IP3 Receptor Binds to and Sensitizes TRPV4 Channel to Osmotic Stimuli via a Calmodulin-binding Site*

Activation of the non-selective cation channel TRPV4 by mechanical and osmotic stimuli requires the involvement of phospholipase A2 and the subsequent production of the arachidonic acid metabolites, epoxieicosatrienoic acids (EET). Previous studies have shown that inositol trisphosphate (IP3) sensitizes TRPV4 to mechanical, osmotic, and direct EET stimulation. We now search for the IP3 receptor-binding site on TRPV4 and its relevance to IP3-mediated sensitization. Three putative sites involved in protein-protein interactions were evaluated: a proline-rich domain (PRD), a calmodulin (CaM)-binding site, and the last four amino acids (DAPL) that show a PDZ-binding motif-like. TRPV4-ΔCaM-(Δ812–831) channels preserved activation by hypotonicity, 4α-phorbol 12,13-didecanoate, and EET but lost their physical interaction with IP3 receptor 3 and IP3-mediated sensitization. Deletion of a PDZ-binding motif-like (TRPV4-ΔDAPL) did not affect channel activity or IP3-mediated sensitization, whereas TRPV4-ΔPRD-(Δ132–144) resulted in loss of channel function despite correct trafficking. We conclude that IP3-mediated sensitization requires IP3 receptor binding to a TRPV4 C-terminal domain that overlaps with a previously described calmodulin-binding site.

TRPV4 channel activity can be sensitized by co-application of different stimuli (9,21,22) or participation of different cell signaling pathways (14), suggesting the presence of different regulatory sites. In this sense, several proteins have been proposed to modulate TRPV4 subcellular localization and/or function: microtubuleassociated protein 7 (23), calmodulin (CaM, (24)), with no lysine protein kinases (25), and PACSIN3 (26). We have also recently reported a close functional and physical interaction between the inositol trisphosphate receptor 3 (IP 3 R3) and TRPV4 that sensitizes the latter to the mechano-and osmotransducing messenger 5Ј-6Ј-epoxieicosatrienoic acid (EET) (14).
In this study, we reveal that IP 3 -mediated sensitization of TRPV4 to EET requires IP 3 R3 binding to TRPV4 and identify the IP 3 R3-binding site in a C-terminal region of the TRPV4 that coincides with a previously described CaM-binding site (24). Besides, our study has provided further evidence of the importance of the TRPV4 N terminus in channel gating as deletion of a proline-rich domain prevents channel activation by different stimuli despite correct localization and oligomerization.

EXPERIMENTAL PROCEDURES
Generation of TRPV4 Mutants-TRPV4 mutations were generated by site-directed mutagenesis (QuikChange II XL site-directed mutagenesis kit, Stratagene) deleting: 13 amino acids (positions 132-144) to generate the mutation lacking the N-terminal PRD domain (TRPV4-⌬PRD); 20 amino acids (positions 812-831) for the mutation lacking the calmodulinbinding domain (TRPV4-⌬CaM); and the last 4 amino acids of the channel (positions 868 -871) for the mutation lacking the C-terminal PDZ domain (TRPV4-⌬DAPL). All TRPV4 constructs were tagged with YFP (or cyan fluorescent protein) at the C terminus, and sequences were verified by PCR.
Cell Transfection-HEK293 and HeLa cells were transiently transfected with ExGen500 (Fermentas MBI) using 8 eq of polyethyl enimine together with 3 g of cDNA of the following plasmids: 1.5 g of rat pcDNA3-IP 3 R3 plasmid and 1.5 g of human TRPV4-YPF plasmids, WT, or mutants. Cells were cultured from 24 to 48 h before use.
Confocal Imaging-HeLa cells transiently transfected with WT and TRPV4 mutants were fixed with 4% paraformaldehyde. The subcellular localization was analyzed under a ϫ63 1.32 oil Ph3 CS objective using an inverted Leica SP2 confocal microscope. Original images were not further processed except for adjustments of brightness, contrast, and color balance.
Co-immunoprecipitation Experiments-HEK293 cells were transiently transfected with IP 3 R3 and TRPV4-YPF plasmids as described before. After 48 h, cells were lysed, total protein was * This work was supported by grants from the Spanish Ministries of Education and Science (Grant SAF2006-04973 and SAF2006-13893-C02-02), Spanish Ministries of Health (Fondo de Investigació n Sanitaria, Red HERACLES RD06/ 0009), Generalitat de Catalunya (SGR05-266), and Fundació la Marató de TV3 (061331). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1  recollected, and Western blot assays were carried out (using 100 g) as described previously (14) to verify the presence of the proteins IP 3 R3 and TRPV4 channel. Direct and reverse co-immunoprecipitation experiments were done to evaluate the interaction between both proteins. For this, 6 g of anti-GFP (direct immunoprecipitation) or anti-IP 3 R3 (reverse immunoprecipitation) were incubated for 1 h with 30 l of G protein (protein G-Sepharose 4B Fast Flow, Sigma-Aldrich) at 4°C. 1500 g of total protein samples from transfected and non-transfected cells were incubated overnight at 4°C, gently mixing, with the previous antibody-G protein complex. Immunocomplexes were washed four times in phosphate-buffered saline ϩ 0.1% Nonidet P-40 solutions, and Western blot was performed. Primary antibodies used were anti-GFP (1:1000, mouse monoclonal antibody, Clontech Laboratories, Inc.) and anti-human TRPV4 (1:250, rabbit polyclonal antibody that recognizes the last 19 amino acids. Its specificity has been tested in both heterologous expression systems (27) and knock-out mouse models (12,15)) for TRPV4-YFP and TRPV4 detection; and anti-IP 3 R3 (1:2000, rabbit antibody, BD Biosciences) for IP 3 R3 detection.
Calcium Imaging-Intracellular calcium was determined in transfected HeLa cells loaded with 4.5 M fura-2AM (Molecular Probes, Leiden, The Netherlands) as described previously (3,27). Cytosolic Ca 2ϩ levels are presented as the ratio of emitted fluorescence (510 nm) after excitation at 340 and 380 nm relative to the baseline. All experiments were registered at room temperature, and cells were bathed in an isotonic solution containing (in mM): 140 NaCl, 5 KCl, 1.2 CaCl 2 , 0.5 MgCl 2 , 5 glucose, and 10 HEPES (ϳ305 mosmol/liter, pH 7.4).
Electrophysiological Recordings-Cationic currents were registered using the patch clamp technique in whole-cell configuration. Patch pipettes were filled with a solution containing (in mM): 20 cesium chloride, 100 cesium acetate, 1 MgCl 2 , 0.1 EGTA, 10 HEPES, 4 Na 2 ATP, and 0.1 NaGTP; 300 mosmol/ liter, pH 7.2-7.3. Pipette solutions containing different concentrations of EET (BIOMOL) and IP 3 (Calbiochem) were also used and indicated in the corresponding figure (Fig. 4, B-F). Cells were bathed in a solution containing (in mM): 140 NaCl, 5 KCl, 1 MgCl 2 , 10 HEPES, and 1 EGTA (ϳ310 mosmol/liter, pH 7.3-7.4). HEK293 were clamped at 0 mV, and ramps from Ϫ140 mV to ϩ100 mV (400 ms) were applied at a frequency of 0.2 Hz. Ramp data were acquired at 10 kHz and low pass-filtered at 1 kHz. All experiments were carried out at room temperature. Only those cells that presented YFP fluorescence were recorded.
Statistics-Data are expressed as the mean Ϯ S.E. of n experiments. Analysis of variance or Student's t test was performed with the Sigma Plot 8.02 (Systat Software, Inc.) and SPSS to determine statistical significance. Bonferroni's test was used for post hoc comparison of means. p Ͻ 0.05 was considered significant.

RESULTS AND DISCUSSION
Localization of Wild Type and TRPV4 Mutants-We evaluated three putative IP 3 R-interacting sites in TRPV4 that may participate in IP 3 -mediated sensitization of TRPV4 to mechanical and osmotic stimuli (Fig. 1A). An N-terminal proline-rich domain (PRD), unique to TRPV4 and not present in other TRPV proteins, has been implicated in the channel interaction with the regulatory proteins PACSINs (28). A shorter C terminus PRD has also been identified in other members of the TRP family (29) and appears to participate in protein-protein interactions including the Homer family proteins and IP 3 R, among others (30). TRPV4 also presents a CaM-binding site on its C terminus (24). CaM-binding sites have been recognized in different TRP channels as important regulatory domains, exerting both inhibitory and facilitatory effects on TRPs upon Ca 2ϩ -CaM binding (31). Besides, there is an overlap between the CaM-and IP 3 R-binding sites in many TRPs (31), leading to the proposal that IP 3 R binding to and displacing CaM from its binding site contributes to the activation of canonical TRP (TRPC) channels (32). CaM binding to ankyrin repeats of TRPV1 has also been identified (33,34), but we do not know at present whether this N-terminal CaM-binding site is also present in TRPV4. Finally, we checked whether the final four amino acids of TRPV4 (DAPL) that present a PDZ-binding motif-like (35), may participate in the interaction with IP 3 R.
Activation of WT and TRPV4 Mutants by Cell Swelling and 4␣PDD-Intracellular [Ca 2ϩ ] measurements using the fura-2 dye indicator showed a clear response to 30% hypotonic solutions in HeLa cells co-expressing TRPV4-WT, TRPV4-⌬CaM, or TRPV4-⌬DAPL constructs with IP 3 R3 (Fig. 2, A, C, and D). On restoration of the isotonic condition and recovering the basal [Ca 2ϩ ] levels, cells also triggered a robust Ca 2ϩ signal in response to 4␣PDD (1 M), a synthetic activator of TRPV4 (19), thereby confirming the channels functional expression. However, cells overexpressing TRPV4-⌬PRD did not respond to either 4␣PDD (Fig. 2B) or hypo- tonic solutions (Fig. 2B, inset) but showed a typical Ca 2ϩ signal in response to the purinergic agonist ATP (100 M).
Altogether, the results indicated that TRPV4-WT, TRPV4-⌬CaM, and TRPV4-⌬DAPL used in this study trafficked to the plasma membrane where they responded to typical TRPV4-activating stimuli. On the other hand, TRPV4-⌬PRD, despite its correct localization, did not elicit Ca 2ϩ signals in response to hypotonicity or 4␣PDD and, therefore, was of no use for the identification of the molecular mechanism underlying IP 3 -mediated sensitization of TRPV4. Nevertheless, TRPV4-⌬PRD revealed an additional piece of evidence suggesting the relevance of the N terminus in TRPV4 gating. A previous report has already pointed to the PRD as a functionally important domain for TRPV4 channel gating (26). D'Hoedt et al. (26) showed that combined mutation of P142A, P143A, and P152A resulted in channels unresponsive to hypotonic shocks, AA, and heat that retained the response to 4␣PDD. Our more drastic approach, deleting the entire PRD, in addition to preventing the response to hypotonicity also induced the loss of 4␣PDD response.
Deletion of the CaM-binding Site of TRPV4 Prevents IP 3 R Binding and Sensitization to Hypotonic Stimulus-We have previously reported the modulatory effect of IP 3 R on TRPV4 activation by mechanical and osmotic stimuli as well as the immunoprecipitation of TRPV4 and IP 3 R3 (not IP 3 R1) (14). We now address the identification of the IP 3 R-binding site on TRPV4 and its relevance for IP 3 -mediated sensitization.
Expression of the proteins of interest in the lysates later used for co-immunoprecipitation studies was probed by Western blotting with anti-GFP (to detect TRPV4) and anti-IP 3 R3 (Fig.  3, A and B). Co-immunoprecipitation experiments found that deletion of the CaM-binding site (V4-⌬CaM) abolished IP 3 R3 binding (Fig. 3C), whereas deletion of the last four amino acids (V4-⌬DAPL) did not affect IP 3 R3 binding (Fig. 3D). These results were further confirmed by reversed co-immunoprecipitation (results not shown).
The functional consequence of the loss of IP 3 R3 binding to TRPV4-⌬CaM on IP 3 -mediated sensitization was first evaluated with the intracellular [Ca 2ϩ ] indicator fura-2. As reported previously (14), concomitant activation of the phospholipase A 2 -AA-EET and phospholipase C-IP 3 pathways using hypotonic solutions and ATP, respectively, resulted in TRPV4 sensitization. Fig. 3E shows Ca 2ϩ signals obtained in HeLa cells expressing TRPV4-WT and IP 3 R3 challenged with a short exposure to 30% hypotonic solution and then exposed to hypotonic solution containing 1 M ATP (a concentration that does not trigger Ca 2ϩ response in HeLa cells (14)). The addition of extracellular ATP potentiated the hypotonicity-induced Ca 2ϩ response. Mean increases in the Ca 2ϩ signal are shown in Fig. 3E (bottom). Potentiation of the hypotonicityinduced Ca 2ϩ signal by ATP was also observed in HEK293 cells  co-expressing TRPV4-WT and IP 3 R3 (not shown). Similar results were obtained examining the TRPV4-⌬DAPL construct that preserves IP 3 R3 binding (Fig. 3G). However, TRPV4-⌬CaM channels retained their ability to respond to hypotonic solutions but showed no enhancement of the Ca 2ϩ signal when challenged a second time with hypotonic solutions containing ATP (Fig. 3F). In fact, similarly to TRPV4-WT-expressing cells exposed to a second hypotonic challenge in the absence of ATP (14), the second hypotonic shock (containing ATP) generated a reduced Ca 2ϩ signal in TRPV4-⌬CaM.
Deletion of the CaM-binding Site of TRPV4 Prevents IP 3 -mediated Sensitization to EET-EET metabolites of arachidonic acid are the final activators of TRPV4 in response to hypotonic shocks (14,36). Accordingly, TRPV4-expressing cells respond to either external (18) or internal EET-containing solutions (14), an effect that can be potentiated by IP 3 (14). In our attempt to characterize the importance of the CaM-binding site in the modulation of TRPV4 activation by osmotic stimulus, we studied the impact of deleting the CaM-binding site on the IP 3 -mediated potentiation of EET-induced TRPV4 currents.
We have previously reported that low concentrations of EET (1 nM) or 30 M IP 3 (added to the intracellular pipette solutions) hardly activate TRPV4 currents. However, the presence of both 1 nM EET and 30 M IP 3 triggered sizeable TRPV4 currents in both native and heterologously expressing cell systems (14), indicative of the IP 3 -mediated sensitization of TRPV4 to the mechano-and osmotransducing messenger EET. When 1 nM EET intracellular pipette solutions were tested on TRPV4-⌬CaM channels, no differences with TRPV4-WT channels were observed (26 Ϯ 13 pA/pF (n ϭ 5) and 27 Ϯ 12 pA/pF (n ϭ 9), respectively; measured at ϩ100 mV; p Ͼ 0.05) (Fig. 4C). A significant difference in current magnitude was detected between the response of TRPV4-WT and TRPV4-⌬CaM channels to the combined presence of 1 nM EET and 30 M IP 3 in the intracellular pipette solutions (Fig. 4D). Considering the reported positive feedback of Ca 2ϩ , via Ca 2ϩ -CaM binding to TRPV4 (24), a possible alternative explanation for the absence of IP 3 -mediated sensitization in TRPV4-⌬CaM channels may be the lost Ca-CaM binding. Although we considered this an unlikely possibility, as whole-cell patch clamp experiments were carried out in the absence of Ca 2ϩ and in the presence of EGTA (see also Ref. 14), this possibility was further ruled out using the CaM inhibitor W-7. The presence of the CaM inhibitor W-7 (30 M) did not affected the IP 3 -mediated sensitization of TRPV4-WT channels to 1 nM EET and 30 M IP 3 (Fig.  4E). Mean responses measured at ϩ100 mV in YFP-, TRPV4-WT-, and TRPV4-⌬CaM-expressing HEK293 dialyzed with 1 nM EET and 30 M IP 3 are shown in Fig. 4F. Mean response of TRPV4-WT cells dialyzed with 1 nM EET, 30 M IP 3 ,and 30 M W-7 are also included.
Conclusion-Altogether, our results show that IP 3 -mediated sensitization of TRPV4 is Ca-CaM independent, that TRPV4 channels lacking the CaM-binding site lose their interaction with IP 3 R3, and consequently, they also lacked the IP 3 -mediated sensitization to osmotic and direct EET stimulation. Our data with the TRPV4-⌬PRD also highlight the relevance of the proline-rich N-terminal region of TRPV4 in channel gating.