Downregulation of PCYT2 correlates with poor prognosis in CRCs. To evaluate the potential role of PCYT2, a key rate-limiting enzyme of PE synthesis pathway in CRC progression, we analyzed PCYT2 protein expression difference between tumor and normal tissue in the publicly available datasets. We found a significant reduction in PCYT2 expression in CRC (Figure 1A) and pancancer (Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.178823DS1). We also observed that low levels of PCYT2 protein in the datasets positively correlated with poor disease-free survival (DFS) and overall survival (OS) (Figure 1, B and C). To further verify PCYT2 expression, we collected 33 paired samples of human CRC tissue and corresponding normal mucosa tissue from West China Hospital, Sichuan University. Western blot and IHC showed significantly lower PCYT2 expression in most CRC samples compared with expression in paired adjacent normal tissues (Figure 1, D and E). We also examined the expression of PCYT2 in CRC cell lines (Lovo, SW480, HCT116, SW620, COLO320, and HT29) and the normal colon cell line NCM460 and observed the highest expression in SW480 and SW620 cells compared with lower expression levels in HCT116 and HT29 cells (Figure 1F). To explore the role of PCYT2 in CRC metastasis, we examined PCYT2 expression in patients with CRC with liver metastases, as the liver is the most common metastasis site in CRC. Intriguingly, IHC revealed significantly lower expression of PCYT2 in live metastatic tumor compared with primary tumor (Figure 1G), implying the possible involvement of PCYT2 in CRC liver metastasis. This possibility is supported by the results of IHC staining in 40 CRC patient tissues, which indicated that patients with CRC with expression of PCYT2 (n = 21) had shorter progression-free survival (PFS) times, whereas patients with high PCYT2 expression (n = 19) had longer PFS times (Figure 1H). Collectively, these results indicate that PCYT2 is critical for tumor development and metastasis and may serve as a tumor suppressor in CRC.

The downregulation of PCYT2 positively correlates with poor prognosis in colorectal cancer. (A) The relative protein expression of PCYT2 in public Clinical Proteomic Tumor Analysis Consortium (CPTAC) datasets exhibited a lower level in primary tumors compared with normal tissues. (B and C) Disease-free survival (DFS) (B) and Overall survival (OS) (C) of colorectal cancer patients in CPTAC datasets. (D) Detection of PCYT2 expression in 16 paired CRC tissues (T) and the corresponding adjacent normal tissues (N) by Western blot. (E) Representative images of PCYT2 immunohistochemical detection in paired patients with CRC tissues and their normal tissues. Representative images and quantification. Scale bar: 100 μm. P value was calculated with 2-tailed paired Student’s t test (****P < 0.0001). (F) Detection of PCYT2 expression in CRC cell lines and the normal intestinal epithelial cell line NCM460 by Western blot. (G) Representative images of PCYT2 IHC detection in patients with CRC paired with primary tumor, normal tissues, liver, and liver metastases. Representative images and quantification. Scale bar: 50 μm. P value was calculated with 1-way ANOVA with Tukey’s multiple-comparison test (*P < 0.05, **P < 0.01, ****P < 0.0001). (H) Estimation of the relationship between PCYT2 expression and PFS times in 40 patients with CRC by Kaplan-Meier analysis. P value was calculated with log-rank test (P = 0.0032).

PCYT2 inhibits CRC migration and EMT in vitro via its cytidine transferase domain. EMT is considered an essential determinant in cancer progression (21). EMT refers to the reversible biological process by which epithelial cells are transformed into mesenchymal phenotypes through specific procedures (22). EMT is driven by Snail1, Snail2, zinc-finger E-box-binding (ZEB), and 2-handed E box binding zinc finger protein SIP1 and basic helix–loop–helix (bHLH) transcription factors, which repress epithelial marker genes and activate genes associated with the mesenchymal phenotype (23–28). To ascertain the function of PCYT2 expression on the metastasis of CRC cells, we constructed HCT116 and HT29 cells stably overexpressing PCYT2 and found that overexpression of PCYT2 led to upregulation of the epithelial marker E-cadherin but downregulation of mesenchymal markers (N-cadherin and Vimentin) and key transcription factors (Snail1, Twist, Snail2, and ZEB1) (Figure 2, A–D). Meanwhile, PCYT2 knockdown in SW480 and SW620 cells decreased epithelial marker expression and increased mesenchymal marker expression (Supplemental Figure 2, A–C). To investigate cell migration, we employed scratch wound and Transwell assays and observed a noticeable reduction in the migratory ability of HCT116 cells stably expressing PCYT2 (Figure 2, E and F). In contrast, the migration abilities were improved in PCYT2-silenced CRC cells (Supplemental Figure 2, D and E). To further examine migration capabilities, we examined the state of focal adhesion formation by measuring paxillin clustering levels. Paxillin is a core component of focal adhesion complexes, and its clustering is a characteristic of focal adhesion formation (29). In HCT116 cells stably overexpressing PCYT2, the number of paxillin fluorescence spots were significantly reduced compared with control (Figure 2G). PCYT2-deficient SW480 cells showed a significant increase in the number of paxillin fluorescence spots (Supplemental Figure 2F), indicating the critical role of PCYT2 in the regulation of CRC migration. Overexpression of PCYT2 increased PE levels, while knockdown of PCYT2 decreased PE levels, indicating that PCYT2 plays a key rate-limiting role in the PE synthesis pathway (Supplemental Figure 2, G–J). Intriguingly, PCYT2 knockdown did not affect the proliferation of CRC cells (Supplemental Figure 3, A–D).

Inhibition of CRC cell migration and EMT in vivo and in vitro by PCYT2 overexpression. (A and B) Detection of epithelial and mesenchymal markers in PCYT2-overexpression HT29 (A) and HCT116 (B) cells by Western blot. (C and D) Transcriptional alteration of epithelial and mesenchymal marker genes in PCYT2-overexpressed HCT116 (C) and HT29 (D) cells detected by quantitative PCR. (E) Representative images of scratch assays in PCYT2-overexpression HCT116 cells and control group. The graph shows the area of a wound evaluated with ImageJ (NIH). Representative images and quantification. Scale bar: 50 μm. (F) Investigation of the cell migratory and invasion ability of PCYT2-overexpression HCT116 cells by Transwell assay. Representative images and quantification. Scale bar: 100 μm. (G) Representative images of immunofluorescence assays for the membrane localization of paxillin in HCT116 cells with ectopic expression of PCYT2. Different colors represent different antibodies or dyes: DAPI (blue), paxillin (green), and phalloidine (red). Representative images and quantification. Scale bar: 20 μm. (H) Schematic diagram of WT PCYT2 and cytidine transferase catalytic domain deletion mutants used in the study. (I) Representative images and quantification of Transwell assays for shPCYT2 cells with ectopic expression of WT PCYT2 or deletion of the cytidine transferase catalytic domain of PCYT2. Statistical chart and representative images. Scale bar: 100 μm. Data represent the mean ± SD (n = 3). Statistical significance was assessed with 2-tailed unpaired Student’s t test (C–G) or 1-way ANOVA with Tukey’s multiple-comparison test (I). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

The PCYT2 protein contains 2 catalytic domains responsible for cytidyl transferase activity: the C-terminal cytidyl transferase domain and the N-terminal cytidyl transferase domain (30). We generated a mutant variant of PCYT2 lacking the functional cytidyl transferase catalytic domain (Figure 2H) to investigate this domain on the migratory capacity of CRC cells. First, we observed that exogenous overexpression of PCYT2 in PCYT2-knockdown HCT116 cells restored their migration ability (Figure 2I). However, only partial restoration of migration ability was observed in PCYT2-knockdown HCT116 cells upon transfection with mutants lacking either the C-terminal transferase domain or the N-terminal transferase domain (Figure 2I). Conversely, no change in migration ability was observed in PCYT2-knockdown HCT116 cells when they were transfected with mutants with deletions in both the C-terminal and N-terminal transferase domains (Figure 2I). Collectively, these results suggest that PCYT2 inhibits the metastasis of CRC in vitro, and the inhibitory effect of PCYT2 on the metastasis of CRC cells in vitro is mediated by its cytidine transferase domain.

The overexpression of PCYT2 suppresses the metastasis of CRC in vivo. To investigate the effect of PCYT2 on in vivo metastasis of CRC cells, we established an orthotopic mouse model by intracecal injection of PCYT2-overexpression and control (vehicle) luciferase-labeled metastatic HCT116 cells (n = 10 per group) (Figure 3A). The progression and distribution of tumors were continuously evaluated by weekly injection of fluorescein substrates using bioluminescence imaging. We found that the fluorescence intensity of the PCYT2-overexpression group mice was weaker than that of the control group mice, indicating a decelerated growth rate of tumor cells in PCYT2-overexpression group mice (Figure 3B). After the fluorescence intensity of in vivo imaging reached its maximum, we dissected all mice to observe tumor growth. Our findings reveal that 60% of the mice in the control group exhibited metastases to distant sites, including the peritoneum, liver, kidney, and spleen, whereas only 20% of the mice from the PCYT2-overexpression group showed peritoneal metastasis (Figure 3, C–E). The pathological H&E staining also demonstrated the presence of tumor cells in the cecum, peritoneum, liver, kidney, and spleen tissues of the control group mice, while tumor cells were only observed in the cecum and peritoneum of PCYT2-overexpression group mice (Figure 3F).

Overexpression of PCYT2 inhibits metastasis of colorectal cancer cells in mice. (A) Schematic diagram of orthotopic mouse model of CRC. (B) Representative images of fluorescence in orthotopic mouse after injection of CRC cells each week. (C) Representative images of abdominal tumors in mice after 30 days of dissection. (D) Representative images of fluorescence in mouse organs. LN, lymph node. (E) The statistics of tumor metastasis rate. (F) Representative images of H&E. The scale bar of the enlarged image is 20 μm, and the scale bar of the full image is 1 mm.

The liver is the most common site for distant metastasis of CRC, followed by the lung as the second most frequent target organ for metastasis of CRC (31, 32). To determine whether PCYT2 also has an inhibitory effect on the metastasis of CRC to the lungs, we also established mouse models by injecting PCYT2-overexpression and control (vehicle) luciferase-labeled metastatic HCT116 cells into BALB/c nude mice (n = 7 per group) (Supplemental Figure 4A). Tumor progression and distribution were evaluated by serial bioluminescence imaging after the injection of the luciferin substrate. Overexpression of PCYT2 significantly reduced the fluorescence in the lungs compared with the vehicle group (Supplemental Figure 4, B–D). Mice injected with PCYT2-overexpression HCT116 cells had a lower rate of metastasis (2 of 7 mice) than those injected with vehicle control cells (6 of 7 mice) (Supplemental Figure 4, E and F). Pathological changes in ectopic foci tissues were detected using H&E and IHC staining. Mice injected with PCYT2-overexpression HCT116 cells showed a significant decrease in the number of metastatic tumor nodules and nodule size on the surface of the lungs compared with mice injected with vehicle control cells (Supplemental Figure 4G). Meanwhile, the IHC results show that PCYT2 was significantly overexpressed in cancer nodules (Supplemental Figure 4H). Collectively, these results suggest that PCYT2 inhibits the metastasis of CRC in vivo.

PCYT2 increases YAP1 phosphorylation and relocates YAP1 to cytoplasm for degradation. To investigate the underlying mechanisms by which PCYT2 suppresses metastasis, we first performed RNA-Seq in SW480 cells with PCYT2 knockdown (Figure 4A) and identified 224 differentially expressed genes, including 136 upregulated and 88 downregulated genes with the threshold of (log2[fold-change]) > 1 and P < 0.01 (Figure 4B) (Supplemental Table 1). Pathway enrichment analysis of the differentially expressed genes was carried out using the Kyoto Encyclopedia of Genes and Genomes (KEGG) public database (http://www.genome.jp/kegg/), suggesting an enrichment for several signaling pathways, including the Hippo and PI3K/AKT pathways (Figure 4C). Other biological processes (BPs), such as extracellular matrix disassembly and cell adhesion, were also enriched (Figure 4C), indicating that PCYT2 is possibly involved in the regulation of CRC cell migration. Among the enriched pathways identified, we particularly focused on the Hippo signaling pathway. The role of the Hippo pathway in EMT has been gradually elucidated (33). The mammalian pathway consists of core kinases (MST1/2 and LATS1/2), adaptor proteins (SAV1 for MST1/2 and MOB1 for LATS1/2), downstream effectors (YAP and its analog TAZ), and nuclear transcription factors (TEAD1-4) (34). The Hippo signaling components LATS1/2, each encoding a tumor-suppressive serine or threonine-protein kinase, phosphorylate their effectors TAZ/YAP, causing YAP1 and TAZ’s cytoplasmic retention and destabilization. This retention in the cytoplasm inhibits TAZ/YAP activity by preventing their translocation to the nucleus, resulting in the transcriptional inhibition of downstream targets such as oncogene c-MYC and mesenchymal genes ZEB1/SNAIL2, which drive metastasis (34–38). During tumor development, YAP1 phosphorylation is inhibited, and its downstream regulators YAP1 and TAZ enter the nucleus, where they increase the transcriptional activation of genes involved in cell proliferation and metastasis (36, 39).

PCYT2 affects the phosphorylation levels of YAP1 and changes the proportion of YAP1 in the nucleus. (A) Heatmap showing upregulated and downregulated genes after PCYT2 knockdown (P < 0.05, [FC Log2] > 1). (B) Volcanic maps of upregulated and downregulated genes with PCYT2 knockdown (P < 0.05, [FC Log2] > 1). Red represents significantly upregulated genes, and green represents significantly downregulated genes. (C) Gene Ontology (GO) terms and KEGG pathways related to PCYT2 knockdown (P < 0.05, [FC Log2] > 1). (D and E) Detection of the YAP1 and p-YAP protein expression levels with PCYT2 knockdown (D) or PCYT2 overexpression (E) CRC cells in cytoplasm and nucleus, respectively, by Western blot. (F and G) Immunofluorescence analysis for the nuclear proportion of YAP1 in HT29, HCT116, and SW480 cells after PCYT2 overexpression (F) or knockdown (G). Different colors represent different antibodies or dyes: DAPI (blue), YAP1 (green), and phalloidine (red). Scale bar: 20 μm. (H and I) Detection of the expression changes of YAP1 in PCYT2-overexpression HCT116 (H) and HT29 (I) cells treated with MG132 by Western blot. (J) The enrichment of YAP1 in CTGF, ZEB1, and Snail2 promoters detected by ChIP assay. (K) Detection of epithelial and mesenchymal markers in CRC cell ectopic expression of vehicle, 3MYC-PCYT2, 3HA-YAP1, and 3MYC-PCYT2/3HA-YAP1 by Western blot. (L) Transwell assay was performed to detect cell migration capacity in HCT116 cell ectopic expression of vehicle, 3MYC-PCYT2, 3HA-YAP1, and 3MYC-PCYT2/3HA-YAP1. Statistical chart and representative images. Scale bar: 100 μm. (M) Detection of YAP1, p-YAP1, and Snail2 protein level in PCYT2- knockdown HCT116 cell ectopic expression of ectopic expression of WT PCYT2 or deletion of the cytidine transferase catalytic domain of PCYT2. Data represent the mean ± SD (n = 3). Statistical significance was assessed with 2-tailed unpaired Student’s t test (J) or 1-way ANOVA with Tukey’s multiple-comparison test (L). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

To investigate the effect of PCYT2 on the Hippo pathway, the expression of its core effectors YAP1/TAZ and the phosphorylation status of YAP1 at Ser127 were determined in PCYT2-knockdown CRC cells. Western blot indicated an increase in YAP1 expression and a decrease in the phosphorylation of YAP1 compared with control, while TAZ expression had a little change (Supplemental Figure 5A). In contrast, a decrease in total YAP1 expression and an increase in the phosphorylation of YAP1 were observed in PCYT2-overexpression CRC cells (Supplemental Figure 5B). YAP1 acts as the core component of the Hippo pathway, and nuclear YAP localization indicates Hippo pathway inactivation (34). Therefore, we detected the amount of YAP1 in the nucleus and cytoplasm by nucleocytoplasmic separation in CRC cells with PCYT2 either knocked-down or overexpressed. We observed the remarkable increase of YAP1 in the nucleus and reduction of phosphorylated YAP1 (p-YAP1) in the cytoplasm of PCYT2-knockdown CRC cells (Figure 4D). Consistently, YAP1 was decreased significantly in the nucleus, while p-YAP1 was increased significantly in the cytoplasm of PCYT2-overexpressed CRC cells (Figure 4E). In line with the results of nucleocytoplasmic separation, the immunofluorescence results also clearly indicated that PCYT2 overexpression led to a marked reduction in YAP1 in the nucleus and to an obvious increase in the cytoplasm (Figure 4F). Meanwhile, YAP1 was translocated into the nucleus after PCYT2 downregulation (Figure 4G). These data indicate the critical role of PCYT2 in the regulation of YAP1 phosphorylation and nuclear localization.

p-YAP1 is ubiquitinated and degraded via the ubiquitin-proteasome pathway (40). Because PCYT2 affected p-YAP1 levels (Figure 4, D and E, and Supplemental Figure 5, A and B), we speculated that changes in PCYT2 expression would eventually lead to altered YAP1 ubiquitin levels. Thus, we investigated the ubiquitination level of YAP1 and found that it was indeed reduced in PCYT2-knockdown CRC cells (Supplemental Figure 5C), while PCYT2-overexpression HCT116 cells tended to increase ubiquitin levels (Supplemental Figure 5D). Simultaneously, we treated PCYT2-overexpression HCT116 and HT29 cells with proteasome inhibitor MG132, and we found that MG132 inhibited the degradation of YAP1 in the PCYT2-overexpression group (Figure 4, H and I); however, it had no effect on PCYT2-knockdown SW480 cells (Supplemental Figure 5E). Since YAP1 phosphorylation represses its transcriptional activity by promoting YAP1 accumulation in the cytoplasm (34, 35), we evaluated the effect of PCYT2 on YAP1’s transcriptional ability in regulating the EMT transcription program. We observed a significant reduction in YAP1 bound to proximal promoter regions of CTGF, ZEB1, and SNAIL2 in HCT116 cells with ectopic PCYT2 expression (Figure 4J). In contrast, loss of PCYT2 markedly increased the enrichment of YAP1 in proximal promoter regions of these genes (Figure 4J). In addition, we examined the effect of YAP1 on EMT and migration and found that YAP1 could promote the migration and EMT of CRC (Supplemental Figure 6, A–E). Importantly, expression of YAP1 in HCT116 cells counteracted the capacity of PCYT2 to repress cell migration (Figure 4, K and L), supporting the idea that YAP1 functions as a target of PCYT2. Simultaneously, we observed that overexpression of PCYT2 effectively reinstated the expression of YAP1, p-YAP1, and the mesenchymal marker Snail2 in PCYT2-knockdown HCT116 cells (Figure 4M). However, overexpression of mutants lacking the cytidylytransferase domain of PCYT2 failed to elicit such restorative effects in PCYT2-knockdown HCT116 cells (Figure 4M). Therefore, these data suggest that YAP1 is one of potential targets repressed by PCYT2 in manipulating the EMT transcription program.

PCYT2 regulates EMT in CRC by inducing YAP1 relocation. Lipids are able to transduce signals in a variety of physiological processes through specific lipid-protein interactions (18, 41). Therefore, we hypothesized that PCYT2 functions in the regulation of EMT by affecting PE synthesis and function in lipid-protein interactions. PEBP1 (also known as Raf1 kinase inhibitory protein [RKIP]) belongs to the family of PE-binding proteins (PEBPs) and has been implicated in various diseases, including cancer, Alzheimer’s disease, diabetes, and retinopathies (42–46). In particular, PEBP1 is reported to suppress cell motility and metastasis development and is downregulated in prostate cancer (45). Previous studies have shown that PE binding to Pebp1 disrupts the association of Pebp1 with Raf-1 and MAPK kinase 1 (MEK1) while enhancing its interaction with kinases upstream of NF-κB (18, 47, 48). However, the specific signaling pathways regulated by PE-PEBP1 interaction in CRC metastasis remain elusive.

Therefore, we wondered whether PCYT2 functions in EMT regulation by affecting the binding of PE with PEBP1. We initially investigated whether the upregulation of PEBP1 leads to the relocation of YAP1 to the nucleus. In concert with the finding of PCYT2-mediated YAP1 relocation, overexpression of PEBP1 in CRC cells led to a significant increase in p-YAP1 accumulation in the cytoplasm and a decrease in YAP1 accumulation in the nucleus (Figure 5A), suggesting that YAP1 may undergo ubiquitination and protein degradation. Conversely, downregulation of PEBP1 resulted in markedly decreased expression of p-YAP1 and increased accumulation of YAP1 in the nucleus (Figure 5B). Meanwhile, the results of immunofluorescence analysis were consistent with those of Western blot, revealing that knockdown of PEBP1 in SW480 cells enhanced the accumulation of YAP1 in the nucleus and that overexpression of PEBP1 reduced the aggregation of YAP1 in the nucleus. (Figure 5, C–E). Additionally, we observed the inhibitory role of PEBP1 in CRC cell migration (Supplemental Figure 7, A–G). In addition, ectopic expression of YAP1 in HCT116 cells counteracted the repressive effect of PEBP1 on cell migration (Figure 5F). Collectively, these results imply that PEBP1 inhibits CRC migration by altering YAP1 distribution in the nucleus, similar to the effect of PCYT2.

PCYT2 regulates EMT in CRC via YAP1 translocation induced by PEBP1. (A and B) Western blot results show the expression of YAP1 and phosphorylated YAP1 (p-YAP1) in the nuclear and cytoplasmic of PEBP1-overexpression HCT116 (A) and PEBP1-knockdown (B) SW480 cells. (C–E) Detection of the expression changes in the nuclear proportion of YAP1 in CRC cells with PEBP1-knockdown (C) or PEBP1-overexpression (D and E) by immunofluorescence assays. Different colors represent different antibodies or dyes: DAPI (blue), YAP1 (green), and phalloidine (red). Scale bar: 20 μm. (F) Detection of the cell migration capacity in ectopic expression of vehicle, 3MYC-PEBP1, 3HA-YAP1, and 3MYC-PEBP1/3HA-YAP1 HCT116 cells by Transwell assay. Statistical chart and representative images. Scale bar: 100 μm. (G) Detection of the expression of Snail2, YAP1, and p-YAP1 in shPEBP1, 3MYC-PCYT2, and shPEBP1/3MYC-PCYT2 HCT116 cells by Western blot. (H) Transwell assays in HCT116 cells described in G. Scale bar: 100 μm. Statistical chart and representative images. (I) Detection of the expression of YAP1, p-YAP1, and epithelial and mesenchymal markers in PEBP1-knockdown HCT116 cells with PCYT2 or kinase inactive PCYT2 mutant. (J) Representative images and of Transwell assays for shPEBP1 cells with ectopic expression of WT PCYT2 or deletion of the cytidine transferase catalytic domain of PCYT2. Statistical chart and representative images. Scale bar: 100 μm. Data represent the mean ± SD (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001 by 1-way ANOVA with Tukey’s multiple-comparison test.

To further clarify the role of the PCYT2/PEBP1 axis on YAP1 relocation, we knocked down PEBP1 in PCYT2-overexpression HCT116 cells. We observed that PEBP1 deficiency compensates the loss of YAP1 caused by PCYT2 overexpression and that YAP1 supplementation partially abolished PCYT2 inhibitory role in CRC migration (Figure 5, G and H). The effects of overexpressing PCYT2 or its cytidyltransferase deletion mutant on the migration phenotype of PEBP1-knockdown HCT116 cells were simultaneously examined. We found that overexpression of full-length PCYT2 restored the expression of YAP1, p-YAP1, and migration ability in PEBP1-knockdown cells (Figure 5, I and J). However, only partial restoration of the expression of YAP1, p-YAP1, and migration ability was observed in PEBP1-knockdown HCT116 cells when they were transfected with mutants lacking either the C-terminal transferase domain or the N-terminal transferase domain alone (Figure 5, I and J). Conversely, overexpression of the deletion of both N-terminal and C-terminal cytidylytransferase domains of PCYT2 did not affect the expression of YAP1, p-YAP1, or migration ability changes in PEBP1-knockdown HCT116 cells (Figure 5, I and J). Therefore, these data suggest that PCYT2 regulates cell migration by affecting the nuclear distribution of YAP1 through PEBP1.

PEBP1 binds to PPP2R1A phosphatase but not YAP1. Next, we explored the mechanism by which the PCYT2/PEBP1 axis affects YAP1 phosphorylation levels and nuclear translocation. Binding of PE to PEBP1 has been shown to disrupt the association between PEBP1 and certain kinases (18, 47). Thus, we speculated that PEBP1 may bind to the upstream kinases of YAP1 to affect the protein phosphorylation cascade in the Hippo pathway. To validate this hypothesis, we initially examined whether PEBP1 interacts with MST1 and LATS1, the upstream kinases of YAP1. However, we did not observe the coimmunoprecipitation (Co-IP) of MST1/LATS1 with PEBP1 in CRC cells (Figure 6A). This result prompted us to speculate the possibility that PEBP1 probably does not function in direct interaction with kinases involved in YAP1 phosphorylation. To further identify the regulatory network of YAP1, we employed affinity purification–mass spectrometry (MS) to uncover the proteins associated with PEBP1. Among 201 candidate proteins identified in GFP-PEBP1–overexpressing cells, PPP2R1A — a subunit of PP2A — attracted our attention (Supplemental Table 2) since PPP2R1A was listed as interacting with YAP1 in the BioGRID database (https://thebiogrid.org/). To confirm the interaction among PEBP1, PPP2R1A, and YAP1, we performed Co-IP using specific antibodies against each protein in CRC cells. We found that PEBP1 indeed interacts with PPP2R1A and YAP1 and vice versa (Figure 6, A–C). The colocalization of these proteins in the cytoplasm detected by immunofluorescence was also observed (Figure 6, D and E). To verify the authenticity of the interactions detected in vivo, we performed an in vitro GST pull-down assay. Indeed, the PEBP1-PPP2R1A and YAP1-PPP2R1A interactions were observed but not the PEBP1-YAP1 interaction (Figure 6, F–H), suggesting that PEBP1 likely regulates YAP1 function by interacting with PPP2R1A. Intriguingly, this mutual interaction between YAP1 and PPP2R1A was significantly reduced following the increase in PEBP1 protein level, implying that PEBP1 can compete with YAP1 to bind PPP2R1A, which in turn stabilizes the YAP1 protein (Figure 6I). This possibility is supported by the finding that the PP2A complex negatively regulates Hippo signaling by dephosphorylating YAP1 (49, 50). These results suggest that PEBP1 may affect the phosphorylation of YAP1 by binding to PPP2R1A. Therefore, we speculate that the change in PCYT2 expression could affect the interaction between PEBP1 and downstream proteins. To test this possibility, we performed Co-IP using a specific antibody against PPP2R1A to detect interactions among PPP2R1A, PEBP1, and YAP1. The results show that PCYT2 knockdown weakened the binding of PPP2R1A to PEBP1, whereas it induced the binding of YAP1 with PPP2R1A (Figure 6J). In contrast, overexpression of PCYT2 enhanced the binding of PPP2R1A to PEBP1, while it weakened the binding of YAP1 to PPP2R1A (Figure 6K). These findings indicate that PCYT2 regulates the phosphorylation of YAP1 by promoting the association of PEBP1 with PPP2R1A in order to impede the interaction of PPP2R1A with YAP1.

PPP2R1A interacts with YAP1 and PEBP1. (A) Detection of the interaction between PEBP1 and MST1/LATS1/YAP1/PPP2R1A in HCT116 and SW480 cells by coimmunoprecipitation. (B) Detection of the interaction between PPP2R1A and PEBP1/YAP1 in HCT116 and SW480 cells by coimmunoprecipitation. (C) Detection of the interaction between YAP1 and PEBP1/PPP2R1A in HCT116 and SW480 cells by coimmunoprecipitation. (D) The colocalization of LATS1/YAP1/PPP2R1A and GFP-PEBP1 in HCT116 cells was detected by immunofluorescence. Scale bar: 20 μm. (E) Representative images of the colocalization of YAP1/PPP2R1A and GFP-PEBP1 in HT29 cells. Scale bar: 20 μm. (F) The validation of the interactions between PEBP1 and PPP2R1A in vitro by GST pull-down assay. (G) The validation of the interactions between YAP1 and PPP2R1A in vitro by GST pull-down assay. (H) The validation of the interactions between PEBP1 and YAP1 in vitro by GST pull-down assay. (I) The validation of the interactions between PEBP1, PPP2R1A, and YAP1 in vitro by GST pull-down assay. (J) Detection of the interaction between PPP2R1A and PEBP1/YAP1 in PCYT2-knockdown SW620 cells by coimmunoprecipitation. (K) Detection of the interaction between PPP2R1A and YAP1/PEBP1 after PCYT2 overexpression in HCT116 cells overexpressing PCYT2 by coimmunoprecipitation.

PPP2R1A facilitates nuclear distribution of YAP1 and cell migration by reducing YAP1 phosphorylation. Since PPP2R1A is functionally annotated as a subunit of phosphatase PP2A, we hypothesize that PPP2R1A could change YAP1 phosphorylation levels in CRC cells. To investigate this hypothesis, we detected YAP1 phosphorylation levels in CRC cells ectopically expressing PPP2R1A. PPP2R1A overexpression remarkably reduced the p-YAP1 in the cytoplasm, while it increased the accumulation of YAP1 in the nucleus (Figure 7A). Conversely, PPP2R1A knockdown greatly increased YAP1 phosphorylation, whereas total YAP1 expression was decreased (Figure 7B). The immunofluorescence assays also showed an enhanced fluorescence signal of YAP1 in the nucleus after the overexpression of PPP2R1A, whereas PPP2R1A knockdown weakened the signal (Figure 7, C and D). Additionally, we found that knocking down PPP2R1A inhibited CRC cell migration, while PPP2R1A overexpression had the opposite effect (Figure 7, E–I). These results indicate that PPP2R1A probably dephosphorylates YAP1 to promote YAP1 nuclear distribution and cell migration.

PPP2R1A dephosphorylates YAP1 and leads to changes in the proportion of YAP1 entering the nucleus. (A) Detection of the YAP1/p-YAP1 expression levels in the nucleoprotein and in cytoplasmic with ectopic expression of vehicle and 3MYC-PPP2R1A HCT116 and HT29 cells by nucleoplasmic separation and Western blot. (B) Detection of the YAP1/p-YAP1 expression in the nucleoprotein and cytoplasmic at SW480 cells infected with control and shPPP2R1A by nucleoplasmic separation and Western blot. (C and D) Detection of the changes in the nuclear proportion of YAP1 in CRC cells with PPP2R1A overexpression (C) or knockdown (D) by immunofluorescence staining. Different colors represent different antibodies or dyes: DAPI (blue), YAP1 (green), and phalloidine (red). Scale bar: 20 μm. (E and F) Transwell assays of HCT116 (E) and SW480 (F) cells demonstrate that cells with higher levels of PPP2R1A had a higher migratory ability than cells with lower PPP2R1A levels. Scale bar: 100 μm. Representative images and statistical chart. (G) Detection of the epithelial and mesenchymal markers with ectopic expression of vehicle and 3MYC-PPP2R1A HCT116 and HT29 cells by Western blot. (H) Representative images of scratch assays in PPP2R1A-overexpression HCT116 cells and control group. The graph shows the area of a wound evaluated with ImageJ. Scale bar: 500 μm. Representative images and quantification. (I) Immunofluorescence assays were performed to determine the membrane localization of paxillin in HCT116 cells with PPP2R1A overexpression. Scale bar: 20 μm. Representative images and statistical chart. Data represent the mean ± SD (n = 3). Statistical significance was assessed with 2-tailed unpaired Student’s t test (E, H, and I) or 1-way ANOVA with Tukey’s multiple-comparison test (F). *P < 0. 05, **P < 0.01, ****P < 0.0001.