The first identification of complete Eph-ephrin signalling in ctenophores and sponges reveals a role for neofunctionalization in the emergence of signalling domains

Animals have a greater diversity of signalling pathways than their unicellular relatives, consistent with the evolution and expansion of these pathways occurring in parallel with the origin of animal multicellularity. However, the genomes of sponges and ctenophores - non-bilaterian basal animals - typically encode no, or far fewer, recognisable signalling ligands compared to bilaterians and cnidarians. For instance, the largest subclass of receptor tyrosine kinases (RTKs) in bilaterians, the Eph receptors (Ephs), are present in sponges and ctenophores, but their cognate ligands, the ephrins, have not yet been detected.Here, we use an iterative HMM analysis to identify for the first time membrane-bound ephrins in sponges and ctenophores. We also expand the number of Eph-receptor subtypes identified in these animals and in cnidarians. Both sequence and structural analyses are consistent with the Eph ligand binding domain (LBD) and the ephrin receptor binding domain (RBD) having evolved via the co-option of ancient galactose-binding (discoidin-domain)-like and monodomain cupredoxin domains, respectively. Although we did not detect a complete Eph-ephrin signalling pathway in closely-related unicellular holozoans or in other non-metazoan eukaryotes, truncated proteins with Eph receptor LBDs and ephrin RBDs are present in some choanoflagellates. Together, these results indicate that Eph-ephrin signalling was present in the last common ancestor of extant metazoans, and perhaps even in the last common ancestor of animals and choanoflagellates. Either scenario pushes the origin of Eph-ephrin signalling back much earlier than previously reported.We propose that the Eph-LBD and ephrin-RBD, which were ancestrally localised in the cytosol, became linked to the extracellular parts of two cell surface proteins before the divergence of sponges and ctenophores from the rest of the animal kingdom. The ephrin-RBD lost the ancestral capacity to bind copper, and the Eph-LBD became linked to an ancient RTK. The identification of divergent ephrin ligands in sponges and ctenophores suggests that these ligands evolve faster than their cognate receptors. As this may be a general phenomena, we propose that the sequence-structure approach used in this study may be usefully applied to other signalling systems where no, or a small number of, ligands have been identified