The signalling molecule Sonic hedgehog (Shh) has been investigated so extensively over the past few years that it is a wonder that there is anything left to discover about its properties. However, as Charron and colleagues reveal in a report in Cell, it still has a few surprises up its sleeve. In addition to its classic role as a morphogen, it can also act as an axon-guidance cue.

In the vertebrate embryo, a group of neurons that originate in the dorsal spinal cord send axonal projections to the ventral midline (or floor plate), where they cross to the other side to generate commissural tracts. Netrin-1 has been identified as an essential chemoattractant that guides these axons towards the ventral midline, but other factors are probably also required.

Shh is expressed in the floor plate of the vertebrate neural tube, and can act at a long range from its source, so it was a potential candidate. Indeed, Charron et al. showed that Shh, like Netrin-1, can reorientate commissural axons in vitro. They also showed that the growth cones of frog spinal axons changed their course and grew towards a source of Shh, further supporting the idea that Shh acts directly as a chemoattractant.

By blocking the function of the Shh receptor Smoothened (Smo) using the specific inhibitor cyclopamine, the authors also obtained insights into the molecular mechanism that underlies the chemoattractant properties of Shh. They found that, at least in vitro, cyclopamine prevented commissural axons from turning towards a source of Shh, but did not affect their response to Netrin-1. To find out whether Shh is required in vivo for commissural axon guidance, they selectively disrupted Smo function in commissural axons. Consistent with Shh acting as a midline chemoattractant, this resulted in defects in commissural axon guidance. Importantly, this also implies that the commissural neurons have a cell-autonomous requirement for Shh signalling.

So, Charron et al. provide compelling evidence that Shh, acting through Smo, functions as an axon-guidance cue at the floor plate of the developing spinal cord. Unlike Netrin-1, it cannot promote axonal outgrowth, but it seems to be an essential component of the molecular mechanism that mediates attraction of axons to the ventral midline. These findings should stimulate investigations into whether other morphogens can 'double up' as axon-guidance molecules once their patterning work is done.