The secreted signalling molecule Hedgehog (Hh) was so baptized in Drosophila because of the effects of its mutation, which turns the ventral surface of the embryo into a lawn of spiky denticles. Hh was subsequently found to belong to a conserved family of secreted signalling molecules, required for growth and patterning during animal development. Cellular responses to Hh are controlled by two transmembrane molecules: Patched (Ptc), a multi-pass membrane protein that binds Hh; and Smoothened (Smo), a seven-pass protein that relays the signal to nuclear effectors. When Hh is not around, Ptc prevents constitutive signalling by Smo — a repression that is relieved when Hh binds to Ptc.

Two recent papers in Cell and Molecular Cell report a new mechanism whereby these proteins regulate the activity of one another post-translationally.

All cells known to respond to the Hh signal express high levels of Smo. Together, both papers show that Ptc causes the levels of Smo protein to go down, unless Hh is present to make them go up. Both effects happen after Smo is transcribed, because Smo transgenes are regulated in the same way as the endogenous gene and Smo regulation is independent of cubitus interruptus (ci), the transcription factor through which the Hh signal is transduced.

What is the post-transcriptional change and how does it affect Smo? Drosophila cells in culture that have been bathed in Hh accumulate a slower migrating, phosphorylated form of Smo — an effect that is mimicked by removing Ptc. Interestingly, as the proportion of phosphorylated Smo rises, levels of Ptc protein begin to drop.

Previous experiments on mammalian homologues had indicated that Ptc and Smo exist as a preformed complex at the membrane; and the binding of Hh to Ptc was presumed to cause the conformational change that nudged Smo into signalling mode. Denef et al. now offer an alternative view: binding of Hh to Ptc induces the internalization of Ptc through intracellular vesicles (probably doomed for the lysosome), removing it from the cell surface, and causes an concomitant accumulation of active, phosphorylated Smo at the plasma membrane. So it seems that Hh treatment can prevent a Patched-dependent alteration in the post-translational modification of Smo.

Alcedo et al. would agree. They propose a ‘self-correcting’ model that imposes a strict interdependence between the presence of Hh and the activities of Smo and Ptc. In their view, Hh signalling has two consequences: the stabilization of Smo protein (independently of Ci), and the transcriptional activation of Ptc that, by silencing Smo, makes signalling entirely dependent on the presence of the ligand. The system is self-correcting because any imbalance is readjusted to an equilibrium that resists small perturbations, where Ptc inhibits Smo as rapidly as possible when Hh is not there.

Because mutations in both Ptc and Smo have been linked to basal cell carcinoma, unravelling the relationship between these molecules has biological implications that extend beyond development. The careful balance of the activities of Hh, Smo and Ptc is therefore needed both to ensure correct development and to circumvent cancer.