The role of receptor-associated E3 ligases in WNT signaling and disease
Summary
The WNT target gene LGR5 was initially identified as a marker of adult intestinal stem cells. Later, it was found to also mark stem cells in other organs, including the pancreas, stomach, mammary gland, liver, and kidney. LGR5 and its homologues LGR4 and LGR6 are receptors for R-spondins. Their WNT signaling-enhancing ability is an essential requirement for WNT-driven adult stem cells. The LGR5/R-spondin complexes function through neutralizing RNF43 and ZNRF3, two transmembrane RING E3 ubiquitin ligases expressed in stem cells that target the WNT receptor Frizzled for degradation. RNF43 and ZNRF3 themselves are also WNT target genes and thus constitute a negative feedback loop. For RNF43 and ZNRF3 to function, they most likely should be able to interact with both R-spondin and the WNT/Frizzled complex. The binding to R-spondin, mediated by their extracellular PA domain has been described in great detail. RNF43/ZNRF3 binding to Frizzled, however, has not been fully studied. Studies performed with PLR-1, the ortholog of RNF43 and ZNRF3 in C. elegans, strongly indicate an involvement of the PA domains of these E3 ligases in Frizzled recognition. Studies performed with lymphocyte-expressed GRAIL, an E3 ligase with similar domain composition, also suggest that the PA domain is required to direct the substrate specificity of the ubiquitin ligase modification. In the case of RNF43 and ZNRF3 this may suggest a competitive binding model for the PA domain, either binding to R-spondin or to Frizzled. Mapping of somatic mutations in the RNF43 gene, found in the COSMIC cancer database, reveals a surprisingly high number within the PA domain. Mutations affecting the level of RNF43 protein or the presumed ability of its PA domain to interact with the WNT/Frizzled complex may result in hyper activated WNT signaling and initiation of cancer. Such phenomena were reported for mouse experiments upon deletion of both E3 ligases. Aberrant WNT signaling due to mutated APC or ß-catenin drives tumorigenesis in almost 90% of the colon cancer patients. RNF43 and or ZNRF3 may function as tumor suppressors and mutated versions may contribute to colon cancer. The same may apply to other tissues, since a high incidence of RNF43 mutations can be detected in tumors of the pancreas, biliary tract, ovary and endometrium.