A Spatial Accommodation by Neighboring Cells Is Required for Organ Initiation in Arabidopsis

See allHide authors and affiliations

Science  10 Jan 2014:
Vol. 343, Issue 6167, pp. 178-183
DOI: 10.1126/science.1245871

Make Way for the Emerging Rootlet

Plant cells are immobilized by their rigid cells walls, and the root endodermal cell layer maintains a impervious perimeter seal made of an indigestible irregular polymer. Despite these mechanical obstacles, lateral root primordia, which initiate in the deep layers of the root, manage to break through to the surface. Vermeer et al. (p. 178; see the cover) used live-tissue imaging and genetics to show that signals are exchanged between the root primordium and the handful of cells overlying it, which then cave in on themselves to open up a channel for the growing root primordium.


Lateral root formation in plants can be studied as the process of interaction between chemical signals and physical forces during development. Lateral root primordia grow through overlying cell layers that must accommodate this incursion. Here, we analyze responses of the endodermis, the immediate neighbor to an initiating lateral root. Endodermal cells overlying lateral root primordia lose volume, change shape, and relinquish their tight junction–like diffusion barrier to make way for the emerging lateral root primordium. Endodermal feedback is absolutely required for initiation and growth of lateral roots, and we provide evidence that this is mediated by controlled volume loss in the endodermis. We propose that turgidity and rigid cell walls, typical of plants, impose constraints that are specifically modified for a given developmental process.

Epithelia are central to multicellular life. Their ringlike paracellular barriers separate different environments, and their polar surfaces mediate selective and vectorial uptake of substances (1). The crucial barrier function of epithelia must be maintained during growth and development. In animals, epithelial remodeling involves a complex, supracellular interplay of force generating cytoskeleton and dynamically remodeled adherens junctions (2). In plants, the root endodermis has a function very similar to that of animal epithelia, yet its independent evolution in the context of a multicellular organism with turgid, wall-bearing, nonmotile cells has led to profoundly different cellular structures (3). Instead of being mediated by direct protein-protein interactions, the paracellular diffusion barrier of the endodermis is set up by the Casparian strips, ringlike, hydrophobic impregnations of the primary cell wall that fuse into a supracellular network between endodermal cells. These impregnations consist of lignin, an inelastic phenolic polymer that is resistant to chemical degradation (4). The Casparian strip establishment is locally guided by the Casparian strip domain proteins (CASPs). These transmembrane proteins, which form a ringlike plasma membrane subdomain, establish a lateral diffusion barrier and recruit biosynthetic enzymes for Casparian strip formation (5, 6). Yet, despite the presence of a Casparian strip, the endodermis has to be remodeled during lateral root formation.

Lateral roots are formed from the pericycle, a cell layer located deep within the primary root, confined between the vascular bundle and the endodermis. As the lateral root primordium grows, it has to traverse the endodermis, cortex, and epidermis in order to emerge from the primary root. The hormone auxin triggers lateral root development and also signals to the overlying cell layers (7). Outer cortex and epidermal cell layers assist lateral root emergence after auxin from the lateral root primoridum triggers degradation of the SOLITARY-ROOT repressor protein, which permits expression of pectinases that disrupt intercellular adhesions (8, 9). The endodermis, however, must be breached first, and its Casparian strip network cannot be degraded by pectinases. Another repressor protein, SHORT HYPOCOTYL 2 (SHY2), mediates endodermis responses (9).

Older electron microscopy studies suggest a profound remodeling of endodermal cell walls during lateral root formation (10, 11). In order to allow for imaging of the deep-lying endodermal cell layer, we specifically developed a set of fluorescent marker lines (12). A transcriptional fluorescent reporter for SHY2 shows expression in two to three endodermal cells overlying newly formed primordia during stage I to stage IV (fig. S1, A to D, and movies S1 to S4). Subsequently, we used markers for the Casparian strip domain (CSD), as well as plasma and vacuole membranes, that we expressed specifically in differentiating endodermal cells. We observed that the CSD marker CASP1-mCherry was locally degraded in two to three endodermal cells overlying stage II and III primordia (Fig. 1A); at the same time, endodermal cells overlying the primordium became thinner (Fig. 1J and fig. S2). We used the plasma membrane–localized soluble N-ethylmaleimide–sensitive factor attachment protein receptor (SNARE) SYP122, CITRINE-SYP122, to visualize the endodermal protoplast during lateral root primordium growth.

Fig. 1 Regulated endodermal cell responses during lateral root formation.

(A) CSD marker CASP1pro::CASP1:mCherry (magenta) gets degraded in endodermal cells overlying the primordium. (B and C) Remodeling of the endodermal protoplast during lateral root formation. Protoplast visualized by CASP1pro::CITRINE-SYP122 (magenta). (D and E) Remodeling of endodermal vacuole during lateral root formation. Vacuoles marked by CASP1pro::γ-TIP:CITRINE (magenta). (A, C, and E) PIN1pro::PIN1:GFP (green) used to label primordia. (F) Integrity of plasma membrane [marked as in (B) and (C)] is maintained in overlying endodermal cells during lateral root formation. Counterstained with propidium iodide (PI) (green). Arrow indicates endodermal cell with absence of nuclear propidium iodide staining. (G) Emerging lateral root pushing away loosened epidermal cells. Note difference to splitting of endodermis in (B). Epidermal plasma membrane marked by UBQ10pro::EYFP:NPSN12. (H and I) Basic fuchsin staining of Casparian strip reveals localized degradation (arrows) of the strip during lateral root formation. (J) Schematic of thinning of endodermal cells (e) (gray) overlying a pericycle or lateral root primordium (p) (yellow). (K and L) Schematic of a surface view of a lateral root primordium (yellow), highlighting differences of growth through the endodermis (gray, K) versus epidermis (blue, L). Seven-day-old seedlings were used. Images (A) to (I): Maximal projections of confocal image stacks. Scale bar, 20 μm.

We observed that the protoplast progressively flattened in parts of the cells, which sometimes led to a donut-shaped structure (Fig. 1, B and C; fig. S3; and movies S5 and S6) that could arise from a flattening of the endodermal cell to the point where plasma membranes from both sides are fusing. Using the tonoplast marker γ-TIP-CITRINE, we observed similar accommodating responses (Fig. 1, D and E; fig. S4; and movies S7 and S8). The vacuole appeared to be remodeled and to fragment into smaller vacuoles but otherwise remained intact. The integrity of the plasma membrane of endodermal cells was at no point compromised during these rearrangements, as indicated by the absence of nuclear propidium iodide stain (used as an indicator of cell death in plants) (Fig. 1F and movie S9).

The protoplast of the endodermis and epidermal cells behaved differently. Epidermal cells do not seem to lose volume, but rather are pushed away after loss of cell-to-cell adherence (compare Fig. 1G with 1B and Fig. 1K with 1L). By contrast, the Casparian strip network keeps the endodermal cells tightly connected, which forces the endodermis to accommodate growth through dramatic volume losses and the required minimal separation of cell walls for lateral root emergence. When we visualized the Casparian strip during lateral root formation, we did not observe a complete disappearance of the strip that would parallel the disappearance of the CASP1 membrane protein (Fig. 1A). Instead, we observed small holes and/or breaking points that allowed for a localized opening of the network in parts of the cell where the primordium penetrates (Fig. 1, H and I).

The SHY2 repressor is thought to regulate endodermal responses during lateral root emergence. Yet SHY2 is widely expressed, and mutants with the dominant shy2-2 allele show many different developmental phenotypes (1315), which makes it difficult to discern primary from secondary defects. To overcome this problem and to specifically block auxin responses in differentiated endodermal cells, we expressed the stabilized form of SHY2, shy2-2, under the control of the CASP1 promoter. In these CASP1pro::shy2-2 plants, the altered leaf shape, severe dwarfism of shoots and roots, and stunted inflorescence phenotypes of the shy2-2 mutant were all absent, although the plants remained smaller than wild type (Fig. 2A). This smaller stature can easily be explained as a secondary consequence of the only other phenotype observed in these plants, which was a complete absence of emerged lateral roots (Fig. 2, B to D). This very strong phenotype became even more apparent under hydroponic growth conditions. Whereas 3-week-old Col-0 plants had a well-developed root system, CASP1pro::shy2-2 plants only showed one long primary root (Fig. 2C). This complete suppression of lateral root emergence was specific for shy2-2 expression in the endodermis, as lines expressing shy2-2 in the cortex (PEPpro) or epidermis (PGP4pro) (16, 17) still had emerged lateral roots (Fig. 2E). The lower rate of emerged lateral roots of the PGP4pro::shy2-2 plants may be due to weak PGP4 promoter expression in endodermis and pericycle (fig. S5).

Fig. 2 Endodermal auxin responses are essential for lateral root formation.

(A) Three-week-old CASP1pro::shy2-2 plants do not display shy2-101 phenotypes but are smaller than Col-0 and CASP1pro::SHY2 controls. (B) CASP1pro::shy2-2 seedling roots grow as controls do but show no emerged lateral roots. CASP1pro::shy2-2 seedlings also do not display clusters of emerged lateral roots (asterisks) sometimes observed in shy2-101. (C) Three-week-old hydroponically grown plants show highly branched root system in wild-type (Col-0). CASP1pro::shy2-2 plants still show no emerged lateral roots. (D) CASP1pro::shy2-2 seedlings [10 days after germination (DAG)] have no emerged lateral roots (LRs). Asterisks indicate significant differences among means. (E) Blocking auxin responses in endodermis inhibits lateral root emergence, whereas blocks in cortical (PEPpro::shy2-2) and epidermal (PGP4pro::shy2-2) responses do not. Asterisks indicate significant differences among means (*P < 0.5 × E–5, **P < 0.5 × E–10, ***P << 0.5 × E–15) by analysis of variance (ANOVA) and Tukey’s test as post hoc analysis (n = 20). All experiments were performed at least three times.

Our experiments suggest that endodermal auxin signaling is crucial for lateral root emergence. We then analyzed at which stage of lateral root formation CASP1pro::shy2-2 plants were blocked. Unexpectedly, even the earliest lateral root primordia stages were absent (fig. S6). We concluded that endodermal auxin signaling already affects initiation of lateral roots in the pericycle. We therefore characterized the expression of the lateral root founder cell marker GATA23pro::NLSGFPGUS (18). Onset of GATA23 expression in CASP1pro::shy2-2 roots was similar to that in wild type, but the staining intensity did not increase to mark initiating and early lateral root primordia as in wild type (Fig. 3, A and B). This supports a very early block in lateral root initiation in CASP1pro::shy2-2 plants, placing it before the first asymmetric divisions of the founder cells.

Fig. 3 Endodermal auxin responses are required for lateral root initiation and emergence.

(A and B) Lateral root founder cell marker GATA23 is misregulated in CASP1pro::shy2-2 seedlings. Staining of GATA23pro::NLSGFP:GUS expression in Col-0 (A) and CASP1pro::shy2-2 (B) seedlings (5 DAG). (C and D) CASP1pro::shy2-2 roots are not resistant to NAA. Four-day-old seedlings were transferred to indicated amounts of NAA for 3 days. Compared to Col-0 (C), CASP1pro::shy2-2 seedlings (D) show no emerged lateral roots after prolonged NAA treatment. (E to H) The auxin signaling reporter DR5pro::N7-3xVENUS shows comparable fluorescence intensity in control situation in both Col-0 (E) and CASP1pro::shy2-2 seedlings (F). Both endodermis and pericycle of CASP1pro::shy2-2 seedlings (H) are less responsive to auxin, compared to Col-0 (G). a.u., arbitrary units. Five-day-old plants were incubated in liquid medium for the indicated time and concentration. (I to L) Blocking endodermal auxin responses severely affects lateral root primordium shape. Four-day-old seedlings expressing the plasma membrane marker UBQ10pro::EYFP:NPSN12 in Col-0 (I and J) and CASP1pro::shy2-2 (K and L) were transferred to auxin-containing plates. (J and L) Orthogonal views of (I) and (K), respectively. Note that induced primordium of CASP1pro::shy2-2 is much broader than Col-0 [compare outlines in (J and L)]. In addition, the endodermis does not become flattened in CASP1pro::shy2-2 roots (highlighted by asterisks). Scale bars: 0.1 mm in (A), (B), (E), and (F); 1 cm in (C) and (D); 50 µm in (G) and (H); and 20 μm in (I) to (L).

We tested whether auxin treatments could rescue the lateral root phenotype of CASP1pro::shy2-2 plants. Primary root growth in both backgrounds was equally sensitive to auxin treatments (Fig. 3, C and D). CASP1pro::shy2-2 plants, however, formed very few lateral roots in response to auxin. High concentrations resulted in some degree of lateral root formation in CASP1pro::shy2-2 plants (fig. S7). Lateral roots formed when 1 μM auxin was present hardly managed to emerge (fig. S7); this was the initially expected emergence phenotype. Auxin treatment of wild-type plants resulted in correctly shaped primordia with normal emergence through the endodermis (Fig. 3, I and J). By contrast, CASP1pro::shy2-2 primordia were often flattened and appeared to have difficulties in growing through the endodermis (Fig. 3, K and L). In addition, the endodermis appeared to stay turgid much longer than wild-type endodermis, where it becomes flattened to the point of being imperceptible (see asterisks, Fig. 3, I and K). These experiments suggest that blocking auxin responses specifically in the endodermis affects lateral root formation at two stages: first, at initiation and, second, at the stage when the primordium needs to emerge. Thus, blocking responses in the endodermis has a non–cell autonomous effect on the ability of auxin to induce cell divisions in the pericycle. This effect was confirmed by using the auxin signaling reporter DR5pro::NLS3xVENUS (19). Under control conditions, there was no obvious difference between the two backgrounds (Fig. 3, E and F), whereas, after short-term auxin treatments, not only the endodermis but also the pericycle responded less in CASP1pro::shy2-2 (Fig. 3, G and H).

Several lines of evidence exclude the possibility that this effect on auxin response competence in the pericycle is due to the direct presence of shy2-2 protein in those cells, either because of leaky promoter activity or protein movement. First, unlike CASP1pro, PGP4pro also displays weak expression in the pericycle (fig. S5), yet it causes a much weaker lateral root phenotype (Fig. 2E). Second, we ruled out movement of shy2-2 protein from endodermis into the pericycle by interfering with plasmodesmatal transport between these two cell layers through endodermal expression of cals3m, which has been shown to obstruct plasmodesmatal connections (20). The phenotype of CASP1pro::shy2-2/CASP1pro::cals3m plants was very similar to CASP1pro::shy2-2 plants (fig. S8), which indicated that protein movement from endodermis to pericycle is not the cause for the non–cell autonomous effect that we observe.

Having excluded direct effects of shy2-2 in the pericycle, we postulated that communication between the endodermis and the primed pericycle cells is required much earlier than previously assumed. We therefore investigated whether remodeling of endodermis and pericycle already occurs at stage I or earlier. We used light sheet fluorescence microscopy to noninvasively image and quantify the dynamics of pericycle and endodermis before and during the first asymmetric cell division. We could show that the transverse area of the pericycle cells increased by ~50% (Fig. 4, A to C, and movie S10). Average increase in cell surface was 22 ± 15 μm2/hour (mean ± SD, n = 9 cells, two biological replicates). This indicates that, before the first division, the pericycle cells swell and the endodermis slightly shrinks or deforms.

Fig. 4 Pericycle cells swell before formative divisions.

(A to C) Four-dimensional light sheet fluorescence microscopy reveals that pericycle cells swell before formative divisions. Seedlings (7 DAG) expressing UBQ10pro::EYFP:PIP1;2 as the plasma membrane marker (gray) and nuclear GATA23pro::NLSGFPGUS (gray) as the lateral root founder cell marker imaged for 30 hours. (A) Single slice of two dividing pericycle cells. Yellow star indicates pericycle; red star indicates endodermis. (B) Orthogonal views of (A) at t = 0 hours and t = 3.58 hours and segmented versions of the images (right). Pericycle is indicated with yellow asterisk and/or filled, endodermis with red asterisk and/or filled. (C) The sum area of the three cells, normalized to its initial value, is plotted as black dots; the yellow and red lines are sliding window averages over six time points. (D) Boxplot showing the quantification of the distribution of pericycle cell widths under control, either 10 μM NAA or 10 μM NPA with 10 μM NAA conditions in Col-0 (white boxes) and CASP1pro::shy2-2 background (gray boxes). Measurements were performed on orthogonal projections of confocal Z-stacks of UBQ10pro::EYFP:NPSN12 roots as depicted in fig. S9. Different letters indicate significant differences among means (P < 0.5 × E–4 by ANOVA and Tukey’s test as post hoc analysis). (n = 25 roots, three independent experiments.) Scale bar, 20 μm.

Under short-term auxin treatments, we observed a significant increase of pericycle cell width in Col-0. However, this increase was absent in CASP1pro::shy2-2 plants (Fig. 4D and fig. S9). This suggests that the ability of pericycle cells to swell is dependent on an early auxin perception in the endodermis. In CASP1pro::shy2-2 plants, pericycle-derived auxin would be unable to induce an early accommodation response of the endodermis, which would make it impossible for the primed pericycle cell to increase in volume and execute the first cell division. Auxin transporter–mediated “reflux” of auxin from the endodermis to the pericycle was reported to assist the progression of founder cells into stage I primordium (21). Yet the weak effects of endodermal auxin transporters on lateral root formation indicate that it cannot account for the strong block of lateral root formation reported here. Nevertheless, we included measurements of pericycle cell widths after simultaneous treatment of 1-naphthylacetic acid (NAA) and an auxin transport inhibitor (22). Inhibition of auxin transport affected neither the auxin-induced changes in pericycle cell width nor the ability of CASP1pro::shy2-2 expression to block this increase (Fig. 4D and fig. S9). Hence, the non–cell autonomous effects observed are not mediated by alterations of auxin transport through the endodermis.

We propose that pericycle cells perceive a nonresponsive endodermis through an increased resistance to its expansion growth. This mechanical stress would then result in an early block of lateral root initiation. Auxin-mediated regulation of the mechanical properties of overlying cell layers was proposed to contribute to the shape of already-formed lateral root primordial (23). Here, we reveal that the requirement for accommodating responses is much more fundamental, being even required for initiation of cell division. Moreover, we show that it is specifically the endodermis that plays a key role in lateral root formation and emergence, because of the strong phenotypes of CASP1pro::shy2-2 plants. Our findings fit with a recent report that auxin-mediated regulation of aquaporins contributes to emergence (24). Clearly, auxin regulation of aquaporins or ion channels might execute the very early accommodating responses of the endodermis that we observe. It will be interesting to specifically manipulate such proteins in the endodermis and study their effects on endodermal shrinkage and lateral root formation. The pericycle-endodermis interaction that we have uncovered now provides a promising system to study how plant cells sense and accommodate to growth of their neighbors—besides lateral root formation, such “mechanical” communication might be of importance to maintain tissue integrity in growing apical meristems or during secondary growth, as well as during the penetration of fungi or pollen tubes into plant tissue or infection thread formation. Generally, the need for coordinated accommodation responses by turgor and/or volume loss might be widespread, wherever there is growth and division of inner cell layers. This might impact on tissue and organ patterning, just as perception of mechanical stresses at surface cell layers shapes growth of the shoot apical meristem (25).

Supplementary Materials

Materials and Methods

Figs. S1 to S9

References (2630)

Movies S1 to S10

References and Notes

  1. Materials and methods are available as supplementary materials on Science on the Web.
  2. Acknowledgments: We thank the Central Imaging Facility (CIF) of the University of Lausanne for technical support, T. Beeckman, M. Bennett, B. De Rybel, Y. Heluriatta and H. Fukaki for sharing published material. We thank T. Beeckman, M. Bennett and T. Goh for helpful comments. We thank B. de Rybel for sharing unpublished results. This work was supported by a Marie-Curie Intra-European Fellowship grant to J.E.M.V., an EMBO long-term fellowship for M.B. Supporting grants were from the Swiss National Science Foundation (SNSF) and the European Research Council (ERC) (to N.G.); the land Baden-Württemberg, the Chica und Heinz Schaller Stiftung, and the CellNetworks cluster of excellence (to A.M.); and the Cluster of Excellence at Frankfurt for Macromolecular Complexes (CEF-MC) and the Deutsche Forschungsgemeinschaft (to D.v.W., E.H.K.S.).
View Abstract

Stay Connected to Science

Navigate This Article