Combining structure and function in the nicotinic superfamily: the single-channel activation mechanism for the prokaryotic model channel ELIC

The purpose of our work is to understand how ion channels function as molecules. The key to this is to know the 3-D structure of the channel (by X-ray crystallography) and to find out how this structure moves when the channel is activated, mainly by recording its electrical activity. It is rarely possible to get both sorts of information for the same molecule. New data mean that this is now possible for the group of channels we work on.
Ion channels are proteins coded in our genome and are essential components of many cells in our bodies. For instance, they allow cells to communicate with each other at cell-to-cell junctions called synapses. This is essential not only in the brain but also to in the rest of our bodies where it allows the appropriate commands to reach muscles in our limbs and to regulate our blood pressure and heart rate. Channels also allow each neurone to process the information it receives from other neurones. Channels are important for normal human physiology and for disease. Mutations in the genes that code for channels can damage channel function and produce inherited human disease, such as cystic fibrosis. In addition to that, many drugs used for common diseases or in anaesthesia act by binding to channels. In particular, the group of channels that we study, the nicotinic superfamily, are targeted by sleeping pills, drugs for epilepsy, nicotine in tobacco and some insecticides.
There are two main sources of information about channels: the first is X-ray crystallography, which provides us with information about the shape of the protein and the second is electrophysiology, which measures the electrical signal the channel produces. In the most advanced form, which is our special expertise, this technique detects the current that passes through a single protein molecule in real time, even though it is very small (more than a billion times smaller than the current in a kettle). This technique is very useful, because it allows us to measure the speed with which molecular events in the function of the channel occur. Hence we can understand channel function precisely as a chemical reaction, quantifying each step, from the binding of the neurotransmitter to the opening of the channel. By doing this in channels in the nicotinic group, we have found how tightly neurotransmitters and drugs bind to the protein when it is active or inactive and why some drugs act more strongly than others.
Ideally we should study the structure and the function of the SAME channel. This is not easy because channels are difficult to crystallize, and so far we have good structures only for three channels in this group (GLIC, ELIC and GluCl). Of these, GLIC produces electrical signals that are too small for good electrophysiology. As for GluCl, we don't know how good its signal is (the structure has literally just been published). A potential problem with GluCl is that it does not open like all other channels in the group do, in response to a neurotransmitter-like compound, but it requires TWO different substances, binding to different places, so we don't know how good a model it will be.
Until now ELIC was thought not to be able to open. Other scientists have now discovered the right substances that activate ELIC, and it turns out to open well and to give an excellent, big signal. We want to apply the single-molecule recording that is our special skill to ELIC, so that we can understand how it functions as a molecule. Once we have that, we can push our understanding much further, because we can refer to precise structural information (available for ELIC) in interpreting the effect of drugs and the effect of mutations in the channel.
This is basic research but it is what is needed if we want to explain what bits of the molecule change and how they move when the channel is activated, where exactly drugs bind to the protein and how we should modify the structure of drugs in order to make them more effective.

In the nicotinic superfamily, high resolution crystal structures are available for two prokaryotic channels, GLIC and ELIC. Neither channel has been studied by detailed functional analysis, such as single-channel kinetics, because GLIC has very low conductance, and the agonist for ELIC was unknown. Recent work by the two European groups that have solved the structure of ELIC (see Letter of Support) showed that ELIC is a non-selective cation channel that opens in response to many small-molecule agonists, including GABA. When open, ELIC has a very high conductance (80 pS) that makes it ideal for single-channel recording. We propose to apply our expertise in single channel kinetics to characterize in detail the activation mechanism of ELIC, by our established method of computational global fitting of mechanisms. This technique is the only one that can estimate microscopic association and dissociation rate constants for the binding of agonists to the channel in its resting, pre-open and open states. It recently allowed us to identify reaction intermediates in the activation of two other nicotinic channels, the glycine receptor and the ACh muscle receptor. These reaction intermediates are the key to understanding differences in agonist efficacy in this superfamily. Applying our methods to ELIC, which is amenable to structural investigation, will allow much more robust interpretation of the effect of mutations, and bring on a new era for the study of structure-function relations in the nicotinic superfamily, making ELIC an extremely attractive model system for this purpose. Once we have obtained a quantitative model for the activation by the full agonists available, we will extend our work to two specific further questions, namely the activation of ELIC by partial agonists (including GABA) and the role of the M2-M3 loop in transducing agonist binding in the extracellular domain to channel opening.

Our aim is to establish by single channel analysis the quantitative activation mechanism of the prokaryotic channel ELIC, a member of the nicotinic superfamily of ligand-gated channels whose structure was solved to high resolution by X-ray crystallography. This will allow the combination of advanced structural and functional data on the same nicotinic channel, a combination that has revolutionised the field of potassium channels.
The results of our work will be useful to all biophysicists and physiologists that work on ligand-gated channels in this superfamily. At the moment we have good structures only for members of the superfamily that are either non-functional (soluble ACh binding proteins) or hard to characterize (such as GLIC). The recent discovery that ELIC can be activated to a high conductance (ideal for advanced electrophysiology) means that we can model quantitatively its behaviour, from the number and nature of the agonist binding steps to the conformational changes it must undergo to open. These results are also essential to pharmacologists. Measuring maximum open probability and agonist efficacy allows reliable interpretation of the more common macroscopic measurement (eg dose-response curves) used to characterise the effects of drugs and that of mutations. Notably, our single-channel method is the only way to estimate the agonist microscopic binding affinity for the receptor resting and activated states. This is key to understanding how agonist activity relates to the structure of the agonist molecule itself and to its sites of interaction with the protein (that may be verified by further structural work).
Having a model for ELIC activation poses the basis of structure-function work in the protein, in which mutational strategies can be designed with firm structural knowledge of the position of each residue, and can be more easily interpreted. ELIC will provide the best model system to monitor the timing of domain motion in the protein by the application of phi analysis, casting light on the dynamics of channel activation in the whole superfamily, a group of channels of great physiological and pharmacological importance.
As modelling the motions of membrane proteins by molecular dynamics becomes more feasible, researchers in this field will find that our results are the ideal experimental counterpart to verify the predictions of computational simulations, as they provide measurements of microscopic ligand affinity for the different states of the receptor, probe the conformational energy landscape of the protein and eventually build a map of the timing of domain motion.
Because of the usefulness of these results to the fields of structure, function and computational modelling, we will take special care to disseminate our results at meetings that are especially designed to be interdisciplinary and bring together scientists with different expertise.
A key aspect in maximising the benefits of our work to the nicotinic field is our collaboration with the group of Prof Dutzler (Zurich), one of the foremost labs working on the structure of prokaryotic channels in this superfamily. They have obtained structures for ELIC and GLIC and showed that ELIC can be activated by a range of agonists (see Letter of Support). Keeping each other up to date on progress on our work on this channel will enable us to identify quickly promising projects for joint work that exploit the complementary points of strength of our labs. This will become possible as soon as the first preliminary characterization of the main features of the activation mechanism become available, because this will then allow us to draw on our expertise on a range of other channels in the superfamily to design appropriate mutational strategies. Further structural characterization is also always a possibility, bearing in mind the difficulty, unpredictability and low success rate for this work in membrane proteins, even of prokaryotic origin.