Regulation of a Transient Receptor Potential (TRP) Channel by Tyrosine Phosphorylation

The recently identified transient receptor potential (TRP) channel family member, TRPV4 (formerly known as OTRPC4, VR-OAC, TRP12, and VRL-2) is activated by hypotonicity. It is highly expressed in the kidney as well as blood-brain barrier-deficient hypothalamic nuclei responsible for systemic osmosensing. Apart from its gating by hypotonicity, little is known about TRPV4 regulation. We observed that hypotonic stress resulted in rapid tyrosine phosphorylation of TRPV4 in a heterologous expression model and in native murine distal convoluted tubule cells in culture. This tyrosine phosphorylation was sensitive to the inhibitor of Src family tyrosine kinases, PP1, in a dose-dependent fashion. TRPV4 associated with Src family kinases by co-immunoprecipitation studies and confocal immunofluorescence microscopy, and this interaction required an intact Src family kinase SH2 domain. One of these kinases, Lyn, was activated by hypotonic stress and phosphorylated TRPV4 in an immune complex kinase assay and an in vitro kinase assay using recombinant Lyn and TRPV4. Transfection of wild-type Lyn dramatically potentiated hypotonicity-dependent TRPV4 tyrosine phosphorylation whereas dominant negative-acting Lyn modestly inhibited it. Through mutagenesis studies, the site of tonicity-dependent tyrosine phosphorylation was mapped to Tyr-253, which is conserved across all species from which TRPV4 has been cloned. Importantly, point mutation of Tyr-253 abolished hypotonicity-dependent channel activity. In aggregate, these data indicate that hypotonic stress results in Src family tyrosine kinase-dependent tyrosine phosphorylation of the tonicity sensor TRPV4 at residue Tyr-253 and that this residue is essential for channel function in this context. This is the first example of direct regulation of TRP channel function through tyrosine phosphorylation.

The recently identified transient receptor potential (TRP) channel family member, TRPV4 (formerly known as OTRPC4, VR-OAC, TRP12, and VRL-2) is activated by hypotonicity. It is highly expressed in the kidney as well as blood-brain barrier-deficient hypothalamic nuclei responsible for systemic osmosensing. Apart from its gating by hypotonicity, little is known about TRPV4 regulation. We observed that hypotonic stress resulted in rapid tyrosine phosphorylation of TRPV4 in a heterologous expression model and in native murine distal convoluted tubule cells in culture. This tyrosine phosphorylation was sensitive to the inhibitor of Src family tyrosine kinases, PP1, in a dose-dependent fashion. TRPV4 associated with Src family kinases by co-immunoprecipitation studies and confocal immunofluorescence microscopy, and this interaction required an intact Src family kinase SH2 domain. One of these kinases, Lyn, was activated by hypotonic stress and phosphorylated TRPV4 in an immune complex kinase assay and an in vitro kinase assay using recombinant Lyn and TRPV4. Transfection of wild-type Lyn dramatically potentiated hypotonicity-dependent TRPV4 tyrosine phosphorylation whereas dominant negative-acting Lyn modestly inhibited it. Through mutagenesis studies, the site of tonicity-dependent tyrosine phosphorylation was mapped to Tyr-253, which is conserved across all species from which TRPV4 has been cloned. Importantly, point mutation of Tyr-253 abolished hypotonicity-dependent channel activity. In aggregate, these data indicate that hypotonic stress results in Src family tyrosine kinasedependent tyrosine phosphorylation of the tonicity sensor TRPV4 at residue Tyr-253 and that this residue is essential for channel function in this context. This is the first example of direct regulation of TRP channel function through tyrosine phosphorylation.
Signaling by hypotonic stress and cell swelling has been well studied in diverse models (reviewed in Ref. 1). Nonetheless, the sensor for hypotonic stress, both at the single cell and organismal levels, remains elusive. A member of the transient recep-tor potential (TRP) 1 family of cation channels has recently been described that fulfills multiple criteria for an osmosensing protein. Based upon homology with OSM-9, an osmosensing TRPlike channel expressed in Caenorhabditis elegans (2), the mammalian candidate osmosensing TRP channel family member, was cloned independently as OTRPC4 (3) and VR-OAC (4). This protein, also identified as TRP12 and vanilloid receptorrelated protein-2 in other contexts, was recently renamed TRPV4 (5).
Although TRPV4 is expressed at the mRNA level in multiple tissues including liver, heart, lung, spleen, and adipose tissue, it is by far most abundant in the kidney (3,4). By in situ hybridization, renal expression was localized principally to the distal convoluted tubule (3). Interestingly, TRPV4 expression was also detected in the specific blood-brain barrier-deficient hypothalamic nuclei responsible for systemic osmosensing and in the mechanosensing hair cells of the inner ear (4). Functional and expression data were both consistent with a role for TRPV4 in osmosensing at both the single cell and organismal levels.
Apart from its gating by tonicity, little is known about TRPV4 regulation. We and others have shown that signaling events in response to hypotonic stress are reminiscent of activation of a receptor-tyrosine kinase (9 -12). We observed that hypotonic stress resulted in rapid tyrosine phosphorylation of TRPV4 in a heterologous expression model and in native murine distal convoluted tubule cells in culture. This tyrosine phosphorylation was sensitive to the inhibitor of Src family cytoplasmic tyrosine kinases, PP1, in a dose-dependent fashion. TRPV4 associated with Src family tyrosine kinases by co-immunoprecipitation studies and confocal immunofluorescence microscopy, and this interaction required an intact SH2 domain. Lyn, a Src family kinase, was activated by hypotonic stress and phosphorylated TRPV4 in an immune complex ki-* These studies were supported by the National Institutes of Health (Grant DK52494 to D. M. C.), by the American Heart Association, and by the Department of Veterans Affairs. 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.
ʈ nase assay and kinase assay with recombinant proteins. Transfection of wild-type Lyn dramatically potentiated hypotonicitydependent TRPV4 tyrosine phosphorylation whereas dominant negative-acting Lyn modestly inhibited it. Phosphorylation occurred principally at residue Tyr-253, and point mutation of this residue abolished tonicity-dependent calcium entry. In aggregate, these data indicate that Src family kinase-dependent tyrosine phosphorylation of Tyr-253 regulates TRPV4 activation in response to hypotonicity.

MATERIALS AND METHODS
Cells and Cell Culture-mIMCD3 cells were maintained and passaged as previously described (13). The murine distal convoluted tubule mDCT cell line (14), kindly provided by Peter Friedman, was maintained in Dulbecco's modified Eagle's medium/F12 supplemented with 10% fetal bovine serum.
Murine TRPV4 Cloning, Transient and Stable Transfection, and Mutagenesis-TRPV4 was cloned from murine kidney poly(A) ϩ RNA via RT-PCR using the SuperScript Preamplification System (Invitrogen) for first strand synthesis, followed by amplification with TRPV4specific primers, 5Ј-AGT ATG GCA GAT CCT GGT GA-3Ј and 5Ј-GTA CAG TGG GGC ATC GTC C-3Ј. Amplified fragment was TA-cloned into pcDNA3.1/V5-His-TOPO (Invitrogen) to create an expression plasmid encoding a carboxyl-terminal V5-epitope-tagged murine TRPV4; identity was confirmed by sequencing (Core Facility, Vollum Institute for Advanced Biomedical Research). Following the demonstration of tyrosine phosphorylation of TRPV4, high probability tyrosine phosphorylation sites were sought using ProSite (us.expasy.org/prosite, Ref. 15) and NetPhos (www.cbs.dtu.dk/services/NetPhos, Ref. 16; see "Results"). The deduced amino acid sequence of the TRPV4 clone in the present study was identical to that of the canonical sequence, NP_071300, with the following exceptions: Leu instead of Arg at position 90, Thr instead of Ile at position 389, and His instead of Tyr at position 860. Leu-90 was previously reported in the context of murine TRP12 and murine VR-OAC. Of note, the extreme carboxyl-terminal Tyr-860, present in NM_022017 and absent from our clone, was predicted to encode an extremely low-probability tyrosine phosphorylation site (probability 0.075, NetPhos). HEK293 cells were transfected with LipofectAMINE Plus in accordance with the manufacturer's directions (Invitrogen). Stable cell lines were derived from pooled clones propagated under selection pressure with Gly-418. Lyn-WT and Lyn-DN were originally described as pSR␣-Lyn wt and pSR␣-Lyn(K-R) (17). Individual point mutants of TRPV4 were generated using the QuickChange mutagenesis system (Stratagene), in accordance with the manufacturer's instructions, using PAGE-purified primers obtained from IDT. Multiple point mutants (e.g. 5ϫ-mutant) were generated through iterated application of this approach. All mutations were confirmed by direct sequencing, as above.
Immunofluorescence Microscopy-Monolayers were grown on LabTek Chamber slides (Nalgene Nunc) and labeled with primary antibody (anti-Lyn (Santa Cruz Biotechnology sc-15), anti-Src (Santa Cruz Biotechnology sc-19), or anti-V5 (Invitrogen 46 -0705)) and secondary antibody (AlexaFluor 594 goat anti-rabbit IgG (Molecular Probes A-11037) or AlexaFluor 488 goat anti-mouse IgG (Molecular Probes A-11001) prior to confocal epifluorescence imaging. Studies were performed on wild-type HEK293 cells and on HEK293 cells stably transfected with TRPV4. The affinity-purified anti-TRPV4 antibody was insufficiently sensitive in the immunofluorescence assay to permit examination of the native (mDCT) system. For confocal imaging, consecutive sections 500 nm apart, bracketing the center of the cell, were scanned alternating between 488 and 568 nm using a Leica TCS SP confocal laser scanning microscope. System settings were held constant for all imaging. Images were digitally captured and then deconvolved using Power HazeBuster TM imaging software (VayTek, Inc., Fairfield, VA). All imaging analysis determining colocalization was carried out using software by IPLab TM (Scanalytics Inc., Fairfield, VA).

RESULTS
Because tyrosine phosphorylation events may accompany hypotonic stress, we sought to determine whether hypotonicity resulted in tyrosine phosphorylation of the tonicity-responsive TRPV4. V5-epitope-tagged TRPV4 was stably overexpressed in HEK293 cells, which were reported to lack this channel (3). By anti-phosphotyrosine immunoblotting of anti-V5 immunoprecipitates, hypotonic stress (150 mosmol/kg H 2 O) resulted in pronounced and transient up-regulation of TRPV4 tyrosine phosphorylation (Fig. 1A). This phenomenon was specific to hypotonic stress as it was not observed in the presence of treatment with the peptide growth factor, epidermal growth factor, or the activator of classical protein kinase C, 12-Otetradecanoylphorbol-13-acetate (TPA) (Fig. 1B). There was no effect upon total TRPV4 abundance in these experiments (data not shown). Two other solute stressors, hypertonic stress and urea stress, both failed to increase anti-phosphotyrosine immunoreactivity in this model (data not shown). Of note, additional less intense bands migrating more slowly than the principal band corresponding to newly tyrosine-phosphorylated TRPV4 were frequently observed in this heterologous expression model. Treatment of lysates with endoglycosidase-F resulted in the complete disappearance of the higher molecular mass bands and enhancement of the principal lower molecular mass band (Fig. 1C). These data were consistent with the presence of forms of TRPV4 exhibiting varying degrees of glycosylation.
We next sought to confirm that: 1) stable heterologous transfection of epitope-tagged TRPV4 resulted in expression of functional TRPV4 and thereby mimicked endogenous expression, and 2) addition of the epitope tag did not interfere with channel function. Calcium transients in response to hypotonicity were examined in untransfected and TRPV4-transfected HEK293 cells. In both untransfected and TRPV4-transfected cells, there was negligible calcium entry following sham (isotonic) treatment ( Fig. 2A). In response to hypotonic stress (150 mosmol/kg H 2 O), there was a robust response in the TRPV4-transfected cells but virtually no response in the wild-type cell line. Next, to confirm that the calcium transients required calcium entry, similar studies were performed with only the TRPV4-transfected cell line in the presence and absence of extracellular calcium (Fig. 2B). The response to hypotonic stress was noted again, in the presence of extracellular calcium ([Ca 2ϩ ] o ϭ 1.25 mM) but was completely absent when calcium was omitted from the assay buffer ([Ca 2ϩ ] o ϳ0 mM). However, the hypotonicitydependent, intracellular calcium signal was restored upon addition of calcium (CaCl 2 , 1.25 mM) to the cuvette. In aggregate, these data indicated that the heterologous expression model functionally mimicked the native system.
To demonstrate that tyrosine phosphorylation of TRPV4 in response to hypotonic stress occurred in a native system, in a fashion similar to the HEK293-based heterologous expression system, a rabbit polyclonal, affinity-purified anti-peptide antibody was generated against the carboxyl terminus of murine TRPV4 (see "Materials and Methods"). The specific peptide was chosen to avoid cross-reactivity with closely related members of the TRP channel family (Fig. 3A); VRL-2, OTRPC4, VR-OAC, and TRP12 are names independently assigned to the protein now designated TRPV4 (5). Murine renal tissue was screened for anti-TRPV4 immunoreactivity. Expression was detected in whole kidney lysate (Fig. 3B), most of which appeared to be conferred by expression in the outer cortex. The distal convoluted tubule, where TRPV4 mRNA had previously been detected (3), is confined to the outer cortex. Several cell lines were also screened for the presence of TRPV4 by anti-TRPV4 immunoblotting (Fig. 3C). Consistent with the reports of others, anti-TRPV4 immunoreactivity was absent in 3T3 and HEK293 cells. It was, however, present, in the renal distal convoluted tubule cell line (14), mDCT, and in the renal inner medullary cell line, mIMCD3. As expected, abundant TRPV4 expression was detected in a HEK293 cell line stably transfected with TRPV4. To assess utility of the anti-TRPV4 antiserum in immunohistochemistry, and to confirm that the TRPV4 overexpression system exhibited membrane-associated immunoreactivity, anti-TRPV4 immunofluorescence microscopy was performed with wild-type HEK293 cells and with the TRPV4/ HEK stable cell line studied in Fig. 2. Anti-TRPV4 confocal immunofluorescence microscopy demonstrated modest nonspecific staining in wild-type HEK293 cells (Fig. 3D) and robust, primarily cell membrane-associated, expression in TRPV4/ HEK stable transfectants (Fig. 3D). The effect of hypotonicity upon TRPV4 tyrosine phosphorylation was next examined in the native model of the mDCT cell line. Hypotonic stress (ϳ 150 mosmol/kg H 2 O) resulted in robust time-dependent tyrosine phosphorylation of TRPV4 in this model (Fig. 3E), confirming that neither heterologous expression nor, potentially, overexpression, was required for this phenomenon.
To ascertain in preliminary fashion the nature of this tyrosine phosphorylation event, a panel of tyrosine kinase inhibitors was applied prior to hypotonic stress. The general tyrosine kinase inhibitor, genistein, only modestly inhibited the effect of hypotonicity whereas the Src family kinase-specific inhibitor, PP1, was more effective. In addition, this effect was dose-de- pendent (Fig. 4). Of note, piceatannol, which is reportedly specific for Syk (23), a related cytoplasmic protein-tyrosine kinase not belonging to the Src family (reviewed in Ref. 24), failed to influence tonicity-dependent tyrosine phosphorylation. These data suggested in a preliminary fashion that tonicity-dependent activation of a Src family kinase may be responsible for TRPV4 tyrosine phosphorylation.
To confirm our inference from the inhibitor data, we first sought a physical association between TRPV4 and members of the Src family of tyrosine kinases. For these studies, the HEK293 heterologous expression model was used primarily because of the clean background afforded by the commercially available monoclonal anti-V5 antibody. (In most experiments, this antibody was covalently linked to Sepharose beads for immunoprecipitation because of the similarity in molecular mass between Src family protein-tyrosine kinases and immunoglobulin heavy chain.) Multiple Src family members were immunodetectable in the anti-V5 (anti-TRPV4) immunoprecipitates (Fig. 5A); the association between TRPV4 and both Lyn and Yes increased with duration of hypotonic stress. Of note, on immunoblots with anti-Src, a faint band co-migrating with bona fide Src was evident following anti-V5 immunoprecipitation even in untransfected HEK cells. The three antibodies in this panel are reported to be free from cross-reactivity with other Src family kinases in murine cells. Reciprocal experiments were performed (Fig. 5B) wherein lysates prepared from untransfected and V5-TRPV4-transfected HEK293 cells were immunoprecipitated with anti-Src family kinase antibodies and then immunoblotted with anti-V5. TRPV4 interaction with Lyn, Src, Fyn, Hck, Lck, and Yes was detected in this fashion. Only Lyn exhibited time-dependent interaction with TRPV4 such that the anti-V5 (i.e. anti-TRPV4) immunoreactivity in anti-Lyn immunoprecipitates was much stronger at 30 min than at time zero. These data suggested a physical interaction between TRPV4 and the Src family tyrosine kinases.
To further support a physical interaction between these proteins, confocal immunofluorescence microscopy was performed. In HEK293 cells stably transfected with TRPV4, near total co-localization of TRPV4 with Lyn was observed (Fig. 6). This co-localization, however, was unaffected by up to 30 min of exposure to hypotonicity. In contrast, only a small fraction of immunodetectable TRPV4 exhibited co-localization with Src (data not shown).
We next sought to confirm the interaction in the native mDCT system and to determine in preliminary fashion the Src   FIG. 3. Tonicity-dependent tyrosine phosphorylation of TRPV4 is evident in a native system. A, alignment of carboxyl termini of TRPV family members (c, C. elegans; h, human; m, murine; r, rat) with TRPV4 (formerly known as VRL-2, OTRPC4, VR-OAC, or TRP12) depicting location of synthetic peptide to which anti-TRPV4 polyclonal antibody was raised. B, immunodetection of TRPV4 in murine kidney tissue lysates prepared from whole kidney (K), outer cortex (OC), inner cortex (IC), medulla (M), and papilla (P). C, anti-TRPV4 immunoblot of whole cell lysates prepared from the indicated untransfected (wild-type) cell types, as well from TRPV4-HEK293 stable transfectants. D, confocal microscopy of subconfluent monolayers of wildtype HEK293 cells (left) and TRPV4/HEK293 stable transfectants (right) labeled with anti-TRPV4 and AlexaFluor-594-conjugated antirabbit secondary antibody (see ''Materials and Methods''). E, anti-phosphotyrosine immunoblot of anti-TRPV4 immunoprecipitates prepared from a murine distal convoluted tubule cell line (mDCT) treated for the indicated interval with hypotonic stress.

FIG. 5. TRPV4 co-immunoprecipitates with Src family tyrosine kinases.
A, anti-V5 immunoprecipitates prepared from wild-type HEK293 cells and TRPV4-HEK293 stable transfectants subjected to hypotonic stress for the indicated interval, and then immunoblotted for the indicated Src family tyrosine kinase. B, anti-V5 immunoblot of immunoprecipitates prepared from wild-type HEK293 cells and TRPV4-HEK293 stable transfectants; cells were subjected to hypotonic stress for the indicated duration and then immunoprecipitated with antibodies specific for the indicated Src family protein-tyrosine kinase.
family tyrosine kinase motif conferring interaction with TRPV4. Fyn was selected for these studies because of its constitutive interaction with TRPV4 by immunoprecipitation and because of the availability of agarose bead-bound deletion mutants of this kinase. Murine Fyn is a 534 (or 537 for FynT) amino acid Src family protein-tyrosine kinase containing three well described structural motifs: an SH3 domain (amino acids ϳ87-141), an SH2 domain (amino acids ϳ149 -231), and a carboxyl-terminal kinase domain (amino acids ϳ268 -517). Agarose-conjugated SH3-SH2 from Fyn (absent the kinase domain) effectively affinity-precipitated TRPV4 from both the heterologous expression (V5-TRPV4-HEK293, Fig. 7A) and the native (mDCT, Fig. 7B) models. In both models, agarose-conjugated Fyn SH2 alone (Fyn-(145-247)) but not agarose-conjugated Fyn SH3 alone (Fyn-(85-139)) reproduced this effect (Fig. 7C). These data suggested that the TRPV4-Src family tyrosine kinase interaction was mediated via the SH2 domain of the latter, presumably in conjunction with one or more cognate phosphotyrosine motifs on TRPV4.
The ability of Src family tyrosine kinases to phosphorylate TRPV4 was next examined. HEK293 cells stably expressing V5-epitope-tagged TRPV4 were transiently transfected with wild-type Lyn, dominant negative-acting Lyn, or with vector alone. TRPV4 was immunoprecipitated with anti-V5 and phosphorylation assessed by anti-phosphotyrosine immunoblotting. Overexpression of Lyn resulted in a marked increase in TRPV4 tyrosine phosphorylation under control conditions, which was even more evident in the presence of hypotonic stress (Fig. 8). These data suggested that Lyn can phosphorylate TRPV4 in intact cells. Dominant negative-acting Lyn modestly but reproducibly suppressed tonicity-dependent phosphorylation of TRPV4 but did not eliminate it entirely, suggesting that Lyn is perhaps a principal kinase involved in this process. These data could not, however, demonstrate whether this effect was direct or indirect.
We performed in vitro kinase assays to determine whether Lyn can directly phosphorylate TRPV4. We first confirmed that Lyn was activated in our model by hypotonic stress using an anti-Lyn immune complex kinase assay with acid-treated enolase as a substrate (Fig. 9A). In a different type of immunoprecipitation kinase assay, Lyn was immunoprecipitated (in conjunction with its associated proteins) from control and hypotonically stressed cells and then subjected to an in vitro kinase assay in the presence of [␥-32 P]ATP, cold ATP, Mg 2ϩ , and Mn 2ϩ . Up-regulated phosphorylation of a band of the appropriate size for TRPV4 was evident in the anti-Lyn immunoprecipitates from hypotonically stressed cells at 30 min of treatment (Fig. 9B). To exclude the possibility that a kinase co-immunoprecipitating with Lyn or TRPV4 was responsible for the TRPV4-directed kinase activity, similar experiments were performed using in vitro transcribed and translated TRPV4 and recombinant LynA (Calbiochem). To facilitate expression in this model and to produce a more soluble protein product, 5 of 6 transmembrane-spanning domains were deleted from murine TRPV4 cDNA to generate ⌬TRPV4 (see "Materials and Methods"). No candidate tyrosine phosphorylation sites were present in this domain of TRPV4. Recombinant Lyn phosphorylated the in vitro transcribed and translated TRPV4 and ⌬TRPV4 producing bands precisely co-migrating with immunodetectable TRPV4 and ⌬TRPV4 (Fig. 9C). Importantly, no substrate phosphorylation was evident in the absence of exogenous Lyn (data not shown). These data suggested that in addition to associating with Lyn, TRPV4 serves as a bona fide Lyn substrate.
We next sought to identify the site of the tyrosine phosphorylation in TRPV4 through site-directed mutagenesis studies. Preliminary sequence analysis of mTRPV4 using a standard protein motif search (ProSite, www.expasy.ch/prosite/) detected a single canonical ((RK)X 2 (DE)X 3 Y or (RK)X 3 (DE)X 2 Y, Refs. [25][26][27] tyrosine phosphorylation site at amino acid Tyr-411. Sequence alignment indicated conservation of the site across all 4 species in which TRPV4 had been cloned. Therefore, of the 36 tyrosines present in VR-OAC, we performed site-directed mutagenesis to replace this residue with phenylalanine (Y411F). When cells were transiently transfected with this mutant (Y411F-TRPV4), however, tyrosine phosphorylation of TRPV4 was as robust as in cells transfected with the wild-type TRPV4 (data not shown).
Subsequent analyses were performed to reconcile the inability of the Y411F mutant to abrogate tyrosine phosphorylation of TRPV4. Guided by a recently described (16) neural net-based contextual amino acid phosphorylation site identification shareware (NetPhos 2.0, www.cbs.dtu.dk/services/NetPhos/), we evaluated the likelihood of encountering regulated tyrosine phosphorylation at each of the 36 tyrosines in TRPV4. Interestingly, the single "classical" tyrosine phosphorylation motif in this protein (upon which we based our initial mutagenesis approach, Tyr-411) was actually a suboptimal candidate for phosphorylation in vivo because of an adjacent bulky tryptophan residue rendering the tyrosine sterically inaccessible. Eight other sites were identified with a high likelihood of undergoing regulated tyrosine phosphorylation. Based upon conservation across species in which TRPV4 has been cloned and upon predicted intracellular localization, we selected five sites for mutagenesis. The 5ϫ-mutant TRPV4 construct included the following mutations (numbered according to accession number NP_071300, the murine consensus TRPV4): Y253F, Y281F, Y415F, Y508F, and Y805F. Hypotonicity-inducible tyrosine phosphorylation was evident in anti-V5 immunoprecipitates of mDCT cells transiently transfected with wildtype TRPV4 but not in cells transfected with the 5ϫ-mutant (Fig. 10A, upper panel). Background nonspecific immunoreactivity co-migrating with TRPV4 (Fig. 10A) precluded detailed analysis of the tyrosine residues undergoing phosphorylation so an additional model system was sought. The COS7 cell line, in which essentially no co-migrating background was detectable, was selected for further study. Consistent with the mDCT model, mutagenesis of the 5 key tyrosine residues of TRPV4 prevented hypotonicity-dependent phosphorylation in the COS7 cells (Fig. 10A, lower panel).
A series of stable COS7 cell lines expressing single Tyr 3 Phe mutants of TRPV4 was generated and compared with cells expressing wild-type TRPV4 (Fig. 10B). Although there was some variability in the level of expression of each of the mutants, it was clear that Tyr-253 accounted for the vast majority of regulated (and constitutive) tyrosine phosphorylation of TRPV4. Interestingly, Tyr-253 is conserved across all species from which TRPV4 has been cloned (mouse, rat, and human), yet is not shared by any of the other TRPV family members or the prototypical C. elegans channel, OSM-9 (Fig. 10C). In addition, it is not shared by any members of the TRPC or TRPM families (data not shown).
Further studies were carried out to determine the importance of the putative tyrosine phosphorylation sites in TRPV4 function in response to hypotonicity. Wild-type HEK293 cells were compared with cells stably transfected with TRPV4 or with TRPV4 harboring mutations (Tyr 3 Phe) in the five key tyrosine residues (Fig. 11A). Again, a robust increase in intracellular calcium was observed in TRPV4-transfected cells but not in untransfected HEK cells. Importantly, the 5ϫ-mutant transfectants essentially lacked hypotonicity responsiveness, with a response indistinguishable from that of untransfected HEK. This effect was not a consequence of lack of expression of the mutant protein or aberrant targeting of the mutant protein because immunodetectable TRPV4 was evident in both whole cell lysates and membrane preparations from the 5ϫ-mutant cell line (Fig. 11A). Because Tyr-253 was the principal target of tyrosine phosphorylation in response to hypotonic stress based upon our mutagenesis data (Fig. 10), we next assessed the effect of the single Tyr-253 point mutant upon TRPV4 channel function (Fig. 11B). The Y253F mutant, similar to the 5ϫmutant, exhibited negligible hypotonicity-dependent calcium response. Again, TRPV4 expression and membrane targeting were comparable to the overexpressed wild-type TRPV4 as demonstrated through immunoblotting and preparation of membrane and cytosolic proteins (Fig. 11B). In addition, the subcellular distribution of the heterologously expressed Tyr-253 was indistinguishable from that of the heterologously ex- FIG. 9. Lyn is activated by hypotonic stress and phosphorylates TRPV4. A, autoradiogram of immune complex kinase assay of anti-Lyn immunoprecipitates prepared from control and hypotonically treated (150 mosmol/kgH 2 O ϫ 15 min) HEK293; arrowhead denotes in vitro phosphorylated Lyn substrate (enolase). B, co-immunoprecipitation kinase assay of anti-Lyn immunoprecipitates prepared from TRPV4-HEK293 stable transfectants; immunoprecipitates were subjected to kinase assay in absence of exogenous substrate. Phosphorylation of band co-migrating with TRPV4 is detected following hypotonic treatment. C, in vitro kinase assay with recombinant LynA and in vitro transcribed/translated TRPV4 or ⌬TRPV4 (a TRPV4 deletion mutant lacking five of six membrane-spanning domains to facilitate expression). A specifically phosphorylated band, co-migrating with either immunodetectable TRPV4 or ⌬TRPV4 was evident. No phosphorylation, specific or nonspecific, was noted in the absence of LynA (data not shown).
pressed wild-type TRPV4 (data not shown). These data indicate that Tyr-253 is essential not only for tyrosine phosphorylation of TRPV4 but also for function of TRPV4 in response to hypotonic stress; they are also consistent with the observation of Liedtke et al. (4) and Watanabe et al. (28) who noted a decrease in the hypotonicity responsiveness of mutant TRPV4 proteins lacking much of the amino terminus. Of note, mutation of a different single tyrosine among the five principal candidates (Tyr-281), upon stable transfection of the resultant plasmid, failed to influence hypotonicity-dependent calcium transients (data not shown), further underscoring the specificity of the Tyr-253 effect. DISCUSSION To the best of our knowledge, these are the first data demonstrating regulated tyrosine phosphorylation of a TRP channel or its association with a Src family protein-tyrosine kinase. Moreover, they are the first to describe regulation of function of a TRP channel through direct tyrosine phosphorylation. We observed the physical and functional interaction between TRPV4 and Src family kinases in both the heterologous overexpression model and in the native expression model afforded by the murine distal convoluted tubule cell line. We show that TRPV4 becomes newly tyrosine-phosphorylated in response to hypotonic stress, and that point mutation of the tyrosine residue conferring essentially all of the de novo phosphorylation virtually abolishes tonicity-dependent calcium mobilization.
In contrast to the present model of direct regulation, several TRP family members are indirectly regulated by upstream activation of a receptor-tyrosine kinase, via intermediary action of a second messenger system. For example, TRPM7 is indirectly inactivated upon receptor-tyrosine kinase activation, an effect mediated through phospholipase C activation and PIP2 hydrolysis (29). TRPC4 and TRPC5, in contrast, are activated in response to receptor-tyrosine kinase activation (30). TRPC3, upon heterologous expression in chicken B-cells, is activated in a Syk-dependent fashion, although the actual effector appears to be phospholipase C ␥ (31).
Although kinase activation and the role of kinase-dependent signaling in anisotonicity has been extensively studied in diverse models (reviewed in Ref. 1), comparatively fewer data describe the role of tyrosine kinase activity in this process. In cell culture models, tyrosine kinase activity is up-regulated by hypotonic stress (9, 10) and is required for immediate-early gene transcription and expression (10,32). Pharmacological inhibition of tyrosine kinase activity blocks, and potentiation of tyrosine kinase activity inhibits, regulatory volume decreaseassociated efflux of cations (9), chloride (33)(34)(35)(36)(37)(38)(39), and the amino acid osmolyte, taurine (39,40).
Several cytoplasmic protein-tyrosine kinases have previously been implicated in the hypotonic stress response. In a marine elasmobranch model of hypotonic stress, Syk-and Lynphosphorylated band 3, an efflux pathway for the osmolyte, taurine (41). In lymphocytes, lck may mediate activation of swelling-induced chloride efflux, based in part upon experiments using p56lck Ϫ/Ϫ mice (35). Lastly, in cells derived from mammalian kidney, chloride channel activity is activated by fyn, although the dependence of this phenomenon upon anisotonicity, per se, remains unexplored (42). Our data extend the range of substrates phosphorylated by Src family kinases in response to hypotonicity.
It is of interest that of the 36 tyrosine residues in murine TRPV4, a single residue (Tyr-411) was identified through a standard approach (ProSite. www.expasy.ch/prosite/) as constituting a candidate phosphorylation site. Mutation of this site, however, failed to influence hypotonicity-dependent tyrosine phosphorylation of TRPV4. A neural net-derived, context-based approach (NetPhos 2.0, www.cbs.dtu.dk/services/NetPhos, Ref. 16) permitted us to identify the eight most likely candidate tyrosine phosphorylation sites, from which five were selected for further study based upon sequence conservation and intracellular location. Interestingly, Tyr-411 was not among these sites. Systematic mutagenesis identified Tyr-253 as the principal site of regulated phosphorylation (Fig. 10B). Consistent with a role in TRPV4 signal transduction in response to hypotonicity, Tyr-253 is conserved across all species from which TRPV4 has been cloned and is not shared by any other members of the TRPV family, or by the prototypical C. elegans osmosensing channel, OSM-9 (Fig. 10C). The apparent requirement for an intact SH2 domain for the Src family proteintyrosine kinase-TRPV4 interaction-suggests perhaps a constitutive level of tyrosine phosphorylation of this protein, which is supported by our anti-phosphotyrosine immunoblot analyses. In addition, the up-regulation in Lyn co-immunoprecipitation with TRPV4 with increasing duration of hypotonic stress suggests a more regulated interaction for this kinase and the possible presence of a positive feedback loop.
The structural role of Tyr-253 in TRPV4 channel gating remains unclear. Tyr-253 resides within the first of three putative ankyrin repeat domains in the TRPV4 amino terminus. Liedtke et al. (4) showed that elimination of the amino-terminal amino acids 1-402 of TRPV4 (including all three ankyrin repeat domains) decreased tonicity-responsiveness of TRPV4. In a similar fashion, Watanabe et al. (28) showed that deletion of residues 235-398 dramatically decreased TRPV4 calcium channel activity in response to heat. These data are in excellent agreement with our assignment of a pivotal role to Tyr-253.
TRPV4 is expressed predominantly in the kidney and much of this expression is confined to the distal convoluted tubule (3). We noted expression in a cell line derived from the renal medulla, as well. In mammals, the distal nephron is intimately involved with both local and systemic osmoregulation. It is tempting to speculate that TRPV4 is instrumental in this process by responding to changes in ambient tonicity in this tightly regulated epithelium. Whether TRPV4 serves to "sense" osmotic changes, as has been proposed by others (3,4), or whether it also serves as an effector of anisotonicity, through post-translational modification by Src family receptor-tyrosine kinases, remains unknown.