The interaction of proteins to form macromolecular complexes is the basis for most biological processes. Approaches have been described that employ artificial constructs to promote such complexes and assess the consequences. For example, a protein interaction scheme has been described that examines the effects of a specific phosphorylation event catalyzed by a protein kinase via the provision of an artificial protein binding interface between a modified version of the kinase and a single substrate. We have generalized this type of approach to form the basis for a genetic selection to identify proteins that exert an activity when recruited to a target protein. The assay uses the leucine zipper domains from the mammalian transcription factors Fos and Jun to force the interaction of two proteins. With a target protein fused to the Jun zipper and a library of open reading frames fused to the Fos zipper, we demonstrate this approach in yeast with both a selection to identify membrane-associated proteins and a selection to identify candidate components of the filamentous growth MAP kinase pathway.