Sodium channels in cell membranes have a crucial role in triggering bioelectrical events that lead to processes such as muscle contraction or hormone release. A crystal structure reveals how one such channel might work. See Article p.353
In 1786, Luigi Galvani famously observed the twitch of dissected frog legs in response to an electrical spark generated during a thunderstorm. What he didn't realize was that when a frog simply feels like jumping, the idea itself begins with a 'spark' — a bioelectrical event. We now know that these events are action potentials caused by a brief influx of positively charged sodium ions into excitable cells, such as neurons and muscle cells1. We also know that the influx of ions is gated by membrane proteins called sodium channels. But, despite more than 50 years of speculation and intense experimentation, the structure of these proteins was unknown. Now, on page 353 of this issue, Payandeh et al.2 report the crystal structure of the sodium channel NavAb from the bacterium Arcobacter butzleri, allowing us to peer inside the protein and see how these ion channels might work.
Sodium channels are members of a large class of voltage-gated ion channels (VGICs) that also includes potassium and calcium channels. They have a special status among VGICs, because almost all action potentials in vertebrates are initiated and caused by the transitory opening of sodium channels in response to a change of potential across the plasma membrane of an excitable cell. Sodium channels distinguish themselves from other VGICs not only in their selectivity for sodium ions, but also in the blazing speed with which they open in response to a stimulus. Another characteristic is that open sodium channels inactivate rapidly if the activating stimulus is maintained. The activation and inactivation gates of sodium channels, either of which may interrupt ion flux, are likely to be located in different parts of the protein3,4.
Like many other VGICs, sodium channels are exquisitely sensitive to small changes of membrane potential — a depolarization as small as 10 millivolts can increase the probability of their being open by a hundredfold5. Because the concentration of sodium ions outside cells is typically ten times that inside, channel opening causes a passive influx of positively charged sodium ions; the resulting depolarization induces more sodium channels to open. This positive-feedback cycle underpins the avalanche-like nature of action potentials, and explains why they can propagate for long distances along the surface of an excitable cell. Action potentials lead inexorably to one of two outcomes, depending on the cell type: the release of a molecule such as a neurotransmitter or hormone, or the contraction of a muscle cell.
VGICs are made up either of four identical subunits, as NavAb (Fig. 1)2 and most potassium channels are, or from a long protein with four structurally similar (but not identical) subunit-like domains, as observed for sodium and calcium channels in animals. All VGICs contain a central pathway for ions and water that is surrounded by a tube known as the pore domain. This domain insulates the ions from the surrounding lipid membrane, thereby reducing the energy barrier to ions traversing the membrane. At the periphery of the pore domain are four voltage-sensing domains, one from each subunit (or subunit-like domain). Each voltage-sensing domain is composed of four transmembrane segments (S1–S4) connected together by three loops. The pore domain comprises the S5 and S6 segments of each subunit (or subunit-like domain), with each S5–S6 pair connected by an intervening loop (see Fig. 1a of the paper2).
The voltage-sensing domains detect changes in membrane potential largely through several positively charged arginine amino-acid residues in the S4 segment. These arginine side chains pull the associated S4 segment back and forth through the electric field across the membrane, in a process akin to a miniaturized form of electrophoresis. The S4 movement is coupled to the activation gate at the intracellular end of the permeation pathway.
In previously reported crystal structures of voltage-gated potassium channels6,7, the activation gate was open. However, in Payandeh and colleagues' structure2 the activation gate of NavAb is shut tight. Interestingly, the authors observed that the S4 segments are in an outward, activated conformation that should favour an open channel. Channel opening in VGICs is typically preceded by outward movement of the S4 segments, and so the authors suggest that their structure reveals NavAb in a 'pre-open' state along the activation pathway of the protein. Another possibility is that the structure represents an inactivated state if the activation gate also serves as the inactivation gate, as seen in hyperpolarization-activated ion channels8.
The structure of NavAb has several fascinating features, three of which I consider here. The first concerns the S4 segment. As in other VGICs, each of its arginine residues is separated from the next by two hydrophobic residues. More notably, in NavAb, part of the S4 segment is wound into a particularly tight helix known as a 310-helix — something that has also been seen in two potassium channels7,9. Consequently, the four voltage-sensing arginine residues near the extracellular ends of the S4 segments of NavAb form a linear array of charged side chains along the S4 axis. This suggests that the S4 segment moves like a piston along its axis, rather than twisting like a helical screw, as was speculated in earlier models of voltage-dependent S4 movement10,11. However, the possibility that S4 undergoes dynamic changes to its secondary structure as it moves cannot be ruled out.
The second notable feature is NavAb's selectivity filter — the structure that allows the channel to select sodium ions, rather than other ions, for passage through the pore. The filter contains a ring of four glutamic acid residues, one from each subunit, which makes it strongly negatively charged. Such a high field-strength anionic site was predicted12 long ago for molecules that selectively bind sodium over potassium ions. Nevertheless, Payandeh and colleagues' structure doesn't close the book on sodium selectivity, because sodium channels in animals have two acidic, one basic and one neutral residue, instead of the four acidic residues of NavAb.
Perhaps the most intriguing features of the crystal structure2, however, are the fenestrations. The NavAb channel has four portals around it, midway through the transmembrane region (see Fig. 4 of the paper2). Such fenestrations are not found in potassium channels. Payandeh et al. observed that the acyl chains of lipid molecules extend through these windows into the middle of NavAb's ion-permeation pathway. As in potassium channels7,13, a phenylalanine residue near the middle of the inner helix of each subunit lining the pore may play a crucial part in ion permeation and gating. In NavAb, these residues sit in the fenestrations, where they can monitor and potentially participate in the movement of molecules between the surrounding membrane bilayer and the central cavity of the pore.
Sodium channels are known to be unusually sensitive to small pore-blocking molecules such as local anaesthetics and related compounds4,14, and these moderately hydrophobic molecules can somehow enter and exit the closed channels — but how they do this is a mystery. The fenestrations may be the explanation. If so, then the portals are ripe for further investigation, to find out how these clinically important drugs interact with the channel. The authors also suggest the tantalizing possibility that the fenestrations open and close depending on the gating state of the channel. Could the fenestrations open when the activation gate closes? Stay tuned!
Hille, B. Ion Channels of Excitable Membranes 3rd edn (Sinauer, 2001).
Payandeh, J., Scheuer, T., Zheng, N. & Catterall, W. A. Nature 475, 353–358 (2011).
Bezanilla, F. & Stefani, E. Annu. Rev. Biophys. Biomol. Struct. 23, 819–846 (1994).
Catterall, W. A. Novartis Found. Symp. 241, 206–218 (2002).
Hirschberg, B., Rovner, A., Lieberman, M. & Patlak, J. J. Gen. Physiol. 106, 1053–1068 (1995).
Long, S. B., Campbell, E. B. & MacKinnon, R. Science 309, 903–908 (2005).
Long, S. B. et al. Nature 450, 376–382 (2007).
Shin, K. S., Maertens, C., Proenza, C., Rothberg, B. S. & Yellen, G. Neuron 41, 737–744 (2004).
Clayton, G. M. et al. Proc. Natl Acad. Sci. USA 105, 1511–1515 (2008).
Catterall, W. A. Annu. Rev. Biochem. 55, 953–985 (1986).
Guy, H. R. & Seetharamulu, P. Proc. Natl Acad. Sci. USA 83, 508–512 (1986).
Eisenman, G. Biophys. J. 2, 259–323 (1962).
Cuello, L. G. et al. Nature 466, 272–275 (2010).
Fozzard, H. A., Lee, P. J. & Lipkind, G. M. Curr. Pharm. Des. 11, 2671–2686 (2005).
About this article
Journal of Materials Chemistry B (2017)
Hydration structure of Na+and K+fromab initiomolecular dynamics based on modern density functional theory
Molecular Physics (2014)
The Journal of Chemical Physics (2013)
Sodium Ion Binding Sites and Hydration in the Lumen of a Bacterial Ion Channel from Molecular Dynamics Simulations
The Journal of Physical Chemistry Letters (2011)