FBXW8-dependent degradation of MRFAP1 in anaphase controls mitotic cell death

Mof4 family associated protein 1 (MRFAP1) is a 14 kDa nuclear protein, which involves in maintaining normal histone modification levels by negatively regulating recruitment of the NuA4 (nucleosome acetyltransferase of H4) histone acetyltransferase complex to chromatin. MRFAP1 has been identified as one of the most up-regulated proteins after NEDD8 (neural precursor cell expressed developmentally down-regulated 8) inhibition in multiple human cell lines. However, the biological function of MRFAP1 and the E3 ligase that targets MRFAP1 for destruction remain mysterious. Here we show, by using an immunoprecipitation-based proteomics screen, that MRFAP1 is an interactor of the F-box protein FBXW8. MRFAP1 is degraded by means of the ubiquitin ligase Cul7/FBXW8 during mitotic anaphase-telophase transition and accumulated in mitotic metaphase. Overexpression of FBXW8 increased the polyubiquitination and decreased the stability of MRFAP1, whereas knockdown of FBXW8 prolonged the half-life of MRFAP1. Moreover, forced expression of MRFAP1 in HeLa cells caused growth retardation and genomic instability, leading to severe mitotic cell death. Thus, Cul7/FBXW8-mediated destruction of MRFAP1 is a regulatory component monitoring the anaphase-telophase transition and preventing genomic instability.


INTRODUCTION
The specific, rapid, and temporally controlled proteolysis of cellular regulators by the ubiquitinproteasome system (UPS) determined the unidirectional progression through the cell cycle [1]. The specificity of this system is mainly determined by E3 ubiquitin ligases [2]. The Cullin-Ring Ligases (CRL) are the well-studied RING E3 ubiquitin ligases, among which the Skp1-Cullin1-Fbox (SCF) complex has been extensively characterized. To date, the Cullin family contains eight mammalian proteins, Cul1, Cul2, Cul3, Cul4A, Cul4B, Cul5, Cul7 and Cul9 [3].

Research Paper www.impactjournals.com/oncotarget
The SCF complexes compose of the scaffold protein Cul1, the RING finger protein Rbx1 and the adaptor protein Skp1 to bridge variable F-box proteins that determine substrate recognition and recruitment to the SCF complexes [4]. The human genome encodes more than 70 F-box proteins that contain a homologous 40-amino-acid F-box domain which can be further divided into three major classes: the FBXW family, the FBXL family, and the FBXO family [5]. Among these F-box proteins, FBXW8 is not only the receptor component of an SCF complex, but also the unique F-box protein known to interact with Cul7 [6]. Cul7 interacts directly with FBXW8 and indirectly with Skp1 in an FBXW8-dependent manner [6]. Interestingly, Cul7 has higher affinity with FBXW8 than Cul1, and the placental phenotypes of FBXW8 -/and Cul7 -/mice are similar, indicating that Cul7 forms a ubiquitin E3 ligase complex with FBXW8 to target substrates for destruction [7]. In agreement with that, Cul7/ FBXW8 ubiquitin ligase-mediated HPK1 (hematopoietic progenitor kinase-1) degradation played a critical role in cell proliferation and differentiation [8].
Mof4 family associated protein 1 (MRFAP1) is highly conserved and only presents in mammals, which is consistent with its specific role in the regulation of mortality factor 4 like 1 (MORF4L1) protein complexes. MRFAP1 maintains the normal histone modification levels by negatively regulating recruitment of the NuA4 histone acetyltransferase complex to chromatin [9]. It also regulates the transcription factors activities of the MRG family (Mas-related genes) via competing with MRGBP for binding to MORF4L1, and plays a role in spermatogenesis, possibly by regulating the hyperacetylation of chromatin on histone H4 [10]. Recently, by quantitative proteomic analysis, MRFAP1 has been shown to be degraded by CRL via the ubiquitin-proteasome system [10,11]. However, which CRL family member involved in this process remains unknown.
In this study, by using an immunoprecipitationbased proteomics screen, we identified that MRFAP1 was an interactor of FBXW8. MRFAP1 was a novel cell cycleregulated protein degraded by Cul7/FBXW8 ubiquitin ligase during mitotic anaphase-telophase transition and accumulated in mitotic metaphase. Failure to degrade MRFAP1 in HeLa cells caused growth retardation and genomic instability, leading to severe mitotic cell death.

Proteomics screen identified MRFAP1 as a potential binding partner of Cul7/FBXW8 E3 ligase complex
To systematically identify the candidate substrates of Cul7/FBXW8, we performed an immunoprecipitationbased proteomics screen in 293T cell stably expressing Flag-FBXW8. MLN4924, a potent and selective inhibitor of NEDD8-activating enzyme, which specifically disrupts CRLs-mediated protein turnover, was used to prevent the degradation of potential FBXW8 substrates and might enhance the protein-protein interactions with FBXW8. The cell lysate from Flag-control or Flag-FBXW8 293T cells with 4 hours MLN4924 treatment were subjected to immunoprecipitation with anti-Flag M2 resin to purify SCF-FBXW8 complexes. The whole IP procedure was monitored by running all separated cell lysate components on SDS-PAGE, followed by Coomassie Brilliant Blue staining ( Figure 1A). The gel was continuously cut and subjected to LC−MS/MS analysis. The results indicate that Cul7 but not Cul1 was the highest score interacting protein of FBXW8 group (data not shown) as expected. Among a number of peptides from putative novel substrates, we recovered two unique peptides from MRFAP1 (EDIASLTR and TQVEASEESALNHLQNPGDAAEGR) ( Figure 1B). The affinity efficiency for Flag-FBXW8 was very high, as almost all exogenously expressed FBXW8 was accumulated in the elution group ( Figure 1A). Moreover, MRFAP1 was detected in the final elution of Flag-FBXW8 group and flow through of both groups ( Figure 1A), suggesting that at least one part of endogenous MRFAP1 was specifically associated with FBXW8.

MRFAP1 is co-localized and associated with FBXW8
To determine the subcellular distribution of FBXW8, and further verify the spatio-temporal possibility that FBXW8 and MRFAP1 might interact with each other in mammalian cells in vivo, we utilized immunofluorescence microscopy on HeLa cells co-expression of Flag-FBXW8 and GFP-MRFAP1. As shown in Figure 2A, FBXW8 was distributed throughout the cytoplasm and less expressed in the nucleus of interphase cells. However, most MRFAP1 was distributed in the nucleus. Nevertheless, in agreement with our biochemistry data, the overlapping between FBXW8 and MRFAP1 in the nucleus was clearly observed ( Figure 2A). Moreover, the interaction between exogenous FBXW8 and MRFAP1 was also confirmed by immunoprecipitation with either GFP or Flag antibodies ( Figure 2B). Taken together, these data suggested that MRFAP1 was associated with FBXW8.

FBXW8 regulates MRFAP1 ubiquitination and degradation
Given the specific interaction between FBXW8 and MRFAP1, we then asked whether FBXW8 regulates the stability of MRFAP1. We found that depletion of FBXW8 by siRNAs consistently increased the basal expression level of endogenous MRFAP1 ( Figure 3A). Furthermore, overexpression of FBXW8 caused the decrease of endogenous MRFAP1 in a dose-dependent manner, which was rescued by proteasome inhibitor MG132 ( Figure 3B), indicating that FBXW8 regulated the expression of MRFAP1 at posttranscriptional level. Indeed, overexpression of FBXW8 significantly promotes the ubiquitination of MRFAP1, suggesting FBXW8 targets MRFAP1 for proteasomemediated degradation ( Figure 3C). Moreover, 293T cells were treated with cycloheximide (CHX) to block protein synthesis. We found that overexpression of FBXW8 decreased the halflife of MRFAP1 ( Figure 4A-4B), whereas knockdown of FBXW8 prolonged it ( Figure 4C-4D). In summary, these data demonstrate that FBXW8 regulates MRFAP1 ubiquitination and degradation.

Cell cycle-dependent degradation of MRFAP1
Because SCF E3 ligase mediates the ubiquitination of several proteins in specific phases of the cell cycle, we also analyzed the expression of MRFAP1 during the cell cycle. Firstly, we created a HeLa cell line stably expressing Flag-MRFAP1. These cells were synchronized by double thymidine arrest, released, and collected at various time points after release. In addition, nocodazole was added to the culture after the release from double thymidine arrest to activate the spindle checkpoint and prevent exit from mitosis. DNA contents of those cells were monitored by flow cytometry analysis (FACS), and lysates of these cells were tested by immunoblotting. As shown in Figure 5A, the protein level of MRFAP1 significantly increased after cells entering into mitosis. However, the protein level of FBXW8 remained unaltered. Interestingly, the increase of MRFAP1 protein level was even early than CyclinB1, which is known to be accumulated in early mitosis. In order to check how MRFAP1 was regulated when cells released from M phase, HeLa cells stably expressing Flag-MRFAP1, which were synchronized by nocodazole block-and-release, were analyzed by FACS and lysates of these cells were tested by immunoblotting ( Figure 5B). As expected, MRFAP1 was highly accumulated in mitosis. However, as cells exited from M phase, MRFAP1 decreased gradually ( Figure 5B). In line with these observations, by using immunofluorescence, we found that MRFAP1 was accumulated in metaphase, but completely disappeared in anaphase and reappeared in telophase  Figure 5C, top panel). However, silencing the expression of FBXW8 prevented the disappearance of MRFAP1 in anaphase ( Figure 5C, bottom panel). Taken together, the data validate that MRFAP1 is a novel cell cycleregulated protein and cell cycle-dependent degradation of MRFAP1was mediated by FBXW8.

Overexpression of MRFAP1 causes mitotic aberrations and cell death
Cell cycle is a precisely regulated process and cell cycle regulated-proteins usually play crucial roles in the regulation processes. Thus, the cell cycle-dependent  degradation of MRFAP1 by FBXW8 during mitosis intrigues us to further investigate its biological function in cell cycle control. Aberrant expression of cell cycle regulated-proteins could lead to genome instability, we therefore tested whether MRFAP1 was involved in the regulation of genome stability. As depicted in Figure 6A, overexpression of MRFAP1 in HeLa cells induced mitotic aberrations including binuclear cells, multi-lobed nuclei cells and cell death. In contrast, all these defects were hardly observed in GFP overexpressed HeLa cells. In agreement, HeLa cells with MRFAP1 overexpression exhibited a growth retardant phenotype ( Figure 6B). Indeed, HeLa cells with prolonged MRFAP1 expression showed severe cell death by the morphological confirmation, FACS assay determination as well as the activation and cleavage of caspase-3 ( Figure 6C-6F), indicating that failure to degrade MRFAP1 in cells produced chromosome instability and eventually lead to cell death.

DISCUSSION
The identification of substrates for ubiquitin E3 ligases by proteomic assay is extremely difficult, as most substrates are rapidly ubiquitylated and degraded as a result of their association with E3 ligase. The administration of proteasome inhibitors such as MG132 prevented the degradation of substrates might help to improve the identification of substrates of E3 ligases, but without sufficient specificity. MLN4924, a potent and selective inhibitor of NEDD8-activating enzyme, specifically disrupts CRLs-mediated protein turnover [12]. Previous studies showed that substrates of CRLs were highly accumulated upon MLN4924 administration, giving the possibility to identify the substrates of CRLs by Mass Spectrum (MS) [13][14][15]. In this study, we performed an immunoprecipitation-based proteomics screen with MLN4924 treatment to identify the candidate substrates of Cul7/FBXW8 and demonstrated MRFAP1 as a bona fide substrate of FBXW8.
After MRFAP1 has been shown to be one of the most up-regulated proteins after NEDD8 inhibition in multiple human cell lines [10], we tested whether MRFAP1 was regulated by FBXW8. The association between MRFAP1 and FBXW8 was confirmed by means of immunoprecipitation-Western blot as well as exogenous indirect immunostaining. Overexpression of FBXW8 increased the polyubiquitination and decreased the stability of MRFAP1, whereas knockdown of FBXW8 prolonged the half-life of MRFAP1, demonstrating that MRFAP1 is the bone fide substrate of Cul7/FBXW8. Substrates of F-box proteins usually involved in cell cycle control and FBXW8 also played a role in cell proliferation, we then examined the expression level of MRFAP1 in cell cycle progress. Interestingly, we found MRFAP1 was a novel cell cycle-regulated protein that accumulated in metaphase, but disappeared thoroughly in anaphase and reappeared in telophase. Importantly, this disappearance could be abolished by FBXW8-depletion, suggesting that FBXW8-mediated MRFAP1 degradation might play a role in the progression of normal cell cycle. Indeed, forced  expression of MRFAP1 in HeLa cells caused growth retardation and genomic instability, and eventually leading to severe mitotic cell death.
Although F-box proteins tend to recognize specific degron sequences within target substrates, prior modifications of these degrons are usually required to trigger the interaction between F-box proteins and their substrate proteins, such as phosphorylation, glycosylation and acetylation [4,16]. To this end, we have also purified MRFAP1 protein and determined the potential phosphorylation sites of it by MS. However, we could not identify any phosphorylation sites (data not shown) and MRFAP1 itself has not been reported to be phosphorylated even in public phosphorylation database, suggesting phosphorylation might not be required for the interaction between MRFAP1 and FBXW8. Nevertheless, further studies are required to clarify the interacting details between MRFAP1 and FBXW8.
Although we have confirmed that MRFAP1 overexpression resulted in decreased cell growth and increased mitotic apoptosis, whether MRFAP1 contributes to DNA rereplication, DNA damage signaling, senescence or autophagy (a phenomena could be observed after MLN4924 treatment) remained completely unknown [17]. Thus, our data has set up the stage for the MRFAP1 function, physiological and pathophysiological significances related researches.
In summary, in this manuscript, by using an immunoprecipitation-based proteomics screen, we identified MRFAP1 was a novel cell cycle-regulated protein and Cul7/FBXW8-mediated destruction of MRFAP1 could be a regulatory component monitoring the anaphasetelophase transition and preventing genomic instability.

CCK8 assay
HeLa cells were inoculated in a 96-well plate 24 hours before transfection. CCK8 (cell counting kit C0038, Beyotime Biotechnology, China) solution was added to each well of the plate 24 hours after transfection. The plate was then incubated for 4 hours in an incubator. The absorbance at 450 nm was measured by using a microplate reader (SpectraMax Paradigm, Molecular Devices, USA). siRNAs FBXW8-specific siRNAs were purchased from RiboBio (Guangzhou, China). The siRNA sequences for human FBXW8 were: No1 5′ GGAGCAUGUUC CUGACACA 3′, No2 5′ GAAGCAAGAUCCUGGUGUA 3′, No3 5′ CAGGAAAGAACUAGGAAGA 3′. 293T-MRFAP1 cell line was plated in a 24-well tissue culture plate for 20 to 24 hours before transfection. For transfection of siRNAs, cells were transfected with 100 nM of siRNA using 1 μL Lipofectamine ® 2000 for each well. Cells were then harvested 48 hours after transfection.

Protein purification, in-gel digestion and Nano-LC−ESI−MS/MS analysis
The whole IP/MS analysis procedure has been described previously [19].

Immunofluorescence microscopy
Cells were cultured on glass coverslips as described above and then fixed with 4% paraformaldehyde in PBS. Fixed cells were washed with PBS and then permeabilized using 1 % Triton X-100 for 10 min followed by washing with PBS. Coverslips were processed for immunolabeling by blocking with 5% BSA in TBST. Cells were incubated with primary antibodies in 5% BSA in TBST on coverslips for 1 hour. The coverslips were washed with PBS for three times. Primary antibodies were detected with Alexa Fluor 488 or Alexa Fluor 594-conjugated secondary antibodies by incubating on coverslips for 30 min in 5% BSA in TBST. Paraformaldehyde, Triton X-100, and BSA were obtained from sinopharm chemical reagent co.,Ltd (China). Alexa Fluor 488 or Alexa Fluor 594-conjugated secondary antibodies were obtained from Invitrogen (U.S.A). Anti-FLAG antibody (1:1000 dilution, F1804) was purchased from Sigma, U.S.A, Polyclonal anti-MRFAP1 (1:500 dilution, 11639-1-AP) was purchased from Ptglab, Wuhan, Hubei, China, Polyclonal anti-FBXW8 (1:500 dilution, sc-167864) was obtained from Santa cruz, U.S.A. Images were captured by fluorescence microscopy (Bx53, Olympus, Japan).

Western blot
Cells were harvested and lysed with ice-cold lysis buffer (62.5 mM Tris-HCl, pH 6.8, 100 mM DTT, 2% SDS, 10 % glycerol). After centrifugation at 12,000 g for 10 min at 4°C, proteins in the supernatants were quantified and separated by 10

Cell cycle synchronization
For nocodazole-thymidine cell cycle synchronization, HeLa cells stably expressing Flag-MRFAP1 were inoculated into a 6-well plate, and treated with 2 mM thymidine for 24 hours, then washed with PBS for 3 times. 0.3 mM nocodazole was added to the culture and the cells were collected at the indicated time for FACS or Western blot analysis. For nocodazole release cell cycle synchronization, HeLa cells stably expressing Flag-MRFAP1 were inoculated into the 6-well plate, and treated with 2 mM thymidine for 12 hours, then washed with PBS for 3 times. After 3 hours, cells were treated with 0.3 mM nocodazole for 12 hours, then washed with PBS for 3 times. After release from the nocodazole block, the cells were collected at the indicated time for FACS or Western blot analysis.

FACS assay
The whole FACS analysis procedure has been described previously [20].

Statistical analysis
The Statistical Package for the Social Sciences software version 15.0.1 (SSPS 15.0.1.) was used for statistical analysis. Values were shown as mean ± SEM. Statistical differences were determined by a Student t test. Statistical significance is displayed as * p < 0.05, ** p < 0.01 or *** p < 0.001.