Identification of targets of the Wnt pathway destruction complex in addition to beta-catenin

Abstract

The proteasomal degradation of beta-catenin mediated by the glycogen synthase kinase 3beta (GSK3beta) and destruction complex is the central step in the canonical Wnt signaling pathway. However, that there are branches of Wnt signaling pathways that do not depend on beta-catenin/Tcf-mediated transcription activation has long been understood. In this study, we hypothesized that there are many more GSK3 and destruction complex-dependent proteolytic target proteins that mediate Wnt signaling in the cell. To test this hypothesis, we have developed and carried out a screen for such candidate proteins using an in vitro expression cloning technique and biochemical reconstitution of Wnt signaling in Xenopus egg cytoplasmic extracts. Forty-two proteins have been identified as potential candidates for GSK3-regulated phosphorylation, proteasomal degradation, or both, of which 12 are strong candidates for Wnt-pathway-regulated degradation. Some of them have been reported to interact with beta-catenin and implicated in the canonical Wnt signaling pathway, and other targets identified include proteins with various cellular functions such as RNA processing, cytoskeletal dynamics, and cell metabolism. Thus, we propose that Wnt/GSK3/destruction complex signaling regulates multiple target proteins to control a broad range of cellular activities in addition to beta-catenin-mediated transcription activation.

Fig. 1.

Examples of the degradation assay of in vitro translated Xenopus cDNA pools used in the screen. Small pools containing 50–100 clones of gastrula stage Xenopus cDNA library were in vitro translated with [³⁵S]methionine. Radiolabeled protein pools were incubated with Xenopus egg extract for 3 h at room temperature and then analyzed by SDS/PAGE and autoradiography. To inhibit GSK3 activity and the destruction complex, 25 mM LiCl or 0.8 μM recombinant GID protein was used, respectively. Filled arrowheads indicate the protein bands that are degraded in the Xenopus egg extract and rescued by LiCl and GID (Li/GID). Open arrowheads indicate protein bands that are degraded in the Xenopus egg extract and rescued by LiCl but not GID (Li-Only). Asterisk indicates protein bands that showed mobility shifts in SDS/PAGE after incubation in the Xenopus egg extract and LiCl blocked the mobility shift.

Fig. 2.

Degradation assay of isolated putative Wnt/GSK3 target proteins identified in the screen. The genes encoding positive protein bands were identified by progressive subselection (see Methods). Positive individual clones were sequenced and verified with the degradation assay. (A and B) Of 42 identified proteins, the degradation of 23 proteins by the Xenopus egg extract was inhibited by LiCl but not GID (A), and 12 proteins were positive for both LiCl and GID (B). (C) Seven proteins showed mobility shift after incubation in the Xenopus egg extract that was inhibited by LiCl. Lanes: −, 1× XB buffer added in the reactive as a negative control; Li, 25 mM LiCl; GID, 0.8 μM GID (GSK3β interaction domain of Axin).

Fig. 3.

Proteins identified in the screen arranged in functional groups. Forty-two phosphorylation or proteolytic targets of Wnt, GSK3, or both are identified by whether their degradation is inhibited by LiCl or GID or their mobility shift is inhibited by LiCl. Also indicated are proteins reported to interact with β-catenin. Known interactions between candidate proteins are indicated by double-headed arrows.

Fig. 4.

Specificity of the degradation of selected candidates for the Wnt/GSK3 signaling pathway destruction complex. (A) GID does not inhibit the activity of GSK3β toward all protein targets in Xenopus egg extracts. GID blocked the degradation of β-catenin but not the mobility shift of Tau or the degradation of Myc protein. Lanes: −, 1× XB buffer added in the reaction as a negative control; Li, 25 mM LiCl; GID, 0.8 μM GID. (B) Selected proteolytic candidates of the Wnt/GSK3 signaling pathway were incubated in the Xenopus egg extract with various inhibitors of the destruction complex. The degradation of Li/GID-positive proteins was rescued by all of the Wnt pathway activators, whereas Li-Only-positive genes were not. Lanes: −, 1× XB buffer added in the reaction as a negative control; LiCl, 25 mM LiCl; GID, 0.8 μM GID; dnGSK3β, 0.3 μM dominant-negative GSK3β (kinase dead); AxΔRGS, 0.2 μM Axin protein missing the RGS domain (which binds to APC); Dsh, 0.2 μM recombinant Dishevelled protein. (C) Wnt-dependent protein stabilization in Xenopus embryos. mRNA encoding one of the target proteins indicated was injected into one of the ventral blastomeres of the 4-cell-stage embryo along with Myc-EGFP mRNA as a nontarget control. Cells were also coinjected with either Wnt8 mRNA or a control mRNA. Filled arrowheads: β-catenin and target proteins; open arrowhead: Myc-EGFP.

Fig. 5.

A model for the expanded canonical Wnt/GSK3 signaling pathway. Binding of Wnt proteins to receptors triggers canonical (β-catenin and destruction complex dependent) or noncanonical (β-catenin and the destruction complex independent) Wnt signaling. In the canonical Wnt signaling pathway, the stability of β-catenin is controlled by destruction complex composed by GSK3, APC, and Axin, etc. In addition to β-catenin, the degradation of many different proteins in various functional classes is regulated by the destruction complex and Wnt receptor signaling.