamate residues in the putative pore region, which are likely to be responsible for the high Ca 2ϩ selectivity (24,25). Electrostatic interaction between these glutamate residues and Ca 2ϩ are a critical determinant of high affinity Ca 2ϩ binding and permeation properties, suggesting that these negative residues could behave as surrogate water molecules to facilitate the passage of dehydrated Ca 2ϩ through the hydrophobic plasma membrane FIG. 9. TRPV6-mediated Ca 2؉ and monovalent Ca 2؉ currents at various holding potentials. A, time course of currents from a representative TRPV6expressing HEK cell dialyzed with an intracellular solution containing 10 mM EGTA (n ϭ 6). The membrane potential was clamped at Ϫ10 or 50 mV as shown on the top. The normal external solution was substituted versus DVF saline, as indicated with the bar. B, I-V relations were recorded before, during, and after depolarization at the time points highlighted in A.  (26). However, the only negatively charged amino acid residues in the putative pore-forming region of TRPV6 are a glutamate residue at position 535 and two aspartate residues at positions 542 and 550. All three amino acids are conserved in the TRPV5 protein from rabbit and especially the aspartate residue at position 542 has been shown to decisively affect Ca 2ϩ permeation (27,28). Substituting aspartate at position 542 versus an alanine in human TRPV6 abolished channel activity (10), which was also the case if this pore mutant was C-terminally EGFP-tagged. 2 However, the very same mutant of mouse TRPV6 conducted monovalent currents under external divalent-free conditions with slightly more outward rectification than in wild-type TRPV6-transfected HEK cells, when intracellular Mg 2ϩ was removed (23). These functional differences may be due to the 10% different amino acid residues between human and mouse TRPV6. Interestingly, the putative pore region of the mouse clone contained an additional aspartate residue at position 547, which is not conserved in human.
The voltage-dependent modulation of TRPV6-mediated Ca 2ϩ currents takes place at physiological [Ca 2ϩ ] i . Changes in current amplitude were recorded at membrane potentials normally encountered under in vivo conditions by cells expressing TRPV6. For instance, pancreatic acinar cells usually have a resting potential of about Ϫ40 mV (29 -31) and muscarinic stimulation results in either depolarization or hyperpolarization, which seems to be species-dependent (29 -31). These changes in the membrane potential are likely to influence TRPV6-mediated Ca 2ϩ entry in a complex fashion since TRPV6 channels are activated at potentials where the driving force for Ca 2ϩ is small. Furthermore the incoming Ca 2ϩ is blocking TRPV6 channel activity and thereby protecting the cell from toxic Ca 2ϩ overload. It will be interesting to investigate whether the voltage-dependent effect on TRPV6-mediated Ca 2ϩ influx occurs in primary cell culture within the range of physiological membrane potentials, thus modulating the [Ca 2ϩ ] i and thereby cellular function such as secretion.