Protein-protein recognition occurs via interacting domains. Protein domains are conserved components within proteins, usually folding independently of other parts of the protein. This property allows modularity in the protein domain architecture. In most cases domains have a distinct function and many domains mediate protein-protein interactions. Combining the modular nature of domain rearrangement and the relative independence of their functionality allows proteins to acquire additional roles by the acquisition of new domains. Indeed, organisms tend to manipulate existing domain architectures to create new proteins, rather than creating new architectures ab initio. The acquired domain can mediate a new protein-protein interaction and therefore increase the protein connectivity in the protein-protein interaction network. This increased connectivity within and between the biological systems of an organism frequently enhances its complexity. Thus, the rate of domain rearrangements rises with the increase in organisms’ complexity during evolution. Here we survey various ways to create new protein domain architectures and the impact of creating new proteins by domain rearrangements on protein connectivity and organisms’ complexity. Within this framework, we highlight the role of domains as mediating protein-protein interactions.