Glandular trichomes as a barrier against atmospheric oxidative stress: relationships with ozone uptake, leaf damage and emission of LOX products across a diverse set of species

Abstract

There is a spectacular variability in trichome types and densities and trichome metabolites across species, but the functional implications of this variability in protection from atmospheric oxidative stresses remain poorly understood. The aim of the present study was to evaluate the possible protective role of glandular and non-glandular trichomes against ozone stress. We investigated the interspecific variation in types and density of trichomes and how these traits were associated with elevated O₃ impacts on visible leaf damage, net assimilation rate, stomatal conductance, chlorophyll fluorescence and emissions of lipoxygenase (LOX) pathway products in 24 species with widely varying trichome characteristics and taxonomy. Both peltate and capitate glandular trichomes played a critical role in reducing leaf ozone uptake, but no impact of non-glandular trichomes was observed. Across species, the visible ozone damage varied 10.1-fold, reduction in net assimilation 3.3-fold and release of LOX compounds 14.4-fold, and species with lower glandular trichome density were more sensitive to ozone stress and more vulnerable to ozone damage compared to species with high glandular trichome density. These results demonstrate that leaf surface glandular trichomes constitute a major factor in reducing ozone toxicity and function as a chemical barrier which neutralizes the ozone before it enters the leaf.

Introduction

Average global ozone (O₃) concentration has approximately doubled during the 20th century and is expected to increase further (Vingarzan, 2004; Fowler et al., 2008; Logan et al., 2012; Hartmann et al., 2013; Oltmans et al., 2013). Currently, the ambient O₃ concentration in the northern hemisphere is around 20-45 nmol mol^-1 and it occasionally reaches 120 nmol mol^-1 or more (Fowler et al., 2008; Lai et al., 2012; Yuan et al., 2015; Gao et al., 2016). Numerous studies have revealed that elevated O₃ negatively impacts plant growth and development and decreases forest productivity and crop yields as well as plant biodiversity (Ainsworth et al., 2012; Leisner & Ainsworth, 2012; Wilkinson et al., 2012; Fares et al., 2013). It is well-established that some plant species are more tolerant to elevated O₃ concentrations than others, but the complexity of underlying mechanisms that lead to enhanced O₃ tolerance is not fully understood (Flowers et al., 2007; Ainsworth et al., 2012; Li et al., 2016; Ainsworth, 2017; Feng et al., 2017).

Plant response to O₃ is a complex process involving several biological levels from molecular to organ and whole plant responses. O₃ enters plants mainly through the stomata, and once inside the leaf, it reacts with organic molecules in the apoplast leading to direct formation of reactive oxygen species (ROS) as well as induction of endogenous formation of ROS that can collectively lead to the onset of cell damage and programmed cell death when the O₃ concentration exceeds critical levels (Wohlgemuth et al., 2002; Pasqualini et al., 2003; Beauchamp et al., 2005; Fiscus et al., 2005; Cho et al., 2011; Ainsworth, 2017; Li et al., 2017; Kanagendran et al., 2017). Generally, progressive and irreversible reductions in photosynthetic activity occur upon acute O₃ exposure, and direct injury to leaf cells in O₃-exposed leaves is also associated with the release of ubiquitous stress-induced volatiles such as volatile products of lipoxygenase (LOX) reactions comprising of various C6 aldehydes and alcohols (Heiden et al., 1999; Beauchamp et al., 2005; Loreto & Schnitzler, 2010; Li et al., 2017). Reductions in the quantum yield of primary photochemistry in the dark-adapted state and emissions of LOX and other volatile organic compounds (VOC) are widely used characteristics to assess plant response to O₃ stress (Nussbaum et al., 2001; Beauchamp et al., 2005; Bussotti et al., 2011; Chutteang et al., 2016; Li et al., 2017).

O₃ sensitivity varies across species and depends on several species-specific biochemical, physiological and morphological traits (Li et al., 2016). Regarding the physiological and biochemical traits, differences in O₃ sensitivity have been related to inherent differences in stomatal conductance and to the leaf antioxidant capacity (Brosché et al., 2010; Fares et al., 2013). Very few studies have been conducted to understand to which extent morphological traits such as trichome morphology and density or differences in exposed mesophyll cell surface or other structural traits such as leaf mass per unit area, leaf nitrogen content and life form may affect interspecific variation in O₃ sensitivity when exposed to acute rather than chronic low to moderate level O₃ (Franzaring et al., 1999; Prozherina et al., 2003; Hayes et al., 2007; Li et al., 2016).

The leaf surface is covered by glandular and/or non-glandular trichomes in many species. Non-glandular trichomes, which are not able to produce or secrete phytochemicals, serve to increase herbivore resistance by interfering mechanically with herbivore movement and/or feeding (Eisner et al., 1998; Corsi & Bottega, 1999; Kennedy, 2003). In addition, by increasing surface reflectance they contribute to reduced solar radiation interception and thus, to enhanced resistance to low water availabilities and photoinhibition stress (Ehleringer, 1981, 1982; Cescatti & Niinemets, 2004; Hallik et al., 2017). However, due to increasing surface area, non-glandular trichomes could potentially participate in chemical reactions on leaf surface as well, e.g. in O₃ adsorption, but little is known of relationships between non-glandular trichome density and species O₃ resistance. However, given that non-glandular trichomes are covered by a wax layer that consists of saturated hydrocarbons that have low reactivity with O₃ (Corsi & Bottega, 1999), it might also be that the effect of non-glandular trichomes in leaf O₃ resistance is minor despite greater surface area.

Glandular trichomes, which store and secrete various secondary metabolites such as flavonoids, monoterpenes or sesquiterpene lactones, are widely distributed on leaf surfaces and may contribute significantly to defense against diverse biotic and environmental challenges (Ehleringer et al., 1976; Huci et al., 1982; Rieseberg et al., 1987; Wagner, 1991; Fahn & Shimony, 1996; Kostina et al., 2001; Heinrich et al., 2002; Amme et al., 2005; Paoletti et al., 2007; Peiffer et al., 2009; Agati et al., 2012; Jud et al., 2016; Lihavainen et al., 2017; Thitz et al., 2017). Two major types of glandular trichomes, peltate and capitate, have been discerned. There are various types of capitate trichomes varying in stalk and head size, and although the basic morphology of peltate trichomes varies less, these is still a significant variation in the shapes and sizes of peltate trichomes as well (Ascensão et al., 1995; Ascensão & Pais, 1998; Baran et al., 2010; Glas et al., 2012). Capitate trichomes are specialized to produce and store a large amount of diterpenes and a wide array of non-volatile or poorly volatile compounds such as defensive proteins, acylated sugars and esters that are directly exuded onto trichome surface (Sallaud et al., 2012; Jud et al., 2016). Their primary function is thought to repel pests. Peltate glandular trichomes, on the other hand, mostly produce and store biogenic volatile or semi-volatile organic compounds related to plant abiotic or biotic stress induced responses (Corsi & Bottega, 1999; Gang et al., 2001; Wagner et al., 2004). Recently, Jud et al. (2016) showed that semi-volatile organic compounds such as various diterpenoids exuded by the glandular trichomes of Nicotiana tabacum act as an efficient chemical protection shield against stomatal O₃ uptake by reacting with and thereby depleting O₃ at the leaf surface. However, the shape and size of glandular trichomes as well as the metabolites stored vary greatly across plant species, and the generality of this finding is currently unclear. Although several studies have demonstrated the important protective role of trichomes in O₃ stress resistance in a given species (Huci et al., 1982; Prozherina et al., 2003; Paoletti et al., 2007; Riikonen et al., 2010), their potential role in O₃ resistance across different plant species, in particular under acute exposure, has not been investigated to date.

Considering leaf surface reactions, it is important to distinguish between O₃ exposure, O₃ uptake by leaf and uptake through the stomata to accurately characterize the thresholds for acute responses. However, this is not fully possible until we discern the extent of O₃ neutralization at the leaf surface (Jud et al., 2016) and determine the extent to which leaf internal O₃ concentrations can build up (Moldau & Bichele, 2002; Niinemets et al., 2014). To fill the gap in understanding the role of trichomes in defense against O₃ stress, we examined interspecific variation in O₃ sensitivity in 24 species with varying densities of non-glandular and glandular trichomes on the leaf surface. The main hypotheses tested were (1) leaves with a high density of glandular trichomes have higher rates of non-stomatal O₃ uptake, resulting in greater O₃ resistance as evidenced by less negative effects on their physiology at the given O₃ exposure concentration; (2) non-glandular trichomes have no significant protective role under O₃ exposure; (3) peltate glandular trichomes have a major impact on non-stomatal O₃ uptake. We demonstrate a broad relationship between O₃ resistance and glandular trichome density across a highly diverse set of species, suggesting that glandular trichomes do serve as an important antioxidative barrier. To our knowledge, this is the first study to provide information on how non-glandular and glandular trichomes can contribute to plant ability to cope with increased O₃ levels and maintain physiological homeostasis under atmospheric oxidative stress.

Materials and Methods

Plant materials and growth conditions

Twenty three herbaceous species, selected for their wide range of trichome characteristics (trichome density, glandular vs. non-glandular, capitate vs. peltate) were chosen for this study (Anchusa officinalis L., Arctium tomentosum Mill, Carduus crispus L., Cucumis sativus L. cv. Libelle F1, Cucurbita pepo L., Erigeron acer L., Erigeron canadensis L., Erodium cicutarium L., Geranium palustre L., Geranium pratense L., Geranium robertianum L., Lavandula angustifolia Mill, Mentha × piperita L., Nicotiana tabacum L. cv. W38, Ocimum basilicum L., Phaseolus vulgaris L. cv. Saxa, Rosmarinus officinalis L., Salvia officinalis L., Silene latifolia Poir, Solanum lycopersicum L. cv. Pontica, Tussilago farfara L., Urtica dioica L. and Verbascum thapsus L.). In addition, although woody species generally exhibit greater tolerance to O₃ than herbaceous species, we chose a single woody species, Betula pendula Roth, was included. Plants of C. sativus, N. tabacum, O. basilicum, P. vulgaris, S. lycopersicum and V. thapsus were grown from seed (seed sources: C. sativus: Seston Seemned OÜ, Estonia; N. tabacum: gift of I. Bichele, University of Tartu; O. basilicum and S. lycopersicum: SC Agrosel SRL, Romania; P. vulgaris: Dalema UAB, Vilnius, Lithuania; the seed of V. thapsus were collected in the field in Tartu, Estonia). After germination, seedlings were replanted in 2 L plastic pots containing a commercial potting mix (Kekkilä Group, Vantaa, Finland) and maintained in a plant room with light intensity at plant level of 400 μmol m^-2 s^-1 (HPI-T Plus 400 W metal halide lamps, Philips) during 12 h photoperiod. The day/night temperatures were maintained at 24/20 °C and daytime humidity at 60%. Seedlings of B. pendula, and plants of G. palustre, G. pratense and U. dioica were collected from the campus of the Estonian University of Life Sciences (EULS), Tartu, Estonia (58°23' N, 27°05' E, elevation 40 m), the plants of E. acer, E. canadensis, E. cicutarium and S. latifolia from Ihaste, Tartu (58°21' N, 26°46' E, elevation 41 m), and plants of G. robertianum from Pühajärve, Estonia (58°03' N, 26°28' E, elevation 131 m). The plants were replanted in 2 L plastic pots containing the same potting mix, and maintained on the roof of the plant biology lab building at the campus of EULS and were exposed to full sunlight under ambient conditions. All plants were fertilized once in two weeks with Biolan NPK (N of 100 mg L^-1, P of 30 mg L^-1, and K of 200 mg L^-1) complex fertilizer with microelements (Biolan Oy, Kekkilä Group, Vantaa, Finland) and they were watered every day to maintain optimal growth conditions.

In the case of N. tabacum and V. thapsus that had leafed petioles making it difficult to achieve a good seal upon enclosure of the leaf to the gas-exchange chamber, we removed the part of the lamina attached to the base of the leaf, while avoiding cutting through first- and second-order veins. In order to minimize the influence of leaf damage on volatile organic compound (VOC) emissions, at least two days were allowed for recovery before conducting measurements. According to previous measurements, the effect of lamina wounding is short-living with the bulk of the stress volatiles released within first 5-10 min after wounding (Portillo-Estrada et al., 2015). We also did not observe any subsequent induction of stress volatiles and volatile isoprenoids after wounding, and leaf photosynthesis measurements with clip-on type gas-exchange system (Walz GFS-3000, Walz GmbH, Effeltrich, Germany) demonstrated that the wounding also did not alter foliage photosynthesis rate and stomatal conductance at 48 h since wounding stress (data not shown).

The leaves of A. officinalis, A. tomentosum, C. crispus, and T. farfara were collected from the campus of the EULS and leaves of C. pepo, L. angustifolia, M. × piperita, R. officinalis and S. officinalis were collected from the gardens of Tartu city. All plants sampled had developed in open areas exposed to full sunlight. Early in the morning, high light exposed terminal shoots with multiple leaves were excised with a sharp razor blade under water and immediately transported to the laboratory. Once in the laboratory, the cut ends of the shoots were recut under water and a representative leaf was selected for measurements. All measurements were conducted with fully-expanded non-senescent leaves.

Gas-exchange system for photosynthesis, stomatal conductance, volatile organic compound and ozone concentration measurements

We used a custom-built gas-exchange system with a temperature-controlled 1.2 L glass chamber described in detail by Copolovici & Niinemets (2010) to fumigate leaves with O₃ and to measure leaf gas exchange characteristics (net photosynthesis and stomatal conductance) and VOC. The ambient air was drawn from outside, passed through a 10 L buffer volume, purified by an O₃ trap and a charcoal filter, humidified by a custom-made humidifier to standard air humidity of 60% and then passed at a flow rate of 1.6 L min^-1 to the glass chamber. The air entering or leaving the chamber was analyzed alternately, with an electronic valve switching between incoming (reference) and outgoing (measurement) air streams. To mix the chamber air and to minimize leaf boundary layer resistance, a fan was mounted inside the chamber. The leaf temperature was maintained at 25 °C, ambient CO₂ concentration at 400 µmol mol^-1 and light intensity at the leaf surface was 1000 µmol m^-2 s^-1 during the experiment.

An infra-red dual-channel gas analyzer (CIRAS II, PP-Systems, Amesbury, MA, USA) operated in absolute mode was used to measure CO₂ and H₂O concentrations, a proton transfer reaction-time of flight-mass spectrometer (PTR-TOF-MS, Ionicon Analytik GmbH, Innsbruck, Austria) to determine VOC concentration and a UV-photometric O₃ sensor (Model 49i, Thermo Scientific, Massachusetts, USA) to measure O₃ concentrations. All instruments were operated continuously and CO₂ and H₂O concentrations and all PTR-TOF-MS signals were recorded every 10 s and the range of O₃ concentrations was recorded for every 15 min interval. Reference measurements (incoming air) were conducted every 10-15 min.

Ozone fumigation treatments

Ozone exposure treatments were carried out using O₃ concentrations that constituted a moderately severe stress leading to visible leaf damage but not resulting in leaf death during more than 20 hours following the experimental treatment. After the leaves were enclosed in the chamber, continuous measurements of photosynthetic characteristics and trace gas exchange were begun immediately. O₃ fumigation was started after stabilization of net photosynthesis and VOC emissions, typically 20-30 min after leaf enclosure. O₃ was generated in a quartz glass reaction chamber (Stable Ozone Generator, Ultra-Violet Products Ltd, Cambridge, UK) under UV light (λ=185 nm) and monitored by a UV-photometric O₃ sensor (Model 49i, Thermo Scientific, Massachusetts, USA). The O₃ enriched air from the reaction chamber was mixed with the air stream entering the leaf chamber. The leaf was fumigated with a step-wise increase in O₃ concentration as follows: 100 ± 5 nmol mol^-1 of O₃ for 30 min, followed by 200 ± 10 nmol mol^-1 for 30 min, and so on progressively, increasing O₃ concentration by 100 nmol mol^-1 in half-hour steps until the final maximum O₃ concentration that lead to a sharp decrease in net assimilation rate and stomatal conductance was reached. This final O₃ concentration varied by plant species. Due to logistic difficulties, PTR-TOF-MS measurements could be carried out in six species with different glandular trichome density (C. sativus, N. tabacum, O. basilicum, P. vulgaris, S. lycopersicum and V. thapsus). In order to further investigate the effect of O₃ concentration rather than O₃ uptake on volatile LOX product emissions, we applied the same O₃ stress with a step-wise increase from 100 to 500 nmol mol^-1 for these six species. In all cases, the duration of exposures at each O₃ concentration was 30 min and the stability of O₃ concentration within the leaf chamber was ± 5-10% of the target. The gas-exchange characteristics and trace gas-exchange were also continuously monitored for 21 h after each of the treatments. The treated leaves were then harvested and their area (S) and the quantitative degree of damage visible as necrotic or chlorotic lesions were estimated using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Operation of the PTR-TOF-MS and online monitoring of the kinetics of volatile release

The main advantages of PTR-TOF-MS are simultaneous detection of all masses rather than predetermined masses, and higher sensitivity, higher mass resolution as well as improved accuracy over the traditional quadrupole PTR (PTR-QMS); this allows quantification of a larger number of organic compounds at trace levels in real time (Jordan et al., 2009; Graus et al., 2010). The PTR-TOF-MS was operated according to the method described in detail in Graus et al. (2010), Brilli et al. (2011) and Portillo-Estrada et al. (2015). The drift tube voltage was kept at 600 V at 2.3 mbar drift pressure and 60°C temperature, corresponding to an E/N ≈ 130 Td in H₃O⁺ reagent ion mode. The PTR-TOF-MS system was calibrated with a calibration standard mixture containing key volatiles from different compound families (Ionimed Analytic GmbH, Austria). The data acquisition software TofDaq (Tofwerk AG, Switzerland) was used to acquire the raw data and the data were further post-processed by PTR-MS Viewer software (PTR-MS Viewer v3.1, Tofwerk AG, Switzerland) (Jordan et al., 2009; Portillo-Estrada et al., 2015; Li et al., 2017).

The total amount of lipoxygenase pathway products (LOX products) emission presented in this study is the sum of the dominant compounds with mass signals m/z 81.070 [hexenal (fragment)], m/z 83.085 [hexenol + hexanal(fragments)], m/z 85.101 [hexanol (fragment)], m/z 99.080 [(Z)-3-hexenal + (E)-3-hexenal (main)], m/z 101.096 [(Z)-3-hexenol + (E)-3 hexenol + (E)-2-hexenol + hexanal (main)] (Brilli et al., 2011; Portillo-Estrada et al., 2015; Li et al., 2017).

Chlorophyll fluorescence measurements

To visualize the distribution of photosynthetic electron transport activity of leaves before and after the O₃ treatment, a portable Imaging-PAM chlorophyll fluorometer with ImagingWin software (Imag-MIN/B, Heinz Walz GmbH, Effeltrich, Germany) was used. The Mini version of the Imaging-PAM M-series has a measurement window area of 24×32 mm and uses a CCD camera (640×480 pixel) for fluorescence imaging and 12 high-power LED lamps to provide actinic light and high-intensity light flashes. After the leaf was fixed in the leaf holder of the Imaging PAM, and dark-adapted for 30 min at 25°C, the minimum fluorescence yield (F₀) was measured, the maximum dark-adapted fluorescence yield (F_m) was then determined by illuminating the leaves with a 500 ms pulse of saturating irradiance of 2700 µmol quanta m^-2 s^-1. The spatially-averaged maximum dark-adapted quantum yield of PSII, F_v/F_m was calculated as (F_m-F₀)/F_m.

Scanning electron microscopy and analyses of density and morphology trichomes

To quantify the trichome characteristics, a Zeiss EVO LS15 Environmental Scanning Electron Microscope (ESEM, Carl Zeiss AG, Jena, Germany) was used. A fresh leaf adjacent to the leaf used for O₃ fumigation was mounted on brass stubs. The samples were viewed and images taken with the ESEM at an acceleration voltage of 15 kV under the low vacuum mode. Images acquired from ESEM were analyzed with the ImageJ software. Trichome density was estimated from the middle zones of both leaf surfaces (adaxial and abaxial) avoiding major veins. Three to ten areas of view were selected for measurements, and all trichomes were counted within 1 mm² areas. The type of glandular trichomes - peltate or capitate - was identified from ESEM images as suggested by Ascensão et al. (1995), Ascensão & Pais (1998), Ko et al. (2007) and Baran et al. (2010), and averages were calculated for all fields of view for the given leaf on both upper and lower surface.

Empty chamber corrections

The background volatile samples were measured from the empty chamber before the plant measurements to correct for possible emission of volatiles adsorbed previously on gas-exchange system components. Although such corrections were minor (≤1%), they were included in the calculations for consistency (Niinemets et al., 2011). O₃ destruction due to surface reactions (“uptake”) by the empty chamber ($C_{{\text{O}}_3}^{\text{chamber}},$ nmol mol^-1) was measured before the plant measurements and calculated as:

$$C_{{\text{O}}_3}^{\text{chamber}}=C_{\text{in}}^{\prime }-C_{\text{out}}^{\prime }$$ Eqn 1

where $C_{\text{in}}^{\prime }$ is the O₃ concentration at the chamber inlet and $C_{\text{out}}^{\prime }$ that at the chamber outlet. The values obtained were subsequently used to correct all measurements of leaf O₃ uptake. Overall, this correction was minor, less than 5% of total O₃ uptake when leaves were enclosed in the chamber.

Calculations of photosynthesis, stomatal conductance, trace gas emission rates and ozone uptake

Net assimilation rate (A_n) and stomatal conductance (g_s) per leaf area were calculated according to von Caemmerer & Farquhar (1981).

Volatile emissions rates (ϕ_x, nmol m^-2 s^-1) were calculated as

$${\varphi }_{\text{x}}=\frac{F}{S}\left[C_{\text{o}}(\text{X})-C_{\text{i}}(\text{X})-C_{\text{c}}(\text{X})\right]$$ Eqn 2

where F is the flow rate through the chamber (1.19×10^-3 mol s^-1), and S is the leaf area enclosed in the chamber (m²), C_o(X) is the concentration (nmol mol^-1) of the target VOC (compound X) measured at the chamber outlet and C_i(X) of that measured at the chamber inlet, C_c(X) is the correction to account for the possible release of the given compound released from the gas-exchange system components (Beauchamp et al., 2005; Li et al., 2017).

The total amount of target VOC emitted over a certain time (Φ_X) (nmol m^-2) was integrated as:

$${\Phi }_{\text{X}}={\displaystyle \sum _{{\text{t}}_{PS}}^{{\text{t}}_{PE}}\Delta t{\varphi }_{\text{X}}}$$ Eqn 3

where t_PS and t_PE are the start and end times of the target VOC emission release, Δt is the measurement time interval (10 s), and ϕ_X is the emission rate of the target VOC measured over this time interval (Beauchamp et al., 2005; Li et al., 2017). The total emission values correspond to the whole experiment, from elicitation to 21 h since the end of the experiment.

The rate of O₃ uptake by the leaf (ϕ_LO3, nmol m^-2 s^-1) was calculated as:

$${\varphi }_{{\text{LO}}_3}=\frac{F}{S}\left(C_{\text{in}}-C_{\text{out}}-C_{{\text{O}}_3}^{\text{chamber}}\right)$$ Eqn 4

where C_in and C_out are the O₃ concentrations in the air entering and exiting the leaf chamber (nmol mol^-1). We calculated C_in and C_out as the average value over the given time interval due to the fluctuations of O₃ concentration produced by the ozone generator and manual adjustment of C_in to account for changes in O₃ uptake during the exposure.

The rate of O₃ uptake by stomata (ϕ_GO3, nmol m^-2 s^-1) was determined by:

$${\varphi }_{{\text{GO}}_3}=\left(C_{\text{out}}-C_{{\text{O}}_{\text{3}}}^{\text{i}}\right)\frac{g_{\text{s}}}{2.03}$$ Eqn 5

where $C_{{\text{O}}_3}^{\text{i}}$ is intercellular O₃ concentration (nmol mol^-1), and 2.03 is the ratio of water vapor to O₃ diffusivities (Li et al., 2017). Given that $C_{{\text{O}}_3}^{\text{i}}$ is typically small compared with C_out, and it cannot be estimated from these measurements, it was assumed to be zero (Laisk et al., 1989; but see Moldau & Bichele, 2002). To be consistent with the C_out calculations, stomatal conductance (g_s) was also determined as the average value over the same time interval.

The total amount of leaf and stomatal O₃ uptake (Φ_Y, where Y stands either for leaf, LO₃, or for stomatal, GO₃, O₃ uptake) over the given exposure period (O₃ dose) was integrated as:

$${\Phi }_{\text{Y}}={\displaystyle \sum _{{\text{t}}_S}^{{\text{t}}_E}\Delta t{\varphi }_{\text{Y}}}$$ Eqn 6

where t_S and t_E are the start and end times of O₃ exposure, Δt is the time interval of the O₃ exposure, and ϕ_Y is either the mean rate of ϕ_LO3 or the mean rate of ϕ_GO3 measured during this time interval (Beauchamp et al., 2005). Ultimately, the total amount of non-stomatal O₃ uptake (Φ_NGO3), i.e. presumably O₃ uptake by the leaf surface was determined by:

$${\Phi }_{{\text{NGO}}_3}={\Phi }_{{\text{LO}}_3}-{\Phi }_{{\text{GO}}_3}$$ Eqn 7

where Φ_LO3 is the total measured amount of O₃ uptake by the whole leaf and Φ_GO3 is the calculated O₃ uptake through the stomata. We note that chemical quenching of O₃ in chamber air by volatiles in the leaf chamber or by volatiles on the leaf surface is also possible and is incorporated in the value of Φ_NGO3.

Quantitative characterization of the protective role of glandular trichomes against O₃ stress

To further characterize the functional role of glandular trichomes in ameliorating changes in leaf physiological characteristics and LOX product emission under O₃ stress, the plant response per unit O₃ taken up by the leaf (γ, O₃ uptake weighted response) was calculated as:

$${\gamma }_{\text{Z}/{\Phi }_{\text{Y}}}=Z/{\Phi }_{\text{Y}}$$ Eqn 8

where Z represents the percentage decrease of A_n, g_s and F_v/F_m or the total amount or maximum rate of LOX products emitted over a certain time (Eqn 3). Analogously, the plant response per unit O₃ taken up by stomata was also calculated for each of these traits.

Data analysis

All O₃ fumigation treatments were conducted in three to five replications (n=3-5) with different plants in each species. Linear- and non-linear regressions were used to analyze the effects of O₃ on foliage traits and relationships between leaf trichome density and physiological traits. All relationships were considered significant at P < 0.05.

Results

Trichome morphology, distribution and density

Non-glandular and glandular trichomes were both present on the leaf surfaces of the 24 studied species (Fig. 1). The density of non-glandular trichomes varied 250-fold across species (Table 1). Two types of glandular trichomes (peltate and capitate) were identified (Fig. 1). Capitate glandular trichomes were present in all species, but peltate glandular trichomes were found only in 12 species belonging to the family Lamiaceae (L. angustifolia, M. × piperita, O. basilicum, R. officinalis and S. officinalis) and Geraniaceae (E. cicutarium, G. pratense and G. robertianum) and in B. pendula, E. canadensis, S. latifolia and U. dioica (Table 1). The density of capitate glandular trichomes varied from 1 mm^-2 in C. pepo to 90 mm^-2 in R. officinalis, and peltate glandular trichome density varied from 5.9 mm^-2 in G. robertianum to 60 mm^-2 in R. officinalis (Table 1). No significant correlation between glandular and non-glandular trichome density was found in the species with only capitate trichomes and with both capitate and peltate glandular trichomes (Fig. S1).

Fig. 1.

Environmental scanning electron microscope (ESEM) micrographs of (a) non-glandular (NG) and peltate (P) and capitate (C) glandular trichomes on the lower surface of Lavandula angustifolia and (b) lower surface of Cucumis sativus containing non-glandular (NG) and glandular capitate trichomes (C). Scale bars on the bottom right.

Table 1.

Mean ± SE densities (number mm^-2) of non-glandular (NGT), peltate (PGT) and capitate (CGT) glandular trichomes in the 24 plant species investigated.

Species	Family	NGT (number mm-2)	CGT (number mm-2)	PGT (number mm-2)
Anchusa officinalis	Boraginaceae	6.8 ± 0.9	12.9 ± 3.5
Arctium tomentosum	Asteraceae	30.1 ± 1.1	60.7 ± 1.3
Betula pendula	Betulaceae	20.1 ± 1.3	10.02 ± 0.41	60.2 ± 1.5
Carduus crispus	Asteraceae	30.1 ± 1.1	14.0 ± 1.2
Cucumis sativus	Cucurbitaceae	15.036 ± 0.021	28.1 ± 0.8
Cucurbita pepo	Cucurbitaceae	50.1 ± 2.0	1.0274 ± 0.0011
Erigeron acer	Asteraceae	26.4 ± 3.2	9.32 ± 0.09
Erigeron canadensis	Asteraceae	15.2 ± 2.3	19 ± 5	7.8 ± 2.2
Erodium cicutarium	Geraniaceae	9.3 ±1.3	20.9 ± 2.0	6.3 ±2.4
Geranium palustre	Geraniaceae	17 ± 10	33 ± 16
Geranium pratense	Geraniaceae	20 ± 7	58 ± 10	6.35 ± 0.35
Geranium robertianum	Geraniaceae	8.13 ± 0.20	15.6 ± 1.7	5.9 ± 2.1
Lavandula angustifolia	Lamiaceae	65.3 ± 1.5	65.2 ± 1.4	30.0 ± 1.2
Mentha × piperita	Lamiaceae	21.0 ± 1.3	37.1 ± 1.1	50.2 ± 1.7
Nicotiana tabacum	Solanaceae	5.172 ± 0.011	70.2 ± 1.2
Ocimum basilicum	Lamiaceae	1.034 ± 0.011	1.215 ± 0.021	9.13 ± 0.05
Phaseolus vulgaris	Fabaceae	2.041 ± 0.007	10.132 ± 0.012
Rosmarinus officinalis	Lamiaceae	250.3 ± 2.2	90.4 ± 2.7	60.3 ± 2.0
Salvia officinalis	Lamiaceae	70.2 ± 1.2	10.114 ± 0.012	45.2 ± 1.1
Silene latifolia	Caryophyllaceae	15.1 ± 3.3	7.9 ± 1.1	8.3 ± 1.3
Solanum lycopersicum	Solanaceae	1.243 ± 0.021	9.035 ± 0.022
Tussilago farfara	Asteraceae	3.053 ± 0.011	37.3 ± 0.9
Urtica dioica	Urticaceae	93 ± 40	5.7 ± 2.0	29 ± 12
Verbascum thapsus	Scrophulariaceae	10.217 ± 0.042	47.03 ± 0.07

Kinetics of O₃ uptake flux and stomatal conductance with rising O₃ concentration

During stepwise increases in O₃ in all species, stomatal conductance (g_s) declined with increasing O₃ exposure (filled symbols, Fig. 2a for representative responses in P. vulgaris, N. tabacum and S. lycopersicum). Relative to the values prior to low-level O₃ fumigation (100 nmol mol^-1), stomatal conductance (g_s) decreased by 27% in P. vulgaris, 66% in N. tabacum and 55% in S. lycopersicum in response to higher-level O₃ exposure treatment (400-700 nmol mol^-1). Despite of sharp decreases in g_s, O₃ uptake by the leaf (ϕ_LO3) increased in all three species (Fig. 2a) although the continuing decrease in g_s at high O₃ concentration eventually caused O₃ uptake to level off and then decline in N. tabacum and S. lycopersicum. In P. vulgaris, the percentage of non-stomatal O₃ uptake increased with rising O₃ concentration (Fig. 2b). The changes in total O₃ uptake and stomatal conductance (g_s) during O₃ fumigation in N. tabacum and S. lycopersicum were generally similar to P. vulgaris up to O₃ concentration of 400 nmol mol^-1. However, when O₃ concentration was raised to 600 or 700 nmol mol^-1 O₃ uptake by the leaf (ϕ_LO3) and stomata (ϕ_GO3) started to decrease because of the lower stomatal conductance (Fig. 2). The percentage of non-stomatal O₃ uptake (i.e., O₃ removal at the leaf surface) in leaves of N. tabacum and S. lycopersicum declined somewhat over the course of the experiment as O₃ concentrations increased (Fig. 2b). These responses were analogous in other species studied (data not shown).

Fig. 2.

Ozone uptake flux by the entire leaf (ϕ_LO3) (a) and via the stomata (ϕ_GO3) (b) normalized to the uptake fluxes at applied O₃ concentration of 100 nmol mol^-1, stomatal conductance (g_s; a) and percentage of O₃ uptake by leaf surface (%; b) in relation to applied O₃ concentration in P. vulgaris (upward pointed triangles), N. tabacum (hexagons) and S. lycopersicum (downward pointed triangles). Open and filled symbols in (a) correspond to whole leaf normalized O₃ uptake flux (ϕ_LO3) and stomatal conductance (g_s), and open and filled symbols in (b) to stomatal O₃ uptake flux (ϕ_GO3) and percentage of O₃ uptake by leaf surface (%), respectively. Error bars indicate ± SE.

The relationships between non-stomatal O₃ uptake and trichome types and density

There was no correlation between non-glandular trichome density and the percentage of non-stomatal O₃ uptake (Fig. 3a), and we also found no correlation between the size (length and width) of non-glandular trichomes with the amount of non-stomatal O₃ uptake (data not shown). In contrast, the glandular trichome density was strongly correlated with the percentage of non-stomatal O₃ uptake (P < 0.0001; Fig. 3b). The correlations were even stronger when species with capitate glandular trichomes only and species with both peltate and capitate type glandular trichomes were examined separately although the latter group had overall greater non-stomatal O₃ uptake (Fig. 3c).

Fig. 3.

Relationships between the percentage of non-stomatal ozone uptake and non-glandular trichome density (a), between the percentage of non-stomatal ozone uptake and glandular trichome density (b), and percentage of non-stomatal ozone uptake and glandular trichome density in species with capitate or both capitate and peltate trichomes (c). In (b) and (c), open symbols represent species with only capitate glandular trichomes and filled symbols species with both capitate and peltate glandular trichomes. The data in (b) were fitted by a non-linear regression (y=17.5+69.3(1-0.982^x)) and in (c) by a linear regression, y=0.667x+17.7, for species with capitate glandular trichomes and by a non-linear regression, y=35.9+63.4(1-0.989^x), for species with capitate and peltate glandular trichomes.

The glandular trichome density showed highly significant correlations with the threshold of total leaf O₃ uptake that induced a sharp decrease in net assimilation rate and stomatal conductance (P < 0.0001; Fig. 4a,c), but it was not correlated with the threshold of total stomatal O₃ uptake (Fig. 4b,d). In the case of induction of volatile LOX product emissions, glandular trichome density showed a non-significant correlation (P = 0.09) with the threshold of total leaf O₃ uptake for all species pooled (Fig. 4e), but a significant correlation (P < 0.01) in the case of species with only capitate glandular trichomes (inset in Fig. 4e), and no significant correlation with the threshold of total stomatal O₃ uptake (Fig. 4f). However, a significant negative relationship (P < 0.0001) between the total amount of volatile LOX pathway products released and glandular trichome density was observed (Fig. 5).

Fig. 4.

Correlations of average glandular trichome density with average physiological thresholds of O₃ uptake by leaf (a, c, e) and stomata (b, d, f). The O₃ uptake thresholds were defined as uptake values inducing a significant decrease of net assimilation rate (T_LO3-An for leaf, and T_GO3-An for stomatal uptake) and stomatal conductance (T_LO3-gs, and T_GO3-gs), and emissions of volatile products of the lipoxygenase pathway (LOX products, also called green leaf volatiles) (T_LO3-LOX and T_GO3-LOX). In (e), the inset does not include O. basilicum that had both capitate and peltate glandular trichomes. The data were fitted by linear or non-linear regressions (y=37.6+1.62x for (a); y=17.9+1.96x for (c), y=22.5x/(1+0.0788x) for (e, main panel) and y=108.1x^0.188 for (e, inset)). B. pendula (filled circles, ●), C. sativus (open circles, ○), E. acer (open squares, □), E. canadensis (filled squares, ■), E. cicutarium (filled diamonds, ◆), G. palustre (open diamonds, ◇), G. pratense (filled hexagons, ⬢), G. robertianum (filled upward pointing triangles, ▲), L. angustifolia (filled downward pointing triangles, ▼), M. × piperita (filled stars, ★), N. tabacum (open hexagons, ⬡), O. basilicum (filled plus symbols, +), P. vulgaris (open upward pointing triangls, △), R. officinalis (filled crosses, ✖), S. officinalis (filled Venus symbols, ), S. latifolia (filled Mars symbols, ), S. lycopersicum (open downward pointing triangls, ▽), U. dioica (filled Christian crosses, ✚), V. thapsus (open stars, ☆). In all cases, open symbols represent species with only capitate glandular trichomes and filled symbols species with both capitate and peltate glandular trichomes. Error bars indicate ± SE.

Fig. 5.

Total LOX product emissions (Φ_LOX) over 21 hours following O₃ exposure in relation to glandular trichome density for six species with different density of glandular trichomes. The y axis is Log₁₀-transformed in the inset. In all cases, O₃ concentration was increased in a stepwise fashion in 30 minute intervals, 100 nmol mol^-1 increments, from 100 to 500 nmol mol^-1. The data were fitted by a non-linear regression (y=-5.79/(1-0.116x) for the main panel, and y=16.8 * 10⁶ /(1+ 10.3 * 10⁵ x) for the inset). Symbols are as defined in Fig. 4.

Changes in physiological characteristics and LOX product emissions in response to O₃ exposure in relation to glandular trichome density

Both net assimilation rate (A_n) and stomatal conductance (g_s) were significantly decreased after O₃ exposure compared to the rates prior to O₃ fumigation. O₃-induced relative reductions in A_n and g_s were strongly determined by glandular trichome density (P < 0.05; Fig. 6a,b). Visible leaf injury was also less in species with greater glandular trichome density at the same amount of leaf O₃ uptake (P < 0.05; Fig. 6c). There was overall reduction of F_v/F_m per ppb O₃ taken up by the leaf (O₃ uptake weighted F_v/F_m, Eq. 8), but the correlation with glandular trichome density was not significant (Fig. 6d). Additionally, no significant correlations were observed between glandular trichome density and relative reduction in stomatal uptake weighted A_n, g_s, F_v/F_m and visible leaf injury (Fig. S2a-d). However, glandular trichome density was negatively correlated with O₃ uptake weighted total LOX product emission and maximum LOX product emission rate, regardless of whether the calculations were based on O₃ taken up by the entire leaf (P < 0.05; Fig. 6e,f) or by stomata (P=0.05 and P=0.06, Fig. S2e,f), and the data were best fitted by nonlinear regressions (Fig. 6e,f and Fig. S2e,f). Across all treatments, visible leaf damage assessed at the end of the O₃ exposure was a good indicator of reductions in foliage physiological characteristics and elicitation of LOX volatiles, except for reductions in stomatal conductance (Fig. 7).

Fig. 6.

(a) The percentage decrease in net assimilation rate (γ_{An decreased / ΦLO3}), (b) the percentage decrease in stomatal conductance (γ_{gs decreased / ΦLO3}), (c) the percentage of leaf area visibly damaged (γ _{damaged / ΦLO3}), (d) the percentage decrease in the dark-adapted maximum quantum efficiency of PSII (F_v/F_m) (γ_{(Fv/Fm decreased / ΦLO3}), (e) the induction of total amount of LOX product emission (γ_{ΦLOX / ΦLO3}) and (f) the maximum rate of induced LOX product emission (γ_{ϕM, LOX / ΦLO3}) per unit leaf O₃ uptake (Eq. 8) in relation to glandular trichome density. As in Fig. 5, the total LOX product emission is the sum of LOX products released since the induction until termination of the experiment at 21 h after the stop of O₃ exposure. Leaf damage after O₃ fumigation was assessed as the presence of necrotic and/or chlorotic lesions on leaf surface. The data were fitted by linear or non-linear regressions (y=0.0662+(2.49/x) for (a); y=0.0755+(2.01/x) for (b); y=0.0871-0.000912x for (c); y=-0.00371+(1.50/x) for (d); y=-0.00332+(1.35/x) for (e) and y=-0.00112+(0.19/x)for (f)). Symbols as defined in Fig. 4. Error bars indicate ± SE.

Fig. 7.

Relative decreases in net assimilation rate (a) and stomatal conductance (b) and maximum dark-adapted quantum yield of PSII (F_v/F_m) (c), and total emission of LOX products (d) and maximum emission rate of LOX products (e) in O₃-exposed leaves in relation to the percentage of leaf area visibly damaged. The data were fitted by linear or non-linear regressions (y=-14.3+0.71.4(1+0.0188x)^-0.366 for (a); y=1.54x-1.99 for (d) and y=0.252x-0.899 for (e)). The Data in the inset of (c) were fitted by linear regressions separately for species with only capitate trichomes (y=12.8+1.110x, open symbols) and with both capitate and peltate glandular trichomes (y=8.09+0.227x, filled symbols). Symbols as in Fig. 4.

Correlations among O₃ uptake and leaf damage and with emissions of LOX products

To estimate O₃ damage across different species, we compared the correlations of modification in leaf physiological traits and leaf damage with O₃ uptake by the whole leaf (Φ_LO3), stomatal O₃ uptake (Φ_GO3) and non-stomatal O₃ uptake (Φ_NGO3). Φ_LO3 was not correlated with the percentage of decrease in A_n, g_s and F_v/F_m, but a significant positive correlation (P < 0.0005) with the percentage of visibly damaged leaf area was observed (Fig. S3a-d). Furthermore, Φ_LO3 was also positively correlated with the total LOX product emission and maximum LOX product emission rate (P < 0.05; Fig. S3e,f). Apart from the protective role of glandular trichomes against O₃ stress, the O₃ dose ultimately entering into the leaf is crucial in understanding the magnitude of reductions in photosynthesis, appearance of visible lesions, and emission of LOX products. Φ_GO3 was strongly correlated with the amount of visible leaf damage (P < 0.05; Fig. 8c), but there was no correlation between Φ_GO3 and the percent decrease in A_n, g_s or F_v/F_m (Fig. 8a,b,d). Furthermore, a positive non-linear relationship was found between Φ_GO3 and both the total LOX product emission and maximum LOX product emission rate across the species (P < 0.05; Fig. 8e,f).

Fig. 8.

The total amount of O₃ uptake by stomata (Φ_GO3) in relation to the percentage decrease of net assimilation rate (a) and stomatal conductance (b), percentage leaf area visibly damaged (c) and percentage decrease of maximum dark-adapted chlorophyll fluorescence yield (F_v/F_m) (d), the total LOX product emission (e) and maximum emission rate of LOX products (f) from O₃-exposed leaves. In (c), (e) and (f), the data were fitted by linear regressions (y=9.03+0.0648x; y=0.166x-3.58 and y=0.0265x-1.06). Symbols as in Fig. 4.

Φ_NGO3 was not correlated with the percentage decrease in A_n, g_s and F_v/F_m, and total amount of visible leaf damage and LOX product emissions (Fig. S4). Finally, the percentage of non-stomatal O₃ uptake showed highly significant correlation with the percentage decrease in F_v/F_m (P < 0.005) but it was correlated with the percentage decrease in A_n and g_s, amount of visible leaf damage and LOX product emissions (Fig. S5).

Discussion

Variation in trichome types and density across species

Non-glandular and glandular trichomes were found on the surfaces of all studied leaves. Specifically, two types of glandular trichomes on the surface were identified: peltate and capitate (Fig. 1). All species had capitate trichomes whereas peltate trichomes were observed only on the surface of 12 species. This agrees with previous observations by Glas et al. (2012), demonstrating more frequent occurrence of capitate glandular trichomes. According to past reports available for some of the species studied here, peltate and capitate trichome densities vary between 1-140 mm^-2 (Wilkens et al., 1996; Ranger & Hower, 2001; Valkama et al., 2004; Kapoor et al., 2007; Lihavainen et al., 2017; Thitz et al., 2017). Correspondingly, the species studied here fell in the same range.

As outlined in the introduction, glandular and non-glandular trichomes play different roles in plant defense. Non-glandular trichomes are linked to mechanical defense against biotic and environmental stressors, while glandular trichomes produce metabolites for the alleviation of those stressors (Levin, 1973; Lihavainen et al., 2017). Glandular and non-glandular trichome density was not correlated suggesting that various selection pressures have occurred independently.

Glandular trichomes affect stomatal O₃ uptake and increase the threshold of acute O₃ responses

Stomata exert a strong control over leaf interior O₃ concentrations and thus, play a major role in protecting leaves from O₃ stress (Beauchamp et al., 2005; Li et al., 2017). Likewise, in this study, the step-wise increases in O₃ concentration induced simultaneous stomatal closure (Fig. 2). In addition, data on non-stomatal O₃ uptake (Φ_NGO3) during the exposure revealed that O₃ was significantly destroyed by leaf surface reactions. Factors that reduce O₃ concentrations at leaf surface will reduce O₃ entry into the leaf interior through the stomata and lower the effective O₃ dose received by the plant. In our study, capitate and peltate glandular trichomes were both related to Φ_NGO3 (Fig 3), whereas peltate glandular trichomes were more strongly related to reduced stomatal O₃ uptake than capitate glandular trichomes, but surprisingly, non-glandular trichomes did not affect at all leaf O₃ uptake.

Across species, a wide variety of metabolites are stored and exuded by glandular trichomes at the leaf surface (Rieseberg et al., 1987; Wagner, 1991; Fahn & Shimony, 1996; Heinrich et al., 2002; Amme et al., 2005; Peiffer et al., 2009; Agati et al., 2012; Jud et al., 2016; Lihavainen et al., 2017). It has been suggested that unsaturated semi-volatile compounds destroy O₃ before it enters the leaf (Jud et al., 2016). Furthermore, high glandular trichome density is often associated with the presence of high concentrations of more volatile terpenoids, but also other non-volatile defense compounds such as acylated sugars and proteins (Gang et al., 2001; Amme et al., 2005). As different types of trichomes may have vastly different composition of secreted compounds (e.g. Fahn & Shimony, 1996; Heinrich et al., 2002; Amme et al., 2005; Kant et al., 2009; Jud et al., 2016), such differences among capitate and peltate glandular trichomes might reflect differences in the chemical reactivity with O₃ of main chemical constituents synthesized in different trichomes. Peltate glandular trichomes are a major site of semi-volatile compound production and contain a much larger oil sac into which these compounds are secreted (e.g. Fahn, 1979; Gang et al., 2001; Amme et al., 2005). Nevertheless, tobacco lacks peltate trichomes, but in this species, the degree of non-stomatal O₃ uptake also correlates strongly with the density of capitate trichomes and their exudates, indicating that capitate trichomes also contribute strongly to surface reactions with O₃ (Jud et al., 2016). In fact, Jud et al. (2016) suggested that diterpenes exuded from tobacco trichomes were responsible for O₃ quenching.

Non-glandular trichomes can form a dense indumentum on leaf surface that serve as a mechanical barrier against herbivores and pathogens (Eisner et al., 1998; Corsi & Bottega, 1999; Kennedy, 2003), but their role in antioxidative responses is less clear. Several species in our study had a massive amount of non-glandular trichomes (Table 1), but we found no correlation between the density of non-glandular trichomes with the amount of non-stomatal O₃ uptake in the present study (Fig. 3a). This likely reflects the circumstance that non-glandular trichomes possess no secretory structures and are covered by a wax layer that consists of long-chained saturated oxygenated and non-oxygenated hydrocarbons (Corsi & Bottega, 1999) that have very low reactivity with O₃.

O₃ dose ultimately entering the plant interior is known to determine the amount of visible leaf damage, but in the case of mild O₃ uptake not exceeding a certain O₃ threshold, neither damage to the leaf biochemical apparatus nor induction of VOC emission have been observed (Nussbaum et al., 2001; Beauchamp et al., 2005; Velikova et al., 2005; Niinemets, 2010). In our study, we found a stronger relationship between glandular trichome density and the threshold for acute responses caused by O₃ taken up by leaf than with that taken up through stomata (Fig. 4), indicating that trichome density promotes O₃ tolerance of plants. However, once O₃ enters the leaf, all species had similar threshold values regardless of trichome density. Moreover, the total emission of volatile LOX products was related to glandular trichome density under the same stepwise increase of O₃ concentration and duration of fumigation (Fig. 5). By combining the correlations between glandular trichome density and non-stomatal O₃ uptake, we have further demonstrated that glandular trichomes directly reduce O₃ concentration at the leaf surface and thus indirectly reduce stomatal O₃ uptake. While previously this has been shown for tobacco (Jud et al., 2016), our results extend this observation to a highly diverse set of plants with a large variability in trichome types, densities and likely also with exudates with highly diverse chemical nature. These results further confirm the previous suggestions that the plants’ acute responses are determined directly by O₃ dose taken up by stomata (Φ_GO3) (Beauchamp et al., 2005; Jud et al., 2016; Li et al., 2017), but also emphasize that these responses are importantly modified by glandular trichome density.

Effects of elevated O₃ concentration on foliage photosynthetic characteristic and LOX product emission

Effects of elevated O₃ concentration on plant growth, development, biodiversity, forest productivity and crop yields have been studied extensively in the last 30 years (Feng & Kobayashi, 2009; Ainsworth et al., 2012; Leisner & Ainsworth, 2012; Wilkinson et al., 2012; Fares et al., 2013; Ainsworth, 2017). The stepwise increase in O₃ concentrations in this study caused reductions in net assimilation rates (A_n), stomatal conductance (g_s) and F_v/F_m and induced formation and emission of LOX products, in agreement with previous reports for single species exposed to a given O₃ concentration (Beauchamp et al., 2005; Chutteang et al., 2016; Li et al., 2017). We observed that there was a stronger relationship between the density of glandular trichomes and the absolute response per unit of O₃ taken up by the whole leaf (i.e., through both surface reactions and via the stomata) than per unit of O₃ taken up through the stomata (Fig. 6, S2), suggesting that trichomes provide protection against O₃ damage on leaf surface. Furthermore, the evidence that visible leaf damage across species correlates more strongly with stomatal O₃ uptake than with whole leaf O₃ uptake (Fig. 8, S3) suggests that plant injuries are more strongly linked to the uptake of O₃ through stomatal pores. These relationships clearly show that glandular trichomes directly reduce stomatal O₃ uptake by depleting O₃ at the leaf surface and thus indirectly alleviate O₃ damage.

Given that glandular trichomes play multiple role in defence, the questions is whether the protection provided by glandular trichomes is directly driven by selection for greater oxidative stress resistance or whether it is primarily coincidential and reflects adaptation to other abiotic and biotic stresses. While the selection for atmospheric pollutant resistance was likely of minor significance in the past, human-driven pollution clearly increases the significance of selection for traits protecting against pollution and thus, the variation in glandular trichome density can importantly alter the fitness of given species in multispecies stands where different species exhibit differences in glandular trichome density. The fact that trichome number is not fixed at the time of leaf emergence and that trichome distribution changes with leaf position also implies that protection by glandular trichomes is a highly adaptive trait (Maffei et al., 1989; Prozhertina et al., 2003; Deschamps et al., 2006; Jud et al., 2016). Thus, the severity of abiotic and biotic stresses during leaf development can importantly alter leaf vulnerability to stress such as O₃ during their lifetime (Jud et al., 2016), and further studies should examine whether long-term increases in atmospheric O₃ concentration alter trichome density and the capacity for protection against atmospheric oxidants.

Although different species exhibited different O₃ tolerance, O₃-induced leaf damage occurred in all species (Fig. 7). The reductions in A_n and F_v/F_m and the induction of LOX product emissions were associated with O₃-induced leaf damage across different species (Fig. 7). Such interspecific relationships have not been well studied, but finding uniform correlations across species suggests that the damage, once induced, alters foliage physiological characteristics similarly in all species. In previous studies, quantitative within-species relationships among LOX product emission and the severity of abiotic (Copolovici et al., 2012; Copolovici & Niinemets, 2016; Li et al., 2017) and biotic stresses (Niinemets et al., 2013) have been observed. The interspecific correlations with the degree of damage observed in our study provide also encouraging evidence that not only modifications in foliage photosynthetic characteristics, but also LOX product emissions can provide a quantitative measure to estimate the deleterious effects of O₃ across species.

Conclusions

Our study provides conclusive evidence that glandular trichomes play a protective role against O₃ stress, and that this protection is associated with the type and density of glandular trichomes across a highly diverse set of species. A stepwise increase in O₃ concentration during fumigation led to severe visible leaf injury, and reductions in maximum quantum yield of PSII, leaf net assimilation rate and stomatal conductance, and increase in emissions of volatile products of lipoxygenase (LOX) pathway. We demonstrated that the presence of both peltate and capitate glandular trichomes are strongly related to reduced stomatal O₃ uptake. Because glandular trichomes reduce O₃ concentrations surrounding the open stomata on leaf surface, the O₃ dose threshold triggering acute responses increased with increasing glandular trichome density. In addition, absolute differences in leaf damage and LOX product emissions caused by leaf O₃ uptake were strongly linked to glandular trichome density. In summary, our study demonstrates that species with low glandular trichome density were more sensitive to O₃ stress compared to species with high trichome density, suggesting that leaf trichominess might importantly drive species dispersal in polluted environments. Further work is needed to gaining an insight into the distribution of species with different level of glandular trichome density in different communities and to understanding how various types of glandular trichomes containing different organic compounds affect plant O₃ resistance.

Summary Statement.

Glandular and non-glandular trichomes are widely distributed on leaf surfaces, but the role of different trichome types and trichome density in protection from atmospheric oxidative stress is poorly understood. This study analyzed ozone stress resistance of photosynthesis and induction of stress volatiles in 24 species with widely varying trichome characteristics and taxonomy, and demonstrated that the presence of glandular trichomes strongly reduced stomatal ozone uptake and ozone-dependent damage. This highlights a key role of glandular trichomes in maintenance and physiological homeostasis under atmospheric oxidative stress.