Post-translational modifications can alter protein structure and facilitate/inhibit interactions with other proteins. The addition of one type of modification may also serve to inhibit or promote the attachment of another type. The addition of post-translational modifications to a protein allows it to expand its interaction repertoire whilst retaining genome size and does not require genomic, transcriptomic and translational regulation.

The best-characterized example of this is the ‘histone code’, where combinations of methylation, acetylation, ubiquitylation and phosphorylation recruit specific binding motifs that can ‘read’ these codes and result in distinct gene expression patterns. In recent years, as the tools to investigate and detect post-translational modifications have improved, evidence of other types of ‘interaction codes’ are emerging: the ‘transcription factor code’, the ‘chaperone code’ and the ‘tubulin code’, where combinations of modifications regulate the functions and interactions of these proteins. It is reasonable to expect the cell may utilize such codes as a general means to regulate protein-protein interactions and we propose this notion to be termed the 'interaction code'.

The examination of protein-protein interaction networks integrated with information on its post-translational modification state and the enzymes responsible for its post-translational modification should enable the discovery of proteins with such interaction codes.

We have examined the yeast protein-protein interaction network and identified the
RNA-binding protein Npl3, as a candidate protein to use an interaction code to modulate its interactions with its partner proteins [1]. The C-terminal domain of Npl3, is known to be phosphorylated and methylated. It is a hub protein with multiple interaction partners. It has a single interaction interface and its interaction with its partners appear to be modulated by the post-translational modification state of its C-terminal domain. Through conditional two-hybrid analysis [2], a technique we have developed, we show that an interaction code does indeed exist for Npl3. Arginine methylation of Npl3 has a positive effect [3], whilst serine phosphorylation has an inhibitory effect on the interaction of Npl3 with its binding partners.