Ozone suppresses soil drying- and abscisic acid (ABA)-induced stomatal closure via an ethylene-dependent mechanism

ABSTRACT

Elevated atmospheric ozone concentrations (70 ppb) reduced the sensitivity of stomatal closure to abscisic acid (ABA) in Leontodon hispidus after at least 24 h exposure (1) when detached leaves were fed ABA, and (2) when intact plants were sprayed or injected with ABA. They also reduced the sensitivity of stomatal closure to soil drying around the roots. Such effects could already be occurring under current northern hemisphere peak ambient ozone concentrations. Leaves detached from plants which had been exposed to elevated ozone concentrations generated higher concentrations of ethylene, although leaf tissue ABA concentrations were unaffected. When intact plants were pretreated with the ethylene receptor binding antagonist 1-methylcyclopropene, the stomatal response to both applied ABA and soil drying was fully restored in the presence of elevated ozone. Implications of ethylene's antagonism of the stomatal response to ABA under oxidative stress are discussed. We suggest that this may be one mechanism whereby elevated ozone induces visible injury in sensitive species. We emphasize that drought linked to climate change and tropospheric ozone pollution, are both escalating problems. Ozone will exacerbate the deleterious effects of drought on the many plant species including valuable crops that respond to this pollutant by emitting more ethylene.

INTRODUCTION

Global tropospheric concentrations of ozone have increased over the last 30 years (Jaffe & Ray 2007) and will continue to do so, particularly in the Northern Hemisphere, as automobile and industrial nitrogen oxide (NOx) emissions increase (e.g. Meehl, Stocker & Collins 2007); ozone is a secondary pollutant formed from the action of sunlight on NOx. Tropospheric ozone is the most phytotoxic air pollutant. It accelerates visible foliar injury, senescence and abscission (e.g. Pell, Schlagnhaufer & Arteca 1997), and reduces growth, biomass production and reproductive capacity (e.g. Isebrands et al. 2001; Bassin, Volk & Fuhrer 2007; Dermody et al. 2008). It has been estimated to be responsible for losses of billions of dollars worth of crops (e.g. Murphy, Delucchi & McCubbin 1999; Ashmore, Toet & Emberson 2006; Morgan et al. 2006). Forests (e.g. Paoletti, Bytnerowicz & Andersen 2007) and other natural communities (e.g. Ashmore 2005) are also considered to be at risk of damage and reduced productivity. Here we present data that reveal another mechanism whereby elevated ozone concentrations seriously impair plant functioning, well before visible leaf injury becomes apparent. We show that ozone impacts on the capacity of stomata to respond to the plant hormone abscisic acid (ABA), thereby affecting the plant's capacity to regulate its water balance. This occurs at concentrations within current UK ambient maxima.

ABA, often termed ‘the drought hormone’, is generally considered the most important chemical regulator of stomatal functioning in plants, particularly when water is in short supply (Wilkinson & Davies 2002). Roots in drying soil synthesize and/or transport ABA to the shoot via the xylem (Davies & Zhang 1991). Here it induces stomatal closure via a well-established network of chemical messengers (e.g. Israelsson et al. 2006; Acharya & Assmann 2008). Not only is ABA involved in the response to soil drying, but it is also an essential mediator of stomatal responses to many other stress factors, such as vapour pressure deficit (VPD) and temperature stress in the aerial environment (Tardieu & Davies 1992; Xie, Wang & Williamson 2006; Wilkinson & Davies 2008). Any factor that prevents ABA from acting to close stomata could have a severe impact on plant water use and, at worst, on plant survival. Drought and tropospheric ozone pollution are concurrently escalating problems under climate change (see above and Bates et al. 2008).

Currently there is a lack of consensus on the effects of ozone on stomatal aperture (see Mansfield 1998; McAinsh et al. 2002; Kwak, Nguyen & Schroeder 2006; Mills et al. 2009). Several authors have described ozone-induced stomatal closure in well-watered plants (e.g. Castillo & Heath 1990; Wellburn 1994; Wittig, Ainsworth & Long 2007), perhaps mediated by effects on photosynthesis (e.g. Farage et al. 1991; Fiscus, Booker & Burkey 2005), direct guard cell ion channel modulation (e.g. Torsethaugen, Pell & Assmann 1999), altered guard cell calcium homeostasis (e.g. Clayton et al. 1999; Evans et al. 2005) or altered hormone production (Overmyer, Kollist & Tuominen 2008). However, contrasting evidence describes ozone-induced stomatal opening and, more frequently, suppression of stomatal closure in response to soil drying (e.g. Pearson & Mansfield 1993; Mansfield 1998; Robinson, Heath & Mansfield 1998; McLaughlin et al. 2007b; Mills et al. 2009) and in response to direct applications of ABA (Mills et al. 2009), as well as in response to other stresses. Such ‘sluggish’ responses of stomata in elevated ozone include those normally induced by changes in photosynthetic flux density (PFD; e.g. Barnes et al. 1990; Paoletti 2005; Grulke et al. 2007) and VPD (e.g. Grulke et al. 2007), as well as by soil moisture deficit. When severe water stress was imposed on individual ozone-pretreated leaves of ozone-sensitive species/mutants by excision from the plant and exposure of cut petioles to air, they exhibited enhanced water loss compared with ozone-untreated leaves (e.g. Paoletti 2005; Vahisalu et al. 2008; Mills et al. 2009), presumably because their stomata remained more open.

Ozone-induced stomatal closure seems to have become a generally accepted cross-species view in some quarters. It has, for example, been built into models (e.g. Sitch et al. 2007) which predict that the potential impact of ozone on atmospheric CO₂ concentrations might be to exacerbate the globally increasing trend (by reducing the capacity of vegetation to ‘mop up’ CO₂). Although ozone-induced stomatal closure is clearly a widespread phenomenon (e.g. Wittig, Ainsworth & Long 2007), its opposing effect to suppress stomatal closure in response to stress will have important consequences for water and CO₂ use in natural and agricultural systems. This may in fact comprise part of the mechanism whereby ozone induces reduced growth, visible injury and enhanced senescence, thereby impacting on plant productivity (McLaughlin et al. 2007a). This effect of ozone also needs to be emphasized.

The gaseous plant hormone ethylene has been implicated in some plant responses to ozone, for example, with regard to lesion development, abscission and senescence (see below). Ethylene has traditionally been regarded as a shoot growth inhibitor and a promoter of ripening, senescence and abscission (Abeles, Morgan & Saltveit 1992). Evidence is now emerging for a role for ethylene at the guard cell in the unpolluted plant, although much of this evidence is contradictory. It has been implicated in stomatal opening (e.g. Madhavan, Chrominiski & Smith 1983), inhibition of ABA-induced closure (e.g. Tanaka et al. 2005) and in closure (e.g. Desikan et al. 2006). It is involved in the control of stomatal opening by auxin and cytokinin (e.g. Merritt, Kemper & Tallman 2001; Tanaka et al. 2006). Sometimes it causes closure and suppression of closure in the same species but under different experimental systems (Desikan et al. 2006). Two lines of evidence in the literature led us to propose the hypothesis, investigated here, that an ozone-induced up-regulation of ethylene could be responsible for the observed reduction in stomatal sensitivity to ABA, drought and other closing stimuli described above. Firstly, it seems very clear that ozone can up-regulate 1-aminocyclopropane-1-carboxylic acid (ACC, the ethylene precursor) and/or ethylene in plants, which can be involved in ozone susceptibility (usually with respect to foliar injury or senescence responses – e.g. Overmyer, Brosché & Kangasjärvi 2003; Tamaoki et al. 2003; Sinn et al. 2004; Diara et al. 2005; Nunn, Anegg & Betz 2005). Secondly (and see above), Tanaka et al. (2005, 2006) have demonstrated that exogenously supplied ethylene, ethylene over-producing mutants and auxin- and cytokinin-induced ethylene generation has an antagonistic effect on ABA- and/or soil drying-induced stomatal closure in Arabidopsis thaliana epidermal peels and/or intact plants. Ethylene prevented stomatal closure. Could this effect be occurring in ozone-exposed plants? The auxin/cytokinin-mediated effect on the ABA response could be reversed when ethylene signalling, perception or biosynthesis was blocked (Tanaka et al. 2006).

In this study, we have exposed calcicolous Leontodon hispidus (Rough Hawkbit) plants to elevated ozone. This forb species is native to the UK and is predominantly found in upland grass sward. It has been identified as being particularly ozone-sensitive. The ozone concentration chosen represents predicted 2050 average background levels (70 ± 15 ppb or nL L⁻¹; see Meehl et al. 2007). This concentration is approximately 30 ppb above current UK average background concentrations, but lies within current UK ambient summer maxima (see Hayes et al. 2007). Control plants have been exposed to unadjusted (lower) ambient greenhouse ozone concentrations (10–35 ppb). We have investigated the stomatal response to ABA in elevated ozone and addressed the following questions: (1) what is the mechanism behind the change in sensitivity to ABA; (2) do plants in elevated ozone produce more ethylene; and (3) when ethylene is prevented from binding to its receptors using 1-methylcyclopropene (1-MCP), is the response of stomata to ABA or to soil drying restored to the level of sensitivity exhibited by control plants in ambient ozone?

MATERIALS AND METHODS

Plant material

Leontodon hispidus plants were raised in 10-cm-tall pots in John Innes No. 2 compost from plugs until the rosette contained at least five leaves of over 5.0 cm in length. This was carried out in a greenhouse under supplemental lighting (provided by 600 W sodium Plantastar lamps, Osram, Germany), giving a photoperiod of 16 h, with a variable day/night temperature. They were watered daily with tap water to the drip point, and every 2 weeks with nutrient solution. At the appropriate stage (see above), plants were transferred to one of four 1.0 m³ growth cabinets in a second greenhouse for ozone exposure (or to one of four identical cabinets without supplemental ozone – controls) for up to 5 weeks, at the end of which time plants comprised up to approximately 30 leaves. Growth cabinets were glass-topped, such that plants were exposed to supplemental lighting as described above, and air-conditioned to temperatures of 21 ± 2.0 °C. Plants were watered as described above. After 2–3 weeks with or without ozone in the growth cabinets, plants were transferred to larger 1.0 L pots.

Ozone exposure

Ozone was generated (TOGB1 generator, Ozonia Triogen, Glasgow, UK) from compressed air which had first been dried (air dryer model TPD50A, Ozonia Triogen). The gas was bubbled through distilled water to remove impurities. A manifold system was used from which four manually controlled flow meters could be manipulated to adjust the flow of gas directly into each of the four growth chambers. Excess ozone and exhaust ozone from the chambers was vented to the exterior of the building. Ozone concentrations in each chamber were measured using an ozone analyser (model 49c, Thermo Environmental Instruments Inc., MA, USA), and the flow of gas to each chamber was adjusted two to three times daily to achieve a concentration of 70 ± 15 ppb (nL L⁻¹) ozone. Unadjusted ozone concentrations in the four control chambers (which were also air-conditioned) were measured daily and varied from 10–35 ppb (nL L⁻¹). These concentrations are lower than those that would have been present in the air outside the greenhouse.

ABA application

Leontodon hispidus plants were sprayed daily for up to 3 d, between 9.00 and 10.00 h (unless otherwise stated), over the entire foliated region, with water or ABA (1.0–5.0 mmol m⁻³) after varying lengths of exposure to ozone gas (or to unadjusted air). It is assumed that the foliar spray penetrates through to the interior of the leaf via ingress through stomatal pores (see Wilkinson & Davies 2008). In some cases, the basal midrib of the measurement leaf was injected between 09.00 and 10.00 h on the day of measurement (unless otherwise stated), with water or ABA (0.2 cm³ water or 1.0–5.0 mmol m⁻³ ABA). The injection point was immediately sealed with lanolin. It is assumed that at least a portion of the injected solution penetrated to the xylem vessels of the leaf (see Wilkinson & Davies 2008).

1-MCP pretreatment

A sprayable formulation of 1-MCP was provided (3.8% active ingredient – a.i.) as a kind gift from Smart-Fresh, AgroFresh Inc, Spring House PA, USA. This was diluted by gentle agitation to 0.1 g L⁻¹ in a 0.05% (v/v) solution of wetting agent (Silwett L-77, De Sangosse Ltd, Cambridge, UK). Once made up, the solution was used within 5 min as 1-MCP is given off as a gas. The solution was immediately sprayed over the entire foliage of L. hispidus plants inside 0.8 × 0.4 m³ plastic boxes in a ventilated room. Boxes were closed and the plants were left for 24 h, after which time they were transferred to the appropriate ozone environment. Control plants were sprayed with 0.05% (v/v) Silwett L-77 only. One treatment of 1-MCP is effective at preventing ethylene from binding to its receptors for at least 1 month (see e.g. Sisler & Serek 2003).

Stomatal conductance measurements

Abaxial stomatal conductance (gs) was measured with a porometer (AP-4, Delta-T Devices Ltd, UK) in five to eight L. hispidus plants per treatment, in two to three of the most recently fully expanded leaves per plant. Where plants had been sprayed or injected with ABA or water, gs measurements were taken a minimum of 3 h later.

Transpiration bioassay

Leaves were detached from ozonated or control plants that had been placed in the dark for 1–2 hours (by covering the growth chambers with black cloths), and the basal leaf blade was re-cut under water before immediate transfer to the treatment solutions. These were 20 cm³ of water +/− 0.1–5.0 mmol m⁻³ ABA in 25.0 cm³ plastic vials covered in aluminium foil and sealed with parafilm to prevent evaporation from the solution surface. Leaves were introduced through slits in the foil so that blade bases were submerged, leaving a blade length of approximately 4–7 cm protruding above the foil surface. The vials containing the leaves were placed under lights (PFD 650 µmol m⁻² s⁻¹) at approximately 11.00 h and were weighed approximately every hour for up to 5 hours, after which time leaf area was measured. Water loss was expressed as mmol per unit leaf area per second.

Soil-drying experiments

For the experiment in which leaf ABA concentrations were measured, water was simply withheld from the droughted plants (5-leaf-stage in 1.0 L pots) for 3 and 6 d. For the 1-MCP experiments, larger plants (30-leaf-stage) in 1.0 L pots were watered to field capacity and weighed. The following day, the pots and plants were re-weighed. Half from each treatment combination were re-supplied with water to field capacity (well-watered), and the other half were only re-supplied with 50% of the water that would have been required to re-establish field capacity (droughted). Over the following 3 days, gs was measured in the morning (10.00 h), and pots were re-watered as described above at 12.00 h. Gs was re-measured at 14.30 h on some days. Soil moisture potentials were measured with a theta probe (Delta-T Devices Ltd, UK) and expressed gravimetrically.

Radio-immunoassay (RIA) for measurement of leaf tissue ABA concentration

Plants in 1.0 L pots were exposed to two ozone concentrations for 8 d. Half of these were watered daily while the soil in the remaining pots was allowed to dry for a further 6 d. Upon harvesting leaf tissue (one leaf from 7–9 plants per treatment), this was immediately frozen in liquid nitrogen. Frozen leaf tissue was freeze-dried for 48 h, finely ground and extracted overnight at 5.0 °C with distilled deionized water using an extraction ratio of 1:50 (gram dry weight : cm³ water). The ABA concentration of the extract was determined using a RIA following the protocol of Quarrie et al. (1988), using [G-³H] (±)-ABA at a specific activity of 2.0 TBq mmol⁻¹ (Amersham International, Bucks., UK), and the monoclonal antibody AFRC MAC 252 which is specific for (+)-ABA.

Gas chromatography method for measurement of ethylene generation from leaf tissue

For the determination of ethylene evolution rate, 1–2 g of fresh leaf tissue (approximately two to three leaves) was weighed and placed in 27 cm³ glass vials (Suba-Seal, SLS, Nottingham, UK) containing saturated filter paper, flushed for 1 min with fresh air from outside the laboratory prior to closure (Suba-Seal, SLS, Nottingham, UK) and incubated for 50–60 min under a lamp (200 µmol m⁻² s⁻¹ PFD). Using a disposable plastic syringe, a 1.0 cm³ headspace sample was withdrawn through the seal and manually injected into a gas chromatograph (6890N, Agilent Technologies UK Ltd, Wokingham, UK), fitted with a J&W HP-AL/S (50 m × 0.537 mm × 15.0 µm) column (HiChrom Ltd, Reading, UK). This was maintained for the first 5 min at 100 °C to resolve ethylene, and then ramped at 15 °C min⁻¹ to 150 °C and held for 1.5 min to drive off any water vapour introduced onto the column by sample injection. The carrier gas was helium at a flow rate of 5.7 mL min⁻¹, and detection was by flame ionization. The rate of ethylene evolution was calculated with reference to peak areas of known ethylene standards (99.995% minimum purity, BOC Special Gases, Manchester, UK) and corrected for tissue fresh weight and time in incubation.

Experimental design and statistical analysis

Data in each figure shows a representative experiment from one of 3–7 repetitions. Means and standard errors were calculated from data from single representative experiments (n = 5–24), and t-tests or Mann–Whitney Rank Sum tests were carried out to assess statistical significance (at P < 0.05, P < 0.01 and P < 0.001).

RESULTS

Effect of ozone on the stomatal response to applied ABA

When intact L. hispidus plants were sprayed with 2.5 mmol m⁻³ ABA once daily for three consecutive days, stomatal conductance (gs) was reduced as expected when plants were exposed to unadjusted control ozone concentrations; however, the effect of ABA to close stomata was suppressed in elevated ozone (Fig. 1a). This effect could be observed within 24 h–7 d (depending on plant age – not shown), and was still occurring after 29 d (longest exposure period imposed). Stomata were able to close normally in response to darkness under both ozone regimes. Figure 1b shows that when 1.0 mmol m⁻³ ABA was injected into leaves on intact eight-leaf-stage plants via the midrib, stomata in control plants were able to respond by closing, although stomata in plants exposed to elevated ozone remained more open. Supplying the ABA to the xylem of detached leaves via the cut petiole and measuring the rate of transpirational water loss from the stomata also revealed a similar effect (Fig. 1c). Stomata in leaves which had been detached from plants exposed to elevated ozone were less responsive to ABA after 2 weeks of exposure. In general, ozone had very little effect on gs/transpiration in the absence of added ABA.

Figure 1. The effect of ozone concentration [control (ctrl) = 10–35 ppb or nL L⁻¹, + ozone = 70 ± 15 ppb or nL L⁻¹] and duration of exposure on the response of stomatal conductance (gs: a, b) or transpiration rate (c) to abscisic acid (ABA) applied as a foliar spray (a: 2.5 mmol m⁻³ ABA, 20-leaf-stage) or injected via the leaf midrib (b: 0.2 cm³ 1.0 mmol m⁻³ ABA, 8-leaf-stage) or supplied via the cut petiole to the xylem of detached leaves (c: 0.5 mmol m⁻³ ABA, 25-leaf-stage) of/to pot-grown Leontodon hispidus plants. ABA was sprayed or injected daily at 09.00–10.00 h on the measurement day (a, b) and on the preceding 2 d (a). On one occasion, gs was measured in plants which had been kept in darkness for 2–3 hours (a). Gs was measured at approximately 14.00 h (a, b). Transpiration rates were measured in plants that had been exposed to ozone for 2 weeks (c). Data are means (a, n = 18; b, n = 5; c, n = 7) ± standard error. Significant differences from minus ABA controls are denoted by * (* = P < 0.05, ** = P < 0.01, *** = P < 0.001).

Effect of ozone on endogenous leaf tissue ABA concentrations

Figure 2 shows that ozone concentration had no effect on the endogenous ABA concentration measured in leaves detached from L. hispidus plants after 11–14 d of exposure, either in the presence or absence of a soil-drying regime. ABA concentrations increased as expected as soil dried.

Figure 2. Effect of ozone concentration (see Fig. 1) on mean abscisic acid (ABA) concentration (n = 7–9 ± standard error) measured in leaves detached from intact pot-grown 5-leaf-stage Leontodon hispidus plants after 11 and 14 d exposure, when the soil around the plants was either kept well-watered (WW), or allowed to dry out (DD). Significant differences between WW and DD treatments are denoted by * (P = <0.05). Gravimetric soil moisture contents (g g⁻¹, n = 8) were 0.633 ± 0.033 (WW) and 0.482 ± 0.052 (DD) on day 3, and 0.643 ± 0.023 (WW) and 0.292 ± 0.011 (DD) on day 6.

Effect of ozone on the generation of gaseous ethylene from detached leaf tissues

Exposure of intact L. hispidus plants to elevated ozone increased the amount of ethylene generated by leaves detached from the plants after 9 d of exposure (Fig. 3) up to 16 d of exposure (not shown), both in the presence and absence of 3.0 mmol m⁻³ ABA applied as a foliar spray for three consecutive days prior to measurement. Ethylene emission was not increased after 4 d of exposure to elevated ozone concentrations (not shown), but neither was the reduced stomatal response to ABA yet apparent. Ethylene concentrations generated by the leaf tissue varied between sampling days (not shown), perhaps due to variations in the aerial microclimate, as experiments were carried out in glasshouses in which light intensity was greatly affected by external ambient conditions.

Figure 3. The effect of exposing intact pot-grown 20-leaf-stage Leontodon hispidus plants to ozone at elevated and control concentrations (see Fig. 1) for 9 d, and spraying them with abscisic acid (ABA; 3.0 mmol m⁻³) or water once daily for the preceding 3 d, on the generation of gaseous ethylene by leaf tissue after detachment from the plants. Data are means (n = 5–6) ± standard error. Significant differences between control and elevated ozone plants are denoted by * (* = P < 0.05, ** = P < 0.01).

1-MCP restores the sensitivity of stomata to ABA in elevated ozone

In the absence of ABA, gs was unaffected by pretreatment of L. hispidus plants with 1-MCP, both in ambient and elevated concentrations of ozone (Fig. 4a– 5 d of exposure). Data shown in Fig. 4b are from plants which had been sprayed with 3.0 mmol m⁻³ ABA for three consecutive days and which had been exposed ± elevated ozone for 8 d. As described above (Fig. 1), ABA reduced gs in ambient ozone but did so less sensitively in elevated ozone concentrations (compare Fig. 4a, b). However, 1-MCP pretreatment fully restored the response to ABA in elevated ozone; stomata in ABA-treated plants were as closed as they were in ambient ozone. This effect was reversible: stomata re-opened both in control- and 1-MCP-pretreated plants 8 d after ABA spraying had ceased, under both ambient and elevated ozone concentrations (not shown).

Figure 4. The effect of pretreating intact pot-grown 25-leaf-stage Leontodon hispidus plants with 1-methylcyclopropene (1-MCP; 0.1 g L⁻¹ of a 3.8% a.i. w/w powdered solid) on the response of stomatal conductance (gs) of plants grown at two ozone concentrations to abscisic acid (ABA; a, −ABA; b, +3.0 mmol m⁻³ ABA). Plants were exposed to control and elevated ozone concentrations (see Fig. 1) for 5 (a) and 8 d (b), respectively, and plants were sprayed with water or ABA on the measurement day and on the two preceding days (b). Data are means (n = 24) ± standard error and significant differences between −ABA and +ABA data points (i.e. between a and b) are denoted by * (b only: * = P < 0.05, ** = P < 0.001).

1-MCP restores the sensitivity of stomata to drying soil in elevated ozone

Plants supplied daily with only 50% of their potential transpirable water (soil drying, DD) closed their stomata in ambient ozone as expected, both in the presence and absence of 1-MCP, compared with plants supplied daily with 100% of their potential transpirable water (well-watered, WW –Fig. 5a). However, stomata in plants growing in elevated ozone (Fig. 5b– 14 d of exposure on day 1 of experiment) closed to a much smaller extent as the soil dried (compare red bars in Fig. 5a,b), although 1-MCP pretreated plants showed reductions in gs in response to soil drying that were as sensitive as those observed in ambient ozone.

Figure 5. The effect of pretreating intact pot-grown 30-leaf-stage Leontodon hispidus plants with 1-methylcyclopropene (1-MCP; as Fig. 4) on the response of stomatal conductance (gs) of plants grown at two ozone concentrations (a, control; b, 70 ± 15 ppb) for 14–16 d to a soil-drying treatment (coloured bars). Data are expressed as percentages of the well-watered (WW) control treatment, and are means (n = 14–16) ± standard error. Significant differences of soil drying (DD) treatments from the appropriate WW treatment are denoted by * (* = P < 0.05, ** = P < 0.01, *** = P < 0.001). Gravimetric soil moisture contents (g g⁻¹, n = 12–20) were 0.531 ± 0.016 (WW) and 0.441 ± 0.023 (DD) on day 1 (pm), and 0.511 ± 0.01 (WW) and 0.327 ± 0.015 (DD) on day 3.

DISCUSSION

We have demonstrated that stomata in plants continuously exposed to elevated ozone concentrations for at least 24 h exhibit a reduced sensitivity to exogenously supplied ABA (Figs 1 & 4, Mills et al. 2009). The ozone concentration that was used (70 ppb) is within the range of current ambient UK summer maxima (see Hayes et al. 2007) and approximates predicted 2050 mean levels (Meehl et al. 2007). These data explain why elevated ozone (as a result of changes in ambient concentrations, or experimentally imposed) has often been reported to reduce stomatal closure in response to stresses that normally act via ABA, such as soil drying (e.g. Fig. 5, Pearson & Mansfield 1993), high VPD (e.g. Paoletti 2005; Grulke et al. 2007), high light intensity (e.g. Grulke et al. 2007), high salinity (Robinson et al. 1998) and severe dehydration following leaf detachment (e.g. Mills et al. 2009). Heggestad et al. (1985) determined that soil moisture stress and ambient ozone independently reduced field-grown soybean yield by just 4 and 5%, respectively, but when experienced together, yield reduction increased to 25%. We have shown that stomatal conductance (gs) exhibits a suppressed response to soil drying under elevated ozone concentrations in L. hispidus (Fig. 5– compare red bars) and other UK upland grass sward species (unpublished work), such that plants lose more water as soil dries when exposed to elevated ozone concentrations, potentially explaining the yield response described by Heggestad et al. (1985). We emphasize that although UK average concentrations are below the threshold for these effects, current peak UK concentrations are now frequently high enough, at least between April and October (see e.g. Hayes et al. 2007), to restrict stomatal responses to stress-induced ABA. Because we know that the effect of elevated ozone on stomatal sensitivity to ABA can arise after just 24 h exposure in some cases (Fig. 1b), and that ambient peaks often occur on several consecutive days (see Hayes et al. 2007), vegetation is likely to be at risk with respect to increased sensitivity to drought for a large part of the summer season over a large portion of the northern hemisphere (Jaffe & Ray 2007). This has the potential to affect many species (see below). Ozone is already believed to be responsible for billions of dollars worth of crop losses at current levels, and climate change predictions are for further global increases in tropospheric ozone concentrations and incidences of drought over the coming decades.

We also demonstrated that L. hispidus plants exposed to elevated ozone concentrations produced more ethylene (Fig. 3), as has been determined for other species (see above and below). This occurred without any change in endogenous leaf tissue ABA concentration (Fig. 2), and both in the presence and absence of applied ABA. Ethylene concentrations were similar to those determined for other species (e.g. Fiorani et al. 2002). ABA concentrations were in the lower range of those determined in other species (e.g. Corlett, Wilkinson & Thompson 1998; Wilkinson & Davies 2008), and they increased as expected as soil dried (see Wilkinson & Davies 2002). Overmyer et al. (2008) showed that ozone increased ABA concentrations in Arabidopsis leaf tissue, although the ozone concentrations employed were more than threefold greater than those used here. Our data do not support a role for changes in endogenous ABA concentrations in the ozone-induced loss of stomatal sensitivity to ABA. Given that stomatal sensitivity to ABA is fully restored in the presence of ozone plus 1-MCP (Figs 4 & 5), it is also unlikely that changes in epidermal cell turgor arising from membrane damage (see Mansfield 1998) were involved.

When ethylene was prevented from binding to its receptors with 1-MCP, the sensitivity of stomata to added ABA or to soil drying was fully restored (Figs 4 & 5), demonstrating that ethylene mediated the ozone-induced reduction in stomatal sensivity to ABA. Emission of ethylene by plants exposed to elevated ozone has long been known to be involved in the induction of the visible injury response (e.g. Mehlhorn, O'Shea & Wellburn 1991), but this is the first time that ethylene has been shown to be involved in ozone-induced impairment of stomatal regulation. Given that many species react to elevated ozone by producing more ethylene, we predict that these will be the species that will demonstrate an increased susceptibility to ABA-mediated stresses such as soil drying as ozone concentrations increase over the next decades. Many of these are commercially important species, or important crops in world agriculture. These include pea and pinto bean (Mehlhorn et al. 1991), potato (Sinn et al. 2004), tomato (Bae et al. 1996), snap bean (Elagoz & Manning 2005) and some wheat cultivars (Tiwari, Agrawal & Manning 2005). Different species, and even different accessions/cultivars within a species, respond to ozone by emitting ethylene to widely differing extents (e.g. Tamaoki et al. 2003; Diara et al. 2005), or by not up-regulating its emission at all. This may explain some of the variability of effects of ozone on stomatal aperture described in the literature (see above). Recently, Overmyer et al. (2008) demonstrated that several Arabidopsis mutants generated ethylene to differing extents in elevated ozone, and that the ozone-sensitive rcd3/slac1 mutant, which exhibited constitutively more open stomata, also exhibited high ethylene production.

As described above, ethylene has been implicated in both stomatal opening and closure in well-watered unpolluted plants, and it will be important to determine the basis of this apparent dual response system (Desikan et al. 2006; Kwak et al. 2006; Acharya & Assmann 2008), as this may also explain differential stomatal responses to ozone. Because ethylene induced closure in intact Arabidopsis leaves, but opening in the presence of ABA in epidermal peels, Desikan et al. (2006) suggest a requirement for intercellular communication during the closure response. We propose here that under oxidizing environments ethylene antagonizes the stomatal response to ABA (Figs 4 & 5), and that this may not occur under less stressful conditions. This model is schematically represented in Fig. 6. Atkinson, Wookey & Mansfield (1991) demonstrated that the oxidizing agents SO₂ and NO₂ also reduced stomatal sensitivity to ABA in spring barley. The data of Overmyer et al. (2008) are also consistent with a promotion, by oxidizing environments, of ethylene ABA antagonism at the guard cell. Parallels can be drawn with the effect of ethylene on plant growth. Ethylene production is involved in the reduction of leaf area expansion and root growth under stress (e.g. Hussain et al. 1999; Sharp 2002; Sobeih et al. 2004). However, more recently, low doses of ethylene have also been shown to stimulate rather than to inhibit growth under some circumstances (Lee & Reid 1997; Fiorani et al. 2002; Pierik et al. 2006).

Figure 6. Schematic representation of the proposed signal transduction model for interactions between abscisic acid (ABA) and ethylene at the guard cell in the presence and absence of oxidative stress. The ‘?’ represents the hypothetical position at which ethylene plus oxidative stress inhibit the stomatal response to ABA. NO, nitric oxide; ROS, reactive oxygen species.

Short-term (2–4 h) exposures to high (100 ppb plus) ozone concentrations have been shown directly (Torsethaugen et al. 1999) or indirectly (Castillo & Heath 1990; Kollist et al. 2007) to involve modulation of guard cell ion channel activity in relation to ozone-induced stomatal closure. These effects are presumably distinct from those which we describe here. It will also be important to examine ion channel activity associated with ozone-induced loss of stomatal sensitivity to ABA. Because we already know that auxin can open stomata via the up-regulation of ethylene (e.g. Tanaka et al. 2006), it is possible that ethylene may modify the activity of guard cell ion channels/pumps which have already been shown to be responsive to both auxin and ABA (see Israelsson et al. 2006; Pandey, Zhang & Assmann 2007). Thus, possible candidates for the locus for the interaction between ABA and ethylene at the guard cell include potassium uptake channels, ion efflux channels and H⁺-ATPases. Recently, elevated concentrations of nitric oxide (NO) have been shown to inhibit K⁺ efflux at the guard cell and to induce stomatal opening, mimicking effects of oxidizing agents (Sokolovski & Blatt 2004). We emphasize again a link between oxidizing environments and open stomata/stomatal insensitivity to ABA (Fig. 6). It must be noted, however, that NO is better known in its opposing role at lower concentrations as part of the signal transduction chain leading to ABA-induced stomatal closure via H₂O₂ production (see Fig. 6 and Neill et al. 2008).

As described above, this work has relevance and implications for hormone signalling in general. The study is among the first (along with Tanaka et al. 2005, 2006) to show ABA ethylene antagonism in adult plant tissues (although Ghassemian et al. 2000 found ethylene to be a negative regulator of ABA responses in Arabidopsis during germination), and the first in plants under oxidative stress. Importantly reactive oxygen species (ROS) such as H₂O₂ are key signalling molecules in ABA-induced stomatal closure in unstressed plants (see Desikan et al. 2006; Acharya & Assmann 2008, Fig. 6). The ozone stress response is also characterized by an induction of ROS including H₂O₂, in the form of a much bigger oxidative burst in planta (see Overmyer et al. 2008). Our finding that ethylene antagonizes the stomatal ABA response under oxidative stress, whereas ozone-induced ROS are directly implicated in ozone-induced stomatal closure (Kollist et al. 2007), indicates that ethylene (plus oxidative stress) suppresses the action of ABA downstream of ROS. This hypothesis is depicted in Fig. 6. Interestingly, the opposing effect of ethylene to mediate stomatal closure in unstressed plants is also dependent on ROS, specifically H₂O₂, generated by the NADPH oxidase AtrbohF (Desikan et al. 2006). It will be important to determine how the guard cell responds differently to ethylene as opposed to ethylene plus oxidative stress. The difference may lie in the concentrations of ROS present at/in the guard cell: greater concentrations will be present under oxidative stress, presumably also increasing NO concentrations. We have already noted that NO apparently induces opposing stomatal responses depending on its concentration (compare Neill et al. 2008 and Sokolovski & Blatt 2004).

Our data suggest that it will be important to re-examine some of the basic stomatal responses in which both ABA and ethylene are implicated: for example, as leaves senesce ethylene is generated (Abeles et al. 1992) and the stomatal response to ABA is suppressed (Atkinson, Davies & Mansfield 1989), and flood-sensitive plants up-regulate ethylene and exhibit more open stomata (Jackson 2002).

Given (1) the established role of ethylene in ozone sensitivity and visible injury (see above and Overmyer et al. 2003) and (2) the observation that ozone sensitive Arabidopsis mutants (slac1) exhibit more ozone-induced visible injury, constitutively more open stomata, reduced stomatal sensitivity to ABA, and more ethylene production (Overmyer et al. 2008; Vahisalu et al. 2008), the ABA ethylene antagonism observed here in elevated ozone is likely to have implications for ozone sensitivity beyond the stomata. More open stomata will undoubtedly give rise to wilt-prone plants with reduced water potentials, at least under soil drying and/or high VPD. This will reduce growth and promote foliar damage and senescence. More open stomata will also predispose plants to further visible ozone damage by allowing increased ozone flux into the leaves during subsequent exposure (see also Overmyer et al. 2008; Vahisalu et al. 2008; Mills et al. 2009). Our data beg the question: are many of the visible symptoms of ozone damage the result of ethylene-induced impairment of stomatal closure?

ACKNOWLEDGMENTS

This work was supported by DEFRA, project AQ3510. We are very grateful to AgroFresh Inc, Spring House, PA, USA (a subsidiary of Rohm and Haas) for supplying the sprayable 1-MCP. The authors thank Dr Julian Theobald for assistance with the analysis of ethylene, Dr Harun Kaman for assistance with stomatal conductance measurements, and Maureen Harrison, David Andrew and Philip Nott for valuable technical support.