Report

A Modular PIP2 Binding Site as a Determinant of Capsaicin Receptor Sensitivity

See allHide authors and affiliations

Science  23 May 2003:
Vol. 300, Issue 5623, pp. 1284-1288
DOI: 10.1126/science.1083646

Abstract

The capsaicin receptor (TRPV1), a heat-activated ion channel of the pain pathway, is sensitized by phosphatidylinositol-4,5-bisphosphate (PIP2) hydrolysis after phospholipase C activation. We identify a site within the C-terminal domain of TRPV1 that is required for PIP2-mediated inhibition of channel gating. Mutations that weaken PIP2-TRPV1 interaction reduce thresholds for chemical or thermal stimuli, whereas TRPV1 channels in which this region is replaced with a lipid-binding domain from PIP2-activated potassium channels remain inhibited by PIP2. The PIP2-interaction domain therefore serves as a critical determinant of thermal threshold and dynamic sensitivity range, tuning TRPV1, and thus the sensory neuron, to appropriately detect heat under normal or pathophysiological conditions.

Charged membrane phospholipids are thought to regulate a variety of ion channels and transporters (1). For example, recent electrophysiological studies suggest a role for the membrane phospholipid PIP2 as a modulator of transient receptor potential (TRP) channels, many of which contribute to the detection of sensory stimuli. TRP channels in the Drosophila eye (2) and TRPV1 channels in the mammalian pain pathway (3, 4) are activated or potentiated when PIP2 is hydrolyzed, whereas the ubiquitously expressed mammalian TRPM7 channel is inhibited by PIP2 cleavage (5). Phospholipase C (PLC) catalyzes the hydrolysis of PIP2 to inositol trisphosphate (IP3) and diacylglycerol (DAG) and has been implicated in the release of TRPV1 from PIP2-mediated inhibition (3), although the underlying mechanism for such regulation has not been elucidated. Despite functional evidence that TRP channels are directly regulated by PIP2, there is now no structural basis to account for this effect. TRPV1 is an especially tractable model for addressing this question, because it can be directly gated by a number of stimuli, including the pungent vanilloid compound capsaicin, extracellular protons (pH < 6.0), or noxious heat (>43°C) (6, 7). Moreover, genetic studies have shown that TRPV1 is an essential component of the signaling pathway through which PLC-coupled receptors increase behavioral sensitivity to heat (3, 8, 9), which makes elucidation of this regulatory pathway of physiological interest.

To investigate the molecular basis of PIP2-dependent regulation, we determined which regions of TRPV1 were required for PLC-mediated potentiation. We reasoned that mutations affecting such domains should render TRPV1 insensitive to PIP2 inhibition, and therefore, we identified mutants exhibiting increased responses to capsaicin or extracellular protons. A TRPV1 mutant lacking a segment of the C-terminal cytoplasmic domain (TRPV1 Δ777–820) produced much larger currents than wild-type channels in response to low doses of capsaicin (250 nM) or protons (pH 5.5), when expressed in Xenopus oocytes (Fig. 1). These enhanced currents could not be attributed to increased cell surface expression, because biotinylation experiments revealed equivalent levels of the mutant and wild-type TRPV1 at the plasma membrane (fig. S1). To determine whether amino acids 777 to 820 were also required for PLC-mediated potentiation, we exposed oocytes expressing TRPV1 channels and the nerve growth factor (NGF) receptor TrkA/p75 to NGF, a treatment that normally elicits a robust potentiation wild-type TRPV (∼20-fold) of responses through activation of PLC-γ (3). Currents from TRPV1Δ777–820 were unchanged, which suggests that these residues are critical for mediating potentiation downstream of PLC-coupled receptor stimulation. Moreover, potentiation of TRPV1Δ777–820 was not observed when other PLC-coupled receptors were activated, including the epidermal growth factor receptor (EGFR) and the G protein–coupled m1 muscarinic acetylcholine receptor (10). Loss of potentiation was not simply due to increased agonist sensitivity of the mutant channel, because the response to either a low dose of capsaicin (100 nM) or protons (pH 6.1) was unchanged after PLC activation (10). In addition, the thermal threshold of TRPV1Δ777–820 was markedly shifted to lower temperatures, and the overall currents were larger than those of wild-type channels (Fig. 1D), a phenotype reminiscent of wild-type channels that have been potentiated by PLC-coupled receptor activation (3).

Fig. 1.

A C-terminal region of TRPV1 is required for PLC-mediated potentiation. (A) Continuous current trace from oocyte expressing TRPV1 and the NGF receptor TrkA/p75 (24). Oocyte was exposed to pH 5.5 buffer for 30 s (⚫), followed by 750 pM NGF for 10 min (gray bar.) After NGF treatment, oocyte was exposed to pH 5.5 again, followed by a saturating (10 μM) dose of capsaicin (◯). (B) Oocytes expressing indicated channel and TrkA/p75 were challenged with pH 5.5buffer. Responses were normalized to current evoked with saturating dose of capsaicin (10 μM) according to the protocol indicated in Fig. 1(A). (C) Oocytes expressing indicated channels and TrkA/p75 were assayed for PLC-mediated potentiation. Results are measured as the fold increase in response to 30-s challenge with pH 5.5 buffer after 10-min treatment with 750 pM NGF. (D) Heat-evoked currents in oocytes expressing TRPV1 (blue traces) or TRPV1Δ777–820 (red traces) were measured in response to temperature ramps and normalized to a saturating dose of capsaicin (10 μM).

If potentiation of TRPV1 involves release from PIP2-mediated inhibition, then regions required for potentiation are potential sites of PIP2 interaction. PIP2-binding domains of ion channels are loosely characterized by clusters of basic residues interspersed with hydrophobic amino acids, an arrangement that is believed to facilitate interactions with the negatively charged head group of the phospholipid (1). In agreement with this, region 777–820 of TRPV1 has a predicted isoelectric point (pI) of 11 and contains eight positively charged residues. To determine which amino acids are most important for mediating potentiation, we generated deletion and point mutations within this region. Removal of as few as 16 residues (777 to 792) completely inhibited potentiation (Fig. 2, A and B). Furthermore, neutralization of as few as two positive charges reduced potentiation by 94 ± 1.1% (Fig. 2B), which suggests that clusters of basic residues are important for mediating TRPV1 potentiation. These mutants not only exhibited increased sensitivity to chemical stimuli, but were also sensitized to heat, as indicated by a shift of activation thresholds to substantially lower temperatures and production of larger currents at suprathreshold temperatures (Fig. 2, C and D). Moreover, as observed for capsaicin- and proton-evoked responses, larger deletions within the 777–820 domain had a more profound phenotype (as indicated by a greater shift in thermal threshold) than small deletions or point mutations, which suggests that residues that affect activation threshold and potentiation are distributed throughout this region.

Fig. 2.

Delineation of a potentiation domain of TRPV1. (A) Representation of amino acid region 777–820 of rat TRPV1. Red boxes indicate basic residues; gray boxes indicate neutralized residues [Lys or Arg replaced by Gln (K→ Q or R→ Q)]; lines indicate deletions. Relative degree of potentiation obtained with each mutant is indicated at right: (+++) maximal potentiation, (–) no potentiation. (B) Responses to a 30-s pulse of pH 5.5 buffer were measured before (gray bars) and after (black bars) exposure to 750 pM NGF for 10 min. Responses from five or six oocytes per construct are normalized to maximal currents obtained with a saturating dose of capsaicin (10 μM). Potentiation is measured as the fold increase in response to a 30-s challenge with pH 5.5 buffer after 10-min treatment with 750 pM NGF: 39.0 ± 9.5 (TRPV1); 0.94 ± 0.9 (TRPV1Δ777–820); 1.24 ± 0.1 (TRPV1Δ777–792); 2.45 ± 0.3 (TRPV1Δ784–792); 3.17 ± 0.5(TRPV1Δ784–789); 2.51 ± 0.4 (TRPV1 R785Q, K788Q); and 2.61 ± 0.5(TRPV1 K788Q, R797Q) (25). (C) Heat-evoked currents were measured in response to temperature ramps and normalized to a saturating dose of capsaicin (10 μM). Traces are representative of responses from four or five oocytes per construct. Light blue (TRPV1); dark blue (TRPV1 R785Q, K788Q); purple (TRPV1Δ784-789); green (TRPV1 K788Q, R797Q); yellow (TRPV1 Δ784-792); orange (TRPV1Δ777–792); and red (TRPV1Δ777–820). Average responses at subthreshold and suprathreshold temperatures are plotted in (D). In all experiments, oocytes also expressed TrkA/p75.

The most well characterized PIP2-binding domains in channels are those of the inward rectifier potassium (IRK) channels. All members of this family are thought to be positively regulated by PIP2 interaction (1), although binding affinities differ among family members (11, 12). Even though PIP2 activates IRKs and inhibits TRPV1, are regulated in opposite directions by PIP2, we asked whether potentiation could be restored to TRPV1Δ777–820 by inserting a region from an IRK family member. Chimeras of TRPV1 were therefore generated in which amino acids 777 to 820 were replaced with residues 207 to 244 from the cytoplasmic C-terminal tail of IRK1 (Fig. 3A), a region shown to be important for PIP2-dependent activity of this channel (11). The resultant chimera (TRPV1.IRK) was responsive to capsaicin, protons, and heat but showed relatively small currents compared with those of wild-type TRPV1 (29.4 ± 3.5% for pH 5.5) (10), even though these channels were expressed at comparable levels, as determined by cell surface biotinylation (fig. S1). This apparent decrease in activity was accompanied by a marked rightward shift in half-maximal effective concentration for capsaicin (EC50= 3.0 ± 0.2 μM for chimera and 0.42 ± 0.03 μM for wild-type TRPV1, Fig. 3B), which suggests that the PIP2-binding domain from IRK1 enhances interaction of the channel with phospholipids, and thus renders it more avidly inhibited than wild-type TRPV1.

Fig. 3.

Functional conservation of a PIP2-binding domain from distinct ion channel families. (A) Comparison of TRPV1 C-terminal regions with putative PIP2-binding domains from IRK1 and GIRK4 (11, 12). Two basic residues required for PIP2 interaction in IRK and GIRK are indicated with black arrowheads (⚫). (B) Agonist-response relationships for TRPV1 and mutants. Responses at each dose of capsaicin were normalized as percentage of maximal activity for each mutant. red, TRPV1Δ777–820; blue, TRPV1; yellow, TRPV1.IRK R218Q; green, TRPV1.IRK. Data were fit to the Hill equation, yielding the following hill coefficients: 1.68 ± 0.18 (TRPV1, n = 8); 2.0 ± 0.31 (TRPVΔ777–820, n = 11); 1.39 ± 0.19 (TRPV1.IRK R218Q, n = 8); 1.54 ± 0.10 (TRPV1.IRK, n = 5). (C) Inhibition of phosphatidylinositol 4-kinase enhances TRPV1-mediated currents. Representative voltage-clamp trace (Ehold = –45mV) from a TRPV1-expressing oocyte challenged with pH 5.5 buffer (30 s; black dots). Protons elicited a small current that was greatly enhanced after 10-min exposure to PAO (30 μM, gray bars). Further treatment with NGF (750 pM, white bar) and PAO (30 μM, 10 min) produced only minimal further enhancement. (D) Replacement of the C-terminal basic domain of TRPV1 with the PIP2-binding site of IRK1 yields channels that can still be potentiated. Oocytes expressing constructs indicated at bottom (along with EGFR, where indicated) were exposed to pH 5.5 buffer for 30 s before and after 5-min treatment with EGF (50 ng/ml) and PAO (15 μM). Potentiation represents the fold increase in current in five to eight oocytes per construct after treatment with EGF and PAO. (E) PIP2 binding affinity correlates with ability of chimeric channels to be potentiated. Oocytes expressing wild-type or chimeric channels, together with EGFR, were exposed to pH 5.5 buffer for 30 s before and after 5-min treatment with EGF (50 ng/ml) and PAO (15 μM). Potentiation represents the fold increase in current in five oocytes per construct after treatment with EGF and PAO. Asterisk indicates P < 0.02 (Student's t test). No statistical difference was seen between TRPV1 and TRPV1.IRK R218Q.

PLC-mediated modulation of IRK1 is nearly undetectable in Xenopus oocytes (13), presumably reflecting the relatively high affinity of this channel for PIP2, and our initial attempts to potentiate the TRPV1.IRK chimera were similarly unsuccessful. We therefore asked whether the presumed high affinity of this chimera for PIP2 could be overcome, by either lowering phospholipid availability or weakening PIP2 affinity. Phenylarsine oxide (PAO) is a cell-permeable, reversible inhibitor of phosphatidylinositol 4-kinase that limits PIP2 resynthesis (14). Treatment of TRPV1-expressing oocytes with PAO potentiated proton-evoked currents, and subsequent activation of TrkA receptors had minimal additional effect on channel activity (Fig. 3C). To further decrease PIP2 levels, we combined PAO treatment with activation of EGFR. Unlike TrkA, this PLC-coupled receptor can be expressed at high levels in oocytes without toxicity, which allows for reliably robust ligand-dependent cleavage of PIP2 (10). Cells expressing wild-type or TRPV1.IRK channels showed significant potentiation (fold increase of 11.3 ± 2.5 and 4.0 ± 0.8, respectively) when treated with EGF and PAO for 5 min, whereas TRPV1Δ777–820 showed no potentiation (fold increase of 0.75 ± 0.02) (Fig. 3D). These results demonstrate that the PIP2-binding domain from IRK1 can substitute for residues 777 to 820 of TRPV1 and support potentiation, so long as PIP2 levels are sufficiently diminished.

To structurally weaken PIP2-channel interactions, we generated a TRPV1.IRK1 chimera harboring a mutation [Arg218 replaced by Gln (R218Q)] that in the context of IRK1 produces a channel with faster kinetics of inhibition, which suggests a decreased affinity for PIP2 (11). This chimera (TRPV1.IRK R218Q) showed significantly greater potentiation than TRPV1.IRK and was indistinguishable from wild-type TRPV1 in this regard (Fig. 3E). Furthermore, the capsaicin dose-response curve for TRPV1.IRK R218Q was closer to that obtained with wild-type TRPV1 channels (Fig. 3B), consistent with a return to a milder form of PIP2-mediated inhibition than found in TRPV1.IRK. Furthermore, consistent with their relative EC50 values for capsaicin, TRPV1.IRK-mediated currents were activated only at very high temperatures (48° ± 0.4°C), whereas the chimera with weakened PIP2 binding affinity (TRPV1.IRK R218Q) had a thermal threshold (44° ± 0.3°C) closer to that of wild-type TRPV1 channels (44°C) (Fig. 4, A and B).

Fig. 4.

PIP2-binding affinity correlates with temperature sensitivity. (A) Representative temperature-response profiles from oocytes expressing TRPV1 or TRPV1.IRK chimeras. Each trace represents a single oocyte expressing TRPV1 (blue), TRPV1.IRK (green), or TRPV1.IRK R218Q (yellow). Currents obtained from water-injected oocytes are indicated in black. (B) The average temperature response threshold (24) for oocytes expressing TRPV1.IRK (47.9° ± 0.4°C) was significantly higher than for those expressing wild-type TRPV1 (44.5° ± 0.6°C, asterisk indicates P < 0.01, Student's t test, for five oocytes per construct). No significant difference was seen between TRPV1 and TRPV1.IRK R218Q (44.3° ± 0.3°C). (C) Mutation of a putative PKC phosphorylation site has no effect on EGFR-mediated potentiation. Oocytes expressing wild-type or S800A mutant channels, together with EGFR, were exposed to capsaicin (250 nM for 30 s) before and after treatment with EGF (100 ng/ml for 5min). Potentiation represents the fold increase in current in eight oocytes per construct after treatment with EGF. No significant difference was seen between TRPV1 and TRPV1S800A (Student's t test). (D) Oocytes expressing wild-type or S800A mutant channels, together with TrkA/p75, were exposed to capsaicin (250 nM for 30 s) before and after treatment with NGF (750 pM for 10 min). Potentiation represents the fold increase in current after treatment with NGF (double asterisk indicates P < 0.02, Student's t test, for nine oocytes per construct).

Several reports have recently suggested that potentiation of TRPV1 is achieved through protein kinase C (PKC)–dependent channel phosphorylation (15, 16). Moreover, two putative phosphorylation sites on TRPV1 have been identified, and mutations at these residues abolish phorbol ester–mediated potentiation of capsaicin-evoked currents (17). One of these sites, S800, is contained within the region that we implicate in PIP2-mediated inhibition, and we therefore asked whether phosphorylation at this site is required for receptor tyrosine kinase–mediated potentiation. Accordingly, we mutated S800 and found that this mutant (TRPV1 S800A) displayed normal capsaicin-evoked currents and normal EGFR-mediated potentiation (fold increase of 6.6 ± 1.2 for wild-type TRPV1 and 5.1 ± 0.5 for S800A, Fig. 4C). TrkA activation also produced robust potentiation of both wild-type and S800A channels, although a statistically significant difference was observed in the absolute magnitude of the effect (fold increase of 17.1 ± 6.1 for wild-type and 8.6 ± 0.5 for S800A, Fig. 4D). Thus, phosphorylation at S800 is not absolutely required for EGFR- or TrkA-mediated potentiation. Additionally, phosphorylation by PKC probably does not contribute to potentiation of the TRPV1.IRK chimeras, because the PIP2-binding domain from IRK1 contains no PKC consensus sites, yet it supports significant receptor-mediated potentiation in the context of TRPV1. Indeed, robust potentiation can also be achieved simply by inhibition of PIP2 synthesis, independent of either PLC or PKC activation (Fig. 3C). However, the reduced level of TrkA-mediated sensitization observed with the S800A mutant raises the possibility that phosphorylation can modulate potentiation efficacy, perhaps by increasing negative charge density at this site and weakening electrostatic interactions between PIP2 and the channel. Alternatively, we have shown that TrkA (but not EGFR) forms a complex with TRPV1 and PLC-γ (3), and thus the reduced sensitization of the S800A mutant by NGF could also reflect a decreased physical interaction and efficiency of local PIP2 hydrolysis.

Our findings provide a structural basis to support the hypothesis that PLC-mediated regulation of TRP channels involves the removal of PIP2 from a binding site on the channel protein. Although PIP2-binding domains in channels exhibit little linear sequence homology, our studies suggest that they function as modular regulatory domains that mediate inhibition or activation by phospholipids depending on their channel context. Several TRP channels can be directly activated in vitro by DAG or other polyunsaturated lipids, such as anandamide in the case of TRPV1 (2, 18, 19). These lipid second messengers could activate or modulate TRP channels by displacing PIP2 from an inhibitory binding site on the channel complex. However, capsaicin and anandamide sensitivity map to a domain that is distinct from the C-terminal region identified here as being important for PLC modulation (20). Furthermore, TRPV1 mutants that lose the ability to be potentiated retain capsaicin sensitivity, which suggests that PIP2 and unsaturated lipid agonists interact with TRP channels at separable sites, at least in the case of TRPV1.

The capsaicin receptor TRPV1 is normally activated with a threshold that corresponds to the onset of tissue damage and pain (∼43°C) (6). Our studies suggest that this physiologically crucial set point is determined, at least in part, by the strength of TRPV1-PIP2 interactions. If this interaction were too strong, then the activation threshold would be too high to serve as an effective warning mechanism, as is illustrated by the VR1.IRK chimera. The more tempered affinity of TRPV1 for PIP2 not only establishes a physiologically relevant activation threshold, but also enables the channel to be dynamically modulated by inflammatory products that activate PLC. Finally, it is interesting to note that the C-terminal domain of TRPV3, a warm-sensitive channel with an activation threshold of ∼35°C (2123), conspicuously lacks a region corresponding to the 777–792 domain of TRPV1, the minimal essential core of the predicted PIP2 binding site. Thus, modification of this PIP2 regulatory domain by genetic, biochemical, or pharmacological mechanisms may have profound effects on sensitivity of primary afferent nerve fibers to chemical and thermal stimuli under normal or pathological conditions.

Supporting Online Material

www.sciencemag.org/cgi/content/full/300/5623/1284/DC1

Materials and Methods

Fig. S1

References

References and Notes

View Abstract

Navigate This Article