Evidence for 14 homeobox gene clusters in human genome ancestry.

The arrangement of Hox genes into physical clusters is fundamental to the patterning of animal body plans. Other homeobox genes are often described as dispersed, with only occasional examples of linkage reported, such as the amphioxus ParaHox and Drosophila 93D/E clusters. This clustering is unlikely to be the derived condition, as the genes of the ParaHox and 93D/E clusters are phylogenetically widespread. To assess whether clustering is retained in mammals, and to infer its history, we considered the distribution of ANTP superclass homeobox genes in human and mouse genomes. We postulate four ancient arrays of ANTP superclass genes in animal genomes, denoted 'extended Hox' (Hox, Evx and Mox), NKL (including NK1, NK3, NK4, Lbx, Tlx, Emx, Vax, Hmx, NK6, Msx), ParaHox (Cdx, Xlox, Gsx) and EHGbox (En, HB9, Gbx). Each of these duplicated in the ancestry of the human genome to yield four Hox, four NKL, four ParaHox and at least two EHGbox clusters or arrays. Two of the human NKL clusters (four in mouse) have subsequently been split by chromosome rearrangement, as has one human EHGbox array. We date all cluster duplications to early chordate evolution and infer that three clusters (Hox, NKL, EHGbox) resided on the same chromosome before duplication.