We are interested in three related problems protein structure prediction, protein binding site characterization and the rational, structure-based design of drug inhibitors. Currently, our main project involves modeling G protein-coupled receptors, a large family of transmembrane receptors which mediates about eighty percent of hormonal signaling and is ubiquitously involved in all aspects of cellular physiology. Our goal is to model the receptor in sufficient detail to allow the design of specific drug inhibitors acting on the intracellular loops of the receptor. Such a new class of agents would block the normal interaction with the G protein, hence interrupting the signalling pathway to downstream effectors. As a first step to this problem, we have designed a new analytical tool which is able to identify, in a protein family, where binding surfaces to other macromolecules are likely to be located. This method, the Evolutionary Trace, has been tested and is now used to guide mutagenesis experiments in three collaborating laboratories. Its generality makes it applicable to a vast number of proteins that might be involved in pathologic processes and where a detailed understanding of the binding domain would be crucial to effective and directed structure-based drug design. For example, we have undertaken as a test to check our method against DNA-binding proteins where the interface is known, so far with good results. Thus our research may have applications to many areas other than G protein signalling. Our efforts are greatly facilitated by the CGL resource. In particular, we use the MidasPlus program on a daily basis.
Showing the most recent 10 out of 508 publications
