SEM picture of Pristionchus pacificus. Courtesy of Jürgen Berger, Max-Planck-Institute für Entwicklungsbiologie, Tübingen, Germany.

A crucial question in evo–devo research is how alterations in gene function result in morphological variation. Plant and animal studies have indicated that the evolution of morphological structures might be caused by mutations in the control regions of regulatory genes, such as transcription factors and their targets. A recent study by Grandien and Sommer lends support to this idea by showing that two orthologous Hox genes in two related nematode species are functionally equivalent, despite having quite divergent coding sequences and clearly distinct developmental roles in the two species.

Nematodes are well suited to comparative morphological studies, as homologous cellular and developmental processes are easily detected in closely related species. A classic example is the development of the nematode vulva — the mating and egg-laying structure — in Caenorhabditis elegans and Pristionchus pacificus. The C. elegans vulva develops from 12 epidermal cells: six form the vulval equivalence group, and descendents of only three of them differentiate into the vulva proper. The Hox gene lin-39 (lineage-39) is required for two steps in vulval development: to establish the vulval equivalence group and, later, for vulval differentiation. Things are quite different in P. pacificus: for example, the vulval equivalence group consists only of three cells; in addition, those cells that do not differentiate into the vulva die, unlike in C. elegans, in which they fuse with the surrounding tissue. The function of lin-39 also differs in P. pacificus — the gene is required to prevent cells from undergoing apoptosis, but is not required later for vulval formation.

What is the genetic basis for the different functions of LIN-39? A predicted MAPK docking and phosphorylation site at the carboxyl terminus of C. elegans (Cel) LIN-39, which is absent from the P. pacificus (Ppa) LIN-39, at first seemed like a convincing candidate, but it was ruled out by functional studies. The lack of any other obvious structural differences between the two LIN-39 proteins gave the first hint that they might be functionally equivalent — a possibility the authors confirmed by transgenic studies. Despite the limited amino-acid similarity between the two proteins, a transgene that contains the Ppa–lin-39 cDNA was found to rescue the egg-laying and vulval cell lineage defects of a Cel–lin-39 mutant but, crucially, only when driven by Cel–lin-39 control regions. The same transgene also restored the presence of the vulval cell neurons, which require lin-39. The conclusion seems to be that the key to the difference between the two proteins lies in their respective control regions.

Follow-up studies will fill in the details, but the message from this paper is that developmental processes can evolve more readily through changes in gene expression than through tinkering with protein function.