A comparative gene circuit approach to study the evolution of segmentation in insects

Why are certain structures, such as jointed legs, very frequently observed in animals while others, such as the wheel, are not? More generally, why do organisms adopt certain types of solution to the pressures of natural selection, and not others? The answer to this question lies in the complex interplay between the processes of development and evolution. Selection works on variation that arises from mutation in the genome, but the effect of this selection acts on the actual characteristics of the organism that arise during the processes of embryonic and juvenile development. Thus, to understand the link between how genes vary, and how organisms evolve, we need to understand the complex processes that interpret the structure of genes to generate the form of an organism. Unfortunately there is no simple one-to-one relationship between a mutation and its effect on an adult character. Changing a single base pair of DNA in the genome of an animal can have very drastic and unexpected effects. For example, a single mutation of the fruit fly can double the number of its wings from two to four. Such effects are hard to predict because of the large number of factors involved in development, and the non-linear interactions among them, which occur at many levels of organisation (genes, proteins, cells, tissues). This means that, despite the massive and rapidly increasing amount of data that we have on genome sequences, and on the molecular mechanisms of development, we still lack a coherent view linking the structure of the genome to the final outcome of development for even the most simple developmental processes. The interactions involved are too complicated for analysis by the human brain alone, and are sometimes counterintuitive. We therefore need the help of computers to keep track of many interactions simultaneously. In this application we exploit modern experimental techniques for monitoring where and when genes synthesise their protein products, together with computer technology, to simulate and analyse a specific developmental process in great detail, at the level of gene regulatory interactions. One process that is simple enough to be studied by these new methods is the early development of insects such as flies, midges and mosquitoes, and specifically the laying down of a repeated pattern of segments during their early stages of embryogenesis. We propose to extend computer models of this segment determination process in the well studied fruit fly Drosophila melanogaster, and to generate equivalent models for two other species of fly, including the malaria mosquito Anopheles, the genome of which has recently been sequenced. This will help us to understand not only how segments arise in these different flies, but also the more general question of how gene networks evolve to create new developmental processes, and through these, the diversity of organisms that we observe in nature.

Over the last ten years, methods have been developed for quantification of gene expression data obtained from fluorescent antibody staining or in situ hybridisation. These methods have enabled a description of the spatial and temporal pattern of segmentation gene expression in wild-type blastoderm stage embryos of Drosophila melanogaster at unprecedented accuracy and temporal resolution. These data have been used for inferring regulatory interactions between the gap class of segmentation genes. This is achieved by fitting models, called gene circuits, to the quantitative expression data by an optimisation method known as simulated annealing. The fitting process, which makes no use of prior knowledge on gene interactions, successfully reconstructs many of the known interactions between these genes, and predicts others that are plausible. Moreover, gap gene circuits have been used to describe dynamic shifts in the position of gene expression domains in the posterior region of the blastoderm. Finally, gene circuit models allow for detailed dynamical analysis of specific regulatory interactions and their effects on gene expression in the wild-type gap gene network. Working with D. melanogaster, we propose to: - Extend the existing quantitative data set for D. melanogaster by collecting data for the product of the gene huckebein (hkb), which acts in the posterior part of the embryo, and by collecting data for a set of five segmentation genes (caudal (cad), Kruppel (Kr), knirps (kni), giant (gt) and even-skipped (eve)) in tailless (tll) mutant embryos. These embryos show a disrupted fate map in which dynamic shifts in spatial patterning are much reduced. - Improve the understanding of dynamic gene regulation in D. melanogaster by simulating mutant expression patterns in tll embryos, and by incorporating transcriptional and translational delays into gene circuit models. To begin a comparative analysis of these gene networks, we will collect data on gap gene expression for two basal dipteran species that show altered dynamic patterning of segmentation. These data will allow us to reconstruct the gap gene networks in these two species, and hence, to compare patterning networks in lower and higher Diptera. To achieve this we will: - Establish the exact timing of cleavage cycles and nuclear divisions in blastoderm stage embryos of the mosquito Anopheles gambiae and moth midge Clogmia albipunctata. - Make use of genomic resources for A. gambiae to generate antibodies against the products of A. gambiae orthologues of eight D. melanogaster segmentation genes (cad, hunchback (hb), Kr, kni, gt, tll, hkb, eve). - Clone gap gene orthologues from C. albipunctata and make antibodies for as many of the same eight gene products as possible. - Collect quantitative protein expression data for these eight genes at an adequate temporal resolution (to be determined empirically) from blastoderm embryos of A. gambiae and C. albipunctata. Analyse these data to characterise dynamic domain shifts and variational properties of domain boundary positioning. - Infer sets of regulatory interactions among maternal and gap genes by fitting of delay gene circuit models to these data. Analyze models to characterise developmental mechanisms involved in gap gene expression in A. gambiae and C. albipunctata. - Perform a comparative analysis to characterise conserved and divergent aspects of gap gene regulation between D. melanogaster, A. gambiae and C. albipunctata. Predict intermediate evolutionary stages by exploration of the parameter space of gap gene circuit models.