Hippo signaling is intrinsically regulated during cell cycle progression by APC/C^Cdh1

Significance

The Hippo signaling pathway is evolutionarily conserved in the animal kingdom and plays essential roles in regulating tissue growth during development and regeneration. We have identified APC/C^Cdh1, a core component of cell cycle control machinery, as an evolutionarily conserved and previously unknown regulator of large tumor suppressor (LATS) kinases, which critically inhibit the YAP/TAZ transcription factors in transducing Hippo signaling. Our results suggest a model that APC/C^Cdh1 destabilizes LATS1/2 kinases in G1 phase of the cell cycle, leading to increased YAP/TAZ activities that promote G1/S transition by upregulating downstream gene expression, including E2F1. Our findings have important implications for a link between cell proliferation and LATS-regulated YAP/TAZ activities.

Abstract

The Hippo-YAP/TAZ signaling pathway plays a pivotal role in growth control during development and regeneration and its dysregulation is widely implicated in various cancers. To further understand the cellular and molecular mechanisms underlying Hippo signaling regulation, we have found that activities of core Hippo signaling components, large tumor suppressor (LATS) kinases and YAP/TAZ transcription factors, oscillate during mitotic cell cycle. We further identified that the anaphase-promoting complex/cyclosome (APC/C)^Cdh1 E3 ubiquitin ligase complex, which plays a key role governing eukaryotic cell cycle progression, intrinsically regulates Hippo signaling activities. CDH1 recognizes LATS kinases to promote their degradation and, hence, YAP/TAZ regulation by LATS phosphorylation is under cell cycle control. As a result, YAP/TAZ activities peak in G1 phase. Furthermore, we show in Drosophila eye and wing development that Cdh1 is required in vivo to regulate the LATS homolog Warts with a conserved mechanism. Cdh1 reduction increased Warts levels, which resulted in reduction of the eye and wing sizes in a Yorkie dependent manner. Therefore, LATS degradation by APC/C^Cdh1 represents a previously unappreciated and evolutionarily conserved layer of Hippo signaling regulation.

Precise growth control in embryonic development and adult tissue regeneration requires tightly regulated cell division and cell loss in response to various changes (1, 2). The Hippo signaling pathway is evolutionarily conserved from nematodes to human and plays essential roles in regulating tissue growth during development and regeneration (1, 3–5). Disruption in Hippo signaling leads to cancer and other devastating diseases (1, 6, 7). Originally identified in Drosophila as one required to maintain precise organ sizes of the eye and wing by controlling both cell proliferation and survival, the Hippo signaling pathway contains Hippo, Salvador, Warts, and Yorkie as core components (8–12) and receives inputs from extracellular environment as well as intracellular pathways to regulate a number of biological processes (13–18). Central to the Hippo signaling cascade is the regulation of the transcription factor Yorkie by Warts-mediated phosphorylation. Increased Yorkie protein levels and nuclear localization due to Warts inactivation result in dramatic tissue overgrowth by activating downstream gene expression to promote cell survival and proliferation. The mammalian Hippo pathway consists of Hippo homologs Ste20-like kinase MST1 and MST2, the scaffolding protein Salvador (SAV, also known as WW45), Warts homologs large tumor suppressor kinase 1/2 (LATS1/2), Yorkie homologs transcription coactivators Yes-associated protein (YAP), and TAZ (also known as WWTR1). Activation of MST1/2 kinase inhibits YAP/TAZ by activating LATS1/2 kinases (3, 19). Phosphorylated YAP/TAZ are sequestered in the cytoplasm via interaction with 14-3-3 and subsequently degraded through β-TrCP–dependent ubiquitination (20, 21).

As cell proliferation is regulated by proper cell cycle progression and Hippo-YAP/TAZ signaling is key to ensure precise growth control, apart from promoting cell proliferation, Hippo-YAP/TAZ signaling may also sense changes in cell proliferation and tissue growth and constantly modify cell cycle progression accordingly. YAP/TAZ are likely critical factors that bridge intrinsic and extrinsic changes with cell cycle progression, because YAP/TAZ can be controlled by both intrinsic and extrinsic stimuli that interact with MST1/2 and/or LATS1/2 kinases (1, 19, 22, 23).

Cell cycle progression is tightly controlled by periodic expression of key components of cell cycle machinery (24). The anaphase-promoting complex/cyclosome (APC/C) is a multisubunit E3 ubiquitin ligase complex that governs cell cycle progression by regulating cyclic degradation of key cell cycle regulators via two adaptor proteins, CDH1 or CDC20 (24–26). Recent reports have suggested critical roles for APC/C in various cellular processes, including genome stability and tumorigenesis (27). CDH1 degrades a number of proteins in cell cycle-dependent manner, many of which are known to mediate its role in negatively regulating cell proliferation and DNA replication. However, Cdh1^−/− mouse embryonic fibroblast (MEF) cells exhibit premature senescence and slow proliferation (28, 29), suggesting that some of Cdh1’s targets may positively regulate cell proliferation. Here, we introduce a cell-intrinsic regulatory mechanism in Hippo signaling by identifying LATS kinases as direct substrates of APC/C^Cdh1. This evolutionarily conserved mechanism links cell cycle progression directly with Hippo signaling in growth control.

Results

APC/C^Cdh1 Is Required for YAP/TAZ Activities.

The potent activity of YAP/TAZ in promoting cell proliferation led us to test whether YAP/TAZ activities are intrinsically regulated during cell cycle progression. We therefore examined YAP/TAZ and Hippo signaling activities in different phases of the cell cycle in the double thymidine block (DTB) assay (Fig. 1A and SI Appendix, Fig. S1A). We found that TAZ and phosphorylated YAP levels (pS127 by LATS kinases) (19, 20, 21, 30) oscillated during cell cycle and, in particular, YAP phosphorylation was reduced, while TAZ protein levels were increased in G1 and S phases (Fig. 1A and SI Appendix, Fig. S1A). Due to the high protein stability of YAP, but not TAZ, TAZ is more sensitive to phosphorylation-regulated degradation in vitro (31). The protein levels of Cyclin B, a degradation target of APC/C, correlated with YAP phosphorylation levels and inversely correlated with TAZ levels during cell cycle progression, suggesting that LATS kinase activities, and hence YAP/TAZ activities, are regulated during cell cycle by APC/C. The APC/C activity requires two activators, CDH1 and CDC20, to ensure timely recognition and subsequent degradation of its diverse substrates. Both CDH1 and CDC20 protein levels oscillate during cell cycle (24, 32), but only CDH1 oscillated similarly to phosphorylated YAP. We next examined YAP/TAZ nuclear localization, which is regulated by phosphorylation in the DTB assay (SI Appendix, Fig. S1 A–C) and found that YAP/TAZ nuclear localization was most pronounced in G1 phase (0 and 14 h), but much reduced in G2/M phase (6 and 8 h). To further confirm that YAP/TAZ nuclear localization changes in different phases of cell cycle under natural and asynchronous conditions, we used the fluorescent ubiquitination-based cell cycle indicator (FUCCI) system, a powerful tool for visualizing cell cycle progression in asynchronous cycling cells (Fig. 1B) (33). FUCCI utilizes the phase-dependent proteolysis of the oscillators Cdt1 and Geminin. Fusion protein of Cdt1 or Geminin with the fluorescent monomeric RFP (Cdt-RFP) or GFP (Geminin-GFP) serves as an indicator of G1 or S and G2 phase, respectively. We found that while YAP/TAZ were clearly localized in the nucleus of cells in the G1 or G1/S phase, their nuclear localization was much reduced in G2 or M phase cells (Fig. 1B). These results indicate that YAP/TAZ activities oscillate during cell cycle and APC/C may regulate YAP/TAZ in a phosphorylation-dependent manner.

Fig. 1.

APC/C^Cdh1 is required for YAP/TAZ activities. (A) HeLa cells were synchronized by double thymidine (2 mM) treatment and then released. Whole-cell lysates were subjected to Western blot analysis at indicated times after release. Western blot analysis was performed using the indicated antibodies. Asy, asynchronous. (B) HeLa cells were transduced with the Premo FUCCI cell cycle sensor and incubated overnight for expression of Geminin-GFP (S/G2/M phase) and Cdt1-RFP (G1 phase). FUCCI-expressing HeLa cells were fixed with 4% fresh PFA and stained with anti-YAP/TAZ antibodies (purple). Nuclei were stained with DAPI. (Original magnification, 600×.) YAP/TAZ nuclear localization were quantified as percentage of cells with YAP/TAZ nuclear localization in the total cells of a specific cell cycle phase. A schematic diagram of FUCCI is shown (Lower Left). (C) Asynchronized HeLa cells were transfected with control siGFP, two siCDH1s or siCDC20s. Levels of TAZ, p-YAP (S127), YAP, CDH1, CDC20, and GAPDH (loading control) were determined by Western blot analysis. (D) YAP/TAZ reporter activities (3xSd-Luc) in asynchronized cells by transfection with the indicated plasmids in the presence of siGFP or siCDH1. n = 3 independent experiments. Error bars represent ±SD. (E) Cdh1^+/+ WT or Cdh1^−/− KO MEFs lysates were subjected to Western blotting analysis with indicated antibodies. (F) Quantitative real-time PCR analysis for the expression of YAP/TAZ-target genes (CTGF and ANKRD1) after knockdown of CDH1 or YAP by siRNA in asynchronized HEK293T cells. The quantities of indicated mRNA were normalized by GAPDH. n = 3 independent experiments. (G) Western blot analysis of CTGF, p-YAP(S127), YAP, and CDH1 in nonsynchronized HeLa cells transfected with the indicated siRNAs. (H) HeLa cells were transfected with control siGFP and siAPC10. Cell lysates were subjected to Western blot analysis with indicated antibodies. In all quantified Western blotting results, representative blots are shown. Data are means ± SD of three biological replicates. **P < 0.01 (two-tailed Student’s t test). ns, not significant.

To further test whether CDH1 or CDC20 of APC/C plays a role in regulating YAP/TAZ protein levels, CDH1 or CDC20 was knocked down by two independent siRNAs in various cell lines. Reduction of CDH1, but not CDC20, led to reduced TAZ protein levels and increased YAP phosphorylation (Fig. 1C and SI Appendix, Fig. S1D). Consistently, reduction of CDH1, but not CDC20, reduced YAP/TAZ transcription activities shown by TEA domain (TEAD)-dependent luciferase reporter activity (34) (Fig. 1D and SI Appendix, Fig. S1E). Furthermore, in Cdh1-deficient (Cdh1^−/−) MEF cells, both TAZ and YAP protein levels were reduced while YAP phosphorylation was markedly up-regulated compared with the wild-type MEF cells (Fig. 1E). Depletion of CDH1 by siRNA also significantly down-regulated expression of YAP/TAZ target genes, such as CTGF and ANKRD1, without altering YAP transcription itself (Fig. 1 F and G). Conversely, overexpressing CDH1 enhanced TAZ-induced reporter activities (SI Appendix, Fig. S1F). These results indicate that CDH1 is required for YAP/TAZ transcription activities. Because CDH1 is known to have APC/C E3 ligase-dependent or -independent function (35–38), we next tested whether regulation of YAP/TAZ by CDH1 depends on the APC/C complex. CDH1 mutants with C-box or Fizzy domain deleted are deficient in interacting with the APC/C complex and they failed to enhance YAP/TAZ activities like the wild-type CDH1 (SI Appendix, Fig. S1 G and H). Furthermore, we found that knocking down APC10, which encodes a necessary subunit of the APC/C complex, led to down-regulation of YAP/TAZ activities, similar to that caused by knocking down CDH1 (Fig. 1H and SI Appendix, Fig. S1I). These results indicate that CDH1 promotes YAP/TAZ activities in an APC/C-dependent manner.

APC/C^Cdh1 Regulates Half-Life of LATS1/2 Kinases.

Our findings that APC/C^Cdh1 regulates YAP phosphorylation and YAP/TAZ activities prompted us to test whether LATS1 and LATS2 are substrates of APC/C^Cdh1, because LATS1/2 directly phosphorylate YAP/TAZ to regulate their protein levels and nuclear localization. LATS1/2 levels were examined during cell cycle progression in a DTB assay and both indeed oscillated in a same pattern as Cyclin B (Fig. 2A). In addition, LATS1/2 and CDH1 levels were inversely correlated in both DTB and thymidine-nocodazole block assay (Fig. 2 A and B and SI Appendix, Fig. S2A), while LATS1/2 levels correlated with phosphorylated YAP levels (Fig. 2B). In contrast, protein levels of an E3 ligase Ddb1-cullin4–associated factor 1 (DCAF1), known to promote LATS degradation (39), did not oscillate during the cell cycle. Neither CDC20 nor DCAF1 inversely correlate with LATS1/2 kinase levels during the cell cycle (Fig. 2A). In addition, when asynchronous HeLa cells were sorted into different cell cycle phases according to the DNA content, we also found that LATS1/2, and phosphorylated YAP levels oscillated during the cell cycle in a same pattern as Cyclin B (SI Appendix, Fig. S2B). Levels of LATS1/2 inversely correlated with YAP/TAZ levels, and expression of AMOTL2, CTGF, and CYR61, transcription targets of YAP/TAZ (SI Appendix, Fig. S2B). Furthermore, CDH1 overexpression or its depletion by small interfering RNA (siRNA) decreased or increased LATS1/2 levels, respectively (SI Appendix, Fig. S2 C and D). Similarly, higher LATS1/2 and lower TAZ levels were found in the Cdh1^−/− MEF cells compared with wild-type MEF cells (Fig. 2C). These results suggest that CDH1 regulates Hippo signaling by promoting LATS1/2 degradation.

Fig. 2.

APC/C^Cdh1 promotes LATS1/2 kinases degradation. (A) HeLa cells were first synchronized by DTB, and then released for the tindicated time. Western blot analysis was performed using the indicated antibodies. (B) HeLa cells were synchronized by thymidine-nocodazole block, and cell lysates were collected at the indicated time point after release. Western blot analysis was performed using the indicated antibodies. (C) Cdh1^+/+ WT or Cdh1^−/− KO MEFs lysates were subjected to Western blot analysis with indicated antibodies. (D) Loss of CDH1 significantly prolongs half-life of LATS1/2. Western blotting analysis of LATS1 and LATS2 proteins in asynchronized HeLa cells treated with cyclohexamide (CHX) for the indicated time (Upper). The line graphs show quantified LATS1 and LATS2 levels at indicated time (Lower). n = 4 independent experiments. Error bars represent ±SD, **P < 0.01 (two-tailed Student’s t test). (E) Myc-CDH1 was transfected in asynchronized HeLa cells, which were incubated with 20 μM MG132 for 8 h. Cell lysates were subjected to Western blot analysis with the indicated antibodies; **P < 0.01 (two-tailed Student’s t test). (F and G) HA-Ubiquitin was transfected with the indicated plasmids in the asynchronized HEK293T cells, 8 h after 20 μM MG132 treatment, LATS proteins were immunoprecipitated from the cell lysates and analyzed by Western blotting. (H) Reconstituted LATS1 expression in the LATS1/2^−/− HeLa cells rescued Hippo signaling defects. Cell lysates were subjected to Western blot analysis with indicated antibodies. (I) siGFP or siCDH1 was transfected into control and LATS1/2^−/− HeLa cells for 72 h. Cell lysates were subjected to Western blot analysis with indicated antibodies.

We then further tested whether LATS1/2 are previously unrecognized substrates of CDH1. LATS1/2 protein stability was examined by treating HeLa cells with the protein synthesis inhibitor cyclohexamide. The half-life of LATS1/2 proteins was significantly prolonged by CDH1 knockdown, while CDH1 overexpression promoted LATS1/2 degradation (Fig. 2D and SI Appendix, Fig. S2E). Similar to the regulation of YAP/TAZ activities (Fig. 1H and SI Appendix, Fig. S1I), knocking down APC10 increased LATS1 and LATS2 (SI Appendix, Fig. S2F), while APC/C binding-deficient mutants of CDH1 failed to promote LATS degradation (SI Appendix, Fig. S2 G and H). APC/C^Cdh1 is known to induce degradation of its substrates through ubiquitin-dependent proteolysis. Indeed, reduction of LATS1/2 protein levels by CDH1 was rescued by treatment with MG132, a proteasome inhibitor (Fig. 2E and SI Appendix, Fig. S2I). In addition, CDH1 overexpression enhanced—while CDH1 knockdown reduced—poly-ubiquitination of LATS1/2 (Fig. 2 F and G and SI Appendix, Fig. S2 J and K), indicating that APC/C^Cdh1 promotes LATS1/2 degradation through the ubiquitin-proteasome pathway.

To further test whether CDH1 up-regulates YAP/TAZ activities by promoting LATS1/2 degradation, we created a LATS1/2 double-knockout (DKO) (LATS1/2^−/−) HeLa cell line using the CRISPR/Cas9 technology (SI Appendix, Fig. S3). The indel mutations in LATS1 and LATS2 abolished LATS1 protein expression and resulted in a nonfunctional LATS2 truncation, respectively (Fig. 2H and SI Appendix, Fig. S3 A–C). In the LATS1/2^−/− cells, YAP phosphorylation was not detected, YAP/TAZ reporter activities and target gene CYR61 expression were highly increased. These defects in Hippo signaling were rescued by reconstituted LATS1 or LATS2 expression (Fig. 2H and SI Appendix, Fig. S3D). However, the LATS1/2^−/− cells still exhibited a distribution in different cell cycle phases, although YAP/TAZ were all localized in the nucleus with no TAZ oscillation (SI Appendix, Fig. S3 E and F). Furthermore, CDH1 reduction in the LATS1/2^−/− cells no longer increased YAP phosphorylation, reduced TAZ levels, or YAP/TAZ transcriptional activities (Fig. 2I and SI Appendix, Fig. S3G). Finally, knocking down CDH1 reduced transcription activities of wild-type TAZ, but not the TAZ-4SA mutant that abolished phosphorylation by LATS1/2 (SI Appendix, Fig. S3H). Taken together, these results indicate that APC/C^Cdh1 enhanced YAP/TAZ activities by promoting LATS1/2 degradation.

Cell cycle regulation of Hippo signaling suggests that cell cycle arrests should lead to changes in Hippo signaling. To test this hypothesis, we induced G1 phase arrest by inhibiting CDK4/6 with palbociclib (40) and partial G2/M arrest by knocking down CDC14B (41). Because YAP/TAZ and LATS1/2 peaked in G1 and G2/M phases (Figs. 1 and 2), respectively, indeed, we found that G1 arrest led to an increase in YAP/TAZ protein levels as well as nuclear localization and target gene expression while LATS1/2 levels were reduced (SI Appendix, Fig. S4 A–D). Partial G2 arrest led to opposite changes and expected increase of β-catenin levels (SI Appendix, Fig. S4 E–H), because it has been shown that Wnt/β-catenin signaling is under cell cycle control and peaks at G2/M phase (42). Parallel FACS analysis confirmed cell cycle phase enrichment (SI Appendix, Fig. S4 B and F). These results further support that Hippo signaling is intrinsically regulated during mitotic cell cycle.

APC/C^Cdh1 Promotes LATS Degradation Through both Evolutionary Conserved A-Box and D-Box.

To investigate whether LATS1/2 are APC/C^Cdh1 substrates, we examined interaction of LATS1/2 with CDH1 by coimmunoprecipitation (co-IP) and proximity-ligation assay (PLA) assays. LATS1/2 and CDH1 bound to each other (Fig. 3 A and B and SI Appendix, Fig. S5A). Furthermore, CDH1 was found to interact with LATS1/2 in both the cytoplasm and nucleus in the PLA assay (Fig. 3C), which was developed to detect protein interactions in close proximity in situ with higher sensitivity and specificity compared with the traditional immunofluorescent staining (43, 44). Among Hippo core components, we also identified SAV1 as a binding partner of CDH1 (SI Appendix, Fig. S5B), but depletion of CDH1 did not change SAV1 abundance (SI Appendix, Fig. S5C). A domain-mapping study revealed that CDH1 and LATS interaction was mediated by the WD-40 repeat domain of CDH1 (Fig. 3D). Substrate recognition by CDH1 requires a variety of degradation motifs, and the most common ones are the destruction box (D-box) and the KEN-box (45, 46). We found that LATS1/2 contain four potential D-boxes, and the third and fourth D-boxes are evolutionary conserved from insect to human (Fig. 4A and SI Appendix, Fig. S6A). Two conserved D-boxes were also found in SAV1, although CDH1 did not regulate its degradation (SI Appendix, Figs. S5C and S6B). To determine whether the D-boxes are responsible for LATS degradation, we generated individual (D1–D4) or all D-box mutants (mut-AllD) of CDH1 and tested LATS–CDH1 interaction. Unexpectedly, we found that all D-box mutants of LATS were still able to bind to and be degraded by CDH1 (Fig. 4 B and C), suggesting that additional recognition motifs are required. Further domain-mapping analyses with serial LATS deletions from both the N and C termini identified two crucial regions (amino acids 601–775 and 881–1130) (Fig. 4D). Because the first binding region (amino acids 601–775) contains the third (D3) and fourth (D4) D-boxes, we used LATS mutant constructs with D-box mutations and C-terminal deletion (LATSΔ881) to test whether D-box and another LATS domain are required for CDH1-dependent proteolysis. Co-IP analysis revealed that LATSΔ881-mD4, but not LATSΔ881-mD3, exhibited impaired interactions with CDH1 (SI Appendix, Fig. S7A). As a result, its poly-ubiquitination and degradation by CDH1 were abolished (SI Appendix, Fig. S7 B and C). Further sequence analysis of the LATS1/2 C-terminal region (amino acids 881–1130) suggested that an evolutionary conserved A-box, a rare destruction motif recognized by CDH1, not CDC20 (47), may be critical for CDH1-regulated degradation (SI Appendix, Fig. S7D). Indeed, mutating both D-box 4 and A-box, but not A-box only, impaired LATS association with CDH1 (Fig. 4 E and F) and mutant LATS with both D-box 4 and A-box mutations (mD4-mAbox) could no longer be regulated by CDH1 (Fig. 4 G and H). As a result, the CDH1-resistant LATS1 mutant showed a longer half-life compared with the wild-type control (Fig. 4I). Because the D and A boxes are located in the LATS kinase domain where LATS activation by phosphorylation occurred, we further found that the mD4-mAbox LATS1 had lost kinase activities as it could not be phosphorylated at T1079 (Fig. 4J) (48). These results indicate that APC/C^Cdh1 promotes LATS degradation by binding LATS through both D-box and A-box, which are also required for LATS kinase activation.

Fig. 3.

LATS interacts with WD40 repeats of CDH1. (A) Flag-LATS1 was transfected with or without Myc-CDH1 into the HEK293T cells. Cell lysates were subjected to co-IP and coprecipitated LATS1 or CDH1 was detected by Western blot analysis. (B) Endogenous interaction between LATS1 and CDH1 detected by co-IP followed by Western blot analysis. (C) Cells were stained with rabbit anti-CDH1 antibody and/or goat anti-LATS1 antibody, and in situ interaction between LATS1 and CDH1 (red dots) was detected with secondary proximity probes as described in Materials and Methods. (Scale bars, 10 μm.) (D) WD40 repeats of CDH1 is required to interact with LATS1. Flag-LATS1 and the indicated Myc-Cdh1 constructs (Right) were expressed in HEK293 T. Twenty-four hours posttransfection, the cells were pretreated with 20 μM MG132 for 8 h before collecting for co-IP and Western blotting assays (Left). WCL, whole-cell lysates.

Fig. 4.

APC/C^Cdh1 requires both evolutionarily conserved A-box and D-box for LATS degradation. (A) Sequence alignment of four putative D-box motifs evolutionary conserved in LATS1 and LATS2 kinases. (B) Myc-CDH1 were transfected with either wild-type Flag- LATS1 or indicated mutant constructs, and then cell lysates were subjected to immunoprecipitation assay. (C) LATS1 mutated in all of four D-box motifs is still degraded by ectopic CDH1 expression. Flag wild-type or mutant LATS1 was transfected into HEK293T cells with or without Myc-CDH1. Cell lysates were subjected to Western blot analysis with the indicated antibodies. Representative blots are shown; error bars represent ±SD of three biological replicates. **P < 0.01 (two-tailed Student’s t test). (D) Western blot analysis of cell lysates and immunoprecipitation derived from HEK293T cells transfected with wild-type LATS1 or truncation mutants with Myc-CDH1 construct. Twenty-four hours posttransfection, the cells were pretreated with 20 μM MG132 for 8 h before collecting (Left). Mapping studies from serial N- or C-terminal deletion reveals that LATS kinase contains two different interacting regions in CDH1. A schematic diagram showed LATS1 deletion mutants used in immunoprecipitation analysis (Right). (E) Schematic illustration of LATS wild-type and D-box, A-box LATS mutants. (F) Myc-CDH1 were transfected with either wild-type Flag-LATS1 or indicated mutant construct, and then cell lysates were subjected to immunoprecipitation assay. (G and H) Both the D-box and A-box are required for the degradation of LATS kinase by APC/C^Cdh1. Myc-Cdh1 or siCDH1 were transfected with wild-type Flag-LATS or indicated mutant constructs, and then subjected to Western blot analysis with indicated antibodies. Representative blots are shown; data are mean D-box and A-box significantly prolonged half-life of LATS1. (I) Western blotting analysis of LATS1 proteins in HeLa cells treated with cyclohexamide (CHX) for indicated time (Upper). The line graphs show quantified LATS1 levels at indicated time (Lower). n = 3 independent experiments. Error bars represent ±SD. *P < 0.05; **P < 0.01 (two-tailed Student’s t test). (J) Mutations in LATS1 disrupted its kinase activity. LATS1 mutants with D-box and/or A-box mutations were transfected in to the LATS1/2 null mutant HeLa cells. LATS1 and YAP phosphorylation were analyzed by Western blotting.

E2F1 Is a Downstream Target Gene of YAP.

Our results that YAP/TAZ activity is up-regulated in G1 phase prompted us to test whether YAP/TAZ promotes S-phase entry. We generated Yap/Taz-deficient MEF cells by infecting the conditional Yap^fl/fl;Taz^fl/fl MEF cells with adenovirus carrying Cre (Ad-Cre) (SI Appendix, Fig. S8A). Loss of Yap/Taz led to reduced cell proliferation, as determined by reduced cell numbers or BrdU labeling (SI Appendix, Fig. S8 B and C) and delayed G1/S transition (SI Appendix, Fig. S8D) with reduced expression of genes involved in G1/S phase transition, such as Cyclin E, Cdc6, and E2f1 (SI Appendix, Fig. S8E). These data suggest that Hippo signaling regulates cell cycle progression and YAP/TAZ activities peaked in G1 promotes G1/S transition. These results are consistent with recent findings that YAP/TAZ/TEAD and AP-1 form a complex that synergistically activates target genes directly involved in S-Phase entry and mitosis (49).

Next, we focused on E2F1 because it is a known critical regulator of G1/S transition (50) and cross-talks with the Hippo-YAP pathway (51–54). Similar to TAZ levels, E2F1 mRNA levels have also been shown to oscillate during cell cycle (55–57). E2F1 expression could be directly regulated by YAP/TAZ and indeed, five TEAD consensus binding sites were found in a region 3-kb upstream of the E2F1 transcription start site (Fig. 5A), and furthermore, these TEAD binding sites have been identified by chromatin immunoprecipitation sequencing (ChIP-Seq) assays and are associated with a transcriptionally active region as determined by the enriched H3K27Ac histone binding (SI Appendix, Fig. S8F). By ChIP assay, we found that these TEAD consensus binding sites, like the ones in the well-known target gene CTGF, were indeed occupied by YAP (Fig. 5B). Consistently, depletion of YAP/TAZ led to robust reduction of E2F1 mRNA and E2F protein expression (Fig. 5 C and D). Because TEAD1–4 are major DNA binding partners of YAP/TAZ, we depleted TEAD1/3/4 by two independent siRNAs, which resulted in reduced expression of E2F1 and other YAP/TAZ target genes CTGF and CYR61 (SI Appendix, Fig. S8 G and H). To further confirm that YAP regulates E2f1 expression in vivo, we analyzed the Mst1/2 DKO livers where Mst1 and Mst2 were deleted in hepatocytes (58). A robust up-regulation of E2f1 was observed in vivo by RNA-seq (59), qPCR, and Western blotting analysis, while expression of other E2f families was less affected (Fig. 5 E and F). In addition, expression of both Yap/Taz and E2f1 target genes were significantly up-regulated in the Mst1/2 DKO mouse livers (SI Appendix, Fig. S8I). Importantly, induction of E2f1 expression in the livers of Mst1/2 DKO mice was completely rescued by genetic removal of one allele of Yap (Fig. 5G). Taken together, these results indicate that E2f1 is a direct downstream target of Yap in vivo.

Fig. 5.

YAP regulates E2F1 transcription. (A) Five putative TEAD binding elements (TBE) are located ∼3 kb upstream of the TSS of E2F1 (Upper). TEAD family transcription factors associate with the indicated motif (underlined, Lower). YAP recognizes and binds consensus sequence (GGAATG) through TEAD. (B) ChIP-qPCR assay was performed at the indicated TEAD biding sites in the promoter of E2F1, and HBB or CTGF was used as a negative or positive control for YAP binding, respectively. n = 3 independent experiments. Error bars represent ±SEM, *P < 0.05, **P < 0.01 (two-tailed Student’s t test). (C) qPCR analysis of E2F1 or YAP/TAZ-target genes (CTGF and CYR61) in cells-depleted YAP/TAZ. n = 3 independent experiments. Error bars represent ±SEM **P < 0.01 (two-tailed Student’s t test). (D) Western blot analysis of endogenous proteins of HEK293T or HeLa cells in the presence of the indicated siRNAs. (E) Heat-map of RNA-seq data showing expression of E2F family genes in indicated mice. (n = 3 mice per genotype). (F) qPCR (Left) or Western blotting analysis (Right) of E2F1 expression in the livers of indicated mice. n = 3 independent experiments. Error bars represent ±SEM, **P < 0.01 (two-tailed Student’s t test). (G) Heterozygous removal of Yap in the livers of Mst1/2 DKOmice restores E2F1 expression. n = 3 independent experiments. Error bars represent ±SEM, **P < 0.01 (two-tailed Student’s t test).

Cdh1 Is Required for Organ Size Control in Drosophila Through a Conserved Mechanism.

Because the Hippo-Yap/Taz signaling pathway has been demonstrated to control organ size in Drosophila (3, 8, 9, 11, 12), we further tested whether our findings that APC/C^Cdh1 promotes YAP/TAZ activities by promoting LATS kinase degradation are evolutionary conserved in vivo. Cdh1 was knocked down in the Drosophila eye imaginal discs by the eye-specific driver GMR-Gal4 (Fig. 6A). Consistent with a role of CDH1 in promoting LATS degradation in mammalian cells, Cdh1 reduction in Drosophila dramatically increased the abundance of coexpressed Myc-tagged Warts (Myc-Wts), the Drosophila ortholog of LATS1/2 (Fig. 6A). Analysis of protein extracted from the eye disk further confirmed increased Myc-Wts protein levels in the Cdh1-depleted imaginal disk (Fig. 6B). In line with these results, the expression of diap1-GFP, a well-established Yorkie (the YAP/TAZ ortholog) target gene (34), was decreased by Cdh1 depletion in the eye disk (Fig. 6 C and D), indicating that loss of Cdh1 led to increased Wts protein levels and down-regulation of Yki transcriptional activity in Drosophila. To further determine the functional relationship between Cdh1 and Wts, we performed a genetic interaction experiment. Suboptimal Cdh1 depletion or Wts overexpression both resulted in slight reduction of eye size, but coexpression of Wts with suboptimal Cdh1-RNAi resulted in a more dramatic reduction of eye size (Fig. 6E). Consistent with our findings in a cell-culture system that CDH1 regulates YAP/TAZ activities via LATS1/2 degradation, Cdh1 knockdown did not suppress eye overgrowth caused by Wts knockdown, indicating that the molecular mechanism for CDH1-mediated LATS degradation is evolutionarily conserved (Fig. 6F). We further tested whether Yki phosphorylation by Wts is required for the effects caused by Cdh1-RNAi in vivo. Overexpression of wild-type Yki, or mutant forms of Yki (YkiS168A and Yki3SA) that could not be phosphorylated by Wts, both caused eye overgrowth. However, down-regulation of Cdh1 by Cdh1-RNAi only suppressed eye overgrowth caused by overexpression of wild-type Yki, but not by mutant Yki forms (YkiS168A and Yki3SA) (Fig. 6 G and H and SI Appendix, Fig. S9A). In addition, Cdh1 knockdown did not suppress eye overgrowth caused by overexpressing a Yki-independent and constitutively activated form of Sd (Sd-GA) (SI Appendix, Fig. S9B), indicating that Cdh1 regulates eye growth through Wts and Yki. We also tested the effects of Cdh1 in wing imaginal discs and found that consistently, Cdh1 RNAi driven by the posterior-compartment–specific gal4 driver hh-Gal4 also led to up-regulation of GFP-Wts (Fig. 6I). Furthermore, depletion of Cdh1 in the posterior wing compartment led to size reduction compared with the control wing that expressed control siRNA (Fig. 6J). Taken together, these results show that Cdh1 in Drosophila also regulates Hippo signaling by promoting Wts degradation in organ size control. The regulatory cascade of Cdh1-Lats-Yap/Taz that we have identified in mammalian cells also operates in an evolutionarily conserved way during growth control in Drosophila development.

Fig. 6.

Cdh1 regulates organ size via the Hippo pathway in Drosophila. (A) Eye discs expressing UAS-Myc-Wts and UAS-GFP (internal control) with or without UAS-Cdh1-RNAi under the control of GMR-Gal4 were immunostained with anti-GFP and anti-Myc antibodies. (B) Extracts from eye discs expressing UAS-Myc-Wts with or without UAS-Cdh1-RNAi under the control of GMR-Gal4 were subjected to Western blot analysis with anti-Myc antibody to detect Myc-Wts. UAS-GFP was coexpressed as an internal control. (C) Eye discs expressing Diap1-GFP with or without UAS-Cdh1-RNAi under the control of GMR-Gal4 were immunostained with anti-GFP and anti-Tubulin antibodies. (D) Extracts from eye discs Diap1-GFP with or without UAS-Cdh1-RNAi under the control of GMR-Gal4 were subjected to Western blot analysis with anti-GFP and anti-Tubulin antibodies. (E) Cdh1 inhibits Wts in organ size control. Side views of a control adult fly eye (GMR > Gal4) or eyes expressing UAS-Cdh1-RNAi, UAS-Wts, or both UAS-Cdh1-RNAi and UAS-Myc-Wts under the control of GMR-Gal4. (F–H) Side views of a control adult fly eye (GMR > Gal4) or eyes expressing both UAS-Cdh1-RNAi and UAS-Wts-RNAi (F) UAS-Yki (G) or UAS-Yki3SA (H) under the control of GMR-Gal4. (I) Wing discs expressing GFP-Wts with or without UAS-Cdh-RNAi under the control of hh-Gal4 were immunostained with anti-GFP and anti-Ci antibodies. Ci marks the anterior compartment. (J) Adult wing expressing UAS-Cdh-RNAi under the control of hh-Gal4 exhibited reduced posterior wing size. Eye or wing surface areas were measured by ImageJ. Error bars represent ±SD, **P < 0.01 n = 4 for each genotype. (Magnification: A, C, and I, 30×; E–H, 10×; J, 5×.) A, anterior wing compartment; P, posterior wing compartment.

Discussion

A key regulatory step in Hippo signaling is the regulation of YAP/TAZ phosphorylation by LATS kinases. In this study, we have identified APC/C^Cdh1, a core component of cell cycle control machinery, as a previously unknown regulator of LATS kinases. We show with both biochemical and genetic approaches that LATS, and therefore YAP/TAZ activities, are intrinsically regulated during mitotic cell cycle by CDH1, an essential component of APC/C^Cdh1. Because CDH1 itself is also regulated by both cell cycle-intrinsic and -extrinsic factors (60), CDH1 regulation of Hippo signaling is likely to be important in many biological processes beyond cell cycle progression. Our results suggest a model that APC/C^Cdh1 destabilizes LATS1/2 kinases in G1 phase of cell cycle, leading to increased YAP/TAZ activities that promotes G1/S transition by up-regulating downstream gene expression including E2F1 (SI Appendix, Fig. S10). In this regard, regulation of YAP/TAZ by LATS can sustain a positive feedback loop in proliferating cells by promoting cell cycle progression. Our findings therefore have important implications for a link between cell proliferation and LATS-regulated YAP/TAZ activities.

It has been shown that mitogenic growth factors can promote YAP/TAZ activities in a LATS-dependent way (23, 61, 62). Our results provide a mechanism whereby mitogenic growth factors synergize with YAP/TAZ activities, by APC/C^Cdh1-mediated LATS degradation through enhanced cell proliferation, which promotes YAP/TAZ activation. Thus, apart from extrinsic factors, intrinsic cell cycle can regulate YAP/TAZ through APC/C^Cdh1. Indeed, both G1 and G2 arrests altered LATS and YAP/TAZ activities (SI Appendix, Fig. S4 A, B, E, and F). Consistent with our observation that LATS1/2 peak at G2/M phase, LATS1 is shown to play a crucial role in controlling mitotic progression by forming a regulatory complex on mitotic apparatus (63). LATS1 can bind to actin and inhibit actin polymerization (64). Importantly, LATS1 regulates actin polymerization that affects cytokinesis through negative modulation of LIMK1 (65). Indeed, there is evidence that LATS1/2 stringently control cytokinesis by regulating CHO1 phosphorylation and mitotic activation of LIMK1 on centrosomes (66). Finally, LATS1 has been suggested to be a component of the mitotic exit network in higher eukaryotes (67), which may explain its abilities to induce G2 arrest and regulate cytokinesis (68–70). These findings suggest that cell cycle regulation of LATS1/2 levels are also linked to their own functions in the cell cycle, which may be independent of YAP/TAZ.

Cell cycle control of proliferating eukaryotic cells involves a complex regulatory network. Our work provides an example and mechanism whereby biological effects of the core engine of the mitotic cycle can be mediated by and synchronized with other important signaling pathways, such as Hippo signaling in development, regeneration, and tumorigenesis. The intercommunication between Hippo signaling and cell cycle machinery indicates that changes in cell cycle machinery and Hippo signaling are interwoven. Our findings are in line with an earlier study that Wnt/β-catenin signaling, also critical in growth control during development, regeneration, and tumorigenesis, is regulated by the cell cycle and peaks at the G2/M phase (42).

CDH1 has been proposed to be an oncosuppressor as it restrains proliferation and maintains quiescent/G1 cells. Our results show that CDH1 in a different context can act as an “oncogene,” as some of the previously unknown degradation targets of CDH1, such as LATS kinases, can inhibit proliferation. It has been reported that nonperiodic activation of APC/C^Cdh1 leads to continuous DNA synthesis uncoupled from mitosis by transcriptionally elevating the E2F1 transcription factor, a target of YAP/TAZ that we have identified in this study (71). In addition, comprehensive analyses (immunohistochemical staining of tissue microarray, clustering, and statistical analysis) of more than 1,600 human benign and malignant tumors revealed that CDH1 accumulates in malignant but not benign tumors (72). Therefore, tumor suppression or enhancement by the same gene can switch in different contexts. Related to this study, E2F1 acts as both a tumor suppressor as well as an oncogene (73). While YAP/TAZ cooperate with E2F1 to promote cell cycle progression and DNA replication (49, 53), E2F1 is not the only YAP/TAZ target mediating their effects in G1/S transition and YAP/TAZ also regulates the cell cycle in E2F-independent manner (49).

A number of studies have reported that Hippo signaling can be regulated by various intra- or extracellular stimuli, including cell–cell contact, mechanical stress, and growth/hormonal factors. Our data identify LATS kinases in Hippo signaling as intrinsic factors of cell cycle machinery directly regulated by APC/C^Cdh1, and such regulation is critical for organ size control, as shown in the fly eye and wing. Therefore, regulation of the Hippo pathway by APC/C^Cdh1 is an evolutionarily conserved growth control mechanism. As a substrate recognition component of the APC/C ubiquitin ligase complex, CDH1 has been implicated in many cellular processes (cell cycle regulation, cell fate determination, and so forth) by regulating distinct substrates, and its role is still expanding, as exemplified by two recent studies showing that Cdh1/Fzr coordinates retinal differentiation with G1 arrest (74) and regulates PCP (75) through targeting Nek2 kinase for degradation. Our study has revealed yet another substrate for CDH1: Wts/LATS1/2 kinase in the Hippo signaling pathway. Supporting our findings in mammalian cells, ablation of Hippo function or overexpressing Yki in Drosophila leads to an increase in E2f1 level and activity (76, 77). Because D-box and A-box mutation both disrupted LATS kinase activities by abolishing its phosphorylation at T1079 (Fig. 4J) (48), it is also possible that CDH1 binding to LATS itself could inactivate LATS by a conformation change that abolishes LATS phosphorylation at T1079. Furthermore, activation of LATS by phosphorylation also alters the conformation of the LATS kinase domain such that they may be resistant to CDH1-mediated degradation. In this regard, activation of LATS1/2 could render them resistant to cell cycle control by CDH1. This possibility is a subject for future investigation, which will be important to further understand how core cell cycle machinery regulates various cellular and tissue processes through Hippo signaling in development and regeneration.

Materials and Methods

Expression Constructs.

Flag-tagged full-length LATS1 or LATS2 were obtained from Addgene or G. Longmore (Washington University School of Medicine in St. Louis, St. Louis, respectively). Deleted forms (amino acids 1–880, 1–600, 601–1130, 775–1130, and 885–1130) and mutant versions of D-box and/or A-box of LATS1 were generated into p2xFlag-CMV2. Deleted form (amino acids 1–775) of LATS1 was a gift from B. M. Gumbiner, University of Virginia, Charlottesville, VA. The 6xMyc-tagged Cdh1 was kindly provided from S. Meloche, Université de Montréal, Montreal. Myc-tagged full-length Cdh1, Myc-Cdh1-N/F1, -N/F2, -N, and -F/WD40 were gifts from J. Lukas, University of Copenhagen, Copenhagen.

Cell Culture, siRNA Knockdown, and Transfections.

HEK293T, HeLa, and Huh7 were maintained in standard conditions (DMEM supplemented with 10% FBS and 100 U/mL penicillin/streptomycin). MEF cells were cultured in standard conditions with 10% nonessential amino acid. siRNA transfection was performed with Lipofectamine RNAi-MAX (Life Technologies) in antibiotic-free medium according to the manufacturer’s instructions. Oligonucleotides of siRNA duplexes were purchased from Life Technologies as follows: 5′- siGFP: 5-’GTTCAGCGTGTCCGGCGAG-3′, siCDH1#1: 5′-GGAACACGCTGACAGGACA-3′, siCDH1#2: 5′-TGAGAAGTCTCCCAGTCAG-3′, siCDC20#1: 5′-CGGCAGGACTCCGGGCCGA-3′, siCDC20#2: 5′-CGGAAGACCTGCCGTTACA-3′, siYAP: 5′-GACATCTTCTGGTCAGAGA-3′, siLATS1: 5′-CACGGCAAGATAGCATGGA-3′, siLATS2: 5′-GCCACGACTTATTCTGGAA-3′. siAPC10#1: ACAAGGCATCCGTTATATCTA, siAPC10#2: AGTACGGGAAATTGGGTCACA, siTEAD1/3/4 #1: ATGATCAACTTCATCCACAAG, siTEAD1/3/4 #2: GATCAACTTCATCCACAAGCT. Plasmid transfections were carried out with PEI (Polyethylenimine; Polysciences Inc.), Lipofectamin 3000 (Life Technologies), or Transit-LT1 (Mirus Bio) according to the manufacturer’s instructions.

Cell Cycle Synchronization.

For G1/S-phase block, HeLa cells were synchronized by DTB. Briefly, cells were incubated with 2 mM thymidine for 18 h followed by 9-h release. Cells were treated with 2 mM thymidine for another 18 h and then released for indicated times by replacing with fresh media. For mitotic block, HeLa cells were synchronized by thymidine-nocodazole block. HeLa cells were treated with 2 mM thymindine for 24 h. Cells were released for 3 h by adding fresh media, and incubated with 100 ng/mL nocodazole for 12 h. Cells were released from mitotic arrest for indicated times.

Cellular DNA Flow Cytomertic Analysis.

The single-cell suspension (1 × 10⁵ to 1 × 10⁶ cells) was prepared in 300 μL PBS, permeabilized by cold 70% ethanol for 30 min at 4 °C, and then washed and resuspended in 500 μL PBS, followed by treatment with 5 μL RNase (DNase-free) at 37 °C for 30 min, chilled on ice, and 50 μL PI (propidium iodide; Roche) treatment in the dark at room temperature for 1 h. DNA contents were acquired using BDCalibur and analyzed using the ModFit v3.3.11 software (Verity Software House).

BrdU Incorporation Assay.

Yap/Taz^fl/fl MEF cells were seeded onto four-well chamber slides. After 48 h, Yap/Taz were removed by Ad-CRE infection. Ad-GFP infection was performed as a control. After 4 d, BrdU incorporation assays were performed using BrdU cell proliferation kit (#2750; Millpore) according to the manufacturer’s protocol.

Quantitative RT-PCR Analysis.

Total RNA was prepared using TRIZOL reagent (Life Technologies) or RNAeasy mini kit (Qiagen) according to the manufacturer’s protocol. cDNA was synthesized from total RNA (1–3 μg) using SuperScript II Reverse Transcriptase (Life Technologies) with random hexamer (Roche). Quantitative real-time PCR were done with SYBR Select Master Mix on StepOnePlus thermal cycler (Applied Biosystem). The threshold cycle (Ct) value for each gene was normalized to the Ct value for GAPDH. The relative mRNA expression was calculated using ΔΔCt method. PCR primers for human samples were: CTGF, forward: AGGAGTGGGTGTGTGACGA; reverse: CCAGGCAGTTGGCTCTAATC. ANKRD1, forward: AGTAGAGGAACTGGTCACTGG; reverse: TGGGCTAGAAGTGTCTTCAGAT. E2F1, forward: GCCACTGACTCTGCCACCATAG; reverse: CTGCCCATCCGGGACAAC. YAP, forward: CCTCGTTTTGCCATGAACCAG; reverse: GTTCTTGCTGTTTCAGCCGCAG. TAZ, forward: GGCTGGGAGATGACCTTCAC; reverse: CTGAGTGGGGTGGTTCTGCT. GAPDH, forward: AGCCACATCGCTCAGACAC; reverse: GCCCAATACGACCAAATCC. Primers for mouse samples were: Ctgf, forward: CTGCCTACCGACTGGAGAC; reverse: CATTGGTAACTCGGGTGGAG. Cyr61, forward: GCTCAGTCAGAAGGCAGACC; reverse: GTTCTTGGGGACACAGAGGA. E2f1, forward: GCCCTTGACTATCACTTTGGTCTC, reverse: CCTTCCCATTTTGGTCTGCTC. Ccnd, forward: AGTGCGTGCAGAAGGAGATT, reverse: CTCTTCGCACTTCTGCTCCT. Ccne, forward: GGAAAATCAGACCACCCAGA; reverse: AGGATGACGCTGCAGAAAGT. Cdc6, forward: AGTTCTGTGCCCGCAAAGTG; reverse: AGCAGCAAAGAGCAAACCAGG. Jag-1, forward: AGAAGTCAGAGTTCAGAGGCGTCC; reverse: AGTAGAAGGCTGTCACCAAGCAAC.

Immunofluorescence.

Cells were seeded in Lab-Tek chamber slide (Thermo Scientific), fixed for 30 min in 4% PFA in PBS, and permeablized for 20 min with 0.5% Triton X-100 in PBS at room temperature. Immunofluorescent staining was done with standard procedures using the anti-CDH1 antibody (#NBP2-15840; Novus) and anti-LATS1 antibody (#sc-9388; Santa Cruz) as primary antibodies. Secondary antibodies are donkey anti-rabbit Alexa-563 and donkey anti-goat Alexa-488 antibodies (Life Technologies). Stained cells were mounted in DAPI mounting medium (H-1200; Vector Laboratories).

Antibodies.

Anti-YAP (4912 or 14074), anti–p-YAP (4911), anti-LATS1 (9153 or 3477), anti-LATS2 (5888), and anti–p-Histone 3 (S10, 9701) were purchased from Cell Signaling. Anti-TAZ (560235) antibody was from BD Transduction Laboratories. Anti-CDH1 (34-2000) and anti-CDH1 (CC43) antibodies were purchased from Life Technologies and Millipore, respectively. Anti-CTGF (ab6992), anti-TEAD4 (ab58310), and anti-CDH1 (ab3242) antibodies were from Abcam. Anti-Flag (M2, F3165), anti-GAPDH (SAB1405848), anti-YAP (WH0010413M1), and anti-TAZ (T4077) antibodies were from Sigma-Aldrich. Anti-YAP (H-125, sc-15407), anti-HA (F-7, sc-7392), anti–c-Myc (9E10, sc-40), anti-E2F1 (KH95, sc-251), anti-APC10 (B-1, sc-166790), anti-TAZ (H-70, sc-48805), anti–p-TAZ (S89, sc-17610) anti-Cyclin B (GNS1, sc-245), anti-SKP2 (H-435, sc-7164), and anti-CDC20 (H-175, sc-8358) antibodies were from Santa Cruz Biotechnology. Anti-HA (11867423001) or DAPI were from Roche or Vector Laboratories, respectively. Anti-mouse and rabbit HRP-conjugated secondary antibodies were from GE Healthcare Life Science. Anti-goat or rat HRP-conjugated secondary antibodies were from Santa Cruz and Sigma-Aldrich, respectively.

In Situ PLA.

In situ PLA experiments were performed as described previously. Primary antibodies (rabbit α-CDH1 and Goat α-LATS1 in blocking solution) was incubated at room temperature for 2 h. Cells were washed for five times for 5 min in PBS plus 0.1% Tween 20. To detect protein–protein interaction between LATS1 and CDH1, secondary proximity probes (Rabbit-PLUS and Goat-MINUS) (Olink Biosciences) were incubated for 90 min at 37 °C. Cells were washed five times for 5 min in 10 mM Tris⋅HCl (pH 7.5) plus 0.1% Tween 20 at 37 °C, then twice for 5 min in PBS plus 0.1%Tween 20. All subsequent steps were performed according to the Duolink proximity ligation assay detection kit protocol (Olink Biosciences). Wide-field images were collected using a Personal DeltaVision system (GE Healthcare) mounted on an inverted Olympus IX71 microscope with a Plan Apo 60×/1.42 oil objective lens. All images were acquired using a pco.edge sCMOS camera with 1 × 1 binning and a 1,024-pixels × 1,024-pixels imaging field. PLA excitation (100%) and DAPI excitation (10%) were collected in emission filters 632/50 and 435/48, respectively. The z-stacks were collected (21 images per stack with a z-interval of 0.1 μ) for each image. All images were deconvolved using an iterative constrained method with 10 cycles, medium noise filtering. Maximum-intensity projections were created for each deconvolved stack. All image capture and postprocessing was done using GE’s SoftWoRx software package v6.0.0.

ChIP Assay.

ChIP was performed by using a modified protocol from the ChIP Assay Kt (17–295; Millipore). Briefly, cells were cross-linked in 1% formaldehyde for 15 min at room temperature, and 0.125 M glycine was added to fixed cells to stop the reaction. Cells were washed, harvested with cold PBS, and resuspended in lysis buffer (1% SDS, 10 mM EDTA, and 50 mM Tris at pH8.1). Nuclei were disrupted by sonication (Misonix XL-2000) by performing eight rounds of 3 × 5-s pulses with at least 20-s rest between pulses and 2-min rest between rounds to obtain fragments of average 200–1,000 bp in size. The sonicated cell supernatant was diluted in ChIP dilution buffer (0.01% SDS, 1.1%Triton X-100, 1.2 mM EDTA, 20 mM Tris at pH8.1 and 167 mM NaCl). Suitable amounts of chromatin were immunopreciptated with specific antibodies overnight. Antibodies used were IgG (#026102; Life Technologies) and YAP (#14074; Cell Signaling). complexes were recovered on ChIP-grade Protein A/G plus agarose bead (#26195; Life Technologies). qPCR was performed using primer sets flanking the predicted TEAD binding sites. Primer sequences used are as follows: TBE1, forward: GCCCAGTGACTGTGGATTTT; reverse: GTCCTGCTGAAGCAGAAAGG. TBE2/3, forward: GTGTTACTGGGACCCTGTGG; reverse: CCCATCCCCTCTCCATAAAG. TBE4, forward: GGACCTACCCCTCAGCTTCT; reverse: AGAAGGGACCACAGAGACCA. TBE5, forward: CCCCTGTCTACTGGACTGTGA; reverse: CTGCAAGTCCCATTTTAGCC. HBB (negative control), forward: GCTTCTGACACAACTGTGTTCACTAGC; reverse: CACCAACTTCATCCACGTTCACC. CTGF (positive control), forward: TGTGCCAGCTTTTTCAGACG; reverse: TGAGCTGAATGGAGTCCTACACA.

Immunoprecipitation and Immunoblotting.

Cells were prepared using a lysis buffer [20 mM Tris (pH 7.4), 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 1 mM EGTA, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM sodium orthovanadate] or RIPA buffer (Santa Cruz Biotechnology), respectively, containing protease inhibitor mixture (Roche). Immnoprecipitates or total cell lysates were analyzed by Western blotting according to standard procedures.

YAP/TAZ Immunofluorescence.

HeLa cells were synchronized by DTB. Cells were stained with anti-YAP/TAZ antibody (CST #8418S) overnight at 4 °C. Samples were then treated with Alexa Fluor 595 Tyramide reagent (Alexa Fluor Tyramide SuperBoost Kit, B40925; Life Technologies) according to the manufacturer’s protocol, and mounted with DAPI mounting medium (H-1200; Vector Laboratories). Images were acquired using Zeiss 510 NLO Meta.

Reporter Assay.

YAP/TAZ transcription activities were examined in cells seeded in 24-well plates that have been cotransfected with Sd (Scalloped, the Drosophila homolog of TEAD)-dependent luciferase reporter construct with pTK-Renilla and effector plasmids (34). The luciferase activities were analyzed using a dual-luciferase reporter assay kit (Promega) according to the manufacturer’s instructions. In knockdown experiments, cells were first transfected with 20 nM siRNA for 48 h before luciferase assay.

Fly Stocks, Transgenes, Immunostaining, and Western Blot.

Transgenic lines used were: UAS-Cdh1-RNAi (VDRC #25553 and #25550); wts^{MI05605-GFSTF.0}, a protein trap insertion transgenic line expressing GFP-Wts (Bloomington Stock #56808); Diap-GFP (34), UAS-Myc-Wts (9), GMR-GAl4, UAS-Yki, UAS-YkiS168A and UAS-SD-GA (34), and UAS-Yki3SA (78). The antibodies used for immunostaining and Western blot analysis were rabbit anti-GFP (Molecular Probes), mouse anti-Myc (Santa Cruz), rat anti-Ci, 2A1 (DSHB), and Cy2-, Cy3-conjugated secondary antibodies (Jackson Immuno Research Laboratories). Images were captured by Zeiss LSM710 confocal microscopy. For protein extract of eye discs, eye imaginal discs were collected from the late third-instar larvae and lysed in Nonidet P-40 cell lysis buffer with protease inhibitor mixture (Roche). Lysates were cleared by centrifugation and then subjected to SDS/PAGE.

Statistical Analysis.

All experiments were performed at least three independent times (unless noted otherwise) and representative data are shown. Statistical analysis between groups was performed by two-tailed Student’s t test in Graphpad Prism 7 to determine significance when only two groups were compared. One-way ANOVA with Tukey’s post hoc tests were used to compare differences between multiple groups. P values of less than 0.05 and 0.01 were considered significant. Error bars on all graphs are SD unless otherwise indicated.