Order: it's in the genes

Thirty years ago, men were men and flies were flies and the two species were thought to share no more than a handful of genes and a partiality to bananas. Then came the discovery that so-called 'homeotic genes' controlled axial patterning, and that their sequence and function was conserved between distant organisms. The unexpected realization that vertebrates and invertebrates were evolutionarily related through such an essential developmental process altered the course of developmental genetics — it turned flies into authorized models for animal development and kickstarted the field of vertebrate molecular embryology.

It was Bateson who got the ball rolling. In 1894, he described natural variants in which body segments had acquired the identity of other regions. The phenomenon, which he called 'homeosis', was shown by T. H. Morgan to be heritable in flies. But it wasn't until Ed Lewis put his mind to it 50 years later that the underlying genetics began to be unravelled. By examining the genotype–phenotype relationship of homeotic mutations at the fly bithorax cluster of genes (BX-C) — for example, those that transformed the third thoracic segment into the second, giving the flies two sets of wings — he reported that the BX-C consisted of distinct genetic elements and that mutations map in an order that corresponds to the anterior–posterior (A–P) order of the body parts that they affect. This feature, called spatial collinearity, would turn out to be a defining feature of both vertebrate and invertebrate homeotic genes. Lewis showed remarkable vision by arguing that the identity of an individual body segment is produced by the particular combination of BX-C genes, and that these were activated in reponse to an A–P gradient.

On the molecular biology front, techniques were speeding up. Two teams, one led by Matthew Scott and the other including William McGinnis, Michael Levine and Walther Gehring, showed in 1984 that homeotic genes share a 184-bp sequence called the 'homeobox'. This is translated into a 60-amino-acid DNA-binding motif — the homeodomain — by which homeobox (or Hox) genes bind to DNA and control transcription. It was only a short time before the homeobox was used in homology searches to pull out more homeotic genes, in flies and farther afield: frogs, chickens, zebrafish, mice and humans all had Hox genes that, like those of flies, were arranged in clusters. In 1989, Duboule and Dollé and the Krumlauf laboratory showed by in situ hydridization that mouse and fly Hox clusters showed a similar spatial and functional organization. Not only did genes in vertebrate and invertebrate clusters have spatial collinearity, but they were expressed in a temporal order that matched their physical order on the chromosome (that is, they showed 'temporal collinearity'). In addition, fly Hox genes — which map to two adjacent clusters, BX-C and ANT-C — could, by and large, be matched up with genes at similar positions on each of the four paralogous vertebrate clusters. This was a remarkable breakthrough and a vindication of Lewis' belief that all homeotic genes are evolutionarily related.

Even before the publication of Lewis' Nobel-prize winning paper, homeotic phenotypes had led to some surprisingly accurate predictions about how homeotic 'selector genes', which Garcia-Bellido insightfully named and described, might specify the development of groups of cells. Three decades on, we have an astonishing amount of information on the genes that pattern the axes and segments of an animal — although precisely how Hox genes combine to produce the numerous, yet stereotypical, morphological patterns that we see still escapes us.