Using the two-hybrid system to detect protein-protein interactions

Abstract Protein-protein interactions are essential in almost all biological processes, including replication, transcription, secretion, signal transduction, and metabolism. Thus a central question in the study of any protein is to determine what other proteins are in contact with it. The answer, in the case of oncogene-encoded proteins for example, has provided enormous insights into the problems of cell cycle control and differentiation. It is, therefore, clear that intense research efforts will continue to focus on identifying proteins that interact with some protein of interest, referred to here as the target protein. Current approaches to detect interacting proteins include traditional methods such as co-immunoprecipitation, crosslinking, and copurification through gradients or chromatographic columns (see Midgley and Lane, Chapter 6 of this volume). These biochemical methods have the major disadvantage that interacting proteins are generally known only as bands of a particular relative mobility on a polyacrylamide gel. To progress from these bands to cloned genes is often a difficult undertaking, involving such methods as the initial purification of sufficient protein for amino acid sequencing or antibody production, followed by additional work to screen a library for the corresponding.