Subcellular localization of ozone-induced hydrogen peroxide production in birch (Betula pendula) leaf cells

Authors

  • Riikka Pellinen,

    1. Institute of Biotechnology, and
    2. Department of Biosciences, Division of Genetics, University of Helsinki, PO Box 56 (Viikinkaari 5D), Helsinki FIN-00014, Finland
    Search for more papers by this author
  • Tapio Palva,

    1. Institute of Biotechnology, and
    2. Department of Biosciences, Division of Genetics, University of Helsinki, PO Box 56 (Viikinkaari 5D), Helsinki FIN-00014, Finland
    Search for more papers by this author
  • Jaakko Kangasjärvi

    Corresponding author
    1. Institute of Biotechnology, and
    2. Department of Biosciences, Division of Genetics, University of Helsinki, PO Box 56 (Viikinkaari 5D), Helsinki FIN-00014, Finland
    Search for more papers by this author

*For correspondence (fax +358 9 191 59 079;
e-mail jaakko.kangasjarvi@helsinki.fi).

Summary

The atmospheric air pollutant ozone (O3) is one of the environmental stresses that induce formation of reactive oxygen species (ROS) in plants. Previously, the toxicity of O3 has been believed to be a result of ROS formation from O3-degradation. Recently, however, it has been shown that O3 induces active ROS production, which suggests that O3-responses may be mechanistically similar to pathogen-induced responses and that O3-damage could be a result of deleterious firing by the ROS of pathways normally associated with the HR. The subcellular localization of O3-induced H2O2 production was studied in birch (Betula pendula). O3 induced H2O2 accumulation first on the plasma membrane and cell wall. Experiments with inhibitors of possible sources for H2O2 in the cell wall suggested that both NADPH-dependent superoxide synthase and the cell wall peroxidases are involved in this H2O2 production. The H2O2 production continued in the cytoplasm, mitochondria and peroxisomes when the O3-exposure was over, but not in chloroplasts. The timing of mitochondrial H2O2 accumulation coincided with the first symptoms of visible damage and, at the same time, the mitochondria showed disintegration of the matrix. These responses may not be directly connected with defense against oxidative stress, but may rather indicate changes in oxidative balance within the cells that affect mitochondrial metabolism and the homeostasis of the whole cell, possibly leading into induction of programmed cell death.

Introduction

Reactive oxygen species (ROS), including hydrogen peroxide (H2O2), superoxide (O2), hydroxyl radicals (OH) and singlet oxygen, are produced in the normal cellular metabolism in several subcellular compartments (Foyer et al. 1994). Several recent studies on plant defense and stress-acclimation have focused on the dual role of H2O2 as a signal molecule, and as a mediator of oxidative damage. H2O2 acts as a signal molecule in the cell collapse during the hypersensitive response (HR), in the systemic acquired resistance (SAR), within the pathogen challenged tissue (Lamb & Dixon 1997), and in the acclimation to photo-oxidative stress (Karpinski et al. 1999). On the other hand, H2O2 can cause direct damage to membranes and proteins, especially through the formation of hydroxyl radicals. To avoid this damage, and to maintain the cellular redox balance, accumulation of ROS is normally kept under control by various enzymatic and chemical scavenging mechanisms (Noctor & Foyer 1998).

The atmospheric air pollutant ozone (O3) is one of the environmental stresses that induce ROS formation in plants. O3 enters plant leaves through stomata and degrades in the apoplast forming various ROS such as O2 and singlet oxygen, which can further react with the components of the cell wall and plasma membrane (Kangasjärvi et al. 1994; Kanofsky & Sima 1995; Runeckles & Vaartnou 1997). It is believed that O3 causes more damage to forest trees than any other air pollutant. Birch has been shown to be an O3-sensitive species. Exposure to even less than 150 nL L−1 of O3 causes HR-like lesions in sensitive birch clones. O3 is also sufficient to induce stress-related defense mechanisms in birch, including an increase in the mRNA levels of defense and energy metabolism-related genes Pal (Tuomainen et al. 1996), Ypr10 (Pääkkönen et al. 1998) and Mpt1 (Kiiskinen et al. 1997), an increase of antioxidative enzyme activities, and single cell death in both resistant and sensitive birch clones (Tuomainen et al. 1996).

Previously, the toxicity of O3 has been believed to be mostly due to loss of membrane integrity and formation of toxic and oxidative products by O3-degradation (Heath 1994). Recently, a new picture has started to emerge with evidence indicating that O3-induced plant responses may be mechanistically similar to pathogen-induced responses including the HR (Kangasjärvi et al. 1994; Sandermann et al. 1998; Schraudner et al. 1998). Active H2O2 (Schraudner et al. 1998) and O2 (Kettunen et al. 1999; Overmyer et al. 1998; Rao & Davis 1999) production has been detected during and after O3-exposure. These results suggest that O3-responses and damage could be a result of deleterious firing of pathways normally associated with the HR and that this would be controlled by the ROS production. The subcellular location and duration of this ROS production in O3-exposed plants, however, is not known.

The aim of this work was to characterize the subcellular localization of O3-induced H2O2 production in birch in order to understand the mechanisms of O3-induced defense response and cell death. It is important to localize the H2O2 production at the subcellular level to better elucidate the mechanisms and processes involved in O3-induced ROS production and its significance in cell death. The results indicate that O3 induces H2O2 accumulation that is localized first on the plasma membrane, continues when the O3-exposure is over, and can be later seen in the cytoplasm, mitochondria and peroxisomes, but not in chloroplasts.

Results

Subcellular localization of H2O2 accumulation with CeCl3 staining

O3 -degradation in the apoplast can lead to the formation of various short-lived ROS that disappear immediately in reactions with the components of cell wall and the plasma membrane. Electron paramagnetic resonance (EPR) studies by Runeckles & Vaartnou (1997) showed that the O3-derived O2 signal disappeared in less than 15 min after removing the plants from O3. Furthermore, as the calculated half-life of O3 in biological tissue is very short (70 × 10−9 sec), O3 cannot penetrate the plasma membrane and will disappear from the tissues almost immediately without continuous flux (Heath & Taylor 1997). To differentiate the H2O2 derived from decomposition of O3 from the endogenous, active H2O2 production by the cells, the plants used in the experiments were removed from O3 at least 25 min before sampling. Samples were infiltrated with CeCl3 and kept further in clean air. Thus the CeCl3 precipitates visible in TEM (Fig. 1) are a result of active H2O2 production by the cells that also continue in the absence of O3, i.e. they are not a result of the low level of direct H2O2 formation from O3-degradation that can theoretically take place (see Heath & Taylor 1997) in the cell wall.

Figure 1.

Subcellular localization of O3-induced H2O2 accumulation in birch leaf cells with CeCl3-staining and TEM.

CeCl3 precipitates with H2O2 forming electron dense cerium perhydroxide precipitates, visible as black spots in the EM-figures. Examples of individual precipitates are indicated by arrows in (c,d,f). Samples were exposed to 150 nL l−1 of O3 for 8 h and further kept in clean air until fixation.

(a) Spongy mesophyll cells in clean air show no H2O2 staining. (b) Vascular tissue in clean air showing H2O2 in the cell walls undergoing lignification. (c) The cell walls of spongy mesophyll cells accumulate H2O2 2 h after the beginning of the O3 exposure. (d) Strong local H2O2 accumulation on the plasma membranes of two adjacent palisade parenchyma cells 10 h after the beginning of the 8 h O3 exposure. (e) H2O2 accumulation on the plasma membrane and cell wall of a palisade parenchyma cell 10 h after the beginning of the exposure. From the cell wall, H2O2 seems to be diffusing into the adjacent epidermal cell. (f) Mitochondrial and peroxisomal H2O2 accumulation 24 h after the beginning of the O3 exposure in the spongy mesophyll cells. (g) H2O2 accumulation is visible in the cell walls, but not in the chloroplasts of O3-exposed plants at 10 h. (h,i) Methyl viologen induces H2O2 accumulation in chloroplasts. (h) 0 μm and (i) 5 μm methyl viologen infiltrated leaf sections after 2 h of illumination under 200 μMol m−2 sec−1. Methyl viologen-induced H2O2 accumulation can be seen as black precipitate spots (indicated with an arrow) scattered on the thylakoids and in the chloroplast stroma in (i). Abbreviations: c, chloroplast; m, mitochondrion; cw, cell wall; s, starch grain; p, peroxisome; ec, epidermal cell; ppc, palisade parechyma cell; lb, lipid body. Bar, 500 nm.

H2O2 accumulation was not detectable in any other cell types of clean air-grown birch (Fig. 1a), except in the cell walls of vascular tissue (Fig. 1b). Accordingly, ROS production in vascular tissues undergoing lignification has been used as a positive control in the localization of H2O2 production (Bestwick et al. 1997; Thordal-Christensen et al. 1997). Two hours after the beginning of the 8 h (150 nL L−1) O3-exposure (all time points for sample analysis are expressed as hours after the beginning of the 8 h O3-exposure), H2O2 accumulation was visible in the cell walls and plasma membranes of spongy mesophyll and palisade parenchyma cells (Figs 1c and 2). By 10 h, 2 h after the ending of the O3-exposure, continuing H2O2 accumulation was clearly evident on the outer surface of plasma membranes and also weakly in the cell walls of the palisade parenchyma cells (Figs 1d and 2). The response of spongy mesophyll cells continued as described for 2 h (Fig. 2). Staining was also observed in the cell walls between the palisade parenchyma and upper epidermal cells, in the epidermal cells, and in their cell walls (Figs 1e and 2). In Fig. 1(e), H2O2 production is visible on the plasma membrane and cell wall of palisade parenchyma cell from where the H2O2 seems to be diffusing into the epidermal cell above.

Figure 2.

Schematic representation of the spreading of H2O2 production in birch leaf cells challenged with 150 nL L−1 of O3 for 8 h.

At 2 h, i.e. during exposure, H2O2 accumulates on the plasma membranes and cell walls of both spongy mesophyll and palisade parenchyma cells. At 10 h, 2 h after the ending of the exposure, spongy mesophyll cells respond similarly to after 2 h. In palisade parenchyma cells, however, the response is stronger and, in addition to the plasma membrane and cell wall located accumulation, H2O2 seems to be diffusing into the epidermal cells above. At 24 h, palisade parenchyma cells still accumulate H2O2 on their plasma membranes and cell walls, but the epidermal cells are not involved any more. In spongy mesophyll cells, H2O2 accumulates in the cytoplasm, mitochondria and peroxisomes, but not in the chloroplasts. DPI = diphenylene iodonium. Red symbols (¦–-) indicate the site of inhibition by the mentioned reagent.

At 24 h, H2O2 accumulation continued on the plasma membranes and cell walls of the palisade parenchyma cells (Fig. 2). At this time point, in spongy mesophyll cells, H2O2 accumulated in the cells. CeCl3 precipitates were detectable in the mitochondria, peroxisomes and cytoplasm of the spongy mesophyll cells (Figs 1f and 2). At 48 h, H2O2 accumulation was still continuing in the mitochondria and cytosol of both spongy mesophyll and palisade parenchyma cells, but not in the cell wall or plasma membrane (data not shown). At the same time, when H2O2 production was shifting from the cell wall to the protoplasm and especially to the mitochondria, disintegration of the mitochondrial matrix was visible. In the O3-exposed plants, 45% of the mitochondria (out of 247 mitochondria scored from 45 to 55 cells) had disintegrated matrix at 24 h, when in clean air grown plants only 15% of the mitochondria had a similar appearance at 0 h.

It has been suggested that O3 induces oxidative stress in the chloroplasts (Pell et al. 1997). In our experiments neither structural changes nor H2O2 accumulation was detected in the chloroplasts (Fig. 1f,g). To confirm that CeCl3 had penetrated the cells and particularly the chloroplasts, and that chloroplastic H2O2 production can be visualized with the CeCl3 staining, leaf sections were infiltrated with methyl viologen that sets off ROS production in the chloroplasts. In these leaves, H2O2 accumulation was visible in the chloroplast’s stroma of the spongy mesophyll cells (Fig. 1i). These leaves also displayed severe ultrastructural damage (data not shown). The appearance of O3- and methyl viologen-induced H2O2 in the cytoplasm, mitochondria, peroxisomes and chloroplasts shows that CeCl3 obviously penetrates biological membranes and can be used to detect intracellular H2O2 production in plant cells in a similar way as in mammalian tissues, where cerium-derived deposits have been detected inside subcellular structures, including mitochondria (Slezak et al. 1995).

Inhibition of possible H2O2 sources in O3-treated birch leaves

To elucidate the possible sources of the H2O2 accumulation in O3-treated birch, leaves were exposed first to 150 nL L−1 of O3 for 8 h and infiltrated with inhibitors of possible sources for radical production in the cell wall and subsequently with CeCl3, as described in Bestwick et al. (1997). Inhibitor effects were analyzed from the palisade parenchyma cells. Results from these inhibitor experiments presented in Table 1 show that precipitation of CeCl3 is H2O2-specific and that both cell wall peroxidases and the plasma membrane NADPH oxidase are possible sources for the H2O2. Catalase removed the staining almost completely, confirming that the precipitate detected is derived from H2O2. DPI, an inhibitor of flavin containing oxidases (Cross & Jones 1986), such as the plasma membrane NADPH oxidase, reduced the percentage of H2O2 producing cells from 70 to 52 and abolished the strong staining completely. KCN and NaN3 were used to inhibit peroxidases (Bestwick et al. 1997). KCN was roughly as effective in inhibiting the O3-induced H2O2 production as DPI, whereas NaN3 abolished H2O2 accumulation detectable with the CeCl3 staining almost completely.

Table 1. . The effect of inhibitors on the cell wall and plasma membrane-localized H2O2 accumulation in O3-exposed birch leaves
Intensity of CeCl3 staining (% of scored cells)
TreatmentStrongMediumFaintNo
  1. Leaves were infiltrated with catalase (100 U mL−1), 8 μm DPI (inhibitor of flavin-containing oxidases), and peroxidase inhibitors 1 mm potassium cyanide (KCN) and 1 mm sodium azide (NaN3). The intensity of CeCl3 staining was estimated from 34 palisade parenchyma cells and classified into categories: strong (as in Fig. 1e), medium (as in Fig. 1d) and faint (as in Fig. 1c).

O3 + buffer11263429
O3 + catalase03196
O3 + DPI014158
O3 + KCN0163054
O3 + NaN305491

Discussion

O3 induces H2O2 accumulation in birch leaves

Schraudner et al. (1998) have recently shown that O3 induces early bursts of active H2O2 production in the cell walls of the O3-sensitive tobacco cultivar Bel W3. The number and distribution of these oxidative bursts correlated with the discrete sites of local cell death and visible symptoms that developed later. Superoxide production has been shown similarly in O3-exposed Arabidopsis thaliana (Kettunen et al. 1999; Overmyer et al. 1998; Rao & Davis 1999). The oxidative burst is one of the earliest events in plant–pathogen interactions. It has recently been shown with H2O2-specific CeCl3 staining that the oxidative burst takes place only in the immediate vicinity of the invading pathogen (Bestwick et al. 1997; Bestwick et al. 1998). The subcellular location of the O3-induced H2O2 and superoxide production, however, has not been analyzed before. Here we show that O3 induces active H2O2 production in birch leaves in two separate phases and in different subcellular locations. The first phase of the H2O2 accumulation is located in the cell wall and on the plasma membrane and the second phase in the cytoplasm, peroxisomes and mitochondria, but not in chloroplasts (Figs 1 and 2).

The first apoplastic phase began shortly after the onset of the O3-exposure and was apparent in both spongy mesophyll and palisade parenchyma cells. We have shown previously that in birch, O3 damage occurs mostly in the palisade parenchyma cells (Tuomainen et al. 1996). The second phase was the intracellular production and accumulation of H2O2 that took place mainly in the cytoplasm, peroxisomes and mitochondria at 24 h, at the same time when the visible damage, clearly evident at 48 h, started to develop. Continuing H2O2 accumulation in birch leaves that can be visualized with DAB (3,3-diaminobenzidine)-staining (Thordal-Christensen et al. 1997) takes place only at the locations that later develop macroscopically visible necrosis (R. Pellinen et al. in preparation). However, as a response to O3, several genes are also activated in the absence of visible damage (M. Korhonen et al. in preparation). This has also been shown in Arabidopsis, where O3 caused differential damage formation, ROS production and gene expression patterns in wild-type, O3-sensitive (Cvi-0) and NahG plants (Rao & Davis 1999). In birch, the cell death-related active H2O2 production that continues in the absence of O3 and leads to H2O2 accumulation that is visualized with CeCl3 (Fig. 1) and DAB staining (not shown), and can be prevented with inhibitors, is triggered only in distinct locations and the O3 response threshold level appears to be higher than the one needed for the increase in transcript levels. This is reminiscent of plant–pathogen interactions where different levels and kinetics of H2O2 production activate different responses; higher concentrations and biphasic ROS production is required for the HR (Bolwell & Wojtaszek 1997; Lamb & Dixon 1997).

Possible sources for the apoplastic H2O2 production

Several different mechanisms for the apoplastic ROS generation have been presented (Bolwell & Wojtaszek 1997). The role of two of these, the plasma membrane NADPH oxidase associated with an extracellular superoxide dismutase (EC-SOD) and extracellular peroxidases, was addressed here with inhibitor experiments (Table 1), which suggest that both are involved in the H2O2 production in the cell walls of O3-exposed birch leaves. The plasma membrane located NADPH oxidase is activated by pathogen or elicitor challenge in various plant systems (Bolwell & Wojtaszek 1997). The superoxide produced by NADPH oxidase has to dismutate very rapidly to H2O2 either spontaneously or by the activity of the EC-SOD (Jabs et al. 1997; Streller & Wingsle 1994).

Ogawa et al. (1997) have proposed a model where an extracellular CuZn SOD is in tight connection with the plasma membrane NADPH oxidase. We have shown an increase in the total cellular SOD activity in O3-exposed birch leaves at 12 h with a maximum at 24 h (Tuomainen et al. 1996). Furthermore, the increase in SOD activity was higher in an O3-sensitive birch clone than in an O3-tolerant clone. It was proposed that the increase in SOD activity was linked to the damage formation: the more sensitive clone had higher SOD activity, which was probably due to a higher ROS level.

The second possible source of the apoplastic H2O2 production, the extracellular peroxidases, are involved in lignin biosynthesis (Halliwell 1978) and catalyze a reaction where H2O2 is produced at the expense of NAD(P)H (Gross et al. 1977). In birch, the quaiacol peroxidase activity was increased by O3 (Tuomainen et al. 1996). Timing of the increase was similar to the increase in SOD activity and the increase was more pronounced in the sensitive birch clone. Altogether, the previous enzyme activity measurements (Tuomainen et al. 1996) and the inhibitor experiments (Table 1) suggest that both plasma membrane NADPH oxidase, in combination with EC-SOD, and cell wall peroxidases could be responsible for the O3-induced H2O2 accumulation seen on the plasma membranes and cell walls of birch cells. It must be remembered, however, that DPI also has a peroxidase inhibiting activity, especially at high concentrations (Bolwell et al. 1998). However, the concentration used in this study, 8 μm, reduced NADPH oxidase dependent H2O2 production in rose cells to less than 25% of the original, when for inhibition of H2O2 production by horseradish peroxidase to 75% of the original, 30 μm DPI was required (Bolwell et al. 1998). Thus, the ability of DPI to partly inhibit peroxidases does not exclude the possibility for involvement of NADPH oxidase as a source of H2O2 in O3-treated birch. This is also substantiated by two different locations for the apoplastic H2O2 accumulation detectable with the CeCl3 staining; in Fig. 1(c,d) CeCl3 precipitation is clearly visible in both internal cell wall regions and on the surface of the plasma membrane. Furthermore, the CeCl3 precipitates on the plasma membrane have a distinct spatial pattern which suggests the presence of a single origin, presumably the oxidase protein complex, for each precipitate (see arrows in Fig. 1c,d). It must be kept in mind, however, that the inhibitors used are not specific and that they may have H2O2 scavenging properties as reported in several recent papers (see Barceló 1998). KCN, NaN3 (Barceló 1998) and DPI (Baker et al. 1998) have an ability to scavenge H2O2 in addition to their inhibitor activity. However, it has been shown in other systems, by combining localization, enzyme activity and inhibitor studies, that the two most likely candidates for stress-induced H2O2 production are plasma membrane NADPH oxidase and cell wall peroxidases (Bolwell et al. 1998). Although the results from the inhibitor experiments (Table 1) must be interpreted in conjunction with the localization of the H2O2 production (Fig. 1) and with enzyme activity measurements (Tuomainen et al. 1996), they provide a good system in dissecting the subcellular sites for ROS production, thus leading the way to more specific studies. Furthermore, the inhibitory effect of DPI on NO (nitric oxide) production and the interaction of NO and ROS production (Delledonne et al. 1998) requires more attention.

Intracellular H2O2 production

In this study we did not see any O3-induced H2O2 accumulation in the chloroplasts. This is probably due to the efficient H2O2 scavenging mechanisms in the chloroplasts of these O3-exposed plants. Accordingly, in chronic O3-stress, overexpression of SOD in the chloroplast reduced leaf damage in tobacco without any increase in H2O2 scavenging activity (Van Camp et al. 1994). It was concluded that in tobacco, chloroplastic SOD activity could be the rate-limiting enzyme for ROS detoxification and that the H2O2 scavenging activity was sufficient for detoxification of the hydrogen peroxide produced by SOD. Methyl viologen-induced superoxide production in the chloroplasts, on the other hand, resulted in H2O2 accumulation (Fig. 1i) that seemed to exceed the detoxification capacity in chloroplasts, since increased CeCl3 precipitation was visible in the chloroplast stroma.

Under high O3-peaks the cytoplasmic radical scavenging systems seem to be more important (Pitcher & Zilinskas 1996) than the chloroplastic ones (Tepperman & Dunsmuir 1990; Torsethaugen et al. 1997) in protecting plants from O3. Recent data also suggest that peroxisomal catalase is a sink also for cytoplasmic H2O2, in addition to the H2O2 produced during photorespiration in peroxisomes (Willekens et al. 1997). This detoxification mechanism seems to be crucial for defense against oxidative stress. The H2O2 accumulation in the cytoplasm, peroxisomes and mitochondria, visible in Fig. 1(f), suggest that peroxisomes (and catalase therein) could be a sink for the O3-induced H2O2 produced in other compartments of the cell, but not for the apoplastic H2O2. Intensity of H2O2 staining in the peroxisomes was lower than in the surrounding cytoplasm, which would be seen if catalase scavenges cytoplasmically synthesized H2O2 in the peroxisomes forming a concentration gradient.

The timing of mitochondrial H2O2 accumulation coincided with the first symptoms of visible damage formation and, at the same time, 45% of the mitochondria showed disintegration of the matrix. Similar changes have also been seen in pathogen infected parsley cells, where the mitochondria were swollen and their tubuli were disintegrated (Naton et al. 1996). A close correlation has been observed between the loss of mitochondrial membrane integrity, ROS accumulation and the death of the infected cells. It has been suggested that in mammals the mitochondria are a major source of ROS in pcd (Jabs 1999) when the mitochondria go through a reduction in transmembrane potential and uncoupling of electron transport from ATP synthesis. The latter causes O2 production and subsequent oxidation of mitochondrial structures.

The role of mitochondria in oxidative stress has not been studied that extensively in plants. Banzet et al. (1998) have shown that oxidative stress induces accumulation of heat shock proteins, especially the mitochondrial HSP22 in tomato cell cultures. Kiiskinen et al. (1997) showed that the transcript levels for the mitochondrial phosphate translocator (Mpt1) were significantly increased by O3 in birch. These responses may not be directly connected with defense against oxidative stress, but rather reflect changes in the oxidative balance of the cell that affect mitochondrial metabolism and the homeostasis of the whole cell. The relationship of these events in plants to mitochondrial permeability transition, subsequent hypergeneration of O2 radicals and following pcd, as in animal cells (Majima et al. 1998), remains to be elucidated.

Experimental procedures

Plant material and treatments

One-year-old birch (Betula pendula Roth) plants were grown and exposed to O3 as described previously (Kiiskinen et al. 1997; Tuomainen et al. 1996). Out of 24 individuals exposed, the eight most severely damaged ones were used for localization of H2O2 production. Fully expanded leaves were collected before (0 h), during (2 and 6 h), and after (10, 24 and 48 h) the single 8 h O3-exposure from the upmost third of the plant, where O3-damage appeared most severely, and were used for the localization of H2O2 production by CeCl3 staining. Methyl viologen was used to induce ROS production in the chloroplasts. Leaves were collected from 1-year-old birch saplings grown in a greenhouse for 2 months in ambient light (22 : 2, light:dark) and temperature conditions (20–25°C). Leaf sections cut from the upmost third of the plant were first infiltrated under vacuum with 5 mm CeCl3 (Sigma, St. Louis, MO, USA) in 50 mm MOPS (Sigma) (pH 7.0), containing 5 μm methyl viologen (Sigma) and placed on 0.8% agar plates containing the same reagents. Sections were then illuminated under 200 μmol m−2 s−1 for 2 h and fixed.

Subcellular localization of H2O2 accumulation by CeCl3 staining

Leaf sections were fixed for transmission electron microscopy with 2.5% glutaraldehyde in 0.1 m Na-phosphate buffer (pH 7.0), first under vacuum and then at +4°C overnight. Samples were then washed with buffer and stored at +4°C. The histochemical localization of H2O2 accumulation by CeCl3 staining and subsequent pre-fixation was performed as described by Bestwick et al. (1997) except that the sections were kept under vacuum until they were fully infiltrated. Cerous ions (Ce3+) react with H2O2 forming electron dense cerium perhydroxide precipitates (Slezak et al. 1995):

H2O2 + CeCl3→ Ce(OH2)OOH

The CeCl3-treated sections and control sections (without staining) were post-fixed in 1% osmiumtetroxide (EMS, Washington, PA, USA), dehydrated in ascending ethanol series and embedded in Epon LX 112 (Ladd Research Industries Inc., Williston, VT, USA) and polymerized. Blocks were sectioned (60 nm) on a Reichert Ultracut microtome using a diamond knife (Diatome, Bienne, Switzerland) and mounted on copper slot grids (2 × 1 mm). Sections were examined with a transmission electron microscope (Jeol JEM-1200EX, Joel Ltd, Tokyo, Japan) at an accelerating voltage 60 kV.

Inhibitor treatments

Twelve 2-month-old birch saplings were exposed to O3 as described above. Fully developed leaves were collected at 0, 10 and 24 h and sectioned. Sections were first vacuum infiltrated with buffer alone (50 mm Mops pH 7.2) or with buffer containing NADPH oxidase and peroxidase inhibitors, as described by Bestwick et al. (1997), for 30 min at RT. H2O2 accumulation was analyzed from palisade parenchyma cells.

Acknowledgements

We would like to thank the Electron Microscopy Unit of the Institute of Biotechnology, University of Helsinki for providing laboratory facilities. This work was supported financially by the Academy of Finland (grant 37995 to J.K.), Biocentrum Helsinki, and the University of Kuopio graduate program in Environmental Physiology, Molecular Biology and Ecotechnology.

Ancillary