GIRK channels allow potassium ions to cross the cell membrane, thereby affecting the electrical status of the cell and so its functioning. Structural data now provide insight into the channels' mode of operation. See Article p.190
Ion channels are the main units responsible for the electrical activity in our body. They constitute a large family of some 400 proteins in humans. A subfamily of these proteins consists of four GIRK channels1, which specialize in converting chemical signals — mostly those of neurotransmitter molecules such as acetylcholine, dopamine, serotonin and adrenaline — into electrical ones in heart cells and neurons. They are therefore essential for controlling heart rate and the activity of neural circuits. In this issue, Whorton and MacKinnon2 (page 190) describe the long-awaited crystal structure of the mammalian GIRK2 channel in complex with two subunits of a G protein (a dimer of the Gβ and Gγ subunits), providing information about their mechanism of openingFootnote 1.
Activation of GIRK channels often begins with stimulation of G-protein-coupled receptors (GPCRs) in the cell membrane (see Fig. 1a of the paper2). For instance, binding of acetylcholine to a muscarinic-type GPCR on a heart cell results in release of the Gα and Gβγ subunits of the G protein that is attached to the GPCR at the intracellular surface of the cell membrane. Gβγ then activates GIRK channels3,4,5, which allow efflux of intracellular potassium ions (K+) from the cell, causing hyperpolarization of the cell membrane (it becomes more negative inside relative to outside) and so reducing the cell's electrical excitability. Acetylcholine thus slows the heart rate.
Following the breakthrough discovery3 that the Gβγ dimer is responsible for opening GIRK channels after GPCR activation, extensive biochemical and electrophysiological studies have focused on the mechanism of activation of these channels and the role of associated modulatory molecules1. These studies, however, fell short of deciphering the exact mode of interaction of the Gβγ subunits with the channel and the structural transitions that lead to channel opening.
Whorton and MacKinnon describe atomic-level interaction of the Gβγ dimer with a GIRK channel consisting of four GIRK2 monomers (Fig. 1a,b). The 3.5-ångström-resolution structure shows that each of the four monomers is bound to a Gβγ dimer, in agreement with previous biochemical evidence. They are also individually bound to a molecule of the phospholipid PIP2 and a sodium ion, both of which are necessary for channel functionality.
The Gβγ dimer and the channel share a relatively small surface area of contact (roughly 700 Å2), compared with the footprint of other known Gβγ interactor molecules such as Gα, the GPCR kinase-2, phospholipase-Cβ and phosducin. Nevertheless, the contact areas of each interactor with Gβγ overlap to various degrees, such that Gβγ cannot bind to more than one interactor simultaneously; this underlines the singularity of the Gβγ-mediated signalling event. The Gβγ dimer seems to interact with the channel at the interface of the channel monomers, including regions that are known to be involved in channel activation, such as the LM loop. The interaction involves both short-range intermolecular forces such as van der Waals forces and hydrogen bonding and long-range electrostatic forces, and is further stabilized by anchoring of Gβγ to the cell membrane through the lipid moiety in the Gγ subunit.
To reveal the structural changes associated with channel activation, Whorton and MacKinnon aligned three structures of GIRK2–PIP2: the normal channel, the channel in complex with Gβγ dimers and an always-active channel mutant. A comparison of the first two structures revealed two main conformational differences. On Gβγ-dimer binding, there was a clockwise (looking from inside the cell) rigid-body rotation of about 4° along the centre axis of the channel, relative to the transmembrane domains. There was also a widening of the bottom of the channel's inner helices — the inner helical gate — on the cytoplasmic side.
The conformational changes that open the inner helical gate are comparable to the widening of a lens aperture by hand-rotating the aperture ring. In the resulting conformation, however, the gate is too narrow to allow hydrated K+ to pass through the channel. So how do Gβγ dimers 'gate' the channel? Adding the always-active channel to the analysis provided an answer. In this structure, the rotation of the cytoplasmic domain relative to the transmembrane domain was more pronounced, causing the inner gate to widen further and permit the passage of hydrated K+.
On the basis of their observations, the authors formulate a Gβγ-dependent gating scheme for GIRK channels (Fig. 1c). Following activation of GPCRs and dissociation of the G protein into Gβγ and Gα, the free Gβγ dimer diffuses to the inner membrane surface and binds to the channel molecule to induce a 'pre-open' state. In this state, rotation of the channel's cytoplasmic domain relative to its transmembrane domain broadens the inner helical gate, although the channel cannot conduct K+. Nevertheless, the pre-open conformation brings the channel to a higher energetic state, allowing it to make frequent random changes to the open conformation, by which K+ conduction can occur. Such frequent conformational changes are consistent with the well-characterized 'bursting' behaviour of the channel that is seen during recordings of single-channel activity.
Whorton and MacKinnon's data relate to homomeric GIRK2 channels, which are present only in selected areas of the brain6. And at least one other GIRK channel, GIRK1, differs from GIRK2 in the length of its amino-acid sequence and in several amino-acid residues involved in Gβγ binding. So it is important to determine the structure of the two most prevalent channel species in the brain and heart — the GIRK1/GIRK2 and GIRK1/GIRK4 heteromers, respectively — in complex with the Gβγ dimer. Although the overall structural transitions associated with gating are likely to be the same, the contact surface of heteromeric GIRKs with the Gβγ dimer and the interaction forces involved could be different. Knowledge of such differences may clarify the different gating behaviours previously seen with channels of varying composition.
Although the intimate interaction of the GIRK channels with Gβγ dimers forms the basis of the channels' gating activity, direct channel interactions with Gα also fine-tunes the gating mechanism7. How Gα provides such control is unknown.
More broadly, it may be possible to design specific molecules that could interfere with channel function by targeting the unique interaction interface of its GIRK2 monomers with Gβγ. Such drugs would be desirable because they would not affect other Gβγ-dependent signalling events.
*This article and the paper under discussion2 were published online on 5 June 2013.
Hibino, H. et al. Physiol. Rev. 90, 291–366 (2010).
Whorton, M. R. & MacKinnon, R. Nature 498, 190–197 (2013).
Logothetis, D. E., Kurachi, Y., Galper, J., Neer, E. J. & Clapham, D. E. Nature 325, 321–326 (1987).
Wickman, K. D. et al. Nature 368, 255–257 (1994).
Reuveny, E. et al. Nature 370, 143–146 (1994).
Lüscher, C. & Slesinger, P. A. Nature Rev. Neurosci. 11, 301–315 (2010).
Rubinstein, M. et al. J. Physiol. (Lond.) 587, 3473–3491 (2009).
About this article
On the mechanism of GIRK2 channel gating by phosphatidylinositol bisphosphate, sodium, and the Gβγ dimer
Journal of Biological Chemistry (2019)
Neuroscience & Biobehavioral Reviews (2016)
The Journal of General Physiology (2013)