Ca2+-dependent potentiation of the nonselective cation channel TRPV4 is mediated by a C-terminal calmodulin binding site.

Most Ca2+-permeable ion channels are inhibited by increases in the intracellular Ca2+ concentration ([Ca2+]i), thus preventing potentially deleterious rises in [Ca2+]i. In this study, we demonstrate that currents through the osmo-, heat- and phorbol ester-sensitive, Ca2+-permeable nonselective cation channel TRPV4 are potentiated by intracellular Ca2+. Spontaneous TRPV4 currents and currents stimulated by hypotonic solutions or phorbol esters were reduced strongly at all potentials in the absence of extracellular Ca2+. The other permeant divalent cations Ba2+ and Sr2+ were less effective than Ca2+ in supporting channel activity. An intracellular site of Ca2+ action was supported by the parallel decrease in spontaneous currents and [Ca2+]i on removal of extracellular Ca2+ and the ability of Ca2+ release from intracellular stores to restore TRPV4 activity in the absence of extracellular Ca2+. During TRPV4 activation by hypotonic solutions or phorbol esters, Ca2+ entry through the channel increased the rate and extent of channel activation. Currents were also potentiated by ionomycin in the presence of extracellular Ca2+. Ca2+-dependent potentiation of TRPV4 was often followed by inhibition. By mutagenesis, we localized the structural determinant of Ca2+-dependent potentiation to an intracellular, C-terminal calmodulin binding domain. This domain binds calmodulin in a Ca2+-dependent manner. TRPV4 mutants that did not bind calmodulin lacked Ca2+-dependent potentiation. We conclude that TRPV4 activity is tightly controlled by intracellular Ca2+. Ca2+ entry increases both the rate and extent of channel activation by a calmodulin-dependent mechanism. Excessive increases in [Ca2+]i via TRPV4 are prevented by a Ca2+-dependent negative feedback mechanism.


Most Ca 2؉ -permeable ion channels are inhibited by increases in the intracellular Ca 2؉ concentration ([Ca 2؉ ] i ), thus preventing potentially deleterious rises in [Ca
] i . In this study, we demonstrate that currents through the osmo-, heat-and phorbol ester-sensitive, Ca 2؉ -permeable nonselective cation channel TRPV4 are potentiated by intracellular Ca 2؉ . Spontaneous TRPV4 currents and currents stimulated by hypotonic solutions or phorbol esters were reduced strongly at all potentials in the absence of extracellular Ca 2؉ . The other permeant divalent cations Ba 2؉ and Sr 2؉ were less effective than Ca 2؉ in supporting channel activity. An intracellular site of Ca 2؉ action was supported by the parallel decrease in spontaneous currents and [Ca 2؉ ] i on removal of extracellular Ca 2؉ and the ability of Ca 2؉ release from intracellular stores to restore TRPV4 activity in the absence of extracellular Ca 2؉ . During TRPV4 activation by hypotonic solutions or phorbol esters, Ca 2؉ entry through the channel increased the rate and extent of channel activation. Currents were also potentiated by ionomycin in the presence of extracellular Ca 2؉ . Ca 2؉ -dependent potentiation of TRPV4 was often followed by inhibition. By mutagenesis, we localized the structural determinant of Ca 2؉ -dependent potentiation to an intracellular, C-terminal calmodulin binding domain. This domain binds calmodulin in a Ca 2؉ -dependent manner. TRPV4 mutants that did not bind calmodulin lacked Ca 2؉ -dependent potentiation. We conclude that TRPV4 activity is tightly controlled by intracellular Ca 2؉ . Ca 2؉ entry increases both the rate and extent of channel activation by a calmodulin-dependent mechanism. Excessive increases in [Ca 2؉ ] i via TRPV4 are prevented by a Ca 2؉ -dependent negative feedback mechanism.
Channels of the TRP family can be divided on the basis of structural features into three subfamilies: classic or canonical (TRPC) channels, melastatin-like (TRPM) channels, and vanilloid receptor-like (TRPV) channels (1)(2)(3)(4). The TRPV subfamily is composed of six members with very different functional properties. TRPV1-4 are Ca 2ϩ -permeable, nonselective cation channels activated by a variety of stimuli but having in common sensitivities to different levels of heat. TRPV1 (VR1) is activated by heat, vanilloids, and protons and is involved in the transduction of noxious heat stimuli in primary sensory neurons (5). TRPV2 (VRL-1, GRC) is activated by high levels of noxious heat (6) and is probably involved in high threshold nociception. In addition, this channel is expressed in other tissues and, in some, is translocated to the cell membrane by stimulators of cell proliferation (7). Recent studies (8 -10) have shown that TRPV3 responds to innocuous levels of heat and is expressed in primary sensory neurons and keratinocytes, suggesting a role of this channel in non-nociceptive thermoreception. TRPV4 was first shown to respond to changes in cell volume or extracellular osmolarity by an unknown mechanism (11)(12)(13)(14). Recently (15), this channel has been shown to be activated by phorbol esters independently of activation of protein kinase C and by heat (16,17). In contrast to TRPV1-4, the other members of this subfamily, TRPV5 and TRPV6, are highly Ca 2ϩ -permeable channels and are involved in epithelial Ca 2ϩ transport in the kidney and gut (18,19).
The activity of many Ca 2ϩ -permeable cation channel types, including voltage-gated Ca 2ϩ channels, cyclic nucleotide-gated channels, and N-methyl-D-aspartate receptors, is regulated by the intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ). This regulation is usually inhibitory and often provides a feedback mechanism to prevent excessive increases in [Ca 2ϩ ] i . In most cases, Ca 2ϩ does not act directly on the channel but through binding to the Ca 2ϩ -binding protein calmodulin (CaM). 1 A number of TRP channel isoforms are modulated by Ca 2ϩ , and some have CaM binding domains in their C termini. All TRPC isoforms have at least one CaM binding domain (20 -22) and bind CaM in a Ca 2ϩ -dependent manner. For TRPC1, CaM has been shown to regulate Ca 2ϩ -dependent feedback inhibition (23). TRPM2 requires intracellular Ca 2ϩ for its activation, although the mechanism of Ca 2ϩ action is unknown (24). Effects of intracellular Ca 2ϩ have also been reported for members of the TRPV subfamily. Rapid desensitization of TRPV1 is dependent on Ca 2ϩ (5), acting at an intracellular site via a Ca 2ϩ -CaM-dependent phosphatase (25). The epithelial Ca 2ϩ channels TRPV5 and TRPV6 are inhibited by intracellular Ca 2ϩ in the submicromolar range (26 -29), and, at least for TRPV6, part of this Ca 2ϩdependent inactivation is CaM-dependent (28). Raising [Ca 2ϩ ] i has also been reported recently (15) to inhibit TRPV4, suggesting that this channel may also undergo Ca 2ϩ -dependent feedback inhibition.
In this paper we show that, in addition to an inhibitory effect, intracellular Ca 2ϩ potentiates currents through TRPV4. Ca 2ϩ entry through the channel is involved in the maintenance of spontaneous activity and increases both the TRPV4 current amplitude and the rate of current activation in response to channel stimulation with hypotonic solutions or phorbol esters. Using mutagenesis, we localized the structural determinant of the stimulatory effect to an intracellular C-terminal domain. Binding experiments showed that the region involved is a Ca 2ϩdependent CaM binding domain. Mutations that prevent CaM binding led to a loss of Ca 2ϩ -dependent potentiation of TRPV4.

Cloning of Human TRPV4 and Generation of a Stably Transfected
Cell Line-For the cloning of human TRPV4, total RNA was prepared from human kidney tissue samples using the TRIzol LS reagent (Invitrogen) according to the manufacturer's protocol. Reverse transcription was performed from 1 g of total RNA using 200 units of Super-Script II reverse transcriptase (Invitrogen) and 5 pmol of the primer (dT) 17 . The human TRPV4 coding sequence was amplified in 30 cycles of PCR under the following conditions: 55°C annealing temperature, extension at 72°C for 3 min, Expand HF polymerase (Roche Applied Science), primers GGAAGCTTGCCACCATGGCGGATTCCAGCGA (sense) and CCGGATCCGCGAGCGGGGCGTCATCAGT (antisense). The PCR product was subcloned into the pcDNA3.1 vector (Invitrogen), and the cDNA sequence was confirmed by DNA sequencing of both strands. A C-terminal green fluorescent protein fusion construct was obtained by HindIII/BamHI digestion of the fragment and ligation into the pEGFP-N1 vector (Clontech). Point mutations were inserted by overlap extension PCR using appropriate sense and antisense primers. C-terminal deletion mutants were amplified using a modified antisense primer.
For the generation of a stably transfected cell line, the HindIII/ BamHI fragment was cloned into the pcDNA4/TO (Invitrogen) vector, and the construct was transfected into the T-REx-293 cell line using the FuGENE 6 transfection reagent (Roche Applied Science). Clonal selection was performed according to the manufacturer's protocol with Zeocin (250 g/ml), and multiple clones were functionally tested for TRPV4 expression.
CaM Pull-down Assays-Human TRPV4 C-terminal fragments were amplified by PCR and subcloned into the pGEX-2TK vector (Amersham Biosciences). The sequences were verified by DNA sequencing, and the glutathione S-transferase fusion proteins were expressed from Escherichia coli BL21 cells. After purification by glutathione-Sepharose according to the standard protocol (Amersham Biosciences), the eluted peptides were subjected to interaction with calmodulin-Sepharose (Amersham Biosciences) for 60 min at room temperature in a buffer containing (in mM) 50 Tris-HCl, pH 7.5, 100 NaCl, 0.1% Triton X-100, and 2 CaCl 2 or 2 EGTA. For quantitative interaction assays, the free calcium concentration was buffered to the desired value using the appropriate chelators (EGTA, HEDTA, nitrilotriacetic acid, or EDTA). Pellets were washed three times with the respective buffers and subjected to SDS gel electrophoresis. Gels were stained with Coomassie Blue dye and dried, and the relative intensities of the bands were analyzed using AIDA software.
Cell Culture and Transfection-HEK293 cells were cultured in minimum essential medium with Earle's salts (Biochrom, Berlin, Germany), supplemented with 10% (v/v) fetal calf serum (Biochrom) and 100 units/ml penicillin and 100 g/ml streptomycin. Cells were plated onto glass coverslips 24 -48 h prior to transfection. The cells were transiently transfected with 1 g of DNA and 6 l of FuGENE 6 transfection reagent (Roche Diagnostics) in 94 l of OptiMEM medium (Invitrogen) per 85-mm dish. Ca 2ϩ measurements and electrophysiological studies were performed 24 -36 h after transfection.
Patch Clamp Recording-Recordings of whole cell currents from single cells were made with an EPC-7 amplifier using Pulse software (HEKA, Lambrecht, Germany) as described previously (12). Experiments were performed using the standard whole cell mode of the patch clamp technique or the perforated patch technique. For the latter, amphotericin B (250 g/ml; Sigma) was used as the ionophore for current recording. Measurements of currents in perforated patch recordings were started when the series resistance approached a plateau at values Ͻ30 megohms. To measure channel currents, cells were held at a potential of 0 or Ϫ20 mV and ramps from Ϫ100 to ϩ100 mV with a duration of 400 ms applied at a frequency of 0.2 or 0.1 Hz. Ramp data were acquired at a frequency of 4 kHz after filtering at 1 kHz.
For measurements of divalent cation permeabilities, currents were fully activated in an NMDG ϩ solution containing the divalent cation (X 2ϩ ) at a concentration of 20 mM. The solution was then exchanged for a Na ϩ solution. The relative permeability was then calculated according to the following equation: Combined Patch Clamp and Ca 2ϩ Measurements-In some experiments, perforated patch recordings were combined with fluorometric recordings of [Ca 2ϩ ] i . Cells were loaded with fura-2/AM (Molecular Probes, Leiden, Netherlands), and Ca 2ϩ measurements were performed using a monochromator/photomultiplier-based system (TILL Photonics, Grä felfing, Germany). Cells were viewed with a ϫ40 oil immersion objective. Fluorescence was excited alternately at 360 and 390 nm for fura-2, and light output was measured from an adjustable rectangular aperture containing the cell of interest. Background fluorescence was obtained from the same region without a cell. The output of the photomultiplier was recorded simultaneously with the current data using Pulse software (HEKA) and the fluorescence signals together with the calculated ratio (F 360 /F 390 ) displayed online.

RESULTS
Properties of Human TRPV4 -The properties of human TRPV4 closely resembled those of the murine ortholog, when studied either after induction of expression in the T-REx-V4 cell line or following transient transfection in HEK293 cells. In whole cell recordings of membrane currents, many cells displayed spontaneous current activity that decayed rapidly in the open whole cell recording mode ( Fig. 1D) but was more stable in perforated patch recordings (Fig. 1A). TRPV4 responds both to decreases in the extracellular osmolarity (11)(12)(13)(14) and to the application of phorbol esters (15). In perforated patch recordings, application of hypotonic solutions resulted within a few seconds in an increase in membrane current that reached a maximum and then decayed (Fig. 1A). Both spontaneous and osmosensitive currents displayed a characteristic outwardly rectifying current-voltage (IV) relation (Fig. 1B). Scaling the IV relation obtained at an osmolarity of 320 mosmol/liter in Fig. 1B reveals the similarity in shape of the IV relations before and during stimulation with the hypotonic solution. The currents differ only in the reversal potential, which is more negative because of the stronger influence of background currents at the lower amplitudes in 320 mosmol/liter. In the open whole cell mode, responses to changes in osmolarity were weak (data not shown). In contrast, responses to phorbol esters were large, even after run-down of spontaneous activity (Fig. 1D), and displayed an identical IV relation to spontaneous currents and currents activated by hypotonic solutions (Fig. 1E).
Divalent Cation Permeability of TRPV4 -Previous studies have shown that TRPV4 is somewhat more permeable to Ca 2ϩ than to Na ϩ (12,15). To estimate the permeability of human TRPV4 to Ca 2ϩ , Ba 2ϩ , and Sr 2ϩ , currents were activated by 4␣-PMA (1 M) in a Na ϩ -free (NMDG ϩ ) solution containing the divalent cation at a concentration of 20 mM. After reaching a maximum, the extracellular solution was changed to an Na ϩ solution. As the influence of background currents became smaller, the reversal potential in the divalent cation solution shifted to more positive potentials during current activation with each of the three cations, before reaching a maximum value (Fig. 1F). Because currents evoked in Ba 2ϩ were much smaller than those in Ca 2ϩ and Sr 2ϩ , the reversal potential in many cells was more negative than those in Ca 2ϩ and Sr 2ϩ . However, in cells with currents of comparable amplitude to those in Ca 2ϩ and Sr 2ϩ more positive reversal potentials were obtained (e.g. Fig. 1F), and only these were used for calculation of the permeability ratio. From the reversal potential in the presence of the divalent cation and the shift in reversal potential upon application of the Na ϩ solution, the permeability ratios P Ca /P Na , P Sr /P Na , and P Ba /P Na were calculated to be 9.3 Ϯ 0.5 (n ϭ 4), 8.7 Ϯ 0.4 (n ϭ 6), and 7.0 Ϯ 0.4 (n ϭ 5), respectively.
Effect of Removal of Permeant Cations on TRPV4 -Replacement of both extracellular Na ϩ and Ca 2ϩ by the large cation NMDG ϩ abolished both the inward current activated by hypotonic solutions (Fig. 1, A and C) or 4␣-PMA and spontaneous currents (data not shown). In addition, as observed for mouse TRPV4 (12), application of the NMDG ϩ solution also resulted in a surprising reduction in the outward current rather than the expected increase (Fig. 1, A and C). This suggests that the channel activity is dependent on either extracellular Na ϩ or Ca 2ϩ . Large reductions in both inward and outward currents also occurred upon application of a Na ϩ -containing, nominally Ca 2ϩ -free extracellular solution during current activation (as shown in Fig. 1D for a phorbol ester-activated current), suggesting that Ca 2ϩ may play a role in TRPV4 activation.
Spontaneous Activity of TRPV4 Is Ca 2ϩ -dependent-The experiments described above suggest that currents through TRPV4 are Ca 2ϩ -dependent. To investigate the regulatory role of Ca 2ϩ in the maintenance of TRPV4 activity, we studied the effect of removal of extracellular Ca 2ϩ on spontaneous channel activity. For this, we combined perforated patch recordings of spontaneous whole cell currents with fluorometric recordings of [Ca 2ϩ ] i using fura-2. Cells were clamped at 0 mV to reduce Ca 2ϩ entry through the spontaneously active channels and prevent steady increases in [Ca 2ϩ ] i that occur at negative holding potentials. Currents were estimated by ramps applied at 5or 10-s intervals. In cells that showed spontaneous TRPV4 activity, replacement of the standard extracellular solution with a nominally Ca 2ϩ -free solution resulted in a complete loss of spontaneous channel activity and a parallel decrease in [Ca 2ϩ ] i (see Fig. 2, A and B and Fig. 3A). The rate of reduction of currents and [Ca 2ϩ ] i was much slower than the solution exchange, which was complete within a few seconds. Readdition of Ca 2ϩ to the extracellular solution resulted in a transient increase in [Ca 2ϩ ] i and currents to levels higher than those before Ca 2ϩ removal ( Fig. 2A). Thereafter, both parameters returned to stable levels similar to those before Ca 2ϩ removal. The similarity in the time course of the currents and the fura ratio upon removal of extracellular Ca 2ϩ suggests that spontaneous TRPV4 activity is dependent on an elevated [Ca 2ϩ ] i resulting from Ca 2ϩ entry through the channel.
We then tested the ability of other permeant divalent cations to substitute for Ca 2ϩ in the maintenance of channel activity. Neither Ba 2ϩ nor Sr 2ϩ (2 mM) were effective replacements for Ca 2ϩ at the same concentration. Replacement of Ca 2ϩ by either of these divalent cations (Ba 2ϩ , n ϭ 8; Sr 2ϩ , n ϭ 5) resulted in reductions of inward and outward currents similar to those seen on Ca 2ϩ removal (Fig. 2, C, D, and E). However, as evidenced by the shape of the IV relation, small TRPV4 currents persisted in Sr 2ϩ but not in Ba 2ϩ (Fig. 2, D and E).
Stimulation of TRPV4 by Ca 2ϩ Release from Intracellular Stores Indicates That Ca 2ϩ Acts at an Intracellular Site-As a permeant cation, Ca 2ϩ could influence channel activity at either an extracellular or an intracellular site. To test whether increases in intracellular Ca 2ϩ can restore TRPV4 activity in the absence of extracellular Ca 2ϩ , we raised [Ca 2ϩ ] i by releasing Ca 2ϩ from intracellular stores. As described above, removal of extracellular Ca 2ϩ abolished channel activity. After [Ca 2ϩ ] i had reached a stable level, in eight from nine cells addition of thapsigargin (1 M) or ionomycin (1 M) resulted in an increase in [Ca 2ϩ ] i and stimulation of an outwardly rectifying current with an identical shape of IV relation to the spontaneous current (Fig. 3, A and B). In Fig. 3, A  activation of an endogenous Ca 2ϩ -activated cation current like TRPM4 (30). In five of these cells, no current was activated by Ca 2ϩ release induced by either ionomycin or thapsigargin in the absence of extracellular Ca 2ϩ , and, in another one, a small current with different properties to those of TRPV4 was observed. The Ca 2ϩ signals observed in response to thapsigargin and ionomycin in control cells were of similar amplitudes to those in TRPV4-expressing cells (e.g. Fig. 3C). Hence, currents through TRPV4 can be potentiated by increases in [Ca 2ϩ ] i .
Stimulation of TRPV4 by Ionomycin in the Presence of Extracellular Ca 2ϩ -In the presence of extracellular Ca 2ϩ , application of ionomycin resulted in large increases in current (from 7.6 Ϯ 2.2 picoamperes/picofarad to 45.3 Ϯ 8.9 picoamperes/ picofarad at Ϫ100 mV, n ϭ 9) and large, often irreversible increases in [Ca 2ϩ ] i (data not shown). Fig. 3, D and E shows the effect of application of ionomycin (1 M) on a cell with large, spontaneous TRPV4 currents. Following addition of ionomycin to the extracellular solution, the current increased transiently before decaying to a level much lower than that of the spontaneous current (Fig. 3D). The IV relation of the ionomycinactivated current had a similar shape to spontaneous TRPV4 currents (Fig. 3E) and to currents activated by 4␣-PMA or reductions in extracellular osmolarity. In control cells, ionomycin did not activate any current in the presence of extracellular Ca 2ϩ (n ϭ 4). Thus, the large increases in [Ca 2ϩ ] i resulting from ionomycin application in the presence of extracellular Ca 2ϩ revealed a dual effect of Ca 2ϩ , an initial potentiation followed by inhibition.

Ca 2ϩ Dependence of Current Activation by Hypotonic Solutions and Phorbol
Esters-To test the role of Ca 2ϩ during current activation by hypotonic solutions or phorbol esters, the stimulus was applied in the absence of extracellular Ca 2ϩ , followed by Ca 2ϩ readdition to the extracellular solution in the continued presence of the stimulus.
In a first series of experiments, we studied the activation of TRPV4 by hypotonic solutions in a Ca 2ϩ -free, Ba 2ϩ -containing solution (Fig. 4A). Reduction of the osmolarity from 300 to 200 mosmol/liter resulted in a slow increase in outward current and a comparatively small increase in inward current. Replacement of Ba 2ϩ by Ca 2ϩ when activation in Ba 2ϩ was near maximal led to an increase in both inward and outward currents (Fig. 4, A  and B). The differences in the time course of current activation by a reduction in osmolarity from 300 to 200 mosmol/liter in Ba 2ϩ and in Ca 2ϩ solutions are illustrated in Fig. 4C. In Ba 2ϩ , currents increased slowly to a plateau and decreased only upon removal of the osmotic stimulus. In contrast, in Ca 2ϩ , currents increased more rapidly to a maximum and began to decay in the continued presence of the stimulus. Following the response, the current decreased to values below those of spontaneous activity before osmotic stimulation and then recovered slowly. The current response in Ca 2ϩ resembles that to ionomycin in the presence of extracellular Ca 2ϩ .
Similar results to those in the presence of Ba 2ϩ were ob- tained in the absence of extracellular Ca 2ϩ . In a nominally Ca 2ϩ -free solution, application of 4␣-PMA (1 M) resulted in a slow but small increase in outward and inward currents to a stable level (see Fig. 4D and Fig. 7A). Addition of Ca 2ϩ (0.5 or 2 mM) to the extracellular solution resulted in a very large, brief, and transient increase in current with the characteristic IV relation of TRPV4 (Fig. 4E). Because of the very rapid current change and the length of the interval between the ramps, the maximum current during the potentiation was often underestimated. After the potentiation of currents upon Ca 2ϩ addition, the TRPV4 current rapidly declined to a low level or disappeared completely. On average, the current in 0.5 mM Ca 2ϩ was about 5-fold larger than that in 4␣-PMA prior to Ca 2ϩ addition (see Fig. 7C).
Localization of a Calmodulin Binding Site in the C Terminus of TRPV4 -The data presented above suggest that intracellular Ca 2ϩ potentiates TRPV4 by acting at an intracellular site and that, in agreement with a previous study (15), Ca 2ϩ also has an inhibitory effect on channel activity. Regulatory actions of intracellular Ca 2ϩ on ion channels are often indirect and mediated by the binding of Ca 2ϩ -CaM to domains in the intracellular C-terminal tail. Direct interactions of Ca 2ϩ are more rare and require a Ca 2ϩ binding pocket. It has been shown previously (28) for another member of the TRPV subfamily, TRPV6, that Ca 2ϩ can regulate channel activity by binding to CaM. The Ca 2ϩ -CaM complex interacts with a distinct site in the intracellular C terminus, facilitating the inactivation of the channel. However, other components of Ca 2ϩ -dependent inactivation were unaffected. In line with this, three sites have been identified in TRPV5 that are involved in Ca 2ϩ -dependent inactivation (29,31). One is located in the intracellular channel loop between S 2 and S 3 (31), and the other two are located in the C terminus (29). The C-terminal Ca 2ϩ -CaM binding motif of TRPV6 does not correspond to any of the known CaM binding consensus sequences but bears some similarity to the protein kinase C-␣ pseudosubstrate site (28). This site includes a number of clustered basic amino acids that are thought to stabilize the complex with CaM. Interestingly, we identified a similar positively charged ␣-helical stretch VGRLRRDRWSSVVPRVV in the C-terminal sequence of TRPV4 starting at position 814 (Fig. 5A).
To investigate the CaM binding properties of this area, we expressed overlapping peptides of the distal TRPV4 C terminus (Fig. 5A) as glutathione S-transferase fusion proteins in E. coli and tested for their interaction with Sepharose-bound CaM (Fig. 5B). Peptides that include the cluster of positively charged amino acids (C1, C3, C5, and C6) bind to CaM in a Ca 2ϩ -dependent manner (Fig. 5B). Peptides lacking this domain (C2 and C4) did not bind to CaM. The smallest of the peptides that binds to CaM (C6) limits the CaM binding domain to a region between amino acids 812 and 831.
To investigate the Ca 2ϩ dependence of CaM binding, the interaction of the C5 peptide with CaM was measured at Ca 2ϩ concentrations from 0.1 M to 100 mM (Fig. 6). The calmodulinbound peptide fraction was quantified by measuring the relative band intensities of the SDS-PAGE gel. In 10 mM EGTA no pull-down was detectable. At a [Ca 2ϩ ] of 0.1 M some peptide binding was already observed, and this increased with [Ca 2ϩ ] to a maximum at about 100 M. Half-maximal CaM binding was observed at a [Ca 2ϩ ] i of ϳ200 nM. As also described for the CaM binding site in TRPV6 (28), higher Ca 2ϩ concentrations resulted in a decrease in CaM binding.
To further define the structural features that are essential for the CaM interaction, point mutations were inserted into the C3 and C5 peptides. The corresponding peptides (Fig. 5A) bearing the deletion of central tryptophan (W822A), a RWSS to AASA exchange (821AASA) or the replacement of the five basic arginine residues with glutamate at position 816 (5R/E) were subjected to a similar pull-down experiment with CaM-Sepharose. All of these mutations led to a loss of Ca 2ϩ -dependent CaM binding (Fig. 5C). Equal Ca 2ϩ -CaM binding properties were observed for C3 and C5 fragments. Thus, the CaM binding site within the C5 peptide could be localized to the tryptophan at position 822 and its positively charged vicinity.
Ca 2ϩ Dependence of C-terminal TRPV4 Mutants-To investigate the functional role of the C terminus and, in particular, the C-terminal CaM binding site, we used two of the mutants of TRPV4 described above (TRPV4 -5R/E and TRPV4-W822A) and two additional mutants. The latter were a deletion mutant lacking the last 70 amino acids including the CaM binding domain (TRPV4-K801term) and a mutant in which the two serines at positions 823 and 824 were mutated to alanine (TRPV4-S823A/S824A). These residues are part of a canonical phosphorylation motif for protein-serine/threonine kinases (32) that is contained within the CaM binding site and might be subject to phosphorylation as described for TRPV6 (28). Following transient transfection in HEK293 cells, all of the mutants, C-terminally fused to green fluorescent protein, displayed a subcellular expression pattern indistinguishable from that of the wild-type channel, with expression in the cell membrane and in intracellular compartments. In whole cell recordings, cells expressing all of the mutants showed spontaneous TRPV4 currents which decayed during the course of the recording like wild-type TRPV4. Removal of extracellular Ca 2ϩ led to a reduction of the remaining TRPV4 current in cells expressing TRPV4-wt and TRPV4-S823A/ S824A but not in cells expressing TRPV4-K801term, TRPV4 -5R/E, or TRPV4-W822A. We then compared the ability of Ca 2ϩ to potentiate 4␣-PMA-activated currents in the mutants with that for TRPV4-wt (see above). For all channels, application of 4␣-PMA in Ca 2ϩ -free solutions resulted in a slow increase in current, as described above. For TRPV4-wt and TRPV4-S823A/ S824A, addition of 0.5 mM Ca 2ϩ resulted in about a 5-fold increase in current at both Ϫ100 and ϩ100 mV (Fig. 7, A and  C). In contrast, TRPV4-K801term, TRPV4 -5R/E, and TRPV4-W822A showed no potentiation (Fig. 7B), and, indeed, TRPV4 -5R/E showed a clear but poorly reversible inhibition upon addition of Ca 2ϩ . Similar results were observed with 2 mM Ca 2ϩ in the extracellular solution. The results are summarized in Fig. 7C. It should be noted that the current density for TRPV4 -5R/E after 4␣-PMA addition in Ca 2ϩ -free solutions was much higher than that for TRPV4-wt and the other mutants. For all isoforms, raising the extracellular Ca 2ϩ concentration resulted in a change in the shape of the current voltage relationship increasing outward rectification as described recently by Voets et al. (33).
If Ca 2ϩ entry potentiates the current response of TRPV4 by a Ca 2ϩ -CaM-mediated interaction with the C-terminal domain, mutations that prevent the interaction of CaM should slow the rate of channel activation. We therefore compared the time course of current activation by 4␣-PMA in solutions containing 2 mM Ca 2ϩ for two mutants that do not bind CaM with the response of TRPV4-wt. The mutants TRPV4-K801term and TRPV4-W822A both responded to 4␣-PMA with a transient increase in current, but current activation was much slower than for TRPV4-wt (Fig. 7D). We quantified the rise time of the current by measuring the time required for the current to increase from 10 to 90% of the maximum. TRPV4-K801term and TRPV4-W822A had similar rise times, but these were significantly longer (on average about 3-fold) than that of TRPV4-wt (Fig. 7E). Interestingly, raising the Ca 2ϩ buffer capacity of the intracellular solution by increasing the EGTA concentration in the pipette solution from 1 to 10 mM led to a similar slowing of current activation to that in the two mutants (Fig. 7E).

DISCUSSION
In this study we have shown that, in addition to having a feedback inhibitory effect, intracellular Ca 2ϩ potentiates TRPV4 currents and plays an important role in accelerating and amplifying the current response to hypotonic solutions and phorbol ester agonists. The potentiatory effect of Ca 2ϩ occurs through an action at an intracellular site in the C terminus of the channel protein and is mediated by CaM binding.
Ca 2ϩ regulates both the spontaneous activity of TRPV4 and the response of the channel to stimulation by phorbol esters and hypotonic solutions as evidenced by the reduction in current upon Ca 2ϩ removal. Our data suggest that this effect of Ca 2ϩ is mediated by an intracellular site and that Ca 2ϩ entering the cell via the channel potentiates TRPV4 currents. The evidence for an intracellular site of action includes the ability of Ca 2ϩ release from intracellular stores to restore TRPV4 currents in the absence of extracellular Ca 2ϩ , the slowing of chan-nel activation with intracellular solutions with stronger Ca 2ϩ buffering, and the loss of potentiation following mutation of putative intracellular domains. Ba 2ϩ was unable to potentiate TRPV4, and Sr 2ϩ was much less effective than Ca 2ϩ in supporting spontaneous channel activity, a sequence of effectiveness shared by a number of Ca 2ϩ -binding proteins including CaM (34). For TRPV4 it is likely, though not proven, that heat and osmotic stimulation lead to the generation of a common messenger that activates the channel and that phorbol esters mimic this endogenous activator (17). Evidence that both osmotic stimuli and phorbol esters can increase TRPV4 currents without an increase in [Ca 2ϩ ] i is an indication that the effect of Ca 2ϩ is as a potentiator rather than a primary activator of the channel.
The C-terminal CaM binding domain involved in Ca 2ϩ -dependent potentiation of TRPV4 is similar but not identical to the domain in TRPV6 that mediates slow channel inactivation (28). Mutations in the domain of TRPV4 like those that reduce inactivation of TRPV6 led to a loss of Ca 2ϩ -dependent potentiation. These mutants also no longer bind CaM, suggesting that the effects of Ca 2ϩ involve its binding to CaM and the subsequent interaction of Ca 2ϩ -CaM with the C-terminal domain. Attempts to pharmacologically prevent the action of CaM with CaM antagonists were unsuccessful because of the instability of recordings in the presence of the substances used (calmidazolium, trifluoperazine, and W7; data not shown). For TRPV6, the interaction of CaM is competitively regulated by protein kinase C-mediated phosphorylation of a threonine residue in the CaM binding domain (28). TRPV4 also has a consensus sequence for protein-serine/threonine kinase phosphorylation within the CaM binding domain. However, mutation of two serines that are prospective phosphorylation sites within this domain in the mutant TRPV4-S823A/S824A were without effect on Ca 2ϩ -dependent potentiation. Thus, the regulation of TRPV4 differs from that of TRPV6 in at least two important aspects. There is the surprising finding that CaM binding domains in similar regions of related proteins have opposite effects on channel function. In addition, for TRPV6, Ca 2ϩ -CaM binding is modified by phosphorylation, but this is not the case for TRPV4.
Activation or potentiation of Ca 2ϩ -permeable channels by Ca 2ϩ is uncommon, particularly for channels in the plasma membrane, because of the positive feedback effect and danger of cellular Ca 2ϩ overload. Other Ca 2ϩ -permeable channels that are potentiated by Ca 2ϩ are those involved in Ca 2ϩ release from intracellular stores and include ryanodine receptors and inositol 1,4,5-trisphosphate receptors. As for TRPV4, the activity of these channels is tightly controlled by Ca 2ϩ -dependent negative feedback mechanisms that involve CaM (35). Polycystin-2 and polycystin-L, cation channels that may be involved in Ca 2ϩ release from intracellular stores or in Ca 2ϩ entry through the plasma membrane, are also activated by increases in cytoplasmic Ca 2ϩ (36). Among the TRP channels, the response of TRPC5, which is modestly Ca 2ϩ -permeable, is potentiated by increases in [Ca 2ϩ ] i (37,38). It has also been shown recently (39,40) that intracellular ADP-ribose-mediated activation of TRPM2, a Ca 2ϩ -permeable nonselective cation channel, is also dependent on Ca 2ϩ entry through the channel (24). Unlike TRPV4, TRPM2 does not appear to have an inhibitory mechanism, and, thus, uncontrolled positive feedback of Ca 2ϩ entry leads to cell death. TRPM4, the other TRP channel that is activated by intracellular Ca 2ϩ , is only poorly Ca 2ϩ -permeable (30). More distantly related channels that, like TRPV4, are activated by a Ca 2ϩ -CaM interaction with a C-terminal domain, include the small conductance Ca 2ϩ -activated K ϩ channels (41). FIG. 7. TRPV4 mutants that no longer bind calmodulin lack Ca 2؉ -dependent potentiation and activate more slowly. A, potentiation of TRPV4-wt by Ca 2ϩ . Currents were activated by 4␣-PMA in the absence of Ca 2ϩ followed by Ca 2ϩ readdition at a concentration of 0.5 mM. B, mutant TRPV4-W822A lacks potentiation by Ca 2ϩ . C, effect of addition of 0.5 mM Ca 2ϩ on TRPV4 and TRPV4 mutants. Bars indicate the current at Ϫ100 mV (downward) and ϩ100 mV (upward) in 0.5 mM Ca 2ϩ relative to the absolute value in 4␣-PMA in 0 Ca 2ϩ immediately prior to Ca 2ϩ addition for TRPV4-wt (n ϭ 10) and the mutants TRPV4 -5R/E (n ϭ 16), TRPV4-S823A/S824A (n ϭ 3), TRPV4-K801term (n ϭ 8), and TRPV4-W822A (n ϭ 5). D, activation of TRPV4 in the presence of Ca 2ϩ is slowed in the mutants TRPV4-K801term and TRPV4-W822A. Currents were activated by 4␣-PMA (1 M) in an extracellular solution containing 2 mM Ca 2ϩ . E, rise time of currents (time from 10 to 90% of the maximum current) through TRPV4-wt with 1 (n ϭ 6) or 10 mM intracellular EGTA (n ϭ 4) and the mutants TRPV4-K801term (n ϭ 6) and TRPV4-W822A (n ϭ 4).
It is likely that Ca 2ϩ -CaM, by binding to the C-terminal domain, induces a conformational change in the TRPV4 channel protein resulting in increased channel activity. Similar mechanisms have been proposed for the activation of a number of channel types, including Ca 2ϩ -activated K ϩ channels and cyclic nucleotide-gated channels, whose activity is controlled by intracellular mediator (Ca 2ϩ , Ca 2ϩ -CaM, or cAMP/cGMP) binding in the C terminus. We could only demonstrate CaM binding to the C terminus of TRPV4 in the presence of Ca 2ϩ . In this respect, the channel differs from small conductance Ca 2ϩactivated K ϩ channels where Ca 2ϩ -dependent and Ca 2ϩ -independent CaM binding sites in their C termini are involved in Ca 2ϩ -induced conformational changes (41).
The biphasic current responses following Ca 2ϩ entry suggest that smaller increases in [Ca 2ϩ ] i potentiate TRPV4, but larger increases are inhibitory. Data on the Ca 2ϩ dependence of CaM binding indicate that Ca 2ϩ -CaM binds to TRPV4 at submicromolar concentrations that will readily be reached at the intracellular membrane surface close to the channel mouth. In a previous study (15), an inhibitory effect of intracellular Ca 2ϩ on 4␣-phorbol dideocanoate-activated currents has been reported for TRPV4 with an IC 50 in whole cell experiments of 406 nM. We also obtained evidence for an inhibitory effect of Ca 2ϩ . Large increases in [Ca 2ϩ ] i , resulting from large currents or the application of ionomycin in the presence of extracellular Ca 2ϩ , resulted in biphasic changes in current. A transient increase in current was followed by inhibition to levels lower than those before stimulation. An additional observation that could be explained by Ca 2ϩ -dependent inhibition of currents is the effect of Ca 2ϩ readdition following a period in Ca 2ϩ -free solutions. This resulted in a current overshoot to levels much higher than the spontaneous activity before Ca 2ϩ removal followed by return to a similar level. This result suggests that currents are partially inhibited by the elevated Ca 2ϩ levels in cells showing spontaneous channel activity and that a decrease in [Ca 2ϩ ] i results in channel disinhibition. Ca 2ϩ -dependent inhibition was clearly observed in the mutant TRPV4 -5R/E, which no longer bound CaM, indicating that the CaM binding site is not involved in channel inhibition. Furthermore, in a few cells expressing TRPV4-K801term or TRPV4-W822A inhibition was observed upon addition of Ca 2ϩ . However, 4␣-PMA-activated currents in most cells expressing these mutants were smaller than for TRPV4-wt or TRPV4 -5R/E, and, in the absence of potentiation, Ca 2ϩ addition will result in smaller increases in [Ca 2ϩ ] i that may be insufficient to inhibit the channel.
The dual regulation of TRPV4 by Ca 2ϩ will shape the cellular Ca 2ϩ response to channel stimuli. The combination of Ca 2ϩ -dependent potentiation and inactivation will result in transient responses with a rapid rising phase. The positive feedback effect of Ca 2ϩ entry on channel activation could result in regenerative Ca 2ϩ responses. Indeed, for mouse TRPV4 we previously observed responses to osmotic stimulation that consisted of a slow Ca 2ϩ increase to a threshold level followed by a very rapid Ca 2ϩ rise (12), and we have seen similar responses of human TRPV4 to osmotic stimuli and phorbol esters. 2 The dual regulation could also support oscillatory responses. Recovery from Ca 2ϩ -dependent inhibition will determine the time from the end of one response until [Ca 2ϩ ] i is sufficient to trigger TRPV4 activity. Ca 2ϩ entry through the open channels will then trigger channel activation. Indeed, oscillatory Ca 2ϩ responses have been observed in cells expressing high levels of TRPV4 (13) and spontaneous increases in current and [Ca 2ϩ ] i observed during perforated patch recordings. 3 Caution should, however, be exercised in applying these results to native tis-sues where expression levels of TRPV4 are likely to be much lower than those in expression systems. Without knowing the physiological role of the channel, it is too early to speculate further on the importance of Ca 2ϩ -dependent potentiation.
Further studies are necessary to determine the physiological role and activation mechanism of this channel. However, TRPV4 is unusual among plasma membrane Ca 2ϩ -permeable channels with moderate to high Ca 2ϩ permeabilities in being potentiated by increases in intracellular Ca 2ϩ concentration. Deleterious increases in intracellular Ca 2ϩ resulting from Ca 2ϩ -activated Ca 2ϩ entry are prevented by Ca 2ϩ -dependent negative feedback.