AMBRA1 phosphorylation by CDK1 and PLK1 regulates mitotic spindle orientation

Abstract

AMBRA1 is a crucial factor for nervous system development, and its function has been mainly associated with autophagy. It has been also linked to cell proliferation control, through its ability to regulate c-Myc and D-type cyclins protein levels, thus regulating G1-S transition. However, it remains still unknown whether AMBRA1 is differentially regulated during the cell cycle, and if this pro-autophagy protein exerts a direct role in controlling mitosis too. Here we show that AMBRA1 is phosphorylated during mitosis on multiple sites by CDK1 and PLK1, two mitotic kinases. Moreover, we demonstrate that AMBRA1 phosphorylation at mitosis is required for a proper spindle function and orientation, driven by NUMA1 protein. Indeed, we show that the localization and/or dynamics of NUMA1 are strictly dependent on AMBRA1 presence, phosphorylation and binding ability. Since spindle orientation is critical for tissue morphogenesis and differentiation, our findings could account for an additional role of AMBRA1 in development and cancer ontogenesis.

Keywords: Cell cycle; Mitotic kinases; Mitotic spindle; NUMA1; Phosphorylation.

Fig. 1

AMBRA1 is phosphorylated during mitosis at sites T1209 and S1223. A–D WB analysis of protein extracts from: A HeLa cells synchronized at G1/S boundary with a double Thymidine block (dTB), and released in the presence of 100 ng/mL Nocodazole, to arrest cells in mitosis; B HeLa cells synchronized at mitosis using 200 ng/mL Nocodazole; C HeLa cells in naturally-occurring mitosis after mitotic shake off; D HeLa cells treated with 200 ng/mL Nocodazole, and released. E HeLa cells synchronized at mitosis with Nocodazole, followed by AMBRA1 IP and in vitro phosphatase assay with lambda protein phosphatase (λPP) on immunoprecipitated protein; phosphatase inhibitors (PI) were used as a control. F HeLa cells transfected with MYC-AMBRA1 and then treated with Nocodazole. In this case protein extracts were separated by Phos-Tag SDS-PAGE and then analyzed by WB. G–J WB on protein extracts from: G, H HeLa cells transfected with MYC-AMBRA1 fragments and treated with Nocodazole. MYC-AMBRA1 is showed at a lower (up) and an higher (bottom) exposure. I HeLa cells transfected with MYC-AMBRA1 full-length (FL) or R1161 and R1195 truncated fragments, followed by treatment with Nocodazole. J HeLa cells transfected with MYC-AMBRA1 wild-type (WT) or phosphosilent (Alanine substitution) for T1201, S1203, T1209, S1223 and T1209/S1223 (AA^1209/1223). Then cells were treated with Nocodazole. Cell cycle phases are specified for each lane in panels A, B, C and D (asyn. = “asynchronous”; interph. = “interphase”). Schematic representation of truncated fragments respect to full-length protein is provided above WB of panels G, H and I. (FL = “full-length”, F1 = “fragment 1”, F2 = “fragment 2”, F3 = “fragment 3”, F3A = “fragment 3A”, F3B = “fragment 3B”, WT = “wild-type”). WD40 domains (51–90, 93–133, 135–175) are highlighted in dark grey. The white arrow indicates AMBRA1 electrophoretic migration, while the black arrow indicates its mobility shift. AMBRA1 hypershift was visualized with low percentage acrylamide gels (5–6%) or with gradient pre-cast gel for blots in G and H. An asterisk marks a MYC-AMBRA1 degradation sub-product. Gel percentages are indicated in each WB panel

Fig. 2

CDK1 phosphorylates AMBRA1 at T1209 and S1223. A Schematic representation of AMBRA1 protein, highlighting T1209 and S1223 with their respective surrounding sequence. CDK1 consensus is underlined. WD40 domains (51–90, 93–133, 135–175) are highlighted in dark grey. B, C WB of immunoprecipitated proteins following Nocodazole treatment, with mouse immunoglobulins used as control (IP ctr): B Endogenous proteins immunoprecipitated with anti-AMBRA1 antibody. C Overexpressed MYC-AMBRA1 immunoprecipitated using anti-MYC antibody. D WB of HeLa cells treated with Nocodazole and then with 9 μM RO-3306, in the last 10 min of treatment. The white arrow indicates AMBRA1 electrophoretic migration, while the black arrow indicates its mobility shift. AMBRA1 hypershift was visualized with a low percentage acrylamide gel (5–6%). E Autoradiography (³²P signal) and WB of an in vitro kinase assay with recombinant CDK1/Cyclin B1 performed on MYC-AMBRA1 WT or AA^1209/1223 and MYC-β-Galactosidase (βGal) as control. Proteins were immunoprecipitated using anti-MYC antibody following HeLa cells transfection with the relative constructs. An asterisk marks a MYC-AMBRA1 degradation sub-product. Gel percentages are indicated in each WB panel

Fig. 3

CDK1-mediated phosphorylation primes PLK1 phosphorylation on AMBRA1. A Schematic representation of AMBRA1 protein, highlighting T1209 and S1223 with their respective surrounding sequence. PLK1 binding motif is underlined. WD40 domains (51–90, 93–133, 135–175) are highlighted in dark grey. B–E WB of immunoprecipitated proteins following Nocodazole treatment: B Endogenous proteins were immunoprecipitated with AMBRA1 antibody. Mouse immunoglobulins were used as control (IP ctr). C–E Overexpressed MYC-AMBRA1 and MYC-β-Galactosidase as control immunoprecipitated using anti-MYC antibody. D HeLa cells treated in the last 10 min of Nocodazole with 9 μM RO-3306. E HeLa cells transfected with MYC-AMBRA1 WT or S1223 phosphosilent mutant (S1223A, Alanine substitution). Quantification, only for mitotic protein extracts, as mean ± s.e.m. of three independent experiments is shown, and significance is **(p < 0.005) for D and *(p < 0.05) for E, by Student’s T test. F Autoradiography (³²P signal) and WB of an in vitro kinase assay with recombinant CDK1/Cyclin B1 and PLK1 performed on MYC-AMBRA1 WT or AA^1209/1223 and MYC-β-Galactosidase (βGal) as control. Proteins were immunoprecipitated using anti-MYC antibody following HeLa cells transfection with the relative constructs. The image represented here derives from the same experiment in which lanes in the middle were cropped. G Model for AMBRA1 phosphorylation at mitosis: AMBRA1 is at first phosphorylated by CDK1/Cyclin B1 on T1209 and S1223, then pS1223 is bound by PLK1 that further phosphorylate AMBRA1 on additional sites. An asterisk marks a MYC-AMBRA1 degradation sub-product. Gel percentages are indicated in each WB panel

Fig. 4

Lack of AMBRA1 phosphorylation causes mitotic defects. A WB of stably AMBRA1-silenced HeLa cells transfected with WT and phosphosilent (AA^1209/1223) MYC-AMBRA1 or with the PLPCX empty vector. An asterisk marks a MYC-AMBRA1 degradation sub-product. Gel percentage is indicated in WB panel. B, C The same cells as in A were grown on coverslips and stained with anti-MYC antibody, to identify transfected cells, and anti-Pericentrin antibody, to identify centrosomes. Nuclei were stained with DAPI (scale bar = 8 μm). Merged images are shown, with 4X magnification shown only in C. White arrows indicate the misaligned chromosomes. The percentage of cells with each defect is represented in the graphs on the right. Bars show mean ± s.e.m. of the percentage of cells which exhibits the indicated defect, and significance is calculated with ordinary one-way ANOVA: * = p < 0.05; *** = p < 0.001; **** = p < 0.0001; n.s. = p > 0.05. Analysis was performed on 100–150 cells for each experiment. D The same cells as in A were grown on coverslips and stained with anti-MYC antibody, to identify transfected cells, and anti-Pericentrin antibody, to identify centrosomes and build the cell division axis. Nuclei were stained with DAPI. On the top left is shown a scheme of how mitotic spindle angle (α) is calculated. On the bottom left, mitotic spindle angle measure (degrees) is shown for all conditions. Bars show mean ± s.e.m. of 50–250 measures, and significance is **** (p < 0.0001) by ordinary one-way ANOVA On the bottom right mitotic spindle angle measure (degrees) is shown as a polar distribution

Fig. 5

AMBRA1 phosphorylation regulates mitotic spindle function through NUMA1 mislocalization from cortex. A Diagram showing significant changing interaction for AMBRA1 upon Nocodazole treatment. HeLa cells, stably transfected with FLAG-AMBRA1, were grown in light (Lys⁰–Arg⁰) and heavy (Lys⁶–Arg¹⁰) SILAC medium. FLAG-AMBRA1 was immunoprecipitated and then eluates were analyzed by mass spectrometry. Here cell cycle regulators are indicated in the green circle while proteins involved in spindle orientation are indicated in the blue circle and specified in the flanking table. NUMA1 is highlighted in red. B, C WB of immunoprecipitated proteins following Nocodazole treatment: B Endogenous proteins were immunoprecipitated with AMBRA1 antibody. Mouse immunoglobulins were used as control (IP ctr). C Overexpressed MYC-AMBRA1 WT or AA^1209/1223 and MYC-β-Galactosidase as control immunoprecipitated using anti-MYC antibody. The image derives from the same experiment in which lanes in the middle were cropped. Quantification, only for mitotic protein extracts, as the mean ± s.e.m. of five independent experiments is shown on the right, and significance is * (p < 0.05) by Student’s T test. D WB of stably AMBRA1-silenced or transfected with WT and phosphosilent (AA^1209/1223) MYC-AMBRA1 HeLa cells. E Analysis of NUMA1 staining for the same cells in D. Merged images of a single confocal plane and 4X magnification are shown for each condition (scale bar = 10 μm). White and yellow arrows indicate NUMA1 cortical staining. The number of cells with an altered NUMA1 staining at the cortex is shown in the graph on the bottom, and significance is n.s. (p > 0.05), *** (p < 0.001) or **** (p < 0.0001) by two-way ANOVA. F NUMA1 line profiles across the cortex from HeLa cells stably expressing GFP-NUMA1 transfected with indicated siRNAs and mCherry-PLPCX or mCherry-AMBRA1 WT or AA^1209/1223. Solid line represents the mean intensity with s.d. represented by dotted lines. N(number of cells, number of independent experiments): siCTR + mCherry (24, 3); siAMBRA1 + mCherry (23, 3); siAMBRA1 + mCherry-AMBRA1 WT (23, 3); siAMBRA1 + mCherry-AMBRA1 AA^1209/1223 (23, 3). G NUMA1 levels at the cortex extracted from (F). Individual cortical intensities with mean ± s.e.m. are plotted. P-values were calculated using Student’s T test and significance is: n.s. (p > 0.05), *** (p < 0.001), **** (p < 0.0001). H The same cells as in (F) were used for quantification of NUMA1 intensity at the poles of mitotic spindles at metaphase. Levels at individual poles are plotted along with mean ± s.e.m. N(number of cells, number of independent experiments): siCTR + mCherry (21, 3); siAMBRA1 + mCherry (23, 3); siAMBRA1 + mCherry-AMBRA1 WT (21, 3); siAMBRA1 + mCherry-AMBRA1 AA^1209/1223 (20, 3). P-values were calculated using Mann–Whitney U test and significance is: n.s. (p > 0.05), **** (p < 0.0001). An asterisk marks a MYC-AMBRA1 degradation sub-product. Gel percentages are indicated in each WB panel

Fig. 6

AMBRA1 depletion impairs PLK1 and CDK1 signaling on NUMA1. A Analysis of LGN and Gαi staining for CTR and AMBRA1 KO HeLa cells. Fluorescent microscopy images are shown on the left (scale bar = 8 μm), while on the right are shown the relative graphs as mean ± s.e.m. of three independent experiments. Significance is n.s. (p > 0.05) by Student’s T Test. B, C AURKA (B) and P AURKA^T288 (C) staining of stable AMBRA1-silenced HeLa cells. The spindle is marked with α-TUBULIN antibody. Fluorescent microscopy images are shown on the left (scale bar = 8 μm), while on the right is shown the relative signal intensity as mean ± s.e.m. of about 40 measures for AURKA (B) and of about 120 measures for P AURKA^T288 (C). Significance is n.s. (p > 0.05) by Student’s T Test. D High-content imaging analysis of PLK1 staining in CTR and AMBRA1 KO HeLa cells treated or not with Nocodazole. Nuclei were stained with DAPI (scale bar = 150 μm). PLK1 signal intensity is shown in the graph on the right. Bars show mean ± s.e.m. of 2000–10,000 measures, and significance is ****(p < 0.0001) by ordinary one-way ANOVA. E, F WB analysis of protein extracts from CTR and AMBRA1 KO HeLa cells treated with Nocodazole. Quantification, as mean ± s.e.m. three (E) and four (F) independent experiments, is shown at the bottom, and significance is * (p < 0.05) (E) by Student’s T test. G Quantification of pole-pole distance increase over metaphase spindle length during anaphase. Individual values with mean ± s.d. are plotted. N(number of cells, number of independent experiments): siCTR (25, 3); siAMBRA1 (5’ UTR) (37, 3); siAMBRA1 (5’ UTR) + WT (25, 3); siAMBRA1 (5’ UTR) + AA (21, 3); siAMBRA1 (5’ UTR) + PXP (24, 3). Significance is n.s. (p > 0.05) and **** (p < 0.0001) by Student’s T Test. Gel percentages are indicated in each WB panel