Mobilization of intracellular Ca²⁺ leads to TRPV1 activation

To establish whether the ‘transients' of intracellular anandamide described above have any functional consequences at the cellular level, we used a Fura-2-AM-based fluorescence method with a Ca²⁺ reconstitution protocol (Figure 2A). We tested whether stimulation of muscarinic receptors leads to TRPV1 activation in hTRPV1-HEK293 cells. In an extracellular Ca²⁺-free medium, carbachol dose-dependently released Ca²⁺ from the ER (Figure 2B) in a way blocked by the nonselective muscarinic receptor antagonist, atropine, the selective IP₃ receptor antagonist, xestospongin C, and the inhibitor of PI-selective PLC, U73122, but not by its inactive analogue U73343 (not shown). Thus, carbachol activates the PLC/IP₃ pathway through muscarinic receptors. It is well known that mobilization of intracellular Ca²⁺ from the ER results in extracellular Ca²⁺ influx (Venkatachalam et al, 2002; Zitt et al, 2002; Clapham, 2003; Putney, 2003). Indeed, upon reconstitution of 2.5 mM Ca²⁺ in the extracellular medium, we observed an immediate rise in intracellular Ca²⁺ levels. This increase in intracellular Ca²⁺ was dependent on the dose of carbachol (Figure 2A–C) and was not observed when the cells were pretreated with atropine and U73122 (Figure 2C). Although significant Ca²⁺ entry was observed also in the absence of any stimuli (Figure 2A, lower trace), likely due to leaky cells, the amount of stimulus-induced Ca²⁺ entry was still very strong. In fact, even after correction for leaky cells (which accounted for ~19% of the total Ca²⁺ entry and about 36–40% of the carbachol-induced Ca²⁺ entry), the stimulus-to-vehicle ratio was still >2 and the maximum change in fluorescence ratio (ΔF_340/380) upon addition of extracellular Ca²⁺ was 1.34±0.15 for 50 μM carbachol (see Materials and methods). Importantly, with 10 μM xestospongin C, a selective IP3 receptor inhibitor, we observed a significant (P<0.05) inhibition of 50 μM carbachol-induced Ca²⁺ influx (Figure 2C). Furthermore, when both atropine (1 μM) and U73122 (1 μM) were given to cells after the carbachol-induced peak of intracellular calcium, they did not significantly affect carbachol-induced Ca²⁺ influx (89.8±4.3 and 88.1±4.7% of control, N=3, P>0.05). This indicates that a brief stimulation by muscarinic receptors and the PLC/IP₃ pathway of intracellular Ca²⁺ mobilization are both necessary and sufficient to induce extracellular Ca²⁺ entry.

Open in a separate window

Figure 2

Mobilization of intracellular Ca²⁺ leads to TRPV1 activation. (A) Typical fluorescence trace of a Ca²⁺ reconstitution experiment. Arrows indicate time point of compound addition (V, vehicle; carbachol (upper trace, CCH, 50 μM, or vehicle, lower trace); CaCl₂ (2.5mM); CPS (0.1 μM) and ionomycin (Iono, 4 μM). (B) Carbachol releases Ca²⁺ from the ER (punctuated line) (n=7–12 for each data point), followed by extracellular Ca²⁺ influx (dashed line) upon Ca²⁺ reconstitution. CPS (0.1 μM) stimulation induces Ca²⁺ influx (solid line, right Y-axis). (C) Effects of pretreatment with atropine (1 and 100 μM), U73122 (5 and 1 μM), U73343 (1 μM), xestospongin (XeC, 10 μM), or (D) of 10 min pretreatment with CPS (0.1 μM) CPZ (10 μM), 5-iodo-resiniferatoxin (I-RTX, 1 nM), SKF96365 (SKF, 30 μM) on 50 μM carbachol-induced Ca²⁺ entry in hTRPV1-HEK293 (ΔF_340/380 (50 μM CCH)=1.34±0.15); wt, wild-type cells. (E, H) TRPV1 antagonists do not reduce calcium influx in wt HEK293 cells. (F) Thapsigargin (1 μM) induced extracellular Ca²⁺ entry in hTRPV1-expressing HEK293 cells (ΔF_340/380 (1 μM Thps)=0.96±0.07) and wt HEK293 cells. The effects of CPZ (10 μM), ruthenium red (RR, 10 μM), IBTU (2.5 μM), 5′-iodo-resiniferatoxin (I-RTX, 10 nM), SKF96365 (SKF, 30 μM) and desensitization with 0.1 μM CPS are also shown. (G) In the presence of extracellular Ca²⁺, thapsigargin (1 μM) increases intracellular Ca²⁺ levels. The effects of selective hTRPV1 antagonists 5-iodo-resiniferatoxin (I-RTX, 10 nM) and 6-iodo-nordihydrocapsaicin (6I-CPS, 10 μM) are shown.

Three lines of evidence suggest that the TRPV1 channel is involved in this carbachol-induced influx of extracellular Ca²⁺. Firstly, carbachol-induced extracellular Ca²⁺ influx was significantly less pronounced in wild-type cells than in hTRPV1-HEK293 cells (Figure 2D). It is noteworthy that basal Ca²⁺ levels were not significantly different in wild-type HEK293 cells from hTRPV1-expressing HEK293 cells: F_340/380=1.26±0.03 (n=7) versus F_340/380=1.36±0.09 (n=12), respectively. Secondly, when TRPV1 was desensitized by pretreatment with its prototypical ligand CPS, before stimulation with carbachol, extracellular Ca²⁺ influx was reduced to levels similar to those observed in wild-type cells (Figure 2D), without affecting carbachol-induced Ca²⁺ transient (see below). Conversely, CPS-induced Ca²⁺ influx was reduced when the cells were pretreated with carbachol (see also below) (Figure 2B). Finally, we found that SKF96365, a nonselective TRPV1 antagonist, completely abolished carbachol-induced Ca²⁺ entry in hTRPV1-HEK293 cells. Furthermore, two other structurally different TRPV1 antagonists, that is, 5′-iodo-RTX (I-RTX) and capsazepine (CPZ), were able to attenuate carbachol-induced Ca²⁺ influx (Figure 2D). This attenuation was not observed in wild-type cells (Figure 2E), thereby excluding nonselective effects of the antagonists.

Similar results were obtained when thapsigargin (1 μM), instead of carbachol, was used to mobilize intracellular Ca²⁺ in hTRPV1-HEK293 (Figure 2F and H). This is in line with the finding that thapsigargin can indirectly produce inward currents in part via TRPV1 (Liu et al, 2003). In fact, pretreatment of cells with the TRPV1 antagonists I-RTX and CPZ and with SKF96365 significantly reduced thapsigargin-induced Ca²⁺ influx after extracellular Ca²⁺ reconstitution. A high concentration (5 μM) of U73122, given to cells before thapsigargin, affected thapsigargin-induced Ca²⁺ influx (43.5±3.6% of control, N=3, P<0.05), but not thapsigargin-induced Ca²⁺ mobilization (83.5±8.6% of control, N=3, P>0.05), in agreement with its capability of directly interacting with channels involved in extracellular Ca²⁺ influx after Ca²⁺ mobilization from intracellular stores. These data reinforce the hypothesis that the mobilization of intracellular Ca²⁺ from the ER is sufficient to activate TRPV1. When Ca²⁺ was present in the extracellular medium since the beginning, and during thapsigargin stimulation, similar results were obtained (Figure 2G), thus possibly suggesting that the presence of leaky cells did not strongly bias the data obtained above with the Ca²⁺ reconstitution protocol. Altogether, these data indicate that TRPV1 is not only a receptor-operated channel, but is also gated following intracellular Ca²⁺ mobilization. Thus, in addition to sensitization or activation by removal of its inhibition by phosphatidylinositolbisphosphate via PLC-mediated hydrolysis (Chuang et al, 2001) or by protein kinase C-catalysed phosphorylation following PLC-mediated diacylglycerol release (Premkumar and Ahern, 2000; Tominaga et al, 2001), TRPV1 can also be activated by increases of intracellular Ca²⁺.

The TRPV1 channels activated by intracellular Ca²⁺ mobilization are likely located on the plasma membrane

Apart from the plasma membrane, a population of TRPV1 receptors was described to be present also on the ER of cells overexpressing this protein (Liu et al, 2003; Turner et al, 2003). Although our calcium fluorescence technique does not allow us to distinguish between these two populations of TRPV1 channels, four lines of evidence strongly suggest that only the plasma membrane population gates Ca²⁺ following Ca²⁺ mobilization by thapsigargin: (1) In the absence of extracellular Ca²⁺, neither CPS (up to 20 μM) nor anandamide (up to 50 μM) released intracellular Ca²⁺ in hTRPV1-HEK293 cells (not shown); we checked with 1 μM thapsigargin or 50 μM carbachol the status of the intracellular stores after addition of CPS (20 μM) or anandamide (50 μM), and in both cases the two former compounds were still capable of strongly releasing intracellular Ca²⁺ (thapsigargin after CPS ΔF_340/380=0.25±0.02, n=3, thapsigargin after anandamide ΔF_340/380=0.24±0.02, n=3, thapsigargin alone ΔF_340/380=0.27±0.04; n=6; carbachol after CPS ΔF_340/380=0.28±0.04, n=3, carbachol after anandamide ΔF_340/380=0.25±0.03, n=3, carbachol alone ΔF_340/380=0.32±0.04; n=6). (2) Only RTX at 4 μM and CPS at 100 μM, that is, at doses that are 100-fold higher than those sufficient for TRPV1 activation at the plasma membrane, mobilized intracellular Ca²⁺. However, pretreatment with 1 μM thapsigargin reduced RTX-induced intracellular Ca²⁺ release by 75% (from ΔF_340/380=0.17±0.03, n=7 to ΔF_340/380=0.04±0.01, n=6). This suggests that the RTX-sensitive Ca²⁺ pool in the ER is overlapping with the thapsigargin-sensitive pool to a great extent. Therefore, emptying the calcium stores with thapsigargin also largely depletes the RTX-sensitive store and cannot lead to its activation. (3) Ruthenium red, a cell-impermeable nonselective TRPV1 antagonist, significantly reduced thapsigargin-induced extracellular Ca²⁺ influx in hTRPV1-HEK293, but not wild-type, cells (Figure 2F). (4) IBTU, a specific TRPV1 antagonist selective for the plasma membrane population of TRPV1 channels (Toth et al, 2004), significantly reduced thapsigargin-induced extracellular Ca²⁺ influx (−35.9±3.8%, n=3; Figure 2F), but not RTX-induced mobilization of intracellular Ca²⁺ in the absence of extracellular Ca²⁺ (+0.33±1.9%, N=3, not shown), in hTRPV1-HEK293 cells. IBTU also significantly reduced carbachol-induced extracellular Ca²⁺ influx by 32.2±4.1% (P<0.05, n=5, not shown).

Formation of intracellular anandamide following calcium mobilization causes TRPV1 activation and calcium influx

Formation of intracellular anandamide, described in the first section of this paper, and Ca²⁺ entry mediated by plasma membrane TRPV1, described in the previous section, both induced by carbachol- and thapsigargin-induced mobilization of intracellular Ca²⁺ stores, might be two independent events. However, two observations indicate that the transient peak of intracellular anandamide observed here causes Ca²⁺ entry by acting at TRPV1: (1) Exogenously added anandamide induces Ca²⁺ influx via TRPV1 in hTRPV1-HEK293 cells (EC₅₀=261±13 nM; n=3, which is in the range of the intracellular concentrations produced by thapsigargin and carbachol stimulation); this effect is blocked by CPZ and I-RTX and is not observed in wild-type HEK293 cells (not shown). (2) Anandamide (100 nM), when injected into the cell, is very efficacious at inducing TRPV1-mediated plasma membrane currents in neurons expressing high levels of this channel (Evans et al, 2004). These latter observations suggest that this lipid induces elevated intracellular Ca²⁺ by activating a predominant population of TRPV1 on the plasma membrane.

If anandamide acts as an intracellular messenger at TRPV1, modulation of its concentration in the cell should affect TRPV1 activity, and hence TRPV1-mediated Ca²⁺ entry. Intracellular anandamide can be inactivated through two concurrent processes, that is, (1) its extrusion from the cell, which is mediated by a specific membrane mechanism selectively inhibited by some synthetic fatty acid amides, and (2) intracellular enzymatic hydrolysis (Piomelli, 2003). Preincubation of cells with two different classes of selective inhibitors of anandamide transport through the plasma membrane with no direct activity at TRPV1, VDM11 and OMDM-1 (De Petrocellis et al, 2000; Ortar et al, 2003) significantly enhanced Ca²⁺ influx in hTRPV1-HEK293 cells (Figure 3A and D). This was not observed in wild-type cells (Figure 3B and C) nor when thapsigargin or carbachol were omitted (not shown), thus demonstrating that anandamide is acting only through TRPV1 and not any other endogenously expressed Ca²⁺-permeable channel. Importantly, VDM11 concomitantly increased the percentage of intracellular anandamide levels after 30 min of stimulation from 28±5 to 63±5% (n=3, P<0.05). These data indicate that the efflux of anandamide from the cell is tonically limiting its actions on TRPV1.

Open in a separate window

Figure 3

Anandamide metabolism and efflux from cell limit TRPV1 activation. (A–D) Carbachol- and thapsigargin-induced Ca²⁺ influx (ΔF_340/380 (50 μM CCH)=1.34±0.15 and ΔF_340/380 (1 μM Thps)=0.96±0.07) in hTRPV1-HEK293 cells is enhanced by pretreatment with inhibitors of anandamide metabolism MAFP (10 μM) and econazole (ECO, 10 μM) and by anandamide transport inhibitors VDM11 (4 μM) and OMDM1 (10 μM), but (B, C) not in wild-type cells. In panel D, extracellular Ca²⁺ was present during thapsigargin treatment. (E) Transient overexpression of FAAH in hTRPV1-HEK293 cells leads to a reduction of thapsigargin-induced extracellular Ca²⁺ influx. Transient expression of 12-lipoxygenase (12-LOX) does not influence thapsigargin-induced Ca²⁺ influx. (F) Thapsigargin-induced desensitization of TRPV1 receptor, as measured by Ca²⁺ influx upon a stimulation with CPS, is abolished when FAAH is transiently overexpressed.

HEK293 cells possess also a fatty acid amide hydrolase (FAAH) that degrades anandamide to arachidonic acid and ethanolamine (Cravatt et al, 1996). To further support our hypothesis that anandamide, produced following store emptying, tonically activates TRPV1, and to rule out the contribution of arachidonic acid metabolites in this process (Hwang et al, 2000; Watanabe et al, 2003), we used an inhibitor of FAAH, methyl-arachidonoyl-fluoro-phosphonate (MAFP). This compound also inhibits the Ca²⁺-dependent phospholipase A₂, the enzyme mostly responsible for arachidonic acid mobilization. MAFP increased carbachol- and thapsigargin-induced extracellular Ca²⁺ influx in hTRPV1-HEK293 cells (Figure 3A and D), but not in wild-type cells (Figure 3B), and did not enhance Ca²⁺ influx in hTRPV1-HEK293 cells in the absence of thapsigargin or carbachol (not shown). Finally, when the cells were pretreated with the cytochrome P450 inhibitor econazole, there was even an enhancement of thapsigargin-induced Ca²⁺ influx (Figure 3D). These data argue against the involvement of arachidonic acid metabolites in TRPV1-mediated Ca²⁺ entry, and again indicate that anandamide is tonically and selectively activating TRPV1 and not any other endogenously expressed TRP channel.

To further demonstrate that anandamide, and not any other putative TRPV1 ligand (e.g. lipoxygenase metabolites or N-arachidonoyldopamine), acts as an intracellular messenger at TRPV1, we used also nonpharmacological approaches. First, we reduced anandamide levels either by overexpression of FAAH or by elimination of arachidonic acid, the ultimate biosynthetic precursor of anandamide. Transient overexpression of rat FAAH into hTRPV1-HEK293 cells increased anandamide hydrolysis from 180±56 to 3385±102 pmol/min/mg protein (n=3, P<0.05) and concomitantly decreased thapsigargin-induced Ca²⁺ influx to the same level as in wild-type HEK293 cells (compare Figure 3E with Figure 2F). The CPS response in the thapsigargin-stimulated, FAAH-transfected cells was comparable to vehicle-stimulated cells (Figure 3F). This indicates that the thapsigargin-induced desensitization of TRPV1 was absent when anandamide degradation was upregulated. Cells cultured for 2 days in the presence of 10 μM CP-24879, a mixed Δ⁵/Δ⁶-desaturase inhibitor that blocks arachidonic acid formation from essential fatty acid precursors (Obukowicz et al, 1998), showed a strongly reduced thapsigargin-induced Ca²⁺ influx from ΔF_340/380=0.96±0.08 to 0.24±0.02 (n=4, P<0.05). At the same time, these cells, when stimulated for 15 min with thapsigargin (1 μM), produced significantly lower intracellular amounts of anandamide with respect to cells cultured for 2 days with vehicle (0.6±0.2 versus 1.4±0.3 pmol/10⁷ cells, P<0.05, n=3). Finally, hTRPV1-HEK293 cells transiently transfected with human 12-lipoxygenase (Figure 1E), although potentially capable of producing compounds active at TRPV1 (Figure 1F), did not exhibit an increased thapsigargin-induced Ca²⁺ influx (Figure 3E). Altogether, these findings indicate that anandamide, and not other putative endogenous TRPV1 agonists, mediates thapsigargin- and carbachol-induced Ca²⁺ influx via TRPV1.

Intracellular anandamide leads to TRPV1 activation in rat sensory neurons

To study whether the formation of anandamide induced by intracellular Ca²⁺ mobilization also leads to endogenous activation of TRPV1 in a native system, we have used neonatal primary sensory neurons from dorsal root ganglia (DRG). These cells constitutively express high levels of TRPV1 and synthesize anandamide in a Ca²⁺-dependent manner (Tognetto et al, 2001; Trevisani et al, 2002; Ahluwalia et al, 2003a, 2003b). As in hTRPV1-HEK293 cells, mobilization of intracellular Ca²⁺ by thapsigargin in rat sensory neurons induced the formation of anandamide whose intracellular concentration was significantly enhanced by pretreatment with VDM11 (Figure 4A). We found, using single-cell Ca²⁺ fluorescence imaging, that thapsigargin-induced extracellular Ca²⁺ entry in rat DRG neurons can be reduced by over 50% by CPZ or I-RTX (Figure 4B). Furthermore, thapsigargin-induced extracellular Ca²⁺ entry was strongly enhanced by VDM11 and MAFP (Figure 4B). Thus, these data indicate that endogenously formed anandamide can activate TRPV1 in rat primary sensory neurons.

Open in a separate window

Figure 4

Mobilization of intracellular Ca²⁺ leads to anandamide formation and subsequent TRPV1 activation in sensory neurons of rat DRG. Intracellular levels of anandamide in rat DRG neurons (25 000 neurons/dish) upon (A) thapsigargin (1 μM) treatment for 30 min of neonatal neurons, and (C) ATP (100 μM) treatment (15 min) of adult neurons. In both cases, an extracellular Ca²⁺-free medium was used, but otherwise the conditions were as described (Ahluwalia et al, 2003b). In panel C, the effects of anandamide transport inhibitor VDM11 (4 μM) and PLC inhibitor U73122 (5 μM) are also shown. (B) Thapsigargin (1 μM)-induced extracellular Ca²⁺ influx in neonatal rat DRG neurons. The effects of CPZ (10 μM), I-RTX (10 nM), MAPF (10 μM) and VDM11 (4 μM) are shown. (D) ATP (100 μM)-induced extracellular Ca²⁺ entry can be attenuated by pretreatment with CPZ (10 μM), I-RTX (10 nM), CPS (10 μM) and RTX (10 nM), and enhanced by pretreatment with VDM11 (4 μM) in adult rat DRG neurons. The effect of VDM11 is not observed when ATP-induced Ca²⁺ release from ER is blocked by U73122 (5 μM). At these doses, CPZ, MAFP and VDM11 did not affect Ca²⁺ entry in the absence of ATP.

ATP has been shown to induce TRPV1-mediated inflammatory hyperalgesia through activation of the G_q/11-coupled P₂Y receptors in primary DRG sensory neurons (Tominaga et al, 2001). Here, we show that anandamide contributes to ATP-mediated Ca²⁺ influx in adult rat DRG neurons (Figure 4C and D). In the absence of extracellular Ca²⁺, ATP mobilized intracellular Ca²⁺ and significantly enhanced intracellular anandamide levels, which were even further increased by blocking its transport out of the neurons with VDM11 (Figure 4C). Upon reconstitution of Ca²⁺, the subsequent Ca²⁺ entry (Figure 4D) could be significantly attenuated by CPZ (−28%) and I-RTX (−93%), and by desensitization with CPS (−77%) or RTX (−90%). This latter process has recently been shown to be specific for TRPV1-expressing neurons (Karai et al, 2004). ATP-induced Ca²⁺ entry was again strongly enhanced by VDM11 (+74%). This latter effect was also blocked by CPZ (not shown). When DRG neurons were pretreated with U73122, the enhanced Ca²⁺ influx by both ATP and VDM11 was inhibited (Figure 4D). Accordingly, anandamide levels were also reduced in ATP-stimulated neurons pretreated with U73122 (Figure 4C). Thus, the production of intracellular anandamide and its subsequent activation of TRPV1 are dependent on a functionally active PLC/IP₃ pathway.

Cell culture and transfection

The hTRPV1-HEK293 and wild-type cells were grown as reported (De Petrocellis et al, 2001). Transient transfections of hTRPV1-HEK293 cells with pcDNA3 vectors containing 12- or 15-lipoxygenase or FAAH were performed with Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. In some experiments, the amount of polyunsaturated fatty acids in hTRPV1-HEK293 cells was reduced by adding 10 μM CP-24879 (Cayman Chemicals) to the medium for 2 days to block biosynthesis of arachidonic acid (Obukowicz et al, 1998). It is noteworthy that serum levels were stepwise reduced in 4 days to 0% before CP-24879 incubation, because it is a source of exogenous polyunsaturated fatty acids.

Anandamide and N-arachidonoyldopamine quantification

Confluent hTRPV1-HEK293 cells or adult rat DRG neurons were washed with phosphate-buffered saline without Mg²⁺/Ca²⁺ and incubated with thapsigargin (1 μM), carbachol (50 μM) or ATP (100 μM) for indicated time in Hanks' balanced salt solution (+2.5 mM EGTA, pH 7.4) without CaCl₂ at room temperature. Atropine and U73122 were added to the cells 5 min before stimulation. In some experiments cells were loaded with BAPTA/AM prior to stimulation. Incubation was terminated by removing the buffer. Lipids were extracted from the cells with Bligh and Dyer method and anandamide, released in the incubation buffer, was recovered by solid phase extraction. Reversed-phase HPLC with isotope dilution mass spectrometry, using d⁸-anandamide (10 pmol) and d⁸-N-arachidonoyldopamine (50 pmol) as internal standard, was applied to quantify anandamide and N-arachidonoyldopamine as described (Di Marzo et al, 2001; Huang et al, 2002).

RT–PCR analysis

The expression of mRNAs for NAPE-PLD in HEK cells, HEK-hVR1 cells and rat DRG neurons was examined by RT–PCR. The PCR primers for NAPE-PLD were selected on the basis of the sequence of the NAPE-PLD mRNA (NCBI accession number AB112352). The NAPE-PLD sense and antisense primers were 5′-TGGACTGGTGGGAGGAG-3′ (nt 689–705) and 5′-GGTTCATAAGCTCCGATGGG-3′ (nt 919–938), respectively. The expected size of the amplicon was 249 bp for NAPE-PLD.

Enzyme assays

Analyses of the activity of FAAH and 12- and 15-lipoxygenase in cell homogenates and PLD with lipid extracts to determine NAPE levels were conducted as described (Di Marzo et al, 2001; Veldhuis et al, 2003).

Measurement of changes in [Ca²⁺]_i

Measurement of intracellular Ca²⁺ fluorescence in cultured rat DRG was made as previously described (Trevisani et al, 2002; Liu et al, 2003). Cells were loaded with Fura-2-AM (3 μM) in Ca²⁺-free buffer solution composed of (mM) KCl 5.4, MgSO₄ 0.4, NaCl 135, D-glucose 5, HEPES 10 with BSA 0.1%, 2.5 EGTA and 0.1% bovine serum albumin, for 40 min at 37°C, at pH 7.4. The dye was excited at 340 and 380 nm to indicate relative [Ca²⁺]_i changes by the F₃₄₀/F₃₈₀ ratio. Changes in fluorescence were monitored after a 10 min stabilization, and in a 0–300 s time interval after neuron stimulation and over 300 s after Ca²⁺ addition. Ca²⁺ fluorescence spectrofluorometry with hTRPV1-HEK293 and wild-type cells was performed in a quartz cuvette with Fura-2-AM or Fluo-3 (in case Ca²⁺ was present during thapsigargin incubation) as described (De Petrocellis et al, 2001). The Ca²⁺ reconstitution protocol was as follows: 300 s pretreatment with vehicle or compound as specified, then store-depleting agent, then 300 s post-treatment (in the case of CPZ), followed immediately by CaCl₂ (2.5 mM), then CPS (0.1 μM) and finally ionomycin (4 μM). Ca²⁺ influx was calculated as the maximum increase in F₃₄₀/F₃₈₀ ratio upon addition of extracellular Ca²⁺ minus the nonselective Ca²⁺ influx as measured with a blank (no agonist/thapsigargin added) to correct for leaky cells. The Ca²⁺ signal in HEK293 cells was calibrated previously (Bisogno et al, 2001) and normalized to total Ca²⁺ influx in Figures 2C–H, ​,33 and ​and4.4. After correction for leaky cells, the maximum change in fluorescence ratio (ΔF_340/380) upon addition of extracellular Ca²⁺ was 1.34±0.15 and 0.96±0.07 for 50 μM carbachol and 1 μM thapsigargin, respectively.

Whole-cell patch-clamp experiments

All recordings were made from the somata of small-diameter DRG neurons (capacitance range was 20–60 pF), at ambient room temperature (20–24°C), with the whole-cell patch-clamp technique using an Axon 200B patch-clamp amplifier and pCLAMP software (Axon Instruments). Patch pipettes of resistance 1.5–2 MΩ were prepared from borosilicate glass using a Sutter Instruments P-87 horizontal micropipette puller (Novato, CA, USA). Pipette capacitance was compensated electronically before breaking into the whole-cell mode, and the series resistance after establishing the whole-cell mode (which ranged from 4 to 7 MΩ), calculated from the time constant of current decay in response to a voltage step, was routinely checked to ensure that it did not increase in the course of an experiment, but was not compensated given the small size of currents recorded. Only one recording was performed on each culture dish in order to ensure that data were not obtained from cells that had been inadvertently exposed to test treatments.

Recordings were performed in an extracellular medium of the following composition (mM): 140 NaCl, 10 CaCl₂, 1 MgCl₂, 4 KCl, 10 Hepes and 4 glucose, neutralized to pH 7.4 with NaOH. Ultrapure water (MilliQ) was used in the preparation of all solutions. Intracellular solution contained the following (mM): 135 KCl, 1.6 MgCl₂, 0.1 EGTA, 2.5 MgATP, 0.2 Li₂GTP and 10 Hepes, neutralized to pH 7.3 with NaOH. Given the slow onset and small size of currents measured, a large number of recordings were performed in order to select only cells with particularly stable holding currents. Holding potential was set at –80 mV, as this increased the stability of the holding current. Experimental protocol after obtaining a whole-cell recording involved first exposure to a control or VDM11-containing solution, followed by application of thapsigargin. At the end of each experiment, a pulse of 500 nM CPS was applied to the recorded neuron to test for functional expression of TRPV1 (see Results).

We thank Drs I Matias, S Petrosino, A Ligresti, T Bisogno and M Valenti for their help, Dr JB Davis (GlaxoSmithKline) for providing HEK293 cells overexpressing the human TRPV1, Dr Larry Marnett for providing human 12-lipoxygenase cDNA constructs and Dr Dale G Deutsch for providing rat FAAH cDNA constructs. This work was partly supported by the Volkswagen Stiftung (Germany, to VDM), by Fondazione Cassa di Risparmio di Modena, Fondazione Cassa di Risparmio di Carpi Italian MIVR (COEIN-PRIN 2004-057339) (Italy, to VV) and by the Wellcome Trust (UK, to PM).