Modeling charge interactions and redox properties in DsbA

Accurate prediction of charge interactions in macromolecules presents a significant challenge for computational biology. A model for the low Cys30 pK(a) and oxidizing power of DsbA (Gane, P.J., Freedman, R. B., and Warwicker, J. (1995) J. Mol. Biol. 249, 376-387) has been investigated experimentally (Hennecke, J., Spleiss, C., and Glockshuber, R. (1997) J. Biol. Chem. 272, 189-195), with substitutions for Glu37 and Glu38 and with residues 38-40 removed. Measured changes in Cys30 pK(a) and redox potential were relatively small and reported to be in contrast to model predictions. It is now shown, particularly with calculations of wild- type:mutant differences for a range of salt concentrations, that the data are consistent with the model and support the key finding that a number of different factors contribute to the oxidizing power of DsbA, so that any particular one need not necessarily be large. A feature of the model is a low protein dielectric, and higher values (which are becoming popular in predictions of pH dependence) are inconsistent with both the difference data and the wild-type Cys30 pK(a).