# Department of Biochemistry, Imperial College, London, SW7 2AZ and § Department of Cell & Molecular Biology, Howard Hughes Medical Institute, Life Sciences Addition #519, University of California, Berkeley CA 94720 USA.
A functional nervous system depends upon the organization and precise connections of a large number of neurons. At least three major mechanisms appear to be operating to guide growth cones in the developing embryo towards their correct target: differential substrate affinities, chemotropism and chemorepulsion. However, the molecular mechanisms by which the axonal projections of neurons can distinguish between a vast number of substrates to find and recognize their targets with great accuracy are poorly understood. To identify the molecules responsible for these mechanisms a genetic screen was performed to identify mutations which result in defects in the normal pattern of axonal trajectories in the central nervous system (CNS).
Initially we have focused on mutations that affect axon guidance at the midline. Recent experiments suggest that the midline of the CNS of invertebrates and vertebrates acts to both attract and repel growth cones. One mutation that affects axon guidance at the midline is roundabout. In roundabout mutants growth cones that would normally follow an ipsilateral projection now cross the midline. The phenotype suggests the gene has a role in maintaining these axons away from the midline. The roundabout gene encodes for a protein containing five immunoglobin domains, three fibronectin type-III domains and an intracellular domain with no homologies to identified proteins. The protein is expressed on the longitudinally projecting axons. The expression pattern and mutant phenotype suggests the protein is required on axons to ensure they follow an ipsilateral projection. Misexpression experiments are being persued to identify if roundabout acts as a receptor for a signal and whether that signal acts on roundabout as a repellent or attractant for axonal extension.