Correlated mutation analyses on super-family alignments reveal functionally important residues

Correlated mutation analyses (CMA) on multiple sequence alignments are widely used for the prediction of the function of amino acids. The accuracy of CMA-based predictions is mainly determined by the number of sequences, by their evolutionary distances, and by the quality of the alignments. These criteria are best met in structure-based sequence alignments of large super-families. So far, CMA-techniques have mainly been employed to study the receptor interactions. The present work shows how a novel CMA tool, called Comulator, can be used to determine networks of functionally related residues in enzymes. These analyses provide leads for protein engineering studies that are directed towards modification of enzyme specificity or activity. As proof of concept, Comulator has been applied to four enzyme super-families: the isocitrate lyase/phoshoenol-pyruvate mutase super-family, the hexokinase super-family, the RmlC-like cupin super-family, and the FAD-linked oxidases super-family. In each of those cases networks of functionally related residue positions were discovered that upon mutation influenced enzyme specificity and/or activity as predicted. We conclude that CMA is a powerful tool for redesigning enzyme activity and selectivity.