Principal Investigator Frank Gibson Project q11

Membrane Biochemistry Group, Machine VP

John Curtin School of Medical Research

Co-Investigator Graeme B Cox

Membrane Biochemistry Group, John Curtin School of Medical Research

A Combined Molecular Biological and Computer Graphics Study of Membrane Proteins

The structure and function of membrane-bound proteins has proved difficult to study by conventional biochemical techniques. In recent years the techniques of molecular biology have been used and, because of the constraints imposed on protein structure by the lipid environment of membranes, it is possible to make reasonable predictions about the structure of the portions of the proteins actually within the membrane bilayer. Our primary interest is in the ATP synthase complex (responsible for the formation of ATP, the `energy currency of all cells') in which there is a background of over 20 years work in our laboratory.

The ATP synthase is very complex and our major interest is in the membrane bound portion, the so-called Fo, which contains three protein subunits (ab2c9) in which the c-subunits are thought to form a ring around a central core of a and b-subunits. Because of the nature of this complex it is very unlikely that any useful X-ray diffraction data will be obtained in the foreseeable future. Information about structure and function will therefore have to be obtained by other methods. Work to date has shown, that computer graphics can play an important role in this work. For example, predicted structures can be modelled to see if they are physically possible and used to predict which amino acid substitutions might be expected to affect function of the complex. There is a constant interaction between computer modelling and site-directed mutagenesis experiments in the laboratory.

Membrane transport systems other than ATP synthase are included in those under investigation. These include phosphate, sulphate and Ca transporters as well as another proton transporter.

What are the basic questions addressed?

What is the nature of the amino acid side chains essential for the transport of water soluble molecules across cell membranes and their spatial relationships?

What are the results to date and the future of the work?

The transmembrane helices of the a-subunit which are postulated to form the major part of the central rotating core of the ATP synthase complex (see Reference) have been modelled as well as the third protein of the Fo, the b-subunit. It is well established that Arginine-210 and Glutamate-219 of Helix 4 and Histidine-245 of Helix 5 of the a-subunit are essential for proton translocation, together with the Aspartate-61 residues of the 9 c-subunits surrounding the central core. A short peptide was modelled to simulate the active region of the c-subunit, added to the a-subunit, a torsion applied to the side chain of Histidine-245 and the resulting structure minimized. The resulting structure showed that a hydrogen bonded system could be formed involving Arginine-210, Histidine-245 and the Aspartate-61 residues. Using the models constructed so far it has been possible to visualize the basic structure of the complete Fo.

Fig 1a Fig 1b

Fig. Ia shows the essential groups of the a-subunit which would face the Aspartate-61 residues of the surrounding ring of c-subunits while Fig. 1b shows the hydrogen bonding system described above.

The work on ATP synthase will be continued for the foreseeable future. There are number of other membrane proteins being studied by the Membrane Biochemistry Group and preliminary modelling has been carried out on a number of these including the phosphate transporter from Escherichia coli, and a transporter of one of the eye pigments of the fruit fly Drosophila melanogaster and a family of proteins from viruses thought to be important in proton translocation and essential for viral replication. Important regions of the `GABA receptor', a calcium channnel in the human brain have also been modelled.

What computational techniques are used and why is a supercomputer required?

The simulation of the structures in which we are interested require extensive calculations for energy minimizations and molecular dynamics. A typical energy minimization on a mutant in the a-subunit would require about 150 hours on our Iris workstation and the use of the supercomputer not only allows many more structures to be examined but frees the Iris for further modelling and computations on smaller molecules.

The commercial molecular modelling packages Insight, Homology and Profiles-3D from Biosym are being used for modelling and their program Discover for energy calculations.