Activation of an oxidative burst is a general feature of sensitive plants exposed to the air pollutant ozone


  • H. Wohlgemuth,

    1. Institute of Biochemical Plant Pathology, GSF National Research Center for Environment and Health, Ingolstädter Landstr. 1, D-85764 Neuherberg, Germany,
    Search for more papers by this author
    • *Both authors contributed equally to this study.

  • K. Mittelstrass,

    1. Institute of Biochemical Plant Pathology, GSF National Research Center for Environment and Health, Ingolstädter Landstr. 1, D-85764 Neuherberg, Germany,
    Search for more papers by this author
    • *Both authors contributed equally to this study.

  • S. Kschieschan,

    1. Institute of Biochemical Plant Pathology, GSF National Research Center for Environment and Health, Ingolstädter Landstr. 1, D-85764 Neuherberg, Germany,
    Search for more papers by this author
  • J. Bender,

    1. Institute of Agroecology, Federal Agricultural Research Centre (FAL), Bundesallee 50, D-38116 Braunschweig, Germany and
    Search for more papers by this author
  • H.-J. Weigel,

    1. Institute of Agroecology, Federal Agricultural Research Centre (FAL), Bundesallee 50, D-38116 Braunschweig, Germany and
    Search for more papers by this author
  • K. Overmyer,

    1. Institute of Biotechnology, University of Helsinki, PO Box 56 (Viikinkaari 5 D), FIN-00014 Helsinki, Finland
    Search for more papers by this author
  • J. Kangasjärvi,

    1. Institute of Biotechnology, University of Helsinki, PO Box 56 (Viikinkaari 5 D), FIN-00014 Helsinki, Finland
    Search for more papers by this author
  • H. Sandermann,

    1. Institute of Biochemical Plant Pathology, GSF National Research Center for Environment and Health, Ingolstädter Landstr. 1, D-85764 Neuherberg, Germany,
    Search for more papers by this author
  • C. Langebartels

    1. Institute of Biochemical Plant Pathology, GSF National Research Center for Environment and Health, Ingolstädter Landstr. 1, D-85764 Neuherberg, Germany,
    Search for more papers by this author

Correspondence: Christian Langebartels. Fax: + 49 89 3187 3383; e-mail:


Ozone exposure stimulates an oxidative burst in leaves of sensitive plants, resulting in the generation and accumulation of hydrogen peroxide (H2O2) in tobacco and tomato, and superoxide (O2–•) together with H2O2 in Arabidopsis accessions. Accumulation of these reactive oxygen species (ROS) preceded the induction of cell death, and both responses co-occurred spatially in the periveinal regions of the leaves. Re-current ozone exposure of the sensitive tobacco cv. Bel W3 in closed chambers or in the field led to an enlargement of existing lesions by priming the border cells for H2O2 accumulation. Open top chamber experiments with native herbaceous plants in the field showed that Malva sylvestris L. accumulates O2–• at those sites that later exhibit plant cell death. Blocking of ROS accumulation markedly reduced ozone-induced cell death in tomato, Arabidopsis and M. sylvestris. It is concluded that ozone triggers an in planta generation and accumulation of H2O2 and/or O2–• depending on the species, accession and cultivar, and that both these reactive oxygen species are involved in the induction of cell death in sensitive crop and native plants.


A plethora of abiotic and biotic factors cause environmental stress in plants, producing visible leaf symptoms and potentially impairing growth and productivity of crop and native plant species (Inzé & Van Montagu 1995). Attempted infection of plants by non-pathogens or avirulent strains of pathogens elicits a rapid collapse of host cells in the so-called hypersensitive response (HR) and transcriptional activation of defence genes in both the challenged and surrounding cells (Goodman & Novacky 1994). HR is a form of programmed cell death (pcd); it is dependent on a specific program of gene expression as demonstrated by the existence of various mutants that form spontaneous lesions (McDowell & Dangl 2000; Glazebrook 2001).

One of the earliest events in the HR is a burst of oxidative metabolism leading to the generation of superoxide anion radicals (O2–•) and hydrogen peroxide (H2O2; Lamb & Dixon 1997; Bolwell 1999). Superoxide disproportionates to H2O2 and O2, either spontaneously, or by activity of superoxide dismutase (SOD). Reactive oxygen species (ROS) are not only directly protective against invading micro-organisms, but they also drive oxidative cross-linking of cell wall components, and induce an array of protective genes (Lamb & Dixon 1997; Wojtaszek 1997). In addition, ROS are important components in regulating cell death in the HR, with H2O2 in tobacco, Arabidopsis and soybean and O2–• in Arabidopsis (Levine et al. 1994; Jabs, Dietrich & Dangl 1996; McDowell & Dangl 2000). Recently, H2O2 has also been implied in systemic signalling (Alvarez et al. 1998).

In addition to pathogen attack, a number of abiotic factors, including air pollutants and UV-B radiation, are thought to affect plants by causing the excess accumulation of ROS (Dat et al. 2000; Langebartels et al. 2002). Tropospheric ozone is the leading component of photochemical air pollution and is believed to cause considerable damage in both natural and cultivated plants (Bergmann, Bender & Weigel 1999; Kley et al. 1999). Ozone concentrations in the troposphere have increased four-fold during the last century and show peak values of 100–400 nL L−1 in urban and suburban areas world-wide (Kley et al. 1999). The critical level for ozone in Europe is based on the accumulated exposure over a threshold of 40 nL L−1 (AOT40), and an AOT40 value of 3000 nL L−1*h was calculated to protect crops and native vegetation (Kärenlampi & Skärby 1996).

Average summer ozone concentrations are sufficient to provoke HR-like lesions and/or chlorotic symptoms on distinct sensitive species although the majority of plants do not exhibit visible leaf injury (Guderian 1985; Langebartels et al. 2002). The ‘classical’ ozone symptoms of the biomonitor plant, tobacco Bel W3, occur as sharply defined, dot-like lesions on the adaxial side of the leaf resulting from the death of groups of palisade cells (Guderian 1985). A recent survey of native European plants revealed that several dicot species, including Malva sylvestris L., Rumex crispus L and Rumex obtusifolius L., also respond to elevated ozone levels with HR-like lesions (Bergmann et al. 1999).

Once ozone has accessed the apoplastic fluid surrounding the leaf cells, it is rapidly converted into ROS such as O2–•, H2O2 and singlet oxygen which can further react with components of the cell wall and the plasma membrane (Rao, Koch & Davis 2000). Work in recent years has indicated that, in addition to the direct ROS formation from ozone, this air pollutant does induce an oxidative burst in model plants, resulting in the accumulation of ROS and activation of signalling pathways that overlap with the HR responses (Sandermann et al. 1998; Rao et al. 2000; Langebartels et al. 2002). The ozone bioindicator tobacco Bel W3 responded with biphasic H2O2 accumulation during (phase I) and after the exposure period (phase II), during post-cultivation in ozone-free air (Schraudner et al. 1998). We have postulated that ozone effects are amplified by cellular production of ROS only in this sensitive cultivar, not in its tolerant counterpart Bel B. Similarly, ultrastructural analysis of birch leaf cells revealed that H2O2 accumulation started in the cell wall and the plasma membrane, while appearing in the cytoplasm during later phases (Pellinen, Palva & Kangasjärvi 1999). Both, O2–• and H2O2 accumulation was reported in ozone-treated Arabidopsis accessions Columbia and Cvi-0 (Rao & Davis 1999) while a front of O2–• production preceded lesion formation in the ozone-sensitive Arabidopsis mutant rcd1 (for radical-induced cell death; Overmyer et al. 2000).

These rather isolated findings with three species prompted us to examine whether an ozone-triggered oxidative burst of plant origin is a common feature of ozone-sensitive plants. Here we show that the predominantly accumulating type of ROS differs between species, cultivars and accessions. ROS accumulation is found in ‘burst initiation sites’ in the vicinity of leaf veins. It is also reported that native and crop plant species accumulate ROS under field conditions, and that this amplification reaction is a major factor in ozone-induced plant cell death.

Materials and methods

Growth and treatment of plants in controlled chambers

Tobacco (Nicotiana tabacum L.) cultivars Bel W3 and Bel B were grown as described previously (Schraudner et al. 1998). Tomato (Lycopersicon esculentum Mill.) cultivars were from the following sources: cv. Ailsa Craig and Piedmont (Syngenta, Bracknell, UK), cv. DRK 2003, Trust and Solairo (De Ruiter, Bergschenhock, The Netherlands), cv. Money Maker (Sutton, UK), cv. Roma (Pötschke, Marburg, Germany) and Nikita (CEAM, Paterna, Spain). Seeds of Malva sylvestris L. were obtained from Bornträger & Schlemmer, Offstein, Germany. The plants were grown in potting substrate in pollutant-free air under 14 h light (0600 to 2000 h, photosynthetic photon flux density, PPFD, 120 µmol m−2 s−1, 25 ± 1 °C)/10 h dark (20 ± 1 °C) cycles and 70 ± 5% relative humidity. Four- (M. sylvestris), 6- (tomato) or 9-week-old plants (tobacco) were used for the treatments. Leaves were numbered from the apex of the plant, with leaf 1 greater than 8 (tobacco, Rumex species) or 13 cm (tomato). Analyses were routinely performed with the ozone-sensitive middle-aged leaves.

Arabidopsis thaliana accessions Columbia (Col-0), Cape Verde Islands (Cvi-0), Jelinka (Jl-1), Lanark (Lan-0), Loch Ness (Lc-0), Landsberg erecta (Ler-1), Nossen (No-0), Nieps (Np-0), Slapy (Sap-0), Shokhdara (Sha) and Wassilewskija (Ws-2), were obtained from the Nottingham Stock Centre. Plants were grown in potting substrate under 14 h light/10 h dark cycles as above, at 22 ± 1 °C and 70 ± 5% relative humidity. Ozone levels during pre-cultivation were below 10 nL L−1. Five plants per 5 cm × 5 cm pot were treated before bolting at the age of 4 weeks (Langebartels et al. 2000).

The plants were exposed for 5 h (from 0900 to 1400 h) to ozone concentrations between 150 and 400 nL L−1 (± 20 nL L−1) or to pollutant-free air in plexiglass chambers (0·8 m3) placed in a walk-in growth cabinet (Langebartels et al. 2000). Ozone was generated from dry oxygen by electrical discharge (Model 500 M; Fischer, Meckenheim, Germany) and was monitored continuously by a UV Photometric Analyzer (CSI 3000; Messer-Griesheim, München, Germany). Recurrent ozone exposures of tobacco cv. Bel W3 were for 4 d, 5 h each day, at 100 ± 10 nL L−1, yielding AOT40 values of 500 to 2000 nL L−1*h. Arabidopsis accessions were treated with one or two 5 h pulses on two consecutive days (300 or 400 nL L−1 ozone, AOT40: 1500 to 4000 nL L−1*h). All time points are expressed as hours after the onset of the first exposure. Visible injury was scored at 48 h post-cultivation in pollutant-free air.

Field and open top chamber experiments

Nine-week-old tobacco cv. Bel W3 and Bel B were grown under shading in the nursery of the Institute for 14 d in August 1999. During the exposure period, daily maxima of 40 to 60 nL L−1 were recorded, yielding an overall AOT40 value of 600 nL L−1*h. Seeds of R. crispus L., R. obtusifolius L. and M. sylvestris were obtained from Bornträger & Schlemmer. Plants were grown in soil in the greenhouse for 3 weeks and then exposed in the open top chambers (26 m3) of the Federal Agricultural Research Centre (FAL) Braunschweig (Bergmann et al. 1999). They were initially supplied with a commercial NPK fertilizer (Wuxal, Aglukon, Düsseldorf, Germany) corresponding to 140 kg N ha−1. The ozone dispensing and monitoring system has been described by Weigel & Jäger (1988). Two treatments were applied: non-filtered air (AOT40: 3820 nL L−1*h) and non-filtered air plus ozone (1·7 × ambient, AOT40: 11480 nL L−1*h). All plants were monitored daily for visible injury and the day of first symptom appearance was recorded. Lesions were scored as the percentage of leaf area independently by two persons.

Histochemical detection of superoxide anions and hydrogen peroxide

Hydrogen peroxide accumulation was analysed using 3,3′-diaminobenzidine 4 HCl (DAB) according to Schraudner et al. (1998). Superoxide anion radicals were detected by nitroblue tetrazolium (NBT) staining according to Jabs et al. (1996). Middle-aged tobacco, Rumex (1 leaf half), M. sylvestris leaves or tomato leaflets (1–2) were immersed (abaxial side up) in 200–300 mL staining solution (0·1% (w/v) DAB, 10 mm 2-(N-Morpholino)ethanesulphonic acid (MES), pH 6·5, or 0·1% (w/v) NBT, 10 mm sodium azide, 50 mm potassium phosphate, pH 6·4) in a desiccator. Infiltration was carried out by building up a vacuum (∼100–150 mbar, for about 1 min) and releasing it twice or three times until the leaves were completely infiltrated. Incomplete infiltration can lead to NBT staining in the non-infiltrated areas. Arabidopsis rosettes were infiltrated in 30–40 mL staining solution in a 50 mL syringe, by pushing and pulling the syringe.

The incubation was for 45 min under lab light (DAB) or for 20 min in the dark (NBT, absolutely necessary). When indicated, catalase (100 U mL−1, EC, bovine liver, Sigma C9322) or SOD (240 U mL−1, EC, horseradish, S4636) were added to the infiltration buffer prior to DAB or NBT staining, respectively. Leaves were destained with 96% (v/v) ethanol, under heating at 40 °C and were stored in 50% (v/v) ethanol (in the dark for NBT). DAB staining was visualized as a red-brown colour, NBT staining by blue formazan formation. To study whether ROS accumulation was due to enzymatic production, the inhibitors diphenylene iodonium (DPI; 2·5–10 µm) and K252a (0·5 µm, Biomol, Hamburg, Germany) were vacuum infiltrated before the ozone treatment. The SOD activity was determined in in-gel assays (Van Camp et al. 1994). Unless stated otherwise, all chemicals were of analytical grade and were purchased from Sigma, Deisenhofen, Germany.

Cross-sections of tobacco and M. sylvestris leaves were treated with DAB as above or 10 µg mL−1 dihydro-2′4,5,6,7,7′-hexafluorofluorescein (H2HFF) coupled to bovine serum albumin (OxyBURST Green H2HFF BSA; Molecular Probes, Eugene, OR, USA) in 10 mm MES buffer, pH 6·5. Fluorescence microscopy was performed with a Zeiss Axioskop with excitation at 488 nm and an emission filter of > 515 nm.

Statistical analysis

All experiments were performed at least twice with three replicate samples per test condition. The data shown are from representative experiments. When indicated, the Tukey multiple range test was used to test for differences among treatment means (at P < 0·05; Statgraphics software; STSC Inc. Rockville, MD, USA).


Spatial occurrence of the oxidative burst and cell death in ozone-treated tobacco

The production of H2O2 was monitored in the ozone-sensitive tobacco cultivar Bel W3 by 3,3′-diaminobenzidine (DAB) staining which yields a brown precipitate produced by peroxidase-catalysed reaction of the dye with H2O2 (Schraudner et al. 1998). Plants accumulated H2O2 after 8 h and developed foliar lesions between 15 and 48 h after the onset of a 5 h pulse treatment with 180 nL L−1 ozone (Fig. 1). As is typical for ozone injury, spot-like lesions occurred predominantly on middle-aged leaves (no. 4 and 5 counted from the apex). Sites of H2O2 accumulation coincided with the pattern of leaf injury, and occurred mainly in the periveinal regions of the second-, third- and fourth-order veins (Fig. 1e & f). Leaf no. 2 showed visible injury only in the tip region where the cells are oldest (Guderian 1985). Accordingly, H2O2 accumulation was evident only in this part of the leaf (Fig. 1c & d). Young leaves of cv. Bel W3 (no 1 and younger; Fig. 1a & b) as well as all leaves of the tolerant counterpart Bel B (Schraudner et al. 1998) did not show any ROS accumulation nor visible injury under similar experimental conditions. As reported by Schraudner et al. (1998), O2–• accumulation detected by nitroblue tetrazolium (NBT) stain was minor in both cultivars.

Figure 1.

Leaf injury and localization of H2O2 production in leaves of ozone-treated tobacco Bel W3. Plants were treated for 5 h with 180 ± 20 nL L−1 ozone and were then post-cultivated in pollutant-free air. Lesions after 48 h (a, c, e) and DAB staining (b, d, f) after 8 h is shown for leaves 1, 2 and 4 from the apex, leaf 1 being larger than 8 cm. (g), (h), Cross-sections of ozone-exposed middle-aged tobacco leaves were stained with DAB (g) or hexafluorofluorescein (H2HFF) covalently linked to bovine serum albumin (h). Leaves or cross-sections were vacuum-infiltrated with 0·1% (w/v) DAB or H2HFF–BSA (10 µg mL−1) in 10 mm MES, pH 6·5, incubated for 15 (H2HFF) or 30 min (DAB) and cleared in 70% (v/v) ethanol (on the slide for sections). Bars in (g), (h) represent 100 µm.

DAB staining of cross-sections of middle-aged leaves revealed that the palisade parenchyma cells were the main site of H2O2 accumulation whereas the spongy mesophyll and epidermal cells showed reactions only at later stages of lesion formation (Fig. 1g). To detect the extracellular release of ROS, we used the fluorescent dye dihydro-2′,4,5,6,7,7′-hexafluorofluorescein covalently linked to bovine serum albumin. As shown in Fig. 1h, H2O2 accumulation occurred extracellularly in the apoplastic fluid of palisade cells.

Ozone-induced oxidative burst and cell death in tomato

We examined ozone-triggered ROS accumulation by DAB and NBT staining in nine tomato cultivars, differing in their sensitivity to ozone. Figure 2 shows that ozone treatment markedly increased the extent of DAB staining in eight cultivars. Accumulation of H2O2 was highest in the most sensitive cultivars, Solairo, Roma and Nikita, and comprised approximately 20% of the leaf area. As an exception, H2O2 accumulation was detected also in the control plants of cultivar DRK 2003. However, it did not increase further upon ozone treatment (Fig. 2).

Figure 2.

Hydrogen peroxide accumulation in leaves of ozone-treated tomato cultivars. Plants from nine tomato cultivars were treated with 200 nL L−1 ozone (black bars) or ozone-free air (grey bars, control) for 5 h. DAB staining was performed after 8 h (during post-cultivation in pollutant-free air) as described in the legend to Fig. 1, and is expressed in percentage leaf area. Means (n = 4) from ozone-treated plants with the same letter are not significantly (P < 0·05) different according to Tukey’s multiple range test.

The H2O2 accumulation after 8 h and visible injury after 24 h took place nearly exclusively in the periveinal regions of the first- and second-order veins in all but one of the tomato cultivars examined. This result is exemplified for cv. Ailsa Craig and Nikita in Fig. 3. Cultivar Trust showed a different behaviour with spot-like H2O2 accumulation and lesions without preference for the vasculature (Fig. 3e & f). Faint NBT staining occurred over the entire leaf surface of all cultivars, but it did not correlate with the sites of lesion formation (data not shown). Superoxide may dismutate to H2O2, either spontaneously or catalysed by SOD activity. However, no differences could be detected in the intensity and the banding pattern of SOD isoforms in ozone-treated and control plants of cv. Ailsa Craig, DRK2003, Nikita and Roma (data not shown).

Figure 3.

Lesion formation and accumulation of reactive oxygen species in ozone-exposed tomato and Arabidopsis. Tomato plants were treated for 5 h with 200 nL L−1 ozone while Arabidopsis accessions received 400 nL L−1 ozone for 5 h each on two consecutive days. Sha and No-0 were exposed to one or two pulses of ozone (300 nL L−1, 5 h), respectively. Lesions 48 h after the onset of exposure. Leaves were detached after 8 (tomato, Sha), 32 or 34 h (Arabidopsis) and were infiltrated with DAB or NBT as described in Materials and Methods. The presence of brown DAB or blue formazan precipitate indicates H2O2 or O2–• accumulation, respectively.

Ozone-induced oxidative burst and cell death in Arabidopsis

In contrast to the Solanaceae species tobacco and tomato, O2–• was the main reactive oxygen species that accumulated in Arabidopsis accessions as a response to ozone. This was demonstrated by the blue formazan precipitate formation from NBT (Figs 3 & 4a), which indicates the sites and extent of O2–• accumulation. Ozone exposure (400 nL L−1, 5 h) for 2 d led to increases of the NBT-stained area of middle-aged leaves in all accessions except Jl-1 and Lan-0 where control levels were already high (Fig. 4a). The accessions Shokhdara (Sha; Fig. 3) and Cape Verde Islands (Cvi-0, data not shown) were most sensitive and reacted with rapid and nearly complete injury of middle-aged leaves under the above conditions. Figure 4a shows that O2–• accumulation in these accessions comprised as much as 40–55% of the leaf area. Interestingly, significant O2–• accumulation could also be detected in the control plants of Cvi-0. Co-inoculation with SOD markedly reduced O2–• accumulation in all but one accession (Sap-0; Inset in Fig. 4a) confirming the involvement of superoxide in formazan formation.

Figure 4.

Superoxide anion and hydrogen peroxide accumulation in Arabidopsis accessions. Plants were treated with 400 nL L−1 ozone (black bars) or ozone-free air (control, grey bars) for 5 h on two consecutive days and were then post-cultivated in pollutant-free air. Leaves were detached after 8 (Cvi-0, Sha) or 32 h and were infiltrated with NBT (a) or DAB (b) as described in Materials and Methods. The presence of brown DAB or blue formazan precipitates indicates H2O2 or O2–• accumulation, respectively. Bars represent ± SE, n= 4. Means from ozone-treated plants with the same letter are not significantly (P < 0·05) different according to Tukey’s multiple range test. Insets: the specificity of NBT and DAB staining was verified by infiltrating ozone-exposed leaves with NBT in combination with superoxide dismutase (SOD, horseradish, 240 U mL−1) or DAB in combination with catalase (bovine liver, 100 U mL−1). Treatments are shown as open bars, controls as shaded bars.

The H2O2 accumulation, visualized by DAB staining, was minor in most of the Arabidopis accessions (Figs 3 & 4b). The addition of catalase reduced DAB staining in all but one accession (Jl-1; Inset in Fig. 4b), hence, DAB staining generally reflected H2O2 accumulation. It was interesting to note that the sites of O2–• and H2O2 accumulation differed between the accessions. They coincided in the O2–• and H2O2 accumulating accession Shokhdara. In contrast, O2–• accumulated at the margins, whereas H2O2 was detected in the central and basal part of middle-aged leaves of Np-0 (Fig. 3). Microscopic analysis of No-0 revealed that H2O2 accumulation mainly occurred in the vicinity of the major leaf vein whereas O2–• accumulation sites scattered over the leaf surface, and were found particularly in the vicinity of trichomes (Fig. 3t & u).

The potential function of the main ROS occurring in Arabidopsis (O2–•) and tomato (H2O2) in subsequent cell death was investigated using diphenylene iodonium (DPI), an inhibitor of NAD(P)H oxidase and other flavin-containing oxidases and K252a, a general protein kinase inhibitor. Pre-treatment of Sha with 10 µm DPI reduced ozone-induced leaf injury and O2–• accumulation in parallel (Fig. 5a). Furthermore, DPI and K252a caused marked reductions in H2O2 accumulation and cell death in the sensitive tomato cultivar Roma (Fig. 5b).

Figure 5.

Role of endogenous ROS production in ozone-induced lesion formation in Arabidopsis (a) and tomato leaves (b). Leaves were infiltrated with a blunt-tip syringe containing diphenylene iodonium (DPI, 10 or 2·5 µm, respectively), an inhibitor of flavin-containing oxidases, or K252a (0·5 µm), a general inhibitor of protein kinases. 3 h later, plants were exposed to 200 (tomato) or 400 nL L−1 ozone (Arabidopsis) for 5 h. Leaves were infiltrated after 8 h with DAB or NBT as above; visible injury was assessed after 48 h.

Ozone-induced ROS accumulation in field-grown plants

We examined whether ROS accumulation in plants also occurs under ambient ozone levels in the field. Tobacco plants were exposed in August 1999 under typical biomonitoring conditions, that is, under shading and for an experimental period of 2 weeks (Guderian 1985). Starting after 4 d and at maximum ozone levels of 40 to 60 nL L−1 (AOT40: 600 nL L−1*h), typical pergament-like lesions were visible on middle-aged leaves (Fig. 6a). These lesions increased in size during the exposure period. When H2O2 accumulation was analysed by DAB staining, a very similar pattern was observed. Apart from background staining, H2O2 accumulation preferentially occurred at the borders of the existing lesions (Fig. 6b & c). As ozone exposure and H2O2 accumulation cannot be resolved temporally in the field, we used closed chambers to test whether daily exposure to ambient ozone would lead to similar patterns of H2O2 accumulation and cell death. As expected, the leaf area covered with ozone-induced lesions increased bet-ween 1 and 4 d of exposure. However, the area of H2O2 accumulation, which was again found predominantly around the existing lesions, did not change significantly during the exposure period of 4 d (data not shown). These results suggest that the existing lesions expand by triggering ozone-induced H2O2 accumulation in the neighbouring cells.

Figure 6.

Lesion formation and accumulation of reactive oxygen species in tobacco and native plant species in the field. (ac), Nine-week-old tobacco plants pre-cultivated in climate chambers were exposed under shading in the nursery of the Institute in August 1999. Daily ozone maxima were between 40 and 60 nL L−1. (a), lesions on middle-aged leaves and (b, c), DAB staining of leaves after 4 d exposure. Leaves were harvested at 1600 h, during the daily ozone maximum. (d–f), Four-week-old M. sylvestris were exposed to 150 nL L−1 ozone for 5 h on two consecutive days in controlled climate chambers. (d), lesions after 48 h, (e, f), NBT staining after 8 h. (g–i), Lesions and ROS accumulation on middle-aged leaves of M. sylvestris plants after 4 weeks of exposure to 1·7 × ambient ozone levels in open top chambers. Leaves were harvested and infiltrated with NBT (h) or DAB (i) at 1500 h on a day with maximum ozone values of 117 nL L−1. (j–l), Ozone-induced anthocyanin formation (j), as well as NBT (k) and DAB staining (l) on leaves of Rumex obtusifolius after 4 weeks of exposure to 1·7 × ozone in open top chambers. Leaves were infiltrated with NBT (k) or DAB (l) at 1500 h on a day with maximum values of 117 nL L−1. (m) NBT staining of a cross-section through an ozone-treated M. sylvestris leaf. Ozone exposure and staining was performed as in (d–f). (n), (o), Effects of DPI inhibition (7 µm) on O2–• accumulation in ozone-treated M. sylvestris. (n), control and (o), DPI-treated leaf.

As shown by Bergmann et al. (1999), several native European plant species are similar to tobacco Bel W3 in their ozone sensitivity. Among these, M. sylvestris as well as Rumex species respond to elevated ozone levels with characteristic symptoms. It was therefore asked whether an oxidative burst of ROS accumulation is also found in these species under controlled conditions and in the field. Ozone exposure of M. sylvestris in closed chambers led to visible symptoms as well as O2–• accumulation preferentially in clusters of 200–300 cells along the leaf veins (Fig. 6d–f). No H2O2 accumulation could be detected (data not shown).

The plants were also exposed to ambient and 1·7 × ambient ozone levels for 4 weeks in open-top chambers in the field (Weigel & Jäger 1988). Elevated ozone (AOT40: 11480 nL L−1*h) gave rise to HR-type lesions in M. sylvestris (Fig. 6g). These lesions were typically located in the vicinity of the leaf veins. When non-symptomatic leaves were analysed for ROS accumulation, a spectacular pattern of NBT staining was detected under 1·7 × ambient (Fig. 6h), but not under ambient (AOT40: 3820 nL L−1*h) ozone. Spots of O2–• accumulation (200–400 cells in size) localized to the veins, and they correlated with the special pattern of visible lesions in older leaves. Again, no H2O2 accumulation was detected by DAB staining (Fig. 6i). The O2–• accumulation commenced in the spongy mesophyll cells (Fig. 6m), demonstrating clear differences in type and location of ozone-triggered ROS accumulation between M. sylvestris and tobacco Bel W3.

Rumex crispus and R. obtusifolius did not react with lesion formation under similar experimental conditions, but R. obtusifolius showed marked anthocyanin accumulation in response to the 1·7 × ambient ozone (Fig. 6j). Both, R. obtusifolius (Fig. 6k) and R. crispus (not shown) accumulated O2–• in the periveinal regions in response to 1·7 × ambient ozone. No DAB staining was found in either species. Treatment of M. sylvestris with 7 µm DPI led to marked reductions in O2–• accumulation, again pointing to active ROS production by the leaf cells (Fig. 6n & o).


The ozone-induced oxidative burst correlates with and precedes the HR-type injury in various plant species

In addition to triggering antioxidant defence responses, ROS are involved, particularly at higher concentrations, in the most drastic plant defence response, the suicide of pathogen-invaded cells (Lamb & Dixon 1997). Here we show that an environmental contaminant, the air pollutant ozone, triggers the accumulation of ROS in ozone-sensitive plant species. Our findings provide evidence that these plant-derived ROS are involved in the formation of HR-type lesions found in plants exposed in climate chambers as well as in the field.

As exemplified with the ozone biomonitor plant, tobacco Bel W3, the patterns of ozone-induced H2O2 accumulation, cell death and visible symptoms were identical with regard to tissue and developmental stage specificity (Fig. 1). Initially, tobacco symptoms were described as ‘weather flecks’, and they were found on the adaxial side of the leaf, in the vicinity of first- and second-order veins (Lucas 1975). The air pollutant ozone was identified as the causal agent for these symptoms in a classical paper by Heggestad & Middleton (1959). Figure 1 shows that the sites of H2O2 accumulation in the ozone-sensitive tobacco Bel W3 correlate with visible symptoms in all developmental stages. Ozone injury in tobacco specifically starts in the palisade parenchyma cells (Guderian 1985). Histochemical analysis of cross-sections of leaves revealed that the palisade cells were also the sites of H2O2 accumulation. Using the non-permeable dye hexafluorofluorescein covalently linked to bovine serum albumin, it could be demonstrated that the accumulation initially occurs extracellularly, that is, in the cell wall and the apoplastic fluid surrounding the palisade cells (Fig. 1). These results confirm the subcellular location of ROS production in ozone-exposed birch leaves in which ROS were initially also detected in the cell wall and on the plasma membrane (Pellinen et al. 1999).

A survey of various species revealed that H2O2 and O2–• accumulation, detected by DAB and NBT staining, respectively, appeared well before the HR-type cell death in ozone-sensitive plants. Sites of H2O2 accumulation correlated spatially with those of cell death in nine tomato cultivars, whereas O2–• sites correlated with lesions in M. sylvestris. All assays were carried out 2–3 h after the end of the exposure period which excludes ozone-derived ROS as the cause for DAB and NBT staining. In addition, pre-treatment of tomato, A. thaliana and M. sylvestris plants with DPI, a suicide inhibitor of flavin-containing oxidases such as the superoxide-producing NADPH oxidase, markedly reduced ROS accumulation and cell death in parallel (Figs 5 & 6). Similarly, DPI abolished endogenous ROS production in pathogen-challenged plants (Levine et al. 1994; Jabs et al. 1996). In addition, the protein kinase inhibitor K252a reduced H2O2 accumulation and cell death in tomato (Fig. 5) suggesting an involvement of protein kinases in the ozone-induced oxidative burst (see also Samuel, Miles & Ellis 2000). Taken together with previous results (Schraudner et al. 1998; Rao & Davis 1999; Overmyer et al. 2000), it appears that ozone triggers the endogenous production of ROS in sensitive plants, thus amplifying the direct effects of ozone in discrete tissues and stages of development. In line with these findings, the onset of cell death in several lesion mimic mutants is tissue- and developmental stage-specific (McDowell & Dangl 2000; Glazebrook 2001).

As summarized by Wojtaszek (1997), the oxidative burst in plant–pathogen interactions is often composed of the more stable H2O2, but also O2–• accumulation is found in certain systems. Accordingly, the type of the predominant ROS differed between the species investigated in this study even though they were exposed under similar climate and pollutant regimes, and were all tested with DAB and NBT staining under identical conditions. Whereas the tobacco and all tomato cultivars preferentially accumulated H2O2 and showed only faint NBT staining (Figs 1, 2, 3 & 6; Schraudner et al. 1998), M. sylvestris and two Rumex species exhibited mainly O2–• accumulation (Fig. 6). The sites of ROS accumulation and lesion formation coincided in all these species, suggesting that both types of ROS may be involved in the HR-type cell death, depending on the plant species.

Arabidopsis thaliana with 10 accessions from three continents showed a different behaviour. The prevailing ROS in response to ozone exposure was superoxide, whereas, in addition, four accessions accumulated H2O2. As shown by co-inoculation with SOD and catalase, NBT and DAB staining reflected O2–• or H2O2 accumulation, respectively, in nine of 10 accessions. NBT staining in Sap-0 and DAB staining in Jl-1 was not reduced by the enzyme treatments which may be due to inactivation by inhibitors in these accessions. The most sensitive accessions found in this study showed either O2–• accumulation (Cvi-0; Rao & Davis 1999) or O2–• together with H2O2 (Shokhdara; Figs 3 & 4). The sites, surprisingly, did not coincide in No-0 and Np-0. It could be that different enzymatic sources are responsible for the formation of both types of ROS, being located at different sites in the leaf. The results obtained by Pellinen et al. (1999) using various inhibitors of enzymatic ROS production also pointed to at least two sources for H2O2 production in birch leaves. Alternatively, the levels of SOD activity, converting superoxide into H2O2, or of antioxidant compounds, may exhibit patterns in constitutive or inducible activity and levels over the leaf surface.

At present, it is not clear which mechanism(s) are responsible for the apoplastic ROS accumulation in ozone-treated plants. Work with Arabidopsis and tomato has revealed that ROS production is dependent on a plasma membrane NADPH oxidase (Keller et al. 1998; Amicucci, Gaschler & Ward 1999; Torres, Dangl & Jones 2002). It was implicated by Rao & Davis (1999) that ozone triggers an oxidative burst by activating a NADPH-dependent oxidase in Arabidopsis. Accordingly, recent results from our laboratory show that ozone exposure of tobacco activates two NADPH oxidase homologs at the transcript level (H. Wohlgemuth, C. Langebartels, unpublished results). Alternative mechanisms for apoplastic and peroxisomal ROS production involving amine oxidases, oxalate oxidase and pH-dependent cell wall peroxidases have been proposed (Bolwell 1999; Corpas, Barroso & del Rio 2001). Given the diverse responses of the species reported in this paper, it seems probable that more than one enzymatic source of ROS formation is responsible for the ozone-activated oxidative burst in different species, possibly also in the (distinctly localized) O2–• and H2O2 accumulation in Arabidopsis.

Reactive oxygen species accumulate in ‘burst initiation sites’ in the vicinity of leaf veins

Ozone-induced ROS accumulation was not homogenously distributed over the leaf surface, but concentrated in the periveinal regions in tobacco, tomato, Malva sylvestris and two Rumex species. Since all cells were directly exposed to the stimulus ozone, clusters of periveinal cells (‘burst initiation sites’; Schraudner et al. 1998) may be particularly sensitive to inducing signals or be otherwise predisposed to activate an oxidative burst. Similarly, several responses in plant–pathogen interactions exhibit preference for the veinal system. Rapid death of periveinal cells was observed in the Cf2 resistance gene-mediated localized response of tomato leaves to race-specific elicitors of Cladosporium fulvum (Hammond-Kosack & Jones 1996). ‘Microbursts’ and resultant ‘micro-HRs’ in Arabidopsis were frequently observed adjacent to the leaf veins (Alvarez et al. 1998). It was suggested by Alvarez et al. (1998) that this pattern reflects the greater exposure of periveinal cells to systemic signals and that, when diffusing out of the veins, the concentration is diluted, so that cells near the vascular bundles are more prone to react.

Accumulation of reactive oxygen species in plants exposed to ozone in the field

Until now, studies on ozone-triggered defense responses in plants were only performed in climate chambers and under ozone concentrations exceeding outdoor levels by up to 500% (reviewed in Langebartels et al. 2002). It was therefore asked whether field-grown tobacco or native plant species also accumulated ROS under near-ambient ozone concentrations. Figure 6 demonstrates that ambient summer ozone levels in Germany lead to typical symptoms in the bioindicator tobacco Bel W3. As known for many years, daily recurrent ozone exposure results in the enlargement of existing lesions rather than formation of new spots (Guderian 1985). DAB staining revealed that the borders of the lesions exhibit a ring of H2O2-accumulating cells. This finding suggests that the lesions release signals to the surrounding cells predisposing them for ROS production. It will be interesting to know whether potential amplifying factors such as ethylene, salicylate, NO (Van Camp, Van Montagu & Inzé 1998) or other signals preferentially are produced or accumulate at these sites.

Several native European plants, including M. sylvestris and Rumex species, are as ozone-sensitive as tobacco Bel W3 and exhibit specific leaf injury (HR-type or chlorotic symptoms; Davison & Barnes 1998; Bergmann et al. 1999). As shown in Fig. 6M. sylvestris as well as R. obtusifolius showed marked accumulation of O2–• under 1·7 × ambient ozone levels, but not under ambient ozone or in pollutant-free air. ROS accumulation occurred in symptom-free leaves, and these sites, after 1 to 2 d, turned into lesions in M. sylvestris. As in the case of tobacco Bel W3, the responses were detected at near-ambient ozone levels.

In conclusion, ozone itself or ozone-derived ROS activate an amplification loop of cellular ROS production in sensitive plants. Thereby, ozone erroneously triggers the pathogen-defence pathway leading to HR-type cell death in crop and native plant species. The known ozone activation of all areas of plant defence (Sandermann et al. 1998; Langebartels et al. 2002), now also including the oxidative burst, together with its occurrence during the vegetation period in wide parts of the industrialized world, makes ozone a primary abiotic elicitor and a modulating factor for disease resistance in the field.


We thank Lucia Gößl and Carina Trenkler for expert technical assistance. Critical reading of the manuscript by our colleagues is gratefully acknowledged. This research was supported by grants from Deutsche Forschungsgemeinschaft, the EU-FAIR project TOMSTRESS, Bayerisches Staatsministerium für Landesentwicklung und Umweltfragen, Fonds der Chemischen Industrie, the Academy of Finland (43671) and the Finnish Centre of Excellence Program (2000-05).