|
Laboratory of Molecular Biophysics
|
Next: Glutamate Receptors, Up: Ion Channels, Return to: Contents.
Potassium channels are a diverse and widespread family of channels. A bacterial K channel, KcsA, for which there is an X-ray structure [2], provides a structural paradigm for the more complex K channels of vertebrate nervous systems. Simulation and modelling studies based on the KcsA structure are being used to address a number of aspects of the atomic resolution physiology of this and related K channels.
|
These studies have continued previous molecular dynamics (MD) simulations of KcsA in a lipid bilayer [3]. In common with other K channels, KcsA is selective for K+ over Na+. Simulations of KcsA in a fully solvated POPC lipid bilayer, starting with either K+ or Na+ ions in the filter of the channel, reveal considerable differences in the nature of the interactions of the two species of ion with the channel protein. In particular, the Na+ ions are seen to remain at their positions in the selectivity filter, as opposed to K+ ions, which are relatively mobile (on a ca. 1 ns timescale). Whereas K+ ions switch between different "sites" in the filter, each site defined by 8 oxygen atoms, Na+ ions prefer to sit close to the centre of a ring of four filter oxygen atoms. In both sets of simulations, small distortions of the selectivity filter are seen, suggesting that it may not be completely rigid on the timescale of ion permeation (ca. 10 ns).
The selectivity of the potassium channel KcsA was investigated by the use of molecular dynamics simulations of both K+ and Na+ along the length of the channel. The simulations comprised of placing each ion at various sampling positions along the length of the pore. Analysis of the dynamics and energetics revealed that even in this relatively simple methodology (compared to e.g. free energy calculations) the selectivity of the channel can be explained quite well by the differences in behaviour between the two ions [4]. Computational studies of selectivity are continuing and, in collaboration with IHS and DPT, the sensitivity of simulation results to ion van der Waals parameters used in simulations is being explored.
Kir6.2 is a mammalian K channel that complexes with a sulphonyl-urea protein
SUR1 to form the KATP channel which controls insulin release in
-pancreatic
cells. The central pore-forming domain of Kir6.2 has been modelled on the basis
of the X-ray structure of KcsA [5]. The model was found to agree well with
experimental data, including cysteine substitution accessibility experiments.
To further test the validity of the model, and to investigate the physical
mechanisms of ion permeation it was subjected to MD simulations under varying
conditions. These enabled us to assess the effects on the model's behaviour of
a number of diverse factors. These included altering the ionisation states of
the sidechains, altering the protein environment (solvated phospholipid bilayer
vs. solvated octane slab - a bilayer mimetic), and changing the hydration level
of the cavity region. MD simulations have also been performed on a homology
model of a mutant channel (in collaboration with F.M. Ashcroft and P. Proks).
The simulations showed that the ability of K+ ions placed in the selectivity
filter to move through the pore in a concerted single-file fashion is
relatively robust to changes in simulation conditions. The selectivity filter
was flexible as the ions and water move through it.
|
One of the main limitations of the studies of Kir6.2 is that the homology model generated based on KcsA is missing both the N and C-terminal regions, which may be expected to be involved both with protein stability and with functionality. In order to understand these aspects of K channel structure we have returned to KcsA. Although this is missing the N- and C-termini from its X-ray structure, coordinates have been derived for these regions from EPR experiments [6]. Using a model of the entire KcsA molecule (coordinates kindly provided by E. Perozo, University of Virginia) simulation studies are being employed investigate e.g. the stability of the KcsA monomer and the nature of the self-assembly process that yields a stable tetramer. In parallel, simulation studies of fragments of KcsA (in collaboration with D. Doyle) will explore their role in the folding and assembly of the intact channel.
The TWIK family of background K channels are involved in a wide range of physiological functions. Such channels are formed by two subunits, each containing two copies of the P-domain fold (i.e. helix/P-loop/helix) found in KcsA and Kir. Combined modelling and simulation studies are being used to study a relatively simple member of the TWIK family from C. elegans, namely Twk-5. It is hoped that development of a plausible model for Twk-5 will provide insights into the structure/function relationships of other members of this important K channel family.
A bioinformatics approach is being used to investigate the regulatory interactions of ion channels. The overall objectives are to identify: (i) regulatory domains/subunits of ion channels and to describe the network of interactions of these and other proteins; (ii) sequences and motifs/folds for these regulatory subunits/domains; and (iii) related motifs/folds in non-channel proteins including bacterial and yeast homologues.
Initial investigations are focussed on K channels. Various techniques from sequence-based and structural bioinformatics are being employed. The former include large-scale sequence searching/retrieval, multiple sequence alignment, secondary structure prediction and domain analysis. The latter cover threading, homology modelling, and refinement of models via molecular dynamics (MD) simulations. The final outcome of the project will be a complete and annotated description of the regulatory subunits/domains of ion channels and a description of the network of interrelationships between these and non-channel proteins.
Next: Glutamate Receptors, Up: Ion Channels, Return to: Contents.