Characterization of ligand-binding domains and the ligand binding site on protein disulphide-isomerase

Proteins carry out most key biological activities in all organisms, and the individual biological activity of each protein is crucially dependent on its specific 3-dimensional structure. One of the most significant discoveries in cell biology in the past 20 years has been the recognition that there is an extensive and complex machinery in cells devoted to ensuring that newly-made proteins fold up and assemble correctly to their specific and unique 3-dimensional structure. This machinery comprises 'folding catalysts' which ensure that proteins fold rapidly and correctly and 'molecular chaperones' which prevent misfolding. This machinery is now quite well-defined -- in terms of the identification of its component parts -- but we know very little about how these folding catalysts and chaperones work in molecular detail. The aim of this proposal is to increase our detailed molecular-level understanding of one part of the cellular protein folding machinery. Protein disulphide-isomerase (PDI) is a folding catalyst and chaperone which has been known for many years. It is absolutely required for the folding and assembly of proteins that contain disulphide bonds. Disulphide bonds provide proteins with additional stability and are found in almost all proteins which are secreted from cells or exposed at the extracellular surface of cells. Since this group of proteins includes most protein hormones and other intercellular messengers, hormone receptors, digestive enzymes, antibodies, blood clotting proteins, and (in other species) venom toxins, plant storage proteins etc., this is a very significant class of proteins. For example, most of the human protein drugs which are currently used in therapy (such as insulin, interferons, growth hormones, antibody fragments, blood clotting factors etc). are disulphide-bonded proteins. To the best of our knowledge, PDI or a closely related member of the PDI family, is required for the correct folding of all such proteins. Consequently, more detailed understanding of PDI would be significant not only for basic cell biology but also for medical, veterinary and biotechnological applications. Surprisingly, after almost 30 years of study, the detailed structure of PDI is not known at the molecular level, and so we cannot picture precisely how it acts to assist newly-made proteins to fold and form correct disulphide bonds. There appear to be some difficulties which have frustrated conventional approaches using x-ray crystallography. Preliminary work that we and our collaborators have done over the past 2-3 years suggests that we now understand the basis of these difficulties and makes it possible to plan how to determine the structure of PDI bit-by-bit. We plan to start with the 'domain' of PDI which is most interesting to us, because we know that it is the key domain for the 'chaperone' properties of PDI. We will determine the structure of this domain alone and in combination with a neighbouring domain, as a step towards determining the whole structure. We will not focus simply on a static picture but also determine the flexibility of these parts of PDI, in order to picture the range of molecular motions they undergo on various timescales. Finally we will study these domains of PDI in combination with some small proteins and even smaller fragments (peptides) aiming to understand how PDI and the proteins on which it acts bind to each other and how each influences the detailed structure and dynamics of the other. This will finally give us some insight into how PDI works, in molecular terms.

We aim to define in detail the interaction between protein disulphide-isomerase (PDI) and its protein substrates, in order to picture its mechanism, illuminate its cellular function and inform work on PDI homologues. We will determine the structure of the b' domain of human PDI (the principal ligand binding domain) in various contexts. Our NMR data on PDI constructs including the b' domain indicate conformational heterogeneity within the b' domain that explains previous failures to solve the structure by x-ray and NMR approaches. Our targets will be constructs whose spectra indicate a single defined structure; the b-b'-x construct comprising all the non-catalytic domains of PDI, and mutants of the wild-type b'-x sequence. We will express them in E.coli and purify 15N or 15N/13C or 15N/13C/2H proteins for high resolution heteronuclear NMR studies. We will assign resonances, generate distance and angle constraints and calculate structures using conventional multi-dimensional NMR approaches. We will explore the dynamic properties of the proteins under study, using relaxation and H/D exchange methods, obtaining site-specific data where possible. We will determine the effects of small unstructured peptide ligands on NMR parameters of these proteins in order to define binding sites and impacts of ligand binding on PDI domain dynamics. We will initiate experiments using incompletely-folded protein ligands, focussed on the structural and dynamic properties of the ligand protein and how these reflect binding to PDI. For this work we will express in E.coli mutants of bovine pancreatic trypsin inhibitor (BPTI) which contain Cys-to-Ser mutations to constrain the disulphides that can be formed. These mutants mimic incompletely folded intermediates on the BPTI disulphide-linked folding pathway. We will use 15N relaxation measurements and transferred NOE approaches to explore the structures of these species when bound non-covalently to PDI and the nature of the interaction.