Why ion pair reversal by protein engineering is unlikely to succeed

Genetic engineering is a powerful tool for exploring correlations between structure and function in proteins, but as yet we are unable to use it for effective protein design. One of the most interesting examples, which would seem to be obvious, is reversing the polarity of an ion pair. Changing a positively charged protein group, that provides a strong binding for negative substrates, to a negative group is expected to provide an effective binding site for a positively charged substrate. But several recent experiments on aspartate aminotransferase, trypsin and aspartate transcarbamoylase (Schachman, H. K. personal communication) have indicated that polarity reversal is not so successful. Here we argue that the same factors that make the enzyme an effective system for the (-+) pair will make it a much less effective system for the (+-) pair. We also point out that the unusually low effective dielectric constant (epsilon approximately equal to 13) for the (-+) interaction is due to its microenvironment and this will destabilize a (+-) arrangement having an entirely different dielectric constant (epsilon approximately equal to 80). The calculations presented here evaluate the energetics of ion pairs in protein active sites on a semiquantitative level. This is particularly important when dealing with strong, functionally important interactions that are difficult to evaluate with macroscopic models.