Regulation of YAP and Wnt signaling by the endosomal protein MAMDC4

Abstract

Maintenance of the intestinal epithelium requires constant self-renewal and regeneration. Tight regulation of proliferation and differentiation of intestinal stem cells within the crypt region is critical to maintaining homeostasis. The transcriptional co-factors β-catenin and YAP are required for proliferation during normal homeostasis as well as intestinal regeneration after injury: aberrant signaling activity results in over proliferation and tumorigenesis. Although both YAP and β-catenin activity are controlled along canonical pathways, it is becoming increasingly clear that non-canonical regulation of these transcriptional regulators plays a role in fine tuning their activity. We have shown previously that MAMDC4 (Endotubin, AEGP), an integral membrane protein present in endosomes, regulates both YAP and β-catenin activity in kidney epithelial cells and in the developing intestinal epithelium. Here we show that MAMDC4 interacts with members of the signalosome and mediates cross-talk between YAP and β-catenin. Interestingly, this cross-talk occurs through a non-canonical pathway involving interactions between AMOT:YAP and AMOT:β-catenin.

Fig 1. MAMDC4 deletion decreases cell proliferation and YAP nuclear localization.

(A) CRISPR/Cas9 Caco2_BBE control and MAMDC4 KO cells were grown on coverslips for 48 hours and labeled with antibodies against pHH3 and counterstained with DAPI. pHH3 positive cells were quantified by immunofluorescence. pHH3 labeling is decreased after MAMDC4 KO, and MAMDC4 overexpression rescues the proliferation phenotype. 500 to 1000 cells were counted for each experimental condition (n = 3). (B, C) CRISPR/Cas9 Caco2_BBE control and MAMDC4 KO cells were analyzed by Western blot. There is a 3-fold increase in p-YAP in MAMDC4 KO cells and no change in levels of the upstream Hippo kinase pathway activators of YAP. Error bars represent SD. *P<0.05, **P<0.01. Statistical significance was determined by unpaired Student’s t-test (n = 3). Residual MAMDC4 is due to persistence of cells that escaped selection. (D) Postnatal day 3 (P3) ileum from MAMDC4 control and MAMDC4 KO mice were labelled with antibodies against YAP. Rectangle denotes areas of inset below. YAP protein is enriched in the nucleus in control ileum (arrowhead) and nuclear labeling is decreased in MAMDC4 KO ileum (arrow). Scale bar: 25μm.

Fig 2. MAMDC4 overexpression increases YAP protein expression.

(A) Western blot analysis of HEK293 cells in the presence and absence of Wnt and with or without MAMDC4 overexpression. (B) Quantification of A shows that MAMDC4 overexpression and Wnt3a treatment results in increased YAP. There is no change in phosphorylation of the upstream Hippo kinase LATS or expression levels of LATS or MST. Error bars represent SD. **P<0.05, **P<0.01, ***P<0.001. Statistical significance was determined using an Ordinary one-way ANOVA with Tukey’s Multiple Comparisons (n = 3).

Fig 3. MAMDC4 deletion inhibits Wnt signaling.

(A) CRISPR/Cas9 Caco2BBE control and MAMDC4 KO cells were analyzed by Western blot. (B) Quantification of Western blots in A. There is a decrease in non-phospho (active) b-catenin and p-LRP6 (Ser 1490) and an increase in Axin1 in MAMDC4 KO cells. Residual MAMDC4 is due to persistence of cells that escaped selection (n = 3). (C) CRISPR/Cas9 Caco2_BBE control and MAMDC4 KO cells were transfected with Super TOPFlash TCF/LEF luciferase reporter and renilla control plasmid and grown for 48 hours. Wnt reporter activity is decreased in MAMDC4 KO cells (n = 5). (D) Enteroids were derived from postnatal day 11 mice, grown in Matrigel and treated for 6hrs with Wnt3a, fixed, and counterstained with DAPI. TCF/LEF:H2B -GFP expression (arrows) was visualized by confocal microscopy. Nuclear TCF/LEF:H2B -GFP expression increases with Wnt stimulation in control enteroids (arrows), There is no TCF/LEF:H2B -GFP expression in MAMDC4 KO enteroids in nuclei in control or Wnt stimulation conditions, indicateing a loss of β-catenin localization to the nucleus in MAMDC4 KO enteroids. Error bars represent SD. *P<0.05, ***P<0.001. Statistical significance was determined by unpaired Student’s t-test Scale bar: 20μm.

Fig 4. MAMDC4 promotes Wnt/β-catenin signaling.

(A) Western blot analysis of HEK293 cells in the presence and absence of Wnt and with or without MAMDC4 overexpression. (B) Quantification of A shows that MAMDC4 overexpression and Wnt3a treatment results in increased non-phospho (active) β-catenin and p-LRP6. There is decreased active β-catenin in Wnt3a stimulated cells over-expressing MAMDC4 relative to Wnt3a treated vector control cells. The are no changes in total LRP6 or Axin1 levels with MAMDC4 overexpression or Wnt stimulation. Both endogenous and exogenous MAMDC4 levels are unchanged with Wnt stimulation (n = 3). (C) HEK293 cells expressing MAMDC4 or vector control were transfected with Super TOPFlash TCF/LEF reporter and renilla control vector. Cells were stimulated with Wnt3a conditioned media for 6 hours. There is an increase in luciferase with MAMDC4 overexpression. In control transfected cells, Wnt3a stimulates luciferase activity 4-fold, however when MAMDC4 is overexpressed, this increase is dampened. Error bars represent SD. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. Statistical significance was determined using an Ordinary one-way ANOVA with Tukey’s Multiple Comparisons. (n = 3).

Fig 5. Increased Wnt signaling is mediated by the MAMDC4 cytoplasmic domain.

(A) Schematic of MAMDC4/Tac cytoplasmic and extracellular chimeras. (B) HEK293 cells expressing vector control, or a chimera containing the MAMDC4 extra-cellular domain or MAMDC4 cytoplasmic domain were transfected with Super TOPFlash TCF/LEF reporter and renilla control vector. There is an increase in reporter activity with the MAMDC4 cytoplasmic domain (n = 4). (C) Schematic of MAMDC4 point mutation constructs and alignment of cytoplasmic domain amino acid sequences. (D) HEK293 cells expressing MAMDC4 constructs with point mutations in the cytoplasmic tail were transfected with Super TOPFlash TCF/LEF reporter and renilla control vector. The T1186A and F1180A mutations show decreased luciferase activity. n = 3 for T1186A, n = 4 for F1180A and D1176A. **P<0.001, ****P<0.0001. Statistical significance was determined using an Ordinary one-way ANOVA with Tukey’s Multiple Comparisons.

Fig 6. MAMDC4 associates with the Wnt signalosome.

(A) HEK293 cells over-expressing MAMDC4 were stimulated with control or Wnt3a conditioned media. Lysates were immunoprecipitated with antibodies against MAMDC4 or control IgG. MAMDC4 associates with LRP6, p-LRP6, total β-catenin, DVL2 and Axin in control and Wnt3a stimulated conditions. There is a substantial amount of LRP6 and p-LRP6 (relative to input) associated with MAMDC4 in both Wnt stimulated and unstimulated conditions. (B) Quantification of signal from immunoprecipitation relative to input. LRP6 and p-LRP6 have a ratio of approximately 1 while the ratio of β-catenin and Axin1 have a ratio of IP to input that is <1/10^th of the amount of LRP6 and p-LRP6 immunoprecipitated. All immunoprecipitations were repeated 3 to 5 times and the input lanes represents 1/50^th of lysate used for the immunoprecipitation.

Fig 7. MAMDC4 is colocalized with the WNT signalosome.

(A) SKCO cells were transfected with MAMDC4 and DVL2. 48 hours post transfection cultures were treated with vehicle or Wnt3a for 1hr. Cells were labeled with antibodies to MAMDC4 (green) and DVL2 (red) and counterstained with DAPI. MAMDC4 and DVL2 are colocalized in puncta regardless of Wnt3a stimulation Scale bar: 10μm. (B) Quantification of Pearson’s Correlation Coefficient (r). Wnt3a treatment does not influence MAMDC4 and DVL2 colocalization. (C) SKCO cells were transfected with MAMDC4 and Axin1. 48 hours post transfection cultures were treated with vehicle or Wnt3a for 5min. Cells were labeled with antibodies to MAMDC4 (green) and Axin1 (red) and counterstained with DAPI. MAMDC4 and Axin1 are colocalized in some puncta after Wnt3a stimulation (arrows). (D) Quantification of Pearson’s Correlation Coefficient (r). There is an increase in MAMDC4 and Axin1 colocalization that is dependent on Wnt3a stimulation. (E) Immunoblot of NeutrAvidin Pull down. (F) Quantification of LRP6 Pull down / Lysate ratio.

Fig 8. MAMDC4 motifs regulate signalosome localization.

SKCO cells were co-transfected with either MAMDC4 or MAMDC4 mutants and DVL2. 48 hours post transfection cells were labeled with antibodies to MAMDC4 (green) and DVL2 (red) and counterstained with DAPI. (A) MAMDC4 and DVL2 are colocalized in puncta, however, DVL2 has diffuse labeling in cells lacking MAMDC4 overexpression (arrow) (B) Quantification of number of cells expressing both MAMDC4 and DVL2, which contain DVL2 puncta. 35 to 45 cells were assessed for each condition, n = 3. **P<0.001, ***P<0.001. Statistical significance was determined using an Ordinary one-way ANOVA. Scale bar: 10μm.

Fig 9. MAMDC4 overexpression disrupts the YAP:AMOT and β-catenin:AMOT interaction.

(A) HEK293 cells were transfected with MAMDC4 or empty vector control. Lysates were immunoprecipitated with antibodies to AMOT and examined by immunoblot. The amount of β-catenin and YAP immunoprecipitated with AMOT is reduced when MAMDC4 is overexpressed. Immunoprecipitation was repeated 3 times and a representative image is shown. A single lane, on all immunoblots was eliminated and this is indicated by a vertical line separating the lysate and IgG lanes. Lysate lanes for both MAMDC4 and β-catenin were scanned at lower intensity to avoid over exposure. (B) Quantification of IP/lysate ratios show a decrease in β-catenin and YAP association with AMOT in MAMDC4 overexpressing cells. (C) Lysates were collected from HEK293 cells expressing control or shRNA targeting MAMDC4. Lysates were immunoprecipitated with antibodies to AMOT and examined by immunoblot with antibodies to AMOT, total β-catenin, MAMDC4, and YAP. (D) Quantification of IP/lysate ratio show increase of β-catenin and YAP with AMOT in MAMDC4 KD cells.