Wnt target IQGAP3 promotes Wnt signaling via disrupting Axin1-CK1α interaction

Abstract

The scaffold protein IQGAP3 is highly upregulated in most epithelial cancers. While recent studies have highlighted its pivotal roles in cancer cell proliferation and metastasis, a deeper mechanistic understanding of IQGAP3 is currently lacking. We have here used TurboID to map IQGAP3 proximity partners and identified the Wnt signaling members Axin1 and CK1α as IQGAP3-interacting proteins. Our functional studies demonstrated that overexpression of IQGAP3 increases β-catenin levels, while IQGAP3 depletion reduces β-catenin levels in gastric cancer cells. Mechanistically, IQGAP3 disrupts Axin1-CK1α interaction, thereby inhibiting β-catenin phosphorylation and ultimately leading to its accumulation. Importantly, we discovered that IQGAP3 itself is regulated by Wnt signaling, suggesting its involvement in a positive feedback loop in Wnt/β-catenin signaling through interactions with Axin1 and CK1α. These findings identify IQGAP3 as a novel mediator of β-catenin stabilization and underscore its potential as a target for cancer therapy.

Fig. 1. Proximity proteomics reveal IQGAP3 interaction with Cdc42, Anillin and with centrosome- and septin-complexes.

A Workflow for proximity-dependent identification of IQGAP3-associated proteins in HEK293 cells treated with either L-cells condition media (−Wnt3a) or L-Wnt3a cells condition media (+Wnt3a) treatment for 4 h, created with BioRender.com. B Venn diagram of biotinylated proteins detected in HEK293 cells treated with L-cells condition media (−Wnt3a) and L-Wnt3a cells condition media (+Wnt3a) (Fold change ≥ 4, p-value ≤ 0.01). C Interactome of −Wnt3a ∩ +Wnt3a (Fold change ≥ 4, p-value ≤ 0.01), node color denote degree distribution, created with Cytoscape. D Gene ontology: Biological Process of shared proximity partners in −Wnt3a and +Wnt3a treated cells. E Gene ontology: Molecular function of shared proximity partners in −Wnt3a and +Wnt3a treated cells. F Gene ontology: Cellular compartment of shared proximity partners in −Wnt3a and +Wnt3a treated cells.

Fig. 2. Wnt3a treatment promotes IQGAP3 localization to non-membrane organelles and cell junctions.

A IQGAP3-biotinylated proteins exclusively found in cells treated with L-cells conditioned media (−Wnt3a). Shown are proximity partners with Fold change ≥ 4, p-value ≤ 0.01, node color denote degree distribution, created with Cytoscape. B IQGAP3-biotinylated proteins exclusively found in cells treated with L-Wnt3a cells conditioned media (+Wnt3a). Shown are proximity partners with Fold change ≥ 4, p-value ≤ 0.01, node color denote degree distribution, created with Cytoscape. C ClueGO analysis for Biological Process enrichment of biotinylated protein in HEK293 cells treated with −Wnt3a and +Wnt3a conditioned media. Size denote number of mapped genes and node color denote p-values. D ClueGO analysis for Cellular Compartment enrichment of biotinylated protein in HEK293 cells treated with −Wnt3a and +Wnt3a conditioned media. Size denote number of mapped genes and node color denote p-values.

Fig. 3. IQGAP3 undergoes phase separation.

A Immunoblot analysis of Doxycycline induction in HeLa Tet-On cells. Cells were induced with Doxycycline for 48 h prior to cell lysis. B Live images of HeLa Tet-On cells expressing EGFP-empty and EGFP-IQGAP3 without and with Wnt3a treatment after 30 min. Cells were induced with Doxycycline for 48 h prior to imaging. Images of untreated and post-treated cells are not the same cells but are the best representative examples. Scale bars, 10 µm. C Timelapse images of HeLa Tet-On cells expressing EGFP-IQGAP3 0–50 min post-Wnt3a treatment. Cells were induced with Doxycycline for 48 h prior to imaging. Images of untreated and post-treated cells are the same cells (refer to timelapse video in Supplementary). Scale bars, 10 µm. D Timelapse images of HeLa Tet-On cells expressing EGFP-IQGAP3 treated with either 1% 2,5-Hexanediol (control) or 1% 1,6-Hexanediol for 0–5 min. Images of untreated and post-treated cells are the same cells (refer to timelapse video in Supplementary). Scale bars, 10 µm. E HeLa Tet-On cells stably expressing EGFP-IQGAP3 were subjected to FRAP bleaching and observed for recovery and fusion events. (For FRAP video refer to supplementary). Scale bars, 2 µm. F Mean normalized standard deviation of FRAP recovery of 5 bleached EGFP-IQGAP3 condensates, half time of recovery (τ1/2) = 30.87 s.

Fig. 4. IQGAP3 co-localizes with Axin1 puncta in the cytosol.

HeLa cells were transfected with (A) mCherry2-empty and Cross-sectional line for Intensity plot measurement from 5 different cells (right). B mCherry-IQGAP1 and Cross-sectional line for Intensity plot measurement from 5 different cells (right). C mCherry2-IQGAP3 and Cross-sectional line for Intensity plot measurement from 5 different cells (right). Scale bars, 10 µm. D Percentile of cell count for IQGAP-distribution, cells with higher membrane intensity were counted as Membrane > Cytosol (Mem > Cyto) vice versa. Cell count for IQGAP distribution, 50 cells per sample. Representative data were collected and are expressed as the mean ± SD from three independent experiments (n = 3). E Cell count for IQGAP-Axin1 co-localization, 50 cells per sample. Representative data were collected and are expressed as the mean ± SD from three independent experiments (n = 3). Student’s t test was performed, with ∗p ≤ 0.05, ∗∗p ≤ 0.01, and ∗∗∗p ≤ 0.001. HeLa cells were transfected with F–H mCherry2-empty with EGFP-empty (negative control), I–K mCherry2-empty with EGFP-Axin1, L–N mCherry2-IQGAP1 with EGFP-Axin1, and O–Q mCherry2-IQGAP3 with EGFP-Axin1. Scale bars, 10 µm.

Fig. 5. IQGAP3 expression reduces Axin1-CK1α interaction.

A Immunoblot of endogenous IQGAP3 immunoprecipitation using an IQGAP3-specific antibody (1:100) and Mouse IgG as a control, with 1 mg of HEK293 cell lysate. Samples were subjected to SDS-PAGE (7.5%) followed by immunoblotting using the indicated antibodies. B HEK293T cells were exposed to Wnt3a conditioned medium at different time points as indicated. Stimulated lysates were subjected to immunoprecipitation using an Axin1-specific antibody and Rabbit IgG as control and were subjected to SDS-PAGE (7.5%) followed by immunoblotting using the indicated antibodies. C HEK293T cells were exposed to Wnt3a conditioned medium at different time points as indicated. Stimulated lysates were subjected to IQGAP3 immunoprecipitation using Myc-Tag antibody and Mouse IgG as control and were subjected to SDS-PAGE (7.5%) followed by immunoblotting using the indicated antibodies. D Immunoblot of Flag-Immunoprecipitated protein expressing HA-Axin1 C-terminal truncations with Flag-IQGAP3 to map the IQGAP3 binding site within Axin1. Samples were subjected to SDS-PAGE (7.5%) followed by immunoblotting using the indicated antibodies. E IQGAP3 reduced exogenous Axin1-CK1 interaction. HEK293T cells expressing Flag-Axin1 were transfected with limiting amounts of HA-CK1α and increasing amounts of Myc-IQGAP3 (5 µg/10 µg/15 µg) as indicated, and the lysates were subjected to anti-Flag immunoprecipitation. Samples were subjected to SDS-PAGE (7.5%) followed by immunoblotting using the indicated antibodies. The relative immunoblot bands of HA and Flag (HA/Flag) from the immunoprecipitation membrane were quantified by densitometry. F IQGAP3 reduced endogenous Axin1-CK1 interaction. HEK293T cells were transfected with increasing amounts of Myc-IQGAP3 (5 µg/10 µg/15 µg) as indicated, and the lysates were subjected to anti-Axin1 immunoprecipitation. Samples were subjected to SDS-PAGE (10%) followed by immunoblotting using the indicated antibodies. The relative immunoblot bands of CK1α and Axin1 (CK1α/Axin1) from the immunoprecipitation membrane were quantified by densitometry.

Fig. 6. IQGAP3 promotes β-catenin activity through its CC and IQ domains.

A TOPFlash assay of β-catenin co-overexpression with Axin1 or increasing IQGAP3 (300 µg, 600 µg) in HEK293T cells. The data is representative of three independent experiments. Student’s t test was performed, with ∗p ≤ 0.05, ∗∗p ≤ 0.01, and ∗∗∗p ≤ 0.001. B TOPFlash assay of β-catenin co-overexpression with siControl or increasing siIQGAP3 (20 nM, 50 nM) in HEK293T cells. The data is representative of three independent experiments. Student’s t test was performed, with ∗p ≤ 0.05, ∗∗p ≤ 0.01, and ∗∗∗p ≤ 0.001. C Immunoblot of IQGAP3 overexpression with and without Wnt3a conditioned media treatment. The data is representative of three independent experiments. D TOPFlash and Immunoblot assays of β-catenin overexpression with a series of IQGAP3 domain deletions. The data is representative of three independent experiments. HeLa cells transfected with E EGFP-empty, F EGFP-Axin1, and G EGFP-CC. Scale bars, 10 µm. H HeLa cells were transfected with EGFP-CC, subjected to FRAP bleaching, and observed for recovery and fusion events. Scale bars, 2 µm. I Mean normalized standard deviation of FRAP recovery of 10 bleached EGFP-CC condensates, half time of recovery (τ1/2) = 7.82 s. HeLa cells transfected with (J–L) EGFP-IQ domain and (M–O) mCherry2-IQGAP3ΔIQ. Scale bars, 10 µm. P Percentile of cell count for protein distribution, cells with higher nuclear intensity were counted as Nuclear > Cytosol (Nuc > Cyto), vice versa. Cell count for protein distribution, 50 cells per sample. Representative data were collected and are expressed as the mean ± SD from three independent experiments (n = 3).

Fig. 7. IQGAP3 overexpression increases β-catenin levels in MKN28 cells.

A Immunoblot of non-cancer/immortalized (HEK293T and HFE145) and gastric cancer cell lines (MKN1, MKN28, AGS, and NUGC3). Samples were subjected to SDS-PAGE (7.5%) followed by immunoblotting using the indicated antibodies. B Immunoblot of MKN28 cells expressing doxycycline-inducible 3×Flag-IQGAP3. Samples were subjected to SDS-PAGE (7.5%) followed by immunoblotting using the indicated antibodies. C GSEA data from RNA-seq of IQGAP3 overexpression in MKN28 cells, showing top 4 pathways. D Gene ontology: Biological Process Enrichment histogram of RNA-seq data from IQGAP3 overexpression in MKN28 cells. Wnt Signaling pathway represented with red bar. E Gene ontology: Molecular function Enrichment histogram of RNA-seq data from IQGAP3 overexpression in MKN28 cells. Wnt-protein and beta-catenin binding are represented with red bars. F KEGG pathway Enrichment histogram of RNA-seq data from IQGAP3 overexpression in MKN28 cells. Wnt Signaling pathway represented with red bar. G IQGAP3 interactome mapped from MKN28 cells stably expressing doxycycline-inducible TurboID-IQGAP3 (Fold change ≥ 2, p-value ≤ 0.05), node color denotes degree distribution, created with Cytoscape.

Fig. 8. Loss of IQGAP3 reduces β-catenin levels in NUGC3 and AGS cells.

A Immunoblot of NUGC3 wild-type and IQGAP3 CRISPR knock-out clones. Samples were subjected to SDS-PAGE (7.5%) followed by immunoblotting using the indicated antibodies. The relative immunoblot bands of β-catenin and β-actin (βcat/βact) were quantified by densitometry. B NUGC3 cells were lysed, and RNA collected were converted to cDNA and subjected to RT-PCR to quantify the amount of Wnt target genes CCND1 and MYC. Representative data were collected and are expressed as the mean ± SD from three independent experiments. Student’s t test was performed, with ∗p ≤ 0.05, ∗∗p ≤ 0.01, and ∗∗∗p ≤ 0.001. C XTT Cell Proliferation Assay of NUGC3 wild-type and IQGAP3 CRISPR knock-out clones. Representative data were collected and are expressed as the mean % growth ±SD from three independent experiments. Two-way ANOVA was performed, with ∗p ≤ 0.05, ∗∗p ≤ 0.01, and ∗∗∗p ≤ 0.001. D Clonogenic Assay of NUGC3 wild-type and IQGAP3 CRISPR knock-out clones. E Colony-Forming Efficiency (%) of NUGC3 wild-type and IQGAP3 CRISPR knock-out clones. F Immunoblot of AGS wild-type and IQGAP3 CRISPR knock-out clones. Image is best representative of three independent experiments. Samples were subjected to SDS-PAGE (7.5%) followed by immunoblotting using the indicated antibodies. The relative immunoblot bands of β-catenin and α-Tubulin (βcat/αTub) were quantified by densitometry. G AGS cells were lysed, and RNA collected were converted to cDNA and subjected to RT-PCR to quantify the amount of Wnt target genes CCND1 and MYC. Representative data were collected and are expressed as the mean ± SD from three independent experiments. Student’s t test was performed, with ∗p ≤ 0.05, ∗∗p ≤ 0.01, and ∗∗∗p ≤ 0.001. H XTT Cell Proliferation Assay of AGS wild-type and IQGAP3 CRISPR knock-out clones. Representative data were collected and are expressed as the mean % growth ±SD from three independent experiments. Two-way ANOVA was performed, with ∗p ≤ 0.05, ∗∗p ≤ 0.01, and ∗∗∗p ≤ 0.001. I Clonogenic Assay of AGS wild-type and IQGAP3 CRISPR knock-out clones. J Colony-Forming Efficiency (%) of AGS wild-type and IQGAP3 CRISPR knock-out clones.

Fig. 9. IQGAP3 is a target of Wnt signaling, sustains Wnt signaling via a positive feedback loop.

A HEK293T cells were treated with wnt3a conditioned media supplemented with 100 ng/ml of rhWnt3a. Cells were harvested, and the RNA collected were converted to cDNA and subjected to RT-PCR to quantify the amount of AXIN2, CCND1, MYC, LEF1, TCF7, TCF7L2 and IQGAP3 transcripts levels. B Time-course of HEK293T cells treated with Wnt3a conditioned media supplemented with 100 ng/ml of rhWnt3a. Image is best representative of three independent experiments. Samples were subjected to SDS-PAGE (7.5%) followed by immunoblotting using the indicated antibodies. The relative immunoblot bands of IQGAP3 and β-actin (IQ3/βact) were quantified by densitometry. C Luciferase assay of 5’ truncation of IQGAP3 promoter region upon treatment with L-cells conditioned or wnt3a conditioned media. D Luciferase assay of Site1 and Site2 within the IQGAP3 promoter region upon treatment with L-cells conditioned or wnt3a conditioned media. E Luciferase assay of Site1, Site1a mutant, Site1b mutant and Site1ab mutant upon treatment with L-cells conditioned or wnt3a conditioned media. F Time-course of HeLa cells treated with 5 µM of XAV939. Samples were subjected to SDS-PAGE (7.5%) followed by immunoblotting using the indicated antibodies. All the above data are representative of three independent experiments.

Fig. 10. Proposed model for IQGAP3 role in Wnt signaling.

We proposed that In IQGAP3-low cells (A), the destruction complex phosphorylates β-catenin, targeting it for degradation. Meanwhile, in IQGAP3-high cells (B), IQGAP3 competes with CK1α for binding to Axin1 or sequesters it away from Axin1 or both, thus, inhibiting CK1α-mediated β-catenin phosphorylation. This inhibition leads to the accumulation of β-catenin, which promotes the expression of Wnt target genes, including IQGAP3, created with BioRender.com.