The discovery of Hox gene clusters, first in Drosophila (a protostome) and then as homologues in vertebrates (deuterostomes), was a major step in our understanding of both developmental and evolutionary biology. Hox genes in both species perform the same overall function: that is, organization of the body along its head-tail axis. The conclusion is that the protostome-deuterostome ancestor, founder of 99% of all described animal species, must already have had this same basic Hox cluster, and that it probably used it in the same way to establish its body plan. A striking feature of Hox genes is the spatial collinearity rule: that order of the genes along the chromosome corresponds with the order of their expression domains along the embryo. For vertebrates, though not Drosophila, there is also the temporal collinearity rule: that order of genes along the chromosome corresponds with timing of Hox expressions in the embryo. Although Hox genes are clearly recognized in pre-bilaterians (Cnidaria), it is only in bilaterians that the characteristic clustered Hox arrangement and function is commonly found. Spatial collinearity in expression is conserved widely throughout Bilateria but temporal collinearity is so far limited to vertebrates, cephalochordates, and some arthropods and annelids. In addition to conserved use of Hox genes to pattern the head-tail axis, some animal groups, particularly lophotrochozoans, have extensively co-opted Hox genes, outside collinearity rules, to regulate development of novel structures. Satisfactory understanding of Hox cluster function requires better understanding of the bilaterian last common ancestor (Urbilateria). Xenacoelomorpha may provide useful living models of the ancestral bilaterian condition.

Arthropods and vertebrates inherited their Hox clusters from an ancestral cluster of at least six genes already present in their last common ancestor, Urbilateria. Clustering and a common transcriptional direction are both likely features of the way that the gene complex first arose in a process of tandem gene duplication. Spatial collinearity (correspondence between ordering of Hox genes along the chromosome and their expression patterns along the head-tail axis) has been conserved in many animal groups and is likely to have been already present in Urbilateria. It is not known why the Hox cluster evolved with spatial collinearity. Four models are discussed. These vary in the significance they place upon Hox chromatin structure, and also on whether they propose that collinearity is primarily concerned with establishment or maintenance of Hox expression. Published proposals to explain spatial collinearity, which invoke enhancer sharing, chromatin closing or chromatin opening, are either problematic or can offer only partial explanations. In an alternative proposal it is suggested here that spatial collinearity evolved principally to maximise physical segregation, and thereby minimise incidence of boundaries, between active and inactive genes within the Hox cluster. This is to minimise erroneous transfer of transcriptional activity, or inactivity, between adjacent Hox genes.

The Hox gene cluster, responsible for patterning of the head–tail axis, is an ancestral feature of all bilaterally symmetrical animals (the Bilateria) that remains intact in a wide range of species. We can say that the Hox cluster evolved successfully only once since it is commonly the same in all groups, with labial-like genes at one end of the cluster expressed in the anterior embryo, and Abd-B-like genes at the other end of the cluster expressed posteriorly. This review attempts to make sense of the Hox gene cluster and to address the following questions. How did the Hox cluster form in the protostome-deuterostome last common ancestor, and why was this with a particular head–tail polarity? Why is gene clustering usually maintained? Why is there collinearity between the order of genes along the cluster and the positions of their expressions along the embryo? Why do the Hox gene expression domains overlap along the embryo? Why have vertebrates duplicated the Hox cluster? Why do Hox gene knockouts typically result in anterior homeotic transformations? How do animals adapt their Hox clusters to evolve new structural patterns along the head–tail axis?

How have animals evolved new body designs (morphological evolution)? This requires explanations both for simple morphological changes, such as differences in pigmentation and hair patterns between different Drosophila populations and species, and also for more complex changes, such as differences in the forelimbs of mice and bats, and the necks of amphibians and reptiles. The genetic changes and pathways involved in these evolutionary steps require identification. Many, though not all, of these events occur by changes in cis-regulatory (enhancer) elements within developmental genes. Enhancers are modular, each affecting expression in only one or a few tissues. Therefore it is possible to add, remove or alter an enhancer without producing changes in multiple tissues, and thereby avoid widespread (pleiotropic) deleterious effects. Ideally, for a given step in morphological evolution it is necessary to identify (i) the change in phenotype, (ii) the changes in gene expression, (iii) the DNA region, enhancer or otherwise, affected, (iv) the mutation involved, (v) the nature of the transcription or other factors that bind to this site. In practice these data are incomplete for most of the published studies upon morphological evolution. Here, the investigations are categorized according to how far these analyses have proceeded.

Developing anatomy along the head-tail axis of bilaterian embryos is specified, to a large extent, by the overlapping patterns of expression of the Hox genes. Hox gene enhancers respond to a variety of signals in order to regulate these discreet domains of expression. For mouse Hoxc8, the 399bp "early enhancer" plays a major role. Activation of this enhancer is now examined using luciferase expression constructs transfected into HepG2 cells. Constructs are activated by the combined actions of Gdf11/Smad and Cdx protein signalling pathways, both of which are functional in early embryos. Each of these pathways alone has little stimulatory effect. Stimulation by the two pathways together exceeds the sum of the effects of each pathway alone, indicating synergistic activity. By mutation analysis, two Smad binding motifs are identified as mediators of the Gdf11 effect and two Cdx binding motifs mediate the Cdx effect. The two Smad motifs and one of the Cdx sites are conserved from fish to mammals. Gdf11 stimulation is partially inhibited by Specific Inhibitor of Smad3, suggesting that Smad3 plays a part in signal transduction. Fgf2 increases luciferase activation by the Hoxc8 enhancer, but not, apparently, by specific interactions with either Gdf11 or Cdx effects.