1 Introduction

Implementing reliable air pollution abatements is an ongoing challenge (Watson et al. 2016). The background concentrations of ground-level ozone (O3) have been steadily increased, even though peak levels have declined in intensity and frequency (Simon et al. 2015), as a consequence of recently appeared O3 “hot spots” and long-distance transport (Bytnerowicz et al. 2013). Because of multi-factorial interactions among the vegetation, atmosphere and O3 precursors, oxidative stress varies in severity from year to year and critical levels are commonly exceeded in various regions throughout the world (Cristofori et al. 2015; Moura et al. 2014a; Sá et al. 2016; Sharma et al. 2016). The prognoses beyond 2040 (Fuhrer et al. 2016; Klingberg et al. 2014) remain uncertain, as a reduction of NOx and other emissions may not sufficiently alleviate the current O3 levels, as a consequence of climate warming (Garthwaite et al. 2009; Lauwaet et al. 2014). The O3 effects are mediated by stomatal influx, which is determined by O3 exposure but also by several other environmental drivers, especially soil water availability and vapour-pressure deficit (Baumgarten et al. 2009). During the exceptionally dry summer of 2003, the O3 uptake—despite higher exposure—was reduced, due to drought-induced closure of stomatal pores (Löw et al. 2006). Given stress response plasticity (Matyssek et al. 2012) and regional environmental peculiarities (Trenberth et al. 2014), ecosystem responses to O3 can be highly variable. Hence, the impact of ground-level O3 on vegetation will remain a lasting concern during the twenty-first century (Mills et al. 2016).

Visible symptoms in foliage, which provide some of the most obvious indications of O3 stress, are better reliable when verified by a combination of approaches including experimental exposure (Dai et al. 2017; Harmens et al. 2017; Novak et al. 2003; Paoletti et al. 2017) and microscopic validation (Alves et al. 2016; Simon et al. 2015; Vollenweider et al. 2013; Moura et al. 2018). These approaches have been used extensively for O3 monitoring and bioindication (Ferretti et al. 2015; Sanz and Calatayud 2011; Schaub et al. 2016). Beech (Fagus sylvatica L.) is a widespread tree species in Central Europe. Visible O3 symptoms usually reported for beech include gradients of interveinal bronzing, dark stippling and less specific leaf discoloration, increasing in severity along yearly shoot increments basipetally. However, the inter- and intraspecific variability of visible injury is astounding and complicates the diagnosis of O3 stress (Sanz and Calatayud 2011; Vollenweider and Günthardt-Goerg 2006).

Microscopic validation can ascertain an O3 injury diagnosis by providing mechanistic understanding of the cause and development of symptoms (Faoro and Iriti 2009; Günthardt-Goerg and Vollenweider 2007; Kivimäenpää et al. 2005). Previous research has demonstrated the close correlation between macro- and micromorphological symptoms (Moura et al. 2011, 2018; Reig-Armiñana et al. 2004; Vollenweider et al. 2013) and has linked the variability of structural injury with that of processes and plant responses triggered or amplified by O3 stress in foliage (Faoro and Iriti 2009; Foyer and Shigeoka 2011; Calderón Guerrero et al. 2013; GünthardtGoerg et al. 1997; Günthardt-Goerg and Vollenweider 2007; Mikkelsen and Heide-Jørgensen 1996; Moura et al. 2014b; Pasqualini et al. 2003). Microscopic injury by O3 stress shows similar traits, irrespective of the taxonomic group (Bussotti et al. 2005; Günthardt-Goerg et al. 2000), and leaf sclerophylly (Calderón Guerrero et al. 2013; Reig-Armiñana et al. 2004). Further, structural markers indicative of other stress factors can show how O3 interacts with several environmental factors such as drought (Calderón Guerrero et al. 2013), light (Paoletti et al. 2010) or coldness during winter (Vollenweider et al. 2013). Microscopic validation is thus particularly insightful for understanding plant responses to O3 in a changing climate. However, such demanding evidence is still limited. Also missing is ultrastructural evidence of responses in different subcellular compartments and cell types, especially in the case of disruptive processes. Furthermore, in most reports (Table 1), only a few markers of O3 injury are described and there is a lack of comprehensive functional understanding regarding the effects of O3 stress at the cell level and for the different plant responses.

Table 1 Review of previous studies (references) of structural markers of oxidative stress and indications of the physiological processes triggered by O3 in symptomatic foliage samples completed by new results with sound pictures Figs. 3, 4, 5, and 6
Full size table

Given the availability of O3-exposed foliage in the CASIROZ experiment performed in a mixed forest on adult beech trees (Southern Germany, Matyssek et al. 2007), the objectives in the present study included to (i) characterize the spectrum of O3 injury within beech foliage by comprehensively analysing symptoms in the leaf blade at the tissue, cell and subcellular levels, (ii) relate the ozone stress markers mechanistically to physiological processes and responses and (iii) analyse the incidence of varying ozone dose on the spectrum of O3 injury and prevalence of physiological responses causing the observed O3 symptoms. Therefore, leaf samples were harvested repeatedly in the sun crowns of trees during field campaigns between 2000 and 2006, focusing on the different types of visible O3-like symptoms. The underlying structural injuries were extensively analysed, combining different methods in bright field light, fluorescence and transmitted electron microscopy (TEM).

2 Material and methods

2.1 The experimental site

This study made use of the Free Air O3 Enrichment (FACE) facility within the framework of the CASIROZ experiment (Carbon Sink Strength of Fagus sylvatica L. in a Changing Environment—Experimental Risk Assessment of Mitigation by Chronic Ozone Impact). In a mixed stand within Kranzberg Forest, 35 km northeast of Munich, Germany (48° 25′ N, 11° 39′ E, 485 m asl), two neighbouring groups of five beech trees each (Fagus sylvatica L.), about 60 years old and 27 m high, were exposed to either the unchanged ambient or experimentally enhanced twice-ambient O3 regime (both naturally fluctuating) from April through October each during the period of 2000 to 2007. The experimental design has been detailed in several papers (Karnosky et al. 2007; Matyssek et al. 2007; Nunn et al. 2002; Werner and Fabian 2002). The overall outcome from the CASIROZ experiment on tree physiology and growth responses to O3 stress is summarized in Matyssek et al. (2010).

Depending on the regime, yearly ozone levels and environmental conditions, the foliage material sampled between June and September and analysed in this study had been exposed to phytotoxic O3 doses above a threshold of 1 mmol m−2 s−1 (POD 1) varying between 5 and 36 mmol m−2 (Fig. 1).

Fig. 1
figure 1

Cumulated phytotoxic ozone dose (POD1, Grünhage et al. 2012) for beech leaf area during the growing seasons of 2002, 2003 and 2006 (May–September) in the top canopy layer of ambient and twice ambient O3 treatment by a free-air fumigation experiment in Kranzberg Forest, Germany. Arrow heads indicate foliage harvest dates with emphasis on (i) visible O3 symptoms 2002 and 2003, (ii) histochemical analyses of visible symptoms 2003 and (iii) analyses of structural and ultrastructural cell injury 2006

Full size image

2.2 Leaf injury assessments and microscopy

At each sampling date (Fig. 1), a leaf sample collection including all O3-like symptom types and asymptomatic material was harvested in the sun crown of beech trees and within both O3 regimes. Hence, we focused on symptom type variability in the most symptomatic part of the canopy, irrespective of each type frequency and recurrence as a function of O3 regime or crown position (microscopic validation approach). In situ, O3 stress effects were diagnosed on the basis of (i) smooth gradients of injury increasing in severity basipetally at branch and leaf levels, (ii) prevalence of injury in foliage (or parts of foliage) exposed to direct solar radiation and (iii) foliar symptoms in the form of leaf discoloration and/or isolated or confluent dark or light stippling (Günthardt-Goerg and Vollenweider 2007; Innes et al. 2001; Schaub et al. 2010). Other types of symptoms incited by various abiotic and biotic stress factors were used for differential diagnosis (Fink 1999; Hartmann et al. 2007; Vollenweider and Günthardt-Goerg 2006). After documenting the symptom position and morphology, disc samples were excised, fixed and processed with a view to histochemical, structural and ultrastructural analysis (Moura et al. 2018; Vollenweider et al. 2013). For relating the micro- to macroscopic injury, the position of sections within sampled discs was determined, assessing the location of analysed macro-and microscopic lesions.

3 Results

3.1 Visible injury

Several types of visible O3-like symptoms, varying between harvests and years, were diagnosed in the upper canopy of beech trees (Fig. 2a). The O3 regimes did not show differences regarding the symptom types. At branch level, the symptoms were distributed along smooth gradients, with their severity increasing basipetally (Fig. 2b). At leaf level, the different types of symptoms included photo-bleaching (Fig. 2e), bronzing (Fig. 2c, e, f), intercostal dark discrete and sometimes confluent stippling (Fig. 2g) and intercostal yellowish dots (Fig. 2h). These symptoms were observed separately (Fig. 2f, g, h) or in combination (Fig. 2e), primarily on the adaxial and light-exposed leaf sides, whilst shaded leaf parts stayed asymptomatic (Fig. 2e). The yellowish dots formed discrete lesions evenly distributed in the leaf blade and also visible abaxially. In 2003, large intercostal brown flecks, as a consequence of aggregated stippling, were additionally observed (not shown).

Fig. 2
figure 2

Visible injury triggered by O3 within foliage of adult beech trees. a Gradients of leaf injury in the canopy, increasing in severity in the better illuminated branches. b Gradients of O3 injury at branch level, increasing in severity basipetally. c–h Singular or combined O3 injury at leaf level: bronzing (b; c, e, f), photobleaching (phb; e), stippling (st; g) and yellowish dots (nd; h). d Asymptomatic

Full size image

3.2 Microscopic injury underlying discoloration symptoms

Whatever the symptom type, the most severe structural injury was detected in mesophyll. At tissue level, the mesophyll cells directly underlying symptoms of photobleaching and bronzing showed degenerative changes in comparison to cells in asymptomatic leaf segments (Fig. 3c versus Fig. 3a). Symptoms were more severely expressed adaxially than abaxially (Fig. 3d versus Fig. 3l). They included an increase in the condensation of cell content, enlargement of vacuome and accumulation of condensed tannins (bronzing symptoms only), especially within the upper palisade parenchyma cell layer (Fig. 3d versus Fig. 3b). At subcellular level, the cytoplasm and nucleus of mesophyll cells showed a coarser granular structure than in asymptomatic material (Fig. 4f versus Fig. 4a). The matrix and inner membrane system of mitochondria were degraded (Fig. 4i, j versus Fig. 4d, e), whilst chloroplasts showed reduced size, reduced grana width and increased size and density of plastoglobules (Figs. 3d and 4g, h versus Figs. 3b and 4b, c). Within the upper palisade cells and with an apparent and concomitant increase in chloroplast illumination, such symptoms were enhanced adaxially (Fig. 4g versus Fig. 4h). Plastoglobules, primarily consisting of lipids (Fig. 5a, c), were excreted into the vacuole where they were either degraded, as indicated by undigested rim residues, or aggregated into lipid inclusions (Fig. 5d, e). In vacuoles which accumulated condensed tannins, the latter inclusions were found within cavities inside sponge-like solid tannin bodies (Fig. 5g, h). Similar but less severe symptoms were observed in spongy parenchyma (Fig. 6g, h versus Fig. 4g–j).

Fig. 3
figure 3

Tissue- and cell-level changes caused by oxidative stress within foliage of beech trees. Changes in segments of leaf blades underlying bronzing (c, d, l), stippling (e, f, i, j, m, o) and yellowish dots (g, h, n) versus asymptomatic material (a, b, k). Bronzing symptoms (c, d): condensed tannins inside vacuoles of upper epidermis (UE, blue) and palisade parenchyma (PP), sometimes in the form of plugs (*), condensation of cytoplasm (c) and nucleus (n), decrease in chloroplast (ch) size and increase in plastoglobule (pg) size and frequency. Stippling symptoms (e, f, i, j, m, o): changes observed in the case of bronzing interspersed with dead, disrupted and partly cytorrhyzed cells in upper palisade parenchyma (e, f); groups of cytorrhyzed, disrupted and condensed (#) cells extending into lower palisade, in the case of stronger reactions (2003; i, j). Less injury in spongy parenchyma (SP) and lower epidermis (LE; l, m, o). Yellowish dots (g, h, n): injury by oxidative stress more severe in abaxial than adaxial tissues, with parts of lower mesophyll disrupted and the cells cytorrhyzed (&; g). Lesions consisting of cell strands either collapsed and condensed (#) or deformed and degenerated (§; h) and containing oxidized condensed tannins (oct; g, h). Leakage of cell debris in the intercellular space (is) of lower leaf blade (n). Other structures: st: starch, oct: oxidized condensed tannins. Technical specifications: staining with Toluidine blue and p-phenylenediamine, observation in bright field microscopy (a, c, e, g, i), using phase contrast (b, d, f, h, j–o)

Full size image
Fig. 4
figure 4

Cellular and subcellular changes caused by oxidative stress within foliage of beech trees. Changes within palisade cells underlying bronzing (f–j), stippling (k–s) and yellowish dots (t–x) versus asymptomatic material (a–e). Bronzing symptoms (f–j): the adaxial chloroplasts (ad ch, g) contained more plastoglobules (pg) but fewer grana than abaxial organelles (ab ch, h). Plastoglobules were larger and grana stacks smaller. Nucleus (n) content was condensed and its envelope electron-opaque (i). j condensed cytosol (c) and injured matrix and inner membrane system (+) within the mitochondria (mt). Stippling symptoms (k–s): the severity of degenerative traits in chloroplasts (l, m), nucleus (n) and cytoplasm (o) was enhanced. k Fewer chloroplasts adaxially and vacuole filled with oxidized tannin bodies (*) often plugging the adaxial cell lumen. Occasionally, partial cell decompartmentation (~), shrinking vacuoles (shv) and cytorrhysis (arrowheads) indicated terminal cell degeneration and cell death. HR-like traits (p–s): cells were cytorrhyzed (arrowheads) and their content disrupted (p), chloroplasts (q, s) together with other organelles and cell structures (s) condensed and deformed, and extensive membrane damage was indicated by poor membrane resolution. Some material had been leaked into the periplasm (pe) through ruptured plasmalemma (pl). r Pyknotic nucleus showing deformation and acute chromatin condensation. Yellowish dots (t–x): cytorrhyzed and condensed (#) versus deformed and degenerated (§) strands (t). Within the latter group of cells (§; u, x), the chloroplasts showed degenerative traits (i.e. thinner grana and higher plastoglobule density). Within the former group of cells (#; v, w), the chloroplasts were condensed and deformed, their grana poorly resolved and the plastoglobules still small. w Pyknotic nucleus. x Intercellular material (im) leaked from adjacent dead cells. Other structures: cw: cell wall, d: dictyosome, l: lipid droplet, re: smooth endoplasmic reticulum, st: starch, v: vacuole. Technical specifications: postfixation using OsO4, contrasting using uranyl acetate and lead citrate, observation in TEM

Full size image
Fig. 5
figure 5

Contribution of plastoglobules to the accumulation of vacuolar lipids. a Accumulation of plastoglobules (in blue) in degenerating cells from palisade (PP) and spongy (SP) parenchyma. b Within stippling symptoms, lipids (encircled) were leaked into the enlarged intercellular spaces (is) between disrupted and cytorrhyzed cells (#). c Accumulation of large plastoglobules within chloroplasts (ch) from palisade parenchyma cells in leaves showing bronzing symptoms. d Accumulation of vacuolar lipid (lpv) after apparent excretion of plastoglobules into the vacuole. e–h Plastoglobule excretion and fate in the vacuole. e The plastoglobules excreted from chloroplasts into vacuoles (v) had their lipids apparently degraded, with an outer shrunken ring as the only remnant. f Accumulation of vacuolar lipids shortly before cell death within disrupting vacuoles (v ~). g Simultaneous accumulation of plastoglobules and condensed tannin (vct) in the vacuole, with lipids filling globular cavities within and gaps between tannin deposits. h Contiguous arrangement of lipids and tannins in the vacuole. The polymerization of condensed tannins in solid bodies was indicated by splits (arrows) and broken pieces (*). The vacuolar lipids formed globular deposits, sometimes encapsulated within tannin bodies and showing evidence of degradation. Other structures: c: cytoplasm; cw: cell wall; g: grana; LE: lower epidermis; n: nucleus; st: starch; UE: upper epidermis. Technical specifications: histochemical revelation of lipids using Sudan Black B (a–d), TEM as in Fig. 3 (e–h)

Full size image
Fig. 6
figure 6

Cellular and subcellular changes caused by oxidative stress in the upper epidermis (a–d), spongy parenchyma (e–h) and lower epidermis (i–j) of symptomatic (c, d, g, h, j) versus asymptomatic (a, b, e, f, i) foliage material. Degenerative changes were indicated by the condensation of cytoplasm (c), decrease of thylakoids and increase of plastoglobules within leucoplasts (lc) or chloroplasts (ch), injury (+) to cristae and matrix of mitochondria (mt) and accumulation of condensed tannins (vct) in the vacuole (v). Notice the lower physiological activity in the asymptomatic epidermis (a, b, i) versus spongy parenchyma (e, f), as indicated by the thinner cytoplasm layer, few organelles, larger vacuome and more condensed nucleus structure (n). Other structures: cu: cuticle; cw: cell wall; g: grana; is: intercellular space; re: smooth endoplasmic reticulum, st: starch, v: vacuole. Technical specifications: TEM as in Fig. 3

Full size image

3.3 Microscopic injury underlying stippling

Structural injury underlying stippling was more severe than that underlying bronzing and photobleaching (Fig. 3, Table 1). Degenerative symptoms in upper palisade cells were aggravated (Fig. 3e versus Fig. 3c). Besides enhanced condensation and degeneration of organelles (Fig. 4l–o versus Fig. 4g–j), cell death was indicated by cytorrhysis, partial decompartmentation of cell content, tonoplast disruption and shrunken vacuoles (Fig. 4k, o). Vacuoles contained a sizable amount of lipid droplets from excreted plastoglobules. Some of these structures had been leaked into the enlarged intercellular space (Fig. 5b). However, some upper palisade cells lacked any degenerative symptom and only showed structural changes indicative of a hypersensitive reaction–like syndrome (HR-like; Table 1). HR-like markers included (i) cytorrhysis (Figs. 3f and 4p), (ii) disruption of plasmalemma and cell content (Figs. 3f and 4p, s), (iii) pyknotic nucleus, after condensation, deformation and shrinkage of nucleic material (Fig. 4r), and (iv) condensation and deformation of chloroplasts. These showed disruption of the lamellar system and chloroplast envelope in the absence of an accumulation of plastoglobules (Fig. 4q, s). The peculiar macroscopic stippling injury observed in 2003 was underlain by more severe structural injury than in other years, with larger groups of necrotic cells showing typical HR-like traits, often extending into the lower palisade parenchyma (Fig. 3i, j). Furthermore, spongy parenchyma showed injury similar to that observed in the case of macroscopic bronzing injury (Fig. 3m, o versus Fig. 3l).

3.4 Microscopic injury underlying yellowish dots

The structural injury underlying yellowish necrotic dots showed traits differing in several instances from those in the case of bronzing and dark stippling (Table 1). The most severe symptoms were found in the lower mesophyll, where entire tissue segments were necrotic and collapsed and the intracellular space was obstructed by leaked material (Figs. 3g, n and 4x). Within the upper mesophyll, strands of cells showing advanced degeneration alternated with others that were cytorrhyzed and disrupted (Figs. 3h and 4t). Advanced degeneration was indicated by (i) cytoplasm and organelle condensation (Fig. 4u, x), (ii) increase in plastoglobule density (Fig. 4u, x) and (iii) accumulation of oxidized tannins in the vacuole (Fig. 3g). Besides cytorrhysis (i), the markers of cell death in cytorrhyzed cells included the (ii) disruption of cell content (Figs. 3h and 4t), (iii) nucleus pyknosis (Fig. 4w) and (iv) degradation of the lamellar system and envelope in chloroplasts (Fig. 4v), whereas only a few and small plastoglobules were observed.

3.5 Microscopic injury in epidermis

Irrespective of the type of visible injury, symptoms in epidermal cells were generally less severe and less variable than those prevailing in mesophyll (Table 1). The micromorphological markers of oxidative stress in the upper and lower epidermis were similar (Fig. 6c, d versus Fig. 6j), but injury was occasionally more severe in the upper epidermis (Fig. 3g, i). At the cell level, the degenerative traits within symptomatic material included the (i) condensation of the cytoplasm (Fig. 6j), (ii) decrease in thylakoid size and apparent enlargement of plastoglobules within leucoplasts (Fig. 6c, d, j), (iii) degeneration of cristae and matrix in mitochondria (Fig. 6d) and (iv) accumulation of condensed tannins (Figs. 3c, g and 6c, d, j)—but not of lipids (Fig. 5b)—in the vacuole. The less severe injuries, as compared to mesophyll cells, may relate to lower physiological activity, as indicated by the thinner cytoplasm layer, fewer organelles, larger vacuome and flattened and slightly condensed nucleus (Fig. 6a, b, i versus Fig. 6e, f).

4 Discussion

4.1 Visible injury

Irrespective of the O3 regime, a broad spectrum of visible O3-like symptoms and symptom combinations was observed in the upper canopy of adult beech trees during the 8-year duration of CASIROZ experiment. Photobleaching, bronzing and dark stippling symptoms were typical of oxidative injury by O3 stress in beech. They were similar to evidence (i) reported from field studies (Bussotti and Ferretti 2009; Hůnová et al. 2011; Vollenweider et al. 2003), (ii) reproduced experimentally (Gerosa et al. 2008; Innes et al. 2001; Novak et al. 2003) or (iii) used as hallmarks of O3 injury in beech during monitoring (Günthardt-Goerg 2013; Sanz and Calatayud 2011). Photobleaching symptoms alone lack specificity with respect to O3 stress diagnosis (Vollenweider and Günthardt-Goerg 2006). The yellowish dots observed in the two O3 treatments were atypical for beech. Apparent on both leaf sides, they thus showed morphological features contrasting with O3 injury common in broadleaved trees. However, Nunn et al. (2002) reported an apparent increase in the frequency of this symptom under the twice-ambient versus ambient O3 regime. Moreover, light-coloured stipples on the adaxial leaf side can be observed in several broadleaved species, in response to elevated O3 levels (Innes et al. 2001; Sanz and Calatayud 2011). These symptoms also resembled mottling injury in conifers (Miller et al. 1996; Vollenweider et al. 2013). Particularly in the case of yellowish dots and photobleaching, microscopic validation was needed, therefore, to confirm the causal link between O3 stress and morphological traits.

Fig. 7
figure 7

Conceptual model of O3 injury variation by the end of vegetation season and in relation to O3 uptake. The spectrum of visible symptoms and microscopic markers of O3 stress late in the vegetation season depends on the yearly O3 uptake (each grey line). Degenerative or disruptive microscopic injury appears at lower O3 dose than the overlying visible symptoms. The term discoloration indicates a change in foliage hue or colour (e.g. photobleaching or bronzing, reddening) as a consequence of chlorophyll degradation or accumulation of compounds from the secondary metabolism. The induction of cellular responses (ACS, HR-like) depends on the species-specific exceedance of O3 uptake threshold, tending to more variable values later in the vegetation season as a consequence of varying foliage phenology. Application range: broadleaved trees and shrubs in temperate and tropical biomes

Full size image

4.2 Cellular responses to O3 stress responsible for the degenerative changes

The broad spectrum of microscopic injury (Table 1) was indicative of the several distinct or overlapping plant responses which O3 triggered in foliage of beech trees from the CASIROZ experiment over the 8-year experiment. Whatever the macroscopic symptom type, chloroplast injury was consistently observed as a distinct indication of oxidative stress, even in the case of mild symptoms (Figs. 3c, d and 4f, g). The production of reactive oxygen species (ROS) is a side effect of photosynthetic activity and further enhanced in the case of O3 stress (Foyer and Shigeoka 2011; Vainonen and Kangasjärvi 2015). Hence, chloroplasts can show incipient (Desotgiu et al. 2010), severe (Vollenweider et al. 2013) and specific injury (Fink 1999; Günthardt-Goerg and Vollenweider 2007) in response to O3 stress. Given photobleaching and bronzing symptoms, the structural changes in chloroplasts appeared basically similar to those during autumnal senescence in beech leaves (Mikkelsen and Heide-Jørgensen 1996; Tevini and Steinmüller 1985). Consistent with degenerative symptoms in other cell compartments, they denoted, therefore, accelerated cell senescence (ACS; Inada et al. 1998; Vollenweider et al. 2003). Further, the increased amounts of oxidized waste lipids being encapsulated within enlarged plastoglobules and transferred to vacuoles (Fig. 5e) were indicative of enhanced lipid trafficking and thylakoid membrane turnover (Bréhélin et al. 2007; Mikkelsen and Heide-Jørgensen 1996). In the case of milder symptoms, however, lipids did not accumulate in the vacuole, probably in relation to the apparent degradation of excreted plastoglobules. Degenerative changes in epidermal cells (Fig. 6c, d, j) were also indicative of ACS.

Some interactions in the vacuole between lipid excretion mediated by plastoglobules and tannin deposition—after formation within the endoplasmic reticulum (Evert 2006)—may explain the peculiar ultrastructure of sponge-like tannin bodies (Franceschi et al. 1998). When both degradation products were observed in the same vacuolar compartment, the lipids appeared segregated, therefore, within spherical cavities of tannin bodies (Fig. 5g). The cavities showing an electron-translucent content might result from subsequent lipid degradation, whilst the polymerized tannin bodies tended to show cracks but never signs of hydrolysis. Large amounts of condensed tannins can be observed in stressed beech foliage (Bussotti et al. 1998; Günthardt-Goerg and Vollenweider 2007; Oksanen et al. 2001), and, together with increasing tannin oxidation adaxially, they form hallmarks of O3 stress (Calderón Guerrero et al. 2013; Hartmann et al. 2007; Vollenweider et al. 2003). Similar to other intra- and intercellular gradients of oxidative injury within the studied material (Figs. 3c and 4k), such markers were indicative of synergistic effects of O3 and photo-oxidative stress (Foyer et al. 1994; Günthardt-Goerg and Vollenweider 2007), as they contributed to visible symptom expression.

4.3 Contrast between degenerative and disruptive traits underlying stippling symptoms

The observation of cell degeneration together with cell death markers within the same mesophyll cells (Fig. 4k) was indicative of the occasional overlap of ACS and HR-like processes within stippling symptoms. Whilst ACS progressively develops during the growing season (Pell et al. 1997), the HR-like processes are triggered once an oxidative stress threshold has been exceeded, causing rapid cell death soon thereafter (Calderón Guerrero et al. 2013; Rao and Davis 2001). In chloroplasts of some cells exhibiting HR-like traits (Fig. 4q), missing degenerative features, such as large and numerous plastoglobules (Lichtenthaler 2013), thus suggested cell death early in the growing season. The disrupted cellular material further contrasted with such undergoing degenerative changes by showing deformation and coagulation, whilst the inner and outer membrane systems were fused, broken or partially hydrolysed (Fig. 4s). Other HR-like hallmarks included nuclear pyknosis (Fig. 4r; Jones 2000; Pasqualini et al. 2003) and incomplete degradation of cellular residues (Fig. 4s; Fukuda 2000). Major injury to cell walls, as observed in other cases of HR-like processes triggered by O3 stress (Calderón Guerrero et al. 2013; Paoletti et al. 2010; Vollenweider et al. 2013), was indicated by leakage of cellular material into the apoplast (Fig. 5b). Regarding the aforementioned accumulation of lipids, and in the case of inclusions contiguous to incompletely degraded structures within dead cells (Figs. 4f and 5d), our observations suggest a decrease in the hydrolytic capacity of vacuoles shortly before cell death. Disruption of the latter hydrolytic compartment forms a terminal degeneration stage of cell content during senescence (Inada et al. 1998). In contrast to ACS, vacuole disruption preceded—maybe impeded—full cell content degradation in the case of HR-like events (Fig. 3j and 4k), which might represent a benchmark trait to differentiate between the slow ACS degeneration process and the quick HR-like cell death.

4.4 Structural indications of O3 injury in yellowish dots

Underlying necrotic dots (Fig. 3g, h, n), strands of cells, either degenerated or disrupted, indicated typical contrasts between ACS or HR-like processes. The absence synergism with photo-oxidative stress effects, together with major injury next to stomata in tissues of the lower leaf blade, suggests oxidative injury by O3 stress alone. Owing to the tissue arrangement within needles, mottling injury by O3 stress in pines is generally closely associated with stomata (Evans and Miller 1972; Soda et al. 2000; Vollenweider et al. 2013). In angiosperms, cell death primarily within the lower mesophyll in response to O3 stress has been observed after exposure to high O3 concentrations (150 ppb; Faoro and Iriti 2005; Pasqualini et al. 2003). The indications of severe O3 stress leading to rapid cell death included (i) necrotic cells also in upper mesophyll (Fig. 3g), (ii) massive leakage of cell debris (Fig. 5b) or (iii) small plastoglobules inside chloroplasts (Fig. 4v). Hence, the singular structural traits underlying the yellowish dots suggest origin from acute O3 stress and locally high O3 concentrations. The severity of injuries peaking in the lower mesophyll also explains that the symptoms become visible on both leaf sides.

4.5 Variability of structural O3 injury in relation to O3 uptake

The variability of visible macroscopic symptoms, underlying microscopic injury and cellular responses to O3 stress observed during the CASIROZ long-term experiment, provided insights on the dependency of macro- and microscopic symptoms by the end of the growing season on the seasonal ozone uptake (Fig. 7). Closely assigning the microscopic injury to visible symptoms during microscopic analyses, the precedence of microscale injury became evident, as macro-scale visibility emerged only once a sizable group of cells had been affected. The concomitance of degenerative and disruptive traits within the same cells can only be understood in terms of a succession, with ACS progressing until a species-specific threshold is reached, so that HR-like processes are triggered, which incite the rapid cell death. In 2003, this succession occurred at an earlier date, as indicated by the low accumulation of tannins and tannin plugs in the mesophyll (Fig. 3i, j). In the case of massive O3 dose, structural injury should mainly relate to HR-like processes. The visible symptom progression commonly observed in the field, with photobleaching symptoms sometimes in summer prior to discoloration and subsequent stippling appears (authors’ unpublished observations) can also be conceptually understood in terms of succeeding ACS and HR-like processes. Hence, by the end of the vegetation season, when most observations of ozone symptoms are realized in the field (Schaub et al. 2010), the symptom types and macro- and micromorphology traits will synthetically reflect the reactions and cellular responses to the experienced ozone dose.

5 Conclusion

In the CASIROZ experiment, a large spectrum of typical as well as atypical O3 macroscopic and microscopic injury was observed during the 8-year period. Variation was encountered especially regarding the microscopic markers of O3 stress from year to year. This variability primarily related to (i) ACS and HR-like plant responses either distinct or overlapping, (ii) interaction between O3 and photo-oxidative stress and (iii) yearly variation in O3 dose and uptake rates. Especially traits of degeneration and disruption showed contrasted structural features. Concerning stippling symptoms, the structural variability was particularly high, as a consequence of interacting ACS and HR-like responses and the initiation of programmed cell death at different times of growing season. Hence, the expression of macroscopic and microscopic symptoms by the end of summer can provide an integrative view of successive plant responses and symptom development during the preceding course of the growing season. The difficulties posed by the symptom variability can be alleviated conceptually, when relating the injuries observed at different scales to the experienced ozone dose and environmental conditions. Within a changing climate and ozone pollution context, farther variability of O3 injury in trees can be expected. Based on the present observations, symptom variability will nevertheless arise from shifting prevalence of typical O3-induced responses and prominent structural markers, as observed so far.