A highly stable engineered disulfide bond in the dimer interface of E. coli orotate phosphoribosyl transferase

Structural disulfides are very rarely found in cytoplasmic proteins of mesophilic organisms. Nevertheless, disulfides can form in the cytosol if these are stabilized sufficiently by the supporting protein structure. To investigate the redox properties of structural disulfide bonds, we introduced disulfide bonds into the cytosolic enzyme from E. coli, orotate phosphoribosyl transferase (OPRT). Because OPRT is a homodimer the introduction of opposing cysteine residues (R44C and D92C) into separate monomers of the enzyme meant that disulfide bond formation could easily be followed by non-reducing SDS-PAGE. Redox titration experiments in DTT buffers revealed that the designed disulfide bond was exceedingly stable with a redox potential of -314 mV, close to that of dithiothreitol (DTT). This interchain disulfide bond improved global thermostability by 5.9 {degrees}C relative to wild type protein and 21.4 {degrees}C above its reduced form. The catalytic activity of the enzyme in either its thiol or disulfide state was not substantially affected, indicating structural integrity. Combining an inactive subunit with the asymmetric nature of the disulfide bond enabled determining the rates of subunit rearrangement and disulfide formation. The R44C/D92C double mutant resulted in the formation of a symmetric homodimer with two interchain disulfides. However, the global stability of the two-disulfide dimer was not substantially increased relative to the dimer with the single disulfide bond, suggesting that the stabilizing effect was gained by preventing dimer dissociation. Finally, we show that the engineered disulfide bonds are formed to a significant degree in the cytosol of yeast cells, allowing us to determine the redox potential of the cytosol in equilibrium with this substrate to -300 mV. This is slightly lower than that previously determined using a GFP-based sensor, rxYFP in yeast. We find that disulfide bond formation is particularly enhanced in mutants lacking the glutathione reductase, Glr1. Thus, the engineered disulfide provides an alternative method for determining intracellular redox potential in living cells with an extended dynamic range relative to GFP-based sensors.