Figure 1. E-Cad is required for the highly cohesive behaviour of PMG cells during migration.

(a) The migrating PMG consists of three groups of cells: PMECs, ICPs and AMPs. (b) Wild-type (WT) embryo stained for E-Cad, arrowheads point to PMECs, arrows to ICPs. The PMG is demarcated by dashed lines. (c) Stills from 20-min movies of ICPs visualized by a membrane-bound GFP driven by insc-Gal4, arrows point to protrusions. (d–j) Stage 12 WT (d,f,h,j) and shotgun (shg) (e,g,i,k) mutant embryos (E-Cad is encoded by the gene shg). The PMG is demarcated by dashed lines in d,e,f,g. (d,e,h,i) GMAGFP is used to label cortical actin11, in both fixed (d,e) and in vivo movies (h,i), arrows point to holes between cells in shg mutants (e); in shg mutants, PMECs show some defects in apicobasal polarity (e and ref. 2). (h,i) Asterisks mark the positions of unmarked germ cells, while gaps between cells appear as larger irregular holes (i, dashed lines). (f,g) PMG cells labelled for Hnt (blue), which is expressed in both PMECs and ICPs, and Inscuteable (Insc) (red), which is expressed in just ICPs, arrows point to mispositioned ICPs (g). (j,k) Transmission electron microscopy images of ICPs show no visible junctions between cells, arrows point to holes in shg mutants, asterisks to lipid droplets. Dashed line in j outlines a cell protrusion tightly aligned with neighbouring cells. Middle and right images are increasing magnifications from the same embryo. Scale bars, 20 μm (b–i); 2 μm (j,k, left and middle); 500 nm (j–k, right).