Wnt3a came into the picture when the group discovered Wnt-dependent Axin2 expression in the tail bud and presomitic mesoderm (PSM). Axin2 expression was cyclic, but oscillated out of phase with the known oscillation of Notch pathway activity, which is also dependent on Wnt3a activity.
Known signaling pathways provide some plausible circuitry. Axin2 produced by Wnt pathway activity should both turn off the Wnt pathway (via a negative feedback loop) and turn on the Notch pathway (by binding to a negative regulator). The instability of Axin2 would later reverse the situation, leading to a continuous cycle.
The gradient idea arose because Wnt3a is made in the tail bud but can direct Axin2 expression throughout the PSM. High concentrations of Wnt3a near the tail bud prevent differentiation. But as the lengthening embryo puts more distance between the tail bud and the anterior, the level of Wnt3a in the anterior PSM drops. Below a certain threshold of Wnt3a, the Notch pathway can take over and direct somite development.
If this gradient was acting alone, the end result would be a steadily moving front of differentiation. But, as Herrmann points out, “the boundary position has to move back in a periodic manner. For that purpose you need the clock.”
The clock operates only above the threshold level of Wnt3a—below this the cells get stuck in the “Notch on (Wnt off)” state. Above the threshold, all cells cycle together. But each time a new “Wnt on” cycle starts, the cells just below the threshold will not be able to join their more posterior neighbors, and for the first time there will be a boundary between “Wnt on” and “Notch (stuck) on.” This boundary is the key: it defines the division between somites. The boundary only arises when the more posterior, above threshold, region cycles back up into a “Wnt on” state, so it is only laid down periodically. ▪