Natural variation in ozone sensitivity among Arabidopsis thaliana accessions and its relation to stomatal conductance


H. Kollist. Fax: +372 7374900; e-mail:


Genetic variation between naturally occurring populations provides a unique source to unravel the complex mechanisms of stress tolerance. Here, we have analysed O3 sensitivity of 93 natural Arabidopsis thaliana accessions together with five O3-sensitive mutants to acute O3 exposure. The variation in O3 sensitivity among the natural accessions was much higher than among the O3-sensitive mutants and corresponding wild types. A subset of nine accessions with major variation in their O3 responses was studied in more detail. Among the traits assayed, stomatal conductance (gst) was an important factor determining O3 sensitivity of the selected accessions. The most O3-sensitive accession, Cvi-0, had constitutively high gst, leading to high initial O3 uptake rate and dose received during the first 30 min of exposure. Analyzing O3-induced changes in stress hormone concentrations indicated that jasmonate (JA) concentration was also positively correlated with leaf damage. Quantitative trait loci (QTL) mapping in a Col-0 × Cvi-0 recombinant inbred line (RIL) population identified three QTLs for O3 sensitivity, and one for high water loss of Cvi-0. The major O3 QTL mapped to the same position as the water loss QTL further supporting the role of stomata in regulating O3 entry and damage.


Ozone (O3) is a phytotoxic air pollutant predicted to increase continually during the 21st century because of climate change and industrial growth in Asia; particularly, high radiation and temperature promote increases in the tropospheric O3 (Bytnerowicz, Omasa & Paoletti 2007). In O3 research, Arabidopsis mutants help to identify the mechanisms underlying differences in plant O3 sensitivity. For example, antioxidants and antioxidative enzymes function to cope with reactive oxygen species (ROS) formed during O3 degradation in the apoplast (Luwe, Takahama & Heber 1993; Musselman et al. 2006). A vtc1 mutant that contained only 30% of the wild-type ascorbate (AA) was hypersensitive to O3, indicating the role of antioxidants in defence against O3 (Conklin, Williams & Last 1996). Because O3 enters the plant leaf mainly via stomata (Kerstiens & Lendzian 1989), stomatal conductance determines the uptake of O3 into the apoplastic space (Kollist et al. 2000; Kangasjärvi, Jaspers & Kollist 2005). An O3-sensitive mutant slac1 with impaired S-type anion channel currents, leading to constitutively high stomatal conductance and restricted stomatal closure under O3 exposure, was recently reported (Vahisalu et al. 2008). Arabidopsis mutants have also been instrumental in defining the role of stress hormones in plant O3 sensitivity. Mutants involved in jasmonate (JA) biosynthesis (fad3/7/8 and jar1) or signalling mutants (coi1) display O3 sensitivity, indicating a protective role for JA in O3-induced cell death (Rao et al. 2000; Tuominen et al. 2004). Studies using the salicylic acid (SA) biosynthesis mutant sid2, the signalling mutant npr1 and transgenic plants degrading SA (NahG) have suggested a dual role for SA in O3 responses; SA is needed to induce defence responses, but elevated SA levels promote cell death (Rao & Davis 1999; Rao et al. 2000; Overmyer et al. 2005; Yoshida et al. 2009). Ethylene promotes O3 damage, and ethylene-overproducing mutants (eto1 and eto3) are O3 sensitive (Rao, Lee & Davis 2002). Furthermore, the hormones are not acting independent of each other; instead, they exhibit both synergism and antagonism, making the study of hormone interaction very complex (Ahlfors et al. 2004a; Tuominen et al. 2004; Kangasjärvi et al. 2005). Abscisic acid (ABA) is a central regulator of stomatal function and abiotic stress responses; however, its role in O3 responses is less established (Kangasjärvi et al. 2005).

Although mutants are very useful in understanding differences in stress sensitivity among plants, they usually focus on only one gene at a time. However, abiotic stress tolerance involves many genes at the same time (Bhatnagar-Mathur, Vadez & Sharma 2008). Arabidopsis has a broad natural distribution throughout the Northern Hemisphere (Alonso-Blanco & Koornneef 2000). Across this geographic range, it experiences a broad range of climatic conditions and selective pressures. Genetic variation between naturally occurring Arabidopsis populations is large, providing a unique source to study the effect of complex genetic variation on stress tolerance. It has been found that the sensitivity of accessions to stress factors, such as increased metal concentrations (Murphy & Taiz 1995), photo-oxidative stress (Abarca et al. 2001), O3 exposure (Rao & Davis 1999; Tamaoki et al. 2003) and pathogen infection (Kover & Schaal 2002), is highly variable. Genetic analyses of natural variation can be performed by quantitative trait loci (QTL) mapping. In Arabidopsis, several recombinant inbred lines (RILs) have been generated from genetically and phenotypically distinct parents, and these have been used for successful QTL mapping for a variety of traits involved in adaptation to local conditions including seed dormancy, flowering time and disease resistance (Alonso-Blanco et al. 2009).

We analysed 93 Arabidopsis accessions and five O3-sensitive mutants (four in Col-0 background, and one in Ws-2 background) for O3-induced visible injury, stomatal conductance and water loss from excised leaves. In addition, nine accessions were studied in more detail for antioxidative capacity, hormone concentrations and O3 uptake characteristics during O3 exposure. Previous studies suggest that there is large variation in O3 response among genotypes even when their geographical origin is restricted to one country (Plantago major, Davison & Reiling 1995) or to one forest (Betula pendula, Oksanen et al. 2005). Thus, a high capacity to adapt to the potentially increasing pollution levels within these populations has been suggested (Oksanen et al. 2005). We asked the following questions: (1) How large is the variation in O3-induced visible injury among Arabidopsis accessions per se, and in comparison with O3-sensitive mutants? (2) Is there any correlation between O3 sensitivity and the origin climatic conditions for accessions? (3) What mechanisms could be related to differential O3 sensitivity? (4) Can QTL mapping identify mechanisms involved in O3 sensitivity?


Analysis of 93 Arabidopsis accessions and five O3-sensitive mutants for O3-induced visible injury

Arabidopsis thaliana accessions used in the initial screening for O3 sensitivity were obtained from the Nottingham Arabidopsis Stock Centre (NASC; Supporting Information Table S1). Accessions were evaluated for stomatal conductance, water loss from excised leaves and O3 sensitivity together with O3-sensitive mutants oji1 (Ws-2 background, Kanna et al. 2003), vtc1, rcd1, re-8, and slac1 (all Col-0 background, Kangasjärvi et al. 2005; Overmyer et al. 2008).

Seeds were sown at high density on a 1:1 v/v mixture of vermiculite and peat (type B2, Kekkilä, Finland), and kept for 2 d at 4 °C for stratification. The plants were grown in controlled growth chambers (Bio 1300, Weiss Umwelttechnik, Germany) under a 12 h photoperiod, 23/19 °C day/night temperature and 70/90% relative humidity. The average photosynthetic photon flux density (PPFD) during the light period was 150 µmol m−2 s−1. When seedlings were 1 week old, they were transplanted into 6 × 6 cm pots at a density of one plant per pot. The plants were sub-irrigated every third day with tap water.

Three-week-old plants were exposed to O3 in growth chambers under the same conditions as they were grown until the experiments. O3 exposure was acute (300–350 nmol mol−1 for 7 h) and started 2 h after light was switched on. Acute, relatively high O3 concentration was selected to get distinct and clearly visible phenotypes to score the plants. Each accession/mutant was grown and exposed to O3 twice (4–11 plants in each experiment), except for Col-0 and Cvi-0, which were included in all experiments as negative and positive controls, respectively. The total number of plants for each accession treated with ozone ranged between 12 and 30 (20 being the most common); the corresponding numbers for Cvi-0 and Col-0 were 117 and 115, respectively. The number of leaves on each plant and the number of O3-injured leaves were counted 24 h after the end of O3 exposure. Photographs of all exposed plants were also taken. Visible injuries of all accessions are presented relative to the negative control Col-0 (i.e. the average number of O3-injured Col-0 leaves on that exposure day was subtracted from the average number of O3-injured leaves of the other accessions exposed on that day).

For the correlation of O3-induced visible injury with habitat temperature and precipitation, the origins of C24, Col-0 and Ler-1 were assigned to their closest natural relatives Co, Gückingen and Landsberg, respectively, as in Hannah et al. (2006). Average habitat temperature and precipitation during vegetation period for each accession were calculated from the data available in the CLIMATE database version 2.1 (W. Cramer, Potsdam, Germany; Vegetation period was defined as time when the mean monthly temperature is ≥5 °C.

Measurements of stomatal conductance, water loss from excised leaves and O3 uptake characteristics

In situ leaf stomatal conductance (gst) was measured with an AP-4 porometer (Delta-T Devices, Cambridge, UK). Both leaf surfaces of three middle-aged rosette leaves were measured from four 23- to 26-day-old plants; each accession was measured on two different days except for Col-0 and Cvi-0 which were always included. The measurements were performed between 10.00 and 14.00, local time.

In order to measure water loss from excised leaves, three middle-aged leaves of six 23- to 26-day-old plants per accession were cut, and their weight was determined immediately after cutting. Detached leaves were subsequently arranged abaxial side up in Petri dishes, and weighed again 1 and 2 h after cutting. Water loss was expressed as the decrease of initial fresh weight (in %).

For monitoring whole-plant gas exchange during ozone treatment, 24- to 26-day-old plants were used. The plants were grown and gas exchange was measured with a custom-made measurement device described in detail in Kollist et al. (2007). The plants were inserted in the device in the morning and O3 exposure (250–500 nmol mol−1 for 4 h) started about 2 h later when gst had stabilized. Experiments were performed at PPFD of 150 ± 3 µmol m−2 s−1 and temperature of 23 ± 1°C. The last values of gst and net CO2 assimilation rate at ambient CO2 concentration (Anet), measured before O3 was applied, characterize the gst and Anet for these accessions. Photographs of plants were taken before the experiment, and rosette leaf area was calculated using ImageJ 1.37v (National Institutes of Health, Bethesda, MD, USA). Stomatal conductance for water vapour, leaf net photosynthesis, O3 uptake rate and cumulative absorbed O3 dose were calculated with a custom-written programme as described in Kollist et al. (2007). We compared our values for gst and Anet to the corresponding values obtained in experiments where irradiance was higher to see the effect of relatively low light levels in our experiment on stomata and photosynthesis. Our values of gst were very similar to those reported by others, but Anet was lower (for Col-0 and Ler, Masle, Gilmore & Farquhar 2005; for Col-0, Tocquin et al. 2006).

Ion leakage

Tissue damage upon O3 treatment was quantified by measuring relative ion leakage (Overmyer et al. 2000, 2005). Samples for ion leakage measurements, and for hormone and antioxidant extractions were taken at 2 and 8 h after the onset of O3 exposure. Clean air controls were sampled at the same time-points. The whole rosette was cut and submerged into 9 mL fresh Milli-Q water (18 MΩ) in 50 mL Falcon tubes. Conductivity (mS cm−1) was measured with a conductivity meter (Mettler Toledo GmbH, Schwerzenbach, Switzerland) after shaking on a rotary shaker with 250 rpm (Infors AG, Bottmingen, Switzerland) for 1 h. Total plant conductivity was measured after the samples were frozen at −20 °C for 24–48 h, melted and shaken as described above. Tissue damage was expressed as the percentage of total ions (conductivity/total conductivity × 100%).

Antioxidant and hormone extraction and quantification

For analysis of plant hormones and antioxidants, whole-rosette plant samples were collected at 2 and 8 h time-points, immediately frozen in liquid nitrogen and stored at −80 °C. For antioxidant determination, frozen leaves were ground to a fine powder and extracted two times with ice-cold extraction buffer containing 6% meta-phosphoric acid (MPA, 33.5–36.5% Fluka, Buchs, Switzerland), 2 mm ethylenediaminetetraacetic acid (EDTA) and 1% insoluble polyvinylpolypyrrolidone (PVPP; Sigma, St Louis, MO, USA), and filtered, as reported in Davey, Dekempeneer & Keulemans (2003). High-performance liquid chromatography (HPLC) analyses were performed on an Agilent 1100 Series HPLC (Agilent Technologies, Palo Alto, CA, USA) using an Alltech Adsorbosphere XL Rocket C18 reverse phase column (3 µm, 53 × 7 mm) fitted with a guard cartridge (SecurityGuard; Phenomenex, Torrance, CA, USA) at a flow rate of 3 mL min−1. Injection volume was 10 µL with needle wash. The mobile phase was a linear gradient from 0 to 30% acetonitrile mix in 3 min, followed by 3 min of 30%, a decrease to 0% in 1 min and a post-run of 3 min. Quantification was performed at 243 nm (AA) and 197 nm (reduced glutathione, GSH), with 360 nm as reference wavelength. Following the subtraction method (Davey et al. 2003), samples were reduced after quantification of AA and GSH to determine the total AA and GSH content. Dehydroascorbate (DHA) and oxidized glutathione (GSSG) were calculated as the difference between the values of AA and GSH before and after reduction.

Plant hormones were analysed using a modified vapour-phase extraction method (Schmelz et al. 2003) as described in Lepistöet al. (2009) without the final silylation step for the analysis of ABA and JA. As internal standards, 50 ng of 13C1–SA, 50 ng of dihydrojasmonic acid (Montesano et al. 2005) and 20 ng of D6–ABA (Icon Isotopes, Summit, NJ, USA) were added to the extraction buffer. Gas chromatography–mass spectrometry (GC–MS) analysis was performed on a Trace-DSQ (Thermo Fisher Scientific Inc., Waltham, MA, USA) in the single-ion monitoring mode on a ZB-5 or a ZB-1 (for SA analysis) capillary GC column (30 m × 0.25 mm × 0.25 µm) with splitless injection and 230 °C injector temperature. The column was held at 40 °C for 1 min after injection, then heated at 15 °C min−1 to 250 °C, held for 4 min, and heated at 20 °C min−1 to 310 °C final temperature (kept for 3 min) with helium as carrier gas (flow, 1 mL min−1). Because no significant differences were detected between 2 h control and the corresponding 8 h control for antioxidant and hormone concentrations, the latter control numbers are not shown.

QTL mapping

The core Col-0 × Cvi-0 RIL population consisting of 164 lines was obtained from INRA Versailles (Simon et al. 2008) with marker and map data ( O3 damage was scored in the population as number of leaves with visible damage after a 6 h exposure to 300 nmol mol−1 O3. This ranged from 0 (Col-0) to 5 (Cvi-0). Water loss was scored from detached leaves and was digitized into either 0 (no phenotype) or 1 (water loss phenotype). QTL detection was undertaken using mixed linear composite interval mapping in QTLNetwork 2.0 (Yang et al. 2005; Yang, Zhu & Williams 2007). Composite interval analysis was undertaken using forward–backward stepwise multiple linear regression with a probability into and out of the model of 0.05 and window size set at 5 cM.

Statistical analysis

Statistical analyses were performed with Statistica, version 7.0 (StatSoft Inc., Tulsa, OK, USA). Analysis of variance (anova) (GLM procedure) was used to assess the effects of accession and O3 exposure on different characteristics. Comparisons between individual means were performed using Tukey HSD test. The significance of linear regressions was tested with GLM simple regression analysis. Data were ln-transformed when necessary. All effects were considered significant at P < 0.05.


Screening of Arabidopsis accessions and O3-sensitive mutants

Treatment with 350 nmol mol−1 of O3 for 7 h caused extensive differences in O3-induced leaf damage between the accessions tested (Fig. 1). Accessions ranged from extremely O3 tolerant to hypersensitive, with Cvi-0 showing the greatest amount of injury, followed by Oy-0. The three most tolerant accessions, which showed hardly detectable damage, were Var2-1, Var2-6 and Ull2-5. The variation in O3 sensitivity was much higher between the natural accessions than between the O3-sensitive mutants and their corresponding wild types. The O3-sensitive mutants were in the seriously injured, O3-sensitive group, but they were not among the most damaged lines.

Figure 1.

Ozone sensitivity of Arabidopsis natural accessions and O3-sensitive mutants. Sensitivity was quantified as O3-induced visible leaf injury at 24 h after the fumigation with 300 nmol mol−1 of ozone for 7 h (average ± SE, n = 12–117). Data are presented as relative to Col-0 which was used as negative control.

To determine the role of stomata in O3 sensitivity, we measured abaxial and adaxial stomatal conductances (gst), and water loss from detached leaves for most accessions. The average ratio of abaxial to adaxial conductances was 1.51 ± 0.06. O3-induced visible injury correlated significantly with both abaxial and adaxial gst and water loss from detached leaves (Table 1). Rather low R2 values (0.09–0.23) may partly be explained by methodological problems; AP4 porometer is not well-suited for measurements of Arabidopsis leaves, and the water loss from detached leaves only indirectly hints at the functioning of intact leaves, but the trends were still obvious. There was no significant correlation between O3-induced leaf injury and average monthly precipitation or temperature during the vegetation period at the geographical origin of accessions (Fig. 2).

Table 1.  Correlations between O3 sensitivity (expressed as O3-induced visible leaf injury or ion leakage) and different studied leaf characteristics in Arabidopsis thaliana accessions
Dependent variablePredictor variablenEquationR2P
Visible injuryWater loss, 2 h89y = 0.58*x − 12.480.23<0.0001
Visible injuryAbaxial gst70y = 0.023*x − 2.380.090.015
Visible injuryAdaxial gst70y = 0.040*x − 4.400.170.0004
Visible injuryIon leakage, 8 h O39y = 1.99*x − 11.210.740.0029
Ion leakage, 8 h O3Total rosette gst9y = 0.066*x + 0.780.680.0062
Ion leakage, 8 h O3JA, 8 h O39y = 0.017x + 5.450.630.010
Ion leakage, 8 h O3Initial O3 uptake rate9y = 0.43x − 0.410.510.031
Ion leakage, 8 h O3O3 dose, 30 min9y = 0.15x + 1.140.490.037
Ion leakage, 8 h O3O3 dose, 60 min9y = 0.086x + 2.090.370.084
Ion leakage, 8 h O3O3 dose, 90 min9Not significant
Ion leakage, 8 h O3O3 dose, 240 min9Not significant
Figure 2.

The relationship between O3-induced visible injury and average monthly precipitation (a) and average monthly temperature (b) during vegetation period at the geographical locations of studied Arabidopsis natural accessions.

Stomatal conductance, photosynthesis and hormone and antioxidant concentrations of nine accessions

Nine accessions with varying O3 sensitivity were selected for further studies. The selection was based on differences in O3 sensitivity, i.e. tolerant, intermediate and sensitive accessions were included, and the availability of genetic tools for some of these accessions (i.e. there are RIL mapping populations available for Col-0, Cvi-0, Ler-1, C24 and Te-0). O3-induced cellular damage can be quantified as ion leakage. Firstly, we analysed whether there was a correlation between O3-induced visible injury and cellular damage measured as changes in ion leakage. Indeed, visible injury and ion leakage upon 8 h O3 exposure were significantly correlated (Table 1). In further correlations, we used O3-induced ion leakage as the dependent variable.

Average whole-rosette stomatal conductance (gst) varied widely among the accessions ranging from 79 mmol m−2 s−1 in C24 to 535 mmol m−2 s−1 in Cvi-0 (Fig. 3). These gst values are averages for the whole rosette, including all leaves irrespective of their age, and petioles. Water loss from detached leaves is often used as an indirect measure for comparing stomatal openness and functionality of different Arabidopsis mutants. Cvi-0 leaves had clearly the highest water loss measured at 1 and 2 h after detachment from plants (Supporting Information Fig. S1). The whole-rosette values of net CO2 assimilation rate (Anet) at ambient CO2 and 150 µmol m−2 s−1 of PPFD were less variable, ranging between 2.4 and 4.3 µmol m−2 s−1. Anet was lower in accessions with low gst (C24 and Est-1), and higher in accessions with high gst (Ler-1 and Cvi-0). The correlation between whole-rosette stomatal conductance and O3-induced ion leakage was significant in the nine accessions; higher stomatal conductance led to higher O3-induced ion leakage (Table 1).

Figure 3.

Whole-rosette stomatal conductance (gst) and net photosynthesis (Anet) (average ± SE) of selected Arabidopsis accessions. Different capital and small letters denote significant differences (P < 0.05, n = 7–29) between accessions for gst and Anet, respectively.

Hormone levels were measured using a method that detects several hormones (SA, JA and ABA) from the same sample simultaneously, which minimizes sample handling and increases the possibility of detecting interactions between the hormones (Schmelz et al. 2003). O3 exposure and accession were both significant factors affecting stress hormone concentrations (Table 2, Fig. 4). There were no significant differences in untreated plants for ABA concentration (2 h control) between the accessions studied. In addition, 2 h of O3 did not have a significant effect on ABA concentration, whereas O3 exposure for 8 h increased ABA concentration in all accessions. However, this increase was not significant in the most O3-tolerant accessions Kin-0 and C24. The highest ABA concentration after 8 h of O3 exposure was detected in Te-0 (20-fold increase), followed by Tsu-1 and Kas-1 (15-fold increase).

Table 2.  Results of analysis of variance (anova) for the effects of accession and ozone exposure on hormone and antioxidant concentrations and redox state of antioxidants
 AccessionO3 exposureAccession*O3
  • *

    The significance (P value of the anovaF test) is indicated as: P < 0.05;

  • **

    P < 0.001;

  • ***

    P < 0.0001.

  • Ns, not significant.

Ion leakage*********
Abscisic acid (ABA)*********
Jasmonic acid*********
Salicylic acid (SA)******Ns
Ascorbate (AA)****Ns
Dehydroascorbate (DHA)**NsNs
Glutathione (GSH)******
Oxidized glutathione (GSSG)*NsNs
Figure 4.

Concentrations of abscisic acid (ABA) (a), jasmonate (JA) (b) and salicylic acid (SA) (c) (average ± SE, n = 4) in Arabidopsis plants treated with 300–350 nmol of O3 mol−1 for 6 h. * denotes a significant difference between 2 h control and corresponding O3 treatment.

In Tsu-1 and Cvi-0, JA concentration increased already after 2 h of O3 exposure, whereas an 8 h O3 exposure was needed for a significant increase in Ler-1, Kas-1 and Te-0 (Fig. 4). O3-tolerant accessions Kin-0, C24 and Col-0, but also O3-sensitive Est-1, showed no increase of JA upon O3 exposure. A significant increase in SA concentration was observed at 2 h O3 exposure in Ler-1 and Te-0, whereas all accessions showed increased SA after 8 h O3; however, in Kas-1 and Est-1 this was not significant.

When O3-induced ion leakage was correlated with 2 h control or O3-induced (8 h) hormone concentrations, only one significant correlation emerged: ion leakage was positively correlated with JA concentration after 8 h O3 exposure (Table 1).

Differences in AA and GSH concentrations and redox states (AA/DHA, GSH/GSSG) between accessions were also detected; the effect of O3 was significant for AA and GSH, but not for DHA and GSSG (Table 2). However, pair-wise comparisons of the corresponding data (Tukey HSD) indicated only a few significant differences. AA concentration in 2 h control of Te-0 was significantly higher compared to the other accessions (data not shown), whereas concentrations of DHA, GSH and GSSG were similar for all accessions. O3 exposure for 8 h significantly increased GSH concentration in Kas-1; however, the effect of O3 exposure on AA, DHA and GSSG concentrations was non-significant when the corresponding pairs were compared (data not shown).

Initial O3 uptake rate into the leaf and dose during the first 30 min of exposure are important for O3-induced leaf damage

Stomatal conductance and O3 uptake characteristics during O3 exposure were followed in the nine accessions selected for more detailed studies. Initial uptake rate of O3 through the stomata and cumulative dose of O3 absorbed by the plant during the first 30 min of O3 exposure were significantly correlated with O3-induced ion leakage (Table 1). Cumulative dose of O3 received during the first hour was also important, whereas the cumulative doses received during 90 and 240 min were not correlated with O3-induced ion leakage. As an example, Fig. 5 shows the time courses of stomatal conductance and O3 uptake characteristics in Cvi-0 and Col-0 during O3 exposure with 350 nmol mol−1 for 4 h. Stomatal conductance before the onset of O3 and initial stomatal uptake rate of O3 were about twofold higher in Cvi-0 (Fig 5a,b), leading to higher cumulative O3 dose received during the first 30 min and, further, to seriously injured leaves (Fig. 5c). Stomata of Cvi-0 closed rapidly under O3 and never recovered; gst and stomatal O3 uptake rate of Col-0 reached the values of Cvi-0 already at the end of the first hour of exposure. As a result, total cumulative O3 dose received by Col-0 was nearly two times larger, but this did not result in leaf injuries (Fig. 5c), because total O3 dose was not correlated with visible injury development (Table 1).

Figure 5.

Time-courses of stomatal conductance and ozone concentration (a), O3 uptake rate into the leaf through stomata and cumulative dose of O3 taken up by the leaf (b) (average ± SE) during 4 h O3 exposure with 350 nmol mol−1 for Cvi-0 (n = 9) and Col-0 (n = 10). Such treatment induced severe lesions in Cvi-0, whereas there was nearly no detectable damage in Col-0 (c).

QTL mapping of O3 sensitivity and water loss in a Col-0 × Cvi-0 RIL population

The O3 sensitivity, high stomatal conductance and water loss from excised leaves of Cvi-0 (Figs 1 & 3, and Supporting Information Fig. S1) made it an attractive choice for QTL mapping to identify molecular constituents of these traits. We screened the core 164 Col-0 × Cvi-0 RIL population (Simon et al. 2008) for O3 sensitivity and water loss phenotypes, and used QTLNetwork 2.0 (Yang et al. 2007) to map the major QTLs (see Materials and methods for details). Three QTLs for O3 sensitivity, and one QTL for water loss were identified (Fig. 6, Supporting Information Table S2). The strongest O3 QTL mapped to the same position as the only water loss QTL, further supporting the correlation between O3 injury and stomatal regulation (Table 1).

Figure 6.

Positions of O3 and water loss quantitative trait loci (QTLs) identified from Col-0 × Cvi-0 recombinant inbred line (RIL) population on the Arabidopsis genome. For QTL mapping, QTLNetwork 2.0 was used, and three QTLs for ozone sensitivity and one QTL for water loss exceeded the threshold F value of 11.7. For comparison, locations of known genes involved in the regulation of O3 responses (SLAC1, RCD1, RETICULATA, VTC1, SID2, NPR1, MPK6, JAR1, COI1, MPK3 ETO1, ETO3; Kangasjärvi et al. 2005) and stomatal regulation (SLAC1, OST1, ABI1, ABI2, CPK, CPK6, MRP5, GORK1, KCO1, AGB1, RBOHF, MPK3, MPK6, MYB61, RBOHD, GPA1, RCN1, RPK1, CPK4 and CPK11; Kwak et al. 2008) are also shown.


Our study revealed a remarkable amount of heritable genetic variation, evident as visible leaf injury, in O3 sensitivity among 93 Arabidopsis thaliana natural accessions. Similar high variation in O3 injury has previously been reported for 38 Medicago truncatula accessions (Puckette, Weng & Mahalingam 2007). In accordance with previous reports, accessions described as O3 sensitive (Cvi-0, Ws, Kas-1, Shakdara; Rao & Davis 1999; Wohlgemuth et al. 2002; Tamaoki et al. 2003) were moderately to very sensitive in our experiment (Fig. 1). The only exception to previous results was Est-1, classified as O3 tolerant by Tamaoki et al. (2003), was sensitive to O3 under our experimental conditions. Most O3-sensitive Arabidopsis mutants come from Col-0 or Ws background, and thus there is a large potential to obtain relevant information on O3 responses by studying other accessions. Understanding the complexes of traits and genes that increase stress tolerance may provide a way to reduce crop loss caused by O3. However, it must be noted that we used acute and relatively high O3 concentration. Recently, it has been shown that the mechanisms of damage resulting from acute and chronic O3 exposures are not identical (Chen, Frank & Long 2009). Therefore, our results are not directly applicable into chronic exposure situations plants face in the field. Still, they suggest the existence of high diversity of O3 sensitivity also in response to chronic treatment.

The lack of correlation between the climatic data of growing season at the geographical origin of accessions and visible O3 injury is partly expected. Plant damage caused by O3 is probably a rather new phenomenon rising from increased industrialization that cannot have been a major factor in the spread of Arabidopsis. In contrast, significant correlations have been found between traits like circadian parameters and day length (Michael et al. 2003) or freezing tolerance and habitat temperature (Hannah et al. 2006). In P. major, changes in O3 resistance can occur within a few generations (Davison & Reiling 1995), thus diminishing the influence of long-term factors. Because no trend between climatic factors, such as average temperature and precipitation, and O3-induced damage was evident (Fig. 2); these factors have not affected the variation in characteristics behind differential sensitivity to acute O3 exposure.

Plant hormones like ABA, SA, JA and ethylene are involved in determining the magnitude of O3 responses (Kangasjärvi et al. 2005). In a study with 20 Arabidopsis accessions, it was concluded that, in general, higher ethylene production correlated with the visible leaf injury; however, there were also O3-sensitive accessions (Cvi and NO) with low ethylene generation (Tamaoki et al. 2003). Treatment with methyl JA protects plants from O3 damage (Overmyer et al. 2000; Rao et al. 2000), and the use of ethylene biosynthesis or perception inhibitors protects plants from O3 damage (Overmyer et al. 2000; Tamaoki et al. 2003). In some Arabidopsis mutants, for example rcd1, O3-induced injuries are associated with hormone balance-related processes and hormone interactions. Thus, it is difficult to dissect exactly which hormone is mainly responsible for the O3 sensitivity (Overmyer et al. 2000, 2005; Ahlfors et al. 2004a). All accessions showed O3-induced changes in hormone concentrations in our experiment. We found that the larger the O3 injury, the higher the JA concentration after 8 h of O3 exposure. JA can attenuate the severity of oxidative stress (Creelman & Mullet 1997) and reduce O3 damage (Rao et al. 2000; Tuominen et al. 2004; Kangasjärvi et al. 2005); accordingly, high JA concentration could be required for attenuation of oxidative stress in the most O3-injured plants. Alternatively, JA biosynthesis is regulated by substrate availability (Wasternack 2007), and plants with extensive O3 damage could have an increase in the release of JA biosynthesis precursors, resulting in high JA levels.

SA plays an important role in plant O3 sensitivity (Rao & Davis 1999). Exposure to O3 increased SA concentrations of Col-0 and Cvi-0 five- to sixfold; the highest O3-induced SA accumulation (12-fold) was detected in Te-0 (Fig. 4). SA is proposed to play a dual role in O3 responses (Rao & Davis 1999); optimal concentrations of SA are required to achieve induction of defence responses, whereas high SA levels activate programmed cell death. Cvi-0 has been suggested to be O3 sensitive because of high SA accumulation combined with JA insensitivity, both processes which would promote increased cell death (Rao et al. 2000). Furthermore, reduction of SA levels in Cvi-0 by transgenic approaches reduced O3-induced cell death (Rao et al. 2000). In our experiments, increase of SA concentration because of O3 exposure was not considerably different in Cvi-0 compared to Col-0, possibly because of different growth conditions or methods for SA determination.

Increased concentrations of ABA stimulate stomatal closure (Assmann & Shimazaki 1999). ABA-insensitive mutants are characterized by higher stomatal conductance and slower stomatal closure under O3 compared to wild-type plants (Ahlfors et al. 2004a). Besides stomatal regulation, ABA may also participate in O3-induced cell death by antagonizing ethylene that is required for continuous ROS production and propagation of cell death (Kangasjärvi et al. 2005). No significant correlation between O3-induced injury and ABA concentration was detected in our experiments. Increase in ABA concentration upon O3 exposure was not significant in the two most O3-tolerant accessions, but in the others ABA concentration increased several-fold, with the highest accumulation detected in Te-0. Interestingly, O3-induced increase of SA was also most prominent in Te-0, while only minor changes in hormone concentrations were evident in C24. These two accessions also differed from others in a study of freezing tolerance, with Te-0 (an accession with origin in Finland) showing enhanced freezing tolerance in non-acclimated plants and after cold acclimation, whereas the freezing tolerance was lowest in C24 (an accession with origin in Portugal) (Hannah et al. 2006). Under control conditions, Te-0 had similar metabolic state as cold-acclimated summer accessions, such as C24, and higher expression of key regulatory transcription factors, CBFs, for cold acclimation (Hannah et al. 2006). Because ABA is a key hormone in cold responses (Chinnusamy, Zhu & Zhu 2006), the high accumulation of ABA in Te-0 following O3 responses could be related to its adaptation to growth in cold climates.

Stomatal conductance is important in determining the uptake of O3 into leaf (Kollist et al. 2000; Kangasjärvi et al. 2005), and is a key factor in O3 deposition models during chronic exposure (Musselman et al. 2006). Indeed, we found that gst, initial uptake rate of O3 into the leaf, and O3 dose accumulated during the first 30 and 60 min of O3 exposure were important in explaining the differences in O3 sensitivity. In contrast, total O3 dose received during a 4 h exposure did not correlate with O3-induced leaf injury. The most O3-sensitive accession, Cvi-0, had the highest gst, and O3-tolerant accession C24 had the lowest gst. Constitutively low gst leading to low photosynthetic rate could be one determinant for the low growth-related characteristics such as leaf area, relative growth rate and rosette weight of C24 found in a study of 24 accessions (Cross et al. 2006). However, all differences in O3 sensitivity cannot be explained by variation in stomatal conductance only. Ler-1 (Landsberg accession with erecta mutation), characterized by high stomatal conductance and reduced water-use efficiency (Masle et al. 2005), had high gst (second highest after Cvi-0; Fig. 3) in our experiment. However, Ler-1 was only moderately O3 sensitive. On the other hand, Est-1, characterized by low gst, was sensitive to O3 in terms of visible injury development. Thus, some accessions were more (Est-1) or less (Ler-1, Col-0) sensitive to O3 than could be expected from their gst and initial O3 uptake rate. Occasionally, we have observed that Est-1 develops lesions in control growth conditions (Supporting Information Fig. S2), similar to some lesion mimic mutants (Lorrain et al. 2003). Thus, Est-1 may have an altered regulation of cell death, which would influence its O3 response.

We have previously proposed a model where O3-induced cell death proceeds in three stages: induction, spreading and containment (Kangasjärvi et al. 2005). The results from the nine accessions suggest that induction of cell death is critically dependent on the initial O3 entry to the plant, that is, the dose during the first 30 min of O3 exposure (but not total dose of O3), which correlated with cell death. This suggests that there is a ‘threshold’ value of O3-induced oxidative stress that needs to be exceeded to initiate cell death. However, the contrasting results described above (Est-1, Ler-1, Col-0) indicate that other processes, such as containment of the cell death, are also important for the outcome of an O3 exposure. O3 detoxification in the leaf is another important factor in determining the plant O3 sensitivity (Conklin et al. 1996). In order to predict vegetation responses to chronic O3 exposure, the effective flux approach, accounting for stomatal uptake and intra-leaf detoxification, was suggested as the best concept (Musselman et al. 2006). Diffusion processes have sometimes been unable to explain the differences in O3 sensitivity, with O3 absorption rate being even higher in tolerant plants (Moldau et al. 1991). Effective detoxification can occur via antioxidant metabolites (AA and GSH) and/or various antioxidative enzymes; consequently, O3 injury may result from inability to induce antioxidative defences and maintain the redox state of GSH and AA (Rao & Davis 1999). Indeed, the low ascorbic acid mutant vtc1 was originally isolated as an O3-sensitive mutant soz1 (Conklin et al. 1996). We found, however, no correlations between AA and GSH concentrations or redox state and O3-induced injury (Table 2). This could be a consequence of our experimental set-up which assesses the significance of short high-concentration O3 pulses, which would act in the apoplastic space. Thus, total leaf antioxidant concentrations and redox capacity might be less significant under these circumstances than the apoplastic antioxidative systems.

Ultimately, to understand the genetic basis of O3 sensitivity, the molecular identity of the genes contributing to O3 sensitivity needs to be identified. Because O3 sensitivity was highest in Cvi-0, we initiated QTL mapping in this accession, taking advantage of a Col-0 × Cvi-0 RIL population (Simon et al. 2008). Using O3-induced leaf damage as the marker, three QTLs for the O3 sensitivity, and in a separate experiment, one QTL for water loss from excised leaves were revealed (Fig. 6). Previous preliminary results of QTL mapping in Cvi-0 placed the major O3 QTL close to COI1 and JAR1, similar to our results (Fig. 6; Rao et al. 2000). The current resolution of our QTL data excludes these two genes as candidates for O3 2–13. Interestingly, the QTL for water loss maps to the same position as the strongest QTL for O3 damage (O3 2–13); thus, it is likely that the correlation we see between O3 sensitivity and gst in Cvi-0 (Figs 1 & 3) is also reflected at the genetic level, and the identification of O3 2–13/WL 2–13 could reveal a new regulator of stomatal aperture. Furthermore, the QTL results give a strong indication that additional processes are involved in determining O3 sensitivity because the QTL O3 3–14 also has a strong effect on O3 damage. O3 2–13 maps close to MPK6, and O3 3–14 maps close to MPK3. Both these mitogen-activated protein (MAP) kinases are regulators of O3 responses (Ahlfors et al. 2004b; Miles et al. 2005), and also proposed regulators of stomatal function (Gudesblat, Iusem & Morris 2007; Xing, Jia & Zhang 2008). The signalling pathway(s) leading to stomatal closure is extremely complex (Kwak, Mäser & Schroeder 2008), and includes ethylene and downstream H2O2 produced through AtRBOHF (Desikan et al. 2006). AtRBOHF is located in the region for O3 1–18, and genes (ETO1 and ETO3) regulating ethylene levels are close to O3 3–14. Fine mapping of the QTLs, measuring ethylene levels and/or determining MAP kinase activity in Cvi-0, could reveal if they contribute to O3 sensitivity. As a model for future research, we propose that O3 2–13/WL 2–13 allows a higher entry of O3 into the plant, and subsequently O3 3–14 regulates a different aspect of plant defences, for example, hormonal interactions, antioxidant scavenging of O3 or regulation of cell death.

To conclude, the variation in sensitivity to acute O3 exposure was strikingly high among natural accessions, much higher than between O3-sensitive mutants and corresponding wild types. Stomatal conductance, initial uptake rate and accumulated O3 dose during the first 30 and 60 min of O3 exposure were all important in explaining the differences in O3 sensitivity between nine accessions. However, all differences in O3 sensitivity among accessions cannot be explained by variation in stomatal openness and initial uptake rate, indicating that also other processes are involved in determining plant O3 sensitivity. Similar results were recently reported for different winter wheat cultivars exposed to chronic O3 treatment (Biswas et al. 2008). Molecular identification of O3 QTLs from Cvi-0 are likely to reveal both a stomatal regulator and the identity of an ‘unknown’ process important for regulating plant defences to O3.


This study was supported by grants from The Estonian Science Foundation (projects no. 7763, 7361 and 6462), and by Estonian Ministry of Education and Research (theme SF0180071s07). Work in the lab of J.K. was supported by the Academy of Finland Centre of Excellence program (2006-11) and Helsinki University Environmental Research Centre. M.B. was supported by an Academy of Finland Post-doctoral grant (decision # 108760).