Protein disulfide isomerase (PDI), a very abundant protein in the endoplasmic reticulum, facilitates the formation and rearrangement of disulfide bends using two nonequivalent redox active-sites, located in two different thioredoxin homology domains [Lyles, M. M., & Gilbert, H. F. (1994) J. Biol. Chem. 269, 30946-30952]. Each dithiol/disulfide active-site contains the thioredoxin consensus sequence CXXC. Four mutants of protein disulfide isomerase were constructed that have only a single active-site cysteine. Kinetic analysis of these mutants show that the first (more N-terminal) cysteine in either active site is essential for catalysis of oxidation and rearrangement during the refolding of reduced bovine pancreatic ribonuclease A (RNase). Mutant active sites with the sequence SGHC show no detectable activity for disulfide formation or rearrangement, even at concentrations of 25 mu M. The second (more C-terminal) cysteine is not essential for catalysis of RNase disulfide rearrangements, but it is essential for catalysis of RNase oxidation, even in the presence of a glutathione redox buffer. Mutant active sites with the sequence CGHS show 12%-50% of the k(cat) activity of wild-type active sites during the rearrangement phase of RNase refolding but <5% activity during the oxidation phase. In addition, mutants with the sequence CGHS accumulate significant levels of a covalent PDI-RNase complex during steady-state turnover while the wild-type enzyme and mutants with the sequence SGHC do not. Since both active-site cysteines are essential for catalysis of disulfide formation, the dominant mechanism for RNase oxidation may involve direct oxidation by the active-site PDI disulfide. Although it is not essential for catalysis of RNase rearrangements, the more C-terminal cysteine does contribute 2-8-fold to the rearrangement activity. A mechanism for substrate rearrangement is suggested in which the second active-site cysteine provides PDI with a way to ''escape'' from covalent intermediates that do not rearrange in a timely fashion. The second active-site cysteine may normally serve the wild-type enzyme as an internal clock that limits the time allowed for intramolecular substrate rearrangements.