The human kinetochore Ska1 complex facilitates microtubule depolymerization-coupled motility

Abstract

Mitotic chromosome segregation requires that kinetochores attach to microtubule polymers and harness microtubule dynamics to drive chromosome movement. In budding yeast, the Dam1 complex couples kinetochores with microtubule depolymerization. However, a metazoan homolog of the Dam1 complex has not been identified. To identify proteins that play a corresponding role at the vertebrate kinetochore-microtubule interface, we isolated a three subunit human Ska1 complex, including the previously uncharacterized protein Rama1 that localizes to the outer kinetochore and spindle microtubules. Depletion of Ska1 complex subunits severely compromises proper chromosome segregation. Reconstituted Ska1 complex possesses two separable biochemical activities: direct microtubule binding through the Ska1 subunit, and microtubule-stimulated oligomerization imparted by the Rama1 subunit. The full Ska1 complex forms assemblies on microtubules that can facilitate the processive movement of microspheres along depolymerizing microtubules. In total, these results demonstrate a critical role for the Ska1 complex in interacting with dynamic microtubules at the outer kinetochore.

Figure 1. A 3 subunit human Ska1 complex localizes to outer kinetochore and microtubules

(A) Images of mitotic cells from clonal human cell lines stably expressing moderate amounts of Ska1, Ska2, and Rama1 as GFP^LAP fusions. Each fusion protein localizes to kinetochores and microtubules throughout mitosis. (B) Purification of a 3-subunit Ska1 complex. GFP^LAP tagged fusions were used to isolate Ska1, Ska2, or Rama1 from the stable cell lines shown in (A) using either one step IPs, or tandem affinity (LAP) purifications. Left, silver stained gel of the Ska1 LAP purification. Right, percent sequence coverage from the mass spectrometric analysis of these samples of proteins identified in the Ska1 complex purifications, but not unrelated controls. (C) Localization of endogenous Rama1. Immunofluorescence using anti-Rama1, anti-HEC1, and anti-GFP (against GFP-CENP-A) antibodies. Rama1 shows a punctuate cellular background staining throughout the cell cycle, but shows pronounced mitotic localization to kinetochores following after nuclear envelope breakdown. (D) Left, merged image showing co-localization of GFP-CENP-A, Rama1, and HEC1. Right, graph showing a linescan of the fluorescent intensity of the highlighted kinetochore. Rama1 co-localizes with Hec1 at the outer kinetochore, and localizes peripherally to CENP-A. (E) Rama1 kinetochore localization depends on the Ndc80 complex, but the Ndc80 complex localizes to kinetochores independently of Rama1. Immunofluorescence of control cells, Rama1 depleted cells, and Nuf2 depleted cells showing localization of GFP-CENP-A, Ndc80/HEC1, and Rama1. Scale bars, 10 μm.

Figure 2. The Ska1 complex is required for kinetochore-microtubule attachments

(A) Overexpression of GFP-Ska1 results in the localization to, and bundling of, interphase microtubules. HeLa cells were transiently transfected with a GFP-Ska1 plasmid and imaged 32 hrs after transfection. (B) Depletion of Ska1 or Rama1 results in a penetrant mitotic arrest. Graph shows percent of G2/M cells based on the fluorescent activated cell sorting and staining against DNA. Visual examination of these cells confirmed that they were arresting in mitosis (not shown). (C) Depletion of Ska1 and Rama1 results in severe chromosome segregation defects. Depleted cells were imaged for microtubules (using anti-tubulin antibodies), DNA, CENP-A (using anti-GFP antibodies in a cell line stably expressing GFP-CENP-A), and Rama1 (using anti-Rama1 antibodies). 1) Class 1 phenotype with the majority of chromosomes aligned at the metaphase plate. Arrows point to a single pair of off axis chromosomes in the depleted cells. 2) Class 2 phenotype with chromosomes severely mis-aligned. 3) Class 3 phenotype with mis-aligned chromosomes and multi-polar spindles. For each condition, 600 mitotic cells were counted to quantify the percent of cells with each of the indicated phenotypes. Scale bars, 10 μm.

Figure 3. Reconstitution of the human Ska1 complex

(A) Coomassie stained gel showing purified GST-Ska1, GST-Ska2, untagged Rama1 (isolated as a GST fusion and cleaved using PreScission protease), a Ska1/Ska2 dimer (isolated using a 6xHis tagged Ska1), and a 3 subunit Rama1/Ska1/Ska2 complex (isolated by mixing the Rama1 subunit with the Ska1/Ska2 dimer). (B) Rama1 binds to the Ska1 complex through the Ska1 subunit. Top, Coomassie stained gel showing a resin-based binding assay for Rama1 to either glutathione agarose beads containing GST (as a control), GST-Ska1, or GST-Ska2, or Ni-NTA agarose beads alone (as a control), or containing the Ska1/Ska2 dimer). Rama1 (arrow) runs at a similar size to GST-Ska1 on SDS-PAGE gels. Bottom, Western blot to visualize Rama1 in these binding assays. The Western blot confirms the binding to GST-Ska1, but not GST-Ska2. (C) Graph showing the elution profile of the full Rama1/Ska1/Ska2 complex or the Ska1/Ska2 dimer on a Superdex 200 size exclusion column (based on OD₂₈₀ absorbance). Arrows indicate the migration of standards with known Stokes radii: Thyroglobulin (85 Å), Ferritin (61 Å), Aldolase (48 Å), and Ovalbumin (30.5 Å). (D) Coomassie stained gel showing the profile of the full Rama1/Ska1/Ska2 complex or the Ska1/Ska2 dimer in a 5-20% sucrose gradient. Arrows indicate the migration of standards with known S values: Chymotrypsinogen A (2.58 S), BSA (4.3 S) and Aldolase (6.5 S). (E) Diagram showing organization of the Ska1 complex based on the reconstitution experiments in this figure.

Figure 4. The Ska1 complex binds cooperatively to microtubules in vitro

(A) Western blot using the indicated antibodies showing the co-sedimentation of 50 nM GST-Ska1, GST-Ska2, or Rama1 with microtubules at the indicated concentrations of polymerized tubulin. At least two samples were examined for each protein. (B) Co-sedimentation of the Ska1/Ska2 dimer, or the full Ska1 complex as in (A) using 50 nM input protein. (C) Graph plotting the quantification of the microtubule binding data shown in (B). The average of multiple samples were plotted for Ska1/Ska2 (n = 3) and Rama1/Ska1/Ska2 (n = 7). Error bars indicate standard deviation. (D) Binding of the Ska1 complex to microtubules requires the acidic C-terminus of tubulin. Treatment of GMPCPP microtubules with subtilisin greatly reduces the binding of the Ska1/Ska2/Rama1 complex based on Western blotting of a microtubule co-sedimentation assay relative to untreated microtubules. (E) Stoichiometry of microtubule binding. Coomassie stained gels showing the pellet fractions containing 0.5 μM microtubules and the indicated concentrations of the Ska1/Ska2 dimer or the full Ska1 complex. Rama1 runs at the same molecular weight as tubulin in SDS-PAGE gels (F) The Ska1 complex shows cooperative microtubule binding behavior. Graph showing microtubule binding as in (C), but with a range of protein input concentrations. (G) Graph showing the binding of the indicated concentrations of the Ska1/Ska2 dimer, the Ska1 complex, or the Ndc80 complex with 5 μM microtubules. Error bars indicate standard deviation. (G) The Ska1 complex bundles microtubules in vitro. Fluorescent images showing rhodamine-labeled microtubules in the presence of the indicated concentrations of the Ska1 complex. Scale bar, 10 μm.

Figure 5. The Ska1 complex forms ring-like assemblies on microtubules

Electron microscopy of negatively stained taxol stabilized microtubules either alone or with Ska1 complex bound. In the bottom image, microtubules with Ska1 complex bound and undecorated microtubules are visible in the same field (indicated by arrows). Scale bars, 100 nm.

Figure 6. The Ska1 complex coated microspheres display microtubule depolymerization driven motility

(A) DIC image (left panel) of a microtubule-bound bead with a corresponding fluorescent image (right panel) showing the stabilized microtubule cap (arrowhead). (B) Example trajectories of moving Ska1 complex-coated beads overlayed with the terminal corresponding DIC image. The trajectory of a bead that is stably attached to the wall of a static microtubule polymer (green) resembles a short arc, because the unattached microtubule plus end swings randomly as the overall microtubule length remains constant. Bead detachment from a microtubule is evidenced by a change of this swinging motion into a random “walk” (blue trajectory). Microtubule depolymerization-driven bead motions (red trajectories) are linearly directed and the amplitude of their swinging motions diminishes as the microtubule shortens. Arrows indicate positions of the coverslip-attached microtubule minus ends. First panel additionally shows an overlayed fluorescent image of the rhodamine-labeled microtubule cap before photo-ablation (arrowhead). (C) Graph showing the percent of beads that moved >0.5 μm and percent of moving beads that reached the microtubule end (full Ska1 complex, n = 112; Ska1/Ska2 dimer, n = 68). (D) Graph showing the mean distances traveled (+/- standard error of the mean - SEM) for Ska1 complex and Ska1/Ska2 dimer coated beads. (E) Graph showing the relative distance between the microsphere and microtubule minus end for the processively moving beads shown in (B). Red lines are the best linear fit for the descending segments of these curves. (F) Histogram distribution showing the motility rates of the Ska1/Ska2 dimer (n = 25) and full Ska1 complex (n = 28) coated microspheres. (G) Graph showing the average rates of bead motions and the time of bead attachment to the shortening microtubule end. Error bars are SEM. (H-I) Speculative model for Ska1 complex assembly and function. (H) The bead motions of the Ska1/Ska2 dimer are similar to Dam1 complex that does not form a ring-like structure, and may be due to the rolling of the beads. The full Ska1 complex moves at a similar rate to ring-like Dam1 complex. In this model, the Ska1 complex is shown as an oligomeric structure on microtubules. The oligomeric nature of Ska1 is based on the strong cooperativity in microtubule binding, the structures visualized on microtubules by EM, and the movement of the Ska1 along microtubules at a rate that suggests it creates an impediment to microtubule depolymerization. (I) Distinct microtubule coupling activities at the kinetochore include a fibril-like structure such as the Ndc80 complex and an oligomeric assembly such as the Dam1 or Ska1 complex.