Abstract

A Model Linking Structure and Function in Mechanistically Diverse Enzyme Superfamilies

Patricia Babbitt
University of California at San Francisco

A critical task in interpreting the information coming out of the genome projects is
assignment of the molecular functions of their protein products. A solution to this problem and that of characterization of functional properties in general requires a better understanding of
how new protein functions evolve from a limited set of structural scaffolds. We have examined very
distantly related protein sequences and structures for clues to understanding how some protein superfamilies evolved to deliver a range of different functions while retaining common elements of function associated with conserved elements of structure. Analysis of these
superfamilies suggest that among the several models currently recognized, chemistry, rather than ability to bind a specific substrate type, is a primary determinant in the evolution of new enzyme functions. Recognition of the fundamental design principles associated with the evolution of function across such mechanistically diverse enzyme superfamilies is important for correct
inference of function for open reading frames for which no close homologs of verified function are present in the databases. This information is also valuable for correction of functions misannotated in sequence databases, prediction of
additional functions for previously characterized proteins, target identification,
and to provide guidance for protein engineering efforts.