Kinesin-6 Klp9 plays motor-dependent and -independent roles in collaboration with Kinesin-5 Cut7 and the microtubule crosslinker Ase1 in fission yeast

Bipolar mitotic spindles play a critical part in accurate chromosome segregation. During late mitosis, spindle microtubules undergo drastic elongation in a process called anaphase B. Two kinesin motors, Kinesin-5 and Kinesin-6, are thought to generate outward forces to drive spindle elongation, and the microtubule crosslinker Ase1/PRC1 maintains structural integrity of antiparallel microtubules. However, how these three proteins orchestrate this process remains unknown. Here we explore the functional interplay among fission yeast Kinesin-5/Cut7, Kinesin-6/Klp9 and Ase1. Using total internal reflection fluorescence microscopy, we show that Klp9 forms homotetramers and that Klp9 is a processive plus end-directed motor. klp9Δase1Δ is synthetically lethal. Surprisingly, this lethality is not ascribable to the defective motor activity of Klp9; instead, it is dependent upon a nuclear localisation signal and coiled coil domains within the non-motor region. We isolated a cut7 mutant (cut7-122) that displays temperature sensitivity only in the absence of Klp9. Interestingly, cut7-122 alone is impaired in spindle elongation during anaphase B, and furthermore, cut7-122klp9Δ double mutants exhibit additive defects. We propose that Klp9 plays dual roles during anaphase B; one is motor-dependent that collaborates with Cut7 in force generation, while the other is motor-independent that ensures structural integrity of spindle microtubules together with Ase1.


Results
Klp9 forms homotetramers. Klp9 belongs to the kinesin-6 family, yet its biophysical characterisation with regards to motor activity has not been reported. In addition, whether Klp9 forms homodimers or homotetramers has not been clarified 37 . Therefore, as a first step, we expressed and purified recombinant full-length Klp9 protein tagged with EGFP (Klp9-EGFP) ( Supplementary Fig. S1). To assess the oligomeric state of Klp9, we imaged the fluorescence spots of single Klp9-EGFP molecules adsorbed onto a glass surface. Although some spots showed two-step photobleaching behaviour, importantly, a population of spots displayed photobleaching in more than two steps (Fig. 1A), suggesting that Klp9-EGFP molecules form homotetramers.
We then quantified the intensities of the Klp9-EGFP spots and compared them with those of well-characterised dimeric kinesin-1-EGFP and tetrameric Eg5-EGFP 53 (Fig. 1B). The intensity profile of Klp9-EGFP is clearly more similar to that of tetrameric Eg5-EGFP than dimeric kinesin-1-EGFP, confirming that Klp9 is tetrameric.
Klp9 is a plus end-directed processive motor. To examine the motor activity of Klp9, we performed in vitro assay with total internal reflection fluorescence microscopy (TIRFM). Purified Klp9 was immobilised onto a glass surface via an anti-His antibody, followed by the addition of paclitaxel (taxol)-stabilised, polarity-marked microtubules (Fig. 1C). This microtubule gliding assay showed that Klp9 could translocate microtubules with the minus end leading at a velocity of 39.0 ± 0.3 nm/s (n = 96, N = 3) (Fig. 1D,E), indicating that Klp9 moved towards the microtubule plus ends.
Next, we added purified Klp9-EGFP proteins to immobilised, polarity-marked microtubules at a concentration of 0.3 nM and observed the behaviour of Klp9 at the single molecule level on the microtubule (Fig. 1F). As shown in Fig. 1G, Klp9 displayed processive motility toward the plus ends of microtubules. Quantification of Klp9 motility showed that Klp9 walked on the microtubule with a velocity of 21 ± 2 nm/s ( Fig. 1H; n = 216, N = 3) and its average run length was 430 ± 29 nm (Fig. 1I). These in vitro results firmly establish that Klp9 is a processive motor with plus end-directed motility.

The Klp9 motor is an important determinant for spindle elongation rate during anaphase B.
We previously isolated a klp9 ts mutant that contains a missense mutation within the N-terminal motor domain (klp9-2, K333M) 26 . In addition, an ATPase-defective klp9 rigor mutant containing a point mutation in the switch II region within the motor domain (Klp9 rigor , G296A) was previously constructed 45 . We examined Klp9 localisation (tagged with YFP under the native promotor) in these mutants and compared it with wild type Klp9 (a microtubule was simultaneously visualised with mCherry-Atb2 (α2-tubulin) 54 ). Interestingly, while Klp9-YFP concentrated on the middle region of spindle microtubules during anaphase B, corresponding to the spindle midzone in which antiparallel microtubules overlap 37 , both Klp9-2-YFP and Klp9 rigor -YFP were localised along spindle microtubules in a more diffuse manner ( Fig. 2A,B). These localisation patterns are consistent with the notion that Klp9 possesses a plus-end motility, thereby accumulating on the spindle midzone. www.nature.com/scientificreports www.nature.com/scientificreports/ We then measured spindle elongation rate in these mutants as well as in a klp9 deletion strain (klp9Δ). In all three mutants examined, spindle elongation rate was significantly reduced compared with wild type cells (~50% reduction, 0.75 ± 0.09 for wild type vs 0.38 ± 0.09-0.43 ± 0.06 for the three klp9 mutants, Fig. 2C,D). By contrast, we did not detect any delay regarding anaphase onset in either klp9 rigor or klp9-2 cells (Fig. 2E). It is of note that, a modest, yet significant delay was observed in klp9Δ cells, consistent with a previous report 45 . This is attributable to a motor-independent role of Klp9 in controlling the timing of anaphase onset (Fig. 2E). Overall, these results indicate that Klp9 is a plus-end motor that is crucial for determining the rate of anaphase B spindle elongation.
Klp9 plays an additional role during anaphase B independent of its motor activity in collaboration with the microtubule crosslinker Ase1. Ase1 is a homologue of vertebrate PRC1 and plays a vital role in spindle assembly and stability as an antiparallel microtubule bundler/crosslinker [49][50][51]55 . Previous work showed that a klp9Δase1Δ double deletion strain is inviable 37 , which we also found (Fig. 3A top). However, it is not established as to how these two MAPs functionally interact and ensure cell viability in collaboration 25,37,45 . As Klp9 is a plus end-directed motor, the simplest scenario would be that Ase1-mediated microtubule crosslinking and Klp9 motor-dependent microtubule sliding activities are required for spindle integrity in a functionally redundant fashion. To interrogate this possibility, we crossed ase1Δ with the motor-defective klp9-2. Intriguingly and rather unexpectedly, we found that the resulting klp9-2ase1∆ double mutant strain was not ts (Fig. 3B). In sharp contrast, cut7∆pkl1∆ double mutants that require motor activity of Klp9 for survival 26 displayed ts lethality in combination with klp9-2 ( Fig. 3B). This result implies that Klp9 motor activity alone does not confer viability of ase1∆ cells.
To substantiate this notion, we crossed klp9 rigor with ase1Δ and found that the double mutant was indeed viable and capable of forming colonies at any temperatures including 36 °C (Fig. 3A, bottom and 3B). This indicates that defective motor activity of Klp9 is not responsible for synthetic lethality between ase1∆ and klp9∆. In order to decipher the essential role of Klp9 in the absence of Ase1, we next investigated the functional domain within Klp9 that is required for viability of ase1Δ cells.
Systematic truncation and mutation analysis of Klp9 highlights the importance of a nuclear localisation signal in the C-terminal non-motor region. It was reported that Klp9 plays one additional, motor-independent role in regulating the timing of anaphase onset (see Fig. 2E), which requires the C-terminal 38 amino acid residues 45 . Crossing ase1∆ with the klp9 mutant lacking this region (klp9-∆38C) showed that the double mutant was still viable ( Fig. 4A and Supplementary Fig. S2A), indicating that the role of Klp9 in anaphase onset is not involved in the functional redundancy with Ase1. We then systematically created a series of C-terminally truncated klp9 mutants at the endogenous locus under the native promotor with GFP in the C-terminus, and crossed them with ase1∆. These included Klp9-∆92C, Klp9-∆133C and Klp9-∆172C, the latter two which lack one of the two coiled coil domains and Klp9-∆234C that does not contain either of coiled coil domains (Fig. 4A) and finally, Klp9-∆Motor that consists of only the C-terminal non-motor domain. It was found that all the truncation mutants except for klp9-∆38C were synthetically lethal with ase1∆ ( Fig. 4A and Supplementary Fig. S2B,C).
We also observed the cellular localisation of individual Klp9 mutants during interphase and mitosis. Previous work showed that full-length Klp9 is localised to the nucleus throughout the cell cycle, but its localisation pattern differs depending upon cell cycle stages; it occupies the entire nucleoplasm from interphase until mid-mitosis, and then translocates to the spindle midzone in late mitosis 37 . First, we confirmed this profile (Fig. 4A,B). Interestingly, Klp9-∆92C, but not Klp9-∆38C or Klp9-∆Motor, lost nucleoplasmic localisation during interphase. We noted that the mitotic localisation of Klp9-∆92C to the spindle midzone was somehow still retained ( Fig. 4A,B), though precise quantification of signal intensities showed a substantial reduction compared to those in wild type cells (Fig. 4C). By contrast, all the other truncation mutants (Klp9-∆133C, -∆172C, -∆234C) did not localise to either the nucleus or the spindle midzone ( Fig. 4A-C). We then measured spindle elongation rate of each mutant during anaphase B. As shown in Fig. 4E, all mutants showed compromised spindle elongation during late mitosis (Fig. 4E).
Nuclear localisation of Klp9 during interphase suggested the presence of a nuclear localisation signal (NLS) 56 . Computational search picked up a putative NLS between 546 th and 578 th that contains three consecutive basic amino acids (573K-574R-575K, Fig. 4D). The replacement of three basic residues with alanines (designated Klp9-KARAKA) abolished the nuclear localisation during interphase despite that Klp9-KARAKA was capable of weakly localising to the anaphase B spindle like Klp9-∆92C ( Fig. 4A-C). Importantly, klp9-KARAKA was synthetically lethal with ase1∆ (Fig. 4A,B and Supplementary Fig. S2B,C); thus klp9-KARAKA recapitulated defective profiles of klp9-∆92C. It should be noted that all the mutants created, except for klp9-∆38C, also exhibited synthetic lethality in combination with cut7∆pkl1∆ ( Supplementary Fig. S2A,C). Collectively, the domain analysis shown here has unveiled the existence of the NLS that is essential for correct localisation and function of Klp9.
Two internal coiled coil regions are crucial for Klp9 function in collaboration with Ase1. Next, we created internal deletion mutants of Klp9. The C-terminal region of Klp9 contains two coiled coil domains (CC1, 421-461 and CC2, 501-541). We deleted each region and integrated it into the endogenous locus with GFP in the C-terminus (designated Klp9-∆CC1 and Klp9-∆CC2 respectively) (Fig. 5A). The cellular localisation of (orange) on a microtubule (blue) in the presence of 1 mM ATP is shown. Scale bar, 5 μm. (H,I) Velocity histogram (H) and cumulative probability (I) of single Klp9-EGFP molecules. The average value and the standard error of the mean values from the three independent experiments are shown (n = 50, 101 and 65). The mean run length value for each experiment was obtained by fitting each data set to 1−cumulative probability, www.nature.com/scientificreports www.nature.com/scientificreports/ these Klp9 mutants showed that neither of them was localised to the spindle midzone in late mitosis (Fig. 5B,C). It is noteworthy that interphase nuclear localisation was retained in these mutants, consistent with the notion that they contain the C-terminal NLS identified earlier (see Fig. 4). Genetic analysis indicated that both deletion mutants were synthetically lethal with either ase1∆ or cut7∆pkl1∆ ( Fig. 5D and Supplementary Fig. S2A,C), implying that both CC1 and CC2 are indispensable for Klp9 function. We also measured spindle elongation rate during anaphase B in these deletion mutants, and found that each exhibited the decrease of spindle growth rate (Fig. 5E). It is worth noting that plotting spindle elongation rate against fluorescence intensities of individual mutants including truncation mutants analysed earlier (Fig. 4C,E) and coiled coil mutants (∆CC1 and ∆CC2, Fig. 5C,E) indicate a proportional linear relationship: cells displaying more intense Klp9-GFP signals on spindle midzone elongate faster during anaphase B (Fig. 5F), which is consistent with results recently reported 57 . Taken together, these results highlight that the C-terminal coiled coil domains are essential for Klp9's plus end-directed motor activity as well as its motor-independent role in cooperation with Ase1.
Kinesin-5 Cut7 collaborates with Klp9 in spindle elongation during anaphase B. In klp9 mutants with defects in motor activity (eg. Klp9-2 and Klp9 rigor ), spindle elongation during anaphase B was substantially www.nature.com/scientificreports www.nature.com/scientificreports/ compromised; however, spindles were still capable of elongating, albeit with a reduced rate (see Fig. 2C,D). Previous studies suggested the contribution of Cut7 in late mitosis 25,26,58 . However, there is no direct evidence for the requirement of Cut7 in anaphase B spindle elongation and this issue remains under debate 37 . One of the reasons for this conundrum is that Cut7 is required in the early stages of spindle assembly. As defects in SPB separation leads to the formation of monopolar spindles at restrictive temperatures, all available cut7 ts mutants do not reach anaphase B 5,18,58 . We posited that if Cut7 is also involved in a later step of spindle assembly, a novel type of conditional mutants could be isolated, in which they are defective only during late mitosis and exhibit the ts phenotype only in the background of klp9∆; in other words, such cut7 mutants would not be ts for growth on their own.
Given this assumption, we introduced PCR-mutagenised cut7 fragments into klp9∆ cells and screened for ts transformants (see Materials and Methods for details). We then backcrossed individual ts isolates with wild type cells and examined how the ts phenotype segregated in progenies; expected cut7 mutants should only display temperature sensitivity in the background of klp9∆. A number of anticipated ts mutants were successfully identified (details will be reported elsewhere). In this study, we present one representative mutant called cut7-122. As shown in Fig. 6A, cut7-122 cells displayed the ts phenotype only in the klp9∆ background, but not in the presence of Klp9. Nucleotide sequencing of the cut7-122 ORF identified two point mutations, which led to amino acid replacements at 134 th and 290 th residues (M134T and T290K) within the N-terminal motor domain (Fig. 6B).
To study the impact of the cut7-122 mutant on mitotic spindle elongation on its own or in combination with klp9∆, we observed the dynamic behaviour of spindle microtubules in cut7-122 or cut7-122klp9∆ cells expressing mCherry-Atb2 (a microtubule marker) 54 and Sid4-mRFP (an SPB marker) 59 . Intriguingly, elongation rate of anaphase B spindles was decreased in a single cut7-122 mutant: 0.67 ± 0.11 μm/min and 0.75 ± 0.09 μm/min for cut7-122 and wild type cells respectively. Furthermore, cut7-122klp9∆ double mutants displayed a very low rate of spindle elongation (0.20 ± 0.06 μm/min, Fig. 6C-E); Cut7 and Klp9 contributed additively to the elongation of anaphase B spindles. It is worth pointing out that a single cut7-122 mutant did not display a significant delay in anaphase onset (Fig. 6F) unlike previously isolated conventional cut7 alleles 5,25,26,60 . Note that the measurement of spindle elongation rate during the period prior to anaphase B indicated a modest decrease in cut7-122 cells compared to wild type cells (0.33 ± 0.09 for wild type vs 0.25 ± 0.06 for cut7-122, Supplementary Fig. S3). This sustains the notion that the major, if not sole, defect in this novel cut7 allele lies in the late stage of mitosis.
Given the synthetic temperature sensitivity between cut7-122 and klp9∆, we next asked which activity of Klp9 is required for this genetic interaction. We found that cut7-122 displays the ts phenotype in combination with klp9 rigor , which is in stark contrast to the result that cut7-122 did not exhibit the ts growth in combination with  (1-8) were dissected on YE5S plates and incubated for 3 d at 27 °C. Representative tetrad patterns are shown. Circles, triangles and squares with green lines indicate wild type, ase1∆ single mutants and klp9 single mutants, respectively. Assuming 2:2 segregation of individual markers allows the identification of lethal klp9∆ase1∆ double mutants (indicated by dashed magenta circles). Circles with magenta lines indicate klp9 rigor ase1∆ double mutants. (B) Spot test. Indicated strains were serially (10-fold) diluted, spotted onto rich YE5S plates and incubated at 27 °C or 36 °C for 2 d. cell conc., cell concentration, temp., temperature. (C) Summary of genetic interactions between the klp9 mutants (klp9∆, klp9 rigor or klp9-2) and cut7∆pkl1∆ or between the klp9 mutants and ase1∆. SL, synthetic lethal. V, viable. ts, temperature sensitive.
www.nature.com/scientificreports www.nature.com/scientificreports/ ase1∆ (Fig. 6A,G). Therefore, the survival of cut7-122 at the restrictive temperature depends upon the motor activity of Klp9. The profiles of genetic interactions obtained for klp9 rigor /klp9-2 and cut7-122 are identical with regards to Ase1 (Figs 3C and 6G). Thus, we conclude that Cut7 functions in late mitosis as well as early mitosis; it acts in concert with the motor activity of Klp9 in spindle elongation during anaphase B. www.nature.com/scientificreports www.nature.com/scientificreports/

Discussion
In this study, we explored the molecular mechanisms by which mitotic spindles elongate during anaphase B in fission yeast. Previous work showed the importance of Kinesin-6 Klp9 and the microtubule crosslinker Ase1 26,37,45 ; however, the molecular details underlying the functional interplay between these two MAPs are not known. In addition, it is undetermined whether Kinesin-5 Cut7 on its own plays any role in late mitosis 25,26 . Using biophysical approaches, we show that Klp9 forms homotetramers and is a processive motor with plus-end directionality. Genetic and cell biological analyses indicate that Klp9 collaborates with both Ase1 and Cut7 during anaphase B in two distinct manners. The collaboration with Ase1 is independent of Klp9 motor activity, while the role in combination with Cut7 relies on outward force generated by Klp9 as a motor. Furthermore, isolation of a novel cut7 mutant allele that is defective in anaphase B spindle elongation has solidified the notion that the Cut7 motor is indeed required for spindle elongation, not only in an early mitotic stage, but also in a later step as well (a model illustrated in Fig. 7).
Kinesin-6 Klp9 is a processive plus end-directed motor. Our in vitro assays including photobleaching and fluorescence intensity measurement of Klp9-EGFP indicate that Klp9 forms homotetramers (Fig. 1A,B). Furthermore, TIRFM assay explicitly shows that Klp9 moves towards microtubule plus ends in a processive manner (Fig. 1G,H). The measured velocity of 21 nm/s is comparable with values reported for plus end-directed Kinesin-5 motors 61-63 . www.nature.com/scientificreports www.nature.com/scientificreports/ Recently, it has become clear that fungal Kinesin-5s including Cut7 are bidirectional motors; they display either plus end-directed or minus end-directed motilities depending upon conditions (eg. the degree of motor crowding on the microtubule lattice) [63][64][65] . Under our experimental conditions, we did not detect a similar bi-directionality. It would be of interest to address where this difference stems from and explore the physiologies behind their distinct properties.

Dual roles of Kinesin-6 Klp9 in spindle elongation and microtubule bundling during anaphase B.
We show that Klp9 does not require its motor activity to carry out essential functions in the absence of Ase1 (Fig. 3). Instead, it depends upon the two domains within the C-terminal non-motor region: two coiled coils and the NLS (Figs 4 and 5). It is of note that it remains established as to whether the sole role of this NLS lies in nuclear import activity or it has additional activity such as microtubule binding. Klp9 forms homotetramers like Cut7 (Fig. 1A,B) 37 and the analogous C-terminal coiled coils in Cut7 and other Kinesin-5 members are responsible for self-tetramerisation [66][67][68] . We envision that coiled coil-dependent homotetramers of Klp9 play a structural role in the bundling of antiparallel microtubules at the spindle midzone together with Ase1. To our knowledge, this is the first report on kinesin molecules that play a motor-independent role in microtubule crosslinking in a physiological context. Whether Kinesin-6 members in other species play an analogous role would be of great interest to be explored. Moreover, it is tempting to speculate that Klp9 is involved in the regulatory processes beyond simple sliding and crosslinking of microtubules; it may ensure proper anaphase B progression through regulating the elasticity and plasticity of microtubule overlapping zones as proposed for vertebrate Kinesin-5 69 .
Roles of Kinesin-5 Cut7 during late mitosis. Using a newly isolated cut7 ts allele, cut7-122, our analysis uncovered the crucial role of Cut7 as a motor in late mitosis. Spindle elongation rate in a single cut7-122 mutant is reduced by 11% compared to wild type cells, while that in a klp9∆ strain drops by 49% and in a cut7-122klp9∆ double mutant, the reduction reaches 73% (Fig. 6E). As the cut7-122 strain is most likely hypomorphic, not a complete loss of function, this degree of contribution from Cut7 to anaphase B spindle elongation is potentially underestimated. To address how these two motors collaborate with each other on the antiparallel microtubules, performing TIRFM 70 in the presence of individual proteins or a mixture of the two as a microtubule pair sliding assay would provide useful information.
Despite its slow velocity, we still detect spindle elongation in the cut7-122klp9∆ double mutant (retaining 27% activity compared to wild type cells). As aforementioned, although the Cut7-122 proteins are unlikely to have completely abolished motor activity, cells might still be capable of undergoing spindle elongation without kinesin-generated motor activities. This could be ascribable to the outward forces elicited by non-motor MAPs, including microtubule polymerisation and/or Ase1-and CLASP/Cls1-mediated microtubule bundling and stabilisation as previously pointed out 25,26 .

Figure 7.
A model of spindle elongation during anaphase B. Spindle elongation during late mitosis is executed by three MAPs, Kinesin-6 Klp9, Kinesin-5 Cut7 and the microtubule crosslinker Ase1. Klp9 plays dual roles in this process; one is motor-dependent and the other is motor-independent. Klp9 motor activity generates outward force, thereby inducing the sliding away of antiparallel microtubules at the spindle midzone (bottom left). Cut7 also generates outward force in an additive manner. This force-generating process collaborates with Ase1 to maintain microtubule bundling of anaphase B spindles and that Klp9 plays a role in this branch independently of its motility (bottom right). The motor-independent role of Klp9 requires the NLS and two coiled coil domains locating at the C-terminal non-motor region. Cut7 may not be involved in this microtubule crosslinking; or at least its role is less important compared with that played by Klp9.
www.nature.com/scientificreports www.nature.com/scientificreports/ Although the requirement of Cut7 for anaphase B spindle elongation has been established in this study, whether Cut7 plays another motor-independent role in microtubule bundling, analogous to Klp9, remains undetermined. We observed that klp9 mutants which cannot form homotetramers are synthetically lethal with ase1∆ regardless of Cut7's presence (Fig. 5A). This raises the possibility that Cut7 may not be involved in microtubule crosslinking or at least its role is not as critical as that of Klp9. Thus, activities and/or properties of microtubule binding may not be identical between Cut7 and Klp9. Further cell biological and in vitro assays are required to clarify whether Cut7 is implicated in spindle integrity during anaphase B in a motor-independent manner.
Cell cycle-dependent cellular localisation of Kinesin-6 Klp9. We have identified a canonical NLS in the C-terminal region of Klp9 that is essential for its mitotic function. Klp9 constructs without a functional NLS could still be loaded onto the spindle midzone within the nucleus, though with reduced levels (Fig. 4A,B). Thus, it is likely that Klp9 is transported into the nucleus in a dual manner: one through its own NLS and the other perhaps through binding to another protein. Ase1, which acts together with Klp9 during anaphase B, might well be such a binding partner; however, two previous reports 37,45 as to whether Klp9 physically binds Ase1 are somewhat controversial. Hence, how Klp9 is localised to the mitotic nucleus in an NLS-independent manner remains undetermined. However, it is worth noting that despite the capability of its NLS-independent nuclear import during late mitosis, levels of Klp9 on spindles are insufficient to support the mitotic role of Klp9 (Fig. 4A). Thus, the NLS within the C-terminal domain of Klp9 is essential for Klp9 function.
Kinesin-5 and Kinesin-6 in spindle elongation in other species. The role of Kinesin-5 in bipolar spindle assembly, in particular during very early mitotic stages when the two centrosomes separate, is conserved across many eukaryotic species 71 . In addition, Kinesin-6 members ubiquitously participate in late mitotic events. Yet, the functional interplay between these two kinesins, if any, remains to be reported 10 . The results presented in the current study are the first to explore this notion. In budding yeast, Kinesin-5 Cin8 concentrates on central spindle microtubules during anaphase and plays a critical role in the maintenance of spindle pole separation 8,72 . However, this organism does not contain Kinesin-6 members. Instead, it has another Kinesin-5 Kip1, which collaborates with Cin8 in spindle elongation. It is likely that this yeast has developed a unique evolutionary strategy by which two different Kinesin-5 members compensate for the functions of Kinesin-6 in anaphase B spindle elongation. It is of note that a recent study showed that Kinesin-8 Kip3 acts in late mitosis together with Kip1 73 .
In higher eukaryotes, multiple Kinesin-6 proteins play collaborative roles in spindle elongation during anaphase B coupled with cytokinesis. In Drosophila melanogaster, Kinesin-5 Klp61F and Kinesin-6 Subito work together in maintaining meiotic bipolar spindles 74 . Subito is known to be an important kinesin for mitotic spindle assembly as well. It would be of great interest to address whether and how the collaboration between Klp61F and Subito in spindle assembly extends into mitosis.

Materials and Methods
Strains, media and growth conditions. Strains used in this study are listed in Supplementary Table S1. Media, cultivation conditions, and manipulations were performed as previously described [75][76][77] . For most of the experiments, cells were grown in rich YE5S liquid media and on agar plates. Spot tests were carried out by spotting 5-10 μl of cells at a concentration of 2 × 10 7 cells/ml after 10-fold sequential dilutions onto rich YE5S plates. Plates were kept at various temperatures from 27 °C to 36 °C as appropriate.
Nucleic acids preparation and manipulation. DNA-manipulating enzymes were used as recommended by the manufacturers (New England Biolabs Inc. Ipswich, MA, U. S. A. and Takara Bio Inc., Shiga, Japan).

Construction of recombinant kinesins.
To construct a full-length Klp9 expressing vector (pcDNA3.4/ SpKlp9FL-EGFP-FLAG-His8), the full-length klp9 cDNA was PCR amplified from a cDNA library (National BioResource Project, pTN-RC5) with a forward primer, ATGATACAGATTTTTCTGCGTGT and a reverse primer, AATTCATTAATATCGATATCAGTTG, and inserted into a modified pcDNA3.4 vector (pcDNA3.4/ EGFP-FLAG-His8) via appropriate oligo nucleotide adaptors for the acceptor vector using the Gibson Assembly method. The resultant linker amino acid between SpKlp9 and EGFP is AAA (one-letter amino-acid code).
To construct Eg5-EGFP (pcDNA3.4/HsEg5FL-EGFP-FLAG-His8), the full-length human Eg5 cDNA was amplified from a cDNA library constructed from HEK293 cells using the SV Total RNA Isolation System (Z3100, Promega Co. Expression and purification of proteins. We essentially followed procedures described previously 78  Cells from 1-L culture were used for two-step purification using Ni-IMAC and FLAG agarose. Protein concentration was determined by the Bradford assay using bovine serum albumin as a standard. Preparation of fluorescently labelled microtubules. Tubulin was purified from pig brain using a high-molarity PIPES buffer (1 M PIPES-KOH pH 6.8, 20 mM EGTA, and 10 mM MgCl 2 ) as described previously 79  www.nature.com/scientificreports www.nature.com/scientificreports/ Quantification of fluorescence signal intensities of purified proteins on a glass surface. Flow chambers were assembled from two coverslips (18 × 18 mm and 24 × 32 mm, Matsunami Glass Ltd., Osaka, Japan) with two slivers of Parafilm as a spacer. The coverslips were heat-sealed on a 98 °C hot plate. GFP-tagged proteins were appropriately diluted in HME buffer (20 mM HEPES pH 7.4 adjusted with NaOH, 2 mM MgCl 2 and 1 mM EGTA), and then perfused into a flow chamber. After checking the adsorption of GFP-labelled kinesins to the glass surface, the flow chamber was washed with HME buffer supplemented with 2 mM dithiothreitol, 25 mM glucose, 21.3 U/ml glucose oxidase and 800 U/ml catalase.
Fluorescence spots were imaged using an objective-type TIRFM based on Ti-E (Nikon) equipped with a 488-nm laser (488-75CDRH, Coherent Inc., Santa Clara, CA, U. S. A.) and a ×60/NA1.49 oil immersion objective lens (CFI Apo TIRF 60xH, Nikon Inc., Tokyo, Japan) at 24 ± 1 °C. Images were magnified by 2.5× TV-adaptor (MQD42120, Nikon Inc., Tokyo, Japan) and projected onto an EMCCD detector (C9100-13, Hamamatsu Photonics K.K., Shizuoka, Japan). The camera was controlled by the Micro-manager software ver. 1.4.2243. Photobleaching behaviour of Klp9-EGFP was analysed by summing intensities from a 9 × 9-pixel region and subtracting the mean intensity of 9-10 regions around each fluorescence spot as a background intensity after four-frame rolling averaging using the Running Z projector plugin for ImageJ. To obtain the intensity profile of GFP-labelled kinesins adsorbed to a glass surface, the TrackMate plugin ver. 3.5.1 for Fiji was used to automatically detect and quantify the intensities of fluorescence spots. The DoG (Difference of Gaussian) detector with an estimated diameter of 9 pixels and a threshold value of 3 were used. Background intensity was measured at 9-10 random background regions in the field of view and subtracted from individual intensities of fluorescence spots. Microtubule gliding was imaged using the same setup as intensity measurement described above except that a 532-nm laser (CDPS532S-020, JDS Uniphase Co., Milpitas, CA, U. S. A.) was used for TIRF illumination. The position of the filament was determined manually by mouse clicking on the tip of the filament in the standalone custom software (Mark2 ver. 2.6) and tracked typically every 10 s. Microtubules that glided for at least 10 s were included in velocity measurement. After the distance travelled were plotted against time, velocities were determined by the slope calculated from the linear regression of each trace.

Microtubule gliding assay.
Single molecule motility assay. We imaged single-molecule behaviour of Klp9 as previously described 78 .
Single-molecule behaviour of GFP-kinesins was imaged using the same setup as intensity measurement described above except that a 632.8-nm laser (30991, Research Electro Optics Inc., Boulder, CO, U. S.A.) was used along with a 488-nm laser (488-75CDRH, Coherent, Inc., Santa Clara, CA, U. S. A.) for TIRF illumination. The image framerate was 100 ms per frame. The custom software (Mark2 ver.2.6) was used to manually pick fluorescence spots and to determine their sub-pixel positions by using automated two-dimensional Gaussian fitting algorithm. All fluorescence spots that moved on a given microtubule were picked but ones that disappeared within 1.2 s or travelled a distance of less than 200 nm were excluded. After the position of the fluorescence spots along the microtubule were plotted against time, velocities were determined by the slope calculated from the linear regression of each trace.
Strain construction, gene deletion and epitope tagging. A simple PCR-based strategy 76,77 was used for complete gene deletion and epitope tagging in the C terminus, thereby all the tagged proteins being expressed under their own endogenous promoter.
Isolation of cut7 temperature-sensitive mutants in the klp9∆ background. The cut7 ts mutants were constructed by PCR-based random mutagenesis. The GFP epitope tag and G418-resistance marker gene cassette (kanR) were first inserted into the 3′ flanking region of the cut7 gene (cut7-GFP-kanR). The cut7-GFP-kanR fragment purified from this strain was amplified/mutagenised with PCR using TaKaRa EX taq polymerase (Takara Bio Inc., Shiga, Japan) and transformed into a klp9∆ strain (deleted by hphR). G418 (and Hygromycin B)resistant colonies were picked up and backcrossed with a wild type strain. After confirming co-segregation between G418/Hygromycin B-resistance and the ts phenotype, nucleotide sequencing was performed to determine the mutated sites within the cut7 gene of these mutants.
Fluorescence imaging experiments were performed with a DeltaVision Elite microscope system (GE Healthcare, Chicago, IL, U. S. A.), which is a wide-field inverted epifluorescence microscope (IX71; Olympus, Tokyo, Japan) equipped with a PLAPON 60× 1.40 NA oil objective lens (Olympus, Tokyo, Japan). Fluorescence images were captured by a charge-coupled device camera (CoolSNAP HQ2; Photometrics, Tucson, AZ, U. S. A.) and the softWoRx 6.5.2 software (GE Healthcare, Chicago, IL). Live cells were imaged in a glass-bottomed culture dish (MatTek Corporation, Ashland, MA, U. S. A.) coated with soybean lectin (L1395, Sigma-Aldrich, St. Louis, MO, U. S. A.) and incubated at 27 °C for most of the strains or at 36 °C for the ts mutants. For ts mutants, cells were cultured in rich YE5S media until mid-log phase at 27 °C and subsequently shifted to the restrictive temperature of 36 °C for further culture until observation. To keep cultures at the proper temperature, cell culture dishes were mounted in a temperature-controlled chamber (Air Therm SMT; World Precision Instruments Inc., Sarasota, FL, U. S. A.). Images of 14-16 sections were taken along the z-axis at 0.2-μm intervals at each time point. After deconvolution, projection images of maximum intensity were obtained. Captured images were processed with Photoshop CS6 (version 13.0; Adobe, San Jose, CA. U. S. A.).

Quantification of fluorescence signal intensities in cells. For quantification of signal intensities
of Klp9-GFP located at the spindle midzone during anaphase B, 14-16 sections were taken along the z-axis at 0.2-μm intervals. After deconvolution and projection images of maximum intensity, a 20 × 20-pixel (2.15-μm square) region of interest (ROI) with maximum sum intensity was determined. After subtracting the mean intensity of three regions around each ROI as background intensities, values of the maximum sum intensities were used for statistical data analysis.
Statistical analysis. The significance of differences in different strains was determined by the two-tailed unpaired Student's t-test. All the experiments were repeated at least twice. Sample numbers used for statistical analysis were presented in the corresponding figures and/or legends. Asterisks shown in the figures indicate the p-values as follows: ****P < 0.0001. P < 0.05 was considered to be statistically significant.