• The cellular events associated with programmed cell death during leaf senescence in rice (Oryza sativa) plants are reported here.
• The cytological sequence of senescence-related changes in rice leaves was studied by transmission electron microscopy, in situ terminal deoxynucleotidyl transferase-mediated dUTP nick end-labelling (TUNEL) assay and DNA ladder assay.
• Cell death in senescing mesophyll cells was marked by depletion of cytoplasm in a tightly controlled manner. However, no apparent morphological feature associated with apoptosis was observed. Nuclear DNA fragmentation was detectable as early as during leaf unfolding and the subsequent developmental and senescent stages. The occurrence of DNA fragmentation correlated well with the size-shift of chromosomal DNA on agarose gel after electrophoresis. However, DNA fragmentation was not accompanied by generation of oligonucleosomal DNA fragments.
• These features of cell death occurring during leaf senescence in monocot rice are quite different from features characteristic of apoptosis in animals. The implications of these results for cellular events associated with rice leaf senescence are discussed.
Leaf senescence is a highly regulated series of events involving disassembly of cellular components in the senescing tissue and recycling of valuable nutrients to other parts of the plant (Buchanan-Wollaston, 1997). This developmental process is under the control of a genetically programmed sequence, and the type of cell death is referred to as programmed cell death (PCD) (Pennell & Lamb, 1997). Many senescence-associated genes have been identified as being up-regulated during the leaf senescence process (Smart, 1994; Buchanan-Wollaston, 1997; Huang et al., 2001; Lee et al., 2001). The encoded gene products are principally involved in degradation or remobilization of biomolecules but are also involved in protecting cell viability for completion of senescence.
Apoptotic cell death is one of the most widely studied forms of PCD in animals and is defined by distinct morphological and biochemical characteristics (Martin et al., 1994). These include chromatin condensation and fragmentation, cytoplasmic condensation and vacuolization, cytoplasmic membrane blebbing and disassembly of the cell into apoptotic bodies that are rapidly engulfed by phagocytes or neighboring cells (Wyllie et al., 1980; Kerr & Harmon, 1991). Apoptosis is usually associated with the activation of nucleases that degrade the chromosomal DNA first into large fragments (50–300 kb) and subsequently into multiple fragments corresponding to the internucleosomal spacing of about 180 bp (Walker & Sikorska, 1994). In plants, selective cell death is necessary during development and also occurs in response to pathogen attack and environmental stress. For example PCD has been reported to occur during endosperm development (Young & Gallie, 1999), diploid parthenogenesis (Havel & Durzan, 1996), anther development (Wang et al., 1999), aleurone cell death (Wang et al., 1998), xylem differentiation (Groover & Jones, 1999), pollination-induced petal senescence (Xu & Hanson, 2000) and the hypersensitive response (Ryerson & Heath, 1996). It is still uncertain to what extent apoptotic cell death occurs in plants. The occurrence of oligonucleosomal DNA cleavage has been reported in some plant tissues undergoing PCD, for example, in the senescent coleoptile of rice seedlings (Kawai & Uchimiya, 2000), and in the coleoptile and initial leaf of wheat seedlings (Kirnos et al., 1997). Grass coleoptile is a particularly short-lived organ that senesces rapidly and is degraded during the early stage of seedling development. In contrast, photosynthetic leaves are relatively long-lived, ranging from weeks to months. More work is needed to clarify whether apoptotic-like cell death is an integral part of the leaf senescence process. Such information should prove valuable in determining research strategies for studying the regulatory mechanisms underlying leaf senescence in different plant species.
In this study, we examine the morphological changes during rice leaf senescence using transmission electron microscopy (TEM) and the biochemical characteristics of nuclear DNA fragmentation using terminal deoxynucleotidyl transferase-mediated dUTP nick end-labelling (TUNEL) assay in situ. We present evidence that cell death during rice leaf senescence is quite different from apoptosis in animals and is morphologically marked by progressive depletion of cytoplasm without loss of cellular compartmentalization until late in the senescence program. Our results also indicate that nuclear DNA fragmentation in mesophyll cells is not accompanied by generation of oligonucleosomal DNA fragments.
Materials and Methods
Rice seeds (Oryza sativa L. cv. Tainong 67) were imbibed in water for 2 d and sown in moist compost mix in a dark growth chamber at 30°C with 85% humidity. The etiolated seedlings were uniformly established on the fifth day and then moved to light for another 5 d until the second true leaves were fully unfolded. For dark-induced leaf senescence studies, 10-d-old-rice seedlings were transferred to continuous darkness as described previously (Lee et al., 2001). The green control was grown under a photoperiod of 14 h light/10 h dark at 30°C. For TEM studies or in situ TUNEL assay, leaf tissues were excised from the distal upper one-fifth portion of the second true leaf blade at 1, 2, 3, and 4 d of dark incubation. The relative Chl contents of the leaf samples at 1, 2, 3, and 4 d of dark treatment were > 90%, 60–70%, 20–30%, and < 5%, respectively. For natural leaf senescence studies, the rice plants were grown in a glasshouse under natural light conditions.
Leaf tissues were excised from the upper one-fifth of the eighth-leaf blade of the main culm at various stages of development and senescence (YR, GR, P, G, S1, S2, S3, S4, and S5) and used for TUNEL assay and total DNA isolation. The developmental and senescent stages were defined according to Lee et al. (2001): YR (young leaf roll before emergence of 1-month-old plant); GR (green leaf roll after emergence of 1-month-old plant); P (partially unfolded green leaf roll of 1-month-old plant), G (fully expanded green leaf of 2-month-old plant; 100% Chl), S1 (senescing leaf of the plant at the stage of panicle development, 80–95% Chl), S2 (senescing leaf of the plant at the flowering stage, 60–80% Chl), S3 (senescing leaf of the plant at the grain-filling stage, 45–60% Chl), S4 (senescing leaf of the plant at the seed maturation stage, 30–45% Chl), and S5 (senescent and withered leaf of 4-month-old plant at the seed maturation stage, < 5% Chl).
Leaf tissues (c. 2 mm2) were fixed in 1.25% (v/v) glutaraldehyde/1.25% (v/v) paraformaldehyde for 12 h at 4°C. Samples were postfixed in 1% (v/v) osmium tetraoxide for 2 h followed by dehydration in a graded series of increasing acetone concentration. The dehydrated samples were progressively embedded in epon-araldite (Agar Aids, Bishop’s Stortford, UK), followed by polymerization in the resin. Ultra-thin sections (70–90 nm) were prepared and mounted on 300-mesh coated copper grids for staining with uranyl acetate and lead citrate. All sections were examined on a Philips CM100 transmission electron microscope (Philips, Eindhoven, The Netherlands).
In situ TUNEL assay
Nuclear DNA fragmentation was identified in situ by TUNEL staining, which detects free 3′ hydroxyl groups of degraded nuclear DNA, as described by Gavrieli et al. (1992). Leaf tissues (c. 0.2 cm2) were fixed in 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS, pH 7.2) containing 0.1% (v/v) Triton X-100 and Tween 20 at 4°C overnight. The samples were dehydrated in a graded series of increasing ethanol concentration. The dehydrated samples were transferred to a series of ethanol and Histoclear mixtures at ratios of 2 : 1, 1 : 1, and 1 : 2, and then to 100% Histoclear for 1 h with two changes of solution. The samples were progressively embedded in paraplast wax to make wax blocks. Sections were cut from the wax-embedded tissues at 7-µm thickness and mounted onto poly L-lysine coated slides for TUNEL staining.
The paraffin sections were dewaxed, rehydrated, and incubated with proteinase K (20 µg ml−1 in 20 mM Tris-HCl, 2 mM CaCl2, pH 7.4) for 30 min at 37°C. The slides were then rinsed and blocked with 3% BSA in PBS buffer containing 0.1% Triton X-100, followed by two washes with PBS. The slides were incubated with TUNEL reaction mixture containing fluorescein-dUTP for 1 h at 37°C in accordance with the manufacturer’s instructions (Boehringer Mannheim GmbH, Mannheim, Germany). The slides were counter-stained with 0.0004% propidium iodide in PBS for 30 min. After a rinsing in PBS, the slides were mounted with VECTASHIELD® mounting medium (Vector Laboratories, Burlingame, CA, USA) for viewing under a fluorescence microscope, Axiovert S100 (Zeiss, Göttingen, Germany). The positive control was made by incubation with 1000 units of DNase I in 50 mM Tris-HCl (pH 7.4), 10 mM MgCl2, 0.1 mM dithioerythritol, and 1 mg mL−1 BSA for 30 min at 37°C to induce DNA strand breaks before proteinase K treatment. The negative control was prepared without terminal deoxynucleotidyl transferase in the reaction mixture.
Leaf total DNA analysis
Rice leaf total DNA at different stages of development and senescence was prepared using the method described by To et al. (1999). Briefly, leaf tissues were ground in liquid N2 and incubated in the preheated extraction buffer (100 mM Tris-HCl, pH 8.0, 50 mM EDTA, 0.5 M NaCl, 1% SDS, 1% Sarkosyl and 40 mM 2-mercaptoethanol) at 60°C for 2 h. They were then subjected to phenol extraction and isopropanol precipitation. Ribonuclease was added to the dissolved DNA to a final concentration of 50 µg ml−1 and incubated at 37°C for 1 h. SDS was further added to make up to 0.5% of the solution, and this was supplemented with proteinase K to a final concentration of 100 µg ml−1, followed by incubation at 37°C for 1.5 h. The DNA solution was further extracted with phenol, pelleted, and dissolved in TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0). For DNA ladder assay, 6 µg DNA of each sample was electrophoresed on a 1.5% agarose gel, transferred to a positively charged nylon membrane (Boehringer Mannheim GmbH, Mannheim, Germany), and hybridized with 32P-labelled rice genomic DNA. To make the probe, rice leaf total DNA was digested with Sau3A1 and labelled with α-[32P] dCTP using a random primed labelling system (NEN Life Science Products, Boston, MA, USA). Visualization of hybridization bands was carried out using PhophorImager with ImageQuant software (Molecular Dynamics, Sunnyvale, CA, USA).
Morphological changes in rice leaf mesophyll cells during the senescence process
The onset and progression of leaf senescence induced by dark incubation of young rice seedlings were characterized by a decline in Chl content of the second leaf. After 36 h of dark incubation, yellowing became slightly visible near the leaf tip. After 3 d, the upper third of the blade turned yellow. During natural leaf senescence of the greenhouse-grown plants, yellowing also occurred near the tip of the blade and gradually spread downward. For ultrastructural studies, leaf tissues from the distal upper one-fifth portion of the second leaf after various time periods of dark incubation were used for TEM. Senescence-associated cellular changes in cross sections of the leaf tips after days 2 and 3 of dark treatment appeared synchronized. The leaf mesophyll cells were marked by condensation and depletion of cytoplasm accompanied by organelle degradation and central vacuole expansion (Fig. 1a–c). The corresponding nuclear features for Fig. 1(a–c) are shown in Fig. 1(d–f), respectively. In green leaf tissue (Fig. 1a), a number of chloroplasts, mitochondria, the nucleus and a considerable amount of cytoplasmic matrix were observed at the periphery of the mesophyll cell. The volume of cytoplasm fell dramatically after 2 d of dark treatment (Fig. 1b). The cytoplasm became a very thin peripheral layer, and the cell wall appeared conspicuously deformed after 3 d of dark treatment (Fig. 1c). At this stage, the organelles remained intact, although their internal structures were lost. The senescing chloroplasts were filled with numerous plastoglobuli, and the thylakoid membranes became dilated (Fig. 1f). In the central large vacuole, many electron-dense inclusion bodies were observed (Fig. 1c), and some of these were recognized as being of chloroplast origin (data not shown).
In green leaf tissue, nuclei contained a nucleolus and some regions of dense chromatin, which appeared relatively unchanged during the first 2 d of dark treatment (Fig. 1d,e). The nuclear matrix became homogenous without a visible region of dense chromatin after 3 d of dark treatment (Fig. 1f). The nuclear envelope persisted, and the nucleolus was no longer observed at this stage. It is likely that the fragmented chromatin was dispersed throughout the senescing nucleus. Thus no condensed chromatin localized to the nuclear periphery was observed in the senescing mesophyll cells. Extensive plasmalemmasome formation was observed in accordance with the collapsed cell wall structure, which is a unique feature of rice leaf senescence (Fig. 1c). This membranous structure, containing numerous double-membraned vesicles, is likely to have derived from invagination of plasma membrane. The possible sequence for the formation of plasmalemmasomes is shown in Fig. 1(g–i). As dark treatment continued, the residual organelles shrank in conjunction with cytoplasmic condensation. Yet cellular compartmentalization persisted until breakdown of the whole cell after 5–6 d of dark treatment (data not shown). No apoptotic bodies were apparent in mesophyll cells undergoing senescence.
Detection of nuclear DNA fragmentation during leaf growth and senescence
Our preliminary work using an in situ TUNEL assay revealed that nuclear DNA cleavage occurred not only in the dark-induced senescing leaf tissue but also in the mature green leaf tissue from the second leaf of young rice seedling (data not shown). To further examine nuclear DNA fragmentation associated with leaf development, more detailed studies using the TUNEL assay were carried out with leaf tissues of greenhouse-grown plants harvested at different stages of development and senescence (Fig. 2). Nuclear DNA cleavage can be detected as yellowish-green fluorescence under blue excitation, whereas nuclear chromatin stained with propidium iodide appears red under UV excitation. The location of nuclei in the mesophyll cells is indicated with arrows. No DNA fragmentation was observed in the tissue of young or green rolled leaf (Fig. 2b,d). Nuclear DNA fragmentation began to be detected during unfolding of the upper leaf portion (Fig. 2f). The positive TUNEL staining was detected in the leaf tissues of the subsequent developmental stages (half-unfolded and fully unfolded leaves) as shown in Fig. 2(h,j), and the senescent stage shown in Fig. 2(l). The DNase I – treated tissue of young leaf roll that revealed yellowish-green fluorescence in nuclei was used as positive control (data not shown). Figure 2(a,c,e,g,i,k) are the corresponding negative controls for the leaf tissues assayed, which all showed negative TUNEL staining.
Lack of DNA laddering in leaf cells during senescence
Rice leaf total DNA isolated from leaf tissues harvested at different stages of development and senescence was subjected to agarose gel electrophoresis (Fig. 3a), followed by Southern analysis using 32P-labelled Sau3A1-digested genomic DNA as probe (Fig. 3b). The overall size of high-molecular-weight DNA began to decrease as green leaf fully unfolded (G stage), and continued to decrease gradually as leaf senescence proceeded from the S1 to S5 stage. This observation correlated with the detection of nuclear DNA fragmentation by TUNEL staining (Fig. 2). However, no accumulation of discrete low-molecular-weight DNA fragments corresponding to multimers of 180 bp was observed in any senescing tissues tested (Fig. 3a). Smearing of degraded DNA of low molecular weight was clearly observed during late stages of senescence. Lack of DNA laddering during the senescence process was further confirmed by Southern analysis (Fig. 3b). Apparently, the cleavage of nuclear DNA into oligonucleosomal fragments that is the hallmark of apoptosis in animal cells does not occur during rice leaf senescence.
Plant cells and tissues undergo various types of PCD. These include lysigeny during aerenchyma formation in corn root induced by hypoxia (He et al., 1996), xylogenesis during tracheary element formation (Fukuda, 1996), autophagy of cultured tobacco cells during sugar starvation (Moriyasu & Ohsumi, 1996) and programmed oncosis during the hypersensitive response to pathogen attack (Bestwick et al., 1997; Jones, 2000). Each type processes and removes cell corpses differently. Perhaps micro tuning creates divergence in cell death mechanisms for different purposes in plant life, with each developmental process dictating its own specific set of morphological and biological variations. The occurrence of apoptosis in plant PCD remains unclear. Apoptotic cell death may be rare during plant PCD since the presence of cell walls precludes the absorption of apoptotic bodies by neighboring cells. The mitochondrion is considered as the central control point of apoptosis in animals, involving the caspase cascade and a series of regulatory molecules that trigger or prevent apoptotic cell death (Desagher & Martinou, 2000). So far functional plant homologues of the mammalian apoptotic machinery or related components have rarely been identified (Sanchez et al., 2000). Our analyses of the ultrastructural changes and the pattern of nuclear DNA fragmentation that occur in rice mesophyll cells during leaf senescence reveal that these features have many differences from those that take place during apoptosis in animal cells.
Ultrastructural studies on leaves induced to senesce by dark incubation reveal that progressive depletion of cytoplasm in mesophyll cells is most obvious during rice leaf senescence. This is largely due to degradation of organelles, in particular, degradation of chloroplasts. Yet the mesophyll cells seem to retain a high degree of membrane integrity and cellular compartmentalization until late into the senescence program. We have also examined the morphological changes of natural senescing leaves (data not shown). Similar to dark-induced senescence, we observed a high degree of structural integrity in the senescing mesophyll cells up to S2–S3 stages when most of the cytoplasmic content was depleted and accompanied by vacuole expansion. Inada et al. (1998) also reported a similar finding during the senescence of rice coleoptiles. The inclusion bodies found in the central vacuole as shown in Fig. 1(c) are presumably of cytoplasmic origin and include degraded chloroplasts, mitochondria, and ribosomes. It has been suggested that the vacuole plays an autophagocytotic role in the clearance of cytoplasmic components in senescent cells (Matile, 1997). Our results show that vacuole integrity is maintained until cytoplasmic components become condensed and nearly depleted. By contrast, PCD in tracheary element differentiation is marked by the rapid collapse of the vacuole leading to autolysis of cellular components during secondary cell wall thickening (Groover & Jones, 1999).
No condensation of nuclear material specifically in the nuclear periphery and no fragmentation of nuclei were observed in mesophyll cells during rice leaf senescence. Condensation of cytoplasm and the nucleus was confined to the peripheral layer at very late senescent stages. This is quite different from what occurs during apoptotic cell death in animals, in which condensed and fragmented nuclear and cytoplasmic materials are enclosed into apoptotic bodies (Wyllie et al., 1980). Extensive formation of plasmalemmasomes was observed to coincide with conspicuous distortion of the cell wall. The functions of these multivesicular bodies have received very little study and they have never been reported to be associated with leaf senescence. Herman & Lamb (1992) proposed a plasmalemmasome pathway for the internalization of the periplasmic matrix for vacuolar-mediated disposal, and this may also apply to rice leaf senescence.
Nuclear DNA fragmentation and DNA laddering have been demonstrated in some plant tissues undergoing PCD. They have been recorded, for example, during endosperm development (Young & Gallie, 1999), anther development (Wang et al., 1999), pollination-induced petal senescence (Xu & Hanson, 2000), and in response to pathogen or toxic compounds (Ryerson & Heath, 1996; Stein & Hanson, 1999). However, some reports of apoptotic-like cell death in plants did not show accumulation of distinct oligonucleosomal DNA fragments (Levine et al., 1996). Using ‘comet assay’, Simeonova et al. (2000) reported detection of nuclear DNA fragmentation during leaf senescence of the monocot Ornithogalum virens and the dicot tobacco, but no evidence for DNA laddering was presented. It seems that the generation of oligonucleosome-sized DNA fragments is not an obligatory stage in the processing of the initial large DNA fragments. Our TUNEL experiments show that nuclear DNA fragmentation occurred in leaf mesophyll cells as early as during leaf unfolding and during the subsequent developmental and senescent stages (Fig. 2). Positive TUNEL staining was also observed in the dark-induced senescing leaf of rice seedling (data not shown). Nevertheless, this nuclear feature did not generate oligonucleosomal DNA fragments (Fig. 3). Although Southern analysis with 32P-labelled probe is a highly sensitive method, the possibility that lack of DNA laddering is due to low population of mesophyll cells undergoing internucleosomal DNA cleavage cannot be ruled out. Our finding is different from what occurred in senescent coleoptile cells of rice seedlings that had germinated under submergence and were subsequently transferred to aerobic conditions. Kawai & Uchimiya (2000) speculated that the initiation of coleoptile cell death might be a response to reactive oxygen species. Yen & Yang (1998) have been able to detect DNA laddering in the senescent leaves of five campus-grown tree species. The discrepancy may result from differences in the plant species and leaf developmental stages being studied. The effect of other factors including pathogen attack, air pollutants and drought should be considered since these trees were grown under the unprotected environment. Furthermore, the leaf senescence program in graminaceous monocot rice may be different from that in woody plants or perennial evergreen plants. Kirnos et al. (1997) mentioned that there was a basipetal age gradient along the blade during leaf development in cereals. They observed in wheat seedlings that DNA fragmentation was initiated when cells in the apical zone of the leaf cease to divide, and then extended basipetally as the leaf grows older. Thomas & Stoddart (1980) have mentioned the induction of a senescence program in the previous leaves of grasses and monocot plants as new leaves continue to grow within the sheaths. The sheaths of the earlier leaves have to accommodate an increasing bulk of younger leaf tissues inside them which becomes an area of stress and may respond by beginning their own senescence sequences. Specific DNA cleavage accompanied by tissue expansion may be an important transition state for senescence program to occur. Thus the senescence program is initiated while the leaf blade is green and healthy.
This work was financially supported by grants from Academia Sinica and the National Science Council of the Republic of China to Dr Shu-Chen Grace Chen, and by the postdoctoral fellowship from Academia Sinica to Dr Ruey-Hua Lee. We thank Miss M. J. Fang, Dr W. N. Jane, and Dr J. H. Tsou for their technical assistance in the microscopic studies.