ABA promotes quiescence of the quiescent centre and suppresses stem cell differentiation in the Arabidopsis primary root meristem

Authors

  • Hanma Zhang,

    Corresponding author
    1. Centre for Plant Sciences, Institute of Integrative and Comparative Biology, Faculty of Biological Sciences, University of Leeds, Leeds LS2 9JT, UK
    Search for more papers by this author
  • Woong Han,

    1. Centre for Plant Sciences, Institute of Integrative and Comparative Biology, Faculty of Biological Sciences, University of Leeds, Leeds LS2 9JT, UK
    2. School of Biotechnology, Kangwon National University, Hyo-Ja 3 Dong, Chuncheon, Gangwon 200-701, South Korea
    Search for more papers by this author
  • Ive De Smet,

    1. Department of Cell Biology, Max Planck Institute for Developmental Biology, D-72076 Tübingen, Germany
    2. Center for Plant Molecular Biology, University of Tübingen, D-72076 Tübingen, Germany
    Search for more papers by this author
    • These authors contributed equally to this work.

    • Present address: Department of Plant Systems Biology, VIB, Ghent University, Technologiepark 927, 9052 Gent, Belgium.

  • Peter Talboys,

    1. Centre for Plant Sciences, Institute of Integrative and Comparative Biology, Faculty of Biological Sciences, University of Leeds, Leeds LS2 9JT, UK
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Rakesh Loya,

    1. Centre for Plant Sciences, Institute of Integrative and Comparative Biology, Faculty of Biological Sciences, University of Leeds, Leeds LS2 9JT, UK
    Search for more papers by this author
  • Amaar Hassan,

    1. Centre for Plant Sciences, Institute of Integrative and Comparative Biology, Faculty of Biological Sciences, University of Leeds, Leeds LS2 9JT, UK
    Search for more papers by this author
  • Honglin Rong,

    1. Centre for Plant Sciences, Institute of Integrative and Comparative Biology, Faculty of Biological Sciences, University of Leeds, Leeds LS2 9JT, UK
    Search for more papers by this author
  • Gerd Jürgens,

    1. Department of Cell Biology, Max Planck Institute for Developmental Biology, D-72076 Tübingen, Germany
    2. Center for Plant Molecular Biology, University of Tübingen, D-72076 Tübingen, Germany
    Search for more papers by this author
  • J. Paul Knox,

    1. Centre for Plant Sciences, Institute of Integrative and Comparative Biology, Faculty of Biological Sciences, University of Leeds, Leeds LS2 9JT, UK
    Search for more papers by this author
  • Myeong-Hyeon Wang

    1. School of Biotechnology, Kangwon National University, Hyo-Ja 3 Dong, Chuncheon, Gangwon 200-701, South Korea
    Search for more papers by this author

For correspondence (fax +44 113 34 33144; email bgyhz@leeds.ac.uk).

Summary

It is well known that abscisic acid (ABA) can halt meristems for long periods without loss of meristem function, and can also promote root growth at low concentrations, but the mechanisms underlying such regulation are largely unknown. Here we show that ABA promotes stem cell maintenance in Arabidopsis root meristems by both promoting the quiescence of the quiescent centre (QC) and suppressing the differentiation of stem cells and their daughters. We demonstrate that these two mechanisms of regulation by ABA involve distinct pathways, and identify components in each pathway. Our findings demonstrate a cellular mechanism for a positive role for ABA in promoting root meristem maintenance and root growth in Arabidopsis.

Introduction

The phytohormone abscisic acid (ABA) plays many important regulatory roles in plant development and in plant adaptation to both biotic and abiotic stresses. Detailed information on its signalling mechanisms has been uncovered at an unprecedented rate over recent years. For example, several recent independent studies in Arabidopsis have led to the identification of an ABA signalling complex that can mediate a complete signalling pathway from the hormonal signal to regulated gene expression. This complex consists of four components: pyrabactin resistance 1(PYR)/ pyrabactin resistance 1-like proteins (PYL)/regulatory component of ABA receptor 1 (RCAR) receptors, phosphotases 2C (PP2Cs), SNF1-related kinases (SnRK2s) and ABA-responsive transcriptional factors (ABFs/AREBs) (Park et al., 2009, Umezawa et al., 2009; Ma et al., 2009; Fujii et al., 2009; Klingler et al., 2010). The signalling process starts when ABA binds to the PYR/PYL/RCAR receptors. This binding disrupts the physical interaction between PP2Cs and their target subfamily 2 SNF1-related kinases (SnRK2s), leaving the latter open to phosphorylation, thereby activating the ABF/AREB transcription factors and leading to the expression of ABA-responsive genes (Klingler et al., 2010). In addition to these components, a large number of other ABA signalling molecules have also been identified (Klingler et al., 2010).

Despite the progresses, there are still significant gaps in our knowledge of the molecular and cellular processes associated with some ABA functions. For example, it is well known that ABA has the ability to maintain meristems in a dormant state without loss of meristem function (Addicott and Carns, 1983; De Smet et al., 2006). When a viable seed of a plant is placed on medium containing high concentrations of ABA under germination-permissive conditions, the shoot apical meristem (SAM) and root apical meristems (RAM) of the embryo inside the seed remain dormant for a long period after the embryo has broken through the seed coat (Lopez-Molina et al., 2001). The functional properties of the meristems are preserved over the dormancy period, and the meristems can resume functioning as soon as ABA is removed. ABA also promotes meristem dormancy during vegetative growth, such as in axillary buds (Shimizu-Sato and Mori, 2001) and newly formed lateral root meristems (De Smet et al., 2003). Functional preservation of meristems is a key aspect of meristem dormancy, but little is known as to how ABA acts in this process.

Another ABA function whose mechanism is not understood is the widely observed positive role of ABA in plant growth (Finkelstein et al., 2002). Although ABA is generally regarded as a growth inhibitor (Zeevaart and Creelman, 1988; Finkelstein et al., 2002), there is strong evidence that it is also required to promote growth under certain (non-stressed) conditions, as indicated by the retarded growth phenotype of many ABA-deficient plants (Ephritikhine et al., 1999; Cheng et al., 2002; LeNoble et al., 2004; Barrero et al., 2005) and the widely observed stimulatory effect on root growth of low levels of exogenously applied ABA (Vartanian et al., 1994; Ephritikhine et al., 1999; LeNoble et al., 2004; Barrero et al., 2005). Very little is known as to how ABA achieves these two opposing roles in plant growth, and, in particular, how it promotes growth.

As a part of our efforts to understand ABA’s roles in meristem dormancy and growth promotion, we investigated the role of ABA in Arabidopsis root meristems, as evidence is available indicating that ABA performs meristem dormancy and growth promotion in roots (De Smet et al., 2003; Cheng et al., 2002; Barrero et al., 2005) and that ABA rescues root meristem failure in a Medicago truncatula mutant (Liang et al., 2007). A good understanding of the functions of ABA in root meristems may provide valuable information on the mechanisms of its functions in meristem dormancy and root growth promotion. Arabidopsis root meristems have a simple and well-defined cellular organization. At the centre lies a small group of stem cells, known as the stem cell niche (Figure S1), which includes the quiescent centre (QC), a group of four cells that divide infrequently, and the surrounding initials (van den Berg et al., 1997; Vernoux and Benfey, 2005; Scheres, 2007; Dolan, 2009). Cells derived from stem cells differentiate strictly according to their positions into various cells/tissues, resulting in well-defined and highly predictable cell/tissue patterns in a root. Maintenance of the stem cells is vital for the continuing function and development of a root. Morphological analyses have shown that this task involves at least two cellular processes: maintenance of QC quiescence and suppression of stem cell differentiation (Ortega-Martínez et al., 2007; Sarkar et al., 2007).

Our results show that ABA promotes stem cell maintenance in Arabidopsis root meristems by both promoting QC quiescence and suppressing stem cell differentiation. We have also identified some aspects of the regulatory mechanisms of these actions. Our findings indicate the cellular mechanism for ABA’s roles in meristem preservation during dormancy and in promoting root growth at low concentrations.

Results

ABA inhibits QC division in Arabidopsis primary root meristems

We observed that blocking ABA biosynthesis in Arabidopsis seedlings using fluridone, a widely used inhibitor of ABA biosynthesis in plants (Schwartz et al., 2003) and other organisms (Nagamune et al., 2008), induced division of QC cells in a significant proportion of roots (Figure 1a,b,d and Figure S2). This effect was dose-dependent (Figure 1n) and occurred in various ecotypes (Figure 4a). Using two QC-specific markers, WOX5p:GFP (Sarkar et al., 2007) (Figure 1d) and QC46 (Sabatini et al., 1999) (Figure 1g), we confirmed that the dividing cells were QC cells (Figure 1e,h and Figure S2). We also confirmed the fluridone-induced QC division on the basis of the positive correlation between QC division and the number of columella cell (CC) layers reported by Ortega-Martínez et al. (2007) (Figure S3). Because fluridone also blocks the biosynthesis of other carotene-derived products, some of which, such as gibberellins and strigolactone (Gomez-Roldan et al., 2008; Umehara et al., 2008), are hormones, we grew seedlings on medium containing both fluridone and ABA to establish whether the QC division was due to reduced ABA biosynthesis, and found that the presence of exogenous ABA over-rides the fluridone-induced effect and restored QC quiescence in a dose-dependent manner (Figure 1o), demonstrating that fluridone-induced QC division is directly linked to ABA deficiency. Consistent with this conclusion, a high level of QC divisions was observed in the ABA-deficient mutants aba1-1, aba2-1, aba2-4 and aba3-2 (Koornneef et al., 1982; Laby et al., 2000) (QC division frequencies of 54, 73, 23 and 27%, respectively; Figure 1j–m) and ABA-insensitive mutants (see Figure 4) on control medium with no fluridone, but no QC division was observed in wild-type roots on the same medium. Based on these observations, we conclude that ABA promotes QC quiescence in Arabidopsis primary root meristems. The effect of ABA or fluridone on QC division occurred within 24 h of treatment (Figure 1p,q and Figure S4), indicating that regulation of QC quiescence by ABA is a rapid response.

Figure 1.

 Fluridone induces and ABA suppresses QC division in Arabidopsis root meristems.
(a–f) Confocal images showing the effects of fluridone and ABA on QC divisions in WOX5p:GFP roots. Note that (a) and (d), (b) and (e), and (c) and (f) are pairs of images from the same root.
(g–i) Light microscopic images showing the effects of fluridone and ABA on QC division in QC46 roots.
(j–m) Confocal images showing QC division in four ABA-deficient mutants grown on control medium.
(n) Dose dependency of the effect of fluridone on QC division frequency in Col roots (= 60).
(o) ABA over-rides the effect of fluridone on QC division in Col roots (= 60).
(p, q) Time course of the effects of fluridone and ABA on QC division. Wild-type (Col) seedlings (= 60) were grown on control (p) or fluridone-containing (q) medium for 7 days, and then transferred to fluridone- (p) or fluridone- and ABA-containing (q) medium, respectively.
The age of the seedlings in (a–o)was 10 days. Error bars in (n–q) are standard deviations. QC, quiescence centre; CSC, columella stem cell; control, base medium with no fluridone or ABA; Flu+, base medium plus 10 μm fluridone; ABA+, base medium plus 500 nm ABA. Scale bars = 10 μm.

Figure 4.

 The effect of ABA on QC division, but not that on stem cell differentiation in the distal part of the root meristem, involves several known ABA signalling proteins.
(a–f) Confocal images showing QC division in abi1-1, abi2-1, abi3-1 and abi5-1 roots, but not abi4-1 or era1-2 roots of 10-day-old seedlings grown on control medium.
(g) Effects of (10 μm) fluridone and (500 nm) ABA on QC division in five ABA-insensitive mutants (abi1-1, abi2-1, abi3-1, abi4-1 and abi5-1) and one ABA-oversensitive mutant (era1-2) and their wild-type lines (Ler, Col & Ws).
(h) Effects of (10 μm) fluridone and (500 nm) ABA on the differentiation of QC, CSC and C1 cells in abi1, abi2, abi3-1 and abi5-1 mutants and their wild-type lines (Ler and Ws).
Data for both (g) and (h) were collected from 10-day-old seedlings and three replicates were used (= 30 for each replicate). Error bars are standard deviations. Asterisks indicate a statistically significant difference (< 0.05) between mutants and their respective wild-type control on the same medium.

ABA suppresses the differentiation of stem cells and their daughters in Arabidopsis primary root meristems

In addition to the regulation of QC division, ABA also suppressed the differentiation of stem cells and their daughters in both the distal and proximal regions of the Arabidopsis primary root meristem. In a normal Arabidopsis root meristem, cells inside the stem cell niche remain undifferentiated. Daughter cells of these stem cells differentiate according to their positions. In the distal part of root meristems (DM), cells produced by the columella initials (also known as columella stem cells, CSCs) undergo terminal differentiation soon after division of CSCs to become columella cells (Figure 2a). In fluridone-treated roots, terminal differentiation (as indicated by the accumulation of starch granules) occurred in QC and CSC cells (Figures 1h and 2c–f and Figure S5). We confirmed that the fluridone-induced QC and CSC differentiation was due to ABA deficiency, as such differentiation did not occur in roots grown on medium containing both 10 μm fluridone and 500 nm ABA (data not shown). These results suggest a role for ABA in regulating stem cell differentiation in the DM. To quantify this ABA function, we monitored the differentiation frequencies of QC, CSC and C1 cells (the first layer of cells below CSC cells in the DM, see Figure 2a) in roots grown on control medium, and on medium containing 10 μm fluridone or 500 nm ABA, and found a negative correlation between the level of ABA and the differentiation of stem cells in the stem cell niche and their descendants in the DM (Figure 2b). To obtain further evidence for this ABA function in the DM, we used the 35S[GVG]:WOX5 line (Sarkar et al., 2007), which over-expresses WOX5 in the presence of dexamethasone (DEX), and, as a result of this WOX5 over-expression, produces extra layers of non-differentiated (stem) cells in the columella root cap (Figure 2l–n) (Sarkar et al., 2007). The presence of ABA increased the number of WOX5-induced stem cell layers (Figure 2n), and fluridone had the opposite effect (Figure 2m,o–q). These results support our conclusion that ABA suppresses stem cell differentiation in the distal part of Arabidopsis root meristems.

Figure 2.

 ABA suppresses cell differentiation in the stem cell niche and the distal region of an Arabidopsis root meristem.
(a) Scheme showing cellular layers in the distal region of an Arabidopsis root meristem.
(b) Fluridone (10 μm) induced differentiation of QC and CSC cells and ABA (500 nm) reduced differentiation of C1 cells (daughters of CSC) in 10-day-old wild-type (Col) roots.
(c–f) Images of 10-day-old QC46 roots on fluridone-containing medium showing various combinations of QC division and stem cell differentiation.
(g–j) Images of 10-day-old 35S[GVG]:WOX5 seedling roots with either no DEX induction (g) or 3 days of DEX induction (1 μm) (i, j) showing the formation of extra cell layers in the columella root cap.
(k–n) Light microscopic images of 10-day-old 35S[GVG]:WOX5 seedling roots showing WOX5-induced formation of extra stem cell layers and the effects of fluridone (10 μm) and ABA (500 nm) on this response.
(o–q) Quantitative data for the effects of fluridone and ABA on induction of stem cells by WOX5.
Scale bars = 10 μm. Error bars indicate standard deviations (= 20). Asterisks indicate a statistically significant difference (< 0.05) between the treatment and its relevant control.

We also investigated whether ABA suppressed cell differentiation in the proximal part of the primary root meristems (PM). In wild-type roots, cells derived from stem cells in the PM continue dividing for a few rounds and then differentiate, a process that is marked by rapid cell expansion in the longitudinal direction (Figure 3a). We used two morphological indicators, the number of cells in the division zone (DiZ) and the number of cells in the transition zone (TZ) (Figure 3), to monitor the rate of cell differentiation in the PM. The former reflects the duration of cells staying in division (the larger the number, the longer the cells remain in division), and the latter reflects the rate at which a cell enters differentiation after cessation of division. Together, these two indicators provide a good estimate of the cell differentiation rate in the PM. We found that the numbers of cells in both the DiZ and TZ were reduced after fluridone treatment and increased when exogenous ABA (50–500 nm) was applied (Figure 3b–d), indicating a negative correlation between ABA and the cell differentiation rate in the PM. To confirm that the increase in cell number in the DiZ of ABA-treated root meristems was not caused by increased cell divisions, we monitored cell division activities in the root meristems of control and ABA-treated seedlings using the CycB1;2p:GUS marker (DiDonato et al., 2004), and found that cell division activity was actually reduced in the ABA-treated root meristems as determined from the number of cells showing GUS activity staining (Figure 3e–g). This observation suggests that the increased cell number in the DiZ of ABA-treated roots is not due to an increased cell division rate, but is more likely due to a reduced cell differentiation rate. Based on these observations, we conclude that ABA also suppresses cell differentiation in the proximal region of Arabidopsis primary root meristems.

Figure 3.

 ABA suppresses cell differentiation in the proximal region of an Arabidopsis root meristem.
(a) Schematic representation showing the various zones along the epidermal cell file. The division zone (DiZ) is defined by cells of small size in the longitudinal dimension due to active division. The transition zone is defined by cells that are approximately double the length of the cells in the DiZ, but do not show rapid elongation (the increment between consecutive cells is <150%). The start of the differentiation zone (DfZ) is defined by cells that are rapidly expanding in the longitudinal direction (the increment between two consecutive cells equals or exceeds 150%).
(b) Representative light microscopic images of the root meristems of 7-day-old wild-type seedlings grown on medium containing 10 μm fluridone (Flu+), medium containing 50 nm ABA (ABA+) or control medium (base medium with no fluridone or ABA). The insets show a higher magnification of cells in similar positions of the root meristems under the three treatments to highlight the difference in cell size.
(c) Mean cell number in the DiZ. Sample sizes: Flu+, = 14; control, = 44; ABA+, = 42.
(d) Mean cell number in the TZ. Sample sizes: Flu+, = 14; control, = 32; ABA+, = 48.
(e, f) Representative GUS staining images of control and 500 nm ABA-treated CycB1;2:GUS roots. Blue spots indicate cells in G2 and M phases.
(g) Mean number of dividing cells per root meristem (= 150).
Error bars in (c, d, g) are standard deviations. Asterisks indicate a statistically significant difference (< 0.05) between treatment and control groups.

Taken together, we have shown that ABA promotes QC quiescence and suppresses cell differentiation in the stem cell niche and the distal and proximal regions of Arabidopsis root meristems. We also obtained evidence that ABA plays similar roles in lateral root meristems. For example, application of exogenous ABA increased the number of cells in the DiZ in lateral root meristems (Figure S6).

ABA regulates QC division independently of ethylene

We used a number of experimental approaches to investigate the possible mechanisms of such regulation in the Arabidopsis primary root meristem. First, we investigated whether ABA acted by modulating ethylene biosynthesis or signalling to promote QC quiescence. Ethylene is known to promote QC division (Ortega-Martínez et al., 2007). We used 2-aminoethoxyvinyl glycine (AVG) (1 μm) and AgNO3 (10 μm) to suppress ethylene biosynthesis and signalling, respectively, in fluridone-treated roots, and did not observe any significant impact of either inhibitor on QC division (Figure S7), indicating that ABA acts independently of ethylene in regulating QC division.

ABA’s regulation of QC division involves ABI1, ABI2, ABI3, ABI5, ERA1, but not ABI4

We then assessed whether regulation of QC quiescence and stem cell differentiation by ABA involved known ABA signalling molecules. We examined QC division in five ABA-insensitive mutants (abi1-1, abi2-1, abi3-1, abi4-1 and abi5-1) (Koornneef et al., 1984; Nambara et al., 1995; Finkelstein, 1994; Finkelstein et al., 2002) and one ABA-hypersensitive mutant (era1-2) (Pei et al., 1998) (Figure 4). Four of the insensitive mutants, abi1-1, abi2-1, abi3-1 and abi5-1, showed high QC division frequencies in root meristems on the control medium (Figure 4a–g), suggesting that ABI1, ABI2, ABI3 and ABI5 play a role in regulation of QC quiescence by ABA. However, these mutants still showed some responsiveness to both fluridone and ABA with respect to QC division (Figure 4g), which suggests that either this regulation involves other ABA signalling components or there is some functional redundancy between these regulators. In contrast, abi4-1 roots showed similar QC division frequencies to wild-type roots on the control, fluridone- and ABA-containing media (Figure 4d,g), suggesting that ABI4 is not involved in this ABA function. era1-2 roots showed a reduced QC division frequency in comparison with wild-type roots when grown on fluridone-containing medium (Figure 4f,g), which suggests that ERA1 plays a negative role in the regulation of QC quiescence by ABA.

We further investigated whether the ABI proteins also play a role in the regulation of stem cell differentiation by ABA, using fluridone-induced stem cell differentiation in the stem cell niche and the ABA-induced increase in cell number in the DiZ to monitor stem cell differentiation in abi1-1, abi2-1, abi3-1 and abi5-1 roots. No significant difference was found between these mutants and their respective wild-type lines for either response (Figure 4h and Figure S8), suggesting that either the ABI proteins do not play a major role in this ABA function or that there are functional redundancies between these proteins.

ABA’s regulation of stem cell differentiation requires the activities of WOX5 and MONOPTEROS

We also investigated whether ABA acted by regulating genes involved in root meristem functions. To this end, we assessed whether ABA regulates the transcription of genes known to be involved in root meristem function or auxin signalling, such as WOX5 (Sarkar et al., 2007), MONOPTEROS (MP) (De Smet et al., 2010; Hardtke and Berleth, 1998) or PLETHORA1–3 (PLT1–3) (Galinha et al., 2007), in the primary root meristem using real-time RT-PCR, and found that ABA induced the expression of WOX5, MP and PLT2, but not that of PLT1 and PLT3 (Figure 5a–e). These results prompted an examination of the possible role of these ABA-induced genes in the regulation of stem cell differentiation by ABA. We use the wox5-1 mutant, which lacks WOX5 function and has disorganized primary root meristems (Sarkar et al., 2007), to investigate the role of WOX5 in this regulation. The mutant roots showed 100% differentiation in C1 cells and CSC and 10% differentiation in QC (Figure 5n) on control medium. Application of either 10 μm fluridone or 500 nm ABA did not alter the differentiation frequencies of the C1, CSC and QC cell layers (Figure 5n), suggesting that the regulation of stem cell differentiation by ABA in the root meristem requires the function of WOX5. To investigate a possible role for MP in the regulation of stem cell differentiation by ABA, we used the 35S:MP line, which constitutively expresses MP (De Smet et al., 2010). 35:MP root meristems showed lower CSC and QC differentiation frequencies than the wild-type root meristems on fluridone-containing medium (Figure 5o), which suggests that MP may also play a role in the regulation of stem cell differentiation by ABA in Arabidopsis root meristems. We also examined the differentiation frequencies of QC, CSC and C1 cells in two homozygous T-DNA insertional plt2 mutants (SALK_089992C and SALK_128164C) grown on control, fluridone-containing (10 μm) and ABA-containing (500 nm) media, and did not find any significant differences between the mutants and wild-type roots in terms of their differentiation frequencies on the various media (data not shown).

Figure 5.

 Evidence for a role of WOX5 and MONOPTEROS in the regulation of stem cell differentiation by ABA in Arabidopsis root meristems.
(a–e) Quantitative RT-PCR analyses showing that ABA induces the expression of WOX5, MP and PLT2, but not that of PLT1 and PLT3. Open bars, no ABA; grey bars, 1 μm ABA.
(f–m) Confocal (f–i) and light microscopic (j–m) images showing that neither fluridone nor ABA treatment affects cell differentiation in the C1, CSC and QC cell layers in the root meristems of the WOX5 null mutant wox5-1.
(n) Differentiation frequencies of the QC, CSC and C1 cell layers in wox5-1 root meristems grown on control medium, medium containing 10 μm fluridone (Flu+) or medium containing 500 nm ABA (ABA+).
(o) Differentiation frequencies of the QC, CSC and C1 cell layers in root meristems of 10-day-old wild-type and 35S:MP seedlings grown on Flu-containing medium.
(p) Summary of the two regulatory roles of ABA in Arabidopsis root meristems and identified regulatory components.
Asterisks indicate a statistically significant difference (< 0.05) between ABA-treated and non-treated roots at the same time point in (a), (b) and (d), or between the Col wild-type and 35S:MP roots in (o).

Discussion

ABA promotes root growth by promoting QC quiescence and suppressing the differentiation of stem cells and their daughters in root meristems

Although it has been widely reported that ABA promotes root growth at low concentrations, the mechanism of this regulation is currently unknown. A recent report suggested that ABA may act in root meristem maintenance (Liang et al., 2007). Here we present experimental evidence showing that ABA promotes stem cell maintenance in root meristems. It acts by both promoting QC quiescence and suppressing the differentiation of stem cells and their daughters (Figure 5p). As stem cell maintenance is vital for the continuing function of root meristems and root growth, our observations suggest a cellular mechanism for the growth-promoting role of ABA in Arabidopsis roots. The levels of ABA required for the promotion of QC quiescence and the suppression of stem cell differentiation are consistent with previous observations that the growth-promoting effect of this regulator in roots occurs at nanomolar concentrations (Finkelstein et al., 2002).

Although our experimental strategy includes the use of fluridone as an inhibitor to block ABA biosynthesis, and this inhibitor also blocks the biosynthesis of other carotene-derived products, including hormones (Gomez-Roldan et al., 2008; Umehara et al., 2008), we obtained additional results demonstrating that the fluridone-induced QC division and stem cell differentiation was due to ABA deficiency, rather than deficiency of other carotene-derived products. For example, we showed that both fluridone-induced effects could be over-ridden by exogenous application of ABA, indicating that these effects are due to a reduced level of ABA. In addition, we also provide genetic evidence to support the role for ABA in regulating QC division, and show that ABA-deficient mutants (aba1-2, aba2-4 and aba3-2) (Figure 1j–m) and ABA-response mutants (abi1-1, abi2-1, abi3-1, abi4-1, abi5-1 and era1-2) (Figure 4) displayed enhanced QC division.

The role of ABA in promoting QC quiescence reflects a general regulatory property of this regulator in suppressing cell division

ABA is known to inhibit cell division in many plants and tissues (Newton, 1977; Robertson et al., 1990). Its role in promoting QC quiescence could be considered as a refection of this general regulatory property. We have shown that ABA also suppresses cell division in the proximal part of Arabidopsis root meristems. Interestingly, the inhibitory effect of ABA on cell division in the QC and the proximal part of Arabidopsis root meristems has opposite consequences on root growth. The reduced cell division in the proximal part of the root meristem represents negative regulation of root growth, and is believed to be the primary cause of the inhibitory effect of ABA on root growth in Arabidopsis (Wang et al., 1998). On the other hand, the suppression of QC division by ABA may be considered as positive regulation of root growth because it promotes stem cell (QC) maintenance. In this way, ABA achieves two opposite morphological effects in Arabidopsis root meristems by the same regulatory mechanism.

Our results suggest that the regulation of cell division by ABA involves the recently uncovered ABA signalling complex (Park et al., 2009, Umezawa et al., 2009; Ma et al., 2009; Fujii et al., 2009; Klingler et al., 2010). For example, among the mutants that show enhanced QC division, abi1-1 and abi2-1 are defective in two PP2Cs and abi5-1 is defective in a member of the ABF/AREB family of transcriptional factors. It will be important to establish the downstream signalling components and target genes involved in this regulation.

It has been proposed that ABA negatively regulates cell division by up-regulating genes that encode negative regulators of the cell cycle, such as the cyclin-dependent kinase inhibitor ICK1 (also known as KRP1) (Wang et al., 1998), which interacts with both Cdc2a (a cyclin-dependent kinase) and CycD3 (a D-type cyclin) and negatively regulates the transition from G1 to S, and down-regulating genes that encode positive regulators, such as chromatin licensing and DNA replication factor 1a (Castellano et al., 2004), topoisomerase I (Mudgil et al., 2002) or a telomerase reverse transcriptase catalytic subunit of Arabidopsis (AtTERT) (Yang et al., 2002), which are required for DNA replication. It will be interesting to establish whether those target genes are also involved in the regulation of QC division by ABA.

ABA and stem cell maintenance in Arabidopsis root meristems

Stem cell preservation is a key aspect of meristem dormancy. Although it is well established that ABA plays a crucial role in promoting meristem dormancy, it is not clear whether it plays a role in stem cell maintenance/preservation. Here we provide evidence for such a role for ABA in Arabidopsis root meristems. We first showed that fluridone induces stem cell differentiation, and that this induction can be over-ridden by exogenously supplied ABA, demonstrating a link between stem cell differentiation and ABA deficiency. We also showed that ABA suppressed and fluridone promoted the differentiation of stem cells in plants over-expressing WOX5. We also found that ABA up-regulated the expression of three genes involved in stem cell maintenance, WOX5, MP and PLT2. Furthermore, we provide evidence that lack of WOX5 function or constitutive expression of MP abolish the regulation of stem cell differentiation by ABA in Arabidopsis root meristems. Therefore, our results present a possible regulatory framework for the action of ABA in stem cell maintenance in Arabidopsis root meristems, i.e. by up-regulating the expression of WOX5 and MP and thereby promoting stem cell maintenance. More analyses are required to establish the details of this regulatory framework and its connections with identified ABA signalling components such as the PYR/PYL family of ABA receptors (Park et al., 2009; Umezawa et al., 2009), the PP2Cs and the SnRKs (Fujii et al., 2007). It is possible that the role of ABA in Arabidopsis root meristems reflects a general regulatory property of this regulator.

Given the critical role of ABI proteins, particularly ABI1 and ABI2, in ABA signalling, it is surprising that no significant alterations in stem cell differentiation were observed in the abi mutants. One possibility is that ABA may act in parallel with other regulators in suppressing stem cell differentiation in Arabidopsis root meristems, and therefore defects in ABA signalling alone do not lead to stem cell differentiation. Consistent with this hypothesis, we have obtained evidence that ethylene plays a similar role to ABA in suppressing stem cell differentiation in Arabidopsis root meristems, and that abi mutants showed increased stem cell differentiation in root meristems compared to their corresponding wild-type controls in the presence of an ethylene biosynthesis inhibitor (T. Chhun, R. Loya, P. Talboys, H. Zhang, unpublished results), suggesting that ABA and ethylene may act redundantly in regulating stem cell differentiation. Another possible explanation for the lack of defects in stem cell differentiation in abi mutants is functional redundancy. For example, there are many PP2C genes in Arabidopsis, of which at least five have been implicated in ABA signalling (ABI1, ABI2, HAB1, AHG1 and AHG3) (Klingler et al., 2010). Similarly, there are 14 PYR/PYL/RCAR receptor-encoding genes in Arabidopsis (Klingler et al., 2010). It is possible that the regulation of stem cell differentiation in root meristems involves specific member(s) of the PP2C or PYR/PYL/RCAR receptor families. However, the fact that abi1 and abi2 are dominant mutations argues against the second explanation.

In conclusion, we have uncovered a cellular mechanism to explain how ABA promotes root growth in Arabidopsis. This mechanism may also explain how ABA helps to preserve the function of root meristems during dormancy or under stress conditions, when maintaining the stem cell population in meristems is critical. The ability of ABA to suppress stem cell differentiation may be a key element of the meristem maintenance/protection process. Further research into the mechanisms of this regulatory role of ABA will not only help our understanding about how meristems are maintained and protected under stress conditions or during dormancy, but may also suggest means to engineer stress-tolerant crops in the future.

Experimental procedures

Plant materials and growth conditions

Mutants and transgenic lines used in the study were obtained from the following sources: abi1-1, abi2-1, abi3-1 and abi5-1 from Ruth Finkelstein (Department of Molecular, Cellular and Developmental Biology, University of California at Santa Barbara, CA), era1-2 from Peter McCourt (Department of Cell & Systems Biology, University of Toronto, Canada), aba1-1 from Maarten Koornneef (Max Planck Institute for Plant Breeding Research, Cologne, Germany), aba2-3 and aba2-4 from Susan Gibson (Department of Plant Biology, University of Minnesota), and abi4-1, wox5-1 and wox5-3 from the Nottingham Arabidopsis Stock Centre (NASC). WOX5p::ERGFP and QC46 were obtained from Ben Scheres and Renze Heidstra (Department of Biology, University of Utrecht, The Netherlands), 35S[GVG]:WOX5 from Keiji Nakajima (Graduate School of Biological Sciences, Nara Institute of Science and Technology, Japan) and CycB1;2:GUS from John Celenza (Department of Biology, Boston University, MA).

Arabidopsis seeds were surface-sterilized as previously described (De Smet et al., 2003). Except for the quantitative RT-PCR experiment (in which half-strength MS medium was used as the base medium), the media for all experiments had the following basic composition: 0.15 mm CaCl2, 0.1 mm NaH2PO4, 0.02 mm MgSO4, 0.4 mm KNO3, 1.8 μm KI, 20 μm H3BO3, 3.0 μm ZnSO4, 0.06 μm CuSO4, 0.4 μm Na2MoO4, 4.0 μm CoCl2, 0.04% v/v ferrous sulfate chelate solution (F-0518, Sigma, http://www.sigmaaldrich.com/), 0.5% w/v 2-(N-morpholino)ethanesulfonic acid (MES), 0.5% w/v sucrose and 10% w/v agar-agar (A-1080, Fisher Chemicals, http://www.fisher.co.uk/). The media were adjusted to pH 5.7–5.8 using KOH, and autoclaved at 121°C for 20 min. Fluridone (59756-60-4, Sigma), ABA, 2-aminoethoxyvinyl glycine (AVG, Sigma) and AgNO3 (MP Biomedicals Europe, http://www.mpbio.com/) were added from filter-sterilized stock solutions after autoclaving and cooling to 50°C. The stock solution of fluridone (20 mm) was prepared in ethanol, and that for ABA (10 μm) was prepared by first dissolving (±)-ABA (A1049, Sigma) in 0.1 mm NaOH, and then neutralizing the solution by adding an equal volume of 0.1 mm HCl. AVG (10 mm) and AgNO3 (1.0 m) stock solutions were prepared in distilled H2O. Seedlings were grown on the surface of the agar-solidified media in vertically placed 9 cm Petri dishes, which were partially sealed with Parafilm (Starlab UK Ltd, http://www.starlab.co.uk/) and kept in a growth chamber at 22°C with a 16 h/8 h light/dark cycle. For treatment with fluridone, ABA, AVG or AgNO3, seedlings were first germinated and grown on basic control medium for 3 days and then transferred to media containing the hormone or chemicals. For DEX induction, the 35S[GVG]:WOX5 seedlings were first germinated and grown for 7 days, and then transferred to media containing 1.0 μm dexamethasone (MP Biomedicals, 02199350.1), with or without 10 μm fluridone or 500 nm ABA, for a further 3 days before scanning laser confocal and light microcopy analyses.

Confocal imaging, light microscopy, starch staining and histochemical GUS activity staining

For confocal imaging, seedling roots were mounted in 10 μg/ml propidium iodide and examined and imaged using a confocal laser scanning microscope (LSM 510 META, Carl Zeiss, http://www.zeiss.com/). For light microscope examination and imaging, seedling roots were fixed in 1% w/v glutaraldehyde and 4% w/v paraformaldehyde dissolved in 50 mm phosphate buffer (pH 7.0), mounted in lactic acid (101384Q, BDH, VWR International Ltd, http://uk.vwr.com/app/Home) on glass slides, and photographed using a Zeiss Axioplan microscope equipped with a CoolSNAP camera (Photometrics, http://www.photomet.co.uk/) and the openlab 4.0.2 imaging software (Improvison Ltd, http://www.improvision.com/). The images were processed using either Photoshop 5 (http://www.adobe.com/) or Microsoft Photo Editor (http://www.microsoft.com/). Starch granules in root meristems were visualized by staining with 1:1 H2O-diluted Lugol solution (OB542827, Merck, http://www.merck-chemicals.com/), and β-glucuronidase (GUS) activity in QC46 roots was visualized by histochemical staining as previously described (De Smet et al., 2003).

Quantitative RT-PCR

Five-day-old seedlings were incubated for the indicated time periods in liquid medium in the presence of 1 μm ABA. RNA was extracted from the apical part of Col-0 seedling roots using an RNeasy plant mini kit (Qiagen, http://www.qiagen.com/), and total RNA samples were treated with DNase I (Fermentas, http://www.fermentas.com/) to remove any residual genomic DNA. First-strand cDNA was synthesized using a RevertAid first-strand cDNA synthesis kit (Fermentas) and oligo(dT) primers. Quantitative PCR was performed on a Chromo4 real-time detector (Bio-Rad, http://www.bio-rad.com/) using Platinum SYBR Green qPCR SuperMix-UDG (Invitrogen, http://www.invitrogen.com/), and the following protocol: 95°C for 10 min, 40 cycles of 95°C for 30 sec, 60°C for 30 sec and 72°C for 30 sec, and then 72°C for 7 min. Targets were quantified using specific primer pairs (Appendix S1). PCRs were performed in triplicate and on at least three biological replicates. Transcript levels were normalized to ACT2 transcript levels.

Acknowledgements

We thank R. Finkelstein (Department of Molecular, Cellular and Developmental Biology, University of California at Santa Barbara, CA) for the abi1-1, abi2-1, abi3-1 and abi5-1 mutants, P. McCourt (Department of Cell & Systems Biology, University of Toronto, Canada) for era1-2, M. Koornneef (Max Planck Institute for Plant Breeding Research, Cologne, Germany) for aba1-1, S. Gibson (Department of Plant Biology, University of Minnesota) for aba2-3 and aba2-4, B. Scheres and R. Heidstra (Department of Biology, University of Utrecht, The Netherlands) for WOX5p:ERGFP and QC46, K. Nakajima (Graduate School of Biological Sciences, Nara Institute of Science and Technology, Japan) for 35S[GVG]:WOX5, J. Celenza (Department of Biology, Boston University, MA) for CycB1;2:GUS and the Nottingham Arabidopsis Stock Centre (NASC) for abi4-1, wox5-1, wox5-3, SALK_089992C and SALK_128164C lines. Financial support was provided by the Royal Society (2006/R1 to H.Z. and M.H.W.) and the UK Biotechnology & Biological Sciences Research Council (research grant 24/16557 to H.Z and a studentship to P.T.). We also thank A. Cumming, B. Davies and D. Pilbeam for valuable comments and editing. I.D.S. was supported by the European Molecular Biology Organization (ALTF 108-2006), and the Marie Curie Intra-European Fellowship scheme (FP6 MEIF-CT-2007-041375).

Ancillary