Structural biology

Ion channel seen by electron microscopy

Structures of the heat-sensitive TRPV1 ion channel have been solved using single-particle electron cryo-microscopy, representing a landmark in the use of this technique for structural biology. See Articles p.107 & p.113

Membrane proteins known as transient receptor potential (TRP) ion channels occur in species ranging from yeast to humans. Members of this receptor family are involved in the perception of an enormous range of stimuli1, including vision (in invertebrates), taste, hot or cold temperatures, pH and physical forces. On page 107 of this issue, Liao et al.2 report the first high-resolution structure of TRPV1, the ion channel responsible for sensing heat. And in a second paper from the same group, Cao et al.3 (page 113) describe the sites at which three ligand molecules bind to TRPV1, and how this binding triggers the opening of the channel.

There are 27 members of the TRP receptor family in humans, each with their own functions and different tissue distribution. Most TRP channels, including TRPV1, are weakly selective for calcium ions. TRPV1 was first identified4 as the receptor for capsaicin — the compound that makes chilli peppers seem hot — in 1997. The channel has four identical subunits, and the modified version used in the present studies has an overall molecular weight of about 300 kilodaltons (bigger than most ion channels). Not only is TRPV1 opened by capsaicin, it is also strongly activated by toxins, such as resiniferatoxin from Euphorbia plants, or 'cysteine-knot' toxins from tarantulas. These chemosensory stimuli are thought to have evolved as protective deterrents against predators, and elicit a burning sensation by usurping normal heat sensing through TRPV1 activation.

To solve the structure of TRPV1, Liao et al. used single-particle electron cryo-microscopy (cryo-EM), with no help from any of the more established methods of structural biology. The authors made full use of several technical advances: they used a slightly truncated rat TRPV1 construct that is biochemically stable; they transferred purified ion channels into a polymeric 'amphipol' framework5 to maintain the channels' stability and solubility in water; and, most importantly, they used a camera that detects electrons directly (minimizing noise and allowing any image blurring during an exposure to be compensated for6,7) and a state-of-the-art computer program that uses statistical methods to estimate particle orientations and to calculate highly accurate reconstructions of structures8.

Liao and colleagues' analysis advanced from an initial EM structure obtained at a resolution of 30 ångströms, through a cryo-EM structure at 8.8 Å resolution, to a final map that includes regions depicted at a resolution of 3.4 Å — good enough for amino-acid side chains and β-sheets to be recognized, and for the polypeptide backbone of the protein to be traced. Previous work9,10 in which cryo-EM was used to study TRP channels reached resolutions of only 15–19 Å. The current work is therefore a landmark both in the evolution of the single-particle cryo-EM method and in its use for tackling the structure of a macromolecular complex that has been difficult to study by other means. Notably, the regions of the map with highest resolution are near the centre of the molecule. More work is needed to discover whether the lower resolution in peripheral regions is caused by the limited accuracy with which orientations of single protein particles can be determined or by flexibility of the TRPV1 channel in these regions.

The authors confirm that the overall architecture of TRP channels is similar to that of members of another group of ion channels, the voltage-gated sodium and potassium channels11. The members of this superfamily all contain a central ion-conducting pore that is made up of two carboxy-terminal transmembrane α-helices (S5 and S6) and a loop containing a short helix (the pore helix) from each subunit (or from each domain, in the case of sodium channels), together with a voltage-sensor module consisting of a bundle of four transmembrane helices (S1 to S4) at the amino terminus (Fig. 1). However, S4 in TRP channels contains only one or two of the characteristic basic amino-acid residues that are responsible for voltage gating (opening of the channel by a potential difference across the membrane). In this respect, TRP channels are similar to a nucleotide-gated channel called MlotiK1 (ref. 12) because it has a compact, four-helix bundle that contains few basic residues, is less sensitive to transmembrane potentials and, as Cao et al. show, does not move during activation. Instead, high temperature or ligand binding triggers channel opening by a mechanism that is less dependent of voltage. But how?

Figure 1: TRPV1 in closed and open states.

The cartoon shows cross sections of the closed and open states of the TRPV1 ion channel, based on structures2,3 obtained using electron cryo-microscopy. The channels form from four subunits, only two of which are shown. Helices in one subunit (S5, S6, the S4–S5 linker, the pore helix and the TRP helix) are labelled. A bundle of helices S1 to S4 is depicted as a single object. a, Two gates are apparent in the closed state, one near the extracellular surface (gate 1) and another deeper within the channel (gate 2). b, Gate 1 opens in response to the binding of spider toxin, whereas gate 2 opens on binding of capsaicin. Another ligand, resiniferatoxin (not shown), can bind at the same site as capsaicin. The arrow indicates the passage of calcium ions through the channel.

Cao et al. find that the binding site for a large spider toxin lies distant from that for the small ligands resiniferatoxin and capsaicin. The spider toxin binds to the extracellular surface of the TRPV1 channel near the pore helix, with each of the toxin's two cysteine-knot domains binding to the junction between TRPV1 subunits. This locks open the extracellular end of the channel (gate 1 in Fig. 1).

The authors propose a binding site for capsaicin that is essentially the same as that for resiniferatoxin, although their identification of this site is less certain. The site consists of a cavity deep within the membrane towards the cytoplasmic side, surrounded by S3, S4, a linker between S4 and S5, and S6 from the adjacent subunit. The bound ligands can induce a structural change to the TRPV1 channel through close interactions with the S4–S5 linker, increasing the pore diameter by shifting S6 (gate 2 in Fig. 1). Thus, the TRPV1 channel seems to have two gates, one at either end of the channel. This dual gating, and the possible complexity of interactions between the two gates in response to different stimuli, is a principal finding of the authors' analysis.

An amino-acid motif adjacent to the C terminus of S6 contains the TRP box that is characteristic of TRP channels. Liao and colleagues' structure shows that the box consists of a short α-helix parallel to the membrane; this helix interacts with both the S4–S5 linker and another helix (the pre-S1 helix), and becomes less well ordered in the activated state.

The current work does not explain how TRPV1 is activated in response to temperature, but this probably depends on finely balanced energy differences between its open and closed structures, which are difficult to address through structural analysis. Nevertheless, the availability of the new structures will surely help simulations of heat activation to be performed. The authors also hint that the methods used in their studies are well suited to trapping different states during heat-evoked gating.

Further improvements in detector efficiency, specimen-preparation methods and image-processing software should bring other advances in the use of cryo-EM for structural biology. With the outstanding success of the current work, the way is open for structural studies of many similar channels. Because TRPV1 and some other ion channels are potential targets for the development of painkilling drugs, the findings may even herald the dawn of cryo-EM as a technique to aid rational drug design.


  1. 1

    Venkatachalam, K. & Montell, C. Annu. Rev. Biochem. 76, 387–417 (2007).

    CAS  Article  Google Scholar 

  2. 2

    Liao, M., Cao, E., Julius, D. & Cheng, Y. Nature 504, 107–112 (2013).

    ADS  CAS  Article  Google Scholar 

  3. 3

    Cao, E., Liao, M., Cheng, Y. & Julius, D. Nature 504, 113–118 (2013).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Caterina, M. J. et al. Nature 389, 816–824 (1997).

    ADS  CAS  Article  Google Scholar 

  5. 5

    Popot, J.-L. et al. Annu. Rev. Biophys. 40, 379–408 (2011).

    CAS  Article  Google Scholar 

  6. 6

    Li, X. et al. Nature Meth. 10, 584–590 (2013).

    CAS  Article  Google Scholar 

  7. 7

    Bai, X. C., Fernandez, I. S., McMullan, G. & Scheres, S. H. W. eLife 2, e00461 (2013).

    Article  Google Scholar 

  8. 8

    Scheres, S. H. W. J. Struct. Biol. 180, 519–530 (2012).

    CAS  Article  Google Scholar 

  9. 9

    Mio, K. et al. J. Mol. Biol. 367, 373–383 (2007).

    CAS  Article  Google Scholar 

  10. 10

    Moiseenkova-Bell, V. Y., Stanciu, L. A., Serysheva, I. I., Tobe, B. J. & Wensel, T. G. Proc. Natl Acad. Sci. USA 105, 7451–7455 (2008).

    ADS  CAS  Article  Google Scholar 

  11. 11

    Long, S. B., Tao, X., Campbell, E. B. & MacKinnon, R. Nature 450, 376–382 (2007).

    ADS  CAS  Article  Google Scholar 

  12. 12

    Clayton, G. M. et al. Proc. Natl Acad. Sci. USA 105, 1511–1515 (2008).

    ADS  CAS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Richard Henderson.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Henderson, R. Ion channel seen by electron microscopy. Nature 504, 93–94 (2013).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.