ERF115 Controls Root Quiescent Center Cell Division and Stem Cell Replenishment

See allHide authors and affiliations

Science  15 Nov 2013:
Vol. 342, Issue 6160, pp. 860-863
DOI: 10.1126/science.1240667

The Root of the Problem

The quiescent center (QC) within the root meristem plays a key role as a stem cell organizer to sustain the root stem cell niche. The QC cells execute a dual role: prevention of the differentiation of neighboring stem cells, and maintenance of the root structure by undergoing only occasional cell division. The mechanisms that account for the low QC proliferation are unclear, although the anaphase-promoting complex/cyclosome (APC/C) E3 ubiquitin ligase is known to suppress QC cell division. Through a systematic functional analysis of APC/C-copurifying proteins, Heyman et al. (p. 860) characterized a transcription factor ERF115 as a rate-limiting factor for QC cell division. ERF115 needs to be destroyed to retain QC cells in a resting state. ERF115 operates in a brassinosteroid-dependent manner and controls QC cell division through transcriptional activation of phytosulfokine signaling.


The quiescent center (QC) plays an essential role during root development by creating a microenvironment that preserves the stem cell fate of its surrounding cells. Despite being surrounded by highly mitotic active cells, QC cells self-renew at a low proliferation rate. Here, we identified the ERF115 transcription factor as a rate-limiting factor of QC cell division, acting as a transcriptional activator of the phytosulfokine PSK5 peptide hormone. ERF115 marks QC cell division but is restrained through proteolysis by the APC/CCCS52A2 ubiquitin ligase, whereas QC proliferation is driven by brassinosteroid-dependent ERF115 expression. Together, these two antagonistic mechanisms delimit ERF115 activity, which is called upon when surrounding stem cells are damaged, revealing a cell cycle regulatory mechanism accounting for stem cell niche longevity.

Plant root growth and development depend on the continuous generation of new cells by the stem cell niche that is located in the proximal zone of the root meristem. Key to the maintenance of the stem cell niche are a small group of organizing cells, the quiescent center (QC) (14). QC cells divide with a frequency lower by a factor of 3 to 10 than mitotically active root cells (2, 57). Combined with the suppression of stem cell differentiation, a low QC proliferation rate is fundamental to maintain root structure and meristem function (7). Whereas inhibition of stem cell differentiation is controlled through the retinoblastoma pathway (8), the molecular components that control the QC cell division rate remain unknown. The Arabidopsis thaliana CELL CYCLE SWITCH 52 A2 (CCS52A2) activating subunit of the anaphase-promoting complex/cyclosome (APC/C), a highly conserved E3 ubiquitin ligase that marks cell cycle proteins for destruction, restrains QC cell division (9). CCS52A2 copurifying proteins identified through tandem-affinity purification (fig. S1) (10) were screened for their ability to promote QC cell proliferation upon ectopic expression. Among these, the ethylene response factor 115 (ERF115) resulted in a QC cell division phenotype that mimicked that of ccs52a2-1 knockout plants (Fig. 1, A to C). Expression of the WOX5-GFP (green fluorescent protein) marker confirmed that it was the QC cells that divided (fig. S2).

Fig. 1 ERF115 drives QC cell division and is a substrate of the APC/CCCS52A2.

(A to C) QC cells of 1-week-old wild-type (Col-0) (A), ccs52a2-1 mutant (B), and ERF115OE (C) plants visualized by confocal microscopy. Cell walls were counterstained with propidium iodide. Scale bars, 20 μm. (D to I) ERF115-GFP protein gel blot analysis with an antibody against GFP on 5-day-old wild-type (D), ccs52a2-1 mutant (F), and ccs52a1-1 mutant (H) seedlings control-treated (–) or treated (+) with the MG132 proteasome inhibitor (100 μΜ). Arrow indicates the position of ERF115-GFP fusion protein. A nonspecific cross-reacting protein was used as a loading control. Quantified proteins levels are corrected for the expression levels [(E), (G), and (I)]. The protein level of ERF115-GFP without MG132 was arbitrarily set to one. Data represent mean ± SD (n = 2 independent pull-down assays). (J) ERF115-GFP protein gel blot analysis with an antibody against GFP on Nicotiana benthamiana leaf spots infiltrated with constructs encoding ERF115-GFP and its D-box mutated variants ERF115-D1 (N-terminal D-box mutated; RVWL→AVWA), ERF115-D2 (C-terminal D-box mutated; RAQL→AAQA), and ERF115-D1/D2 (both D-boxes mutated). LC, loading control of Coomassie-stained nonspecific bands.

ERF115 (At5g07310) belongs to the ETHYLENE RESPONSE FACTOR family of transcription factors that control the transcription of genes linked to various biological processes related to growth and development. Biochemical data validated that ERF115 is a proteolytic target of APC/CCCS52A2. The proteasome inhibitor MG132 stabilized the chimeric ERF115-GFP reporter in a CCS52A2-dependent manner (Fig. 1, D to G, and fig. S3). In contrast, knockout of the paralogous CCS52A1 gene, which controls the timing of cell cycle exit of the root cells within the cell elongation zone through cyclin destruction (6, 11), did not affect proteolysis of ERF115 (Fig. 1, H and I). ERF115 has two putative destruction (D)–box sequences (amino acids 115 to 118 and 150 to 153) that are recognized by the APC/C (fig. S4A). Inactivation of the proximal D-box stabilized ERF115, whereas its stability was increased by mutation of the second D-box (Fig. 1J and fig. S4B).

In agreement with ERF115 being a proteasome target, within translation reporter lines, ERF115-GFP fluorescence could only be detected upon MG132 treatment, revealing a QC cell–specific accumulation pattern (fig. S5). Correspondingly, ERF115 promoter activity was observed in the QC cells (Fig. 2A), albeit only in 11.7% of the examined roots (n = 60 root tips). As observed previously (6), a modest temperature increase promoted QC cell division (31.0% at 24°C versus 15.0% at 21°C; n = 20 and 29 roots, respectively), coinciding with a temperature-dependent rise in pERF115:GUS-positive QC cells (Fig. 2C), of which 32.3%, corresponding to the QC cell division frequency at 24°C, showed signs of a recent cell division, as indicated by the presence of two adjacent blue cells (Fig. 2B). When grown with the cell cycle inhibitory drug hydroxyurea, plants had fewer pERF115:GUS-positive QC cells (Fig. 2C). Thus, ERF115 expression marks dividing QC cells.

Fig. 2 ERF115 expression correlates with QC cell division and is regulated by brassinosteroids.

(A) pERF115:GUS activity in QC cells of a 1-week-old seedling. Scale bar, 20 μm. (B) pERF115:GUS-positive cells marking a divided QC cell (arrow). (C) Quantification of pERF115:GUS-positive QC cells in control, hydroxyurea-treated (0.75 mM HU), and brassinolide-treated (0.5 nM BL) 1-week-old seedlings, grown at 21°C and 24°C. Red represents the percentage of plants with unstained QC cells, whereas green, blue, and yellow indicate the percentage of root tips with either one, two, or more positive QC cells (n > 57 root tips). (D) Root tip organization of 1-week-old wild-type (Col-0) and ERF115SRDX seedlings control-treated or treated with 0.5 nM brassinolide (BL+). QC cells are visualized by the WOX5-GFP marker. Cells were counterstained with propidium iodide. Scale bar, 50 μm.

Ethylene plays a putative role in QC cell division (5) and regulates some members of the ERF gene family. However, the frequency of pERF115:GUS-positive QC cells did not vary upon treatment with the ethylene precursor 1-aminocyclopropane-1-carboxylic acid (ACC) or ethylene itself, nor upon treatment with the ethylene inhibitor silver nitrate (fig. S6), suggesting that ERF115 is not involved in ethylene perception or signaling. Brassinosteroids also promote QC cell division (12). Correspondingly, ERF115 expression appeared to depend on brassinosteroids, because treatment with brassinolide increased the number of pERF115:GUS-positive QC cells (Fig. 2C and fig. S7) and reached up to 86.6% (n = 82 root tips) at 24°C. Because of the link between ERF115 expression and QC cell division, we investigated whether the brassinosteroid-dependent QC cell proliferation phenotype was ERF115 dependent. QC cells of erf115KO lines still divided in response to brassinosteroid treatment, perhaps due to gene redundancy in the 122-member ERF gene family. To circumvent this problem, we converted ERF115 into a dominant negative form by fusing it with the SUPERMAN repression domain (SRDX) (13). Wild-type plants treated with brassinolide displayed disorganized root meristems due to hyperproliferation of the QC. In contrast, ERF115SRDX plants treated with brassinolide showed reduced QC hyperproliferation (Fig. 2D and fig. S8). Thus, the brassinosteroid-induced QC divisions depend at least in part on the ERF115 activity. Contrasting, in line with the absence of transcriptional control of ERF115 by ethylene, no effect on ethylene-dependent QC cell division was observed in the ERF115SRDX plants (fig. S9), indicating that brassinosteroids and ethylene control QC cell division through different pathways.

To identify the mechanism by which ERF115 drives QC cell division, we screened for genes controlled by ERF115 through transcriptome analysis on ERF115OE plants and tandem chromatin-affinity purification analysis (Fig. 3A). Compared with the wild type, 259 genes were activated in ERF115OE root tips (table S1), of which 20 were bound by the ERF115 transcription factor (tables S2 and S3). In this group, we found one of the five PHYTOSULFOKINE PRECURSOR–encoding genes of Arabidopsis thaliana, which give rise to a sulfonated pentapeptide hormone molecule, known to control root growth, cell proliferation, cell elongation, and callus formation (1417). PSK5 was more than 8-fold transcriptionally up-regulated and was bound by ERF115 in its promoter region (Fig. 3B and fig. S10). PSK5 induction in the ERF115OE line was confirmed through quantitative reverse transcription polymerase chain reaction analysis (Fig. 3C). In the dominant negative ERF115SRDX roots, PSK5 expression was more than 60% repressed, supporting a role for ERF115 as transcriptional activator of this particular PSK gene (Fig. 3C). Increased PSK5 transcript levels were also observed in the ccs52a2-1 mutant, in which the ERF115 protein is stabilized, validating the PSK5 transcriptional activation by ERF115 (Fig. 3D). Similar to the ERF115 gene, upon brassinolide treatment, PSK5 transcription is activated in a manner that depends on the brassinosteroid-insensitive 1 receptor (Fig. 3E), in agreement with the recent observation that PSK signaling is brassinosteroid dependent (18). As PSK peptides are predominantly sensed by the leucine-rich repeat receptor kinase PSKR1 (19, 20), we introduced the ERF115OE construct into pskr1-3 plants. Absence of a functional PSKR1 receptor resulted in a normal QC cell division phenotype (Fig. 3F and fig. S11), demonstrating that the ERF115-induced QC cell divisions depend on PSK signaling.

Fig. 3 ERF115 controls QC cell division through PSK5 signaling.

(A) Venn diagram depicting the overlap between the 259 genes up-regulated in the ERF115OE root tip (yellow) and the 608 genes bound by ERF115 (red). This overlap is higher than expected by chance (P < 0.05; hypergeometric distribution). (B) Graphical representation of the reads resulting from tandem chromatin-affinity purification sequencing mapping to the PSK5 gene. Gray and black bars represent reads mapped to the forward and reverse strand, respectively. (C) Relative PSK5 expression levels in 5-day-old wild-type (Col-0), ERF115OE, and ERF115SRDX mutant root tips. Expression levels of the wild type were arbitrarily set to one. Data represent mean ± SE (n = 3 independent RNA extractions). (D) Relative PSK5 expression levels in 5-day-old wild-type (Col-0) and ccs52a2-1 mutant roots. Expression levels of the wild type were arbitrarily set to one. Data represent mean ± SE (n = 3 independent RNA extractions). (E) Relative ERF115 (black) and PSK5 (gray) expression levels in 1-week-old wild-type (Ws-2) and brassinosteroid-insensitive 1 receptor (bri1-5) mutant roots control-treated (BL-) or treated with 0.5 nM brassinolide (BL+). Expression levels of the wild type were arbitrarily set to one. Data represent mean ± SE (n = 3 independent RNA extractions). (F) Confocal microscopy image of 1-week-old wild-type (Col-0), ERF115OE, and ERF115OE pskr1-3 mutant root meristems. Scale bar, 50 μm.

Our data suggest that the APC/CCCS52A2 and brassinosteroid pathways regulate the QC cell division through their antagonistic effect on the ERF115 abundance, a mechanism that might account for the low proliferation rate of the QC cells, in which APC/CCCS52A2 dampens the activity of brassinosteroid-induced ERF115 activity through its proteolytic turnover (fig. S12). Next to brassinosteroids, QC cell division is observed upon treatment with other stress signaling hormones (5, 21). Reminiscent to the law of Bergonié and Tribondeau stating that the reproductive activity of cells is proportional to their sensitivity toward genotoxic signals (22), we postulated that the low division rate of QC cells would mark them as less vulnerable to stress in comparison with other stem cells and that the QC cells might represent a reservoir of cells that are called upon to replenish damaged stem cells. As verified under our conditions, treatment of Arabidopsis roots for 24 hours with the radiomimetic drug bleomycin triggered programmed cell death of stem cells neighboring the QC, as visualized by the uptake of the cell death marker propidium iodide (23) (Fig. 4, A and B). In contrast, cell death was suppressed in the QC cells (Fig. 4B and fig. S13), which might in part be attributed to the significant enriched expression of DNA damage repair genes (P < 0.001) (table S4). Upon retransfer of the bleomycin-treated plants to drug-free medium, the QC expression domain increased through cell division (Fig. 4C), displacing dead cells from the stem cell niche within 2 to 4 days (Fig. 4, D and E). In plants hemizygous for ERF115SRDX, the WOX5 expression domain expanded as well but through ectopic acquirement of QC cell identity rather than cell division (Fig. 4, F and G), in agreement with the observation that ERF115 activity is required for QC cell division. As a consequence, the original QC cells lost their cell identity, whereas the newly formed QC cells remained in direct contact with dead cells, even after transfer to drug-free medium (Fig. 4, H to J). This phenotype was more notable in homozygous ERF115SRDX plants that lost meristem organization completely within 2 days on recovery medium (fig. S14). Root-growth measurement illustrated that clearance of all dead stem cells in wild-type plants allowed roots to resume growth after release from the stress, whereas in ERF115SRDX growth was inhibited (fig. S15). These data are reminiscent of pioneering work in maize showing that root apex regeneration upon x-ray radiation occurs through proliferation of the QC cells (24) and indicate that the ERF115-PSK signaling pathway not only may account for the low QC cell proliferation rate but also contributes to ensuring the longevity of the stem cell niche. Coexistence of quiescent and actively dividing adjoining stem cells is not plant-specific, but rather appears to be a general principle, found as well within hair follicle, gut, and bone marrow stem cell niches (25). Therefore, maintaining a stem cell subpopulation that is used to replace damaged stem cells might represent a general mechanism to maintain a functional stem cell niche under stress conditions.

Fig. 4 ERF115 maintains the QC identity domain upon damage of the stem cell niche.

(A to E) Wild-type (Col-0) root tips before (A) and after treatment with 0.6 μg/mL bleomycin for 24 hours (B) and upon recovery on bleomycin-free medium for 24 hours (C), 48 hours (D), and 72 hours (E). (F to J) ERF115SRDX root tips before (F) and after treatment with 0.6 μg/mL bleomycin for 24 hours (G) and upon recovery on bleomycin-free medium for 24 hours (H), 48 hours (I), and 72 hours (J). QC cells are visualized by the WOX5-GFP marker. Cells were counterstained with propidium iodide. Scale bars, 50 μm.

Supplementary Materials

Materials and Methods

Figs. S1 to S15

Tables S1 to S5

References (2638)

References and Notes

  1. Acknowledgments: The authors thank M. Sauter for sharing the pskr1-3 mutant and A. Bleys and M. De Cock for help in preparing the manuscript. This work was supported by Ghent University (Multidisciplinary Research Partnership Bioinformatics: From nucleotides to networks); theInteruniversity Attraction Poles Programme (IUAP P7/29 MARS), initiated by the Belgian Science Policy Office; Ghent University-Bijzonder Onderzoeksfonds (to D.V.D.S.); and Research Foundation-Flanders (grants G.029809 and G.022.10N to D.V.D.S. and L.D.V., respectively). K.S.H. is indebted to the Agency for Innovation by Science and Technology for a predoctoral fellowship. T.C., F.V., and J.V.L. are Postdoctoral Fellows of the Research Foundation-Flanders. Microarray and sequencing data have been deposited in the Gene Expression Omnibus (GEO) database, under accession nos. GSE48836 and GSE48793, respectively. J.H. and L.D.V. are inventors on a patent application filed by Ghent University and Vlaams Instituut voor Biotechnologie that covers the use of ERF115 to modulate plant growth through phytosulfokine gene expression.
View Abstract

Navigate This Article