Histological and cytological evidence of where and when apoptotic cells occur in Pc-2/Vb oat cells treated with victorin was obtained by observing DNA strand breaks at both light (LM) and electron microscope (EM) levels using TUNEL techniques. DNA from leaf segments that had been floated on victorin solution with the abaxial epidermis removed showed typical ladders on agarose gels. Nuclear chromatin condensation, followed by cell collapse, started in the mesophyll cells closest to the victorin solution. LM-TUNEL was positive in the non-collapsed cells but not in the collapsed cells in the treated leaves. However, the EM-TUNEL assay was positive in the nuclei of the non-collapsed as well as the collapsed cells where nuclear fragments dispersed into the cytoplasm, and the immunogold density was much higher than that in the cells killed by a high concentration of H2O2, suggesting that the victorin-treated collapsed cells are in the last stage of apoptotic cell death. The immunogold labelling in the victorin-treated non-collapsed cells was restricted to condensed heterochromatin, indicating that chromatin condensation is associated with DNA cleavage. Pharmacological studies indicated that proteases and nucleases may play a role in the apoptotic response. However, the EM-TUNEL assay indicated that EGTA co-incubated with victorin blocked DNA cleavage, but failed to prevent chromatin condensation. Moreover, protein kinases were involved in chromatin condensation, but did not affect DNA digestion, suggesting that chromatin condensation and DNA cleavage are differentially regulated in the death process in oats.
The term ‘apoptosis’ was originally introduced to describe a type of mammalian cell death with a broad significance in tissue homeostasis that exhibited cardinal morphological features, in contrast to necrotic cell death which is a passive, catabolic process (Kerr et al., 1972). Cells undergoing apoptosis shrink in size, the chromatin condenses into large granular masses (Earnshaw, 1995), and the genomic DNA is cleaved into multiple inter-nucleosomal fragments referred to as DNA ladders (Wyllie, 1980). It has been demonstrated that plant cells undergoing death in certain circumstances show similar morphological and biochemical changes to those that occur in apoptotic animal cells. Apoptosis is now recognized as a common cellular default mechanism of programmed cell death not only in animals, but also in plants (Dangl et al., 1996; Gilchrist, 1998; Greenberg, 1997; Heath, 1998; Lam et al., 1999).
The most common methods used to characterize apoptosis in plants have been light or electron microscopy to detect morphological changes, electrophoresis to detect DNA ladders, and LM-TUNEL analysis to label DNA strand breaks in situ. However, the LM-TUNEL assay is known to reveal necrotic cells under certain conditions (McCabe et al., 1997; Renvoize et al., 1998), and it has been emphasized that classification of cell death in a given model should always include morphological examination coupled with at least one other assay (Renvoize et al., 1998). Also, it is generally known that intracellular organelles, especially mitochondria, retain morphological integrity during apoptosis, although some abnormalities such as a reduction in size, a hyper-density of the matrix, transition from an orthodox to a condensed conformation, and the release of cytochrome c as a result of the mitochondrial membrane becoming permeable were recently confirmed in some cases (Desagher and Martinou, 2000). By contrast, in necrotic cells, mitochondria dilate, other organelles dissolve and the plasma membrane ruptures (Desagher and Martinou, 2000; Hayashi et al., 1998; Kerr et al., 1972). Thus, the morphological integrity of the organelles, in conjunction with DNA cleavage or other biochemical changes, is taken as a marker of apoptosis. Based on these considerations, an electron microscopic (EM) TUNEL assay as used in animal cells could be an additional method for the verification of apoptosis in plants, as EM permits simultaneous assessment of the TUNEL reaction and chromatin condensation, and enables evaluation of the morphological integrity of cytoplasmic organelles (Goping et al., 1999; Hayashi et al., 1998). The assay may also be useful for analysis of the fate of fragmented DNA during nuclear degradation and apoptotic cell death in plants.
Victorin is a host-selective peptide toxin that is essential for the pathogenicity of Cochliobolus victoriae on oat plants carrying the dominant Vb gene for toxin sensitivity (Mayama et al., 1986; Scheffer and Livingston, 1984; Wolpert et al., 1985). Furthermore, the gene Vb conferring victorin sensitivity and the Pc-2 gene controlling crown rust resistance are either very tightly linked or the same gene (Luke et al., 1966; Mayama et al., 1995). Victorin is also reported to function as a specific elicitor for anti-microbial phytoalexin accumulation in Pc-2/Vb oats (Mayama et al., 1986). The resistant response in Pc-2/Vb oats is accompanied by DNA laddering in the primary leaves of oats in response to both crown rust infection and elicitors such as chitin, chitosan oligomers and victorin (Mayama et al., 2001; Navarre and Wolpert, 1999; Tada et al., 2001). In this system, the induction of DNA laddering in the victorin-treated oat leaves was consistently associated with cell death and paralleled the expression of a newly formed nuclease p28 (Tada et al., 2001). Interestingly, the production of phytoalexin was mediated differently to the induction of DNA laddering (Tada et al., 2000; Tada et al., 2001).
In this report, we examine serial morphological changes in the nucleus and other organelles during apoptotic cell death induced by victorin in the primary leaves of oat. The methods focused on the EM-TUNEL assay coupled with the common methods of LM-TUNEL and DNA laddering. We provide cytological evidence for the occurrence of an apoptotic cell response in oats during victorin treatment, and report that the chromatin condensation and DNA cleavage that occur during apoptotic cell death in this system are mediated by different signalling pathways in victorin-treated oat leaves. As far as we know, TUNEL staining for DNA cleavage at the ultrastructural level has not been reported before for plants.
Where do apoptotic cells occur in oat leaves treated with victorin?
Inter-nucleosomal cleavage of DNA, detected as DNA ladders, was recently found in primary leaves of oats treated with victorin, suggesting the involvement of apoptosis in the response of oats to victorin (Navarre and Wolpert, 1999; Tada et al., 2001). However, no cytological data were reported to establish where and when apoptosis occurs in the victorin-treated oat leaves. In the present study, cytological and morphological analysis using both light and electron microscopy coupled with immunogold labelling were carried out to address this question.
The epidermis was peeled from the abaxial leaf surface of oat leaves. The leaf segments were then floated on 5 ng ml−1 victorin solution or water for 2–12 h by placing the peeled side down. Leaf cross-sections were stained with toluidine blue to show the general cytological state. All cells in the peeled leaf segments treated with water were intact throughout the experimental period (Figure 1a), indicating that little physical damage occurred on the exposed cells by peeling off the epidermis. However, in the victorin-treated leaves, the mesophyll cells most proximal to the peeled side appeared abnormal at 2 h, as indicated by heavier staining of chloroplasts with toluidine blue (Figure 1b). By 6 h, the first two or three cell layers in the exposed surface had collapsed in the treated leaf segments, as shown by the less-stained and degraded chloroplasts, and could be distinguished clearly from the upper non-collapsed cells (Figure 1c). The chloroplasts in the non-collapsed cells were more intensely stained and exhibited an abnormal appearance like the cells most proximal to the peeled side shown in Figure 1(b). At 12 h, most of the mesophyll cells had collapsed except those proximal to the upper epidermis (Figure 1d). These data illustrate several aspects of a continuum of changes in the mesophyll cells of the peeled leaf segments in response to victorin.
Using the same toxin treatment technique, we then carried out experiments to cytologically detect DNA cleavage by the LM-TUNEL assay. The leaf segments treated with victorin for 6 h, which contain both the collapsed and non-collapsed cells as observed with toluidine blue, were used for this study. In the non-collapsed cells distributed in the upper portion of the leaf tissue, the nuclei showed distinct fluorescent signals by both Hoechst 33342 (Figure 1f) and TUNEL (Figure 1h) staining, indicating that nuclei were intact but DNA cleavage had occurred within the nucleus by 6 h. In the collapsed cells near the exposed layer, however, few cells were stained by Hoechst or LM-TUNEL staining (Figure 1f,h). On the other hand, the nuclei in the water control were stained with Hoechst 33342 throughout the cross-section of the leaf segment (Figure 1e) but there was no detectable fluorescent signal from the LM-TUNEL reaction (Figure 1g), indicating that the DNA in the nuclei was intact in these cells.
The observations from light microscopy were further confirmed in the same samples when observed by electron microscopy. As illustrated in Figure 2, heterochromatin in the water control was evenly scattered within the nucleus, and cytoplasmic organelles remained intact for the duration of the experiment (Figure 2a). In the 2 h victorin-treated leaves, however, heterochromatin had aggregated into a few large masses in the first layer of the exposed cells and the nuclear envelope was grossly indented, but other organelles still remained morphologically intact (Figure 2b). These cells collapsed by 4 h after treatment. By contrast, in the non-collapsed cells adjacent to and above to the collapsed cells, chromatin condensed and the morphological structures of mitochondria and chloroplasts were well preserved (Figure 2c). Similar results were observed in the 6 h victorin-treated leaf tissues although the collapsed zone became larger. In the non-collapsed cells, chromatin condensed significantly, accompanied by relatively well preserved organelles (Figure 2d,e), whereas nuclei and other organelles lost their integrity in the collapsed cells (Figure 2f). These ultrastructural changes confirm the continuum of cell responses to victorin as observed by light microscopy.
Cytological and ultrastructural features of H2O2-induced cell death
A number of chemical agents such as H2O2 or CuSO4 have been used to induce necrotic cell death in plant cells (Levine et al., 1996; O'Brien et al., 1998; Ryerson and Heath, 1996). It has been reported in tobacco cells that increasing concentrations of H2O2 delivered over the same time periods resulted in either apoptosis defined by chromatin condensation as an early event, or, at higher concentrations of H2O2, necrosis without passing through an apoptotic stage (O'Brien et al., 1998). In order to use H2O2 as a necrosis-inducing agent in oat leaf cells, as opposed to victorin, we examined the effects of several dilutions of H2O2 on the morphological changes during cell death in oat leaves.
The peeled leaf segments were treated with 1, 25, 100 or 300 mm H2O2 for 6 h and subjected to toluidine blue staining and electron microscopy observation. The response of oat cells to H2O2 treatments differed depending on the concentrations (Figure 3). The first layer of mesophyll cells treated with 1 mm H2O2 showed some abnormality in their chloroplasts, which were stained differently by toluidine blue, and could be distinguished from the other cells in the upper layers (Figure 3a). However, no obvious change was found in terms of the integrity of organelles, including the nucleus, in the first cell layer (Figure 3d). In the 25 mm H2O2 treatment, the first two layers of mesophyll cells became abnormal, as shown by more faint staining with toluidine blue than those in the upper cell layers (Figure 3b), while chromatin condensation occurred in both the first and second layers of mesophyll cells on the peeled side. In the first cell layer, the nuclear membrane was disrupted in some places and chloroplasts were already degraded (Figure 3e), whereas organelles, including mitochondria and chloroplasts, had not yet degraded in the second layer (Figure 3f).
In the 300 mm H2O2 treatment, the mesophyll cells had degraded thoroughly in the treated leaf segments (Figure 3c), and the organelles, including nuclei, mitochondria and chloroplasts, had burst (Figure 3h). Similar results were obtained by treatment with 100 mm H2O2 (data not shown). Even as early as 2 h after treatment with 300 mm H2O2, the membranes of nuclei and cytoplasmic organelles had ruptured, and unusual condensed heterochromatin and a high electron-density of euchromatin were observed (Figure 3g). These features were clearly different from those induced by 25 mm H2O2 (Figure 3f) and victorin (Figure 2b), indicating that the high concentration of H2O2 could induce acute necrotic cell death in oat leaves. Thus, we used 300 mm H2O2 as a chemical agent to induce necrotic death in oat leaf cells.
Detection of nuclear DNA cleavage by the EM-TUNEL assay
To detect DNA strand cleavage in the victorin-induced apoptotic cells, immunogold labelling of DNA breaks at the 3′-OH ends (EM-TUNEL) was carried out using the LR White embedding method (Goping et al., 1999). Nuclear DNA cleavage was assessed by electron microscope observation in both the non-collapsed and collapsed cells in the peeled leaf segments during victorin treatment. As shown in Figure 4(a), intense labelling of DNA breaks with immunogold particles was found in the condensed heterochromatin region, but not in the euchromatin region of the nuclei in the non-collapsed cells of 6 h victorin-treated leaf segments, indicating a close association of chromatin condensation and DNA cleavage with cell death. We also examined the collapsed cells that were proximal to the peeled side of 6 h victorin-treated leaves by EM-TUNEL. In contrast to the LM-TUNEL assay (Figure 1h), the EM-TUNEL assay enabled us to detect fragmented chromatins labelled with immunogold particles not only in the nuclei but also in the cytoplasm of the collapsed cells (Figure 4b,c). This clearly demonstrated that the nuclei had burst and nuclear materials were scattered through the cytoplasm at this stage of apoptotic cell death. As described earlier, the non-collapsed cells of the leaf segments appeared to be undergoing changes similar to the collapsed cells as a continuum of the cell responses to victorin treatment (Figure 1). In fact, at earlier stages, such as 2 h after victorin treatment, immunogold labelling was also detected in the condensed chromatin of the nuclei in the mesophyll cells most proximal to the peeled side (Figure 4d). By contrast, only a few immunogold particles were found in the nuclei by EM-TUNEL in the water-treated leaf segments used as a control (Figure 4e). These results indicate that chromatin condensation and DNA cleavage precede apoptotic cell death.
To evaluate the frequency of DNA cleavage in victorin-induced apoptotic cell death, as opposed to H2O2-induced necrotic cell death, statistical analysis was performed on the data obtained in the EM-TUNEL assay. The density of immunogold labelling in the cells treated with victorin for 6 h was significantly higher than that in the cells treated with 300 mm H2O2 for the same period, not only in the non-collapsed but also in the collapsed cells where the condensed chromatin had fragmented into the cytoplasm (Figure 5). This suggests that victorin-induced cell death is apoptotic, in contrast to the necrotic death induced by a high concentration of H2O2. Interestingly, the non- collapsed cells treated with 25 mm H2O2 for 6 h showed slightly higher immunogold labelling than did the collapsed cells treated with 300 mm H2O2. This result suggests that the lower concentration of H2O2 (25 mm) could induce a distinct type of cell death in the non-collapsed cells that was different from the acute cell death induced by 300 mm H2O2 in the collapsed cells. These two types of H2O2-induced cell death were also morphologically distinguishable (Figure 3), which appears to suggest that the death associated with lower concentrations of H2O2 is apoptotic. Nevertheless, the density of labelling in those non-collapsed cells treated with H2O2 was significantly lower than that in the non-collapsed or even collapsed cells in the victorin-treated leaves, indicating a strong induction of nuclear DNA cleavage and apoptotic cell death in the victorin-treated cells.
Relationship between DNA laddering and chromatin condensation
To investigate the relationship between morphological changes in nuclei and DNA laddering induced by victorin, plant DNA was extracted from oat leaves treated with 5 ng ml−1 victorin for various periods of time. Equivalent leaf segments were then prepared for EM analysis. As shown in Figure 6(a), DNA laddering was detected at 6–12 h after victorin treatment in the victorin-sensitive oat cultivar X469, but no fragmentation was observed in the victorin-resistant cultivar X424. To assess the frequency of chromatin condensation in the victorin-treated tissues, we counted a total of 150 nuclei in both collapsed and non-collapsed mesophyll cells by electron microscopy. Only the cells typified by large masses of condensed chromatin and accompanied by well-preserved organelles as shown in Figure 2(b–e) were counted to maintain a fixed standard of chromatin condensation and ensure accuracy as there was a variation in the degree of condensation among the affected cells. As indicated earlier, there is a continuum of change in chromatin condensation so this approach should give a conservative estimate of the cells with this apoptotic morphology. At 2 h after victorin treatment, about 5% of the nuclei in mesophyll cells mainly located at the exposed side had the typical condensed chromatin (Figure 6b), even though DNA laddering could not be detected at this time (Figure 6a). Therefore, chromatin condensation is an earlier indicator of the initial phases of cell death than the detection of DNA laddering. By 6 h, about 30% of the apoptotic cells showed condensed chromatin (Figure 6b). They were mainly found in the non-collapsed cells restricted to the upper portion of the leaf segments. The other 70% of mesophyll cells at this time point were mainly collapsed cells with degraded chromatin and non-collapsed cells in which heterochromatin showed variance in the degree of condensation. At 12 h after victorin treatment, however, over 95% of the mesophyll cells had collapsed, the nuclei had degraded and chromatin condensation was no longer visible. A few typical condensed chromatin was found only at the cells most proximal to the upper epidermis. However, the DNA ladder was detected at 12 h as clearly as in the 6 h victorin treatment. This DNA cleavage was supported by the coincidental labelling of the nuclear material with immunogold particles in the collapsed cells as observed in the EM-TUNEL assay (Figure 4b,c). The data again suggest that the collapsed cells of oats induced by victorin are the last stage of apoptotically programmed cell death and can be detected by EM-TUNEL.
Proteases and nucleases involve in chromatin condensation and DNA cleavage
In this study, various inhibitors of the signal mediators were used to study the relationship between DNA cleavage and chromatin condensation. When leaf segments treated with victorin were co-incubated with the serine protease inhibitor aprotinin and the nuclease inhibitor aurintricarboxylic acid, chromatin condensation was completely blocked as shown by electron microscopy (Table 1). The cysteine protease inhibitor E-64 seemed to be less effective in blocking chromatin condensation and DNA laddering at 300 µm (Table 1). At 500 µm, E-64 completely prevented chromatin condensation and DNA laddering, suggesting that cysteine proteases were associated with DNA laddering and chromatin condensation. Treatment of the cells with these inhibitors not only blocked DNA laddering but also blocked the immunogold labelling in both the LM- and EM-TUNEL assays (data not shown). As in animals, proteases and nucleases may play a central role in the initiation and completion of apoptotic oat cell death followed by chromatin condensation and DNA cleavage during victorin treatment.
Table 1. Effects of inhibitors on victorin-induced chromatin condensation and DNA ladders
a Epidermis-peeled leaf segments were co-incubated with 5 ng ml−1 victorin (or water for the water control) and the indicated inhibitors for 6 h.
b For each treatment, 150 mesophyll cell nuclei were counted. The criteria for chromatin condensation are described in Experimental procedures. The results are means ± SD of three independent experiments.
DNA was extracted with the same samples. +, DNA ladder present; ±, DNA ladder weakly present; –, DNA ladder absent.
Influence of Ca2+ depletion on chromatin condensation and DNA cleavage
To investigate the influence of calcium depletion on chromatin condensation in nuclei, we used chemical agents such as EGTA (an extracellular calcium chelator), ruthenium red (a blocker of organellar calcium channels) and zinc chloride (a calcium antagonist). As shown in Table 1, zinc chloride prevented chromatin condensation, and ruthenium red was partially effective in preventing the induction of nuclear morphological change. These results completely paralleled the data on inhibition of DNA ladders.
DNA laddering induced by victorin also could be prevented by EGTA (Table 1). However, EGTA seemed to enhance the effect of victorin on chromatin condensation (Table 1; Figure 7c). In fact, treatment with EGTA at 20 mm for 6 h caused an initial condensation in the absence of victorin (Table 1), but these changes were reversible at 12–18 h after the treatment (data not shown). When victorin-treated segments were co-incubated with EGTA for 6 h, the collapse of cells in the exposed side of victorin-treated tissues (Figure 1c) was clearly suppressed (Figure 7a). It is also noteworthy that EGTA treatment caused cytoplasmic shrinkage. Hoechst staining revealed that nuclei remained intact up to at least 6 h (Figure 7b). In addition, the nuclei did not react in the LM-TUNEL assay, but EM observation indicated that chromatin condensation was induced in the cells co-incubated with victorin and EGTA (Figure 7c and Table 1). The immunogold assay was then used to examine DNA strand breaks. Surprisingly, far fewer immunogold particles were found in the intensely condensed chromatin area in the cells treated with victorin and EGTA (Figure 7d) compare with those treated with victorin alone. These data indicate that DNA fragmentation is not always concomitant with chromatin condensation and that extracellular calcium may not be involved in the induction of chromatin condensation, although it appears to be involved in DNA cleavage as would be the case if Ca2+-activated endonucleases were involved in plant as well as animal apoptosis (Mittler and Lam, 1995).
Protein kinase is involved in chromatin condensation but not DNA cleavage
Protein kinase, which is an important component of signal transduction pathways in plants, has been associated with the hypersensitive response (Stone and Walker, 1995; Suzuki et al., 1999). A pharmacological test on victorin-induced apoptotic cell death in oats indicated, however, that protein kinase was not involved in the induction of DNA laddering (Table 1; Tada et al., 2001).
To further study the role of protein kinases in induction of chromatin condensation, we simultaneously treated the leaf segments with victorin and the protein kinase inhibitors K-252a and staurosporine. K-252a prevented the victorin-induced collapse of the cells except for a few cells in the first layer of exposed surface (Figure 7e). Most of the nuclei were clearly stained with Hoechst 33342 (Figure 7f) but not by LM-TUNEL (data not shown), suggesting that the nuclei were intact. Although no chromatin condensation in the nuclei was observed by EM (Figure 7g; Table 1), significant numbers of gold particles were observed by the EM-TUNEL assay in the discrete chromatin (Figure 7h), indicating that DNA cleavage occurred in these nuclei without chromatin condensation. In fact, clear DNA laddering was previously demonstrated in oat tissue treated with victorin and K-252a (Tada et al., 2001). DNA cleavage in the discrete chromatin was also observed with staurosporine treatment (Table 1). Taken together, these findings indicated that protein kinases play an important role in inducing chromatin condensation but are not directly involved in inter-nucleosomal cleavage induced by victorin in oats.
Identification of apoptotic cells by a combined histochemical and ultrastructural immunogold assay
Despite several recent reports describing apoptosis induction in plants in response to various stresses, none have demonstrated cytological and morphological characteristics specific to apoptosis in plants using EM-TUNEL coupled with LM-TUNEL and DNA laddering assays. We analysed in detail a serial set of morphological changes that dying oat cells undergo in response to victorin, ranging from collapsed cells to non-collapsed cells contiguous to the collapsed cells in the region of contact with victorin. We compared the morphological and cytological changes in both of the afore-mentioned types of victorin-treated cells to those in other leaf cells treated with toxic chemicals that induce death by a necrotic mechanism. The results demonstrate that the non-collapsed cells of the victorin-treated leaf show the major hallmarks of apoptosis characterized in animals, and were most likely apoptotic cells based on the following facts. The nuclei of the cells showed typical chromatin condensation in response to victorin (Figure 2), and intense immunogold labelling of nuclear DNA at 3′-OH ends, indicative of inter-nucleosomal DNA cleavage, was localized to the condensed heterochromatin of the nuclei (Figure 3). Conversely, DNA cleavage was much less in the chemically induced necrotic cells (Figure 5). Mitochondria in the non-collapsed cells treated with victorin remained morphologically intact during chromatin condensation (Figure 2c,d), which is another hallmark of apoptosis, whereas mitochondria in the necrotic cells rapidly disintegrated (Figure 3g,h). We also observed that the chromatin condensation, immunogold labelling and DNA laddering in the victorin-treated leaf segments can be blocked with chemical agents that inhibit proteolytic cleavage by cysteine and serine proteases (Table 1; Tada et al., 2001), which play essential roles in apoptosis induction (Lam et al., 1999; del Pozo and Lam, 1998; Solomon et al., 1999; Thomberry and Lazebnik, 1998). The most critical aspects differentiating the victorin-induced apoptotic cells from the chemically induced necrotic cells were the immunogold labelling density coupled with the sustained morphological integrity of nuclei, mitochondria and chloroplasts throughout the early stages of the apoptotic cell response as described above. From these observations, we conclude that the non-collapsed cells contiguous to the collapsed cells in the victorin-treated oat leaves are in the early stages of apoptotic cell death in a continuum of cell responses to victorin that culminates in the total collapse of the affected cells.
On the other hand, the cells that collapsed during victorin treatment in the peeled leaf segments are considered to be indicative of later or terminal stages of the apoptotic cell response (Figures 1d and 2f). At this point, the EM-TUNEL assay revealed intense immunogold labelling in disintegrated chromatin fragments dispersed into the cytoplasm of the collapsed cells (Figure 4b,c), and the labelling density was much higher than that in the necrotic cells killed by high concentrations of H2O2 (Figure 5). The decreased amount of gold labelling in the victorin-induced collapsed cells is most likely due to the complete degradation of condensed chromatin during cell collapse (Figure 4). Unlike the victorin-induced cell death, treatment with 300 mm H2O2 caused rapid cell death with extreme organelle lysis, especially in mitochondria and chloroplasts (Figure 3) with a concomitant limited amount of gold labelling. Thus, we believe that the collapsed cells in victorin treatment are post-apoptotic cells, although the apoptotic bodies that are sometimes reported in animal cells were not observed.
Chromatin condensation and DNA cleavage during apoptotic cell death are mediated by different signals
In this study, we provide evidence that the signal mediators for chromatin condensation and DNA cleavage are not always the same, although they always occur in parallel in the oat leaf cells in response to victorin. Inhibitors of signal mediators such as proteases, nucleases, oxidative stress and Ca2+ suppressed the DNA fragmentation and chromatin condensation simultaneously, with the exception of EGTA as an extracellular Ca2+ chelator and K-252a as an inhibitor of serine/threonine kinase (Table 1; Figure 7). The treatment of EGTA plus victorin effectively blocked DNA digestion as assessed by DNA laddering and immunogold labelling (Figure 7d), but failed to prevent chromatin condensation (Figure 7c), suggesting that extracellular calcium influx may not be required for chromatin condensation, at least when cell death is induced by victorin. Conversely, DNA cleavage could be detected by EM-TUNEL, but chromatin condensation did not occur when protein kinase inhibitors were applied with victorin (Figure 7h; Table 1), indicating that protein kinase may play an important role in regulating chromatin condensation, but is not required for DNA fragmentation. It also has been reported that kinase inhibitors could effectively suppress programmed cell death triggered by oxidative stress or an elicitor (Levine et al., 1996). Taken together, these data clearly demonstrate that chromatin condensation and DNA cleavage during apoptotic cell death are mediated by different, but potentially converging, signalling pathways in victorin-treated oat leaves. In addition, these studies also indicate that EM-TUNEL is very useful when signal pathways for chromatin condensation and DNA cleavage are analysed in relation to induction of apoptotic cell death in plants.
Proteases and nucleases involved in chromatin condensation and DNA cleavage
In the oat–victorin system, we also showed that nuclease and protease inhibitors effectively suppress DNA laddering (Table 1; Tada et al., 2001). Further, we confirmed that protease inhibitors prevented chromatin condensation and immunogold labelling indicating DNA cleavage in the nuclei of the apoptotic cells induced by victorin, strongly suggesting that proteases play a central role in the apoptotic nuclear disassembly and that activation of proteases precedes activation of nucleases as is the case in animal cells undergoing apoptosis (Solomon et al., 1999; Thomberry and Lazebnik, 1998). In tobacco and cowpea plants, a caspase-like activity seemed to be involved in hypersensitive cell death, a programmed type of cell death in those plants (D'Silva et al., 1998; del Pozo and Lam, 1998). It has been shown in animal cells that chromatin condensation is associated with irreversible proteolytic degradation of the nuclear lamins that are believed to play a role in nuclear envelope integrity and the organization of interphase chromatin (Lazebnik et al., 1995; McConkey, 1996). We previously showed that the inhibition of cysteine protease activity suppresses the activation of a nuclease that occurs during the process of cell death and appearance of DNA laddering, suggesting the involvement of a cysteine protease in production of the nuclease (Tada et al., 2001). Although the different signalling mediators for both chromatin condensation and DNA cleavage were currently shown, these are major enzymes in a common pathway triggering the apoptotic degradation of oat cells. The elucidation of detailed mechanisms of chromatin condensation, DNA cleavage and their relationship are essential targets in the study of apoptotic cell death in plants.
Victorin/Vb oat system as a system to study the role of mitochondria in apoptotic cell death
It has been reported that victorin binds to the mitochondrial glycine decarboxylase complex and inhibits the photorespiratory cycle, which is then linked to mitochondria, chloroplasts and peroxisomes (Wolpert et al., 1994). This is reminiscent of the observations in animal apoptosis where mitochondria play a major role in the induction of apoptosis (Desagher and Martinou, 2000). It has been shown in animals that the mitochondrial changes could be the critical phase in commitment to both apoptotic and necrotic types of cell death although this remains controversial (Saraste and Pulkki, 2000). Recent reports in plants also suggest that cytochrome c and reactive oxygen species appear to be released from mitochondria and may participate in the induction of cell death in plant cells (Lacomme and Santa Cruz, 1999; Maxwell et al., 1999). As presented in the current and previous papers (Navarre and Wolpert, 1999; Tada et al., 2001), victorin induced an apoptotic cell response with the hallmarks of apoptosis as described above. In addition, it has been reported that leaf slices of oats treated with victorin undergo specific proteolytic cleavage of the Rubisco large subunit in chloroplasts and chlorophyll loss; thus, victorin induces a apoptotic/senescence-like response (Navarre and Wolpert, 1999). However, it is not clear how the molecular malfunction of mitochondria induced as a result of the inhibition of glycine decarboxylase complex by victorin relates to the induction of the apoptotic cell response but these data are consistent with the hypothesis that mitochondrial stress leading to the release of cytochrome c and activation of apoptosis could function in plants as it does in animals (Sun et al., 1999). The victorin/Vb oat system is definitely one of the best systems to evaluate the role of mitochondria in the induction of apoptosis in plants.
Advantage of the EM-TUNEL assay for identification of apoptosis
Our current results show that the EM-TUNEL assay is a very effective method to simultaneously analyse nuclear DNA cleavage and ultrastructural changes in organelles at the cellular level. In fact, the EM-TUNEL assay could detect DNA cleavage within condensed chromatin as early as 2 h after victorin treatment in the victorin-sensitive oat leaves (Figures 2b and 4d), whereas clear DNA laddering was found in 6 h victorin-treated leaf segments (Figure 6a). These results indicate that detection of chromatin condensation and DNA cleavage by EM-TUNEL is a more sensitive and earlier indicator of the initial phases of the apoptotic response. In fact, as described in our previous study, the induced p28 nuclease was identified as early as 1 h after victorin treatment (Tada et al., 2001). The current data therefore support the idea that activation of p28 precedes the appearance of inter-nucleosomal cleavage of chromosomal DNA. Even more significant is the fact that we were able to use the EM-TUNEL assay to detect DNA fragmentation when other evidence was difficult or impossible to obtain (Figure 7d,h). As mentioned above, the EM-TUNEL assay was crucial to demonstrate that the victorin-induced collapsed cells were apoptotically responding cells. The LM-TUNEL assay could detect the apoptotic markers only in the non-collapsed apoptotic cells in the induction stage, but not in those that had already collapsed (Figure 1h). It is not known why LM-TUNEL could not detect the markers when cells had transited to the post-apoptotic collapsed stage. However, this lack of marker resolution is most likely due to the complete disruption of nuclei and dispersal of fragmented chromatin throughout the cytoplasm leading to a decrease in the labelling density of DNA ends in these cells. Moreover, we also noted that LM-TUNEL did not detect the nuclei in the cells co-treated with victorin and K-252a, where the inhibitor prevented chromatin condensation (Figure 7g) and cell collapse (Figure 7e), but did not inhibit DNA cleavage (Figure 7h). Therefore, one possible conclusion that can be drawn from these data is that LM-TUNEL can be successfully used only when both chromatin condensation and DNA fragmentation exist, as rapid DNA degradation could not be detected by the assay in uncondensed nuclei or condensed chromatin with intact DNA. This could be one of the reasons for false identification by the LM-TUNEL assay of apoptotic cells in plants as suggested elsewhere (Fath et al., 1999).
Finally, several methods for identification of apoptotic oat cells have been described here. None of them used alone are sufficient to unambiguously discriminate between healthy, apoptotic and necrotic cells (Renvoize et al., 1998), but when applied together or in select combinations it is possible to confirm apoptotic cells. We believe that a combination of assays including EM-TUNEL is important and especially useful for pharmacological and cytological approaches to study the apoptotic mechanisms leading to cell death in relation to the hypersensitive defence response in plants.
The near-isogenic oat lines X424 (victorin-insensitive) and X469 (victorin-sensitive) were grown in vermiculite in a growth chamber under a 16 h photoperiod at 20°C, as described previously (Mayama et al., 1986). Seven-day-old primary leaves were used for victorin treatments and pharmacological tests.
Victorin C, a host-selective toxin produced by Cochliobolus victoriae, was kindly provided by T.J. Wolpert (Oregon State University, Corvallis, Oregon, USA) and applied at concentrations of 5 ng ml−1. Three calcium inhibitors were used as aqueous solutions in this study. EGTA (Nacalai, Kyoto, Japan), a calcium chelator; ruthenium red (Sigma), an inhibitor of intracellular calcium movement; and zinc chloride (Nacalai), a calcium antagonist, were used at 20 mm (10 mm MOPS pH 7.2), 100 µm and 20 mm, respectively. Other inhibitors employed for pharmacological tests included protein kinase inhibitors (5 µm K-252a (Sigma) and 100 nm staurosporine (Sigma)); a cysteine protease inhibitor (300 or 500 µm E-64); a serine protease inhibitor (150 µg ml−1 aprotinin (Sigma)) and a nuclease inhibitor (500 µm aurintricarboxylic acid). The concentration of each inhibitor was determined according to our previous report (Tada et al., 2001). Exogenous H2O2 also was used to induce death at various concentrations (1–300 mm).
Victorin and inhibitor treatments
The lower epidermis of primary oat leaves was carefully peeled off, and 5 cm segments, taken at 1–6 cm from the leaf tip, were floated on 3 ml victorin solution or water in glass Petri dishes with the peeled surfaces in contact with the liquid. Inhibitors were co-incubated with victorin (or water for the controls) at the concentrations described above. After incubation for 2–24 h (6 h for inhibitor tests) under light conditions in the growth chamber, the leaf segments were used for biochemical and cytological analyses.
LM- and EM-TUNEL procedures
The treated leaf segments were cut into small pieces (2 × 3 mm2) and immediately fixed in 4% paraformaldehyde, 0.2% glutaraldehyde with a 1/15 m phosphate buffer, pH 7.4, overnight at 4°C. Two different embedding methods were applied to generate the light microscopic samples. For staining nuclei with Hoechst 33342, the fixed materials were embedded in a 4% agar solution (50°C) to make agar blocks, then cut at 15–35 µm thickness with Microslicer (DSK, Kyoto, Japan). The sliced samples were stained with 5 µg ml−1 Hoechst 33342 directly on the slides for 10 min at room temperature. For the terminal deoxynucleotidyl transferase-mediated dUTP nick end-labelling (TUNEL) assay, a modified LR White embedding method was used. The LR White method may used to detect double-strand breaks in DNA at both the LM and EM levels (Goping et al., 1999). After an overnight wash in phosphate buffer, the samples were dehydrated through a graded series of ethanol and infiltrated with medium grade LR White (London Resin Company, Basingstoke, UK) at room temperature for one day, and then polymerized (50°C) in gelatin capsules containing LR White resin for 24 h. Semi-thin sections were cut at 2.5 µm thickness and mounted on slides. One group of slides was stained with toluidine blue for tissue observation; another group was prepared for the TUNEL assay.
The ApopTag® Plus Fluorescein in situ Apoptosis Detection Kit (Intergen) was used to observe apoptotic cells from LR White semi-thin sections. The specimens were washed with PBS buffer (50 mm sodium phosphate, 200 mm NaCl, pH 7.4) and equilibration buffer before adding terminal deoxynucleotidyl transferase (TdT). The reaction mixture containing TdT was incubated in a humidified chamber at 37°C for 1 h and the reaction was stopped by adding stop/wash buffer. Specimens incubated without TdT were employed as a negative control. The specimens were then mounted and incubated with anti-digoxigenin–fluorescein antibody conjugate for 45 min at room temperature. After washing three times with PBS, the specimens were viewed by a fluorescence microscope (Olympus® BX-FLA) with an FITC filter. For electron microscopy observation, the LR White blocks were cut with a diamond knife (Diatome, Bienne, Switzerland) on a Reichert-Nissei Ultracut (Leica AG, Austria) to obtain ultrathin sections (85 nm thick), which were then collected onto nickel grids, etched in saturated sodium metaperiodate for 30 sec, and rinsed in distilled water. The sections were labelled with digoxigenin using TdT in the same manner as described above. After applying stop/wash buffer at 37°C for 30 min, the TdT mixture-labelled sections were washed three times with 0.02 m Tris buffer (pH 7.4) containing 0.1% bovine serum albumin, and then incubated with sheep anti-digoxigenin antibody conjugated to 10 nm colloidal gold (British BioCell, Cardiff, UK) at room temperature for 1 h. Negative controls were prepared by incubation of TdT-untreated specimens with the colloidal gold. After three washes with distilled water, the ultrathin sections were stained briefly with uranyl acetate and lead citrate, then examined and photographed in a transmission electron microscope (Hitachi-7100, Tokyo, Japan) at an accelerating voltage of 75 kV.
Image analysis for free 3′-OH DNA ends
Ultrathin cross sections of the treated oat leaves were prepared from three different LR White blocks. Twenty-five mesophyll cell nuclei were photographed in each of the treatments with a final magnification of × 45 500. To determine the nuclear area (µm2) and the numbers of immunogold particles in the nuclei, an image analysis computer (MOP videoplan, Kontron, Munich, Germany) was used as described by Park and Unno (1999). The density of immunogold particles is shown as mean values with standard deviation.
The leaf segments were pre-fixed in 2.5% glutaraldehyde with 0.1 m cacodylate buffer and post-fixed with 1% osmium tetroxide. The samples were further dehydrated in ethanol and embedded in Epon 812 resin. Then the ultrathin sections were stained with 2% uranyl acetate (20 min) and afterwards by lead citrate (10 min). The criteria for chromatin condensation were based on the EM observations where the heterochromatin was typically aggregated into large masses and the other organelles remained structurally intact as shown in Figure 2(b–e).
DNA extraction and analysis
Plant DNA was extracted by the cetyltrimethylammonium bromide (CTAB) method as described previously (Tada et al., 2001). Treated oat leaf segments were ground with liquid nitrogen into a fine powder. The samples were incubated for 30 min at 65°C in 2% CTAB solution (100 mm Tris–HCl, pH 8.0, 1.4 m NaCl, 20 mm EDTA, 2% CTAB) and then mixed with an equal volume of chloroform:isoamyl alcohol mixture (24:1). After gently shaking, the mixture was centrifuged for 15 min at 10 000 g. The chloroform:isoamyl alcohol extraction was repeated twice. Total DNA was precipitated by the addition of 1% CTAB solution (50 mm Tris–HCl, pH 8.0, 10 mm EDTA, 1% CTAB) to the supernatant. After centrifugation for 10 min at 8000 g, the pellet was dissolved in 1 m CsCl, and subjected to alcohol precipitation with a twofold volume of 100% ethanol. DNA was recovered by centrifugation for 10 min at 10 000 g, followed by washing with 70% ethanol, and dissolved in Tris–EDTA buffer containing 0.1 mg ml−1 RNAase. After incubation at 37°C for 30 min, phenol extraction and ethanol precipitation were performed to recover DNA. The DNA samples stained with 0.5 µg ml−1 ethidium bromide were detected on a 2% agarose gel. DNA ladders were photographed with an Image Saver AE-6905C (ATTO Corporation, Tokyo, Japan).
We are grateful to Dr David Gilchrist, University of California (Davis, California, USA) for his critical reading of the manuscript and advice. This work was supported in part by a Grant-in-Aid for Scientific Research (No. 12052215) from the Ministry of Education, Science, Sports and Culture of Japan.