Hox genes are metazoan pattern forming genes—genes that are universally associated with defining the identities of regions of the body. There are multiple Hox genes present, and one of their unusual properties is that they are clustered and expressed colinearly. That is, they are found in ordered groups on the chromosome, and that the gene on one end is typically turned on first and expressed at the head end of the embryo, the next gene in order is turned on slightly later and expressed further back, and so on in sequence. That the tidy sequential order on the chromosome is associated with an equally tidy spatial and temporal pattern of expression in the body has always been one of the more fascinating aspects of these genes, and they are one of the few cases where we see an echo of phenotypic form comprehensibly laid out in the DNA.

However, there are some exceptions to the tidy clustering, and they occur right in two animals that have been central to developmental/genetic research, Caenorhabditis elegans and Drosophila melanogaster. These animals have broken clusters. Almost everywhere else, the Hox genes are ordered in one place, but in two of the most common research organisms, they’ve been split apart into two groups…so what’s going on? We have what looks a little bit like a universal rule in genetic organization, and then it gets violated with seemingly little consequence. How do worms and flies get away with it?

One way to find out is to look for more exceptions to Hox ordering, and here’s a doozy: an animal, the tunicate Oikopleura dioica, has blown its Hox gene clusters to flinders and scattered the individual Hox genes all over its genome, with no detectable linkage between them.

The authors studied the genomic organization of this tiny, rapidly-developing tunicate and could find no trace of clustering of the nine Hox genes—there was no linkage, and many commonly expressed non-Hox genes are scattered between them. Nine genes, and nine clusters. Another ascidian, Ciona intestinalis has its Hox cluster split apart into 5 chunks. In addition, O. dioica has jettisoned all of its middle Hox genes, retaining only 3 anterior genes (Hox 1, 2, and 4) and a full complement of 6 posterior (or tail) genes (Hox 9A, 9B, 10, 11, 12, and 13). It’s something of a surprise to see all those Hox genes floating free, uncoupled from one another.

The next surprise, though, is that despite losing all genomic order, they are still expressed in the expected anterior-posterior order in the tissues.

The schematic organization of each tissue is drawn in blue, from the posterior end of the trunk (left) to the tail tip (right). Hox gene expression domains are represented in red for anterior genes and in green for posterior genes.

This diagram of the expression domains of the O. dioica Hox genes is a bit complicated, because the domains differ in different tissues. The key point, though, is that in each tissue the order, from anterior at the left to posterior at the right, is still in the same sequence we would see in chordates with unfractured clusters—we don’t see Hox13 expressed in front of Hox4, for example.

Does breaking up the cluster have any effect? Yes, the genes have lost their temporal ordering. O. dioica is an organism that develops very, very rapidly, completing a life cycle in 4 days. It also follows a determinative pattern of early development; that is, it is mosaic with a fixed pattern of cleavages and cell fate assignments, instead of the more flexible (or sloppier) pattern other chordates, like us, follow.

What about those other model organisms, C. elegans and Drosophila, that also have broken clusters? They are also animals characterized by very rapid and tightly determined patterns of early development. What this suggests is that the arrangement of Hox genes in a cluster is important for the order of development, and that organisms that have so compressed the timing of development that order is irrelevant, or use maternal determinants to define spatial position, have lost any constraint on clustering and the groups of genes can begin to break up.

Discrete changes of Hox gene complements in chordates. The chordate ancestor gained a rich set of posterior genes, which were inherited in the three subphyla but partly lost in ascidians. Central genes were gradually lost in tunicates, with larvaceans keeping anterior and posterior genes only. Whereas the Hox cluster was multiplied in vertebrates (with subsequent losses of a few paralogues in some clusters), the cluster degenerated in tunicates, and ultimately disappeared in larvaceans. The loss of central genes and of the Hox cluster coincides with a partition of Hox expression domains, which largely overlap in cephalochordates and vertebrates (ascidian data are still lacking). The motor of both events might be the decrease in size and transition to determinative development.

What this means for our cousin chordates is that lineages that have evolved towards small size, rapid development, and fixed or determinative mechanisms of development, have shed any requirement for the temporal sequence of expression conferred on Hox genes by their linear order, and have dismantled the clusters to varying degrees.

Seo H-C, Edvardsen RB, Maeland AD, Bjordal M, Jensen MF, Hansen A, Flaat M, Weissenbach J, Lehrach H, Wincker P, Reinhardt R, Chourrout D (2004) Hox cluster disintegration with persistent anterioposterior order of expression in Oikopleura dioica. Nature 431:67-71.