Gene duplication, co-option and recruitment during the origin of the vertebrate brain from the invertebrate chordate brain.

The brain of the basal chordate amphioxus has been compared to the vertebrate diencephalic forebrain, midbrain, hindbrain and spinal cord on the basis of the cell architecture from serial electron micrographs and patterns of developmental gene expression. In addition, genes specifying the neural plate and neural plate border as well as Gbx and Otx, that position the midbrain/hindbrain boundary (MHB), are expressed in comparable patterns in amphioxus and vertebrates. However, migratory neural crest is lacking in amphioxus, and although it has homologs of the genes that specify neural crest, they are not expressed at the edges of the amphioxus neural plate. Similarly, amphioxus has the genes that specify organizer properties of the MHB, but they are not expressed at the Gbx/Otx boundary as in vertebrates. Thus, the genetic machinery that created migratory neural crest and an MHB organizer was present in the ancestral chordate, but only co-opted for these new roles in vertebrates. Analyses with the amphioxus genome project strongly support the idea of two rounds of whole genome duplication with subsequent gene losses in the vertebrate lineage. Duplicates of developmental genes were preferentially retained. Although some genes apparently acquired roles in neural crest prior to these genome duplications, other key genes (e.g., FoxD3 in neural crest and Wnt1 at the MHB) were recruited into the respective gene networks after one or both genome duplications, suggesting that such an expansion of the genetic toolkit was critical for the evolution of these structures. The toolkit has also increased by alternative splicing. Contrary to the general rule, for at least one gene family with key roles in neural crest and the MHB, namely Pax genes, alternative splicing has not decreased subsequent to gene duplication. Thus, vertebrates have a much larger number of proteins available for mediating new functions in these tissues. The creation of new splice forms typically changes protein structure more than evolution of the protein after gene duplication. The functions of particular isoforms of key proteins expressed at the MHB and in neural crest have only just begun to be studied. Their roles in modulating gene networks may turn out to rival gene duplication for facilitating the evolution of structures such as neural crest and the MHB.