An endoplasmic reticulum response pathway mediates programmed cell death of root tip induced by water stress in Arabidopsis


  • Yunfeng Duan,

    1. The State Key Laboratory of Plant Cell & Chromosome Engineering, Center of Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, 286 Huaizhong Road, Shijiazhuang, Hebei 050021, China
    Search for more papers by this author
    • The authors contributed equally to this work.

  • Wensheng Zhang,

    1. The State Key Laboratory of Plant Cell & Chromosome Engineering, Center of Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, 286 Huaizhong Road, Shijiazhuang, Hebei 050021, China
    Search for more papers by this author
    • The authors contributed equally to this work.

  • Bao Li,

    1. The State Key Laboratory of Plant Cell & Chromosome Engineering, Center of Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, 286 Huaizhong Road, Shijiazhuang, Hebei 050021, China
    Search for more papers by this author
    • The authors contributed equally to this work.

  • Youning Wang,

    1. The State Key Laboratory of Plant Cell & Chromosome Engineering, Center of Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, 286 Huaizhong Road, Shijiazhuang, Hebei 050021, China
    2. Graduate University, Chinese Academy of Sciences, Beijing 100039, China
    Search for more papers by this author
  • Kexue Li,

    1. The State Key Laboratory of Plant Cell & Chromosome Engineering, Center of Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, 286 Huaizhong Road, Shijiazhuang, Hebei 050021, China
    Search for more papers by this author
  •   Sodmergen,

    1. College of Life Sciences, Peking University, Beijing 100871, China
    Search for more papers by this author
  • Chunyu Han,

    1. The State Key Laboratory of Plant Cell & Chromosome Engineering, Center of Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, 286 Huaizhong Road, Shijiazhuang, Hebei 050021, China
    Search for more papers by this author
  • Yizhang Zhang,

    1. The State Key Laboratory of Plant Cell & Chromosome Engineering, Center of Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, 286 Huaizhong Road, Shijiazhuang, Hebei 050021, China
    Search for more papers by this author
  • Xia Li

    1. The State Key Laboratory of Plant Cell & Chromosome Engineering, Center of Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, 286 Huaizhong Road, Shijiazhuang, Hebei 050021, China
    Search for more papers by this author

Author for correspondence:
Xia Li
Tel: +86 0311 85871744


  • Drought induces root death in plants; however, the nature and characteristics of root cell death and its underlying mechanisms are poorly understood. Here, we provide a systematic analysis of cell death in the primary root tips in Arabidopsis during water stress.
  • Root tip cell death occurs when high water deficit is reached. The dying cells were first detected in the apical meristem of the primary roots and underwent active programmed cell death (PCD). Transmission electron microscopic analysis shows that the cells undergoing induced death had unambiguous morphological features of autophagic cell death, including an increase in vacuole size, degradation of organelles, and collapse of the tonoplast and the plasma membrane. The results suggest that autophagic PCD occurs as a response to severe water deficit.
  • Significant accumulation of reactive oxygen species (ROS) was detected in the stressed root tips. Expression of BAX inhibitor-1 (AtBI1) was increased in response to water stress, and atbi1-1 displayed accelerated cell death, indicating that AtBI1 and the endoplasmic reticulum (ER) stress response pathway both modulate water stress-induced PCD.
  • These findings form the basis for further investigations into the mechanisms underlying the PCD and its role in developmental plasticity of root system architecture and subsequent adaptation to water stress.


Programmed cell death (PCD) is an active process of cellular suicide essential for development and immune responses in eukaryotes. In plants, PCD plays critical roles in plant development and survival. It is involved in leaf senescence, development of xylem tracheary elements, and regulation of lace plant leaf shape during development (Lam et al., 2001; Gunawardena et al., 2004; Rogers, 2005). Localized PCD, also termed hypersensitive response (HR), is known to be a mechanism mediating plant responses during pathogen attack (Lam et al., 2001; Lam, 2004; Greenberg & Yao, 2004). Recent researches have shown that diverse environmental stresses, such as salt stress, drought and nutrient starvation, are able to induce PCD in plant root tips (Huh et al., 2002; Liu et al., 2009). The results indicate that this active program is highly conserved and is involved in development and response to external stimuli.

Programmed cell death has been classified into three categories based on the morphological features during PCD: apoptotic, autophagic and nonlysosomal PCD. Apoptosis results in breakdown of cell fragments by another cell, and autophagic PCD occurs within the dying cells Nonlysosomal PCD means that the cells kill themselves by inhibiting some major biosynthetic pathways, by destabilizing the membranes or in other unknown ways (Van Doorn, 2005; Van Doorn & Woltering, 2005; Hatsugai et al., 2006). Current evidence shows that, in plants, both developmental PCD and stress-induced PCD, such as that caused by nutrient deficiency, are forms of autophagic PCD (Patel et al., 2006). The mechanisms underlying this active suicide strategy remain largely unknown. However, the current evidence indicates that this strategy may play a crucial role in both the execution of morphogenetic decisions based on perceived information and the subsequent adaptation to heterogeneous environments.

Plants, because of their sessile nature, have evolved varying strategies to adapt to and tolerate unfavorable conditions. The most important strategy used by plants is to have greater plasticity in postembryonic growth and development. It is assumed that PCD plays an important role in plant developmental plasticity and subsequent adaptation to various adverse environmental stimuli. However, our understanding of plant PCD and its underlying mechanisms is still in its early stages. The most extensively studied type of plant PCD is HR-PCD. The timely activation of HR-PCD is considered to be essential for containment of a pathogen near the infection site (Liu et al., 2005; Patel et al., 2006). In dealing with abiotic stress it has been reported that PCD in plant root apical meristem cells is induced by salt stress (Huh et al., 2002). Other examples of PCD have been found in the formation of constitutive aerenchyma in roots of bulltongue arrowhead (Sagittaria lancifolia) (Schussler & Longstreth, 2000) and in the formation of inducible aerenchyma in maize (Zea mays) roots (Gunawardena et al., 2001). Programmed cell death has also been implicated in root tip death in maize and pea (Pisum sativum) exposed to flooding stress (Subbaiah & Sachs, 2003). Recently, several important genes mediated plant cell PCD have been identified, and endoplasmic reticulum (ER) stress-mediated programmed cell death has been found to be involved in cell death progression induced by both biotic and abiotic stresses (Watanabe & Lam, 2006, 2008; Costa et al., 2008). For example, BAX inhibitor-1 (BI-1) located in the ER was identified as a key attenuator of cell death in eukaryotes. In plants BI-1 mRNA level is increased during leaf senescence and under several types of environmental stimuli, such as salt, heat, etc.). Loss of function in AtBI1 results in hypersensitivity of the mutants to the stresses, whereas overexpression of the gene suppressed the plant’s cell death induced by abiotic and biotic stresses (Watanabe & Lam, 2006). Recent genetic and pharmacological results also showed a pivotal role of AtBI1 during ER stress and ER stress-mediated cell death (Watanabe & Lam, 2008). AtZIP28 and AtZIP17, the ER stress-induced leucine zipper transcription factor genes anchored to the ER membrane, have also been found to be key players in stress responses. They regulate the target genes of unfolded protein response (UPR) in ER and subsequent ER homeostasis and the proper folding and maturation of secretory proteins (Liu et al., 2007a,b; Costa et al., 2008).

One of the least-known forms of PCD is that which occurs in response to drought. Drought is the most severe adverse environmental factor limiting plant productivity in both natural and agricultural systems (Verslues et al., 2006). Whenever and wherever drought occurs, plant water deficit develops. Because the plant’s roots regulate water intake, soil water status can alter plant root depth and root length density in a pronounced way. Water stress not only affects primary root growth, but also alters lateral root initiation and elongation (Van der Weele et al., 2000; Deak & Malamy, 2005). Importantly, extensive studies of the effects of drought on root mortality have shown that severe soil water deficit and/or prolonged drought induce premature root death. Drought-induced root death is considered quite common among plant species, as it has been found in various crops and forest trees (Deans, 1979; Taylor, 1983; Goss & Watson, 2003). Despite the fact that drought-induced root death is thought to be very important for plant response to water deficit, its physiological roles and the mechanisms of the process are poorly understood. In particular, we wanted to research whether or not the drought-induced root cell death is PCD, exactly which cells die, and the morphological features of cells undergoing PCD. Such an analysis would provide an important foundation for further research on mechanisms regulating the PCD in root tips induced by water deficit, and their roles in the developmental plasticity of root system architecture.

In this study, we chose Arabidopsis as a model for the study of cell death induced by water stress that is simulated by polyethylene glycerol (PEG-8000; Amresco, Solon, OH, USA). We report the identification and description of the PCD that occurred in root tips under water stress. We further examined the ultracellular features of the PCD using transmission electron microscopy analysis. We also analysed the expression of the genes in ER response signaling and the responses of atbi1 mutants to water stress. We demonstrated that AtBI1 is a critical mediator of water stress-induced cell death, and the ER stress signaling is involved this process.

Materials and Methods

Plant materials, growth conditions and root growth measurements

Arabidopsis thaliana (L.) ecotype Columbia (Col-0) ecotype was used as the wild-type. Atbi1 mutant in Col-0 background was kindly provided by Dr Eric Lam (Rutgers University, NJ, USA). Seeds were surface sterilized and placed onto agar plates containing MS salt (Sigma), 2.5% sucrose (w : v) and 0.8% agar (Sigma) (w : v), and the pH was adjusted to 5.7. The plates were placed at 4°C in dark for 2 d to synchronize germination, and then transferred to a growth room (22°C; 14-h photoperiod).

The PEG-infused plates were prepared as described by Van der Weele et al. (2000) with some modifications. The full-strength MS salt and 2.5% sucrose (w : v) were added to the media. The PEG solutions were made by dissolving solid PEG of molecular weight 8000 (PEG-8000; Amerasco) in liquid basal media containing full-strength MS salt (Sigma), 2.5% sucrose with 2 mM 2-(N-morpholino) ethanesulfonic acid (MES) buffer (pH 5.7). A 1% solution was defined as 1 g of PEG per 100 ml final volume. After autoclaving, the PEG solution was then poured onto agar-solidified basal media (3 : 2, v : v). The plates were then kept in a fume hood for 48 h to allow the PEG to infiltrate into agar media at room temperature. The excess PEG solution was removed and water potentials (see the Supporting Information, Table S1, Fig. S1) of the PEG-infused plates were measured using a Dewpoint Meter (Model WP4; DECAGON Devices Inc., Pullman, WA, USA).

The lengths of roots were measured from root tips to hypocotyl–root junction sites 10 d after stress treatments. Each data point was obtained by analysing at least 20 seedlings within a single experiment. We performed two independent experiments to confirm our data.

Detection of cell death with Trypan Blue and propidium iodide (PI) staining

The red fluorescent dye PI and Trypan Blue staining have been widely used as indicators of dead cells. Application of Trypan Blue and PI staining was performed as described by previous reports (Leite et al., 1999; Ning et al., 2002). In brief, Trypan Blue (Bio Basic Inc., Markham, Ontario, Canada) or PI (Sigma) was dissolved in distilled water at a final concentration of 10 mg ml−1 or 1 μg ml−1, respectively. The seedlings were then immersed in the solutions and incubated for 15 min or 1 min, respectively, at room temperature and then washed with PBS several times. The staining was then observed by light microscopy for the former and under fluorescence microscope (Axioskop; Carl Zeiss) for the latter. Only whole root tip stained a dark color in Trypan Blue and/or PI staining analyses were counted as a death event when death rates of root tips were scored.

Transmission electron microscopy

Tissue samples from the stressed root tips at different stages were fixed as described by Gunawardena et al. (2005). Root tips were collected and kept in 2% glutaraldehyde in 0.05 M sodium cacodylate buffer, pH 6.9, overnight under vacuum (137.9 KPa). After washing with the same buffer, the samples were postfixed in 2.5% aqueous osmium tetroxide for 4 h at room temperature, and then were dehydrated in a graded ethanol.

Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay

To detect cell apoptosis caused by PEG stress, an in situ apoptosis detection kit (Takara Bio Inc., Shiga, Japan) was used. The TUNEL assay was performed according to the manufacturer’s instructions with a few modifications. In brief, the whole seedlings were fixed in 4% paraformaldehyde/PBS (pH 7.4) solution at 4°C overnight. After being washed with 1× PBS solution three times, the samples were immersed into 70% ethanol for at least 24 h at −20°C. After the samples were washed three times with PBS, they were subjected to a permeabilization buffer for 15 min on ice and followed by a PBS wash. Then the samples were transferred into 50 μl Reaction Buffer (TdT Enzyme 5 μl + Labeling Safe Buffer 45 μl) for a 90-min incubation at 37°C. The labeling procedure was stopped by washing with a PBS solution. The image was acquired with a Zeiss LSM 510 Confocal laser scanning microscope with a 488 nm excitation line and a 530 nm emission filter.

4,6-Diamidino-2-phenylindole (DAPI) staining

For DAPI staining, the seedlings of the wild type and/or the mutant atbi1-1 were immerged in 1 μg ml−1 (w : v) DAPI (Sigma) in 1× PBS buffer with 1% (v : v) Triton X-100 for 10 min at room temperature, followed by washing several times in PBS. The DAPI stain image was viewed microscopically with UV filter.

Assay of reactive oxygen species (ROS)

Histochemical detection of ROS in the root tips was first performed using 3,3-diaminobenzidine (DAB; Sigma) staining. Six-day-old Col-0 seedlings were exposed to water stress (40% PEG-8000) for 0, 3, 12 and 24 h. Roots were incubated in 1 mg ml−1 solution of DAB (pH 5.5, 1× PBS buffer) for 80 min at room temperature. The seedlings were washed three times in PBS buffer and then immersed in boiling ethanol (70% v : v) for 10 min. The deep brown polymerization product in the root tips were photographed under a microscope (Axioskop; Carl Zeiss). Experiments were repeated at least three times and similar results obtained.

Water stress-induced ROS accumulation in root tips was then imaged using DCFH-DA and confocal microscopy. Six-day-old Col-0 seedlings were exposed to water stress (40% PEG-8000) for 0, 3, 12 and 24 h. Roots were incubated in 10 μM DCFH-DA (Beyotime, Jiangsu, China) in 1× PBS buffer for 30 min at 37°C, and washed three times in PBS buffer, and then viewed microscopically (Zeiss LSM 510, excitation 488 nm, emission 530-nm). Experiments were repeated at least three times and similar results obtained.

DNA isolation and electrophoresis

To detect apoptotic DNA cleavage, a DNA fragmentation assay was performed. The 100 mg root tip samples (< 1 cm in length to root tip) grown under certain conditions and harvested at certain time points were frozen in liquid nitrogen immediately after sampling and ground to a fine powder. Isolation of DNA was performed using the Apop-Ladder EXTM DNA fragmentation assay kit (TaKaRa) according to the manufacturer’s instructions. To visualize DNA fragmentation, 3 μg of genomic DNA of each sample was run with a standard 100 bp DNA ladder on a 2% agarose (w : v) gel containing ethidium bromide at a constant voltage of 50 V for 2 h.

Quantitative real-time reverse-transcription polymerase chain reaction (RT-PCR)

Total RNA was isolated from root tissues of 6-d-old seedlings using RNAVzol (Vigorous, Beijing, CHN) according to the manufacturer’s instructions. The PCR amplification was performed on an ABI7000 using a SYBR PrimeScript RT-PCR Kit (TaKaRa). Each reaction was performed in a total reaction system of 10 μl according to the manufacturer’s protocol. To verify the amplification specificity, a melting curve analysis of the PCR products was conducted and the expression levels of genes tested were standardized to the constitutive expression level of GAPC which encodes cytosolic GADPH (C subunit, At3g04120). In each experiment, there are at least three replicates. The relative expression level of each gene, corresponding to the GAPC level, was calculated using the method described by Livak & Schmittgen (2001). The specific primers for each gene are shown in Table S2.


Root growth and the mortality rate of Arabidopsis young seedlings are sensitive to water stress

To understand how plant roots sense and response to drought signals during root development, we performed systematic studies using Arabidopsis wild-type Col-0. Water stress was simulated by treating an agar medium with various concentrations of PEG-8000. Six day-old young seedlings of Col-0 germinated on MS medium were transferred on to the PEG-infused media with different levels of water stress, and the root growth was evaluated 10 d after treatment. As shown in (Fig. 1a,b), root elongation of primary roots was very sensitive to water stress. The root elongation was inhibited by water deficit. For convenience we will use the percentage of PEG8000 used instead of the water potential (Table S1, Fig. S1) when describing levels of water stress. When exposed to 10% PEG primary root length was reduced by 62.2% during a 10-d period. At 20% and 30% water potentials the root growth inhibition was 23.5% and 5.7%, respectively, compared with the control. No root growth was observed at 40% or higher concentrations of PEG. The root grew slowly under mild stress conditions during a prolonged period, whereas root growth was completely inhibited under severe stress (Fig. 1a,b).

Figure 1.

 Low water potentials inhibit growth of the primary root and induce root tip death in Arabidopsis. Six-day-old wild type Col-0 seedlings were grown for 10 d on MS control media or MS media infused with different concentrations of polyethylene glycerol (PEG-8000). (a) Growth of one representative seedling on PEG-infused plates. Bars, 2.5 mm. (b) Root growth repression under low water potentials was quantified by measuring the length of the primary roots. (c) The time-course of root tip death induced by severe water stress (40% (open bars) and 50% (closed bars) PEG-infused agar plate, respectively). The root tips which showed complete and solid blue in Trypan Blue staining analysis were considered. The data are means (±SD) from at least 20 seedlings.

Cessation of growth of the primary roots under high levels of water deficit did not appear to be a temporary growth-inhibitory effect, because the roots did not regrow even when the stress conditions were removed. To test the possibility that the specific growth cessation might be caused by root death, we performed a Trypan Blue staining analysis to check membrane integrity of treated and untreated plants. Trypan Blue can only penetrate the membranes of dead cells. The staining analysis showed that at 40% PEG, 18% of tested root tips were dead during a 1-h treatment period, and the death rate of root tips of the primary roots reached 100% at 24 h after treatment (Fig. 1c). When treated with 50% PEG for 1 h death rate was 62%, and increased to 100% of the primary root tips at 12 h of treatment. By contrast, no blue staining was detected in the root tips of the control plants and the stressed seedlings under mild water stress (PEG ≤ 30%, water potential < −1.4 MPa) within a 7-d period (data not shown).

Progression of water stress-induced cell death during root development

In order to identify the pattern and progression of water stress-induced cell death in root tips, we stained the roots treatment with 40% PEG using PI, a nucleic acid stain that intercalates into double-stranded nucleic acids. It is excluded by viable cells but can penetrate cell membranes of dying or dead cells (Ning et al., 2002). As shown in (Fig. 2b), untreated control plants exhibited either a sporadic pattern of stained cells, which were spread over the entire root, or no stained cells at all. However, following treatments of Col-0 seedlings with severe water stress, the distinct pattern of the stained cells appeared at the root apical meristematic cells in the upper part of the meristem c. 1 h after treatment (Fig. 2b). The rest of root remained unstained. The stained area was then enlarged, and the number of PI-positive cells and staining intensity was significantly increased after prolonged treatment. At 12 h nearly the entire meristem of the stressed root tip was strongly stained. At this time root cap cells seemed to be unaffected, and stained cells in the elongation zone of the roots were still sporadic. After 24 h, the whole root tip, including meristematic cells and root cap cells, were all strongly stained. Eventually, the collumela cells began to be stained, but the cells in the elongation zone and the rest of root still remained nearly unstained (not shown). Staining analysis using Trypan Blue showed similar pattern of cell death progression (Fig. 2c). The results indicate that water stress-induced cell death proceeds rapidly and primarily occurs in elongation zone and meristematic cells in the root tip.

Figure 2.

 Progression of cell death of Arabidopsis root during polyethylene glycerol (PEG)-infused water stress. Six-day-old seedlings were treated with agar plates which are infused with 40% PEG-8000 and subjected to propidium iodide (PI) (b) or Trypan-blue (c) staining at 0, 1, 12, 24 and 60 h after treatment, respectively. (a) Arabidopsis root tip under microscope at specified time-point. Bars, 100 μm.

Water stress-induced root cell death is programmed cell death

To further determine whether the root cell death induced by severe water stress is a kind of PCD, we investigated the alterations in nucleus morphology, nucleus fragmentation and DNA cleavage, all of which are usually used as diagnostic markers for programmed cell death (Huh et al., 2002; Gunawardena et al., 2005). To visualize cleavage of nuclear DNA and nuclear morphology we performed a TUNEL assay on whole-mount tissue samples. There was a very weak TUNEL signal detected in the untreated young seedling root tips and nearly no TUNEL-positive nuclei were observed (Fig. 3a). In sharp contrast there was a gradual increase in the number of TUNEL-positive nuclei from 0 to 24 h after exposure to severe water stress in the meristem of the stressed roots (data not shown), and gradually spreading to the whole meristem of the root at c. 12 h. At this time, TUNEL-positive cells in other parts of the root, such as elongation zone, were sporadic and faint. As a control, DAPI staining of the nonstressed and stressed roots was also performed. Strong fluorescence intensity, uneven staining and inhomogeneous intensity pattern of the meristematic cells in the stressed root tips were detected, indicating occurrence of chromatin condensation at the time (Fig. 3a).

Figure 3.

 Water stress induces chromatin condensation and DNA fragmentation in primary roots of the stress Arabidopsis seedlings. (a) 4,6-Diamidino-2-phenylindole (DAPI) and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining in cells of the primary roots of 6-d-old seedlings after growth in MS medium or 50% polyethylene glycerol (PEG)-infused agar plates for 0, 12 and 24 h. Insets: close up observation of chromatin condensation. Bars, 100 μm. (b) Detection of nucleosomal DNA in a 2% agarose gel. DNA was extracted from root tissues of wild-type Col-0 seedlings treated with 50% PEG-infused agar plates for the time indicated. M, molecular marker (DL 2000).

To verify the nature of the dying cells, the DNA ladder profile was also analysed in plants having PCD and the control plants. Genomic DNA was isolated and separated by agarose gel electrophoresis (Fig. 3b). DNA smearing was detected as early as 12 h after treatment, and extensive DNA smearing was observed after prolonged water stress treatment. However, no clear ‘laddering’ of DNA degradation of the genomic DNA into internucleosomal fragments of multiples of c. 180 bp were detected. Collectively, these data suggest that the primary root tip cell death in response to water stress is an active suicide process that might be an adaptive mechanism by which root systems are modulated. The reason that DNA laddering was not detected in the water-stressed root tips is possibly because the cells undergoing PCD comprise only very small fraction of the tissues collected for gel electrophoresis, which has been shown previously (Gunawardena et al., 2003).

Cell ultrastructure during the stress-inducing PCD

Cells from the untreated root apical meristem have few small vacuoles with large, rounded nuclei (Fig. 4a,b,c,m). The nuclear chromatin is evenly distributed throughout the nucleus. The organelles, such as mitochondria, endoplasmic reticulum and Golgi apparatus are normal and clearly located in the cytoplasm and the plasma membrane is intact. However, the cells in the first stage showed many abnormalities. Many small vesicles or vacuoles formed between the plasma membrane and nucleus. As the PCD progressed, the cells appeared more advanced in all respects with more and larger vacuoles, degrading organelles and early signs of chromatin condensation (Fig. 4d–f).

Figure 4.

 Transmission electron micrographs of Arabidopsis meristematic cells from the root tip during water stress-induced cell death. (a–c) Meristematic cells from nonstressed root tip, showing normal cells with tonoplast, nuclei, plasma membrane, Golgi and mitochondria. (d–f) Stage I of cell death. Many small vesicles appear within the vacuolar space, and chromatin condenses irregularly (arrows). (g–i) Stage II dying cells showing disrupted cellular ultrastructures and appearance of large vacuoles (arrows). (j–l) Stage III tissue and cells experiencing death. The various organelles disappear though the nucleus is still intact, and the cytoplasm shrinks (arrows). (m–o) Nonstressed control root tip (M) and root tips experiencing cell death at middle (n) and late (o) stages. Bars: (a,d,g,j,m–o), 10 μm; (b,c,e,f,h,i,k,l), 2 μm.

At the middle stage, disrupted cellular ultrastructure was clearly observed. The nuclei of the cells were distorted and contained large clumps of chromatin. Small vesicles or vacuoles were collapsed, and several larger vacuoles had formed. Higher magnification showed that most membrane vesicles possessed double or multiple membrane boundaries. These morphological features clearly reflect classic autophagic characteristics. Therefore, the observed double/multiple-membrane vesicles were autophagosomes and the larger vacuoles were the autophagic vacuoles. In addition, degraded cytoplasm was less electron-dense, suggesting that degradation of its contents had begun. However, the plasma membrane still was appressed to the cell wall (Fig. 4g–i,n).

The dying cells at late stage showed a complete breakdown of cellular structures. The various organelles disappeared, although some of the nucleus was still intact, although somewhat deformed, with very condensed chromatin. The cytoplasm was extremely light, as if degraded. In some cells vacuoles occupied the major cellular space, and the cell had essentially no intact cytoplasmic contents except for several compact dense bodies. The plasma membrane had retracted from the cell wall (Fig. 4j–l). At the final stage degradation of the nucleus was also observed but the cell wall of dead cells was not dramatically affected (Fig. 4j,o).

Water stress-induced lateral and adventitious roots are tolerant to water stress

As the dying cells are clustered in the meristem of the stressed root tip, which is responsible for making new cells and planning root system architecture, we sought to elucidate the link between cell death and developmental plasticity of root system, and to assess whether such an active suicide program in the root tip might be an adaptive response to severe water stress. Thus, the influence of PCD on the lateral roots and adventitious roots and the subsequent tolerance of these roots to water stress were examined.

When the growth of the primary roots completely ceased and root tip death occurred under severe water deficit (40% PEG), new root initiation was not observed during the first 2–3 wk. After this time, lateral and adventitious root primordia started to occur on the tap roots, mostly in the upper part of the tap root (Fig. 5a–d). The new roots elongated very slowly, but these water stress-induced roots, regardless of lateral and adventitious roots, were much thicker and tuberized (Fig. 5). The average widths of the water stress-induced roots were approximately twice that of the normal lateral roots (data not shown). Under prolonged water stress, the short and tuberized roots remained alive and the number of lateral and adventitious roots per tap root substantially increased (Fig. 5c,d). The new lateral roots were formed at the lower part of the tap root. However, these newly formed roots, induced by severe water deficit, exhibited some variability in initiation sites and number of lateral and/or adventitious roots per tap root. As a control, we also tested the stress tolerance of lateral and adventitious roots induced by removal of the growth tip of the primary roots. Growth tips of 6-d-old seedlings grown on MS medium were cut off to stimulate lateral and adventitious roots, and the seedlings were then subjected to severe water stress (40% PEG). The lateral and adventitious roots induced by decapitation of the primary roots were as sensitive as the primary root to water stress and the root tip death occurred immediately (data not shown).

Figure 5.

 Root system development is regulated by water status. Six-day-old Arabidopsis Col-0 seedlings were transferred to 40% polyethylene glycerol (PEG)-infused plates. Growth of the primary roots ceased. Initiation (a) and elongation (b–d) of lateral roots and adventitious roots (d) were observed during prolonged severe water stress. The short and tuberized lateral roots resumed growth upon rehydration (e,f). Arrows indicate lateral root primordia (a), lateral roots, and adventitious roots (b–d). Thick arrows in (e,f) show new lateral roots after rehydration. Bar, 3 mm.

Importantly, these water stressed-induced roots were able to reinitiate elongation within 24–48 h after the seedlings were transferred to a nonstressed medium. The elongated lateral and adventitious roots emerging after rehydration were morphologically similar to those roots occurring under normal conditions (Fig. 5e,f). The fact that stress-induced lateral roots and adventitious roots had enhanced tolerance to water deficit suggests that promotion of lateral and adventitious roots mediated by primary root tip death under severe water deficit is an adaptive mechanism.

Water stress induces ROS accumulation in root tips

Reactive oxygen species, such as hydrogen peroxide (H2O2), superoxide ion, and nitric oxide (NO) are well-recognized triggers of cell death (Jabs, 1999; Van Breusegem & Dat, 2006). Accumulation of ROS can be detected in the plants under drought stress (Elstner & Osswald, 1994; Foyer & Noctor, 2005). To examine whether ROS, such as H2O2, could trigger PCD induced by water stress, formation of ROS in plant root cells were measured by staining analysis using DAB and the fluorescent probe DCFH-DA (Beyotime). Under normal conditions, there was a very little amount of ROS in the root tips of Arabidopsis seedlings (Fig. 6a,b). However, significant generation of ROS in the root of Col-0 was detected at 12 h after exposure to water stress. The ROS mainly formed in the meristem and elongation zone of the stressed roots where the cell death occurred. The results suggest that ROS may be involved in activation of PCD in plant root tips during water stress.

Figure 6.

 Induction of reactive oxygen species (ROS) and expression of the genes in the endoplasmic reticulum (ER) stress response pathway. The ROS accumulation in root tips after exposure to water stress (40% polyethylene glycerol (PEG 8000)) was detected by 3,3′-diaminobenzidine (DAB) (a) and 2,7-dichlorofluorescin diacetate (DCFH-DA) (b) staining. Relative expression levels of the genes localized in the ER (c) were detected in the stressed root tips of the 6-d-old Col-0 seedlings at the given time-points. GapC was used as a reference gene in the analysis. Bars, 100 μm.

AtBI1 Expression is induced by water stress

AtBI1 is a critical survival factor of stress tolerance in plants. To determine whether water stress-induced cell death is also mediated by disruption of ER homeostasis and ER response, we analysed the expression of AtBI1 transcript under water stress. As shown in (Fig. 6c), rapid increase of AtBI1 mRNA was observed at 1 h after exposure to water stress. A similar gene expression pattern was also detected for AtBip2, which has been used as an ER marker gene for UPR activation in eukaryotes (Koizumi et al., 2001; Martinez & Chrispeels, 2003). AtBip2 belongs to the HSP70 family, and is required for the transport and secretion of proteins in the ER. Strikingly, strong and rapid induction in the transcript level of a PR1 gene that encodes a pathogenesis-related protein lying in downstream of ER signaling pathway was also observed after treatment, which is consistent with the previous results (Jelitto-Van Dooren et al., 1999; Durrant & Dong, 2004; Watanabe & Lam, 2008). The expression of AtBI1, BiP2 and PR1 genes reached the highest levels at 3 h after water stress treatment, and there was an approx. 90-fold increase in PR1 gene expression at the time (Fig. 6c). The results suggest that water stress activates the UPR via accumulation of unfolded and/or misfolded protein and induces PCD.

AtBI1 mutant exhibits enhanced sensitivity to water stress

To further investigate the role of AtBI1 for plant tolerance to water stress, we analysed the responses of 6-d-old Arabidopsis seedlings of the wild type Col-0 and a loss of function mutation in AtBI1 mutant (atbi1-1) to water stress. The atbi1-1 plants displayed significant increased sensitivity in both root elongation and shoot growth to water stress. Strikingly, loss of function in AtBI1 resulted in substantially inhibited lateral root development under water stress, indicating that ATBI1 is required for lateral root development in response to water stress (Fig. 7a,b). We then examined whether loss of function in AtBI1 causes accelerated PCD in root tip PCD under water stress. As shown in (Fig. 7c), the death rate of atbi1-1 root tips was significantly higher that the wild type. At 8 h after treatment with 40% PEG, the death rate of the mutant roots was c. 80%, whereas, the percentage of died roots in the wild type was < 40% (Fig. 7c). The results indicate that the cessation of root growth of atbi1-1 is caused by enhanced cell death progression rather than growth arrest. By using DAPI and TUNEL staining analyses, we found that root tip cell death of atbi1-1 mutant under water deficit also underwent PCD, with typical nuclear chromatin condensation and TUNEL-positive staining in the root meristem, and the chromatin condensation status of the mutant root tips at 8 h after treatment was similar to that of the wild type at 12 h after stress occurred in the mutant roots (Fig. 7d). The results demonstrate that AtBI1 modulates activation of PCD and is an important determinant for survival of plant cell under water stress.

Figure 7.

AtBI1 mutant seedlings are hypersensitive to water stress. Six-day-old Arabidopsis Col-0 and atbi1-1 seedlings were exposed to various levels of water stress. (a,b) atbi1-1 mutants were sensitive to water stress (Col-0, open bars; atbi1-1, closed bars). (c) Root tip mortality of the mutant seedlings was more sensitive to water stress on the medium containing 40% polyethylene glycerol (PEG). The data are means (±SD) from at least 20 seedlings (Col-0, open bars; atbi1-1, closed bars). (d) The root tip cell death in atbi1-1 showed the characteristics of programmed cell death (PCD), and occurred earlier. The picture shows 4,6-diamidino-2-phenylindole (DAPI) and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining analyses of the root tips under normal and 8 h after water stress. Insets: close up images of chromatin patterns. Bar, 100 μm.


Severe water stress is the primary cause of root death in nature. This phenomenon has been reported in various plant species, although there are large variations in root mortality among species (Deans, 1979; Chiatante et al., 1999;Huang & Gao, 2000; Goss & Watson, 2003). It is assumed that severe water stress-induced root cell death might be an adaptive response. However, how exactly root death occurs under severe water stress, its biological role and the underlying mechanisms remain largely unknown. Here we report the first detailed description of the existence, pattern and morphological characteristics of PCD induced by water stress, which might represent an important mechanism of plant drought tolerance at the organ level.

Root growth and mortality is sensitive to water availability

Because of the complexity of plant drought responses, one obstacle to studying plant drought avoidance and tolerance is the difficulty of setting up a system in which there are reproducible differing levels of low water potential. To understand root death induced by severe water stress, we have modified previous procedures (Van der Weele et al., 2000; Verslues et al., 2006) and were able to successfully produce severe water stress (water potential as low as −4.95 MPa) (Fig. S1). Our experimental protocol for creating low water potential can largely mimic plant or root response to soil drying. Root elongation is greatly reduced upon exposure to low water potentials (Fig. 1a,b) and severe stress induces the formation of short and tuberized roots (Fig. 6), which is consistent with results from Arabidopsis plants treated with severe soil drying reported by Vartanian et al. (1983, 1994).

Our sensitivity analysis showed that mild and moderate water stress retarded root elongation (Fig. 1). Once the water deficit in the plate had reached a threshold (30% PEG), root growth of Col-0 seedling was completely inhibited; but the roots were still alive. Growth inhibition by water stress may occur through the regulation of cell cycle progression leading to the arrest of cell division of meristematic cells in root apical meristem (Sacks et al., 1997). When subjected to more severe water stress (40% PEG and above), root tip death of the primary root occurred over short periods of time in individuals, and we observed increases in root mortality among the treated seedlings owing to continued exposure (Fig. 1c). Root mortality in response to severe water deficit was conditional and dependent upon the growth potential of the seeds and young seedlings and the severity of water deficit. Our study also demonstrated that root mortality varied with ecotypes (data not shown). Variation in root mortality among ecotypes indicates that water stress-induced root death is genetically controlled.

The physiological mechanisms controlling root mortality under water stress are not yet understood. Some studies propose that water stress disrupts cell membrane integrity resulting in severe leakage of organic solutes from root cells, which could be related directly to the root death of tall fescue cultivars during soil drying (Huang & Gao, 2000). Several researchers working with woody plants have suggested that carbohydrate depletion is the primary cause of root death during drought stress (Marshall, 1986; Kosola & Eissenstat, 1994). The fact that carbon allocation to surface fine roots of citrus was reduced 80% when roots were exposed to localized soil drying (Kosola & Eissenstat, 1994) supports this hypothesis. However, some other studies found that carbohydrate depletion is not closely related to root mortality in the loblolly pine (Pinus taeda) seedlings under drought stress. Thus, the physiological causes of root death in response to water deficit are still controversial. Many questions remain to be answered, such as: Do the physiological factors influencing root mortality under water stress vary among species? and What are the key physiological factors inducing root death in Arabidopsis?

Progression and the nature of water stress-induced root cell death

The most extensively studied form of plant cell death is the HR to infection (Lam, 2004). Hypersensitive response cell death is localized to the site of infection during pathogen attack and occurs rapidly within 12 h (Goodman & Novacky, 1994). Recent studies have demonstrated that the localized cell death involves host-mediated PCD, which plays a critical role in determining both immunity and disease progression in plants (Greenberg & Yao, 2004; Lam, 2004).

In this study, we present evidence that indicates that the water stress-induced cell death in Arabidopsis roots is also localized and undergoes PCD. Cell death was monitored using the fluorescent exclusion dye PI (Fig. 2), which is widely used to detect dying cells and labels the nucleus in dying cells lacking an intact plasma membrane. Upon exposure to severe water stress, root cell death processes occur extremely early and continue under prolonged water stress. The greatest cell death always occurs in the basal region of the elongation zone and the meristematic cells of the Arabidopsis root tip (Figs 2, 3). It is obvious that water stress-induced cell death is largely localized within cells that have the capacity to divide. It is likely that cells are more sensitive to the stress stimuli when they are dividing. In the root apical meristem, the central cells divide much more slowly than the surrounding cells. This might explain the progression pattern of the cell death induced by severe water deficit. Because the root apical meristem produces only the primary root, cell death in root apical meristem results in cessation of growth in the primary roots of the stressed plants.

Autophagic PCD in root primary tip is induced by water stress

By conducting the ultrastructural study using transmission electron microscopy, we showed that water stress-induced cell death appears to be autophagic cell death. Based on the major ultrastructural characteristics, we have divided the process into four stages. In stage I, numerous vesicles appear within the vacuolar space, indicating alteration of the tonoplast permeability. Chromatin condenses irregularly (Fig. 4d–f). In stage II, disrupted cellular ultrastructures appear (Fig. 4g–i). Large vacuoles form with double membrane boundaries, suggesting that water stress-induced PCD is autophagic. In stage III, the various organelles disappear, although the nucleus is still intact, and the cytoplasm shrinks (Fig. 4j–l). In the last stage, the nucleus breaks down and the collapsed primary cell wall is left behind without degradation (Fig. 4j,o).

Autophagic cell death is a separate pathway of PCD distinctly different from apoptosis. The differing mechanism for degradation of dying cells is the clearest distinction between autophagic cell death and apoptosis. During autophagic cell death, DNA fragmentation occurs and autophagic vacuoles form that are used for the destruction of cytoplasm (Lee & Baehrecke, 2001). Compelling evidence indicates that autophagic PCD reflects a high degree of flexibility in a cell’s response to changes in environmental conditions (Beers & McDowell, 2001; Kuriyama & Fukuda, 2002; Patel et al., 2006). Recently, Liu et al. (2009) reported that autophagy is required for salt and osmotic tolerance, the autophagy-related gene AtATG18a is induced by abiotic stresses and loss of function in AtATG18a resulted in increased sensitivity of the mutants to salt and osmotic stresses. Further functional analysis of autophagy-related genes in response to water stress will provide insights into the role of autophagy in PCD induced by water stress in plant root tip.

Water stress-induced PCD of root tip modifies root system architecture and enhances drought tolerance

As root traits are closely related to drought tolerance of plants (Lynch, 1995; Huang & Gao, 2000; Deak & Malamy, 2005), it is important to determine whether the water stress-induced PCD of primary root tips might provide a protective mechanism by modifying the root architecture system in response to severe water deficit, and how this dedicated active process increases drought tolerance and enables plants to survive a period of water stress.

When the Arabidopsis root is seriously stressed, the meristem cells respond to severe water deficit by activating a suicide mechanism, while the mature root cells remain alive (Fig. 3). During a prolonged period of water stress, the lateral roots emerge from the primary roots above the elongation zone, although there are some variations in the site of emergence and the number of lateral roots among the seedlings (Fig. 5). Interestingly, these newly formed lateral roots are short and show a tuberized shape. These highly differentiated lateral roots with altered root anatomy are also observed when soil-grown Arabidopsis seedlings are subjected to progressive water stress (Vartanian et al., 1983; Marshall, 1986; Vartanian et al., 1994). The threshold at which water deficit induces short roots and the position of lateral root initiation varies in the different Brassicaceae species (Vartanian et al., 1994) and ecotypes of Arabidopsis (data not shown), indicating that the water stress-induced root system modification is genetically controlled. Further studies are needed to examine whether water stress affects the initiation or outgrowth of the lateral roots and the underlying molecular mechanisms.

Could the altered root system architecture confer enhanced tolerance to severe water stress or could such drought rhizogenesis in Arabidopsis directly determine their capability to withstand a prolonged drought period? Our results show that water stress-induced short roots remain alive although they grow extremely slowly under prolonged exposure to severe stress. However, the lateral roots formed under normal conditions or induced by excision of primary roots are as sensitive as the primary roots in response to the same level of drought stress (data not shown). The findings indicate that the meristem cells of these short roots induced by water stress are more tolerant to extreme drought. Another important feature is that these short roots, upon rehydration, rapidly recover their elongation, forming a new normal branched root system (Fig. 5d,e). These findings are in agreement with the observations in the soil stress treatment (Vartanian et al., 1994). Thus, development of the short lateral roots during drought is an important adaptive strategy, which is essential for plant survival and post-stress recovery.

AtBI1-modulated ER stress mediates root cell death induced by water stress

In eukaryotic cells, the ER is responsible for the synthesis, post-translational modification and delivery of proteins to their appropriate target sites within the secretory pathway and the extracellular space. Disruption of any of these processes in the ER triggers ER stress and activates programmed cell death. The ER has been found to be very sensitive to various environmental stresses. Recently, it has shown that BI-1 genes from mammalian and plants function as critical survival factors to protect cells from ER stress-induced programmed cell death (Watanabe & Lam, 2008). For example, loss of function AtBI-1 Arabidopsis lines (atbi1-1 and atbi1-2) exhibited accelerated cell death during heat shock, and pathogen infection, and chemical induced ER stress, whereas overexpression of AtBI-1 attenuates ER stress-mediated programmed cell death (Watanabe & Lam, 2008). Here we provide evidence that AtBI1 also functions as a cell death attenuator for water stress-induced PCD in root tips of Arabidopsis. Water stress may activate UPR in ER and induce AtBI1-mediated PCD. Loss of function in AtBI1 results in hypersensitivity to water stress (Fig. 7). In addition, expression of AtBI1 is induced by water deficient (Fig. 6c). This genetic and molecular evidence supports the hypothesis that AtBI1 may suppress ER stress-induced PCD to allow the cell to re-establish the ER homeostasis. Previously, it has been showed that high salinity can also induce the unfolded protein response and subsequently activate programmed cell death in the root tips of the stressed Arabidopsis young seedlings (Huh et al., 2002). It appears that the ER stress-mediated plant PCD is a highly conserved adaptive mechanism to allow eukaryotic cells to survive the changing environment conditions.

We also showed here that water stress markedly induces ROS accumulation in the root tip of the Arabidopsis wild-type seedlings (Fig. 6a,b). Oxidative stress caused by high concentrations of ROS induced by various adverse stimuli can provoke destructive protein modification, protein–DNA crosslinks and many cellular functions (Beckman & Ames, 1997; Berlett & Stadtman, 1997; Montillet et al., 2005). Under water stress, significantly high amounts of H2O2 in root meristematic cells may disrupt normal ER functions resulting in ER stress that leads to cell death. Given that there are spatial and temporal correlations between ROS generation and PCD occurrence in Arabidopsis roots under severe water stress, we proposed that ROS, for example, H2O2, play key roles in triggering a PCD program. The ROS level in atbi1-1 root cells in response to water stress was significantly higher than that in wild-type root cells (data not shown), indicating that AtBI1 mediates ER stress-mediated ROS production in response to water stress. Our results provide insights into ROS-mediated cell death of root tips induced by severe water stress, however, further studies are needed in order to better understand how water stress induces various ROS and cellular redox changes, and how ROS signals are perceived and transduced to trigger PCD in root tips.

In summary, we have provided detailed information about water stress-induced PCD in the primary root tips of Arabidopsis and have presented a description,, of what happens at the ultrastructural level during plant adaptation to water stress. Our data suggests that autophagic PCD in plant root tip is induced by water stress and that this highly controlled and ordered mechanism is essential for the developmental plasticity of the root system in response to drought. It might also allow plants to confront water deficit more efficiently, thereafter conferring enhanced drought tolerance. Modification of root system architecture by autophagic programmed death for root tip cells may be a conserved adaptive mechanism that is essential for survival of higher plants under various abiotic stresses, such as heavy metal stress (Pan et al., 2001), salt stress (Katsuhara & Kawasaki, 1996; Katsuhara, 1997; Huh et al., 2002), etc. Clarification of the exact molecular mechanisms underlying such a complex adaptive phenomenon and an analysis of how it regulate stress-induced developmental root plasticity are two major issues that should be addressed in future research.


The authors thank Yingchun Hu, School of Life Sciences, Peking University, for assistance with transmission electron microscopy. The authors also thank Dr E. Lam and N. Watanabe for providing the atbi1 mutants, B. Wakeland and Jon Y. Wong for copy editing, and Yunqiang Gao for discussion of the manuscript. This work is supported by the ‘Hundred Talent Program’ of Chinese Academy of Sciences.