Mechanisms of membrane pore formation

The amino acid sequence of a protein, determined by its genetic code, defines the structure and function of the protein. Usually, the native (correct) protein structure is unique, and is the most stable form for that particular protein. There are two distinct environments in which proteins are stable: water-soluble proteins are found in solution in cells and tissues, but membrane proteins are found in the oil-like layers of cellular membranes. Soluble proteins are much easier to work with and are much better understood than membrane proteins. Although membrane proteins play important roles in many key cellular functions such as detection and response to signals from the environment, communication between cells, and uptake of nutrients, only a tiny number of three-dimensional structures of membrane proteins are known in comparison to those of soluble proteins. The surface properties of soluble and membrane proteins are very different. A very interesting class of proteins break the general rule that proteins are either water-soluble or inserted into membranes. The bacterial toxins and certain proteins of the immune system are synthesised as individual, water-soluble proteins but in the course of their function, they assemble into rings that penetrate cell membranes and puncture holes through their target membranes. In the case of bacterial toxins, the role of such toxins is to release nutrients for the bacteria, incidentally killing the host cell. In the immune system, pore forming proteins are secreted in the course of immune surveillance when infected or cancerous cells are detected, in order to kill them. Thus, pore-forming proteins have evolved as part of the 'armaments race' between organisms and their pathogens. In this project, we are studying the structures of the pores formed by the bacterial toxin pneumolysin, an important factor in pneumonia and other diseases, and also by the immune system protein perforin, essential for the immune response to maintain the health of the organism. The pores are bound to model membranes, and we can study them by recording images of rapidly frozen suspensions of these membranes in an electron microscope. With suitable computer image processing we can determine their three-dimensional structures, which will give us an understanding of how these proteins change their shape and properties so dramatically, and how the membrane is punctured.

We plan to study the mechanisms of membrane pore formation by the bacterial toxin pneumolysin and by the immune system protein perforin using cryo electron microscopy (EM) and single particle analysis of pores in membrane vesicles. Pneumolysin is a member of the family of cholesterol dependent cytolysins, and we have recently determined the structures of membrane surface-bound (prepore) and pore forms by imaging these ring shaped complexes in liposome membranes. Docking the atomic structures of the domains of the closely related perfringolysin into our cryo EM density maps has revealed dramatic protein conformational changes and membrane bending associated with pore formation. We plan to build on this work, which is in collaboration with Professor P Andrew at Leicester University, to gain a better understanding of the mechanism of pore formation. We will attempt to improve the resolution of the pore structure by using erythrocyte membrane vesicles, which are more robust than artificial liposomes. In order to probe the pathway of the conformational change, we will study the structure of an intermediate between prepore and pore states, which is formed in a helical assembly. Mutant forms of the protein will also be analysed to probe the pore formation pathway. Perforin: Professor J Trapani of the Peter MacCallum Cancer Institute in Melbourne, Australia, has recently developed a successful eukaryotic expression system for the membrane-disruptive protein perforin, which is released by cytotoxic T lymphocytes and natural killer cells to kill infected or transformed target cells by osmotic lysis or by entry of granzymes to stimulate apoptosis. We have found that our methods for imaging membrane bound pores are also applicable to perforin pores. Therefore we plan to accumulate a data set of perforin pore images from vitrified liposomes and determine their 3D structure. Now that a reproducible expression system is available, it is likely that atomic structure information will become available for at least parts of the perforin molecule during the course of this project. Our maps will be combined with any other structural information available in order to enhance their interpretation. New approaches to protein-membrane interactions: In order to address the protein-membrane interactions in more detail we also plan to explore two new approaches. With Professor M Overduin of Birmingham University we will study the membrane docking interaction of pneumolysin by NMR spectroscopy of its C-terminal domain with bound phospholipid/cholesterol/detergent. This domain has an immunoglobulin fold and is stable as an independent structure. Prof Overduin has developed NMR structure determination methods for protein-lipid complexes. If successful, this analysis will enable us to define the protein-lipid interactions in membrane docking and to analyze the sequence specificity of different CDC family members. In another new collaboration, we plan to explore novel ideas of protein/peptide design to probe the dramatic conformational change involved in pore insertion. Prof D Woolfson will apply his experience in protein engineering and sequence/structure analysis to designing peptides, based on the transmembrane inserting regions of the CDC family, that will self assemble in solution and in the presence of lipid/detergent to see if we can model the structural conversion and transmembrane beta-barrel formation. All the structural information obtained from these experiments will be interpreted in the context of the pore structure in the lipid bilayer.