Ion channels catalyze the diffusion of inorganic ions down their electrochemical gradients across cell membranes. Because the ionic movements are passive, ion channels would seem to be extraordinarily simple physical systems, yet they are responsible for electrical signaling in living cells. Among their many functions, ion channels control the pace of the heart, regulate the secretion of hormones into the bloodstream, and generate the electrical impulses underlying information transfer in the nervous system. My research is aimed at understanding the physical and chemical principles underlying ion channel function.
Ion Conduction and Selectivity
Ion channels, like enzymes, have their specific substrates: potassium, sodium, calcium, and chloride channels permit only their namesake ions to diffuse through their pores. Ion selectivity refers to the ability of channels to discriminate among ions. Without ion selectivity, electrical signal production in biology as we know it would be impossible. My laboratory used mutational analysis to show that potassium channels are tetramers of identical subunits and that specific "signature sequence" amino acids are responsible for potassium selectivity. The signature sequence is conserved in all potassium channels throughout nature and forms a structural unit called the selectivity filter.
To understand how the selectivity filter conducts potassium ions rapidly and selectively, we determined the x-ray structures of numerous potassium channels, including the KcsA potassium channel at a resolution higher than 2.0 Å. The selectivity filter is a narrow, 12-Å-long segment of the pore that is lined with carbonyl oxygen atoms, as shown in Figure 1. These atoms act as surrogate water molecules, allowing potassium ions to shed their hydration shell and enter into the pore. We showed that two potassium ions bind in the selectivity filter at once, most likely with a single intervening water molecule. Entry of a third potassium ion on one side of the filter is associated with concerted potassium ion exit from the opposite side, giving rise to efficient potassium ion conduction. Electrostatic repulsion between ions inside the filter prevents the potassium ions from binding too tightly, thus allowing them to conduct at a high rate.
We have developed methods for chemical synthesis and folding of a functional potassium channel in collaboration with Tom Muir (Rockefeller University). Using these methods, we synthesized potassium channels with main-chain and side-chain chemical groups that do not occur in natural proteins. These synthetic potassium channels have taught us that ester carbonyl oxygen atoms are tolerated in the potassium-binding sites, but not without energetic consequence. The chemical synthesis of "unnatural" potassium channels has taught us that a specific conformation of the selectivity filter is important for preventing sodium conduction when potassium ions are not present in solution. When potassium ions are present, they prevent sodium conduction through competition for binding sites.
Chloride channels mediate the conduction of chloride ions across cell membranes. To understand how nature mediates anion selectivity, we determined the atomic structures of two prokaryotic members of the ClC chloride channel family at resolutions of 3.0 Å and 2.5 Å, respectively. These anion-selective transport proteins are built on a completely different architectural plan than that of potassium channels, but they nevertheless use many of the same basic physical principles to catalyze ionic movement across the membrane. These principles include oriented alpha-helical-end charges to stabilize ions inside the membrane and short selectivity filters containing multiple ions in a queue to mediate conduction, as shown in Figure 2.
Ion Channel Gating
Ion channel gating refers to opening and closing of the ion conduction pore in response to a specific stimulus. Certain channels open when ligands bind (ligand-dependent channels); others open in response to membrane voltage (voltage-dependent channels).
MthK is a ligand-gated potassium channel that opens in response to calcium binding. To understand this form of ligand-dependent gating, we cloned and expressed MthK, characterized its function in membranes, and determined its crystal structure. Because this channel structure was determined in the presence of calcium, the pore is open. Through comparison of the KcsA and MthK channel structures, we can infer the conformational changes that underlie potassium channel opening. We have discovered that many potassium channels have a "gating hinge" that allows large helical motions within the membrane to open the pore.
Voltage-dependent gating allows "excitable" cells to produce electrical impulses called action potentials, which underlie muscle contraction and nervous system function. Our laboratory has sought to understand the structural basis of voltage-dependent gating. We first identified an archaebacterial potassium channel, KvAP, that contains the canonical voltage-sensor domain of higher eukaryotic voltage-dependent potassium (Kv) channels. We used electrophysiological methods to characterize the functional properties of KvAP and x-ray crystallography to determine several atomic structures. We recognized these structures as nonnative conformations, but through a combination of structural, biochemical, and functional analysis we posited several fundamentally new ideas: (1) voltage-sensor domains are loosely attached to the pore at its perimeter, (2) they contain a helix-turn-helix element that we named the voltage-sensor paddle, and (3) the voltage-sensor paddle moves at the protein-lipid interface, carrying charges within the membrane electric field.
In the next stage of the voltage-dependent-gating project, we determined the atomic structures of mammalian voltage-dependent potassium channels called Kv1.2 (a Shaker-family potassium channel) and paddle chimera (a mutant) at high resolution (2.4 Å) in a lipid membrane-like environment (movie in Figure 3). These large mammalian membrane protein complexes (containing an aldo-keto reductase beta-subunit) show that the voltage sensors take the form of nearly independent structural domains and that voltage-sensor paddles are indeed located at the membrane-protein interface. The structures also reveal the mechanical connections that enable conformational changes within the voltage sensor to open and close the pore and the atomic interactions that enable charged amino acids to reside within the membrane and sense voltage changes.
My laboratory is also working to understand the structure and function of a different family of eukaryotic potassium channels, known as inward-rectifier (Kir) channels. Different members of this family fulfill important biological functions such as the control of heart rate (Kir3) and the regulation of insulin secretion (Kir6). Because of these functions, Kir channels are important to many disease states. So far we have determined the atomic structure of a chimeric Kir3 channel, one part prokaryotic and another part eukaryotic, and more recently we have determined the structure of a eukaryotic "strong" inward rectifier Kir2.2.
Our work on the structure and function of various ion channels has led us to the realization that membrane protein function, at least for channels, is tied intimately to the lipid composition of the cell membrane. A developing line of research in our laboratory is aimed at understanding the interplay between the membrane proteins and the lipid membrane.
