Introduction
Colorectal cancer (CRC) is one of the leading cancer types resulting in new cancer cases and deaths worldwide [
1]. Increasing evidence, including ours, shows that dysregulation of Hippo/YAP signaling contributes to tumorigenesis, including CRC [
2,
3]. YAP drives target gene expression by forming complexes with multiple transcription factors, which are required to drive tumor initiation and progression [
4]. Phosphorylation of YAP, a major downstream transducer of the Hippo pathway [
5], is a key event in YAP signaling. Both cytoplasmic and nuclear localization of YAP could be regulated by YAP phosphorylation at different sites by different kinases [
6,
7]. In addition to these, lysine methylation is another important post-translational modification involved in YAP activation and location [
8]. However, the new factors and the concise mechanisms for regulation of subcellular localization and activation of YAP are still poorly known.
Long non-coding RNAs (lncRNAs) are transcripts longer than 200 nucleotides that have no or limited protein-coding capacity. A large body of evidence has demonstrated that lncRNAs are engaged in the signaling pathways of CRC. LncRNAs are important versatile molecules involved in a variety of tumorigenic processes and diseases via interactions with DNA, RNA, or proteins. LncRNAs execute molecular functions as archetypes of decoys, signals, guides, and scaffolds [
9]. For instance, XIST is one of first functionally annotated lncRNAs that plays a critical role in X inactivation by recruiting multiple factors [
10]. HOTAIR is a lncRNA of the HOXC locus, which forms RNA-DNA-DNA triplexes with predicted target sites in mesenchymal stem cells [
11]. LncRNA nuclear enriched abundant transcript 1 (NEAT1) has a profound effect on cross-regulation between paraspeckles and mitochondria by altering the sequestration of mito-mRNAs in paraspeckles [
12]. LncRNA growth arrest-specific 5 (GAS5) induces apoptosis by binding to the domain of the glucocorticoid receptor [
13]. Recently, we found that lncRNA uc.134 inhibits YAP downstream target genes by inhibiting CUL4A to ubiquitinate LATS1 and increasing pYAP
S127 expression [
14]. Nevertheless, whether lncRNAs can regulate YAP activation by direct interaction or post-translationally modifying YAP protein remains to be elucidated.
The N6-methyladenosine (m
6A) RNA modification, as the most abundant internal epi-transcriptomic modification in eukaryotic messenger RNAs (mRNAs), is introduced by the m
6A methyltransferase complex, which have been designated as a “writer,” and can be deleted by m
6A demethylases, such as fat mass and obesity-associated protein (FTO) and ALKBH5. Factors interpreting specific modifications have been identified as “readers,” such as YTHDF1/2/3 and YTHDC1/2 [
15]. FTO is the first m
6A demethylase that is highly expressed in acute myeloid leukemia (AML), and it plays a critical oncogenic role [
16]. YTHDF3 facilitates translation of protein synthesis in synergy with YTHDF1 and affects decay of methylated mRNA mediated through YTHDF2 [
17]. Although the m
6A is reported to be important in cancer progression, whether lncRNAs regulate the m
6A modification and the role of m
6A in lncRNA transcripts in cancer progression remain unknown. Here, we aimed to investigate the functional links between lncRNAs and the m
6A modification in YAP signaling in CRC.
Materials and methods
Cell lines, cell culture, and transfection
DLD1, LOVO, SW480, SW620, LS174T, HCT116, RKO, and HT29 cell lines were obtained from the Cell Bank of Type Culture Collection (Guangzhou Cellcook Biotech Co., Ltd., Guangzhou, China). The cell line authentication report showed the cell lines to be considered as identical to the reference cell line in the ATCC STR database. The cells were cultured in incubators containing 5% CO
2 at 37 °C and were maintained in RPMI 1640 supplemented with 10% FBS. The transfection of plasmids or small interfering RNAs (siRNAs) was performed using jetPRIME (Polyplus, Strasbourg, France), according to the manufacturer’s instructions. As we previously described, 24 h after transfection, cell lysates were subjected to western blot, and the western blot data were quantified using the ImageJ software [
14]. Detailed descriptions of antibodies, oligonucleotide sequences and primers can be found in the Additional file
2.
RIP sequencing
An RIP experiment was performed according to the instructions of the Magna RIP RNA Binding Protein Immunoprecipitation Kit (Millipore, MA, USA). Briefly, lysate was prepared in a lysis buffer containing protease inhibitor cocktail and RNase inhibitor. Then, protein A/G magnetic beads were prepared for incubation with 5 μg of purified antibodies per immunoprecipitation with rotation for 30 min at room temperature. Further, to precipitate RNA-binding protein-RNA complexes, and the mixture was incubated with rotation for 3 h overnight at 4 °C. Finally, RNA was purified using proteinase K buffer and an Agilent 2100 Bioanalyzer (Agilent, CA, USA). A NanoDrop 2000 (Thermo Fisher, MA, USA) was used to analyze the total RNA quality and quantity. Further RNA purification in an immunoprecipitation may be pursued by deep sequencing or quantitative reverse transcription polymerase chain reaction (qRT-PCR). The cDNA libraries were sequenced on the Illumina sequencing platform by Genedenovo Biotechnology Co., Ltd. (Guangzhou, China).
MeRIP sequencing
Total RNA was extracted using Trizol reagent (Takara, Dalian, China). An Agilent 2100 Bioanalyzer (Agilent, CA, USA) and NanoDrop 2000 (Thermo Fisher, MA, USA) were used to analyze the total RNA quality and quantity. More than 50 μg of total RNA is sufficient following RNA fragmentation and immunoprecipitation according to the instructions of the Magna MeRIP™ m6A Kit (Merck, Darmstadt, Germany). Briefly, the total cell RNA is fragmented into ~ 100-nt-long oligonucleotides using fragmentation buffer under elevated temperature. Then the post-fragmentation size distribution is validated by an Agilent 2100 Bioanalyzer with an Agilent RNA 6000 Kit. The Magna ChIP Protein A/G Magnetic Beads were incubatated for 30 min at room temperature with m6A-specific antibody in immunoprecipitation buffer. The mixture was then incubated with the MeRIP reaction mixture for 2 h at 4 °C. Then eluted RNA and MeRIPed RNA were analyzed by deep sequencing on an Illumina Novaseq™ 6000 platform at the LC-BIO Bio-tech ltd (Hangzhou, China) following the vendor’s recommended protocol.
In vivo model
BALB/c male mice 6–8 weeks old were purchased from Guangdong Medical Laboratory Animal Center, China. Mice were raised under pathogen-free conditions. All in vivo experiments were done according to approved protocols from the Institutional Animal Care and Use Committees, according to national and institutional guidelines. All procedures were performed essentially as previously described. Briefly, for the subcutaneously injected tumor model, 2 × 106 viable cells were subcutaneously injected into the flanks of mice. Tumor volume was assessed as (L × W2/2), where L and W represent the length and the width of the tumor, respectively. After 4 weeks, the tumors were embedded in paraffin and stained for in situ hybridization (ISH) or immunohistochemistry (IHC). For the lung metastasis model, 2 × 106 viable cells were injected into the tail veins of mice. The mice were monitored for 6 weeks for lung metastasis by hematoxylin-eosin (H&E) staining. The metastatic foci were calculated using Dmetrix software by combining the number and area of lung metastatic nodules in individual mice.
Tissue samples, immunohistochemistry (IHC), and in situ hybridization (ISH) staining
Formalin-fixed paraffin-embedded (FFPE) colon cancer tissues and adjacent noncancerous tissues were retrieved from the Department of Pathology at Sun Yat-sen Memorial Hospital, Sun Yat-sen University (Guangzhou, China). The study was approved by the Human Research Ethics Committees of Sun Yat-sen University. The study is compliant with all relevant ethical regulations for human research participants, and informed consent was obtained from all subjects.
All procedures were performed essentially as previously described [
14]. Briefly, for H&E staining, sections of tissue were deparaffinized in xylene and then stained with hematoxylin and eosin according to standard histological procedures. For IHC staining, after the sections were deparaffinized and re-hydrated, the specimens were incubated in EDTA buffer (1 mM, PH 8.0) for antigen retrieval using a high-pressure method. Then, tissue sections were incubated overnight at 4 °C with primary antibodies, including anti-YAP, anti-YTHDF3, and anti-Ki 67. 3,3′-diaminobenzidine (DAB) solution (ZSGB-BIO, Beijing, China) was used to detect target proteins, which were conjugated with a peroxidase enzyme to form a brown precipitate. For ISH staining, lncRNA GAS5 expression was measured in paraffin-embedded samples according to the instructions of the ISH Kit™ (BOSTER, Wuhan, China). Briefly, after the sections were deparaffinized and re-hydrated, the specimens were incubated with proteinase for 10 min. at 37 °C. After washing twice in PBS, the hybridization mix was applied, and hybridized samples were incubated overnight at 40 °C. Then, the sections were incubated with blocking solution for 30 min and anti-DIG reagent was applied for 60 min. Then, the sections were incubated with AP substrate 4-nitro-blue tetrazolium and 5-bromo-4-chloro-3′-indolylphosphate (NBT-BCIP) for 30 min at 37 °C. The sections were mounted with Nuclear Fast Red. A blue stain in the samples indicated a positive signal by NBT-BCIP. For DAB-stained samples, a brown precipitate showed a positive signal, and the slides were then counterstained with hematoxylin.
The staining scores were evaluated by two individuals in a blinded fashion. A quick scoring system from 0 to 12 that combined the intensity and percentage of the positive signal was used as described previously [
14,
18]. Briefly, a signal of 0 indicated no staining, 1 indicated weak staining, 2 indicated intermediate staining and 3 indicated strong staining. Percentage scores were assigned as follows: 0 corresponded to 0%, 1 to 1–25%, 2 to 26–50%, 3 to 51–75%, and 4 to > 75%. The median value of total staining scores was identified as the optimal cut-off value. If the evaluated score was lower than the median, the indicator expression of in those CRC samples was classified as low; otherwise, it was classified as high.
RNA-pulldown assay
Biotin-labeled RNAs were transcribed in vitro with the Biotin RNA Labeling Mix and T7 RNA polymerase (Roche, Basel, Switzerland). Biotinylated RNAs were mixed with streptavidin agarose beads (Life Technologies, Gaithersburg, MD) at 4 °C overnight. Total cell lysates were freshly prepared and added to each binding reaction with Protease/Phosphatase Inhibitor Cocktail and RNase inhibitor, and then the mixture was incubated with rotation for 1 h at 4 °C. After washing thoroughly three times, the RNA–protein binding mixture was boiled in SDS buffer and the eluted proteins were detected by western blot or mass spectrometry.
Bio-layer interferometry (BLI) analysis
A bio-layer interferometry (BLI) experiment was carried out using the Octet system (ForteBio, Fremont, CA). Streptavidin sensors were used for immobilization of biotin-labeled lncRNA GAS5. A five-point concentration series was assayed for purified YAP protein (10-nM to 1-μM range). Wells with assay buffer only were used as reference wells. Reference biosensors with no immobilized ligand was used to avoid nonspecific binding to GAS5 during the binding event. The procedure of Octet System Data Acquisition software was followed to analyze the kinetics. Dissociation (Kd) and association rate constants (Ka) were determined with the Octet Data Analysis Software, as a result of a global fit considering the entire step times, and assuming a 1:1 binding model.
RNA FISH and immunofluorescence co-staining
The locked nucleic acid-modified oligonucleotide probe targeting GAS5 (Exiqon, Vedbaek, Denmark) was used for RNA fluorescent in situ hybridization (FISH). We detected YAP protein in situ in CRC cells with immunofluorescence assays. The anti-YAP1 antibody (Alexa Fluor 647) (abcam, MA, USA) was used to detective the YAP protein with a confocal microscope (Leica, Wetzlar, Germany) (shown in red), and DAPI was used for labelling nuclear DNA (shown in blue). The RNA signal was detected by incubation with biotinylated conjugated anti-DIG antibodies, and the signals were amplified using SABC – FITC (shown in green).
Chromatin immunoprecipitation (ChIP) and luciferase reporter assay
The chromatin immunoprecipitation (ChIP) procedure was performed using SimpleChIP® Enzymatic Chromatin IP Kit (Cell Signaling Technology, MA, USA) following the manufacturer’s instructions. Briefly, cells are fixed with formaldehyde to cross-link histone and non-histone proteins to DNA. Then, chromatin is digested with micrococcal nuclease into 150- to 900-bp DNA/protein fragments. Antibodies specific to YAP proteins are added and the complex co-precipitates are captured by Protein G magnetic beads. Finally, cross-links are reversed, and the level of enrichment of the target DNA sequence is purified, at which point it is ready for PCR. One tenth of the input chromatin was also treated in the same way and purified. The enriched DNA fragments were presented as a percentage of input chromatin. Luciferase reporter assay was measured using the Dual-Glo Luciferase Assay System (Promega, WI, USA) following the manufacturer’s instructions. Briefly, YAP plasmids were co-transfected with YTHDF3 promoter-luciferase vector and pRL-TK Vector. The pGL3-basic vector was transfected as a negative control. After 24 h, prepared cell extracts were used to measure the luciferase activity on a Spark multimode microplate reader (TECAN, Mannedorf, Switzerland).
Quantification and statistical analysis
All statistical analysis was carried out using GraphPad Prism version 7 (GraphPad Software, CA) for Windows to assess the differences between experimental groups. The data were analyzed by analysis of variance tests or Student’s t-tests (*P < 0.05, **P < 0.01, ***P < 0.001). A multi-way classification analysis of variance tests was performed to assess data obtained from the CCK8 assays and tumor growth. Survival curves were plotted based on the Kaplan–Meier curves and log-rank tests. Correlations among GAS5 expression, YAP, and YTHDF3 were analyzed with a Spearman rank correlation. P < 0.05 was considered to indicate a significant difference. Each experiment was repeated independently with similar results at least three times.
Data availability
The RIP-sequencing, lncRNA-sequencing, and MeRIP-sequencing data discussed in this paper have been deposited in NCBI’s Gene Expression Omnibus [
19] and are accessible through GEO Series accession numbers GSE129535, GSE129624 and GSE129716. The data will become public when this article is published online (Additional files
4,
5 and
6).
Discussion
YAP sequestered in the nucleus is essential for YAP-mediated target gene transcription and tumor progression. TEAD family transcription factors are major nuclear partners that effect transcriptional activation [
21]. Serine/threonine-protein kinase LATS1, a core component of the Hippo pathway, is the key regulator of YAP phosphorylation and facilitates its cytoplasmic localization [
22]. Therefore, targeting the Hippo-YAP pathway may provide new approaches for cancer therapy [
4,
23,
24]. In addition to these, c-Abl, a 140-kDa proto-oncoprotein, directly phosphorylates YAP at position Y357 to stabilize YAP in a c-Abl kinase-dependent manner [
25]. The β-catenin-YAP1-TBX5 transcriptional complex is reported to be essential for tumor survival and tumorigenesis [
26]. YAP is O-GlcNAcylated by O-GlcNAc transferase at serine 109 in a LATS1-dependent manner [
27]. The ARID1A-containing SWI/SNF complex inhibits YAP transcription activity by blocking the association between YAP and TEAD [
28]. Non-coding RNAs, such as microRNAs and circular RNAs also have been reported to play vital roles in targeting YAP. For example, miR-15a and miR-16-1 directly bind with YAP1 3’UTR to regulate YAP1 expression [
29]. CircFAT1 is reported to abundantly sponge miR-375 to up-regulate YAP1 expression [
30]. To date, lncRNAs have been found to interact with the core components of the Hippo-YAP pathway at different levels. For example, a ROR1-HER3-LncRNA MAYA signaling axis was reported to modulate the Hippo-YAP pathway by methylating Hippo/MST1 at Lys59 [
31]. A previous study showed that lncARSR binds with YAP to impede LATS1-induced YAP phosphorylation and facilitates YAP nuclear translocation in propagation of renal tumor-initiating cells [
32]. Nevertheless, the key non-coding RNAs involved in YAP signaling, especially those that directly interact with YAP, remain largely unclear. Therefore, their functions in cancer progression, including CRC, are also not well-characterized.
Here, we identified several candidates for YAP-interacting lncRNAs that may be key regulators for YAP signaling. We established a systematic strategy to screen YAP-interacting lncRNA by RIP-seq, RNA pull-down, BLI analysis, RNA FISH, and immunofluorescence co-staining assays. At present, there is no systematical study to identify YAP-interacting lncRNAs. Using the interaction between lncRNA GAS5 and YAP as a model, we also developed a new method to assess binding affinity of lncRNAs and protein using BLI analysis. Although RNA pull-down is one of the most common methods for detecting the interaction between lncRNAs and proteins, it cannot evaluate the binding affinity of RNA and proteins quantitatively. BLI analysis has been widely used in compound screening, protein interaction, and virus titer analysis [
33]. To the best of our knowledge, ours is the first study to use BLI analysis to investigate the binding affinity of lncRNAs and protein complexes. Our study indicated that BLI analysis may offer a promising approach for examining the binding affinity of the lncRNA-protein complex. Our method attempts to preserve the RNA hairpin structure as much as possible to delete GAS5 mutants, while enabling the analysis of interactions between GAS5 and YAP. In particular, we identified nucleotides 262–480 of GAS5, which directly bonded with the WW domain of YAP to facilitate its cytoplasmic retention. Despite numerous studies, the RNA structural domains and the functional role in RNA and protein interaction remain largely unknown. Structural analysis of the lncRNA-protein complex will be of great importance for illuminating the concise mechanism for YAP-GAS5 interactions.
Moreover, our study also revealed that GAS5 directly binds to YAP to trigger YAP phosphorylation at Ser127 in CRC cells, which helps YAP cytoplasmic localization and facilitates its ubiquitin-mediated degradation. Functional analyses and a mouse xenograft tumor model showed that overexpression of GAS5 significantly suppressed the proliferative and metastasis capacity of CRC cells, whereas the exogenous YAP expression could successfully reverse GAS5-mediated inhibition of CRC tumor progression in vitro and in vivo. A previous study showed that lncARSR binds with YAP to impede LATS1-induced YAP phosphorylation and facilitates YAP nuclear translocation in propagation of renal tumor-initiating cells [
32]. Whether lncRNAs can contribute to YAP protein stability remains largely unclear. Here, our study demonstrated that GAS5 works as an RNA scaffold to promote degradation of YAP in CRC progression. Our data uncovered a key mechanism by which GAS5 inhibited CRC proliferation and metastasis through directly binding with YAP and facilitates its phosphorylation and subsequently ubiquitin-mediated degradation. In the absence of GAS5, YAP translocated in the nucleus and activated its target gene transcription, which in the context of CRC would promote a more malignant phenotype. It will be of great interest to identify common lncRNAs that regulate YAP protein stability.
More importantly, we identified YTHDF3 as a novel target of YAP, which plays a key role in CRC progression. YAP signaling is essential for cancer progression; therefore, the findings of the novel and key targets will be of great importance for understanding the role and the mechanisms of YAP pathway in cancer. In our study, YTHDF3 is significantly elevated in CRC tumor tissues compared with that in counterpart normal tissues. Functionally, gain-of-function and loss-of-function experiments showed that YAP significantly promoted the proliferation, invasion, and metastasis of CRC in vitro and in vivo. In contrast, YTHDF3 knockdown reversed YAP-mediated promotion of CRC tumor progression, and co-transfection of YTHDF3 and GAS5 also obtained similar results. Therefore, our data indicated that YTHDF3, as a novel target of YAP, plays a key role in CRC progression in vitro and in vivo, which may provide new insights into CRC therapy.
Ultimately, we established a key role of YTHDF3 in m
6A-modified GAS5 and degradation of GAS5 transcription, uncovering a negative feedback loop between YAP-interacting lncRNA GAS5 and the m
6A reader YTHDF3 in Hippo/YAP signaling and tumor progression of CRC. We found that YTHDF3 knockdown significantly prolonged the decay rate of GAS5 as a result of accumulation of m
6A modifications in GAS5 and therefore decreased YAP protein expression. Numerous studies indicated that expression of GAS5 was repressed in many types of malignant tumors [
34,
35]. Nevertheless, very little is known about its mechanism for downregulation or degradation. It has been suggested that GAS5 transcription may be degraded by nonsense surveillance, also known as the nonsense-mediated RNA decay pathway [
36]. Interestingly, a recent study showed that lncRNA XIST is highly methylated by RBM15 and METTL3, which is required for XIST-mediated transcriptional repression [
37]. This study suggested that METTL3-mediated m
6A modification is important for lncRNA expression. Another study also indicated that the internal m
6A modification of linc1281 is required for mESC differentiation [
38]. In our study, we clarified one mechanism by which GAS5 decayed through m
6A modification, suggesting a novel lncRNA regulatory mechanism. Extensive studies should be done to analyze m
6A modification leading to lncRNA destabilization and shed light on future studies of YTHDF3.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.