SEARCH FOR SITES WITH FUNCTIONALLY IMPORTANT SUBSTITUTIONS IN SETS OFRELATED OR MUTANT PROTEIN

A new method is proposed for search of sites containing functionally i mportant substitutions in sets of related or mutant proteins with know n activities or other properties. The method is based on calculating a novel structure-activity correlation (determination) coefficient, SAD C, from the values of activity variation in groups of proteins with si milar site structure. The value of SADC is given by the squared multip le correlation coefficient between protein activity and various site c haracteristics. Five protein families were analyzed: desintegrins, alp ha-interferons, luciferases, Kunitz protease inhibitors, and M2 protei ns of influenza virus A. SADC values calculated for linear, discrete, and spatial sites allowed identification of the positions where differ ences in the amino acid residues correlate with the activities or othe r properties of the proteins or with the phenotype of the respective o rganism. The positions found are located within the known functionally important sites of the proteins or close to them. The new method belo ngs to a series of the quantitative methods of protein analysis used i n the program ProAnalyst (ftp://ftp.ebi.ac.uk/pub/software/dos/panalys t) and can be used in studies of structure-function relationships in p roteins and in constructing mutated proteins with desired changes in a ctivity.