As an increasing number of genomes are sequenced, scientists gain information on what proteins might exist but are still limited in understanding what these proteins do, or where and when they perform their unknown functions. Previous attempts to uncover the roles of new proteins have been based on their similarity to other known proteins, but this approach can fail for a variety of reasons.
John Gerlt and colleagues now apply computational docking of potential enzyme substrates into a structural model to assign the function of BC0371, a protein that had previously been misassigned based on its close relationship with another protein. The authors found that allowing the amino acid side chains in the protein to vary their position during the experiments resulted in improved docking and pointed towards a specific group of compounds as likely substrates. Experimental confirmation of these preferences allowed the team to reassign the protein with its proper function. Application of these ideas should lead to similar advances for other proteins, particularly those in protein superfamilies.
John A. Gerlt (University of Illinois, Urbana, IL, USA)
Abstract available online.
(C) Nature Chemical Biology press release.
Message posted by: Trevor M. D'Souza