Fig. 3. Slit signaling is required for ganglionic branch turning at the midline. (A-M) Stage 16 embryos stained for the tracheal lumen in blue and the GB1 cell marker DSRF (A-E) or longitudinal connectives (F-K) in brown. In wild-type embryos (A), GBs turn posteriorly at the ventral midline but in robo and slit mutants (arrows in B and D) they cross it. The fascicles of the longitudinal connectives are separated in wild type (F), but cross the midline in robo mutants (arrows in G); in slit mutants (I), the fascicles are fused at the midline (arrows). (L) Transgenic tracheal expression of UAS-robo (anti-Robo staining in brown) in robo mutants rescues GBs from crossing the ventral midline (arrowhead), whereas GBs with no detectable Robo expression still cross (arrow). The axon phenotype in robo mutant embryos is partially rescued by transgenic expression of UAS-robo in all neurons with the elav-GAL4 driver (K), but GBs still cross the ventral midline (arrow in K). (C) In robo2 mutant embryos, the GBs fail to enter the CNS (asterisk) and several of the GBs that do enter stall or become misrouted (arrowhead). (H) The longitudinal axon tracts are disrupted (arrows) and the outer fascicles fuse with the medial fascicles in robo2 mutants. (M) The GB failure to enter the CNS in robo2 mutants can be rescued by transgenic tracheal expression of UAS-robo2 with the SRF-GAL4 driver. (E) GBs in the double mutant robo, robo2 fail to enter the CNS (asterisk) and cross the ventral midline (arrow) and the longitudinal axons collapse along the midline (J). These phenotypes are very similar to the phenotypes of slit mutants (D,I). Scale bar: 20 µm in A-M.