Extracellular protein interactions are crucial to the development of multicellular organisms because they initiate signalling pathways and enable cellular recognition cues. Despite their importance, extracellular protein interactions are often under-represented in large-scale protein interaction datasets because most high-throughput assays are not designed to detect low affinity extracellular interactions. Due to the lack of a comprehensive dataset, the evolution of extracellular signalling pathways has remained largely a mystery. We investigate this question using a combined dataset of physical pairwise interactions between zebrafish extracellular proteins, mainly from the IgSF (immunoglobulin) and LRR (leucine-rich repeat) families, and their spatiotemporal expression profiles. We took advantage of known homology between proteins to estimate the relative rates of changes of four parameters after gene duplication, namely extracellular protein interaction, expression pattern, and the divergence of extracellular and intracellular protein sequences. We show that change in expression profile is a major contributor to the evolution of signalling pathways, followed by divergence in intracellular protein sequence; whereas extracellular sequence and interaction profiles are relatively more conserved. Rapidly evolving expression profiles will eventually drive other parameters to diverge more quickly because differentially expressed proteins get exposed to different environments and potential binding ligands. This allows homologous extracellular receptors to attain specialised functions and become specific to tissues and/or developmental stages.