Wnt/β-catenin signaling is critically important to development and disease. How this signaling pathway functions normally in vivo remains unclear. Although a standard model is widely accepted (Fig. 1), it has been established without direct observation in the steps between receptor activation and the final step, the transcriptional activation by β-catenin (β-cat; Armadillo, Arm in Drosophila). To explain events in this ‘black box’ (Fig. 2) several diverse models have been postulated, based on protein overexpression in highly artificial in vitro assays. In our recent work, we performed a functional analysis of the Axin complex, the central regulator in the pathway, which functions to limit or turn off signaling (Fig.1B). One central finding was that this multi-protein complex is largely resistant to inactivating mutations because tripartite interactions within the complex result in robust assembly and function (Peterson-Nedry, 2008). We also demonstrated that Wnt signaling is not controlled primarily by changing Axin levels, excluding a popular model (Peterson-Nedry, 2008).
Fig. 1. (A) A genetic view of the pathway. (B,C) Representation of a core pathway model and (D) an accessory module. (B) In the absence of Wnt, the ‘destruction complex,’ contains as its core components the scaffold Axin, APC, and the kinase Shaggy (Sgg; Drosophila GSK3), which phosphorylates Armadillo (Arm; Drosophila β-catenin, β-cat), targeting it for degradation. (C) Wingless (Wg), the principal Drosophila Wnt, binds to Frizzled and Arrow (LRP5&6 in vertebrates) receptors. In a poorly understood way, the receptors activate the cytoplasmic transducer Dishevelled (Dsh, a protein without known enzymatic activity), initiating inhibition of the destruction complex. This allows Arm to accumulate and activate transcription in the nucleus, completing the cascade. (D) Model for an accessory module, allowing signal amplification in the shallow part of a Wg/Wnt ligand gradient to generate a more linear intracellular signaling gradient. Wnt receptor activation promotes the phosphorylation of Arrow by Gish (Drosophila Casein Kinase Iγ, CKI γ), enabling Arrow to bind Axin. This interaction potently inhibits the destruction complex, generating a greatly increased signal compared to (C) (see also Fig. 2; Baig-Lewis, 2007).
Fig. 3. The signal amplification function of Arrow (Fig. 1D) is mimicked by torDArr construct, which strictly depends on available Wg ligand to potentiate the intracellular signal. (A, B) torDArr expressing clones (green) in the wing imaginal disc also exhibit a cell-autonomous increase in the expression of the Wg/Wnt target gene Distal-less (Dll, red). Within the clone, Dll expression diminishes in a graded fashion away from the source of Wg at the dorsoventral (D/V)boundary (indicated by arrows), reflecting the Wg dependence of signal potentiation by torDArr. (C) The endogenous Dll gradient profile matches the Dll gradient generated in a clone expressing torDArr. In order to accommodate the three-dimensional nature of this tissue, Dll staining in the region indicated in panel A was determined and the graph indicates expression levels as a function of the distance from the endogenous signal peak (at the D/V boundary, arrowhead in A). As apparent in panel A, a sharply defined edge of high-level expression is visible at the border of the clone facing the D/V boundary. The gradual increase of Dll expression levels reflects averaging across the curved boundary of the clone and the slanted angle of the epithelium. (C′) A vertical section reveals that tissue folds position the GFP expressing clone at an angle similar to the Dll profile observed in panel C and explains why the Dll increase appears gradual. This vertical section through the Z-stack along the middle of the box shown in panel A; green, GFP; red, Dll; the blue horizontal line marks the level of the section shown in panels A and B (for details see Baig-Lewis, 2007).
The main obstacle to observing dynamic Wnt signaling is that Wnt pathway components typically have additional functions outside the Wnt pathway. Because the sub-pool of molecules functioning in Wnt signaling is not evident, it cannot be observed by conventional methods (immunohistochemistry, epitope-tagging or co-localization).