Ozone foliar symptoms in woody plant species assessed with ultrastructural and fluorescence analysis


  • Filippo Bussotti,

    Corresponding author
    1. Department of Plant Biology, University of Florence. Piazzale delle Cascine 28, 50144 Firenze, Italy;
      Author for correspondence:Filippo Bussotti Tel: +39 0553288389 Fax: +39 055360137 Email: filippo.bussotti@unifi.it
    Search for more papers by this author
  • Giovanni Agati,

    1. Istituto di Fisica Applicata ‘Nello Carrara’– IFAC, Consiglio Nazionale delle Ricerche, via Madonna del Piano, 50019 Sesto Fiorentino, Firenze, Italy
    Search for more papers by this author
  • Rosanna Desotgiu,

    1. Department of Plant Biology, University of Florence. Piazzale delle Cascine 28, 50144 Firenze, Italy;
    Search for more papers by this author
  • Paolo Matteini,

    1. Istituto di Fisica Applicata ‘Nello Carrara’– IFAC, Consiglio Nazionale delle Ricerche, via Madonna del Piano, 50019 Sesto Fiorentino, Firenze, Italy
    Search for more papers by this author
  • Corrado Tani

    1. Department of Plant Biology, University of Florence. Piazzale delle Cascine 28, 50144 Firenze, Italy;
    Search for more papers by this author

Author for correspondence:Filippo Bussotti Tel: +39 0553288389 Fax: +39 055360137 Email: filippo.bussotti@unifi.it


  • • This paper compares the responses to ozone in five woody species: Fagus sylvatica (FS), Acer pseudoplatanus (AP), Fraxinus excelsior (FE), Viburnum lantana (VL) and Ailanthus altissima (AA). The hypothesis being tested was that the strategies that plants adopt to resist oxidative pressure are species-specific.
  • • The study was carried out on field grown plants in an area in Northern Italy characterized by elevated levels of ozone pollution. The observations were made both at ultrastructural (using light and electronic microscopy) and physiological (using chlorophyll a transient fluorescence and microspectral fluorometry) level.
  • • Common responses were: the hypersensitive response (i.e. the death of palisade mesophyll cells) and the formation of callose layers separating injured from healthy cells. FS and AP were capable of thickening the palisade mesophyll cell walls. This thickening process involved changes in cell wall chemical structure, evidenced by the accumulation of yellow autofluorescence compounds. Species-specific behaviours were observed with the fluorescence analysis, with special reference to the photochemical de-excitation constant (Kp). This value increased in FE and AP, and decreased in AA.
  • • The observed responses are interpreted as adaptative strategies against the ozone stress. The increase of Kp indicates that the reaction centres were working as more effective quenchers.


High ozone uptake in the leaves of sensitive plants induces a wide variety of symptomatic manifestations and an equally wide-ranging variability in cell responses, both ultrastructural (Vollenweider et al., 2003) and physiological (Gravano et al., 2004). The types of response that give rise to visible and clearly recognisable injury may be grouped into two categories: hypersensitive response (HR), accompanied by processes of oxidative burst (OB), and pigment accumulation (Wohlgemuth et al., 2002; Vollenweider et al., 2003). A HR to ozone consists of the programmed cell death (PCD), that is the death of singular or clusters of palisade mesophyll cells (Schraudner et al., 1997; Pell et al., 1997; Wohlgemuth et al., 2002), and produces stipples on the upper (adaxial) surface of the leaf, indicating the presence of necrotic areas. The presence of stipples may, however, equally be the outcome of the localised oxidation of the cell contents (OB), and/or the accumulation of pigments, such as anthocyanins, which exert a detoxifying role (Guderian et al., 1985; Steyn et al., 2002). In some species, instead of being localised, the pigmentation may affect most of the mesophyll and epidermis, in an extensive way. In those cases the stipples are not clearly recognisable; one observes, instead, an overall change in the leaf colour.

As far as photosynthetic activity is concerned, the net photosynthesis is reduced in the symptomatic leaves with respect to the nonsymptomatic ones (Zhang et al., 2001). During the early stages of the injury, however, in Fraxinus excelsior (HR species) chlorophyll efficiency may be positively stimulated (Gravano et al., 2004) thanks to the activation of metabolic responses to compensate for the reduced number of photosynthetically active cells. Conversely, in Viburnum lantana and Prunus avium (which respond with leaf reddening) an increase of the dissipation processes was observed (Gravano et al., 2004). These different physiological responses were described by analysing the direct fluorescence of chlorophyll a (fluorescent transient O-J-I-P, cf. Strasser et al., 2000, 2004).

The aim of this paper is to test the hypothesis that there is a relationship between ultrastructural and physiological responses, and these responses define species-specific strategies to ozone stress. Five woody species were tested (Ailanthus altissima Swingle – AA; Fraxinus excelsior L. – FE; Acer pseudoplatanus L. – AP; Fagus sylvatica L. – FS and Viburnum lantana L. – VL), all of them growing spontaneously in areas with high tropospheric ozone concentrations.

Materials and Methods

Study area

The study was carried out on young trees growing spontaneously around the margins of the Moggio beechwood, in the Northern Italian Alpine region (Lat. +455441; Long. +093017; altitude 1250 m asl; mean yearly precipitations: 1500 mm; mean yearly mean temperature: 8°C), 100 m from the permanent monitoring plot belonging to the Italian intensive forest monitoring program CON.ECO.FOR. (Allavena et al., 2000), and around the open field meteorological station equipped with passive samplers for ozone. Mean ozone concentrations were measured, in the years 2001–03, by means of the passive samplers developed at the University of Munich, Department of Forest Bioclimatology and Immission Research (Werner, 1992; Hangartner et al., 1996). The samplers were installed at a height of 2 m at the open field meteorological station. Paper disks were collected weekly and sent to a central laboratory for visual colorimetric analysis (Buffoni, 2002).

The study area is subjected to high ozone exposure levels, as a result of the transport of precursors generated by the large conurbation of Lumbardy (Sandroni et al., 1994; Gerosa et al., 1999; Vecchi & Valli, 1999). Ozone mean concentrations for the period April to September were (in µl l−1): 57.8 (2001); 67.2 (2002) and 73.8 (2003). The corresponding AOT40 (Accumulated Ozone above the Threshold of 40 ppb) levels (modelled from mean concentrations, according to Gerosa et al., 2004) were, respectively (in ppmh): 51.7 (2001); 53.7 (2002) and 96.9 (2003). The unusually high exposure levels in 2003 correspond to an exceptionally hot summer: mean temperature from May to August was 18.3°C in 2003, compared with 15.6°C the year before.

Visible symptom assessment and microscopy

Four woody species were sampled at the edge of the beechwood (FE, AP, FS and VL), in full light conditions (eccept VL, that grows in the understore). A fifth species (AA) was sampled from the suburbs of the nearby urban centre of Moggio (890 m asl). Five isolated young trees (1–3 m tall) per species were considered for all measurements made in this study. The assessments were carried out on three different dates during the summer of 2003: 8 July, 29 July and 1 September. The exposure levels, calculated starting from the full leaf emission (end of April), were, respectively, 44.8 ppmh; 59.3 ppmh and 78.2 ppmh. The foliar symptoms were recognised according to the available field manuals (Innes et al., 2001; Working Group on Air Quality, 2003). For FE, AP, VL and FS, foliar symptoms were validated by comparison with the results on the same species obtained in the open-top chamber experiment in southern Switzerland (Canton Ticino, about 50 km from Moggio; VanderHeyden et al., 2001; Novak et al., 2003; Gravano et al., 2004). For AA the comparison was done with the finding of a gradient study in central Italy (Gravano et al., 2003). Microscopical analysis (Vollenweider et al., 2003) gave further evidence that ozone was the causal agent of the observed symptoms (see Results section).

Three leaf samples (representing the mean foliar intensity of symptoms on each tree; Working Group on Air Quality, 2003) were collected from each tree (five trees per species), for microscopy. The observations by light and fluorescent microscopy were performed with a Zeiss Axioplan (Oberkochen, Germany) microscope on fresh sections (30 µm thick) cut with a freezing microtome (Reichert-Yung, Wien, Austria), and on sections embedded in historesin (2 µm thick) cut with an ultramicrotome Ultracut S (Reichert-Yung, Wien, Austria). The tests performed (O’Brien & McCully, 1981) included: toluidine blue, pH 4.4 for the structure and a first explorative survey; aniline blue, on 30 and 2 µm thick sections observed with UV filter 350–390 nm for callose; PAS (Periodic Acid Schiff) and fluorescent PAS and calcofluor for polysaccharides; acidified vanilline for tannins.

Samples destined for observation by Transmission Electron Microscope (TEM) were prefixed in phosphate buffer (pH 7.2) containing 2.5% glutaraldehyde +4% paraformaldehyde. After 20 h at 5°C, samples were rinsed twice (2 × 10 min) in the same buffer, then postfixed (2 h) in 2% osmium tetroxide prepared in the same buffer. Subsequently, samples were dehydrated in an increasing ethanol series (10 min at each stage of the fixation series). Finally, after two 5-min rinses in propylene oxide (100%), the samples were embedded in resin, according to Spurr's procedure (Spurr, 1969). A Reichert Ultracut S (Leica, Heerhrugg, CH) microtome was used to cut ultra-thin sections (0.09 µm) with a diamond knife. These sections were stained with uranyl acetate (500 mg in 10 ml of distilled water) and lead citrate (saturated soluion), and then observed with a EM-300 Philips (Amsterdam, the Netherlands) microscope.

Analysis of fluorescence transient O-J-I-P

Chl a fluorescence transients of intact attached leaves were measured in vivo with an ADC FIM 1500 (ADC Bioscientific Ltd, Hodderston, UK) fluorometer (direct fluorescence) at the 1 September assessment. For each species 40 different leaves representing different levels of symptom diffusion (foliar symptom diffusion was expressed as percentage of the affected leaf area, according to a 5% step scale), were randomly assessed from the five selected trees. Before each measurement, the leaves were dark-adapted for 30 min with leaf clips. The rising transients were induced by a red light (peak at 650 nm) of 600 Wm−2 provided by an array of six light-emitting diodes; they were recorded for 3 s, starting from 50 µs after the onset of illumination, with 12 bit resolution. Data acquisition was performed every 10 µs for the first 2 ms, every 1 ms until 1 s, and every 100 ms from 1 to 3 s. On a logarithmic time scale, the rising transient from F0 (F at 50 µs, when all the reaction centres of the PSII are open, i.e. when the primary acceptor quinone QA is fully oxidised) to FP (where FP = FM under saturating excitation light, when the excitation intensity is high enough to ensure the closure of all reaction centres of PSII, that is the full reduction of all reaction centres) had a polyphasic behaviour (Strasser & Govindjee, 1992a,b; Strasser et al., 1995). Analysis of the transient took into consideration fluorescence values at 50 µs (F0, step 0), 100 µs (F100), 300 µs (F300), 2 ms (step J), 30 ms (step I) and maximal (FM, step P). This method is called JIP-test (Strasser et al., 2000). Table 1 summarizes the technical parameters of curves necessary for further elaborations, as well as the selection of the JIP-test parameters used in this study.

Table 1.  Explanation of the technical data of the O-J-I-P curves and the selected JIP-test parameters used in this study
Technical fluorescence parameters
AreaArea between fluorescence curve and FM
F0F50 µs, fluorescence intensity at 50 µs
F100 µsFluorescence intensity at 100 µs
F300 µsFluorescence intensity at 300 µs
FJFluorescence intensity at the J-step (at 2 ms)
FIFluorescence intensity at the I-step (at 30 ms)
FMMaximal fluorescence intensity
FV : F0(FM − F0) : F0 = Kp : Kn
V : Δt)0 or M0Slope of the curve at the origin of the fluorescence rise. It is a measure of the rate of the primary photochemistry. M0 = 4(F300 − F0) : (FM − F0)
VJRelative variable fluorescence at 2 ms. VJ = (F2 ms – F0) : (FM − F0)
VIRelative variable fluorescence at 30 ms. VI = (F30 ms – F0) : (FM − F0)
Quantum efficiency or flux ratios
φP0 or TR0 : ABSTrapping probability, or Quantum yield efficiency. Expresses the probability that an absorbed photon will be trapped by the PSII reaction centre. φP0 = (FM − F0) : FM
ψ0 or ET0 : TR0Expresses the probability that a photon trapped by the PSII reaction centre enters the electron transport chain. ψ0 = 1 − VJ
De-excitation constants (for their relative formulae, Havaux et al., 1991; Strasser et al., 2000)
KpPhotochemical de-excitation constant
KnNonphotochemical de-excitation constants, summing up KH (for heat dissipation) and KF (for fluorescence emission)
SumKKp + Kn
Density of reaction centres
RC : CS0Gives the number of active RCs to one inactive RC for a PSII cross – section. RC : CS0 = φP0 (VJ : M0)F0
Performance index
PIABSSee text. PIABS = (RC : ABS) [φP0 : (1 − φP0)][ψ0 : (1 − φP0)]

Multispectral fluorescence microimaging and microspectrofluorometry

This analysis was performed on symptomatic leaves of AP and FS collected on 1 September 2003. Leaves were maintained at about 4°C, attached to 20-cm-long shoots, wrapped in wet paper and analysed within 2 d. Cross sections (50-µm thick) of fresh leaf tissue, cut with a vibratory microtome (Vibratome 1000 Plus, Vibratome, St. Louis, MO, USA), or leaf pieces, hand cut with a razor blade, were mounted in phosphate buffer (pH 6.8) with the addition of 1% NaCl (w/v) and observed by the inverted epi-fluorescence microscope previously described (Agati et al., 2002). The acquisition system consisted of both a charge-coupled device (CCD) camera for acquisition of fluorescence images and a CCD multichannel spectral analyser to record fluorescence spectra in the 400–800 nm range. Switching between detectors was allowed by a mobile mirror permitting a rapid sequential acquisition of images and spectra on the same sample area. The excitation wavelength was selected by using 10 nm bandwidth interference filters, namely 365FS10-25 and 436FS10-25 (Andover Corporation, Salem, NH, USA) and dichroic mirrors ND400 and ND510 (Nikon, Japan) for the 365 and 436 nm excitation wavelengths, respectively. The CCD camera (Chroma CX260, DTA, Italy) was equipped with a motorised filter wheel carrying eight different interference filters that permitted the sequential acquisition of fluorescence images on specific spectral bands. Recording fluorescence spectra was useful for compound identification and to drive the choice of acquisition bands in the multispectral fluorescence microimaging.

Under UV excitation three fluorescence images were sequentially acquired on narrow (10 nm) bands centred at 470, 546 and 680 nm, with integration times between 2 and 8 s, selected by the 470FS10-25, 546FS10-25 and 680FS10-25 interference filters (Andover Corporation, Salem, NH, USA), respectively. With blue excitation only the yellow (546 nm) and red (680 nm) bands were measured. Monochrome images were then recombined after band-colour assignment in a single multicolour image using the RGB technique by the Image-Pro Plus v.4.0 software (Media Cybernetics, Silver Spring, MD, USA). The blue, green and red colours were attributed to the 470, 546 and 680 nm fluorescence images, respectively. The RGB (RG for the 2-band acquisition under 436 nm excitation) visualisation mode showed the colocalisation of compounds with different fluorescence properties. Image spatial calibration using a × 10 Plan Fluor (NA = 0.3) objective was 0.79 µm per pixel.

Profiles of fluorescence along the leaf cross sections were obtained by plotting the mean intensity of each row of pixels vs a longitudinal direction. When fluorescence spectra were recorded, residual excitation light was removed by GG400 and GG475 long-pass filters (Schott Glas, Mainz, Germany) for the 365 and 436 nm excitation wavelengths, respectively. Fluorescence spectra were corrected for the transmission spectra of optics and filters, and reported in a smoothed form.

Leaf cross sections were also observed after incubation in 0.1% (w/v) 2-amino ethyl diphenyl boric acid (Naturstoff reagent, NR) in phosphate buffer (pH 6.8) with addition of 1% NaCl (w/v), as reported by Hutzler et al. (1998). Staining time was 5 min and excess NR solution was removed by 2 min washing with phosphate buffer. Leaf cross sections extracted with MeOH for 1 h were also analysed.


Data recorded by direct fluorescence were processed through Biolyzer 3.0 software (1999–2001, by Ronald M. Rodriguez, Bioenergetics Laboratory, Geneva, CH). Combined elaborations were performed by relating fluorescence data and data on leaf injury to each individual leaf. All analysis were processed by means of Statistica 6.0 (Statsoft 2001, Tulsa, OK, USA). Statistical elaborations included the analysis of the regressions between the foliar extension of leaf injury and the JIP-test parameters. A multivariate statistical technique, the Principal Factors Analysis, was used to research and reveal structures of variability and multiple correlations between variables within the data. The selection of the number of factors was based on eigenvalues (over 1).


Symptom development and ultrastructure analysis

On the date of the first sampling (8 July), only AA and – to a lesser extent – FE displayed visible symptoms, in the interveinal area of the upper (adaxial) surface of the leaf. AA presented whitish stipples (that later were to turn brownish), while in FE they were brown. On the date of the second sampling (28 July), the symptoms were extensive in both above-mentioned species and began to be visible also in FS and AP (brown stipples in both cases) and in VL (reddish stipples). By the last sampling date (1 September), extensive brown patches were visible in all species, and reddening in VL.

The detailed findings of the ultrastructure observations in FS and AP are reported in Figs 1 and 2, respectively. The features of the other species are displayed in Fig. 3. More detailed information has been reported in previous papers: Gravano et al. (2003) for AA, and Gravano et al. (2004) for FE and VL. For each date, 15 samples per species were examined: we report the prevalent microscopical features for each species. In all investigated species, the development of symptoms was associated with the production of callose in the palisade mesophyll cells. Callose was visible on the walls of the cells around the areas beneath the stipples (Figs 1a, 2a and 3b). Another feature common to all investigated species was the development of HR in the upper mesophyll cells (Fig. 1e, 2c, 3c, 3f,g). These responses were preceded by the degeneration of the cytoplasm and the cell content (Figs 2d and 3b). The presence of collapsed cells was evident in AA, FE and AP from the very first appearance of symptoms, whereas it only became visible in FS and VL at the end of the growth season. In two species (FS and AP) the development of symptoms was accompanied also by thickening of the palisade mesophyll cell walls, due to the accumulation of cellulose layers, which reduced the inner space of the cells (Fig. 1b,f,g,h and Fig. 2f,g). The synthesis of structural compounds, occurring in the development of the cell wall, calls for an intensive metabolic activity, especially on the part of the Golgi apparatus, as Fig. 2e shows. The cell wall thickening, which was particularly marked in the samples collected at the end of the summer, displayed an irregular trend. Cell extrusions of polysaccharide nature were observed on the walls of the spongy mesophyll in VL. No alterations were observed on stomata. Except for FS, the species investigated had no tannins among their vacuolar contents. In FS the tannins were normally abundant within the vacuoles of palisade mesophyll cells, but undergo degenerative processes in the areas where stipples were formed (Fig. 1c).

Figure 1.

Ultrastructural features of Fagus sylvatica. (a) Aniline blue stained section, observed with UV fluorescence. Arrows indicate callose apposition on the walls. (b) PASS reaction. Arrows indicate a local thickening of the wall, by polysaccharide compounds. (c) Vanilline stained section. Asterisks indicate the lower response of phenols in the damaged cells. (d) TEM. Healthy chloroplast and thin wall in an undamaged leaf. The vacuole is filled with tannins. (e) TEM. Cells from the palisade mesophyll in the damaged leaves. Cells are partly collapsed. Arrows indicate the thickening of the walls. Asterisks indicate the absence of tannins in the vacuoles. (f) TEM. Simple arrows indicate a particular of the thickened wall. Double arrows indicate the apposition of callose. (g) Calcofluor stained section, observed with UV fluorescence. Arrows indicate appositions of cellulose to the wall. (h) TEM. Asterisks indicate the thickening of the primary cell wall (cw).

Figure 2.

Ultrastructural features of Acer pseudoplatanus. (a) Aniline blue stained section, observed with UV fluorescence. Arrows indicate callose appostion on the walls. (b) TEM. Healthy chloroplast in an undamaged leaf. (c) PASS reaction. Collapsed cells are visible. Arrows indicate starch accumulation in the damaged cells. (d) TEM. Chloroplast in a damaged leaf; cw = cell wall. (e) TEM. Cytoplasm in palisade cells of damaged leaves, rich in vesicle of the Golgi apparatus (see upper-right square). (f) TEM. Thickened cell walls in the palisade mesophyll; cw = cell wall. Arrows indicate the median lamella. (g) TEM. Strongly thickened cell walls in the palisade mesophyll, in contact with the upper epidermis; cw = cell wall. (h) TEM. Particular of (g); cw = cell wall. Arrows indicate the lack of the median lamella.

Figure 3.

Ultrastructural features of Viburnum lantana, Fraxinus excelsior and Ailanthus altissima. (a–e) Ultrastructural features of Viburnum lantana. (a) Toluidine blue stained section. Collapsed cells are visible in the upper epidermis and the palisade tissue. (b) TEM. Chloroplasts in damaged cells. Asterisk indicates the presence of callose. (c) Collapsed cells in the palisade tissue, without any thickening of the wall. (d) PASS reaction. Arrows indicates polysaccharidic extrusions from the cell wall. (e) TEM. Particular of a polysaccharidic extrusion. (f–h) Ultrastructural features of Fraxinus excelsior. (f) Toluidine blue stained section. Collapsed cells are visible in the palisade tissue (arrow). (g) TEM. Apposition of callose. (h) TEM. Particular of collapsed cell (g) Ailanthus altissima. Toluidine blue stained section. Collapsed cells are visible in the palisade tissue (arrow).

Multispectral fluorescence microimaging and microspectrofluorometry

Figure 4 shows the autofluorescence microscopic analysis of an injured AP adaxial leaf area. The autofluorescence images of Fig. 4a,b,c are the acquisitions at 470, 546, and 680 nm, respectively, under UV excitation (365 nm). Figure 4d shows the RGB recombination of the three (A, B, C) monochrome images after blue, green, and red colour assignment to the 470, 546, and 680 bands, respectively. The stippled area was characterised by a blue-green emission under UV excitation. The healthy tissue surrounding the damaged zone showed the typical red autofluorescence of chlorophyll (Chl). Blue-green autofluorescence appeared located on the cell walls, although we can not rule out contributions from the cell vacuoles. Autofluorescence spectra shown in Fig. 4e indicate the reduction of Chl inside the injured zone, as demonstrated by both the lower red fluorescence intensity and the change in the Chl fluorescence spectral shape. In fact, the lower ratio between the red and far-red Chl bands in healthy tissues, compared with damaged tissues, was due to a higher reabsorption effect because of a higher Chl concentration (Agati et al., 1993; Gitelson et al., 1998). Blue dominant emitting zones of damaged tissues peaked at about 470 nm under UV excitation, while blue excitation induced a yellow band at about 540 nm.

Figure 4.

Multispectral autofluorescence microimaging and microspectrofluorometry of an Acer pseudoplatanus adaxial leaf area injured by ozone stress. (a,b,c) autofluorescence images under UV excitation (365 nm) recorded at 470, 546, and 680 nm, respectively. (d) RGB recombination of the three (a,b,c) monochrome images with blue, green, and red assigned to the 470, 546, and 680 bands, respectively; bar = 50 µm. (e) Autofluorescence spectra of injured tissues under UV- (blue and green lines) and blue- (yellow line) excitation and healthy tissues (red dashed line) under UV excitation. Fluorescence intensity of UV- and blue-excited spectra is reported on the left and right y-axis, respectively.

Further features of the fluorescence microanalysis are reported in Fig. 5 for AP and FS leaf cross sections. In AP, the fluorescence image recorded at 546 nm (F546) under blue excitation (436 nm) shows that there was an increase of yellow autofluorescence in necrotic palisade areas compared with surrounding photosynthetically active mesophyll tissues (Fig. 5a). The corresponding image at 680 nm (F680) indicates the almost complete disappearance of Chl in the damaged palisade, while intact chloroplasts were present in the palisade cells next to the necrotic ones and in the spongy parenchyma even beneath the injured palisade. The profiles of both F546 and F680 signals acquired along a longitudinal axis (dotted arrows in Fig. 5a) are plotted in Fig. 5b. They show that the Chl fluorescence decreased by about eight times on average, going from the healthy to the damaged palisade, while the yellow autofluorescence increased twice in the same zones. It is worth noting that the Chl degradation is strictly spatially overlapped to the yellow autofluorescence increase. This is also well represented in the RG image merging of Fig. 5c, where green and red colours correspond to the yellow autofluorescence and Chl fluorescence, respectively. Microspectrofluorometry (Fig. 5d) confirms multispectral imaging showing that fluorescence signals, integrated over a palisade injured spot of the yellow and Chl fluorescence, were four times higher and lower, respectively, than those of the proximal normal tissue. Figure 5e,f shows fluorescence images and spectra of AP leaf cross sections treated with NR in order to check if there was an accumulation of flavonoids within the injured area. Figure 5e shows that green-coloured fluorescence corresponding to the 546 nm band of the necrotic palisade was more intense than the corresponding autofluorescence in untreated samples (Fig. 5c). This is also evident in the fluorescence spectra under both UV and blue excitation of the necrotic palisade reported in Fig. 5f. Here, however, we observe that there was no change in the spectral shape with or without NR treatment. This evidence suggests that no accumulation of flavonoids occurred within the injured tissues. In fact, if this were the case, one would expect a shift towards longer wavelengths in the fluorescence spectrum of NR-stained sections because of the orange fluorescence contribution of flavonoids (Agati et al., 2002). A further comparison between autofluorescence and NR-induced fluorescence images of Fig. 5c,e, respectively, revealed an enhanced intensity of fluorescence at 546 nm on the adaxial cuticle of NR-treated sections. This increase occurred in the epidermal layer above both damaged and healthy palisade. In FS (Fig. 5g), the accumulation of yellow autofluorescence in the necrotic zone is not as evident as in AP. On the other hand, a marked increase in fluorescence at 546 nm was observed in the cuticular layer above the damaged parenchyma (Fig. 5g,h). Similarly to AP, no flavonoid-related response to NR-staining was observed in the injured tissues of FS. For both AP and FS, the microfluorometric analysis on cross sections extracted for 1 h in MeOH revealed that the yellow autofluorescence remained after MeOH extraction, while Chl was almost completely disappeared even from normal tissues. This result suggests that the intense yellow autofluorescence of necrotic tissues derives from cell wall-bound compounds.

Figure 5.

Multispectral fluorescence microimaging and microspectrofluorometry of cross sections from Acer pseudoplatanus (a–f) and Fagus sylvatica (g,h) leaves injured by ozone stress. Bars = 100 µm. (a) Fluorescence images at 546 (F546) and 680 (F680) nm bands under 436 nm excitation of an A. pseudoplatanus leaf cross section. (b) Fluorescence intensity profiles for F546 and F680 along the longitudinal direction indicated by dotted arrows of (a). (c) RG two-bands merging of the fluorescence images of (a), green and red colours are assigned to the 546 and 680 nm bands, respectively. (d) Autofluorescence spectra under blue excitation of healthy and ozone damaged palisade tissues, the inset represents a scale enlargement of the green-yellow band. (e) RG two-bands merging of 546 and 680 nm fluorescence images of a blue-excited cross section stained with Naturstoff reagent (+ NR) in phosphate buffer. (f) Comparison between autofluorescence and Naturstoff-induced fluorescence spectra under both UV and blue excitation of a necrotic palisade. Fluorescence intensity of UV- and blue-excited spectra is reported on the left and right y-axis, respectively. (g) RG two-bands merging of 546 and 680 nm fluorescence images of an UV-excited F. sylvatica cross section. (h) Autofluorescence spectra under UV excitation of healthy and ozone damaged epidermal layers selected by a × 40 objective lens.

Photosynthetic parameters

The relationship between extension of visible symptoms and photosynthetic parameters, calculated using the JIP-test, was investigated by means of PCA (Table 2) and simple regressions (Table 3).

Table 2.  Factorial analysis of the JIP-test parameters and leaf injury diffusion, for each considered species
 Fagus sylvaticaViburnum lantanaAcer pseudoplatanusFraxinus excelsiorAilanthus altissima
Fact 1Fact 2Fact 3Fact 1Fact 2Fact 1Fact 2Fact 3Fact 1Fact 2Fact 1Fact 2
  • Loadings of each variable for the respective factors (Equamax rotation) are reported. Loadings  0.60 are reported in bold. The last row reports the percent of variance explicated from each factor.

  • 1

    VJ was not considered in the elaboration because autocorrelated to ψ00 = 1 − VJ)

Injury0.75–0.06 0.060.75 0.380.88–0.11 0.21 0.87 0.190.86–0.26
FM 0.90–0.40−0.01 0.83–0.52 0.85 0.50−0.010.97–0.08 0.09 0.99
F0–0.040.98–0.16 0.010.99 0.96 0.23−0.010.96 0.010.94 0.17
Fv : Fo 0.90 0.40 0.14 0.92 0.31 0.26 0.94 0.140.96–0.07 0.95 0.18
M00.60–0.510.600.96–0.24−0.250.86 0.33 0.07 0.960.98–0.11
VI∠0.26−0.170.910.91–0.06−0.330.76 0.44−0.19 0.780.96–0.05
φP0 0.90 0.37 0.07 0.90 0.29 0.27 0.90 0.300.96 0.08 0.93 0.15
ψ0 0.26 0.17 0.91 0.91 0.06 0.33 0.76–0.44 0.190.78 0.96 0.05
SumK 0.03 0.98 0.17−0.09 0.990.96–0.23–0.12 0.97–0.19 0.91–0.36
Kn0.93 0.31 0.030.82 0.490.72–0.58–0.30 0.95–0.23−0.080.99
Kp 0.32 0.93 0.17 0.20 0.970.98 0.05 0.03 0.96–0.15 0.93–0.27
RC/CS0 0.87 0.20 0.37 0.84–0.44 0.80 0.55−0.110.98–0.14 0.71 0.68
PIABS 0.60 0.42 0.65 0.94 0.19 0.28 0.88–0.310.86–0.48 0.90 0.06
Var. Expl. (%)423219593145371169216921
Table 3.  Figures of regressions between leaf injury diffusions and the selected JIP-test parameters
 Fagus sylvaticaAcer pseudoplatanusAilanthus altissimaFraxinus excelsiorViburnum lantana
  1. Regressions were calculated with the normalised values of the JIP-test parameters. The mean value for each JIP-test parameter in the not symptomatic leaves was considered as 1 and, then, the relative variations respect to 1, were calculated in the symptomatic leaves at the different levels of symptoms diffusion.

  2. P indicates the level of significance (in bold are evidenced the regressions with P < 0.05).

  3. b indicates the slope of the regression. Italics indicate the species were the slope for each parameter was steeper.

F00.3210.143 0.0010.9470.000–0.0080.7340.029 0.0130.9740.000–0.0080.1220.442–0.002
Fv : Fo0.7330.007–0.0030.3440.126–0.0020.7310.0300.0090.8880.001–0.0080.5560.054–0.003
M00.6450.016 0.0050.0270.697 0.0010.8350.011 0.0420.0250.737 0.0000.6450.030 0.006
VJ0.0180.750 0.0000.0030.890 0.0000.9340.002 0.0050.1130.461 0.0000.5290.064 0.002
VI0.0040.876 0.0000.1280.384–0.0010.9100.003 0.0030.2320.274 0.0000.3770.142 0.001
ψ00.0180.750 0.0000.0030.890 0.0000.9340.0020.0050.1130.461 0.0000.5290.064–0.002
SumK0.3290.137–0.0010.9200.000 0.0210.6820.0430.0070.9560.000 0.0180.1770.347 0.003
Kn0.3610.115 0.0030.7730.004 0.0340.6290.060 0.0020.8910.001 0.0390.5340.062 0.007
Kp0.6360.018–0.0020.9430.000 0.0170.6970.039–0.0080.9860.000 0.0120.0570.607 0.001
RC : CSo0.7110.009–0.0030.8270.002–0.0080.7640.023–0.0050.9470.0000.0100.5750.048–0.004

PCA revealed that in all investigated species the extent of leaf injury was included in Factor 1. Leaf injury was inversely related to the density of reaction centres (RC : CS0) in all species; with ϕP0 in FS, FE, AA and VL; with Ψ0 in VL and AA and with PIABS in VL, FE and AA. In FS, PIABS was in Factor 3, directly associated with ϕP0 and, inversely, with the kinetic parameters (M0 and VI) of the transient, indicating the accumulation of electron acceptors at different stages of the rising curve. In AP, PIABS is found in Factor 2, associated directly with both Ψ0 and ϕP0. FM was inversely associated with leaf injury in all species except for AA. F0, on the other hand, displayed a species-specific behaviour and was inversely associated with leaf injury in both FE and AP, while it was directly associated in AA. The behaviour of F0 was associated inversely with the sum of de-excitation constants (SumK), which includes the constant of photochemical de-excitation (Kp) and the nonphotochemical de-excitation (Kn). Kp increased as the extension of symptoms in FE and AP increased, whereas it decreased in AA. Kn was directly associated with symptom development in all investigated species except for AA.

The extent of responses in each different species, and for each parameter, was determined by means of the slope of the regression between injury extent and the considered JIP-test parameters (Table 3). AA was the species that showed the most marked responses for the parameters F0, M0, Vi (their values increased as the extent of injury increases) and Fv : F0, ϕP0, Ψ0, PIabs (which decreased). AP presented the most marked response in relation to SumK and Kp (their values increased), while FE reacted the most to F0, RC : CS0 (negatively) and to Kn (positively).


AOT indices are considered not suitable to define the dose of ozone actually assumed from leaves, and they should be replaced by CUO (Cumulative Uptake of Ozone) indices (Karlsson et al., 2004). However, the assessment of foliar fluxes of ozone at site level, implies the availability of micrometeorological and physiological parameters (stomatal conductance to water – gw), very infrequent even in long-term forest monitoring programmes (Gerosa & Anfodillo, 2003). Matyssek et al. (2004) and Karlsson et al. (2004) found, in alpine and in well watered conditions, very strong correlations between the two indices; even CUO was more effective at low ozone levels. Gerosa et al. (2002) demonstrated that, in the subalpine region of Lumbardy, there are no water shortage problems for forest vegetation, even during the dry season.

Species-specific differences in gw determine differences in CUO among species growing in a same environment, but does not explain the differences in the onset, diffusion and intensity of foliar symptoms (Zhang et al., 2001). Previous studies carried out on young trees of some of the same species considered in this paper, evidenced gw gradients: according to Zhang et al. (2001), gw(AP) < gw(VL) < gw(FS) < gw(FE); according to Gerosa et al. (2003), gw(FS) < gw(FE); according to Gravano et al. (2004), gw(FE) < gw(VL). The capacity of detoxification and other active response to cope with oxidative stress are also factors influencing the onset and diffusion of visible symptoms.

In the present research, the species in which symptoms appeared earliest (already by the beginning of the summer) were AA and FE. In VL, AP and FS the onset of symptoms occurred later. Levels of AOT40 recorded at the dates of first appearance of symptoms were in all cases considerably higher than those defined in the literature for these same species (VanderHeyden et al., 2001; Gravano et al., 2003; Novak et al., 2003; Gerosa et al., 2003). In the above-mentioned literature, considerable differences were found, even with reference to the same species, as to the levels of exposure required for the onset of symptoms. This may be explained (apart from genotype and soil moisture conditions; Schaub et al., 2003) by the different degree of foliar reactivity to ozone during the growth season. In the stage immediately after sprouting, high ozone concentrations can influence the development of foliar morphology (Pääkkönen et al., 1997; Bussotti et al., 2005) giving the leaf a greater resistance potential. Furthermore, the detoxifying ability of the leaf varies over the course of the season (Polle et al., 2001). Although at higher exposure levels, the temporal gradient of symptom onset in different species has been confirmed (Vanderheyden et al., 2001; Gravano et al., 2003; Novak et al., 2003; Gerosa et al., 2003) in all species except for VL. The onset of symptoms in this last species occurred very early in the OTC experiments (Novak et al., 2003), but rather late in the field. The difference in behaviour may be explained by the fact that shrubs usually grow in shady settings where photosynthesis activity is not optimal.

AA, FE and AP displayed a very clear HR, with collapsed cells in the palisade mesophyll beneath the visible stipples. In FS and VL, HR occurred later, after the onset of the visible symptoms. In FS the symptoms are associated with the degeneration of the cell contents (oxidative burst), including also the oxidisation of the vacuolar tannins (Vollenweider et al., 2003), whereas in VL they correspond to the increase of anthocyanins content. The formation of anthocyanins is to be considered an acclimatisation to oxidative stress (Steyn et al., 2002) and is induced primarily by solar irradiation. Orendovici et al. (2003) described this behaviour as an early response to ozone in many species. In all cases, the leaves compartmentalised the affected cells by developing callose layers on the walls of surrounding cells, to ensure that they could continue to exert their normal metabolic functions.

The two least sensitive species, FS and AP, are both capable of reacting to ozone stress by thickening the walls of the palisade cells (Günthard-Goerg et al., 1997, 2000). The purpose of this response may be to increase the cells’ mechanical resistance, or to increase the detoxifying processes linked to the wall's enzyme activity (Wieser et al., 2003). The construction of the wall requires an intense metabolic activity, especially involving the functions of the Golgi apparatus. A more accurate analysis of the characteristic features of the response on these two species (AP and FS) was carried out using multispectral fluorescence microimaging and microspectrofluorometry. One of the most evident findings was the accumulation of autofluorescent compounds in the palisade cells beneath the stipples. Accumulation of autofluorescence compounds has been widely reported as a symptom of HR and oxidative burst to pathogen infection (Bennett et al., 1996), as well as a response to ozone stress (Pasqualini et al., 2003). Usually the analysis of this autofluorescence signal involves only the visual microobservation of colour under UV or blue excitation, which does not easily allow for the identification of the responsible compounds. A phenolic nature was attributed to the autofluorescent deposits, mainly by using different histochemical markers (phloroglucinol, Mirande-, Neu- and Wilson-reagents, Dai et al., 1995). However, only a limited number of reports on the biochemical identification of autofluorescent secondary metabolites in infected plants has been published. It has been suggested that the coumarin scopolin may be responsible for the blue fluorescence observed under UV excitation in necrotic areas generated by pathogen attacks in tobacco (Mock et al., 1999; Chong et al., 2002). Bennett et al. (1996) associated the yellow autofluorescence under blue excitation in lettuce undergoing HR to caffeic acid and syringaldehyde released from the vacuoles and esterified on to cell walls. Nevertheless, the in vitro spectral properties of caffeic acid and other hydroxycinnamates indicate that these compounds do show autofluorescence, but peaked in the blue (around 450 nm) and only when excited in the UV region (Lang et al., 1991; Lichtenthaler & Schweiger, 1998). Therefore, the Bennett's assignment (Bennett et al., 1996) of blue-excited yellow-emission can be valid only assuming that the binding to cell walls markedly modifies the structure of phenolics, inducing a shift to longer wavelengths of both their absorption and fluorescence spectra. A similar reason could explain our fluorescence analysis of cells under ozone stress. The results of the autofluorescence microscopic examination of the leaf surface were similar to those previously reported for hybrid poplar exposed to ozone fumigation (Strohm et al., 1999). The autofluorescence response was substantially similar in both species examined, although much more marked in AP than in FS. In particular, while in AP the cells beneath the stipples were totally devoid of chlorophyll, in FS the fluorescence response of affected cells still revealed the presence of this pigment.

The extension of symptoms was associated with a reduced efficiency of processes related to the capture of light energy. The reduced density of reaction centres (RC : CS0) is a response to ozone that is common to all species described in the literature (Soja et al., 1998; Manes et al., 2001; Nussbaum et al., 2001; Gravano et al., 2004). This means that many reaction centres respond to the excessive oxidative pressure by functioning as energy dissipators (silent centres, Strasser et al., 2004). Reduced RC : CS0 was in most cases associated with a reduction in ϕP0. The reduction in the values of the JIP-test parameters linked to light process efficiency (capture and retention of light energy), that is RC : CS0 and ϕP0, suggests the importance of photo-inhibition processes and is the main response observed in FS in relation to symptom development. In VL and AA, as symptoms develop, dark processes, linked to the reduced efficiency in electron transport, take on special importance. This behaviour is highlighted by the reduction of the Ψ0 value and by the increase of parameters linked to curve kinetics, that is M0, VJ and VI. This suggests that there is an accumulation of electron transporters underway, both in the initial events of the transient (M0) and in the different steps of the curve (VJ variable fluorescence at 2 ms; and VI variable fluorescence at 30 ms). In AA irreversible damage to the reaction centres is shown by an increase in the value of F0 as symptoms progress (Krause, 1988) and highlighted by an overall reduction of the values of the de-excitation constants (SumK). In FE and AP, on the other hand, a reduction in F0 was observed alongside an increase in the de-excitation constant values (SumK), with special reference to the photochemical constant (Kp). Thus, reaction centres appear to act as more efficient quenchers (Krause, 1988).


The findings of this research suggest that the foliar symptoms, with their visible and ultrastructural characteristics, represent a spy of the active responses (species-specific) that the different plant species act (or not) to cope with oxidative stress. The common trait of all the considered species was the hypersensitive response (HR), with the programmed cell death (PCD), accompanied by the compartimentation of the injured cells with a layer of callose. Other responses observed imply physiological and ultrastructural changes at the level of the cell walls structure, the effectiveness of the residual chlorophyll and/or the accumulation of pigments. Cell wall thickening facilitates detoxification processes and confers upon the cells themselves a greater mechanical resistance to avert the risk of collapse. The role of the cell wall in the mechanisms of ozone response, especially the accumulation of phenolic compounds, was also highlighted by means of microspectral fluorometry. The increase of the photochemical constant of de-excitation (Kp) in the reaction centres of Chl a (Photosystem II), suggests that the electron transporter in the luminous phase of photosynthesis are fed. That could be considered as a compensative mechanism which involves the chlorophyll of the healthy cells, and allows the plant to maintain a certain level of detoxifying capacity of the symplast. Finally, the pigmentation due to anthocyanin accumulation in VL suggests an interaction between separate oxidative stresses (ozone plus light irradiation). The plant responds to these stresses by applying mitigation strategies to oxidative pressure.


This research was carried out within the INFOGESO (Influenza dell’ozono sulla gestione sostenibile del sistema agricolo e forestale Lombardo) programme, co-ordinated by ERSAF (Ente Regionale per i Servizi all’Agricoltura e Foreste), Lombardy, and cofinanced by the Regione Lombardia.

The measurements of ozone concentrations and levels were carried out within the CON.ECO.FOR. programme (National Forest Services, Division V). We thank Armando Buffoni for providing the data and Giacomo Gerosa for the AOT40 estimation.