Skp-cullin-F box E3 ligase component FBXL2 ubiquitinates Aurora B to inhibit tumorigenesis

Abstract

Aurora B kinase is an integral regulator of cytokinesis, as it stabilizes the intercellular canal within the midbody to ensure proper chromosomal segregation during cell division. Here we identified that the ubiquitin E3 ligase complex SCF(FBXL2) mediates Aurora B ubiquitination and degradation within the midbody, which is sufficient to induce mitotic arrest and apoptosis. Three molecular acceptor sites (K¹⁰², K¹⁰³ and K²⁰⁷) within Aurora B protein were identified as important sites for its ubiquitination. A triple Lys mutant of Aurora B (K¹⁰²/¹⁰³/(²⁰⁷R)) exhibited optimal resistance to SCF(FBXL2)-directed polyubiquitination, and overexpression of this variant resulted in a significant delay in anaphase onset, resulting in apoptosis. A unique small molecule F-box/LRR-repeat protein 2 (FBXL2) activator, BC-1258, stabilized and increased levels of FBXL2 protein that promoted Aurora B degradation, resulting in tetraploidy, mitotic arrest and apoptosis of tumorigenic cells, and profoundly inhibiting tumor formation in athymic nude mice. These findings uncover a new proteolytic mechanism targeting a key regulator of cell replication that may serve as a basis for chemotherapeutic intervention in neoplasia.

Figure 1

Ectopic expression of FBXL2 specifically induces Aurora B degradation. (a–c) MLE cells (2 × 10⁵) were plated on 35 mm glass bottom tissue culture dishes for 48 h, cells were then transfected with control plasmid lacZ or other V5-tagged F-box plasmids (4 μg/dish), including FBXL (L) (a), FBXW (W) (b) and FBXO (O) (c) family members. Cells were collected and cell lysates were analyzed for V5, Aurora B and β-actin immunoblotting (n=2 experiments). (d) Cells were transfected with increasing amounts of FBXL2 plasmid. Cells were collected and cell lysates were analyzed for V5, Aurora B and β-actin immunoblotting (n=2 experiments). After 24 h, cell lysates were collected and processed for Aurora B, and β-actin immunoblotting. (e) Aurora B protein half-life determination after FBXL2 overexpression (upper panel) or a control (CON) plasmid, or FBXL2 knockdown using siRNA (middle lower panel) or a control, scrambled (CON) RNA (n=2 experiments). In overexpression studies, MLE cells were transfected with a plasmid encoding FBXL2 or a Lacz plasmid (4 μg/dish) and 24 h later cells were collected for V5, Aurora B and β-actin immunoblotting. For siRNA studies, 1 × 10⁶ MLE cells were transfected using Lipofectamine 2000 with 10 μg of scrambled RNA or Aurora B siRNA and collected after an additional 48 h. Cell lysates were used for V5, Aurora B and β-actin immunoblotting. Lower graph shows levels of Aurora B quantified densitometrically based on immunoblots

Figure 2

Ectopically expressed FBXL2 depletes Aurora B within the midbody during mitosis and causes tetraploidy. (a) MLE cells (2 × 10⁵) were plated on 35 mM glass bottom tissue culture dishes for 48 h, cells were then washed with PBS and fixed with 4% paraformaldehyde for 20 min. Cells were co-immunostained for FBXL2 and Aurora B. Nuclei were counterstained using DAPI. Green: FBXL2, red: Aurora B, blue: DAPI. White scale bar indicate 10 μm. (b) Cells were transfected with either vehicle or FBXL2 plasmid on 35 mM glass bottom tissue culture dishes for 48 h. Cells were then washed with PBS and fixed with 4% paraformaldehyde for 20 min. Cells were co-immunostained for FBXL2 and Aurora B. Nuclei were counterstained using DAPI. Green: FBXL2, red: Aurora B, blue: DAPI. (c) Cells were transfected with either control or FBXL2 shRNA on 35 mm glass bottom tissue culture dishes for 48 h. Cells were then washed with PBS and fixed with 4% paraformaldehyde for 20 min. Cells were co-immunostained for FBXL2 and Aurora B. Nuclei were counterstained using DAPI. Green: FBXL2, red: Aurora B, blue: DAPI. (d) FBXL2 ectopic expression versus a control (CON) plasmid leads to furrow regression (2 h) in MLE cells expressing H2B-mCherry and MyrPalm-mEGFP. (e) Multi-nucleate cells in c were quantified and graphed (n=150 cells, *P<0.05 versus control)

Figure 3

Ectopically expressed FBXL2 depletes Aurora B, causes G2/M arrest and inhibits tumorigenesis. (a–b) MLE cells were transfected with increasing amounts of FBXL2 plasmid, transfected cells were processed by BrdU uptake and 7-AAD staining, followed by FACS cell cycle analysis (a), 2N, 4N and 8N DNA histograms were quantitated and graphed in b (n=2 experiments). (c) Quantification of FACS analysis showing levels of apoptotic human adenocarcinoma (A549) cells after FBXL2 or control (CON) plasmid overexpression, (n=3 experiments, *P<0.01 versus empty). (d) Effect of stable expression of FBXL2 plasmid or a control vector (CON) on growth of A549 tumor implants in nude mice, n=4 mice/group. The upper panel showed representative images of variable sizes of xenografts in three nude mice after expression of an empty vector or FBXL2 plasmid (arrows). Lower panel: tumors from four controls and four FBXL2 treated A549 implants in mice were collected at the end point, and assayed for Aurora B and FBXL2 proteins by immunoblotting

Figure 4

SCF^FBXL2 targets Aurora B for polyubiquitination. (a) A549 cells were collected, followed by immunoprecipitation of endogenous Aurora B and then immunoblotting for several E3 ligase subunits. Immunoprecipitated proteins normalized by input were calculated and graphed (right panel). (b) In vitro ubiquitination assays. Purified SCF complexes were incubated with V5-Aurora B and the full complement of ubiquitination reaction components. The reaction products were then used for Aurora B immunoblotting. (c) Diagram of Aurora B deletion mutants constructed. Red lines indicate putative ubiquitin receptor sites or potential residues important for FBXL2 binding. (d) In vitro ubiquitination assays. Purified SCF complexes were incubated with WT V5-Aurora B or various Aurora B deletion mutants and the full complement of ubiquitination reaction components. The reaction products were then used for immunoblotting. (e) In vitro ubiquitination assays. Purified SCF complexes were incubated with Aurora B point mutants and the full complement of ubiquitination reaction components. The reaction products were processed for immunoblotting as above. (f) Aurora B protein half-life determination after the expression of WT V5-Aurora B or point mutants (n=2 experiments). For half-life studies, cells were treated with cycloheximide (40 μg/ml) at different time points after transfection of WT or point-mutant Aurora B plasmids. Cell lysates were collected and processed for Aurora B and β-actin immunoblotting. Below are decay curves, graphically showing levels of WT or mutant proteins over time after bands were quantitated on immunoblots

Figure 5

Expression of Aurora B stable mutant-induced apoptosis. (a) MLE cells were transfected with a control empty plasmid and plasmids encoding WT or a triple Aurora B mutant (3K/R) in cells co-expressing mCherry-tagged histone H2B and MyrPalm-mEGFP as markers to assess cytokinesis. Expression of an Aurora B triple K/R mutant led to delays in anaphase onset in cells expressing H2B-mCherry and MyrPalm-mEGFP. (b) Quantitative analysis of mitosis in MLE cells of videos taken in G (n=15 cells in each condition). (c) Quantification of metaphase time in b. (d) Quantification of FACS analysis showing levels of apoptotic cells after WT or Aurora B triple K/R mutant overexpression, (n=3 experiments, *P<0.01 versus control)

Figure 6

BC-1258 induces apoptosis in leukemic cells by causing mitotic arrest and tetraploidy. (a) Peripheral blood mononuclear cells (PBMCs) from five controls, and AML and ALL subjects were cultured in RPMI medium for 18 h. Cells were then collected, lysed and assayed for FBXL2 and Aurora B by immunoblotting. (b) Structure of BC-1258. (c–e) Human leukemia cells (U937, K562 and THP1 cells) were treated with BC-1258 at different concentrations for 16 h. Cells were collected and assayed for Aurora B, cyclin D2, cyclin D3, β-actin and FBXL2 immunoblotting. (f–h) MLE cells were treated with BC-1258 at different concentrations for 16 h, cells were processed by BrdU uptake and 7-AAD staining, followed by FACS cell cycle analysis (f), 2N, 4N and 8N DNA histograms were quantitated and graphed in g. (h) Quantification of FACS analysis showing levels of apoptotic MLE cells after BC-1258 treatment at each time point. *P<0.05

Figure 7

BC-1258 inhibits tumorigenesis. (a–e) Effect of BC-1258 and related FBXL2 inducers on growth of U937 tumor implants in nude mice (n=4 mice/group) with drug concentrations at 30 μg/ml in the drinking water. (a) Representative images of variable sizes of xenograft in three nude mice (arrows) after drug treatment. (b) Tumor volume measurements over time (n=4 mice/group, *P<0.05 versus con). (c–d) Tumors were imaged, weighed and graphed (n=4 mice/group, *P<0.05 versus control). (e) Tumors from three controls and three drug-treated U937 implants in mice were collected at the end point, and were assayed for Aurora B, cyclin D2 and FBXL2 proteins by immunoblotting. (f–i) Serum of each mouse were collected at the end point and processed for creatinine, lactate dehydrogenase (LDH), alanine aminotransferase (ALT) and creatine kinase activity