Role of disorder in protein interaction networks
Peter Tompa

Intrinsically unstructured proteins (IUPs) defy the classical structure-function paradigm in that they do not have a well-defined 3D structure but exist in a highly flexible, open conformational state. This special feature enables them to occupy unique functional niches not accessible to ordered, globular proteins. The lack of a stable structure confers special functional advantages, such as a large interaction surface or the ability to adapt to different partner molecules. Here it is shown that these features contribute to the functioning of IUPs in protein interaction networks. Subtle functional details of a range of IUPs unveil that some of these proteins elicit opposing – inhibitory and activatory – action on different partners or even the same partner molecule. We argue that this mechanism might serve as a basic device for increasing network complexity via moonlighting, when the same protein fulfils more than one function. In keeping with this notion, the analysis of the protein interaction networks of S. cerevisiae, D. melanogaster, C. elegans and H. sapiens shows that hub proteins, which determine network topology via interacting with a large number of neighboring proteins, contain significantly more structural disorder than other proteins. Thus, hubs are structurally suited for having multiple interactions, which calls for a re-interpretation of the issue of emergence of scale-free topology in protein interactomes. Our suggestion is that an evolutionary specialization process, in which certain components of the protein interactome raised their fitness for binding by accruing regions of disorder, has taken place and explains network organization as seen today.