Study of GSK-3 had an inauspicious beginning rooted in intermediary metabolism. However, owing to the fortuitous convergence of several disparate areas of biology, the enzyme now offers unique opportunities for study of the control of a variety cellular processes. While at first sight a role in transcriptional regulation appears unlikely for a protein first identified as acting on glycogen synthase, it is even more surprising that the same protein should be functionally interchangeable with a fruit fly homeotic gene. Such understandable scepticism, however, is based on teleological bias. Glycogen synthase is a critical enzyme regulating glucose storage. The c-Jun oncoprotein may have the potential to transform cells but this does not excuse it from similar mechanisms of control to glycogen synthase. Likewise, homeotic genes play a crucial role in setting up the body plan of an embryo but must also be subject to control. The main difference is that when such control is lost, the result is rather graphic. It is, therefore, only to be expected that regulatory protein kinases will surface in superficially quite unrelated areas and that many of their targets will be 'housekeeping' proteins. Perhaps the most difficult aspect of protein phosphorylation research is the linking of physiological substrates with particular protein kinases, hence reconstructing pathways. No matter how compelling in vitro data appear, there must be demonstration that the protein is targeted by the specific protein kinase in cells, an extremely difficult process. Most progress in this respect has been made using genetic analysis in lower organisms, especially yeast. Here another problem arises: demonstration of biochemical linkages underlying genetic interactions which requires function to be ascribed to genes identified by a gross effect. The challenge is to co-ordinate these two approaches, a strategy currently being employed to further unravel the biological role of GSK-3.