Arthropod segmentation: Revisiting the Drosophila paradigm

Subdivision of the body into repeated units, or segments, is a key aspect of the organisation of several groups of animals, including vertebrates, annelids and arthropods. Over the last thirty years, the genetic analysis of segmentation in the fruit fly Drosophila has provided detailed insight into a cascade of gene interactions that pattern segment boundaries in the embryo. The study of these gene interactions has provided a textbook example for how genes are regulated during animal development, and revealed key aspects of gene organisation - for example the modular logic of genetic control. Despite this, fundamental aspects of the segmentation process are still not fully understood. By modelling the process of segment pattern formation in silico, we have shown that the current model for Drosophila segment patterning is incomplete. Maturation of the segment pattern, and in particular the transition from double segment ("pair-rule") to single segment patterning, depends on a switch in the nature of regulatory interactions between transcription factors, mediated by an extrinsic timing signal. There is also clear data that the Drosophila segmentation cascade exhibits dynamic behaviour at the level of the gap genes, and that this has consequences for pair rule gene expression. Our models suggest that dynamic behaviour plays a functional role in segment patterning, and that gap gene-mediated stripe movements work in concert with expression dynamics driven by regulatory feedback from within the pair rule gene network itself.

Intriguingly, these observations make it easier to relate the network logic of Drosophila segmentation to processes that have recently been shown to pattern segment formation in other, so called "short germ", arthropods. These species make their segments sequentially, using oscillatory cycles of gene expression, rather than simultaneously, as in Drosophila. Short germ segmentation is clearly ancestral to the insects, and to arthropods as a whole, but up to now little vestige of this ancestral process has been apparent in the Drosophila segmentation cascade. However, our current view of Drosophila segmentation suggests that at least parts of the gene regulatory logic of a pair-rule feedback circuit still persist in the Drosophila embryo, and that they may be responsible for limited, but key aspects of Drosophila segment development. We will examine the dynamic behaviour of the pair rule network in Drosophila, test the extent to which this gene circuit resembles the gene circuits that have recently been shown to drive segmentation in the beetle Tribolium and other short germ arthropods, and so gain insight into the process of gene regulatory network evolution.

The genetic cascade patterning segmentation in Drosophila is one of the best understood gene networks controlling development, but there is still no adequate model for certain key aspects of its function. It has recently become clear that interactions between gap genes drive dynamic change of spatial patterns, and that this dynamic behaviour affects pair-rule gene patterning.
Taking a systems level approach to the existing Drosophila data, we have identified limitations in the current model for segment pattern maturation, which lead us to propose:
(i) that timing of segment pattern maturation is not intrinsically controlled by the interactions between pair rule genes, but is mediated by extrinsic factors that are regulated independently of the segmentation cascade.
(ii) that the regulatory interactions between segmentation genes undergo a key transition at late cellularisation, just before gastrulation, which initiates the process of frequency doubling from pair-rule to single segment patterning, and that this transition may be mediated by the Zic family protein Odd-paired.
(iii) that components of a feedback circuit persist in the early regulatory logic of the Drosophila pair rule genes, and that interactions of this circuit pattern the last parasegment of the germ band, in a manner which may be analogous to the patterning of all abdominal segments in the trunk of short germ insect ancestors.
(iv) that dynamic behaviour of the pair rule network may also play a key role in the specification of the parasegment boundaries of other segments.

We will test these hypotheses using genetic manipulations in Drosophila, including CRISPR/Cas9 genome editing and the MS2 system for live imaging of RNA expression. If correct, these ideas imply that there are more persisting commonalities between the process of segmentation in Drosophila and in its short germ ancestors than are currently apparent.

The principal short and medium term (1-5 year) impacts of this work will be within the academic community, particularly in terms of showing how a system wide approach with in silico modelling can provide fresh insight even in a long studied biological system. It also illustrates the value that comparative biology has in stimulating new ways of thinking about even the best studied systems. We detail in the Pathways to Impact statement how we will communicate this message to professional, student and public audiences.

We cannot predict a direct impact of our work on applied insect biology. However, it is important to stress that insects and other arthropods are the most successful and biodiverse group of animals on earth. They have major economic impact on man through their effects on agriculture and food storage, and they are among the most significant of disease vectors. They are currently being proposed as a valuable food source, and many molecules synthesised by arthropods have practical application, including venoms, silks, elastic proteins and others. Their nervous systems provide practical models for robotic vision and movement control.

There are many well documented areas where research on Drosophila has led to practical applications in species of economic or medical interest. Such areas include the control of sex determination, the mechanism of olfaction, and hormonal control of the life cycle. More broadly, studies in Drosophila have made a major contribution to the whole of biomedical science. The PIs own earlier work on the control of segment determination in Drosophila ultimately contributed to the discovery of Hox genes and the whole family of homeobox transcription factors, although this would certainly not have sounded credible if written into a grant application at the time. The work then seemed to be of purely academic interest.

Similarly, we believe that having a better understanding of the dynamic behavior of gene networks in a model such as Drosophila, with such a wealth of accumulated data and detailed developmental knowledge, is likely to have unforeseen consequences much more widely, both in the applications of systems level thinking to a wide range of problems in biology, and more specifically in understanding how what we know in such detail from Drosophila relates to the biology of the many arthropod species that are of practical concern.

We will communicate our results to the broadest possible community, through presentations at academic meetings, through review articles, and through outreach events such as those organized during the Cambridge Science Festival, and through the Cambridge Museum of Zoology. We will also use our results to update the teaching of science undergraduates, stressing the value of explicit modeling in challenging existing orthodoxy and generating and testing novel hypotheses.