Although it is not yet practical to design and synthesize new proteins from their constituent amino acids, existing protein molecules can serve as convenient starting points for the construction of new active sites. Such molecules can be redesigned by using either recombinant techniques or site-selective chemical modification. For example, we have changed the serine protease subtilisin into an artificial selenoenzyme by chemically converting the active site serine residue into a selenocysteine. The resulting protein, selenosubtilisin, has a number of remarkable properties. The modified enzyme catalyzes both the hydrolysis and the aminolysis of activated esters. Indeed, acyl transfer to amines is four orders of magnitude more efficient than with native subtilisin. This result, and the fact that selenosubtilisin does not hydrolyze peptides, suggests that the modified enzyme could be a practical peptide ligase, useful in the convergent synthesis of proteins. Because selenium has several accessible oxidation states, the redox chemistry of selenosubtilisin is also interesting. The artificial enzyme efficiently catalyzes the oxidation of thiols by alkyl hydroperoxides (see Scheme 1), mimicking the action of glutathione peroxidase, an important enzyme that protects mammalian cells from oxidative damage.
Site-directed mutagenesis, kinetic analyses,1H- and 77Se-NMR spectroscopy, and crystallography are integral to our effort to understand how the active site microenvironment influences the intrinsic reactivity of the selenium prosthetic group and to our attempts to optimize the chemical efficiency of selenosubtilisin. Extension of these studies to new protein templates and new prosthetic groups may make a wide range of tailored protein catalysts readily available.