Biomedical Engineering Reference
In-Depth Information
More complex potassium channels have large extra-cellular domains instead of
these relatively short sequences. The structure of some of these domains has been
solved separately, for example the T1 and beta domains of a voltage gated potas-
sium channels [54]. A key missing piece at the moment is a structure of the other
transmembrane helices of voltage-gated channels with 6 helices per subunit. The
structures of KcsA currently known are all in the closed state. At the moment one
structure of an open potassium channel is known: MthK is a calcium-gated potas-
sium channel that was trapped in the open state by the presence of calcium [65].
Other channel structures that are known experimentally are two chloride channels
[41] and two mechanosensitive channels [9, 29]. Several modelling and simula-
tion studies have considered the large conductance mechanosensitive channel MscL
[17, 45, 55, 106], but to date no simulation studies of the chloride channels have
appeared. All of these channels are from bacteria, but the potassium and chloride
channels have significant homology to eukaryotic channels. Homology modelling
techniques may be used to build molecular models of at least parts of these channels.
For example, as mentioned above two out of the six helices that make up the trans-
membrane domain of the voltage-gated Kv channels, share homology with KcsA,
and models of the inner two helices of such channels have been build (recently re-
viewed by [25]). Similarly, the pore parts of voltage-gated sodium channels and
calcium channels are also likely to be homologous to KcsA. If we try to link con-
ductance properties to such models an additional degree of uncertainty is introduced
by the use of homology models rather than 'real' structures. For this reason I do not
consider homology models in this chapter, although clearly the proteins modeled are
physiologically very important, and often are important drug targets. As more tem-
plate structures become available and computational procedures improve, this type
of modelling is expected to become increasingly important and useful.
A final important class of channel proteins for which structural information is
available but which is not related to the potassium/sodium/calcium channels or the
chloride channels is the class of ligand-gated channels. These include neurotrans-
mitter gated channels such as the nicotinic acetylcholine receptor, GABA receptors,
and glycine and serotonin receptors. A lower resolution structure of the nicotinic
acetylcholine receptor has been obtained from electron crystallography studies [117].
This protein is a hetero-pentamer consisting of a mixture of homologous subunits.
It probably has 4 transmembrane helices per protein, and has large domains out-
side the membrane. The extracellular domain is highly homologous to a recently
discovered water-soluble acetylcholine binding protein, the structure of which was
solved by crystallography [22]. The pore-lining helices by themselves aggregate
into a channel with conductance properties reminiscent of those of the full channel
[84]. Several models of this peptide channels have given an idea of what it looks
like [71, 80]. Combined with the low-resolution structure of the full protein and the
high-resolution structure of the acetylcholine-binding domain, models/structures of
the full channel as well as homologous proteins are likely to appear soon.
Most simulation and modelling systems have used a set of well-characterized
model systems, ranging in complexity from an infinitely long featureless cylinder
to the actual potassium channel KcsA. The literature on these channels is quite ex-
Search WWH ::
Custom Search