The vertebrate brain is so complex that tracing its origins among invertebrates is problematic. It must have evolutionary roots somewhere, but where? On page 289 of this issue, Pani et al. point the way with their finding that acorn worms have gene-expression signatures very similar to those that direct vertebrate brain development (A. M. Pani et al. Nature 483, 289–294; 2012).

Credit: A. M. PANI

The closest relatives of vertebrates are the urochordates (tunicates, or sea squirts) and cephalochordates (the superficially fish-like lancelet). More distant on the evolutionary tree lie the hemichordates, which include acorn worms such as Saccoglossus kowalevskii (pictured), and the echinoderms (including starfish and sea urchins). Together, these groups comprise the deuterostomes, among which the elaborate structural and genomic innovations of vertebrates stand out rather like a cuckoo chick in a nest of willow warblers.

One such feature is the signalling centres — physical regions of a developing embryo in which a specific set of secreted proteins drives the formation of a particular body section. These centres are present in the vertebrate neuroectoderm, an embryonic tissue that gives rise to the brain and nervous system. They are modified or completely absent in tunicates and lancelets, suggesting that many aspects of brain development are unique to vertebrates and arose de novo in the vertebrate lineage.

Pani and colleagues' data contradict this theory. They show that a gene-expression program similar to that in three vertebrate neuroectodermal signalling centres is also present in S. kowalevskii. Although acorn worms do not have anything like a brain, Pani et al. show that these genetic programs are used to pattern (direct the development of) the ectoderm of the embryonic animal. The genes expressed are equivalent to those found in vertebrates, and they are deployed at similar times and places in vertebrate and acorn-worm embryos.

The researchers propose that these signalling centres were part of an ancient gene-regulation 'scaffold' that was present in the common ancestor of all deuterostomes, and that was retained in hemichordates and vertebrates but lost in the evolutionary branches that formed lancelets and urochordates. Such an idea implies that the common ancestor was unexpectedly complicated. This does not mean that it had a complex brain, but that the genetic programs that were eventually modified for use in patterning the vertebrate brain already existed in an ancestral creature that lived perhaps more than 600 million years ago.