Essential roles of Drosophila inner centromere protein (INCENP) and aurora B in histone H3 phosphorylation, metaphase chromosome alignment, kinetochore disjunction, and chromosome segregation

Abstract

We have performed a biochemical and double-stranded RNA-mediated interference (RNAi) analysis of the role of two chromosomal passenger proteins, inner centromere protein (INCENP) and aurora B kinase, in cultured cells of Drosophila melanogaster. INCENP and aurora B function is tightly interlinked. The two proteins bind to each other in vitro, and DmINCENP is required for DmAurora B to localize properly in mitosis and function as a histone H3 kinase. DmAurora B is required for DmINCENP accumulation at centromeres and transfer to the spindle at anaphase. RNAi for either protein dramatically inhibited the ability of cells to achieve a normal metaphase chromosome alignment. Cells were not blocked in mitosis, however, and entered an aberrant anaphase characterized by defects in sister kinetochore disjunction and the presence of large amounts of amorphous lagging chromatin. Anaphase A chromosome movement appeared to be normal, however cytokinesis often failed. DmINCENP and DmAurora B are not required for the correct localization of the kinesin-like protein Pavarotti (ZEN-4/CHO1/MKLP1) to the midbody at telophase. These experiments reveal that INCENP is required for aurora B kinase function and confirm that the chromosomal passengers have essential roles in mitosis.

Figure 1

Map of the DmINCENP locus and constructs used in this study. DmINCENP (CG12165) is located on P1 clone, AA005425 and contains six exons. A gene encoding a bromodomain protein (CG1845) is transcribed in the opposite direction to the DmINCENP gene; the ORFs are separated by ∼700 nucleotides. INCENP NH₂- and COOH-terminal constructs were expressed to make antigens for immunization, and full-length INCENP was expressed for microtubule binding and DmAurora B binding experiments. (Sequence data are available from GenBank/EMBL/DDBJ under indicated accession nos.)

Figure 2

(A) Immunoblot of total fly embryo protein extract probed with antibody R801, raised against GST-INCENP_1–348. A single band of 110 kD is visible. (B–Q) Localization of DmINCENP in embryos and cultured cells. All panels are stained for INCENP with R801 (red), α-tubulin (green), and DNA (blue). (B) Syncytial prophase (note that DmINCENP is distributed along the arms of the condensing chromosomes as well as at centromeres); (F) syncytial metaphase and early anaphase; (J) syncytial anaphase. (C, G, and K) Enlargements of selected mitotic figures from B, F, and J, respectively. (D, H, and L) Corresponding INCENP staining; (E, I, and M) the corresponding DNA staining. Arrows show centromere staining in H and I, and midzone staining in K. (N) Mitotic domain of cellularized embryo showing cells at different stages of mitosis. m, metaphase; a, anaphase; t, telophase. (O–Q) Dmel2 cells at metaphase, anaphase, and telophase, respectively. Bars, 5 μm.

Figure 3

(A–L) Immunolocalization of DmAurora B in embryos and cultured cells. DNA (blue), DmAurora B (red), and tubulin (green) are shown. (A) Syncytial nuclei entering mitosis. Black arrow indicates a nucleus (enlarged in C) just ahead of the mitotic wave, lacking detectable aurora B staining. White arrow points to nucleus (enlarged in panel D) entering prophase, with DmAurora B accumulating at the centromere. (E–H) Localization of DmAurora B in syncytial nuclei at various mitotic stages. Note that DmAurora B is located solely at the centromeres from prophase to metaphase. DmAurora B redistributes to the spindle midzone at anaphase. (B) Mitotic domain from a cellularized cycle 14 embryo showing DmAurora B at prometaphase (pm), metaphase (m), late anaphase (a), and telophase (t). (I–K) Dmel2 staining at metaphase, anaphase, and telophase, respectively. White arrows point to centromeric (I) and midbody (K) staining. (L) Immunostaining of Aurora B is prevented by coincubation with recombinant Aurora B protein (Black arrows show absence of staining). Bars, 5 μm.

Figure 4

(A) DmINCENP binds directly to microtubules. (lane M) Molecular weight markers. (lanes 1–5) Coomassie blue–stained gel of microtubule pellet after sedimentation through a sucrose cushion. 64% of the added INCENP (lane 4) but no GST (lane 5) cosediments with microtubules. INCENP does not sediment in the absence of microtubules but remains in the supernatant (lanes 2 and 7). (lanes 6–10) Coomassie blue–stained gel of the supernatant fraction. 36% of the INCENP, but all the GST, remained unbound to microtubules (lanes 9 and 10). (B) DmINCENP binds DmAurora B in vitro. Coomassie blue–stained gel showing DmAurora B input (lane 1, 5% loaded). (lanes 2–3) GST-bound beads incubated with buffer (lane 2) or DmAurora B (lane 3, 50% loaded). No DmAurora B associates with GST beads. (lanes 4–5) GST-INCENP–bound beads incubated with buffer (lane 4) or DmAurora B (lane 5, 50% loaded). 19% of the added DmAurora B associates with GST-INCENP.

Figure 5

Summary of effects of RNAi on cell growth and mitosis. (A′ and A″) Efficacy of RNAi was assessed by immunostaining (INCENP and aurora B RNAi) and by immunoblotting (INCENP RNAi only). Graph shows percentage of cells lacking detectable INCENP or aurora B staining (n = 75). Partially depleted cells were not scored as nulls. INCENP immunoblots were quantitated in NIH Image and normalized with respect to a tubulin-loading control, and levels were expressed as a percentage of t = 0 levels. (B) Growth curve of control and dsRNA-treated cells. (C) Time course of percentage of polyploid (multinucleate and abnormally large) cells. (D–F) Histograms showing the percentage of control, INCENP null, or aurora B null mitotic cells at different mitotic stages. For INCENP at t = 0, 24; and aurora B at t = 0, the whole mitotic population was scored due to the scarcity of null cells. For each time point, n = 50.

Figure 6

INCENP and DmAurora B are mutually dependent for their correct localization. (A) Control metaphase showing INCENP (red) in discrete centromeric spots. (B) Rare metaphase from DmAurora B dsRNA-treated cells. INCENP is more diffusely distributed on the chromosome arms than in controls. Boxes show INCENP staining alone. (C) After DmAurora B RNAi, INCENP does not transfer to the spindle midzone or midbody (arrowhead) at telophase. (D) Control telophase with DmAurora B (red) at the midbody. After INCENP RNAi, DmAurora B is completely delocalized from the chromosomes (E) or from the midbody at telophase (F). Throughout, microtubules are green; DNA, blue). Bars, 5 μm.

Figure 7

INCENP and DmAurora B are required for histone H3 phosphorylation, but not binding of kinetochore protein CENP-A/Cid. (A–F) Phospho-H3 (green) and CENP-A/Cid (red). (A) Condensed chromosomes lacking detectable phospho-H3 (two cells from the same image, INCENP RNAi, 36 h). (B–D) Normal mitotic chromosomes labeled for CENP-A/Cid and phospho-H3: metaphase (B), early anaphase (C), late anaphase (D). (E) Condensed chromosomes lacking detectable phospho-H3 (two cells from the same image, aurora B RNAi, 36 h). (F) Dumpy chromosomes with decreased phospho-H3 (aurora B RNAi, 36 h). (G) There is only a weak correlation between levels of detectable phospho-H3 (green boxes) and local chromatin condensation (blue circles). Note the extreme scatter in the levels of phospho-H3. Ordinate: average pixel intensity. Abscissa: each paired blue circle and green box represent the average for a different cell of three background-corrected measurements of the pixel intensity at 457 and 617 nm, respectively. 457-nm measurements were sorted in ascending order based on pixel intensity. (H and I) Similar levels of chromatin condensation in mitotic cells expressing (H) or lacking (I) Dm Aurora-B (green, both from aurora B RNAi, 36 h). (J and K) High levels of cyclin B protein (red) in prometaphase cells: control RNAi (J) and aurora B RNAi (K). ′, indicates merge; ″, indicates phospho-H3 or aurora B; ‴, indicates DNA (DAPI). Bars, 5 μm.

Figure 9

Removal of DmINCENP causes cytokinesis defects. (A, D, and D′) Telophase figures from untreated cells. All other panels show cells after INCENP RNAi treatment. (A–C) Merged images with INCENP (red), tubulin (green), and DNA (blue). After INCENP RNAi, INCENP (red) is absent from the midbody at telophase in B (arrowhead) compared with untreated cell in A. (C) Example of microtubule structure remaining between two nuclei after a failed cleavage. (D–F) Merged images with actin (red) and DNA (blue). Actin accumulation at the cleavage furrow in late telophase is unaffected by removal of INCENP (compare D with E). However, actin is not present between nuclei in binucleate cells (F). (D′–F′) The corresponding INCENP images. INCENP is efficiently removed by the RNAi procedure. (G) Grossly polyploid cell 72 h after INCENP RNAi. Radial arrays of microtubules extend to the cell cortex in these greatly enlarged cells. Bars, 5 μm.

Figure 8

Aberrant centromere disjunction and lagging chromatin in the INCENP and aurora B RNAi. (A and B) Aberrant anaphases with paired CENP-A/Cid spots (red) at poles and lagging chromatin (INCENP RNAi, 36 h). (C and D) Aberrant telophases with Cid spots at the midbody/midzone (INCENP RNAi, 36 h). (E) Metaphase and telophase/G1 cell, the latter showing an elongate nucleus with a bipolar microtubule array (INCENP RNAi, 36 h). The metaphase cell is still expressing INCENP, which is not detected in the aberrant telophase/G1 cell. (F and G) Banana-shaped nuclei similar to those shown in E, showing well-separated clusters of Cid spots joined by a continuum of chromatin (aurora B RNAi, 36 h). ′, indicates merge (green, α-tubulin); ″, indicates CENP-A/Cid with double arrows showing paired Cid spots; ‴, indicates DNA (DAPI). Bars, 5 μm.

Figure 10

Pavarotti-KLP distribution is normal after DmINCENP or DmAurora B RNAi. (A) Wild-type cell showing PAV-KLP (red) at the midbody, tubulin (green), and DNA (blue). (B) PAV-KLP (arrow) is present at the midbody after INCENP RNAi. (C and D) Aberrant telophases after DmAurora B RNAi. (C) A highly abnormal polyploid cell with two midbodies, both positive for PAV-KLP (arrowhead). (D) A binucleate cell, after a failed cytokinesis, in which a spot PAV-KLP is visible at an internalized midbody remnant (arrowhead). (E) Histogram showing the percentage of telophase cells with positive midbody staining for DmAurora B or PAV-KLP in control or dsRNA-treated cells. Although DmAurora B RNAi removes DmAurora B staining from 80% of cells (left), PAV-KLP staining is removed from only 10% of telophase cells compared with control cells (right, red, n = 30). In the INCENP RNAi, PAV-KLP is detected at the midbody in 94% of telophase cells (blue bar, n = 95). Bars, 5 μm.