Determinants of 4α-Phorbol Sensitivity in Transmembrane Domains 3 and 4 of the Cation Channel TRPV4*

TRPV4, a Ca2+-permeable member of the vanilloid subgroup of the transient receptor potential (TRP) channels, is activated by cell swelling and moderate heat (>27 °C) as well as by diverse chemical compounds including synthetic 4α-phorbol esters, the plant extract bisandrographolide A, and endogenous epoxyeicosatrienoic acids (EETs; 5,6-EET and 8,9-EET). Previous work identified a tyrosine residue located in the first half of putative transmembrane segment 3 (TM3) as a crucial determinant for the activation of TRPV4 by its most specific agonist 4α-phorbol 12,13-didecanoate (4α-PDD), suggesting that 4α-PDD interacts with the channel through its transmembrane segments. To obtain insight in the 4α-PDD-binding site and in the mechanism of ligand-dependent TRPV4 activation, we investigated the consequences of specific point mutations in TM4 on the sensitivity of the channel to different chemical and physical stimuli. Mutations of two hydrophobic residues in the central part of TM4 (Leu584 and Trp586) caused a severe reduction of the sensitivity of the channel to 4α-PDD, bisandrographolide A, and heat, whereas responses to cell swelling, arachidonic acid, and 5,6-EET remained unaffected. In contrast, mutations of two residues in the C-terminal part of TM4 (Tyr591 and Arg594) affected channel activation of TRPV4 by all stimuli, suggesting an involvement in channel gating rather than in interaction with agonists. Based on a comparison of the responses of WT and mutant TRPV4 to 4α-PDD and different 4α-phorbol esters, we conclude that the length of the fatty acid moiety determines the ligand binding affinity and propose a model for the interaction between 4α-phorbol esters and the TM3/4 region of TRPV4.

TRPV4, a Ca 2؉ -permeable member of the vanilloid subgroup of the transient receptor potential (TRP) channels, is activated by cell swelling and moderate heat (>27°C) as well as by diverse chemical compounds including synthetic 4␣-phorbol esters, the plant extract bisandrographolide A, and endogenous epoxyeicosatrienoic acids (EETs; 5,6-EET and 8,9-EET). Previous work identified a tyrosine residue located in the first half of putative transmembrane segment 3 (TM3) as a crucial determinant for the activation of TRPV4 by its most specific agonist 4␣-phorbol 12,13-didecanoate (4␣-PDD), suggesting that 4␣-PDD interacts with the channel through its transmembrane segments. To obtain insight in the 4␣-PDD-binding site and in the mechanism of ligand-dependent TRPV4 activation, we investigated the consequences of specific point mutations in TM4 on the sensitivity of the channel to different chemical and physical stimuli. Mutations of two hydrophobic residues in the central part of TM4 (Leu 584 and Trp 586 ) caused a severe reduction of the sensitivity of the channel to 4␣-PDD, bisandrographolide A, and heat, whereas responses to cell swelling, arachidonic acid, and 5,6-EET remained unaffected. In contrast, mutations of two residues in the C-terminal part of TM4 (Tyr 591 and Arg 594 ) affected channel activation of TRPV4 by all stimuli, suggesting an involvement in channel gating rather than in interaction with agonists. Based on a comparison of the responses of WT and mutant TRPV4 to 4␣-PDD and different 4␣-phorbol esters, we conclude that the length of the fatty acid moiety determines the ligand binding affinity and propose a model for the interaction between 4␣-phorbol esters and the TM3/4 region of TRPV4. TRPV4 2 is a widely expressed cation channel of the "transient receptor potential" (TRP) superfamily (1)(2)(3)(4)(5)(6). TRP chan-nels are intrinsic membrane proteins with six putative transmembrane spans (TM) and a cation-permeable pore region formed by a short hydrophobic stretch between TM5 and TM6 (7). TRPV4 functions as a Ca 2ϩ entry channel that exhibits a surprising gating promiscuity. The channel can be activated by physical stimuli (cell swelling and moderate warmth (Ͼ27°C) (3, 8 -13)), by the synthetic phorbol ester 4␣-PDD (14), by the active compound of an extract from the plant Andrographis paniculata, bisandrographolide A (BAA) (15), and by EETs derived from arachidonic acid (AA) (3,16). Cell swelling activates TRPV4 by means of the phospholipase A 2 -dependent formation of AA, which is subsequently metabolized to various EETs via cytochrome P450 epoxygenases (2). Phorbol esters and heat operate via a distinct, phospholipase A 2 -and cytochrome P450 epoxygenase-independent pathway, which critically depends on an aromatic residue at the N terminus of TM3, Tyr 556 (referred to as Tyr 555 in a previous publication) (17). In mouse aorta endothelial cells, modulation of CYP2C expression enhanced the responses to AA and cell swelling, whereas responses to other stimuli remained unaffected (18). Relatively few reports describe the functional impact of TM1-4 segments on the regulation of the TRP channel function (6). The first insights came from a study on TRPV1 (19), suggesting that Tyr 511 in the intracellular loop between TM2 and TM3 interacts with the vanillyl moiety of capsaicin on the cytosolic face of the membrane. Nearby residues, such as Ser 512 or Arg 491 , may interact with capsaicin via hydrogen bonds, whereas lipophilic residues in TM3 can be involved in stabilization via hydrophobic interactions with the aliphatic moiety of capsaicin within the plane of the membrane (19). This model was refined by Gavva et al. (20), who demonstrated that Tyr 550 , Trp 549 , and Met 547 in TM4 might be involved in the interaction with the vanilloid moiety, whereas the aliphatic tail of capsaicin binds to Tyr 511 . Nevertheless, these two models do not explain the loss of capsaicin sensitivity by mutations of N-and C-terminal residues, Arg 114 and Asp 761 , in TRPV1 (21,22) and still require additional biochemical and structural data for validation.
In the present study, we identify residues in TM4 of TRPV4 that are critically involved in ligand binding. Moreover, by comparing the effects of 4␣-phorbol with different aliphatic side chains, we obtained the first insights into the orientation of these ligands and their interaction with TM3/TM4 region of TRPV4.
Site-directed Mutagenesis-All mutants were obtained by the standard PCR overlap extension method (23) using mTRPV4 from pCAGGS/IRES-GFP vector. Accuracy of all mutant sequences was verified by sequencing. Expressions of mutant channels were verified by cell surface biotinylation and immunodetection with anti-TRPV4 antibodies (see supplemental Fig. 1).
Electrophysiological Recordings-Whole-cell membrane currents were measured with an EPC-10 (HEKA Elektronik, Lambrecht, Germany) using ruptured patches with samples rate of 20 kHz, and currents were filtered at 2.9 kHz. Patch electrodes had a DC resistance between 2 and 4 megaohms when filled with intracellular solution. An Ag-AgCl wire was used as a reference electrode. Capacitance and access resistance were mon-itored continuously. Between 50 and 70% of the series resistance was electronically compensated to minimize voltage errors. We applied a ramp protocol consisting of a voltage step from the holding potential of 0 mV to Ϫ100 mV followed by a 400-ms linear ramp to ϩ100mV. This protocol was repeated every 2 s.
Step protocol consists out of 20-mV steps from Ϫ100 mV to ϩ100 mV. Mean cell membrane capacitance value was 7.65 Ϯ 0.1 picofarads for n ϭ 78 and very similar for all HEK cells; therefore, current densities are not calculated.
Measurement of Intracellular Ca 2ϩ Concentrations-Intracellular [Ca 2ϩ ] i was measured with a monochromator-based imaging system described in detail elsewhere (24). For every condition, at least 20 cells from at least three independent experiments were assayed. The calibration procedure was described previously in (24). For experiments with responses to heat, K eff for changes in temperature was corrected (25).
Data Analysis-Electrophysiological data were analyzed using the PATCHMASTER and FITMASTER programs (HEKA Elektronik). For statistical analysis and data display, the Origin 7.1 software package was used (OriginLab Corp., Northampton, MA). Dose-response relationships were fitted with a Hill equation of the form, where C represents the agonist concentration, EC 50 represents the concentration for half-maximal response, and n H represents the Hill coefficient. Data are expressed as mean Ϯ S.E., and significance (p Ͻ 0.01) between individual groups was tested using unpaired Student's t test.

RESULTS
Sequence alignment of the putative TM3 and TM4 of TRPV4 (residues 547-570) with the corresponding region of TRPV1 displayed a high level of similarity (60% homology/40% identity) (Fig. 1). Importantly, Leu 584 , Trp 586 , and Met 587 in TM4 of TRPV4 aligned with residues in TRPV1 that were previously described to contribute to vanilloid binding (20). Moreover, Asn 588 , Tyr 591 , and Arg 594 were also highly conserved in other TRPV members. To assess contribution of these residues to TRPV4 activation, we constructed a series of point mutations within TM4 and tested the response of the mutant channelexpressed HEK293 cells to the different known TRPV4 stimuli using intracellular Ca 2ϩ ([Ca 2ϩ ] i ) measurements and wholecell patch clamp recordings.
Mutations to the Central Part of TM4 Affect Responses to Phorbol Esters and Heat-When compared with wild type TRPV4, mutants L584M and W586A displayed significantly reduced [Ca 2ϩ ] i increases and inward TRPV4 currents in response to 5 and 10 M 4␣-PDD (Fig. 2, A, B, and D), together with an ϳ10-fold increase in EC 50 (Table 1). In contrast, single mutations to the nearby residues Met 587 and Asn 588 (mutants M587A and N588A) mutants had no obvious effect on the [Ca 2ϩ ] i responses to 4␣-PDD ( Fig. 2A and Table 1). However, the combined mutant W586A,M587A exhibited a higher EC 50 value than the W586A single mutant ( Fig. 2C and Table 1). APRIL 27, 2007 • VOLUME 282 • NUMBER 17

JOURNAL OF BIOLOGICAL CHEMISTRY 12797
Tyrosine 556 in TM3 was previously identified as a critical residue for activation of TRPV4 by 4␣-PDD and heat, based on the finding that substituting a nonaromatic amino acid at this position abolished the response to 1 M 4␣-PDD and heating to 42°C (17). However, we found that 4␣-PDD concentration of 5 M and higher evoked significant [Ca 2ϩ ] i responses and whole-cell currents in cells expressing Y556A and determined an EC 50 of 6.3 M (Fig. 2, A and B, and Table 1). Interestingly, combining mutation Y556A with mutations in TM4 led to a further decrease (mutant Y556A,M587A) or even complete loss (mutant Y556A,W586A) of 4␣-PDD sensitivity ( Fig. 2E and Table 1). Taken together, these results indicate that residues in both TM3 and TM4 cooperate in the 4␣-PDD-dependent stimulation of TRPV4.
BAA, derived from the plant A. paniculata and used in traditional Chinese medicine as an antiinflammatory, immunostimulant, and antihypertensive compound, was recently described as a new agonist of TRPV4 that activates the channel in a membrane-delimited manner (15). We tested whether the above described mutants with reduced 4␣-PDD sensitivity were also affected in their responses to BAA. We found that mutants L584M and W586A were fully unresponsive to BAA over a broad range of concentrations (Fig. 3, B and C). In contrast, the response of Y556A to BAA was indistinguishable from that of WT TRPV4, with an EC 50 of about 1 Ϯ 0.2 M (Fig. 3C).
In the absence of a clear stimulus, TRPV4 displays some basal activity that can be estimated from the spontaneous single-channel openings in cell-attached patches (8), the increased basal whole-cell currents shortly after establishment of the whole-cell configuration (12,26), and the higher basal Ca 2ϩ levels in TRPV4 transfected cells (8,16,17). We found that basal [Ca 2ϩ ] i levels in HEK cells expressing the mutants L584M, W586A, Y556A, Y556A,W586A, W586A,M587A, and Y556A,M587A were similar to that of nontransfected cells, which contrasts to the elevated basal [Ca 2ϩ ] i levels in TRPV4-transfected cells (Fig. 4A). Moreover, none of these mutants showed a [Ca 2ϩ ] i response to a heat stimulus of 42°C (Fig. 4B).
Two lines of evidence indicate that the strong impairment of the responses to 4␣-PDD and heat upon mutating Leu 584 ,

4␣-Phorbol-binding Site in TRPV4
Trp 586 , or Tyr 556 was not due to an effect on channel expression/targeting. First, cell surface biotinylation experiments confirmed that all these three mutant channels are expressed in the plasma membrane (supplemental Fig. 1). Second, and more importantly, mutants L584M, W586A, Y556A, Y556A,W586A, Y556A,M587A, and W586A,M587A all responded to hyposmotic cell swelling and to exogenously applied AA (10 M) or 5,6-EET (500 nM), and the amplitudes of the [Ca 2ϩ ] i responses and inward currents were similar to those of wild type TRPV4 (Fig. 5). Note that all mutants displayed similar IV relationships as WT-TRPV4, indicating that the pore properties are unaltered (supplemental Fig. 2). Taken together, these results indicate that residues Leu 584 and Trp 586 at the C-terminal part of TM4 are selectively involved in activation by 4␣-PDD, BAA, and heat.
Activation of TRPV4 by Different 4␣-Phorbol Esters-The finding that residues both at the intracellular end of TM3 (Tyr 556 ) and at the extracellular part of TM4 (Leu 584 and Trp 586 ) are crucial for 4␣-PDD activation may suggest that 4␣-PDD interacts with TRPV4 in a region between TM3 and TM4. This raises the question how 4␣-PDD is oriented with respect to the channel. To approach this question, we examined how the length of the aliphatic side chains attached to the 4␣-phorbol moiety determines TRPV4 activation. We compared potencies of 4␣-PDBu, 4␣-PMA, 4␣-PDA, and 4␣-phorbol (Fig. 6A) on activation of wild type and mutant TRPV4 channels. In the presence of extracellular Ca 2ϩ , application of all these 4␣-phorbol derivatives to HEK cells transfected with WT TRPV4 resulted in dose-dependent increases of [Ca 2ϩ ] i , albeit with lower potencies (i.e. higher EC 50 values) than for 4␣-PDD (Fig. 6B). No [Ca 2ϩ ] i increase was observed in nontransfected cells or when extracellular Ca 2ϩ was omitted (data not shown). Note that the steepness of the dose-response curves also differed significantly between the 4␣-phorbols, reflected in the variation in n H values (4.5 for 4␣-PDD, 1.8 for 4␣-PMA, and 0.9 for 4␣-PDBu). Although in theory, differences in n H may be indicative of altered cooperativity of ligand binding and/or altered number of binding sites, we do not think that we can make such conclusions from our data. Indeed, the steepness of the dose-response curves is not only determined by the channel-ligand interaction but also by factors such as Ca 2ϩ dye saturation and the time course of channel desensitization.
Since TM3-4 residues participate in 4␣-PDD sensitivity of TRPV4, we tested the potency of the different 4␣-phorbols to activate Y556A and W586A mutants. In comparison with WT TRPV4, mutant Y556A had a lower sensitivity to 4␣-PDD, 4␣-PDBu, and 4␣-PMA (Fig. 7, A and C, and Table 2). However, 4␣-PDA and 4␣-phorbol activated WT TRPV4 and the Y556A mutant with similar potency (Fig. 7C and Table 2), suggesting that Tyr 556 is only involved in channel activation by 4␣-phorbols with longer aliphatic chains. In contrast, mutant W586A not only displayed a significantly reduced sensitivity to both 4␣-PDD and 4␣-PDBu but was also fully unresponsive to 4␣-PDA, 4␣-PMA and 4␣-phorbol (Fig. 7, D and F, and Table  2). This indicates that this residue at the extracellular part of TM4 is involved in channel activation by all 4␣-phorbols.
Mutations to the C-terminal Part of TM4 Affect General Activation of TRPV4-In contrast to the above described mutants in the central part of TM4, mutating residues Tyr 591 and Arg 594 in the C-terminal part of TM4 did not reduce the basal channel activity as evidenced by the elevated [Ca 2ϩ ] i levels in HEK cells expressing the Y591A and R594A mutants. In fact, the basal    APRIL 27, 2007 • VOLUME 282 • NUMBER 17 [Ca 2ϩ ] i level in cells expressing the Y591A was significantly higher than in those expressing WT TRPV4 (Fig. 8A), suggesting a higher level of basal activity. This was confirmed in wholecell current measurements showing robust basal currents (mean amplitude at Ϫ80 mV for nontransfected ϳϪ39 Ϯ 8 pA; WT TRPV4 ϳϪ75 Ϯ 7 pA and Y591A transfected HEK cells ϳϪ191 Ϯ 25 pA and p Ͻ 0.01) for Y591A, which were blocked in a voltage-dependent manner by the TRPV4 blocker Ruthenium Red (2 M) (Fig. 8B). The increases in [Ca 2ϩ ] i and wholecell current measured in response to all tested stimuli (4␣-PDD, heat, cell swelling, AA, 5,6-EET) were always lower in cells expressing Y591A than in those expressing WT TRPV4 (Figs. 8,  C, D, and G, and 9). These lower responses are most likely a consequence of the higher basal activity of the Y591A mutant as the absolute [Ca 2ϩ ] i at the peak of the response to these different stimuli was not significantly different from those obtained with WT TRPV4 (Fig. 8H).

4␣-Phorbol-binding Site in TRPV4
In contrast, mutant R594A showed a normal response to heat stimulation (Fig. 8G) but was fully unresponsive to 4␣-PDD (Fig. 8, C  and D), cell swelling, AA, or 5,6-EET (Fig. 9). To test whether the positive charge was an important determinant of TRPV4 activation, Arg 594 was mutated to either a lysine or a glutamine. Although mutant R594Q showed an identical response profile as R594A, the R594K showed a significant response to 4␣-PDD, albeit with a significantly higher EC 50 than WT TRPV4 (Fig. 8, E and F, and Table 1), and normal responses to cell swelling, AA, or 5,6-EET (Fig. 9). These data indicate that a positively charged residue at position 594 is crucial for the detection and/or transduction of the chemical stimuli 4␣-PDD and 5,6-EET but dispensable for heat sensitivity.

DISCUSSION
Several members of the TRP superfamily act as chemosensors, activating in response to a multitude of endogenous and exogenous ligands. At present, relatively little is known about how these ligands interact with TRP channels and how this interaction regulates channel gating. In this study, we provide evidence suggesting that the TM3-4 region of TRPV4 forms an important site for channel activation by phorbol esters. In particular, we found that mutations at positions Tyr 556 in TM3 or Leu 584 and Trp 586 in TM4 strongly affected channel activation by the most potent and selective agonist 4␣-PDD.
Several possible mechanisms could explain a reduced sensitivity by 4␣-PDD or BAA in these TRPV4 mutants. Firstly, these mutants may endure a trafficking problem to the plasma membrane. However, this possibility can be excluded given that these mutants could still be normally activated by hypotonic cell swelling, AA, and 5,6-EET and that they could be detected as biotinylated proteins at the cell surface (supplemental Fig. 1). Secondly, 4␣-PDD binding may occur through an unknown accessory 4␣-PDD-binding protein. The effects of mutations in TM3-4 may then reflect an altered interaction between channel and accessory protein. Although we cannot formally exclude this possibility, we consider it as relatively unlikely. If we accept that TRPV4 is structurally related to six TM K ϩ channels, the crucial TM3-4 residues point toward each other at the center of the TM1-4 region, where interactions with accessory proteins are unlikely to occur. Thirdly, these residues

4␣-Phorbol-binding Site in TRPV4
may form the structural element necessary for the transduction of the agonist-binding signal to the opening of the pore. An indication against such a mechanism is that opening of the mutant channels is normal upon stimulation with other ligands, such as 5,6-EET or, in the case of Y556A, BAA. Therefore, at present, we prefer the fourth possible mechanism, namely that these residues in TM3 and TM4 are crucial for interaction with 4␣-PDD and possibly part of 4␣-PDD-binding site.
To obtain better insight in the activation mechanism of TRPV4 by 4␣-PDD, we compared the potency of 4␣-phorbol and 4␣-phorbol esters with different aliphatic side chains. For WT TRPV4, we found an inverse correlation between the length of the aliphatic side chains and the EC 50 value for channel activation (Fig. 6C), suggesting that a hydrophobic side chain enhances the interaction with the binding site. Interestingly, the Y556A mutant displayed a reduced sensitivity to stimulation by 4␣-phorbol esters with longer aliphatic side chains, such 4␣-PDD, 4␣-PMA, and 4␣-PDBu, whereas responses to 4␣-PDA and 4␣-phorbol remained unaffected. This suggests that Tyr 556 is involved in the interaction between TRPV4 and the aliphatic part of 4␣-phorbol esters. In contrast, mutating Trp 586 impaired channel activation by all 4␣-phorbols, suggesting that this residue interacts with the 4␣-phorbol moiety of 4␣-PDD. Based on these data, and in analogy to what has been proposed for the interaction between capsaicin and TRPV1 (19,20), we propose a tentative model in which the pocket between TM3 and TM4 constitutes a binding site for a 4␣-PDD (Fig. 10). In this model, 4␣-PDD is positioned such that the aliphatic side chains point toward the cytosol, where they can interact with Tyr 556 , whereas the 4␣-phorbol moiety is positioned closer to the extracellular side of the plasma membrane, where it can interact with residues Trp 586 and Leu 584 . The higher EC 50 values for double mutants Y556A,M587A and W586A,M587A when compared with the single mutants Y556A and W586A, is suggestive of an indirect role of Met 587 in the interaction with 4␣-phorbols, although more evidence is required to establish this point. Clearly, structural data will be required to validate the location of the binding sites in TRPV4 for 4␣-PDD and related compounds. The orientation of the ligand between TM3-4 of TRPV4 is similar to a putative model for a capsaicinbinding site of TRPV1 proposed by Gavva et al. (20) in which the aliphatic chain of capsaicin points toward the intracellular side of the membrane, whereas the vanilloid part is positioned closer to the extracellular side of the plasma membrane. Similarly, a recent study provides evidence for a direct interaction   APRIL 27, 2007 • VOLUME 282 • NUMBER 17 between the cooling agent menthol and the cold-and mentholsensitive channel TRPM8 (28). It should be stressed that although these models adequately explain the functional results, they are still speculative and require additional biochemical and structural information for validation.

4␣-Phorbol-binding Site in TRPV4
At room temperature (ϳ25°C), TRPV4 exhibits a basal activity that is manifested by an increased basal [Ca 2ϩ ] i level in TRPV4-overexpressing cells (17) (Fig. 4A). Interestingly, cells expressing the different mutants with reduced 4␣-PDD did not exhibit such elevated basal [Ca 2ϩ ] i levels and no longer responded to a heat stimulus. In contrast to other heat-sensitive TRPV channels, which are directly activated by temperature changes, heat activation of TRPV4 does not occur in cell-free inside-out patches, which may indicate that soluble messengers play a role in this process (9,13,27). One possibility would be that increases in temperature lead to the generation of a ligand activator of TRPV4. Our finding that several mutations in the TM3-4 region equally affect sensitivity to heat and 4␣-PDD would then suggest that the heat-induced messenger and 4␣-PDD have overlapping binding sites. We reasoned that sphingomyelin metabolites, which are generated upon heat shock and serve as messengers of this stress condition (29), could play the role of heat-inducible TRPV4 agonists, but direct application of sphingosine-1-phosphate, sphingosine, or ceramide did not evoke any response in TRPV4-expressing cells (data not shown). Moreover, we found that mutant R594A was unresponsive to all tested TRPV4 activation stimuli with the    whereas Met 587 is rather indirectly involved (yellow). Leu 584 and Trp 586 (orange) are highly important components in the interaction with BAA. Tyr 591 and Arg 594 (green) located at the C terminus of TM4 are rather involved in general openings mechanism of the channel.

4␣-Phorbol-binding Site in TRPV4
exception of heat stimulation. Thus, at this point, the mechanism underlying heat activation of TRPV4 remains elusive. Two conserved residues located closer to the C terminus of TM4 (Fig. 10) appear to play a more general role in activation of TRPV4. Mutation of Tyr 591 to an alanine resulted in a channel with high basal activity, as evidenced by the increased basal [Ca 2ϩ ] i and robust unstimulated whole-cell currents in cells expressing Y591A. This mutant channel exhibited significant but reduced responses to all tested stimuli, which most likely reflects the already elevated basal activity of the mutant channel. Mutating the highly conserved basic residue Arg 594 to alanine or glutamine impaired responsiveness of the TRPV4 channel to 4␣-PDD, cell swelling, AA, and 5,6-EET, whereas activation by moderate heat remains unchanged. Substituting a lysine for Arg 594 yielded a mutant channel that exhibited similar responses as WT TRPV4, suggesting that a charged residue at this position is required for the transduction of chemical stimuli.
In conclusion, we propose that the TM3-4 region plays a central role in the activation of TRPV4 by thermal and chemical stimuli and may form a possible interaction side for BAA and 4␣-phorbols.