The chromosomal passenger complex activates Polo kinase at centromeres

Abstract

The coordinated activities at centromeres of two key cell cycle kinases, Polo and Aurora B, are critical for ensuring that the two sister kinetochores of each chromosome are attached to microtubules from opposite spindle poles prior to chromosome segregation at anaphase. Initial attachments of chromosomes to the spindle involve random interactions between kinetochores and dynamic microtubules, and errors occur frequently during early stages of the process. The balance between microtubule binding and error correction (e.g., release of bound microtubules) requires the activities of Polo and Aurora B kinases, with Polo promoting stable attachments and Aurora B promoting detachment. Our study concerns the coordination of the activities of these two kinases in vivo. We show that INCENP, a key scaffolding subunit of the chromosomal passenger complex (CPC), which consists of Aurora B kinase, INCENP, Survivin, and Borealin/Dasra B, also interacts with Polo kinase in Drosophila cells. It was known that Aurora A/Bora activates Polo at centrosomes during late G2. However, the kinase that activates Polo on chromosomes for its critical functions at kinetochores was not known. We show here that Aurora B kinase phosphorylates Polo on its activation loop at the centromere in early mitosis. This phosphorylation requires both INCENP and Aurora B activity (but not Aurora A activity) and is critical for Polo function at kinetochores. Our results demonstrate clearly that Polo kinase is regulated differently at centrosomes and centromeres and suggest that INCENP acts as a platform for kinase crosstalk at the centromere. This crosstalk may enable Polo and Aurora B to achieve a balance wherein microtubule mis-attachments are corrected, but proper attachments are stabilized allowing proper chromosome segregation.

Figure 1. INCENP colocalizes with Polo in early mitosis and interacts with Polo at the inner centromere.

(A) Early prophase, Polo is localized at centrosomes (asterisks), nuclear envelope (blue arrowhead), and starts appearing on chromosomes (white arrows), while INCENP is not yet clearly associated with chromatin. (B) Late prophase, INCENP appears associated with heterochromatin but not yet concentrated at the inner centromere (2, arrows), while Polo is already concentrated in dots colocalizing with INCENP-positive regions (3, white arrows). (C) Metaphase, INCENP at the inner centromere region stretching between Polo-positive kinetochores (arrows). (D) Early anaphase, INCENP decorates thread-like structures joining the segregating Polo-positive sister kinetochores (arrows). Green, Polo-GFP; red, INCENP. Scale bar = 1 µm. (E) In vitro pulldown assays. GST or GST-tagged INCENP were incubated with in vitro translated ³⁵S labelled Polo, Aurora B, or Luciferase. U, unbound fraction; B, bound fraction. Right panel, Coomasie stained gel showing the proteins used for the pulldown. (F) Immunoprecipitation assays. Protein extracts from Polo-GFP (upper panel) or Aurora B-GFP (lower panel) stably transfected cell lines were used in IP experiments using anti-GFP or IgG (Input, whole cell extract, 1% of total loaded; IP, bound fraction, 20% of total loaded). (G–K) Proximity ligation assay (PLA) using antibodies against (G) INCENP and Aurora B (positive control); (H) INCENP and Polo; (I) INCENP and Polo^T182Ph; (J) INCENP and γTubulin (negative control); and (K) Polo and CID (Negative control). Zoomed panels show colocalization of the PLA signal (red) with Aurora B-GFP. Scale bar = 1 µm.

Figure 2. INCENP is required for the activation of Polo kinase at the inner centromere.

(A) Sequence alignment showing the conservation of Thr182 across species. (B) Immunoblot of cell lines stably expressing wild-type, T182A, or T182D Polo in the absence or presence of okadaic acid. The phosphoespecific antibody anti-Plk1^T210Ph recognises Drosophila Polo^T182Ph, both Myc-tagged and endogenous (e). Both endogenous and Myc-tagged Polo^T182Ph were detected as a doublet, suggesting that they can be modified at another site. The asterisk indicates a non-specific band that does not disappear after Polo RNAi. This band increases following okadaic treatment, and therefore could correspond to a non-specific phospho-epitope. (C–F) Control or RNAi-treated DMel-2 cells stably expressing Polo-GFP showing colocalization of INCENP and Polo/Polo^T182Ph. Arrows point to chromosomes shown in zoomed images. Linescans show fluorescence intensity across the kinetochores (dashed lines). (C) Control prometaphase. Polo^T182Ph is visibly enriched at the inner centromere compared to Polo (arrows). Linescans show both Polo and Polo^T182Ph are present at the inner centromere (double-ended arrows show difference in intensity with respect to background levels: green, Polo blue, Polo^T182Ph). (D) Control metaphase. Asterisks point to centrosomes, and Polo^T182Ph is virtually undetectable in the inner centromere (linescan; note that intensity drops to background level). (E,F) INCENP RNAi-treated cells. Asterisks point to Polo^T182Ph on centrosomes (note absence from kinetochores, also shown in linescans). Zoomed images show localization at the centromere/kinetochore of the indicated proteins.

Figure 3. Polo T-loop phosphorylation is required for mitotic progression but not for mitotic entry.

(A) Stable cell lines allowing the copper-inducible expression of Polo^WT-GFP or Polo^T182A-GFP (or control DMel-2 cells) were treated with CuSO₄ for 1 d and transfected with Polo 3′UTR dsRNA the next day in presence of CuSO₄. Four days later, cells were analyzed by immunoblotting. (B) The same cells were analyzed by immunofluorescence to measure the mitotic index (± S.E.M.) using anti-phospho-Histone H3 staining. Note that the expression of Polo^WT-GFP, but not Polo^T182A-GFP, rescued the mitotic index in cells depleted of endogenous Polo. (C) Quantification of the phase distribution of mitotic cells after staining for α-tubulin and DNA (± S.E.M.). Cells expressing Polo^T182A-GFP accumulate in prometaphase/metaphase compared with Polo^WT-GFP cells. (D) Cells expressing Polo^T182A-GFP and depleted of endogenous Polo (D) showed aberrant prometaphase/metaphase figures, with scattered chromosomes, whereas cells expressing Polo^WT-GFP progressed into a normal metaphase (E). Scale bar = 10 µm. (F) Quantification of defects in chromosome alignment and bipolar spindle formation in the different experimental conditions, where all cells were depleted of endogenous Polo by 3′UTR dsRNA. Error bar = SEM.

Figure 4. Aurora B activity is required for the activation of Polo kinase at the inner centromere.

(A) Aurora B phosphorylates Polo kinase in vitro. Bacterially expressed HIS-Polo or HIS-Polo^T182A (which is catalytically inactive and therefore unable to autophosphorylate) were incubated with (or without) Drosophila Aurora B in complex with a fragment of INCENP (residues 654–755) in presence of ³²P-g-ATP, in triplicate. Reaction products were resolved by SDS-PAGE transferred to nitrocellulose and analyzed by autoradiography (AR) and anti-Polo Western blot (WB). Quantitative measurements of signals were obtained (see Materials and Methods), and the ratios were calculated for each reaction (AR/WB, A.U.: arbitrary units). Right, average values for the relative phosphorylatin of Polo^WT and Polo^T182A by Aurora B. Error bars, SEM. (B–D) DMel-2 cells stably expressing Polo-GFP treated with (B) DMSO or (C–D) Binucleine-2, immunostained for INCENP, Polo, and Polo^T182Ph (insets: zoomed images of kinetochores). In (C–D) asterisks indicate centrosomes. Merged images show INCENP/Polo/DNA. Zoomed images in (C–D) insets show examples of kinetochore pairs showing decreased levels of Polo^T182Ph. (E) Dot plot showing the quantification of INCENP/Polo/Polo^T182Ph signal intensity at the kinetochore (t test: *** p<0.0001; n.s., not significant; p = 0.4028). Signal intensities for individual kinetochores were measured using the SoftWorx Data Inspector tool; average background was subtracted; data was plotted using KaleidaGraph software. (F) RNAi depletion of Aurora B, but not Aurora A, strongly reduces Polo^T182Ph levels in DMel-2 cells treated with okadaic acid. Cells were transfected with the indicated dsRNAs for 4 d, and 100 nM okadaic acid added for 4 h before immunoblotting to improve visualization of phosphorylated Polo. A dsRNA against the Kanamycin resistance bacterial gene was used as a negative control. Asterisks: non-specific bands. Both bulk Polo and PoloT^182ph appear as doublets. (G) RNAi depletion of Aurora B or INCENP, but not Aurora A, reduces Polo^T182Ph levels at centromeres/kinetochores. Cycling cells were treated with the indicated dsRNAs for 3 d (immunoblots are shown in Figure S6B) and Polo^T182Ph was detected by immunofluorescence. Levels of Polo^T182Ph at centromeres/kinetochores in prometaphase and metaphase cells were measured at individual kinetochores using Image J, subtracting background (Kt-bkd). Asterisks indicate centrosomes. Error bars = S.E.M.

Figure 5. The centromeric activation of Polo in mitosis depends on INCENP and Aurora B in vivo.

(A–E) Immunostaining analysis of the phosphorylation of Polo^T182 in third instar neuroblasts of (A) wild type larvae (Canton-S), (B) incenp ^QA26 mutant larvae, and (C–D) wild type larvae treated with the Aurora B-specific inhibitor Binucleine 2. INCENP (2, green), Polo^T182Ph (3, red). Arrows point to INCENP blocks characteristic of the incenp ^QA26 and Binucleine-2 treatment phenotype. Asterisks indicate centrosomes. (E) Dot plot showing the quantification of INCENP/Polo^T182Ph signal intensity at the kinetochore (t test: *** p<0.0001; ** p = 0.003). Signal intensities for individual kinetochores were measured using the SoftWorx Data Inspector tool; average background was subtracted; data was plotted using KaleidaGraph software. (F–I) Levels of Polo kinase are not affected by defects in Incenp or Aurora B function (F) wild type larvae (Canton-S), (G) incenp ^QA26 mutant larvae, and (H) wild type larvae treated with the Aurora B-specific inhibitor Binucleine 2. INCENP (5, red), Polo (6, green). (I) Dot plot showing the quantification of INCENP/Polo signal intensity at the kinetochore (t test: *** p<0.0001). Signal intensities for individual kinetochores were measured using the SoftWorx Data Inspector tool; average background was subtracted; data were plotted using KaleidaGraph software.

Figure 6. RNAi depletion of Aurora B or INCENP in human cells strongly reduces Plk1^T210Ph levels at kinetochores.

(A) Control and Aurora B-depleted cells were fixed and immunostained with α-Aurora B (green), α-Plk1 (red), α-Plk1^T210Ph (red), and DNA (blue). (B) Quantification graph of Plk1 and Plk1^T210Ph levels at centromeres in Control and Aurora B-depleted cells. Fluorescence intensities are in Arbitrary Units (A.U.). (C) Control and INCENP-depleted cells were fixed and immunostained with α-INCENP (green), α-Plk1 (red), α-Plk1^T210Ph (red), and DNA (blue). Scale bar = 10 µm. (D) Quantification graph of Plk1 and Plk1^T210Ph levels at centromeres in Control and INCENP-depleted cells. Fluorescence intensities are in Arbitrary Units (A.U.) t test: *** p<0.0001.

Figure 7. Model for the interactions between the CPC and Polo kinase at the centromere/kinetochore.

See text for details.