The human transient receptor potential vanilloid 3 channel is sensitized via the ERK pathway

The transient receptor potential vanilloid 3 (TRPV3) channel is a Ca2+-permeable thermosensitive ion channel widely expressed in keratinocytes, where together with epidermal growth factor receptor (EGFR) forms a signaling complex regulating epidermal homeostasis. Proper signaling through this complex is achieved and maintained via several pathways in which TRPV3 activation is absolutely required. Results of recent studies have suggested that low-level constitutive activity of TRPV3 induces EGFR-dependent signaling that, in turn, amplifies TRPV3 via activation of the mitogen-activated protein kinase ERK in a positive feedback loop. Here, we explored the molecular mechanism that increases TRPV3 activity through EGFR activation. We used mutagenesis and whole-cell patch clamp experiments on TRPV3 channels endogenously expressed in an immortalized human keratinocyte cell line (HaCaT) and in transiently transfected HEK293T cells and found that the sensitizing effect of EGFR on TRPV3 is mediated by ERK. We observed that ERK-mediated phosphorylation of TRPV3 alters its responsiveness to repeated chemical stimuli. Among several putative ERK phosphorylation sites, we identified threonine 264 in the N-terminal ankyrin repeat domain as the most critical site for the ERK-dependent modulation of TRPV3 channel activity. Of note, Thr264 is in close vicinity to a structurally and functionally important TRPV3 region comprising an atypical finger 3 and oxygen-dependent hydroxylation site. In summary, our findings indicate that Thr264 in TRPV3 is a key ERK phosphorylation site mediating EGFR-induced sensitization of the channel to stimulate signaling pathways involved in regulating skin homeostasis.

The transient receptor potential vanilloid 3 (TRPV3) channel is a Ca 2؉ -permeable thermosensitive ion channel widely expressed in keratinocytes, where together with epidermal growth factor receptor (EGFR) forms a signaling complex regulating epidermal homeostasis. Proper signaling through this complex is achieved and maintained via several pathways in which TRPV3 activation is absolutely required. Results of recent studies have suggested that low-level constitutive activity of TRPV3 induces EGFR-dependent signaling that, in turn, amplifies TRPV3 via activation of the mitogen-activated protein kinase ERK in a positive feedback loop. Here, we explored the molecular mechanism that increases TRPV3 activity through EGFR activation. We used mutagenesis and whole-cell patch clamp experiments on TRPV3 channels endogenously expressed in an immortalized human keratinocyte cell line (HaCaT) and in transiently transfected HEK293T cells and found that the sensitizing effect of EGFR on TRPV3 is mediated by ERK. We observed that ERK-mediated phosphorylation of TRPV3 alters its responsiveness to repeated chemical stimuli. Among several putative ERK phosphorylation sites, we identified threonine 264 in the N-terminal ankyrin repeat domain as the most critical site for the ERK-dependent modulation of TRPV3 channel activity. Of note, Thr 264 is in close vicinity to a structurally and functionally important TRPV3 region comprising an atypical finger 3 and oxygen-dependent hydroxylation site. In summary, our findings indicate that Thr 264 in TRPV3 is a key ERK phosphorylation site mediating EGFR-induced sensitization of the channel to stimulate signaling pathways involved in regulating skin homeostasis.
Transient receptor potential vanilloid 3 (TRPV3) 3 is a thermosensitive ion channel widely expressed in epithelial tissues of the skin, where it plays critical roles in epidermal proliferation, differentiation, hair growth, sensory thermotransduction, itch sensation, and the development of cutaneous pain. A number of stimuli have been shown to activate TRPV3, including warm or noxious temperatures (1)(2)(3); endogenous compounds such as farnesyl pyrophosphate or nitric oxide; the naturally occurring aromatic monoterpens carvacrol, menthol, and camphor; and synthetic drugs, of which the most frequently used is 2-aminoethoxydiphenyl borate (2-APB) (see Refs. 4 -7 for recent comprehensive reviews). The activity of TRPV3 can be further sensitized by a number of signaling molecules downstream of phospholipase C, including unsaturated fatty acids (8), Ca 2ϩ / calmodulin (9,10), protein kinase C (11), and phosphatidylinositol 4,5-bisphosphate (12). Moreover, among other thermosensitive TRP ion channels, TRPV3 has a unique intrinsic property; it is strongly up-regulated by its own activity (13,14). This, together with a remarkable gating promiscuity (15), makes the channel extremely sensitive to any type of conformational change and, as a consequence, it is difficult to determine the mechanisms of TRPV3 modulation at the molecular level.
In man, several gain-of-function mutations in TRPV3 that lead to enhanced channel activity have been linked to severe channelopathy, congenital Olmsted syndrome, accompanied by strong skin defects (16,17). Also, the genetic deletion of the trpv3 gene in mice leads to an impaired epidermal barrier structure and strong deficits in responses to innocuous and noxious heat (18). A significant breakthrough in our understanding of how aberrant TRPV3 activity relates to skin pathophysiology has been made by Cheng et al. (19), who elaborated an observation that the defects in hair morphogenesis in mice deficient in trpv3 strikingly resemble the mouse phenotype bearing mutations in the epidermal growth factor receptor (EGFR) and its ligand, transforming growth factor-␣ (TGF-␣). Their study demonstrated that in keratinocytes, the TRPV3 channel forms a signaling complex with EGFR, whereby its activity is regulated by EGFR-coupled signaling pathways. A mechanism has been proposed in which a weak constitutive activity of TRPV3 raises the intracellular concentration of calcium, which in turn releases TGF-␣ that binds to and activates EGFR. A further increase in intracellular calcium downstream of EGFR activation then results in a potentiation of TRPV3. As this potentiating effect could be blocked by inhibitors of phospholipase C (PLC) and mitogen-activated protein kinase ERK, these experiments have indicated that the PLC-and ERK-mediated regulation of TRPV3 could be a primary mechanism required for EGFR-dependent functions (19). Whereas the effects of the PLC-signaling cascade on TRPV3 have been studied intensively at the single-cell level (12,20), the importance of ERK in mediating TRPV3 potentiation has not been established, and it is unclear whether ERK by itself is sufficient to influence TRPV3.
The activation of EGFR by TGF-␣ triggers the activation of downstream kinases known as the mitogen-activated protein kinase (MAPK) signaling pathway leading to the final effectors, the extracellular signal-regulated protein kinases 1 and 2 (ERK1/2). These two kinases are generally referred to as ERK because they both exhibit a functional redundancy and no evidence has been provided for a difference in substrate specificity between these two isoforms (21). Although ERK is considered a primary effector of EGFR signaling, it may also integrate other signaling pathways and phosphorylate various ion channels and regulate their gating properties, as is the case with the Cav2.2, Nav1.7, Kv4.2, and TRPC6 channels (22)(23)(24)(25).
The sequence analysis of human TRPV3 reveals eight putative phosphoacceptor sites for ERK kinase (S/TP), all present at the cytoplasmic N-terminal end (supplemental Table S1). This part of the protein, as in other members of the vanilloid TRP channel subfamily, features a characteristic ankyrin repeat domain (ARD) consisting of six ankyrin repeats (AR). These ARs are formed by motifs of ϳ33 amino acid residues that give rise to a structure consisting of two anti-parallel ␣-helices connected by ␤-hairpin loops called "fingers." The ARD of TRPV3 has been recently structurally characterized (amino acids Arg 118 -Ile 365 ) and a functionally important finger from which the TRPV3 channel can be decisively modulated was identified (26). The fact that two of the predicted ERK phosphorylation sites, Thr 264 and Thr 343 , are situated within this relatively well defined region raises an interesting possibility that the channel may be modulated by direct phosphorylation at its N terminus. Here we set out to explore whether the activation of EGFR signaling in keratinocytes results in the potentiation of TRPV3 responses through ERK phosphorylation, and to determine the contribution of the predicted phosphorylation sites in this process.

EGF treatment promotes potentiation of TRPV3 currents in keratinocytes via MAPK signaling pathway
We used human keratinocyte HaCaT cells and the whole-cell patch clamp technique to study the effects of EGF on the activity of endogenously expressed TRPV3 channels. To identify TRPV3-expressing cells, a combination of two TRPV3 agonists 2-APB and carvacrol at saturating concentrations (200 and 250 M) was first briefly applied to a cell (Fig. 1A). If a cell displayed a detectable current response at a holding potential of Ϫ70 mV (ϳ40% of cells), a combination of subsaturating concentrations of 2-APB and carvacrol (80 and 100 M) was applied for 30 s and orange squares, n Ն 3; PD98059: cyan diamonds, n Ն 7). In cells treated with U0126 or PD98059, the degree of potentiation of the seventh response after the 14-min presence of EGF was only 38 Ϯ 4% (p ϭ 0.003, compared with the cells treated with EGF alone; n ϭ 4) and 47 Ϯ 5% (p ϭ 0.008; n ϭ 7), respectively. In the presence of U0126, a prolonged application of EGF (30 min) caused a further sensitization to 53 Ϯ 5%, which was comparable with the relative amplitude obtained in control cells (48 Ϯ 4%; p ϭ 0.639; n ϭ 3). Data are expressed as mean Ϯ S.E. **, p Ͻ 0.01.

Functional regulation of TRPV3 via phosphorylation
then repeatedly for two 30-s periods at 5-min intervals. Whereas the responses to the latter combination of agonists subsequently increased, the maximum response to a combination of higher concentrations of agonists remained largely unchanged, and was regularly tested for rundown correction. This protocol enabled us to assess the degree of potentiation induced by EGF that runs in parallel with the use-dependent sensitization that is intrinsic to the TRPV3 channel (14). TRPV3 responses were recorded either under control conditions or in the continuous presence of EGF, applied with or without one of the two different selective mitogen-activated protein kinase (MEK) inhibitors U0126 or PD98059. The first (b), seventh (c), and tenth responses (e) were normalized to a subsequent maximal saturating response (corresponding to b/a, c/d, and e/f in Fig. 1A). Under control conditions, the first response to the combination of subsaturating agonists reached 35 Ϯ 2% of the maximum response and was potentiated upon the seventh and tenth application to 46 Ϯ 3 and 48 Ϯ 4% (Fig.  1B), respectively. The cells treated with EGF (20 ng/ml, 14 min) exhibited a seventh response that reached 72 Ϯ 6% of the maximum response. Prolonged incubation with EGF (30 min) did not cause any further increase (72 Ϯ 2%). Consistent with previous reports (19), the amplitude of TRPV3 responses induced by saturating concentrations of the agonists was not affected by the presence of EGF.
To determine whether the potentiating effect of EGF is mediated by the MAPK signaling pathway, two different MEK inhibitors, U0126 (10 M) and PD98059 (30 M), were tested together with EGF. The treatment of the cells with either of the two inhibitors completely abolished the potentiating effect of EGF on TRPV3 currents (Fig. 1B). In cells treated with U0126 or PD98059, the degree of potentiation of the seventh response after the 14-min presence of EGF was only 38 and 47%, respectively. In the presence of U0126, a prolonged application of EGF (30 min) caused a further potentiation to 53 Ϯ 5%, which was comparable with the relative amplitude obtained in control cells. Together, these experiments confirm previous observations obtained in mouse primary keratinocytes (19) and suggest that the potentiating effect of EGF on TRPV3 is mediated via MEK kinases, whose direct substrate ERK1/2 kinase may phosphorylate TRPV3.

Thr 264 represents a potential ERK phosphorylation site
To examine whether the potentiating effect of EGF on TRPV3 responses in HaCaT cells can be mediated by the phosphorylation of TRPV3, we used several prediction servers to assess potential phosphorylation sites for ERK (supplemental Table S1). In total we identified eight ERK1/2 consensus sites (PX(S/T)P) at the N terminus of human TRPV3, four of them containing a PXSP motif (Ser 60 , Ser 85 , Ser 88 , and Ser 102 ) and the next four a PXTP motif (Thr 35 , Thr 96 , Thr 264 , and Thr 343 ). We individually replaced all these serines and threonines with aspartic acid to mimic the negative charge of the phosphate group. The mutants were transiently transfected to HaCaT cells and assayed electrophysiologically using repeated stimulations with a low concentration of 2-APB (50 M), applied four times for 30 s with 30-s intervals and followed by a 10-s application of a combination of 2-APB (200 M) with carvacrol (150 M) to obtain a reference saturating response ( Fig. 2A).
In wild-type TRPV3, the inward currents subsequently increased upon repeated stimulation with 2-APB, on the fourth application reaching about 5% of the saturating response (Fig.  2B). The T264D mutant produced significantly larger first responses than the wild-type channels. When 2-APB currents were normalized to the reference saturating response for a given cell, two additional mutants, S85D and S102D, were identified to produce significantly larger first responses than the wild-type channels (supplemental Fig. S1). Two of the mutant channels, T264D and T343D, did not exhibit any sensitization upon repeated stimulation with 2-APB. In T264D, the first response reached about 7% of the maximum response, which was not significantly different from the fourth response. In contrast to the currents through wild-type channels, the majority of T264D responses to 2-APB apparently reached a plateau during 30 s of application. The 2-APB currents through T343D were barely detectable, being significantly smaller than those of wildtype channels and the other mutants. The fourth response in T343D remained at the initial level of 0.3% of the maximum, indicating functional rather than phosphorylation-mimicking effects. The T35D construct yielded also a significantly smaller initial response to 2-APB but exhibited sensitization comparable with wild-type channels. Apart from S85D, for all the mutants tested, maximal inward currents elicited by the combination of saturating agonists were not statistically different from wild-type channels.
Taken together, these findings identify threonine 264 as the most likely candidate to be a phosphorylation site for ERK1/2, because the phosphomimicking mutation sensitized the channel without affecting its maximum responses, as would be expected for the potentiating effect of MEK kinases observed in our experiments. Therefore, we set out to further characterize the effects of other substitutions at this key residue to assess its role in TRPV3 channel functionality.

Phosphomimicking/phosphonull mutations at Thr 264 affect TRPV3 functioning
In further experiments, we used the HEK293T expression system to avoid a possible contribution of endogenous TRPV3 channels that are expressed in HaCaT cells, and to examine whether the effects of mutations at Thr 264 are independent of the expression system used. We measured currents from the T264D mutant channels (Fig. 3A) using an analogous protocol to that described above. Moreover, we examined two additional constructs: the alternative phosphomimetic mutation T264E and the phosphonull mutation T264A. T264D and T264E exhibited an identical phenotype to T264D in keratinocytes, dramatically larger responses to 2-APB that did not sensitize upon repeated applications (50 M, 30 s with 30-s intervals). In these two constructs, the fourth response to 2-APB was not statistically different from the initial response (Fig. 3B), whereas that for wild-type channels increased 3-fold and T264A about 2-fold. In T264D, all four 2-APB responses were significantly larger compared with wild-type TRPV3. There was no significant difference in the maximum currents induced by a combination of saturating agonists (200 M 2-APB and 150 M carvacrol, 10 s) between the mutants and the wild-type channels. These results indicate that mutations at threonine 264 affect

Functional regulation of TRPV3 via phosphorylation
TRPV3 functioning in a manner consistent with persistent or lacking phosphorylation.

Overexpression of ERK1 supports the specific role of Thr 264 in TRPV3 phosphorylation
To further explore whether the ERK kinase may functionally affect TRPV3, ERK1 was transfected to HEK293T cells together with the wild-type TRPV3 channel. The presence of ERK1 in cells caused a strong increase in 2-APB responses (Fig. 4A). Upon repeated stimulation, the currents did not exhibit statistically significant sensitization, and their fourth responses reached 11.5 Ϯ 3.3% of the maximum amplitude induced by saturating concentrations of 2-APB and carvacrol. The saturating responses trended toward lower values in the presence of ERK1, although the difference did not reach statistical significance (p ϭ 0.085). Next, the cells were cotransfected with ERK1 Wild-type responses to 2-APB were markedly smaller compared with that induced by the mixture of agonists. Therefore two current scales were used. The holding potential was Ϫ70 mV. B, summary plot of average amplitudes (in logarithmic scale) evoked by four (1-4) consecutive 2-APB applications followed by the mixture of agonists corresponds to panel A recordings for wild-type (white bars) and aspartic acid mutants (colored bars). The T35D construct yielded a smaller initial response to 2-APB (p Ͻ 0.041, see also supplemental Fig. S1) but exhibited potentiation of the fourth response not significantly different from wild-type channels (3.3 Ϯ 0.5-fold; p ϭ 0.606). The T264D mutant produced significantly larger first responses than the wild-type channels (one-way analysis of variance with Dunnett's post hoc test, p Ͻ 0.001). In this mutant, the first response reached 6.7 Ϯ 3.2% of the maximum response, which was not significantly different from the fourth response, which reached 7.1 Ϯ 3.1% of the maximum amplitude (p ϭ 0.697; paired t test). The 2-APB currents through T343D were barely detectable, being significantly smaller than those of wild-type channels and the other mutants (Student's t test, p ϭ 0.03). The fourth response in T343D remained at the initial level of 0.3% of the maximum (p ϭ 0.987; paired t test), indicating functional rather than phosphorylation-mimicking effects. Apart from S85D (2.7 Ϯ 0.8 nA; one-way analysis of variance, p ϭ 0.007), for all the mutants tested, maximal inward currents elicited by the combination of saturating agonists were not statistically different from wild-type channels (4.8 Ϯ 0.9 nA). Data are shown as mean Ϯ S.D. *, p Ͻ 0.05. n are indicated in brackets. See also supplemental Fig. S1.  Fig. 2A. In T264D and T264E, the fourth response to 2-APB was not statistically different from the initial response (p ϭ 0.303 and 0.232, paired t test), whereas that for wild-type channels the response increased 3-fold and for T264A the response was about 2-fold. In T264D, all four 2-APB responses were significantly larger compared with wild-type TRPV3. There was no significant difference in the maximum currents induced by the combination of saturating agonists between the mutants and the wild-type channels (one-way analysis of variance, p ϭ 0.119). The data represent mean Ϯ S.D., **, p Յ 0.01. n are indicated in brackets. C, threonine 264 is predicted to be accessible for phosphorylation. The Thr 264 residue (blue) shown in the context of ankyrin repeats 3 and 4 (AR3 and AR4) of the mouse TRPV3-ARD structure (Protein Data Bank code 4N5Q), finger 3 is shown in yellow. The key hydrophobic residues responsible for the bending of finger 3 are depicted (Phe 249 , Phe 259 , and Phe 261 , Leu 268 from the inner helix of repeat 4). The hydroxylation site Asn 242 is shown.

Functional regulation of TRPV3 via phosphorylation
together with the phosphonull T264A mutant (Fig. 4B). The 2-APB currents through T264A coexpressed with ERK1 exhibited amplitudes not significantly different from cells transfected with T264A alone. Likewise with wild-type channels, the maximum response induced by a saturating combination of agonists was slightly, but not significantly smaller (p ϭ 0.230). Thus, the wild-type channels in the presence of ERK1 kinase exhibit larger initial responses to 2-APB that do not further sensitize, similar to those in the phosphomimetic mutants T264D and T264E (compare Fig. 4A and Fig. 3, A  and B). Taken together, these results support a possible role of Thr 264 in the ERK1-mediated phosphorylation of the human TRPV3 channel.

T264A mutation results in a substantial decrease in EGF-mediated potentiation
To further confirm the role of Thr 264 in EGF-induced sensitization, wild-type or T264A were transiently transfected to HaCaT cells and assayed electrophysiologically using repeated stimulations with a low concentration of 2-APB, applied four times for 10 s with 30-s intervals, first in the absence and then in the continuous presence of EGF (20 ng/ml; Fig. 5). EGF was pre-applied for 1 min before the fifth 2-APB application. At the end of the series, a combination of 2-APB (200 M) with carvacrol (150 M) was applied for 10 s to obtain a reference saturating response. Compared with the initial 2-APB response, the 8th response increased about 6-fold in WT and 4-fold in T264A. Whereas the fourth response through TRPV3 measured in the presence of EGF reached 11% of the maximum response, the mutant channels exhibited only 2% increase, thus supporting a key role of Thr 264 in EGF-induced potentiation.

Discussion
In this study, we identify threonine 264 in the N-terminal ARD as a putative site for the ERK1-dependent modulation of TRPV3. This conserved residue is located at the beginning of the inner helix of repeat 4, forming part of the general conserved TPLA consensus sequence of the repeat. Although the overall fold of TRPV3-ARD is shared with other TRPV channels, recent structural analysis revealed an apparent uniqueness of this domain (26). In TRPV3, finger 3 atypically bends over toward finger 2 and is stabilized by interactions with the inner helix of repeat 3 (Fig. 3C). The authors of the study demonstrated that mutations disrupting hydrophobic interactions stabilizing the finger 3 conformation severely affect TRPV3 sensitivity to 2-APB stimulation. This characteristic region immediately precedes threonine 264, thus implying a possibility that interactions between this residue and ERK1 may alter the conformation of this important loop and powerfully influence the channel's activation. Besides this, finger 3 contains a motif including asparagine 242, through which the TRPV3 channel can be effectively regulated by oxygen-dependent hydroxylation (27). Although we cannot exclude the possibility that mutations at Thr 264 themselves disturbed a local structure near the functionally important finger 3, there are several lines of evidence supporting our conclusion that TRPV3 could be a substrate for ERK-phosphorylation with Thr 264 as one of the possible targets. First, our experiments obtained from HaCaT cells confirm previous studies in mouse primary keratinocytes (19) and suggest that the potentiating effect of EGF on TRPV3 can be inhibited by two different selective MEK inhibitors, U0126 or PD98059, and hence it is likely mediated via MEK

Functional regulation of TRPV3 via phosphorylation
kinases. Second, phosphomimetic mutations at Thr 264 markedly promoted the activity of TRPV3 (Figs. 2 and 3), whereas the unphosphorylatable T264A mutant exhibited slightly (p ϭ 0.112) lower fourth responses to the subsaturating agonist (Fig.  3B). Third, the overexpression of ERK1 with TRPV3 significantly increased 2-APB responses and prevented their sensitization upon repeated stimulation. This increase was most likely not caused by changes in the level of TRPV3 expression (supplemental Fig. S2). The presence of ERK1, on the other hand, did not affect the nonphosphorylatable T264A mutant. Fourth, T264A mutation resulted in a substantial decrease in EGF-induced potentiation (Fig. 5).
We cannot rule out the presence of some additional phosphorylation site(s) on TRPV3, because a mild increase in 2-APB current responses was observed in cells co-transfected with T264A and ERK1 (Figs. 4B and 5). The secondary potential ERK phosphorylation site could be Ser 102 or less likely Ser 85 (supplemental Fig. S1). Both of these residues are located at the beginning of the N terminus, outside the ARD, and thus highly accessible to the aqueous environment. Although the phosphomimetic mutations at both these serines increased the first normalized 2-APB response resembling a constitutively phosphorylated channel, the S85D mutation reduced maximal inward currents elicited by the combination of saturating agonists (Fig. 2B), which would not be expected from the EGFmediated effects observed in our experiments (Fig. 1).
Recent mass spectrometric-based large-scale surveys (Phos-phoSitePlus database (28)) have identified several in vivo phosphorylation sites in human or mouse TRPV3 (Ser 374 , Thr 378 , Ser 387 , Thr 421 , and Thr 427 ). However, none of these sites fulfills the consensus sequence for ERK. To map putative in vivo ERK phosphorylation sites, we subjected TRPV3 purified from HaCaT cells to mass spectrometry (MS) analysis (supplemental Figs. S3-S5). Using TRPV3-transfected HaCaT cells, this approach identified two endogenous phosphorylation sites, Thr 35 and Thr 37 , within the peptide sequence RPAEITPTKK. Whereas Thr 37 does not fulfill the consensus sequence for ERK, the phosphomimetic mutation at Thr 35 in our experiments shown in Fig. 2B led to a significant decrease in initial response to 2-APB but, upon repeated stimulation, the sensitization was comparable with wild-type channels. Therefore, these sites are probably not involved in the observed functional changes. Next, we have purified TRPV3 from HaCaT cells transfected with TRPV3 together with ERK1 in a ratio of 1:1. This situation corresponds to our experiments shown in Fig. 4. The MS analysis did not reveal any clear phosphorylation site, suggesting that the excess ERK1 kinase may change the phosphorylation status of TRPV3. In the sequence comprising Thr 264 (YQHEGFYF-GETPLALAACTNQPEIVQLLMEHEQTDITSR) the phosphorylation has been detected neither in TRPV3-expressing cells nor in TRPV3-ERK1-expressing cells. Clearly more work is needed before the signaling pathways regulating TRPV3 can be fully understood in the context of EGFR-generated signals. In skin tissues and HEK293 cells, EGFR can directly or indirectly associate with TRPV3, and the EGFR activation results in the tyrosine phosphorylation of TRPV3 (19). Thus, other possible phosphorylation sites in addition to Thr 264 are most likely involved in EGFR-mediated sensitization.
In the study by Cheng et al. (19) the authors explored the regulatory mechanisms of TRPV3 activity via the activation of the EGFR receptor in keratinocytes. Based on Ca 2ϩ measurements, they suggested that EGFR activation leads to the potentiation of TRPV3 activity, and that this effect is mediated by PLC-dependent and partly by MAPK signaling pathways. In our study, we support these observations by using EGF treatment and inhibitors of MEK activation and the MAPK cascade in a human keratinocyte-derived HaCaT cell line. Moreover, we demonstrate that in the absence of Ca 2ϩ , the inhibitors of MEK completely prevent the EGF-induced potentiation of TRPV3 responses (Fig. 1B), suggesting that the MAPK signaling pathway may contribute to TRPV3 sensitization independently of the PLC signaling pathway.
In conclusion, the data presented here identify threonine 264 in the N-terminal ankyrin repeat 3 of TRPV3 as a potential

Functional regulation of TRPV3 via phosphorylation
target for ERK-mediated phosphorylation. The close vicinity of this residue from the structurally and functionally important region comprising an atypical finger 3 and oxygen-dependent hydroxylation site adds further complexity to the cellular signaling pathways involved in the regulation of skin homeostasis via growth factor downstream signaling.

Cell culture and transfection of HaCaT cells
Human immortalized keratinocytes, HaCaT cells (CLS Cell Lines Service GmbH, Germany), were grown in calcium-free DMEM (GIBCO, Invitrogen) supplemented with 10% fetal bovine solution (FBS) in an atmosphere of 95% air and 5% CO 2 at 37°C. Cells were split twice a week in a 1:10 ratio using TrypLE express enzyme solution (Thermo Fisher Scientific). In some experiments (Figs. 2 and 5) HaCaT cells were transiently transfected with 150 ng of human TRPV3 (in pcDNA5/FRT vector, kindly provided by Ardem Patapoutian) or its mutated variant, and 300 ng of GFP (Takara, Japan) using Lipofectamine 2000 (Invitrogen). Cells were grown on cover glasses (24 mm diameter) for 24 -48 h.

Cell culture and transfection of HEK293T cells
HEK293T cells were cultured in Opti-MEM I medium (Invitrogen) supplemented with 5% FBS as described previously (29). The day before transfection, cells were plated in 24-well plates (2 ϫ 10 5 cells per well) in 0.5 ml of medium and became confluent on the day of transfection. The cells were transiently co-transfected with 300 ng of cDNA plasmid encoding wildtype or mutant TRPV3 and with 300 ng of GFP plasmid using the magnet-assisted transfection technique (IBA GmbH, Goettingen, Germany) and then plated on poly-L-lysine-coated glass coverslips. In the experiments with ERK1, cDNA plasmid encoding wild-type or mutant TRPV3 was co-expressed with ERK1 (in pEGFP-C1 vector, Addgene, number 14747) and GFP in the cDNA ratio of 300:300:150. At least three independent transfections were used for each experimental group. The wildtype channel was regularly tested in the same batch as the mutants. The cells were used 24 -48 h after transfection.

Mutagenesis
The mutants were generated by PCR using a QuikChange II XL site-directed mutagenesis kit (Agilent Technologies) and confirmed by DNA sequencing (GATC Biotech, Germany).

Patch clamp recording
Whole-cell membrane currents were recorded by employing an Axopatch 200B amplifier and pCLAMP 10 software (Molecular Devices, Sunnyvale, CA). Patch electrodes were pulled from a glass tube with a 1.5-mm outer diameter. The tip of the pipette was heat-polished, and its resistance was 3-5 megaohm. Series resistance was compensated by at least 70% in all recordings. The experiments were performed at room temperature (23-25°C). Only one recording was performed on any one coverslip of cells to ensure that recordings were made from cells not previously exposed to chemical stimuli.
The extracellular control solution used for transiently transfected HaCaT and HEK293T cells contained: 160 mM NaCl, 2.5 mM KCl, 1 mM CaCl 2 , 2 mM MgCl 2 , 10 mM HEPES, 10 mM glucose, 320 mosmol, adjusted to pH 7.3 with NaOH. To obtain measurable current responses from endogenously expressed TRPV3 channels in HaCaT cells, CaCl 2 was replaced with 1 mM EGTA (Fig. 1). The pipette solution for HEK293T cells contained: 125 mM Cs-glucono-␦-lactone, 15 mM CsCl, 5 mM EGTA, 10 mM HEPES, 0.5 mM CaCl 2 , 2 mM MgATP, 280 mosmol, adjusted to pH 7.3 with CsOH. The pipette solution for HaCaT cells contained: 125 mM Cs-glucono-␦-lactone, 15 mM CsCl, 5 mM EGTA, 10 mM HEPES, 0.5 mM CaCl 2 , 1 mM MgCl 2 , 2 mM MgATP, 0.3 mM NaGTP, 290 mosmol, adjusted to pH 7.3 with CsOH. Experimental solutions containing TRP channel agonists or signaling pathway modulators were always prepared fresh from stock solutions dissolved in DMSO. EGF was dissolved in 10 mM acetic acid at a concentration of 1 mg/ml and then diluted in 10% FBS in DMEM to its final concentration of 10 g/ml. Bovine serum albumin in FBS served as a carrier of EGF and prevented the adsorption of this peptide onto the plastic material. A system for rapid superfusion of the cultured cells was used for drug application (30). All of the chemicals except for MEK1/2 inhibitor U0126 (Cell Signaling) were purchased from Sigma.

Data analysis
Electrophysiological data were analyzed in pCLAMP10 (Molecular Devices). Curve fitting and statistical analyses were done in SigmaPlot 10 (Systat Software). Statistical significance was determined using Student's t test or one-way analysis of variance followed by Dunnett's post hoc test, as appropriate. All data are presented as the mean Ϯ S.D. unless stated otherwise.