Fig. 3.

Magi acts through RASSF8. (A-F) Pupal eye at 44 h APF stained for E-Cad showing ommatidial organization. Arrowheads highlight extra IOCs compared with wild type. Whereas the supernumerary IOC phenotype in Magi mutants (A) is completely rescued by introducing a wild-type Magi transgene (B), only partial rescue is seen with a MagiΔWW transgene (C). Similarly, only a wild-type RASSF8 transgene (E), and not a RASSF8^YA transgene mutated in its Magi-interacting motif (F), is able to fully rescue the extra IOC phenotype in RASSF8 mutants (D). (G) Quantification of the number of IOCs per ommatidium in A-F (see Materials and Methods). s.e.m. is shown; ***P<0.001, **P<0.01, *P<0.05 (unpaired t-test). (H) Strength of rescue obtained after normalization. Normalization was obtained by dividing the number of IOCs in the different rescue experiments by the number of IOCs in the corresponding Magi or RASSF8 mutant background. (I-M) Scanning electron microscopy of Drosophila adult eyes following GMR-Gal4-driven overexpression. Full-length Magi overexpression produces a mild rough eye (K) compared with wild type (I), while MagiΔWW has a weaker effect (J) (all transgenes are inserted at the same locus ensuring equivalent expression levels). This effect of Magi is strongly suppressed by removing one copy of RASSF8 (RASSF8⁶/+; L) but is enhanced by co-expressing RASSF8 (M). (I′-M′) High magnification of adult eyes corresponding to I-M. Note the almost wild-type packing after MagiΔWW (J′) overexpression, and the very disorganized eye structure after Magi and RASSF8 co-overexpression (M′). (N) Quantification of the effects observed in I′-M′, highlighting the suppression of the effects of Magi overexpression by RASSF8 mutants. n, total number of ommatidia scored per genotype. Scale bars: 100 µm in I; 50 µm in I′.