The preprophase band of microtubules controls the robustness of division orientation in plants

See allHide authors and affiliations

Science  14 Apr 2017:
Vol. 356, Issue 6334, pp. 186-189
DOI: 10.1126/science.aal3016

Refined understanding of the preprophase band

Because plant cells do not move, plant tissues are constructed according to how they place the divisions of their constituent cells. Schaefer et al. found a mutation in the model plant Arabidopsis that abolishes a visible precursor of cell division, the preprophase band. Despite loss of the band—previously thought essential to define the division plane—the general orientations of cell division planes in the roots of these plants were normal. However, individual division orientations showed more variance than normal. Thus, the preprophase band serves to focus and refine the final orientation of the nascent cell division plane.

Science, this issue p. 186


Controlling cell division plane orientation is essential for morphogenesis in multicellular organisms. In plant cells, the future cortical division plane is marked before mitotic entry by the preprophase band (PPB). Here, we characterized an Arabidopsis trm (TON1 Recruiting Motif) mutant that impairs PPB formation but does not affect interphase microtubules. Unexpectedly, PPB disruption neither abolished the capacity of root cells to define a cortical division zone nor induced aberrant cell division patterns but rather caused a loss of precision in cell division orientation. Our results advocate for a reassessment of PPB function and division plane determination in plants and show that a main output of this microtubule array is to limit spindle rotations in order to increase the robustness of cell division.

Because of the cell wall, plant cells have no capacity to migrate, and their three-dimensional (3D) organization is essentially established by the timing and orientation of mitoses. Precise control of the orientation of cell division is thus required to ensure the reliability of developmental processes. In plant cells, cytokinesis is achieved by de novo synthesis of a cell plate, initiated at telophase and growing centrifugally toward the cortex, where it ultimately fuses with the parental plasma membrane. The cortical division zone (CDZ), which subsequently narrows down to specify the cortical division site (1), is marked just before mitotic entry by the preprophase band (PPB), a dense ring of microtubules encircling the nucleus at the cortex (fig. S1 and movie S1). This microtubule array is proposed to play a role in CDZ definition by inducing cortical modifications that are later recognized by the growing cell plate (2). However, support for this theory has been hindered by the difficulty in interfering with PPB function without impeding other microtubule arrays, as in mutants such as ton1 and fass (36). In this study, characterization of a mutant impaired in PPB formation allowed us to assess the function of the PPB in mitotic Arabidopsis cells.

The TTP (TON1-TRM-PP2A) complex is a protein network involved in the spatial organization of cortical microtubules, including the PPB (7) (fig. S2). Within the complex, the TRM superfamily plays a role in complex assembly and targeting to the cytoskeleton (7, 8). By transcriptome analysis, we identified TRM7 as being regulated by the cell cycle (9), with its transcription peaking at the G2/M transition (Fig. 1A). Fusions between the TRM7 gene and a GUS reporter cassette in Arabidopsis plants confirmed that the TRM7 promoter drives cyclic expression in dividing cells (Fig. 1, B to E, and fig. S3) and that TRM7 is regulated by proteasome-mediated degradation through motifs located within its first 71 residues (fig. S4).

Fig. 1 TRM7 is a specific PPB marker.

(A) Absolute expression levels of TRM6, TRM7, TRM8, and two typical G2/M expressed genes (POK1 and KNOLLE) in aphidicolin-synchronized cell culture (9). The right y axis relates to KNOLLE expression only, and the left one relates to the four other genes. (B to E) GUS staining patterns in 4-day-old seedlings expressing the pTRM7::TRM71-71-GUS construct: whole plantlet (B), close-up on the shoot (D) and root (E) meristems, and cotyledon epidermis (C) (GUS reflection in blue in PI-stained tissue). (F) Root tip of a seedling expressing the TRM7-3xYFP construct. (G) Localization of the TRM7-3xYFP signal in root tip cells at the PPB, spindle, and phragmoplast (phragm.) stages, using a Cer-TUA6 marker to identify mitotic stages (table S1 and fig. S14A). White, no signal; light blue, cytoplasmic and cortical; dark blue, cytoplasmic only. No interphasic cell expressing the fusion was detected in 16 independent roots observed. (H and I) Cross section (H) and surface view (I) of a cell at the PPB stage expressing the TRM7-3xYFP (gray) and the Cer-TUA6 (blue) markers. Scale bars, 2 mm (B), 20 μm [(C) and (F)], 50 μm (E), 10 μm [(H) and (I)].

We constructed a functional fluorescent fusion by recombineering (recombination-mediated genetic engineering) a 3xYFP (yellow fluorescent protein) marker at the C terminus of TRM7 within a 63-kb genomic fragment (fig. S5). In root tips, only G2/M cells expressed the fusion, showing a diffuse cytoplasmic signal together with a well-defined cortical ring colocalizing with PPB microtubules (Fig. 1, F to I, and fig. S6). Observation of TRM7-3xYFP in dividing root tip cells showed that at PPB stage, most cells (87%) displayed a dual cytoplasmic and cortical signal (Fig. 1G). At spindle stage, most of the fluorescence was only cytoplasmic (72% of cells), whereas at phragmoplast stage, the signal was no longer detectable in 95% of the cells. TRM7 thus has a narrow expression window and specific subcellular localization at the PPB, defining a G2/M specific isoform of the TTP complex.

In Arabidopsis, TRM7 is part of TRM group 2, together with TRM6 and TRM8 (8). Neither TRM6 nor TRM8 display cell-cycle regulation (9) (Fig. 1A). In root tip cells, in accordance with transcriptome data, TRM6-3xYFP was not detected, whereas TRM8-3xYFP displayed a constitutive expression as a diffuse cytoplasmic signal, except at the G2/M stage, where the TRM8 fusion labeled the PPB (fig. S6B).

To further characterize the function of TRM6-8 in dividing cells, we recovered mutant lines for each gene (fig. S7). All homozygous mutant combinations (single, double, and triple) were viable and fertile (fig. S8). Using tubulin immunolocalization, we assessed microtubule array formation in root and shoot apical meristematic cells (Fig. 2 and fig. S9). In the root meristem, mutations in TRM6 or TRM8 did not cause PPB defects, whereas TRM7 disruption partially suppressed PPB formation, with only 57% of mutant cells forming a wild-type PPB. When combined with the trm6 or trm8 mutation, these defects were enhanced, and the triple trm678 mutant was essentially devoid of normal PPB both in the root and the shoot meristems. Interphasic microtubules (fig. S10) and noncortical mitotic arrays (fig. S11) appeared normal in mutant lines. Thus, TRM7 activation in late G2 is necessary for proper PPB formation, TRM6 and TRM8 being present constitutively. The trm678 mutant hence provides a tool to analyze PPB function.

Fig. 2 Mutations in the TRM6-8 group affect PPB formation.

(A to C) Coimmunolocalization in wild-type (A), trm78 (B), and trm678 (C) root-tip cells using an antibody to KNOLLE as a G2/M marker (magenta) and an antibody to tubulin as a microtubule marker (green). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (blue). Scale bars, 5 μm. (D) Quantification of normal PPBs [white, as in (A)], abnormal PPBs with sparse microtubules at the cortex, and a high density of perinuclear microtubules [light blue, as in (B)] and absence of PPB [dark blue, as in (C)] in combinations of trm6, trm7, and trm8 mutations (see also table S2 and fig. S14B).

To assess the phenotypic consequences of PPB loss, root tips from trm678 and wild-type plants were stained with propidium iodide (PI) and imaged in 3D. In mutant roots, tissue organization was globally normal, although wall placement was less regular and some cell files were locally distorted (Fig. 3, A and B; fig. S12A; and movies S2 and S3), in contrast with previously characterized cell division mutants (fig. S12, B to I). Cell differentiation and the continuity of cell layers were not affected in mutant roots, as judged from markers labeling the endodermis, quiescent center, or protophloem (fig. S12, N to P). Therefore, PPB impairment in the trm678 mutant did not compromise the anatomical organization of the root.

Fig. 3 PPB absence impairs the robustness of cell division orientation.

(A) Longitudinal sections of PI-stained wild-type and trm678 roots. Arrowhead indicates local periclinal cell file duplication. (B) Three-dimensional views of a 100-μm-long section above the root cap of 3D-segmented cell layers of wild-type and trm678 (blue, epidermis; magenta, cortex; yellow, endodermis). (C) Combined Tukey’s box plot/violin plot diagram of division angles in cells of epidermis, cortex, and endodermis. The measured angles correspond to the angle between transverse walls (green) and the red line joining cell centroids (left). (D) Recently divided sister cells were selected from a longitudinal root section (left), were 3D-segmented (middle), and their volume ratios were measured. Combined Tukey’s box/violin plot diagram of the volume ratio between sister cells (right). (E) Orientation of metaphase plates in PI/DAPI-stained root tip cells. The four upper panels show anticlinal divisions. The two lower panels illustrate a >90° rotation of the metaphase plate [longitudinal (left) and transverse (right) view]. Combined Tukey’s box/violin plot diagram of angles between the metaphase plate and the cell file’s axis (right). In (C), (D), and (E), wild-type distributions are in red and trm678 ones are in blue; the mean is indicated as a red dot. Statistical details are in table S3. Scale bars, 25 μm.

We quantified the orientation of symmetric cell division, focusing on the three external root cell layers (Fig. 3C). For all three layers, the mean angle between the transverse wall and the cell file’s axis was identical between the mutant and the wild type. However, in all three tissues, the variance of angles showed a significant (2- to 3.2-fold) increase in mutant cells as compared with the wild type. The symmetry of cell division was further investigated by measuring the volume ratio of recently divided sister cells in the root meristem (Fig. 3D). The mean volume ratio was similar between the wild type and the mutant, but again, its variance showed a 1.6-fold increase in the mutant (table S3B). Such effects on variance without affecting the mean are characteristic of a process acting as a stabilizer rather than as a primary determinant (10).

In trm678 cells entering mitosis, ring-shaped accumulation of cortical microtubules constituting the hallmark of PPB formation was replaced by a dense concentration of perinuclear microtubules (fig. S9) that evolved into a spindle without apparent formation of polar caps that, in wild-type cells, establish spindle bipolarity just before nuclear envelope breakdown (fig. S11). This prompted us to investigate the effect of the trm678 mutation on spindle orientation (Fig. 3E). Spindle rotations are common in wild-type cells but are generally corrected at the phragmoplast stage (11). In mutant roots, 10% of cells displayed anticlinal to periclinal rotation of the metaphase plate to the point that an angle could not be measured, leading to plane switches likely at the origin of local cell file duplications. Such rotations were never observed in wild-type cells. After removal of these extreme cases, comparison of the remaining data points showed that, although the mean orientation was similar, the variance was increased 3.7-fold in the mutant. These results show that PPB disruption has a major effect on spindle positioning.

From the spatial correlation between PPB position and the final site of division (12), it is assumed that the PPB defines the CDZ by targeting molecules that preserve its positional information throughout mitosis (2). However, trm678 mutant plants did not exhibit the highly disorganized cellular topology observed upon disruption of CDZ factors such as phragmoplast-orienting kinesin (POK) (fig. S12) (13). Therefore, we assessed CDZ establishment in trm678 cells by analyzing the dynamics of POK1 localization in dividing root cells (Fig. 4). In the wild type, POK1 localization shifted from cytoplasmic at PPB stage to a ring-shaped cortical distribution at phragmoplast stage, as described in (13). In mutant cells, targeting of YFP-POK1 to the cortex was delayed (Fig. 4H). At preprophase stage, YFP-POK1 was mainly cytoplasmic. At spindle stage, 63.5% of cells displayed a mixed localization, and at phragmoplast stage, 91% exhibited a ring-shaped cortical signal, looking normal in 73% of the cases (table S4). Thus, in trm678 mutant cells, even though the timing and efficiency of POK1 targeting from the cytoplasm to the cortex appear altered in PPB absence, POK1 retains its capacity to translocate to the cortex and to form a cortical ring that most probably corresponds to a CDZ.

Fig. 4 PPB-less cells retain their capacity to target a CDZ factor to the cortex.

(A to C) Images of Col0 (left) and trm678 (right) cells at preprophase (A), spindle (B), or phragmoplast (C) stages (gray, POK1-YFP fluorescence; blue, antibody to tubulin). (D to G) Abnormal POK1 localization in trm678 cells. Most trm678 preprophase cells displayed cytoplasmic localization of POK1 (D). When at the cortex at later stages, the POK1 signal was sometimes discontinuous—i.e., present on one side of the cell but not on the other [arrowheads in (E) and (G)] or more loosely defined [asterisks in (E) and (F); compare to (C)]. Scale bars, 10 μm. (H) Distribution of POK1 localization in Col0 and trm678 root tip cells at the preprophase (noted PPB for Col0 and PP for trm678), spindle, and phragmoplast (Phrag.) stages. Cells were visually classified into five categories depending on the fluorescence signal ratio between the cortex and the cytoplasm (from white for only cytoplasmic to dark blue for only cortical) (table S4 and fig. S14, C and D).

In contrast to core-TTP mutants that display severe dwarfism and distorted organs (fig. S12) (4, 5), PPB disruption in trm678 did not lead to major developmental defects or lethality (fig. S13). Mutant plants grew in soil and produced all vegetative and reproductive organs, including fertile flowers and viable seeds, showing that plant morphogenesis is tolerant to PPB disruption. However, trm678 mutant plants were smaller, they produced 20% fewer seeds, and invariant traits like the number of cotyledons, sepals, or petals became variable in the mutant, indicating a general loss of growth capacity and developmental robustness.

Taken together, our results advocate for a reassessment of PPB function in symmetric cell division. Rather than a causal determinant of the cell division plane, the PPB appears as a noise-reducing module that aligns to a predefined spatial cue, thereby providing a sturdy equatorial reference for orientation and bipolarity of the spindle, as already proposed (14). Such a function in spindle positioning is reminiscent of the role of centrioles and astral microtubules during mitosis of centrosomal cells (15) and may be linked to the similarity of many TTP proteins to animal centrosomal proteins (8). In turn, spindle stabilization allows one to filter out most of the noise in cell division plane positioning and to increase the robustness of plant development. Spindle stability thus emerges as an important parameter of division plane specification in plants, as in other eukaryotes (16).

Supplementary Materials

Materials and Methods

Figs. S1 to S14

Tables S1 to S4

Movies S1 to S3

References (1748)

References and Notes

  1. Acknowledgments: We are grateful to S. Müller, J. Alonso, G. Jürgens, T. Schmülling, B. Scheres, and S. Huang for sharing materials and resources. E.S. was the recipient of a Ph.D. fellowship from the French Ministry for Research, and Y.D. was funded by the Agence Nationale de la Recherche (ANR-08-BLAN-0056). The Institut Jean-Pierre Bourgin benefits from the support of the LabEx Saclay Plant Sciences (ANR-10-LABX-0040-SPS) and from the région Île-de-France. The supplementary materials contain additional data.
View Abstract

Navigate This Article