Lateral to end-on conversion of chromosome-microtubule attachment requires kinesins CENP-E and MCAK

Abstract

Background: Proper attachment of chromosomes to microtubules is crucial for the accurate segregation of chromosomes. Human chromosomes attach initially to lateral walls of microtubules. Subsequently, attachments to lateral walls disappear and attachments to microtubule ends (end-on attachments) predominate. While it is known in yeasts that lateral to end-on conversion of attachments occurs through a multistep process, equivalent conversion steps in humans remain unknown.

Results: By developing a high-resolution imaging assay to visualize intermediary steps of the lateral to end-on conversion process, we show that the mechanisms that bring a laterally bound chromosome and its microtubule end closer to each other are indispensable for proper end-on attachment because laterally attached chromosomes seldom detach. We show that end-on conversion requires (1) the plus-end-directed motor CENP-E to tether the lateral kinetochore onto microtubule walls and (2) the microtubule depolymerizer MCAK to release laterally attached microtubules after a partial end-on attachment is formed.

Conclusions: By uncovering a CENP-E mediated wall-tethering event and a MCAK-mediated wall-removing event, we establish that human chromosome-microtubule attachment is achieved through a set of deterministic sequential events rather than stochastic direct capture of microtubule ends.

Figure 1

Lateral Kinetochores Are Gradually Converted into End-on Kinetochores (A–C) Fluorescence micrographs showing lateral kinetochores on microtubule walls (A) and levels of Mad2 (B) and astrin (C) on lateral kinetochores in monastrol-treated HeLa cells immunostained with antibodies as indicated and stained with DAPI for DNA. Scale bar measurements are indicated. Insets show 3× magnifications. (D) Graph of percentage of lateral kinetochore pairs with both, one, or neither (none) of the pair displaying Mad2 or astrin in monopolar spindles of monastrol-treated or untreated cells as indicated. (E) Deconvolved single-plane images of a Z stack time-lapse movie of a lateral kinetochore (red) and associated microtubule bundle (green) (column 1) in monastrol-treated HeLa^{YFP-Tub; Cen-Red} cells. Trace of tubulin signal (column 2) illustrates shrinkage and growth of the microtubule bundle. Note the synchronous movement of kinetochore and MT end after end-on attachment (arrows mark kinetochore-attachment site). Images in column 3 are 270° Z-rotated view of column 1. Models in column 4 mark the lateral (L) or end-on (E) kinetochore being tracked. Scale bars represent 2 μm. (F) Graph of total intensity of tubulin and CENP-B signals along the length of rectangular segment (dashed white) (see column 3 of E). Values in purple indicate ratio of tubulin intensities measured at positions before (blue circle) and after (gray circle) the site of kinetochore interaction with microtubule. (G) Illustration of changes in positions of kinetochore (ΔKT) or microtubule-end (ΔMT) between consecutive time frames (T₁ and T₂) and distance between the kinetochore attachment site and the microtubule plus end (KT-MT end). (H) Graph of difference between ΔKT and ΔMT through time during end-on conversion of a lateral kinetochore. T₀ indicates first time point of end-on attachment. Asynchronous (asyn) and synchronous (syn) behaviors indicate greater and lesser than 0.25 μm, respectively, of ΔKT-ΔMT absolute values. (I) Deconvolved single-plane images of a Z stack time-lapse movie of a lateral kinetochore (red) and the associated microtubule bundle (green) (left panels) in monopolar spindles of HeLa^{YFP-Tub; Cen-Red} cells soon after nuclear envelope breakdown (NEBD). Models (right panels) mark lateral (L) or end-on (E) kinetochore being tracked. (J) Temporal evolution of end-on conversion in seven lateral kinetochores selected at random. Fractional contribution of ΔMT (height of green box) and ΔKT (height of red box) to change in KT-MT end distance in μm (green) is shown. + and – indicate direction of kinetochore movement toward plus and minus end of the MT, respectively. (K) Model showing two distinct events, microtubule shrinkage and kinetochore gliding, for bringing the lateral kinetochore and the microtubule end closer to each other. Error bars represent the SEM across cells. See also Table S1 and Figure S1.

Figure 2

Lateral Kinetochores Seldom Detach during End-on Conversion (A) Graph of transition between attachment states (D, detached; L, lateral; and E, end-on) in lateral kinetochores, within a 50 s window, counted from time-lapse movies of monastrol treated or untreated monopolar spindles. (B) Graph showing the percentage of kinetochores that are detached or attached in a lateral or end-on fashion in monastrol-treated and HEC1^Ndc80-depleted cells expressing RNAi-resistant YFP fused to HEC1^Ndc80 -WT or HEC1^Ndc80-Δloop. Lateral kinetochores were scored using tubulin and HEC1 signals in fixed cells immunostained with antibodies against HEC1^Ndc80, β-tubulin, and CREST antisera. The green dashed line marks the degree of uncertainty in distinguishing lateral kinetochores in monopolar spindles, despite a large majority of detached kinetochores, as deduced from Figure S1B. (C) Representative immunofluorescence images of cells depleted of endogenous HEC1^Ndc80 and expressing siRNA-resistant HEC1^Ndc80-WT-YFP or HEC1^Ndc80-Δloop-YFP. Cells were treated with monastrol for 1 hr prior to fixation, immunostained with antibodies against GFP (to detect YFP-HEC1^Ndc80), astrin, and tubulin, and stained with DAPI for DNA. Cropped images are 3× magnified. Insets 1 and 2 are lateral and end-on kinetochores, respectively. Scale bars represent 5 μm (insets, 2 μm). (D) Graph of percentage of lateral kinetochore pairs with both, one, or neither (none) of the pair displaying astrin in cells treated as in (C). (E) Representative images from time-lapse movie of HeLa cells expressing RNAi-resistant HEC1^Ndc80-WT-YFP or HEC1^Ndc80-Δloop-YFP (in green) and mKate-tubulin (in red) as indicated, transfected with HEC1^Ndc80 siRNA, and filmed in the presence of monastrol. White arrowheads mark kinetochores. (F) Graph of percentage of lateral kinetochores that remained lateral (L to L) or converted into end-on (L to E) or detached (L to D), within three consecutive time frames, in movies of HEC1^Ndc80-depleted cells expressing RNAi-resistant HEC1^Ndc80-Δloop-YFP. Error bars represent the SEM values across three experimental repeats. p values representing significance level were derived with the Mann-Whitney U test (A) and proportion test for others. ^∗ and # indicate significant and insignificant differences, respectively.

Figure 3

MCAK and CENP-E Activities Are Required for Lateral to End-on Conversion (A and B) Fluorescence images (A) and graph (B) showing the incidence of lateral kinetochores in cells transfected with siRNA oligos as indicated and treated with monastrol alone prior to immunostaining with antibodies against β-tubulin and HEC1^Ndc80 and CREST antisera and stained with DAPI for DNA. HEC1^Ndc80 images are not shown in (A) but were used for generating (B). Insets are 3× enlarged. Green dashed line marks the degree of uncertainty in distinguishing lateral kinetochores, as deduced from Figure S1B. (C and D) Fluorescence images showing the incidence of lateral kinetochores in cells transfected with either plasmids encoding MCAK wild-type or dead mutant “hypir” (C) or siRNA oligos (D) as indicated and treated with monastrol for 3 hr, prior to immunostaining with antibodies against β-tubulin, either HEC1^Ndc80 (D) or GFP (C), and CREST antisera and stained with DAPI for DNA. Insets are 3× enlarged. Yellow arrows in (C) mark lateral kinetochores with MCAK signal in the outer kinetochore along microtubule fiber, and yellow asterisks mark end-on kinetochores with MCAK signal in the inner centromere. (E) Images showing lateral kinetochores in monastrol-treated cells in the presence or absence of CENP-E inhibitors (GSK923295) and immunostained as in (A). Scale bar measurements are indicated. Error bars represent the SEM values across three experimental repeats. p values representing significance level were obtained with the Mann-Whitney U test. ^∗ and # indicate significant and insignificant differences, respectively. See also Figures S2 and S3.

Figure 4

End-on Conversion Failure in Bipolar Spindles of CENP-E- or MCAK-depleted Cells (A–C) Fluorescence images (A) and graphs (B and C) showing lateral kinetochores in congressed (B) and uncongressed (C) chromosomes of bipolar spindles. HeLa cells were transfected with siRNA oligos as indicated, treated with MG132 for 90 min prior to fixation, and then immunostained with antibodies against β-tubulin or HEC1^Ndc80 and CREST antisera and stained with DAPI for DNA. Scale bar measurements are indicated. Insets are 3× magnified. Yellow triangles mark lateral kinetochores of CENP-E- and MCAK-depleted cells. Error bars represent the SEM values across three experimental repeats. The green dashed line marks the degree of uncertainty in distinguishing lateral kinetochores in bipolar spindles, as deduced from Figure 5E. (D and E) Fluorescence images (D) showing loss of kinetochore-bound Mis12 after Mis12 siRNA treatment and a graph (E) showing lateral, syntelic, bioriented, and detached kinetochores in cells treated with control or Mis12 siRNA. siRNA-treated cells were arrested in metaphase with MG132 for 90 min prior to fixation. Fixed cells were immunostained with antibodies against β-tubulin and Mis12 and CREST antisera and stained with DAPI for DNA. The dashed red line marks the degree of uncertainty in determining lateral kinetochores, in the presence of a large majority of detached kinetochores. Error bars represent the SEM values across two experimental repeats. p values representing significance level were obtained with the Mann-Whitney U test. ^∗ indicates significant difference.

Figure 5

Kinetochores Retain Contact with Lateral Walls of Microtubules in MCAK-Depleted Cells (A) Representative still images of time-lapse movie of HeLa^{YFP-Tub;Cen-Red} cells transfected with siRNA oligos as indicated and filmed in the presence of monastrol. Microtubule tips (yellow arrows) remain distal to kinetochore attachment site in MCAK siRNA-treated cells. Kinetochores (white arrowheads) detach from microtubule walls in CENP-E siRNA treated cells. Lateral, end-on, or no attachment (detached) statuses are marked as L, E, and D respectively. T marks the tumble of kinetochore axis >30°. Scale bars represent 5 μm. (B) Representative graph of ΔKT, ΔMT, and the difference between ΔKT and ΔMT through time in a lateral kinetochore assessed from movie stills as in (A). (C) Graph of attachment state transitions experienced by percentage of lateral kinetochores, within a 50 s window, measured from time-lapse movies of cells transfected with siRNA oligos as indicated and treated with monastrol for 1 hr prior to imaging. (D) Graph of interkinetochore distances describing fate of three randomly chosen kinetochores from conditions described in (A). (E) Cumulative graph and T₅₀ table showing lifetime of lateral kinetochores in HeLa cells treated with either control or MCAK siRNA or HeLa cells expressing tetracycline-inducible siRNA-resistant HEC1^Ndc80-Δloop-YFP as indicated. The dashed line marks T₅₀, the period when 50% of lateral kinetochores had ceased to remain laterally attached. (F) Representative immunofluorescence images of control or MCAK siRNA-transfected cells treated with monastrol for 1 hr prior to fixation and immunostained with antibodies against astrin and tubulin and CREST antisera and stained with DAPI for DNA. Insets are 3× magnified. Scale bars represent 5 μm (insets, 2 μm). (G) Graph of the percentage of lateral kinetochore pairs with both, one, or none of the pair displaying astrin in cells treated as in (F). (H) 3D-rendered images are rotated (as indicated) showing a lateral kinetochore in MCAK siRNA-treated cells immunostained with antibodies against astrin (red) and β-tubulin (green) and CREST antisera (blue). The blue arrow marks kinetochore attachment site and pink arrow marks the plus end of laterally interacting microtubules. Scale bars represent 2 μm. p values representing significance levels were obtained with the Mann-Whitney U test (C), proportion test (D and F), and paired sample t test (E). ^∗ and # indicate significant and insignificant differences, respectively. See also Figure S4.

Figure 6

CENP-E Is Required for Tethering Specifically to Lateral Walls of Microtubules (A) Graph of percentage of detachment episodes observed in end-on or lateral kinetochores during the 5 min duration of time-lapse movies of HeLa^{YFP-Tub;Cen-Red} cells transfected with control or CENP-E siRNA. Detachment episodes that lasted at least for three consecutive time frames alone were considered. (B) Immunofluorescence images of control or CENP-E siRNA-transfected cells treated with monastrol for 1 hr prior to fixation and immunostained with antibodies against Mad2 and tubulin and CREST antisera and stained with DAPI for DNA. Cropped images are 3× magnified. Scale bars represent 5 μm (insets, 2 μm). (C) Graph of percentage of lateral kinetochores with both, one, or none of the kinetochores displaying Mad2 in cells treated as in (B). (D and F) Immunofluorescence images of control or CENP-E siRNA-transfected cells treated with monastrol for 1 hr prior to fixation and immunostained with antibodies against astrin and tubulin and CREST antisera and stained with DAPI for DNA. Cropped images are 3× magnified showing lateral (D) and end-on (F) kinetochores. Scale bars represent 5 μm (insets, 2 μm). (E) Graph of percentage of lateral kinetochore pairs with both, one or none of the kinetochores displaying astrin in cells treated as in (D). (G) Representative still images from time-lapse movie of lateral kinetochore in cells treated or untreated with CENP-E inhibitor. Scale bars represent 5 μm. (H and I) Cumulative graph and T₅₀ table showing the lifetime of lateral kinetochores (H) and the percentage of lateral kinetochores that detached or converted into end-on (I) in cells treated with monastrol for 1 hr prior to imaging in the presence or absence of the CENP-E inhibitor GSK923295. Detachment episodes that lasted at least for three consecutive time frames alone were considered. (J) Representative still images from a time-lapse movie of HeLa cells expressing siRNA-resistant HEC1^Ndc80-Δloop-YFP (in green) and mKate-tubulin (in red) as indicated, treated with siRNA as indicated, and filmed in the presence of monastrol. Arrowheads mark lateral kinetochores. Scale bars represent 2 μm. (K) Illustration of a multistep end-on conversion process involving (1) kinetochore tethering to microtubule walls, (2) bringing kinetochore and microtubule plus end proximal to each other for removing wall contact, and (3) kinetochore tethering to dynamic microtubule ends. CENP-E’s role in tethering kinetochores to lateral walls (4) and MCAK’s role in eliminating microtubule walls interacting with kinetochore (5) and HEC1^Ndc80-loop-dependent tethering of kinetochores to microtubule ends (6) have been assigned sequential steps. Error bars represent the SEM values across three experiments. p values representing significance levels were obtained with the proportion test, except in (H), where the paired sample t test was used. ^∗ and # indicate significant and insignificant differenced, respectively. See also Figure S5.