A web of possibilities: network-based discovery of protein interaction codes

Many proteins, including p53, the FoxO transcription factors, RNA polymerase II, pRb, and the chaperones, have extensive post-translational modifications (PTMs). Many of these modifications modulate protein-protein interactions, controlling interaction presence/absence and specificity. Here we propose the notion of the interaction code, a widespread means by which modifications are used to control interactions in the proteome. Minimal interaction codes are likely to exist on proteins that have two modifications and two or more interaction partners. By contrast, complex interaction codes are likely to be found on "date hub" proteins that have many interactions, many PTMs, or are targeted by many modifying and demodifying enzymes. Proteins with new interaction codes should be discoverable by examining protein interaction networks, annotated with PTMs and protein-modifying enzyme-substrate links. Multiple instances or combinations of phosphorylation, acetylation, methylation, O-GlcNAc, or ubiquitination will likely form interaction codes, especially when colocated on a protein's single interaction interface. A network-based example of code discovery is given, predicting the yeast protein Npl3p to have a methylation/phosphorylation-dependent interaction code.