|
Laboratory of Molecular Biophysics
|
Return to: Contents.
Membrane proteins are central to many, if not all, cellular processes.
In particular they play key roles in signalling between cells, in
transport across cell membranes, and in energy transduction processes.
From a post-genomic perspective, their importance can be assessed from
the observation that ~25% of genes code for membrane proteins. Thus
there are ~8000 membrane proteins encoded within the human genome. From
a biomedical perspective, one may note that membrane proteins
constitute ~50% of possible targets for novel drugs, ranging from drugs
acting on the nervous systems to novel antimicrobial therapies.
Despite the importance of membrane proteins, they remain under-explored territory. High resolution structures are known for ~50 membrane proteins, in contrast to ~22,000 structures for water soluble proteins. However, the number of structures of membrane proteins is starting to grow exponentially. Computational methods play an essential role in understanding the relationship between structure and function of membrane proteins. Such studies can be divided into three broad categories:
Ion channels play a central role in the physiology of both excitable
(e.g. neurones, muscle, heart) and non-excitable (e.g. pancreas,
bacteria) cells. They allow rapid (~107 ions s-1)
flux of selected ions through a transmembrane pore. Channels are gated,
i.e. regulated, by a number of factors including e.g. transmembrane
voltage and/or binding of ligands to extra-membranous domain. The
overall aims of computational studies of ion channels in my research
group are:
Our research embraces a number of channels and related membrane
proteins. In particular, the current focus is on:
 |
| Figure 1. Water and ions within the
pore of the nicotinic acetylcholine receptor (image courtesy of Oliver
Beckstein). |
In this area, our research has focussed on the properties of ions and
water when confined within transmembrane pores of sub-nanometer
dimensions. In particular, we are interested in how the dynamic
behaviour of water and ions within pores may differ from that in the
bulk state. Such basic information is fundamental to an accurate (and
hence predictive) understanding of the relationship between single
channel structure and function.
The main research challenge here is to develop conceptual and
computational approaches than enable us to bridge several timescales,
from sub-ns protein motions that are essential to smoothing the energy
landscape for ion permeation, to ms timescale motions involved in
channel gating. In this way it should be possible to construct a bridge
from atomic resolution structure to single channel (i.e. patch clamp)
physiological measurements. This in turn provides a paradigm for a
rigorous approach linking structure with function for a range of more
complex transport proteins. Potassium channels provide a focus for
developing this approach as a wealth of physiological data is
available, in addition to a number of (bacterial) structures.
The majority of membrane protein structures are of bacterial homologues
of membrane proteins. It is therefore essential to develop accurate
models of the corresponding human membrane proteins and to assess the
extent to which simulations may be used to relate (model) structure to
physiological function. Having achieved some success in this area with
K channels we wish to build upon this approach, extending it to a wider
range of membrane proteins. A component of this will involve
development of a high-throughput pipeline for modelling and simulation
of whole families of human membrane proteins based on bacterial
homologue structures.
Neurotransmitter receptors provide a focus for studying the gating of
ion channels. Two aspects of gating are currently being investigated:
A number of transporters (e.g. ABC transporters, lactose permease,
AcrAB) appear to operate via a mechanism whereby access to (i.e.
gating) of a central pore is coupled to solute translocation. The
conformational changes underlying this are relatively slow (~1 ms)
compared to the timescales of conventional molecular dynamics
simulations (~10 ns). A number of more advanced simulation techniques,
ranging from steered simulations to coarse-grained simulation models,
will be required to address the mechanisms of these transporters.
Ion channels and related membrane proteins create nanopores in
biological membranes. Therefore theoretical and computational studies
of such proteins provide insights into the design principles of
nanopores in general. One aspect of current research is to explore the
interface between membrane protein biology and nanoscience, with a
longer term view of aiding the development of membrane protein based
nanotechnologies, such as receptor and channel based biosensors.
Simulation and theoretical studies are being used to understand the
physical behaviour of water and electrolytes within nanopores (see
above). In collaboration with colleagues undertaking experimental
studies (H. Bayley, Chemistry, Oxford) my group is exploring how
changes in the pore dimensions and the chemical nature of the pore
lining alter the behaviour of water and ions within nanopores.
In order for membrane proteins to be exploited in nanotechnological applications we require an improved understanding of their nanomechanical properties. Simulations are being used to explore the behaviour of proteins under nanomechanical stress (e.g. compression in simulated AFM experiments, in collaboration with J. Davis, Chemistry, Oxford), and under non-native conditions (e.g. in non-bilayer environments). These studies will lead to a better understanding of the robustness of selected membrane proteins to changes in their physical environment.
Bacterial outer membrane proteins (OMPs) have a diversity of roles
(pores, transporters, enzymes) in the biology of Gram negative
bacteria. They are potential targets for novel anti-microbial drugs and
vaccines. A substantial number of crystallographic structures, and some
from NMR, are known for OMPs and so they provide an opportunity for
comparative simulation studies of membrane protein conformational
dynamics in relation to both function and stability.
The structures of a number of smaller OMPs (OmpA, OmpX, PagP) have been
determined both by protein crystallography and by NMR of mixed
protein/detergent micelles. Extended MD simulations are being used to
explore how the conformational dynamics of these proteins changes
between the experimentally relevant environments (crystal or micelle
see fig 2) and the biological environment (membrane). These studies are
importance in aiding extrapolation from experimentally determined
structures to biological function.
 |
| Figure 2. Self-assembly of
a protein-detergent micelle by 'brute force' simulation (image courtesy
of Peter Bond). |
The OMPs provide an opportunity for large scale comparative studies of
conformational dynamics within a rather disparate family of membrane
proteins. To this end, high quality (multi-nanosecond) simulations are
being performed on all OMPs for which structures have been determined
(currently ~25 OMPs). Comparative analyses will exploit the BioSimGRID
technology (see below) and will initially focus on e.g. protein
dynamics vs. secondary structures, and on lipid/protein interactions.
The complexities of transport across bacterial membranes make the outer
membrane transport systems a viable test case for a systems approach to
biomolecular simulation. However, to address emergent phenomena on
larger length and time scales than is possible via 'brute force'
atomistic simulations, some degree of coarse-graining will be necessary
in order to develop suitable mesoscale models. Very large scale outer
membrane outer membrane simulations are being performed (using the
facilities at HPCx). These will be analysed with a view to
parameterisation of mesoscale models.
The increase in number of membrane protein structures enables
structural bioinformatics studies in order to discover general
principles of membrane protein structure. Current studies are focussed
on transmembrane (TM) helices for which ~400 structures are available,
and include:
A combined database and comparative simulations approach is being used
to probe the nature of interactions between membrane proteins and lipid
molecules. A particular focus is on how amphipathic aromatic and basic
residues help to 'lock' membrane proteins into their preferred
transbilayer orientation.
A number of simulation approaches (implicit solvent models, micelle
simulations, replica exchange) are being combined to develop MD
simulations as a 'brute force' approach to predicting TM helix packing
as a component of ab initio fold prediction for (simple)
membrane proteins.
We ate heading a UK consortium to develop a prototype distributed
database for biomolecular simulation data (BioSimGRID). This
project will exploit the developing UK GRID/e-science infrastructure,
and involves collaboration between 6 UK biomolecular simulation
laboratories and 2 computer science departments.
We also heading a newly established collaboration between 4 UK
simulation laboratories and 2 computer science departments to implement
multiscale (from QM/MM through to mesoscale) simulations of membrane
bound enzymes.
A number of membrane protein databases (membrane protein structure,
lipid/protein interactions, ABC transporters, glutamate receptors etc.)
are being developed in my laboratory. The intention is to use e-science
technology to integrate these to provide a single resource for membrane
protein structural bioinformatics and simulation data.
There is a pressing need to develop methodologies that enable vertical
integration between biomolecular and higher-level 'systems'
descriptions of membrane proteins and their biological roles. A project
is being initiated to explore different algorithmic approaches to
bridging the gap between biomolecular and more conventional systems
level descriptions of membrane biology. The initial applications will
be to prokaryotic transport and signalling, and these are somewhat more
data-rich. However, the technologies to be developed will also be
applicable to integrated approaches to ion channels and related
transport proteins in the context of a computational systems biology
approach to cardiac function (work in collaboration with D. Noble, P.
Hunter et al., Physiology, Oxford and Univ. of Auckland).
Return to: Contents.