Protein disulphide-isomerase and the formation of native disulphide bonds.

Disulphide bonds are found in practically every class of extracellular protein, and the formation of disulphide bonds must be regarded as a key post-translational modification of secretory proteins. Despite this, and despite the fact that the existence of disulphide bonds has been known for many years, the mechanism of disulphide bond formation during protein biosynthesis and secretion is not well established, especially by comparison with more recently discovered modifications also discussed in this Colloquium, such as proteolysis of signal sequences or y-carboxylation of glutamyl residues. This is because the problem lies on the periphery of three distinct fields : (i) post-translational modification, (ii) protein biosynthesis, translocation and processing, and (iii) protein folding. Most work on the formation of protein disulphide bonds has derived from an interest in the chemistry of protein folding and so has involved model studies, rather than studies aimed at characterizing the process as it occurs in the cell. However, this line of work has been remarkably informative. The classic work of Anfinsen and his colleagues established that small, single-domain disulphidebonded proteins, such as ribonuclease, after reduction in denaturing conditions, would recover in high yield their native disulphide-bond pairing, conformation and activity on removal of the denaturant and reductant. (Epstein et al., 1963; Anfinsen, 1972). This result established that the native tertiary structure is defined by the amino acid sequence and that the disulphide bonds do not carry structural information. However, it does not imply that disulphide bond formation is irrelevant to the process of protein folding or the maintenance of the native state. Both insulin and collagen cannot be directly folded to their native state. Both are in a metastable state in a conformation achieved by folding of a disulphide-bonded precursor (Freedman & Hawkins, 1977). Thus in both cases appropriate disulphide bond formation is essential either to generate, or to maintain, the biologically active conformation. Recent work has thrown more light on the mechanism of disulphide bond formation in model studies in vitro. Firstly, the use of disulphide oxidants (rather than the dissolved O2 used in Anfinsen’s studies) has clarified the chemistry of the process and established that protein disulphide bond formation occurs via a series of thiol :disulphide interchange reactions (Wetlaufer & Ristow, 1973). Secondly, the isolation by Creighton of specific disulphide-bonded intermediates of the folding and oxidation process has permitted a full kinetic and thermodynamic description of the pathway, in the case of bovine pancreatic trypsin inhibitor, a small protein containing three disulphide bonds. (Creighton, 1978). This work has established that there is a kinetically favoured folding route, but that the native disulphide bonds are not simply introduced in turn; species