Identification of new dynactin subunits in S. pombe

By searching the S. pombe genome database, we identified homologs of three dynactin subunits in S. pombe: Arp1 (encoded by arp1 SPBC1347.12), Mug5 (encoded by mug5 SPAC14C4.08), and Jnm1 (also known as Mug1; encoded by jnm1 SPCC11E10.03). Arp1 is a well‐conserved homolog of the actin‐related protein Arp1 in vertebrates, whereas Jnm1 is a homolog of p50/dynamitin in vertebrates and Jnm1 in S. cerevisiae (McMillan and Tatchell, 1994). Mug5 shows weak similarity to p24 in vertebrates and Ldb18 in S. cerevisiae (Amaro et al., 2008; Moore et al., 2008). The expression of arp1, mug5 and jnm1 was highly induced in diploid cells starved for nitrogen, indicating that these genes were expressed exclusively in meiotic cells (supplementary material Fig. S1A). The expression profiles were similar to that of ssm4, which encodes the dynactin subunit p150^Glued (Yamashita et al., 1997). mug5 and jnm1 have been shown to be upregulated during meiosis (Martín‐Castellanos et al., 2005), and jnm1 transcripts have been shown to be selectively eliminated by the Mmi1‐DSR system during vegetative growth (Harigaya et al., 2006).

The deletion mutants of the genes encoding the dynein subunits and Ssm4 are known to produce asci carrying irregular number of spores on a sporulation medium, while most wild‐type cells produce four‐spored asci (Fujita et al., 2010; Miki et al., 2002; Yamamoto et al., 1999; Yamashita et al., 1997). Therefore, we examined phenotypes of the deletion mutants of the identified dynactin genes. All three disruptants (arp1Δ, jnm1Δ and mug5Δ) showed no obvious phenotypic alterations during vegetative growth on a rich medium and did not exhibit sensitivity to the microtubule‐depolymerizing drug thiabendazole (15 µg/ml) similar to in the ssm4Δ mutant (supplementary material Fig. S1B), suggesting that all three gene products were dispensable for mitotic growth. By contrast, these mutants generated aberrant asci similar to dynein and dynactin mutants; although more than 97% of asci from homothallic wild‐type cells had four spores, the arp1Δ, jnm1Δ and mug5Δ strains frequently produced asci containing less or more than four spores (Fig. 1A,B). The frequency of abnormal asci was similar to that of the ssm4Δ strain (Fig. 1B).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Dynactin is required for proper meiotic progression.Dynactin is required for proper meiotic progression. (A) Zygotic asci of wild‐type (wt, JY450) and arp1Δ (JT291) strains. Cells were induced into sporulation on sporulation medium, fixed with 50% ethanol, and stained with Hoechst 33342 to visualize chromosomal DNA. Scale bar: 5 µm. (B) Frequencies of asci with irregular spores in wild‐type (JY450) and dynactin‐mutant (JX524, JT500, JT499 and JT501 for ssm4Δ, arp1Δ, jnm1Δ, and mug5Δ, respectively) strains. Mean±s.d. values of three independent experiments (n = 200 each) are shown. (C) Typical tracks of the SPB movements in wild‐type (JT502) and arp1Δ (JT503) zygotes from karyogamy to meiosis I. Fluorescent images for representative time points are shown in supplementary material Fig. S1C.

Next, we examined the horse‐tail nuclear movement during meiotic prophase in wild‐type and the dynactin mutant zygotes. To chase the movement of chromosomes and the SPB, we used strains carrying histone H2B tagged with CFP (Htb1–CFP) and the SPB component Sfi1 tagged with mCherry (Sfi1–mCherry). In wild‐type zygotes, the SPB started to move vigorously between two poles of a cell immediately after fusion of the two SPBs until the cell entered into metaphase I (Fig. 1C; supplementary material Fig. S1C). In arp1Δ zygotes, vigorous oscillation of the SPB was not observed during the entire meiotic prophase (Fig. 1C; supplementary material Fig. S1C), and the SPB appeared to wander, without apparent direction, around the center of the cell in arp1Δ zygotes. jnm1Δ and mug5Δ zygotes also showed a similar phenotype (supplementary material Fig. S1D).

Two‐hybrid assays indicated that Arp1 directly interacted with Mug5 and Jnm1 (supplementary material Fig. S2A). Ssm4 also interacted with Mug5 and Jnm1, but there was no interaction between Ssm4 and Arp1 (supplementary material Fig. S2B). The immunoprecipitation of Ssm4 brought down Arp1 (supplementary material Fig. S2C, lane 6), and the interaction between Arp1 and Ssm4 was abolished in mug5Δ and jnm1Δ cells (supplementary material Fig. S2C, lanes 7, 8). These results suggest that Arp1 and Ssm4 form a complex through both Mug5 and Jnm1 (supplementary material Fig. S2D).

From these data, we concluded that Arp1, Mug5 and Jnm1 were components of the dynactin complex in S. pombe and were essential for the horse‐tail nuclear movement.

Dynein accumulates at the ends of shrinking microtubules

Next, we observed the behavior of the dynein complex along microtubules in wild‐type cells. To visualize the dynein complex, we used a strain carrying the dynein light intermediate chain Dli1 tagged with three tandem copies of mCherry (Fujita et al., 2010). GFP‐tagged Mal3, a homolog of EB1 that localizes to plus ends of growing microtubules, was simultaneously observed (Sandblad et al., 2006). In cells undergoing nuclear oscillation, dynein and Mal3 were found to be present at different locations (Fig. 2A). Dynein localized to the SPB and microtubules at the forefront of the nuclear movement, whereas Mal3 tended to localize towards the rear of the movement. This biased distribution of Mal3 is consistent with a previous report showing that microtubules elongate mainly in the opposite direction to the nuclear movement (Ding et al., 1998). The mutually exclusive localization of dynein and Mal3 implies that dynein accumulation occurs on nongrowing microtubules. This is contrast to the localization of dynein in budding yeast, in which dynein accumulates at the plus ends of both growing and shrinking microtubules (Lee et al., 2003; Sheeman et al., 2003).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Dynein accumulates at the ends of shrinking microtubules.Dynein accumulates at the ends of shrinking microtubules. (A) Time‐lapse images of dynein (Dli1, magenta) and Mal3 (green) in the wild‐type strain (JT521). (B) Left, an image of dynein (Dli1, red), microtubules (Atb2; green) and Sad1 (blue) in the wild‐type strain (JT504). Right, time‐lapse images of the microtubule bundle shown in the boxed region in the left image. ‘a’ indicates the point of microtubule catastrophe. (C) Time‐lapse images of dynein (Dli1, magenta) and microtubules (Atb2, green) in the wild‐type strain (JT504). (D) Magnified time‐lapse images of the boxed region in C. Scale bars: 5 µm.

We next observed dynein with microtubules and the SPB using the strain carrying Dli1–3mCherry, α‐tubulin tagged with GFP (GFP–Atb2) (Sato and Toda, 2007), and the SPB component Sad1 tagged with CFP (Sad1–CFP). When growing microtubules made contact with a cell end, dynein began to localize at the plus ends of the microtubules, which immediately initiated catastrophe (Fig. 2B, arrowhead labeled a). The plus ends continued to contact the cell cortex, and the accumulated dynein remained at the cortical sites, while the microtubules shortened (Fig. 2B). Accordingly, the SPB was pulled towards the cell end where dynein was accumulated. These observations suggest that dynein at the plus ends of the shrinking microtubules is anchored to the cell cortex and generates a pulling force on the shrinking microtubules at the anchoring sites. Microtubules frequently showed a lateral interaction with the cell cortex and dynein foci were observed on such microtubules (Fig. 2C). The signal intensity of microtubules passing through dynein foci was weak, suggesting that a proportion of the microtubules in a bundle stop growing and that dynein foci might localize to the plus ends of such microtubules.

Additionally, we sometimes observed that microtubules continued to grow along the cell cortex and eventually broke up into segments (Fig. 2C). In these cases, dynein accumulated at both ends of shrinking fragments of broken microtubules (Fig. 2D), indicating that dynein accumulation can be induced by the depolymerization of microtubules.

Dynein is not anchored to the cell cortex in arp1Δ cells

We observed dynein in arp1Δ cells simultaneously with Mal3. In arp1Δ cells, microtubules emanated from the SPB towards both sides of cells and Mal3 was observed at the plus ends of growing microtubules in both sides (Fig. 3A). This is consistent with lack of oscillatory nuclear movement, which is coupled with precisely regulated microtubule dynamics. Dynein appeared at plus ends of microtubules just after Mal3 disappeared from the plus ends (Fig. 3A,B, arrowheads labeled a,c). Unlike wild‐type cells, accumulated dynein did not stay at the cell cortex, but instead moved towards the SPB in arp1Δ cells (Fig. 3B). We frequently observed that dynein was extruded and deviated towards the cell cortex by a plus end of a newly growing microtubule that came in close proximity to the accumulated dynein (Fig. 3B, arrowhead labeled b). Such plus‐end‐directed scattering of dynein was also observed in wild‐type and num1Δ cells (supplementary material Fig. S3A). When simultaneously visualized with microtubules, dynein moving towards the SPB was found at plus ends of shrinking microtubules (Fig. 3C). Unlike wild‐type cells, mutant cells exhibited microtubules that made contact with the cell cortex and then immediately detached from the cell cortex (Fig. 3C). Detachment of microtubules was also observed in num1Δ, jnm1Δ and mug5Δ cells (supplementary material Fig. S3B). The signal intensity of dynein at plus ends in arp1Δ cells was higher than that in wild‐type and num1Δ cells (Fig. 3D). These results indicate that the three newly identified dynactin subunits Arp1, Jnm1 and Mug5 are essential for anchoring of dynein to the cell cortex and pulling of shrinking microtubules at the cell cortex.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Dynein is not anchored to the cell cortex in the mutant.Dynein is not anchored to the cell cortex in the arp1Δ mutant. (A) Time‐lapse images of dynein (Dli1, magenta) and Mal3 (green) in the arp1Δ strain (JT522). (B) Magnified time‐lapse images of the boxed region in A. (C) Top, an image of dynein (Dli1, red), microtubules (Atb2, green) and Sad1 (blue) in the arp1Δ strain (JT505). Bottom, time‐lapse images of a microtubule bundle in the boxed region shown in the top image. ‘a’ and ‘c’, dynein appearing at plus ends of microtubules just after Mal3 disappeared; ‘b’, a plus end of a newly growing microtubule that comes in close proximity to the accumulated dynein. (D) Signal intensity of Dli1–3mCherry at plus ends of shrinking microtubules and the SPB in wild‐type (wt, JT504), arp1Δ (JT505), and num1Δ (JT508) strains. Mean±s.e.m. are shown. *P0.01 (Student's t‐test). Scale bars: 5 µm.

Arp1 localizes to dynein‐anchoring sites at the cell cortex

To examine the localization of Arp1, we constructed a plasmid expressing N‐terminally HA–GFP‐tagged Arp1 as described in the Materials and Methods. During the horse‐tail movement, Arp1 transiently appeared as foci at the cell cortex where dynein accumulated (Fig. 4A). Arp1 was sometimes observed faintly in the cytoplasm where dynein accumulated (Fig. 4A, 2.5 min). Although dynein and Ssm4 localized intensely to the SPB during nuclear movement (Fujita et al., 2010; Miki et al., 2002; Niccoli et al., 2004; Yamamoto et al., 1999), localization of Arp1 to the SPB was very weak, if at all. Colocalization of Arp1 and Ssm4 was prominent at the cortex but faint at the SPB (Fig. 4B). Mug5 and Jnm1 showed localization patterns similar to that of Arp1 (Fig. 4C). These localization patterns suggest that the dynactin complex containing Ssm4 and the other subunits mainly assemble at the cortical anchoring sites, but not at the SPB. In mug5Δ cells, localization of Arp1 at the cell cortex or plus ends was totally abolished (Fig. 4D). This was not due to the reduction in Arp1 protein in mug5Δ cells because HA–GFP–Arp1 was expressed equally in wild‐type and mug5Δ cells (Fig. 4E). In contrast, Arp1 colocalized with dynein at the plus ends of microtubules in num1Δ cells (Fig. 4D), suggesting that colocalization of Arp1 with dynein was independent of Num1.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Arp1 colocalizes with dynein at the cell cortex.Arp1 colocalizes with dynein at the cell cortex. (A) A wild‐type (JT516) strain expressing Dli1–3mCherry (red) and Sad1–CFP (blue) was transformed with the pHGL41‐arp1 plasmid. Transformants were cultured on MM medium to induce the expression of HA–GFP–Arp1 (green) and then shifted onto SPA medium to induce into meiosis. Arrowheads indicate colocalization of Arp1 and dynein. An asterisk indicates colocalization at the plus end detached from the cell cortex. (B) Colocalization of Ssm4–3mCherry (red) and HA–GFP–Arp1 (green). A wild‐type strain expressing Ssm4‐3mCherry (JT438) was transformed with the pHGL41‐arp1 plasmid and observed as in A. The signal intensity of Ssm4–3mCherry and HA–GFP–Arp1 along a microtubule (dotted line) is shown at the bottom. Asterisks indicate localization of Ssm4 at the SPB. (C) Localization of dynein (Dli1, magenta) and Mug5 (in JT593, top, green) or Jnm1 (in JT594, bottom, green) during horse‐tail nuclear movement. Arrowheads and an asterisk indicate colocalization of Mug5 or Jnm1 with dynein at the cell cortex and the plus end detached from the cell cortex, respectively. (D) Localization of HA–GFP–Arp1 (green) with Dli1–3mCherry (red) and Sad1–CFP (blue) in mug5Δ (JT588) and num1Δ (JT589) strains. An asterisk in the mug5Δ strain indicates plus‐end accumulation of dynein without Arp1. Arrowheads in the num1Δ strain indicate colocalization of Arp1 and dynein. (E) Western blot analysis of HA–GFP–Arp1 expressed from the pHGL‐arp1 plasmid in wild‐type (wt, JT875) and mug5Δ (JT876) strains. Cdc2 was used as a loading control. Scale bars: 5 µm.

Ssm4 mediates microtubular localization and cortical anchoring of dynein

We previously showed that Ssm4 enhanced microtubular localization of dynein depending on the microtubule‐binding activity of Ssm4 (Fujita et al., 2010). Unlike arp1Δ cells, ssm4Δ cells exhibit a severe impairment of microtubular localization and no accumulation of dynein at plus ends. Additionally, Ssm4 binds directly with the dynein light chain Dlc1, which is a homolog of Tctex‐1 (also known as DYNLT1) (Niccoli et al., 2004). We located the Dlc1‐binding domain within the middle region of Ssm4 (amino acids 301–400) using two‐hybrid binding assays (supplementary material Fig. S4A). We also mapped the Mug5‐binding domain in Ssm4 to its C‐terminal region (amino acids 541–630; supplementary material Fig. S4B). We constructed a mutant strain designated ssm4‐dC540, in which Ssm4 completely lacked the Mug5‐binding region (Fig. 5A). Immunoprecipitation assays showed that Ssm4‐dC540 lost the ability to bind to Mug5 and interact with Arp1 (Fig. 5B; supplementary material Fig. S4C). In both wild‐type and ssm4‐dC540 strains, Ssm4 colocalized with dynein at the SPB and microtubules, suggesting that Ssm4‐dC540 retained its microtubule‐binding activity and function in the microtubular localization of dynein (Fig. 5C). However, microtubules did not maintain contact with the cell cortex, and the SPB did not show obvious oscillatory movement. Dynein accumulated at plus ends of microtubules detached from the cell cortex in ssm4‐dC540 mutant cells, as observed in arp1Δ, mug5Δ and jnm1Δ cells (Fig. 5C).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

The C‐terminal region of Ssm4 mediates cortical anchoring of dynein.The C‐terminal region of Ssm4 mediates cortical anchoring of dynein. (A) Schematic diagram of Ssm4 and Ssm4‐dC540. (B) Co‐immunoprecipitation of Mug5 with Ssm4. Exponentially growing diploid cells (JT862, JT863 and JT864) were cultured in MM liquid medium without a nitrogen source for 4 h to induce expression of Ssm4–3GFP and Mug5–13myc. Native cell extracts were prepared from these cells and immunoprecipitated (IP) with anti‐GFP antibodies. Ten percent of the extracts and immunoprecipitates were immunoblotted (IB) using anti‐Myc and anti‐GFP antibodies. GFP, Ssm4–3GFP; dC, dC540–3GFP. (C) Time‐lapse images of dynein (Dli1–3mCherry) and Ssm4–3GFP (in JT518, left) or Ssm4‐dC540–3GFP (in JT519, right) along microtubule bundles. Scale bars: 5 µm.

These results indicate that the C‐terminal region of Ssm4, the binding site for Mug5, is unnecessary for microtubular localization of dynein, but indispensable for anchoring of dynein to the cell cortex.

Num1 binds to dynein independently of dynactin and is indispensable for avoiding dynein sliding along the cell cortex

Num1 is known to anchor dynein to the cell cortex (Saito et al., 2006; Yamashita and Yamamoto, 2006). Thus, we next investigated the molecular and functional connection between dynactin and Num1. Because S. cerevisiae Num1 is known to bind the dynein intermediate chain Pac11 (Farkasovsky and Küntzel, 2001), we examined the physical interactions between these proteins in S. pombe by immunoprecipitation (Fig. 6A). The dynein intermediate chain Dic1 (Niccoli et al., 2004) was brought down by immunoprecipitation of Num1 (Fig. 6A, lane 6). This physical interaction between Dic1 and Num1 was also observed in arp1Δ and ssm4Δ strains, although there was a modest decrease in the amount of binding (Fig. 6A, lanes 7, 8), indicating that dynactin is not essential for binding between dynein and Num1.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Num1 binds to the dynein intermediate chain independently of dynactin and is required for static positioning of dynein at the cell cortex.Num1 binds to the dynein intermediate chain independently of dynactin and is required for static positioning of dynein at the cell cortex. (A) Co‐immunoprecipitation of Dic1 with Num1. Native cell extracts from diploid cells [JT584 (wt), JT585 (wt), JT586 (arp1Δ) and JT587 (ssm4Δ)] were prepared as described in Fig. 5B and immunoprecipitated (IP) with anti‐GFP antibodies. For each sample, 2.5% of the extract or immunoprecipitate was immunoblotted (IB) using anti‐HA and anti‐GFP antibodies. (B) Kymographic view of time‐lapse images of Dli1–4GFP foci along the cell‐peripheral regions boxed in the left panels in wild‐type (JT590) and num1Δ (JT592) cells. Time‐lapse images were taken every 5 s. (C) Distance of the movement of dynein cortical foci in wild‐type (JT590), arp1Δ (JT591), num1Δ (JT592) and arp1Δ num1Δ (JT861) cells. Time‐lapse images were taken as in D, and the moving distance of dynein foci along the cell periphery for every 5 s was measured. (D) Localization of Num1 tagged with 3mCherry (magenta), coronin (Crn1–GFP; green), and microtubules (CFP–Atb2; blue) during meiotic prophase in a wild‐type zygote (JT879). (E) The kymographic view of time‐lapse images inside the boxed region in C. Num1–3mCherry and Crn1–GFP are shown in magenta and green, respectively. Scale bars: 5 µm.

Next, we analyzed movement of dynein foci at the cell cortex in arp1Δ and num1Δ mutants. Dynein was not observed at the cell cortex by our previous single time‐point observations in num1Δ cells (Yamashita and Yamamoto, 2006). In num1Δ cells, microtubules started to shrink when they made contact with the cortex (Yamashita and Yamamoto, 2006). However, using detailed, live observations, we found that dynein was localized to the cell cortex for a short time before detachment, although dynein was less stably focused during this period (supplementary material Fig. S3A). In num1Δ cells, dynein drifted on the cell cortex, but its localization was relatively static in wild‐type cells (Fig. 6B). Faint streaks of dynein suggested that these cortical foci were dynein on microtubules. We tracked foci of Dli1–4GFP that existed at the cell cortex and measured the moving distance of the foci along the cell cortex for each 5‐s interval (Fig. 6C). In wild‐type cells, dynein foci hardly moved along the cell cortex (1.28±0.09 µm/min; mean±s.e.m.). By contrast, dynein foci in num1Δ cells tended to slide along the cell cortex (2.50±0.10 µm/min). In arp1Δ cells, movement of dynein foci was also observed, but appeared to be less dynamic than that in num1Δ cells (1.88±0.09 µm/min). In arp1Δ num1Δ double mutant cells, the behavior of dynein foci was similar to that in num1Δ cells (2.37±0.08 µm/min). Consistent with this, Num1, which localizes to the cell cortex forming many foci (Saito et al., 2006; Yamashita and Yamamoto, 2006), hardly moved on the cell cortex, in contrast to the F‐actin‐binding protein coronin (Crn1–GFP) (Petersen et al., 1998), which exhibited rapid changes in localization on the cell cortex (Fig. 6D,E). These data suggest that Num1 has a role in avoiding dynein sliding along the cell cortex, whereas Arp1 makes a weaker contribution to this process.

The motor activity of dynein is essential for establishment of dynein anchoring

To investigate the contribution of dynein motor activity to its anchoring, we constructed a dhc1‐P1 mutant strain, which had an amino acid mutation from Lys to Thr at position 1896 in the walker‐A motif of the first of six AAA⁺ domains of the dynein heavy chain Dhc1. This type of mutation is predicted to cause loss of dynein motor activity (Kon et al., 2004). The dhc1‐P1 strain formed aberrant asci similar to in the dhc1Δ strain (Fig. 7A).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7.

Loss of dynein motor activity affects localization and cortical anchoring of dynein and microtubular dynamics.Loss of dynein motor activity affects localization and cortical anchoring of dynein and microtubular dynamics. (A) Frequencies of asci with irregular spores in the dhc1‐3GFP (dhc1⁺; JT897), dhc1Δ (JT896), and dhc1‐P1‐3GFP (JT513) strains (n>200). (B) Localization of Dhc1‐P1 tagged with 3GFP (JT513). Asterisks and arrowheads indicate localization of Dhc1‐P1 around the SPB and to the plus‐ends of microtubules, respectively. (C) Localization of dynein (Dli1, red), microtubules (Atb2, green), and Sad1 (blue) in wild‐type (wt, JT512) and dhc1‐P1 mutant (JT511) strains. (D) Time‐lapse images of a microtubule bundle in the dhc1‐P1 strain (JT512) in the boxed region shown in C. (E) Localization of dynein (Dli1, red) and Sad1 (blue) around the MTOC (SPB) in the wild‐type (JT504), arp1Δ (JT505), and dhc1‐P1 mutant (JT511) strains. (F) Box plots of microtubule shrinkage rates in the wild‐type (JT504) and mutant strains (JT505, JT506, JT507, JT508, JT509, JT511, and JT510 for arp1Δ, jnm1Δ, mug5Δ, num1Δ, arp1Δ num1Δ, dhc1‐P1, and ssm4Δ, respectively). The diamonds represent the mean, the vertical lines in the boxes represent the median, and the boxes represent the interquartile range. The whiskers extend to the most extreme data point which is no more than 1.5 times the interquartile range from the box. The crosses represent outliers. More detailed information is provided in supplementary material Table S1. (G) The number of microtubule bundles per cell in wild‐type (wt, JT504), num1Δ (JT508), arp1Δ (JT505), jnm1Δ (JT506), mug5Δ (JT507), ssm4Δ (JT510), and dhc1‐P1 (JT511) strains. At least 20 cells were counted for each strain. Error bars indicate s.d. *P<0.005, **P<0.001 (Student's t‐test). Scale bars: 5 µm (B–D), 2 µm (E).

Dhc1‐P1 tagged with 3GFP localized to the SPB and microtubules (Fig. 7B). In dhc1‐P1 mutant cells, the dynein subunit Dli1 showed a similar localization pattern to that of Dhc1‐P1 (Fig. 7C). In this strain, dynein was accumulated at plus ends of shrinking microtubules, which detached from the cell cortex, as observed in the arp1Δ strain (Fig. 7D). These data indicate that the motor activity of dynein is necessary for the establishment of dynein anchoring to the cell cortex.

Interestingly, the localization of dynein to the SPB in the dhc1‐P1 strain was different from that in the wild‐type and the arp1Δ strains. Dynein was excessively accumulated near the SPB in dhc1‐P1 cells (Fig. 7D,E). Sad1, which was observed as a single dot on the SPB during the horse‐tail period in wild‐type cells, scattered around the microtubule‐organizing center [MTOC; 54% (13/24) in the dhc1‐P1 strain, 8% (2/26) in the wild‐type strain, and 15% (4/26) in the arp1Δ strain]. In dhc1‐P1 cells, dynein seemed to be dispersed along microtubules in a broader region than Sad1 (Fig. 7E). Scattering and reclustering of the Sad1 complex in the early stage of meiosis is a key process that regulates telomere clustering during meiotic prophase and is partly mediated by dynein (Yoshida et al., 2013). Excess accumulation of dynein in the dhc1‐P1 mutant might perturb this process.

The motor activity of dynein affects microtubular dynamics

Loss of dynein has been reported to affect microtubular dynamics (Yamamoto et al., 2001). Therefore, we recorded time‐lapse image sequences and measured growth and shrinkage rates of microtubules during the horse‐tail movement in various mutants. We could not detect significant differences in growth rates among the mutants (supplementary material Table S1) or in shrinkage rates between wild‐type and num1Δ cells, as reported previously (Fig. 7F) (Yamashita and Yamamoto, 2006). However, shrinkage rates in the arp1Δ, jnm1Δ, mug5Δ, arp1Δ num1Δ and dhc1‐P1 mutant strains were significantly lower than those in wild‐type and num1Δ cells (Fig. 7F; supplementary material Table S1). The shrinkage rate in the ssm4Δ strain, in which dynein was extruded from microtubules, was greater than those in the wild‐type strain and other strains.

The dynein motor is thought to enhance bundling of microtubules (Amos, 1989). To examine this function of dynein, we measured number of bundles per cell in each strain (Fig. 7G). Whereas wild‐type and num1Δ cells had approximately five microtubule bundles on average, arp1Δ, jnm1Δ, mug5Δ, ssm4Δ, and dhc1‐P1 mutant cells had approximately seven microtubule bundles.

These data suggest that dynein motor activity on microtubules, as well as dynactin subunits Arp1, Jnm1 and Mug5, regulated microtubule shrinkage and enhanced their bundling during the horse‐tail movement, independently of Num1.

The dynactin complex and Num1‐mediated cortical anchoring of microtubules

In our current study, we identified the dynactin subunits Arp1, Mug5 and Jnm1 and showed that they were essential for anchoring of dynein to the cell cortex. Although both deficiencies in dynactin and Num1 led to detachment of dynein from the cell cortex, the mechanisms through which they regulate dynein anchoring seem to be different. Num1 associated with Dic1 independently of dynactin, suggesting that Num1 captured dynein directly. Moreover, dynein foci tended to slide along the cell cortex in num1Δ cells. Accordingly, Num1 appeared to capture dynein at the cell cortex and inhibit sliding of dynein along the cell cortex. The relatively static positioning of Num1 foci supports this interpretation.

One of the essential functions of dynactin is to localize dynein to microtubules (Fujita et al., 2010; Niccoli et al., 2004). In addition to demonstrating this, we also showed that Arp1, Mug5, Jnm1 and the C‐terminal region of Ssm4, to which Mug5 binds, were unnecessary for microtubular localization of dynein, but indispensable for dynein anchoring and microtubule pulling. Although Ssm4 localized to the SPB, microtubules and foci on the cell cortex, similar to dynein, Arp1, Mug5, and Jnm1 localized mainly at the foci on the cell cortex. This finding suggests that some dynactin subunits, namely Arp1, Mug5 and Jnm1, are regulated through a different mechanism than that of Ssm4. The possibility that dynactin subunits could form subcomplexes with discrete functions was proposed in S. cerevisiae (Moore et al., 2008; Woodruff et al., 2009).

In addition, Arp1 foci appeared when microtubules contacted the cell cortex and dynein started to accumulate, suggesting that assembly of the dynactin complex was locally regulated at the cell cortex in vivo. Because the colocalization of Arp1 with dynein was independent of Num1, other cortical factors might regulate the assembly of the dynactin complex.

Regulation of dynein motor activity and microtubule dynamics

In the dhc1‐P1 mutant, in which dynein localized to microtubules but lacked its motor activity, microtubules were not anchored to the cell cortex. These data suggest that dynein is unable to withstand the force generated by microtubule shrinkage without its motor activity, leading to its detachment from the cell cortex. Therefore, regulation of the motor activity of dynein might be the key for the establishment of microtubule anchoring.

Recent studies have demonstrated that barrier‐ or bead‐conjugated dynein tethers microtubule plus ends and slows microtubule shrinkage in vitro (Hendricks et al., 2012; Laan et al., 2012). We found that the microtubule shrinkage rate was increased in ssm4Δ cells, in which dynein was extruded from microtubules (Fujita et al., 2010), similar to the result in dhc1Δ cells (Yamamoto et al., 2001), suggesting that dynein slowed microtubule shrinkage in vivo. The role of this function in the horse‐tail movement has not been fully elucidated; however, we hypothesized that this function might contribute to the generation of coordinated oscillation by balancing the dynamic instability of microtubules. In addition, bundling of microtubules by dynein and dynactin might also play an important role in mediating efficient movement by enhancing unidirectional pulling forces.

In our study, we also observed intense accumulation of dynein in arp1Δ cells, implying that more dynein stalled at plus ends without being directed towards minus ends in arp1Δ cells than in wild‐type and num1Δ cells. In contrast to ssm4Δ and dhc1Δ cells, the microtubule shrinkage rate was lower in arp1Δ cells than in wild‐type cells. In this case, accumulation of immotile dynein in the shrinking plus ends might disturb depolymerization of tubulin. Interestingly, microtubule shrinkage was also inhibited in the dhc1‐P1 mutant, in which the motor‐dead dynein was accumulated at plus ends, supporting the hypothesis that immotile dynein at plus ends may inhibit microtubule shrinkage. The accumulation of dynein at plus ends has been also reported in dynactin mutants in Aspergillus nidulans and S. cerevisiae (Moore et al., 2008; Sheeman et al., 2003; Xiang et al., 2000). Moreover, one study suggested that the plus‐end accumulation might be a consequence of dynein inactivation due to the absence of dynactin (Sheeman et al., 2003). Collectively these data and our current findings of the localization of the dynactin subunits suggest that temporary assembly of the dynactin complex might be the key step for the local and timely activation of dynein, which is probably important for the generation of integrated forces.

Regulation of the microtubular localization of dynein

Our time‐lapse observations have revealed that S. pombe dynein intensely localized to plus ends of microtubules. Although there was some localization and movement of dynein along growing microtubules, S. pombe dynein did not localize to the growing plus ends, but rather localized to the shrinking ends, unlike well‐known plus‐end‐tracking proteins (+TIPs), such as EB1 and CLIP‐170 (also known as MAPRE1 and CLIP1, respectively) (Akhmanova and Hoogenraad, 2005). In addition, S. pombe dynein accumulated at both ends of shrinking microtubules. Characteristics of shrinking ends of microtubules, such as curling protofilaments (Howard and Hyman, 2003), might collect dynein as it disperses along microtubules.

In dhc1‐P1 mutant cells, dynein was accumulated around the SPB. A previous in vitro study has shown that Lys‐to‐Thr mutation in the P‐loop 1 of the Dictyostelium dynein motor domain, which corresponds to the Lys1896‐to‐Thr mutation in S. pombe Dhc1 (Dhc1‐P1), leads to hyperaffinity for microtubules (Kon et al., 2004). Given that dynein was loaded to microtubules from the SPB, accumulation of dynein around the SPB might be a consequence of dynein stalling at its initial loading sites as a result of the increased affinity of dynein to microtubules.

Despite the observation that dynein acts as a minus‐end‐directed motor, dynein accumulated at the plus end of a shrinking microtubule seemed to be transferred to another growing microtubule and dragged or scattered towards the plus end of the growing microtubule. A recent study has visualized single dynein molecules in high temporal resolution and found that dynein molecules diffuse along microtubules after attaching to microtubules within the cytoplasm (Ananthanarayanan et al., 2013). Although diffusive motion of single dynein molecules is thought to be bidirectional, our time‐lapse imaging of Dli1–3mCherry localization exhibited an apparent tendency to orientate towards plus ends. In S. cerevisiae, dynein is transported to plus ends in a manner that is dependent upon the CLIP‐170 homolog Bim1 (Caudron et al., 2008). In S. pombe, however, dynein accumulation at plus ends is independent of the CLIP‐170 homolog Tip1 (our unpublished observations). Therefore, accumulation of dynein at plus ends may facilitate efficient contact of microtubular dynein with the cell cortex.

In conclusion, we propose a model of sequential regulation of dynein localization and anchoring in the horse‐tail movement, whereby dynein is captured by Num1 at the cortex and the dynactin subunits then assemble locally and activate dynein (Fig. 8). Our future studies will seek to elucidate the molecular mechanisms through which dynein is targeted to plus ends in S. pombe and to investigate the spatial regulation of dynactin assembly.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 8.

Model of cortical anchoring of cytoplasmic dynein in the horse‐tail movement.Model of cortical anchoring of cytoplasmic dynein in the horse‐tail movement. (A) Dynein localizes to the SPB independently of dynactin. (B) Dynein, originating from the cytoplasm or being released from the SPB, attaches to microtubules. (C) Ssm4 facilitates efficient binding of dynein to microtubules and allows dynein to disperse along microtubules, biasing towards plus ends. (D) Dynein accumulates at shrinking plus ends and is captured by Num1 at the cell cortex. (E) Arp1, Mug5, and Jnm1 assemble with the dynein–Ssm4 complex at the cell cortex and activate dynein. (F) Assembly of dynactin and Num1 establishes stable anchoring of dynein to the cell cortex, enabling dynein to generate a microtubule‐pulling force.