The School of Molecular and Cellular Biology at the University of Illinois at Urbana-Champaign

With the availability of complete sequences for the genomes of numerous eubacteria, archaea, and eukaryotes, parallel structure/function studies of enzymes derived from a common progenitor is the best strategy for elucidating structure/function relationships for enzyme-catalyzed reactions. This approach allows an efficient and precise identification of essential structure/function relationships that are important in catalysis. It also allows an understanding of the strategies used by Nature to evolve "new" enzymes and, therefore, provides design principles for the in vitro design of novel enzymes that catalyzed unnatural reactions. We expect that our studies will provide the ability to predict the functions of "unknown" proteins discovered in genome projects.

We are studying two groups of enzymes that are derived from common ancestors, both of which share the ubiquitous (β/α)₈-barrel fold. The members of the enolase superfamily catalyze different overall reactions initiated by abstraction of the a-proton of a carboxylate anion to form an enediolate anion intermediate that must be stabilized by the active site. The members of the orotidine 5'-monophosphate (OMP) decarboxylase suprafamily catalyze different reactions that do not share any mechanistic features; our discovery of this suprafamily supports the hypothesis that Nature is opportunistic and can both identify and utilize functionally versatile active site templates in the evolution of new enzymatic activities.

The active sites of members of the enolase superfamily contain a conserved binding site for a catalytically essential divalent metal ion. The acid/base groups are located on the rims of the active sites "cups" located at the C-terminal ends of (β/α)₈-barrel domains, with one functional group located at the end of each of the eight β-strands. Members of the superfamily can be functionally promiscuous, i.e., in addition to their natural reaction they also catalyze an "accidental" reaction. Understanding how conserved structure delivers different functions allows us to 1) predict the functions of unknown homologues in the sequence databases; and 2) redesign "old" enzymes to catalyze "new" reactions.

Our studies of the recently discovered OMP decarboxylase (OMPDC) suprafamily are focusing on enzymes that catalyze reactions involving enolate anion chemistry rather than the "vinyl anion" chemistry that superficially describes the mechanism of the OMPDC reaction, including 3-keto-L-gulonate 6-phosphate decarboxylase, D-arabino-hex-3-ulose 6-phosphate synthase, and D-ribulose 5-phosphate 3-epimerase. We are performing detailed mechanistic studies of each of these enzymes so that we can understand how the "same" active site residues can be used to catalyze different reactions.

Analyses of the protein databases reveal that the enolase superfamily and the OMP decarboxylase suprafamily contain many members of unknown function-their sequences have been determined in genome projects without regard to biological and biochemical function. In fact, functional assignment of proteins of unknown function is a major problem in postgenomic biology. We are undertaking a major new project in which functional biology (mechanistic enzymology), structural biology (x-ray crystallography), and computational biology (bioinformatics, modeling, and ligand docking) are brought together to address the problem of functional assignment of unknown proteins, with emphasis on members of the enolase superfamily and the OMP decarboxylase suprafamily. The multidisciplinary approaches utilized in this project reflect the intellectual and practical demands of postgenomic biology.

These functional studies involved in these projects provide training in molecular biology, protein purification, and mechanistic analyses, including syntheses of substrates, substrate analogs, and isotopically labeled substrates, as well as product characterization by NMR and MS. The x-ray structure determinations of our proteins are conducted in collaboration with Professors Hazel Holden and Ivan Rayment at the University of Wisconsin. The computational aspects are conducted in collaboration with Professors Patsy Babbitt, Matt Jacobson, Andrej Sali, and Brian Shoichet at the University of California, San Francisco.