Defining the domain boundaries of the human protein disulfide isomerases

The protein disulfide isomerase (PDI) family of folding catalysts are constructed from combinations of redoxactive and redox-inactive domains, all of which are probably based on the thioredoxin fold. To understand the function of each domain in the variety of catalytic reactions that each family member can perform (to differing extents), the domain boundaries of each family member must be known. By using a technique based on sequence alignments and the known structure of the a and b domains of human PDI, we generated a large number of domain constructs for all six redox-active human PDIs: PDI, PDIp, ERp72, ERp57, P5, and PDIr. The ability to generate significant amounts of soluble protein in E. coli from most of these domain constructs strongly indicates that the domain boundaries are correct. The implications for these domain boundaries on the tertiary structure of the human PDIs are discussed.