Endogenous Localizome Identifies 43 Mitotic Kinesins in a Plant Cell

Endogenous localizome identifies 43 mitotic kinesins PNAS PLUS in a plant cell

Tomohiro Mikia,b, Haruko Naitoa, Momoko Nishinaa, and Gohta Goshimaa,b,1

aDivision of Biological Science, Graduate School of Science, Nagoya University, Nagoya 464-8602, Japan; and bMarine Biological Laboratory, Woods Hole, MA 02543

Edited by Ronald D. Vale, Howard Hughes Medical Institute and University of California, San Francisco, CA, and approved February 11, 2014 (received for review June 14, 2013) Kinesins are microtubule (MT)-based motor proteins that have available (12). Critically, the rate of homologous recombination been identified in every eukaryotic species. Intriguingly, land is unusually high in P. patens, such that gene disruption and plants have more than 60 kinesins in their genomes, many more epitope tagging are easy to perform. GFP or other fluorescent than that in yeasts or animals. However, many of these have not proteins can be fused to the endogenous protein, so that the yet been characterized, and their cellular functions are unknown. tagged proteins are expressed under the native promoter at the Here, by using endogenous tagging, we comprehensively deter- native chromosomal locus. P. patens spends most of its life cycle mined the localization of 72 kinesins during mitosis in the moss in a haploid state; therefore, the tagged protein substitutes the Physcomitrella patens. We found that 43 kinesins are localized to untagged protein (13, 15). The GFP-based localization deter- mitotic structures such as kinetochores, spindle MTs, or phragmo- mined in this way would thus likely represent the endogenous plasts, which are MT-based structures formed during cytokinesis. protein localization. Another key advantage of the P. patens Surprisingly, only one of them showed an identical localization system is the ease of live cell microscopy in caulonemal stem cells pattern to the animal homolog, and many were enriched at un- that actively divide every 5–6 h (13, 15). Caulonemal cells are expected sites. RNA interference and live-cell microscopy revealed found in protonemata, which are filamentous structures that ap- postanaphase roles for kinesin-5 in spindle/phragmoplast organi- pear after spore germination, and experience less autofluorescence- zation, chromosome segregation, and cytokinesis, which have not related interference derived from chloroplasts; therefore, they been observed in animals. Our study thus provides a list of MT- are suitable for the live imaging of GFP- or red fluorescent based motor proteins associated with the cell division machinery protein (RFP)-tagged markers during cell division (14, 19, 20). in plants. Furthermore, our data challenge the current generaliza- Previous studies have characterized the process of mitotic spin- tion of determining mitotic kinesin function based solely on studies using yeast and animal cells. dle and phragmoplast formation in caulonemal cells by using GFP-tubulin, MAP65-GFP (anti-parallel MT-bundling protein), or EB1-GFP (MT plus-end-tracking protein) (14, 19, 22). Cau- ammalian kinesins have been classified into 14 subfamilies – lonemal cells, like many other plant cell types, do not have Maccording to sequence similarities (kinesin-1 kinesin-14) centrosomes or other discrete MT organizing centers; instead, (1, 2). A subset of kinesins plays critical roles during cell division, they generate MTs predominantly through amplification medi- where spindle bipolarity establishment, chromosome alignment ated by a protein complex called augmin (14, 23). Moreover, to the metaphase plate, spindle microtubule (MT) sizing, and unlike most other cell types in land plants, caulonemal cells do cytokinesis are ensured by the coordinated actions of kinesins, not form the preprophase band (PPB), a MT-based structure including the kinesin-4, -5, -6, -7, -8, -10, -12, -13, and -14 sub- that transiently appears at the early stages of mitosis in the cell families. These mitotic kinesins have been identified through cortex and determines the future division plane (24–26). Thus, functional analyses and localization studies in several organ- mitotic spindle/phragmoplast assembly in caulonemal cells appears isms. For example, GFP-tagging revealed four of six yeast to be a highly self-organized process that involves multiple force- kinesins localized to the kinetochores (3). Comprehensive generating factors; these features are also typically observed in RNA interference (RNAi) showed that 8 of 25 kinesins play acentrosomal spindles in animals (8, 27). a role in mitosis in the Drosophila S2 cell line (4). In humans, ∼20 of 45 kinesins have been identified as mitotic kinesins by RNAi and localization analyses using several cell types (2, 5, Significance 6). Most kinesins exert forces toward MTs or chromosomes; therefore, the identification of kinesins localized to the mi- It has been a mystery as to why plants have many more > totic apparatus has significantly contributed to the under- kinesins than do animals (e.g., 45 in humans, 60 in land standing of the mechanisms underlying mitosis, such as spindle plants). To address this issue, the experimental analysis of each bipolarity establishment, spindle elongation, or chromosome move- kinesin is critical. However, it has not been trivial to analyze > ment (7–9). 60 kinesins in a comprehensive manner in plants. Here, by Interestingly, recent whole genome sequencing has revealed using endogenous YFP tagging, we found that as many as 43 kinesins are localized to the mitotic apparatus in the moss that land plants possess many more kinesins in their genomes than Physcomitrella patens do animals. Arabidopsis thaliana has 61 kinesins, some of which . All but one kinesin showed a localization pattern distinct from the animal homologues, indicating cannot be assigned to the conventional 14 subfamilies (2, 10, 11). the necessity of studying individual kinesin functions in plants. Why they have so many kinesins is an open question, and this

question cannot be addressed without experimentation. However, Author contributions: T.M. conceived the project; T.M. and G.G. designed research; T.M., it has been challenging to analyze more than 60 kinesins in a com- H.N., and M.N. performed research; T.M., H.N., M.N., and G.G. analyzed data; and T.M. prehensive manner in plants. To date, we have limited knowledge and G.G. wrote the paper. on the functions of the individual plant kinesins or a global func- The authors declare no conflict of interest. tional view of the plant kinesin superfamily. This article is a PNAS Direct Submission. The moss Physcomitrella patens is an emerging model land 1To whom correspondence should be addressed. E-mail: [email protected].

plant, for which a variety of molecular techniques can be used This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. PLANT BIOLOGY (12–21). It has 27 chromosomes, and the genome sequence is 1073/pnas.1311243111/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1311243111 PNAS | Published online March 3, 2014 | E1053–E1061 Downloaded by guest on September 28, 2021 In this study, we attempted to identify mitotic kinesins by determining the localization of all of the endogenous kinesins during mitosis of caulonemal cells in P. patens.Among60 kinesins for which expression was detected, 43 exhibited chromosome, spindle, or phragmoplast localization. Surprisingly, all but one kinesin showed a localization pattern distinct from the animal homologs. We further identified postanaphase mitotic functions of kinesin-5 by using RNAi. Results Overview of P. patens Kinesins. In this article, we followed the nomenclatures proposed by Shen et al. (28), who surveyed kinesin gene sequences in the P. patens genome and named individual kinesins based on ref. 1. Our independent BLAST search was ba- sically consistent with Shen et al. For example, kinesin-1, -3, -6, -10, and -11 were missing in the genome of P. patens. However, we additionally found two kinesin-like genes that were not listed in (28) (we named them kinesin 12-Ip and kinesin 12-IId). In total, we identified 78 kinesins, 69 of which could be classified into 15 subfamilies (kinesins 1–14 and kinesin-ARK). The remaining nine kinesins that did not belong to any subfamilies were called “orphan” kinesins (28) (Tables S1 and S2).

Endogenous Kinesin Localization During Mitosis. We attempted to place the citrine [a GFP variant (29)] gene in-frame at the C-terminal end of all 78 kinesins (Fig. 1 A and B). We chose the kinesin C terminus for Citrine tagging, because N-terminal tagging is technically more difficult. One caveat of our strategy, however, is that tagging of Citrine to the C terminus might perturb the kinesin structure, and therefore mislocalization may be induced. Although this may be the case, this approach enables a first-pass genome-wide approach for analyzing kinesin localization. Additionally, C-terminal tagging has been successful for Fig. 1. Flowchart of mitotic kinesin localizome. Experimental procedure to many kinesins in a number of different systems (3, 6, 30–32). comprehensively determine the mitotic localization of endogenous kinesins Using PCR, we confirmed proper alteration of the kinesin loci in P. patens.(A) The mCherry-tubulin–expressing cell line was used as the via homologous recombination for 69 kinesins (Fig. 1C, Left). parental cell line. A cell in metaphase is displayed (red, mCherry-tubulin; For these, we selected two independent transgenic lines for the green, autofluorescence of chloroplasts). (B) The C terminus of the endog- localization analysis in caulonemal cells (Fig. 1D). Endogenous enous kinesin gene was tagged with the citrine gene by homologous re- tagging could not been confirmed for three kinesins (kinesin combination. The citrine gene was followed by the terminator of Agrobacterium tumefaciens nopaline synthase and the G418-resistant cassette (nptII). Black 4-Id, 12-Ij, and 12-IIb), because two sets of the PCR primers we – boxes, untranslated region; white boxes, exons; black lines, introns; red designed were not functional. For those lines, 10 13 indepen- arrows, PCR primers used for strain confirmation. (C) Strain confirmation by dent lines were observed; given the high efficiency of homolo- PCR and immunoblotting. The representative images of gel electrophoresis gous recombination, we expected that Citrine would be inserted after PCR (Left) and anti-GFP immunoblotting (Right) are displayed (note at the designed location for at least one line for these genes. For that anti-GFP can recognize Citrine protein). Arrows indicate the predicted those lines of which Citrine expression was detected by micros- molecular weight of the Citrine fusion proteins, and asterisks indicate non- copy, we performed immunoblotting against the whole cell ex- specific bands. (D) Live imaging of caulonemal cells with spinning-disk con- tract to ensure that the signals were not derived from the focal microscopy. A cell expressing KINID1a/kinesin Orph-IIb-Citrine (green) truncated proteins (Fig. 1C, Right). In total, we performed time- and mCherry-tubulin (red) is displayed as an example. Time 0 corresponds to NEBD. (Scale bar, 10 μm.) lapse imaging of the Citrine lines for 72 kinesins. We could not obtain Citrine lines for the remaining six kinesin-like proteins, because the database did not indicate their C termini. were rare; only four kinesins were found attached to a complete We present each kinesin localization during mitosis by pro- chromosome or kinetochore (colocalization of kinesin 7-III with viding time-lapse movies (Movie S1 presents the complete dataset, a conserved kinetochore protein Mis12 is shown in Fig. S1). whereas the localization of individual kinesins is displayed in Most of the kinesins were observed to be diffused in the cyto- higher resolution as Movies S2–S11). The results are summa- plasm during prophase; however, because of substantial auto- rized in Figs. 2 and 3, and Table S3. Some kinesins showed specific localization during interphase (e.g., cell tip); however, fluorescence caused by chloroplasts in the cytoplasm, it was these data are not presented in this manuscript. Overall, 40 difficult to determine the precise localization of these kinesins kinesins in eight subfamilies, and 3 unclassified kinesins were in prophase. localized to the mitotic apparatus at one or multiple phases during cell division. The most frequently observed location was Relationship Between Sequence Homology and Localization Similarity. the spindle/phragmoplast midzone, where some MTs are cou- The comprehensive survey of kinesins localization allowed us pled in an antiparallel manner (19, 22). In anaphase and telo- to correlate the localization similarity to protein homology. In phase, nearly 30 kinesins were localized to this region. Spindle/ a simple way, we created a phylogenetic tree of all P. patens phragmoplast localization in regions other than the midzone was kinesins and placed it side-by-side with the localization summary also frequently seen (e.g., kinesin 7-IV), and some kinesins oc- data (Fig. 3). We also analyzed every pair of kinesins within cupied the entire length of spindle/phragmoplast MTs (e.g., a subfamily; the full-length, motor domain, or nonmotor domain kinesin-13s). On the other hand, chromosome-localized kinesins was subjected to comparison (Fig. S2). Both analyses indicated

E1054 | www.pnas.org/cgi/doi/10.1073/pnas.1311243111 Miki et al. Downloaded by guest on September 28, 2021 PNAS PLUS

Fig. 2. Kinesin localization during cell division. The representative images are displayed for cells in prophase (A), metaphase (B), anaphase (C), and cytokinesis (D). Green, kinesin-Citrine and autofluorescence derived from the chloroplast; red, mCherry-tubulin. As a control, cells expressing only mCherry-tubulin are displayed (Left). The localization and name of the kinesin gene are described at the top and bottom of each image, respectively. Arrowhead indicates the phragmoplast edge location. (Scale bar, 10 μm.) See also Movies S1–S11.

that the localization and homology of the proteins were generally and is shown to play a critical role in spindle bipolarization in well correlated. For example, four kinesin-5s showed identical animals, fungi, and Arabidopsis (33–35). However, the localiza- localization throughout mitosis, and their mutual protein ho- tion of this protein is slightly different among cell types. Frog mologies were higher than most of the other pairs (Fig. 3 and Eg5 is enriched near the spindle pole by dynein-dependent Fig. S2A, red). Fifteen pairs of six kinesin-7s were divided into transport (36), whereas fly Klp61F is uniformly localized on the two subgroups, showing a similar or distinct localization pattern, spindle MTs in the S2 cell line or embryos (30, 37). AtKRP125c and this grouping was consistent with the amino acid sequence (At kinesin 5-c), one of four Arabidopsis kinesin-5s, and tobacco homology (Fig. S2A, green). An exception was found for kinesin- TKRP125 are abundantly localized to all MT arrays throughout 12, however, in which the correlation between sequence ho- the cell cycle (33, 38). Despite the diverse localization pattern, mology and localization similarity was poor (Fig. S2B). the dominant concept regarding the general function of kinesin-5 is that it slides apart antiparallel MTs at the spindle midzone, Kinesin-5 Is Depleted in the Midzone of the Spindle and Phrag- because this can best explain its depletion phenotype in many moplast. The localization survey has provided several interesting cell types: spindle monopolarization (39–41).

candidate proteins for further analysis. We analyzed the function P. patens has four kinesin-5s that are highly homologous to PLANT BIOLOGY of kinesin-5 in the present study. Kinesin-5 forms a homotetramer one another (kinesin 5-a to 5-d; Fig. 4 A and B). Immunoblotting

Miki et al. PNAS | Published online March 3, 2014 | E1055 Downloaded by guest on September 28, 2021 Fig. 3. Summary of the kinesin expression and localization during cell division. (Left) The metaphase spindle or phragmoplast was subdivided into three regions. 1′ and 4′, midzone; 2′ and 5′, intraspindle/phragmoplast; 3′ and 6′, polar region. (Right) A heat map showing a summary of the localization of 78 kinesins in each phase of mitosis. Squares indicate whether the Citrine signal was detected clearly (green), weakly (dark green), or undetected (black). No information was available for six genes (gray) to which Citrine could not be tagged. The rightmost column indicates whether the kinesin is categorizedas mitotic (M), nonmitotic (NM), or not expressed (—). The bottom row indicates total numbers of kinesins detected at each location. Subfamily assignments are designated on the right of the phylogenetic tree. (Center) Phylogenetic tree of P. patens kinesins. A bar at the bottom indicates amino acid substitutions per site. (Bottom Left) Numerical summary of P. patens kinesins in this study. See also Movies S1–S11 for the actual data and Table S3.

E1056 | www.pnas.org/cgi/doi/10.1073/pnas.1311243111 Miki et al. Downloaded by guest on September 28, 2021 PNAS PLUS

Fig. 4. Nonuniform distribution of kinesin-5 in the spindle and phragmoplast. (A) Phylogenetic analysis of the kinesin-5 family of proteins from Phys- comitrella patens (Pp), Arabidopsis thaliana (At), and Marchantia polymorpha (Mp). The numbers on the branches represent the local bootstrap probability. The local bootstrap values with 1,000 replicates are shown on branches. The horizontal branch length is proportional to the estimated evolutionary distance. A bar indicates the amino acid substitutions per site. (B) Amino acid sequence identity among P. patens or A. thaliana kinesin-5s (determined by ClustalW2). (C) Immunoblotting of Citrine-tagged kinesin 5-a to 5-d in protonemal cell extracts. Relative amount (%) of kinesin 5-a to 5-d was derived from band intensity. Coomassie-stained gel serves as a loading control. Note that anti-GFP can recognize Citrine protein. (D) Kinesin 5-c localization during mitosis. Time 0 corresponds to NEBD. (Scale bar, 10 μm.) See also Movie S3.(E) Signal intensity plot along metaphase spindle (Left) or phragmoplast (Right) MTs. Boxed regions (white) were analyzed. Means of adjacent three pixels were plotted. Kinesis 5-c signals (green) sharply decreased at the equatorial region at which MTs (red) are present. (Scale bar, 10 μm.) (F) Kinesin 5-c localization in the phragmoplast. Poleward movement of the kinesin 5-c-Citrine signal was detected in the kymograph; the molecular basis of this movement is unknown. (Scale bars: horizontal bars, 10 μm; vertical bar, 3 min.)

indicated that kinesin 5-d was most abundantly expressed in the exclusion of kinesin 5-c from the equator; instead, we de- protonemata, whereas kinesin 5-a expression was low (Fig. 4C). tected poleward flow of the Citrine signals (Fig. 4F). We con- Live cell imaging at 30-s intervals showed identical localization cluded that kinesin-5 is depleted in the spindle/phragmoplast patterns for kinesin 5-a to 5-d, although signal intensities varied midzone and rather enriched in the region where parallel MTs (Movie S3). The signals were seen in cytoplasmic MTs at pro- predominate. phase, including those around the nucleus. After nuclear enve- lope breakdown (NEBD), they were localized at the spindle and Kinesin-5 Is Required for Postanaphase Spindle/Phragmoplast Organi- phragmoplast MTs (Fig. 4D). Surprisingly, the signal intensity zation and Chromosome Segregation. To investigate the function of was weakest at the central region of the spindle/phragmoplast P. patens kinesin-5s in mitosis, we performed phenotypic analysis throughout mitosis, although this corresponds to the region using a conditional RNAi system. In this system, the addition of β enriched with antiparallel overlaps (Fig. 4E) (19, 22). Next, we -estradiol to the culture medium induces the production of PLANT BIOLOGY imaged telophase cells at shorter intervals (3 s) and confirmed dsRNAs that are homologous to (multiple) target genes (14). We

Miki et al. PNAS | Published online March 3, 2014 | E1057 Downloaded by guest on September 28, 2021 Fig. 5. Postanaphase roles of kinesin-5. (A) Six RNAi constructs were used against four kinesin-5s. Gray and red/green bars indicate cDNAs and dsRNAs, respectively. The quadruple construct (first construct) was designed such that 311- to 424-bp sequences of each kinesin-5 were fused tandem. The nucleotide sequence identities of the targeted region were high among four kinesin-5 genes (72–89%). (B) qRT-PCR demonstrates knockdown of kinesin 5-a to 5-d transcripts. The value relative to the control line (=1.0) is plotted with the SEM of two independent experiments. Note that kinesin 5-d protein predominates in the protonemata (Fig. 4C), such that a reduction of its mRNA would most critically affect the total amount of kinesin-5. (C) Summary of RNAi line selection. We selected 30 candidate RNAi lines by chromosomal integration of the RNAi vectors and observed multinucleated cells in 20 lines. This hit rate is similar to that in a previous study of seven constructs targeting three genes required for mitosis (14); importantly, we never observed such a phenotype in controls in 41 independent time-lapse image sets (note that our RNAi line selection method does not assess whether the RNAi vector is entirely integrated) (14). It is therefore possible that the lines that did not give rise to a phenotype had not properly integrated the RNAi unit. (D and E) Cells with multiple nuclei were observed for all six kinesin-5 RNAi constructs. Error bars indicate the SD of three to five independent experiments. (Scale bar, 50 μm.) (F) Time-lapse imaging of cells expressing GFP-tubulin (green) and HistoneH2B-mRFP (red) using wide-field microscopy. Caulonemal cells of a control and a transgenic RNAi line (first construct, clone #15) are displayed. Images were acquired every 3 min. Time 0 corresponds to NEBD. (Scale bar, 50 μm.) See also Movie S12.(G)Time-lapseimaging of a caulonemal cell expressing GFP-tubulin (green) and HistoneH2B-mRFP (red) with spinning-disk confocal microscopy. The representative control and kinesin-5 RNAi (first construct, clone #15) cells are displayed. Time 0 corresponds to the timing of sister chromatid separation in the top and middle panels, whereas ar- bitrary time is indicated for the cell in the bottom. Arrowheads indicate two split phragmoplast-like structures. (Scale bar, 10 μm.) See also Movie S13.

E1058 | www.pnas.org/cgi/doi/10.1073/pnas.1311243111 Miki et al. Downloaded by guest on September 28, 2021 designed a “quadruple” RNAi construct that contained four POK2, which control division plane orientation at the PPB and PNAS PLUS sequences targeting kinesins 5-a to 5-d fused in tandem and also cell cortex (52). five other constructs that target each of the kinesin 5-b to -d Other than the kinesin 12-I subfamily, the gene number is genes; three nonoverlapping regions of the kinesin-5 genes were generally comparable between P. patens and Arabidopsis (Table covered by these constructs (Fig. 5A). Because the four kinesin-5 S2); however, a notable difference is the presence of highly ho- genes are highly homologous to each other, the RNAi construct mologous genes within the subfamily in P. patens. For instance, designed to target one of them is expected to knock down multiple there are four kinesin-5 genes in Arabidopsis and P. patens. genes simultaneously (14, 22, 42). Quantitative real-time RT-PCR However, except for At kinesin 5-c and 5-d (69%), the amino showed that this is indeed the case; the transcripts of multiple acid sequences of each At kinesin-5 are more diverged (39–45%) kinesin-5 genes were reduced for each construct (Fig. 5B). than in P. patens (72–77% identity among Pp kinesin 5-a to 5-d), We selected multiple stable transgenic RNAi lines for each suggesting that the function of each kinesin-5 is more diverged in construct and performed long-term time-lapse imaging of β-estra- Arabidopsis (25). Indeed, in Arabidopsis, mutation in a single diol–treated cells using low-magnification lens (3-min intervals, kinesin-5 (AtKRP125c) causes a dramatic spindle defect (33). >5 h). We identified multinucleated cells in multiple transgenic An interesting finding in this comprehensive study is that ex- lines for all six RNAi constructs, but never in the control (Fig. 5 cept for just one kinesin (kinesin 4-Ic), none of the remaining 42 C–E;inFig.5E,47–140 tip cells were analyzed for RNAi lines in kinesins that were identified in this study showed the identical three to five experiments and >200 cells for the control). The ob- localization pattern to those of better-studied metazoan kinesins. servation of mitotic chromosomes suggested that multinucleation For example, the orthologs of the well-studied kinetochore kine- was caused by the generation of lagging chromosomes during sins (kinesin-8s and kinesin-13s) did not show such localization anaphase and/or by cytokinetic failure (Fig. 5F and Movie S12). and instead were more concentrated at the spindle midzone. The Next, we performed imaging of mitosis using spinning-disk con- accumulation of kinesin-5 was specifically diminished at the mid- focal microscopy. The observations confirmed defective spindle/ zone, which has been identified as the most important functional phragmoplast morphology and chromosome segregation during site for this motor by previous studies in animals and yeasts (note anaphase and cytokinesis (Fig. 5G and Movie S13). Furthermore, that according to the figure presented in ref. 33, Arabidopsis metaphase spindles appeared longer than normal and were oc- AtKRP125c may also be similarly localized to Pp kinesin-5, al- casionally associated with misaligned chromosomes. The most though costaining of spindle MTs and AtKRP125c have not been frequently observed phenotype was that of lagging chromosomes reported). The absence of kinesins at these expected sites would during anaphase, accompanied by slacked spindle MTs (Fig. 5G). not be solely attributed to the lack of these sites in the moss In some cases, phragmoplasts were later normally formed and spindle, because other kinesins were found at those sites (e.g., expanded, after which cytokinesis occurred (Fig. 5G, Middle), kinesin 7-III was localized at the kinetochore and several kinesin-13s whereas in other cells, phragmoplast MTs failed to align properly, and kinesin-14s were enriched at the spindle midzone). The dis- and cytokinesis was incomplete (Fig. 5G, Bottom). However, we crepancy would be therefore attributed, at least partly, to the difference in kinesin protein sequences. In other cases, we observed did not observe defects in spindle bipolarity before anaphase. that kinesin 12-Ia, -Ib, and kinesin 7-IV have pole-proximal accu- Discussion mulation during anaphase chromosome segregation and that nine kinesins accumulate at the metaphase spindle midzone. To the best Endogenous Localizome of Plant Kinesins in Mitosis. In this study, we of our knowledge, these localizations have not been reported for identified 43 kinesins that are localized to the mitotic spindle the animal homologs. Our study may thus urge us to reconsider and/or phragmoplast. Remarkably, this represents 72% of the each motor’s conserved function during mitosis in plants and in total 60 kinesins for which expression was detected in caulonemal animals. Plant kinesins might have evolved to acquire new cellular cells. Plants may have increased the number of kinesins that functions. Alternatively, the function might be conserved in ani- participate in mitotic cell division, which should be executed mals but has been overlooked for some reason such as presence of without the help of centrosomes or another MT-based motor redundant mechanisms in animal cells. Because most of the dynein. The data also suggest a redundancy of kinesins, because kinesins are widely conserved in plants, follow-up studies will also in many subfamilies, highly homologous members showed in- be conducted in other plant species, such as the popular model distinguishable localizations throughout cell division. However, organism A. thaliana. loss-of-function analysis using gene disruption of each kinesin is critical to address these hypotheses. Roles of Kinesin-5 in Chromosome Segregation. We carried out func- The kinesins previously studied in other plant species, such as tional studies on kinesin-5 by using an inducible RNAi system. Arabidopsis or tobacco, generally show similar localization [e.g., This system allows phenotypic characterization after simulta- – kinesin 7-II (43, 44); kinesin 12-II (45); kinesin 14-I (46 48)]. neous knockdown of multiple paralogues in a reasonable period This similarity may not be surprising, given that essential features (∼3 mo) and is applicable to any genes of interest, including of the mitotic apparatus, such as a lack of centrosomes, are those essential for moss viability (compared with gene disrup- conserved between mosses and flowering plants (14, 23, 49, 50). tion approach which cannot be used for an essential gene). However, the total number of kinesins is lower in Arabidopsis (61 After kinesin-5 RNAi induction, we predominantly observed genes) than in the simpler P. patens (78 genes). Phylogenetic defects in chromosome segregation, spindle/phragmoplast orga- comparison attributed this to the presence of extremely large nization, and cytokinesis in postanaphase. This result is un- numbers of kinesin 12-I subfamily genes in P. patens (18 genes) expected because kinesin-5 inhibition in many other cell types, compared with Arabidopsis (three genes) (Table S2) (28, 51). including Arabidopsis root cells (33), primarily causes spindle Our study suggests that most kinesin 12-I in P. patens functions at collapse during prometaphase/metaphase. A recent study using the spindle/phragmoplast equator. It is tempting to speculate that mammalian cells and a chemical inhibitor of kinesin-5 explicitly the lack of the PPB or prophase polar caps in moss caulonemata, showed that anaphase-specific inhibition of kinesin-5 results in which could help spindle bipolarization in flowering plants (25, an abnormal spindle elongation rate but does not alter spindle 26), necessitates motor proteins that cross-link antiparallel MTs morphology (53). P. patens kinesin-5 might have evolved a new at the equator to ensure bipolarity. Three P. patens kinesin 12-Is function. However, it is also possible that flowering plant kine- (12-Ia, 12-Id, and 12-Ie) showed punctate localization at the sin-5 has similar postanaphase function; the preanaphase defect

cell cortex, where the phragmoplast then extends. This locali- might have prevented specific assessment of the postanaphase PLANT BIOLOGY zation is reminiscent of two Arabidopsis kinesin 12-Is, POK1 and phenotype in the Arabidopsis kinesin-5 mutant (33). Inefficient

Miki et al. PNAS | Published online March 3, 2014 | E1059 Downloaded by guest on September 28, 2021 RNAi inhibition may account for the absence of monopolariza- confocal unit CSU-X1 (Yokogawa) and an electron-multiplying charge-coupled tion in our system; Pp kinesin-5 may retain spindle bipolarization device camera (ImagEM; Hamamatsu) was used. Images were acquired every function but our RNAi-based study might have overlooked it. An 30 or 60 s. For long-term imaging of the kinesin-5 RNAi lines, fluorescence ’ × × alternative possibility is that kinesin-5 is dispensable for spindle imaging was performed with Nikon sTimicroscope(10 0.45-NA or 10 0.50- bipolarization in this particular cell type. We think that the latter NA lens) equipped with an electron-multiplying charge-coupled device camera (Evolve; Roper or iXon3; Andor). Images were acquired at multiple sites every idea is viable, given that 16 other kinesins, including the kinesin-12 3 min. The microscopes were controlled by the Micro-Manager software (59), family that has been shown to function redundantly with kinesin-5 and image data were analyzed with ImageJ. in bipolarizing the spindle in mammals (54, 55), are concentrated at the spindle midzone during metaphase, whereas Genotyping by PCR. Genomic DNA for PCR was prepared as follows. Pro- kinesin-5s are not. Gene disruption of kinesin-5s will clarify their tonemata cultured on BCDAT plates for 1–2 wk were flash frozen by liquid

role in preanaphase in the future. nitrogen. A 10× PCR buffer [160 mM (NH4)2SO4, 670 mM Tris·HCl, 0.1% In either case, our results indicate that the integrity of the Tween-20, pH 8.8) was added, and the cells were squashed with a pipette postanaphase spindles and phragmoplasts depends on kinesin-5. tip. Following incubation at 65 °C for 10 min and centrifugation, the su- The appearance of misaligned chromosomes in preanaphase and pernatant was collected. The PCR amplification was performed with KOD FX lagging chromosomes in anaphase suggests that kinetochores and Neo (TOYOBO), following the manufacturer’s instructions. PCR was per- MTs are improperly attached because of the reduced kinesin-5 formed against 15 lines for each kinesin gene. level. Furthermore, the appearance of split miniphragmoplast structures (Fig. 5G, Bottom) suggests that kinesin-5 might organize Quantitative Real-Time RT-PCR. Total RNA was prepared from protonemal cells using the RNeasy Plant Mini Kit (Qiagen), followed by DNase treatment. the phragmoplast through parallel bundling of the adjacent First-strand cDNA was synthesized with the PrimeScript II first-strand cDNA MTs, which was recently proposed to be a critical mechanism for Synthesis Kit (TAKARA). Quantitative RT-PCR (qRT-PCR) was performed in phragmoplast expansion (50). triplicate using Power SYBR Green PCR Master Mix and the 7500 Real-Time PCR system (model 7500; Applied Biosystems). The results were normalized to Conclusion. In summary, to our knowledge, this study represents those for TUA1 (α-tubulin) (14). Primers for qRT-PCR are listed in Dataset S1. the most comprehensive analysis of kinesin localization during mitosis in a plant cell. The gene expression under the native Immunoblotting. Protonemata cultured on BCDAT plates for 2–4 wk were promoter at the native chromosomal locus of a WT multicellular flash frozen by liquid nitrogen, followed by preparation of the cell extracts organism (i.e., not a cultured cell line) makes our dataset rare with an SDS-containing buffer. Immunoblotting was performed with anti- and valuable. Furthermore, several genes are interesting targets GFP (JL8, which also recognized Citrine; Clontech). for further analysis; as with kinesin-5, such analysis could provide insights into motor function during mitosis. Finally, the trans- Sequence Analysis. The deduced amino acid sequences were aligned using genic lines produced in this study would serve as valuable tools to MAFFT ver. 7.043 (60, 61) and then revised manually with MacClade ver. 4.08 dissect the function of heavily amplified motor proteins in the OSX. In Fig. 3, 189 amino acid residues were used to calculate evolution distances for 66 genes (989 amino acid residues for 9 genes in Fig. 4A) using land plant lineage. the Jones-Taylor-Thornton (JTT) model (62) to construct a neighbor-joining Materials and Methods tree by the MEGA5 software (63). Statistical support for internal branches by bootstrap analyses was calculated using 1,000 replications. The Arabidopsis Plasmids. The PCR primers used in this study are listed in Dataset S1.The Information Resource (www.arabidopsis.org) and Phytozome (www.phyto- α mCherry and -tubulin genes were amplified by PCR and subcloned into zome.net) (for Selaginella moellendorffii) were used as the database. The thevectorthatcontainsthericeactin promoter (56), the rbcS terminator, sequence of Marchantia polymorpha kinesin-5 was provided by Ryuichi the modified aph4 cassette, and the genomic fragment of the PIG1 locus. Nishihama and Takayuki Kohchi (Kyoto University, Kyoto, Japan). For the C-terminal fusion of Citrine to kinesins, integration plasmids were constructed, in which the ∼1-kb C terminus and ∼1-kb 3′-UTR region sequences Gene ID. A. thaliana genes used in this study are as follows (51): At kinesin of the kinesin genes were flanked by the citrine gene, the nopaline synthase 5-a (AT2G37420, AtKRP125a), At kinesin 5-b (AT2G36200, AtKRP125b), At polyadenylation signal (nos-ter), and the nptII cassette that confers G418 kinesin 5-c (AT2G28620, AtKRP125c), and At kinesin 5-d (AT3G45850). Gene resistance (57). The start codon (ATG) of the citrine gene was changed to IDs of P. patens are described in Table S1 and Dataset S1. TTG. Six kinesin-5 RNAi constructs were made by using the Gateway system (Invitrogen) with pGG626 (14) as the destination vector. ACKNOWLEDGMENTS. We thank Ken Kosetsu for sharing several plasmids; Yuji Hiwatashi for help with the phylogenetic analysis and plasmids; Moss Culture, Transformation, and Microscopy. We essentially followed the Takashi Murata and Mitsuyasu Hasebe for providing plasmids; Yuki Tsukada methods described in ref. 14. In brief, BCDAT medium was used for regular for the image analysis; Yuki Nakaoka for valuable advice; Ryuichi Nishihama culturing of protonemata at 25 °C under continuous white light. Protonemal and Takayuki Kohchi for M. polymorpha sequence information; Yasunori cells, which had been cultured on glass-bottom (MatTek) or six-well glass- Machida, Akiko Tomioka, Ayako Watanabe, Kimiko Kameda, and Rie Inaba bottom (IWAKI) plates with BCD agar medium at 24–25 °C, were imaged. To for technical assistance; Ryota Uehara and Tomoko Nishiyama for valuable observe RNAi phenotypes, we cultured protonemata for 6–8 d in the presence discussion and advice; Fumiko Esashi and Tomoko Nishiyama for critical read- of 1 μM β-estradiol that turns on the promoter (58) and thereby expression of ing of the manuscript; Futamura Chemical Industries for the cellophane; Kyowa Hakko Kogyo for the Driselase; and Leica Inc. for the use of their micro- dsRNAs. Transformation was performed by the standard polyethylene glycol– scope at the Marine Biological Laboratory. This work was supported by the Next mediated method. mCherry-tubulin/kinesin-Citrine lines were generated Generation grant (Japan Society for Promotion of Science), Human Frontier – from the same parental mCherry-tubulin expressing line. For the high- Science Program, and James A. and Faith Miller Memorial Fund (to G.G.). magnification time-lapse microscopy of the kinesin-Citrine or RNAi lines, the T.M. is a recipient of a predoctoral fellowship of the Japan Society for Nikon TE2000 microscope (100× 1.40-NA lens) equipped with the spinning-disk Promotion of Science.

1. Lawrence CJ, et al. (2004) A standardized kinesin nomenclature. JCellBiol167(1):19–22. 7. Scholey JM, Brust-Mascher I, Mogilner A (2003) Cell division. Nature 422(6933): 2. Miki H, Okada Y, Hirokawa N (2005) Analysis of the kinesin superfamily: Insights into 746–752. structure and function. Trends Cell Biol 15(9):467–476. 8. Walczak CE, Heald R (2008) Mechanisms of mitotic spindle assembly and function. Int 3. Tytell JD, Sorger PK (2006) Analysis of kinesin motor function at budding yeast Rev Cytol 265:111–158. kinetochores. JCellBiol172(6):861–874. 9. Goshima G, Scholey JM (2010) Control of mitotic spindle length. Annu Rev Cell Dev 4. Goshima G, Vale RD (2003) The roles of microtubule-based motor proteins in mitosis: Biol 26:21–57. Comprehensive RNAi analysis in the Drosophila S2 cell line. JCellBiol162(6):1003–1016. 10. Vale RD (2003) The molecular motor toolbox for intracellular transport. Cell 112(4): 5. Zhu C, et al. (2005) Functional analysis of human microtubule-based motor proteins, 467–480. the kinesins and dyneins, in mitosis/cytokinesis using RNA interference. Mol Biol Cell 11. Reddy AS, Day IS (2001) Kinesins in the Arabidopsis genome: A comparative analysis 16(7):3187–3199. among eukaryotes. BMC Genomics 2:2. 6. Maliga Z, et al. (2013) A genomic toolkit to investigate kinesin and myosin motor 12. Rensing SA, et al. (2008) The Physcomitrella genome reveals evolutionary insights into function in cells. Nat Cell Biol 15(3):325–334. the conquest of land by plants. Science 319(5859):64–69.

E1060 | www.pnas.org/cgi/doi/10.1073/pnas.1311243111 Miki et al. Downloaded by guest on September 28, 2021 13. Prigge MJ, Bezanilla M (2010) Evolutionary crossroads in developmental biology: 38. Asada T, Kuriyama R, Shibaoka H (1997) TKRP125, a kinesin-related protein involved PNAS PLUS Physcomitrella patens. Development 137(21):3535–3543. in the centrosome-independent organization of the cytokinetic apparatus in tobacco 14. Nakaoka Y, et al. (2012) An inducible RNA interference system in Physcomitrella BY-2 cells. J Cell Sci 110(Pt 2):179–189. patens reveals a dominant role of augmin in phragmoplast microtubule generation. 39. Sharp DJ, et al. (1999) The bipolar kinesin, KLP61F, cross-links microtubules within Plant Cell 24(4):1478–1493. interpolar microtubule bundles of Drosophila embryonic mitotic spindles. J Cell Biol 15. Cove D (2005) The moss Physcomitrella patens. Annu Rev Genet 39:339–358. 144(1):125–138. 16. Senger T, et al. (2005) A multifunctional lipoxygenase with fatty acid hydroperoxide 40. Kapitein LC, et al. (2005) The bipolar mitotic kinesin Eg5 moves on both microtubules cleaving activity from the moss Physcomitrella patens. JBiolChem280(9):7588–7596. that it crosslinks. Nature 435(7038):114–118. 17. Bezanilla M, Pan A, Quatrano RS (2003) RNA interference in the moss Physcomitrella 41. Kashina AS, et al. (1996) A bipolar kinesin. Nature 379(6562):270–272. patens. Plant Physiol 133(2):470–474. 42. Vidali L, Augustine RC, Kleinman KP, Bezanilla M (2007) Profilin is essential for tip 18. Bezanilla M, Perroud PF, Pan A, Klueh P, Quatrano RS (2005) An RNAi system in growth in the moss Physcomitrella patens. Plant Cell 19(11):3705–3722. Physcomitrella patens with an internal marker for silencing allows for rapid identi- 43. Nishihama R, et al. (2002) Expansion of the cell plate in plant cytokinesis requires – fication of loss of function phenotypes. Plant Biol (Stuttg) 7(3):251 257. a kinesin-like protein/MAPKKK complex. Cell 109(1):87–99. 19. Hiwatashi Y, et al. (2008) Kinesins are indispensable for interdigitation of phragmo- 44. Strompen G, et al. (2002) The Arabidopsis HINKEL gene encodes a kinesin-related – plast microtubules in the moss Physcomitrella patens. Plant Cell 20(11):3094 3106. protein involved in cytokinesis and is expressed in a cell cycle-dependent manner. 20. Vidali L, Rounds CM, Hepler PK, Bezanilla M (2009) Lifeact-mEGFP reveals a dynamic Curr Biol 12(2):153–158. apical F-actin network in tip growing plant cells. PLoS ONE 4(5):e5744. 45. Lee YR, Giang HM, Liu B (2001) A novel plant kinesin-related protein specifically as- 21. Vidali L, et al. (2009) Rapid formin-mediated actin-filament elongation is essential for sociates with the phragmoplast organelles. Plant Cell 13(11):2427–2439. – polarized plant cell growth. Proc Natl Acad Sci USA 106(32):13341 13346. 46. Ambrose JC, Cyr R (2007) The kinesin ATK5 functions in early spindle assembly in 22. Kosetsu K, de Keijzer J, Janson ME, Goshima G (2013) MICROTUBULE-ASSOCIATED Arabidopsis. Plant Cell 19(1):226–236. PROTEIN65 is essential for maintenance of phragmoplast bipolarity and formation of 47. Ambrose JC, Li W, Marcus A, Ma H, Cyr R (2005) A minus-end-directed kinesin with the cell plate in Physcomitrella patens. Plant Cell 25(11):4479–4492. plus-end tracking protein activity is involved in spindle morphogenesis. Mol Biol Cell 23. Doonan JH, Cove DJ, Lloyd CW (1985) Immunofluorescence microscopy of micro- 16(4):1584–1592. tubules in intact cell lineages of the moss, Physcomitrella patens. I. Normal and CIPC- 48. Chen C, et al. (2002) The Arabidopsis ATK1 gene is required for spindle morpho- treated tip cells. J Cell Sci 75:131–147. genesis in male meiosis. Development 129(10):2401–2409. 24. Müller S, Wright AJ, Smith LG (2009) Division plane control in plants: New players in 49. Hotta T, et al. (2012) Characterization of the Arabidopsis augmin complex uncovers its the band. Trends Cell Biol 19(4):180–188. critical function in the assembly of the acentrosomal spindle and phragmoplast mi- 25. Bannigan A, Lizotte-Waniewski M, Riley M, Baskin TI (2008) Emerging molecular crotubule arrays. Plant Cell 24(4):1494–1509. mechanisms that power and regulate the anastral mitotic spindle of flowering plants. 50. Murata T, et al. (2013) Mechanism of microtubule array expansion in the cytokinetic Cell Motil Cytoskeleton 65(1):1–11. phragmoplast. Nat Commun 4:1967. 26. Lloyd C, Chan J (2006) Not so divided: The common basis of plant and animal cell 51. Zhu C, Dixit R (2012) Functions of the Arabidopsis kinesin superfamily of microtubule- division. Nat Rev Mol Cell Biol 7(2):147–152. based motor proteins. Protoplasma 249(4):887–899. 27. Goshima G, Kimura A (2010) New look inside the spindle: Microtubule-dependent 52. Müller S, Han S, Smith LG (2006) Two kinesins are involved in the spatial control of microtubule generation within the spindle. Curr Opin Cell Biol 22(1):44–49. cytokinesis in Arabidopsis thaliana. Curr Biol 16(9):888–894. 28. Shen Z, Collatos AR, Bibeau JP, Furt F, Vidali L (2012) Phylogenetic analysis of the 53. Collins E, Mann BJ, Wadsworth P (2013) Eg5 Restricts Anaphase B Spindle Elongation Kinesin superfamily from physcomitrella. Front Plant Sci 3:230. in Mammalian Cells (Cytoskeleton, Hoboken, NJ). 29. Griesbeck O, Baird GS, Campbell RE, Zacharias DA, Tsien RY (2001) Reducing the 54. Tanenbaum ME, et al. (2009) Kif15 cooperates with eg5 to promote bipolar spindle environmental sensitivity of yellow fluorescent protein. Mechanism and applications. – J Biol Chem 276(31):29188–29194. assembly. Curr Biol 19(20):1703 1711. 30. Goshima G, Vale RD (2005) Cell cycle-dependent dynamics and regulation of mitotic 55. Vanneste D, Takagi M, Imamoto N, Vernos I (2009) The role of Hklp2 in the stabili- – kinesins in Drosophila S2 cells. Mol Biol Cell 16(8):3896–3907. zation and maintenance of spindle bipolarity. Curr Biol 19(20):1712 1717. 31. West RR, Malmstrom T, Troxell CL, McIntosh JR (2001) Two related kinesins, klp5+ and 56. McElroy D, Zhang W, Cao J, Wu R (1990) Isolation of an efficient actin promoter for – klp6+, foster microtubule disassembly and are required for meiosis in fission yeast. use in rice transformation. Plant Cell 2(2):163 171. Mol Biol Cell 12(12):3919–3932. 57. Nishiyama T, Hiwatashi Y, Sakakibara I, Kato M, Hasebe M (2000) Tagged mutagenesis and – 32. Garcia MA, Koonrugsa N, Toda T (2002) Two kinesin-like Kin I family proteins in fis- gene-trap in the moss, Physcomitrella patens by shuttle mutagenesis. DNA Res 7(1):9 17. β sion yeast regulate the establishment of metaphase and the onset of anaphase A. 58. Kubo M, et al. (2013) System for stable -estradiol-inducible gene expression in the Curr Biol 12(8):610–621. moss Physcomitrella patens. PLoS ONE 8(9):e77356. 33. Bannigan A, et al. (2007) A conserved role for kinesin-5 in plant mitosis. JCellSci 59. Edelstein A, Amodaj N, Hoover K, Vale R, Stuurman N (2010) Computer control of 120(Pt 16):2819–2827. microscopes using μManager. Curr Protoc Mol Biol 14:Unit 14.20. 34. Hagan I, Yanagida M (1990) Novel potential mitotic motor protein encoded by the 60. Katoh K, Misawa K, Kuma K, Miyata T (2002) MAFFT: A novel method for rapid multiple fission yeast cut7+ gene. Nature 347(6293):563–566. sequence alignment based on fast Fourier transform. Nucleic Acids Res 30(14):3059–3066. 35. Sawin KE, LeGuellec K, Philippe M, Mitchison TJ (1992) Mitotic spindle organization 61. Katoh K, Standley DM (2013) MAFFT multiple sequence alignment software by a plus-end-directed microtubule motor. Nature 359(6395):540–543. version 7: Improvements in performance and usability. MolBiolEvol30(4): 36. Uteng M, Hentrich C, Miura K, Bieling P, Surrey T (2008) Poleward transport of Eg5 772–780. by dynein-dynactin in Xenopus laevis egg extract spindles. JCellBiol182(4): 62. Jones DT, Taylor WR, Thornton JM (1992) The rapid generation of mutation data 715–726. matrices from protein sequences. Comput Appl Biosci 8(3):275–282. 37. Cheerambathur DK, Brust-Mascher I, Civelekoglu-Scholey G, Scholey JM (2008) Dy- 63. Tamura K, et al. (2011) MEGA5: Molecular evolutionary genetics analysis using namic partitioning of mitotic kinesin-5 cross-linkers between microtubule-bound and maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol freely diffusing states. J Cell Biol 182(3):429–436. Biol Evol 28(10):2731–2739. PLANT BIOLOGY

Miki et al. PNAS | Published online March 3, 2014 | E1061 Downloaded by guest on September 28, 2021