The USP46 complex deubiquitylates LRP6 to promote Wnt/β-catenin signaling

Abstract

The relative abundance of Wnt receptors plays a crucial role in controlling Wnt signaling in tissue homeostasis and human disease. While the ubiquitin ligases that ubiquitylate Wnt receptors are well-characterized, the deubiquitylase that reverses these reactions remains unclear. Herein, we identify USP46, UAF1, and WDR20 (USP46 complex) as positive regulators of Wnt signaling in cultured human cells. We find that the USP46 complex is similarly required for Wnt signaling in Xenopus and zebrafish embryos. We demonstrate that Wnt signaling promotes the association between the USP46 complex and cell surface Wnt coreceptor, LRP6. Knockdown of USP46 decreases steady-state levels of LRP6 and increases the level of ubiquitylated LRP6. In contrast, overexpression of the USP46 complex blocks ubiquitylation of LRP6 by the ubiquitin ligases RNF43 and ZNFR3. Size exclusion chromatography studies suggest that the size of the USP46 cytoplasmic complex increases upon Wnt stimulation. Finally, we show that USP46 is essential for Wnt-dependent intestinal organoid viability, likely via its role in LRP6 receptor homeostasis. We propose a model in which the USP46 complex increases the steady-state level of cell surface LRP6 and facilitates the assembly of LRP6 into signalosomes via a pruning mechanism that removes sterically hindering ubiquitin chains.

Fig. 1. The USP46 complex is a positive regulator of Wnt signaling.

A–C Overexpression of the USP46 complex stimulates Wnt activation and stabilizes β-catenin levels. HEK293 STF cells were transfected and treated with recombinant Wnt3a (10 ng/ml) for 24 h as indicated. A Individual and B pairwise overexpression of USP46, FLAG-tagged WDR20, and FLAG-tagged UAF1 in HEK293 STF cells in the presence of Wnt3a leads to lower Wnt reporter (TOPFlash) activity compared to overexpression of all three members. The asterisk in (A) indicates a longer exposure for the USP46 blot. C Overexpression of the USP46 complex (Tri46) in HEK293 STF cells potentiates Wnt3a-mediated TOPFlash and promotes β-catenin stabilization. For transfections, the amount of total DNA was kept constant by the addition of vector plasmid. p values compare vector-transfected with USP46 complex-transfected cells. Tubulin is loading control. D, E Knockdown of USP46 and UAF1 by siRNA inhibits Wnt signaling and decreases β-catenin levels. HEK293 STF cells were transfected with non-targeting control (NT) or two independent (D) USP46 or (E) UAF1 siRNAs and treated with recombinant Wnt3a. Immunoblotting confirmed the knockdown of USP46 and UAF1 proteins. GAPDH is loading control. All graphs show mean ± SD of TOPFlash normalized to vector control in the presence of Wnt. p values compare NT treated to USP46 or UAF1 siRNA-treated cells. p ≥ 0.05 is not significant (ns). Significance was analyzed by one-way ANOVA followed by Tukey’s multiple comparisons test. Graphs all show a representative of n = 3 biologically independent experiments. All immunoblots are representative of at least three independent experiments. Source data are provided as a Source data file.

Fig. 2. The USP46 complex regulates Wnt signaling in Xenopus and zebrafish.

A, B The USP46 complex induces secondary axis formation and Wnt target gene expression in Xenopus embryos. A Xenopus embryos were injected ventrally at the 4-cell stage with Xwnt8 mRNA (10 pg), individual USP46 members (1 ng), or a 1:1:1 mixture (0.33 ng each). The percentage of axis duplication is shown with absolute numbers on the top of each bar and representative images. p-values for injection of individual components versus the complex and for injected embryos compared to wildtype is <0.0001. Scale bar = 200 mm. B RT-PCR show induction of Xenopus Wnt target genes, Xnr3 and Chordin. Expression is shown as a ratio of Ornithine decarboxylase (Odc) normalized to Xwnt8 injected animal caps. Control is un-injected. p-values compare injections of individual components versus the USP46 complex. C, D Knockdown of the USP46 complex in zebrafish induces a cyclopic phenotype and reduction of Wnt target gene transcripts, which are rescued by co-injecting homologous human mRNAs. C Embryos were injected at the single-cell stage with Morpholino oligonucleotides (MO, 3 ng) and rescued with mRNAs (0.8 ng). The percentage of cyclopic embryos is shown with absolute numbers on the top of each bar. Arrows show developing eyes. p-values compare MO-injected versus rescue. p-values for all injected embryos compared to wildtype is <0.0001. Scale bar = 200 mm. D mRNAs were isolated, and Lef1 and CyclinD1 levels were quantified by RT-PCR. Gene expression is graphed as a ratio to β-actin control and normalized to un-injected embryos. p-values shown compare MO-injected versus un-injected embryos and morpholino-injected versus corresponding rescued embryos. B, D Graphs show mean ± SEM, n = 3 independent pools of embryos. Significance was analyzed by two-tailed Student’s t test. Source data are provided as a Source data file.

Fig. 3. The USP46 complex acts at the level of the Wnt receptor, LRP6.

A, B USP46 is upstream of the β-catenin destruction complex. A USP46 depletion does not block the stabilization of β-catenin upon AXIN1 siRNA knockdown. HEK293 cells were transfected with USP46 and AXIN1 siRNAs as indicated and immunoblotted for β-catenin. Immunoblotting for AXIN1 and USP46 confirmed their knockdown. B The tankyrase inhibitor, XAV939, which stabilizes AXIN to promote β-catenin degradation, inhibits β-catenin stabilization mediated by the USP46 complex. Cells were transfected and treated with Wnt3a in the absence or presence of XAV939 (1 μM) as indicated and immunoblotted for β-catenin, USP46, FLAG-UAF1, and FLAG-WDR20. C, D Overexpression and knockdown of the USP46 complex increases and decreases steady-state levels of LRP6, respectively. Cells were transfected as indicated, incubated in the absence or presence of Wnt3a, and immunoblotted for LRP6. C HEK293 cells were transfected with the USP46 complex (Tri46) in the presence of Wnt3a. D HEK293 cells were transfected with non-targeting (NT) or USP46 siRNAs. Graph shows mean ± SD of TOPFlash normalized to non-targeting control in the presence of Wnt. p-values compare cells incubated with NT versus USP46 siRNA in the presence of Wnt3a. Significance was analyzed by two-tailed Student’s t-test. The graph shows a representative of n = 3 biologically independent experiments. E Wnt signaling promotes the association between the USP46 complex and endogenous LRP6. HEK293 cells were transfected with USP46 complex components and treated with Wnt3a as indicated, FLAG-UAF1 (*) and FLAG-WDR20 (**) immunoprecipitated (IP) with anti-FLAG conjugated beads, and co-immunoprecipitated LRP6 detected by immunoblotting. WCL whole cell lysates. F The USP46 complex increases cell-surface levels of LRP6. HEK293 cells were transfected with USP46 complex components and treated with Wnt3a, as indicated. Cells were then surface biotinylated, lysates subjected to neutravidin-pulldown, and immunoblotted for endogenous LRP6 and insulin (IR, control) receptors. WCL whole cell lysates. Tubulin and GAPDH are loading controls. All immunoblots are representative of at least three independent experiments. Source data are provided as a Source data file.

Fig. 4. The USP46 complex associates with large complexes independent of LRP6 signalosomes in the presence of Wnt.

A Wnt signaling induces the formation of high molecular weight USP46 complexes as assessed by size exclusion chromatography. HEK293 cells were treated in the absence or presence of Wnt3a, lysates prepared, and high-speed (100,000 × g) supernatants passed over a Superdex 200 FPLC column. Fractions obtained were subjected to TCA precipitation followed by immunoblotting. B The USP46 complex does not co-fractionate with high molecular weight LRP6 aggregates (signalosomes) on sucrose gradient centrifugation. HEK293 cells were incubated in the absence or presence of Wnt3a, Triton X-100 lysates were collected, and sucrose density gradient (15–40%) sedimentation was performed. Fractions were precipitated with chloroform-methanol extraction and analyzed by immunoblotting. The predicted kDa is based on the elution profile of a set of protein standards. Immunoblots show a representative of three independent experiments. Source data are provided as a Source data file.

Fig. 5. The USP46 complex deubiquitinates LRP6 and opposes the action of the Wnt receptor E3 ligase, RNF43.

A, B The catalytic activity of USP46 is required for its regulation of LRP6 levels. Cells expressing the wild-type or mutant USP46 complexes were incubated in the absence or presence of Wnt3a, and immunoblotting was performed. A Overexpression of the USP46 complex containing the catalytically dead USP46^C44S (Tri46^C44S) does not increase LRP6 levels. B Overexpression of the USP46 complex containing the USP46 binding mutant, FLAG-UAF1^S170Y, does not increase LRP6 levels. GAPDH is loading control. C–E His-ubiquitylation assays. HEK293 cells were transfected as indicated, lysed under denaturing conditions, and His-Ub (hexahistidine-tagged ubiquitin) modified proteins isolated by nickel affinity purification. LRP6 and FLAG-tagged proteins were detected by immunoblotting with anti-LRP6 and anti-Flag antibodies, respectively. C Overexpression of FLAG-RNF43 promotes endogenous ubiquitylation of LRP6, which is opposed by overexpression of the USP46 complex. D Overexpression of FLAG-RNF43 in LF203 cells similarly promotes LRP6-FLAG ubiquitylation, which is opposed by overexpression of the USP46 complex. E Knockdown of USP46 enhances LRP6-FLAG ubiquitination. WCL = whole cell lysates. F The sensitivity of intestinal organoids to USP46 depletion is RSPO-dependent. Exogenous RSPO is essential for culturing intestinal organoids in vitro. Intestinal organoids infected with a control lentivirus or two independent lentiviruses expressing USP46 shRNAs were grown with decreasing RSPO conditioned media (CM). Viability was assessed by Cell-Titer Glo. p ≥ 0.05 is not significant (ns). Significance was analyzed by one-way ANOVA followed by Tukey’s multiple comparisons test. Graphs show mean ± SD (n = 3 independent experiments). GAPDH is loading control. All immunoblots are representative of at least three independent experiments. Source data are provided as a Source data file.

Fig. 6. Model of USP46 complex regulation of LRP6 in the Wnt pathway.

A The USP46 complex catalyzes the deubiquitylation of LPR6, opposing the activities of RNF34 and ZNFR3 (and potentially other E3 ligases) to maintain low homeostatic levels of LRP6. B Upon Wnt ligand binding to Frizzled and LPR6, the USP46 complex is recruited to and deubiquitylates LPR6, allowing LRP6 to be assembled into the Wnt signalosome.