α-Catenin doesn't link cadherin to actin (top) but does inhibit actin polymerization by Arp2/3 (bottom).
WEIS/ELSEVIER
Weis's original goal was ambitious—a structural understanding of adherens junctions. First came biochemistry. “It was only when we attempted to reconstitute [the junctions],” he says, “that we found that nobody had actually done it before.” The textbook model “was all based on binary interactions.”
The group now reports that α-catenin can either bind as a monomer to β-catenin and thus cadherin, or bind as a homodimer to actin. But the binding events are mutually exclusive so there is no direct link from cadherin to actin. Consistent with this, cadherins and actin have very different dynamics.
Weis stresses that their two model systems—clustered soluble fragments of cadherin and membrane patches—are something less than a full-blown adherens junctions. But such imprecision may be reasonable. In cells, says Weis, although “there is certainly local [cadherin] clustering, I don't think there's precise geometry.”
In the dimeric state α-catenin inhibits actin polymerization by Arp2/3, perhaps favoring formin's ability to create stable actin cables over Arp2/3′s ability to promote protrusive actin branching. The membrane patches did not, however, reconstitute assembly of actin filaments, suggesting that other proteins are missing. These other proteins may link actin to other membrane proteins, or they may reinforce the actin-modifying behavior of α-catenin.
“People are very intrigued and excited” by the new results, says Weis. But the new textbook is clearly a work in progress. One of the biggest mysteries is how a system with no direct linkages generates force during morphogenesis. “Maybe all you need to do is to organize the gel state of actin correctly, and that organization will support the mechanical function of the junction,” says Weis. But “with constriction, you clearly need to be linked to something.” 
References:
