Ion channels controlled by sound underlie the sense of hearing. Having long eluded researchers, the first such mammalian channel has now been identified in the mouse inner ear.
When you hear a sound or shake your head, the hair cells of your inner ear convert movement into the electrical currency that the brain understands. But how such transduction from mechanical to electrical signals works on a molecular scale remains unclear. The problem is that the ear's hair cells are relatively few in number, and the molecules involved are sparse, so progress has been slow. On page 723 of this issue, however, Corey and colleagues1 suggest that a key component of the vertebrate ‘mechano-transducer’ is an ion channel called TRPA1.
Sound waves or head movements deflect rod-like projections called stereocilia on hair cells. These stereocilia stand out a few micrometres from the cell surface, and when they are rocked, their relative sliding motion pulls on a series of extracellular linking proteins that run between their tips — the ‘tip links’ (Fig. 1a). The pulling makes ion channels in the stereocilia open, causing the hair cells to send signals to the brain. The identity and precise mechanism of action of these mechano-transducing channels remain obscure, however.
There have been hints from studies of non-mammalian species that the transducers might belong to the TRP superfamily of ion channels. Initially described in fruitflies (Drosophila), this diverse family of membrane proteins now boasts members throughout the animal kingdom that are implicated in intracellular calcium release, osmotic regulation, temperature sensitivity and pain mechanisms2,3.
When TRP-family members were first found in specialized mechanosensory cells of Drosophila and zebrafish, they caused considerable excitement. The mechanosensitive bristles of Drosophila, for example, involve a TRP channel protein called NOMPC (alternatively named TRPN1)4, and a second TRP channel, Nanchung, is found in the fly's hearing organ5. The structure of NOMPC suggested that it was a good candidate for a mechanosensory channel in other organisms, but it turned out to have no direct counterpart in most vertebrate genomes, and the search for the vertebrate transducer faltered.
Corey et al.1 now argue that at least one component of the vertebrate mechano-transducer is indeed another member of the TRP superfamily, TRPA1. Described only recently, TRPA1 (also known as ANKTM1)6 is a close relative of NOMPC. It is found in a subset of neurons, and is associated with sensing painful cold stimuli, being activated just below room temperature. It is also activated by the contents of kitchen cupboards (such as mustard oil and cinnamon) and bathroom cabinets (including icilin from cold creams), and, even more tantalizingly, by cannabinoids7.
Corey et al. make their case using several lines of argument. Indeed, there is a tendency in this field to use a variety of species in a single report, and the authors studied hair cells in the hearing and balance organs of mice, but derived corroborating details from the larger cells of frogs and from genetic manipulation of zebrafish. By exhaustively screening all 33 mouse TRP channels, Corey et al. found that TRPA1 is expressed in hair cells only when the cells become functional mechano-transducers. When the gene encoding TRPA1 is shut down, mechano-transduction is reduced, as detected either by electrophysiological measurements or by dye permeation through the channel. Moreover, turning off the zebrafish relative of the TRPA1 gene produces hair cells that do not work well.
It seems unlikely that the hair-cell transducer channel is made up of only TRPA1 protein subunits. TRP channel proteins are promiscuous and form assemblies with other proteins to produce a wide range of functional channel properties. In the Drosophila hearing organ, two members of the fly TRPV subfamily, Nanchung and Inactive, are required to form the channel pore8. One might expect something similar in vertebrate hair cells. There are indeed other TRP candidates in hair cells, including TRPV1 (ref. 9) and TRPML3 (ref. 10), and these may form complexes with TRPA1.
In fact, the mix ‘n’ match habits of TRP channels might even be necessary to explain some cochlear biophysics. The kinetics of the hair-cell transducer varies along the cochlea, as the sounds being sensed change from low to high frequency11,12, and it is possible that the gradation comes about as different TRPs form complexes. All the evidence suggests that stereocilia form extremely dynamic structures, with a high turnover of membrane and organellar constituents — just the sort of assembly ground for a transducer channel matched to hair-cell diversity.
The structure of the whole transducer complex still needs to be understood. How is this mechanosensitive channel controlled? An attractive hypothesis concerns a feature present in both NOMPC and TRPA1 — an iterating series of amino-acid motifs termed ankyrin repeats, found at one end (the amino-terminal end) of the proteins. There are 29 ankyrin repeats in NOMPC, and 17 in TRPA1. In NOMPC, the repeats form a whole turn of a helical spring at the cytoplasmic surface, which might couple displacement of the stereocilia to control of the channel13 (Fig. 1b). The structure of the vertebrate tip link suggests that most of the transducer compliance resides in cytoplasmic portions of the channel, not in the link1, so the control spring is probably on the inside of the cell. The fewer ankyrin repeats in TRPA1 may indicate that it too has a helical spring at the cytoplasmic surface, but one that has only a partial turn. Such a spring would exert a torque when distorted, and it is not hard to see that this could rapidly control the channel when pulled or pushed by a tip link.
The highest-resolution electron micrographs show tip links splitting into two or three strands at the stereocilial tip14. Does this imply that several subunits of the channel are directly connected to the link, or are there further intermediates of the complex yet to be found15? The conundrum of which molecules associate together at the stereocilial tip remains. The presence of TRPA1, with its distinct structure, however, reopens the Pandora's box of the transduction field — the possibility that when the channel is opened, extracellular calcium entering through it directly alters the mechanics of the channel by acting on the helix structure. This local feedback might, if (and it's a big ‘if’) the phasing is right, even contribute to amplifying the very sounds that stimulate the hair cells in the first place. In any case, it is clear that we've not heard the last of TRPA1 and its associates.
Corey, D. P. et al. Nature 432, 723–730 (2004).
Beech, D. J., Muraki, K. & Flemming, R. J. Physiol. (Lond.) 559, 685–706 (2004).
Moran, M. M., Xu, H. & Clapham, D. E. Curr. Opin. Neurobiol. 14, 362–369 (2004).
Walker, R. G., Willingham, A. T. & Zuker, C. S. Science 287, 2229–2234 (2000).
Kim, J. et al. Nature 424, 81–84 (2003).
Story, G. M. et al. Cell 112, 819–829 (2003).
Jordt, S. E. et al. Nature 427, 260–265 (2004).
Gong, Z. et al. J. Neurosci. 24, 9059–9066 (2004).
Zheng, J. et al. J. Neurophysiol. 90, 444–455 (2003).
Di Palma, F. et al. Proc. Natl Acad. Sci. USA 99, 14994–14999 (2002).
Ricci, A. J., Crawford, A. C. & Fettiplace, R. Neuron 40, 983–990 (2003).
He, D. Z., Jia, S. & Dallos, P. Nature 429, 766–770 (2004).
Howard, J. & Bechstedt, S. Curr. Biol. 14, R224–R226 (2004).
Kachar, B., Parakkal, M., Kurc, M., Zhao, Y. & Gillespie, P. G. Proc. Natl Acad. Sci. USA 97, 13336–13341 (2000).
Weil, D. et al. Hum. Mol. Genet. 12, 463–471 (2003).
About this article
Physiological Reviews (2020)
Phonetics and Speech Sciences (2012)
Journal of Evolutionary Biochemistry and Physiology (2009)