Developmental Genetics

The Developmental Genetics Laboratory examines the molecular processes that underlie animal development, in particular, the mechanisms used to generate the complex architecture of an organism. We are particularly interested in the process of segmentation - the formation of repeated units along the anteroposterior axis. Although developmental regulators and pathways have been highly conserved during evolution, the mechanism of segmentation differs considerably in the two systems we study - Drosophila and vertebrates. In Drosophila, segmentation occurs while the embryo is a single, multinucleate cell. In vertebrates and most other animals, segmentation occurs in a growing, multinucleate embryo, and requires neighbouring cells to signal between each other in order to coordinate their behaviours.

In Drosophila, we are studying how molecular motors contribute to one mechanisms for establishing organismal and tissue asymmetry: the localisation of mRNAs and other cargoes within cells. These studies provide insights into how the specificity of intracellular motors is determined, and how motor activity is regulated during patterning and in other developmental contexts. In vertebrates, we are analysing the circuitry and targets of the segmentation clock, the molecular oscillator that drives formation of successive intersegmental boundaries. We are also studying how stem/progenitor cells drive axial growth of the embryo.

DNA damage induces nuclear transport in Drosophila embryos

The first stages of Drosophila embryogenesis comprises 13 nuclear divisions that occur every 9-13 min, but in the absence of associated cytokinesis. This process generates a multinucleate syncytium, the single-celled blastoderm embryo, in which up to 6.000 nuclei are arranged as a layer around the cortex. Invaginating membranes then enclose each individual nucleus to form the somatic cells of the future larva and adult.

The rapidity of nuclear cleavage precludes most forms of DNA repair and checkpoint controls, and so the embryo has adopted a novel method of dealing with the consequences of mitotic disruptions. Damaged nuclei are excluded from the future organism by internalisation into the yolk region of the embryo. We have been examining this form of organelle transport in vivo by following the motion of internalising nuclei labelled with fluorescent histones (Histone-GFP or -mRFP).

We find that damage-induced internalisation is due to transport along microtubules (MTs): depolymerisation of MTs blocks internalisation, and inactivation of the blocking agent allows the resumption of transport (Figure 1a). Because MTs are orientated in blastoderm embryos with their minus-ends apical, one might expect nuclear internalisation to involve plus-end-directed transport. Our results indicate that this is indeed the case, and that internalisation is driven by the Kinesin-1 motor complex: timelapse imaging shows that internalisation is compromised in embryos mutant for components of this complex.

We are also investigating the signalling pathways that trigger internalisation. We have confirmed previous reports that DNA damage-induced internalisation is completely dependent on the activity of the chk2 check point gene, and are investigating (maternal) requirements for the upstream Rad9-Rad1-Hus1 (9-1-1) complex that is directly involved in sensing DNA damage. Internalisation is only delayed in hus1^- embryos, implying that multiple pathways are involved in detecting DNA damage.

lunatic fringe during segmentation

During vertebrate segmentation, embryos generate bilateral pairs of new somites at regular intervals. In the mouse, 65 somite-pairs are generated, one every two hours. This repetitive process is controlled by the segmentation clock, whose action is revealed by oscillatory transcription of various genes in the presomitic mesoderm (PSM), the precursor tissue of the somites. The molecular composition of the clock is unclear although several pathways, including Notch, Wnt and Fgf signalling, have been implicated in regulating cyclic gene expression.

We have been analysing the role of lunatic fringe (lfng), a gene whose expression is activated by Notch signalling, and which encodes a potential inhibitor of Notch signalling. lfng transcription occurs cyclically in the posterior PSM, and non-cyclically in a stripe at the anterior of the PSM where new somites are generated.

The relative importance of these two domains for segmentation is not known. Negative feedback of lfng activity could contribute to the segmentation clock in the posterior PSM. The anterior lfng stripe might have a role in generating the boundary that defines the new, forming somite.

To analyse the individual functions of each lfng domain, we made use of a promoter element that drives expression in only the anterior stripe domain. We generated transgenic mice in which this is the only lfng expression domain, i.e. that lack posterior cycling expression, and examined whether such selective lfng expression can rescue segmentation.

Unexpectedly, we find that requirements for oscillating lfng vary along the body axis. Anterior somites (numbers 1-30) require the cycling lfng activity for normal somite formation, but more posterior somites require only the striped expression. We also find that a small domain (somites 31-34) forms normal somites independent of lfng activity (Figure 1b). These results indicate that the circuitry defining somite boundary formation is modified during the course of axis elongation.

For a list of refereed research papers, see Publications (in navigation on left).