A melange of chemoattractants and repellants guides growing axons to their targets. Axons that cross the midline of the spinal cord on their way toward the nascent brain, however, have the complicated task of having to move past local attractants at the midline to complete their journey.
Midline cells also produce repulsive molecules called Slits; but by keeping levels of the Slit receptors Robo1 and 2 low, axons approaching the midline can ignore Slits and instead respond only to attractants. As axons cross the midline, they up-regulate Robo1 and 2, become more sensitive to the repellants, and so are booted out the other side. Precrossing axons also express Robo3, which silences Robo1 and 2. But, curiously, Robo3 is also found on postcrossing axons, when Robo-mediated repulsion is at its strongest. “That was a puzzle,” Tessier-Lavigne says. And it's that puzzle they've resolved in the current study.
While sequencing a collection of Robo3 cDNAs, the authors discovered two isoforms, one with and one without the two terminal exons. “We fell off our chairs when we did the immunohistochemistry,” Tessier-Lavigne says. Robo3.1 was expressed on the precrossing axonal segment, whereas Robo3.2 was found on the postcrossing segment. Overexpression of 3.1 facilitated crossing but led many axons to recross. Robo3.2 overexpression prevented most axons from crossing all together, but those that did never recrossed.
The results suggest that unlike Robo3.1, Robo3.2 is not a silencer of the Slit receptors, but instead cooperates with them. “Our hypothesis is that Robo3.1 and 3.2 are tied together to give you an all-or-nothing switch,” says Tessier-Lavigne. The mechanisms controlling this switch are yet to be determined.