Mechanisms for focusing mitotic spindle poles by minus end-directed motor proteins

Abstract

During the formation of the metaphase spindle in animal somatic cells, kinetochore microtubule bundles (K fibers) are often disconnected from centrosomes, because they are released from centrosomes or directly generated from chromosomes. To create the tightly focused, diamond-shaped appearance of the bipolar spindle, K fibers need to be interconnected with centrosomal microtubules (C-MTs) by minus end-directed motor proteins. Here, we have characterized the roles of two minus end-directed motors, dynein and Ncd, in such processes in Drosophila S2 cells using RNA interference and high resolution microscopy. Even though these two motors have overlapping functions, we show that Ncd is primarily responsible for focusing K fibers, whereas dynein has a dominant function in transporting K fibers to the centrosomes. We also report a novel localization of Ncd to the growing tips of C-MTs, which we show is mediated by the plus end-tracking protein, EB1. Computer modeling of the K fiber focusing process suggests that the plus end localization of Ncd could facilitate the capture and transport of K fibers along C-MTs. From these results and simulations, we propose a model on how two minus end-directed motors cooperate to ensure spindle pole coalescence during mitosis.

Figure 1.

Capture and transport of kinetochore fibers along centrosomal-nucleated microtubules. (A) A time-lapse image sequence of GFP-tubulin in an untreated (wild-type) S2 cell. Lower panels are enlarged images, and minus ends of K fibers are marked by colored dots. Kinetochore microtubule bundles (K fibers) are cross-linked near their minus ends (e.g., purple dots), followed by focusing to the spindle pole by transport along C-MTs (four dots are gradually focused to the pole during 99–234 s). See also Videos 1 and 2. Bars, 5 and 1 μm. (B) K fiber unfocusing upon C-MTs depolymerization. By replacing culture medium with serum-free medium (SFM), C-MTs are rapidly depolymerized. K fibers gradually shorten and become unfocused. Image acquisition began 30 s after SFM addition. See Fig. S1 for quantitation and see also Video 3 (left). Bar, 5 μm. (C) K fiber focusing upon C-MTs recovery. After 10 min of the serum-free medium (SFM) treatment, large amount of conditioned medium (CM) was added. C-MTs gradually reappeared and accordingly, K fibers became focused. Images acquisition began 60 s after conditioned medium addition (CM). See also Video 3 (right). Bar, 5 μm.

Figure 2.

Distinct effects of Ncd and cytoplasmic dynein on pole coalescence. (A) Representative spindle morphology after RNAi of indicated genes. Green, tubulin; blue, DNA. Bar, 10 μm. (B) Immunoblot showing knockdowns of indicated proteins. Samples were collected at day 7 after RNAi (dsRNA was treated at day 0 and day 4). Note that single and double RNAi result in similar level of reduction of Ncd or Dhc64C. (C) Quantitation of K fiber unfocusing. Relative mean width of the K fiber minus ends (green line) are shown with standard errors bars (control no RNAi cells; 2.2 ± 1.2 μm [mean ± SD; n = 94]). Experiments were performed twice and combined data is presented. Statistically significant increase was seen after Ncd (n = 57), EB1 (n = 91) and Ncd/Dhc64C (n = 59) double RNAi (asterisk; P < 0.0001 for each experiment), whereas Dhc64C single RNAi (n = 73) did not produce significant effects at this sample size (P > 0.05, t test). The frequency of metaphase spindles with multiple poles or unfocused K fiber(s) was also scored and is described at the bottom (n >50 for each treatment). EB1, Ncd, and double Ncd/Dhc64C RNAi significantly increased the spindles with multiple poles or unfocused K fibers. (D) Quantitation of centrosome detachment in the metaphase spindle. The gap distance between centrosome (stained by γ-tubulin) and the minus end of K fiber that is most closely located to the centrosome was measured in three independent experiments (blue line), and relative average distance to accompanying control sample (average 1.2 ± 0.7 μm [mean ± SD; n = 307]) is shown with standard error bars after single Ncd (n = 63), Dhc64C (n = 188), EB1 RNAi (n = 109), or double Ncd/Dhc64C RNAi (n =128). In each spindle, the pole with the wider gap distance was chosen for measurement. We confirmed that the severe detachment seen after Dhc64C RNAi was not due to prolonged metaphase (Goshima and Vale, 2003), as RNAi of APC/cyclosome component Cdc16 or double Cdc16/Ncd also accumulated in metaphase but such a defect was not observed. Reduction of Ncd often leads to formation of the spindles with multiple asters (Goshima and Vale, 2003). However, we selected the cells whose spindle had overall bipolar structure (i.e., chromosomes are aligned and visible in a single line) and single aster on either side.

Figure 3.

FRAP studies of K fiber-associated Ncd-GFP. (A) Comparison of expression levels of Ncd-GFP with endogenous Ncd. A control wild-type cell (1) and a Ncd-GFP–expressing cell in which endogenous Ncd was selectively knocked down by RNAi against UTR region of endogenous Ncd gene (2) were stained with an anti-Ncd antibody. The immunofluorescence intensity of Ncd-GFP in 2 was similar to endogenous Ncd (1). Bar, 5 μm. (B) FRAP experiment reveals dynamic interaction of Ncd with K fiber. (Left) Ncd-GFP fluorescence in the subregion of K fiber (white square) was bleached. (Right) Fluorescence recovery after photobleaching of Ncd-GFP on K fiber. Mean value of relative GFP intensity is shown by square dots with standard deviation (n = 9). Relative GFP intensity was plotted at each time point after normalization using NBA as reference. GFP intensity before bleaching was adjusted to 1.00 (Materials and methods). Immobile fraction was <10% and half time of equilibrium was 2.5 ± 1.0 s, indicating a fast turnover of Ncd-GFP. A FRAP result of GFP-tubulin (subregion of K fiber) is also plotted (open square) and significant fluorescence recovery is not seen in this time range.

Figure 4.

EB1-dependent, microtubule plus end tracking of Ncd. (A, top) A still image of Ncd-GFP in the spindle. (Bottom) Time-lapse imaging of the inset region shows tip association of Ncd-GFP on a growing astral microtubule (yellow arrowheads). Red arrowheads indicate the position of the punctate signals detected at time 0. See also Video 5. (B) Ncd-GFP in the spindle in a cell depleted of EB1 by RNAi. Fewer growing microtubules were observed and Ncd-GFP did not accumulate at the tips (arrowheads). See also Video 8. (C, top) A still image of a cell expressing Ncd-GFP-NES. (Bottom) Time-lapse sequences of the inset region. Enrichment of the signals at the growing tip is seen (yellow arrowheads). Red arrowheads indicate the position of the tip signals detected at time 0. See also Video 6. (D) An Ncd-GFP-NES cell after EB1 RNAi (day 7). Accumulation at the tips of growing MTs of Ncd-GFP-NES is not detected (yellow arrowheads). Red arrowheads indicate the position of the tip signals detected at time 0. See also Video 7. Bars of A–D are 10 μm (top images) or 1 μm (bottom images). (E) In vitro interaction of Ncd tail with COOH terminus EB1 fragment. Coomassie staining after GST pull-down experiment is shown. Binding reaction was performed in 200 μl volume with 12 μM dEB1C and 1 μM GST-Ncd-tail (lanes 2–6) or GST (lane 1). GFP-hAPC-C was preincubated with dEB1C at 1:10, 1:3, 1:1, and 2:1 molar ratio for competition (lanes 3–6). Bound fractions were 18-fold loaded compared with unbound fractions.

Figure 5.

Computer simulation of pole focusing by minus end–directed motor proteins. (A) Five K fibers and one aster of dynamic microtubules were simulated to study the pole formation in the mitotic spindle by dynein (green), Ncd (red), and EB1 (yellow) proteins. The K fibers were 10 times stiffer than single astral microtubule, and arranged in a parallel manner with the plus end upward, from which they were held in place firmly, fixed both in position, and rotation (e.g., immobile chromosomes, large white discs). The interaction of K fibers and microtubules was mediated by complexes able to cross-link any two fibers together (see zoomed regions a–c). The time sequence shows the process of K fiber focusing during this simulation. (B) Definition of K fiber distance in the simulation. (C) Influence of dynein dwell-time after reaching the microtubule minus end. Dynein is simulated alone, varying its detachment rate after reaching the minus end, and the other characteristics of the motor (Table I). Because these different parameters are chosen independently, the random sampling can be used to assess the influence of the dwell-time on the focusing. The quality of the focusing is measured likewise experimentally, and varies between 5 μm for nonfocused K fibers to ∼2 μm for a good focusing (see Materials and methods). The points are derived from 2,135 individual simulations and collected in bins based upon different minus end dwell times. The average and standard deviation of the data from these simulations are shown with open symbols (in order to simplify the graph, at dwell-time below 50 ms, circles represent the mean of more than one dwell time). These simulations reveal that the dwell-time has a strong influence on focusing, with a smooth transition around 100 ms. Dynein with a dwell time of 80 ms or more were able to focus the ends very well in some runs. See also Video 9. (D and E) Effect of plus end–localized Ncd–EB1 complexes on K fiber distance (see Table I for each parameter). In the simulations presented in D, the motor domain of Ncd–EB1 complex binds to C-MTs and the nonmotor domain to the K fiber, whereas it is reversed in E (Ncd motor domain binds to K fibers). Each point represents results of K fiber focusing from two simulations: one for dynein alone and one in which dynein was augmented by Ncd–EB1 (various motor parameters used, but the same set of dynein parameter was compared between plus and minus Ncd–EB1). The K fiber focusing result obtained for dynein alone defines the x-axis coordinate, while the focusing obtained after the addition of Ncd–EB1 defines the y-axis coordinate. Points on the diagonal show that Ncd–EB1 did not affect the outcome of dynein-mediated K fiber focusing. Points below the diagonal reveal a positive contribution of Ncd–EB1 to the focusing, while points above reflect a detrimental effect. These simulation reveal that Ncd–EB1, in a configuration where the Ncd motor domain interacts with C-MTs and the nonmotor domain interacts with K fibers, contributes positively to dynein-mediated K fiber focusing (points tend to be below the diagonal), and that this effect is particularly dramatic for simulations in which dynein was less effective for focusing on its own (larger values on the x-axis). In contrast, Ncd–EB1 complex has a strong negative effect on K fiber focusing in the configuration where the motor domain binds to K fibers. See also Video 10.

Figure 6.

A model for three-step pole focusing by minus end–directed motors. (A) During spindle assembly process in early mitosis, or even during metaphase when steady-state length of bipolar spindle is maintained, C-MTs, and kinetochore microtubule bundles (K fibers) are often discontinuous. (B) Inter-K fiber cross-linking by minus end–directed motors (predominantly Ncd in S2 cells). (C) Search and capture of K fibers by Ncd, which is recruited to the microtubule plus end. Note that EB1 protein is required for plus end tracking of Ncd but is not described in this cartoon figure. Because Ncd has two microtubule binding domains, Ncd enables C-MTs to “search” for, then “capture,” and generate force upon K fibers. Our computer simulation supports the herein described configuration of Ncd where the nonmotor domain of Ncd binds to K fibers. (D) Next, minus end–directed motors (primarily processive dynein motor) transport the K fibers along C-MTs to the pole (bottom). Our simulations suggest that the search and capture mechanism in C can potentially facilitate such motor-mediated transport. However, in the case where a C-MT locates spontaneously closely to a K fiber, dynein can cross-link and transport the K fiber without the help of plus end–tracking Ncd (top).