The E3 ubiquitin ligase component, Cereblon, is an evolutionarily conserved regulator of Wnt signaling

Immunomodulatory drugs (IMiDs) are important for the treatment of multiple myeloma and myelodysplastic syndrome. Binding of IMiDs to Cereblon (CRBN), the substrate receptor of the CRL4CRBN E3 ubiquitin ligase, induces cancer cell death by targeting key neo-substrates for degradation. Despite this clinical significance, the physiological regulation of CRBN remains largely unknown. Herein we demonstrate that Wnt, the extracellular ligand of an essential signal transduction pathway, promotes the CRBN-dependent degradation of a subset of proteins. These substrates include Casein kinase 1α (CK1α), a negative regulator of Wnt signaling that functions as a key component of the β-Catenin destruction complex. Wnt stimulation induces the interaction of CRBN with CK1α and its resultant ubiquitination, and in contrast with previous reports does so in the absence of an IMiD. Mechanistically, the destruction complex is critical in maintaining CK1α stability in the absence of Wnt, and in recruiting CRBN to target CK1α for degradation in response to Wnt. CRBN is required for physiological Wnt signaling, as modulation of CRBN in zebrafish and Drosophila yields Wnt-driven phenotypes. These studies demonstrate an IMiD-independent, Wnt-driven mechanism of CRBN regulation and provide a means of controlling Wnt pathway activity by CRBN, with relevance for development and disease.

shows that CRBN co-precipitates with multiple components of the β-catenin destruction complex after WNT3A stimulation. This is really interesting, and of course raises the question of what other substrates does CRBN have in the complex, and how is this regulated. Some of this is beyond the scope of the paper, but raises the question of how direct the effect on CK1 is. This gets us to They tell us that when you knockdown the scaffold, the protein on the scaffold becomes less stable. The role of CRBN in this process is not addressed. I suggest these experiments could be expanded (e.g. knockdown AXIN1 or inhibit GSK3 -/+ CRBN knockdown).
Most confusing to me, Fig 3D shows that inhibition of GSK3 stimulates CK1α disappearance. Since GSK3 is downstream of CK1α while WNT3A treatment is upstream, it's not clear to me what mechanism related to WNT3A and CRBN is implicated in CK1α disappearance here. This needs to be clarified. Fig 3E -the effect of CRBN knockdown is modest, and with p only 0.05, so this result is not compelling. First, please show all the data points and not bar graphs, as this is a key experiment. Second, as an experimental suggestion, the WNT3A stimulation might be maxed out and so reductions are harder to see. The reduction of signaling by CRBN knockdown might be better tested at submaximal (not 450-fold, but instead 20-fold) activation by WNT3A.
Line 171: "Wnt stimulation also decreased Ck1α levels in organoids". This is an important implication of the prior data and the results in Fig 3F. However, the data presented in extended data 3B, lane 2 compared with lane 1 does not strongly support that claim. I get that organoid westerns are difficult, but extended figure 3B is not strong. It is clear that shCrbnA-D makes CRBN decrease, and CK1α increase.
Did the authors try to correlate the changes in CK1α mediated by CRBN with changes in β-catenin S45 phosphorylation?
Minor points addressable in text.
Plunger plots. e.g. Figures 1D, 1E, 2E, 3E, 3Fvi, 4, should be replaced. Experiments with small samples sizes should use scatter-plots or similar methods that allow evaluation of the distribution of the data. The manuscript by Shen et al. provides an interesting relationship between CRBN and Wnt-mediated signalings. After discovering CRBN as a target of thalidomide and its analogs, many reports have shown that CRBN works as a CRL4 E3 ligase and recognizes certain neosubstrates in the presence of various ligands such as thalidomide. However, the function of drug-unbound CRBN remains largely unclear. In this study, the authors suggest that Wnt3a induced the degradation of CK1a by CRL4CRBN. The authors also found that CRBN interacted with the CK1a-containing destruction complex and demonstrated that CRBN was required for Wnt signaling using the Organoid, Zebrafish, and Drosophila. The study may provide an important clue for the understanding of the original CRBN role in Wnt signaling. Therefore, the study is valuable, but there are many concerns, and the authors should address my comments listed below.
Major comments: 1. My biggest concern in this manuscript is that the authors have not clarified that the target of Wntdependent degradation by CRBN is indeed CK1a and not other subunits of the destruction complex. In Figs. 3B and C, the authors showed that loss or inhibition of other subunits of the destruction complex led to reduction of CK1a. There remains the possibility that CK1a is not a direct CRBN substrate. Decrease in CK1a might be resulted from the secondary effect of the breakdown of the real CRBN substrate. The authors should conduct the rescue experiments using a CRBN-binding defective mutant of CK1a and show it is sufficient to confer resistance to Wnt signaling by CRBN-mediated regulation.
2. The authors mentioned that Wnt signaling promotes degradation of a subset of protein substrates by stimulating CRBN activity ( Figs. 1 and 2). However, it is unclear whether Wnt signaling increases CRBN activity or promotes susceptibility of the destruction complex to degradation by CRBN, or both.
3. The ubiquitination assays (Fig. 1F, G, and Fig. 2D) are lacking essential information and appropriate controls that are described below. (i) The positions of molecular weight markers are required. (ii) The blots of input are required. (iii) For in vivo ubiquitination assays, the blots of anti-ubiquitin (or anti-HA for Fig. 1F) are required. (iv) The appropriate negative controls are required. For Fig. 1F, I recommend HA-Ub minus with Wnt3a as a negative control. For Fig. 1F and G, it is better to perform these experiments using CRBN -/-cells. For Fig. 2D, I recommend ATP (-) as a control. (v) Fig. 2D also requires the properly exposed images of CK1a that confirm the same amount of CK1a protein subjected to each reactions. (vi) For Fig. 1F, it is difficult to compare the ubiquitination level of CK1a between 8 h and the others because that is combined image gathered from different locations or exposure times. The authors should provide the original images help to evaluate this result. (vii) I feel Fig. 1G and 2D are poor and need to be much cleaner to support author's claim. 4. The authors should validate the knockdown phenotypes using another siRNA or perform corresponding experiments using knockout cells. These are necessary to avoid off-target effects of siRNA. figure 2D, the authors used mouse NIH3T3 cells. Previous studies have shown that mouse CRBN did not bind any substrates such as Ikaros, Aiolos, and CK1a in the presence of lenalidomide because the critical valine was replaced with isoleucine (I391) in mouse CRBN. If Wnt signal induced the interaction between mouse CRBN and mouse CK1a, this is a very important finding for understanding the physiological role of CRBN. The authors should examine the interaction between mouse CRBN and mouse CK1a using mouse cells such as Figure 3A and also the effects of lenalidomide on the interaction. 6. In this manuscript, the role of the destruction complex for degrading CK1a proteins by Wnt treatment is unclear. How does the destruction complex contribute to CRBN-mediated Ck1a degradation? 7. In the final paragraph, the authors claimed that "these studies show that CRBN is a novel, evolutionarily conserved regulator of the Wnt signaling pathway, whose Wnt-dependent regulation results in the ubiquitin-dependent proteasomal degradation of a subset of substrates." However, the conservation is not proved in fly and zebrafish at the mechanistic level. In order to justify the present title of this manuscript, the authors should investigate whether the Crbn-dependent ubiquitination of CK1a accounts for the observed phenotypes in fly and zebrafish.

In the Extended
8. Exogenous Wnt stimulation attenuated intestinal organoid differentiation and knockdown of CRBN rescued the Wnt-induced pathological phenotype (Fig. 3F). The authors should consider gut organoid differentiation markers and examine whether CK1α expression affects the phenotype rescued by CRBN silencing.
Other comments: 1. Fig. 1B requires negative control, such as housekeeping genes. Fig. 2C, the data are preliminary, and the authors should conduct additional experiments. (1) The immunoblot against CRBN is necessary. (2) The data do not rule out that CK1a was downregulated by the mechanism different from the authors' suggestion. To enhance the authors' hypothesis, the authors should check if the overexpression of CK1a-binding deficient CRBN mutants does not decrease CK1a. 5. In fig 4A v-vii, some Sens signal is detectable in En+ area. The expression levels of Sens need to be quantitatively measured, and the quantification methods have to be described. It would be nice to have some control experiment where known Wnt regulator(s) is inactivated to evaluate the severity of the Crbn-knockdown phenotype. Scale bars are missing for all panels.

In
6. In fig 4B, according to the current guidelines for MO use in zebrafish (Stainier, PLoS Genet 2017), MOs need to be validated by using knock-out mutant. The validation is particularly important, when the authors describe a new phenotype that appears by MO injection. The authors should make sure that MO does not produce cyclopia when the MO is injected into crbn KO fish. It seems that the cyclopic phenotype was examined at 2 dpf. 2 dpf is too late as abnormalities in eye size and distance could arise as a secondary effect. These phenotypes would better be analyzed in early 1 dpf (for example, described in Pei et al., Developmental Biology 2009). Scale bars should be added to Bii-iv.
Reviewer #3 (Remarks to the Author): In this study, Shen et al described an interesting role of CRBN in regulating WNT signaling. Authors showed that WNT recruits CRBN to the beta-catenin destruction complex and somehow activates the activity of CRBN. This leads to ubiquitination and degradation of CK1a and other known substrates of CRBN. Consistent with the negative role of CK1a in WNT signaling, CRBN-dependent degradation of CK1a promotes WNT/beta-catenin signaling. Loss of function studies in Drosophila and zebrafish support a role of CRBN in regulating WNT signaling in vivo. Overall, this is a nice study and it would certainly be interesting for people studying WNT signaling and CRBN. I have following suggestions for authors to improve the manuscript. 1. The finding that Wnt3a induces CK1a degradation is very interesting. However, Wnt3a recombinant protein is not 100% pure. To confirm the activity of Wnt3a recombinant protein is mediated by Wnt3a, authors should check whether FZD8-CRD can block the effect of Wnt3a recombinant protein on CK1a expression. 2. The extent of Wnt3a-induced CK1a degradation seems to vary in different experiments, at least based on quantifications provided in figures. Treatment of cells with 250ng/ml Wnt3a for 24 hrs decreased CK1a expression by 7 fold in Fig. 1A, but only 1 fold in Fig. 1E and Extended Fig. 1A. Authors should comment on this. 3. Fig. 2F. In CRBN-based glue degrader field, it is standard to use CRBN CRISPR KO cells to demonstrate the effect of glue degrader is mediated by CRBN. Authors should use HEK293 CRBN KO cells to demonstrate the effect of Wnt3a on CK1a and other CRBN substrates is dependent on endogenous CRBN. 4. Overexpression of CRBN in zebrafish induced eye loss (Fig. 4B). Does overexpression of CRBN increase WNT signaling in HEK293 cells? 5. Fig. 2C. Authors showed that overexpression of CRBN decreased the expression CK1a. Is this mediated by increased ubiquitination and degradation of CK1a? 6. Petzold et al suggested that CRBN does not bind to CK1a in the absence of lenalidomide. Have authors tested the direct binding between CRBN and CK1a? What is the degron of CK1a that mediates WNT and CRBN-dependent degradation of CK1a? Is the same beta-hairpin loop of CK1a shown in the CK1a-lenalidomide-CRBN structure (Petzold et al) important for WNT and CRBN-dependent degradation of CK1a? Can authors mutate critical residues involved in CK1-CRBN binding based on the crystal structure? 7. The finding that Wnt3a increases the activity of CRBN is intriguing. Does Wnt3a affect posttranslational modification of CRBN? Without solving the molecular mechanism, authors should at least provide some speculations in the discussion section.
The authors have two stories in one paper with a potential connection. They identify Cereblon (CRBN) as a regulator of Wnt/β-catenin signaling in a cell line, organoids, drosophila and zebrafish. This part is strong and novel. However, it lacks robust mechanistic insight. In addition, they start the manuscript with the observation that CRBN can regulate the ubiquitylation and abundance of CK1α, an important regulator of β-catenin S45 phosphorylation. There are several weak points in the connecting of the dots.
We thank the reviewer for their insightful comments, which have led to a significantly improved manuscript. To better illustrate the connection between CRBN and CK1a, we strengthened our data showing that CRBN ubiquitinates and degrades CK1a in response to Wnt activation and validated that CRBN regulates Wnt signaling in a CK1a-dependent manner across various model systems. Answers to their specific questions are shown below: 1. The CRBN knockdowns are done with a single siRNA without rescue. siRNA experiments need better controls.
In addition to previously used smart-pool siRNA, we have now used additional distinct individual siRNA or shRNAs targeting CRBN. These additional reagents were used to validate the effect of CRBN on CK1a degradation and Wnt activity (see revised Figures  2. The model predicts that effects of CRBN knockdown should be able to be rescued by overexpression of CK1α, but this was not tested. Our model predicts that CRBN degrades CK1a. Therefore, its knockdown leads to increased levels of CK1a and, as a result, decreased Wnt activity (see revised Figure 5A, 7A and Supplementary figures 5A, 7A), making the proposed experiment challenging. In contrast, our model also predicts that CRBN overexpression reduces CK1a levels and hence decreases its ability to inhibit Wnt signaling (see revised Figures 5B, 7A and Supplementary figures 5B, 7A-B). Thus, we took advantage of this latter observation by using the established CK1a agonist, pyrvinium, to rescue the effect of CRBN overexpression on Wnt activity. Such rescue experiments were successfully performed in both HEK293STF cells and in zebrafish (revised Figures 5B, 7Biii-iv, 7D).
3. It's not clear to me that the effect of CRBN on CK1α abundance is direct, versus indirect.
To address whether the effect of CRBN on CK1a abundance is direct, we performed a number of experiments: a) We utilized an in vitro binding assay to show that CRBN isolated from Wnt-stimulated cells exhibits an increased ability to associate with purified recombinant CK1a (revised Figures  2D-E).
b) We utilized an in vitro ubiquitination assay to show that CRBN isolated from Wnt-stimulated cells exhibits an increased ability to ubiquitinate purified recombinant CK1a (revised Figures  2D, 2F). c) We used a CK1a mutant, which exhibits decreased binding to CRBN (doi: 10.1038/nature16979), to show that CK1a is resistant to Wnt-mediated degradation when it cannot associate with CRBN (revised Figure 2C). d) Regarding arguments that CK1a degradation might be a secondary effect due to the disruption of other destruction complex subunits, we now show that: i. While Wnt3a induces a decrease in CK1a levels within 6 hours, it has little effect on the steady-state levels of other components of the b-catenin destruction complex over the same period of time. This observation is not consistent with CRBN acting on CK1a indirectly via disruption of the destruction complex (revised Figure 4A). ii. Knockdown of CRBN rescues the Wnt-stimulated decrease in CK1a protein levels without affecting the stability of other destruction complex components, inconsistent with CRBN acting on CK1a indirectly via disruption of the destruction complex (revised Figure  4B). iii. CK1a degradation induced by disruption of the destruction complex (APC or AXIN1 knockdown) remains CRBN-dependent (revised Figure 4D).
4. WNT3A seems to cause mono-, rather than poly-ubiquitylation of CK1α (Figs 1G), and the evidence for direct action of CRBN on CK1α is weak (Fig 2D).
We have optimized our CK1a ubiquitination assays and validated that Wnt3a induces a polyubiquitination of CK1a in cells (revised Figures 1F and Supplementary figure 1D). In addition, we now show that purified, recombinant CK1a protein binds to and is ubiquitinated by Wntexposed CRBN in vitro, supporting the direct action of CRBN on CK1a (revised Figures 2D-F). The answer to point #3 above also addresses the direct action of CRBN on CK1a.
5. How WNT3A signaling regulates association of CRBN with the destruction complex is unclear.
We have shown that upon Wnt stimulation, CRBN is recruited to the destruction complex (revised Figures 4C). Further, we show that the Wnt stimulated degradation of CK1a occurs prior to the dissociation of the destruction complex indicated by the level of scaffolding proteins Axin and APC (revised Figure 4A). However, when the integrity of the destruction complex is disrupted by the knockdown of AXIN1 or APC, in the absence of Wnt3a stimulation, CK1a is also degraded. This latter mechanism of CK1a degradation is also CRBN-dependent (revised Figure 4D). To explain these various results, we now present a model in which the destruction complex protects CK1a from CRBN induced degradation in the absence of Wnt. However, upon Wnt stimulation, the destruction complex serves as a scaffold to recruit CRBN, where it binds to CK1a and ubiquitinates it (see revised Figure 8).
6. GSK3 inhibition, acting downstream of CK1α, also causes a decrease in CK1α protein. Is this regulated by CRBN?
Although GSK3 requires the upstream b-catenin phosphorylation mediated by CK1a in order to phosphorylate b-catenin, GSK3 plays other roles in the destruction complex. For example, in the Wnt-off state, GSK3 regulates the interaction between Axin and APC to promote b-catenin destruction (doi: 10.7554/eLife.08022). In this case, there is no clear epistatic relationship between GSK3 and CK1a. Therefore, we cannot ensure that the inhibition of GSK3 in our previous experiments occurs downstream of CK1a. Thus, although we have new data showing that CK1a degradation in response to GSK3 inhibition is attenuated by knockdown of CRBN, we no longer show any GSK3 inhibition data-in order to better focus our manuscript. If the reviewer preferred such data be included, we would be happy to add it back.
Specific questions for the authors: 1. Fig 2D purports to show that adding Flag-CRBN in vitro "enhances" CK1α ubiquitylation. This figure is not compelling data...
We have further optimized our in vitro ubiquitination assay, which now better illustrates that Wntstimulated CRBN significantly increases the poly-ubiquitination of purified, recombinant CK1a (revised Figure 2F).
2. The authors rely extensively on a single CRBN siRNA. siRNA has many off-target effects, and so needs to be better supported....
As suggested, we have now utilized multiple distinct siRNA and shRNA to target CRBN, confirming the effect of CRBN on CK1a abundance and Wnt activity (revised Figures 3B, 3D They tell us that when you knockdown the scaffold, the protein on the scaffold becomes less stable. The role of CRBN in this process is not addressed. I suggest these experiments could be expanded (e.g. knockdown AXIN1 or inhibit GSK3 -/+ CRBN knockdown).
We now show that even though other destruction complex components are degraded by Wnt activation, only CK1a degradation is CRBN-dependent (revised Figure 4B). These findings are consistent with CK1a being the relevant CRBN substrate in the destruction complex. Also, as suggested, we now show that the CK1a degradation resulting from the knockdown of AXIN1 or APC is CRBN-dependent. These studies have helped clarify the role the destruction complex plays in CRBN mediated CK1a degradation (see Figure 8).
4. Most confusing to me, Fig 3D shows  See the answer to point #6 above.
5. Fig 3E -the effect of CRBN knockdown is modest, and with p only 0.05, so this result is not compelling. First, please show all the data points and not bar graphs, as this is a key experiment… In our revised Figure 5A, we now show that CRBN knockdown decreases Wnt reporter activity by approximately 75%, with a p value < 0.01 (**). As requested, we also now show all data points in the graph. In addition, we have clarified and expanded our definition of statistical significance in the methods.
6. Line 171: "Wnt stimulation also decreased Ck1α levels in organoids". This is an important implication of the prior data and the results in Fig 3F. However, the data presented in extended data 3B, lane 2 compared with lane 1 does not strongly support that claim. I get that organoid westerns are difficult, but extended figure 3B is not strong. It is clear that shCrbnA-D makes CRBN decrease, and CK1α increase.
We have shown in our previous publication (doi: 10.1126/scisignal.aak9916) that Wnt signaling decreases the protein level of CK1a in organoids, and now show that this Wnt-dependent CK1a decrease can be rescued by knocking down Crbn. The relevant text has now been revised to clarify this point (Line 204-207).
7. Did the authors try to correlate the changes in CK1α mediated by CRBN with changes in βcatenin S45 phosphorylation?
As now shown in revised Figure 4A, the change in CK1a protein levels correlate with the change in β-Catenin S45 phosphorylation.
Minor points addressable in text.
1. Plunger plots. e.g. Figures 1D, 1E, 2E, 3E, 3Fvi, 4, should be replaced. Experiments with small samples sizes should use scatter-plots or similar methods that allow evaluation of the distribution of the data.
As suggested, we now show all data points in the relevant revised figures.
2. Fig 1 legend: "HEK cells were transfected with HA-tagged ubiquitin (HA-Ub)". I expect they mean that the cells were transfected with a plasmid driving expression of HA-tagged ubiquitin?
We have corrected this mistake.
3. Figures 1F, 1G, 2D need molecular weight markers for proper interpretation. Fig 1F and G, is the predominant species mono-ubiquitylated CK1α? If it's mono-ubiquitylation, what are implications for proteasome? It would be interesting to compare the results of CK1α with βcatenin in these assays, since β-catenin is polyubiquitylated inversely with CK1α ubiquitylation.
We now show, in revised Figure 1F, 2F, and Supplementary figure 1D, that CK1a is polyubiquitinated in response to Wnt stimulation. In addition, CK1a is ubiquitinated in a manner that is inversely correlated with that of β-catenin (revised Supplementary figure 1D). In addition, we note that though a portion of CK1a appears mono-ubiquitinated in our in vitro experiment (revised Figure 2F and Supplementary figure 1D), this was not the case in our experiments looking at ubiquitination of endogenous CK1a (see Figure 1F). We thank the reviewer for pointing this issue out, which we have now clarified in the text, figure legend and methods.

Any role for FAM83F in this process?
While this is a very interesting question, in order to keep our manuscript focused, we have not addressed this question here.
Reviewer #2 (Remarks to the Author): The manuscript by Shen et al. provides an interesting relationship between CRBN and Wntmediated signalings. After discovering CRBN as a target of thalidomide and its analogs, many reports have shown that CRBN works as a CRL4 E3 ligase and recognizes certain neosubstrates in the presence of various ligands such as thalidomide. However, the function of drug-unbound CRBN remains largely unclear. In this study, the authors suggest that Wnt3a induced the degradation of CK1a by CRL4CRBN. The authors also found that CRBN interacted with the CK1a-containing destruction complex and demonstrated that CRBN was required for Wnt signaling using the Organoid, Zebrafish, and Drosophila. The study may provide an important clue for the understanding of the original CRBN role in Wnt signaling. Therefore, the study is valuable, but there are many concerns, and the authors should address my comments listed below.
We thank the reviewer for their generous and insightful comments. To address their remaining questions, we have added a significant number of new and revised experiments, revised our text, and clarified various existing experiments. For responses to their specific comments, please see our responses below.
Major comments: 1. My biggest concern in this manuscript is that the authors have not clarified that the target of Wnt-dependent degradation by CRBN is indeed CK1a and not other subunits of the destruction complex. In Figs. 3B and C, the authors showed that loss or inhibition of other subunits of the destruction complex led to reduction of CK1a. There remains the possibility that CK1a is not a direct CRBN substrate. Decrease in CK1a might be resulted from the secondary effect of the breakdown of the real CRBN substrate. The authors should conduct the rescue experiments using a CRBN-binding defective mutant of CK1a and show it is sufficient to confer resistance to Wnt signaling by CRBN-mediated regulation.
To address this important question, we now show that: a) While Wnt3a induces a decrease in CK1a levels within a 6-hour period of time, it has little effect on the steady-state levels of other components of the destruction complex over the same period of time. This observation is not consistent with CRBN acting on CK1a indirectly via degradation of other components of the destruction complex (revised Figure  4A). b) Knockdown of CRBN prevents the Wnt-driven decrease in CK1a protein levels but does not affect the stability of other destruction complex components. This finding is also inconsistent with CRBN acting on CK1a indirectly via degradation of other components of the destruction complex (revised Figure 4B). c) CK1a degradation induced by disruption of the destruction complex (APC or AXIN1 knockdown) remains CRBN-dependent (revised Figure 4B). d) We also used a CK1a mutant, which exhibits decreased binding to CRBN (doi: 10.1038/nature16979) to show that CK1a is resistant to Wnt-mediated degradation when it cannot associate with CRBN (revised Figure 2C).
2. The authors mentioned that Wnt signaling promotes degradation of a subset of protein substrates by stimulating CRBN activity (Figs. 1 and 2). However, it is unclear whether Wnt signaling increases CRBN activity or promotes susceptibility of the destruction complex to degradation by CRBN, or both.
We have shown that upon Wnt stimulation, CRBN is recruited to the destruction complex (revised Figure 4C). Further, we show that the Wnt-stimulated degradation of CK1a occurs prior to the dissociation of the destruction complex indicated by the level of scaffolding proteins Axin and APC (revised Figure 4A). However, when the integrity of the destruction complex is disrupted by the knockdown of AXIN1 or APC, CK1a is degraded in a Wnt3a-independent manner. This latter mechanism of CK1a degradation is however still CRBN-dependent (revised Figure 4D). To explain these various results, we now present a model in which the destruction complex protects CK1a from being destabilized by CRBN in the absence of Wnt. However, upon Wnt stimulation, the destruction complex serves as a scaffold to recruit CRBN, where it binds to CK1a and ubiquitinates it (see revised Figure 8).
3. The ubiquitination assays (Fig. 1F, G, and Fig. 2D) are lacking essential information and appropriate controls that are described below. (i) The positions of molecular weight markers are required. (ii) The blots of input are required. (iii) For in vivo ubiquitination assays, the blots of antiubiquitin (or anti-HA for Fig. 1F) are required. (iv) The appropriate negative controls are required. For Fig. 1F, I recommend HA-Ub minus with Wnt3a as a negative control. For Fig. 1F and G, it is better to perform these experiments using CRBN -/-cells. For Fig. 2D, I recommend ATP (-) as a control. (v) Fig. 2D also requires the properly exposed images of CK1a that confirm the same amount of CK1a protein subjected to each reactions. (vi) For Fig. 1F, it is difficult to compare the ubiquitination level of CK1a between 8 h and the others because that is combined image gathered from different locations or exposure times. The authors should provide the original images help to evaluate this result. (vii) I feel Fig. 1G and 2D are poor and need to be much cleaner to support author's claim.
We thank the reviewer for their comments and suggestions. We have now modified our ubiquitination assays and data presentation as suggested (revised Figures 1F, 2F and Supplementary Figure 1D).
4. The authors should validate the knockdown phenotypes using another siRNA or perform corresponding experiments using knockout cells. These are necessary to avoid off-target effects of siRNA.
In addition to previously used smart-pool CRBN siRNA, we now show results using additional distinct siRNA and shRNA targeting CRBN, which further validate the effect of CRBN on CK1a degradation and Wnt activity (revised Figures 3B, 3D, 4E-G, 5A and Supplementary figures 2C-D, 2F-G, 4A, 4G, 5A). figure 2D, the authors used mouse NIH3T3 cells. Previous studies have shown that mouse CRBN did not bind any substrates such as Ikaros, Aiolos, and CK1a in the presence of lenalidomide because the critical valine was replaced with isoleucine (I391) in mouse CRBN. If Wnt signal induced the interaction between mouse CRBN and mouse CK1a, this is a very important finding for understanding the physiological role of CRBN. The authors should examine the interaction between mouse CRBN and mouse CK1a using mouse cells such as Figure 3A and also the effects of lenalidomide on the interaction.

In the Extended
We now show that Wnt3a induces an interaction between CRBN and CK1a in mouse fibroblasts, while lenalidomide does not (revised Supplementary figure 3B).
6. In this manuscript, the role of the destruction complex for degrading CK1a proteins by Wnt treatment is unclear. How does the destruction complex contribute to CRBN-mediated Ck1a degradation?
We have shown that upon Wnt stimulation, CRBN is recruited to the destruction complex (revised Figure 4C). Further, we show that the Wnt stimulated degradation of CK1a occurs prior to the dissociation of the destruction complex indicated by the level of scaffolding proteins Axin and APC (revised Figure 4A). However, when the integrity of the destruction complex is disrupted by the knockdown of AXIN1 or APC, CK1a is degraded in a Wnt3a-independent manner. This latter mechanism of CK1a degradation is, however, still CRBN-dependent (revised Figure 4D). To explain these various results, we now present a model in which the destruction complex protects CK1a from being destabilized by CRBN in the absence of Wnt. However, upon Wnt stimulation, the destruction complex serves as a scaffold to recruit CRBN, where it binds to CK1a and ubiquitinates it (see revised Figure 8).
7. In the final paragraph, the authors claimed that "these studies show that CRBN is a novel, evolutionarily conserved regulator of the Wnt signaling pathway, whose Wnt-dependent regulation results in the ubiquitin-dependent proteasomal degradation of a subset of substrates." However, the conservation is not proved in fly and zebrafish at the mechanistic level… As suggested, we conducted additional experiments to support the stated conservation of function. We injected zebrafish embryos with crbn mRNA and showed corresponding changes in Ck1a protein levels (revised Figure 7E and Supplementary figure 7C). In addition, we now show that the established Ck1a agonist, pyrvinium, can rescue the eye loss phenotype induced by crbn mRNA (revised Figure 7Biii-iv, 7D, and Supplementary figure 7Aii-iv). Together, these experiments are consistent with Crbn primarily regulating Wnt activity via destabilization of Ck1a in zebrafish.
Other comments: 1. Fig. 1B requires negative control, such as housekeeping genes.
In the original figure, gene expression of the indicated genes were already normalized to that of the housekeeping gene GAPDH, indicated in the y axis legend. We now have added TBP expression as an additional housekeeping control (revised Figure 1B) and revised the legend to clarify this point. Fig. 2C, the data are preliminary, and the authors should conduct additional experiments.

In
(1) The immunoblot against CRBN is necessary. (2) The data do not rule out that CK1a was downregulated by the mechanism different from the authors' suggestion. To enhance the authors' hypothesis, the authors should check if the overexpression of CK1a-binding deficient CRBN mutants does not decrease CK1a.
As another reviewer commented that this figure lacked some controls, we no longer show this figure. Instead, as suggested, we have used a CK1a mutant that exhibits decreased binding to CRBN (doi: 10.1038/nature16979) and now show that this mutant is not degraded by Wnt3a (revised Figure 2C). As suggested, we have modified the text to clarify this definition (Line 1, 80-81, 118-119, 125, 146-148).

The authors claimed that
5. In fig 4A v-vii, some Sens signal is detectable in En+ area The expression levels of Sens need to be quantitatively measured, and the quantification methods have to be described. It would be nice to have some control experiment where known Wnt regulator(s) is inactivated to evaluate the severity of the Crbn-knockdown phenotype. Scale bars are missing for all panels.
As requested, we added quantitation of Sens reduction following RNAi-mediated knockdown of ohgata, or the negative control gene (yellow), in the posterior compartment of the wing disc (revised Supplementary figure 6D). To quantitate, we measured the length of the posterior region in which Sens was reduced as a fraction of the total length of the posterior compartment. The quantitation method is described in the methods and the associated figure legends. In addition, as requested, we provide new RNAi-mediated knockdown data of dishevelled, an essential Wingless pathway component (revised Supplementary figure 6B). The knockdown of dishevelled also decreased Sens in the posterior compartment, with more penetrant phenotypes compared to ohgt knockdown (revised Supplementary figure 6D). Scale bars have been added.
6. In fig 4B, according to the current guidelines for MO use in zebrafish (Stainier, PLoS Genet 2017), MOs need to be validated by using knock-out mutant. The validation is particularly important, when the authors describe a new phenotype that appears by MO injection. The authors should make sure that MO does not produce cyclopia when the MO is injected into crbn KO fish. It seems that the cyclopic phenotype was examined at 2 dpf. 2 dpf is too late as abnormalities in eye size and distance could arise as a secondary effect. These phenotypes would better be analyzed in early 1 dpf (for example, described in Pei et al., Developmental Biology 2009). Scale bars should be added to Bii-iv.
To validate our crbn MO-induced cyclopia phenotype we rescued the phenotype via the coexpression of crbn mRNA (revised Figure 7C and Supplementary figure 7Ai). Consistent with the crbn MO acting in a specific manner, zebrafish co-injected with crbn mRNA and MO do not exhibit any eye phenotype (revised Figure 7C and Supplementary figure 7Ai). In addition, we have now added 1 dpf images for WT and crbn MO phenotypes (revised Supplementary figure  7B). Scale bars have also been added to figures as suggested.
Reviewer #3 (Remarks to the Author): In this study, Shen et al described an interesting role of CRBN in regulating WNT signaling. Authors showed that WNT recruits CRBN to the beta-catenin destruction complex and somehow activates the activity of CRBN. This leads to ubiquitination and degradation of CK1a and other known substrates of CRBN. Consistent with the negative role of CK1a in WNT signaling, CRBN-dependent degradation of CK1a promotes WNT/beta-catenin signaling. Loss of function studies in Drosophila and zebrafish support a role of CRBN in regulating WNT signaling in vivo. Overall, this is a nice study and it would certainly be interesting for people studying WNT signaling and CRBN. I have following suggestions for authors to improve the manuscript.
We thank the reviewer for their supportive and insightful comments. As suggested, we have improved our analyses and modified our text accordingly. For specific comments, please see our responses below.  figure 1B).
2. The extent of Wnt3a-induced CK1a degradation seems to vary in different experiments, at least based on quantifications provided in figures. Treatment of cells with 250ng/ml Wnt3a for 24 hrs decreased CK1a expression by 7 fold in Fig. 1A, but only 1 fold in Fig. 1E and Extended Fig. 1A. Authors should comment on this.
The reviewer is correct in noting that although we always see decreases in CK1a in response to Wnt3a, there is some variation in the extent of this decrease. For the most part, the reasons behind these variations are not clear to us but likely depend on the batch and activity of recombinant Wnt3a and levels of the confluence of the cells used. We have modified the text to note this variance (Line 275).
3. Fig. 2F. In CRBN-based glue degrader field, it is standard to use CRBN CRISPR KO cells to demonstrate the effect of glue degrader is mediated by CRBN. Authors should use HEK293 CRBN KO cells to demonstrate the effect of Wnt3a on CK1a and other CRBN substrates is dependent on endogenous CRBN.
We requested HEK CRBN KO cell lines from two different groups but have yet to receive them. Alternatively, we generated and used our own HEK shCRBN knockdown cell lines. These cell lines allowed us to further demonstrate that CK1a is not degraded by Wnt3a in the absence of CRBN (revised Supplementary figure 2C).
We now show that overexpression of CRBN increases Wnt reporter activity in HEK293STF cells (revised Figure 5B) and does so in a manner that can be rescued using the established CK1a agonist, pyrvinium. Together, these results further support that CRBN regulates Wnt activity in a CK1a-dependent manner.
5. Fig. 2C. Authors showed that overexpression of CRBN decreased the expression CK1a. Is this mediated by increased ubiquitination and degradation of CK1a?
As another reviewer commented that this figure lacked some controls, we no longer show this figure. Instead, we have focused on validating our finding that CRBN ubiquitinates and degrades CK1a by refining our existing assays and adding additional control experiments (revised Figures  1F, 2C Figure 2D-E). However, it remains possible that a bridging molecule is co-purified with CRBN and acts to facilitate CRBN and CK1a binding. This result, along with this caveat, are now shown and discussed in the revised text (Line 139-142). b) We also mutated one of the critical residues in CK1a involved in lenalidomide-induced CRBN-CK1a binding, based on the Petzold crystal structure, Gly40. We note that this CK1a mutant is also resistant to Wnt-induced degradation. These results are consistent with the Gly40 beta-hairpin loop also being important for Wnt-driven degradation of CK1a (revised Figure 2C).
7. The finding that Wnt3a increases the activity of CRBN is intriguing. Does Wnt3a affect posttranslational modification of CRBN? Without solving the molecular mechanism, authors should at least provide some speculations in the discussion section.
We have not noticed a significant change in gene expression or protein levels of CRBN in response to Wnt activation. However, we did observe increased binding of CRBN to CK1a and its subsequent ubiquitination in response to Wnt induction. Based on such results, we speculate that Wnt likely regulates CRBN via some post-translational mechanism. As CRBN has been reported to autoubiquitinate in the presence of some immunomodulatory drugs (10.1038/nature13527), we have added text in the discussion speculating that Wnt3a might regulate CRBN autoubiquitination in order to activate it (Line 251-253).