|Home | About | Journals | Submit | Contact Us | Français|
Directional cell expansion in interphase and nuclear and cell division in M-phase are mediated by four microtubule arrays, three of which are unique to plants: the interphase array, the preprophase band, and the phragmoplast. The plant microtubule-associated protein MAP65 has been identified as a key structural component in these arrays . The Arabidopsis genome has nine MAP65 genes, and here we show that one, AtMAP65-3/PLE, locates only to the mitotic arrays and is essential for cytokinesis. The Arabidopsis pleiade (ple) alleles are single recessive mutations, and we show that these mutations are in the AtMAP65-3 gene. Moreover, these mutations cause C-terminal truncations that abolish microtubule binding. In the ple mutants the anaphase spindle is normal, and the cytokinetic phragmoplast can form but is distorted; not only is it wider, but the midline, the region where oppositely oriented microtubules overlap, is unusually expanded. Here we present data that demonstrate an essential role for AtMAP65-3/PLE in cytokinesis in plant cells.
There are nine Arabidopsis MAP65 genes, and a divergent member of this family is AtMAP65-3 . We raised an antiserum specific for AtMAP65-3 and used it together with anti-tubulin to double stain Arabidopsis cells throughout the cell cycle (Figure 1). The results show that AtMAP65-3 does not locate to the interphase cortical array. During cell division anti-AtMAP65-3 stains the preprophase band. No staining of microtubules in the prophase, metaphase, or early anaphase spindles was observed, although cytoplasmic staining was evident and suggested that the AtMAP65-3 protein is present at these stages. Binding of AtMAP65-3 to the microtubules during cell division reappears in the region of overlap between the half spindles in late anaphase after chromosome separation and persists in the cytokinetic phragmoplast at the midline, where the plus ends of microtubules overlap and where the cell plate forms (Figure 1).
Here we show that AtMAP65-3 is synonymous with PLEIADE. The Arabidopsis pleiade mutant alleles were isolated in genetic screens for defects in root morphogenesis . The ple alleles are recessive, and each has a short, irregular expanded root phenotype. These phenotypes are orchestrated by enlarged multinucleated cells with incomplete crosswalls, indicating that the defect is in cytokinesis (shown for ple-1 in Figures 2A and 2B). Cells in the ple mutants may accumulate up to 32 nuclei, indicating that five rounds of division have taken place and that karyokinesis is complete but cytokinesis is incomplete. All root tissues are affected, but some multinucleated cells are present in the embryonic roots and hypocotyls of mature embryos . A consequence of the enlarged multinucleated cells is irregular cell expansion, but this must be indirect because not all the multinucleated cells have lost their ability to expand anisotropically; for example, epidermal cells are still able to differentiate into root hair cells.
We have positionally cloned PLE by mapping the gene to an interval of 46 kb on the bacterial artificial chromosome K17N15 at the bottom of chromosome 5 (Figures 3A and 3B). We used the closest molecular markers to identify seven overlapping binary cosmid clones for complementation analyses of ple-1 and ple-2. Only the 15 kb genomic fragment of one clone, N16 (Figure 3B), reverted the ple phenotypes in 10 of 12 independent transformants. If one takes into consideration the overlaps from the noncomplementing cosmids, the PLE locus could be further narrowed to 10 kb, and this included three genes, At5g51600, At5g51610, and At5g51620 (Figure 3B). By using a combination of heteroduplex analysis and sequencing, we confirmed the At5g51600 open reading frame. PLE has 11 introns (Figure 3C); the transcriptional start site was determined by primer extension PCR, and this identified the 5′ UTR intron, which was also absent from the full-length cDNA clone. The predicted protein has a size of 707 amino acids and a molecular weight of 80.3 kDa and is synonymous with AtMAP65-3. A base substitution (G to A) at the 3′ splice site of the second intron of ple-1 leads to inefficient mRNA processing. RT-PCR analyses and sequencing revealed that the majority of transcript is unspliced and encodes a protein of 61 amino acids (Figures 3D and 3E). The mutations of ple-5 (CGA to TGA) and ple-6 (CAA to TAA) introduce stop codons and infer truncated proteins of 377 and 57 amino acids, respectively (Figure 3C). These data demonstrate the isolation of the first MAP65 mutants with a distinct cytokinesis defect. Interestingly, AtMAP65-3/PLE is expressed in all organs analyzed (see the Supplemental Data available with this article online), but the cytokinesis defect is only observed in root tissues , indicating that there is likely to be redundancy in function between members of the AtMAP65 gene family in different plant organs.
The facts that PLE is the AtMAP65-3 gene and mutations in this gene affect cytokinesis prompted us to analyze the phragmoplast microtubule arrays in the ple mutants. By using anti-tubulin immunofluorescence and crossing ple-1 and ple-6 with an in vivo microtubule marker line harboring MAP4:GFP , we observed all four microtubule arrays in ple cells. Closer examination of the phragmoplasts in these mutants revealed significant size differences shown by the histograms in Figure 4B. Whereas in the wild-type the midline of the phragmoplast is very narrow (0.3 μm ± 0.1), in the ple mutants it is about three times broader: 0.93 ± 0.23 μm in ple-1 and 0.8 ± 0.27 μm in ple-6 (Figures 4A and 4B) when visualized with the MAP4:GFP marker. Similar results were obtained with anti-tubulin immunofluorescence, i.e., from 0.25 ± 0.05 μm in the wild-type to 0.48 ± 0.09 μm and 0.67 ± 0.16 μm in ple-1 and ple-5, respectively. Moreover, the total width of the phragmoplast was also increased in the ple mutants, i.e., from 2.75 ± 0.62 μm in the wild-type to 5.64 ± 1.64 μm and 5.3 ± 0.58 μm in ple-1 and ple-6, respectively. From these data, together with the complete karyokinesis and the incomplete cytokinesis defects in the ple mutants, we conclude that AtMAP65-3/PLE is essential for phragmoplast function.
The three ple mutations truncate the AtMAP65-3/PLE protein at the C terminus. The shortest truncation predicts a protein of 377 aa. We generated the equivalent recombinant protein fragment (residues 1–377) and tested its ability to bind microtubules compared to the full-length recombinant protein. The data show (Figure 4C) that the truncated protein does not cosediment with polymerized microtubules upon centrifugation but that the full-length protein does. These data suggest that the loss of function in the ple mutants is due to a lack of microtubule binding.
In this paper we have shown that AtMAP65-3/PLE only locates to M-phase microtubule arrays and, in particular, to the midzones of the anaphase spindle and the cytokinetic phragmoplast. The ple mutants are defective in cytokinesis because the phragmoplast is distorted, but karyokinesis is unaffected. The PLE gene was cloned and found to encode AtMAP65-3. Moreover, recombinant protein that mimics the predicted ple-5 mutation, i.e., a truncated protein of 377 residues, cannot bind microtubules, indicating that the defect in cytokinesis is caused by the loss of microtubule binding in the ple mutants.
This localization of AtMAP65-3/PLE is similar to the previously characterized tobacco MAP65, NtMAP65-1 , with the exception that AtMAP65-3/PLE does not localize to the interphase cortical array. Both tobacco and carrot MAP65 proteins have been shown to bind and bundle microtubules in vitro [5-8], and their localization to the interphase array has suggested an involvement in directional cell expansion as well as in the organization of antiparallel microtubules in the midzones of the mitotic spindle and cytokinetic phragmoplast [5, 9]. The localization of AtMAP65-3/PLE suggests a role in cell division but not in cell expansion. It is interesting to note that anti-AtMAP65-3/PLE stains the metaphase spindle faintly but clear microtubule staining is only apparent in late anaphase. This may explain why karyokinesis is not affected in the ple mutants. AtMAP65-3/PLE, although present, may not be active in the metaphase spindle and is only activated, by an as-yet-unknown process, after the chromosomes are separated in late anaphase when the phragmoplast begins to form. These data would place MAP65 as a functional equivalent of the divergent family  of ‘midzone MAPs’ that includes mammalian PRC1  and yeast Ase1p .
The phragmoplast is composed of antiparallel microtubules that meet and overlap at the midline. Vesicles transported along these microtubules to the midline carry wall materials for the cell plate. The cell plate grows centrifugally outwards until it meets the parental cell walls, where it fuses. Microtubules remain at the boundary of the cell plate as it grows and are continually turning over. At the midline, the new tubulin subunits are incorporated at the plus ends of overlapping microtubules . Anti-tubulin and the MAP4-GFP do not usually stain this midline, giving the appearance of a clear zone. AtMAP65-3 locates to the midline or clear zone. Curiously, in the phragmoplasts of the ple mutants the clear zones are wider, as are the phragmoplasts themselves. This emphasizes the importance of AtMAP65-3/PLE for the integrity of the phragmoplast in cytokinesis in plants. In cells underexpressing PRC1, the human member of the midzone MAP family, the microtubules at the midzone of the anaphase spindle were not inter-digitated, the two half spindles separated, and the mid-body was not formed . Recent data from live-cell microscopy of ase1-deleted cells have demonstrated that Ase1p is essential for the slow phase of spindle elongation in anaphase B . By comparison, the anaphase spindle appears normal, and the cytokinetic apparatus in plants is not abolished in the ple mutants but is distorted, with a wider clear zone and a wider phragmoplast.
This distorted phragmoplast in the ple mutants could arise in either of two ways. First, the absence of AtMAP65-3/PLE could have destabilized the region of overlap at the midline by abolishing crosslinked, antiparallel microtubules and, as a consequence, the two halves of the phragmoplast may have drifted apart. However, it is clear that the drift is even and does not disrupt the shape of the phragmoplast, so the two halves must be held in place by another force. Second, AtMAP65-3/PLE may work in cooperation with other microtubule-interacting proteins that promote the flux of tubulin through the plus ends. In the absence of AtMAP65-3, the flux of tubulin through the plus ends (possibly through the actions of EB1 [13, 14] and/or MOR1/GEM1 ) is increased, allowing further polymerization and expansion of the clear zone. The increased overall width of the phragmoplast may result from increased microtubule sliding by plus end-directed kinesins [1, 16, 17]. In conclusion, these data demonstrate that AtMAP65-3/PLE is essential for maintaining the integrity of the overlapped microtubules in the phragmoplast and is essential for completing cell division in roots.
For raising the AtMAP65-3/PLE antiserum in mice, a 73 residue recombinant polypepetide corresponding to amino acids 554–627 of the AtMAP65-3/PLE open reading frame was used as an immunogen. Three mice were immunized, and the serum from all three mice identified a single band of relative molecular weight 80 kDa on immunoblots of Arabidopsis cell suspension culture total protein extracts. All three antisera gave identical staining patterns throughout the cell cycle. Arabidopsis suspension culture cells were collected at the exponential growth phase (3- to 4-day-old culture) and fixed with either paraformaldehyde as described  or with 100% methanol for 15 min at −20°C followed by 100% acetone for 10 min at −20°C. They were then treated with the mixture of cell wall digestion enzymes supplemented with 1 mM PMSF, 10 μg/ml leupeptin, and 10 μg/ml Pepstatin A for 5 min. In the case of methanol/acetone fixation, cells were rehydrated in PBS for 30 min before treatment with cell wall digestion enzymes. The cells were settled onto poly-L-lysine-coated coverslips and stained with rat monoclonal anti-tubulin clone YL1/2 (Serotec, Oxford, UK) diluted 1:100 and with anti-AtMAP65-3 diluted 1:500. Anti-mouse TRITC conjugates preadsorbed with rat immunoglobulins and anti-rat FITC conjugates preadsorbed with mouse immunoglobulins (Jackson Immunoresearch, West Grove, PA) were used as secondary antibodies. The staining was repeated four times for paraformaldehyde fixation and two times for methanol/acetone fixation. An average of 74% of the cell population gave staining with anti-AtMAP65-3. In the remaining cells no signal was observed with either anti-tubulin or anti-AtMAP65-3, suggesting that these cells were not processed successfully for immunostaining. Of the stained cells, 107 interphase cells and 154 dividing cells were scored in the six independent experiments. The staining with anti-AtMAP65-3 was identical for all stained cells: 24 preprophase bands, 41 prophase and metaphase spindles, ten early anaphase spindles, 12 late anaphase spindles, and 67 phragmoplasts. The specificity of the secondary antibodies was checked by application of anti-mouse antibody to cells stained only with YL1/2 antibody and vice versa.
The whole root of the ple mutants was scored compared to controls on vertical nutrient agar plates containing 1× Murashige and Skoog salt mixture and 4.5% sucrose (pH 5.7) in 1% agar. The nuclei were stained with YO-PRO (Molecular Probes, Leiden, the Netherlands) on fixed roots and analyzed with a confocal scanning laser microscope (Leica TCS SP2) as described . Microtubules were visualized in vivo with the microtubule marker MAP4:GFP. The phragmoplast measurements were carried out on 4- to 14-day-old seedlings (when the root meristem is fully active) via the microtubule MAP4:GFP marker line (a kind gift from H. Höfte) and confirmed by immunolocalization with YOL1/34 tubulin antibodies . In total, 37 wild-type roots were analyzed (27 preprophase bands, 14 spindles and 38 phragmoplasts were scored, and all phragmoplasts were measured), and 49 ple mutant root meristems were analyzed (33 preprophase bands, 22 spindles, and 25 phragmoplasts were scored, and all phragmoplasts were measured).
Genomic DNA was isolated from seedlings by a modified CTAB method . Fine mapping was done with duplex analysis markers  (CIC4G5L, mfg13not, CIC6A5L, KG8sp, mioph2c, F12A17sp; http://www.boku.ac.at/zag/Arabidopsis/at_index.htm), a RFLP marker (mio24), and CAPS markers (CIC7B9r and K17N15/49) on approximately 1000 F2 plants of crosses between the Columbia alleles ple-1 and ple-2 and the accession Wassilewskija. The cosmid contig of the PLE locus was established by screening the genomic abi cosmid library (kind gift of E. Grill) with labeled PCR probes (K17N15/49, K17N15/60, F12A17sp, and mio24). Transformation of wild-type, ple-1, and ple-2 was done via the floral dip method . Complementing transformants were confirmed in the T2 generation by cosegregation analysis. Duplex analysis  was used to screen for allele-specific polymorphism at the PLE locus. Sequencing was done on an ABI Prism system with BigDye terminator sequencing chemistry (Applied Biosystems, Warrington, UK). A full-length cDNA clone was isolated from the CD4-15 lambda library . The transcriptional start site was confirmed by a primer extension method. Exon/intron borders and splicing defects were analyzed by RT-PCR and sequencing. Total RNA was prepared with the RNeasy Plant Mini kit (Qiagen, Crawley, UK) or Trizol (Life Technologies, Paisley, UK), and cDNA was synthesized with moloney murine leukemia virus (MMLV) SUPERSCRIPT I (Invitrogen, Paisley, UK) and oligo d(T)18 primers. Analysis of the splicing defect was done with primers K17N15_63R and K15_624F. For the 5′ primer extension experiments, primer K15N624F was used for cDNA synthesis. The sequences of all primers are listed in Table S1 (see the Supplemental Data available with this article online).
Full-length AtMAP65-3/PLE and the truncated fragment (residues 1–377) corresponding to the PLE-5 mutant protein were subcloned into NdeI/XhoI sites of vector pET28a (Novagen). The recombinant proteins containing 6xHis tag on the N termini were expressed in E. coli strain BL21 (DE3) Rosetta and purified under denaturing conditions. The proteins were refolded by dialysis against MTSB buffer (0.1 M PIPES [pH 6.8], 2 mM EGTA, 2 mM MgSO4, 2 mM DTT, 50 mM NaCl, and 10% glycerol).
Tubulin was isolated from porcine brain as described . Tubulin, and centrifuging full-length AtMAP65-3/PLE, and truncated (1–377 residues) PLE-5 mutant protein at 150 000 × g for 15 min at 2°C removed protein aggregates. Microtubules were polymerized with 10 μM taxol, mixed in the appropriate samples with recombinant proteins, incubated for 10 min at 37°C, and centrifuged at 100 000 × g for 15 min. When the AtMAP65-3/PLE and the PLE-5 mutant protein were not mixed with microtubules, the mixture was supplemented with MTSB buffer containing 10 μM taxol. The protein samples were separated on a 7.5% one-dimensional SDS-PAGE gel.
We would like to thank I. Kreuzer, K. Guger, A. Redweik, N. Schlager, W. Öhr, and G. Resch for technical assistance. We are specially obliged to Farhah Assaad and Wolfgang Lukowitz for providing the ple-5 and ple-6 alleles. This work was funded by grants of the FWF Austrian Science Fund to M.T.H. (project numbers P14477-B04 and P16410-B12). S.M. was financed by the Austrian National Bank (Jubiläumsfondprojekt 5598) and the European Grant PL-960217. V.W. is funded by the DOC fellowship of the Austrian Academy of Sciences. A.S. and P.J.H. are funded by the Biotechnology and Biological Research Council UK.
Supplemental data are available with this article online at http://www.current-biology.com/cgi/content/full/14/5/412/DC1/.