Protein structure to function: insights from computation

Computation plays an important role in functional genomics. THEMATICS is a computational method that predicts chemical and electrostatic properties of residues in enzymes and utilizes information contained in those predictions to identify active sites. The only input required is the three-dimensional structure of the query protein. The identification of residues involved in catalysis and in recognition is discussed. The two serine proteases Kex2 from Saccharomyces cerevisiae and subtilisin from Bacillus subtilis are used as examples to illustrate how the method finds the catalytic residues for both enzymes. In addition, Kex2 is specific for dibasic sites and THEMATICS finds the recognition residues for both the S1 and S2 sites of Kex2. In contrast, no such recognition sites are found for the non-specific enzyme subtilisin. The ability to identify sites that govern recognition opens the door to better understanding of specificity and to the design of highly specific inhibitors.Received 22 July 2003; received after revision 16 September 2003; accepted 20 October 2003