Regulation of Ca2+-dependent Desensitization in the Vanilloid Receptor TRPV1 by Calcineurin and cAMP-dependent Protein Kinase*

The vanilloid receptor TRPV1 is a polymodal nonselective cation channel of nociceptive sensory neurons involved in the perception of inflammatory pain. TRPV1 exhibits desensitization in a Ca2+-dependent manner upon repeated activation by capsaicin or protons. The cAMP-dependent protein kinase (PKA) decreases desensitization of TRPV1 by directly phosphorylating the channel presumably at sites Ser116 and Thr370. In the present study we investigated the influence of protein phosphatase 2B (calcineurin) on Ca2+-dependent desensitization of capsaicin- and proton-activated currents. By using site-directed mutagenesis, we generated point mutations at PKA and protein kinase C consensus sites and studied wild type (WT) and mutant channels transiently expressed in HEK293t or HeLa cells under whole cell voltage clamp. We found that intracellular application of the cyclosporin A·cyclophilin A complex (CsA·CyP), a specific inhibitor of calcineurin, significantly decreased desensitization of capsaicin- or proton-activated TRPV1-WT currents. This effect was similar to that obtained by extracellular application of forskolin (FSK), an indirect activator of PKA. Simultaneous applications of CsA·CyP and FSK in varying concentrations suggested that these substances acted independently from each other. In mutation T370A, application of CsA·CyP did not reduce desensitization of capsaicin-activated currents as compared with WT and to mutant channels S116A and T144A. In a double mutation at candidate protein kinase C phosphorylation sites, application of CsA·CyP or FSK decreased desensitization of capsaicin-activated currents similar to WT channels. We conclude that Ca2+-dependent desensitization of TRPV1 might be in part regulated through channel dephosphorylation by calcineurin and channel phosphorylation by PKA possibly involving Thr370 as a key amino acid residue.

The vanilloid receptor TRPV1 is a polymodal nonselective cation channel of nociceptive sensory neurons involved in the perception of inflammatory pain. TRPV1 exhibits desensitization in a Ca 2؉ -dependent manner upon repeated activation by capsaicin or protons. The cAMP-dependent protein kinase (PKA) decreases desensitization of TRPV1 by directly phosphorylating the channel presumably at sites Ser 116 and Thr 370 . In the present study we investigated the influence of protein phosphatase 2B (calcineurin) on Ca 2؉ -dependent desensitization of capsaicin-and proton-activated currents. By using site-directed mutagenesis, we generated point mutations at PKA and protein kinase C consensus sites and studied wild type (WT) and mutant channels transiently expressed in HEK293t or HeLa cells under whole cell voltage clamp. We found that intracellular application of the cyclosporin A⅐cyclophilin A complex (CsA⅐CyP), a specific inhibitor of calcineurin, significantly decreased desensitization of capsaicin-or proton-activated TRPV1-WT currents. This effect was similar to that obtained by extracellular application of forskolin (FSK), an indirect activator of PKA. Simultaneous applications of CsA⅐CyP and FSK in varying concentrations suggested that these substances acted independently from each other. In mutation T370A, application of CsA⅐CyP did not reduce desensitization of capsaicin-activated currents as compared with WT and to mutant channels S116A and T144A. In a double mutation at candidate protein kinase C phosphorylation sites, application of CsA⅐CyP or FSK decreased desensitization of capsaicin-activated currents similar to WT channels. We conclude that Ca 2؉ -dependent desensitization of TRPV1 might be in part regulated through channel dephosphorylation by calcineurin and channel phosphorylation by PKA possibly involving Thr 370 as a key amino acid residue.
The capsaicin receptor TRPV1, a nonselective cation channel expressed predominantly in nociceptive sensory neurons, transduces and integrates various stimuli such as noxious heat (Ͼ42°C), capsaicin, protons, (1,2), the endogenous cannabinoid anandamide (3), lipoxygenase products, and other lipids related to arachidonic acid (4) and ethanol (5). Studies on TRPV1 gene knock-out mice suggest that TRPV1 is essential for the development of thermal hyperalgesia following inflammation or local injection of bradykinin and nerve growth factor (6 -8).
Activation of TRPV1 leads to Ca 2ϩ influx into nociceptive sensory neurons, resulting in membrane depolarization and release of proinflammatory neuropeptides from primary afferent nerve terminals (9). Prolonged or repeated activation of TRPV1 results in desensitization and insensitivity of the receptor to subsequent stimuli (10,11). The physiological role and importance of TRPV1 desensitization is unknown but speculated to be a process of adaptation and regulation of the peripheral nervous system for the perception of pain. Comparable with other ion channels, desensitization of TRPV1 is at least in part a Ca 2ϩ -dependent process (10,11). There is growing evidence for the involvement of Ca 2ϩ -dependent phosphorylation and dephosphorylation processes to regulate desensitization and excitability of TRPV1. Previous studies in rat dorsal root ganglion neurons have demonstrated that desensitization is reduced in the presence of inhibitors of the Ca 2ϩ -and calmodulin-dependent protein phosphatase 2B (calcineurin) (12). Conversely, phosphorylation of TRPV1 by Ca 2ϩ -calmodulin-dependent kinase II (CaMKII) 1 seems to be a prerequisite for activation of TRPV1 by capsaicin (13).
Another candidate involved in the mechanisms of Ca 2ϩ negative feedback and Ca 2ϩ -dependent inactivation in many ion channels is the Ca 2ϩ sensor calmodulin (CaM) itself. There is growing evidence that multiple regions of TRPV1 indeed may bind CaM (14,15). TRPV1 is also a target for cAMP-dependent protein kinase (PKA)-and protein kinase C (PKC)-dependent phosphorylation. Phosphorylation by PKA sensitizes the channel to heat (16) and capsaicin (17) and reduces Ca 2ϩ -dependent desensitization of capsaicin-and proton-activated currents (18,19). Amino acids residues Ser 116 and Thr 370 are the major substrates for PKA-dependent phosphorylation, although other putative PKA phosphorylation sites might be involved as well. Phosphorylation by PKC sensitizes the channel to capsaicin, protons, and heat (20 -22). Here, residues Ser 502 and Ser 800 are the major substrates for PKC-dependent phosphorylation.
In the present study we investigated the influence of cal-cineurin on Ca 2ϩ -dependent desensitization of capsaicin-and proton-activated currents and examined the interactions of calcineurin and PKA and PKC phosphorylation pathways. We found that Ca 2ϩ -dependent desensitization of TRPV1 might be in part regulated through channel dephosphorylation by calcineurin and channel phosphorylation by PKA possibly involving Thr 370 as a key amino acid residue.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis and Transient Transfection-Mutagenesis of rat TRPV1-cDNA was performed with rTRPV1-pcDNA3 by means of the transformer site-directed mutagenesis kit (BD Biosciences Clontech, Palo Alto, CA) as described previously (19). Human embryonic kidney (HEK) 293t cells or cells of a human adenocarcinoma-derived cell line (HeLa cells) were transfected with wild type or mutant plasmid (0.75 or 10 g, respectively) along with reporter plasmid (CD8-pih3m, 1 g) by the calcium phosphate precipitation method. After incubation for 12-15 h, the cells were replated in 35-mm culture dishes. Transfected cells were used for experiments within 2-3 days. Transfection-positive cells were identified by immunobeads (anti-CD-8 Dynabeads; Dynal Biotech, Oslo, Norway). Transfection efficiency was ϳ50 -70% on average for TRPV1-WT and mutant channels.
Chemicals and Solutions-Capsaicin (8-methyl-N-vanillyl-6-nonenamide) and cyclosporin A (CsA; both Sigma-Aldrich) were dissolved in absolute ethanol to give stock solutions of 10 mM. Forskolin (FSK; Calbiochem-Novabiochem GmbH, Bad Soden, Germany), phorbol 12myristate 13-acetate (PMA; Calbiochem-Novabiochem GmbH), and okadaic acid (OA; Alomone Labs, Ltd., Jerusalem, Israel) were dissolved in dimethyl sulfoxide to give stock solutions of 10, 1, and 1 mM, respectively. Human brain CaM and N-(6-aminohexyl)-5-chloro-1naphthalenesulfonamide hydrochloride (W-7; both Calbiochem-Novabiochem GmbH) were dissolved in double distilled water to give stock solutions of 100 mM. Cyclophilin A (CyP; Sigma-Aldrich) was dissolved in Tris-Cl, pH 7.4, containing HEPES, 1,4-dithio-DL-threitol, phenylmethanesulfonyl fluoride, and sodium azide to give a stock solution of 20 M. All of the stock solutions were stored at Ϫ20°C. All of the control and test solutions were applied with a polytetrafluorethylen glass multiple-barrel perfusion system. Standard bath solutions contained 70 mM NaCl, 70 mM choline Cl, 5 mM KCl, 2 mM MgCl 2 , 2 mM CaCl 2 , 10 mM HEPES, and 10 mM glucose (adjusted to pH 7.4 with tetramethylammonium hydroxide). The NaCl/Choline Cl composition was used to reduce the amplitude of the WT currents. Choline Cl did not have any influence on WT or mutant channels. Pipette solution contained 140 mM KCl, 2 mM MgCl 2 , 5 mM EGTA, and 10 mM HEPES (adjusted to pH 7.4 with KOH).
Electrophysiological Technique and Data Analysis-Currents were recorded at room temperature with the whole cell configuration of the patch-clamp method. Holding potential was Ϫ60 mV. Patch pipettes were pulled from boroslilicate glass tubes (TW150F-3; World Precision Instruments, Sarasota, FL) and heat-polished at the tip to give a resistance of 0.8-1.2 M⍀. The currents were recorded with an Axopatch 200B patch-clamp amplifier (Axon Instruments, Union City, CA), filtered at 1 kHz, and sampled at 2 kHz. pCLAMP 8.0.1 software (Axon Instruments) was used for acquisition and analysis of currents. Origin 6.1 software (OriginLab Corporation, Northampton, MA) was used to perform least squares fitting and to create figures. The data are presented as the means Ϯ S.E. or fitted value Ϯ S.E. of the fit. An unpaired Student's t test (SigmaStat; SSPS Science, Chicago, IL) was used to evaluate the significance of changes in mean values. p values Ͻ0.05 were considered statistically significant.

Inhibition of Calcineurin Decreases Ca 2ϩ -dependent Desensitization of Capsaicin-activated TRPV1-WT Currents-TRPV1
channel exhibits desensitization in a Ca 2ϩ -dependent manner (10,11). It has been suggested that a rise in cytosolic Ca 2ϩ level caused by TRPV1 activation results in the activation of Ca 2ϩ / calmodulin-dependent protein phosphatases that mediate channel desensitization (12). To test this hypothesis for TRPV1-WT transiently expressed in HEK293t cells, we first studied the effect of various protein phosphatase inhibitors on Ca 2ϩ -dependent desensitization of capsaicin-activated TRPV1-WT currents, specifically on the decreasing current response to successive stimulation (tachyphylaxis). We applied a series of brief (ϳ5 s long) pulses of 1 M capsaicin at 2-min intervals in Ca 2ϩ -containing solution (2 mM) and measured the current response. Under control conditions, TRPV1-WT showed strong tachyphylaxis (Fig. 1A). Most of the tachyphylaxis occurred between the first and second application, as described previously (19). The mean current amplitudes at the second and fourth capsaicin application were 5.1 Ϯ 2.7 and 3.1 Ϯ 0.4% of that of the first application, respectively (Fig. 1F). Pretreatment of cells for 10 min with 1 M OA in the pipette solution, an inhibitor of protein phosphatase 1 and 2A, did not lead to any change in Ca 2ϩ -dependent tachyphylaxis of capsaicin-activated currents (Fig. 1B). Here, current amplitudes at the second and fourth capsaicin application were 2.5 Ϯ 1.6 and 3.7 Ϯ 2.4% of that of the first application, respectively (Fig. 1F). Similarly, OA was without any effect on Ca 2ϩ -dependent tachyphylaxis when used in both lower or higher concentrations (0.01-100 M, data not shown). In contrast, tachyphylaxis was significantly decreased when cells were pretreated for 10 min . After a whole cell voltage clamp was established, the cells were dialyzed for 10 min before the first capsaicin application. The intervals between capsaicin applications were 2 min. F, mean amplitudes of currents Ϯ S.E. measured in experiments as described for A-E. The amplitudes were normalized to the current amplitude obtained with first capsaicin application. * indicates a statistically significant difference in the mean amplitude compared with that obtained under control conditions. with the immunosuppressive drug cyclosporin A (CsA; 14 nM) together with the "immunophilin" cyclophilin A (CyP; 17 nM) in the pipette solution (Fig. 1D). CsA and CyP are known to form a drug/immunophilin complex, which associates with and thus inhibits protein phosphatase 2B (calcineurin) (23). In the presence of CsA/Cyp, current amplitudes at the second and fourth capsaicin application were 58.4 Ϯ 6.9 and 37.2 Ϯ 6.2% of that at the first application, respectively (Fig. 1F). This is in good agreement with previously reported results obtained in rat dorsal root ganglion neurons using similar concentration of CsA and CyP in the pipette solution (12). Higher concentrations of CsA (up to 1 M) along with CyP (up to 1 M) did not lead to any further decrease in tachyphylaxis compared with that observed with 14 nM CsA plus 17 nM CyP (data not shown). Pretreatment of cells with CsA alone up to concentrations of 100 M in the pipette solution did not lead to any change in Ca 2ϩ -dependent tachyphylaxis of capsaicin-activated currents compared with control conditions, indicating that indeed CsA complexed to CyP is the active form that specifically targets calcineurin ( Fig. 1, C and F). Pretreatment of cells for 10 min with 14 nM CsA and 17 nM CyP along with 1 M OA in the pipette solution did not lead to any further decrease in tachyphylaxis compared with that observed with CsA⅐CyP ( Fig. 1, E and F). Pretreatment of cells with OA, CsA, or CsA⅐CyP in the pipette solutions did not have any significant effect on capsaicin-activated peak currents of TRPV1-WT (Table I).
We also investigated responses of TRPV1-WT channels to prolonged applications (30-s) of 1 M capsaicin in Ca 2ϩ -containing (2 mM) bath solution. TRPV1-WT currents had peak amplitudes in the range of 2.6 -14.4 nA with a mean of 7.1 Ϯ 1.8 nA. Currents reached their peak at 1.6 Ϯ 0.5 s after beginning of activation. The currents then began to decrease rapidly during continuous capsaicin application and reached values of 0.41 Ϯ 0.18, 0.34 Ϯ 0.06, and 0.21 Ϯ 0.03 nA after 10, 20, and 30 s, respectively ( Fig. 2A). This type of desensitization has been termed acute desensitization, consistent with previous reports (10,11,19).
Pretreatment of cells for 10 min with 14 nM CsA plus 17 nM CyP in the pipette solution also led to a significant decrease in acute desensitization of capsaicin-activated TRPV1-WT cur-rents. The currents had peak amplitudes in the range of 2.4-8.9 nA with a mean of 5.0 Ϯ 1.1 nA, which was not significantly different from control conditions ( Table I). The currents reached their peak at 1.9 Ϯ 0.7 s after beginning of activation. Then the currents began to decrease during continuous capsaicin application and reached values of 2.18 Ϯ 0.62, 0.64 Ϯ 0.08, and 0.59 Ϯ 0.05 nA after 10, 20, and 30 s, respectively (Fig. 2B). To describe and compare acute desensitization quantitatively for control conditions and in the presence of CsA⅐CyP, we measured the areas under the current curves over a time of 30 s and normalized them to an idealized (rectangular), nondesensitizing current of a respective size. The data are given in Fig. 2C.
Calmodulin Is Not Required for the Decrease in Ca 2ϩ -dependent Desensitization by Inhibition of Calcineurin-Calcineurin exerts its phosphatase activity in a Ca 2ϩ -and calmodulin-dependent manner. CaM itself is a dominant Ca 2ϩ sensor for Ca 2ϩ -dependent inactivation in many ion channels (24 -26). There is accumulating evidence that multiple regions of TRPV1 may bind CaM. One putative region was identified in the C-terminal (14), and another was identified in the N-terminal segment (15). We investigated the functional role of CaM in Ca 2ϩ -dependent desensitization of capsaicin-activated  TRPV1-WT channels. The experiments were performed as described for those shown in Fig. 1. Pretreatment of cells with 100 M W-7 in the pipette solution, which is a potent noncompetitive antagonist of calmodulin, did not lead to any change in Ca 2ϩ -dependent tachyphylaxis of capsaicin-activated TRPV1-WT currents (Fig. 3, A and C). W-7 also did not have any significant effect on capsaicin-activated peak currents of TRPV1-WT (Table I). This is in good agreement with an earlier report about the ineffectiveness of W-7 on TRPV1 channel desensitization (14). Tachyphylaxis was significantly decreased when cells were pretreated with 100 M W-7 along with 14 nM CsA plus 17 nM CyP in the pipette solution (Fig. 3B). Here, current amplitudes at the second and fourth capsaicin application were 71.7 Ϯ 12.0 and 42.2 Ϯ 8.1% of that of the first application, respectively (Fig. 3C). The decrease in channel tachyphylaxis was comparable with that observed in the presence of CsA plus CyP in the pipette solution (Figs. 1F and 3C). This indicates that inhibition of calcineurin alone is sufficient to inhibit TRPV1-WT channel desensitization. CaM alone up to concentrations of 100 M in the pipette solution did not lead to any change in tachyphylaxis of capsaicin-activated TRPV1-WT currents compared with control conditions (Fig. 3, D and F). Pretreatment of cells with CaM up to concentrations of 100 M along with 14 nM CsA and 17 nM CyP in the pipette solution led to a significant decrease in tachyphylaxis of capsaicin-activated TRPV1-WT currents (Fig. 3E). Here, the results were quantitatively similar to those obtained with CsA⅐CyP alone or CsA⅐CyP along with W-7 in the pipette solution (Figs. 1F and 3, C and F).
Phosphorylation by PKA and Dephosphorylation by Calcineurin Regulates Ca 2ϩ -dependent Desensitization of TRPV1-In the resting state, TRPV1 is highly phosphorylated, at least when heterologously expressed in CHO-K1 cells (18). PKA is able to phosphorylate TRPV1. However, PKA phosphorylation only becomes obvious in the desensitized state (18). Phosphorylation by PKA partly rescues TRPV1 from desensitization (18,19). Because inhibition of calcineurin decreases desensitization of capsaicin-activated TRPV1 currents to a similar extent like activation of PKA, we investigated the interplay between PKA activation and calcineurin inhibition and their effect on Ca 2ϩ -dependent desensitization of capsaicin-activated TRPV1 currents. As demonstrated previously, pretreatment of cells for 10 min with 10 M FSK, an activator of adenylate cyclase and thus an indirect PKA activator, led to a significant decrease in channel tachyphylaxis (Fig. 4A) (19). Current amplitudes at the second and fourth capsaicin application were 62.7 Ϯ 7.8 and 42.8 Ϯ 9.8% of that at the first application, respectively (Fig. 4C). FSK pretreatment did not have any significant effect on capsaicin-activated peak currents in TRPV1-WT (Table I). Pretreatment of cells for 10 min with 14 nM CsA plus 17 nM CyP in the pipette solution along with 10 M FSK in the external solution did not lead to any significant further decrease in tachyphylaxis compared with that in the presence of 10 M FSK alone (Fig. 4, B and C). To exclude the possibility that the decrease in desensitization in the presence of CsA⅐CyP is caused by PKA phosphorylation caused by disinhibition of PKA by CsA⅐CyP, we investigated the effect of CsA⅐CyP on channel tachyphylaxis in the presence of the PKA inhibitor KT5720. In these experiments the decrease in desensitization was similar to that obtained with CsA⅐CyP alone (data not shown).
To address the question of whether the regulation of desensitization by FSK and CsA⅐CyP are of an additive or synergistic nature, we also measured the effect of submaximal concentrations of FSK and CsA⅐CyP on channel tachyphylaxis (Table II). Pretreatment of cells for 10 min with 1.4 nM CsA plus 1.7 nM CyP in the pipette solution did not lead to any significant decrease in tachyphylaxis. With 7 nM CsA plus 8.5 nM CyP in the pipette solution, current amplitudes at the second and fourth capsaicin application were 21.9 Ϯ 2.6 and 17.5 Ϯ 2.4% of that at the first application, respectively (Fig. 4E). Comparable effects were obtained when cells were pretreated with 0.  (Fig. 4F).
These results demonstrate that the concentration dependence for the CsA⅐CyP-mediated decrease in desensitization is rather steep, as expected for effects that require a cascade of reactions rather than simple one-to-one reactions. Thus, we refrained from performing a more detailed quantitative assessment for the interactions of CsA⅐CyP with FSK. However, the results obtained by simultaneous application of FSK and CsA⅐CyP exclude a subadditive action on channel desensitization and are in favor of an additive action for two reasons. First, the effect on tachyphylaxis caused by a submaximal concentration of FSK was unaltered by application of CsA⅐CyP in a concentration that alone was too weak to effect desensitization. Second, simultaneous application of CsA⅐CyP and FSK both in concentrations that alone caused a significant but submaximal decrease in channel tachyphylaxis led to a maximal decrease in tachyphylaxis, meaning a decrease that could not be further enhanced by higher concentrations of CsA⅐CyP or FSK.
Among several putative PKA phosphorylation sites of TRPV1, amino acid residues Ser 116 and Thr 370 seem to be the most critical ones for PKA-dependent modulation of TRPV1 (18,19). Substitution of Ser 116 or Thr 370 with either alanine or aspartate led to mutant channels that could not be modulated by PKA (19). We now investigated the effect of calcineurin inhibition on Ca 2ϩdependent desensitization in TRPV1 mutant channels S116A and T370A. Experiments were performed as described for those shown in Fig. 1. Under control conditions, TRPV1-S116A clearly showed some tachyphylaxis that was however less pronounced compared with TRPV1-WT (Fig. 5A). Here, the current amplitudes at the second and fourth capsaicin application were 22.6 Ϯ 8.7 and 11.0 Ϯ 4.0% of that of the first application, respectively, and were significantly larger than those for TRPV1-WT under   E, and H). After a whole cell voltage clamp was established, the cells were dialyzed for 10 min before the first capsaicin application. The intervals between capsaicin applications were 2 min. C, F, and I, mean amplitudes of currents Ϯ S.E. measured in experiments as described for  A, B, D, E, G, and H. The amplitudes were normalized to the current amplitude obtained with the first capsaicin application. * indicates a statistically significant difference between the mean values obtained without and with CsA⅐CyP treatment. control conditions (Figs. 1F and 5C). Pretreatment of TRPV1-S116A for 10 min with 14 nM CsA and 17 nM CyP in the pipette solution, however, led to a significant decrease in tachyphylaxis compared with control conditions (Fig. 5B). Current amplitudes at the second and fourth capsaicin application were 66.1 Ϯ 4.1 and 38.2 Ϯ 7.9% of that of the first application, respectively (Fig.  5C). This indicates that dephosphorylation by calcineurin of a residue other than Ser 116 might significantly contribute to desensitization.
As described before (19), TRPV1-T370A is one of the least desensitizing alanine mutation at putative PKA phosphorylation sites under control conditions (Fig. 5D). Here, current amplitudes at the second and fourth capsaicin application were 83.2 Ϯ 5.4 and 59.5 Ϯ 7.9% of that of the first application (Figs. 1F and 5F). Pretreatment for 10 min with 14 nM CsA and 17 nM CyP in the pipette solution, however, had no statistically significant effect on channel tachyphylaxis (Fig. 5E). Here, current amplitudes at the second and fourth capsaicin application were 82.1 Ϯ 3.1 and 56.7 Ϯ 5.5% of that of the first application, respectively (Fig. 5F). Because mutation T370A constitutively shows only weak tachyphylaxis, it could be speculated that inhibition of calcineurin will not produce any measurable effect on tachyphylaxis, even if the channel was a substrate for calcineurin. To exclude this possibility, we investigated the effect of calcineurin inhibition on mutation T144A, which is another channel exhibiting impaired tachyphylaxis (19). In this mutation, pretreatment for 10 min with 14 nM CsA and 17 nM CyP in the pipette solution significantly decreased tachyphylaxis compared with control conditions (Fig. 5, G-I). Pretreatment of cells with CsA⅐CyP did not have any significant effect on the peak amplitudes of capsaicin-activated currents in any of these mutant channels (Table I). These observations support the idea that amino acid residue Thr 370 might be a key site for calcineurin-induced dephosphorylation of TRPV1.
Phosphorylation by PKC⑀ Does Not Modulate Ca 2ϩ -dependent Desensitization of Capsaicin-activated TRPV1 Currents-In addition to PKA, the ⑀-isoform of PKC was demonstrated to directly modify TRPV1 and to sensitize heat-and capsaicinactivated currents (20 -22). Here, residues Ser 502 and Ser 800 were suggested to be the major substrates for PKC-dependent phosphorylation (22). However, there is no evidence for a PKCdependent modulation of TRPV1 desensitization so far (19). We were interested in whether or not there is any interplay between PKC phosphorylation and calcineurin-modulation of TRPV1. Experiments were performed as described for those shown in Fig. 1.
As demonstrated before, pretreatment of cells with PMA (0.1 M), an activator of PKC, did not have any effect on channel tachyphylaxis (Fig. 6A). PMA pretreatment also did not have any significant effect on the peak amplitudes of capsaicinactivated currents (Table I). Pretreatment for 10 min with 14 nM CsA and 17 nM CyP in the pipette solution along with 0.1 M PMA in external buffer significantly decreased channel tachyphylaxis (Fig. 6, B and C).
Calcineurin inhibition by CsA⅐CyP also significantly decreased tachyphylaxis in double mutation S502A/S800A, in which putative PKC phosphorylation sites were disrupted (Fig.  6, D, E, and G). In the same double mutation S502A/S800A, FSK decreased tachyphylaxis to a similar extent like CsA⅐CyP (Fig. 6, F and G). Pretreatment of cells with CsA⅐CyP or FSK did not have any significant effect on peak amplitudes of capsaicin-activated currents in mutation S502A/S800A (Table I). These results confirm that phosphorylation of TRPV1 by PKC⑀ indeed is not involved in the channel desensitization process.
Activation of PKA Increases Capsaicin Sensitivity of TRPV1-To determine the effect of calcineurin inhibition by CsA⅐CyP and PKA activation by FSK on the sensitivity of TRPV1-WT, TRPV1-S116A, and TRPV1-T370A toward capsaicin, we measured the concentration dependence of capsaicin responses in Ca 2ϩ -free bath solutions and determined the halfmaximal activating concentrations (EC 50 ) before and after pretreatment with 10 M FSK in the external solution or 14 nM CsA with 17 nM CyP in the pipette solution. Under control conditions, EC 50 for TRPV1-WT was 242 Ϯ 2 nM, the Hill coefficient (h) was 1.9 Ϯ 0.1 (Fig. 7A). These values are in reasonable agreement with values found previously for TRPV1 expressed in HEK293t cells (19). The capsaicin concentrationresponse curve was significantly shifted leftward by ϳ4.5-fold after pretreatment with FSK (EC 50 ϭ 52 Ϯ 2 nM; h ϭ 1.8 Ϯ 0.1; Fig. 7A) but was not significantly changed after pretreatment with CsA⅐CyP (EC 50 ϭ 190 Ϯ 11 nM; h ϭ 1.8 Ϯ 0.2; Fig. 7A).
These results confirm earlier reports that the PKA pathway not only regulates desensitization of TRPV1 but also sensitizes the channel to capsaicin. These results also imply that under resting conditions, the channel is only partly phosphorylated at PKA sites, unlike suggested earlier (18).
These values are in reasonable agreement with values found previously for TRPV1-S116A (19). Comparable with WT, in this mutation the capsaicin concentration-response curve was significantly shifted leftward by ϳ3-fold after pretreatment with FSK (EC 50 ϭ 67 Ϯ 5 nM; h ϭ 1.2 Ϯ 0.1; Fig. 7B) but was not significantly changed after pretreatment with CsA⅐CyP (EC 50 ϭ 180 Ϯ 8 nM; h ϭ 1.2 Ϯ 0.1; Fig. 7B). Earlier, we have demonstrated that in mutation TRPV1-S116, there was a slight reduction in tachyphylaxis after pretreatment with FSK, which, however, was not statistically significant (19). The data in this study confirm that one or several other residues in addition to Ser 116 are very likely involved in the mechanism of PKA-dependent modulation of TRPV1.
For TRPV1-T370A, sensitivity toward capsaicin was significantly lower under control conditions compared with WT and TRPV1-S116A (EC 50 ϭ 41 Ϯ 3 nM; h ϭ 1.2 Ϯ 0.1; Fig. 7C). In this mutation, pretreatment with FSK or CsA⅐CyP did not have any significant effect on the EC 50 values or Hill coefficients. These observations support the idea that amino acid residue Thr 370 might be a key site for PKA-mediated phosphorylation of TRPV1. Concentration-effect experiments were performed in Ca 2ϩ -free solutions to prevent channel desensitization. This way, the ineffectiveness of calcineurin in these experiments can be explained.
Inhibition of Calcineurin Decreases Ca 2ϩ -dependent Desensitization of Proton-activated TRPV1-WT Currents-TRPV1 is a multimodal sensor that in addition to vanilloids is also activated by protons and heat (1, 2), anandamide (3), ethanol (5), various lipoxygenase products, and other lipids related to arachidonic acid (4). Although not unequivocal, activation of TRPV1 by protons has been demonstrated to lead to channel desensitization in a Ca 2ϩ -dependent manner (11). This desensitization can be partly rescued by PKA activation as well (18).
We studied the effect of calcineurin inhibition on desensitization of proton-activated TRPV1 currents. As HEK293t cells were shown to endogenously express an acid sensing ion channel (hASIC1a) (27), channels were transiently expressed in HeLa cells for these experiments.
Under control conditions, TRPV1-WT showed pronounced proton-induced tachyphylaxis (Fig. 8, A and C) that was qualitatively similar to capsaicin-induced tachyphylaxis. Proton-induced tachyphylaxis was significantly decreased when cells were pretreated for 10 min with 14 nM CsA plus 17 nM CyP in the pipette solution (Fig. 8, B and C). Similarly, CsA⅐CyP decreased acute desensitization of TRPV1-WT induced by prolonged application (30 s) of protons (Fig. 8, D-F). Peak amplitude of protonactivated currents were unaffected by CsA⅐CyP (Table I). DISCUSSION In this study, we show that specific inhibition of calcineurin (protein phosphatase 2B) significantly decreases Ca 2ϩ -dependent desensitization of capsaicin-and proton-activated TRPV1 currents. This effect is qualitatively and quantitatively similar to but independent from that obtained by extracellular application of FSK and cannot be further enhanced by simultaneous application of FSK and CsA⅐CyP. In mutation T370A, but not mutations S116A and T144A, desensitization properties are unaffected by calcineurin inhibition. In double mutation S502A/S800A, in which putative PKC phosphorylation sites are disrupted, both calcineurin inhibition and PKA activation decrease desensitization of capsaicin-activated currents similar to WT channels. We conclude that Ca 2ϩ -dependent desensitization of TRPV1 might be in part regulated through channel dephosphorylation by calcineurin and channel phosphorylation by PKA, possibly involving Thr 370 as a key amino acid residue.
Dephosphorylation and Desensitization-Protein phosphorylation and dephosphorylation is a major mechanism in mam- Increasing capsaicin concentrations were applied to cells expressing TRPV1-WT or mutant channels in Ca 2ϩ -free external bath solution in the absence (f; n ϭ 8 for WT, n ϭ 6 for S116A, and n ϭ 5 for T370A) or presence of 10 M FSK in the external solution (OE; n ϭ 10 for WT, n ϭ 6 for S116A, and n ϭ 6 for T370A) or 14 nM CsA ϩ 17 nM CyP in the pipette solution (E; n ϭ 6 each for WT, S116A, and T370A). The intervals between applications were 1 min. The peak amplitudes of capsaicin-activated currents were measured, normalized to the maximum response measured in each cell, and plotted against the capsaicin concentration. The lines represent the fits of the data to the Hill equation. EC 50 values and Hill coefficients (h) are given in the figure. malian cells for regulating structure and function and responding to external stimuli. In a previous study, 32 P labeling and immunoprecipitation revealed that TRPV1 is highly phosphorylated in the resting state, at least when heterologously expressed in CHO-K1 cells, and that phosphorylation could be significantly reduced by application of a desensitizing concentration of capsaicin (18). In another study, dephosphorylation could be prevented by coapplication of capsaicin and FK-506, another calcineurin inhibitor. Moreover, evidence was presented that phosphorylation of TRPV1, presumably at a CaMKII consensus phosphorylation site (Ser 502 /Thr 704 ), is a prerequisite for the capsaicin binding capacity of TRPV1 (13).
The decrease in desensitization by the CsA⅐CyP complex found in this study confirms previously reported results obtained in rat dorsal root ganglion neurons (12) and strongly suggests that dephosphorylation by calcineurin indeed comprises channel activity and thus elicits desensitization. Because calcineurin is a Ca 2ϩ -and calmodulin-dependent phosphatase, this mechanism could account for the Ca 2ϩ dependence of desensitization. Similar mechanisms have been suggested to underlie the desensitization of the ionotropic receptor P2X3 (28), the neuronal nicotinic acetylcholine receptor (29) and the N-methyl-D-aspartate receptor (30,31). Interestingly, CsA alone had no effect on desensitization, indicating that HEK293t cells might not contain sufficient cyclophilin A to allow formation of the inhibitory complex, at least not under our experimental conditions. Calcineurin inhibition not only decreased channel tachyphylaxis but also decreased acute desensitization elicited by a 30-s-long capsaicin application, an effect that could not be observed with PKA activation in an earlier study (19). This disparity in effects between calcineurin inhibition and PKA activation suggests that dephosphorylation might be a faster process as compared with rephosphorylation. Moreover, dephosphorylation might not require a closed or ligand-free channel, as hypothesized for the process of rephosphorylation.
Calcineurin inhibition also decreased desensitization of proton-activated TRPV1 currents under our experimental conditions. Unlike suggested by others (13), we conclude that activation by capsaicin and activation by protons most probably initiate comparable mechanisms of Ca 2ϩ -dependent desensitization.
The Role of CaM-Calcineurin acts in a Ca 2ϩ -and calmodulindependent manner. CaM is a dominant Ca 2ϩ sensor for Ca 2ϩdependent inactivation in many ion channels (24 -26). There is accumulating evidence that multiple regions of TRPV1 may bind CaM. One putative region was identified in the C-terminal (14), and another was identified in the N-terminal segment (15).
In our experiments both CaM and the calmodulin antagonist W-7 did not have any effect on TRPV1 desensitization or peak amplitudes of capsaicin-activated currents. The ineffectiveness of the CaM inhibitor is in good agreement with an earlier report (14) and might suggest that CaM is not involved in Ca 2ϩ -dependent desensitization of capsaicin-activated TRPV1 currents. However, CaM inhibitors would only be expected to inhibit desensitization if CaM acted as a free molecule (32,33). This, however, might not be the case as suggested by recent reports (15,22).
CaM was demonstrated to mediate Ca 2ϩ inhibition of TRPV1 in inside-out excised patches of Xenopus oocytes and HEK293 cells expressing TRPV1. In that study, CaM was applied to the intracellular site of the channel together with 50 M free Ca 2ϩ (15). We cannot exclude that the ineffectiveness of CaM in our study is due to the rather uncontrolled intracellular free Ca 2ϩ concentration under our experimental conditions. Thus, from our data, we cannot derive profound evidence for or against a role of CaM in TRPV1-WT channel desensitization.
Functional Coupling of Calcineurin and PKA-The decrease in desensitization of capsaicin-activated currents by calcineurin inhibition in this study was qualitatively and quantitatively similar to that obtained by PKA activation. Simultaneous application of FSK and CsA⅐CyP in submaximal and maximal concentrations suggest additive actions of FSK and CsA⅐CyP rather than subadditive or synergistic actions. In mutation T370A, in which a putative PKA phosphorylation site is disrupted, desensitization properties were unaffected by calcineurin inhibition. Calcineurin inhibition, however, did reduce desensitization in both mutations S116A, in which another critical PKA phosphorylation site is disrupted (18,19), and mutation T144A, which shares similar desensitization properties with mutation T370A (19). These results suggest that Ca 2ϩ -dependent desensitization of TRPV1 might be in part regulated through channel dephosphorylation by calcineurin and channel phosphorylation by PKA possibly involving Thr 370 as a key amino acid residue. A functional coupling of calcineurin and PKA was proposed before in mouse ventricular myocytes to control Ca 2ϩ influx through Ca 2ϩ channels and Ca 2ϩ release through ryanodine receptors (34). In these cells, immunofluorescence also revealed colocalization of calcineurin and PKA.
We hypothesize that a similar mechanism could control excitability of nociceptive sensory neurons by regulating desensitization and thus channel availability. The interplay between After a whole cell voltage clamp was established, the cells were dialyzed for 10 min before the first capsaicin application. The intervals between proton applications were 2 min. C, mean amplitudes of currents Ϯ S.E. measured in experiments as described for A and B. The amplitudes were normalized to the current amplitude obtained with the first proton application. * indicates a statistically significant difference in the mean amplitude compared with control. D and E, whole cell current responses of TRPV1 channels evoked by a 30-s-long application of protons (pH 5.0) in Ca 2ϩ -containing (2 mM) bath solution without (D) and with 14 nM CsA ϩ 17 nM CyP in the pipette solution (E). F, the areas under the current curves were measured and normalized to an idealized, nondesensitizing current of respective size. The bars represent the mean values Ϯ S.E. * indicates a statistically significant difference compared with control.
CaMKII and PKA remains to be characterized in future studies. TRPV1 channel phosphorylation by PKA (16) and PKC (22) to control activation thresholds and TRPV1 channel phosphorylation/dephosphorylation by PKA (18,19), CaMKII (13), and calcineurin to regulate desensitization/availability might allow fine tuning of the nociceptor in response to a noxious environment.