The Microtubule-Binding and Coiled-Coil Domains of Kid Are

RESEARCH ARTICLE The microtubule-binding and coiled-coil domains of Kid are required to turn off the polar ejection force at anaphase Shou Soeda1,2,3,4,*, Kaori Yamada-Nomoto1,5,* and Miho Ohsugi1,2,3,*,‡

ABSTRACT throughout mitosis. In addition, the forces generated on the Mitotic chromosomes move dynamically along the spindle chromosome arms contribute to the congression and alignment of microtubules using the forces generated by motor proteins such as the chromosomes during prometaphase and metaphase. The chromokinesin Kid (also known as KIF22). Kid generates a polar chromokinesin Kid (also known as KIF22) is a plus end-directed ejection force and contributes to alignment of the chromosome arms motor protein characterized by a DNA-binding domain (Tokai et al., during prometaphase and metaphase, whereas during anaphase, 1996). During prometaphase and metaphase, Kid is implicated in Kid contributes to chromosome compaction. How Kid is regulated chromosome congression and alignment through generation of a and how this regulation is important for chromosome dynamics remains polar ejection force (PEF) on the chromosome arms (Brouhard and unclear. Here, we address these questions by expressing mutant Hunt, 2005), although the degree of the physiological importance of forms of Kid in Kid-deficient cells. We demonstrate that Cdk1-mediated Kid-mediated PEF varies among species and cell types. Kid in Xenopus phosphorylation of Thr463 is required to generate the polar ejection has been shown to be essential for metaphase chromosome Xenopus force on Kid-binding chromosomes, whereas dephosphorylation alignment in egg extract; however, there is no evidence of of Thr463 prevents generation of the ejection force on such the involvement of Kid in chromosome congression in mouse chromosomes. In addition to activation of the second microtubule- oocytes and zygotes (Funabiki and Murray, 2000; Antonio et al., binding domain through dephosphorylation of Thr463, the coiled-coil 2000; Ohsugi et al., 2008; Kitajima et al., 2011). Moreover, in domain is essential in suspending generation of the polar ejection force, somatic cell lines, mammalian Kid generates the PEF and contributes preventing separated chromosomes from becoming recongressed to driving chromosome movement toward the spindle equator during during anaphase. We propose that phosphorylation of Thr463 switches prometaphase (Levesque and Compton, 2001; Tokai-Nishizumi the mitotic chromosome movement from an anti-poleward direction to et al., 2005; Magidson et al., 2011; Iemura and Tanaka, 2015). – a poleward direction by converting the Kid functional mode from At the metaphase anaphase transition, loss of cohesion between polar-ejection-force-ON to -OFF during the metaphase–anaphase sister chromatids leads to a change in direction toward the spindle transition, and that both the second microtubule-binding domain and poles. In addition to the loss of sister chromatid cohesion, the the coiled-coil domain are involved in this switching process. change of direction of chromosome movement requires a reduction of the PEF, which occurs through degradation of Xenopus Kid KEY WORDS: Cell division, Chromosome dynamics, Kinesin, degradation in Xenopus egg extract (Funabiki and Murray, 2000). In Phosphorylation status contrast, the majority of Kid is maintained during anaphase and contributes to the tight clustering of anaphase chromosomes INTRODUCTION (anaphase chromosome compaction), which prevents the The accuracy of chromosome segregation during mitosis is formation of blastomeres with multiple nuclei in mouse zygotes guaranteed by strictly regulated processes. Failures in orchestrating (Ohsugi et al., 2008). Therefore, the effects of reduced PEF in these processes can result in aneuploidy, tumorigenesis or cell death. anaphase chromosome segregation in mammalian cells remains The dynamic movement of mitotic chromosomes along the spindle is unclear; if a reduction in PEF does occur, it must be achieved by achieved by forces generated by microtubule dynamics and several means other than through protein degradation. motor proteins, including members of the cytoplasmic dynein and Although previous studies have revealed different functions of kinesin superfamilies (Cross and McAinsh, 2014). Forces acting on Kid, how transitions between these multiple functions are regulated the kinetochores through the attached microtubules are the primary remains unclear. One plausible regulatory mechanism is Cdk1– and indispensable driving forces for chromosome movement cyclin-B-mediated phosphorylation of Thr463 on Kid (Ohsugi et al., 2003). Besides the motor domain (the first microtubule- binding domain), Kid has a second microtubule-binding domain 1 Division of Oncology, Institute of Medical Science, The University of Tokyo, Minato- (MTBD) between the motor and coiled-coil (CC) domains ku, Tokyo 108-8639, Japan. 2Department of Life Science, Graduate school of Arts and Sciences, The University of Tokyo, Meguro-ku, Tokyo 153-8902, Japan. (Shiroguchi et al., 2003). We have previously shown that the 3Department of Biophysics and Biochemistry, Graduate school of Science, The microtubule-binding activity of Kid mediated by the MTBD is University of Tokyo, Bunkyo-ku, Tokyo 113-0032, Japan. 4Japan Society for the Promotion of Science Research Fellow, Chiyoda-Ku, Tokyo 102-0083, Japan. regulated negatively by phosphorylation on Thr463 (Ohsugi et al., 5Department of Obstetrics and Gynecology, The University of Toyama, Toyama-shi, 2003). Amino acid substitution of Thr463 to Ala increases the Toyama 930-0194, Japan. affinity of Kid for microtubules, leading to its sequestration to *Present address: Department of Life Science, Graduate school of Arts and Sciences, The University of Tokyo, Meguro-ku, Tokyo, 153-8902, Japan. spindle microtubules and a failure in chromosomal localization of the protein during prometaphase and metaphase. This indicates that ‡ Author for correspondence ([email protected]) phosphorylation on Thr463 is essential for the proper localization of M.O., 0000-0003-0288-3428 Kid on chromosomes (Ohsugi et al., 2003). However, whether phosphorylation of Thr463 is essential for Kid to generate PEF, and,

Received 30 March 2016; Accepted 17 August 2016 conversely, whether dephosphorylation of Thr463 is sufficient to Journal of Cell Science

3609 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3609-3619 doi:10.1242/jcs.189969 suppress PEF remain unproven. Many kinesin-family members of anaphase. To address this possibility, we used mouse oocytes, form oligomers through the CC domain when they move along the which offer several advantages. First, oocytes are arrested at microtubules. However, Kid is purified as a monomer from cells metaphase of the second meiosis (metaphase II), and the arrested in prometaphase (Shiroguchi et al., 2003) and is known to metaphase–anaphase transition can be induced artificially through be motile as a monomer (Yajima et al., 2003). The CC domain of parthenogenetic activation. Second, the importance of Kid during Kid has been shown to be important for the regulation of spindle anaphase and telophase is prominent (Ohsugi et al., 2008). length (Tokai-Nishizumi et al., 2005), but its involvement in PEF Moreover, exogenous protein expression is achieved easily in generation has remained obscure. metaphase-II-arrested oocytes by the injection of mRNA that has In the current study, we expressed a series of Kid mutants in Kid- been transcribed in vitro. deficient cells to analyze the importance of each domain and the To visualize chromosomes, mRFP-fused histone H2B was phosphorylation status of Thr463 for Kid functions, as well as in expressed in oocytes from mice wild-type for Kid or in which Kid switching the direction of chromosome movement at the had been knocked out (Kid-KO mice). In addition to H2B-mRFP, metaphase–anaphase transition. GFP-tagged wild-type or mutant forms of Kid were expressed in Kid-KO oocytes. The expression levels of GFP-tagged Kid proteins RESULTS were almost equivalent to each other (Fig. S1A). Oocytes were Thr463 phosphorylation, but not the CC domain of Kid, is subjected to live-cell imaging immediately after parthenogenetic essential for generating PEF activation. In wild-type oocytes, one set of segregated sister In order to clarify whether phosphorylation of Thr463 (Fig. 1A) of chromatids was extruded from oocytes as a second polar body, and Kid was required for the generation of PEF as well as for its proper the other formed a single female pronucleus (fPN) (Fig. 2A and B; chromosomal localization, we determined whether a mutant of Kid Movie 1). In contrast, although chromosomes were segregated and a in which Thr463 was replaced by Ala (Kid-T463A) generated PEF. polar body was emitted, about 75% of the Kid-KO oocytes In the absence of Kid, chromosomes on a monopolar spindle are exhibited fragmentation of the fPN (Fig. 2A and C), as reported centered (Levesque and Compton, 2001). Taking advantage of this previously (Ohsugi et al., 2008). Hereafter, we refer to this as the effect, we expressed GFP, GFP–Kid-WT (GFP-tagged wild-type multi-fPN phenotype. Exogenously expressed GFP–Kid-WT construct) or GFP–Kid-T463A in Kid−/− mouse embryonic almost completely rescued the multi-fPN phenotype of Kid-KO fibroblasts and cultured these cells with or without an inhibitor of oocytes, whereas the Kid mutant lacking the DNA-binding domain Eg5 (also known as kinesin-5 and KIF11) (Fig. 1B and C). (GFP–Kid-ΔDBD), which failed to localize on anaphase Chromosome distributions on monopolar spindles in Eg5-inhibitor- chromosomes, did not (Fig. 2A,D and E). This demonstrated that treated cells were examined by examining Hoechst signal intensities GFP-tagged human Kid protein compensates for the lack of along a 12-µm line centered on a spindle pole in GFP-positive endogenous mouse Kid protein in oocytes. monopolar mitotic cells (Fig. 1D). In Kid−/− cells expressing GFP We examined the importance of dephosphorylation of Thr463 in alone or GFP–Kid-T463A, which was preferentially localized on the post-metaphase processes by expressing phosphorylation- a spindle (Fig. 1B and C), chromosomes were centered on a defective (T463A) or phosphorylation-mimetic (T463D) mutants monopolar spindle. In contrast, chromosomes were ejected from a of Kid in Kid-KO oocytes. GFP–Kid-T463A accumulated on pole of a monopolar spindle in cells that could be rescued with anaphase chromosomes and rescued the multi-fPN phenotype as GFP–Kid-WT (Fig. 1B and E). Because Kid also contributes to the effectively as GFP–Kid-WT (Fig. 3A and B). This suggested that maintenance of spindle size during metaphase (Tokai-Nishizumi phosphorylation of Thr463 is not necessary for Kid function after et al., 2005), we also compared spindle sizes in these cells and found the onset of anaphase. In contrast, when Kid-KO oocytes expressing no significant differences (Fig. 1F). These results indicate that GFP–Kid-T463D entered anaphase, chromosomes initially moved differences in chromosome distribution were not due to differences toward each pole, but then a portion of or all separated in the spindle length, and that phosphorylation at Thr463 is essential chromosomes moved back to the spindle equator (Fig. 3A and C; for Kid to generate PEF during prometaphase. Movie 2). When one or more anaphase chromosomes moved toward We next sought to examine the involvement of the MTBD and the spindle equator, we classified the oocytes as undergoing CC domain in PEF generation. To this end, we constructed Kid ‘recongression’. The anti-poleward movement of anaphase mutants that lacked part of the MTBD (amino acids 445–451, chromosomes was never observed in Kid-KO oocytes, suggesting ΔMTBD) or CC domain (amino acids 471–496, ΔCC). Deletion of that this chromosome-recongression phenotype was distinct from the MTBD led to preferential localization on chromosomes, even fragmentation of the fPN caused by the loss of Kid. Furthermore, the with the T463A mutation (Fig. 1B and C). This finding is consistent expression of GFP–Kid-ΔMTBD in Kid-KO oocytes resulted in a with our previous results that, in the absence of Thr463 recongression of anaphase chromosomes, even when Thr463 was phosphorylation, the MTBD enhances the affinity of Kid for replaced with Ala (Fig. 3A,D and E; Movie 3). These results suggest microtubules (Ohsugi et al., 2003; Shiroguchi et al., 2003), whereas that dephosphorylation of Thr463 activates the MTBD, which is GFP–Kid-ΔCC showed a similar localization to that of GFP–Kid- essential for chromosome segregation during anaphase. WT (Fig. 1B and C). Chromosomes were ejected from the pole of monopolar spindles under expression of either protein (Fig. 1G). The phosphorylation status of Thr463 acts as a switch to These results indicate that the CC domain is dispensable for Kid- convert Kid between the PEF-ON and PEF-OFF modes mediated PEF generation, and that PEF generation is also dependent We further characterized the anaphase chromosome phenotypes on the suppression of the MTBD by phosphorylation of Thr463. induced by GFP–Kid-T463D and GFP–Kid-ΔMTBD. Despite the fact that Kid generates PEF on prometaphase–metaphase Activation of the MTBD through dephosphorylation of Thr463 chromosomes in somatic cells, Kid is dispensable for chromosome is required for chromosome segregation in anaphase alignment during oocyte meiosis (Ohsugi et al., 2008; Kitajima The results described above raise the possibility that the MTBD and et al., 2011). Expression of neither the GFP-tagged Kid-WT nor the

CC domain are essential for Kid to function during or after the onset mutant forms caused the redistribution of already aligned Journal of Cell Science

3610 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3609-3619 doi:10.1242/jcs.189969

Fig. 1. See next page for legend. Journal of Cell Science

3611 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3609-3619 doi:10.1242/jcs.189969

Fig. 1. Phosphorylation of Thr463 is required for Kid to generate anaphase, dephosphorylated Kid not only halts the generation of PEF. (A) Schematic diagram of the functional domains of the Kid protein. − − PEF but also antagonizes the PEF generated by these Kid mutants. (B,C) Kid / mouse embryonic fibroblasts expressing GFP-tagged Kid-WT, Kid-T463A, Kid-ΔMTBD, Kid-ΔMTBD-T463A, Kid-ΔCC or GFP alone were treated with (left panels) or without (right panels) Eg5 inhibitor III, fixed and The CC domain is required to convert unphosphorylated Kid stained for α-tubulin, GFP and DNA (Hoechst). Representative images to the PEF-OFF mode (deconvolved and z-stacked) are shown. Scale bars: 10 µm. GFP signal Next, we examined the importance of the CC domain after the onset intensities on chromosomes relative to those on microtubules were measured. of anaphase. Kid-KO oocytes expressing either the Kid mutant All of the quantification measurements were performed on 15 cells. Error bars lacking the CC domain (Kid-ΔCC) or with a disrupted CC domain P P ’ represent mean±s.e.m. ** <0.01; n.s., not significant ( >0.05), Student s (Kid-8D or 7S, in which Val475, Glu476, Glu477, Lys478, Glu491, t-test. (D,E) The signal intensity of Hoechst staining was measured in 30 cells along a 12-µm line centered on a spindle pole of z-projected images of GFP- Lys492, Glu493, Ala494' were replaced with Asp or `Val475, positive monopolar cells. All data points were normalized to their maximum Glu476, Glu477, Lys478, Lys492, Glu493, Ala494' were replaced value, and the average was plotted. Error bars represent mean±s.e.m. with Ser residues, respectively; see Fig. S3; Lupas, 1996) entered (F) Spindle microtubule length judged by using α-tubulin staining. The graph anaphase. However, the segregated chromosomes congressed again shows each value and box-and-whisker plots. The top and bottom of the box to the spindle equator, similar to in Kid-KO oocytes that expressed show upper and lower quartile values, respectively. The median is indicated Kid-T463D or Kid-ΔMTBD (Fig. 4A,B; Movie 4, Fig. S3A–D). with line inside the box and the whiskers denote the range of the data. (G) The The rigor-inducing mutation also bypassed the requirement of the signal intensity of Hoechst staining was measured in 30 cells expressing GFP–Kid-ΔMTBD, GFP–Kid-ΔMTBD-T463A, or GFP–Kid-ΔCC and analyzed CC domain in order to arrest PEF generation (Fig. 4A and C). These as described in E. MT, microtubule. results indicate that the CC domain of Kid is required to halt PEF generation during anaphase. Because Thr463 is located close to the CC domain, it is possible metaphase-II chromosomes, suggesting that the recongression that loss of the CC structure interferes with dephosphorylation of phenotype caused by GFP–Kid-T463D or GFP–Kid-ΔMTBD was Thr463 and that the mutants act as constitutively phosphorylated not due to their influence on the chromosomes during the phases of forms, similar to Kid-T463D. To test this possibility, we meiosis that occur before anaphase II. We also addressed the constructed and expressed a phosphorylation-defective form of possibility that the recongression phenotype had resulted from Kid-ΔCC (Kid-T463A-ΔCC) in Kid-KO oocytes. Although defective spindle elongation during anaphase caused by the Kid GFP–Kid-ΔCC showed no chromosomal accumulation during mutants. To this end, we expressed mCherry–α-tubulin in oocytes, anaphase, GFP–T463A-ΔCC showed preferred localization to the and measured the pole-to-pole distance of the spindle during chromosomes rather than to the microtubules (Fig. 4B and D). The metaphase II and anaphase II. In this assay, oocytes were treated with difference in localization between these two mutants suggests that an inhibitor of actin polymerization to prevent the anaphase spindle the deletion of the CC domain affects T463 dephosphorylation. from being rotated and bent during chromosome segregation and However, the anaphase chromosome-recongression phenotype polar body emission. Under this condition, the spindles elongated was still observed in oocytes expressing Kid-T463A-ΔCC rapidly, reached their maximum length within 10 min after the onset (Fig. 4A and D), indicating that the recongression phenotype of anaphase, and then shortened gradually (Fig. S1B). The spindle- caused by Kid-ΔCC was not dependent on Thr463 phosphorylation. length kinetics were not significantly affected by the expression of This suggests that, in addition to the MTBD, the CC domain Kid mutants (Fig. S1B and C), indicating that the recongression is essential for unphosphorylated Kid to be set to the PEF-OFF phenotype could not be attributed to defective anaphase spindle mode. elongation. Upon the onset of anaphase, the chromosomes were Immunofluorescence analysis demonstrated that while Kid-WT, pulled apart toward the poles as the spindle elongated. However, the Kid-T463D, and Kid-ΔMTBD showed a preferential localization on chromosomes kept moving and became fully segregated in Kid-KO anaphase chromosomes, Kid-ΔCC localized on both central spindle oocytes expressing Kid-WT, whereas the chromosomes in oocytes microtubules and chromosomes (Fig. S3E). To examine whether expressing Kid-T463D or Kid-ΔMTBD traveled only a short Kid-ΔCC used MTBD to ectopically localize on the spindle, the distance and then began to move back to the spindle equator as double deletion mutant Kid-ΔMTBD-ΔCC was constructed and the central microtubules were beginning to develop between the expressed in Kid-KO oocytes. GFP-Kid-ΔMTBD-ΔCC showed segregated chromosomes (Fig. S1D). These results suggest that similar localization as Kid-ΔCC and caused the recongression the chromosomes were pulled apart by the forces generated by the phenotype (Fig. 4B and E). Furthermore, introducing the rigor-type spindle; however, they were then recongressed by Kid-T463D or mutation, which enhances the affinity for microtubules, restored the Kid-ΔMTBD along the central spindle. When motor activity was localization of Kid-ΔCC (Fig. 4C). These results suggest that the suppressed by introducing a rigor-inducing mutation (Thr34 to mislocalization of Kid-ΔCC was not due to enhanced affinity for Asn substitution) into Kid-ΔMTBD (Meluh and Rose, 1990; microtubules. Together, these results indicate that the CC domain Ohsugi et al., 2003), the resulting GFP–Kid-Rigor-ΔMTBD did may facilitate or maintain Kid accumulation on anaphase not induce the anaphase chromosome-recongression phenotype chromosomes by controlling the association between the DNA- (Fig. 3A and F). This result supports the model that the chromosome- binding domain of Kid and the chromosomes, thereby enabling recongressing force is driven by Kid-ΔMTBD motor activity, dephosphorylated Kid to function in anaphase chromosome although we cannot exclude the possibility that enhanced compaction but not in PEF generation. microtubule-binding affinity resulting from the rigor-inducing Previous in vitro assays showed that the presence of the MTBD mutation bypasses the requirement for the MTBD. The expression and CC domain enhanced the affinity of Kid for the same of GFP–Kid-T463D in Kid-KO zygotes also resulted in the anaphase microtubule along which the motor domain moved, but did not chromosome-recongression phenotype at the first cleavage division reduce the velocity (Shiroguchi et al., 2003). Therefore, it is likely (Fig. S2A). Furthermore, the anaphase chromosome-recongression that Thr463-unphosphorylated Kid retains motility but cannot phenotype was rarely observed in wild-type oocytes that expressed generate sufficient force on the binding chromosomes to transport

Kid-T463D or Kid-ΔMTBD (Fig. S2B). This suggests that, during them along the microtubule. Together, these findings suggest that Journal of Cell Science

3612 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3609-3619 doi:10.1242/jcs.189969

− − − − Fig. 2. Expression of GFP–Kid-WT rescues the multi-fPN phenotype in Kid / oocytes. Kid / or Kid+/+ mouse oocytes during meiosis II and expressing H2B–mRFP or GFP-tagged proteins were subjected to parthenogenetic activation and live-cell imaging. (A) The results of expression experiments using GFP-tagged BAF (barrier to autointegration factor) (-) [n=47 oocytes (+/+), 102 oocytes (−/−)], Kid-WT (WT, n=105 oocytes) or Kid-ΔDBD (ΔDBD, n=46 oocytes). The numbers in the bars indicate the number of oocytes that were classified into each phenotype group. The x-axis indicates these as percentages of the total number of cells. (B–E) Time-lapse images of representative oocytes from the experiment described in A. The numbers above each image indicate the time after the onset of anaphase (min). Scale bars: 25 µm. The contrast in H2B–mRFP images was adjusted by linear transformation (peak-enhanced). The arrowheads indicate the second polar bodies. the MTBD and CC domain are the key domains for switching Kid oocytes in Kid-KO oocytes (Fig. 2A), chi-squared test] or Kid- from the PEF-ON to the PEF-OFF mode upon the onset of anaphase Rigor-ΔCC [P<0.01 (Fig. 4A) with control (−) in Kid-KO oocytes in response to the phosphorylation status of Thr463. (Fig. 2A), chi-squared test]. To clarify the contribution of the MTBD, CC and motor domains to Kid-mediated fPN formation, we The motor domain but not motor activity is essential for Kid expressed GFP-tagged Kid-Rigor (containing only the rigor- to ensure formation of a single fPN inducing mutation) or Kid-ΔMot (lacking the Kid motor domain) When GFP–Kid-Rigor-ΔMTBD or GFP–Kid-Rigor-ΔCC was constructs into Kid-KO oocytes. Despite similar localization on expressed, only ∼44% of Kid-KO parthenogenetic zygotes anaphase chromosomes, oocytes expressing GFP–Kid-ΔMot showed the multi-fPN phenotype (Figs 3A and 4A). This eventually formed multi-fPNs at a high frequency (73%) (Fig. 5A indicates that the multi-fPN phenotype was rescued partially by and B) [P=0.84 with control (−) BAF in Kid-KO oocytes (Fig. 2A),

Kid-Rigor-ΔMTBD [P<0.001 (Fig.3A) with control (−) in Kid-KO chi-squared test]. In contrast, expression of GFP–Kid-Rigor in Journal of Cell Science

3613 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3609-3619 doi:10.1242/jcs.189969

Fig. 3. Activation of the MTBD of Kid is required for Kid to suspend PEF generation. (A) The results of Kid-mutant expression experiments with GFP-tagged Kid-T463A (T463A, n=40 oocytes), Kid-T463D (T463D, n=38 oocytes), Kid-ΔMTBD (ΔMTBD, n=25 oocytes), Kid-ΔMTBD-T463A (ΔMTBD-T463A, n=26 oocytes) and Kid-Rigor-ΔMTBD (Rigor-ΔMTBD, n=29 oocytes). All the oocytes that showed chromosome recongression were classified as having the chromosome-recongression phenotype. The other oocytes that did not show the chromosome-recongression phenotype were further categorized into two groups – multi fPN or single fPN. The numbers in the bars indicate the number of oocytes that were classified into each phenotype group. The x-axis indicates these as percentages of the total number of cells. (B–F) Time-lapse images of representative oocytes from the experiments shown in A. The numbers above each image indicate the time after the onset of anaphase (min). Scale bars: 25 µm. H2B–mRFP signals were peak-enhanced. The arrowheads indicate the second polar bodies. Journal of Cell Science

3614 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3609-3619 doi:10.1242/jcs.189969

Fig. 4. The CC domain is required to suspend PEF generation. (A) The results of Kid-mutant expression experiments with GFP-tagged Kid-ΔCC (ΔCC, n=27 oocytes), Kid-CC-8D (CC-8D, n=25 oocytes), Kid-Rigor-ΔCC (Rigor-ΔCC, n=32 oocytes), Kid-T463A-ΔCC (T463A-ΔCC, n=40 oocytes) and Kid-ΔMTBD-ΔCC (ΔMTBD-ΔCC, n=22 oocytes). Representative images of CC-8D are shown in Fig. S3D. The numbers in the bars indicate the number of oocytes that were classified into each phenotype group. The x-axis indicates these as percentages of the total number of cells. (B–E) Time-lapse images of representative cells from the experiments shown in A. The numbers above each image indicate the time after the onset of anaphase (min). Scale bars: 25 µm. H2B-mRFP signals are peak- enhanced. The arrowheads indicate the second polar bodies. oocytes reduced the formation of multi-fPNs to 37% (Fig. 5A and phosphorylation status of Thr463 of Kid functions as a switching C) [P<0.001 with control (−) BAF in Kid-KO oocytes (Fig. 2A), mechanism to alter the functional modes (PEF-ON and PEF-OFF) chi-squared test]. These results indicate that the motor domain, but of Kid on mitotic chromosomes (Fig. 5D). Based on our results, and not motor activity, is essential for Kid to ensure formation of a single previous knowledge, we propose the following model (Fig. 5D). fPN. Because there were no significant differences in the formation (1) During prometaphase and metaphase, Cdk1–cyclin-B of multi-fPNs among zygotes expressing Kid-Rigor, Kid-Rigor- phosphorylates Kid on Thr463. This phosphorylation suppresses ΔMTBD or Kid-Rigor-ΔCC (P=0.72, chi-squared test), the the MTBD-mediated affinity of Kid for microtubules, and enables significance of the MTBD and CC domain in Kid-mediated Kid to localize and generate PEF on the chromosome arms. (2) anaphase chromosome compaction remain unclear. Upon the onset of anaphase, Thr463 is dephosphorylated and the In conclusion, the results of the current study show that both the MTBD becomes functional. (3) Then, the MTBD together with the MTBD and CC domain are dispensable for PEF generation but are CC domain functions so that Kid can no longer generate PEF but essential for halting PEF generation after the onset of anaphase. The rather resist this force. Journal of Cell Science

3615 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3609-3619 doi:10.1242/jcs.189969

Fig. 5. The motor domain but not motor activity of Kid is essential for the formation of a single fPN. (A) The results of Kid-mutant expression experiments with GFP–Kid-ΔMot (ΔMot, n=48 oocytes) and GFP–Kid-Rigor (Rigor, n=63 oocytes). The numbers in the bars indicate the number of oocytes that were classified into each phenotype group. The x-axis indicates these as percentages of the total number of cells. (B,C) Time-lapse images of representative cells from the experiments shown in A. The numbers above each image indicate the time after the onset of anaphase (min). Scale bars: 25 µm. H2B–mRFP signals were peak- enhanced. The arrowheads indicate the second polar bodies. (D) A model for Kid functional switching at the metaphase/anaphase transition.

DISCUSSION previous report that shows that anaphase chromosomes are pulled In this study, we provide evidence that Cdk1–cyclin-B-mediated back to the spindle equator in a Kid-dependent manner when cells phosphorylation on Thr463 of Kid is essential for generating the retain Cdk1 activity owing to the expression of non-degradable PEF on the chromosome arms in mammalian cells during cyclin B1, (Wolf et al., 2006). Our data further indicate that among prometaphase and metaphase. Furthermore, phosphorylation- the numerous sites on many proteins that would be phosphorylated mimetic substitution of Thr463 to Asp resulted in ectopic by Cdk1–cyclin-B, phosphorylation on Kid Thr463 is sufficient to generation of PEF during anaphase. This is consistent with a generate ectopic PEF on anaphase chromosomes and realign them at Journal of Cell Science

3616 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3609-3619 doi:10.1242/jcs.189969 the spindle equator. Based on these results, we propose that the mediated switching between the ON and OFF modes of Kid- phosphorylation status of Thr463 functions as a switch for Kid- mediated PEF is not accomplished simply through association with mediated PEF generation, and that dephosphorylation promotes the and dissociation from the chromosomes by changing the affinity for dynamic directional change of mitotic chromosome movement from microtubules, but rather by changing the mode of action of Kid. anti-poleward to poleward upon the metaphase–anaphase transition. The CC domain of Kid is shorter than those of other kinesin- Our study provides further support for the concept that, in family motor proteins that function as oligomers (Verhey and addition to sister chromatid disjunction, Kid-mediated PEF must be Hammond, 2009). Although we cannot exclude the possibility that suspended when cells enter anaphase in order to ensure Kid forms an oligomer under certain conditions, such as a high chromosome segregation, despite the fact that Kid-mediated PEF concentration on chromosomes or microtubules, there is no is dispensable for metaphase chromosome alignment in mammalian evidence of CC-mediated oligomerization both in vivo and cells (Levesque and Compton, 2001; Tokai-Nishizumi et al., 2005; in vitro (Ohsugi et al., 2003; Shiroguchi et al., 2003). Unlike the Ohsugi et al., 2008). Expression of a constitutively PEF-generating majority of other motile kinesin-family proteins, the CC domain mutant such as Kid-T463D, Kid-ΔMTBD or Kid-ΔCC in Kid- appears to be dispensable for Kid to transport its cargo (mitotic deficient oocytes induced anaphase chromosome recongression in chromosome) along microtubules. Instead, the CC domain, as well more than 70% of the cells, whereas only 26% or fewer cells showed as the MTBD, was required to halt the congressing force on anaphase chromosome recongression when expressed in wild-type chromosomes after anaphase onset. In addition to the failure in oocytes (Fig. S2B). These results suggest that the cells can tolerate turning off PEF, deletion of the CC domain resulted in the failure of the presence of some phosphorylated Kid after the onset of chromosomal accumulation. Instead, even in anaphase cells, Kid- anaphase. Although termination of PEF is achieved by the ΔCC localized on both the chromosomes and spindles, as in anaphase-promoting complex (APC/C)-proteasome-mediated prometaphase–metaphase cells. Further, because Kid-ΔMTBD- degradation of Xenopus Kid in Xenopus egg extracts (Funabiki ΔCC showed anaphase spindle localization and Kid-Rigor, which and Murray, 2000), in mammalian cells, this termination is achieved shows increased microtubule-binding activity (Ohsugi et al., 2003), indirectly by the APC/C proteasome through degradation of cyclin did not induce preferred localization on anaphase spindles (Fig. 5C; B followed by dephosphorylation of Kid, contributing to the Ohsugi et al., 2003), the mislocalization caused by ΔCC was formation of proper daughter nuclei (Ohsugi et al., 2008). Thus, probably due to decreased affinity for anaphase chromosomes rather APC/C-proteasome-mediated protein degradation both triggers and than increased affinity for microtubules. Therefore, we speculate maintains the poleward movement of anaphase chromosomes by that the CC domain is involved in anaphase chromosome eliminating both sister chromatid cohesion and Kid-mediated PEF. accumulation by changing the mode of association with Because a substantial amount of Xenopus Kid has been reported to chromosomes, from binding them as cargo to binding them in remain on anaphase chromosomes in a Xenopus somatic cell line order to cluster them, after anaphase onset, leading to a difference in (Antonio et al., 2000) and the amino acid sequences of the MTBD the mobility of chromosome-binding Kid between metaphase and and CC domain, as well as Thr463, are highly conserved between anaphase (Tahara et al., 2008). We speculate that the chromosome- Xenopus Kid and mammalian Kid (Shiroguchi et al., 2003), the bound fraction of Kid-ΔCC generates PEF, similar to the action of dephosphorylation-mediated PEF-OFF mechanism could also be the chromosome-bound fraction of Kid-WT on prometaphase– used in Xenopus during non-embryonic cell division. metaphase chromosomes. In addition to the Kid-T463A mutant, other Kid mutants that Although the essential difference of conformation of lacked the MTBD or the CC domain also exhibited the chromosome- microtubule-Kid-chromosome complex between the PEF-ON and recongression phenotype. Although we cannot completely exclude PEF-OFF modes of Kid remains unclear, Kid might bind to mitotic the possibility that the chromosome recongression was due to chromosomes and/or microtubules in distinct ways during defects in other functions of Kid rather than the (re-)activation of metaphase and anaphase as a result of changes in its affinity for PEF, we believe that the most reasonable interpretation of the results microtubules and mitotic chromosomes through the MTBD and the is that these mutations afford constitutive PEF generation even after CC domain, respectively. Further exploration of the function of the anaphase onset, resulting in the recongression phenotype. A CC domain will help to establish the precise mechanism underlying previous study has demonstrated that bacterially expressed, and the functional switching of Kid. Our data also revealed that motility thus unphosphorylated, Kid515GSdC, in which the C-terminal along the microtubule, and thus the PEF generation, was not region containing the DNA-binding domain of Kid is replaced with a essential for Kid function in ensuring nucleus formation in the portion of the gelsolin protein, is motile and transports an actin mouse zygote. Further research should address whether the PEF- fragment along the microtubules (Shiroguchi et al., 2003). An OFF state of Kid is coupled to Kid-mediated anaphase chromosome in vitro study has further demonstrated that the presence or absence compaction. of the amino acid region 442–515, which includes both the MTBD and CC domain, as well as artificial dimerization of the Kid motor MATERIALS AND METHODS domain does not affect the velocity (Shiroguchi et al., 2003; Yajima Cell culture and retrovirus infection − − et al., 2003). Therefore, it is unlikely that the MTBD and CC domain Kid+/+ and Kid / mouse embryonic fibroblasts (MEFs) obtained from act simply as a brake system when Thr463 is dephosphorylated. The embryos at embryonic day (E)14 (Ohsugi et al., 2008) were cultured in failure of PEF generation by Kid-T463A in prometaphase cells Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% could be explained by the reduction of Kid protein on the heat-inactivated fetal calf serum and serially passaged more than 28 times chromosome arms due to sequestration to the spindle microtubules according to the 3T3 protocol (Todaro and Green, 1963), and were then used for experiments. Retrovirus-mediated expression of Kid mutants was (Fig. 1; Ohsugi et al., 2003). However, despite the predominant performed as described previously (Ohsugi et al., 2003). For retrovirus localization on anaphase chromosomes, Kid-WT and Kid-T463A production, pMX-puro vector (Onishi et al., 1996) encoding GFP, GFP– arrested PEF and mediated anaphase chromosome compaction, Kid-WT, GFP–Kid-T463A, GFP–Kid-ΔMTBD, GFP–Kid-ΔMTBD- whereas Kid-T463D and Kid-ΔMTBD ectopically generated PEF T463A or GFP–Kid-ΔCC was transfected into PlatE packaging cells.

(Figs 2B, 3B–D; Fig. S3E). These results imply that MTBD- Cells were incubated for 24 h in retrovirus-containing DMEM, treated with Journal of Cell Science

3617 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3609-3619 doi:10.1242/jcs.189969

100 nM Eg5 inhibitor III (Calbiochem) for an additional 24 h post projected images of α-tubulin, the areas with higher signals than the mean transfection and then fixed and subjected to immunofluorescent staining. value were judged as spindle microtubule regions. The diameters of these regions were measured as monopolar spindle diameters. Collection of oocytes and mRNA injection +/+ −/− Female Kid and Kid mice having BalbC/A×C57BL6/J F1 background Antibodies and staining reagents (6 weeks–6 months old) (Ohsugi et al., 2008) were induced to superovulate Primary antibodies against α-tubulin (Santa Cruz, YL1/2, 1:500 or Sigma- with intraperitoneal injections of 5 IU pregnant mare serum gonadotropin Aldrich, T9026, 1:500), β-tubulin (Sigma-Aldrich, 2-28-33; 1:500) and (PMSG) and 5 IU human chronic gonadotropin (hCG) at 48-h intervals. The GFP (MBL, 598, 1:500 or BioAcademia, 60-001, 1:500) and secondary unfertilized oocytes (MII oocytes) were collected 16–17 h post hCG goat antibodies against mouse IgG (conjugated to Alexa-Fluor-555) administration. The cumulus cells were removed through exposure to M2 (Invitrogen, A-21428, 1:800), rat IgG (conjugated to TRITC) (Invitrogen, medium (Sigma-Aldrich) containing 100 μg/ml hyaluronidase for less than A18870, 1:800), rat IgG (conjugated to Alexa-Fluor-488) (Invitrogen, 5 min, followed by pipetting into fresh M2 medium. The oocytes were A-21208, 1:800) and rabbit IgG (conjugated to Alexa-Fluor-488) cultured in M16 medium (Sigma-Aldrich) at 37°C under 5% CO2 before and (Invitrogen, A-11008, 1:800) were used for immunofluorescent staining. after mRNA injection. The mRNAs were in vitro transcribed using For DNA staining, Hoechst 33342 dye (Invitrogen) was used. RiboMax™ Large Scale RNA Production System-T7 (Promega) 7 supplemented with Ribo m G Cap Analog (Promega), as described Acknowledgements previously (Yamagata et al., 2005). The template plasmids for in vitro We thank K. Yamagata for materials and kind advice on the time-lapse observations mRNA production were constructed by inserting DNA fragments coding of mouse embryos. We also thank Hiroshi Kimura (Tokyo Institute of Technology, GFP–Kid mutants, GFP–BAF, GFP–α-tubulin or H2B–mRFP into Yokohama, Kanagawa, Japan) for the cDNA encoding histone H2B, Toshio pcDNA3.1-polyA vector, which was a gift from Kazuo Yamagata (Kinki Kitamura (The Institute of Medical Sciences, The University Tokyo, Minato-Ku, University, Kinokawa, Wakayama, Japan). mRNAs (10–100 ng/μl) were Tokyo, Japan) for the retrovirus vector systems, and N. Tokai-Nishizumi and T. Yamamoto for helpful discussions. injected into oocytes at metaphase II and cultured for 4 h before the oocytes were subjected to live-cell imaging or fixation. The experiments with Competing interests animals were performed in accordance with the guidelines for animal use The authors declare no competing or financial interests. issued by the Committee of Animal Experiments, Institute of Medical Science, The University of Tokyo. Author contributions S.S. and M.O. designed the research. K.Y.-N. performed expression experiments in Parthenogenetic activation Kid-KO oocytes. S.S. performed all other experiments and data analysis. S.S. and For parthenogenetic activation, oocytes were transferred into activation M.O. wrote the manuscript. medium (ActM; M16 medium containing 5 mM SrCl2 and 5 mM EGTA) with or without 0.25 ng/ml cytochalasin B (Sigma-Aldrich) 4 h after mRNA Funding injection. They were then subjected to live-cell imaging or were cultured at This work was supported by Japan Society for the Promotion of Science KAKENHI [grant number 14J10315 to S.S.]; Ministry of Education, Culture, Sports, Science, 37°C under 5% CO2 until subjected to immunofluorescent staining. and Technology KAKENHI [grant numbers 20055004, 20570161, 26116505 and 15H05971 to M.O.]; the Ministry of Education, Culture, Sports, Science and Live-cell imaging Technology Grant Research Program of Innovative Cell Biology by Innovative mRNA-injected oocytes were transferred into a parthenogenetic activation Technology; and the Precursory Research for Embryonic Science and Technology medium drop and covered with mineral oil in a glass-bottomed dish and program by Japan Science and Technology Agency. observed using a fluorescent microscope (IX70, OLUMPUS, 20×0.85 NA oil objective lens, CoolSNAP HQ camera, Roper Scientific) that was Supplementary information controlled using Delta Vision SoftWorx (Applied Precision) (Fig. 2A; Supplementary information available online at Fig. S2A), or a laser spinning disc confocal microscopy system utilizing a http://jcs.biologists.org/lookup/doi/10.1242/jcs.189969.supplemental Yokogawa CSU22 instrument (IX71, OLUMPUS, 40×1.30 NA oil or 20×0.85 NA oil objective lens, iXon DU897E-CSO-#BV camera, ANDOR) References controlled by Metamorph Software (Universal Imaging) (all figures except Antonio, C., Ferby, I., Wilhelm, H., Jones, M., Karsenti, E., Nebreda, A. R. and Fig. 2A and Fig. S2A ) equipped with a CO2 microscope stage incubator. Vernos, I. (2000). Xkid, a chromokinesin required for chromosome alignment on Samples were scanned from their bottom to their top (21 optical sections, 4 the metaphase plate. Cell 102, 425-435. μm optical section spacing, ActM imaging medium). Brouhard, G. J. and Hunt, A. J. (2005). Microtubule movements on the arms of mitotic chromosomes: polar ejection forces quantified in vitro. Proc. Natl. Acad. Sci. USA 102, 13903-13908. Immunofluorescent staining Cross, R. A. and McAinsh, A. (2014). Prime movers: the mechanochemistry of MEFs were fixed with ice-cold methanol for 10 min at −20°C. The mRNA- mitotic kinesins. Nat. Rev. Mol. Cell Biol. 15, 257-271. injected oocytes were denuded of their zona pellucida with acidic Tyrode’s Funabiki, H. and Murray, A. W. (2000). The Xenopus chromokinesin Xkid is solution, washed with 0.5% polyvinyl pyrrolidone in PBS and fixed with essential for metaphase chromosome alignment and must be degraded to allow ice-cold methanol for 10 min at −20°C. Fixed samples were incubated with anaphase chromosome movement. Cell 102, 411-424. a primary antibody solution and a secondary antibody solution sequentially. Iemura, K. and Tanaka, K. (2015). Chromokinesin Kid and kinetochore kinesin Observations were performed at 25°C with a fluorescent microscope (IX70, CENP-E differentially support chromosome congression without end-on Nat. Commun. OLUMPUS, 100×1.35 NA oil objective lens, CoolSNAP HQ camera, attachment to microtubules. 6, 6447. Kitajima, T. S., Ohsugi, M. and Ellenberg, J. (2011). Complete kinetochore Roper Scientific) that was controlled using Delta Vision SoftWorx (Applied tracking reveals error-prone homologous chromosome biorientation in Precision). GFP-positive MEF samples were scanned from their bottom to mammalian oocytes. Cell 146, 568-581. their top (60 optical sections, 0.2 μm optical section spacing, 80% glycerol Levesque, A. A. and Compton, D. A. (2001). The chromokinesin Kid is necessary with 1% N-propyl gallate imaging medium). For DNA distribution for chromosome arm orientation and oscillation, but not congression, on mitotic measurements, non-deconvolved image stacks were quick-projected, and spindles. J. Cell Biol. 154, 1135-1146. the projected images were analyzed with ImageJ software. For Lupas, A. (1996). Prediction and analysis of coiled-coil structures. Methods quantification of the GFP signal, images were deconvolved, and the Enzymol. 266, 513-525. ’ image stacks were quick-projected. Oocyte samples were scanned from the Magidson, V., O Connell, C. B., Loncarek, J., Paul, R., Mogilner, A. and μ Khodjakov, A. (2011). The spatial arrangement of chromosomes during bottom of their spindles to the top (60 optical sections, 0.2 m optical prometaphase facilitates spindle assembly. Cell 146, 555-567. section spacing, 0.1% bovine serum albumin in PBS imaging medium). Meluh, P. B. and Rose, M. D. (1990). Kar3, a kinesin-related gene required for yeast

Images were deconvolved, and the image stacks were quick-projected. In the nuclear fusion. Cell 60, 1029-1041. Journal of Cell Science

3618 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3609-3619 doi:10.1242/jcs.189969

Ohsugi, M., Tokai-Nishizumi, N., Shiroguchi, K., Toyoshima, Y. Y., Inoue, J.-I. Tokai, N., Fujimoto-Nishiyama, A., Toyoshima, Y., Yonemura, S., Tsukita, S., and Yamamoto, T. (2003). Cdc2-mediated phosphorylation of Kid controls its Inoue, J. and Yamamoto, T. (1996). Kid, a novel kinesin-like DNA binding distribution to spindle and chromosomes. EMBO J. 22, 2091-2103. protein, is localized to chromosomes and the mitotic spindle. EMBO J. 15, Ohsugi, M., Adachi, K., Horai, R., Kakuta, S.,Sudo,K.,Kotaki,H.,Tokai-Nishizumi, 457-467. N.,Sagara,H.,Iwakura,Y.andYamamoto,T.(2008). Kid-mediated chromosome Tokai-Nishizumi, N., Ohsugi, M., Suzuki, E. and Yamamoto, T. (2005). The compaction ensures proper nuclear envelope formation. Cell 132, 771-782. chromokinesin kid is required for maintenance of proper metaphase spindle size. Onishi, M., Kinoshita, S., Morikawa, Y., Shibuya, A., Phillips, J., Lanier, L. L., Mol. Biol. Cell 16, 5455-5463. Gorman, D. M., Nolan, G. P., Miyajima, A. and Kitamura, T. (1996). Applications Verhey, K. J. and Hammond, J. W. (2009). Traffic control: regulation of kinesin of retrovirus-mediated expression cloning. Exp. Hematol. 24, 324-329. motors. Nat. Rev. Mol. Cell Biol. 10, 765-777. Shiroguchi, K., Ohsugi, M., Edamatsu, M., Yamamoto, T. and Toyoshima, Y. Y. Wolf, F., Wandke, C., Isenberg, N. and Geley, S. (2006). Dose-dependent effects (2003). The second microtubule-binding site of monomeric kid enhances the of stable cyclin B1 on progression through mitosis in human cells. EMBO J. 25, microtubule affinity. J. Biol. Chem. 278, 22460-22465. 2802-2813. Tahara, K., Takagi, M., Ohsugi, M., Sone, T., Nishiumi, F., Maeshima, K., Yajima, J., Edamatsu, M., Watai-Nishii, J., Tokai-Nishizumi, N., Yamamoto, T. Horiuchi, Y., Tokai-Nishizumi, N., Imamoto, F., Yamamoto, T. et al. (2008). and Toyoshima, Y. Y. (2003). The human chromokinesin Kid is a plus end- Importin-beta and the small guanosine triphosphatase Ran mediate chromosome directed microtubule-based motor. EMBO J. 22, 1067-1074. loading of the human chromokinesin Kid. J. Cell Biol. 180, 493-506. Yamagata, K., Yamazaki, T., Yamashita, M., Hara, Y., Ogonuki, N. and Ogura, A. Todaro,G.J.andGreen,H.(1963). Quantitative studies of the growth of mouse embryo (2005). Noninvasive visualization of molecular events in the mammalian zygote. cells in culture and their development into established lines. J. Cell Biol. 17, 299-313. Genesis 43, 71-79. Journal of Cell Science

3619