Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Oct 3.
Published in final edited form as: Sci Signal. 2022 Aug 23;15(748):eabo2820. doi: 10.1126/scisignal.abo2820

PI(4,5)P2-stimulated positive feedback drives the recruitment of Dishevelled to Frizzled in Wnt-β-catenin signaling

Jacob P Mahoney 1, Elise S Bruguera 1, Mansi Vasishtha 1, Lauren B Killingsworth 1, Saw Kyaw 1, William I Weis 1,*
PMCID: PMC9528458  NIHMSID: NIHMS1834245  PMID: 35998232

Abstract

In the Wnt-β-catenin pathway, Wnt binding to Frizzled (Fzd) and LRP5 or LRP6 (LRP5/6) coreceptors inhibits the degradation of the transcriptional coactivator β-catenin by recruiting the cytosolic effector Dishevelled (Dvl). Polymerization of Dvl at the plasma membrane recruits the β-catenin destruction complex, enabling the phosphorylation of LRP5/6, a key step in inhibiting β-catenin degradation. Using purified Fzd proteins reconstituted in lipid nanodiscs, we investigated the factors that promote the recruitment of Dvl to the plasma membrane. We found that the affinity of Fzd for Dvl was not affected by Wnt ligands, in contrast to other members of the GPCR superfamily for which the binding of extracellular ligands affects the affinity for downstream effectors. Instead, Fzd-Dvl binding was enhanced by increased concentration of the lipid PI(4,5)P2, which is generated by Dvl-associated lipid kinases in response to Wnt and which is required for LRP5/6 phosphorylation. However, binding to Fzd did not promote Dvl polymerization, which is required for signaling downstream of Fzd. Our findings suggest a positive feedback loop in which Wnt-stimulated local PI(4,5)P2 production enhances Dvl recruitment and further PI(4,5)P2 production to support Dvl polymerization, LRP5/6 phosphorylation, and β-catenin stabilization.

Introduction

The transcriptional coactivator β-catenin controls cell differentiation and tissue regeneration, and dysregulated β-catenin signaling can cause several diseases (1). β-catenin activity is stimulated by members of the Wingless/Int1 (Wnt) family of secreted growth factors. In the absence of Wnt, β-catenin is bound in a “destruction complex” containing glycogen synthase kinase-3 (GSK-3), casein kinase 1 (CK1), and the scaffolding proteins Axin and Adenomatous Polyposis Coli (APC). This complex phosphorylates β-catenin to promote its ubiquitination and proteolytic degradation. Wnt binding to its coreceptors, the seven-transmembrane Frizzled (Fzd) and the single-transmembrane low-density lipoprotein receptor-related protein 5 or 6 (LRP5/6), inhibits β-catenin destruction. In this pathway, Fzd recruits the cytoplasmic protein Dishevelled (Dvl), which in turn recruits the destruction complex by binding to Axin, ultimately promoting phosphorylation of the LRP5/6 cytoplasmic tail by GSK-3. Phosphorylated LRP5/6 inhibits GSK-3 activity (2, 3), thereby slowing β-catenin degradation and allowing β-catenin to translocate to the nucleus. The phosphorylation of LRP5/6 also requires a Wnt-stimulated increase of the plasma membrane lipid phosphatidyl inositol 4,5-bisphosphate [PI(4,5)P2], mediated by lipid kinases that associate with Dvl (4–6). PI(4,5)P2 has been suggested to support LRP5/6 phosphorylation by several mechanisms downstream of Dvl recruitment, including displacement of proteins occluding LRP5/6 phosphorylation, recruitment of additional cytoplasmic proteins that aid in Axin recruitment, or promotion of receptor clustering in clathrin-coated pits (7–11).

Dvl is a multifunctional scaffolding protein that contains three structured domains: DIX, PDZ, and DEP. The Dvl PDZ and DEP domains bind to, and the DEP domain stimulates the activity of, the lipid kinase PI4KIIα, whereas the DIX domain binds and activates the lipid kinase PIP5K1, consistent with the notion that recruitment of Dvl to the plasma membrane can produce an increase in local PI(4,5)P2 concentration (4–6). The DIX domain also mediates dynamic polymerization of Dvl, which provides the avidity necessary to directly recruit the destruction complex by binding to a homologous domain in Axin, designated DAX (12–14). Dvl polymers can be crosslinked by dimerization mediated through a domain-swapping interaction between neighboring DEP domains (15, 16), but it is unclear whether DEP dimerization is triggered upon membrane recruitment or occurs solely due to the increased local concentration of Dvl. It is thought that Wnt-stimulated recruitment of Dvl to the plasma membrane occurs by a direct Fzd-Dvl interaction (16–22). Both the PDZ and DEP domains of Dvl bind weakly (dissociation constants (KD) in the tens of μM range) to peptides derived from Fzd (23, 24) and may interact with membrane phospholipids in support of signaling (25–28). Deletion of the DEP, but not PDZ, domain impairs Dvl membrane localization and Wnt-β-catenin signaling when mutant Dvl is present in endogenous amounts, and the isolated DEP, but not PDZ, domain can be recruited to the plasma membrane by Fzd (16, 19–21, 29). These findings suggest that Wnt signals recruit Dvl to the plasma membrane by enhancing the Fzd-DEP interaction.

The ten human Frizzled proteins, along with the closely related receptor Smoothened, form a distinct phylogenetic group (class F) within the G protein–coupled receptor (GPCR) superfamily (30). Class F GPCRs feature an extracellular ligand-binding cysteine-rich domain (CRD) connected to the transmembrane domain. Multiple factors, including agonist binding and the local membrane environment, promote conformational changes within GPCRs that enable their interactions with downstream transducers (31–35). Each Fzd can be stimulated by a subset of the 19 human Wnt proteins, and the secreted transforming growth factor β (TGF-β) family protein Norrin stimulates β-catenin signaling selectively through Fzd4. Wnts and Norrin stabilize conformational changes within Fzd, perhaps in support of G protein coupling (36–38). However, G proteins appear to be dispensable for Wnt-β-catenin signaling (39), and a cryo-electron microscopy structure of a Wnt8-FZD5 complex did not reveal interactions between the Wnt-bound extracellular domain and the transmembrane domain of FZD5 (40). Thus, it is unclear whether ligands shift the conformational equilibrium of Fzds to allosterically enhance Fzd-DEP affinity.

Here, we investigated factors that promote Dvl recruitment to the plasma membrane. Membrane PI(4,5)P2 and phosphatidyl inositol 3,4,5-trisphosphate [PI(3,4,5)P3] selectively enhanced the binding affinity between purified Fzd and the Dvl DEP domain, and depletion of PI(4,5)P2 diminished Dvl membrane recruitment in cells. The Fzd-DEP interaction was unaffected by extracellular ligand binding to monomeric or dimeric Fzd, and reconstitution of Fzd with the coreceptor LRP6 did not affect Fzd-DEP binding affinity in the absence or presence of ligand. Furthermore, binding of the DEP domain to Fzd did not alter the oligomeric state of the DEP domain. These results indicate that Dvl accumulation at the membrane is not triggered by ligand- or coreceptor-dependent allosteric enhancement of the Fzd-DEP interaction, nor by receptor-catalyzed DEP domain swapping and Dvl dimerization. Instead, our findings support a model in which Dvl-dependent PI(4,5)P2 production, which may be initiated by receptor clustering, creates positive feedback to enhance Dvl membrane residency and oligomerization in support of Wnt-β-catenin signaling.

Results

PI(4,5)P2 is required for Dvl binding and recruitment to Fzd

Based on the homology between Fzds and classical GPCRs, and the ability of the Dvl DEP domain to interact with Fzd and with anionic phospholipids, we hypothesized that the Fzd-DEP interaction would be stimulated by ligand binding to Fzd and modulated by the local membrane environment. We purified full-length mouse Fzd4 and inserted it into reconstituted high-density lipoprotein particles (nanodiscs). We assembled nanodiscs using conditions that favor incorporation of monomeric GPCRs (fig. S1A) (41), and the resultant nanodisc preparations showed an average stoichiometry of 1.1 Fzd4 per nanodisc (fig. S1, B to D). We purified the DEP domain of mouse Dvl2 and monitored its binding to Fzd4, testing a range of nanodisc lipid compositions based on lipids previously reported to support Dvl membrane recruitment (fig. S2, A and B) (25–27). DEP domain binding to Fzd4 was markedly enhanced when PI(4,5)P2 was incorporated into the nanodiscs (Fig. 1, A and B; fig. S2C). We did not detect binding between Fzd4 and purified Dvl2 PDZ domain, regardless of lipid composition (fig. S2B), indicating that PI(4,5)P2 selectively enhanced the Fzd-DEP interaction.

Fig. 1. PI(4,5)P2 enhances Dvl DEP domain binding to multiple Fzd subtypes.

Fig. 1.

Open in a new tab

(A) Purified Fzd4 was inserted into nanodiscs composed of the indicated molar ratios of 16:0–18:1 phosphatidylcholine (POPC) and cholesterol, supplemented with charge-matched amounts of 16:0–18:1 phosphatidylglycerol (POPG), phosphatidylserine (POPS), phosphatidic acid (POPA), or porcine brain PI(4,5)P2. Binding of Dvl2 DEP domain to Fzd4 was monitored using biolayer interferometry (BLI). Shown are representative association and dissociation traces (from n=3 independent experiments) of 10 μM Dvl2 DEP domain binding to Fzd4 inserted into nanodiscs containing the indicated molar percentages of lipids. (B) Representative BLI traces of DEP binding to Fzd4 in nanodiscs containing 75% POPC, 20% cholesterol, and 5% PI(4,5)P2. (C) Concentration-response curves of DEP binding to Fzd4 in nanodiscs prepared with increasing amounts of PI(4,5)P2. Data points are shown as mean ± SEM from n=8 independent experiments with 5% PI(4,5)P2 and n=3 independent experiments for all other amounts of PI(4,5)P2. (D) Fzd4-DEP binding affinities determined from the experiments plotted in (C). (E) Effect of increasing nanodisc PI(4,5)P2 content on the association and dissociation rate constants (Kon and Koff) for DEP binding to Fzd4, determined from the same set of experiments as (C). (F) Fzd4 was inserted into nanodiscs containing the indicated molar ratios of POPC and cholesterol along with 18:0–20:4 PI, PI(4,5)P2, PI(3,4,5)P3, or porcine brain PI(4)P (5% PI(4,5)P2, n=4; 3% PI(3,4,5)P3, n=4; 5% PI(3,4,5)P3, n=3 independent experiments). (G and H) Representative binding traces (from n=3 independent experiments) of 1 μM DEP domain binding to FZD5 (G) or FZD7 (H) in nanodiscs with or without 5% PI(4,5)P2. All data shown in this figure were replicated using Fzd and DEP from at least two independent purifications. All rate constants and affinity values are provided in the Supplementary Materials (table S2).

PI(4,5)P2 is localized on the cytoplasmic leaflet of the plasma membrane and accounts for approximately 1% of the membrane lipid, with density estimates ranging from 4,000–60,000 molecules/μm2 (42). Because Wnt stimulation increases membrane PI(4,5)P2 content (4, 5), we measured Fzd4-DEP affinity in nanodiscs containing 1–20% PI(4,5)P2. These nanodiscs are expected to have an average of 0.4–8 molecules of PI(4,5)P2 in the “cytoplasmic” leaflet of the nanodisc membrane, corresponding to a PI(4,5)P2 density of ~9,000–180,000 molecules PI(4,5)P2/μm2 (table S1). Increasing the amount of PI(4,5)P2 across this range progressively enhanced DEP association and slowed its dissociation to increase the observed Fzd4-DEP affinity: DEP bound to Fzd4 with approximately 50-fold higher affinity in discs containing 20% PI(4,5)P2 (KD = ~40 nM) compared to discs containing 1% PI(4,5)P2 (KD = ~2 μM) (Fig. 1, C to E; fig. S2C; table S2). Background binding of DEP to nanodiscs lacking Fzd4 was identical with and without PI(4,5)P2 (fig. S3, A and B), and we could not co-sediment DEP with lipid vesicles containing 3% PI(4,5)P2 (fig. S3C). These results are consistent with previous work suggesting that the Dvl DEP domain has low affinity for phosphoinositides (25) and demonstrate that the DEP binding we observed was not solely the result of DEP binding to PI(4,5)P2. In subsequent experiments, we utilized Fzd reconstituted in nanodiscs containing 5% PI(4,5)P2, an amount near the physiological PI(4,5)P2 concentration and which maximizes the likelihood that each Fzd-containing nanodisc will contain at least one PI(4,5)P2 molecule in the “cytoplasmic” leaflet of the membrane (table S1).

In cells, PI(4,5)P2 is primarily synthesized by lipid kinases that phosphorylate the headgroup of phosphatidylinositol (PI) to produce multiple phosphoinositide lipid species. Each of these lipids is typically built upon a distinct acyl chain backbone (18:0–20:4) (43), different than the 16:0–18:1 lipids used for the other components of nanodisc formation. Therefore, we sought to determine whether the acyl chain or headgroup of PI(4,5)P2 was responsible for enhancing Fzd-DEP binding affinity by comparing a series of phosphoinositides bearing 18:0–20:4 acyl chains. Incorporation of 18:0–20:4 phosphatidylinositol (PI) did not stimulate Fzd4-DEP binding, and although phosphatidylinositol-4-phosphate (PI(4)P) weakly supported DEP binding, the apparent Fzd4-DEP KD was still >10 μM in this case. Both 18:0–20:4 PI(4,5)P2 and PI(3,4,5)P3 enhanced DEP binding affinity to a comparable extent (Fig. 1F and fig. S4, A to D). These results indicate that the effect of PI(4,5)P2 on Fzd-DEP affinity depends principally on phosphorylation of the inositol ring rather than acyl chain–dependent effects on Fzd conformation or change in bilayer fluidity. However, because the composition of the nanodiscs does not perfectly mimic that of the plasma membrane, it is possible that the local acyl chain composition of the membrane modulates the Fzd-Dvl interaction in cells.

The Dvl DEP domain contains two conserved elements: a proposed Fzd-binding loop (the β1/β2 loop) and a cluster of positively charged residues suggested to interact with phosphatidic acid (25)(fig. S5, A and B). Mutations in either region of DEP diminish Dvl membrane recruitment and β-catenin signaling in cells (21, 25, 29) and weakened the Fzd4-DEP interaction in our assays (fig. S5, C to E), suggesting that Dvl membrane recruitment is driven by interaction of DEP with Fzd and the PI(4,5)P2 headgroup, although we cannot rule out that PI(4,5)P2 may affect Fzd conformation. Fzd4 belongs to one of four phylogenetic subgroups of Fzd proteins, so we sought to determine whether DEP-binding sensitivity to PI(4,5)P2 is a more general property of Fzds. Purified human FZD5 or FZD7, which are members of distinct class F GPCR subgroups (30), did not bind DEP when reconstituted in nanodiscs lacking PI(4,5)P2. However, both bound DEP with an affinity comparable to that of Fzd4 when PI(4,5)P2 was included (Fig. 1, G and H; fig. S6, A to D), indicating that PI(4,5)P2 is a general modulator of Fzd-Dvl interactions.

To determine whether PI(4,5)P2 contributes to Fzd-mediated membrane recruitment of Dvl in cells, we used confocal microscopy to examine the effect of PI(4,5)P2 depletion on the subcellular localization of Dvl2 fused to green fluorescent protein (Dvl2-GFP) in HEK293T cells. To rapidly deplete PI(4,5)P2 from the plasma membrane, we used a PI(4,5)P2 degradation system in which a yeast PI(5)-phosphatase fused to cyan fluorescent protein and FK506-binding protein (CFP-FKBP-Ins54P) is recruited to a plasma membrane–targeted FKBP-rapamycin binding domain (Lyn-FRB) upon application of rapamycin (44) (fig. S7). As previously reported (16, 17, 19, 29), Dvl2-GFP showed punctate cytoplasmic localization in the absence of co-transfected Fzd4, and apparent plasma membrane localization upon overexpression of Fzd4 (Fig. 2A). Rapamycin treatment increased the number of Dvl2-GFP puncta in cells transfected with Fzd4, indicating an increase in cytoplasmic Dvl concentration and that plasma membrane PI(4,5)P2 contributes to Fzd-mediated membrane translocation of Dvl (Fig. 2, A and B).

Fig. 2. PI(4,5)P2 contributes to Fzd-dependent Dvl recruitment to the plasma membrane.

Fig. 2.

Open in a new tab

(A) HEK293T cells were transfected with Lyn-FRB, CFP-FKBP-Ins54P, and Dvl2-GFP, with or without HA-Fzd4. Cells were incubated with rapamycin to deplete the plasma membranes of PI(4,5)P2 prior to fixation and staining with an antibody specific for Dvl2. Shown are representative images from three independent transfections. Scale bar is 30 μm. (B) Number of Dvl2 cytoplasmic puncta per cell in each of the conditions represented in panel (A). Data points from each of n=3 independent experiments are shown as blue squares, orange circles, or green triangles, with the total number of cells listed below each distribution. Median values with 95% confidence intervals are indicated with black lines (median and [95% CI] values are: −Fzd4/−Rap, 54 [50, 60]; −Fzd4/+Rap, 40 [30, 48]; +Fzd4/−Rap, 0 [0, 1]; +Fzd4/+Rap, 8 [5, 13]). The median values for all conditions were found to be significantly different using a Mann-Whitney test with Bonferroni correction (p<0.0001). The decrease in number of puncta per cell upon rapamycin treatment of cells lacking transfected Fzd4 (95% CI = [−25, −5] puncta per cell) may partially mask the increase in puncta per cell upon rapamycin treatment of cells transfected with Fzd4 (95% CI = [5, 13] puncta per cell). To account for this effect of rapamycin, we determined a 95% confidence interval for the difference of differences in number of puncta for cells with or without Fzd4 (95% CI = [13, 35]).

Fzd ligands do not allosterically alter affinity for Dvl

After establishing a baseline Fzd-DEP affinity in the absence of agonist, we sought to determine whether this interaction could be allosterically enhanced by Fzd agonists. Pre-saturating Fzd4 with purified human Norrin (fig. S8, A and B) had no effect on Fzd4-DEP affinity in nanodiscs containing 5% PI(4,5)P2 (Fig. 3A), nor did Norrin stimulate DEP binding in discs lacking PI(4,5)P2 (fig. S2B), even though Norrin stimulated activity of a β-catenin–responsive luciferase reporter (TopFlash) in cells (fig. S8C). We also purified complexes of Fzd4 and FZD5 bound to Xenopus laevis Wnt8 (xWnt8) (40) and inserted these complexes into nanodiscs containing 5% PI(4,5)P2 (fig. S9, A to C). Some Wnt was lost during the reconstitution procedure, leaving the reconstituted complexes with an occupancy of 30–40% Wnt-bound Fzd (fig. S9, D to H). Even though the final Wnt:Fzd stoichiometry was sub-saturating, any effect of Wnt on DEP affinity should still be apparent: intermediate (3–30 fold) enhancement of DEP binding affinity should produce a shallow binding curve with Hill slope < 1, whereas a large (≥100-fold) shift in DEP affinity should produce a biphasic binding curve (fig. S10A). Purified xWnt8 produced a TopFlash response in Fzd-knockout HEK293T cells transfected with either Fzd4 or FZD5 (fig. S10B), but bound xWnt8 had no effect on Fzd4-DEP or FZD5-DEP affinities (Fig. 3, B and C; fig. S10, C to H).

Fig. 3. Fzd ligands do not allosterically enhance binding of the Dvl DEP domain.

Fig. 3.

Open in a new tab

(A) Binding of Dvl2 DEP domain to monomeric Fzd4 in nanodiscs containing 5% PI(4,5)P2 with or without pre-saturation with Norrin (100 nM), measured by biolayer interferometry (BLI). (Fzd4 alone, n=7; Norrin + Fzd4, n=4 independent experiments performed with Fzd4, Norrin, and DEP from at least two independent purifications.) (B and C) Fzd4 (B), FZD5 (C), or pre-formed xWnt8-Fzd complexes were inserted into nanodiscs containing 5% PI(4,5)P2 prior to measuring DEP binding affinity by BLI. (Fzd4 alone n=3; xWnt8 + Fzd4 n=3; performed with DEP from two independent purifications and a single preparation of Fzd4 and xWnt8. FZD5 alone, n=4; xWnt8 + FZD5, n=4; performed with FZD5, xWnt8, and DEP from two independent purifications.) (D) The effect of ligands on DEP affinity for homodimeric Fzd4 was assessed as described for the monomeric preparations in panels A-C, but instead pre-saturated with 0.15 nM Norrin for 3 h to reach an approximate 1:1 ratio of bound Norrin dimer to Fzd dimer (Fzd4 dimer, n=6; Norrin + Fzd4 dimer, n=3; xWnt8 + Fzd4 dimer, n=3; performed using two preparations of Fzd4 and DEP and a single preparation of Norrin and xWnt8). (E) The Fzd4-LRP6 heterodimer was inserted into nanodiscs containing 5% PI(4,5)P2 and pre-bound to Norrin (100 nM) on the BLI sensor prior to DEP binding (n=3; performed using DEP from two independent purifications and a single reconstitution of Fzd4/LRP6). All rate constants and affinity values are provided in the Supplementary Materials (table S2).

Structural investigations of isolated Fzd CRDs have raised the possibility that Wnts promote Fzd dimerization or activate a pre-formed Fzd dimer (45–47). Likewise, Norrin is a constitutive dimer (48, 49) that might promote receptor dimerization. Therefore, we sought to determine if Fzd4 dimerization was required to stabilize conformational changes that allosterically enhance binding of the Dvl DEP domain. Using Fzd4 constructs individually fused to complementary fragments of split GFP, we purified a Fzd4 homodimer and reconstituted it into nanodiscs (50). The DEP domain bound monomeric and dimeric preparations of Fzd4 with comparable affinities (table S2), and neither Norrin nor xWnt8 enhanced DEP affinity for the Fzd4 dimer (Fig. 3D and fig. S11, A to D). We used a similar split GFP–based purification and reconstitution strategy to test whether the presence of the coreceptor LRP6 allosterically enhanced the interaction of Fzd4 with Dvl DEP upon ligand binding (50). Again, Norrin did not enhance DEP binding to Fzd4 co-reconstituted with LRP6 in nanodiscs (Fig. 3E and fig. S11, E and F). Taken together, our results show that binding of the Dvl DEP domain to Fzd was not enhanced by allosteric coupling to extracellular ligand binding, ligand-induced Fzd dimerization, or coreceptor-dependent conformational changes. We cannot rule out the possibility that other regions of Dvl may engage Fzd in a ligand-dependent manner, but the DEP domain has been shown to be sufficient for Fzd-dependent membrane recruitment (16, 20).

Fzd does not promote Dvl DEP domain dimerization

Dimerization of Dvl driven by DEP domain swapping after Dvl has bound to Fzd has been proposed as a requirement for enhancing Dvl oligomerization during Wnt-β-catenin signaling (15, 16). In this interaction, helix 1 of the DEP domain is traded from one Dvl molecule to another, and the DEP β1/β2 loop adopts an extended conformation, forming β-strand interactions with the β1/β2 loop of the neighboring DEP. This DEP dimer is kinetically trapped and incapable of binding Fzd (16), so we tested whether Fzd catalyzes DEP dimerization by stabilizing DEP helix 1 rearrangement upon Fzd-DEP binding. If so, DEP helix 1 may exist in an equilibrium between “closed” and “open” states, and binding to Fzd would stabilize an “open” conformation that favors DEP dimerization (Fig. 4A). In this case, movement of DEP helix 1 should be energetically coupled to Fzd binding, and restricting helix 1 motion would be expected to reduce Fzd-DEP binding affinity. However, DEP bearing a G436P mutation, which blocks Dvl dimerization by limiting helix 1 motion (16), bound monomeric Fzd4 with an affinity comparable to that of wild-type DEP (Fig. 4, B and C).

Fig. 4. Fzd4 does not catalyze DEP domain dimerization.

Fig. 4.

Open in a new tab

(A) Cartoon showing a potential mechanism for Fzd-catalyzed DEP dimerization. In this model, Fzd binding to the DEP β1/β2 loop (in purple) stabilizes an “open” conformation of helix 1, thereby facilitating DEP dimerization through domain swapping. After the DEP dimer has been formed, the residues of the β1/β2 loop are engaged in intermolecular contacts within the dimer and are no longer accessible for Fzd binding (16). The dimerization process is expected to be enhanced by receptor dimerization or higher-order clustering. (B) Structure of mouse Dvl1 DEP domain (PDB 1FSH) showing the location of Gly436 (orange sphere; the β1/β2 loop is colored in purple). Mutation of this residue to proline has been shown to decrease DEP dimerization (16). (C) Binding of wild-type or G436P mutant DEP to Fzd4. Data are plotted as mean +/− SEM from n=3 experiments using Fzd4 and DEP from two independent purifications. (D and E) DEP domain was incubated with 3 M urea before SEC on Superdex 75 as previously described (15). (F and G) DEP domain (100 μM) was incubated with 10 nM nanodiscs containing homodimeric Fzd4 bound to Norrin prior to SEC on Superdex 75.

Receptor clustering, or perhaps Fzd dimerization, might also promote DEP dimerization by creating a high local concentration of DEP domain (16). Using size exclusion chromatography to monitor DEP oligomeric state, we found that although urea treatment weakly induced dimerization of the DEP domain (15), incubation of the same concentration of DEP with nanodiscs containing Norrin-bound Fzd4 homodimer failed to do so (Fig. 4, D to G). These findings indicate that Fzd does not catalyze DEP domain swapping. It remains possible, however, that local concentrations of Fzd and Dvl are higher in cells than in nanodiscs and that higher concentrations are necessary to promote DEP dimerization.

Discussion

Wnt-β-catenin signaling appears to depend upon Dvl membrane recruitment driven by increased Fzd-Dvl affinity, and local enrichment of Dvl at the membrane allows for its limited oligomerization to further increase membrane residence time (22). If extracellular Fzd ligands do not allosterically enhance Fzd-DEP affinity, then what is responsible for Wnt-stimulated Dvl membrane accumulation? Our data suggest that Wnt-stimulated production of PI(4,5)P2 (4) provides the enhancement of Fzd-Dvl affinity required for Dvl membrane recruitment. The Fzd4-DEP affinity was 1.2 μM in nanodiscs containing 3% PI(4,5)P2 (Fig. 1C; table S2), a PI(4,5)P2 density that mimics the basal amounts of PI(4,5)P2 in the cytoplasmic leaflet of the plasma membrane (table S1). This affinity is consistent with the estimated 4–20 μM affinity derived from live-cell imaging (22), suggesting that the amount of PI(4,5)P2 in the plasma membrane is sufficient to allow weak Fzd-Dvl association even in the absence of Wnt stimulation. Assuming ~3 μM Fzd-DEP affinity and ~140 nM cellular Dvl concentration (22), occupancy of Fzd by Dvl will be low (< 5% at equilibrium). Ligand binding to Fzd or the Fzd-LRP complex did not enhance Fzd-DEP affinity (Fig. 3, A to E), but increasing the PI(4,5)P2 content of Fzd4-containing nanodiscs progressively increased Fzd4-DEP affinity by affecting both the association and dissociation rates of DEP binding (Fig. 1, C to E; fig. S2C). Wnt stimulation increases total cellular PI(4,5)P2 content 2- to 3-fold (4), increases membrane-localized Dvl ~4-fold (22), and slows Dvl dissociation from the plasma membrane ~2-fold (22). Our binding assays predict this fold-change in cellular PI(4,5)P2 should increase apparent Fzd-DEP affinity 3- to 10-fold, depending on the Fzd-proximal concentration of PI(4,5)P2 in cells, which would substantially enhance Fzd-Dvl occupancy (~12–30% at equilibrium).

Based on these data, we propose that Wnt-stimulated, Dvl-mediated PI(4,5)P2 generation promotes Wnt-β-catenin signaling by initiating a positive feedback loop that stimulates Dvl membrane accumulation (Fig. 5, A to D). The ability of Dvl to recruit Axin depends on binding of the Dvl DIX domain to the Axin DIX domain (DAX). Given the low cytoplasmic concentrations of Dvl and Axin (~140 and ~150 nM, respectively) (22, 51) and the weak DIX-DAX affinity (KD = 10–20 μM) (13, 14), interaction of these two proteins is minimal in the absence of Wnt stimulation. A PI(4,5)P2-dependent increase in local Dvl concentration at the membrane would favor the limited DIX domain oligomerization (KD = 5–20 μM) thought to provide avidity for efficient Axin binding (13, 14) and subsequent LRP5/6 phosphorylation, the critical step for inhibiting destruction complex-resident GSK-3. Thus, rather than stabilizing conformational changes in Fzd that enhance Dvl affinity, an important role of Wnts and other agonists of the β-catenin pathway may be to trigger Dvl-dependent PI(4,5)P2 generation.

Fig. 5. Model for PI(4,5)P2-dependent membrane accumulation of Dvl.

Fig. 5.

Open in a new tab

(A) In the absence of Wnt, the basal amount of PI(4,5)P2 in the plasma membrane supports a weak, transient interaction between Fzd and Dvl. (B) Wnt binding does not allosterically enhance Fzd-DEP affinity to promote Dvl membrane accumulation. Rather, Wnt stimulates PI4KIIα and PIP5K1 activity through the small pool of associated Dvl to increase PI(4,5)P2 concentration in the vicinity of the receptors, possibly by receptor clustering. (C) The local generation of PI(4,5)P2 initiates a positive feedback loop in which PI(4,5)P2 promotes Dvl membrane residency by enhancing DEP recruitment and slowing its dissociation from Fzd. The increase in Dvl membrane residence would support PI4KIIα and PIP5K1 activity, keeping local PI(4,5)P2 concentration high and promoting additional Dvl recruitment in support of Dvl oligomerization through its DIX domain. For clarity, only the DIX domains of additional Dvl molecules are shown in the cartoon. (D) PI(4,5)P2-dependent accumulation of Dvl at the membrane enhances avidity for Axin, whose DAX domain binds Dvl DIX, thereby recruiting the destruction complex, allowing for LRP5/6 phosphorylation and inhibition of GSK-3.

Although ligand-stimulated PI(4,5)P2 generation is required for LRP5/6 phosphorylation (4), an increase in PI(4,5)P2 concentration and Dvl recruitment may not be sufficient for β-catenin stabilization in the absence of extracellular ligand binding. Indeed, Fzd overexpression increases cellular PI(4,5)P2 content without exogenous Wnt treatment (4), but unlike overexpression of LRP5/6 or Dvl, overexpression of Fzd does not activate β-catenin signaling in the absence of ligand (18, 21, 52). The heterodimerization model of Wnt agonism proposes that Wnts, Norrin, or Wnt surrogate ligands stimulate β-catenin signaling by crosslinking Fzd and LRP5/6 (40, 52–56), so the absence of pathway activation upon Fzd overexpression may indicate a failure to efficiently recruit LRP5/6 to the Fzd-Dvl complex. For example, in the absence of ligand, Dvl binding to Fzd promotes the ubiquitination of Fzd (catalyzed by ZNF3 or RNF43), thereby enhancing Fzd internalization and degradation (57). Endocytosis of Fzd is enhanced by a direct interaction between Dvl and the clathrin adapter AP2, and PI(4,5)P2 is known to enhance AP2 recruitment to membrane cargos and their subsequent endocytosis (58). Therefore, overexpression of Fzd and the accompanying increase in PI(4,5)P2 synthesis and Dvl membrane recruitment likely accelerates the clearance of Fzd from the plasma membrane, decreasing the odds of Fzd encountering LRP5/6 unless a ligand is present to promote receptor heterodimerization. It is also possible that overexpression of Fzd may create a non-optimal Fzd:LRP5/6 stoichiometry that hampers signalosome formation. In any case, PI(4,5)P2 likely works in concert with ligand-mediated recruitment of LRP5/6 to increase the local concentration of Dvl in the vicinity of LRP5/6, supporting Dvl- and Axin-mediated LRP5/6 phosphorylation. The mechanism by which extracellular ligands trigger enhanced PI(4,5)P2 generation remains to be elucidated. Given the positive feedback provided by enhanced PI(4,5)P2 synthesis, signaling may be exquisitely sensitive to a small increase in local Dvl concentration, which would in turn increase the local concentration of Dvl-associated PI4K and PIP5K and thereby promote longer-lived Fzd-Dvl interactions in the signalosome. An increase in local Dvl concentration could be initiated by ligand-stimulated receptor clustering. Ligand-dependent clathrin-mediated clustering of receptors at the cell surface has been proposed to be essential for Wnt-β-catenin signaling, perhaps driven by an interaction between Dvl or the LRP5/6 tail with the clathrin adaptor AP2 (11, 16, 58, 59), although this has been challenged by the finding that deletion of AP2 or knockdown of clathrin heavy chain has no effect on Wnt-β-catenin signaling in embryonic stem cells (60). In vertebrates, Amer1 (APC membrane recruitment protein 1, also known as WTX) also aids in receptor clustering and destruction complex recruitment, but its recruitment appears to depend on ligand-stimulated PI(4,5)P2 generation (7, 9). PI(4,5)P2 would also be expected to promote clustering by aiding in recruitment of β-arrestin and AP2 (61–63). Dimerization of the Fzd extracellular domain mediated by Wnt-associated lipids has also been proposed as a mechanism for receptor clustering (45–47), and Norrin is a dimer, but this does not explain the agonist activity of monomeric Wnt surrogate agonists or a monomeric Norrin mutant that would be expected to form a ternary complex with Fzd and LRP5/6 (48, 53–55). Clustering could additionally be mediated by interactions between LRP5/6 molecules. The LRP6 extracellular domain exhibits considerable conformational flexibility in the hinge between the second and third β-propeller–EGF repeats (64, 65), which raises the possibility that ligand binding to either portion shifts the conformational equilibrium of LRP6 to expose sites that enable interaction between LRP6 molecules. Rather than receptor clustering, it is also possible that an LRP5/6-binding protein triggers recruitment or activation of PI4K and PIP5K, or inhibits local degradation of PI(4,5)P2 to allow its accumulation (8). Finally, given that PI(4,5)P2 recruits several actin-binding proteins to the plasma membrane and promotes actin filament formation (66), it is likely that our findings will be relevant to Fzd signaling in the planar cell polarity pathway (67).

Materials and Methods

Expression and purification of Frizzled

Mouse Fzd4 (residues 41–537; 99% sequence identity with human Fzd4), human FZD5 (residues 27–585), or human FZD7 (residues 33–574) were subcloned into pVL1393. Each expression construct contained an N-terminal influenza hemagglutinin signal peptide followed by a FLAG tag. Baculovirus stocks were generated using the BestBac system according to the manufacturer’s instructions (Expression Systems). Spodoptera frugiperda Sf9 cells (Expression Systems) were grown to a density of 2–3 × 106 cells/ml, at which point the cultures were supplemented with Chemically Defined Lipid Extract (1:100 v/v dilution; ThermoFisher) and infected with baculovirus (1:1000 to 1:100 v/v dilution, depending on viral titer).

Cultures were harvested 72 h post-infection and centrifuged for 10 min at 1500 g. Pelleted cells were resuspended in 500 ml lysis buffer (50 mM HEPES-NaOH pH 8.0, 50 mM NaCl, 1 mM EDTA, pH 8.0) supplemented with protease inhibitors (1 μM leupeptin, 0.15 μM aprotinin, 200 μM phenylmethylsulfonyl fluoride, 90 μM N-p-Tosyl-L-phenylalanine chloromethyl ketone, 90 μM Nα-Tosyl-L-lysine chloromethyl ketone) and lysed by nitrogen cavitation (Parr Instrument Company). The lysate was centrifuged for 15 min at 1000 g to pellet nuclei and cell debris. The supernatant was collected and centrifuged for 40 min at 300,000 g. Membrane pellets were collected in 200 ml wash buffer (50 mM HEPES-NaOH pH 8.0, 300 mM NaCl, protease inhibitors) and resuspended using a Dounce homogenizer. Homogenized membranes were centrifuged for 40 min at 300,000 g and the pellets were resuspended in a Dounce homogenizer using 100 ml of 50 mM HEPES-NaOH pH 8.0, 150 mM NaCl supplemented with protease inhibitors. Homogenized membranes were flash frozen by slowly pouring into liquid nitrogen and stored at −80 °C until protein purification.

Frozen membranes were thawed and diluted to 6.25 mg/ml protein as determined by Bradford assay. Glycerol was added to 10% (v/v) final concentration. To solubilize membrane proteins, a mixture of n-dodecyl-β-D-maltopyranoside (DDM; Anatrace) and cholesteryl hemisuccinate (CHS; Anatrace) was added from a 10%/1% (w/v) DDM/CHS stock solution to reach a final concentration of 1%/0.1% DDM/CHS and the solution (final protein concentration of ~5 mg/ml) was incubated with gentle stirring for 1 h at 4 °C. Solubilized membranes were centrifuged for 40 min at 300,000 g and the supernatant (DDM-soluble fraction) was collected and supplemented with CaCl2 to a final concentration of 4 mM.

The DDM-soluble fraction was applied to anti-FLAG sepharose resin. The column was washed with five bed volumes of 20 mM HEPES-NaOH pH 8.0, 500 mM NaCl, 4 mM CaCl2, 0.1% DDM, 0.01% CHS, followed by five bed volumes of 20 mM HEPES-NaOH, 100 mM NaCl, 4 mM CaCl2, 0.05% DDM, 0.005% CHS. One bed volume of elution buffer (20 mM HEPES-NaOH pH 8.0, 150 mM NaCl, 0.025% DDM, 0.0025% CHS, 5 mM EDTA, 100 μM FLAG peptide) was applied to the sealed column, and the column was incubated for 1 h at 4 °C with gentle rotation. Eluted Frizzled was concentrated to 0.5 ml using a VivaSpin 100-kDa cutoff spin concentrator (Sartorius), applied to a Superdex 200 Increase 10/300 GL (Cytiva), and eluted in SEC buffer (20 mM HEPES-NaOH pH 8.0, 150 mM NaCl, 0.012% DDM, 0.0012% CHS). Fractions containing monodisperse Frizzled were pooled and concentrated to 5–10 mg/ml (measured by A280). The concentrated sample was supplemented with glycerol to a final concentration of 10% and aliquots were frozen in liquid nitrogen and stored at −80 °C.

Expression and purification of Fzd4 homodimers and Fzd4/LRP6 heterodimers was performed using a split GFP strategy, as described (50).

Expression and purification of Dishevelled PDZ and DEP domains

The PDZ and DEP domains from mouse Dvl2 (residues 264–353 and 416–510, respectively) were each subcloned into a modified pCDFduet vector. The final expression construct contained N-terminal 6xHis and MBP tags followed by a TEV protease cleavage site and the PDZ or DEP sequence. The PDZ or DEP expression constructs were expressed in Escherichia coli strain BL21(DE3)-RIL (Agilent). Cultures were grown in terrific broth at 37 °C until OD600 reached 0.2, at which time the incubator temperature was lowered to 18 °C. When OD600 reached 0.8, cultures were induced with 0.5 mM IPTG and incubated shaking overnight at 18 °C. The following day, cells were collected by centrifugation at 5000 g for 30 min. Cell pellets were washed once in PBS, flash frozen in liquid nitrogen, and stored at −80 °C until purification.

Frozen cell pellets were thawed and resuspended in lysis buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 1 mM EDTA, 0.01% Tween-20, 5 mM DTT, protease inhibitors). Cells were lysed by passing twice through an Emulsiflex system (Avestin) with pressure pulsed to 15,000 psi. The lysate was centrifuged at 50,000 g for 30 min and the clarified lysate was loaded onto Amylose resin (New England Biolabs). The column was washed with ten bed volumes of wash buffer (20 mM HEPES-NaOH pH 8.0, 150 mM NaCl, 1 mM EDTA, 2 mM DTT). After washing, the column was capped, one bed volume of wash buffer and 0.5 mg TEV protease were added, and the column was incubated overnight at 4 °C with gentle rotation. The column was drained and if needed, the TEV eluate was concentrated using a 3-kDa cutoff Amicon spin concentrator until an appropriate loading volume for size exclusion chromatography (SEC) was reached. Concentrated protein was loaded onto a Superdex 75 Increase 10/300 GL or HiLoad Superdex 75 26/600 (Cytiva), depending on the size of the prep. The PDZ domain was eluted using a buffer composed of 20 mM HEPES-NaOH pH 8.0, 150 mM NaCl, and the DEP domain was eluted in the same buffer supplemented with 0.1 mM TCEP. Peak fractions were pooled, concentrated to ~5 mg/ml protein using a 3-kDa cutoff spin concentrator, and supplemented with glycerol to 10% v/v final concentration. Aliquots of protein were frozen in liquid nitrogen and stored at −80 °C.

Expression and purification of Norrin

Human Norrin bearing an N-terminal MBP tag and 3C protease recognition site (MBP-3C-Norrin) was subcloned into pAcGP67 and baculovirus was generated using the BestBac system according to the manufacturer’s protocol. Cultures of Sf9 cells were grown to a density of 2–3 × 106 cells/ml and infected with Norrin baculovirus (1:1000 v/v dilution). After 72 h, the culture was centrifuged for 10 min at 2000 g and the conditioned media was filtered through a 0.2-μm filter. Filtered media was loaded onto Amylose resin (1 ml packed resin bed per 100 ml culture). The column was washed with five bed volumes of wash buffer (20 mM HEPES-NaOH pH 8.0, 150 mM NaCl, 1 mM EDTA, 5% glycerol) and MBP-3C-Norrin was eluted using the same buffer supplemented with 10 mM maltose. The eluate was concentrated using a 30-kDa Amicon spin concentrator, loaded onto a Superdex 200 Increase 10/300 GL, and eluted in wash buffer. Peak fractions were pooled and used for binding or signaling assays without further concentration. To remove the MBP tag prior to use in assays, MBP-3C-Norrin was diluted to 1 μM and incubated with 3C protease for 30 min at room temperature.

Expression and purification of xWnt8

Xenopus laevis Wnt8 (xWnt8) was expressed and purified as previously described (40). Drosophila melanogaster S2 cells were grown in Schneider’s Drosophila Medium (ThermoFisher) supplemented with 1x glutamine/penicillin/streptomycin (GeminiBio) and 10% v/v heat-inactivated fetal bovine serum (MilliporeSigma). Cells were transfected with mouse Fzd8 CRD fused to an Fc tag (Fzd8CRD-Fc) and untagged xWnt8 expression plasmids (both gifts from Dr. K. Christopher Garcia) using the CalPhos Transfection Kit (Takara). Stable transfectants were selected in the presence of 10 μg/ml blasticidin (ThermoFisher). Following culture expansion, conditioned medium was harvested when cultures reached a density of 7–10 million cells/ml. Conditioned medium was loaded onto HiTrap Protein A Sepharose (Cytiva). The column was washed with 10 bed volumes of 10 mM HEPES-NaOH pH 7.2, 150 mM NaCl followed by 10 bed volumes of 10 mM HEPES-NaOH pH 7.2, 500 mM NaCl. To elute xWnt8 from mFzd8CRD, the column was washed with 10 mM HEPES-NaOH pH 7.2, 500 mM NaCl, 0.1% (w/v) DDM. Fractions containing eluted xWnt8 were pooled and loaded onto Concanavalin A agarose (Vector Laboratories) for detergent exchange. The column was washed with five column volumes of 10 mM HEPES-NaOH pH 7.2, 500 mM NaCl, 0.1% (w/v) DDM, followed by exchange into glyco-diosgenin (GDN; Anatrace) using a series of five-column-volume washes containing the following detergent concentrations (each in 10 mM HEPES-NaOH pH 7.2, 500 mM NaCl): 0.09% DDM and 0.01% GDN, 0.05% DDM and 0.05% GDN, 0.01% DDM and 0.09% GDN, 0.1% GDN, 0.01% GDN. xWnt8 was eluted with a buffer containing 10 mM HEPES-NaOH pH 7.2, 150 mM NaCl, 0.01% GDN, 250 mM methyl α-D-glucopyranoside, and 250 mM methyl α-D-mannopyranoside. Fractions containing eluted xWnt8 were stored at 4 °C.

Formation of xWnt8-Fzd complexes

xWnt8-Fzd complexes were prepared as described for xWnt8-FZD5 (40). Prior to addition of xWnt8, GDN was added to a 10 μM stock of mouse Fzd4 or human FZD5 to achieve a final concentration of 0.1% GDN. Following a 1-hour incubation on ice, Fzd was mixed with a 1.2-fold molar excess of xWnt8 and the mixture was diluted in 10 mM HEPES-NaOH pH 7.5, 150 mM NaCl to reduce the final GDN concentration to 0.005%. This mixture was incubated overnight at 4 °C and concentrated using a 100-kDa cutoff VivaSpin concentrator for SEC. The complex was loaded onto Superdex 200 Increase 10/300 GL and eluted in 10 mM HEPES-NaOH pH 7.5, 150 mM NaCl, 0.0025% GDN, 0.00025% CHS. Fractions containing the xWnt8-Fzd complex were pooled and concentrated using a 100-kDa cutoff VivaSpin concentrator before reconstitution into nanodiscs.

To assemble a complex between xWnt8 and Fzd4 dimer, the Fzd4 dimer was thawed and incubated with 2% LMNG at room temperature for 30 minutes before injection onto a Superose 6 SEC column. Dimer peak fractions were concentrated to 37 μM dimer in a 100 kDa-cutoff spin concentrator and incubated with 0.1% GDN on ice for 1 hour. For reconstitution with xWnt8, Fzd4 dimer was incubated with 1.5x molar excess xWnt8 at room temperature for 2 hours, then concentrated to 25 μM dimer for reconstitution.

Expression, purification, and biotinylation of MSP1D1 and MSP1E3D1

His6-TEV-MSP1D1 or -MSP1E3D1 (from the laboratory of Dr. Stephen Sligar; Addgene plasmids #20061 and #20066) were both expressed in E. coli strain BL21(DE3)-RIL. Cultures were grown in terrific broth at 37 °C until OD600 reached 3. Expression of MSP1D1 or MSP1E3D1 was induced with 1 mM IPTG and cultures were incubated for 1 h at 37 °C before lowering the incubator temperature to 30 °C. After an additional 2.5 h of incubation, cells were harvested and cell pellets were stored at −80 °C until protein purification.

Cells were lysed in 1x PBS supplemented with 1% Triton X-100 and protease inhibitors. Lysate was centrifuged for 30 min at 50,000 g and the clarified lysate was loaded onto Ni-NTA resin. The column was washed with ten column volumes of 40 mM Tris-HCl pH 8.0, 300 mM NaCl, 1% Triton X-100; ten column volumes of 40 mM Tris-HCl pH 8.0, 300 mM NaCl, 50 mM sodium cholate, 20 mM imidazole; and ten column volumes 40 mM Tris-HCl pH 8.0, 300 mM NaCl, 20 mM sodium cholate, 50 mM imidazole. Protein was eluted using 20 mM HEPES-NaOH pH 8.0, 300 mM NaCl, 5 mM sodium cholate, 400 mM imidazole. Fractions containing eluted protein were pooled and dialyzed overnight against 20 mM HEPES-NaOH pH 8.0, 100 mM NaCl, 1 mM EDTA, 5 mM sodium cholate.

Dialyzed protein was concentrated using a 10-kDa cutoff Amicon spin concentrator and centrifuged for 10 min at 20,000 g. Protein concentration was determined by A280 and an equimolar amount of NHS-PEG4-biotin (ThermoFisher) was added. Following a 30 min incubation at room temperature, the biotinylation reaction was quenched by addition of 1 M Tris-HCl, pH 8.0 to reach 20 mM final Tris concentration. The mixture was subjected to SEC on Superdex 200 HiLoad 16/600 (Cytiva) in a buffer composed of 20 mM HEPES-NaOH pH 8.0, 100 mM NaCl, 1 mM EDTA, 5 mM sodium cholate. Fractions containing MSP were pooled and concentrated to approximately 14 mg/ml (MSP1D1) or 12 mg/ml (MSP1E3D1). Extent of MSP biotinylation was assessed using the Pierce Biotin Quantitation Kit (ThermoFisher). Aliquots of protein were flash-frozen in liquid nitrogen and stored at −80 °C until use.

Nanodisc reconstitutions

Insertion of Fzd into recombinant high-density lipoprotein particles (nanodiscs) was performed according to previously published protocols (41, 68). 16:0–18:1 phosphatidylcholine, phosphatidylglycerol, phosphatidylserine, or phosphatidic acid, 18:0–20:4 phosphatidylinositol, brain phosphatidylintositol-4-phosphate, brain phosphatidylintositol-4,5-bisphosphate, 18:0–20:4 phosphatidylinositol-4,5-bisphosphate, and 18:0–20:4 phosphatidylinositol-3,4,5-bisphosphate were purchased from Avanti Polar Lipids. Cholesterol was purchased from MilliporeSigma. Reconstitutions were performed using a 1:20 molar ratio of Fzd:MSP1D1 (1:10 Fzd:nanodisc ratio) to favor incorporation of monomeric Fzd. Appropriate amounts of each lipid (from stock solutions of lipid dissolved in chloroform, or 20:9:1 v/v chloroform:methanol:water for PIP lipids) were combined in a glass test tube and solvent was evaporated under a stream of argon gas. Lipids were dried under vacuum for 1 h and the dried lipids were resuspended in HNE buffer (20 mM HEPES-NaOH pH 8.0, 100 mM NaCl, 1 mM EDTA) supplemented with 50 mM sodium cholate. Following initial solubilization, lipids were diluted with HNE buffer and purified Fzd and MSP1D1 were added to reach final concentrations of 18.5 mM sodium cholate, 6.5 mM lipid, 0.1 mM MSP, and 5 μM Fzd. The reconstitution mixture was incubated for 1 h on ice and incubated with BioBeads SM-2 (BioRad; 80 mg beads per nmol of lipid used) overnight at 4 °C. The following day, nanodisc samples were recovered from BioBeads. The beads were washed twice with HNE buffer and the washes were collected with the nanodisc samples. Nanodiscs were subjected to size-exclusion chromatography on Superose 6 Increase 10/300 GL (Cytiva) in HNE buffer and fractions were analyzed by SDS-PAGE.

Superose 6 fractions containing reconstituted Fzd were pooled, supplemented with CaCl2 to achieve 1 mM final concentration, and loaded onto a column containing 0.1 ml (packed bed volume) of M1 anti-FLAG sepharose resin. The column flowthrough was collected and re-loaded twice to ensure complete capture. The column was washed with 20 mM HEPES-NaOH pH 8.0, 150 mM NaCl, 1 mM CaCl2, 1 mg/ml BSA and eluted in 20 mM HEPES-NaOH pH 8.0, 150 mM NaCl, 5 mM EDTA, 0.2 mg/ml FLAG peptide, 1 mg/ml BSA. To determine the concentration of Fzd-containing nanodiscs, nanodisc samples were subjected to SDS-PAGE alongside a MSP1D1 standard curve, and the amount of protein was determined by densitometry.

For insertion of the Fzd4 homodimer or Fzd4-LRP6 heterodimer into nanodiscs, lipids were dried and re-solubilized as described above before addition of receptors and MSP to reach final concentrations of 18 mM sodium cholate, 6 mM lipid, 0.07 mM MSPE3D1 or 0.1 mM MSP1D1, and 7 μM receptor dimer. Isolation of receptor-containing nanodiscs was performed using SEC and anti-GFP nanobody affinity resin as described (50).

Biolayer interferometry

Biolayer interferometry assays were performed using Octet RED384 (Sartorius) or GatorPrime (Gator Bio) instruments. Biosensors were hydrated in binding buffer (20 mM HEPES-NaOH pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mg/ml BSA) for at least 15 min prior to use. All assay steps below were performed at 25 °C with shaking at 1000 rpm. Data were collected at 5 or 10 Hz.

Biotinylated nanodiscs were diluted to 30 nM in binding buffer and captured on streptavidin-coated biosensors for 2 min, yielding a final interference shift of ~3 nm. The loading parameters were optimized to ensure sufficient DEP-binding signal while avoiding kinetic artifacts due to crowding of nanodiscs on the biosensor surface. For each Fzd-containing disc sample to be analyzed, a reference sensor loaded with a matched sample of empty nanodiscs was run in parallel to define non-specific binding. Following loading, sensors were re-equilibrated for 10 min in binding buffer, then dipped into wells containing binding buffer for 30 s to establish a baseline. Sensors were dipped into wells containing ligand to monitor association signal and returned to wells containing binding buffer to monitor ligand dissociation. For experiments monitoring DEP domain binding, repeated association-dissociation cycles were performed using progressively higher concentrations of DEP in each association step. Association steps were 5 min each, and dissociation times were adjusted to ensure complete DEP dissociation before the next association-dissociation cycle commenced. The first association-dissociation cycle, used to define baseline drift, was performed in wells containing binding buffer without ligand. For experiments monitoring Norrin binding, a single association-dissociation cycle was performed, where each sensor monitored binding of a single Norrin concentration.

Initial data processing was performed in Octet Data Analysis 10.0 or Gator 1.7 software. Non-specific binding signal was removed by subtracting the reference sensor ligand-binding signal from the signal obtained using sensors loaded with Fzd-containing nanodiscs. Baseline drift was removed by subtracting the first association-dissociation cycle from subsequent cycles. Corrected binding traces were exported to Prism 9 (GraphPad) for curve fitting, using single-phase association and dissociation models for kinetic data and a one-site binding model for equilibrium data. To determine Kon, we plotted the observed association rate constant as a function of DEP concentration, using data from DEP concentrations that were sufficiently slow to produce reliable curve fits (typically 0.1–3 μM DEP), and verified that the rate constants increased linearly with DEP concentration. Koff was determined by averaging the rate constants derived from fitting the dissociation traces obtained with multiple concentrations of DEP. Equilibrium affinities were determined by plotting the association signal plateau as a function of the logarithm of DEP concentration and fitting to a one-site binding model in GraphPad Prism 9.

Confocal Microscopy

HEK293T cells (ATCC CRL-3216) were maintained in DMEM supplemented with 10% v/v fetal bovine serum and were plated into four-well chamber slides (Lab-Tek) at a density of 1×106 cells/well. The following day, cells were transfected with 150 ng Lyn-FRB (from Tobias Meyer, Addgene plasmid #20147), 80 ng CFP-FKBP-In54P (from Tobias Meyer, Addgene plasmid #20155), 100 ng GFP-PLCδ PH domain (from Tobias Meyer, Addgene plasmid #21179) or 300 ng Dvl2-GFP, and 400 ng HA-tagged Fzd4 plasmids per well using Lipofectamine 2000 according to the manufacturer’s instructions. The following day, cells were treated with 100 nM rapamycin (EMD Millipore) for 30 min at 37 °C. Cells were washed with HBSS (ThermoFisher) and fixed with 2% paraformaldehyde for 10 min. Blocking was done for 1 h at room temperature in HBSS supplemented with 1% BSA. Cells were incubated overnight with anti-HA antibody (BioLegend #901501, RRID: AB_2565006; diluted 1:5000) at 4 °C, washed, and incubated with Alexa Fluor 647-labeled goat anti-mouse IgG antibody (Abcam #150115, RRID: AB_2687948; diluted 1:5000) for 2 h at room temperature. Cells were washed in HBSS and one drop (60–80 uL) of mounting solution (Ibidi) was applied directly to the cells. For experiments designed to examine Dvl2 localization, cells were fixed as described above, blocked and permeabilized for 1 h at room temperature in HBSS containing 1% BSA and 0.1% Triton X-100, and slides were stained overnight with anti-Dvl2 antibody (Cell Signaling Technology #3216, RRID: AB_2093338; diluted 1:100). After washing, slides were incubated with Alexa Fluor 488-labeled goat anti-rabbit IgG antibody (ThermoFisher #A-11034, RRID: AB_2576217; diluted 1:200) for 2 h at room temperature. Cells were imaged at 40X magnification using a Leica SP8 inverted confocal system equipped with WLL. Images were collected and processed using LAS X v3.523 software (Leica Microsystems).

To assess the subcellular localization of Dvl2-GFP, cells in confocal images from three independent transfections were identified using the AlexaFluor 488 fluorescence signal and Dvl puncta in each cell were counted manually. The median numbers of puncta per cell in each treatment condition were compared using a Mann-Whitney test in GraphPad Prism 9 and the resultant p values were adjusted with a Bonferroni correction. The confidence intervals on the medians were determined by an empirical bootstrapping method in which 10,000 synthetic datasets were constructed by drawing N observations, with replacement, from the dataset for each treatment condition (where N is the number of cells counted in that treatment condition). To calculate confidence intervals for the difference in number of Dvl puncta between treatment conditions, a similar bootstrapping method was employed in which two synthetic datasets were constructed for the conditions being compared and the difference in median values between the two datasets was determined. This process was iterated times to obtain 10,000 difference values, and the 95% confidence interval was determined from this distribution.

Vesicle co-sedimentation

To prepare vesicles composed of 67% POPC, 20% cholesterol, 10% DOPS, and 3% porcine brain PI(4,5)P2, lipids (dissolved in chloroform, or 20:9:1 v/v chloroform:methanol:water for PI(4,5)P2) were combined in a glass test tube and solvent was evaporated under a stream of argon gas. The lipid film was further dried under vacuum overnight at room temperature. Lipids were resuspended in TBS (20 mM Tris pH 7.4, 140 mM NaCl) with vortexing and subjected to ten freeze-thaw cycles using liquid nitrogen. Unilamellar vesicles were prepared using an extruder (Avanti Polar Lipids) with filter size of 100 nm.

Vesicles were diluted in TBS and Dvl2 DEP domain was added to achieve a mixture containing 1 mM total lipid (30 μM PI(4,5)P2) and 3 μM DEP. Following 30 min incubation at room temperature, the reaction mixture was centrifuged for 40 min at 200,000 g. The supernatant was removed by pipetting and the pellet was resuspended in SDS-PAGE sample buffer. Aliquots of reaction mixture, supernatant, and pellet were subjected to SDS-PAGE and proteins were visualized by stain-free imaging (BioRad).

DEP dimerization

Urea-induced dimerization was performed as described (15). DEP domain (100 or 250 μM) was incubated in 20 mM HEPES-NaOH pH 8.0, 150 mM NaCl, 0.1 mM TCEP, 3 M urea for 8 h at room temperature. The sample was centrifuged for 5 min at 20,000 g and subjected to SEC on Superdex 75 Increase 10/300 GL in the same buffer lacking urea. Gel filtration standards (BioRad) were used to determine the approximate molecular weight of eluted DEP.

Testing for Fzd-induced DEP dimerization was performed by combining 10 nM Fzd4 homodimer-containing nanodiscs, 10 nM MBP-3C-Norrin, and 100 μM DEP in 20 mM HEPES-NaOH pH 8.0, 150 mM NaCl, 1 mM EDTA, 1 mg/ml BSA. The reaction mixture was incubated 1 h at 37 °C, centrifuged for 5 min at 20,000 g, and applied to Superdex 75 Increase 10/300 GL.

TopFlash assays

Fzd1/2/4/5/7/8-knockout HEK293T cells (69) were maintained in DMEM supplemented with 10% v/v fetal bovine serum. Cells were seeded into white 96-well plates (PerkinElmer) and transfected with 80 ng SuperTopFlash plasmid (laboratory of Dr. Randall Moon; Addgene plasmid #12456), 20 ng CMV-driven LacZ plasmid, and 1 ng FLAG-tagged Fzd4 or FZD5 plasmid per well using Lipofectamine 2000 according to the manufacturer’s instructions (ThermoFisher). Purified xWnt8 or Norrin (with MBP tag removed as described above for BLI) were added 24 h post-transfection and cells were incubated for 16–18 h with ligand. Cells were lysed and luciferase and β-galactosidase signals were quantified on a BioTek Synergy 2 plate reader using the Dual-Light Reporter System (ThermoFisher) according to the manufacturer’s instructions.

Supplementary Material

Supplementary Material

Acknowledgments:

We thank Drs. Bradley Efron and Balasubramanian Narasimhan for advice on bootstrap methods and Amy Wang for performing bootstrap calculations. We also thank Dr. K. Christopher Garcia for providing expression constructs for xWnt8, and Drs. Sachdev Sidhu and Stephane Angers (University of Toronto) for providing the anti-Fzd scFv. Research reported in this publication was supported by the National Center For Advancing Translational Sciences of the National Institutes of Health under Award Number UL1TR003142. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health

Funding:

This work was funded by National Institutes of Health grant R35 GM131747 (WIW), National Institutes of Health postdoctoral fellowship F32 GM126642 (JPM), National Institutes of Health predoctoral fellowship F31 EY031947 (ESB), Stanford ChEM-H Undergraduate Scholars Program (SK), and Stanford BioX Summer Research Program (SK).

Footnotes

Competing interests: The authors declare that they have no competing interests.

Supplementary Materials

Figs. S1 to S11.

Tables S1 and S2.

Data and materials availability:

All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials. Newly created expression constructs are available from the corresponding author upon request and will be deposited in Addgene.

References and Notes

  • 1.Nusse R, Clevers H, Wnt/β-Catenin Signaling, Disease, and Emerging Therapeutic Modalities. Cell. 169, 985–999 (2017). [DOI] [PubMed] [Google Scholar]
  • 2.Cselenyi CS, Jernigan KK, Tahinci E, Thorne CA, Lee LA, Lee E, LRP6 transduces a canonical Wnt signal independently of Axin degradation by inhibiting GSK3’s phosphorylation of -catenin. Proc. Natl. Acad. Sci. 105, 8032–8037 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Piao S, Lee S-H, Kim H, Yum S, Stamos JL, Xu Y, Lee S-J, Lee J, Oh S, Han J-K, Park B-J, Weis WI, Ha N-C, Direct Inhibition of GSK3β by the Phosphorylated Cytoplasmic Domain of LRP6 in Wnt/β-Catenin Signaling. PLoS One. 3, e4046 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Pan W, Choi S-C, Wang H, Qin Y, Volpicelli-Daley L, Swan L, Lucast L, Khoo C, Zhang X, Li L, Abrams CS, Sokol SY, Wu D, Wnt3a-Mediated Formation of Phosphatidylinositol 4,5-Bisphosphate Regulates LRP6 Phosphorylation. Science. 321, 1350–1353 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Qin Y, Li L, Pan W, Wu D, Regulation of Phosphatidylinositol Kinases and Metabolism by Wnt3a and Dvl. J. Biol. Chem. 284, 22544–22548 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Hu J, Yuan Q, Kang X, Qin Y, Li L, Ha Y, Wu D, Resolution of structure of PIP5K1A reveals molecular mechanism for its regulation by dimerization and dishevelled. Nat. Commun. 6, 8205 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Tanneberger K, Pfister AS, Brauburger K, Schneikert J, V Hadjihannas M, Kriz V, Schulte G, Bryja V, Behrens J, Amer1/WTX couples Wnt-induced formation of PtdIns(4,5)P 2 to LRP6 phosphorylation. EMBO J. 30, 1433–1443 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Kim W, Kim SY, Kim T, Kim M, Bae D-J, Choi H-I, Kim I-S, Jho E, ADP-ribosylation factors 1 and 6 regulate Wnt/β-catenin signaling via control of LRP6 phosphorylation. Oncogene. 32, 3390–3396 (2013). [DOI] [PubMed] [Google Scholar]
  • 9.Kříž V, Pospíchalová V, Mašek J, Kilander MBC, Slavík J, Tanneberger K, Schulte G, Machala M, Kozubík A, Behrens J, Bryja V, β-Arrestin Promotes Wnt-induced Low Density Lipoprotein Receptor-related Protein 6 (Lrp6) Phosphorylation via Increased Membrane Recruitment of Amer1 Protein. J. Biol. Chem. 289, 1128–1141 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Kim M, Kim S, Lee S-H, Kim W, Sohn M-J, Kim H-S, Kim J, Jho E-H, Merlin inhibits Wnt/β-catenin signaling by blocking LRP6 phosphorylation. Cell Death Differ. 23, 1638–1647 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Kim I, Pan W, Jones SA, Zhang Y, Zhuang X, Wu D, Clathrin and AP2 are required for PtdIns(4,5)P2-mediated formation of LRP6 signalosomes. J. Cell Biol. 200, 419–428 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Schwarz-Romond T, Fiedler M, Shibata N, Butler PJG, Kikuchi A, Higuchi Y, Bienz M, The DIX domain of Dishevelled confers Wnt signaling by dynamic polymerization. Nat. Struct. Mol. Biol. 14, 484–492 (2007). [DOI] [PubMed] [Google Scholar]
  • 13.Yamanishi K, Fiedler M, Terawaki S, Higuchi Y, Bienz M, Shibata N, A direct heterotypic interaction between the DIX domains of Dishevelled and Axin mediates signaling to β-catenin. Sci. Signal. 12, 1–9 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Kan W, Enos MD, Korkmazhan E, Muennich S, Chen D-H, Gammons MV, Vasishtha M, Bienz M, Dunn AR, Skiniotis G, Weis WI, Limited dishevelled/Axin oligomerization determines efficiency of Wnt/β-catenin signal transduction. Elife. 9, 1–33 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Beitia GJ, Rutherford TJ, Freund SMV, Pelham HR, Bienz M, V Gammons M, Regulation of Dishevelled DEP domain swapping by conserved phosphorylation sites. Proc. Natl. Acad. Sci. 118, e2103258118 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Gammons MV, Renko M, Johnson CM, Rutherford TJ, Bienz M, Wnt Signalosome Assembly by DEP Domain Swapping of Dishevelled. Mol. Cell. 64, 92–104 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Yang-Snyder J, Miller JR, Brown JD, Lai C-J, Moon RT, A frizzled homolog functions in a vertebrate Wnt signaling pathway. Curr. Biol. 6, 1302–1306 (1996). [DOI] [PubMed] [Google Scholar]
  • 18.Cong F, Schweizer L, Varmus H, Wnt signals across the plasma membrane to activate the β-catenin pathway by forming oligomers containing its receptors, Frizzled and LRP. Development. 131, 5103–5115 (2004). [DOI] [PubMed] [Google Scholar]
  • 19.Axelrod JD, Miller JR, Shulman JM, Moon RT, Perrimon N, Differential recruitment of Dishevelled provides signaling specificity in the planar cell polarity and Wingless signaling pathways. Genes Dev. 12, 2610–2622 (1998). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Pan WJ, Pang SZ, Huang T, Guo HY, Wu D, Li L, Characterization of Function of Three Domains in Dishevelled-1: DEP Domain is Responsible for Membrane Translocation of Dishevelled-1. Cell Res. 14, 324–330 (2004). [DOI] [PubMed] [Google Scholar]
  • 21.Gammons MV, Rutherford TJ, Steinhart Z, Angers S, Bienz M, Essential role of the Dishevelled DEP domain in a Wnt-dependent human-cell-based complementation assay. J. Cell Sci. 129, 3892–3902 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Ma W, Chen M, Kang H, Steinhart Z, Angers S, He X, Kirschner MW, Single-molecule dynamics of Dishevelled at the plasma membrane and Wnt pathway activation. Proc. Natl. Acad. Sci. 117, 16690–16701 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Wong H-C, Bourdelas A, Krauss A, Lee H-J, Shao Y, Wu D, Mlodzik M, Shi D-L, Zheng J, Direct Binding of the PDZ Domain of Dishevelled to a Conserved Internal Sequence in the C-Terminal Region of Frizzled. Mol. Cell. 12, 1251–1260 (2003). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Tauriello DVF, Jordens I, Kirchner K, Slootstra JW, Kruitwagen T, Bouwman BAM, Noutsou M, Rudiger SGD, Schwamborn K, Schambony A, Maurice MM, Wnt/ -catenin signaling requires interaction of the Dishevelled DEP domain and C terminus with a discontinuous motif in Frizzled. Proc. Natl. Acad. Sci. 109, E812–E820 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Simons M, Gault WJ, Gotthardt D, Rohatgi R, Klein TJ, Shao Y, Lee H-J, Wu A-L, Fang Y, Satlin LM, Dow JT, Chen J, Zheng J, Boutros M, Mlodzik M, Electrochemical cues regulate assembly of the Frizzled/Dishevelled complex at the plasma membrane during planar epithelial polarization. Nat. Cell Biol. 11, 286–294 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Chen Y, Sheng R, Källberg M, Silkov A, Tun MP, Bhardwaj N, Kurilova S, Hall RA, Honig B, Lu H, Cho W, Genome-wide Functional Annotation of Dual-Specificity Protein- and Lipid-Binding Modules that Regulate Protein Interactions. Mol. Cell. 46, 226–237 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Sheng R, Kim H, Lee H, Xin Y, Chen Y, Tian W, Cui Y, Choi J-C, Doh J, Han J-K, Cho W, Cholesterol selectively activates canonical Wnt signalling over non-canonical Wnt signalling. Nat. Commun. 5, 4393 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Capelluto DGS, Zhao X, Lucas A, Lemkul JA, Xiao S, Fu X, Sun F, Bevan DR, Finkielstein CV, Biophysical and Molecular-Dynamics Studies of Phosphatidic Acid Binding by the Dvl-2 DEP Domain. Biophys. J. 106, 1101–1111 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Paclíková P, Bernatík O, Radaszkiewicz TW, Bryja V, Mol. Cell. Biol, in press, doi: 10.1128/MCB.00145-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Schulte G, International Union of Basic and Clinical Pharmacology. LXXX. The Class Frizzled Receptors. Pharmacol. Rev. 62, 632–667 (2010). [DOI] [PubMed] [Google Scholar]
  • 31.Dawaliby R, Trubbia C, Delporte C, Masureel M, Van Antwerpen P, Kobilka BK, Govaerts C, Allosteric regulation of G protein–coupled receptor activity by phospholipids. Nat. Chem. Biol. 12, 35–39 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Yen H-Y, Hoi KK, Liko I, Hedger G, Horrell MR, Song W, Wu D, Heine P, Warne T, Lee Y, Carpenter B, Plückthun A, Tate CG, Sansom MSP, Robinson CV, PtdIns(4,5)P2 stabilizes active states of GPCRs and enhances selectivity of G-protein coupling. Nature. 559, 423–427 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Strohman MJ, Maeda S, Hilger D, Masureel M, Du Y, Kobilka BK, Local membrane charge regulates β2 adrenergic receptor coupling to Gi3. Nat. Commun. 10, 2234 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Mizumura T, Kondo K, Kurita M, Kofuku Y, Natsume M, Imai S, Shiraishi Y, Ueda T, Shimada I, Activation of adenosine A 2A receptor by lipids from docosahexaenoic acid revealed by NMR. Sci. Adv. 6, 1–15 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Damian M, Louet M, Gomes AAS, M’Kadmi C, Denoyelle S, Cantel S, Mary S, Bisch PM, Fehrentz J-A, Catoire LJ, Floquet N, Banères J-L, Allosteric modulation of ghrelin receptor signaling by lipids. Nat. Commun. 12, 3938 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Wright SC, Cañizal MCA, Benkel T, Simon K, Le Gouill C, Matricon P, Namkung Y, Lukasheva V, König GM, Laporte SA, Carlsson J, Kostenis E, Bouvier M, Schulte G, Hoffmann C, FZD5 is a Gαq-coupled receptor that exhibits the functional hallmarks of prototypical GPCR. Sci. Signal. 11, eaar5536 (2018). [DOI] [PubMed] [Google Scholar]
  • 37.Bang I, Kim HR, Beaven AH, Kim J, Ko S-B, Lee GR, Kan W, Lee H, Im W, Seok C, Chung KY, Choi H-J, Biophysical and functional characterization of Norrin signaling through Frizzled4. Proc. Natl. Acad. Sci. 115, 8787–8792 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Xu L, Chen B, Schihada H, Wright SC, Turku A, Wu Y, Han G-W, Kowalski-Jahn M, Kozielewicz P, Bowin C-F, Zhang X, Li C, Bouvier M, Schulte G, Xu F, Cryo-EM structure of constitutively active human Frizzled 7 in complex with heterotrimeric Gs. Cell Res. (2021), doi: 10.1038/s41422-021-00525-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Bowin C-F, Inoue A, Schulte G, WNT-3A–induced β-catenin signaling does not require signaling through heterotrimeric G proteins. J. Biol. Chem. 294, 11677–11684 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Tsutsumi N, Mukherjee S, Waghray D, Janda CY, Jude KM, Miao Y, Burg JS, Aduri NG, Kossiakoff AA, Gati C, Garcia KC, Structure of human Frizzled5 by fiducial-assisted cryo-EM supports a heterodimeric mechanism of canonical Wnt signaling. Elife. 9, 1–20 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Whorton MR, Bokoch MP, Rasmussen SGF, Huang B, Zare RN, Kobilka B, Sunahara RK, A monomeric G protein-coupled receptor isolated in a high-density lipoprotein particle efficiently activates its G protein. Proc. Natl. Acad. Sci. 104, 7682–7687 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Hilgemann DW, Local PIP2 signals: when, where, and how? Pflügers Arch. - Eur. J. Physiol. 455, 55–67 (2007). [DOI] [PubMed] [Google Scholar]
  • 43.Dickson EJ, Hille B, Understanding phosphoinositides: rare, dynamic, and essential membrane phospholipids. Biochem. J. 476, 1–23 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Suh B-C, Inoue T, Meyer T, Hille B, Rapid Chemically Induced Changes of PtdIns(4,5)P 2 Gate KCNQ Ion Channels. Science. 314, 1454–1457 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Nile AH, Mukund S, Stanger K, Wang W, Hannoush RN, Unsaturated fatty acyl recognition by Frizzled receptors mediates dimerization upon Wnt ligand binding. Proc. Natl. Acad. Sci. 114, 4147–4152 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.DeBruine ZJ, Ke J, Harikumar KG, Gu X, Borowsky P, Williams BO, Xu W, Miller LJ, Xu HE, Melcher K, Wnt5a promotes Frizzled-4 signalosome assembly by stabilizing cysteine-rich domain dimerization. Genes Dev. 31, 916–926 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Hirai H, Matoba K, Mihara E, Arimori T, Takagi J, Crystal structure of a mammalian Wnt–frizzled complex. Nat. Struct. Mol. Biol. 26, 372–379 (2019). [DOI] [PubMed] [Google Scholar]
  • 48.Ke J, Harikumar KG, Erice C, Chen C, Gu X, Wang L, Parker N, Cheng Z, Xu W, Williams BO, Melcher K, Miller LJ, Xu HE, Structure and function of Norrin in assembly and activation of a Frizzled 4–Lrp5/6 complex. Genes Dev. 27, 2305–2319 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Chang T-H, Hsieh F-L, Zebisch M, Harlos K, Elegheert J, Jones EY, Structure and functional properties of Norrin mimic Wnt for signalling with Frizzled4, Lrp5/6, and proteoglycan. Elife. 4, 1–27 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Bruguera ES, Mahoney JP, Weis WI, Reconstitution of purified membrane protein dimers in lipid nanodiscs with defined stoichiometry and orientation using a split GFP tether. J. Biol. Chem. 298, 101628 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Tan CW, Gardiner BS, Hirokawa Y, Layton MJ, Smith DW, Burgess AW, Wnt Signalling Pathway Parameters for Mammalian Cells. PLoS One. 7, e31882 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Hua Y, Yang Y, Li Q, He X, Zhu W, Wang J, Gan X, Oligomerization of Frizzled and LRP5/6 protein initiates intracellular signaling for the canonical WNT/β-catenin pathway. J. Biol. Chem. 293, 19710–19724 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Holmen SL, Robertson SA, Zylstra CR, Williams BO, Wnt-independent activation of β-catenin mediated by a Dkk1-Fz5 fusion protein. Biochem. Biophys. Res. Commun. 328, 533–539 (2005). [DOI] [PubMed] [Google Scholar]
  • 54.Janda CY, Dang LT, You C, Chang J, de Lau W, Zhong ZA, Yan KS, Marecic O, Siepe D, Li X, Moody JD, Williams BO, Clevers H, Piehler J, Baker D, Kuo CJ, Garcia KC, Surrogate Wnt agonists that phenocopy canonical Wnt and β-catenin signalling. Nature. 545, 234–237 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Miao Y, Ha A, de Lau W, Yuki K, Santos AJM, You C, Geurts MH, Puschhof J, Pleguezuelos-Manzano C, Peng WC, Senlice R, Piani C, Buikema JW, Gbenedio OM, Vallon M, Yuan J, de Haan S, Hemrika W, Rösch K, Dang LT, Baker D, Ott M, Depeille P, Wu SM, Drost J, Nusse R, Roose JP, Piehler J, Boj SF, Janda CY, Clevers H, Kuo CJ, Garcia KC, Next-Generation Surrogate Wnts Support Organoid Growth and Deconvolute Frizzled Pleiotropy In Vivo. Cell Stem Cell. 27, 840–851.e6 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Tao Y, Mis M, Blazer L, Ustav M, Steinhart Z, Chidiac R, Kubarakos E, O’Brien S, Wang X, Jarvik N, Patel N, Adams J, Moffat J, Angers S, Sidhu SS, Tailored tetravalent antibodies potently and specifically activate Wnt/Frizzled pathways in cells, organoids and mice. Elife. 8, 1–16 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Jiang X, Charlat O, Zamponi R, Yang Y, Cong F, Dishevelled promotes wnt receptor degradation through recruitment of znrf3/rnf43 e3ubiquitin ligases. Mol. Cell. 58, 522–533 (2015). [DOI] [PubMed] [Google Scholar]
  • 58.Yu A, Xing Y, Harrison SC, Kirchhausen T, Structural Analysis of the Interaction between Dishevelled2 and Clathrin AP-2 Adaptor, A Critical Step in Noncanonical Wnt Signaling. Structure. 18, 1311–1320 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Grainger S, Nguyen N, Richter J, Setayesh J, Lonquich B, Oon CH, Wozniak JM, Barahona R, Kamei CN, Houston J, Carrillo-Terrazas M, Drummond IA, Gonzalez D, Willert K, Traver D, EGFR is required for Wnt9a–Fzd9b signalling specificity in haematopoietic stem cells. Nat. Cell Biol. 21, 721–730 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Rim EY, Kinney LK, Nusse R, β-catenin-mediated Wnt signal transduction proceeds through an endocytosis-independent mechanism. Mol. Biol. Cell. 31, 1425–1436 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.Gaidarov I, Krupnick JG, Falck JR, Benovic JL, Keen JH, Arrestin function in G protein-coupled receptor endocytosis requires phosphoinositide binding. EMBO J. 18, 871–81 (1999). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Chen W, ten Berge D, Brown J, Ahn S, Hu LA, Miller WE, Caron MG, Barak LS, Nusse R, Lefkowitz RJ, Dishevelled 2 Recruits ß-Arrestin 2 to Mediate Wnt5A-Stimulated Endocytosis of Frizzled 4. Science. 301, 1391–1394 (2003). [DOI] [PubMed] [Google Scholar]
  • 63.Merrifield CJ, Kaksonen M, Endocytic Accessory Factors and Regulation of Clathrin-Mediated Endocytosis. Cold Spring Harb. Perspect. Biol. 6, a016733–a016733 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64.Matoba K, Mihara E, Tamura-Kawakami K, Miyazaki N, Maeda S, Hirai H, Thompson S, Iwasaki K, Takagi J, Conformational Freedom of the LRP6 Ectodomain Is Regulated by N-glycosylation and the Binding of the Wnt Antagonist Dkk1. Cell Rep. 18, 32–40 (2017). [DOI] [PubMed] [Google Scholar]
  • 65.Cheng Z, Biechele T, Wei Z, Morrone S, Moon RT, Wang L, Xu W, Crystal structures of the extracellular domain of LRP6 and its complex with DKK1. Nat. Struct. Mol. Biol. 18, 1204–1210 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 66.Saarikangas J, Zhao H, Lappalainen P, Regulation of the Actin Cytoskeleton-Plasma Membrane Interplay by Phosphoinositides. Physiol. Rev. 90, 259–289 (2010). [DOI] [PubMed] [Google Scholar]
  • 67.Butler MT, Wallingford JB, Planar cell polarity in development and disease. Nat. Rev. Mol. Cell Biol. 18, 375–388 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 68.Ritchie TK, Grinkova YV, Bayburt TH, Denisov IG, Zolnerciks JK, Atkins WM, Sligar SG, in Methods in Enzymology (2009; https://linkinghub.elsevier.com/retrieve/pii/S0076687909640118), vol. 464, pp. 211–231. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 69.Voloshanenko O, Gmach P, Winter J, Kranz D, Boutros M, Mapping of Wnt-Frizzled interactions by multiplex CRISPR targeting of receptor gene families. FASEB J. 31, 4832–4844 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 70.Steinhart Z, Pavlovic Z, Chandrashekhar M, Hart T, Wang X, Zhang X, Robitaille M, Brown KR, Jaksani S, Overmeer R, Boj SF, Adams J, Pan J, Clevers H, Sidhu S, Moffat J, Angers S, Genome-wide CRISPR screens reveal a Wnt-FZD5 signaling circuit as a druggable vulnerability of RNF43-mutant pancreatic tumors. Nat. Med. 23, 60–68 (2017). [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material
NIHMS1834245-supplement-Supplementary_Material.docx (1.7MB, docx)

Data Availability Statement

All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials. Newly created expression constructs are available from the corresponding author upon request and will be deposited in Addgene.

RESOURCES