Predicting enzyme function by docking high-energy intermediates of potential substrates

COMP 371

Peter Kolb, kolb@blur.compbio.ucsf.edu1, Johannes C. Hermann, joha@blur.compbio.ucsf.edu1, Ricardo Marti-Arbona2, Eman Ghanem2, John J. Irwin, jji@cgl.ucsf.edu1, Frank M. Raushel, raushel@mail.chem.tamu.edu2, and Brian K. Shoichet, shoichet@cgl.ucsf.edu1. (1) Department of Pharmaceutical Chemistry, University of California, San Francisco, 1700 4th Street, San Francisco, CA 94158, (2) Department of Chemistry, Texas A & M University, P.O. Box 30012, College Station, TX 77842
Due to many genome sequencing and structural genomics initiatives, there is an abundance of enzymes of known 3D structure but unknown function. To use in vitro experiments to annotate them is clearly not tractable. Moreover, bioinformatics methods are not applicable in all cases, especially when sequence similarity is low.

As an alternative computational approach and a complementary method in general, docking of high-energy intermediates of substrates is an emerging technology. Its basic idea is to dock structures mimicking reaction intermediates instead of the groundstates of molecules, because enzymes are preorganized to recognize and stabilize the intermediate states.

We have applied this approach to the functional annotation of amidohydrolases, a large enzyme superfamily which catalyzes hydrolysis reactions. In multiple cases, we have successfully predicted the correct catalyzed reaction without any prior knowledge of the substrate class.