Ion channels are integral membrane proteins that generate the electrical activity of the nervous system by opening and closing a pore at their centres to control the flow of ions into and out of cells1. For instance, opening of the potassium-ion (K+) channel Slo1 leads to a flow of K+ out of cells that decreases neural excitability. Slo1 is of special interest to neuroscientists because it can conduct more ions per second than any other K+ channel2. Moreover, it can be synergistically activated by both an increase in intracellular levels of calcium ions (Ca2+) and an increase in voltage across the cell membrane — as such, it is vital for controlling processes that involve coupling between Ca2+ and voltage3, including muscle contraction and neuronal excitation. Online in Nature, two papers4 now address the mechanisms underlying Slo1's high K+ conductance and dual activation, using cryo-electron microscopy to determine the first full-length, atomic-resolution structures of the channel isolated from the sea slug Aplysia californica.
In the first paper, Tao et al.4 (page 46) describe the open structure of Slo1 in the presence of Ca2+ and magnesium ions (Mg2+), which can also activate the channel. In the second paper, Hite et al.5 (page 52) resolve Slo1 in the absence of these ions, and compare the two structures. Slo1 comprises four identical subunits, and the structures show that voltage-sensor domains from each subunit surround a central pore–gate domain (PGD). Both the voltage sensors and the PGD are embedded in the cell membrane, and are docked to a large cytoplasmic 'gating' ring consisting of four Ca2+ sensors (Fig. 1). The voltage sensors regulate opening and closing (gating) of the PGD in response to membrane voltage, controlling passage of K+ through the channel. The Ca2+ sensors are thought to pull open the PGD in response to Ca2+ binding3,6,7.
The structures reveal a surprising number of potential transduction pathways that couple the various sensors and PGD. Short linkers connect the voltage sensors and PGD; long linkers connect the Ca2+ sensors and PGD; there is direct contact between the voltage sensors and PGD; and there is a direct interface between the Ca2+ and voltage sensors at the top of the gating ring. Hite et al. posit that interfaces between the voltage and Ca2+ sensors allow the two to gate the channel through shared pathways. The convergence of linkers from the voltage and Ca2+ sensors onto the PGD also points to pathway coupling.
The existence of shared transduction pathways could provide a mechanistic basis for the previous observation3 of cooperative activity and coupling between the sensors. For instance, the gating ring and a linker between it and the PGD have been shown6 to modulate voltage-based Slo1 activation by acting as a passive spring in the absence of Ca2+. It had not been determined whether it was the linker alone, the gating ring alone, or the linker–ring complex that formed the spring. The current structures show that the linker is stiff, so it is the gating ring that has spring properties.
Previous studies introduced mutations into the gene encoding Slo1, and analysed the channel's gating using electrophysiological techniques to define the roles of its various protein domains3. This provided evidence for two Ca2+-binding sites and one Mg2+-binding site per subunit8, but crystal structures of isolated gating rings identified no bound Mg2+ and only one Ca2+ per subunit7. Moreover, it has been shown9 that two Ca2+ bound to each subunit act cooperatively to increase channel activity over that observed when two Ca2+ bind different subunits. The current studies detect all three binding sites for each subunit, in the locations the electrophysiological work predicted8. This, in addition to Hite and colleagues' comparison of Ca2+-bound and Ca2+-free structures, provides an explanation for cooperativity — Ca2+ ions at both binding sites act jointly to expand the gating ring to its open structure.
The current papers also provide insight into why Slo1 has such high conductance. Previous studies2,7 suggested that there are three major contributing factors: a wide, central ion-conducting pore that leads from a central cavity to the selectivity filter (which selects for K+); side portals and a wide ion-conducting pathway through the centre of the gating ring that supply K+ to the central cavity; and negative charges to attract K+ to the entrance of the conducting pore. The full-length structures expand on these data, showing that the selectivity filter is the only narrow part of the conducting pathway, that there is a large build-up of negative charge at the surfaces lining the conducting pathway, and, most importantly, that the selectivity filter is almost identical to that of a related but lower-conductance K+-conducting pore in a channel dubbed Kvchim. Thus, the large ion-conduction pathways and central-cavity negativity are primarily responsible for the high conductance of Slo1.
Slo1 channels pass through multiple states over time during gating. These kinetic states are typically attributed to the time that the channel spends in different structural conformations10. Hite et al. observed four different structures for Slo1 at zero millivolts and without Ca2+ or Mg2+ — conditions in which the channels would be expected to be closed, but transitioning between a subset of conformational states. This raises the intriguing possibility that Hite et al. have identified some of the structures associated with different kinetic states, providing insight into which conformational states might be adopted by the closed channel. Future work should determine the conformations of the protein at intermediate Ca2+ levels, when different numbers of Ca2+ sensors would be activated, and at different voltages.
The full-length structures will provide invaluable information about how best to mutate the Slo1 gene in future electrophysiological studies, both to determine the contributions of various sensors to gating through different transduction pathways, and to gain insight into the mechanisms by which various naturally occurring mutations in Slo1 channels lead to defective channel gating and diseases such as epilepsy and dyskinesia11. Indeed, Hite and colleagues include a movie (Supplementary Video 1) that illustrates how Slo1 might morph between closed and open states during Ca2+ activation, and the transduction pathways involved. The authors consider the possibility that a major component of the voltage activation may be transduced from voltage sensors to an interface with the gating ring, and then via linkers to the PGD. This idea seems to contradict previous proposals that voltage and Ca2+ sensors are only weakly coupled3,12. If the hypothesis is confirmed, it may be necessary to rethink current models for dual activation of Slo1.