Fig 2: CCNK stimulates YAP activation in HCC.(A) YAP signaling is increased in HCC with high CCNK expression. NES, normalized enrichment score. (B and C) YAP translocates into the cytoplasm in CCNK-depleted cells by immunofluorescence staining (B) and cell fractionation (C). Scale bars, 20 μm. (D and E) Analysis of YAP signaling in CCNK-depleted Huh7 cells. (F) Analysis of TEAD-responsive reporter (8×GTIIC-luc) in HEK-293T cells upon CCNK ablation. (G and H) Immunoblot and ubiquitination assay evaluating 14-3-3/YAP interaction (G) and ubiquitination (H) in cells with or without CCNK depletion. MG132, 20 μM for 6 hours. (I and J) Evaluation of YAP subcellular distribution in CDK12-depleted cells. Scale bars, 20 μm. (K and L) Analysis of YAP activation in cells upon CDK12 loss. (M) Analysis of TEAD-responsive reporter in HEK-293T cells with or without CDK12 ablation. (N and O) Evaluation of YAP localization in Huh7 cells following SR-4835 treatment (0.5 μM) for 8 hours. Scale bars, 20 μm. (P and Q) Evaluation of YAP activity in cell upon SR-4835 treatment (0.25 and 0.5 μM). (R) Analysis of TEAD-responsive reporter in HEK-293T cells treated with or without SR-4835 treatment (0.5 μM) for 8 hours. (S and T) Evaluation of 14-3-3/YAP interaction and ubiquitination in cells treated with SR-4835 (0.5 μM, 8 hours). MG132, 20 μM for 6 hours. (U) Correlation analysis of CCNK expression with YAP and its targets in the TCGA dataset. (V) Heatmap demonstration of correlation between Cyclin K and YAP expression in HCC tissues (n = 90). Scale bars, 50 μm. The color scale indicates the IHC score of YAP or Cyclin K for each HCC tissue. For [(E), (F), (L), (M), (Q), and (R)], P values were determined by one-way ANOVA. For [(U) and (V)], the Pearson’s correlation test was used to analyze the link. Data are presented as means ± SD. *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig 3: Targeting CDK12/Cyclin K–YAP signaling addiction provides therapeutic vulnerability in HCC.(A to C) Control or YAP-depleted Huh7 cells were transfected with or without siRNA targeting CCNK and subjected to immunoblotting (A), qPCR analysis (B), and reporter activity analysis (C). (D to F) Control or YAP-depleted Huh7 cells were treated with or without SR-4835 and subjected to immunoblotting (D), qPCR analysis (E), and reporter activity analysis (F). (G) Control or YAP-depleted Huh7 cells with or without CCNK overexpression were implanted into nude mice. Gross image and quantification of xenograft weight are shown. Scale bar, 10 mm. (H) Correlation analysis of YAP with CCNK or CDK12 in HCC cell lines. Transcriptional datasets were download from the GEO (GSE97098). The Pearson’s correlation test was used to analyze the link. (I) Cluster analysis of CDK12, CCNK, and YAP coexpression. JHH2, MHCC97H, SNU398, and CLC40 cell lines were indicated. The color scale indicates the level of each gene presented as reads per kilobase per million mapped reads (RPKM). (J) Protein levels of CCNK, YAP, and CDK12 in indicated HCC cell lines. (K) Quantification of cellular sensitivity (IC50) to Cyclin K molecular glue degraders in indicated cell lines. IC50 values are shown. (L) Determination of the mRNA and protein levels of CCNK, YAP, and CDK12 in Huh7 and PDX models. (M) Evaluation of tumor growth of vehicle and SR-4835 (5 mg/kg per day) treatment in the PDX mouse model. Gross xenografts and quantification of xenograft weight are shown. Scale bars, 10 mm. (N) Histological analysis of xenografts from indicated PDX models. Scale bars, 50 μm. For [(B), (C), (E), (F), and (G)], P values were determined by one-way ANOVA. The unpaired t test was used in (M) to determine statistical significance. Data are presented as means ± SD. *P < 0.05, **P < 0.01, and ***P < 0.001; n.s., not significant.

Fig 4: CDK12/Cyclin K physically associates with YAP.(A) Coomassie Brilliant Blue staining showing Cyclin K immunoprecipitates. (B) List of candidate interactors related to canonical YAP signaling. (C) Co-IP of endogenous Cyclin K and Hippo signaling components. (D) YAP interacts with Cyclin K and CDK12. (E) Immunostaining showing the colocalization of YAP and Cyclin K in Huh7 cells. Left: Huh7 cells transfected with exogenous HA-Cyclin K and Flag-YAP. Scale bars, 10 μm. Quantification of the fluorescence intensity along the dashed line indicated in the merged image is shown (middle). Right: Costaining of endogenous YAP with Cyclin K or CDK12 in Huh7 cells with low or high density. Scale bars, 10 μm. (F) Immunostaining for YAP and CDK12 in human HCC tissues. Scale bars, 10 μm. (G) GST pull-down assay of in vitro interaction between Cyclin K and YAP. Arrows indicate the proteins at the expected molecular weight. (H) Cyclin K was required for YAP/CDK12 interaction. (I and J) Mapping the region(s) of YAP or Cyclin K that mediates the interaction. Diagram shows the domains of YAP or Cyclin K protein. aa, amino acids. (K) Cyclin K interacts with YAP via two PPxY motifs. Individual PPxY mutant, PPxY-M1 or PPxY-M2. Double mutant, PPxY-2M. (L to N) YAP activity or TEAD-responsive reporter analysis in Huh7 cells with vector, WT CCNK, or PPxY-2M. (O) Cell fractionation evaluating YAP localization in Huh7 cells with vector, WT CCNK, and PPxY-2M. (P and Q) Determination of 14-3-3/YAP interaction and YAP ubiquitination in HEK-293T cells with vector, WT CCNK, or PPxY-2M. (R) Quantification of xenograft weight in indicated groups (n = 5). Gross images of xenografts are shown. Scale bar, 5 mm. For [(L), (M), and (Q)], P values were determined by one-way ANOVA. Representative data are presented as means ± SD. *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig 5: YAP co-occupies CDK12/Cyclin K LLPS.(A) IDR prediction across Cyclin K and CDK12 proteins. (B) Purified EGFP-CCNK before (left) and after (middle) the addition of indicated concentration of PEG-8000. The liquid droplets were visualized using fluorescence microscopy. Scale bars, 5 μm. (C) FRAP images of EGFP-CCNK condensates. Representative images were acquired after photobleaching the indicated ROI with fluorescence signal recovery curves. The dotted circle indicates the region of photobleaching. Scale bars, 2 μm. (D and E) Endogenous Cyclin K and CDK12 formed condensates in Huh7 cells were treated with 0.2 M sorbitol for 20 min. Scale bars, 10 μm. (F) CCNK-EGFP formed nuclear foci after sorbitol treatment for indicated times. Scale bars, 20 μm. (G) Time-lapse imaging of the fusion of nuclear CCNK-EGFP droplets. Scale bars, 1 μm. (H) FRAP images of nuclear CCNK-EGFP condensates. Scale bars, 1 μm. Quantification of the FRAP curve of CCNK-EGFP condensates are shown (n = 7). (I) Confocal images of Huh7 cells expressing either WT CCNK (left) and CDK12 (right) or an IDR region truncated protein with or without sorbitol treatment (0.2 M) for 20 min. Scale bars, 10 μm. (J) Representative images of CDK12 and Cyclin K condensates (top) and individual CCNK, YAP (middle), or both CCNK/YAP (bottom). Nuclei were counterstained with DAPI. Scale bars, 10 μm. (K) Purified EGFP-CCNK and mCherry-YAP proteins were incubated with 20% PEG-8000. Images were acquired under fluorescence and bright-field illumination. Scale bars, 5 μm. (L) Condensates of Flag-CDK12, HA-CCNK, and Myc-YAP in Huh7. Quantification of fluorescence intensity along the dashed line indicated in the merged image is shown. Scale bars, 5 μm. (M) Condensates of CDK12 and YAP in Huh7 cells with or without Cyclin K depletion. Scale bars, 10 μm. Data are presented as means ± SD.