Cancer Biology

Targeting the Hippo pathway in cancers via ubiquitination dependent TEAD degradation

Department of Discovery Oncology, Genentech, Inc., South San Francisco, CA 94080, USA
Department of Discovery Chemistry, Genentech, Inc., South San Francisco, CA 94080, USA
Department of Structural Biology, Genentech, Inc., South San Francisco, CA 94080, USA
Department of Microchemistry, Proteomics & Lipidomics, Genentech, Inc., South San Francisco, CA 94080, USA
Department of Bioinformatics, Genentech, Inc., South San Francisco, CA 94080, USA
Department of Immunology Discovery, Genentech, Inc., South San Francisco, CA 94080, USA
Department of Cheminformatics, Genentech, Inc., South San Francisco, CA 94080, USA
Department of Biochemical and Cell Biology, Genentech, Inc., South San Francisco, CA 94080, USA
Department of Translational Oncology, Genentech, Inc., South San Francisco, CA 94080, USA
Oregon Health and Science University, Portland, OR 97210, USA
Helen Diller Family Comprehensive Cancer Center, University of California San Francisco, San Francisco, CA 94158, USA

https://doi.org/10.7554/eLife.92450.1

Open access
Copyright information

Figures and data

TEAD transcription factors are ubiquitinated and targeted for proteasomal dependent degradation.
(A) TEADs are ubiquitinated. HEK 293 cells were transfected with TEADs-Myc and V5-ubiquitin plasmids for 48 hours, followed by immunoprecipitation with an anti-V5 antibody. TEADs expression and ubiquitinated TEADs were detected with anti-Myc antibody (n=3). (B) MG132 induces accumulation in TEAD ubiquitination. HEK 293 cells were transfected with TEADs-Myc and V5-ubiquitin plasmids for 48 hours followed by immunoprecipitation with anti-V5 antibody after additional treatment of 10 µM of MG132. TEADs expression and ubiquitinated TEADs were detected with anti-Myc antibody (n=3). (C-E) Ubiquitination sites on TEAD4 were mapped using KGG motif immunoprecipitation, followed by LC-MS/MS. For schematic in C, TEAD4-Myc was overexpressed in HEK293 cells for 48 hours with or without 10 µM of MG132 for 6 hours. KGG motif immunoprecipitation was then performed followed by LC-MS/MS. * represents each ubiquitination site. In D, predicted Lysine sites of ubiquitination is in the summary table. In E, ubiquitination for TEAD4 across conditions (red line) represents the aggregate of individually quantified K-GG peptides (black lines) using a linear mixed-effect model. Each quantification event represents the non-redundant, absolute peak area (converted to log2) under a given condition for a confidently assigned K-GG peptide spectral match.

RNF146 is a putative ubiquitin ligase for TEADs.
(A) RNF146 was identified from our E3 ligase screen using siRNA libraries targeting ubiquitin-related genes. siRNA libraries targeting 1088 ubiquitin-related genes were transfected to assess the effects on MCF7-TEAD reporter activity (n=2) as described in the method. All seven siRNAs of RNF146 were identified from the screen. The x-axis represents the calculated robust Z-score and y-axis represent the number of siRNAs per bin. (B) RNF146 is a negative regulator of the Hippo pathway. Achilles dependency network of YAP/TAZ (Nodes are significantly correlated genes, genes that are significantly correlated are connected by a line. Colors of nodes denote 2 distinct clusters based on random walk algorithm). An example of the RNF146 Chronos score is negatively correlated with the Chronos score of WWTR1 and positively correlated with Chronos scores of NF2. (C) TEAD ubiquitination is reduced upon knockdown of RNF146. After 24 hours of siRNF146 treatment, TEADs-Myc and V5-ubiquitin were transfected into HEK293 cells for an additional 48 hours and in vivo ubiquitination assay was performed as described in Figure 1. TEADs expression and ubiquitinated TEADs were detected with anti-Myc antibody. RNF146 knockdown was confirmed via quantitative RT-PCR (n=3). (D) RNF146 negatively regulates the Hippo signature. OVCAR8 cells were transfected with siRNAs for RNF146, YAP1 as indicated for 48hrs. RNA-seq was performed and Hippo signature was examined. (E) TEAD interacts with RNF146. HEK 293 cells were transfected with TEAD4-Myc for 48 hours, TEAD4 pull-down was then performed with an antibody that recognizes Myc-tag. TEAD4 expression and RNF146 interaction were detected using Myc antibody and RNF146 antibody respectively.

The ubiquitination of TEADs is dependent on their PARylation state.
(A) TEAD transcription factors interact with multiple members of the PARP family including PARP1 and PARP9. IP Mass-spec was done in Patu-8902 cell line using a pan-TEAD antibody which detects and pulls down all four TEAD isoforms. IgG antibody was used as a control for immunoprecipitation. The numbers represent peptide counts. Condition 1 was in the shNTC cell line, condition 2 was in the shYAP1/WWTR1 cell line and the interaction was detected in both conditions. (B) TEAD4 interacts with PARP1. HEK 293 cells were transfected with TEADs-Myc plasmid for 48 hours followed by immunoprecipitation with anti-Myc antibody. PARP1 expression was detected with an anti-PARP1 antibody (n=3). (C) TEADs are PARylated. HEK 293 cells were transfected with duo tagged TEAD-Myc-Flag and V5-ubiquitin plasmids for 48 hours, followed by immunoprecipitation with anti-Flag antibody. TEAD2 and TEAD4 expression and PARylation of TEADs were detected with anti-Myc antibody and anti-PAR antibody respectively (n=3). (D) TEAD ubiquitination is significantly reduced upon knock down of PARP1. HEK 293 cells were transfected with specific guide RNAs to transiently knock out PARP1. The control cell line and PARP1 KO cell line were further transfected with TEAD4Myc and V5-ubiquitin plasmids for 48 hours, followed by immunoprecipitation with an anti-V5 antibody. TEAD4 expression and ubiquitinated TEAD4 were detected with anti-Myc antibody (n=3). (E) PARylation of TEAD is located in the DNA binding domain (Asp70 site on TEAD4). Structure of TEAD4 DNA binding domain in complex with DNA (Shi et al., 2017) with labeled sites for PARylation (green) and ubiquitination (magenta). (F) TEAD4(D70A) mutation affects PARylation. HEK 293 cells were transfected with duo tagged TEAD-Myc-Flag and V5-ubiquitin plasmids for 48 hours, followed by immunoprecipitation with anti-Flag antibody. TEAD2, TEAD4 and TEAD4(D70A) expression and PARylation of TEADs were detected with anti-Myc antibody and anti-PAR antibody respectively (n=2). (G) The ubiquitination of TEAD is significantly reduced in the absence of PARylation. HEK293 cells were transfected with TEAD4-Myc or TEAD4(D70A)-Myc and V5-ubiquitin plasmids for 48 hours, followed by immunoprecipitation with an anti-V5 antibody. TEADs expression and ubiquitinated TEADs were detected with anti-Myc antibody (n=3).

Overexpression of RNF146 E3 ligase antagonizes the overgrowth phenotype caused by the Hpo mutant.
(A) PARylation site for TEAD is conserved in humans and Drosophila melanogaster as shown in the alignment of TEAD1/2/3/4and Sd protein sequences. (B, C) Overexpression of RNF146 can rescue the overgrowth phenotype caused by hpo RNAi. In B, the first panel represents the expression of a control RNAi (ey:: Gal4>UAS:: luciferase RNAi) (n=9). The second panel represents knockdown of Hpo (w1118; ey::Gal4/+: UAS::hpoRNAi/+), which results in a significantly larger eye compared to luciferase RNAi control (n=10). The third panel represents overexpression of RNF146 (ey::Gal4>UAS::RNF146) (n=9). In the fourth panel, overexpression of RNF146in a hyperactive Yki background (hpo RNAi) rescues the over-growth phenotype resulting in a significantly smaller eye (n=13). C is the quantification of B. Means +/- S.D. of eye area quantified per genotype are shown in arbitrary units (A.U.). One-way ANOVA with Sidak’s multiple comparisons. Numbers (n) of eye area quantified and P-values are indicated above. (D, E) Overexpression of RNF146 can rescue the over-growth phenotype caused by Hpo loss of function mutation. (D) Wild-type control clones were generated using the eyFLP system (genotype: w1118; neoFRT42D, +/neoFRT42D, GMRHid; ey::Gal4, UAS::FLP/+) (Newsome et al., 2000) (n=11, both eyes). hpo^KC202 mutant clones (genotype: w¹¹¹⁸; FRT42D, hpoKC202/FRT42D, GMRHid; ey::Gal4,UAS::FLP/+) result in a large, folded eye with cuticle overgrowth around the eye. This increase in eye area was significant compared to eyes of WT clones. (p<0.0001; n=9, both eyes). Clones overexpressing RNF146 showed no phenotype compared to control (genotype: w¹¹¹⁸; FRT42D, + /FRT42D, GMRHid; eyGal4, UAS::FLP/^UAS::RNF146^OE; n=11, both e^{yes). Combini}ng hpo^K^C202 and RNF146 overexpression (genotype: w¹¹¹⁸; FRT42D, hpoKC202/FRT42D, GMRHi^d; ey::Gal4,UAS::FLP/ UAS::RNF146^OE) resulted in a signifi^{cantly smalle}r eye as compared to the hpo^KC202 mutant only. (p=0.0027; n=6, both eyes). E is the quantification of eye area of all of the eyes in D. Means +/- S.D. of eye area quantified per genotype are shown in arbitrary units (A.U.). One-way ANOVA with Sidak’s multiple comparisons. Numbers (n) of eye areas quantified and P-values are indicated above.

TEAD knockdown induces robust Hippo pathway suppression.
(A) Cell viability assessed by Cell Titer-Glo upon siRNA treatment of TEAD1-4 in Hippo/TEAD sensitive (NF2null/YAP1 amplified) and insensitive cell lines (no YAP1/WWTR1). Cells were treated with siNTC (dark gray bars), siTox (light gray bars) and siTEAD1-4 (red bars) for 72hr before cell viability assay. (B) Tumor growth (measured as tumor volume) in xenograft mice upon TEAD1-4 knockdown in MDA-MB-231cells in presence of sucrose (red) and doxycycline (green). shNTC is control non-targeting shRNA (black and blue). (C) Chemical structures, biochemical data and cellular activity of all CIDEs screened. (D) Crystal structure of compound A bound to TEAD. (E) Protein abundance measurements made using tandem mass tag quantitative mass spectrometry and plotted as a log fold change. The scatterplot depicts the change in relative protein abundance inMDA-MB-231 cells when treated with compound D (x-axis) and compound E (y-axis) relative to protein abundance in compound A treated cells. Cells were treated with 0.5mM of compound for 8 hours before harvesting.

CIDEs mediated TEAD degradation is dependent on CRBN and the ubiquitin-proteasome pathway
(A) Immunoblot for pan-TEAD, PDE68, YAP, TAZ and β-Actin in MDA-MB-231 after 24h of compound D treatment at the indicated concentrations (0.128, 0.8, 5 µM). (B) Immunoblot for pan-TEAD, YAP, TAZ and β-Actin in OVCAR-8 after 16h treatment of compound A/compound E/compound D at the indicated concentrations of 0.1μM, 1µM and 10 µM. (C) Immunoblot for pan-TEAD, YAP, TAZ and β-Actin in OVCAR8 upon co-treatment with DMSO/ compound E/compound D (1 µM) for 5 hours and MG132 (10 µM) for 6 hours. (D) Immunoblot for pan-TEAD, YAP, TAZ and b - Actin in MDA-MB-231 cells upon co-treatment with DMSO/ compound D/ compound E (1 µM) for 5 hours and MLN-7243 (1 µM) for 6 hours. (E) Immunoblot for pan-TEAD, YAP, TAZ and β-Actin in MDA-MB-231 cells upon treatment with DMSO/compound E/compound D (0.1 µM and 0.5 µM) for 16 hours in siRNA (CRBN and control) treated cells. siRNA treatment was done for 3 days followed by degrader treatment.

TEAD degradation inhibits cell proliferation and downstream Hippo pathway signaling.
(A, B) Cell viability assessed using CellTiter-Glo in (A) OVCAR-8, (B) SK-N-FI after 6 days of treatment with compound D (red), compound E (green), compound A (purple), compound B (blue) and DMSO (black). Error bars represent the standard deviation of five technical replicates. (C, D) Crystal violet growth assay in OVCAR-8 (Hippo dependent) and SK-N-FI (Hippo independent) after 6 d treatment with compound D at indicated concentrations. (E, F) RNA-seq of Hippo signature genes after compound D/compound A treatment in MDA-MB-231 cell line. Hippo signature genes exhibit significant downregulation upon treatment with compound D in a concentration dependent manner. Cells were treated with 0.1 µM and 0.5 µM of compoundA/D for 24 hours before harvesting.

Compound D specifically reduces chromatin accessibility at TEAD motifs.
(A) Differential analysis of ATAC-seq of OVCAR-8 cells treated with DMSO or Compound D for 48 hours performed by DiffBind. (B) Genomic annotations of the differential ATAC-seq peaks by ChIP-Annotate (C) UCSC browser view of ATAC-seq at known targets of TEAD, ANKRD1 and CCN1. (D) Lost ATAC-seq peaks were visualized by Deeptools and analyzed for motif enrichment by HOMER. (E) Lost peaks were assigned to genes using known distal enhancer-gene target links in Poly-Enrich and the 10 most significant pathways are shown.

Sign up for email alerts