METTL3 stabilization by PIN1 promotes breast tumorigenesis via enhanced m6A-dependent translation

Methyltransferase-like 3 (METTL3) is the catalytic subunit of the N6-adenosine methyltransferase complex responsible for N6-methyladenosine (m6A) modification of mRNA in mammalian cells. Although METTL3 expression is increased in several cancers, the regulatory mechanisms are unclear. We explored the regulatory roles of peptidyl-prolyl cis-trans isomerase NIMA-interacting 1 (PIN1) in METTL3 stability and m6A modification of mRNA. PIN1 interacted with METTL3 and prevented its ubiquitin-dependent proteasomal and lysosomal degradation. It stabilized METTL3, which increased the m6A modification of transcriptional coactivator with PDZ-binding motif (TAZ) and epidermal growth factor receptor (EGFR) mRNA, resulting in their efficient translation. PIN1 knockout altered the distribution of TAZ and EGFR mRNA from polysomes into monosomes. Inhibition of MEK1/2 kinases and PIN1 destabilized METTL3, which impeded breast cancer cell proliferation and induced cell cycle arrest at the G0/G1 phases. METTL3 knockout reduced PIN1 overexpression-induced colony formation in MCF7 cells and enhanced tumor growth in 4T1 cells in an orthotopic mouse model. In clinical settings, METTL3 expression significantly increased with tumor progression and was positively correlated with PIN1 expression in breast cancer tissues. Thus, PIN1 plays a regulatory role in mRNA translation, and the PIN1/METTL3 axis may be an alternative therapeutic target in breast cancer.


INTRODUCTION
The cis-trans conformational change of proline induced by peptidyl-prolyl cis-trans isomerase (PPIase) regulates protein conformations in cells [1,2]. The human genome encodes approximately 30 PPIases belonging to three families, i.e., cyclophilin, FK506-binding protein (FKBP), and parvulin [3]. PPIases act as molecular chaperones and play a crucial role in the assembly and disintegration of ribonucleoprotein complexes [4]. The proper packaging of mRNAs into ribonucleoprotein complexes is essential for various post-transcriptional events, including epitranscriptomic modifications, alternative splicing, nuclear export, mRNA degradation, and translation [5]. Several PPIases were identified as intrinsic components of the spliceosome and mRNA degradation machinery [4]; however, peptidyl-prolyl cis-trans isomerase NIMAinteracting 1 (PIN1) is the only member that binds to-and isomerizes-specific phosphorylated serine/threonine-proline (pSer/Thr-Pro) motifs [3]. PIN1 expression is increased in several cancer types, including breast cancer [2,6]. PIN1 promotes breast tumorigenesis by inducing conformational changes in several transcription factors and epigenetic modifiers, including p53, c-Myc, nuclear factor kappa B, β-catenin, hypoxia-inducible factor 1α, signal transducer and activator of transcription 3, and SUV39H1 to regulate their function [2]. High-throughput transcriptome sequencing revealed a strong correlation between the expression of PIN1 and genes involved in mRNA splicing, translation, and ribosome biogenesis [7]. This has triggered interest in the study of PIN1 as a regulator of RNA processing events.
The regulatory role of PPIase in RNA processing is supported by several lines of evidence. For instance, parvulin 14 facilitates the splicing of pre-rRNA to generate 18 S and 28 S rRNA [8]. Furthermore, PIN1 regulates the stability of mRNA containing AU-rich elements (ARE) through prolyl isomerization of AREbinding proteins [9]. Similarly, PIN1 regulates the stability of parathyroid hormone mRNA in a rat model of secondary hyperparathyroidism [10]. The interest in studying the role of PPIase in mRNA translation is based on an early discovery wherein the immunosuppressive drug rapamycin bound to prolyl isomerase FKBP12 and suppressed protein synthesis via inhibition of the mechanistic target of rapamycin complex [11]. Further studies showed that PIN1 directly binds to-and promotes-ribosomal S6 kinase phosphorylation, which may facilitates protein synthesis in mouse embryonic fibroblast cells as well as in different cancer cell lines [12]. In addition, PIN1 facilities the ubiquitination of cytoplasmic polyadenylation element-binding proteins (CPEB) to restore the translation of cyclin B1 mRNA in Xenopus oocyte [13]. Therefore, PPIases including PIN1 act as crucial regulators of RNA splicing, mRNA stability, and translation to regulate physiological processes. However, most studies were performed in noncancerous models and the role of PIN1 in RNA processing during carcinogenesis remains unclear.
Recently, epitranscriptomic modifications of mRNA such as N 6methyladenosine (m 6 A) have emerged as critical factors in regulating post-transcriptional processing of mRNAs in human cancers, including breast cancer [14][15][16]. The deposition of the m 6 A modification is mediated by a nuclear methyltransferase complex comprising three core proteins, methyltransferase-like protein 3 (METTL3), METTL14, and Wilms tumor 1-associated protein (WTAP) [17]. In this complex, METTL3 serves as the catalytic subunit. METTL3-deposited m 6 A modifications regulate mRNA stability and translation efficiency to promote carcinogenesis [18]. For instance, enhanced m 6 A modification of transcriptional coactivator with PDZ-binding motif (TAZ) and epidermal growth factor receptor (EGFR) mRNA enhances their translation efficiency to promote lung tumorigenesis [19]. Interestingly, METTL3 promotes the translation of TAZ and EGFR via enhanced polysome formation, a function which is independent of its methyltransferase activity [19]. Notably, TAZ and EGFR expression is frequently upregulated in breast cancer patients and is associated with poor prognoses [20,21]. However, it is unclear whether PIN1 regulates the abundance of specific oncoproteins such as TAZ and EGFR via regulation of m 6 A-related mRNA processing. This requires examining protein-protein interactions between PIN1 and proteins involved in m 6 A modifications. In this study, we identified the novel METTL3-PIN1 interaction that has important implications in the regulation of m 6 A-dependent mRNA translation. Our results indicated that the PIN1/METTL3 axis may be a potential therapeutic target for breast cancer.

RESULTS
Increased PIN1 expression is associated with METTL3 overexpression in breast cancer As PIN1 expression is significantly increased in breast cancer [2], we examined the expression of several genes correlated with PIN1 expression in The Cancer Genome Atlas (TCGA) Pan-Cancer Atlas. Functional classification of the most significant hits (p > 0.5, q-value <0.1) was performed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) (https:// david.ncifcrf.gov/, version 6.8). Genes involved in RNA processing, splicing, translation, transcription, apoptosis, and protein transport significantly correlated with PIN1 expression in breast cancer patients (Fig. 1a, b, Supplementary Table 1). Epitranscriptomic modification via RNA methylation has emerged as a crucial regulator of RNA processing [17]. METTL3 protein expression most significantly correlated with PIN1 expression out of all the other RNA methyltransferases examined in the TCGA Pan-Cancer Atlas (Fig. 1c, d and Supplementary Table 2). In addition, a negative CERES dependency score indicated that both METTL3 and PIN1 are essential for breast cancer cell survival and showed that these proteins function as oncogenes (Fig. 1e and Supplementary Table  3). PIN1 and METTL3 expression was significantly increased in breast tumors compared to that in paracancerous tissue following immunohistochemistry (IHC) (Fig. 1f-h). Furthermore, METTL3 expression increased with tumor progression (Fig. 1i). The linear trend between IHC scores of PIN1 and METTL3 positively correlated with METTL3 and PIN1 expression in the breast cancer tissue (Fig. 1j). Together, these data indicated that increased PIN1 expression was associated with METTL3 overexpression in breast cancer.
PIN1 interacted with METTL3 in a phosphorylation-dependent manner Potential interactions between PIN1 and m 6 A writers and erasers were screened using mammalian two-hybrid (M2H) analysis. The relative luciferase activity resulting from hybridization significantly increased between PIN1 and m 6 A writer proteins, but not with the m 6 A demethylase fat mass-and obesity-associated protein; METTL3 exhibited the strongest hybridization (Fig. 2a). M 6 A writer proteins are often present in the form of a methyltransferase complex; therefore, in vitro nickel-nitrilotriacetic acid (Ni-NTA) pulldown assays were performed to identify the direct binding partners of PIN1. METTL3, but not METTL14 or WTAP physically interacted with PIN1 (Fig. 2b). This was confirmed using an in vitro binding assay (Fig. 2c). A glutathione S-transferase (GST) pulldown assay using GST-PIN1 revealed that PPIase (peptidyl-prolyl cis-trans isomerase) domain of PIN1 interacted with METTL3, while the WW domain did not (Fig. 2d). PIN1 catalyzes the isomerization of a proline residue and is dependent on a phosphorylated serine/ threonine-proline (S/T-P) motif of the substrate protein [3]. Dephosphorylation of METTL3 by alkaline phosphatase abolished binding of METTL3 to PIN1 (Fig. 2e), showing that a phosphorylation-dependent interaction was required. Furthermore, FBS stimulation increased the METTL3-PIN1 interaction in a time-dependent manner indicating that it is dependent on FBSinduced METTL3 phosphorylation (Fig. 2f). METTL3 phosphorylation sites critical for PIN1 binding were identified by replacing the serine (S) or threonine (T) residues that were followed by a proline (P) with alanine (A) (denoted as S43A, S50A, T106A, T348A, and S525A). A Ni-NTA pulldown assay with His-PIN1 revealed that the S525A mutation abolished the METTL3-PIN1 interaction, indicating that the SP site located at S525 in METTL3 is essential for the interaction with PIN1 (Fig. 2g). Furthermore, the SP site at S525 is evolutionarily conserved across different species (Fig. 2h). circRNAbased reporter assays showed that mitogen-activated protein (MAP) kinase/extracellular signal-regulated kinase (ERK) kinase 1/2 (MEK1/2), c-Jun N-terminal kinase 2 (JNK2), and MAP3K8 regulate METTL3 function [22]. Treatment with PD98059, an inhibitor of MEK1/2 significantly reduced the METTL3-PIN1 interaction, suggesting that this interaction is modulated by the activity of the MEK/ERK pathway (Fig. 2i). Furthermore, FBS stimulation increased the interaction between endogenous PIN1 and METTL3 (Fig. 2j), a finding that supported the in vitro binding results. Finally, bimolecular fluorescence complementation (BiFC) analysis revealed increased complementation in MCF7 cells transfected with pBiFC-METTL3 wild-type (WT) through a significant increase in the counts of mVenus-positive cells; however, this was not observed for mutant METTL3-S525A and pBiFC-VC-PIN1 after FBS stimulation (Fig. 2k, l). The complementation was predominantly localized in the nucleus (Fig. 2k). Together, these results suggested that METTL3 phosphorylation at S525 induced by MEK/ERK kinases promoted the interaction of METTL3 with the PPIase domain of PIN1 in the nucleus.
PIN1 promoted METTL3 protein stability PIN1 overexpression increased METTL3 expression, whereas PIN1 knockout using sgPIN1 reduced METTL3 expression in MCF7 cells ( Fig. 3a and Supplementary Fig. S1a). In addition, METTL3 expression decreased in PIN1 -/mouse embryonic fibroblasts with METTL3 via its PPIase domain in a phosphorylation-dependent manner. a Screening of protein-protein interactions using the mammalian two-hybrid assay. pACT-PIN1 and pBIND plasmids were co-transfected with pG5luc plasmid into HEK293 cells. Firefly luciferase activity in the lysate was measured and normalized against Renilla activity. pBIND-SGK1 was included as a known binding partner for PIN1. Values are presented as the mean ± SD using one-way ANOVA (N = 3), **p < 0.01, ***p < 0.001. b, c PIN1 interacted with METTL3. His/Xpress-PIN1 expressed in HEK293 cells was pulled down using Ni-NTA resin and incubated with lysates from HEK293 cells expressing FLAG-METTL-3, -14, or WTAP, and the precipitated proteins were detected by immunoblotting (IB) using the indicated antibodies. b Immunoprecipitation (IP) of HEK293 cells overexpressing FLAG-METTL3 and His/Xpress-PIN1 using FLAG (top) or Xpress (bottom) antibodies, followed by IB with the indicated antibodies. Normal mouse IgG was used as an isotype control (c). d The PPIase domain of PIN1 bound to METTL3. Lysates from HEK293 cells expressing FLAG-METTL3 were incubated with GST, GST-PIN1-WT, -WW, or -PPIase. Bound proteins were detected by IB using anti-FLAG antibody. GST fusion proteins were stained with Coomassie Blue. e, f PIN1 interacted with METTL3 in a phosphorylation-dependent manner. Lysates from HEK293 cells expressing FLAG-METTL3 were treated with 0.5 or 1 unit of CIAP per microgram of protein at 37°C for 1 h and used for the GST pulldown assay. Proteins bound to GST-PIN1 were detected by IB using anti-FLAG antibody. GST fusion proteins were stained with Coomassie Blue (e). MCF7 cells were treated with 10% FBS for the indicated times, and the lysate (either GST or GST-PIN1) was used for the GST pulldown assay. Proteins in the pulldown fraction and cell lysate were detected by IB using the indicated antibodies. GST fusion proteins were stained with Coomassie Blue (f). g Schematic representation of the Ser(S)/Thr (T)-Pro (P) motif in METTL3 (upper panel). Lysates from HEK293 cells expressing FLAG-METTL3-WT, -S43A, -S50A, -T106A, -T348A, -S525A, and His/ Xpress-PIN1 were incubated with Ni-NTA resin followed by IB with anti-FLAG antibody. Cell lysates were analyzed by IB with the indicated antibodies. h Sequence alignment of the S/T-P motif in METTL3. i Inhibition of MEK1/2, but not JNK1/2 and TPL2, suppressed METTL3-PIN1 binding. MCF7 cells were pretreated with PD98059 (MEK1/2 inhibitor, 10 µM), SP600125 (JNK inhibitor, 20 µM), or TKI (TPL2 inhibitor, 20 µM) in serum-free media for 24 h and exposed to 10% FBS for 30 min. Cell lysates were used for Ni-NTA pulldown assays and the proteins were detected using anti-FLAG antibody. Proteins in cell lysates were detected using the indicated antibodies. j Binding between endogenous METTL3 and PIN1 induced by FBS. IP of cell lysate from MCF7 cells serum-starved for 24 h, then exposed to 10% FBS using PIN1 antibody, followed by IB with the indicated antibodies. k, l Visualization of the METTL3-PIN1 interaction with BiFC. MCF7 cells transfected with the indicated BiFC plasmids were serum-starved for 24 h, followed by treatment with 10% FBS. Fluorescence examination of mVenus (green) and DAPI to detect nuclear DNA (blue) (k). The population of mVenus-positive cells was counted using Fiji software (http://fiji.ac) (l). Values are presented as the mean ± SD using one-way ANOVA (N = 3), *p < 0.05, **p < 0.01.
(MEFs) compared to that in MEF-WT, which was rescued by the ectopic expression of human PIN1 ( Supplementary Fig. S1b, c). Real-time quantitative polymerase chain reaction (qPCR) revealed that PIN1 did not affect the levels of METTL3 mRNA (Fig. 3b), indicating that PIN1 may regulate METTL3 stability. FBS-induced METTL3 expression was significantly increased in PIN1overexpressing MCF7 cells, suggesting that PIN1 enhanced the FBS-induced expression of METTL3 (Fig. 3c, d). The relative expression of METTL3-WT was higher than that of METTL3-S525A after the transient transfection of corresponding plasmids into MCF7 cells (Fig. 3e). In addition, FBS starvation significantly reduced the expression of both forms, and FBS stimulation rescued the expression of METTL3-WT, but not that of METTL3-S525A (Fig. 3e). These results indicated that METTL3 phosphorylation at S525 was essential for regulating FBS-induced METTL3 expression. The half-life of METTL3-S525 was significantly decreased compared with that of METTL3-WT after inhibiting protein synthesis in MCF7 cells using cycloheximide (Fig. 3f, g). Similarly, PIN1 knockout in MCF7 cells reduced METTL3 stability following a cycloheximide chase ( Supplementary Fig. S1d). Given that protein stability is regulated via proteasomal and lysosomal degradation pathways, we examined the precise mechanism underlying METTL3 stability regulated by PIN1. METTL3 expression increased following treatment with MG132, a proteasome inhibitor, in a time-dependent manner suggesting that METTL3 was degraded via the proteasomal pathway ( Supplementary  Fig. S1e, upper). In addition, MG132 treatment elevated the METTL3 levels in PIN1-overexpressing MCF7 cells (Fig. 3h, left panel). In contrast, reduced METTL3 expression resulting from the PIN1 knockout was rescued by MG132 treatment, indicating that PIN1 prevented the proteasomal degradation of METTL3 to increase its stability (Fig. 3h, right panel). As MCF7 cells have increased lysosomal function, we examined the lysosomemediated degradation of METTL3. Interestingly, treatment of MCF7 cells with chloroquine, a lysosome inhibitor, increased the METTL3 levels in a time-dependent manner indicating that METTL3 is also degraded via the lysosomal pathway (Supplementary Fig. S1e, lower). Chloroquine treatment further upregulated METTL3 expression in PIN1-overexpressing MCF7 cells (Fig. 3i, left panel). In contrast, reduced METTL3 expression in PIN1-knockout cells was augmented by chloroquine treatment (Fig. 3i, right panel). Simultaneous inhibition of proteasomal and lysosomal degradation by co-treatment with MG132 and chloroquine synergistically enhanced the expression of METTL3 in MCF7 cells (Fig. 3j) which was augmented in PIN1-overexpressing cells (Fig. 3k). These data indicated that PIN1 reduced the proteasomal and lysosomal degradation of METTL3 to increase its stability. FBSinduced METTL3 expression was enhanced upon co-treatment with MG132 and CHQ (Fig. 3l). Proteasomal degradation is mediated through protein ubiquitination [23]. The knockout of PIN1 in MCF7 cells increased METTL3 polyubiquitination, suggesting that PIN1 prevented the ubiquitination of METTL3 (Fig. 3m). In contrast, PIN1 overexpression reduced METTL3 polyubiquitination of METTL3 ( Supplementary Fig. 1g). Furthermore, decreased METTL3 expression in response to GFP-LC3 overexpression was inhibited upon ectopic expression of PIN1, suggesting that PIN1 prevented the lysosomal degradation of METTL3 facilitated by autophagosomes (Fig. 3n).
PIN1 enhanced METTL3-mediated deposition of the m 6 A modification on TAZ and EGFR mRNA METTL3 is associated with the m 6 A modification of mRNA [17]. PIN1 overexpression in MCF7 cells increased the m 6 A levels based on enzyme-linked immunosorbent assay (ELISA) and dot blotting of poly(A) RNA ( Fig. 4a and Supplementary Fig. S2a), whereas the METTL3 knockout attenuated m 6 A modification induced by PIN1 overexpression (Fig. 4b and Supplementary Fig. S2b). Additionally, FBS stimulation synergistically increased the m 6 A levels in PIN1overexpressing MCF7 cells, which was inhibited by the METTL3 knockout ( Fig. 4c and Supplementary Fig. S2c). These data indicated that PIN1 increased the m 6 A levels by enabling the enhanced stability of METTL3. Examination of mRNAs containing m 6 A in a previous report [24] revealed several proteins that were positively correlated with PIN1 expression using the DepMap dataset. TAZ and EGFR were both modified by m 6 A and their protein levels were positively correlated with PIN1 expression (Supplementary Fig. S2d and Supplementary Table S4). In addition, analysis of previously published data on lung cancer cells [25] revealed that METTL3 enhanced the translation efficiency of TAZ and EGFR ( Supplementary Fig. S2e). Therefore, TAZ and EGFR were selected for further studies. Immunoprecipitation of methylated RNA followed by reverse transcription (RT)-PCR (m 6 A-RIP-RT-PCR) revealed that PIN1 overexpression in MCF7 cells increased the levels of m 6 A modification in TAZ and EGFR mRNA ( Fig. 4d and Supplementary Fig. S2f), whereas the PIN1 knockout in MCF7 cells reduced the levels of the m 6 A modification ( Fig. 4e and Supplementary Fig. S2g). The METTL3 knockout abolished the increased levels of m 6 A modification of TAZ and EGFR induced in response to PIN1 overexpression ( Fig. 4f and Supplementary Fig.  S2h). Rescue of METTL3-WT or METTL3-S525A in METTL3 knockout MCF7 cells showed that ectopic expression of METTL3-WT restored the levels of m 6 A modification in TAZ and EGFR mRNA, whereas METTL3-S525A failed to restore the levels of this modification (Fig. 4g, h). Furthermore, treatment with serum significantly increased the relative m 6 A modification of TAZ and EGFR mRNAs in PIN1-overexpressing MCF7 cells compared to mock-transfected cells (Fig. 4i, j), indicating that the phosphorylation-dependent stabilization of METTL3 by PIN1 increased the levels of m 6 A modification in TAZ and EGFR mRNA.
PIN1 increased the efficiency of m 6 A modification-mediated TAZ and EGFR translation M 6 A modification is known to regulate mRNA stability as well as translation efficiency [19,25,26]; therefore, we first analyzed the effect of PIN1-METTL3 axis on the stability of TAZ and EGFR mRNA. Chasing of mRNA levels after actinomycin D treatment of MCF7 Fig. 3 PIN1 enhanced METTL3 stability by reducing its ubiquitination and lysosomal degradation. a, b PIN1 increased protein, but not mRNA levels of METTL3. MCF7 cells transfected with mock or XP-PIN1 (left) and sgCtrl or sgPIN1 (right) were detected by IB using the indicated antibodies (a). mRNA expression levels in MCF7 cells transfected with mock or XP-PIN1 and sgCtrl or sgPIN1 were measured using real-time PCR with primers targeting the indicated genes (b). Values are presented as the mean ± SD, N = 3. c, d PIN1 enhanced METTL3 expression induced by FBS stimulation. MCF7 cells transfected with mock or XP-PIN1 were serum-starved for 24 h and exposed to different percentages of FBS for 30 min, followed by IB with the indicated antibodies (c). Band densities were quantified using ImageJ and normalized to those of β-actin (d). Values represent the mean ± SD using the unpaired t test (N = 3); **p < 0.01 (mock versus XP-PIN1, corresponding FBS percentages). e FBS stimulation regulated WT METTL3 expression, but not the S525A mutant. MCF7 cells transfected with FLAG-METTL3-WT or FLAG-METTL3-S525A were grown in complete media or serum-free media for 24 h. Serum-starved cells were then exposed to 10% FBS for 30 min. Proteins in cell lysates were detected using the indicated antibodies. f, g METTL3-S525A had a shorter half-life than that of METTL3-WT. MCF7 cells expressing FLAG-METTL3-WT or FLAG-METTL3-S525A were treated with cycloheximide (CHX, 100 µg/mL) for the indicated times. Proteins in cell lysates were detected by IB using the indicated antibodies (f). The band densities were measured using ImageJ and normalized to those of β-actin (g). Values are the mean ± SD using the unpaired t test (N = 3), *p < 0.05 (WT versus S525A, corresponding time). Half-life (t1/2) value indicated in the inset (h, i) PIN1 overexpression inhibited the proteasome-mediated-and lysosomal degradation of METTL3. MCF7 cells transfected with mock or XP-PIN1 (left) and sgCtrl or sgPIN1 (right) and treated (+) or untreated (-) with MG132 (20 µM) for 3 h were detected by IB using the indicated antibodies (h). MCF7 cells transfected with mock, XP-PIN1, sgCtrl, or sgPIN1 and treated (+) or untreated (-) with chloroquine (CHQ, 100 µg/mL) for 3 h were detected by IB using the indicated antibodies (i). j Synergistic enhancement of METTL3 stability by the dual inhibition of proteasomal-and lysosomal degradation pathways. MCF7 cells treated (+) or untreated (-) with MG132 or CHQ for 3 h were detected by IB using the indicated antibodies. k PIN1 overexpression rescued METTL3 stability by the simultaneous inhibition of lysosomal-and proteasomal degradation of METTL3. MCF7 cells transfected with XP-PIN1 were treated (+) or untreated (−) with MG132 and CHQ for 3 h. Proteins in the cell lysates were detected by IB using the indicated antibodies. l Synergistic enhancement of METTL3 stability induced by FBS stimulation and the dual inhibition of the proteasomal-and lysosomal degradation pathways. MCF7 cells were serum-starved for 24 h and pretreated with MG132 and/or CHQ for 3 h, followed by stimulation with 10% FBS for 30 min. Proteins in the cell lysates were detected by IB using the indicated antibodies. m PIN1 prevented METTL3 polyubiquitination. MCF7 cells co-transfected with sgPIN1 and HA-Ub were pretreated with MG132 (20 µM) for 3 h, followed by IP with anti-METTL3 and IB with HA-HRP antibodies. Cell lysates were detected by IB using the indicated antibodies. n PIN1 inhibited the autophagosome-mediated lysosomal degradation of METTL3. MCF7 cells were co-transfected with GFP-LC3 and/or XP-PIN1, followed by IB using the indicated antibodies.
cells revealed that knockout of PIN1 does not affect the stability of TAZ and EGFR mRNA ( Supplementary Fig S3a-c). Next, we examined the effects of PIN1 on the m 6 A-dependent translation of TAZ and EGFR. Analysis of previously published m 6 A-RIP-RNAseq data [24] revealed that m 6 A sites were enriched in the 3ʹ UTR and around the stop codons of TAZ and EGFR mRNA (Fig. 5a, integrative genomics viewer (IGV) plot). Translation reporter constructs were designed by inserting the consensus wild-type (denoted as WT) "GGAC" m 6 A modification sites of TAZ and EGFR into the 3ʹ UTR of the Renilla luciferase gene (RLuc) present in the psiCHECK3 vector (Fig. 5a). In addition, adenosine (A) of "GGAC" was substituted with thymidine (T) to construct m 6 A-deficient mutant (denoted as Mut) reporter vectors (Fig. 5a). PIN1 overexpression in MCF7 cells increased the translation of Renilla luciferase gene containing WT m 6 A sites derived from TAZ and EGFR compared to mock-transfected groups but failed to increase the translation of the Mut sites (Fig. 5b). Furthermore, stimulation of MCF7 cells with FBS significantly increased the translation of the Renilla luciferase gene, which was attenuated upon PIN1 knockout, suggesting that PIN1 promoted the translation of TAZ and EGFR mRNA via the m 6 A modifications (Fig. 5c). Nascent protein synthesis in PIN1-knockout MCF7 cells was detected by evaluating Values are presented as the mean ± SD using the unpaired t test (N = 3), *p < 0.05. b PIN1 increased the m 6 A modification via METTL3. PIN1-overexpressing MCF7 cells were transfected with sgCtrl or sgMETTL3. Global m 6 A levels were measured using ELISA. Values are presented as the mean ± SD using one-way ANOVA (N = 3), ***p < 0.05. c PIN1 enhanced m 6 A levels induced by FBS stimulation via METTL3. PIN1-overexpressing MCF7 cells were transfected with sgCtrl or sgMETTL3, serum-starved for 24 h, then treated with 10% FBS for 30 min. Global m 6 A levels were measured using ELISA. Values are presented as the mean ± SD using oneway ANOVA (N = 3); *p < 0.05, ***p < 0.001. d-f PIN1 increased m 6 A modification of TAZ and EGFR induced by METTL3. Immunoprecipitation of total RNA isolated from MCF7 cells transfected with mock or XP-PIN1 using normal IgG or anti-m 6 A antibody. Gene expression in the input or IP fractions was analyzed by real-time PCR with primers targeting the indicated genes (d). Immunoprecipitation of total RNA isolated from MCF7 cells transfected with sgCtrl or sgPIN1 using normal IgG or anti-m 6 A antibody. Gene expression in the input or IP fractions was analyzed by real-time PCR with primers targeting the indicated genes (e). PIN1-overexpressing MCF7 cells were transfected with either sgCtrl or sgMETTL3. Total RNA isolated from cells was immunoprecipitated using anti-m 6 A antibody, and gene expression in the input and IP fractions was analyzed by real-time PCR with primers targeting the indicated genes (f). For figures (d and e), values are presented as the mean ± SD unpaired t test (N = 3); **p < 0.01, ***p < 0.001. For figure (f), values are presented as the mean ± SD using one-way ANOVA (N = 3); ***p < 0.001. g, h WT METTL3, but not S525A, induced the m 6 A modification of TAZ and EGFR. METTL3-knockout MCF7 cells transfected with mock, FLAG-METTL3-WT, or FLAG-METTL3-S525A constructs. Total RNA isolated from cells was immunoprecipitated using anti-m 6 A antibody, and gene expression in the input and IP fractions was analyzed by semi-quantitative RT-PCR (g) and real-time PCR (h) with primers targeting the indicated genes. METTL3 expression was detected by IB using an anti-METTL3 antibody. Values are presented as the mean ± SD using one-way ANOVA (N = 3); **p < 0.01, ***p < 0.001. i, j PIN1 enhanced m 6 A modification of TAZ and EGFR upon FBS stimulation. MCF7 cells overexpressing PIN1 were serum-starved for 24 h and exposed to 10% FBS for 30 min. Total RNA isolated from cells was immunoprecipitated using anti-m 6 A antibody. Gene expression in the input and IP fractions was analyzed by semi-quantitative RT-PCR (i), and real-time PCR (j) using primers targeting the indicated genes. Values are presented as the mean ± SD using one-way ANOVA (N = 3), *p < 0.05. the expression of puromycilated proteins via immunofluorescence to verify the regulatory role of PIN1 in translation. The PIN1 knockout significantly reduced the incorporation of puromycin into nascent proteins which was recovered by ectopic expression of METTL3 (Fig. 5d, e). Furthermore, reduced puromycilation of TAZ and EGFR in PIN1-knockout MCF7 cells was recovered by overexpression of METTL3, indicating that PIN1 effectively promoted the translation of TAZ and EGFR in a METTL3dependent manner (Fig. 5f, g). As METTL3 is known to directly bind to m 6 A-modified mRNAs to facilitate mRNA loop formation and polysome assembly [19,25]; we examined the interaction of METTL3 with TAZ and EGFR mRNA in PIN1-knockout MCF7 cells (Fig. 5h). TAZ and EGFR mRNA tagged with bacteriophage MS2 hairpin loop were precipitated with His-MS2BP using Ni-NTA resin to detect the mRNA-protein interactions. PIN1 knockout reduced the binding of METTL3 to TAZ and EGFR mRNA (Fig. 5i, j). Furthermore, polysome analysis revealed that the PIN1 knockout significantly reduced the global translation, as there was a simultaneous reduction in peaks corresponding to 40/60/80 S (Fig. 5k). In addition, the distribution of TAZ and EGFR mRNA was shifted from polysomes to monosomes in PIN1-knockout MCF7 cells, suggesting that PIN1 promotes the translation efficiency of these mRNAs (Fig. 5k, l). Moreover, immunoblotting of the fractionated samples showed that the m 7 G cap-binding protein eIF4E, translation initiation factor eIF3 core subunit eIF3b, and METTL3 shifted from polysomal to monosomal fractions in PIN1knockout MCF7 cells, suggesting a facilitatory role of PIN1 in METTL3-mediated polysome assembly (Fig. 5k). Finally, FBS treatment increased the expression of TAZ and EGFR in wild type but not in PIN1-knockout MCF7 cells, whereas overexpression of METTL3 restored TAZ and EGFR level in PIN1-knockout cells (Fig. 5m). These data indicated that PIN1 enhanced the METTL3induced translation, resulting in increased TAZ and EGFR expression in breast cancer cells.
Combinatorial treatment with MEK1/2 and PIN1 inhibitors synergistically reduced TAZ and EGFR translation and inhibited breast tumorigenesis As the METTL3-PIN1 interaction was modulated by the MEK/ERK pathway, we aimed to use a combination of MEK1/2 and PIN1 inhibitors as a therapeutic strategy for breast cancer. Cycloheximide chasing revealed that simultaneous inhibition of MEK1/2 with PD98059 or trametinib and PIN1 with all-trans retinoic acid (ATRA) resulted in reduced METTL3 stability (Fig. 6a, b). Furthermore, combinatorial treatment with PD98059 and ATRA resulted in significantly reduced global m 6 A levels in poly(A) RNA (Fig. 6c), reduced m 6 A modification of TAZ and EGFR mRNA, and concomitantly decreased protein levels (Fig. 6d). The translation reporter assay showed that co-treatment with PD98059 and ATRA resulted in synergistically reduced the translation efficiency of TAZ and EGFR (Fig. 6e), which was consistent with reduced METTL3 expression in Fig. 6a. Co-treatment of MCF7 cells with PD98059 and ATRA resulted in significantly reduced incorporation of BrdU into nascent DNA, indicating that this drug combination inhibited cell proliferation (evaluated using ELISA) (Fig. 6f), and induced significant cell cycle arrest in the G0/G1 phase (Fig. 6g, h). METTL3 overexpression released the MCF7 cells from the PD98059 and ATRA-induced arrest in the G0/G1 phase (Fig. 6i, j), and increased the proliferation of MCF7 cells (based on the significant recovery in the population of Ki67-positive cells) (Fig. 6k, l). These data suggested that the combinatorial treatment with PD98059 and ATRA exerted anticancer effects against breast cancer cells by reducing METTL3 stability.

PIN1 promoted Mettl3-induced breast tumorigenesis in vivo
Increased colony formation in PIN1-overexpressing MCF7 cells was attenuated upon knocking out METTL3, suggesting that PIN1 promoted tumorigenesis by enabling METTL3 stabilization (Fig. 7a,  b). The PIN1-METTL3 axis-induced breast tumorigenesis was studied in vivo using an orthotopic mouse model. Mouse breast cancer cells from the 4T1 cell line stably overexpressing human PIN1 were transfected with either sgCtrl or two separate guide RNAs targeting mouse Mettl3 (sgMettl3-1 and -2), followed by injection into the mammary glands of BALB/c mice (Fig. 7c). PIN1 overexpression significantly increased weight and volume of the tumors formed by the 4T1 cells (Fig. 7c, d). In contrast, CRISPR/Cas9-mediated knockout of Mettl3 using separate guide RNAs consistently reduced the tumor size and volume in PIN1-overexpressing 4T1 cells (Fig. 7c, d). Increased tumor size is correlated with development of the necrotic core [27]. Histopathological staining of the tumor sections using hematoxylin and eosin (H&E) revealed that the necrotic area was larger in PIN1-overexpressing 4T1 mice compared with that in the mock 4T1 mice. Furthermore, Mettl3 knockout resulted in reduced tumor necrosis in PIN1-overexpressing 4T1 mice, suggesting that it reduced necrotic cell death to suppress tumor formation in PIN1overexpressing 4T1 mice (Fig. 7e). Dot blotting using an m 6 A antibody showed that PIN1 overexpression significantly increased the global m 6 A levels in mouse tumors. M 6 A levels decreased in the Mettl3 knockout (Fig. 7f). Furthermore, m 6 A-RIP using RNA isolated from tumor samples revealed that PIN1 overexpression increased TAZ and EGFR mRNA methylation with a concomitant increase in protein expression in a METTL3-dependent manner (Fig. 7g, h). These data indicated that PIN1 increased breast tumorigenesis in vivo by increasing the stability of METTL3. Furthermore, PIN1 overexpression was associated with increased m 6 A-dependent translation of Taz and Egfr mRNA in mouse tumors.  6 A-dependent translation of TAZ and EGFR via increased polysome assembly. a Schematic representation of the psiCHECK3 translation reporter construct (left). Integrative genomics viewer (IGV) plots of m 6 A peaks for TAZ (right, upper) and EGFR (right, lower) mRNA obtained from the published dataset. The blue boxes, blue lines, red histogram, and blue histogram represent exons, introns, the m 6 A IP sample, and the input sample, respectively. The nucleotide sequence marked with arrowheads was PCR-amplified and cloned into psiCHECK3 vector. The wild-type (WT) m 6 A modification motifs "GGAC" and methylation-deficient mutant (Mut) "GGTC" are underlined. b PIN1 promoted the m 6 A-dependent translation of TAZ and EGFR mRNA in MCF7 cells. MCF7 cells were transfected with psiCHECK3, psiCHECK3-TAZ-WT, -TAZ-Mut, psiCHECK3-EGFR-WT, or -EGFR-Mut with or without XP-PIN1. Renilla luciferase activity in lysates from transfected cells was measured and normalized to Renilla luciferase (RLuc) mRNA level. The representative agarose gel image indicates the mRNA level of RLuc as determined using RT-PCR. Values are presented as the mean ± SD using one-way ANOVA (N = 3), **p < 0.01, ***p < 0.001. c FBS stimulation promoted m 6 A-mediated translation of TAZ and EGFR in a PIN1-dependent manner. MCF7 cells were transfected with psiCHECK3, psiCHECK3-TAZ, or psiCHECK3-EGFR with or without sgPIN1 in FBS-free media. After 24 h, cells were exposed to 10% FBS for 1 h, and Renilla luciferase activity in the cell lysates were measured and normalized to RLuc mRNA level. The representative agarose gel image indicates the mRNA level of RLuc as determined using RT-PCR. Values are presented as the mean ± SD using one-way ANOVA (N = 3), ***p < 0.001. d, e PIN1 knockout reduced nascent protein synthesis in MCF7 cells. MCF7 cells transfected with sgCtrl or sgPIN1 with or without FLAG-METTL3 were treated with puromycin (2 µg/mL) for 2 h. Puromycilated nascent proteins were detected by immunofluorescence using anti-puromycin antibody (red). Cells were visualized by phase-contrast microscopy (d). Relative fluorescence intensity produced by cells was quantified using Fiji software (e). Values are presented as the mean ± SD using the unpaired t test (N = 6), **p < 0.01, ***p < 0.001. f, g PIN1 knockout reduced puromycilation of TAZ and EGFR proteins. MCF7 cells transfected with FLAG-TAZ, Myc-METTL3, sgCtrl, or sgPIN1 were treated with puromycin (2 µg/mL) for 1 h. Lysates from treated cells were immunoprecipitated with anti-puromycin antibody, followed by IB with anti-FLAG antibody (f). MCF7 cells transfected with Myc-EGFR, FLAG-METTL3, sgCtrl, or sgPIN1 were treated with puromycin (2 µg/µL) for 1 h. Cell lysates were immunoprecipitated with antipuromycin antibody, followed by IB with anti-EGFR antibody (g). The cell lysate was detected by IB using the indicated antibodies. h Schematic diagram of the MS2-tagged RNA affinity purification. PIN1 knockout reduced the association of METTL3 with TAZ-and EGFR mRNA. MCF7 cells transfected with FLAG-TAZ-MS2 (i) or Myc-EGFR-MS2 (j), and His-MS2BP were irradiated with UV light. The ribonucleoprotein complex was pulled down using Ni-NTA resin, and the protein and mRNAs in the pulldown and input fractions were analyzed with IB and RT-PCR using the indicated antibodies and primers, respectively. k, l The PIN1 knockout reduced polysome assembly and distribution of TAZ and EGFR mRNA in the polysome fraction. Lysates prepared from MCF7 cells transfected with sgCtrl or sgPIN1 were subjected to sucrose-gradient centrifugation. Fractionated samples were analyzed by IB and RT-PCR using the indicated antibodies and primers, respectively (k). mRNA levels of TAZ and EGFR in each fraction were analyzed using real-time PCR with primers targeting indicated genes (l). Values are presented as the mean ± SEM. N = 3. m PIN1 increased TAZ and EGFR protein expression induced by FBS. PIN1-knockout MCF7 cells were transfected with FLAG-METTL3, serum-starved for 24 h, and then exposed to 10% FBS for 1 h. Protein levels in cell lysates were detected by IB using the indicated antibodies.

DISCUSSION
Epitranscriptomic RNA modification is an important mechanism for regulating gene expression [28]. To date, over 130 chemical modifications have been reported in RNA, with m 6 A being the most common modification in mRNA [28]. m 6 A regulates various aspects of mRNA biology, including stability, splicing, and translation efficiency, to mediate important physiological processes, such as neurodevelopment, immune response, and gametogenesis [29,30]. Upregulation of the key m 6 A methyltransferase (METTL3) promotes the progression of various human cancers, including breast cancer [16,31]. However, the mechanisms underlying the regulation of METTL3 expression and function remain unclear. In this study, we identified PIN1 as a positive regulator of METTL3 stability in breast cancer cells. PIN1 promoted the translation efficiency of TAZ and EGFR in an m 6 A-dependent mechanism to promote breast tumorigenesis.
METTL3 is expressed at higher levels in breast cancer cells than in normal cells [16,31]. Moreover, METTL3 overexpression in breast cancer is associated with the promotion of tumor cell invasion, metastasis, and resistance to chemotherapy, phenotypes that are often associated with high-grade breast cancer [15,32,33]. However, examination of gene expression at the mRNA level did not reveal a significant increase in METTL3 expression in breast tumors, suggesting that enhanced protein stability may be responsible for the increased levels of METTL3 [14]. Regulation of protein stability depends on protein phosphorylation and subsequent changes in protein ubiquitination [34]. METTL3 phosphorylation-induced by proline-directed ERK1/2promotes its stability by reducing polyubiquitination [22]. However, regulation of METTL3 ubiquitination remains poorly understood. Furthermore, the role of lysosomal degradation in determining METTL3 stability has not been elucidated. Here, we found that PIN1 binds to METTL3 through its PPIase domain. Although the WW domain of PIN1 has a 10-fold-higher binding affinity for peptides with the pSer/Thr motif, the PPIase domain alone binds with-and induces cis/trans isomerization around the motif [35]. We found that METTL3 phosphorylation at S525 was crucial for its interaction with PIN1, a result that was in agreement with the results of a previous study showing that METTL3 phosphorylation at S525 increases METTL3 stability [22]. This suggests that the PIN1-METTL3 interaction may play an important role in regulating the stability of METTL3. We showed that PIN1 reduces the levels of METTL3 polyubiquitination and its lysosomal degradation, thereby increasing its stability and half-life in breast cancer cells. Furthermore, METTL3 stabilization by PIN1 was associated with increased global levels of m 6 A mRNA modification.
It is essential to identify specific mRNA substrates to understand the precise mechanism by which m 6 A modifications promote breast tumorigenesis. Several studies have shown that TAZ and EGFR play critical roles in breast tumorigenesis [21,[36][37][38]. Interestingly, both TAZ and EGFR undergo METTL3-mediated deposition of the m 6 A modification in lung cancer tissues [19]. Our work showed that PIN1 increased the METTL3-mediated deposition of the m 6 A modification on TAZ and EGFR mRNA in MCF7 cells. Increased m 6 A modification of mRNA potentially regulates gene expression through several mechanisms, including the recruitment of m 6 A reader proteins, such as YTH domaincontaining proteins (YTHDF1/2/3), or by recruiting other indirect m 6 A readers, such as eIF3b [17]. The outcome of m 6 A deposition largely depends on the location of m 6 A residues in the mRNA. For instance, the presence of m 6 A in the 5ʹ UTR promotes capindependent translation via eIF3b recruitment, whereas m 6 A located within the coding sequence promotes translation by recruiting elongation factors such as eEF2 in a YTHDF1-dependent manner [39,40]. In contrast, m 6 A occurs around stop codons in TAZ and EGFR, and promotes translation by facilitating polysome formation in a METTL3-dependent manner [19]. Using translation reporter constructs containing m 6 A sites derived from TAZ and EGFR mRNA, we showed that PIN1 enhanced the efficiency of TAZ and EGFR translation. Previous studies have shown that METTL3 directly binds to m 6 A-modified mRNA in the cytosol, thereby promoting mRNA looping and translation via the formation of the METTL3-eIF3h bridge [25]. Furthermore, METTL3 bound to the 3ʹ UTR of an mRNA that recruits the initiation factor eIF3b to eIF4E to initiate mRNA translation [19]. We detected a strong interaction between METTL3 and TAZ or EGFR mRNA which was substantially reduced upon knocking out PIN1. Consequently, polysome formation and ribosome loading onto TAZ and EGFR mRNAs were significantly reduced in PIN1-knockout breast cancer cells.
As PIN1 interacts with METTL3 in a phosphorylation-dependent manner, we hypothesized that the dual inhibition of PIN1 and the upstream kinase pathway responsible for METTL3 phosphorylation may be an alternative therapeutic strategy. The MEK/ERK pathway phosphorylates METTL3 at S525 which facilitates its interaction with PIN1 [41,42]. Here, we showed that the dual inhibition of MEK1/2 and PIN1 using PD98059 and ATRA synergistically reduced METTL3 protein stability to suppress the m 6 A-dependent translation of TAZ and EGFR. The combination of these compounds induced cell cycle arrest at the G0/G1 phase, reduced cell proliferation, and decreased expression of Ki67 cell proliferation biomarker. Interestingly, METTL3 overexpression attenuated the anticancer effects of ATRA and PD98059. PD98059 alone is not effective against breast cancer and the sensitivity of cancer cells to ATRA is variable [42][43][44]. Here, we showed that combinatorial treatment with PD98059 and ATRA inhibited METTL3-induced translation, thereby offering a unique possibility to tackle breast tumorigenesis. The combination of ATRA and PLX4032 (a BRAF Fig. 6 Combinatorial treatment with MEKi and ATRA reduced the m 6 A-dependent translation of TAZ and EGFR. a, b Treatment with MEK inhibitors and ATRA accelerated METTL3 degradation. MCF7 cells were treated with PD98059 (10 µM, left) or trametinib (10 µM, right) alone, or in combination with ATRA (20 µM) for 24 h, followed by treatment with cycloheximide (100 µg/mL) for 6 h. Protein levels in cell lysates were detected by IB using the indicated antibodies (a), and the METTL3 band intensities were measured with ImageJ and normalized to those of β-actin (b). c Co-treatment with PD98059 and ATRA synergistically reduced the global m 6 A levels in MCF7 cells. MCF7 cells were treated with PD98059 (10 µM) alone, or in combination with ATRA (20 µM) for 24 h. m 6 A levels in poly(A) RNA were detected by dot blotting with an antim 6 A antibody. Methylene blue staining was used as the loading control. d Co-treatment with PD98059 and ATRA reduced the m 6 A modification of TAZ and EGFR mRNA in MCF7 cells. Total RNA from MCF7 cells treated with PD98059 (10 µM) and/or ATRA (20 µM) for 24 h was immunoprecipitated using anti-m 6 A antibody. mRNA levels in the input and IP fractions were analyzed by RT-PCR using the indicated primers. TAZ and EGFR expression in cell lysates was detected by IB using the respective antibodies. e Co-treatment with PD98059 and ATRA reduced the m 6 A-dependent translation of TAZ and EGFR. MCF7 cells were transfected with psiCHECK3, psiCHECK3-TAZ, or psiCHECK3-EGFR, followed by treatment with PD98059 (10 µM) and/or ATRA (20 µM) for 24 h. Renilla luciferase activity in cell lysates was measured and normalized to Renilla luciferase (RLuc) mRNA level. The representative agarose gel image indicates the mRNA level of RLuc as determined using RT-PCR. Values are presented as the mean ± SD using one-way ANOVA (N = 3); *p < 0.05, **p < 0.01, ***p < 0.001. f MCF7 cells were treated with PD98059 (10 µM) and/or ATRA (20 µM) for 24 h, and cell proliferation was determined using a BrdU incorporation assay. Values are presented as the mean ± SD, N = 3, ***p < 0.001. g, h Co-treatment with PD98059 and ATRA resulted in a G0/G1 phase arrest. Histogram showing the cell cycle distribution of MCF7 cells treated with PD98059 (10 µM) and/or ATRA (20 µM) for 24 h (g), and the G0/G1-arrested population (h). Values are presented as the mean ± SD using one-way ANOVA (N = 3), ***p < 0.001. i, j Ectopic METTL3 expression relieved MCF7 cells from PD98059/ ATRA-induced cell cycle arrest. Cell cycle distribution of MCF7 cells overexpressing FLAG-METTL3, followed by treatment with PD98059 (10 µM) and ATRA (20 µM) for 24 h (i), and the population of cells arrested at the G0/G1 phase (j). Values are presented as the mean ± SD using one-way ANOVA (N = 3), ***p < 0.001. k, l Combined treatment with ATRA and PD98059 reduced the expression of cell proliferation biomarker Ki67 in a METTL3-dependent manner. METTL3-overexpressing MCF7 cells were treated with PD98059 (10 µM) and ATRA (20 µM) for 24 h and stained with anti-Ki67-PE/7-AAD. Cell distribution was analyzed using Muse cell analyzer (k) and the percentage of Ki67-positive cells was determined (l). Values are presented as the mean ± SD using one-way ANOVA (N = 3), *p < 0.05.
V600E kinase inhibitor) has been reported to synergistically reduce EGFR expression in melanoma cells [45]. The Mettl3 knockout reduced growth of tumors formed by PIN1-overexpressing 4T1 cells in an orthotopic mouse model. This was consistent with the in vitro results and implied that PIN1-induced breast tumorigenesis depends on METTL3 stabilization.
Emerging evidence suggests that the dysregulation of protein synthesis is a central feature of cancer cell plasticity, a key challenge for anticancer chemotherapy [46]. Specifically, regulating protein synthesis via the epitranscriptomic modification of mRNA has attracted significant interest from a therapeutic perspective. However, the molecular mechanisms underlying the Fig. 7 Mettl3 knockout suppressed PIN1-induced breast tumorigenesis in vivo. a, b METTL3 knockout inhibited colony formation induced by PIN1 overexpression. Representative images of soft agar colonies formed by MCF7 cells stably overexpressing PIN1 followed by the transfection of two separate sgMETTL3 vectors (a), and quantification of colony size and numbers (b). Values are presented as the mean ± SD using one-way ANOVA (N = 6), ***p < 0.001. c, d Mettl3 knockout reduced PIN1-induced tumorigenesis in vivo. Representative images of tumors formed by 4T1 cells stably overexpressing human PIN1 transfected with sgCtrl or sgMettl3 (c). Tumor weight and volume at the time of sacrifice (d). Values are presented as the mean ± SD using one-way ANOVA; N = 5 (mock or XP-PIN1), 8 (sgMettl3-1 and -2); *p < 0.05, **p < 0.05, ***p < 0.001. e Histopathological examination of the tumor. Representative images of tumor sections stained with hematoxylin and eosin. The necrotic area is shown below each figure. f, g PIN1 increased the levels of m 6 A modification in Taz and Egfr mRNA via Mettl3 in vivo. Global m 6 A levels in poly(A) RNA samples isolated from mouse tumors were detected by dot blotting with anti-m 6 A antibody (f). Total RNA from mouse tumor samples was immunoprecipitated with anti-m 6 A antibody and mRNA levels in the input and IP fractions were analyzed using RT-PCR with primers targeting the indicated genes (g). h PIN1 increased Taz and Egfr protein levels via Mettl3 stabilization in vivo. Immunoblotting of lysates from mouse tumors using the indicated antibodies.
regulation of epitranscriptome-related proteins remains poorly understood. In this study, we revealed the regulatory role of PIN1 in mRNA translation through the regulation of METTL3 stability in breast cancer cells. In vivo studies and clinical trials using a combination of MEK and PIN1 inhibitors in breast cancer are warranted to determine their feasibility for clinical application. In addition, our data showed that PIN1 interacts with METTL3 within the SAM-binding motif which is essential for the methyltransferase activity of METTL3. Particularly, recovery of METTL3 (S525A) failed to induce the m 6 A modification of TAZ and EGFR mRNAs implying that interaction with PIN1 might also be important for METTL3 function. Further studies are required to understand whether PIN1 also regulates the catalytic function of METTL3. Furthermore, structural analysis of METTL3 and PIN1 binding interfaces may lead to the development of a small-molecule inhibitor of the METTL3-PIN1 interaction for therapeutic purposes. In conclusion, we showed that PIN1-mediated METTL3 stabilization is a critical regulator of breast tumorigenesis.

MATERIALS AND METHODS
All materials and methods, including the cell culture protocols and establishment of stable cells, antibodies, reagents, CRISPR/Cas9 knockout, mammalian expression vectors, the mammalian two-hybrid assay, polysome fractionation, cell proliferation assay (BrdU Incorporation), GST pulldown assay, Ni-NTA pulldown assay, BiFC and immunofluorescence microscopy, protein immunoblotting and immunoprecipitation, immunohistochemistry, poly(A) RNA dot blotting, meRIP-PCR, puromycin labeling and immunoprecipitation, luciferase assay, cell cycle analysis, Ki67 staining, MS2-tagged RNA pulldown assay, anchorage-independent cell transformation (soft agar assay), bioinformatic analyses, and statistical analyses can be found in the supplementary information of this article.

DATA AVAILABILITY
The proteomic analysis results published here are wholly-or partly based upon the data generated through The Cancer Dependency Map (DepMap) Project at the Broad Institute (The Cancer Target Discovery and Development screening project) and TCGA. The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials.