Polarity precedes adhesion

Epithelial cells distinguish top from bottom even before they stick to their neighbors, according to results on page 135 from Harris and Peifer. Figure Apical restriction of DE-Cadherin (green, left) is lost in baz mutants (right). Polarized epithelia are characterized by complexes that provide tight adhesions between neighboring cells. These complexes, known as adherens junctions (AJs), contain transmembrane cadherins that link adjacent cells. Cadherin is anchored to the actin cytoskeleton through the AJ proteins Armadillo and α-catenin. The building of AJs is normally considered the earliest known event in establishing apicobasal polarity. Looking for polarity in the absence of AJs, however, is tricky; sheet architecture falls apart if AJs are removed, so the cells are no longer epithelial. But Harris and Peifer find that, in the developing fly, enough epithelial cells stay in place to determine that a polarized epithelium can develop without AJs. The signs of polarization come from two apical scaffold complexes that also help establish polarity. One apical complex is defined by Bazooka (Baz), which recruits Crumbs, a member of the second complex. The authors see that both complexes attain apical localization in Armadillo mutants. In these flies, an early epithelium is built, and portions of the embryo undergo morphogenesis. Without Baz, on the other hand, AJ components are not correctly localized, suggesting that Baz is the earlier polarity factor. AJs are needed to maintain polarity, as most epithelial structures later broke down in their absence. These defects probably reflect failures in cell–cell adhesion, but AJs also had specific effects on late polarity events, such as the segregation of the basolateral protein Discs Large, which was retained apically the residual epithelial structures of the mutants. The earliest polarity cue is still a mystery, but cells may somehow sense the outside of the embryo and define it as apical. Actin and microtubules were polarized in AJ mutants, suggesting that they may transport Baz to the apical domain. Large complexes of Baz oligomers may then recruit AJ proteins, but as yet it is unclear exactly how Baz is linked to AJ formation.