Armadillo/beta-catenin plays an essential role in both Wnt signaling and adhesion. beta-catenin mediates Wnt signaling by serving as an essential transcriptional coactivator upon binding to Tcf/LEF-1 transcription factors. beta-catenin also associates with the intracellular domain of cadherins to mediate adhesion. Three putative beta-catenin homologs have been identified in C. elegans: WRM-1, BAR-1 and HMP-2. WRM-1 functions in a Wnt pathway that is different from the pathway in flies and vertebrates. Instead of interacting directly with the Tcf homolog POP-1, WRM-1 interacts with a kinase, LIT-1. The activities of WRM-1 and LIT-1 ensure that POP-1, which probably acts as a repressor in this pathway, is downregulated in the posterior daughter cells of many anterior/posterior cell divisions. Differences in POP-1 expression levels may subsequently specify differences in fate. In addition to this putative repressor function, we show that POP-1 can also promote target gene transcription. POP-1 binds to the consensus Tcf target site and activates transcription of a Tcf reporter gene in combination with Armadillo. We asked which of the three worm beta-catenins can bind directly to POP-1. Only BAR-1 binds to POP-1 in yeast-two-hybrid and co-immunoprecipitation experiments. Furthermore, only the combination of POP-1 and BAR-1 activates transcription of a Tcf reporter gene. This suggests that POP-1 and BAR-1 have a function that is similar to that of Tcf and beta-catenin. To investigate this further, we asked if the POP-1/BAR-1 interaction is required for the expression of a putative Wnt target gene,
mab-5. The Hox gene
mab-5 is required for the posterior migration of the descendants of QL (QL.d) and mutation of
mab-5 results in the anterior migration of these cells. Mutations in the Wnt pathway components
egl-20/Wnt,
lin-17/Fz and
bar-1 also result in QL.d migration defects. These mutations show reduced or absent
mab-5 expression, suggesting that a Wnt pathway regulates the expression of this Hox gene. To test if the interaction between POP-1 and BAR-1 is also required for the expression of
mab-5, we inducibly expressed a
pop-1 mutant (DN-
pop-1) that lacks the BAR-1 interaction domain and acts as a strong dominant negative mutation in a Tcf reporter gene assay. We find that overexpression of DN-
pop-1 results in the anterior migration of the QL.d and the inhibition of
mab-5 expression. This demonstrates that BAR-1 and POP-1 are indeed required for the expression of a Wnt target gene, which suggests that BAR-1 and POP-1 function in a Wnt pathway that is similar to that in flies and vertebrates. We find that the third worm beta-catenin, HMP-2, is the only beta-catenin homolog that interacts with the single cadherin, HMR-1. The interaction of HMP-2 with HMR-1 was anticipated from the previously reported colocalization of HMP-2 with the alpha-catenin HMP-1 and HMR-1 in adherens junctions. The absence of an apparent adhesion function for BAR-1 and a signaling function for HMP-2 is however unexpected. We propose that the functions of Armadillo/beta-catenin have been distributed over three different beta-catenins in C. elegans: WRM-1 and BAR-1 function in Wnt signaling, whereas HMP-2 functions specifically in adhesion