Combined Acute Ozone and Water Stress Alters the Quantitative Relationships between O3 Uptake, Photosynthetic Characteristics and Volatile Emissions in Brassica nigra

Ozone (O3) entry into plant leaves depends on atmospheric O3 concentration, exposure time and openness of stomata. O3 negatively impacts photosynthesis rate (A) and might induce the release of reactive volatile organic compounds (VOCs) that can quench O3, and thereby partly ameliorate O3 stress. Water stress reduces stomatal conductance (gs) and O3 uptake and can affect VOC release and O3 quenching by VOC, but the interactive effects of O3 exposure and water stress, as possibly mediated by VOC, are poorly understood. Well-watered (WW) and water-stressed (WS) Brassica nigra plants were exposed to 250 and 550 ppb O3 for 1 h, and O3 uptake rates, photosynthetic characteristics and VOC emissions were measured through 22 h recovery. The highest O3 uptake was observed in WW plants exposed to 550 ppb O3 with the greatest reduction and poorest recovery of gs and A, and elicitation of lipoxygenase (LOX) pathway volatiles 10 min–1.5 h after exposure indicating cellular damage. Ozone uptake was similar in 250 ppb WW and 550 ppb WS plants and, in both treatments, O3-dependent reduction in photosynthetic characteristics was moderate and fully reversible, and VOC emissions were little affected. Water stress alone did not affect the total amount and composition of VOC emissions. The results indicate that drought ameliorated O3 stress by reducing O3 uptake through stomatal closure and the two stresses operated in an antagonistic manner in B. nigra.


Introduction
Tropospheric O 3 is a phytotoxin and considered as one of the most damaging greenhouse gases; it is mainly formed via photochemical reactions involving nitrogen oxides (NO x ) and volatile organic compounds (VOCs) [1][2][3]. According to recent studies, tropospheric O 3 concentration is beginning to stabilize [4][5][6]; nevertheless, the current O 3 levels reduce plant photosynthesis and growth rates and decrease potential agricultural productivity in many areas of the world [7][8][9]. Average O 3 concentrations vary between 14.5 and 70.1 ppb during the growing season in the forests of Europe [10], but during heat waves O 3 concentrations can be even higher [11,12].
The impact of O 3 is often studied using short-term acute treatments when plants are exposed to high O 3 for several minutes to some hours, and resultant visible damage symptoms on leaves and modifications in photosynthetic activity are monitored [13][14][15]. O 3 enters in the leaves through stomata and the O 3 exposure inhibits stomatal conductance to water vapor (g s ) [16] and leads to reductions in net assimilation rate (A), whereas plants may or may not recover depending on the O 3 dose [17]. O 3 also activates the production of reactive oxygen species (ROS) [18,19], which in turn activate various defense reactions Next to Arabidopsis thaliana, the classical brassicaceous model, another species, black mustard (Brassica nigra (L.) W. D. J. Koch) is giving valuable knowledge to the glucosinolate pathway [64][65][66][67]. Brassica nigra is a large plant, 0.5-2 m tall with highly competitive capacity in rural sites; it has a more complex genome and greater tolerance to several environmental stresses than A. thaliana [66][67][68]. In a previous study, B. nigra had surprisingly high heat stress tolerance and a complex heat stress response [47]. In particular, the heat stress response in B. nigra was characterized by different responses of constitutive and stress-induced specialized and generic volatiles and photosynthetic characteristics, including major emissions of glucosinolate breakdown products [47].
The aims of the current study were to investigate how acute and moderate O3 exposures alter foliage photosynthetic characteristics and VOC emissions and whether drought modifies O3 effects in an interactive manner. We hypothesized that (1) acute O3 treatment leads to major reductions in photosynthetic characteristics (A, gs); (2) emissions of specific glucosinolate breakdown products and LOX compounds are enhanced in the recovery phase; (3) the rates of these volatile emissions depend on stomatal O3 uptake; (4) drought and O3 affect plant volatile emissions in an interactive manner.
The highest O3 uptake was observed in WW plants exposed to 550 ppb O3 at all timepoints, at the beginning (ca. 5 min from the start of fumigation), in the middle (0.5 h after start of fumigation) and at the end (1 h after the fumigation), reaching 317 ± 25 µmol m −2 by the end of the treatment (Figure 1). O3 uptake of WW plants exposed to 250 ppb O3 and WS plants exposed to 550 ppb O3 was similar ( Figure 1). O3 uptake in these plants varied by only around 16 to 23 µmol m −2 at the beginning of the fumigation and around 100 to 140 µmol m −2 at the end of the fumigation (Figure 1; O3 uptake was significantly greater for WW plants exposed to 550 ppb O3 than in the plants in the other two treatments).  At different times during the exposure period, average (±SE) surface O 3 uptake rate (including uptake by surface and ozone quenching by plant-emitted reactive hydrocarbons) of WW plants exposed to 550 ppb O 3 (41 ± 5 nmol m −2 s −1 ) for the whole exposure period was higher than that of WW plants exposed to 250 ppb O 3 (13 ± 1 nmol m −2 s −1 ; Table 1). For WS plants exposed to 550 ppb O 3 , the surface O 3 uptake rate was intermediate (25.3 Table 1). Table 1. Average surface and stomatal O 3 uptake rates and relative uptake rates at the beginning (~5 min), in the middle (0.5 h) and at the end (1 h) of fumigation with 250 ppb or 550 ppb O 3 for well-watered (WW) and water-stressed (WS) Brassica nigra plants. Stomatal O 3 uptake rate in WW plants exposed to 550 ppb O 3 was higher than in the other two treatments and reached up to 53 ± 4 nmol m −2 s −1 ( Figure 2, Table 1). Stomatal O 3 uptake rates of WW and WS plants exposed to 250 and 550 ppb O 3 ranged from 12 to 17 nmol m −2 s −1 through the exposure period (Table 1). Within treatments, there were only minor time-dependent variations in surface and stomatal O 3 uptake rate and the ratios of surface and stomatal O 3 uptake rates to whole leaf uptake rate (Table 1). O 3 exposure concentration increased surface and stomatal O 3 uptake rates through all treatments, but the share of surface vs. stomatal uptake rate did not differ significantly among O 3 concentrations (Table 1). Strong positive relationships between the stomatal O 3 uptake and g s were observed for both 250 ppb and 550 ppb O 3 fumigations (p < 0.001; Figure 2).

Changes in Photosynthetic Characteristics Upon Ozone Exposure
O3 exposure reduced A in all cases (significant O3 effect in Figure 3a) with the reduction observed already 10 min after the exposure to O3, followed by a more gradual reduction between 10 min and 4.5 h, and a minor recovery between 7.5 and 22 h for WS plants exposed to 550 ppb O3 and WW plants exposed to 250 ppb O3, whereas A continuously decreased in WW exposed to 550 ppb O3 (significant time and O3 x time effects in Figure  3a). The reduction was greater in plants that took up more O3 (Table 1). In particular, at the time point of maximum photosynthetic reduction at 4.5 h, the reduction in A was greater in WW plants exposed to 550 ppb O3 than that in WS plants exposed to 550 ppb and greater than in WW plants exposed to 250 ppb O3 (Figure 3a).   Figure 3A) with the reduction observed already 10 min after the exposure to O 3 , followed by a more gradual reduction between 10 min and 4.5 h, and a minor recovery between 7.5 and 22 h for WS plants exposed to 550 ppb O 3 and WW plants exposed to 250 ppb O 3 , whereas A continuously decreased in WW exposed to 550 ppb O 3 (significant time and O 3 x time effects in Figure 3A). The reduction was greater in plants that took up more O 3 (Table 1). In particular, at the time point of maximum photosynthetic reduction at 4.5 h, the reduction in A was greater in WW plants exposed to 550 ppb O 3 than that in WS plants exposed to 550 ppb and greater than in WW plants exposed to 250 ppb O 3 ( Figure 3A).

Changes in Photosynthetic Characteristics Upon Ozone Exposure
O3 exposure reduced A in all cases (significant O3 effect in Figure 3a) with the reduction observed already 10 min after the exposure to O3, followed by a more gradual reduction between 10 min and 4.5 h, and a minor recovery between 7.5 and 22 h for WS plants exposed to 550 ppb O3 and WW plants exposed to 250 ppb O3, whereas A continuously decreased in WW exposed to 550 ppb O3 (significant time and O3 x time effects in Figure  3a). The reduction was greater in plants that took up more O3 (Table 1). In particular, at the time point of maximum photosynthetic reduction at 4.5 h, the reduction in A was greater in WW plants exposed to 550 ppb O3 than that in WS plants exposed to 550 ppb and greater than in WW plants exposed to 250 ppb O3 (Figure 3a).  Differently from A, g s showed a decreasing trend only in WW plants exposed to 550 ppb O 3 , although there was a globally significant O 3 effect across all treatments ( Figure 3B). In WS plants exposed to 550 ppb O 3 , g s changed little with a significant, but minor reduction, only observed at 7.5 h after O 3 treatment ( Figure 3B). Taken together, 550 ppb O 3 exposure effect on g s was greater for WW than for WS plants ( Figure 3B). In the case of WW plants exposed to 250 ppb, g s increased during recovery, and g s was greater than that in control leaves at the end of the measurements ( Figure 3B). Both WW plants exposed to 250 and 550 ppb O 3 had a higher C i /C a at 22 h recovery point than the controls ( Figure 4A). C i /C a in WS plants exposed to 550 ppb was almost constant with a minor reduction at 7.5 h after treatment ( Figure 4A). During the recovery period, C i /C a ratio was positively correlated with g s in WW plants exposed to 250 ppb O 3 and WS plants exposed to 550 ppb O 3 ( Figure 4B). A values during recovery also correlated positively with C i /C a ratio in WW and WS plants exposed to 550 ppb O 3 (p = 0.02 for both; p = 0.09 for WW plants exposed to 250 ppb O 3 ; data now shown). Differently from A, gs showed a decreasing trend only in WW plants exposed to 550 ppb O3, although there was a globally significant O3 effect across all treatments ( Figure  3b). In WS plants exposed to 550 ppb O3, gs changed little with a significant, but minor reduction, only observed at 7.5 h after O3 treatment (Figure 3b). Taken together, 550 ppb O3 exposure effect on gs was greater for WW than for WS plants (Figure 3b). In the case of WW plants exposed to 250 ppb, gs increased during recovery, and gs was greater than that in control leaves at the end of the measurements (Figure 3b). Both WW plants exposed to 250 and 550 ppb O3 had a higher Ci/Ca at 22 h recovery point than the controls (Figure 4a). Ci/Ca in WS plants exposed to 550 ppb was almost constant with a minor reduction at 7.5 h after treatment (Figure 4a). During the recovery period, Ci/Ca ratio was positively correlated with gs in WW plants exposed to 250 ppb O3 and WS plants exposed to 550 ppb O3 ( Figure 4b). A values during recovery also correlated positively with Ci/Ca ratio in WW and WS plants exposed to 550 ppb O3 (p = 0.02 for both; p = 0.09 for WW plants exposed to 250 ppb O3; data now shown).

Drought and O 3 Impacts on Total Volatile Emissions and Emissions of Different Volatile Groups
In control plants, the total VOC emissions were dominated by saturated aldehydes and geranyl diphosphate (GDP, monoterpenes) or geranylgeranyl diphosphate (GGDP) pathway compounds (carotenoid breakdown products); most WW and WS plants before O 3 treatment showed similar total VOC emission rates ( Figure 5). O 3 exposure and water stress (for plants exposed to 550 ppb O 3 ) modified the share of different volatile groups depending on treatment ( Figure 6, see below), but total emissions were not affected ( Figure 5).
O 3 exposure affected all groups of volatiles except glucosinolate breakdown products and saturated aldehydes, but the effects were treatment-specific ( Figure 6). LOX compound emissions in WW plants exposed to 250 ppb O 3 at 10 min and 22 h measurements were enhanced compared to control leaves ( Figure 6A). In 550 ppb exposed plants, the LOX compound emissions were greater in WW than in WS plants through the recovery ( Figure 6A). For GDP and GGDP compound emissions, O 3 enhanced emissions, but the emissions at different recovery time points differed little ( Figure 6C,D). Only in WW plants exposed to 250 ppb O 3 , the emission of GGDP compounds was greater at 10 min recovery time point compared with the control and other recovery times ( Figure 6D). Water stress inhibited GDP and GGDP compound emissions in plants exposed to 550 ppb O 3 ( Figure 6C,D). For saturated aldehydes, the treatments generally had a minor effect. Saturated aldehydes emission was only enhanced in WW plants exposed to 250 ppb O 3 right after O 3 exposure, and it decreased in WS plants exposed to 550 ppb O 3 22 h after treatment ( Figure 6E).

Drought and O3 Impacts on Total Volatile Emissions and Emissions of Different Volatile Groups
In control plants, the total VOC emissions were dominated by saturated aldehydes and geranyl diphosphate (GDP, monoterpenes) or geranylgeranyl diphosphate (GGDP) pathway compounds (carotenoid breakdown products); most WW and WS plants before O3 treatment showed similar total VOC emission rates ( Figure 5). O3 exposure and water stress (for plants exposed to 550 ppb O3) modified the share of different volatile groups depending on treatment ( Figure 6, see below), but total emissions were not affected (Figure 5).  . Emission rates within treatments at different time-points during recovery were compared by paired sample t-tests, and significant differences are indicated by lowercase letters (p < 0.05). The effect of O3 exposure, recovery time and their interaction and the effect of water stress (WS vs. WW) for plants exposed to 550 ppb O3 were evaluated by repeated measures ANOVA. ns denotes non-significant effects (p > 0.05). The list of compounds belonging to different volatile groups is shown in Table 2. O3 exposure affected all groups of volatiles except glucosinolate breakdown products and saturated aldehydes, but the effects were treatment-specific ( Figure 6). LOX compound emissions in WW plants exposed to 250 ppb O3 at 10 min and 22 h measurements were enhanced compared to control leaves (Figure 6a). In 550 ppb exposed plants, the LOX compound emissions were greater in WW than in WS plants through the recovery (Figure 6a). For GDP and GGDP compound emissions, O3 enhanced emissions, but the emissions at different recovery time points differed little (Figure 6c,d). Only in WW plants exposed to 250 ppb O3, the emission of GGDP compounds was greater at 10 min recovery . Emission rates within treatments at different time-points during recovery were compared by paired sample t-tests, and significant differences are indicated by lowercase letters (p < 0.05). The effect of O 3 exposure, recovery time and their interaction and the effect of water stress (WS vs. WW) for plants exposed to 550 ppb O 3 were evaluated by repeated measures ANOVA. ns denotes non-significant effects (p > 0.05). The list of compounds belonging to different volatile groups is shown in Table 2.

Impact of Glucosinolate Breakdown Products on the O 3 -Induced Smell Bouquet
All control and stressed plants released a similar blend of glucosinolate breakdown products 22 h after the O 3 treatments (Monte Carlo permutation test, p > 0.05, data not shown) ( Table 2). Specific compounds like dimethyl disulfide and methanethiol were characteristic to WW plants exposed to 250 ppb O 3 and WS plants exposed to 550 ppb O 3 , but they were rare in WW plants exposed to 550 ppb O 3 ( Table 2). The rest of the volatile glucosinolates, cyclohexyl isocyanate, cyclohexyl isothiocyanate, 2-propenenitrile, tetramethylthiourea and tetramethylurea, were characteristic to the control-as well as the stressed plants of different treatments ( Table 2).

O 3 Exposure Effects on Other Volatiles
Among GDP pathway compounds, at 4.5 h recovery point, α-pinene emission was decreased in WW plants exposed to 250 ppb O 3 and at 22 h recovery point, 3-carene emission was enhanced in WW plants exposed to 550 ppb O 3 compared to the values in control plants ( Table 2). Among saturated aldehydes, decanal, heptanal, nonanal and octanal emissions were enhanced in WW plants exposed to 250 ppb O 3 right after exposure compared with the emissions in controls ( Table 2). Heptanal emission was greater at 4.5 h recovery time in WS plants exposed to 550 ppb O 3 than in the controls ( Table 2).

How Drought and Different O 3 Levels Affect Leaf Photosynthetic Characteristics in B. nigra
The entry of gaseous pollutants into the leaf interior is controlled by the openness of stomatal pores and stomatal density that together determine the stomatal conductance and the diffusion of the given pollutant into leaf interior (Figure 1) [69][70][71]. Before O 3 treatment, drought stress strongly reduced the stomatal conductance (g s , Figures 2 and 4), and the ratio of intercellular CO 2 concentration (C i ) to ambient CO 2 concentration (C a ) (C i /C a ) ( Figure 4) in water-stressed (WS) plants compared with well-watered (WW) plants, highlighting the classical plant response to drought that allows conservation of water use [41,[72][73][74]. Apart from drought effects, O 3 typically triggers a rapid stomatal closure and the production of reactive oxygen species (ROS) [16,70,[75][76][77]. In our study, in WW plants exposed to 550 ppb O 3 , stomatal closure was already observed at 10 min after O 3 treatment, and the stomata remained closed through the recovery period ( Figure 3B). Despite stomatal closure, due to higher initial g s and high O 3 exposure dose, the highest stomatal O 3 uptake rate was observed in this treatment (Figures 1-3, Table 1). In contrast, there was almost no effect of O 3 on g s in WS plants exposed to 550 ppb O 3 , except at 7.5 h of the recovery period and there was a moderate increase in WW plants exposed to 250 ppb O 3 at the end of the recovery period ( Figure 3B). Small to moderate impacts of O 3 treatments on g s in WS plants exposed to 550 ppb and WW plants exposed to 250 ppb indicate that O 3 constituted a mild stress in these treatments. These data collectively suggest that B. nigra is a relatively ozone-tolerant species as generally observed for cruciferous plants [68,78,79].
Abscisic acid (ABA) is considered to be the most important chemical regulator of stomatal functioning in plants [30], and these differences among treatments might be related to differences in ABA accumulation and stomatal sensitivity to ABA. Under drought stress, the stomatal closure is strongly dependent on ABA, and the ABA-sensitivity increases with the severity of drought stress [30,80]. O 3 , in turn, reduces the stomatal sensitivity to ABA [81], and this can ultimately lead to a failure to close stomata, also called stomatal sluggishness [40], in response to environmental stimuli such as reduced humidity or reduced light characteristically decreasing stomatal conductance [82]. Thus, the reduction of ABA sensitivity might explain why in WS plants g s was reduced only to a minor degree upon O 3 exposure, and the stomatal response was delayed until 7.5 h after O 3 exposure, indicative of stomatal sluggishness (Figures 3B and 4B). Despite the reduction in stomatal sensitivity, water stress-driven stomatal closure strongly reduced the amount of O 3 entering through the stomata (Table 1, Figure 2). Similar results have been reported in water-stressed and O 3 -treated Phaseolus vulgaris [44] and Quercus ilex [83]. As the result of limited O 3 entry, water stress diminishes O 3 damage in plants [36,84,85]. Yet, in the long term, water stress inevitably reduces photosynthetic production and plant biomass, and this can be exacerbated by ozone stress [83].
Our study demonstrated a classic reduction of photosynthetic rate (A) by water stress due to diminished g s and CO 2 availability (Figures 3 and 4) as observed in numerous studies [26,36,86], and this reduction was only moderately enhanced by O 3 exposure ( Figure 3A). This further emphasizes that reduced g s prior to O 3 exposure protects photosynthetic machinery and results in a lower effective O 3 dose as observed also in other studies [23]. Similarly to WS plants exposed to 550 ppb O 3 , exposure of WW plants to 250 ppb O 3 caused only a minor reduction in A ( Figure 3A). Although the reductions in A and g s were apparently correlated (Figure 3) as demonstrated by strong positive relationships C i /C a ratio and g s ( Figure 4B), g s and A changed differently during recovery after O 3 exposure. Indeed, the C i /C a ratio increased by the end of the recovery period in WW plants ( Figure 4A), but due to different reasons for plants exposed to 250 ppb and 550 ppb O 3 . In the case of 250 ppb O 3 fumigation, A remained at the control level and g s was increased, whereas in the case of 550 ppb O 3 fumigation, both A and g s were reduced, but the reduction in A was greater (Figure 3). This indicates that non-stomatal factors also affected A in WW plants exposed to the higher O 3 concentration. Such non-stomatal factors responsible for the decrease of A can be the reduction in ribulose 1,5-bisphosphate (RuBP) carboxylase/oxygenase activity, inactivation of photosynthetic electron transport and concomitant reduction in RuBP regeneration rate [87,88]. Analogously to our study, a reduction of A partly independent of g s has been shown in other species [89,90]. Furthermore, as a reduction in A can also cause a reduction in g s [91,92], we cannot rule out the non-stomatal reduction in A in earlier phases of recovery in WW plants exposed to 550 ppb O 3 .

Water Stress Effects on Non-Ozonated Plant Volatile Emissions
Plant physiological and metabolic processes critically depend on water stress duration and severity [40,[93][94][95]. Mild water stress can have only limited effects on constitutive VOC emissions, but moderate water stress might enhance VOC emissions, and severe water stress can decrease the emissions [35,40,96,97]. Although water stress might alter the constitutive VOC emissions, it typically does not result in induction of stress volatile emissions [98]. In our study, water stress did not affect significantly the total average VOC emission rate and the emission rate and chemical composition of volatiles was similar among non-ozonated WW and WS B. nigra plants ( Figure 5, Table 2). Given that the constitutive VOC emissions occur at a low level in B. nigra ( Figure 5, Table 2) [47], the metabolic energy and photosynthetic substrate requirements for VOC synthesis are also low. Therefore, it is plausible that the reduction in photosynthate production rate in waterstressed plants did not limit the substrate availability for constitutive volatile synthesis in B. nigra.

O 3 Effects on Total VOC Emissions in Well-Watered and Water-Stressed Plants
In the present study, both WW and WS plants released stress VOCs after the treatment with 550 ppb O 3 ( Table 2). The blend of released volatiles was similar to that in the heat-stressed B. nigra [47], yet the different treatments in our study did not affect the total emissions of VOCs (see Results). Similarly, limited effects of chronic low-level O 3 treatments on VOC emissions in B. nigra have been observed in other studies [57,99]. Nevertheless, the emission of VOCs from plants is affected by the stress dose (severity) and also depends on plant species [40,47,93,94,100,101]. In another O 3 treatment study, B. nigra was most ozone tolerant compared to Sinapis alba, S. arvensis and B. napus [57].

Effects of O 3 and Water Stress Treatments on Emissions of Specific LOX Pathway Compounds
Presence of LOX compounds in the emission blend typically indicates severe stress; LOX emissions can be induced upon mechanical damage, heat, ozone, and herbivory stresses, and the rate of LOX compound emissions often scales with the severity of stress [20,[102][103][104]. O 3 exposure enhanced LOX compounds emission across all treatments, but comparison of WS and WW 550 ppb O 3 -treated plants, demonstrated that WW plants reacted more strongly (significant water stress effect; Figure 6A). In WS plants, less O 3 entered the leaves compared to WW plants exposed to 550 ppb O 3 and, as a result, the O 3 -dependent increase in total LOX emissions was absent, although stress indicator compounds were detected (Tables 1 and 2, Figure 6A). There is also a probability that we might have lost some LOX compound emissions during the first hour of O 3 treatment in WW plants because the earliest LOX emission burst can occur a few minutes to 30 min after the damage of plant cells [105][106][107]. On the other hand, LOX-related volatiles emerge 1-2 h [23] or in some cases even 5 h after the O 3 fumigation [20]. Clearly, we have been able to detect the delayed LOX emission response in WW plants exposed to 250 and 550 ppb O 3 ( Figure 6A).
In our study, hexanal and pentanal were the main LOX volatiles forming the vast amount of the LOX compounds emitted upon O 3 exposure ( Table 2). The same LOX compound composition was observed in O 3 -treated N. tabacum [89]. The detection of various LOX pathway compounds in the emissions of WW or WS B. nigra plants during recovery confirms the presence of cell membrane damage [108,109]. LOX pathway volatiles like (Z)-3-hexen-1-ol, 1-hexanol, 1-penten-3-ol, 1-penten-3-one together with 2-ethylfuran ( Table 2) were characteristic only to the WW and WS plants exposed to 550 ppb O 3 . Next to typical LOX compounds, we have observed previously that 2-ethylfuran is one of the signals of a severe stress as its emissions increased in N. tabacum and B. nigra after a threshold temperature was exceeded [47,104]. It is proposed that 2-ethylfuran is formed from (E)-2-hexenal and co-emitted with (Z)-3-hexenol [110,111], and 2-ethylfuran was observed in WW 550 ppb O 3 -treated plants in our study as well. This evidence collectively suggests that at a given O 3 exposure, higher stomatal O 3 uptake caused a more severe stress of WW B. nigra plants (Tables 1 and 2; Figures 1 and 2).

O 3 Effects on Species-Specific Glucosinolate Degradation Products
Glucosinolates and/or their degradation products are characteristic compounds of the Brassicaceae that are considered as human health-promoting compounds and having an important role in plant protection [112,113]. The results of this study showed that O 3 or water stress did not significantly affect emissions of glucosinolate degradation products ( Figure 6B). Nevertheless, we detected multiple glucosinolate breakdown products in WW and WS plants during the recovery period, including damage-related dimethyl disulfide and even methanethiol [114,115] (Table 2). Lack of a quantitative scaling of glucosinolate breakdown product emission and O 3 exposure might indicate that O 3 treatments did not cause extensive cell damage that would have led to enhanced activation of the glucosinolate-myrosinase system [49,116]. This result is different from heat-stressed B. nigra, where emissions of glucosinolate breakdown products and LOX compounds occurred similarly [47]. Over the longer term, O 3 treatment and water-deficit can lead to modifica-tions in glucosinolate content [34,54,96], and this could further contribute to alterations in the release of glucosinolate breakdown products under O 3 or other stresses.

O 3 Effects on Other Volatile Groups
The plastidial methylerythritol phosphate (MEP) pathway is responsible for the biosynthesis of GDP-derived compounds and their emission depends on the supply of photosynthates [117,118]. Being highly reactive, monoterpenes could improve plant thermoor O 3 -tolerance by reacting with stress-generated oxidative compounds [119,120]. In the present study, the emission of GDP compounds through recovery was similar within O 3 treatments ( Figure 6C). Such a limited time-dependent response might seem surprising given the strong changes in photosynthetic rate (Figure 3). However, this might indicate the overall low GDP compound emission rate in B. nigra such that the carbon flux going into GDP compound synthesis is small and not suppressed by the O 3 -driven reductions in photosynthesis. Decreased GDP emissions have been previously found in O 3 treated N. tabacum [89] and B. napus [121], and relatively stable GDP emissions in O 3 treated B. nigra plants [56,89,[121][122][123].
Part of our results, especially high GGDP compounds emission at 10 min recovery time in WW plants exposed to 250 ppb O 3 , support earlier findings of increased emission of GGDP compounds upon O 3 treatment [38,124,125]. It has been suggested, that in B. nigra the release of GGDP compounds is related to the oxidative cleavage of carotenoids [126,127]. Carotenoids are continuously produced and degraded in green plants [128,129], but their primary role is to protect the photosynthetic apparatus against the oxidative stress [130,131].
The release of saturated aldehydes is probably related to several pathways functioning simultaneously with the LOX pathway such as the conversion of saturated fatty acids into aldehydes or the opposite [109,132,133]. Emission of saturated aldehydes could be partly controlled by stomatal conductance [43,134], but in our study there was no evidence of the direct effect of stomata on the emissions of these volatiles through recovery in different treatments and an overall limited effect of different treatments on saturated aldehyde emissions ( Figure 6E). Decanal and nonanal are commonly found in the odor of B. nigra leaves and flowers and more importantly, their emission is associated with plant O 3 tolerance and O 3 dose, and the degree of damage [135][136][137]. For example, B. nigra plants treated with 70 ppb or 120 ppb O 3 , showed diminished emissions of nonanal and decanal, while the treatment of B. napus leaves with~140 ppb O 3 increased the release of aforementioned saturated aldehydes [56,121]. The insensitivity of saturated aldehyde emissions to water stress and O 3 exposure further underscores the high O 3 tolerance of B. nigra.

Plant Material
Brassica nigra seeds of local source were purchased from the Department of Entomology, University of Wageningen, the Netherlands. Seeds were sown in 0.8 L plastic pots filled with commercial soil (Biolan Oy, Eura, Finland) and quartz sand (1:1 mixture) (AS Silikaat, Tallinn, Estonia). The soil was fertilized with a slow-release mineral fertilizer at an optimum level (Biolan Oy, Eura, Finland). Plants were grown at a light intensity of 400 µmol m −2 s −1 (HPI-T Plus 400 W metal halide lamps, Philips, Brussels, Belgium), relative humidity of 60%, day length of 12 h and day/night temperatures of 24/20 • C in a plant growth room. During the first three weeks, the plants were watered every other day and at the beginning of the fifth week. Then the plants were randomly divided between 'well-watered plants' (WW) and 'water-stressed plants' (WS; Figure 7). Watering of WW plants was continued as described above, but WS plants were watered once a week, resulting in a mild water stress as plant water potential was −0.5 to −0.8 MPa (measured with a PMS 600 pressure chamber, PMS Instrument Company, Albany, OR, USA). The experiment was conducted with mature fully-expanded leaves (ca. 3 week old leaves). growth room. During the first three weeks, the plants were watered every other day and at the beginning of the fifth week. Then the plants were randomly divided between 'wellwatered plants' (WW) and 'water-stressed plants' (WS; Figure 7). Watering of WW plants was continued as described above, but WS plants were watered once a week, resulting in a mild water stress as plant water potential was −0.5 to −0.8 MPa (measured with a PMS 600 pressure chamber, PMS Instrument Company, Albany, OR, USA). The experiment was conducted with mature fully-expanded leaves (ca. 3 week old leaves).

Experimental Set-Up and Gas Exchange Measurements
A custom-made gas-exchange system was used for foliage photosynthesis and transpiration measurements and volatile sampling. The system has a 1.2 L temperature-controlled double-walled glass chamber that is connected to an infra-red dual-channel gas analyzer (CIRAS II, PP-Systems, Amesbury, MA, USA) for measuring CO 2 and H 2 O concentrations at the chamber in-and outlets [138]. Chamber temperature is regulated by a controlledtemperature water bath that circulates water between the double glass walls of the gasexchange chamber [138]. All gas-exchange measurements were conducted at standard conditions of light intensity of 800 µmol m −2 s −1 at leaf surface, air temperature inside the chamber of 25 • C (leaf temperature was ± 1 • C of chamber temperature) and relative humidity 60%. The temperature inside the gas-exchange chamber was monitored by a thermistor (NTC thermistor, model ACC-001, RTI Electronics, Inc., St. Anaheim, CA, USA). The air was taken from outside, passed through a 10 L buffer volume, an HCl-activated copper tubing for scrubbing O 3 and further through a humidifier. Inside the chamber, the CO 2 concentration was 380-400 µmol mol −1 and the air flow rate was set to 1.6 L min −1 . Turbulent conditions inside the chamber were achieved by a fan installed in the chamber. Three fully-expanded mature upper canopy leaves were inserted in the chamber, and 20-30 min after plant enclosure, when leaf gas-exchange rates had stabilized, net assimilation rate (A) and stomatal conductance to water vapor (g s ) were recorded (Figure 7). Simultaneously with the gas-exchange measurements, volatiles were collected as described below. Values of A and g s were calculated according to [139].

Ozone Stress Application
O 3 concentration in chamber in-and outlets was monitored by a UV photometric O 3 detector (Model 49i, Thermo Fisher Scientific, Waltham, MA, USA). In the ambient air entering the chamber, O 3 concentration was less than 2 ppb. O 3 was generated with a Stable Ozone Generator SOG-2 (LLC-Upland, CA, USA) and mixed with the air entering the chamber to the desired concentration. Plants were exposed to O 3 for 1 h as follows: 250 ± 10 ppb for 6 WW plants and 550 ± 25 ppb for 5 WW plants and 4 WS plants (Figure 7). Exposure with 250 ppb O 3 was not tested for WS treatment as preliminary experiments showed no effect on volatile emissions due to too low O 3 uptake. A and g s were recorded 10 min, 1.5 h, 4.5 h, 7.5 h and 22 h after O 3 exposure, and volatiles were collected 10 min, 1.5 h, 4.5 h and 22 h after O 3 treatment. O 3 uptake rate by stomata (nmol m −2 s −1 ), leaf surface (nmol m −2 s −1 ) and the total amount of O 3 uptake (Φ O 3 , µmol m −2 ) by the leaves were calculated at the beginning (~5 min), in the middle (0.5 h) and at the end (1 h) of O 3 application by using the O 3 binary diffusion coefficient and equations from [23]. "Surface" ozone uptake also includes quenching of ozone by reactive hydrocarbons emitted by leaves [89,140].

Volatile Sampling and GC-MS Analysis
Multi-bed stainless steel cartridges filled with three different carbon-based adsorbents were used for collecting the volatiles (VOCs) [141]. A portable 210-1003MTX air sampling pump (SKC Inc., Houston, TX, USA) was connected to the chamber outlet with a T-piece and the air enriched with plant VOCs was drawn from the chamber through the cartridge with a constant flow rate of 200 mL min −1 for 20 min. Each day before the plant measurements, VOCs in the empty chamber were collected and subtracted from measurements with plants.
The cartridges were analyzed with a combined Shimadzu TD20 automated cartridge desorber connected to a Shimadzu 2010 Plus gas chromatograph-mass spectrometer (GC-MS) (Shimadzu Corporation, Kyoto, Japan). Adsorbent cartridges were heated to 250 • C and back-flushed with high purity He (99.9999% AGA, The Linde Group, Tallinn, Estonia) at a flow rate of 40 mL min −1 for 6 min. During this period, desorbed VOCs were collected onto a cold trap filled with Tenax TA at −20 • C. In the second stage, the trap was heated to 280 • C and during 6 min, VOCs were carried with He into a Zebron ZB-624 fused silica capillary column (0.32 mm i.d., 60 m, 1.8 µm film thickness, Phenomenex, Torrance, CA, USA). In GC, the flow rate of the carrier gas (He) was 2.9 mL min −1 and VOCs were separated according to the following program: oven temperature kept at 40 • C for 1 min, then increasing the temperature by 9 • C min −1 to 120 • C, held for 5 min, then increasing the temperature by 2 • C min −1 to 190 • C, held for 2 min, and finally increasing the temperature by 5 • C min −1 to 250 • C, held for 5 min. The Shimadzu QP2010 Plus mass spectrometer (MS) was operated in the electron impact mode. The transfer line temperature was 255 • C and ion-source temperature 170 • C. The GC-MS system was calibrated as explained in [47] and in [141]. VOCs (Table 1) were identified by comparing their mass spectra to the spectra of authentic standards and to those in the NIST library (NIST05). Volatile emissions rates per leaf area were calculated according to [142].

Statistical Analyses
For statistical analyses, net assimilation rate (A) and stomatal conductance (g s ) were log-transformed. The controls for WW and WS plants (values prior to O 3 exposure) were compared with the values of treated plants at different time-points by paired sample t-tests. The differences of A and g s between the control and treated plants were calculated as V c -V t (t), where V c is the average trait value of control and V t is the average trait value of O 3 -treated plants at time t after O 3 treatment. Paired sample t-tests were used to compare the paired values of photosynthetic characteristics, ozone uptake characteristics and log-transformed total VOC emission rates and average emission rates at different time-points during recovery within the given treatment.
The effects of O 3 , recovery time, their interaction and the effect of water stress (for WW and WS plants exposed to 550 ppb O 3 ) on A, g s , total VOC emission, VOC groups and O 3 uptake rates by stomata, leaf surface and total O 3 uptake were evaluated by repeated measures ANOVA. Linear regression analyses were used to explore the relationships of stomatal O 3 uptake with g s , and the relationship of the C i /C a ratio with g s during recovery. All these statistical analyses were conducted with STATISTICA 7 (StatSoft Inc., Tulsa, OK, USA). All statistical effects were considered significant at p < 0.05. Differences in volatile emissions for different combinations of O 3 exposure and water stress were evaluated by principal component analyses (PCA) [143] using CANOCO 5.0 software (ter Braak and Smilauer, Biometris-Plant Research International, the Netherlands). In total, 71% of data variance was explained by the first and the second principal components (PCA 1 46%, PCA 2 25%) (data not shown). Before the analyses, the data were mean-centered and log-transformed. Redundancy data analysis (RDA) with the Monte-Carlo permutation test was used to test for statistical differences in emission blends across the treatments and no statistical differences were observed (p > 0.05, data not shown).

Conclusions and Outlook
This study demonstrated classic reductions in stomatal conductance (g s ) and net assimilation rate (A) in water-stressed B. nigra, but the impact of water stress on constitutive volatile emissions was minor, nor did water stress elicit emissions of stress volatiles. In well-watered plants, acute O 3 stress (550 ppb exposure for 1 h) led to major reductions in g s and A with limited recovery by the end of the experiment at 22 h. In these plants, acute O 3 stress also led to induction of emissions of lipoxygenase pathway (LOX) volatiles between 10 min and 1.5 h after exposure. Due to lower g s , O 3 uptake was reduced under water stress, and as a result, soil water limitation strongly ameliorated the impact of acute O 3 stress, as evidenced in much lower reductions in g s and A and lack of induction of LOX volatiles. Ozone uptake was similar in well-watered plants exposed to a lower O 3 concentration of 250 ppb and in water-stressed plants, and the responses of foliage physiological characteristics to O 3 exposure were also similar, except that g s and A were maintained at a higher level in well-watered plants. This suggests that water stress did not result in the priming of plants to subsequent O 3 stress. Nevertheless, O 3 exposure resulted in surprisingly minor modifications in volatile profiles in the annual plant B. nigra compared with observations for perennial species. This might be indicative of high ozone tolerance of B. nigra, but also reflect the general strategy of annual species that have a rapid leaf turnover and, instead of responding stronger after the stress and repairing the damages, might sacrifice the heavily damaged leaves to form new leaves [144]. In the current study, we looked at O 3 responses right after exposure and through 22 h recovery, but O 3 exposure can have longer-term effects, including initiation of acclimation responses that improve plant tolerance to sustained mild chronic O 3 stress [145,146]. We suggest that future studies should look at stress memory and priming effects including alterations in the physiological characteristics of O 3 -stressed leaves and new leaves formed after O 3 stress.
Author Contributions: K.K., A.K., E.T. and Ü.N. designed the research; K.K performed the experiment and analyzed the data; K.K. and A.K. interpreted the data and wrote the first draft of the article; E.K. interpreted the data; K.K., E.K., A.K. and Ü.N. wrote, reviewed, edited, and approved the final version of the article. All authors have read and agreed to the published version of the manuscript.

Funding:
We acknowledge funding from the European Commission through European Research Council (advanced grant 322603, SIP-VOL+) and the European Regional Development Fund (Centre of Excellence EcolChange), and from the Estonian Research Council (team grant PRG537) and the Estonian University of Life Sciences (grant P180273PKTT).

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable. Data Availability Statement: Data available in a publicly accessible repository after the manuscript is published.