The Kinesin Klp2 Mediates Polarization of Interphase Microtubules in Fission Yeast

See allHide authors and affiliations

Science  08 Jul 2005:
Vol. 309, Issue 5732, pp. 297-300
DOI: 10.1126/science.1113465


Fission yeast (Schizosaccharomyces pombe) cells grow longitudinally in a manner dependent on a polarized distribution of their interphase microtubules. We found that this distribution required sliding of microtubules toward the cell center along preexisting microtubules. This sliding was mediated by the minus end–directed kinesin motor Klp2, which helped microtubules to become properly organized with plus ends predominantly oriented toward the cell ends and minus ends toward the cell center. Thus, interphase microtubules in the fission yeast require motor activities for their proper organization.

The linear growth of fission yeast cells relies on transport from the cell center to the cell ends of “end markers” along cytoplasmic microtubules (MTs) (1, 2). MTs lie parallel to the long cellular axis in bundles spanning half of the cell length and overlapping at the cell's center; these are termed interphase microtubule assemblies (IMAs) (3). MTs are apparently uniformly polarized within IMAs, with minus ends concentrated at the cell center and plus ends facing the cell periphery (3, 4). This polarization is thought to result from central MT nucleation from interphase microtubule organizing centers and the spindle pole body (36), followed by MT plus-end growth toward the cell periphery and MT alignment parallel to the main cellular axis (7). However, MTs can be nucleated elsewhere in the cytoplasm (8).

We examined IMA turnover and dynamics in cells expressing green fluorescent protein (GFP) fused to α-tubulin (GFP-Atb2) (9, 10). In a population of interphase cells (n = 89), each cell contained three or four IMAs on average (Fig. 1A) (35). This could result from each cell having a different, fixed number of IMAs. However, IMA number within each cell varied continually through time (Fig. 1B). One cause for this variation was MT nucleation both close to (0.25 events/min) and away from (0.06 events/min) the nucleus (Fig. 1C) (table S1). Another cause was the occurrence of “separation/fusion” (S/F) events, in which either a single IMA gave rise to two independent assemblies (0.06 events/min) or two IMAs fused into one (0.18 events/min) (Fig. 1D) (fig. S1, movie S1, and table S1). Fusion occurred with arbitrary direction of alignment between the two IMAs, and IMAs could be seen to invert their direction of alignment during the transit from one IMA into another (Fig. 1, E to G) (movie S2). These findings suggested that IMAs may contain MTs that are not uniformly polarized in the way conventionally thought (35).

Fig. 1.

Dynamics of IMA formation and maintenance. (A) Distribution of IMA number per cell in interphase GFP-Atb2 cells. (B) Variation of IMA number inside a single cell (left, IMA number as it fluctuates; right, “temporal” distribution of IMA number in the cell). (C) De novo IMA assembly (whole-cell projection). Arrowhead, nucleation; arrows, longitudinal alignment. (D) Image sequence (single focal plane) revealing IMA “separation” (S) and “fusion” (F). (E) Whole-cell time series showing a GFP-Atb2 structure inverting orientation as it transits from one IMA into another (white arrows). (F and G) Whole-cell image series of cells expressing GFP-Mal3 [a protein that decorates the MT lattice and accumulates at MT plus ends (29)] revealing parallel (F) and antiparallel (G) MT “fusion” (“F”) events. Arrows denote MT orientations inferred from the GFP-Mal3 signal (stronger at plus ends). Scale bars, 5 μm.

To investigate this possibility, we determined MT plus-end localization in vitro in permeabilized cells. Elongation of endogenous GFP-MTs with rhodamine-labeled porcine tubulin led to the decoration of most IMA tips facing cell ends, indicating the presence there of MT plus ends (Fig. 2A). We then examined the polarity of individual MTs within IMAs by “hook decoration,” a method in which preexisting MTs are decorated with hook-shaped tubulin appendages (11). Electron microscopic (EM) cross sections revealed cells with bundles of two to five MTs decorated by hooks; some bundles contained MTs labeled with hooks of opposite handedness, corresponding to MTs of opposite orientation (Fig. 2B) (fig. S2). Three-dimensional reconstruction from serial sections revealed that bundles of oppositely oriented MTs could be found near the cell tip (Fig. 2, B and C) (movie S3). However, the bundles contained a majority of MTs with plus ends directed toward the cell tips (Fig. 2C) (movie S3). Thus, although antiparallel MTs can coexist within IMAs, MTs within IMAs are mostly uniformly polarized, as predicted by in vivo studies (35).

Fig. 2.

Microtubule polarization within IMAs. (A) Two-color experiment (left) revealing incorporation of exogenous tubulin (center) onto endogenous GFP-Atb2 yeast IMAs (right). Arrowheads, IMA tips; arrows, end of overlap region. Scale bar, 5 μm. (B) Serial EM sections showing a bundle of three hook-decorated MTs. Arrowhead indicates MT antiparallel (with opposite hook handedness) to the other two. (C) Three-dimensional reconstruction stereopair of a hook-decorated MT bundle near the cell tip. Green, MT with plus end toward the cell tip; red, MT with opposite polarity; white/gray surface, cell cortex. Bundle length shown, ∼1.8 μm.

How could a generally uniform polarization of MTs in IMAs be maintained if MTs are able to undergo random “fusion”? One way would be if random association due to fusion were only transient and followed by a reorganization of the MTs leading to uniform polarization. Such a reorganization would be revealed by corresponding variations in MT content within IMAs. Examination of GFP-Atb2 fluorescence intensity revealed that IMAs often contained multiple, brighter MT overlap regions that gradually became “focused” near the cell's midplane (average velocity 8.75 ± 1.15 μm/min, 8 events; Fig. 3A and movie S4), suggestive of MT movement within IMAs. To test this idea, we photobleached GFP-IMAs and observed their fluorescence recovery after photobleaching (FRAP) by kymograph analysis (5). The dark (photobleached) zones of IMAs often regained fluorescence due to outward MT polymerization (Fig. 3B) (movie S5). However, we also detected fluorescence recovery from the periphery inward (Fig. 3C) (movie S6). This was most obvious in the kymographs, where short MTs (Fig. 3D) were seen to enter and exit the photobleached regions of the IMAs (only short MTs could be unequivocally tracked as single MT units). The average velocity of these MTs within IMAs (8.22 ± 1.67 μm/min, 22 events, 250 IMAs filmed) was identical to that of overlap region focusing and differed from individual MT shrinkage and growth velocities [15.45 ± 2.95 μm/min, 30 events, and 3.54 ± 0.63 μm/min, 22 events, respectively (5)]. Moreover, the MTs shrank, grew, or kept a constant length as they crossed the photobleached marks. For this reason, and because the velocity of movement was more than double that of MT growth, we concluded that the movement was not due to MT “treadmilling” (12) and was due to MT translocation (“sliding”) along the IMA (12). Furthermore, in all cases the movement was directed inward (i.e., toward the main MT overlap region of the IMA), indicating a motion directed toward MT minus ends.

Fig. 3.

Klp2-mediated sliding of microtubules within IMAs. (A) Whole-cell image sequence revealing overlap region focusing within a single IMA (arrowheads). (B) Kymograph of “outward” IMA fluorescence recovery (single focal plane). Arrows denote “outward” growing MT. (C and D) Single focal plane image sequence (C) and corresponding kymograph (D) showing “inward” IMA fluorescence recovery. Arrowheads and arrow denote “inward” sliding MT. (E) Kymographs of IMAs photobleached in cells lacking klp4/tea2 (top) or klp5 and klp6 (bottom). Arrows denote “inward” fluorescence recovery. (F) Kymographs from cells lacking dhc1, pkl1, and klp2 (top) or klp2 (bottom). Note lack of “inward” MT sliding after photobleaching. White arrows, MT growth from the cell tip inward; black arrows, MT shrinkage until cell tip (inferred MT minus and plus end localizations marked with – and +). Brackets at left indicate photobleaching sites. Scale bars, 5 μm.

To investigate whether MT-associated motor protein(s) were involved in this sliding movement, we constructed GFP-Atb2 strains carrying deletions of one or several genes encoding fission yeast MT motors. Deletions of the genes klp3 (13), tea2/klp4 (14), klp5 (15), klp6 (15), and klp8, corresponding to putative plus end–directed kinesin-like proteins (KLPs), had no effect on motility (15 events, 214 IMAs filmed; Fig. 3E); sliding occurred with the same frequency, velocity, and directionality as in wild-type cells. However, in cells deleted for dhc1 [dynein heavy chain (9)], klp1/pkl1 (16), and klp2 (17), the genes encoding three putative minus end–directed motors, sliding was abolished (0 events, 202 IMAs filmed). Sliding took place in cells singly deleted for dhc1 and pkl1 but was abolished in klp2-deleted (klp2Δ) cells (0 events, 110 IMAs filmed), suggesting that klp2 mediates MT sliding.

Kymograph analysis of IMAs in klp2Δ cells revealed features consistent with a defect in MT (minus end) focusing at the cell center. These included lack of fusion of MT overlap regions and features indicative of MT minus-end localization at the cell periphery, such as MT growth from the cell end inward and MT depolymerization (“shrinkage”) all the way to the cell tip (Fig. 3F). klp2Δ cells also displayed decreased IMA stability. Although the frequencies of IMA nucleation and fusion in klp2Δ cells were similar to those in the wild type, the frequency of IMA “separation” and catastrophe was increased by a factor of 3 in klp2Δ cells (0.20 events/min and 0.26 events/min, respectively; Fig. 4A) (table S1). In addition, IMAs in klp2Δ and wild-type cells were unstable in the presence of the MT-depolymerizing drug carbendazim (MBC) but, unlike in control cells, the number of MBC-stable MT remnants in klp2Δ cells decreased with time of exposure to the drug (Fig. 4B), and IMAs reassembled after drug removal frequently failed to occupy the whole cell length (Fig. 4C).

Fig. 4.

Role of Klp2 in interphase microtubule stabilization and polarization. (A) Whole-cell sequences of a GFP-Atb2 klp2Δ cell showing two “separation” events (arrowheads, IMA “separation” sites). (B) Number of MT remnants in wild-type and klp2Δ cells after 15 min of treatment with MBC. (C) MT repolymerization during recovery from 10 min of MBC treatment (samples taken after 0, 1.5, and 3 min at 25°C). Arrows indicate an IMA failing to occupy the cell length. (D) Top: Whole-cell kymographs of Tea1-GFP–expressing wild-type (left) and klp2Δ (right) cells (average cell size, 9.5 ± 1.1 μm and 9.5 ± 0.9 μm, respectively). Each track corresponds to an individual Tea1-GFP particle. Arrowheads: “inward” moving Tea1-GFP. Bottom: Percentages of Tea1-GFP moving “outward” (left) and “inward” (right). (E) Tea1-GFP kymographs (top) and quantitation (bottom) using HU-elongated wild-type (left) and klp2Δ (right) cells (average cell size, 21.5 ± 1.7 μm and 21.7 ± 1.6 μm, respectively). Scale bars, 5 μm.

To determine whether Klp2 was involved in the generation of uniformly polarized MTs, we examined the movement of the GFP-labeled cell end marker Tea1 in wild-type and klp2Δ cells. Tea1 particles associate with MT plus ends and travel outward from the cell center (1, 14, 18), and so act as a marker of MT polarization. In kymographs of wild-type cells, 90% of Tea1-GFP moved outward and 10% inward (231 particles, 25 cells; Fig. 4D), indicating a factor of 9 excess of MT plus-end orientation toward the cell tips. In klp2Δ cells, the corresponding values were 75% outward and 25% inward (310 particles, 24 cells; Fig. 4D), indicating only a factor of 3 excess of MT plus-end orientation toward the cell tips. This difference was enhanced in elongated cells generated by treatment for 5 hours with the DNA synthesis inhibitor hydroxyurea (HU). In HU-treated wild-type cells, 71% of Tea1-GFP moved outward and 29% inward (565 particles, 29 cells; Fig. 4E and movie S7), whereas in HU-treated klp2Δ cells the values were 57% outward and 43% inward (695 particles; 29 cells; Fig. 4E and movie S8). These indicate, respectively, a factor of 2.5 and a factor of 1.3 excess of MT plus-end orientation toward the cell tips. A value of 1.3 is close to random (i.e., 50% outward and 50% inward), indicating that in elongated cells Klp2 is required to maintain normal MT polarization with MT plus ends oriented toward the cell tips.

Klp2 is a member of the conserved Kar3/Ncd family of minus end–directed KLPs (19) that generally act in mitotic and meiotic spindles. Its Drosophila melanogaster homolog Ncd exerts an inward force on spindle poles by cross-linking and sliding interpolar MTs (20), and moves along MTs at ∼16 μm/min in vitro (21). In the fission yeast, Klp2 regulates spindle size (17) and localizes to MTs during interphase (17). We suggest that, like bipolar spindles (2227), fission yeast interphase MTs require motor activity for their proper organization. We propose that Klp2 mediates minus end–directed sliding of cytoplasmic MTs relative to each other, which is necessary to maintain the uniform polarization of interphase microtubular arrays (fig. S3). Interestingly, interphase microtubule organizing center integrity can be lost upon disruption of minus-ended motor complexes in mammalian cells (28). Hence, minus end–directed MT sliding may contribute to interphase MT polarization in other eukaryotes.

Supporting Online Material

Materials and Methods

Figs. S1 to S3

Table S1

Movies S1 to S8

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article