The ancestry of Antennapedia-like homeobox genes

Abstract

I show that broad evolutionarily significant sub-groups of Antennapedia-like homeobox genes can be distinguished by consideration of 4 amino acids. The absence of proline at position 26 is a synapomorphy of the class; HOX-like homeodomains contain a 19,30 salt-bridge of inverted polarity with respect to the same residues in typical NK-like homeodomains, and residue 28 is highly conserved within but not necessarily between orthologous groups. None of these residues has an obvious role in sequence specific DNA recognition. The EH1 and hexapeptide sequence motifs outside the homeodomain are not well correlated with sub-type. From the discovery of a hexapeptide motif in a sponge NK-like gene, and identification of new instances of longer engrailed-like (EH2) variants of the hexapeptide, I infer that scattered motif distribution is unlikely to be due to convergent evolution, but rather multiple independent loss events. I reconcile these features and the species distribution of current genes to propose a scheme for the ordering of duplication events in the cnidarian-bilaterian stem group.