A Spatiotemporal DNA Endoploidy Map of the Arabidopsis Root Reveals Roles for the Endocycle in Root Development and Stress Adaptation

Abstract

Somatic polyploidy caused by endoreplication is observed in arthropods, molluscs, and vertebrates but is especially prominent in higher plants, where it has been postulated to be essential for cell growth and fate maintenance. However, a comprehensive understanding of the physiological significance of plant endopolyploidy has remained elusive. Here, we modeled and experimentally verified a high-resolution DNA endoploidy map of the developing Arabidopsis thaliana root, revealing a remarkable spatiotemporal control of DNA endoploidy levels across tissues. Fitting of a simplified model to publicly available data sets profiling root gene expression under various environmental stress conditions suggested that this root endoploidy patterning may be stress-responsive. Furthermore, cellular and transcriptomic analyses revealed that inhibition of endoreplication onset alters the nuclear-to-cellular volume ratio and the expression of cell wall-modifying genes, in correlation with the appearance of cell structural changes. Our data indicate that endopolyploidy might serve to coordinate cell expansion with structural stability and that spatiotemporal endoreplication pattern changes may buffer for stress conditions, which may explain the widespread occurrence of the endocycle in plant species growing in extreme or variable environments.

Figure 1.

Peak Expression Distribution of Genes in 28 Clusters across Root Tissues and Sections. Displayed is the proportion (%) of genes in each cluster that are peak-expressed in any given marker line (top) or at any given developmental stage (slices, bottom) of the Arabidopsis root. Only clusters with expression patterns peaking at a single endoploidy level (indicated on top) are shown. For marker lines that do not mark all developmental stages of a particular tissue, the slices marked are given in parentheses. Tissues are abbreviated as follows: lateral root cap (LRC), columella (Colu), quiescent center (QC), atrichoblast (NHC), trichoblast (HC), cortex (Cor), endodermis (Endo), xylem pole pericycle (XPP), phloem pole pericycle (PPP), lateral root primordia (LRP), xylem (X), phloem (P), and phloem companion cells (PCC).

Figure 2.

Schematic Representation of the Mathematical Modeling Approach Used to Predict the Root Endoploidy Map. Shown is an example of how a root endoploidy map is predicted from the endoploidy-specific and spatiotemporal expression patterns of a single gene g. Initially, the endoploidy boundaries are randomly positioned on the map. Simulated annealing-based optimization is then used to shift endoploidy boundary positions along the longitudinal root axis and find the root endoploidy map that generates the best possible fit between the simulated and observed spatiotemporal expression profiles for the investigated gene (lower panels). The boxes in the simulated expression panels symbolize the slice s and marker line m for which root cells in the upper panels have been colored according to their predicted endoploidy levels (no real data). See Supplemental Movie 1 for a time-lapse movie of an example optimization run for a single gene.

Figure 3.

The Predicted Root Endoploidy Map and Experimental Validation. (A) The endoploidy map of the Arabidopsis root using a balanced set of 332 marker genes. (B) Difference map between the simulated endoploidy map in (A) and the map derived from flow cytometry measurements. Color hues indicate the difference in the number of endoreplication rounds between both maps for each cellular position. Hair cell slices 1 to 6 are unresolved in the flow cytometry map (gray).

Figure 4.

Experimental Mapping of Ploidy Borders in Atrichoblast and Trichoblast Cell Files. DAPI fluorescence was measured in 12 trichoblast (A) and atrichoblast (B) cell files. The fluorescence measurements across cell files were normalized so that the average fluorescence of the last three measured cells in each cell file equals 1600 (for trichoblasts) or 800 (for atrichoblasts), and the average of the first three measured cells equals 200, reflecting the expected endoploidy levels at the beginning and end of the measured cell files. A model with 3 (4) endoploidy boundaries and average DAPI fluorescence levels of 200, 400, 800, and 1600 for 2C, 4C, 8C, and 16C cells, respectively, was fit to the atrichoblast (trichoblast) data using a simulated annealing optimization routine in Matlab R2014b. Open circles indicate the average measured fluorescence (after normalization) across all cell files at any given cell number, and the black contours indicate the corresponding standard deviations (standard deviations <10, caused by insufficient numbers of data points at a few early and late cell numbers, were manually set to 100 to avoid technical optimization problems).

Figure 5.

Ploidy Map of Lateral Root Cap Cells. (A) Representative longitudinal mid-section of a DAPI-stained root meristem (of three independent roots imaged). (B) Extraction of nuclear outlines of the lateral root cap cells. Note that this image has been processed using MorphoGraphX to highlight the LRC nuclei only (see Methods). (C) DAPI fluorescence (arbitrary units) heat map. Nuclei fall into three categories likely corresponding to 2C (dark blue, n = 7), 4C (light blue-green, n = 16), and 8C (red, n = 2) nuclei. Bar = 20 μm. Data on independent roots are given in Supplemental Data Set 9.

Figure 6.

Temporal-Spatial Expression Patterns of the SIM and SMR1 Endocycle Markers. Cross sections of SMR1:GFP-GUS (A) and SIM:GUS (B) marker lines. The positions of the cross sections along the longitudinal root axis are indicated by red lines. The colored cross sections on the right depict the endoploidy levels predicted by the root endoploidy map shown in Figure 3A for the tissues marked by SMR1 (atrichoblast, cortex, xylem, and procambium; [A]) or SIM (trichoblast, atrichoblast, cortex, and phloem; [B]). Predicted 4C and 8C cells are colored blue and green, respectively. Bars for total root and cross sections are 100 and 20 μm, respectively.

Figure 7.

Functional Enrichment of Transcripts Peaking at a Given Endoploidy Level for Hormone and Stress Response GO Categories. (A) Hormones. (B) Stress conditions. Cells are colored according to their enrichment P value after Benjamini and Hochberg false discovery rate correction (only for corrected P ≤ 0.05). Columns, endoploidy level; rows, GO categories.

Figure 8.

Predicted and Validated Effects of Stress on the Root Endoploidy Content. (A) The predicted effect of stress conditions on root endoreplication levels. The x axis represents 149 publicly available stress/control expression profiling data sets, and the y axis depicts the sum of squares (SSQ) of the endoploidy weight differences in the stress and control endoploidy distributions (scaled to the range [0,1]). Red and green circles indicate that endoreplication is suppressed or promoted, respectively. The number between parentheses represents the stress condition index in Supplemental Data Set 12. The dotted line indicates a cutoff below which the conditions are not annotated on the panel. Condition keywords indicate: salt (117) = 150 mM NaCl, 6 h treatment; salt (119) = 150 mM NaCl, 12 h treatment; salt (121) = 150 mM NaCl, 24 h treatment; gold (1183) = 0.125 mM KAuCl₄, 6 h treatment; KNO₃ (413) = 5 mM KNO₃, 3.5 h treatment; Pi (1165) = Pi starvation, 6 h treatment; UV-B (182) = 15 min 1.18 W/m² Philips TL40W/12, 24 h after treatment; osmotic (206) = 300 mM mannitol, 12 h treatment; osmotic (1226) = 300 mM mannitol, 3 h treatment; osmotic (208) = 300 mM mannitol, 24 h treatment; salt (381) = 140 mM NaCl, 1 h treatment; salt (115) = 150 mM NaCl, 3 h treatment; sulfate (339) = sulfate limitation, 4 h treatment; boron (1203) = 5 mM boric acid, 12 h treatment; auxin (1176) = 1 µM IAA, 12 h treatment; auxin (1181) = 1 µM IAA, 8 h treatment; hot (218) = 38°C, 3 h treatment, 1 h recovery; UV-B (172) = 15 min 1.18 W/m² Philips TL40W/12, 0.5 h after treatment; cold (109) = 4°C, 24 h treatment; iron (379) = iron deficiency, 72 h treatment; osmotic (204) = 300 mM mannitol, 6 h treatment; genotoxic (166) = 1.5 µg/ml bleomycin + 22 µg/mL mitomycin, 12 h treatment; osmotic (202) = 300 mM mannitol, 3 h treatment; pH (11) = pH 4.6, 6 h treatment; hot (216) = 38°C, 3 h treatment, 0 h recovery. The dotted line indicates a cutoff below which the conditions are not annotated on the panel. (B) Predicted and validated endoreplication indices (EI) under stress conditions. Five-day-old seedling were transferred to either control medium or medium holding salt (140 mM NaCl), auxin (1 μM IAA), or low pH (4.6). Ploidy levels were measured in 5-mm root tips 48 h after treatment. Data represent mean ± sd (2 replicates with >200 seedlings per replicate used; *P < 0.05, two-sided Student’s t test).

Figure 9.

SMR1 Expression Precedes Cell Expansion. (A) Representative confocal image of a SMR1:GFP-GUS root counterstained with propidium iodide, displaying atrichoblast expression. (B) GFP expression heat map. (C) Cell volume heat map. The asterisks indicate the position of the first GFP signal. Note that for the purpose of visualization, MorphographX was used to digitalize the images. The confocal image was converted into a working stack to edit the stacked images, deleting the layers over and under the epidermis to enable visualization of the boundaries between epidermal cells and the presence of GFP. Bar = 50 μm.

Figure 10.

Volcano Plot Showing Transcripts That Are Up- and Downregulated in SMR1:GFP-GUS Fluorescent Protoplasts of Control (Col-0) versus smr1 (Col-0) Plants. Red, green, and blue dots correspond to cell wall modifying, chromatin-linked, and cell cycle genes, respectively. Gene IDs are given at the right.

Figure 11.

smr1 Mutant Cells Display Altered Shape. (A) Serial block-face scanning electron microscopy radial sections through the elongation zone of Col-0 and smr1 mutant plants. Trichoblasts, atrichoblasts, and cortex cells are false-colored in yellow, blue, and green, respectively. Bar = 25 μm. (B) Circularity of control (Col-0) and smr1 trichoblast cells within the transition zone. Data represent mean ± se (n > 20, *P ≤ 0.05, two-sided Student’s t test).