Fig 2: Endocytosis inhibitors cause a decrease in the level of Armadillo/ß-catenin both in stimulated and unstimulated cells. (A) The level of Armadillo increased in S2R+ cells that had been treated with Wingless-conditioned medium (lane 2) or SB-216763 (lane 5). This was prevented by treatment with Dyngo-4a (lanes 3 and 6) or Dynasore (lanes 4 and 7). Lamin, Actin and Syntaxin levels were unaffected. (B) Dyngo-4a reversibly reduced the level of Armadillo in uninduced cells. (C) Dyngo-4a reduced signalling-induced accumulation of ß-catenin in RKO cells. Cells were pre-incubated with SB-216763 to activate signalling and then exposed to a mixture of SB-216763 and Dyngo-4a. The total time of treatment with either drug is indicated. As in Drosophila cells, Dyngo-4a caused a decrease in ß-catenin levels in unstimulated cells. A progressive decrease can be seen after 0.5, 1 and 2 hours of treatment with Dyngo-4a (no SB-216763) in lanes 9–11. (D) The effect of Dyngo-4a and SB-216763 on the level of various components of the Wnt pathway. LRP6, GSK3ß and CK1a were largely unaffected, whereas the levels of APC and Axin1 dropped markedly 30–60 minutes after treatment with Dyngo-4a. SB-216763 caused an increase in the amount of ß-catenin. This correlated with a decrease in phosphorylated ß-catenin (pT42/S37/S33 ß-catenin; lane 3), as expected because phosphorylated ß-catenin reflects the activity of the degradation complex (Hernández et al., 2012). By contrast, the (mild) decrease in ß-catenin caused by Dyngo-4a (lanes 7 and 8) is paralleled by a similar decrease in phosphorylated ß-catenin, suggesting that Dyngo-4a impacts on the level of ß-catenin through a mechanism that is independent of the destruction complex.

Fig 3: The C-terminal DAX domain of AXIN2 controls CTNNB1 transcriptional activity.(A) GT insertions in LRP6 (top histogram) and AXIN2 (bottom histogram) mapped for the sorted cells from the WNT positive regulator, low stringency screen (Figure 1E). The histograms depict the number of GT integrations in the sense (blue) or antisense (red) orientation, relative to the coding sequence of the gene, within consecutive 500 base pair (bp) intervals along the length of each gene. Due to the directionality of the splice acceptor in the GT, typically only sense GT insertions in introns disrupt the gene, whereas GT insertions in exons generally disrupt the gene regardless of orientation. RefSeq gene tracks for LRP6 and AXIN2 are shown beneath each histogram following the University of California, Santa Cruz (UCSC) genome browser display conventions: coding exons are represented by thick blocks, UTRs by thin blocks, and introns by horizontal lines connecting the blocks. Both genes are displayed with their 5’ ends to the left, and encompass chromosome 12, bps 12267499–12116000 for LRP6, and chromosome 17, bps 65561999–65528500 for AXIN2 (hg18). (B) Schematic representation of the human AXIN2 protein drawn to scale in the horizontal dimension. Amino acid numbers are indicated below, and arrows show the sites at which truncations were made by CRISPR/Cas9-mediated genome editing in the indicated cell lines. Known domains, regions and motifs (based on UniProt annotation) are depicted in gray (TB, tankyrase-binding motif). Exon 11, eliminated by GT insertions found in cells sorted during the WNT screens for positive regulators, is delineated by a thinner white block. (C) Fold-induction in WNT reporter (median EGFP fluorescence from 20,000 cells) following treatment with 50% WNT3A CM or 10 µM of the GSK3 inhibitor CHIR-99021. Each circle represents a unique clonal cell line (determined by genotyping, Supplementary file 2), and the average of three to four independent clones for each genotype is indicated by a horizontal line. For each treatment, percentage reporter activation relative to WT cells is also indicated above each group of circles to facilitate comparisons. Significance was determined by one-way ANOVA, and is indicated as ****(p<0.0001) or ns (not significant). (D) Fold-induction in soluble CTNNB1 protein (average CTNNB1 intensity normalized to ACTIN intensity from duplicate immunoblots) following treatment with 50% WNT3A CM or 10 µM CHIR-99021. Each circle represents a unique clonal cell line, and the average of two independent clones for each genotype is indicated by a horizontal line. Significance was determined by unpaired t-test with Welch’s correction. Representative immunoblots used for quantification of CTNNB1 and ACTIN are shown in Figure 3—figure supplement 1C. (E) Nuclear CTNNB1 (average nuclear fluorescence per unit area from three fields of view) in single clonal cell lines of the indicated genotypes was quantified as described in Materials and methods. Cells were treated with 50% WNT3A CM where indicated. For each cell line, the fold-increase in CTNNB1 nuclear accumulation following treatment with WNT3A, expressed as percentage of WT, is also indicated above the bars to facilitate comparisons. For WNT3A-treated cells, differences in nuclear CTNNB1 between WT and mutant cells were not statistically significant as determined by one-way ANOVA. Examples of confocal sections used for quantification of nuclear CTNNB1 are shown in Figure 3—figure supplement 2B–2D. (F) WNT reporter activity (median EGFP fluorescence from 10,000 cells), soluble CTNNB1 protein (average CTNNB1 intensity normalized to ACTIN intensity from duplicate immunoblots), and nuclear CTNNB1 protein (average nuclear fluorescence per unit area from 2 to 3 fields of view), depicted as percentage of WT, for cells treated with 50% WNT3A CM. Each circle represents a unique clonal cell line, and where applicable the average of two independent clones is indicated by a horizontal line. (G) Adult D. melanogaster wings expressing Axin-V5 (top images) or Axin?C-V5 (bottom images) under the control of the c765-Gal4 driver. Loss of sensory bristles and tissue at the wing margin, indicative of impaired Wg signaling, is shown (arrow) in the higher magnification view of the delineated area. While loss of Notch signaling can also result in wing margin defects due to a requirement of Notch signaling for Wg expression at the dorso-ventral boundary in the wing imaginal disc (Diaz-Benjumea and Cohen, 1995), we ruled out that possibility by confirming intact Wg expression in the wing imaginal disc of flies expressing Axin?C-V5 (Figure 3—figure supplement 3B). Scale bars = 20 µm. (H) Percentage of flies with wing margin defects. 4.3% of flies expressing Axin-V5 exhibited loss of bristles at the wing margin, but no loss of wing tissue; 46.6% of flies expressing Axin?C-V5 exhibited loss of bristles at the wing margin as well as loss of wing tissue.DOI: http://dx.doi.org/10.7554/eLife.21459.007

Fig 4: Reporter-based, forward genetic screens in haploid human cells identify negative, attenuating and positive regulators of WNT signaling.(A) Schematic model of canonical WNT signaling, highlighting the main pathway components and regulatory events in the absence (left panel) and presence (center panel) of ligands, and other known regulators relevant to this work (right panel). When the pathway is off, the transcriptional co-activator ß-catenin (CTNNB1) is constitutively targeted for proteasomal degradation by the destruction complex, composed of the scaffold proteins adenomatous polyposis coli (APC) and AXIN, and the kinases glycogen synthase kinase 3 (GSK3) and casein kinase 1a (CSNK1A1). The T-cell-specific transcription factor (TCF)/lymphoid enhancer-binding factor (LEF) family of transcription factors, together with transducin like enhancer of split (TLE) and histone deacetylases (HDAC), repress WNT target genes. Binding of WNT to its co-receptors frizzled (FZD) and low-density lipoprotein receptor-related protein 6 (LRP6) leads to the assembly of a receptor complex that inactivates the destruction complex through a mechanism involving recruitment of AXIN by LRP6 and the adapter protein dishevelled (DVL). Consequently, CTNNB1 accumulates in the cytoplasm, translocates to the nucleus and promotes WNT target gene transcription in cooperation with TCF/LEF and other co-activators such as CREB-binding protein (CREBBP) and B-cell CLL/lymphoma 9 protein (BCL9). R-spondins (RSPOs) are secreted proteins that potentiate the response of stem cells to WNT ligands by blocking the degradation of FZD and LRP6 receptors. RSPO binds to leucine-rich repeat-containing G-protein-coupled receptors (LGRs) and neutralizes two transmembrane E3 ubiquitin ligases, ZNRF3 and RNF43, that clear WNT receptors from the cell surface. Other regulatory mechanisms include modulation of AXIN levels by the poly ADP-ribosylation-dependent E3 ubiquitin ligase RNF146, and recruitment of DOT1L and MLLT10, two proteins involved in histone H3 K79 methylation, to WNT target genes. (B) Schematic of WNT reporter-based forward genetic screens in haploid human cells using a GT-bearing retrovirus for mutagenesis, followed by phenotypic enrichment by FACS. LTR, long terminal repeats; SA, splice acceptor; pA, polyadenylation signal. (C–F) Circle plots depicting genes enriched for GT insertions in screens for negative (C), attenuating (D) and positive (E and F) regulators of WNT signaling. Two independent screens for positive regulators were performed at low (E) and high (F) selection stringencies by sorting for cells with the lowest 10% and 2% WNT reporter fluorescence, respectively. The y-axis indicates the significance of GT insertion enrichment in the sorted vs. the control cells (expressed in units of -log10FDR-corrected p-value) and the x-axis indicates genes (in random order) for which GT insertions were mapped in the sorted cells. Genes with FDR-corrected p-value<0.01 are labeled and colored in light blue if they encode a known pathway component, or in pink if their product has not been previously implicated as a regulator of canonical WNT signaling. The diameter of each circle is proportional to the number of unique inactivating GT insertions mapped in the sorted cells, which is also indicated next to the gene name for the most significant hits with FDR-corrected p-values<10-4.DOI: http://dx.doi.org/10.7554/eLife.21459.003

Fig 5: APC knockout in CRC cells impairs Wnt-induced ß-catenin destruction complex localization. (a) APC staining in RKO and RKO-APCKO cells. (b–d) Representative images of RKO-APCKO cells treated with UB or WB. Representative images and scoring shown for (b,e) ß-catenin, (c,f) Axin1, and (d,g) GSK-3ß. Data averaged from three independent experiments (n = 25 cells per condition). Error bars, SEM; Statistical analysis by t-test: *P < 0.05; **P < 0.01; ***P < 0.001. Scale bar, 10 µm.