Have you ever considered the paradox of how a tightly bound link between two surfaces can be both very secure and reversible? Probably not, I guess. But the people at the Velcro companies did and came up with the idea for ‘loop and hook’ fasteners that have found so many applications in recent years — so much so that the name of the company has become a generic term for this type of fastening. A similar, somewhat paradoxical, issue exists in the field of axon guidance that seeks to uncover the mechanisms leading to the precise spatial patterning of neural connections. Here, it is thought that repellent signalling, one of the basic mechanisms of axon guidance, is mediated by direct cell–cell contact. So how does cell adhesion lead to repulsion and how is this mechanism controlled so that repulsion is transient rather than permanent?

An answer is provided by two studies published in Science. In the first study Hattori, Osterfield and Flanagan describe how contact-mediated axon repulsion is achieved by an interaction between the metalloprotease Kuzbanian, and both the axon guidance molecule ephrin and the Eph receptors. Ephrin ligands are membrane-bound proteins that bind tightly to their respective tyrosine kinase receptors of the Eph family. The growth cone of an extending process contains Eph receptors that bind to ephrin ligands, located on the surfaces of other cells and this promotes adhesion of the two cells. Hattori and colleagues showed that ephrin-A2 forms a stable complex with Kuzbanian in the absence of Eph receptors. However, once the ephrin ligand binds to the Eph receptor, Kuzbanian is activated and cleaves ephrin-A2 from the cell surface in a localized reaction. A key experiment showed that a mutation of ephrin-A2 that blocked ephrin cleavage did not prevent growth-cone collapse but did delay axon withdrawal.

The importance of metalloproteases in axon guidance was further emphasized by Galko and Tessier-Lavigne. These authors showed that the effects of another guidance molecule — the axonal chemoattractant, netrin-1 — could be potentiated by metalloprotease inhibition. This effect may result from stabilization of the netrin receptor DCC (Deleted in Colorectal Cancer) on the axon.

These two papers show that metalloprotease proteolytic activity is a critical regulator of axon guidance and that, when it comes to finding new solutions to sticky problems, nature often finds an elegant solution.