TRPchannel--cell sensors
DAVID E. CLAPHAM
Howard Hughes Medical Institute, Pediatric Cardiology, Children's Hospital of Boston, Department of Neurobiology, Harvard Medical School, Enders 1309, 320 Longwood Avenue, Children's Hospital, Boston, Massachusetts 02115, USA
Correspondence and requests for materials should be addressed to the author (dclapham@enders.tch.harvard.edu).
TRP channels are the vanguard of our sensory systems, responding to temperature, touch, pain, osmolarity, pheromones, taste and other stimuli. But their role is much broader than classical sensory transduction. They are an ancient sensory apparatus for the cell, not just the multicellular organism, and they have been adapted to respond to all manner of stimuli, from both within and outside the cell.
The human genome encodes hundreds of channels that broker the passage of charged ions across impermeable lipid bilayers1. While energy-requiring pumps labour to build charge and concentration gradients across the membrane, ion channels spend this stored energy, much as a switch releases the electrical energy of a battery. Small conformational changes cause channels to open, allowing over ten million ions to flow per second through each channel. Ca2+ ions are particularly important in cellular homeostasis and activity, and the surface of each cell holds thousands of channels that precisely control the timing and entry of Ca2+ ions.
Transient receptor potential (TRP) channels were first described in Drosophila, where photoreceptors carrying trp gene mutations exhibited a transient voltage response to continuous light2, 3. Unlike most ion channels, TRP channels are identified by their homology rather than by ligand function or selectivity, because their functions are disparate and often unknown. They have been called store-operated channels (SOCs), but this description is theoretical and related to a poorly understood phenomenon.
The known functions are diverse. Yeast use a TRP channel to perceive and respond to hypertonicity4, 5. Nematodes use TRP channels at the tips of neuronal dendrites in their 'noses' to detect and avoid noxious chemicals6. Male mice use a pheromone-sensing TRP channel to tell males from females7. Humans use TRP channels to appreciate sweet, bitter and umami (amino acid) tastes8, and to discriminate warmth, heat and cold. In each of these cases, TRPs mediate sensory transduction, not only in a classical sense, for the entire multicellular organism, but also at the level of single cells. Almost all mammalian TRP channel genes are now known. Here I summarize the common characteristics of the diverse mechanisms of TRP channel activation, highlighting major questions that remain to be answered. More details can be found in other reviews9-15.
What are TRP channels?
Mammalian TRP channels comprise six related protein families with sequence identity as low as 20% (Fig. 1). All TRP channels are putative six-transmembrane (6TM) polypeptide subunits that assemble as tetramers to form cation-permeable pores (Box 1). In general, they are almost ubiquitously expressed and most have splice variants. So most cells have a number of TRP channel proteins.
Figure 1 Mammalian TRP family tree. Full legend
High resolution image and legend (77k)
It will take time to elucidate all of the diverse functions of TRP proteins. To date, the most informative approaches have been ligand-specific expression cloning, and targeted (global) gene inactivation in mice. In the near future, further information is likely to be gleaned from tissue-specific and developmental-stage-specific gene targeting.
The TRPC (canonical TRP) subfamily
All mammalian TRPC proteins appear to be analogous to the TRP involved in Drosophila phototransduction, in that they function as receptor-operated channels. They are activated by stimulation of G-protein-coupled receptors (GPCRs) and receptor tyrosine kinases (Box 2).
TRPC1, the first mammalian TRP reported16, forms heteromeric channels with TRPC4 and/or TRPC5 (ref. 10). The properties of the heteromultimers are distinct from those of TRPC4 and TRPC5 homomultimers (see Table 1 and Fig. 2). TRPC5, but not TRPC1, is present in hippocampal growth cones and modulates neurite extension17. Mice lacking TRPC4 have defects in agonist-induced vasoregulation and lung microvascular permeability18, 19.
Figure 2 Representative transmembrane currents flowing in response to a set voltage (I–V relation) of various TRP channels. Full legend
High resolution image and legend (61k)
TRPC3, TRPC6 and TRPC7 proteins share 75% identity, have relatively low selectivity for Ca2+ over Na+, and are sensitive to the intracellular concentration of Ca2+ ([Ca2+]i). Diacylglycerol (DAG) analogues20 potentiate their activity, but not through protein kinase C activation. TRPC3 has been investigated extensively as a putative inositol-1,4,5-trisphosphate (InsP3) receptor (IP3R)-binding SOC, with conflicting results21. All of the TPRC3, TRPC6 and TRPC7 subfamily are highly expressed in smooth and cardiac muscle cells, making them candidates for the receptor-activated nonselective cation channels known to exist in these sites. In support of this idea, TRPC6 is an essential part of the 1-adrenoreceptor-activated cation channel in rabbit portal vein myocytes22. They may also have roles in the regulation of vascular tone, airway resistance and cardiac function.
TRPC2 appears to be a pseudogene in humans, but its rat orthologue encodes an important sensor localized to neuronal microvilli in the vomeronasal organ23. Trpc2-deficient mice display abnormal mating behaviour, consistent with a role for this channel in pheromone signalling7.
The TRPV (vanilloid receptor, osm9-like) subfamily
TRPV1 was identified by expression cloning using the 'hot' pepper-derived vanilloid compound capsaicin as a ligand. TRPV1 is a Ca2+-permeant channel24, 25 that is potentiated by heat (>43 °C) and decreased pH, and inhibited by intracellular phosphatidylinositol-4,5-bisphosphate (PIP2)24, 26. Its thermal sensitivity is enhanced by bradykinin and nerve growth factor, which appear to act via phospholipase C (PLC) to hydrolyse PIP2, releasing inhibition of the channel26. Trpv1-/- mice are defective in nociceptive, inflammatory and hypothermic responses to vanilloid compounds, supporting the interpretation that TRPV1 contributes to acute thermal nociception and hyperalgesia after tissue injury27. TRPV1 also participates in mechanically evoked purinergic signalling by the bladder urothelium28.
TRPV2, which is 50% identical to TRPV1 (ref. 29), may mediate high-threshold (>52 °C) noxious heat sensation, perhaps through lightly myelinated A nociceptors. Interestingly, TRPV2 translocates from intracellular pools upon insulin growth factor stimulation of transfected cells30. Stretch reportedly increases TRPV2 translocation, and cardiac-specific transgene expression of TRPV2 results in Ca2+-overload-induced cardiomyopathy31. But it is not surprising that overexpression of a Ca2+-permeant channel induces cardiomyopathy, because such TRP channels are deleterious to many cells, including neurons. In fact, the mechanism of pain-relieving topical capsaicin is due, in part, to neuronal cell death.
Increased temperature also activates TRPV3 (>31 °C)32-34 and TRPV4 (>25 °C)35. The neuronal distribution of TRPV3 overlaps with TRPV1, raising the interesting possibility that they may heteromultimerize32. TRPV3 is also highly expressed in skin, tongue and the nervous system, possibly explaining the activity of 'warm-sensitive' neurons. The effect of temperature on rates of biological processes is expressed as the 10 °C temperature coefficient1: Q10 = rate(T + 10 °C)/rate(T). Most ion channels and enzymes have gating Q10 values of 3–5, but the Q10 of TRPV3 gating is >20 (ref. 33), and for TRPV1 and TRPV4 it is estimated to be 10–20 (ref. 36). TRPV4 current is potentiated by hypotonicity (cell swelling)37, 38. Trpv4-/- mice have a marginally impaired renal response to hypertonicity, probably due to abnormal central control of antidiuretic hormone secretion39. Hypotonicity increases TRPV4-mediated current in primary afferent nociceptive nerve fibres, an effect that is enhanced by the hyperalgesic inflammatory mediator prostaglandin E2 (ref. 40). Expressed TRPV4 may be gated by epoxyeicosatrienoic acids41.
TRPV5 and TRPV6 comprise a distinct subfamily of homomeric and heteromeric channels found in transporting epithelia of the kidney and intestine. They show strong inwardly rectifying currents and are the most Ca2+-selective TRP channels (permeability ratio PCa/PNa > 100)42, 43, suggesting that they mediate Ca2+ uptake. Both are inactivated by [Ca2+]i (ref. 44); TRPV6 shows voltage-dependent intracellular Mg2+ blockade45.
The TRPM (melastatin) subfamily and TRPA
TRPM1 (melastatin) was initially identified as a transcript that showed decreased expression in highly metastatic versus non-metastic melanoma cells46. It is widely expressed in normal tissues, but its function and electrophysiological properties have not been described.
TRPM2 (ref. 47) forms a Ca2+-permeant channel48 that is gated by binding of ADP ribose (EC50 100 µM) and nicotinamide adenine dinucleotide (NAD; 1 mM)48, 49 to a carboxy-terminal NUDT9 Nudix hydrolase domain. ADP ribose is a breakdown product of NAD, CD38, cyclic ADP ribose (a Ca2+-release messenger) and protein de-acetylation (O-acetylated ADP ribose50), but the TRPM2 domain itself is an ineffective hydrolase51. The channel is regulated by signalling pathways responsive to H2O2 and tumour-necrosis factor-, suggesting that it may act as a sensor of intracellular oxidation/reduction52, possibly during the oxidative burst of neutrophils53.
TRPM3 (refs 54, 55) forms a Ca2+-permeant nonselective channel that is constitutively active when heterologously expressed. Its activity is increased by hypotonicity (200 mOsm per litre), but there is little homology to TRPV4 that might suggest a common mechanism of activation54. TRPM4 is expressed primarily in kidney distal-collecting-duct epithelium and in the central nervous system.
TRPM4 and TRPM5 are the only monovalent-selective ion channels of the TRP family. They are widely distributed and may account for observed Ca2+-activated 20–30 pS nonselective channel activities56, 57. They are activated through GPCRs coupled to PLC-dependent endoplasmic reticular Ca2+ release, perhaps by direct Ca2+ binding to the channel. However, relatively high [Ca2+]i is required to activate these channels56, 57, suggesting that they localize close to sites of Ca2+ release or that other modulators are important. Although their instantaneous I–V relationships are linear, the fraction of open channels increases at positive potentials. This voltage dependence is not mediated by divalent cation binding, suggesting an intrinsic voltage-sensing mechanism57, 58.
TRPM5 is found in cells expressing taste receptors59. In an in vivo study in TrpM5-/- mice, it was shown that taste receptors T1R and T2R share a common signalling pathway involving PLC2 and TRPM5, to produce sweet, umami and bitter taste sensations8. The authors concluded that InsP3, Ca2+ and thapsigargin-mediated store depletion did not activate TRPM5. However, it is possible that PIP2 or other molecules modulate its sensitivity to [Ca2+]i.
TRPM6 and TRPM7 are unique among ion channels because they also contain functional kinase domains60. TRPM7 passes little inward current under physiological conditions, is permeant to both Ca2+ and Mg2+, and is inhibited by 0.6 mM intracellular free Mg2+(refs 61, 62). In contrast to other GPCR-activated TRP channels, TRPM7 current increases slowly under whole-cell recording conditions and is inactivated by PIP2 hydrolysis by PLC or PLC62. The function of the kinase domain is poorly understood and its substrates have not been identified. In contrast to our original report, it is not required for channel activation62, 63. The catalytic core of the kinase domain is similar to that of other eukaryotic protein kinases and to enzymes with ATP-grasp domains64. The sensitivity of TRPM7 to physiological Mg–ATP levels has been suggested to have a central role in metabolic sensing61 or to serve as a mechanism to adjust cellular Mg2+ homeostasis63. But a spontaneous human mutation in TRPM6 results in familial hypomagnesaemia with secondary hypocalcaemia, suggesting that TRPM6 may be important for Mg2+ uptake in the kidney and intestine.
TRPM8 was identified as a messenger RNA that was upregulated in prostatic and other cancers. Its sensory role was recognized when it was isolated by expression cloning of a menthol receptor from trigeminal neurons65, 66. TRPM8 is a nonselective, outwardly rectifying channel that can be activated by cold (8–28 °C) and enhanced by 'cooling' compounds such as menthol and icilin. TRPM8 is widely expressed, but thought to function specifically as a thermosensor in TrkA+, small-diameter primary sensory neurons.
ANKTM1, a Ca2+-permeant, nonselective channel homologous to Drosophila painless67, is distinguished by 14 amino-terminal ankyrin repeats. It is activated by noxious cold temperature (<15 °C) but bears little similarity to menthol-sensitive TRPM8 (ref. 68). It is found in a subset of nociceptive sensory dorsal root ganglion neurons, in the company of capsaicin-sensitive TRPV1, but not TRPM8. Interestingly, the Drosophila orthologue of ANKTM1 responds to warming (>27 °C) rather than to cooling when expressed in Xenopus oocytes69. These observations are consistent with sensitivity to the surrounding membrane environment, but might also be reconciled if the lowest energy state of the mammalian channel is the open configuration. Other TRP channels have not been systematically tested for temperature sensitivity, but such a comparison would clarify this issue.
The TRPP (polycystin) and TRPML (mucolipin) subfamilies
Polycystic kidney disease proteins PKD2, PKD2L1 and PKD2L2 are 6TM Ca2+-permeant channels called TRPP2, TRPP3 and TRPP5, respectively. The much larger TRPP1, polycystin-REJ and polycystin-1L1 proteins are 11TM proteins that contain a C-terminal 6TM TRP-like channel domain. TRPP1 is not known to form a channel by itself, but it complexes with TRPP2 to form a Ca2+-permeable nonselective cation channel70. Autosomal dominant polycystic kidney disease is caused by mutations in TRPP1 or TRPP2, leading to alterations in the polarization and function of cyst-lining epithelial cells. Trpp1-/- and Trpp2-/- mice die in utero with cardiac septal defects and cystic changes in nephrons and pancreatic ducts71, 72. The mouse orthologue of TRPP3 is deleted in krd mice, resulting in defects in the kidney and retina73.
TRPP proteins have another role in development. Normal body asymmetry appears to arise from leftward extracellular flow generated by motor-protein-dependent rotation of monocilia on the ventral surface of the embryonic node. Motile monocilia generate nodal flow, and non-motile TRPP2-containing cilia sense nodal flow, initiating an asymmetric Ca2+ signal at the left nodal border74. TRPP1 and TRPP2 both appear to be targeted to primary cilia cells of renal epithelia, where the channel complex is gated by fluid flow75.
The mucolipins (MCOLN1, MCOLN2 and MCOLN3) are 6TM channels that are probably restricted to intracellular vesicles. Mutations in MCOLN1 (TRPML1) are associated with mucolipidosis type IV, a neurodegenerative lysosomal storage disorder76. The defect appears to be in sorting or transport in the late endocytic pathway. Mutations in a Caenorhabditis elegans TRPML1 homologue, cup-5, cause excess lysosome formation and apoptosis in all cell types77. TRPML3 is present in the cytoplasm of hair cells and the plasma membrane of stereocilia. TRPML3 is mutated in the varitint-waddler mouse, resulting in deafness and pigmentation defects78.
Theories of TRP channel gating
The gnawing mystery of TRP channels is their elusive mechanism of gating. Many TRP channels show some constitutive activity in overexpression systems, but only a few have been studied in their native environment or at physiological temperatures. Our poor understanding of TRP channel gating may reflect our ignorance of potential intra- or extracellular ligands. But a more interesting possibility is that there might be a common TRP gating mechanism. Attempts to understand TRP channel activation have given rise to several theories, discussed below.
Receptor-operated theory This is the most likely mechanism for TRPC channels and the Drosophila photoreceptor TRP (Box 2). All mammalian TRPC channels can be activated by GPCRs. These include muscarinic type 1 (TRPC1, TRPC4, TRPC5 heteromers, or presumed TRPC4, TRPC5 homomers), histaminergic type 1 (TRPC3, TRPC6) and purinergic receptors (TRPC7). GPCRs are often attached to multimolecular complexes by scaffolding proteins, adding an extra level of complexity. To date, investigators have identified only the proteins that interact prior to PLC activation. Unfortunately, PLC activation generates DAG and free cytoplasmic Ca2+ (via InsP3), both of which can unleash legions of active signal transduction molecules. The major challenge of the receptor-operated theory is to find a receptor-activated messenger that directly binds and specifically activates a TRP channel. To establish this theory for a subset of TRP channels, it will be necessary to identify the native receptors that activate a particular channel, the downstream messengers mediating activation, and the physiological roles of these proteins and messengers in the context of the tissues that normally express them. One potential messenger, PIP2, is described in more detail.
PIP2 comprises 1% of anionic phospholipids in cells and is much more abundant than its signalling relative PIP3. PIP2 regulates at least seven types of ion channel and transporter79. Because several TRPs (TRPC3, TRPC4, TRPM7) are known to bind PLC and/or PLC, PLC in part determines the PIP2 concentration that TRPs encounter. PIP2 inhibits Drosophila TRP and TRPL80 and mammalian TRPV1 (ref. 26), but increasing PIP2 lowers the temperature activation threshold of TRPV1 (ref. 81) and probably other TRP channels. Constitutive TRPM7 activity is increased by PIP2 and inactivated by PIP2 hydrolysis at 22 °C (ref. 62). These results suggest that PIP2 commonly interacts with positively charged regions of ion channels to alter the energy required for gating. Is PIP2 a common denominator of TRP activation?
The head group of PIP2, carrying about four negative charges, may 'snorkel' into the aqueous phase above other phospholipids (inset to Box 2 figure). Positively charged peptides containing basic residues can hug the bilayer if they contain aromatic amino acids (particularly Phe and Trp). If the peptide also contains clusters of basic residues (4), it can electrostatically fence in PIP2, forming basins of concentrated PIP2 (ref. 82). Indeed, such a cluster in the mid-C terminus is required for inhibition of TRPV1 by PIP2 (ref. 81), but is missing in the PIP2-insensitive TRPV3 (S. Ramsey, unpublished work). Interestingly, all TRPs contain a Trp/Phe segment with basic lysine and arginine residues (TRP box) just distal to the S6 gating helix/cytoplasmic junction, but there may not be enough positive charge in this region to sequester PIP2. Considering the data to date, PIP2 is likely to modulate gating of some TRP channels, but it is not a unifying mechanism of TRP channel activation.
Store-operated calcium entry hypothesis Putney proposed that emptied Ca2+ stores (primarily in the endoplasmic reticulum) somehow gate the entry of external Ca2+ to replenish the deficit (for a review, see ref. 83). This mechanism is called capacitative Ca2+ entry, or store-operated Ca2+ entry (SOCE). The physiological hallmark of SOCE is a large, receptor-mediated, transient [Ca2+]i increase followed by a prolonged high [Ca2+]i plateau that is dependent on [Ca2+]o (ref. 84). Thapsigargin, an inhibitor of smooth endoplasmic reticulum Ca2+-ATPase (SERCA) pumps, is often used to examine this phenomenon, with the assumption that the drug is specific for SERCAs. The SOCE hypothesis has intrigued many scientists and has led to the search for an ion channel and a mechanism that could link store depletion with Ca2+ entry. ICRAC (calcium-release-activated current) is the best candidate for a store-depletion-responsive current, but it does not appear to be mediated by any TRP protein. None of the TRP channels exhibits the requisite high Ca2+ selectivity (PCa/PNa 1,000), low single-channel conductance (<0.1 pS), and pharmacological enhancement by 1–5 µM 2-APB85. The three main theories for SOCE are a direct coupling mechanism, a diffusible messenger, and store-depletion-mediated fusion of a vesicle containing a Ca2+-permeant channel (Fig. 3).
Figure 3 Three theories of store-operated Ca2+ entry. Full legend
High resolution image and legend (89k)
Although dozens of papers tout the link between SOCE and TRP channels, it is worth re-examining the data. The most common standard of proof has been to show, either by Ca2+ imaging or by electrophysiology, that a particular heterologously expressed TRP channel enhances store-depletion-mediated entry. However, a better standard for this activity would be total abrogation of the phenomenon by elimination of a protein or gene, either in vivo or in a faithful model of the native system.
Nilius and colleagues18 studied Trpc4-/- mice and concluded that TRPC4 was an essential component of SOCE channel activation. However, electrophysiological characterization suggested that the defect resulted from a decrease in ICRAC, and TRPC4 alone has few properties characteristic of CRAC channels. Our laboratory suggested that the Ca2+-selective TRPV6 might be a component of a CRAC channel42, and a dominant-negative TRPV6 subunit in Jurkat cells was interpreted as suppressing endogenous ICRAC86. But TRPV6 alone does not account for several properties of ICRAC42, 87. In a careful Ca2+ imaging and electrophysiological study using double-stranded RNA to target endogenous TRPC1-containing channels in Chinese hamster ovary cells, it was concluded that TRPC1 is an essential component of SOCE88. Despite the use of knockout or knockdown methods, none of these proteins is universally accepted as a store-operated Ca2+ channel. In my opinion, the major molecules comprising CRAC remain unknown.
Almost every publication concluding that a TRP is a SOC is countered by a paper stating that the same TRP is not a SOC. Most of the data based on Ca2+ imaging are pro-SOCE, whereas most of the electrophysiological data are con-SOCE. Ca2+ imaging is prone to false-positives and electrophysiology is prone to false-negatives. Ca2+ imaging has an advantage in that cells can be kept in a fairly normal environment. But Ca2+ imaging is a very indirect assay of channel function because it reflects accumulated free [Ca2+]i, regardless of the source. Addition of thapsigargin without accounting for constitutive Ca2+ entry leads to falsely high [Ca2+]i. SERCA pump inhibition can also artificially increase the Ca2+ signal by eliminating the endoplasmic reticulum as a Ca2+ buffer and reducing the effective Ca2+ volume of the cell. Ca2+ entry into the unbuffered cytoplasm may induce Ca2+-activated channels, Ca2+ transporters, or other unrelated Ca2+-dependent processes that affect the results (this problem can be reduced by substitution of extracellular Ca2+ with Ba2+). Most importantly, voltage levels, the driving force for Ca2+ entry into the cytoplasm, are uncontrolled.
Patch clamp is a direct assay of channel activity that overcomes most of these limitations. In patch clamp, the channel can be bathed in defined solutions both inside and outside the cell, and voltage is controlled. Excellent time and current resolution allows the dissection of independent processes. Test compounds and proteins can readily be applied to either side of the channel. However, in standard patch-clamp recordings, intracellular contents are perfused, rapidly replacing small molecules and diffusible proteins. This problem can be circumvented by perforated patch recording, in which molecules larger than ions are restricted. Unfortunately, although it is not difficult to perform, it has rarely been used for these studies. Finally, electrophysiologists routinely record at 22 °C, and the few TRP channels that have been tested are highly sensitive to temperature. These problems could easily be rectified, and if they are, electrophysiology will most probably become the standard for SOC identification.
In my opinion, Ca2+-permeant TRPs, like all Ca2+-permeant channels, contribute to [Ca2+]i and thus affect the SOCE process. But they have not been proved to be SOCs in any direct sense. The SOCE hypothesis for TRPs will not be settled until there is a consensus on the molecular identity of SOCs and the endoplasmic-reticulum-dependent signal that activates them. In the interim, the use of the term 'store-operated channel' for TRPs is confusing and should be avoided.
Vesicle fusion hypothesis A Ca2+ entry channel was proposed to fuse with the plasma membrane to mediate SOCE, but neither the channel mediating this event nor the mechanism was identified89, 90. Independent of the SOCE theory, do TRPs rapidly translocate to the plasma membrane in response to stimuli? The mucolipins (TRPML) are involved in intravesicular trafficking, but little more is known of their function or whether they can also be present in the plasma membrane of native cells. Interestingly, TRPC1 and TRPC3–TRPC6 are all present in rat brain synaptosomes91. Recently, C. elegans sperm TRP (TRP-3), required for sperm–egg interaction, was localized in intracellular vesicles until fertilization competence required TRP-3 translocation from the vesicles to the plasma membrane92.
These observations raise the question of whether some TRP channels are held in reserve and are then rapidly placed in the cell membrane in response or in adaptation to a stimulus. If so, what mechanism links the stimulus to the translocation of TRP-containing vesicles? And are the TRP channels in vesicles required for vesicle trafficking, swelling and/or the fusion process itself?
Cell sensory hypothesis Block93 beautifully encapsulated the underlying physical principles of sensory transduction. At the heart of sensation is the ability to distinguish input from a photon (vision, heat, electromagnetic force), phonon (sound), chemical (for example, odorant), or mechanical force (stretch, osmolarity, gravity) from background thermal energy or extraneous inputs (noise). Evolution has enhanced these abilities not just by sensitive detection, but also by amplification and signal processing. Built into almost all TRPs is the ability to conduct Ca2+ ions in a parsimonious manner; their I–V relations, nonselective character and low density dictate that they admit relatively little Ca2+ per second compared with, for example, a voltage-gated Ca2+ channel (CaV). Second, they are active at resting membrane potentials where cells spend most of their time. These characteristics can be harmful to a cell if TRPs are overexpressed, as is commonly observed in expression systems or in some cancer cells. As for most Ca2+-permeant channels, the majority of TRPs are inherently self-inactivating by virtue of their [Ca2+]i sensitivity. These properties provide a first level of signal processing common to sensory mechanisms, but what about detection?
Sensation requires the detection of force. The primary mechanisms for TRP activation involve mechanical force, intracellular ligand binding (signal transduction molecules) and temperature. How do these forces translate into channel gating? The Boltzmann equation gives the equilibrium distribution of channels in the open and closed states; the fraction of open channels is [1 + exp(G/NkBT)]-1, where G is the free energy of transition between the closed and open state, kB is Boltzmann's constant, N is Avogadro's number and T is the absolute temperature. The free energy, G, is equal to w–, where w is the conformational energy increase upon gating the channel and represents the sum of energies transduced to the channel from external forces. Voltage-sensitive channels are usually held in a higher-energy, closed state at resting membrane potentials; removal of this energy (depolarization to 0 mV) allows the channel protein to relax into its lower-energy structural configuration. To sense this electromotive energy, voltage-sensitive ion channels have charged amino acid 'solenoids' that drive conformational changes. Most TRP channels lack these attributes and at the resting membrane potential are contentedly immune to the transmembrane field. What then is the for TRPs?
The detection limit for an open TRP channel is imposed by thermal energy (0.6 kcal mol-1 at 37 °C). This energy is tiny, less than the energy of a single photon of visible light per molecule. But other cellular noise sources must be overcome, probably requiring channel gating for the cell to record a significant stimulus. The gating energy for the membrane-tension-sensitive bacterial channel in a liposome is 10 kcal mol-1 (ref. 94), only 20 times the thermal limit. Mechanosensation—the basis of hearing, osmolar sensing, stretch and flow sensing—easily provides these levels of energy, especially if the mechanical advantage provided by cilia is taken into account95. TRPC2 (ref. 23), OSM-9 (ref. 6), TRPP2 (PKD)75 and Nan96 are all found in ciliated structures. TRPs are also common within ciliated cells that sense flow in kidney epithelia, vascular endothelia, lung and intestine, and in the hearing and vestibular apparatus, taste cells and odorant-sensing cells. Anchoring TRPs to mechanical forces is not without its risks to the cell. Prolonged Ca2+ entry due to abnormal membrane tension, such as occurs with loss of membrane-stabilizing proteins in muscular dystrophy, may result in muscle degeneration. Prolonged stretch in cardiac muscle leads to Ca2+ overload, hypertrophy and cardiac failure.
For signal-transduction-gated TRPs, the energies imparted by intracellular ligand binding are also sufficient for gating, depending on the binding affinity of the specific interaction (dissociation constant values of 1 µM imply 1–10 kcal mol-1). But the most obscure mechanism for activation of TRPs is temperature. Oddly, all TRP recordings published to date fail to reach steady state and do not saturate within the range of practical recordings. This prevents estimates of gating energy, but temperature-sensitive TRP channels have estimated Q10 values up to ten times larger than is typical for enzymes or channels. Temperature-dependent TRPV3 current is reversible upon cooling, but exhibits a pronounced hysteresis; the current increases with increasing temperature but abruptly collapses when cooling begins33. Perhaps heat induces lipid bilayer rearrangements to alter membrane tension. Alternatively, the protein may denature (melt) and rapidly refold when energy is removed. Finally, TRPs could be gated by temperature-dependent cooperative binding by second messengers.
What's next?
The large number of TRP subtypes, their overlapping electrophysiological characteristics, broad expression patterns, heteromultimerization, lack of specific blockers, and poorly understood mechanisms of activation have made their study difficult for TRP channel researchers. But the field is now progressing beyond simple overexpression and Ca2+ imaging, and more specific tools are being used. In the near future, a wealth of data from genetically altered animals will emerge, as well as useful molecular tools and assays. The next few years should determine whether any of the TRPs are indeed SOCs, or whether this term has simply substituted for a poorly understood activation mechanism. TRPs are sensory molecules, but not just in the usual organismal definition of sensation. TRP channels are intrinsic sensors of the cellular environment.
