The tomato ethylene receptor LE-ETR3 (NR) is not involved in mediating ozone sensitivity: causal relationships among ethylene emission, oxidative burst and tissue damage

Authors


Author for correspondence: Annamaria Ranieri Tel: +39 (0)50 2216605 Fax: +39 (0)50 2216630 Email: aranieri@agr.unipi.it

Summary

  • • The causal relationships among ethylene emission, oxidative burst and tissue damage, and the temporal expression patterns of some ethylene biosynthetic and responsive genes, were examined in the Never ripe (Nr) tomato (Lycopersicon esculentum) mutant and its isogenic wild type (cv. Pearson), to investigate the role played by the ethylene receptor LE-ETR3 (NR) in mediating the plant response to ozone (O3).
  • • Tomato plants were used in a time-course experiment in which they were exposed to acute O3 fumigation with 200 nl l−1 O3 for 4 h.
  • • The pattern of leaf lesions indicated similar sensitivities to O3 for cv. Pearson and Nr. In both genotypes, O3 activated a hydrogen peroxide (H2O2)-dependent oxidative burst, which was also ethylene-driven in Nr leaves. Ozone induced some ethylene and jasmonate biosynthetic and inducible genes, although with different timings and to different extents in the two genotypes.
  • • The overall data indicate that Nr retains partial sensitivity to ethylene, suggesting only a marginal role of the NR receptor in mediating the complex response of tomato plants to O3.

Introduction

Tropospheric ozone (O3) is the major component of photochemical air pollution. As a result of the rapid urbanization and increase in human activities that occurred in many regions of Europe in the last century, the O3 concentration has increased from 2- to 5-fold, reaching peak values of 100–400 nl l−1, which threaten native and cultivated ecosystems (Inclán et al., 1999; Kley et al., 1999).

Short-term exposure to high O3 concentrations causes visible leaf lesions, the appearance of which resembles the hypersensitive response to an avirulent pathogen attack. The similarity between these two processes is not only external. These phenomena share many physiological and molecular features (reviewed in Kangasjärvi et al., 1994; Rao et al., 2000; Rao & Davis, 2001), such as the induction of a biphasic oxidative burst and the activation of a signalling pathway. Recent studies have presented evidence that, in addition to primary radical production in the apoplast through the chemical reactivity of O3 in aqueous solutions, active generation of reactive oxygen species (ROS) may occur in the affected cells, resulting in the activation of signalling pathways that overlap with the hypersensitive response (Pellinen et al., 1999, 2002; Langebartels et al., 2002; Pasqualini et al., 2002; Wohlgemuth et al., 2002; Ranieri et al., 2003; Diara et al., 2005; Kangasjärvi et al., 2005). Although ROS are believed to act as key molecules eliciting the plant response to O3, recent findings underline the importance of the complex interactions between ROS and several phytohormones such as ethylene (ET), salicylic acid (SA) and jasmonic acid (JA) in modulating the plant response to the pollutant (Overmyer et al., 2000; Rao et al., 2000; Rao et al., 2002; Kangasjärvi et al., 2005). ET, SA and JA act as second messengers in oxidative signal transduction and regulate the induction and the spread of oxidative stress symptoms (Moeder et al., 2002; Rao et al., 2002; Vahala et al., 2003). Mutual antagonistic or synergetic interactions between these phytohormones have been reported. In general, ET and SA are believed to amplify the oxidative signal and are needed for the development of lesions, whereas JA contains and limits O3-induced damage, acting as a component of anti-cell death pathways (Overmyer et al., 2000, 2005; Watanabe et al., 2001; Langebartels et al., 2002; Moeder et al., 2002; Tuominen et al., 2004; Kangasjärvi et al., 2005).

Ethylene biosynthesis proceeds via the conversion of S-adenosyl-L-methionine to 1-aminocyclopropane-1-carboxylic acid (ACC). The reaction is catalysed by ACC synthase (ACS), which is encoded by a multigene family and is often considered to be the rate-limiting step in ET synthesis (Kende, 1993). ACC, in turn, is oxidized to ET by ACC oxidase (ACO), which is also encoded by a small family of genes (Kende, 1993) and is considered to be less rigorously controlled (Wang et al., 2002). Modulation of ET action can also occur at the level of perception. Six tomato (Lycopersicon esculentum) ET receptors, designated LE-ETR1, LE-ETR2, LE-ETR3 (NR), LE-ETR4, LE-ETR5 and LE-ETR6, have been isolated and characterized (Klee & Tieman, 2002; Klee, 2004). While several studies have demonstrated the importance of functional ET perception in determining the degree of O3-induced lesion formation, little is known about the roles of the different receptors in modulating the plant response to the pollutant. In tomato, an ET-insensitive mutation (Never ripe, Nr), arising from a single base substitution in the N-terminal coding region of the LE-ETR3 gene (NR), homologous to ETR1 of Arabidopsis (Wilkinson et al., 1995), has been identified as a spontaneous naturally occurring mutation leading to delayed and incomplete fruit ripening (Lanahan et al., 1994).

In the present work, we attempted to elucidate the role played by the ET receptor LE-ETR3 (NR) in mediating the plant response to O3, by evaluating whether and how the Nr mutation influences O3 sensitivity. For this purpose, the causal relationships among ET emission, oxidative burst and tissue damage, as well as the temporal expression patterns of some ET biosynthetic and responsive genes, were assessed in the Nr tomato mutant and its isogenic wild type (cv. Pearson) exposed to a single pulse of 200 ppb O3 for 4 h. Moreover, employing a pharmacological approach involving the use of aminoethoxyvinylglycine (AVG), an inhibitor of ACS activity, we explored whether a block in ET synthesis could also influence O3 sensitivity.

Materials and Methods

Plant material

Sterilized seeds of tomato (Lycopersicon esculentum L. cv. Pearson and the isogenic line Never ripe (Nr), which has a mutation in the ET receptor LE-ETR3) were germinated in a sand–soil mixture for 7–10 d. The seedlings were then transplanted and grown for c. 30 d in a glasshouse at 17/25°C (night/day) and a relative humidity of 60–80% under a 14 : 10 h light:dark photoperiod. Only uniform plants with eight fully expanded leaves were selected. Leaves were numbered from 1 to 8 starting from the plant apex. For all the biochemical, molecular and histochemical analyses, leaflets from fully expanded middle-aged leaves (numbers 3–5) were used. For each genotype, a minimum of three plants per time-point per treatment were used.

Fumigation treatment

O3 fumigation was performed as described in Diara et al. (2005). Briefly, plants preadapted to the chamber conditions for 48 h were exposed to acute fumigation with 200 ppb O3 for 4 h and left to recover in charcoal-filtered air for up to 24 h. Leaves were collected before (0 h), during (0.5, 1, 2 and 4 h) and after (5, 7 and 24 h) exposure to the pollutant. Times of sampling refer to hours after the onset of fumigation. Untreated plants were kept in charcoal-filtered air chambers under the same conditions and used as controls.

AVG treatment

To explore whether a block in ET synthesis could influence O3 sensitivity, 24 h before fumigation cv. Pearson and Nr plants were randomly divided into two groups. Leaves of one group were painted on both the abaxial and adaxial surfaces with 1 mm AVG, an inhibitor of ACS activity, while the other group was treated with H2O and used as control for AVG treatment.

Cell viability measurement

Cell viability was assessed using Evans blue dye, which is usually excluded from cells that have intact membranes. Leaf discs of known area (1.13 cm2) were vacuum-infiltrated for 1 h with 0.25% Evans blue solution with continuous shaking (Schraudner et al., 1998). After infiltration, discs were rinsed with distilled water to remove any excess dye and photographed.

Ion leakage

Cell damage was assessed by measuring ion leakage from leaf discs. Ten discs of known area (1.13 cm2) were incubated with shaking in 5 ml of Milli-Q water (Millipore Corporation, Billerica, MA, USA) for 3 h at room temperature. The conductivity of the incubation medium was recorded by a conductivity meter (Jenway, Dunmow, UK). The results are expressed as percentage of total conductivity, measured after overnight incubation of the boiled discs.

Histochemical detection of hydrogen peroxide (H2O2)

H2O2 accumulation was analysed using 3,3′-diaminobenzidine (DAB) which, upon reaction with H2O2, develops a red-brown colour. Leaflets were vacuum-infiltrated (at −65 kPa; three cycles of 1 min each) with 0.1% DAB in 10 mm 2-(N-morpholino)ethanesulfonic acid (MES), pH 6.5. After 45 min of incubation under ambient light, leaflets were destained with 96% ethanol at 40°C and stored in 50% ethanol (Wohlgemuth et al., 2002). DAB staining was measured as percentage of leaf area and the data were calibrated with a portable area meter (LI300C; Li-Cor, Lincoln, NE, USA).

In situ localization of H2O2 accumulation

H2O2 production was assessed cytochemically via determination of cerium perhydroxide formation after reaction of CeCl3 with endogenous H2O2 (Bestwick et al., 1997). Freshly harvested leaves were cut into slices (1–2 mm2) which were incubated for 1 h in 5 mm CeCl3 in 50 mm 3-(N-morpholino)propanesulfonic acid (MOPS), pH 7.2, fixed for 1 h in 1.25% (weight/volume (w/v)) glutaraldehyde and 1.25% (w/v) paraformaldehyde in 50 mm Na-cacodylate buffer (CAB), pH 7.2, and washed twice in CAB for 10 min. After postfixation for 2 h in 1% (w/v) osmium tetroxide in 50 mm CAB, pH 7.2, samples were washed twice in the same buffer, dehydrated in a graded ethanol series (25, 50, 75, 90 and 100%, v/v), transferred into propylene oxide and gradually embedded in Epon-Araldite. Thin sections of embedded tissues were obtained using a Reichert-Ultracut microtome (Leica, Heidelberg, Germany), mounted on uncoated copper grids and observed using a transmission electron microscope (Hitachi 300; Hitachi, Tokyo, Japan) at 75 kV.

Determination of ACC content

Leaf material was ground in liquid nitrogen and extracted with 2% metaphosphoric acid. Free ACC and total ACC released by acid hydrolysis (2 N HCl for 3 h at 120°C) were oxidized to ET by adding 100 µl of 10 mm HgCl2 and 200 µl of saturated NaClO/NaOH solution (2 : 1, v/v) to appropriate aliquots of the supernatant in a reaction volume of 1 ml. The reaction was carried out within sealed containers and the ET produced was evaluated as described in the next section. The amount of conjugated ACC was calculated by subtracting the amount of free ACC from that of total ACC (Diara et al., 2005).

Ethylene determination

Fifteen minutes after excision, leaves were incubated within sealed containers at room temperature and, after 1 h, 2-ml samples were withdrawn with a hypodermic syringe. Ethylene evolution was measured by injecting samples into a gas chromatograph equipped with a dual flame ionization detector and a metal column (150 × 0.4 cm internal diameter) packed with alumina (70–230 mesh). The column and detector temperatures were 70 and 350°C, respectively. Nitrogen (N2) was used as a carrier at a flow rate of 40 ml min−1 (Mensuali Sodi et al., 1992).

Enzyme extraction and activity

Frozen leaves were ground in liquid nitrogen with 10% (w/w) polyvinylpolypirrolydone (PVPP) and homogenized in 100 mm Tricine-KOH buffer (pH 8.0), 20 mm MgCl2, 50 mm KCl, 10 mm ethylenediaminetetraacetic acid (EDTA), 0.1% (v/v) Triton X-100, 1 mm dithiothreitol (DTT), and 0.50 mm phenylmethylsulfonyl fluoride (PMSF). After centrifugation at 12 000 g for 30 min at 4°C, the supernatant was collected and dialysed overnight against diluted extraction buffer. For the APX assay, 50 mm Na-ascorbate was added to the extraction and dialysis buffers to avoid APX inactivation during the extraction procedure.

APX activity was determined by following the decrease in absorbance at 290 nm (A290) that resulted from the oxidation of ascorbic acid in the first 30 s of the reaction, using the extinction coefficient of 2.8 mm−1 cm−1 for ascorbate. The reaction medium contained 0.5 mm Na-ascorbate, 0.1 mm H2O2, 1 mm EDTA and 0.1 m Hepes-KOH buffer (pH 7.8) (Diara et al., 2005).

The rate of NAD(P)H oxidation was measured by following the decrease in A340 of a reaction medium containing 40 mm Na-acetate (pH 5.5), 250 mm sucrose, 1 mm MnCl2, 100 µm salicylhydroxamic acid, 100 µm NADH and the extract aliquot (Diara et al., 2005). The activity was calculated using the extinction coefficient of 6.22 mm−1 cm−1 for NADH.

The protein content of the extracts was measured spectrophotometrically at 595 nm according to Bradford (1976), using bovine serum albumin as standard.

RNA isolation and analysis

Total RNA was extracted from frozen, homogenized leaf tissue (0.1–0.15 g fresh weight (FW)) of control and O3-treated tomato plants with the NucleoSpin® RNA Plant (Macherey-Nagel, Düren, Germany) according to the manufacturer's instructions. A given amount of total RNA (1–2 µg) was reverse-transcribed for 1 h at 42°C using 200 units of SuperscriptII RT (Invitrogen, Carlsbad, CA, USA) with 1 × corresponding buffer, 10 mm DTT, 0.4 mm of each dNTP and 0.5 µg µl−1 oligo dT (5′-T(25)-VX-3′) primer (Invitrogen). Gene-specific fragments were amplified by semiquantitative reverse transcriptase–polymerase chain reaction (RT-PCR) using 1 unit of Taq polymerase (Amersham Bioscience, Uppsala, Sweden), 1 × corresponding buffer, 0.2 mm of each dNTP and 10 µm of each specific primer (ribosomal protein L2, RPL2; ACC synthase 2, ACS2; ACC synthase 6, ACS6; ACC oxidase 1, ACO1; ACC oxidase 3, ACO3; β-1,3-glucanase, GLUB; allene oxide synthase 2, AOS2; allene oxide cyclase, AOC; proteinase inhibitor II, PINII) (Invitrogen). The primer sequences of all genes analysed, the annealing temperature and the number of cycles are given in Table 1. The cycle number is dependent on the amount of template RNA and the abundance of the target transcript. To ensure that the analysis was conducted in the linear range of the amplification assay, a range of cycle numbers were initially used for each gene, and quantification was conducted at a cycle number within the linear range.

Table 1.  Oligonucleotide primer sequences, polymerase chain reaction (PCR) conditions and number of cycles used for analysis by semiquantitative reverse transcriptase–PCR
GenePrimersPCR annealing temperature (°C)Cycle sample number (forward and reverse)
  1. ACC, 1-aminocyclopropane-1-carboxylic acid.

ACC synthase 2 (ACS2)5′-TTAGCGGCAATGCTATCGGAC-3′6230
5′-AACAAACTCGGAACCACCCTG-3′  
ACC synthase 6 (ACS6)5′-GTTTGGCTGATCCTGGTGATG-3′6030
5′-GTTGACTAAACGCGGTTGCTG-3′  
ACC oxidase 1 (ACO1)5′-CAAGGGACTCCGCGCTCATAC-3′63.430
5′-GGCTTGAAACTTGAGTCCAGC-3′  
ACC oxidase 3 (ACO3)5′-TGGTGAACCATGGGATTCCAC-3′63.728
5′-ACCTCAAGCTGGTCACCAAGG-3′  
1,3-glucanase (GLUB)5′-AATCAGCCCTGTTACTGGCAC-3′63.228
5′-TGTTGCACCAAAAGCACCAGC-3′  
Allene oxide synthase 2 (AOS2)5′-ACGAATATGCCACCGGGACCATTC-3′62.428
5′-TCAATGCCGGACCATCAGTTCCGAG-3′  
Allene oxide cyclase (AOC)5′-TGCAAGAGCCAGAGCACCTCAACAG-3′58.525
5′-ACACAACAATTCAGATGGCAGACC-3′  
Proteinase inhibitor II (PINII)5′-TGTTGATGCCAAGGCTTGTAC-3′56.330
5′-AACCCTTGTACCCTGTGCAACACG-3′  
Ribosomal protein L2 (RPL2)5′-ACAGAGGTGATTCACGATCC-3′5926
5′-TCCAGGAGGTGCATCACGAC-3′  

The amplification steps were 94°C for 75 s, 94°C for 5 s, annealing for 20 s, 72°C for 45 s, and a final extension at 72°C for 7 min. Amplification products were analysed by 1.2% agarose gel electrophoresis, stained with ethidium bromide and quantified with Java2HTML Image software (http://rsb.info.nih.gov/ij). For standardization, cDNAs were equalized on the basis of their content of the RPL2 housekeeping gene which was amplified using the primers reported in Table 1. The corrected cDNA concentrations were successively used for the semiquantitative amplification of specific genes. The PCR products were purified using a Jet-Quick spin column (Genomed, Löhne, Germany) and directly sequenced using a BigDye terminator kit (Applera, Norwalk, CT, USA) according to the supplier's instructions. Sequencing reactions were run on an ABI Prism 310 Genetic Analyzer (Perkin Elmer Life and Analytical Sciences, Boston, MA, USA).

Experimental design and statistical analysis

Data reported in the figures refer to one of three fumigation experiments with similar results. All the analyses were carried out on leaflets from leaf numbers 3–5 collected from three individual plants (for assessment of cell viability, ion leakage, ET evolution and H2O2 accumulation) or pooled from the three plants (for ACC quantification and RNA isolation). Cell viability and ion leakage were determined in six replicate samples, each represented by 10 leaf discs. ET evolution was measured in three replicate samples, each comprising two leaflets per plant, and H2O2 accumulation was estimated in six replicate individual leaflets. Estimation of free and conjugated ACC content and RT-PCR analyses were carried out on six or four replicate tissue samples, respectively, each comprising a pool of leaflets collected from the three plants.

The statistical significance of differences between control and O3-treated leaves was assessed using Student's t-test (P = 0.01).

Results

Lesion development

Exposure to 200 ppb of O3 for 4 h led to leaf injury on middle-aged leaves of both cv. Pearson and Nr tomato plants. The extent of damaged leaf area was similar in the two genotypes, ranging from 9 to 11% of total leaf area. Leaf injury was not evident during the exposure period but developed between 24 and 48 h after the beginning of the fumigation. The first visual symptoms were detectable as translucent spot-like lesions which developed in small necrotic interveinal areas detectable after 24 h (Fig. 1a). Pretreatment of leaves with AVG almost completely prevented lesion development in both cv. Pearson and Nr O3-treated leaves (Fig. 1a).

Figure 1.

Lesion development (a) and evaluation of the integrity of plasma membranes using vital nonpermeable Evans blue staining (b) and ion leakage from leaf discs (c) of cv. Pearson and Never ripe tomato (Lycopersicon esculentum) plants exposed to filtered air (control) or to 200 nl l−1 ozone (O3-treated) for 4 h. Times refer to hours after the start of fumigation. (a) To examine ethylene involvement in the progression of O3-induced lesions, leaves were pretreated with H2O or with an inhibitor of ethylene biosynthesis, aminoethoxyvinylglycine (AVG), 24 h before O3 fumigation. (b) Stained cells are indicative of injury occurring at the plasma membrane level. (c) Black bars indicate the duration of exposure. Data represent the mean ± standard error. Statistically significant differences according to the t-test between O3-treated samples (closed circles) and the respective controls (open circles) at each time-point are indicated with an asterisk (*) (P < 0.01, n = 6).

The ability of plasma membranes to exclude the vital nonpermeable dye Evans blue is often employed as an indicator of membrane integrity. From 1 h of O3 exposure onwards, small clusters of blue-stained cells were evident in cv. Pearson as well as in Nr O3-treated leaves. In both genotypes, the number of injured cells progressively increased during O3 exposure and reached a maximum during the postfumigation period, leading to large blue-stained areas 24 h after the initiation of treatment (Fig. 1b). Similarly to observations for visible leaf symptoms, no difference in the size or number of injured cells was detected between cv. Pearson and Nr O3-treated samples.

The integrity of cell membranes was also assessed by measuring ion leakage from leaf discs. A slight, but significant, increase in membrane permeability was observed in both genotypes from 0.5 h of O3 exposure onwards (+34 and +25% in cv. Pearson and in Nr, respectively; Fig. 1c). Never ripe displayed a maximum increase in ion leakage at 1 h (+43%), after which the increase in ion leakage remained significantly higher than the value detected in control leaves, with percentages of increase ranging from 15 to 27% (see Fig. 3 below). A similar behaviour was shown by cv. Pearson, which, in addition to the peak recorded at 1 h (+48%), exhibited a second peak at 5 h (+55%; Fig. 1c).

Figure 3.

Evaluation of hydrogen peroxide (H2O2) accumulation using 3,3′-diaminobenzidine (DAB) staining in leaves of cv. Pearson and Never ripe tomato (Lycopersicon esculentum) plants exposed to filtered air or to 200 nl l−1 ozone (O3) for 4 h. Times refer to hours after the start of fumigation. Leaves were pretreated with H2O or with an inhibitor of ethylene biosynthesis, aminoethoxyvinylglycine (AVG), 24 h before O3 fumigation, as described in the text. The percentage of leaf area stained by DAB is reported above each leaf.

Evolution of ethylene and free and conjugated ACC concentrations

Both genotypes were found to increase ET evolution following O3 exposure and exhibited maximum ET emission after 2 h of fumigation (+488% and +612% in cv. Pearson and Nr in comparison to the respective controls). However, different patterns of emission were observed in the two genotypes (Fig. 2a,b). In cv. Pearson, ET evolution increased earlier than in Nr, with increases of +36% at 30 min and +87% at 1 h of exposure, and remained higher than control values for the duration of O3 exposure (Fig. 2a), whereas in Nr ET evolution started to increase at 2 h of exposure and, although it progressively decreased, it remained significantly higher than control values until 3 h of recovery in pollutant-free air (Fig. 2b). In both genotypes, ET evolution was reduced by more than 97% following leaf pretreatment with AVG (data not shown).

Figure 2.

Leaf ethylene emission (a, b) and measurement of free (c, d) and conjugated (e, f) pools of 1-aminocyclopropane-1-carboxylic acid (ACC) of cv. Pearson and Never ripe tomato (Lycopersicon esculentum) plants exposed to filtered air (control) or to 200 nl l−1 ozone (O3-treated) for 4 h. Times refer to hours after the start of fumigation and black bars indicate the duration of exposure. Data represent the mean ± standard error. Statistically significant differences according to the t-test between O3-treated samples (closed circles) and the respective controls (open circles) at each time-point are indicated with an asterisk (*) (P < 0.01; ethylene emission, n = 3; ACC measurement, n = 6).

The concentrations of free ACC were found to increase in both cv. Pearson and Nr leaves following O3 exposure, although, similarly to the situation for ET emission, the time-course of ACC accumulation differed between the two genotypes. In particular, in cv. Pearson, the free ACC concentration had already increased slightly after the first 30 min of O3 exposure (+43%) and reached a maximum 2 h after the start of fumigation, when it was about 5-fold higher than the control value (Fig. 2c). The pool of free ACC returned to the same levels as at the beginning of the exposure at 4 h of exposure, and exhibited two further slight increases after 1 and 20 h of recovery in filtered air (+76% and +38%, respectively). In Nr, the increase in the free ACC pool was evident only after 2 h of exposure to the pollutant, being about 6-fold higher than the control value. The free ACC pool further increased to 11-fold that of the control at 4 h and, after a sharp decline during the first hour of the postfumigation period (although it was still significantly higher than the control value at +64%), it increased markedly at the end of the recovery period (5.5-fold; Fig. 2d).

The conjugated pools of ACC were also found to be positively affected by O3 in both cv. Pearson and Nr, but the O3-induced increase was smaller than that of the free form. The conjugated pools of ACC exhibited similar behaviours in the two genotypes during the 4-h fumigation period, with increases over the respective controls reaching a maximum at 2 h (+142 and +148% in cv. Pearson and Nr, respectively; Fig. 2e,f). After a transient decline to control levels at 5 h, both genotypes exhibited similar increases in conjugated ACC concentration at the end of the recovery period (+97 and +108% in cv. Pearson and Nr, respectively; Fig. 2e,f).

In situ localization of O3-induced H2O2 accumulation

H2O2 accumulation detected by DAB staining was visible as dark-brown precipitates produced by the peroxidase-catalysed reaction of H2O2 with the dye. Both genotypes showed local H2O2 accumulation, which was almost completely prevented by pretreatment with the ET synthesis inhibitor AVG (Fig. 3). DAB staining was first detectable after 1 h of O3 exposure, when it accounted for 7.3 and 4.8% of leaf area in cv. Pearson and Nr, respectively. In both genotypes, the dark-brown precipitates increased during the fumigation period (18 and 12.9% of leaf area at 4 h in Nr and cv. Pearson, respectively) and were particularly abundant during the recovery period, with maximum values at 24 h (28.6 and 19.9% of leaf area in Nr and cv. Pearson, respectively; Fig. 3).

H2O2 accumulation at the cellular level was monitored using a cytochemical assay based on the reaction of H2O2 with CeCl3, which produces electron-dense insoluble precipitates of cerium perhydroxide at sites at which H2O2 has accumulated. In control leaves of both genotypes, no or very few electron-dense precipitates were detectable in the cell walls of palisade parenchyma or spongy mesophyll cells (Fig. 4a,c). Faint electron-dense precipitates were evident in both genotypes from 1 h of O3 exposure onwards (data not shown). The intensity of CeCl3 staining increased at the end of the fumigation period, when leaves of cv. Pearson exhibited faint precipitates in the spongy mesophyll tissue, at the sites of connection between adjacent cell walls close to the intercellular spaces (Fig. 4b; arrows), whereas in Nr CeCl3 staining was visible in the cell walls of both spongy mesophyll and palisade parenchyma cells (Fig. 4d–f; arrows). In these palisade parenchyma cells, however, the diffuse precipitates were fainter and smaller and were detectable only in the cell walls facing the spongy mesophyll (Fig. 4e,f; asterisks). In both genotypes, very strong precipitation was observed at the end of the recovery time (24 h), mainly in the cell walls of the spongy mesophyll (Fig. 4g,h).

Figure 4.

In situ localization of hydrogen peroxide (H2O2) accumulation in leaves of cv. Pearson (a, b, g) and Never ripe (c–f, h) tomato (Lycopersicon esculentum) plants exposed to filtered air or to 200 nl l−1 ozone (O3) for 4 h. Leaf spongy mesophyll cells of cv. Pearson and Nr exposed to filtered air showed few electron-dense precipitates (a, c). Leaf spongy mesophyll cells of cv. Pearson and Nr 4 h after the start of the O3 fumigation (b, d), palisade parenchyma cells of Nr at 4 h (e, f) and spongy mesophyll cells of cv. Pearson and Nr at the end of the 24-h recovery period (g, h) exhibited strong H2O2 accumulation on cell walls between adjacent cells. Bars, 1 µm. Nu, nucleus; s, starch grain; chl, chloroplast; m, mitochondrion. Arrows and asterisks (*) indicate H2O2 accumulation.

Enzyme activity

The APX activity of cv. Pearson was significantly decreased by O3 throughout the period of exposure (ranging from −69% at 30 min to −29% at 4 h) as well as after 1 h of recovery in nonpolluted air (−41%). This trend was reversed at the end of the recovery period, when APX activity significantly increased over that of the control (+54%) (Fig. 5). In Nr, APX activity was negatively affected by O3 only in the very early phase of exposure (−32% at 30 min) and remained unchanged in comparison to the control throughout the O3 fumigation period. (Fig. 5). A slight but significant increase in APX activity was then recorded at 1 and 3 h of recovery in pollutant-free air (+20 and +14%, respectively; Fig. 5).

Figure 5.

Ascorbate peroxidase (APX) activity and NAD(P)H oxidation rate of cv. Pearson and Never ripe tomato (Lycopersicon esculentum) plants exposed to filtered air (control) or to 200 nl l−1 ozone (O3-treated) for 4 h. Times refer to hours after the start of fumigation. Data represent the mean ± standard error. Statistically significant differences according to the t-test between O3-treated samples (closed bars) and the respective controls (open bars) at each time-point are indicated with an asterisk (*) (P < 0.01, n = 6).

As far as NAD(P)H oxidation rate is concerned, the two tomato genotypes showed quite different behaviours, with differences between them being greatest during the fumigation period. In particular, in cv. Pearson the activity was unchanged until 1 h of recovery, while in Nr a marked stimulation was already detectable at 30 min (+181%). The activity remained significantly higher than that of the control until 2 h (+138 and +78% at 1 and 2 h, respectively; Fig. 5) and then returned to control values at 4 and 5 h. In both genotypes, the NAD(P)H oxidation rate was significantly increased in the later stages of the postfumigation period (+84% at 7 h and +325% at 24 h in cv. Pearson, and +68% at 7 h and +111% at 24 h in Nr; Fig. 5).

Ozone induced ethylene biosynthetic genes

Both ACS and ACO are encoded by multigene families in most plant species, and these genes are regulated differently at the transcriptional level. In our experiment, LE-ACS6 was constitutively expressed at a higher level than LE-ACS2 in control plants, particularly in cv. Pearson (Fig. 6). The O3 treatment induced increases in the expression levels of both genes, although the trends of induction were different. In cv. Pearson, the LE-ACS6 mRNA content had already increased at 0.5 h; the transcript abundance remained elevated during the 4 h of exposure, and then during the recovery period it declined markedly to near control levels. In Nr, LE-ACS6 was induced 30 min after the initiation of O3 exposure and reached a maximum at 2 h, after which the transcript abundance decreased slightly, although it remained significantly higher than that of unfumigated plants. In cv. Pearson, LE-ACS2 expression increased after 1 h of exposure and peaked at 4 h, and after 3 h of recovery its levels were comparable to those measured in unfumigated plants (Fig. 6). In Nr, LE-ACS2 was induced later than in cv. Pearson and reached a maximum at 5 h, declining thereafter, and after 24 h its levels were comparable to those of the control. Ozone exposure also caused a rapid increase in the expression of both the LE-ACO1 and LE-ACO3 genes. LE-ACO1 expression was strongly increased after 1 and 2 h of O3 treatment in cv. Pearson and Nr, respectively. In both genotypes, transcript abundances remained elevated until 7 h and then decreased to near control values (Fig. 6). Also, LE-ACO3 was strongly up-regulated following O3 fumigation, with similar trends in cv. Pearson and Nr plants (Fig. 6). Interestingly, during the first hour of fumigation in the mutant, we found a significant accumulation of LE-ACS6 transcript only; in contrast, in cv. Pearson all tested genes were up-regulated. When the plants were pretreated with AVG the transcript abundance for all ET biosynthetic genes was not significantly influenced by O3 fumigation (Fig. 7), consistent with the measured decrease in O3-induced ET emission and tissue damage.

Figure 6.

Expression levels of Lycopersicon esculentum 1-aminocyclopropane-1-carboxylic acid (ACC) synthase 6 (LE-ACS6), LE-ACS2, ACC oxidase 1 (LE-ACO1) and LE-ACO3 mRNA in leaves of cv. Pearson and Never ripe tomato (Lycopersicon esculentum) plants during ozone (O3) fumigation (200 nl l−1 for 4 h) and during the recovery period. Times refer to hours after the start of fumigation. Total RNA (1–2 µg) was reverse-transcribed and amplified by semiquantitative reverse transcriptase–polymerase chain reaction (RT-PCR). The cDNAs were standardized against the ribosomal protein L2 (RPL2) gene of tomato. The relative intensities of the signals (mean ± standard error; n = 4) are shown. The significance of differences between controls (open bars) and O3-treated (closed bars) leaves was determined with a t-test, and an asterisk (*) indicates a difference significant at the P < 0.01 level. A photograph of a representative gel is shown below each graph. Leaf samples were extracted at 0.5, 1, 2, 4, 5, 7 and 24 h after the start of fumigation from control (lanes 1–7) and O3-treated (lanes 8–14) plants.

Figure 7.

Expression levels of Lycopersicon esculentum 1-aminocyclopropane-1-carboxylic acid (ACC) synthase 6 (LE-ACS6), LE-ACS2, ACC oxidase 1 (LE-ACO1) and LE-ACO3 mRNA in leaves of cv. Pearson and Never ripe tomato (Lycopersicon esculentum) plants during ozone (O3) fumigation (200 nl l−1 for 4 h) and during the recovery period. To explore whether a blockage in ethylene synthesis could influence O3 sensitivity, 24 h before O3 fumigation leaves of the two genotypes were painted with 1 mm aminoethoxyvinylglycine (AVG), an inhibitor of ACS activity, and then fumigated or exposed to O3-free air (control). Total RNA (1–2 µg) was reverse-transcribed and amplified by semiquantitative reverse transcriptase–polymerase chain reaction (RT-PCR). The cDNAs were standardized against the ribosomal protein L2 (RPL2) gene of tomato. The relative intensities of the signals (mean ± standard error; n = 4) are shown (control, open bars; O3-treated, closed bars). A photograph of a representative gel is shown below each graph. Leaf samples were extracted at 0.5, 1, 2, 4, 5, 7 and 24 h after the start of fumigation from control (lanes 1–7) and O3-treated (lanes 8–14) plants.

The Nr mutation did not affect the JA pathway

We used analysis of GLUB mRNA in wild-type and Nr leaves to test whether ET insensitivity affects an ET/JA-inducible pathogenesis-related protein under acute O3 fumigation. Ozone significantly induced GLUB mRNA accumulation in cv. Pearson plants (Fig. 8). Interestingly, we observed two peaks: the first within 1–2 h of the start of fumigation and the second at 24 h. In Nr, induction of GLUB by O3 was weaker and later than in cv. Pearson; we observed an increase in the amount of mRNA 7 and 24 h after the start of fumigation (Fig. 8). To explore whether the Nr mutation influenced JA biosynthesis we investigated the expression of the AOS and AOC genes, which play a central role in JA biosynthesis. The AOS2 mRNA analysis showed a low level of expression in the control plants of both genotypes (Fig. 8). Following O3 exposure, in cv. Pearson AOS2 mRNA content increased at 7 and 24 h. In Nr, AOS2 was induced at the end of fumigation (4 h) and peaked at 7 h, but, unlike the situation in cv. Pearson, at 24 h the mRNA content was similar to that detected in control unfumigated plants (Fig. 8). Allene oxide, formed by AOS, is cyclized by AOC to 12-oxo-phytodienoic acid. Comparing the expression profiles of AOC in cv. Pearson and Nr, we observed a marked induction in both genotypes during the recovery period. The activation of the JA pathway in response to O3 treatment in tomato plants was confirmed by the expression profile of the JA-inducible gene PINII. Results from PINII mRNA analyses showed a very low level of PINII expression in unfumigated plants of both genotypes and late O3-induced PINII mRNA accumulation at 7 and 24 h in Nr and cv. Pearson, respectively (Fig. 8).

Figure 8.

Expression levels of β-1,3-glucanase (GLUB), allene oxide synthase 2 (AOS2), allene oxide cyclase (AOC) and proteinase inhibitor II (PINII) mRNA in leaves of cv. Pearson and Never ripe tomato (Lycopersicon esculentum) plants during ozone (O3) fumigation (200 nl l−1 for 4 h) and during the recovery period. Total RNA (1–2 µg) was reverse-transcribed and amplified by semiquantitative reverse transcriptase–polymerase chain reaction (RT-PCR). The cDNAs were standardized against the ribosomal protein L2 (RPL2) gene of tomato. The relative intensities of the signals (mean ± standard error; n = 4) are shown. The significance of differences between controls (open bars) and O3-treated (closed bars) leaves were determined using a t-test, and an asterisk (*) indicates differences significant at the P < 0.01 level. A photograph of a representative gel is shown below each graph. Leaf samples were extracted at 0.5, 1, 2, 4, 5, 7 and 24 h after the start of fumigation from control (lanes 1–7) and O3-treated (lanes 8–14) plants.

Discussion

In recent years, ET evolution by O3-treated plants has been recognized as one of the fastest responses to this pollutant. It has been shown that lesion propagation and the extent of cell death are under the control of ET and that functional ET perception is a fundamental step in triggering the signal transduction machinery and the consequent biological response (Overmyer et al., 2000; Moeder et al., 2002; Rao et al., 2002; Tuominen et al., 2004; Kangasjärvi et al., 2005). However, the roles played by the different ET receptors have not been investigated to any great extent. In an attempt to understand the relative importance of the ET receptor LE-ETR3 (NR) in perceiving ET and, consequently, in determining O3 sensitivity, we employed the tomato mutant Nr, which carries a semidominant mutation in LE-ETR3 and exhibits a number of pleiotropic effects indicative of ET insensitivity throughout the plant (Lanahan et al., 1994).

Here, we have demonstrated that the Nr mutant did not differ from cv. Pearson in O3 sensitivity, as suggested by the absence of substantial differences between the two genotypes in the appearance of O3-induced leaf lesions. This lack of a difference in O3 sensitivity, which was also observed by Moeder et al. (2002), was confirmed by the similar degrees of plasma membrane alteration detected in the two genotypes following Evans blue staining and ion leakage measurement. As far as this latter parameter is concerned, cv. Pearson exhibited a second peak 1 h into the recovery period, which, however, did not reflect a difference in the extent of membrane damage and was not related to major ET production and/or a difference in the degree of H2O2 accumulation in comparison to the mutant, suggesting that this behaviour could be ascribed to quantitative fluctuations in the general trend of increased ion leakage, unless there is involvement of other factors not investigated in the present experiment. In fact, despite the absence of a second peak, ion leakage of O3-treated samples in Nr was also always higher than in the respective controls.

As treatment of plants with the inhibitor of ET synthesis (AVG) almost completely prevented the development of tissue damage on O3-fumigated leaves of both cv. Pearson and Nr, ET was also undoubtedly involved in the progression of O3-induced lesions in the mutant genotype, which thus appeared to be able to sense ET despite the defect in the receptor machinery.

Never ripe was found to evolve ET following O3 stress, in agreement with the findings reported by Moeder et al. (2002), although in the present experiment the mutant exhibited delayed and prolonged emission compared with cv. Pearson. Lanahan et al. (1994) reported similar rates of ET production by Nr and cv. Pearson following infiltration of either wild-type or mutant leaves with ACC or with the bacterial pathogen Pseudomonas syringae, suggesting that Nr is not impaired in any step of ET biosynthesis (Lanahan et al., 1994). We have shown here that O3 induced the ET biosynthetic enzymes ACS and ACO at the transcriptional level, and that inhibition of ACS activity by AVG reversed the increase in the transcript abundances of all tested ET biosynthetic genes and led to increased O3 tolerance in both cv. Pearson and Nr.

Increased transcription of specific members of the ACS family has been reported in different O3-treated species (Tuomainen et al., 1997; Moeder et al., 2002; Nakajima et al., 2002; Diara et al., 2005), and in some Solanaceae species a sequential expression of ACS genes has been observed in response to both biotic and abiotic stresses (Schlagnhaufer et al., 1997; Nakajima et al., 2001; Moeder et al., 2002). Our results show that LE-ACS6 and LE-ACS2 were transiently induced under O3 and that, in agreement with the literature (Nakajima et al., 2001; Moeder et al., 2002), LE-ACS6 was induced earlier than LE-ACS2. Tobacco (Nicotiana tabacum) plants transformed with LE-ACS6 in antisense orientation were less sensitive to O3 and evolved ET to a lesser extent than the untransformed controls (Nakajima et al., 2002), suggesting a primary role for LE-ACS6 in triggering O3-induced ET evolution. Conversely, the LE-ACS2 antisense plants did not differ in their O3 sensitivity when compared with the untransformed plants (Moeder et al., 2002), emphasizing the importance of ET produced in the initial phases of the exposure as a promoter of the O3-induced ET-dependent responses. However, post-translational modification of ACS proteins by phosphorylation has also been recognized as a regulatory mechanism capable of modulating ET biosynthesis before ACS gene activation (Tuomainen et al., 1997, and references in Kangasjärvi et al., 2005).

Interestingly, in the present experiment, LE-ACS6 expression in Nr did not undergo the sharp reduction detected in cv. Pearson and reported previously in tomato (Nakajima et al., 2001; Moeder et al., 2002), remaining higher than control values throughout the recovery period. The sharp decline in LE-ACS6 mRNA detected in cv. Pearson can be attributed to an autoinhibitory mechanism related to ET accumulation (Kim et al., 1997) that is probably impaired in Nr by its ET insensitivity. LE-ACS2 and LE-ACO1 in tomato fruit have been reported to be under positive feedback regulation by ET (Nakatsuka et al., 1998). Intriguingly, we found that these genes were induced by O3 in both genotypes, although in Nr their induction was delayed by c. 1 h with respect to the wild type, suggesting that the insensitivity of the mutant to ET is only partial and that the threshold for induction of biosynthetic genes is achieved in Nr later than in the wild type because of the defect in ET perception.

As a result of the enhanced transcription of the two ACS genes, a marked increase in ACC concentration was detected during O3 exposure in both tomato genotypes. In this context it is noteworthy that in O3-treated leaves of Nr, which exhibit a prolonged induction of LE-ACS6 and a delayed activation of LE-ACS2, the pool of free ACC displayed a second peak in the recovery period. However, although ET evolution declined more slowly than in cv. Pearson, still remaining significantly higher than control values until 3 h of recovery, its production at 20 h of recovery was negligible. ACC is known to have two possible destinies: to undergo ET convertion by ACC oxidase, or to undergo conjugation to metabolically inert compounds. At 20 h of recovery the conjugated pools of ACC underwent a substantial increase in both genotypes, suggesting similar conjugation abilities, but, as ACO transcript levels, and probably the enzyme activity, declined to control values, the higher amount of ACC synthesized by Nr following the prolonged ACS stimulation could not be converted to ET, thus leading to free ACC accumulation.

Recent studies presented evidence that O3 activates an oxidative burst and may induce a hypersensitive response mechanistically similar to the pathogen-induced response (reviewed by Baier et al., 2005; Kangasjärvi et al., 2005). In this context, ROS would act as early signals capable of eliciting the diverse effects that constitute the plant response to O3 (Pellinen et al., 1999, 2002; Langebartels et al., 2002; Pasqualini et al., 2002, 2003; Wohlgemuth et al., 2002; Ranieri et al., 2003; Diara et al., 2005). In the O3-sensitive tobacco cultivar BelW3, the oxidative burst was accompanied by a transient increase in the transcript abundance of the hypersensitive response marker PR1a, increased protease activity and chromatin condensation, indicating that ROS-induced programmed cell death is initiated by O3-induced oxidative stress (Pasqualini et al., 2003). In the present experiment, in both cv. Pearson and Nr leaves, O3 activated a H2O2-dependent oxidative burst which, consistent with findings in O3-sensitive species or clones (Schraudner et al., 1998; Pellinen et al., 1999), displayed a second oxidative peak during the postfumigation period. As indicated by the dramatic reduction in DAB staining observed when ACS activity was inhibited by AVG treatment, ET also appeared to be intimately involved in the amplification of ROS production in Nr leaves. Moreover, in the mutant, H2O2 production also spread to the palisade parenchyma cell walls and, although very faintly, to the chloroplasts, probably sustained by the prolonged ET emission during the postfumigation period. A multitude of ROS-generating systems during the O3-induced oxidative burst have been reported. In addition to direct generation following O3 degradation in the apoplast or the interaction of the pollutant with organic molecules containing C–C bonds, H2O2 production can originate from the activity of the plasma membrane NAD(P)H oxidase complex, extracellular pH-dependent peroxidases (PODs), oxalate oxidase and diamine and polyamine oxidases, as well as from a reduced scavenging capacity (Bolwell & Wojtaszek, 1997; Sebela et al., 2001; Langebartels et al., 2002; Ranieri et al., 2003; Diara et al., 2005). In the two tomato genotypes, different H2O2-producing and -scavenging mechanisms were responsible for H2O2 accumulation. In Nr, until 2 h of O3 exposure, active H2O2 generation occurred at the expense of NAD(P)H oxidation, whereas the reduced APX activity made a contribution only in the very early stage of fumigation (until 30 min). Conversely, in cv. Pearson, H2O2 accumulation until 5 h seemed rather to originate from the substantial decrease in the activity of APX, a key enzyme in the process of H2O2 scavenging (Foyer et al., 1994). Then, from 7 h onwards, both genotypes actively produced H2O2. However, it should be noted that cv. Pearson displayed higher constitutive activities of APX, and that at the end of the recovery period Nr was not able to stimulate APX, thus allowing the oxidative burst to spread also to the palisade parenchyma cell walls and to the chloroplasts. However, because of the complexity of the genesis of the O3-induced oxidative burst, the involvement of other producing and scavenging systems not investigated in the present experiment must be taken into account.

Previous studies have implicated ET in the control of the expression of pathogenesis-related (PR) genes in response to biotic and abiotic elicitors of the defence response (Chappell et al., 1984; Broglie et al., 1986; Ecker & Davis, 1987; Roby et al., 1991). However, many PR proteins are under the control of a combination of signalling molecules. The gene encoding GLUB was demonstrated to be induced by ET (Gu et al., 2000; Grimmig et al., 2003) or by JA (Wu & Bradford, 2003). Our results show a marked induction of this gene in cv. Pearson within 1–2 h of the start of fumigation and at 24 h into the recovery period. Induction of GLUB by O3 was weak in Nr, and limited to the recovery period. The early GLUB induction detected in cv. Pearson can be correlated with the endogenous level, or perception of, ET. The lack of induction of GLUB found in Nr during the fumigation could be attributed to a desensitization of the mutant to ET, whereas the GLUB induction observed in both the wild type and Nr in the recovery period could be a result of JA accumulation. Although we did not measure JA concentrations, we did observe an O3-induced increase in the transcript abundances of the JA biosynthetic genes AOS2 and AOC (which have been reported to be strictly correlated to JA concentration; Creelman & Mullet, 1997; León & Sánchez-Serrano, 1999; Turner et al., 2002) in both genotypes only during the recovery period. Moreover, in the postfumigation period, the JA-responsive gene PINII was also actively transcribed in both cv. Pearson and Nr, suggesting that the defective ET perception by the mutant did not impair biosynthesis and perception of JA. Recently, Fisahn et al. (2004) showed that an increase in cytosolic Ca2+ is required for induction of JA biosynthesis and PINII gene expression. The earlier induction of PINII and of genes involved in JA biosynthesis detected in Nr suggested altered timings in the O3-induced Ca2+ flux and in the mitogen-activated protein kinase (MAPK) cascade, which requires further investigation.

Overall, the data obtained indicate that the Nr mutant retains a partial sensitivity to ET, thus suggesting only a marginal role of the LE-ETR3 (NR) receptor in mediating the complex response of tomato plants to O3 and implying the involvement of other ET receptors. There is published evidence of strong up-regulation of the LE-ETR1 gene by O3 (Moeder et al., 2002). Moreover, disruption of ET perception by transformation of birch (Betula pendula) with the dominant negative mutant allele etr1-1 of the Arabidopsis ET receptor gene ETR1 partially prevented O3-induced cell death (Vahala et al., 2003), suggesting a major role for this ET receptor. Conversely, the NR gene seems to be heavily involved in the response to pathogen infection, as indicated by the significantly higher tolerance of the Nr mutant to a virulent strain of Xanthomonas campestris (Lund et al., 1998; Ciardi et al., 2000).

Further experiments are thus necessary to elucidate the roles played by the different ET receptors in mediating the plant response to O3 and also to unravel the relative importance of these receptors in regulating the cell responses to different kind of stresses sharing similar signalling routes.

Accession numbers

Sequence data from this article can be found in the NCBI/GenBank data libraries under the following accession numbers: X59139 (ACS2), AF179249 (ACS6), X58273 (ACO1), Z54199 (ACO3), M80608 (GLUB), AF230371 (AOS2), AF384374 (AOC), AY129402 (PINII) and X64562 (RPL2).

Acknowledgements

We are indebted to Dr Jennifer Petersen (Tomato Genetic Resource Center, University of California, Davis, CA, USA) for providing the tomato seeds. This research was supported by a grant from MIUR (National Project) Rome, Italy, and by funds from the University of Pisa.

Ancillary