How are cells instructed to form somite boundaries by the zebrafish segmentation clock?

In vertebrate embryos, the elongating body axis is patterned via the sequential and rhyth-mic production of segments from a posterior unsegmented tissue called the presomitic mesoderm (PSM). This process is controlled by a population of cellular oscillators termed the segmentation clock.

Cellular oscillators are locally coupled by Notch signaling and they generate spatiotem-poral waves of gene expression that sweep the tissue from posterior to anterior, reiterating during the formation of each segment. These anteriorly-moving waves are thought to in-teract with a wavefront, regressing posteriorly through the PSM, that records the cellular oscillations as it passes. This interaction transforms a temporally periodic signal into a permanent spatially periodic pattern, which prefigures the location of future somite boundaries. The molecular details of how this interaction occurs remain elusive.

In this work, using the zebrafish as a model, I first studied how the clock's wave pattern and the wavefront interact during the formation of the anterior-most somite boundaries. Using lightsheet microscopy, I imaged the first onset of the segmentation clock using a transgene reporting her1 expression, and another transgene reporting on the dynamics of the wavefront gene tbx6.

I found that Tbx6, oscillates in individual cells and travels as waves of gene expression in the PSM during early somitogenesis. Our results suggest that the her1 wave pattern is re-quired for the generation of Tbx6 waves. Two waves of Tbx6 expression were found to arrest at positions that prefigure the locations of the anterior boundaries of the first and the second somite of the embryo, respectively.

I then addressed the question of what level of synchrony in the wave pattern is required for the formation of correct somite boundaries. In zebrafish Notch pathway mutants, a small number of segments are properly formed in the anterior, followed by defective segments in the posterior. This phenotype can be explained by the fact that, in absence of Notch sig-naling, neighboring noisy oscillators slowly drift out of phase, ultimately resulting in the disruption of the wave pattern and consequently the formation of defective segments. De-fects in somite boundaries are thought to result in malformations of the spine, a condition termed congenital scoliosis in the clinic. Drugs blocking Notch signaling can desynchro-nize the segmentation clock, mimicking Notch mutants. Upon drug washout, the segmen-tation clock can gradually resynchronize and, once the wave pattern is restored, normal segments form again.

We imaged the resynchronization of the segmentation clock in live embryos. We observed that defects in somite boundaries were often spatially restricted to a specific part of the boundary, e.g. dorsal versus ventral. By back-tracking cells from defective boundaries and looking at their Her1 oscillatory traces, we found that the level of synchrony between a cell and its neighbors was spat