Fig. 3.

Fat is required for planar polarized junctional remodeling. (A-C) Images from movies of wild-type (WT) Drosophila embryos expressing β-catenin:GFP. AP edges, blue lines; DV edges, pink lines. Colored dots mark individual cells. The three types of junctional remodeling leading to neighbor exchange are shown. (A) An AP edge between the yellow and orange cells shrinks and a new AP edge forms between the red and green cells (58% of neighbor exchange events, n=90). (B) An AP edge shrinks and a new DV edge forms (26% of neighbor exchange events). (C) A DV edge shrinks and a new AP edge forms (16% of neighbor exchange events). DV edges that shrink and form DV edges between new neighbors were not observed. Ventral views, anterior left. Scale bar: 5 μm. (D) Junctional remodeling events leading to cell rearrangement plotted by column in WT and fat^Grv/fat^NY1 in stages 13-16. DV edges contacted cells of the same column; AP edges contacted cells posterior to a given column. There were no significant differences between WT and fat. (E) Junctional remodeling in all columns of WT and fat^Grv/fat^NY1. There were more shrinking AP edges than shrinking DV edges (P<0.002) and more forming AP edges than forming DV edges in WT (P<0.001). This bias was retained in fat mutants (P=0.05 for shrinking edges and P=0.03 for forming edges). (F) Junctional remodeling in column 3 of WT and fat^Grv/fat^NY1. fat mutants had more shrinking DV edges (P<0.004) and more forming DV edges (P<0.03) in column 3 compared with WT. No significant differences in AP edge behaviors were observed. A single value was obtained for each column in each denticle belt by normalizing the number of cell rearrangements to the number of cells (32-40 cells/column in nine denticle belts from five embryos for WT, 18-25 cells/column in five denticle belts from five embryos for fat^Grv/NY1). The mean ± s.e.m. of these values is shown. ^*P<0.05.