An aggregon in conductin/axin2 regulates Wnt/β-catenin signaling and holds potential for cancer therapy

Abstract

The paralogous scaffold proteins axin and conductin/axin2 are key factors in the negative regulation of the Wnt pathway transcription factor β-catenin, thereby representing interesting targets for signaling regulation. Polymerization of axin proteins is essential for their activity in suppressing Wnt/β-catenin signaling. Notably, conductin shows less polymerization and lower activity than axin. By domain swapping between axin and conductin we here identify an aggregation site in the conductin RGS domain which prevents conductin polymerization. Induction of conductin polymerization by point mutations of this aggregon results in enhanced inhibition of Wnt/β-catenin signaling. Importantly, we identify a short peptide which induces conductin polymerization via masking the aggregon, thereby enhancing β-catenin degradation, inhibiting β-catenin-dependent transcription and repressing growth of colorectal cancer cells. Our study reveals a mechanism for regulating signaling pathways via the polymerization status of scaffold proteins and suggests a strategy for targeted colorectal cancer therapy.

Fig. 1

The conductin RGS domain prevents DIX-mediated polymerization. a Schematic to scale representation of axin and conductin (Cdt) with the domains interacting with APC (RGS), GSK3 (GSK), and β-catenin (β), and the polymerization domain (DIX). Percentage similarity (sim.) and identity (id.) are shown for each domain. b, d GFP fluorescence in U2OS cells transfected with indicated GFP-tagged axin or conductin constructs. Scale bars: 20 µm. c Schematic representation of chimeric axin-conductin proteins with axin parts shown in black and conductin parts in gray, and deletion mutants of axin and conductin used in b; not to scale. Distribution (Distrib.) is indicated on the right. Red lines mark the protein part (RGS domain) which determines distribution. e Percentage of transfected cells showing puncta formation of indicated constructs. Per construct, 1500 cells of three independent experiments as in b were analyzed. Results are mean ± SEM (n = 3). **p < 0.01, ***p < 0.001 (Student’s t-test). Source data are provided as a Source Data file

Fig. 2

An aggregation site within the RGS domain prevents conductin polymerization. a Aggregation propensity score calculated by the TANGO algorithm for amino acids in the RGS domains of conductin (Cdt), axin, and mutated conductin (Cdt mut). b Clustal Omega alignments of conductin and axin sequences for the three aggregation sites predicted in a. Identity (*) and conservation between amino acid groups of strongly (:) and weakly (.) similar properties are indicated. Mutated key residues are highlighted. c GFP fluorescence in U2OS cells transfected with indicated GFP-tagged constructs. Scale bar: 20 µm. d Percentage of transfected cells showing puncta formation of indicated constructs. Per construct, 1500 cells of three independent experiments as in c were analyzed. Results are mean ± SEM (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001 (Student’s t-test). e 3D structural model of the conductin RGS domain calculated by SWISS-MODEL using the crystal structure 1DK8 of the highly similar axin RGS domain (Fig. 1a) as a template^,. The Global Model Quality Estimate (GMQE) reached 0.80. Amino acids mutated within aggregation site III are indicated in red. N- and C-terminus are labeled with N and C, respectively. f Immunofluorescence staining of APC (red) in U2OS cells transfected with GFP-tagged conductin QV-PS (green) together with APC. Scale bar: 20 µm. g Western blotting for indicated proteins after GST-pull down from an extract of HEK293T cells expressing GFP-SAMP which was aliquoted and supplemented with equal amounts of either GST, GST-RGS Cdt, or GST-RGS Cdt QV-PS. Source data are provided as a Source Data file

Fig. 3

Aggregation site III is a functional interaction site (aggregon) at the RGS domain surface. a Western blotting for GFP under native (upper panel) or denaturing conditions (lower panel) in lysates of U2OS cells transfected with indicated GFP-tagged constructs. b Western blotting under native (upper panel) or denaturing conditions (lower panel) for recombinant GST-RGS Cdt and GST-RGS Cdt QV-PS purified from bacteria. c Western blotting for GFP in lysates (L) of HEK293T cells transfected with GFP-Cdt 2–345 (Cdt 2–345) or the GFP-Cdt 2–345 QV-PS mutant (QV-PS 2–345) together with GFP, or in fractions of these lysates prepared via ultracentrifugation through a sucrose density gradient (fractions 2 [low density] to 20 [high density]). Distribution of GFP shows comparable fractionation of both samples. d Coomassie Brilliant Blue staining of proteins extracted from bacteria using SDS-containing (lanes 1–4) or Triton X-100-containing (lanes 5–8) lysis buffers, or purified by pulldown on glutathione beads (lane 9). Source data are provided as a Source Data file

Fig. 4

Loss of RGS aggregation increases inhibition of Wnt/β-catenin signaling by conductin. a Immunofluorescence staining of endogenous β-catenin (red) in SW480 cells transfected with indicated GFP-tagged constructs (green). In GFP panels, insets are shown enlarged at the upper right. Dashed lines mark transfected cells. Scale bar: 20 µm. b Quantification of β-catenin fluorescence intensities in one out of five representative experiments as in a. Results are mean ± SEM (n = 50). c Quantification of GFP intensities in cells analyzed in b. Results are mean ± SEM (n = 50). Data distribution is shown in Supplementary Fig. 7. d Upper panel: Luciferase activity (TOP/FOP) in HEK293T cells transfected with indicated plasmids without or with (+) Wnt3a treatment. Results are mean ±  SEM (n = 5). Lower panel: Western blotting for GFP in extracts of HEK293T cells which were transfected with indicated GFP-tagged constructs at equal ratios as for the luciferase assay, and lysed directly in SDS-containing sample buffer due to differences in protein solubility. Loading control: α-tubulin. e MTT absorbance reflecting the number of viable SW480 (left panel) or DLD1 cells (right panel) expressing GFP-tagged conductin (Cdt, gray line) or the QV-PS mutant (black line) 0, 72, 96, and 120 h after seeding. One out of three representative experiments is shown. Results are mean ± SEM of six replicates (n = 6). **p < 0.01, ***p < 0.001 (Student’s t-test). Source data are provided as a Source Data file

Fig. 5

Saturation of the aggregon induces puncta formation of conductin. a Schematic illustration of conductin RGS aggregation with inhibited DIX-mediated polymerization, which can be shifted towards high-order DIX-mediated polymerization by (i) QV-PS mutation (red x), (ii) RGS co-expression, or (iii) the small peptide P^182–195 (red triangle). b Immunofluorescence staining for Flag (red) in U2OS cells transfected with Flag-tagged conductin (Flag-Cdt) either together with conductin 2–345 or conductin 2–345 QV-PS tagged with GFP (green). Scale bar: 20 µm. c Percentage of transfected cells exhibiting Cdt puncta. Per bar, 1500 cells of three independent experiments as in b were analyzed. Results are mean ±  SEM (n = 3). d Western blotting for indicated proteins in U2OS cell extracts. e Immunofluorescence staining of conductin or its M3 mutant in U2OS cells transfected with HA-tagged conductin or conductin M3 either alone, or together with P^182–195, or the QV-PS mutant of the peptide. Scale bar: 20 µm. f Percentage of transfected cells exhibiting Cdt or CdtM3 puncta. Per bar, 900 cells of three independent experiments as in e were analyzed. Results are mean ± SEM (n = 3). g, j Dot blot: Detection of GFP(-tagged) proteins binding to membrane pieces which were spotted with H₂O (−) or 8 nmol of R₉, P^182–195-R₉ or the QV-PS mutated peptide (QV-PS-R₉) prior to incubation with lysates of HEK293T cells transfected with indicated plasmids (g) or with in vitro translated GFP-RGS (j). h Western blotting for GFP in the lysates used in g. Loading control: α-tubulin. i, l 2D densitometry quantification of dot blots in g and j, respectively. Results are presented relative to the GFP/R₉ (i) or the GFP-RGS/R₉ (l) combination as mean ± SEM of five independent dot blots (n = 5). k Coomassie Brilliant Blue staining of in vitro translated GFP-RGS (+DNA template lane, arrowhead) used in j. Bands also present without template DNA (−) show purified kit components. m Western blotting for HA under native conditions in lysates of HEK293T cells transfected with HA-Cdt 2–345 alone (−) or together with P^182–195. *p < 0.05, **p < 0.01, ***p < 0.001 (Student’s t-test). Source data are provided as a Source Data file

Fig. 6

P^182–195 induces proteasomal degradation of β-catenin. a Immunofluorescence staining of endogenous β-catenin (green) in SW480 and SW480 AXIN2 knockout cells co-transfected with P^182–195 or its QV-PS mutated analog together with mScarlet-tubulin (red) to visualize transfected cells. Scale bar: 20 µm. b, c Quantification of nuclear β-catenin (b) or mScarlet (c) fluorescence intensities in individual cells of four independent experiments as in a. Results are mean ± SEM (n = 80); ***p < 0.001 (Student’s t-test). Data distribution is shown in Supplementary Fig. 10. d, e Western blotting for GFP and α-tubulin (loading control) in lysates of DLD1 (d) or HEK293T cells (e) transfected with indicated constructs, which were untreated or treated with the proteasome inhibitor MG132 (DLD1: 10 µM for 6 h; HEK293T: 2.5 µM for 2 h). Given nanograms (ng) of the peptide refer to the transfection of a 12 well. f Western blotting for GFP and HA in lysates (Input) of HEK293T cells transfected with indicated constructs, and after precipitation of HA-ubiquitin conjugated proteins from these lysates (IP α HA). Arrowheads point to polyubiquitinated GFP-β-catenin. HA blots show similar overall ubiquitination (Input) and similar precipitation of ubiquitinated proteins (IP) in both samples. Source data are provided as a Source Data file

Fig. 7

P^182–195 inhibits Wnt signaling and blocks growth of colorectal cancer cells. a–e, g, m Luciferase activity (TOP/FOP) in SW480 cells transfected with indicated amounts of P^182–195 or the QV-PS mutated control (a, b) in SW480 cells which were untreated or treated with indicated concentrations of the synthetic peptides R₉ (control), P^182–195-R₉, or its QV-PS mutant (QV-PS-R₉) for 48 h (c), in SW480 cells transfected with P^182–195 together with siRNA against GFP (control) or against axin2 (d), in parental SW480 cells (WT #1) and two AXIN2 knockout clones (−/− #1, 2) which were untreated or treated with 0.1 µM P^182–195-R9 for 48 h (e), in parental SW480 cells (WT #1), a WT axin2 control clone (WT #2), and two axin2 QM-PS mutated clones (QM-PS #1, 2) transfected with P^182–195 (g), in SW480 cells transfected with P^182–195 and/or treated with 50 nM of the tankyrase inhibitor G007-LK overnight (m). Western blot below d shows efficient axin2 knockdown. Results are mean ± SEM (n = 5 [a, g], n = 4 [b, c, d], n = 3 [e, m]). f Western blot shows absence of axin2 in the AXIN2 knockout clones used in e. h Relative mRNA expression of the β-catenin target genes LGR5, MYC, and AXIN2 normalized to GAPDH in SW480 cells or SW480 AXIN2 knockout cells which were untreated or treated for 48 h with 10 µM of indicated peptides. Results are mean ± SEM (n = 3). i, k Cell colonies grown for 96 h from SW480 cells which were transfected with increasing amounts of P^182–195 or its QV-PS mutated analog together with GFP, sorted by GFP expression and sparsely plated (i), or from SW480 cells or SW480 AXIN2 knockout cells which were treated with 10 µM of indicated synthetic peptides (k). Cells were stained by ethidium bromide incorporation and visualized with UV light. Scale bar: 0.5 cm. j, l Automated quantification of colony numbers from three independent experiments as in i (j) or k (l). Results are mean ± SEM (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001 (Student’s t-test). Given nanograms (ng) of the peptide refer to the transfection of a 12 well. Source data are provided as a Source Data file