Thomas E. Creighton

Catalysis by protein-disulphide isomerase of the unfolding and refolding of proteins with disulphide bonds

Abstract Purified protein-disulphide isomerase has been examined for effects on the pathway and kinetics of the unfolding and refolding which accompanies disulphide bond breakage and reformation in bovine pancreatic trypsin inhibitor and bovine ribonuclease A. The intermediates of the pathways were not altered, although some interconversions which normally are not significant became so in the presence of the isomerase. The rate of every step involving both substantial protein conformational changes and protein disulphide bond formation, breakage or rearrangement was found to be increased significantly, but only when the conformational changes were rate-determining. The protein-disulphide isomerase appears to be a true catalyst of protein unfolding and refolding involving disulphide bond breakage, formation or rearrangement.

A three-disulphide intermediate in refolding of reduced ribonuclease A with a folded conformation.

The interaction between cysteine residues of an unfolded protein to form disulphide bonds under appropriate conditions is the most useful current experimental probe of the large confomrational transitions which take place during folding to the stable, folded conformation, as it may be controlled experimentally, and the normally unstable and transient intermediates may be trapped in a stable form using the disulphide bonds [ 11. This approach has produced the first detailed folding pathway of a globular protein, that of BPTI [2]. A similar study of the larger protein, RNase, the classic subject of protein folding [3-71, extended the earlier studies and trapped and isolated the large number of intermediates that accumulate transiently [8,9]. The 1-disulphide intermediates were found to be essentially a random collection of the 28 possible species, and there were too many 2-, 3and 4-disulphide intermediates for any of the cysteine residues paired in disulphide bonds to be identified, although these intermediates were not a full random collection. Virtually all of the trapped intermediates had no significant amount of stable, folded conformation on the basis of their spectral and hydrodynamic properties, binding to antibodies against unfolded, reduced RNase, but not to antibodies against native RNase, the absence of an unfolding transition induced by urea, and the absence of enzymic activity [9,10]. The sole exception was a species isolated with the incorrectly refolded, 4disulphide intermediates, which had substantial enzymic activity and the spec-

Counting integral numbers of amino groups per polypeptide chain

Proteins have integral numbers of each of the 20 amino acids, but all the currently accepted methods of determining this number measure only a nonintegral ratio of moles of ammo acid per mole of protein. This value is rarely found to be close to an integer, due to experimental error and uncertainty about the molecular weight of the protein. An alternative method which gives correct integral values was demonstrated recently for cysteine residues [ 11; a very similar procedure had been proposed earlier [2,3]. The procedure is independent of any other information about the protein, including its molecular weight. Here we demonstrate a related procedure for counting the integral number of ammo groups, which gives the number of lysine residues plus the o-amino group.