Protein engineering of domains in flavoprotein disulphide oxidoreductases: contributions to folding and assembly.

Introduction The flavoprotein disulphide oxidoreductases constitute a growing family of homologous dimeric enzymes, with an important and diverse range of functions in vivo. Among its members are dihydrolipoyl dehydrogenase, glutathione reductase, mercuric reductase, thioredoxin reductase, trypanothione reductase and a related enzyme, NADPH peroxidase (for recent reviews of their structures and properties, see [1,2]). The threedimensional structures of glutathione reductase (GR) from human erythrocytes [3,4] and Escherichia coli [ 5 ] have been determined, as have the structures of hotobacter vinelandii and Pseudomonas Juorescens dihydrolipoyl dehydrogenases [6,7], Bacillus sp. mercuric reductase [8], E. coli thioredoxin reductase [9], Crithidia fasciculata and Typanosoma cruzi trypanothione reductases [ 10121 and Streptococcus faecalis NADPH peroxidase [13]. In keeping with a measure of similarity in their amino acid sequences, these proteins exhibit an overall structural similarity, although with some important and interesting differences [2]. In general, each subunit (Mr approx. 55 000) consists of four well-delineated domains: an FAD-binding domain and an NAD(P)H-binding domain (both Rossmann folds), as well as a smaller central domain and a large interface domain that contains most of the inter-subunit contacts (Figure 1).