PELO facilitates PLK1-induced the ubiquitination and degradation of Smad4 and promotes the progression of prostate cancer

PLK1 and Smad4 are two important factors in prostate cancer initiation and progression. They have been reported to play the opposite role in Pten-deleted mice, one is an oncogene, the other is a tumor suppressor. Moreover, they could reversely regulate the PI3K/AKT/mTOR pathway and the activation of MYC. However, the connections between PLK1 and Smad4 have never been studied. Here, we showed that PLK1 could interact with Smad4 and promote the ubiquitination and degradation of Smad4 in PCa cells. PLK1 and PELO could bind to different domains of Smad4 and formed a protein complex. PELO facilitated the degradation of Smad4 through cooperating with PLK1, thereby resulting in proliferation and metastasis of prostate cancer cell. Changes in protein levels of Smad4 led to the alteration of biological function that caused by PLK1 in prostate cancer cells. Further studies showed that PELO upregulation was positively associated with high grade PCa and knockdown of PELO expression significantly decreased PCa cell proliferation and metastasis in vitro and vivo. PELO knockdown in PCa cells could enhance the tumor suppressive role of PLK1 inhibitor. In addition, blocking the interaction between PELO and Smad4 by using specific peptide could effectively inhibit PCa cell metastasis ability in vitro and vivo. Overall, these findings identified a novel regulatory relationship among PLK1, Smad4 and PELO, and provided a potential therapeutic strategy for advanced PCa therapy by co-targeting PLK1 and PELO.


INTRODUCTION
Prostate cancer (PCa) is one of the leading cancer types in men in the world. Even though the survival rate of localized PCa is high, the lethal rate will become higher when it relapses and transformed into metastatic tumor cells [1,2]. Polo-like kinase 1 (PLK1) plays an important role in regulating cell cycles, such as centrosome maturation, spindle assembly, mitotic exit and cytokinesis. The increased PLK1 expression in Pten-deleted mice could induce prostate cancer formation [3]. Targeting PLK1 inhibits the activity of the PI3K/AKT/mTOR signaling pathway [4]. Higher PLK1 expression has been identified positively associated with high grade PCa [5] and is considered as a therapeutic target in multiple cancer types [6]. Thus far, PLK1 inhibitors have already been applied to treat multiple cancers in clinical trials [7]. Unlike PLK1, Smad4 was identified as a tumor suppressor in human PCa, Smad4 deletion drives progression of Pten-deficient mice prostate tumor to highly aggressive prostate cancer [8]. Interestingly, PLK1 induces MYC phosphorylation and protein accumulation [9], whereas Smad4 is a transcription suppressor of MYC in cancer cells [10]. Whether there are any relationships between the oncogenic factor PLK1 and the tumor suppressor Smad4 in prostate cancer is still unclear.
PELO is an indispensable gene in the cell meiotic division process and mice embryonic development [11]. PELO knockout in mice results in early embryonic lethality because it plays a crucial role in cell mitosis. The significant role of PELO in ribosome-associated mRNA quality control mechanisms has been first found in yeast [12]. Although it was reported to be a tumor suppressor in breast cancer [13], the function of PELO in prostate cancer initiation, progression and metastasis has not yet been studied.
In the present study, we investigated PLK1 evoked oncogenic transformation related events. We identified the crucial role of PLK1 in regulating the ubiquitination and degradation of Smad4. Our findings supported that PELO was an important factor to facilitate PLK1-induced the ubiquitination and degradation of Smad4, thereby resulting in proliferation and metastasis of prostate cancer cell. We further proved that PELO was highly expressed in PCa tissues, knockdown of PELO markedly inhibited cell growth and metastasis ability. These findings expanded our understanding of regulatory relationship among PLK1, Smad4 and PELO in prostate cancer, and provided new insights into potential PCa treatments targeting PLK1 and PELO.

PLK1 promoted the ubiquitination and degradation of Smad4 in prostate cancer
To explore the relationship between PLK1 and Smad4, we first investigated whether there was interaction between PLK1 and Smad4 through exogenous and endogenous Coimmunoprecipitation experiments. The results indicated that PLK1 strongly interacts with Smad4 (Fig. 1A, B). Next, we found that the expression level of Smad4 was gradually decreased when the amount of PLK1 increased in PCa cells ( Fig. 1C and Supplementary Fig. 1A). Accordingly, PLK1 inhibitor RO3280 (RO) [14] could elevate the protein level of Smad4 in PC3 and LNCaP cells in a dose-dependent manner ( Fig. 1D and Supplementary  Fig. 1B). To investigate the molecular mechanism by which PLK1 regulated Smad4 protein expression level, we assessed the halflife of Smad4 protein. PLK1 overexpression PC3 cells or control cells were treated with the protein synthesis inhibitor cycloheximide (CHX) at different time points. The results showed Smad4 and Flag-tagged PLK1 expression plasmids as indicated, whole cell lysates were extracted and immunoprecipitated with anti-Flag antibodies. B PC3 cells extracts were immunoprecipitated with Smad4 antibody or normal rabbit IgG, and then subjected to immunoblot with PLK1 or Smad4 antibody. C PC3 cells were transfected with increasing amounts of Flag-PLK1 expression plasmids as indicated. The Smad4 protein level was assessed by immunoblotting. GAPDH was used as a loading control. D PC3 cells were treated with increasing amounts of RO3280 as indicated. The protein levels of PLK1 and Smad4 were analyzed by immunoblotting. E, F PC3 cells transfected with Flag-PLK1 or control vector were treated with 50 μg/mL cycloheximide (CHX) for the indicated times and then analyzed by western blot. Representative western blot (E) and quantification of relative protein levels (F) are shown. G-H, PC3 cells were treated with RO3280 for 12 h, followed by treatment with 50 μg/mL CHX for the indicated times. The protein levels of PLK1 and Smad4 were assessed by western blot. Representative western blot (G) and quantification of relative protein levels (H) are shown. I PC3 cells transfected with Flag-PLK1 or control vector were treated with or without MG132 (20 μM) for 8 hours. Cell lysates were subjected to immunoblot with Flag or Smad4 antibody. J HEK293T cells were transfected with increasing amounts of Flag-PLK1 together with HA-Ub and Myc-Smad4. Immunoprecipitation was performed with anti-Myc antibody, and subjected to western blot with anti-HA, Myc or Flag antibody. K HEK293T cells were treated with various amounts of RO3280 for 12 h. Then cells were harvested for immunoprecipitation and immunoblot analysis using the indicated antibodies. Ubiquitin antibody was used to detect endogenous ubiquitinated Smad4. L Phosphorylation sites (S, serine; T, threonine) of smad4 by PLK1 (http://gps. biocuckoo.cn/online.php). M HEK293T cells were transfected with HA-Ub and Myc-Smad4 Wild-Type or mutant plasmids together with Flag-PLK1 or control vectors. Immunoprecipitation was performed with anti-Myc antibody, and subjected to western blot with anti-HA, Myc or Flag antibody. GAPDH was used as a loading control. N HEK293T cells were transfected with Myc-Smad4 Wild-Type or Myc-Smad4 (T197A) together with Flag-PLK1 or control vectors. Immunoprecipitation was performed with anti-Myc antibody, and subjected to western blot with anti-p-Thr, Myc or Flag antibody. GAPDH was used as a loading control. Data are represented as means ± S.D of three independent experiments. *p < 0.05; **p < 0.01; two-way ANOVA test (E-H).
that PLK1 overexpression promoted the protein degradation of Smad4 and shortened the half-life of Smad4 protein (Fig. 1E, F). However, PLK1 inhibitor RO3280 in PC3 cells led to the reduced degradation of Smad4 protein. (Fig. 1G, H). We then found that the degradation of Smad4 protein caused by PLK1 was obviously inhibited by the ubiquitination-proteasome pathway inhibitor MG132, suggesting that PLK1 attenuated Smad4 protein stability through the ubiquitination-proteasome pathway (Fig. 1I).
To further validate the results, we co-expressed HA-ubiquitin, Myc-Smad4 or Flag-PLK1 in HEK293T cells. Smad4 protein was immunoprecipitated by MYC antibody, and the ubiquitinated protein was detected with HA antibody. The results indicated that PLK1 overexpression could strongly strengthen the ubiquitination of Smad4 protein in a dose-dependent manner (Fig. 1J). Then we treated PC3 cells with increasing amount of RO and measured the endogenous ubiquitin Smad4 protein that was immunoprecipitated by the Smad4 specific antibody. We found that RO led to the decreased ubiquitination of Smad4 protein in a dose-dependent manner (Fig. 1K). The results further validated that PLK1 accelerated the ubiquitination and degradation of Smad4 through the ubiquitination-proteasome pathway.
To explore how does PLK1 increase the ubiquitination and degradation of Smad4, we performed bioinformatic analysis using GPS. group based prediction system (http://gps.biocuckoo.cn/ online.php) and found that four amino acid sites at the linker region of Smad4 (Ser171, Ser187, Ser191, Thr197) might be the potential phosphorylation site of PLK1 (Fig. 1L). Then we mutated these four sites and performed ubiquitination experiments, respectively. We observed that PLK1 could significantly promote the ubiquitination and degradation of Smad4 wild type, Smad4 S171A, Smad4 S187A, Smad4 S191A, respectively, but PLK1induced the ubiquitination and degradation of Smad4 T197A were obviously inhibited (Fig. 1M). Moreover, the phosphorylation experiment was also performed, the results demonstrated that PLK1 could strongly enhance the phosphor-threonine status of Smad4 wild type, however, PLK1-induced the phosphor-threonine status of Smad4 T197A was significantly diminished compared with wild type Smad4 (Fig. 1N). Taken together, these results suggested that PLK1 probably promoted the ubiquitination and subsequent degradation of Smad4 by phosphorylating Smad4 at Thr197.

Smad4 reversed the biological function of PLK1 in prostate cancer
To determine whether the amount of Smad4 protein is important for PLK1 to perform its oncogenic role in PCa, we performed rescue experiment. PC3 and LNCaP cells were treated with RO for 12 h respectively, followed by transfection with Smad4 siRNA. Then cell proliferation was measured by XTT colorimetric assays. We found that RO significantly decreased the cell proliferation both in PC3 and LNCaP cells, but knockdown of Smad4 rescued cell proliferation ( Fig. 2A, B and Supplementary Fig. S2E). Consistent with these results, PLK1 overexpression evidently enhanced the cell proliferation both in PC3 and LNCaP cells, but Smad4 elevation rescued cell proliferation (Fig. 2C, D). Next, we measured the migration ability of the PC3 and Du145 cells with the similar treatment as the above proliferation assay and proved that Smad4 could reverse the impact of PLK1 on the cell migration ability (Fig. 2E-H and Supplementary Fig. S2A-F). We further measured the expression levels of epithelial-mesenchymal transition markers, cell growth and migration related genes in PC3 cells. We found that Ectopic expression of Smad4 could reverse the mRNA levels of E-Cadherin, N-Cadherin, Vimentin, MMP9 and CCND1 in PLK1 overexpressed PC3 cells (Fig. 2I). These findings proved that PLK1 plays critical role in cell proliferation and migration by regulating Smad4 expression levels in PCa.

PLK1 formed a protein complex with Smad4 and PELO in PCa
To explore the Smad4-interacting proteins in prostate cancer cells, we initially performed an IP-MS assay by using specific anti-smad4 antibody to identify Smad4-binding proteins in PC3 cells (Fig. 3A, Supplementary Table 1). Then, we performed Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis to find the main functional pathways of the Smad4-interacting proteins. Among these Smad4-interacting proteins, 70 proteins are enriched in Spliceosome pathway, 24 proteins enriched in mRNA surveillance pathway and 22 proteins enriched in Ribosome pathway (Fig. 3B). Protein-protein interaction (PPI) networks result showed these proteins connect the pathways of mRNA surveillance, RNA transport, RNA degradation and Biosynthesis of amino acids together ( Supplementary Fig. S3A). One of these proteins, PELO, belongs to mRNA surveillance pathway but has fewer connections with other proteins in this pathway ( Supplementary  Fig. S3B). Then we ectopically expressed Myc-Smad4 and Flag-PELO in HEK239T cells and performed immunoprecipitation experiment using antibodies specifically identify Myc or Flag. The results proved that Smad4 could interact with PELO ( Fig. 3C, D). Glutathione S-transferase (GST) pull down assay further confirmed that Smad4 could directly interact with PELO (Supplementary Fig. S3C). To avoid the artifactual interactions, we then performed endogenous co-immunoprecipitation experiment, the result further proved that Smad4 could interact with PELO in cells (Fig. 3E). Considering both PLK1 and PELO could interact with Smad4, we then explored the associations between PLK1 and PELO and verified that PELO could interact with PLK1 ( Fig. 3F).
To determine which region of Smad4 is required for PLK1 or PELO binding, Myc-tagged full-length Smad4 and deletion mutants ( Fig. 3G) were co-transfected into HEK293T cells with PLK1 or PELO expression vector. Immunoprecipitations were performed using anti-Myc antibody. The results showed that functional domain Linker of Smad4 was essential for the binding of PLK1, whereas PELO could bind to the MH2 domain of Smad4 (Fig. 3H, I). In this way, our results indicated the associations between PLK1, PELO, and Smad4 occur at non-overlapping regions of the proteins. These data prompted us to hypothesize that Smad4 might form a complex with PLK1 and PELO. Smad4specific antibody was used to immunoprecipitated PC3 cells extracts and the immunoprecipitated proteins were analyzed by Western blotting. As expected, PLK1 and PELO could indeed be coimmunoprecipitated (Fig. 3J). Together, these results indicated that PLK1 and PELO could bind to Smad4 simultaneously through different regions, forming a protein complex. However, the role of PLK1/PELO/Smad4 complex in PCa needs to be further explored.

PELO facilitated PLK1-induced the ubiquitination and degradation of Smad4
To investigate the role of PELO in this proteins complex, we next designed short hairpin RNAs that could specifically knockdown PELO expression. Interestingly, similar to PLK1, knockdown of PELO also markedly elevated the protein level of Smad4 (Fig. 4A). Then we evaluated whether PELO influences PLK1-induced degradation of Smad4. As shown in Fig. 4B, C, PELO knockdown blocked PLK1-induced protein degradation of Smad4 in PC3 cells to a significant extent.
To explore whether PELO regulates the interaction between PLK1 and Smad4, PC3 cells transfected with PELO overexpression lentivirus or PELO shRNA lentivirus, and immunoprecipitations were performed with Smad4 antibody. The results showed that overexpression of PELO enhanced the interaction between PLK1 and Smad4 (Fig. 4D), whereas knockdown of PELO attenuated the interaction between PLK1 and Smad4 (Fig. 4E). Next, we evaluated whether PELO influences PLK1-induced ubiquitination of Smad4. We found that overexpression of PELO could enhance PLK1induced the ubiquitination and degradation of Smad4 in a dosedependent manner (Fig. 4F). Consist with this result, knockdown of PELO partially blocked PLK1-induced ubiquitination of Smad4 in PC3 cells (Fig. 4G). These results indicated that PELO was essential for mediating the interaction between PLK1 and Smad4, and PLK1-induced the ubiquitination and degradation of Smad4.

Increased expression of PELO is positively associated with high risk of PCa
To determine the role of PELO in PCa initiation and progression, we first examined the expression level of PELO in clinical prostate cancer samples. Hence, we analyzed the mRNA level of PELO in published prostate cancer sample cohorts. PELO was highly expressed in high Gleason score in samples from GSE134051 (Fig.  5A). Importantly, PELO was found highly expressed in high stage and metastatic PCa samples compared with primary samples in several cohorts (Fig. 5B, C and Supplementary Fig. S4D, F, G). Interestingly, we found that PELO was highly expressed in liver metastasis samples compared with lymph node metastasis samples in cohort GSE74685 (Supplementary Fig. S4E). We also were treated with RO3280 for 12 h, followed by transfection with Smad4 siRNA. Then cell proliferation was measured by XTT colorimetric assays (mean ± SD of triplicate experiments). PC3 cells (C) or LNCaP cells (D) stably expressing control or PLK1 plasmid were transiently transfected with Smad4 plasmid. Then cell proliferation was measured by XTT colorimetric assays (mean ± SD of triplicate experiments). Quantification analysis of migration assays (E) and representative trans-well data of migration assays (F) in PC3 cells treated with RO3280 for 12 h, followed by transfected with Smad4 siRNA. Quantification analysis of migration assays (G) and representative trans-well data of migration assays (H) in PC3 cells stably expressing control or PLK1 plasmid, followed by transfected with Smad4 plasmid. I Real-time RT-PCR analysis of mRNA level of different EMT markers and CCND1 in the indicated cells. Data are represented as means ± S.D. *p < 0.05; **p < 0.01; ***p < 0.001. P values were assessed using two-tailed Student's t tests (A-G).
found that PELO expression was correlated with the overall survival ( Fig. 5D) of PCa patients in the TCGA cohorts. Then, we collected two cohorts of prostate samples, one cohort includes 3 normal prostate tissues, 50 adjunct normal samples and 94 tumor samples, the other involves 42 adjunct normal samples and 95 primary tumor samples. By performing immunohistochemistry (IHC) analysis on the PCa tissue chips, we observed that PELO was significantly upregulated in tumor samples in comparison with the normal and adjunct normal prostate tissue samples in those two cohorts (Fig. 5E, F and Supplementary Fig.  S4A-C). Moreover, the protein level of PELO was highly expressed in high Gleason score 7-9, a vital parameter to evaluate PCa progression, compared with low Gleason score 6 PCa tissues (Fig. 5G). These results indicated that PELO expression was strongly elevated during PCa progression and associated with shorter survival of patients.
We next explore the biological function of PELO in PCa cells. The cell proliferation rate was significantly decreased when diminishing PELO protein level in PCa cell lines such as LNCaP, PC3 cells and Du145 cells (Fig. 5H, I and Supplementary Fig. S5C-E) even though knockdown of PELO in RWPE1 cells has no effect on cell growth ( Supplementary Fig. S5A-B). PC3 and Du145 cells A Coomassie blue staining of coimmunoprecipitated mixtures separated by SDS-PAGE; bands were cut from the gel as indicated, digested with trypsin, and subjected to LC/MS. B Molecular function categories for potential Smad4 interacting proteins.HEK293T cells were transfected with Myc-tagged Smad4 and Flag-tagged PELO expression plasmids as indicated, whole cell lysates were extracted and immunoprecipitated with anti-Flag antibodies (C) or anti-Myc antibodies (D). E Lysates of PC3 cells were immunoprecipitated with Smad4 antibody or normal rabbit IgG, and then subjected to immunoblot with PELO or Smad4 antibody. F HEK293T cells were transfected with V5-PELO and Flag-PLK1 expression plasmids as indicated, and whole cell lysates were extracted and immunoprecipitated with anti-Flag antibodies. G The schematic illustration of Smad4 and its truncated fragments were shown. Mapping the binding domain of Smad4 with PELO (H) or PLK1 (I). HEK293T cells were transiently co-transfected with the expression vectors as indicated. After 24 h, cell lysates were prepared and immunoprecipitations were performed with anti-Myc antibody. Immunoprecipitants were analyzed with anti-Flag antibody and anti-Myc antibody. J PC3 cells extracts were immunoprecipitated with Smad4 antibody, and the immunoprecipitated proteins were analyzed by Western blotting. belong to advanced PCa cell lines, knockdown of PELO expression in those two cell lines could strongly restrain cell migration and invasion ability (Fig. 5J, K and Supplementary Fig. S5E-J). Inhibition of PELO expression in PC3 cells could also apparently decrease cell colony formation ability ( Supplementary Fig. S5K-L). To investigate in vivo metastatic capacity of PELO, we further validate the impact of PELO on organ colonization ability of human prostate tumor cells in NSG mouse model. We injected the control PC3 cells and PELO deletion PC3 cells into 5 NSG mice through the tail veins, respectively. All mice were killed 40 days after cells injection, except for three mice in the control group that were killed earlier due to their poor condition. Histopathological analysis showed that the average size and amount of tumor lesions in the lungs and livers in mice injected with PC3-shPELO subline cells were significantly smaller and lower than in the control mice (Fig. 5L-N). Taken together, these data demonstrated that PELO deletion could strongly suppress cell growth and metastasis in PCa cells, and PELO might be a novel therapeutic target for the treatment of advanced PCa in clinical.
Knockdown of PELO enhanced the tumor suppressive function of PLK1 inhibitor RO To determine whether PELO could cooperate with PLK1 to regulate prostate cancer cell growth, PC3 cells were transfected with PELO shRNA lentivirus or control lentivirus, followed by treatment with RO and tested protein level of Smad4, cell proliferation and metastasis. The results showed that inhibiting the expression of PLK1 and PELO at the same time obviously led to the higher protein level of Smad4 compare with inhibition of the expression of PLK1 or PELO (Fig. 6A). As expected, either RO or knockdown of PELO in PC3 and Du145 cells reduced cell growth and metastasis, but knockdown of PELO in PC3 cells treated with RO significantly restrained cell growth (Fig. 6B, C) and metastasis (Fig. 6D-G) to a significant extent. These results suggested that PELO cooperates with PLK1 to enhance prostate cancer cell growth and metastasis.
Blocking the interaction between PELO and Smad4 by specific peptide inhibited PCa cell metastasis To determine which domain of PELO is required for binding Smad4, full-length PELO and deletion mutants of PELO were cloned into vector pCMV-FLAG, expressing Flag-tagged proteins. HEK293T cells were co-transfected with Myc tagged-Smad4 and PELO deletion mutants or control vector. Immunoprecipitations

of PELO reduces PCa cell proliferation measured by XTT colorimetric assays (mean ± SD of triplicate experiments) in LNCaP cells (H) and PC3 cells (I). Quantification analysis of migration (J) and invasion (K) assays in PC3 cells transfected with
PELO shRNA (PELO shRNA1, PELO shRNA2) or control shRNA. L Representative H&E images of lung and liver metastases in all five groups of mice at 5× magnification, respectively. Dotted circles represent metastatic lesions. Scale bars, 500 μm. Bar graph represents an average area in percentage of metastatic lesions within the lung (M) and liver (N) in the aforementioned groups, n = 5. *p < 0.05; **p < 0.01; ***p < 0.001. In H-K, M, N, P values were assessed using two-tailed Student's t tests. In F, G, P values were assessed using one-way ANOVA test. Fig. 6 Knockdown of PELO enhanced the tumor suppressive function of PLK1 inhibitor RO. A PC3 cells were transfected with PELO shRNA lentivirus or control lentivirus, followed by treatment with RO3280 and then analyzed by western blot. GAPDH was used as a loading control. B Cell proliferation assays in PC3 cells treated with RO3280 or stably transfected with PELO shRNA lentivirus as in (B). C Cell proliferation assays in Du145 cells treated with RO3280 or stably transfected with PELO shRNA lentivirus as in (C). Quantification analysis (D) and representative pictures (E) of migration assays in PC3 cells treated with RO3280 or stably transfected with PELO shRNA lentivirus as in (D, E). Quantification analysis (F) and representative pictures (G) of migration assays in Du145 cells treated with RO3280 or stably transfected with PELO shRNA lentivirus as in (E, F). *p < 0.05; **p < 0.01; ***p < 0.001. In B, C, D, F, P values were assessed using two-tailed Student's t tests.
were performed with anti-Flag antibody. As shown in (Fig. 7A), the 137-268 amino acids of PELO were required for Smad4 binding. To determine whether blocking the interaction between PELO and Smad4 influenced cell metastasis in vivo, we designed several peptides targeting the 137-268 amino acids of PELO and picked one peptide which could effectively block the interaction between PELO and Smad4 in PC3 cells (Fig. 7B). Interestingly, blocking the interaction between PELO and Smad4 apparently increased the protein level of Smad4 (Fig. 7C). To investigate whether the peptide suppresses tumor metastases in vivo, PC3 cells were injected into 15 NSG mice. 7 mice were injected with PBS and the others were injected with peptide through tail vein injection every two days. Mice will be killed when they were under poor conditions and calculated as death. There were no mice left in control group and 5 mice left in the peptide injection group at day 50 after the cell injection. We found that the surface of the lungs is smoother and the white spots at the surface of the liver are less in the mice injected with peptide compared with mice injected with PBS (Fig. 7D). We also observed that mice (n = 8) that were injected with peptide via tail vein at a dose of 10 mg/kg remarkably increased survival rate compared with mice were injected with PBS (n = 7) (Fig. 7E). Moreover, Histopathological analysis showed a remarkably decreased burden of tumors in lungs and livers in the mice treated with peptide compared with mice injected with PBS ( Fig. 7F-I). Taken together, these results indicated that the peptide injection could significantly prevent the tumor cells from metastasizing to the lung and liver in mice.

DISCUSSION
In Pten-deleted mice, high expression level of PLK1 induces cancer formation, whereas high expression level of Smad4 performs the role of tumor suppressor. However, the connections between those two important factors in prostate cancer initiation and progression have never been reported. In this study, we first deeply investigated the relationship between PLK1 and Smad4 thoroughly. Our results proved that PLK1 and PELO could directly bind to different domains of Smad4 and formed a protein complex. PLK1 promotes the ubiquitination and degradation of Smad4 by phosphorylating Smad4 at Thr197 in PCa cells with the help of PELO. The biological impact of PLK1 inhibitor on cell growth and metastasis could be restored by knockdown of Smad4. PLK1 and Smad4 are important factors for the initiation and progression in multiple cancer types [15][16][17]. PLK1 was found to act upstream of PI3K/AKT/mTOR signaling pathway [4], whereas another study proved that Smad4 could deactivate the PI3K/mTOR pathway by interacting with the P85 subunit of PI3K in ovarian cancer [18]. PLK1 and Smad4 oppositely regulate MYC activity [9,10]. Moreover, the Smad4 mutated HNSCC cell lines have shown the resistance characteristics for PLK1 inhibitors [19]. All these previous studies suggested that there is an antagonistic function between Smad4 and PLK1 and build strong research basis for our present study. Here, our findings further unveiled the reason of why Smad4 mutated cells have shown resistance to PLK1 inhibitor. We provided strong evidence to support that PLK1 could indeed induce the ubiquitination and degradation of Smad4 with the help of PELO. The degradation of Smad4 should be crucial for PLK1 to switch on its downstream events. Our present study supplies more clues for applying the PLK1 inhibitors in clinical trials.
The cellular role of TGF-β/Smad signaling pathway is a paradox in cancer. The TGF-β exerts tumor suppressor function in early stages, whereas it enhances tumor invasiveness and metastasis in the late stages of cancer. The molecular mechanism of this duality in tumor development is still unclear. Smad4, the tumor suppressor in prostate cancer, is one of the important mediators in the TGF-β signaling pathway. The phosphorylation of Smad4 at different sites has different effects on the activity of TGF-β signaling pathway. One study reported that the phosphorylation of Smad4 at Thr276 could increase TGF-β-induced nuclear accumulation and transcriptional activity of Smad4 [20]. Another study found that Wnt could prevent the phosphorylation of Smad4 at Thr265, Thr269, Thr273 and potentiate TGF-β signaling when in presence of FGF [21]. In this study, we found that PLK1 could phosphorylate Smad4 at Thr197, promote the ubiquitination and subsequent proteasomal degradation of Smad4, leading to prostate cancer progression by inhibiting the expression of E-Cadherin and increasing the expression of N-Cadherin, Vimentin, MMP9 and CCND1. However, the impact of phosphorylation of Smad4 at Thr197 on the activity of TGF-β signaling pathway needs to be further investigated.
In this study, we also explored the role of PELO in PCa initiation and progression for the first time. In contrast to the biological role in breast cancer, PELO is positively associated with advanced PCa. We further proved that PELO was highly expressed in prostate cancer tissues, especially in metastatic tumors. Importantly, knockdown of PELO markedly decreases the metastasis ability of advanced PCa cells in vitro and vivo. Blocking the interaction of PELO and Smad4 could prevent the degradation of Smad4 and inhibit cell metastasis in vitro and in vivo. Interestingly, unlike in tumor cells, depletion of PELO expression in normal prostate epithelial cells had no impact on cell proliferation. Recent studies indicated that the inactivation of PELO in mammalian tissues disrupts the translational control and activates mTOR signaling to compensate for impaired or reduced availability of ribosomes [22]. These findings suggested that the signaling pathways of PELO in normal and tumor cells are quite different. Exploring the role of PELO during tumor initiation will provide us more clues to understand the transformation of cells from normal to tumor status.
In the last decades, inhibitors and peptide therapeutics developed fast in clinical trials and pharmaceutical market [23][24][25]. Several PLK1 inhibitors have been developed and applied in clinical trials [26,27]. However, preclinical success with PLK1 inhibitors has not translated well into clinical success, the specificity of the inhibitors probably still need to be improved [28,29]. Overall, our studies pinpointed a novel mechanism of Smad4 degradation in prostate cancer and proved that a combination of PLK1 inhibitor with inhibition of PELO could synergistically enhance the effect of inhibiting tumor cell growth and metastasis. All these findings provided a potential therapeutic strategy for advanced PCa treatment by co-targeting PLK1 and PELO in future studies.

MATERIALS AND METHODS Cell culture
All cell lines were originally purchased from the American Type Culture Collection (ATCC) and none of these cell lines were found to be contaminated with mycoplasma during our study. HEK293T cells were grown in DMEM (Invitrogen, catalog no. 10569010); RWPE-1 cells were cultured with keratinocyte serum-free medium (Invitrogen, catalog no. 10724011) with bovine pituitary extract and human recombinant EGF; LNCaP and DU145 cells were grown in RPMI1640 (Merck, catalog no. R8758); PC3 cells were cultured in F-12 Kaighn's Modification (HyClone, catalog no. SH30526.01). All media were supplemented with 10% fetal bovine serum (Gibco, catalog no. 10099-141C) and with antibiotics (Thermo Fisher Scientific, catalog no. 15140122).

Immunoprecipitation
Immunoprecipitation was performed as described previously [30]. Briefly, for transfection-based co-immunoprecipitation assays, HEK293T cells were transfected with the indicated plasmids using Lipofectamine 2000 (Invitrogen, catalog no. 11668019) lysed and immunoprecipitated with anti-FLAG antibody (Sigma-Aldrich, catalog no. F3165) or anti-MYC antibody (Thermo Fisher Scientific, catalog no. MA1-980) combined Fig. 7 Blocking the interaction between PELO and Smad4 by specific peptide markedly inhibited PCa cell metastasis. A HEK293T cells were co-transfected with Smad4 and the deletion mutants of PELO with Flag tag. After 24 h of transfection, cell lysates were prepared and immunoprecipitations were performed with anti-Flag antibody. Immunoprecipitants were analyzed with anti-Flag antibody and anti-Myc antibody. The schematic illustration of PELO and its truncated fragments were shown in the right upper panel. B Peptide localized at PELO-2 domain could block the interaction between Smad4 and PELO in PC3 cells. C PC3 cells were treated with Peptide and the protein level of Smad4 was analyzed by western blot. D Mice (n = 8) that were injected with peptide via tail vein at a dose of 10 mg/kg significantly suppressed lung and liver metastases compared with mice were injected with PBS (n = 7). E Mice (n = 8) that were injected with peptide via tail vein at a dose of 10 mg/kg remarkably reduced survival rate compared with mice injected with PBS (n = 7). Survival Differences were evaluated with the log-rank test. F Representative H&E images of lung metastases in two groups of mice at 5× magnification. Dotted circles represent metastatic lesions. Scale bars, 500 μm. G Bar graph represents an average area in percentage of metastatic lesions within the lung in the aforementioned groups, peptide (−) group: n = 7; peptide (+) group: n = 8. H Representative H&E images of liver metastases in two groups of mice at 5× magnification. Dotted circles represent metastatic lesions. Scale bars, 500 μm. I, Bar graph represents an average area in percentage of metastatic lesions in liver in the aforementioned groups, peptide (−) group: n = 7; peptide (+) group: n = 8. **p < 0.01. In G, I, P values were assessed using two-tailed Student's t tests.
agarose beads, Cells lysed and immunoprecipitated with anti-Smad4 antibody (Abclonal, catalog no. A19116) combined agarose beads for 12 h at 4°C. The beads were washed three times with the lysis buffer and eluted in SDS sample loading buffer. The eluted proteins were separated by SDS-PAGE, transferred to PVDF membranes (Millipore, catalog no. ISEQ00010) and detected in western blots with appropriate primary antibodies coupled with HRP-conjugated secondary antibody by chemiluminescence (Pierce, catalog no. 32106). For detection of the interaction of endogenous proteins, whole cell extracts were immunoprecipitated with specific antibodies overnight at 4°C. Normal IgG was used as a negative control. After extensive washes, the immunoprecipitates were resolved with SDS sample loading buffer for Western blot analysis.

Ubiquitination of Smad4
HEK293T cells were transfected with the indicated constructs. 36 h after transfection, cells were harvested and lysed in lysis buffer. Anti-HA (ABclonal: AE008) antibody was used to detect polyubiquitinated Smad4. For detecting the endogenous ubiquitination of Smad4, PLK1 overexpression in PC3 cells were lysed in lysis buffer. Then, the ubiquitination of Smad4 was analyzed by immunoprecipitated with anti-Smad4 (ABclonal: A19116) antibody and immunoblotted with anti-Ub (ABclonal: A19686) antibody or anti-Smad4 antibody.

Cycloheximide (CHX) chase assay
Cells were incubated in 50 μg/ml cycloheximide (Cell Signaling Technology, catalog no. 2112) for indicated times. Then, cells were harvested and lysed in 2 × Loading sample buffer. Protein level was determined by western blot analysis.

Immunohistochemistry (IHC) assays
To explore the expression of PELO in PCa tissues and adjacent tissues, two PCa tissue chips (Catalog No. HProA 150CS01, HProA150PG02) were purchased from Outdo Biotech (Shanghai, China). We performed the experiment according to the protocols approved by the ethical committee of Shaanxi Normal University (2021-030). We first deparaffinized and rehydrated the tissue chips, retrieval antigen 5 min at 121°C, blocked endogenous enzyme activity 30 min at room temperature using 3% H 2 O 2methanol, blocked the chips 1 h with 5% BSA. We incubated the chips with the PELO primary antibody (Abcam, catalog no. ab140615) (diluted at 1:50) overnight at 4°C. Then we incubated the tissue chips with EnVision-HRP secondary antibody for 1 h, the signals were detected by diaminobenzidine followed by hematoxylin counterstaining. Aperio ScanScope slide scanner was used for viewing the signals. Of note, the result was assessed according to the intensity of staining (0, 1+, 2+, and 3+) and the percentage of positive cells (0 (0%), 1 (1-25%), 2 (26-50%), 3 (51-75%), and 4 (76-100%)) by two experienced pathologists. Finally, the final staining score was calculated and analyzed.

Xenograft tumor model and in vivo metastasis analysis
30 male athymic nude mice at 4 weeks of age were used according to the protocols approved by the ethical committee of Shaanxi Normal University (2021-030). Briefly, for in vivo metastasis analysis, male nude mice (randomly choose 5 mice as one group) were injected i.v. via the tail vein with 1 × 10 6 PC3 sublines (control shRNA, shPELO-1 and shPELO-2). Six weeks later, nude mice were sacrificed, and the lung and liver metastases were examined by H&E staining. For the peptide experiment,

RNA isolation and real-time PCR
Total RNA was extracted using RNeasy Mini kit (Tiangen, catalog no. DP430) according to the manufacturer's instructions. First strand cDNA was synthesized using 1 ug of total RNA and Superscript II (Thermo Fisher Scientific, catalog no. 4368814). Real-time quantitative PCR was performed in triplicates with SYBR Green qPCR mix (Thermo Fisher Scientific, catalog no. 4472908). The relative amount of specific mRNA was normalized to GAPDH.

Cell viability and proliferation assays
Cells treated with PELO shRNA or control shRNA lentivirus (1×10 3 per well) were seeded in 96-well plates. Cell viability and proliferation were determined with XTT assays (Roche, catalog no. 11465015001) at designed time points by measuring the absorbance at 450 nm, following the manufacturer's instructions. Values were obtained from three replicate wells for each treatment and time point. The results are representative of three independent experiments. Significance was calculated by a twotailed t test.

Transwell invasion and migration assay
Standard 24-well Boyden invasion chambers (BD Biosciences) were used to assess cell invasiveness following the manufacturer's suggestions. Briefly, cells were trypsinized, rinsed twice with PBS, resuspended in serum-free medium at 2.5 × 10 5 cells/ml. 200 µl of cell suspension was transferred into 8-μm Transwell inserts (Corning Costar, catalog no. 3422) with (invasion) or without (migration) 100 µl Matrigel (diluted with serum free medium to 250µg/ml) coating (BD Biosciences, catalog no.356230). 10% serumcontaining medium was added to the lower chambers. Following 48 h incubation, chambers were fixed in 3.7 % formaldehyde, after permeabilized with methanol, stained with Wright-Giemsa (Sigma-Aldrich, catalog no. WG16) for manual counting. A two-tailed t-test was employed to perform statistical analysis from three replicate inserts.

Lentiviral constructs, lentivirus production and infection
The shRNA constructs targeting PELO were ordered from Merck or designed using the online tool (https://rnaidesigner.thermofisher.com/ rnaiexpress/design.do). In detail, HEK293T cells were trypsinized and seeded into 3.5-cm plates with normal DMEM medium containing 10 % FBS, 1 % penicillin and streptomycin. The normal DMEM medium was replaced with 2 ml low glucose DMEM (Invitrogen, catalog no. 21885025) before transfection. For second generation lentiviral production, 1.5 μg shRNA construct, 0.375 μg pMD2.G (envelope plasmid), and 1.125 μg psPAX2 (packaging plasmid) were diluted to 100 ul using Opti-MEM (Invitrogen, catalog no. 11058-021) and combined with 100 ul Opti-MEM containing 8 µl Lipofectamine 2000 (Invitrogen, catalog no. 11668019). This mixture was incubated at room temperature for 20 min and added to the 2 ml medium covering the 293T cells. The medium was changed to fresh medium after 24 h and the virus-containing medium was harvested every 24 h up to three times. Lentivirus was passed through 0.45 μm filter unit and stored at −80°C. For viral transduction, Viral supernatants containing 8 μg/ml polybrene (Merck, catalog no. H9268) were added to PCa cells that were seeded 24 h before infection at 70-80% confluence. For lentivirusmediated knockdown experiment, virus was removed and replaced by normal medium with antibiotics after 24 h, after which cells were selected with 1 μg/ml puromycin (Merck, catalog no. P9620). When uninfected control cells were completely killed, the target cells were cultured in normal growth medium with 0.5 μg/ml puromycin for further experiments.
GST pull-down assay PELO full length was cloned in frame with the glutathione S-transferase (GST) gene of pGEX-6P-1 and the resulting GST fusion proteins were expressed in Escherichia coli BL21, induced by 0.1 mM isopropyl-b-Dthiogalactopyranoside (IPTG) and solubilized from bacteria in lysis buffer (1% Triton X-100 in PBS) by sonication. After centrifugation for 10 minutes at 4°C, the supernatant was incubated with the glutathione-Sepharose 4B beads (Thermo Fisher Scientific, catalog no. 16100) at 4°C overnight. Beads were harvested through centrifugation and washed three times in cold PBS to remove nonspecific binding. GST and GST-fusion proteins were bound to glutathione-Sepharose beads followed by SDS-PAGE. The PC3 cell lysates were treated with DNase I (Thermo Fisher Scientific, catalog no. 18047019) for 30 min at 37°C to remove the genomic DNA contamination before being mixed with GST fusion protein which adsorbed to Sepharose beads. The binding reaction was carried out overnight at 4°C in 1 mL binding buffer (20 mmol/L Tris-HCl at pH 8.0, 150 mmol/L NaCl, 1 mmol/L EDTA, 10% glycerol and 0.1% NP-40). After extensive washing, specifically bound proteins were subjected to 10% SDS-PAGE followed by western blot analysis.

Soft agar assays
About 400 PC3-control, PC3-shPELO-1 and PC3-shPELO-2 cells were suspended in 0.2 ml of prewarmed 0.35% SeaPlaque agarose/MEM/5% fetal bovine serum and pipetted on top of 0.3 ml of 0.7% SeaPlaque agarose/MEM/5% fetal bovine serum gelled onto a well of a 24-well plate, then 0.4ml medium was added on top of the cells. Cells were incubated for 15 days, at 37°C and 5% CO 2 , and fresh medium was changed every three days. The colony spots were counted through eye vision.

Statistical analysis
All statistical analyses were performed using GraphPad Prism version 5.0 software (GraphPad, San Diego, CA, USA) or RStudio (v. 1.2.5033) with R (v. 3.6.3). Data were obtained from the cBioPortal for Cancer Genomics [31,32], and GEO database [33,34]. Statistical tests for patients with gene expression were calculated using Mann-Whitney U test. Kaplan Meier survival analysis was used to assess the impact of gene expression levels on prostate cancer prognosis and survival. P value < 0.05 was considered to be statistically significant, all error bars in graphical data represent mean ± SD. Statistical tests and figures were generated in R (3.6.3)